F A C U L T Y O F S C I E N C E

U N I V E R S I T Y O F C O P E N H A G E N

Academic advisor: Søren Johannes Sørensen & Mette Burmølle

Submitted: 07/02/2014

PhD thesis

Dawei Ren

Synergistic Interactions in Multispecies Biofilms

1

2

Synergistic Interactions in Multispecies Biofilms

Ph.D. Thesis

Dawei Ren

Supervisors:

Professor: Søren Johannes Sørensen

Associate professor: Mette Burmølle

Section for Microbiology, Department of Biology

Faculty of Science

University of Copenhagen

Denmark

February, 2014

This thesis has been submitted to the PhD School of The Faculty of Science,

University of Copenhagen

3

Cover picture

Illustration of interspecific interactions in multispecies biofilms (Burmølle et al., 2013).

Top left: co-metabolism or niche generation

Bottom left: coaggregation

Top right: horizontal gene transfer (HGT; conjugation)

Bottom right: quorum sensing (QS)

Burmølle M, Ren D, Bjarnsholt T, Sørensen SJ (2013) Interactions in multispecies biofilms: do they actually

matter? Trends in microbiology. doi: 10.1016/j.tim.2013.12.004

4

Preface

This thesis embraces all the efforts that I have put into exploring synergistic interactions in

multispecies biofilms during the last three years as a PhD student in Molecular Microbial Ecology

(MME) group, Department of Biology, University of Copenhagen. This study was funded by

Danish Council for Independent Research and China Scholarship Council.

First and foremost, I would like to express my deepest appreciation to my supervisors Prof. Søren J.

Sørensen and Associate Prof. Mette Burmølle. They always give me invaluable guidance, priceless

advice and generous support. Thank you, Søren, for your optimism to create a highly motivated

group, for your humor to make us feel relaxed and passionate in work and for your encouragement

to boost my confidence in the face of difficulties. Thank you, Mette, for your patience when guiding

my work, for your thoughtfulness when making an outline for my clear understanding and for your

concern when helping me arrange journey and accommodation in Britain. I also wish to thank Karin

Vestberg. She is the most professional, hardworking and helpful technician I have ever met. Her

white hair, affable smile and enthusiasm have left a deep impression on me. Thank you, Karin, for

your promptness in helping me to find the protocols and reagents for the experiment, for your

carefulness in assisting me to use the new instruments and for your kindness in correcting the

occasional improprieties in my lab work.

I am especially grateful for the biofilm group members, including Jonas Madsen, Henriette Røder,

Lea Hansen, Jakob Herschend, Wenzheng Liu and Jakob Russel. Thank you, Jonas, for your

constructive comments on my manuscripts. Thank you, Lea, for your excellent work on

transcriptomic analysis. Thanks to all of you for your active contribution in the biofilm group

meeting that truly inspired me in work. I also want to give thanks sincerely to, Lasse Bergmark, for

your great help in primers design and quantitative PCR work; Waleed Abu Al-Soud and Lars

Behrendt, for your valuable suggestions in RNA extraction from biofilms; Sten Struwe and

Annelise Kjøller, for your incredible kindness to Chinese students that have warmed my heart so

much! Dr Jeremy S. Webb, Dr Robert P. Howlin and Caroline Duignan are appreciated for giving

me the opportunity to have a wonderful experience in Southampton, Britain. I would like to

especially mention Caroline Duignan for her generosity in donating time and sharing resource to

guide me in the lab and take care of my life. Her genuine friendship is a precious gift to me!

A big thank you to all the other MME group members: Lars Hansen, Bo Jensen, Anders Prieme,

Tim Evison, Leise Riber, Stefan, Anette, Gisle, Martin Hansen, Zhuofei, Trine, Samuel, Witold,

Tue, Jonas Stenbæk, Luo, Martin Mortensen, Peter, Tom, Claudia, Michael, Ines and the members

who have left: Barbara, Luisa, Shanshan, Lili and Analia. You are so lovely, amazing and make my

life in Denmark so enjoyable. Special thanks to all my Chinese friends, especially to Lili for her

care throughout my first few months in Denmark, to Shanshan and Luo for being my listeners and

sharing the joy and pain in my life here.

Finally, I would like to extend my sincerest thanks to my parents. Without your unconditional

support and love, these past three years have been impossible. As my best friends, you always

timely enlighten me in spite of thousands of miles between us. This thesis is also dedicated to my

5

dear grandfather who passed away when I struggled with my work here. Your instruction and

encouragement will never be forgotten.

I appreciate the past three years spent in this beautiful country, full of laughter and tears and this

will be precious treasure in my future life and engraved in my memory forever.

Dawei Ren

January 2014 in Copenhagen

6

Publications list

1. Ren D, Madsen JS, de la Cruz-Perera CI, Bergmark L, Sørensen SJ, Burmølle M (2013)

High-Throughput Screening of Multispecies Biofilm Formation and Quantitative PCR-

Based Assessment of Individual Species Proportions, Useful for Exploring Interspecific

Bacterial Interactions. Microbial ecology. doi: 10.1007/s00248-013-0315-z Manuscript 1

2. Ren D, Madsen JS, Sørensen SJ, Burmølle M (2013) High prevalence of biofilm synergy

among bacterial soil isolates in co-cultures indicates bacterial interspecific cooperation.

Submitted to The ISME journal. Manuscript 2

3. Hansen LBS+, Ren D

+, Sørensen SJ, Burmølle M (2014) Metatranscriptome analysis of

multispecies biofilms indicates strain- and community- dependent changes in gene

expression. In preparation. Manuscript 3

+Shared first authorship

4. Ren D, Ekelund F, Sørensen SJ, Burmølle M (2013) Effects of grazing by flagellate

Neocercomonas jutlandica on mono- and multi-species biofilms. In preparation.

Manuscript 4

5. de la Cruz-Perera CI, Ren D, Blanchet M, Dendooven L, Marsch R, Sørensen SJ, Burmølle

M (2013) The ability of soil bacteria to receive the conjugative IncP1 plasmid, pKJK10, is

different in a mixed community compared to single strains. FEMS Microbiol Lett

338(1):95–100. Manuscript 5

6. Burmølle M, Ren D, Bjarnsholt T, Sørensen SJ (2013) Interactions in multispecies biofilms:

do they actually matter? Trends in microbiology. doi: 10.1016/j.tim.2013.12.004

7

English summary

The coexistence of hugely diverse microbes in most environments highlights the intricate

interactions in microbial communities, which are central to their properties, such as productivity,

stability and the resilience to disturbance. Biofilm, in environmental habitats, is such a spatially

structured aggregation consisting of multiple species of bacteria whose function relies on a complex

web of cooperative and/or competitive interactions between community members, indicating that

research in “whole-entity” should not be based on the assembled results from “mono pieces”. As

one of the best multispecies biofilm models, oral microbial community, also known as “dental

plaque” is thoroughly investigated as a focal point to describe the interspecies interactions [1].

However, owing to the lack of a reliable high throughput and quantitative approach for exploring

the interplay between multiple bacterial species, the study to elucidate the impact of interaction

networks on the multispecies biofilms in natural ecosystems, especially in soil, is still at an early

stage. The diverse patterns of interactions within the mixed communities as well as the predator-

prey relationship between protozoa and biofilm are summarized in Sections 1, 2 and 3 of this thesis,

where the state-of-the-art techniques developed to exploit such interactions, including precisely

quantifying the numbers of individual species by quantitative PCR (qPCR) and monitoring gene

expression changes during interactions by transcriptomic analysis are also presented.

Due to the poor reproducibility of most biofilm quantification assays, the first part of my work is to

develop a rapid, reproducible and sensitive approach for quantitative screening of biofilm formation

by bacteria when cultivated as mono- and multispecies biofilms, followed by species specific qPCR

based on SYBR Green I fluorescence to measure the relative proportion of individual species in

mixed-species biofilms. The reported approach was described in Manuscript 1 which can be used

as a standard procedure for evaluating interspecies interactions in defined microbial communities.

By use of this valuable tool, a more than 3-fold increase in biofilm formation and dominance of

Xanthomonas retroflexus and Paenibacillus amylolyticus over the other two species

Stenotrophomonas rhizophila and Microbacterium oxydans were demonstrated, indicating the

strong synergistic interactions in this four-species biofilm model community.

Manuscript 2 presents the further application of this developed approach on evaluating the

synergistic/antagonistic interactions in multispecies biofilms composed of seven soil isolates. 63%

of the four-species biofilms were found to interact synergistically, indicating a prevalence of

synergistic interaction in biofilm formation among these strains. Hereafter, the population dynamics

in a multispecies biofilm composed of Stenotrophomonas rhizophila, Xanthomonas retroflexus,

Microbacterium oxydans and Paenibacillus amylolyticus, was assessed using qPCRs with species

specific primers. Despite of the high prevalence of X. retroflexus (> 97% of total biofilm cell

number), the presence of the three other strains was indispensable for the strong synergism that

occurs in this mixed-species biofilm. The dramatically increased cell numbers of each strain at 24 h

proved all the individual strains gained benefits in the multispecies biofilms compared with in

monospecies biofilms, that is, they would rather cooperate than compete with each other.

8

The significant synergistic interaction observed in the biofilm consisting of four soil bacteria make

this consortium a powerful model to study development and interactions in multispecies biofilms.

In Manuscript 3, the gene expression profile of Xanthomonas retroflexus in a single-species

biofilm was compared to its expression profiles in dual-species biofilms with Stenotrophomonas

rhizophila, Microbacterium oxydans or Paenibacillus amylolyticus as well as in a four-species

biofilm. The strongest change in expression profile was observed in the dual-species biofilms of X.

retroflexus and P. amylolyticus, while a distinct expression pattern (non-linear response) was

detected in the four-species biofilm, indicating the significant effect of interspecies interactions on

gene expression. This is consistent with the results presented in manuscript 2 where each species

was demonstrated to be indispensable for the synergistic interactions in the biofilm formation. 70

genes were found differentially expressed when co-culturing X. retroflexus with other species,

which include genes involved in membrane bound efflux system and MazE/MazF toxin-antitoxin

system, suggesting the enhanced resistance of multispecies biofilms.

Despite of the widespread existence of biofilms and protozoa in nature, the predator-prey relation

between biofilms and protozoa is still poorly studied. Moreover, this relationship could be affected

by interspecies interactions within multispecies biofilms. The study presented by Manuscript 4 was

to test whether these interactions in the developed multispecies biofilm model are involved in the

defense mechanism of bacterial biofilms against protozoan grazing. The presence of the flagellate

Neocercomonas jutlandica was shown to increase or reduce the bacterial abundance in biofilms,

depending on the co-cultured bacterial prey, which suggests the grazing ability is closely related

with the predator-prey interactions, whereas, the synergistic interactions in the multispecies biofilm

model did not confer more protection against predation compared with single-species X. retroflexus

biofilm. The same ratio of cell numbers between three species regardless of protozoan grazing

suggests they were spatially arranged in integrated communities in multispecies biofilm. However,

these conclusions are based on the assumption that this flagellate predator prefers surface attached

cells which needs to be confirmed by further studies.

Horizontal gene transfer by conjugation occurs more efficiently in biofilms. The connection

between plasmid host range and composition of the recipient community was investigated in

Manuscript 5 by comparing plasmid permissiveness in single populations and in a microbial

community composed of 15 soil strains. By use of flow cytometry (FCM) and 16S rRNA gene

sequencing, the IncP1 plasmid, pKJK10, was found only to transfer from Pseudomonas putida to

Stenotrophomonas rhizophila in a diparental mating. However, when hosted by Escherichia coli,

transfer of this plasmid occurred only in the mixed community, with Ochrobactrum rhizosphaerae

as the dominating plasmid recipient. This study demonstrates that the plasmid host range can be

greatly affected by the surrounding bacterial community. This needs to be taken into account as

many antibiotic resistance and virulence determinants are plasmid-encoded, which can spread

further and raise antibiotic-resistant bacteria in soil.

9

Dansk resumé

I de meget diverse mikrobielle samfund i naturen er der nogle yderst komplekse interaktioner, som

er meget centrale med hensyn til både produktivitet, stabilitet og modstandsdygtighed overfor

forandringer i miljøet. Biofilm i naturlige habitater er en rumlig samling af bakterier bestående af

flere arter, hvis funktion er afhængig af et indviklet net af kooperative og / eller

konkurrencemæssige relationer mellem dem, hvilket indikerer, at forskning i en biofilm ikke bør

være baseret på resultater fra enkelt bakterier. Som en af de bedste biofilm modeller med flere arter

er det orale mikrobielle samfund, også kendt som " plak ," der er grundigt undersøgt med fokus på

at beskrive interspecies interaktioner. Men manglen på nyere forskning med kvantitativ tilgang til at

udforske samspillet mellem flere bakteriearter for at belyse konsekvenserne af interaktionen på

mange arts biofilm i naturlige økosystemer, især i jord, gør at forskningen stadig er på et tidligt

stadium. Forskellige mønstre af interaktioner i de blandede samfund såvel som predator-prey

forhold mellem protozoer og biofilm er sammenfattet i afsnit 1, 2 og 3 i denne afhandling, hvor

state-of- the- art teknikker udviklet til at udnytte sådanne interaktioner, herunder præcist at

kvantificere antallet af enkelte arter ved kvantitativ PCR (qPCR) og overvågning af genekspression

ændringer i interaktioner med transkriptomic analysis, også er præsenteret.

På grund af den ringe reproducerbarhed af de fleste biofilm kvantificerings analyser er den første

del af mit arbejde gået ud på at udvikle en hurtig, reproducerbar og følsom metode til kvantitativ

screening af biofilm dannelse af bakterier, når de dyrkes som mono-og flere arts biofilm, efterfulgt

af artsspecifikke qPCR baseret på SYBR Green I fluorescens til at måle den relative andel af de

enkelte arter i blandede arts biofilm. Resultaterne er beskrevet i Manuscript 1, og kan bruges som

standard procedure for evaluering af interspecies interaktioner i definerede mikrobielle samfund.

Ved brug af dette værdifulde værktøj, blev der påvist en mere end 3-fold stigning i biofilm dannelse

og dominans af Xanthomonas retroflexus og Paenibacillus amylolyticus over de to andre arter

Stenotrophomonas rhizophila og Microbacterium oxydans, med angivelse af de stærke

synergistiske interaktioner i dette fire-arts biofilm model samfund.

Manuskript 2 præsenterer den videre anvendelse af den ovenfor beskrevne metode til at evaluere de

synergistiske / antagonistiske interaktioner i flere arts biofilm bestående af syv jord isolater. 63% af

de fire arts biofilm interagerede synergistisk , hvilket indikerer en prævalens for synergistisk

interaktion i biofilmdannelse blandt disse stammer . Populationsdynamik i en flere arts biofilm

bestående af Stenotrophomonas rhizophila, Xanthomonas retroflexus , Microbacterium oxydans og

Paenibacillus amylolyticus blev estimeret ved hjælp qPCR med artsspecifikke primere. På trods af

den høje forekomst af X. retroflexus (> 97% af det samlede antal biofilm bakterier) , var

tilstedeværelsen af de tre andre stammer essentiel for den kraftige synergi, der opstår i denne

blandede arts biofilm . De dramatisk øgede celletal for hver stamme efter 24 timer viste, at alle de

individuelle stammer opnåede fordele i flere arts biofilm sammenlignet med i monospecies biofilm,

det vil sige at de hellere ville samarbejde end konkurrere med hinanden .

Den signifikante synergistiske interaktion, der blev observeret i en biofilm bestående af fire

jordbakterier gør dette konsortium til en vigtig model til at studere udviklingen og interaktioner i

10

flere arts biofilm. I Manuscript 3 blev genekspression profilen for Xanthomonas retroflexus i en

enkelt arts biofilm sammenlignet med dens ekspressions profil i dual- arts biofilm med

Stenotrophomonas rhizophila , Microbacterium oxydans eller Paenibacillus amylolyticus såvel som

i en fire- arts biofilm . Den største ændring i ekspressions profilen blev observeret i dual- arts

biofilm med X. retroflexus og P. amylolyticus , mens et klart ekspressionsmønster ( ikke-lineær

respons) blev detekteret i fire arts biofilm , hvilket indikerer en signifikant effekt af interspecies

interaktioner på genekspression . Dette er i overensstemmelse med de resultater, der præsenteres i

manuskript 2, hvor hver art viste sig at være essentiel for de synergistiske interaktioner i

biofilmdannelse. Der blev fundet 70 gener, som blev udtrykt når X. retroflexus blev dyrket med

andre arter, der omfatter gener involveret i et membranbundet efflukssystem og i Maze / MazF

toksin - antitoxin systemet, hvilket tyder på forbedret resistens i flere arts biofilm .

På trods af den udbredte forekomst af biofilm og protozoer i naturen er predator-.prey relationen

mellem biofilm og protozoer stadig dårligt undersøgt. Endvidere kan dette forhold blive påvirket af

interspecies interaktioner indenfor flere arts biofilm. Undersøgelsen præsenteret i Manuscript 4 var

at teste, om disse interaktioner i den udviklede flere arts biofilm model er involveret i en

forsvarsmekanisme hos den bakterielle biofilm mod protozo græsning. Tilstedeværelsen af

flagellaten Neocercomonas jutlandica viste sig at øge eller reducere den bakterielle forekomst i

biofilm, afhængig af bakterien, hvilket tyder på, at græsningsevnen er nært beslægtet med predator-

prey interaktioner; de synergiske interaktioner i flere arts biofilm giver ikke mere beskyttelse mod

prædation end enkelt - arts X. retroflexus biofilm. Det samme forhold af antal celler mellem de tre

arter uanset protozo græsning antyder, at de var rumligt placeret i et integreret samfund i flere arts

biofilm. Men disse konklusioner er baseret på den antagelse, at den benyttede flagellat foretrækker

overflade vedhæftede celler; dette forhold skal bekræftes af yderligere undersøgelser.

Horisontal genoverførsel ved konjugering forekommer mere effektivt i biofilm. Forbindelsen

mellem plasmid host range og sammensætningen af recipient samfundet blev undersøgt i

Manuscript 5 ved at sammenligne plasmid tolerance i populationer af en enkelt art og i et mikrobielt

samfund bestående af 15 jord-stammer. Ved brug af flowcytometri (FCM) og 16S rRNA-gen

sekventering, blev IncP1 plasmidet pKJK10 kun vist at blive overført fra Pseudomonas putida til

Stenotrophomonas rhizophila i en diparental mating. Når Escherichia coli var vært skete overførsel

af dette plasmid kun i det blandede samfund, med Ochrobactrum rhizosphaerae som dominerende

plasmid recipient. Denne undersøgelse viser, at plasmidets værtsspektrum kan blive kraftigt

påvirket af den omgivende bakterielle samfund. Dette skal der tages hensyn til, da mange

antibiotikaresistens og virulensdeterminanter er plasmidkodede, og kan sprede sig yderligere og øge

antibiotikaresistente bakterier i jord.

11

12

Table of Contents

1 Interactions in multispecies biofilms .......................................................................................... 14

1.1 Biofilm................................................................................................................................. 14

1.2 Multispecies biofilms in soil ............................................................................................... 14

1.3 Interactions in multispecies biofilms ................................................................................... 16

1.3.1 Coaggregation, cross-species protection and co-metabolism ...................................... 17

1.3.2 Chemical signaling systems ......................................................................................... 18

1.3.3 Lateral gene transfer ..................................................................................................... 20

1.3.4 Synergism or antagonism/Cooperation or competition ............................................... 20

2 How to study multispecies biofilms? .......................................................................................... 24

2.1 In vitro biofilm models ........................................................................................................ 24

2.2 Quantitative PCR ................................................................................................................. 25

2.3 Transcriptomics ................................................................................................................... 27

3 Biofilms and protozoa ................................................................................................................. 31

3.1 Protozoa ............................................................................................................................... 31

3.2 Biofilms- the response of cell consortia to protozoan grazing ............................................ 31

3.3 Protozoa and biofilms- reservoirs of pathogenic bacteria ................................................... 34

4 Where are we going with biofilms? - In the context of microbial ecology ................................ 35

5 References .................................................................................................................................. 38

6 Manuscripts ................................................................................................................................. 50

Manuscript 1

Manuscript 2

Manuscript 3

Manuscript 4

Manuscript 5

13

14

1 Interactions in multispecies biofilms

1.1 Biofilm Already in the late 1600s, van Leeuwenhoek had observed biofilm in the plaque on his own teeth.

The first appearance of “biofilm” theory can be traced to 1987 when Costerton et al. described

biofilm as adherent population consisting of single cells and microcolonies of sister cells all

embedded in a highly hydrated, predominantly anionic matrix of bacterial exopolymers and trapped

extraneous macromolecules [2]. Over the course of the past 25 years, this concept has evolved to

include not only the irreversible cell attachment but also physiological attributes including altered

growth rate and gene transcription [3].

Biofilms, prevalent on the most inert or living surfaces [4], are the dominant communities on planet

earth. It is estimated that 99% of bacteria in nature exist in biofilms, while biofilms account for

more than 65% of hospital infections [5, 6]. And the same kind of bacteria are known to show

profoundly different characteristics when they are in a biofilm compared with in planktonic cultures.

The high resistance to harsh conditions, including pollutants [7] , desiccation [8], protozoan grazing

[9], antimicrobial agents [10] in nature and host defenses [11] in chronic infections, is some/one of

the most important features of biofilms. This resilience can be explained by several mechanisms

which are summarized in Table 1. Moreover, the heterogeneity within biofilms offers the possibility

of the joint action of these multiple resistance mechanisms in a single community. Despite the

intensive research in batch cultures, extrapolations of these results in bulk to that in biofilm cells

seems unwise due to the biofilm-specific physiological properties.

Table 1 Summary of resistance mechanisms of biofilms.

Biofilms Mechanisms Adverse factors References

Escherichia coli

Pseudomonas aeruginosa

Staphylococcus epidermidis

Reduced

permeability of

matrix

β-lactam antibiotic

desiccation

metal

host defenses

[12] [13, 14] [15]

Pseudomonas aeruginosa

Escherichia coli Slow growth Ciprofloxacin [16]

Escherichia coli

Salmonella sp.

Pseudomonas aeruginosa

Biofilm-specific

phenotype

Ciprofloxacin

host defenses

protozoa

[17] [18] [19]

Pseudomonas aeruginosa Persister cells Ofloxacin, metal [20] [14]

Serratia marcescens

Pseudomonas aeruginosa Quorum sensing protozoa [21] [19]

1.2 Multispecies biofilms in soil Biofilms in nature habitats are complex communities where various types of microorganisms (e.g.

bacteria, archaea, protozoa, fungi and algae) are held together and protected by self-excreted

extracellular polymeric substance (EPS) [22]. For example, soil is such a potential environment

where the population density and diversity of bacteria may be up to 109 cells/g soil and 10

6

15

species/g soil, respectively [23, 24]. Because of the spatial variability in nutrient concentration,

microbial cells are not uniformly distributed through the soil [25]. Bacteria living close to the

nutrient sources i.e., plant roots or decaying organic matter, are prone to attach various available

surfaces (e.g. roots, litter or soil particles) and develop into multispecies biofilms.

By organizing into a biofilm community, bacteria could gain highly resilience to adverse soil

conditions. Water is by far the largest component of the biofilm matrix which can account up to

97%, whereas the remaining are 2-5% microbial cells, 3-6% EPS (polysaccharides, proteins, nucleic

acids and lipids) and ions [26]. The highly hydrated matrix could therefore buffer the biofilm cells

against desiccation stress which is an important challenge met by soil bacteria. Chang et al.

provided the direct evidence that alginate production by Pseudomonas putida contributed to a

hydrated microenvironment which protected residents from water-limiting stresses [27]. Moreover,

biofilm has a great capacity for heavy metal biosorption and toxic compound degradation which has

a significant impact on bioremediation [28, 29]. Additionally, the widespread exposure to

antibiotics makes biofilm formation more favorable in soil. The results from Walker et al. suggested

that upon root colonization, Pseudomonas aeruginosa gained resistance against root-secreted

antibiotics by forming a biofilm [30]. Apart from these advantages, biofilms can also function as

protective barriers against protozoan grazing which is a major mortality factor faced by bacteria in

the soil environment [31]. Section 3 (Biofilms and protozoa) will be devoted to a coherent

introduction of the relationship between biofilms and protozoa.

These improved biofilm-associated fitnesses mentioned above suggest that the preferred mode of

bacterial growth is in a biofilm. By being encased in the recalcitrant matrix, the bacteria grow in a

relatively stable environment called microbial homeostasis [32], reflected not by the characteristics

of resident individuals but by the balance imposed by the numerous microbial interactions,

including examples of quorum sensing (QS) and horizontal gene transfer (HGT). By means of

quorum sensing, the sessile cells in the biofilms can “talk” to each other. Due to the increased

population density and constrained diffusion, the quorum sensing molecules are concentrated. Once

reaching a threshold level, these quorum sensing molecules modulate the transcription of certain

genes and trigger phenotypic changes, including swarming motility, biofilm formation and the

production of virulence factors [33-35]. This issue will be elaborated further in the next part of

Section 1.The dramatically increased horizontal transfer of plasmid-borne antibiotic resistance

determinants was observed by Savage et al. in the Staphylococcus aureus biofilm [36]. Since many

antibiotic resistance determinants are plasmid-encoded, this further spread of antibiotic resistance

genes among bacteria allows us to conclude that soil represents a reservoir of antibiotic resistance

genes [37] which probably increases the current arsenal of antibiotic resistance mechanisms in

pathogens when gene transfer occurring from soil bacteria to pathogenic bacteria. This was

confirmed by Forsberg et al. [38] with the finding that multidrug-resistant soil bacteria, containing

resistance cassettes against five classes of antibiotics, have perfect nucleotide identity to genes from

diverse human pathogens. Therefore, the enhanced efficiency of gene transfer in biofilms has a

profound impact on the pathogenesis, persistence and hence the treatment of human disease.

16

Despite of the notorious resistance to various common antibiotics and host defenses, soil biofilms

can also be exploited for their diverse application in agriculture. The biofilmed inocula can be used

as biofertilisers (BFBF) to promote and stimulate plant growth as well as aid in disease control [39].

Furthermore, in the biofilm formed by bacteria and fungi, another natural inhabitant in soil, there is

often generated synergistic interactions with possible consequences of a significant increase in

nutrient acquisition and uptake of phosphorus, nitrogen and metal ion [40]. The fungal-bacterial

biofilms (Penicillium frequentans and Bacillus mycoides) resulted in a 14-fold increase in the

biodegradability of degradable polyethylene by P. frequentans [41]. The co-culture of

Pseudomonas fluorescens and a mushroom fungus (Pleurotus ostreatus) increased the endophyte

colonization of tomato by 1000% compared to inoculation with P. fluorescens alone [42]. A

bradyrhizobial-fungal biofilm showed nitrogenase activity, whereas the bradyrhizobial strain alone

did not, which improved the shoot and root growth, nodulation and nitrogen accumulation of

soybean and directly contributed to soil nitrogen fertility in the long term [43]. Additionally,

anaerobic degradation of complex organic matter into methane and carbon dioxide requires the

progressive action of numerous species of microorganisms [44]. Biofilms can provide such an ideal

environment for the interaction of these metabolically cooperative organisms, owing to their highly-

organized structure enhancing the nutrient availability as well as removal of potentially toxic

metabolites.

In summary, as the dominant growth form for bacteria in soil, mix-species biofilms play an

essential role in maintaining the ecological balance, whereas from an evolutionary perspective, this

role is further strengthened by the selective pressures which favor bacteria capable of forming

biofilms in versatile soil environment.

What is real is rational -- what is rational is real.

--Hegel, 1821, Basic Outline of the Philosophy of Right

1.3 Interactions in multispecies biofilms Different species, exhibiting different growth and survival properties, encased in an extracellular

polymeric network could lead to the spatial and functional heterogeneity within biofilms. Even in a

single-species biofilm, the physical, chemical (e.g. gradients of nutrients, waste products and

signaling compounds) and biological (distinct metabolic pathways and stress responses)

heterogeneity can develop [45]. In environmental habitats, diverse bacteria and in many cases fungi,

algae and protozoan, do not live independently in their local microenvironments. The interactions

among these microorganisms and with the external environment critically influence the

development, structure and function of the biofilm and conversely, the spatial heterogeneity and

biodiversity clearly have a dramatic effect on the communication between different biofilm

components, allowing for the development of a complex multispecies community (Figure 1).

17

Figure 1 Communication in a natural multispecies biofilm [46]. Biofilm communities are shaped by various

interactions between microbial species, including (1) competition between bacteria populations and their

neighbours such as fungi (A), bacteria (B) and microalgae (C), (2) quorum sensing derived from clonal

growth in microcolonies (D) which may induce the protection against protozoa (E) and (3) interactions with

second colonizers such as macroalgae spores (F) and invertebrate larvae (G). Furthermore, the bacteria-host

interaction should also be taken into consideration in medical environment.

1.3.1 Coaggregation, cross-species protection and co-metabolism

Various types of interactions within biofilms include coaggregation, cross-species protection, co-

metabolism, quorum sensing (QS) and genetic exchange. Coaggregation, defined as the specific cell

to cell recognition among genetically distinct bacteria [47], is vital for both biofilm formation and

the existence of certain bacterial species. This has been well described in numerous studies of the

oral biofilms. An example is provided by Fusobacterium nucleatum which can coaggreate with

species that can not bind to each other thus serves as bridge between early and late colonizers in a

sequential process [48]. Another example is the ability of Escherichia coli O157:H7 to adhere and

persist in a capillary flow cell which requires the colonizing partner Pseudomonas aeruginosa

PAO1. Development of E. coli microcolonies occurred only along the outer 200 μm edge of the

flow cell after P. aeruginosa migrating away into the center of the flow path, indicating that the

conditioned surface by the later may facilitate attachment of the former [49]. In addition,

coaggregation confers cross-species protection of anaerobic species from oxygen and of susceptible

species from antimicrobials [50, 51]. Cross-species protection is also likely derived from

extracellular polymer and this was demonstrated by the study from the mixed fungal-bacterial

biofilm where the extracellular polymer produced by Staphylococcus epidermidis RP62A could

inhibit fluconazole penetration and conversely, the presence of Candida albicans in this biofilm

appeared to protect the slime-negative Staphylococcus against vancomycin [52].

Co-metabolism, where one species utilizes a metabolite produced by a neighboring species,

presumably plays a major role in biodegradation of organic molecules and is advantageous to the

entire microbial community. Boonchan et al. proved that inoculation of fungal-bacteria co-cultures

18

resulted in significantly improved co-metabolic degradation of polycyclic aromatic hydrocarbons

(PAHs) in soil [53]. It has also been demonstrated that the acceleration of the remediation of

chlorophenol- and phenol-contaminated groundwater by a sequencing batch biofilm reactor was

probably due to the co-metabolism [54]. The efficient degradation when multiple species are

present can be derived from the optimized substrate availability when growing attached to a surface

and the close proximity of enzymes involved in degradation that may be retained in the biofilm

matrix [55, 56]. Additionally, metabolic communications were also reported between bacteria

within the oral cavity. One of them is the metabolic interaction through arginine between two oral

bacteria Actinomyces naeslundii and Streptococcus gordonii, where S. gordonii genes involved in

arginine biosynthesis and transport were induced when coaggregated with A. naeslundii, otherwise

S. gordonii could not grow without sufficient arginine [57].

1.3.2 Chemical signaling systems

A cell-to-cell signaling mechanism known as quorum sensing (QS) has been shown in many studies

to play crucial roles in biofilm development as mentioned in Section 1.2. And due to the social

behaviors of bacteria, the term “sociomicrobiology” was introduced by Parsek MR et al. in 2005 to

vividly describe the inextricable link between biofilms and quorum sensing [58]. It has been shown

that QS can be induced by a few thousand bacteria, which size is analogous to the number of

bacteria found in biofilm microcolonies [59]. Many bacterial behaviors are regulated by chemical

autoinducer molecules that are produced and used by bacteria to sense one another. Bacteria can

communicate both intraspecifically and interspecifically via autoinducers which alter gene

expression and allow bacteria to respond coordinately to their environments, in a manner that is

comparable to behavior and signaling in higher organisms. In gram-negative bacteria, acylated

homoserine lactones (AHLs) are the most intensively investigated signal molecules and have been

well described in Pseudomonas aeruginosa. There is also report that two different chemical

languages: N-acyl homoserine lactones (AHLs) and cis-2-unsaturated fatty acids were utilized to

control biofilm formation and virulence in Burkholderia cepacia complex (Bcc) [33]. Despite of the

species specificity of AHL systems, the cross-species talk was reported in a biofilm composed of

cystic fibrosis-associated P. aeruginosa and B. cepacia [60, 61]. In addition, Bacillus sp. and

Variovorax paradoxus were reported to degrade AHLs and interfere with quorum sensing of other

species [62, 63]. In gram-positive bacteria, such as Staphylococcus aureus, peptides operate

generally by binding to receptors on the cell surface rather than diffusing back into the cell like

AHLs [64, 65]. The signaling communication in multispecies biofilms are mainly mediated by

autoinducer 2 (AI-2), which is synthesized by the enzyme LuxS and found in both gram-negative

and –positive bacteria [1, 66]. AI-2 has been shown to promote the biofilm formation of two oral

bacteria Actinomyces naeslundii T14V and Streptococcus oralis 34. Whereas, AI-2 of

Fusobacterium nucleatum was reported to differentially regulate biofilm growth of two oral

streptococci by producing a stimulatory effect on Streptococcus gordonii and an inhibitory effect on

S. oralis [67]. Two redundant quorum sensing systems were verified in Vibrio harveyi, with AHL

for intraspecies communication and AI-2 for interspecies cell–cell signaling [68]. In spite of the

apparent universality of luxS (present in more than 40 bacterial species), the difficulties of obtaining

purified AI-2 from species expressing AI-2 activity raise the doubts that whether AI-2 is a universal

19

signal or just may be a byproduct of the activated methyl cycle (AMC) [69]. Recently, Santiago‐

Rodriguez et al. reported luxS sequences in 25- to 40-million-year-old bacteria, such as Bacillus

schakletonii and B. aryabhattai, two extant bacterial species that had not been previously reported

as carrying luxS [21]. This in turn raises new questions on the specific role of luxS in ancient

microorganisms and whether it is involved in the regulation of metabolism in amber bacteria.

Another quorum sensing signal-diffusible signal factor (DSF), identified in Burkholderia

cenocepacia [70] and Pseudomonas aeruginosa [71], was reported recently to be involved in

interspecies communications by altering biofilm formation, architecture and resistance to antibiotic

[72-74]. Although the underlying mechanism of DSF in mixed communities remains to be

elucidated, this signal may play a crucial role in mediating cell-to-cell interactions in parallel with

AI-2 and AHL signals, owing to its widespread existence in various species and niches. Despite the

significant advances regarding to bacterial quorum sensing and group behaviors mentioned here,

expounding the functional consequences of QS in multispecies biofilms is still a challenge.

Although quorum sensing (QS) can be considered as diffusion sensing (DS) as QS induction or

repression is based on the interaction between the diffused signal and the cognate receptor [75], QS

enables bacteria to coordinate their behaviors for group benefit while DS depends on individual

fitness benefits [76]. These conflicting concepts are unified by an alternate hypothesis efficiency

sensing (ES), which suggests the role of autoinducers relies on both cell density and spatial

distribution and thus is favored by both group and individual benefits [77]. Future research towards

uncovering the genetic network induced by QS that deterministically controls biofilm adaption will

undoubtedly provide new insights into biofilm manipulation.

Another chemical signaling system 3', 5'-cyclic diguanylic acid (c-di-GMP), as a ubiquitous

secondary messenger, is found in diverse bacteria. The levels of c-di-GMP are mediated in the cell

by diguanylate cyclase (DGC) activity involved in c-di-GMP synthesis and phosphodiesterases

(PDE) activity involved in c-di-GMP degradation. The well characterized example is in Vibrio

cholera where 62 genes are predicated to encode proteins capable of producing or degrading c-di-

GMP and influence many phenotypes including motility, biofilm formation and virulence [78, 79].

Moreover, c-di-GMP was reported to reciprocally control biofilm formation and virulence with

quorum sensing in V. cholera. These two signaling function antagonistically to regulate biofilms

and synergistically to repress virulence factor expression, which are fundamental for V. cholerae

survival ex vivo and in vivo, respectively [80]. The increased production of c-di-GMP could

enhance biofilm formation and decrease swarming motility were also observed in Pseudomonas

aeruginosa [81]. George O’Toole elaborated upon how c-di-GMP influenced biofilm formation via

sensing environmental input, phosphate levels, in Pseudomonas fluorescens. This involved two

proteins, LapA and LapD. Whereas LapA was required for stable surface attachment and biofilm

formation, LapD served as a unique c-di-GMP effector protein that utilized an inside-out signaling

to regulate LapA localization and thus surface commitment [82]. Despite the diverse components

and molecular processes involved in biofilm formation throughout the bacterial kingdom, c-di-GMP

signaling seems to represent a common principle, which suggests that the enzymes that controlling

cellular c-di-GMP levels may be promising targets for anti-biofilm drugs.

20

1.3.3 Lateral gene transfer

Lateral gene transfer (LGT), also termed horizontal gene transfer (HGT), refers to the gene

exchange among bacteria cells in a manner other than traditional reproduction, which can occur

among conspecific strains [83] and strains in different species in biofilms [84]. There are two LGT

mechanisms: transformation and conjugation. Transformation - the uptake of free DNA from the

environment by a bacterial cell- requires exogenous DNA which can be easy to meet as a result of

the large amount of extracellular DNA in biofilms. The other mechanism, conjugation, occurs when

there is direct cell-to-cell contact or a bridge-like connection and transfers small pieces of DNA,

usually plasmids which often carry virulence and antimicrobial resistance genes. Efficient

transformation and conjugation in microbial consortium have been found in both natural and

artificial environments. For example, lateral gene transfer of AHL synthase gene could facilitate

cross talk between Burkholderia spp. and Pseudomonas spp. [85], whereas lateral gene transfer of

ring-hydroxylating-dioxygenase (RHD) gene may improve aromatics’degradation by spreading

the gene among different species [86]. Burmølle et al. reported a conjugative plasmid pOLA52,

which confers resistance to olaquindox and other antimicrobial agents through a multidrug efflux

pump, can also promote biofilm formation in Escherichia coli [87]. Also, there is example that

transfer efficiency of plasmids in Pseudomonas putida biofilm depended on the type of antibiotics,

suggesting biofilm bacteria may “sense” antibiotics to which they are resistant and enhance the

spread of that resistance [88]. Overall, efficient gene transfer is both the cause and consequence of

biofilm development. On the one hand, LGT is facilitated within biofilms as a result of the presence

of extracellular DNA, close spatial juxtaposition of bacterial cells and stable habitat provided by

EPS matrices [89]. On the other hand, both DNA transfer processes seem to have positive effects on

biofilms, such as enhanced resistance against predators, toxins and antibiotic factors and the

stabilization of biofilm structure mediated by conjugative pili which may act as cell adhesins in

hydrodynamic biofilm systems [90].

1.3.4 Synergism or antagonism / Cooperation or competition

The interactions responsible for synergism in biofilms as exemplified above, often operate in

concert and have been demonstrated to strengthen the protective effects of biofilms when multiple

species are present compared with single species communities. This was verified by a substantial

increase in the chlorine tolerance of a multispecies biofilm from drinking water, regarding to the

planktonic cultures and monospecies biofilms [91]. Likewise, it was recently reported that a

reproducible mixed-species biofilm comprising Pseudomonas aeruginosa, Pseudomonas protegens

and Klebsiella pneumonia was more resistant to the antimicrobials sodium dodecyl sulfate and

tobramycin than the single-species biofilms. Moreover, such community level resilience was found

to come from the protection offered by the resistant species rather than selection for the resistant

species, suggesting that the community-level interactions, such as the sharing of public goods, are

unique to the structured biofilm community [92]. Nevertheless, some antagonistic interactions

between bacteria also have been documented in the dental biofilm [93] as well as in marine [94] and

soil environments [95]. Microbial antagonism can be caused by inhibition of microbial growth by

diffusible antibiotics, toxins or biosurfactants [96], competition for colonization sites and nutrients

[97] and degradation of quorum sensing molecules [98]. Diverse physical interactions between

21

bacteria and fungi have been associated with reduced fungal viability due to the antifungal

molecules secreted by bacteria into the local environment. An example is Acinetobacter baumannii-

Candida albicans in chronic infections. It was shown that A. baumannii could inhibit several

important virulence determinants of C. albicans, including hyphae and biofilm formation via

polymicrobial infection and conversely, the viability of A. baumannii is reduced when C. albicans

cells adopt the quorum mode in a biofilm environment [99]. Thus, exploiting the mechanisms used

by competing microorganisms could potentially contribute to combating detrimental biofilms in

medical, industrial and natural environments.

According to West et al. [100], interactions can be roughly classified into cooperation or

competition, based on the effect of the microbial social behavior on each population in a binary

system (actor and recipient). When recipient benefits from the presence of actor, the interaction is

termed cooperation, which can be further subcategorized into mutualism (beneficial for both actor

and recipient) and altruism (beneficial for recipient but costly to actor). On the contrary, when

recipient is negatively affected, the interaction is identified as competition which is grouped into

selfishness (beneficial for actor) or spite (negative effect on both recipient and actor). In addition,

two other closely related terms are also frequently used to describe the social behavior in biofilms,

i.e. synergism and antagonism. Synergism has been defined as the cooperative action of two or

more organisms where the effect of their collective effort is greater than it would be by their

individual effect [101]. In contrast, antagonism is the relationship in which one species of an

organism is inhibited or adversely affected by another species in the same environment. When these

definitions are applied to soil microorganisms, synergism, also known as protocooperation, is

described as a facultative phenomenon that both populations can survive on their own but the

association provides some mutual benefits [102]. In this thesis, we use “synergism” to refer to the

enhanced overall productivity (biomass) and fitness (resistance against protozoa grazing) of the

multispecies community as a whole compared with the individual species. Competition often

indicates the active competition for nutrients and space. Thus, in a narrower sense, competition can

also be defined as “the injurious effect of one organism on another because of the removal of some

resource of the environment” [103]. While, antagonism may be used broadly to include the

competition for limited substrate, the inhibition by antibiotics or metabolites produced by another

organism, or exploitation which is either predation or direct parasitism [104]. When used practically,

however, these definitions are not clearly defined and may vary somewhat. For example, Foster et

al. applied the evolutionarily stringent definition of cooperation [97] when analyzing the

productivity of two-species mixtures grown in aquatic microcosms, that is, only the interactions that

can cause the increased productivity of both species in co-cultures, can be termed as cooperation.

Moreover, in spite of the usefulness of this binary system mentioned above in defining interaction,

the natural communities are far more complex than expected where more than one type of

interactions probably simultaneously occurs and more than two species are involved. Hence, it may

seem trivial to place the interaction within a biofilm in one category while exclude others [105]. A

typical example described by Hansen et al. [106] showed that the increased biomass of the dual-

species biofilm in total and of one member (Pseudomonas putida) is at the expense of another

member (Acinetobacter sp.) due to intensified competition for oxygen. Therefore, both the biomass

22

and/or function of the integrated multispecies community and each individual member need to be

evaluated to define whether the cooperative or competitive interactions shape the community. This

is what we have done in manuscript 1 and 2 [107]. The quantitative PCR developed in our study

was applied to measure the absolute cell numbers of each species in a four-species biofilm and

hence identified the dominance of cooperation interaction as each member could benefit, with

respect to biomass, in this multispecies biofilm compared with when they grow alone.

What we know from the research focusing on the link between genetic population structure and

social behavior are still very limited. And most studies support the idea that while cooperation will

occur within the same genotype, competition should prevail between different genotypes [108-110].

However, this prediction is drew with the implicit assumption that there is competition for resources

among these interacting genotypes which may not always be the case in nature environment,

exemplified by efficient biodegradation through co-metabolism in multispecies communities

mentioned above. Zhang et al. demonstrated that the switch between synergy and competition is

closely related to the flow conditions. Limited resource replenishment favors competition under

low-flow conditions while high flow favoring synergy by providing greater resource and making

biofilm easier to shear from the surface [111]. Moreover, cooperation between different species also

may prone to emerge in communities with low niche overlap and high relatedness within each

member. From the social evolution point of view, cooperation is challenging to explain as

cooperative phenotypes are susceptible to exploitation by rapidly growing, non-cooperative cells.

However, Nadell et al. reported that this condition can be reversed if cooperative cells are

segregated in space and preferentially interact with each other, indicating the cooperation may

evolve readily than naively expected [112]. The proposition that biofilm promotes altruism was

evidenced by an individual-based model simulations which showed the often-observed structural

organization into microcolonies and shedding of single cells in biofilms are necessary for the origin

and maintenance of the altruistic strategy [113]. Moreover, a recent report by [114] pointed that

intraspecies variation, despite the enhanced individual fitness of the generated morphotypic variants

compared with their parental strains, was reduced in the mixed species biofilms, suggesting this

sacrifice of self-produced diversity in the presence of other species probably represents altruism in

multispecies biofilms [115]. The “Black Queen Hypothesis (BQH)” was put forward recently to

better explain how and why cooperation in a multispecies community can evolve [115]. It presents a

scenario whereby a division of labor in microbial communities is likely to occur by receivers

deleting costly pathways that are provided by the surrounding bacteria which unavoidably produce

public resources, thus bacteria might often develop interdependent cooperative interactions. This is

opposite to the “Red Queen Hypothesis” which states the evolutionary conflict between co-

inhabitant organisms. Therefore, BQH provides a new framework for looking at interactions in the

ecological context where these organisms are evolving, which needs to be carefully considered

when applying social evolution theory to microbial communities. Viewed in this way, the conflict

between group-level selection and individual-level selection, which favor cooperation and

competition respectively, should be assessed on a case-by-case basis. Instead of discerning which is

the ‘right’ model, it should be noted, however, that different types of selection are likely involved in

shaping the evolutionary adaption of biofilm, leading to different types of molecular strategies

23

involved in regulating surface associated communities according to various environmental cues

[116].

Since the majority of microorganisms present in the environment remain unculturable, the diversity

of complex bacterial communities is inevitably underestimated. We can not exclude that other

bacteria out there might depend to a great extent on cooperation with partners, and perhaps just this

is one of the reasons why we failed so far to isolate them in the laboratory [117]. This also

highlights the importance of the in situ studies of microbial interactions, which is a prerequisite for

deeper understanding of social behavior in multispecies biofilms and hence provide better treatment

for biofilm-associated infections and exploit biofilms for beneficial applications, such as waste

water treatment, N2 fixation, pollutant degradation and so on.

24

2 How to study multispecies biofilms? The study of biofilms has proliferated in the past 15 years. And it is widely acknowledged that

biofilm formation has a close correlation with the species and/or intra-/interspecies communication

that is present, as well as the environmental conditions involved.

2.1 In vitro biofilm models In vitro biofilm experiments were brought into being in the mid-1980s [118, 119], followed by the

development of several biofilm-forming devices, including Modified robbins device [120],

Microtiter plates [121], Calgary device [122], Flow cell [123] and BioFilm ring test [124], which

allow to explore the adhesion ability of different microorganisms and the effect of various

environmental conditions on biofilm development with tightly controlled parameters. Using 96-well

microtiter plates, the effects of coaggregation and quorum sensing molecules on biofilm formation

were assessed respectively [125, 126]. Although in vitro studies have the advantages in evaluating

the influence of pre-determined environmental, physiological and genetic variables on biofilm

formation, large amounts of reproducible data under different conditions are still missing as the

insufficient versatility and the difficulty to control all the complex variables especially when a

heterotrophic consortium, such as a multispecies biofilm, is studied. Considering the lack of

reproducibility, possibly due to the researcher’s dependent variable, the comparison of results

obtained when using different protocols and biofilm growth systems, particularly between different

research groups, is problematic. This is what we hopefully are contributing to in manuscript 1 [107].

A standard procedure was developed for evaluating interspecies interactions in defined microbial

communities by comparing the reproducibilities of 96 Well Cell Culture Plate and Nunc-TSP lid

system (also referred to Calgary methods)-based biofilm formation. The Calgary Biofilm Device

(CBD) was first described and applied by Ceri et al. [127] for high-throughput determination of

antibiotic susceptibilities of bacterial biofilms. The biofilm is formed on the pegs of a modified

microtiter lid in this device instead of at the bottom of a well which can avoid non-specific bacterial

sedimentation that may not accurately relate to biofilm, though the difference between this “upside-

down” biofilm and the biofilm formed at the bottom has not been addressed. Apart from the broad

and robust applicability for many microorganisms, CBD, moreover, seems to be more amenable for

combination with microscopy, e.g. epifluorescence microscopy or confocal laser scanning

microscopy (CLSM), which makes it possible to analyze the structural heterogeneity of biofilms

under diverse exposure conditions [128]. Thus, by introducing microscopy into biofilm studies,

many parameters such as biofilm biomass, total and active number of cells as well as associated

physiological activity, extracellular matrix and overall structure can be assessed. Especially for

multispecies biofilms, the spatial organization of different species is very important which plays a

vital role in determining biofilm function, including antimicrobial resistance [92] and driving

metabolic reactions [129].

In order to gain a more comprehensive and in-depth understanding of complex interactions that

drive biofilm development, the multispecies consortia should be investigated nondestructively, in

real time and in situ which also enable studies of the species that are unculturable in the laboratory.

However, it is still challenging until now to mimic the microenvironment as much as possible

where microorganisms live in close proximity and interact with each other, especially for soil

25

biofilms, owing to the complex structure of soil environment and the strikingly high bacterial

numbers. Therefore, the use of in vitro and in vivo biofilm model systems under controlled

conditions is indispensable for the study of the complex communities. For instance, biofilms

exposed to shear stress are studied in flow cells, and when combined with Fluorescence in situ

hybridization (FISH) and CLSM, both quantitative data and interactions information can be

provided [130]. FISH allows the detection not only of cultivable microorganisms but also of

fastidious or uncultured species. However, the main target molecules, 16S rRNAs, could somewhat

cause perplexity when measure viability and metabolic activity, as ribosomes can remain intact

even in recently nonviable and/or non-metabolically active cells. This can be surmounted by

targeting the short-lived intergenic space region (ISR) between the 16S and 23S rRNA segments,

called Spacer-FISH [59]. The recent combination of Raman-FISH and secondary ion mass

spectrometry can be applied for functional studies of biofilm microbes on a single-cell level by

using stable isotopes as labels [131]. Nevertheless, the widely used FISH technique also has

limitations, such as laboriousness and unsuitability for high-throughput screening. An alternative

device, BioFlux, was presented recently which comprises disposable microplates with embedded

microfluidic channels and a distributed pneumatic pump providing rate-controlled fluid flow [132].

BioFlux 1000, which integrates a high performance microscopy workstation, allows real-time,

automated image capture. In combination with strain specific markers such as reporter genes,

fluorescence-labeled antibodies and probes, flow cells and BioFlux are capable of exploring the

structural organization and dynamics of biofilms.

Despite the direct visualization of biofilm structures using microscopy techniques, this technique

can only provide high quality but semi-quantitative results. Other molecular techniques, such as

quantitative PCR and transcriptomic analysis using next-generation sequencing, are promising tools

in quantitative assessment and have led to a clearer depiction of the patterns of transcriptional

regulation of biofilm-specific genes and the signaling network employed by biofilms at various

stages in their growth.

2.2 Quantitative PCR Quantitative PCR (qPCR) has been widely used to quantify microorganisms and measure functional

gene markers in complex communities due to its accuracy, high sensitivity, specificity and speed.

In contrast to the traditional end-point PCR, the amplification of the PCR products are recorded in

“real-time” via a corresponding increase in fluorescent signals. By detecting the accumulation of

amplicons during the early exponential phase of the PCR, gene/transcript numbers are quantified

when these are proportional to the starting amount of nucleic acid, while the levels of expressed

genes are measured when combined with a preceding reverse transcription reaction (RT-qPCR).

There are two commonly used reporter systems, namely, SYBR Green assay and TaqMan probe

assay. Since SYBR Green binds to all double-stranded DNA by intercalating between adjacent base

pairs, including nonspecific PCR products, the target may be overestimated if the primers are not

highly specific to their target sequence. Hence, a post-PCR melting (dissociation) curve analysis is

needed to verify that the fluorescence signal only comes from target templates. However, the

introduction of the reporter probe can significantly increase specificity of the detection. TaqMan

26

probes are dual labeled hydrolysis probes, incorporating a fluorescent reporter molecule at the 5'

end and a quencher molecule at the 3' end [133]. During template extension, the reporter-quencher

proximity is broken by the 5' to 3' exonuclease activity of the Taq polymerase, thereafter

unquenched emission of fluorescence is detected after excitation with a laser. By using multiple

TaqMan probes and primer sets, highly similar sequences can be differentiated in different qPCR

assays [134]. Furthermore, TaqMan probes labeled with different fluorophores enable that different

targets are amplified and quantified within a single reaction which is called multiplex qPCR [135].

In spite of the additional specificity afforded by Taqman probe, the relatively high cost of labeled

probe limits its use to some extent. In addition, TaqMan amplicons need to be longer as additional

conserved sites are required for designing probes, whereas identification of three conserved regions

for primer set and probe within the short amplicon may not be always possible, especially for

divergent gene sequences. But this dilemma can be alleviated to a certain extent by using TaqMan

MGB probe as the DNA probes with conjugated minor groove binder (MGB) groups form

extremely stable duplexes with single-stranded DNA targets, allowing to design shorter probes

[136]. However, using MGB probes will further increase the running cost. Maeda et al. reported

that both TaqMan and SYBR Green assays showed sufficient sensitivity and specificity for

quantification of bacteria species in dental plaque [137]. Taking into account the lower running cost,

ease in primer design and assay set-up, SYBR Green assay may be suitable for routine clinical

examinations.

As a powerful, convenient tool, quantitative PCR assay has been increasingly employed in the past

few years to detect and quantify target bacteria in biofilms, or to perform gene expression analysis

during biofilm development. By using SYBR Green based qPCR, the population dynamics of

pathogenic salmonellas during 4-week period in both water and biofilm samples had been followed

[138]. Zhang et al. identified three genes involved in Pseudomonas aeruginosa biofilm-specific

resistance to antibiotics by performing mRNA-based qPCR to compare the expression of these

genes in planktonic- and biofilm-grown cells [139]. The data from differential transcript levels

coupled with quantification of DNA release and cell densities in mixed cultures of three

streptococci provided new insights into ecological factors that influence the competition between

pioneer colonizing oral species in oral biofilm. In addition, in combination with propidum

monoazide (PMA), qPCR assay can overcome its main limitation of the inability to discriminate

between live and dead cells [140, 141]. This is achieved through PMA penetrating the membranes

that have lost their integrity and then binding to the dsDNA which prevents the use of dsDNA-PMA

complex as a template for PCR reaction. The first research using qPCR-PMA technique together

with TaqMan probe for analyzing the live and dead cells present in a multispecies oral biofilms was

reported recently [142]. Moreover, by use of laser capture microdissection microscopy (LCMM), a

small group of cells can be harvested at spatially resolved sites within biofilms, thus reflect

heterogeneity inherent to biofilms. Combined LCMM with multiplex RT-qPCR, Franklin presented

a stratified biofilm which showed the high amount of housing keeping, acpP, in the top 30 μm of

the biofilm, with little or no mRNA at the base of the biofilms, suggesting that the transcription of

individual genes varies dramatically in different regions of the biofilms [143].

27

In order to make valid comparisons between different samples, there are a number of factors that

should be taken into account before performing the qPCR assay. One of the important factors is the

choice of method used for nucleic acid extraction which is a major determinant on the final

quantification. As the extraction efficiencies vary significantly between different methods as well as

different types of environmental samples [144], it is problematic to make direct comparison of

absolute cell numbers between studies without ensuring that the same extraction procedure is used

for each sample. The DNA extraction protocol optimized in our study, which allows DNA

extraction from both the gram-negative and -positive cells in the multispecies biofilm with the same

efficiency [107], is likely to have a wide application in DNA-based biofilm research. Additionally,

PCR inhibitors are often found in environmental samples and interfere with the following qPCR

performance. Hence, the equivalent amplification efficiencies between the environmental templates

and external standard curves are necessary for absolute quantification. Other potential variables in

qPCR assays include preparation and quantification of the standard curve, the subsequent qPCR

efficiency, as well as different qPCR reagents and analysis software that are used [145, 146].

Therefore, only the ‘absolute’ numbers generated from the same single qPCR assay can be

compared when using the same standard curve [147].

Despite of the unparalleled specificity and sensitivity provided by qPCR-based approaches to target

the sequences from a mixed community sample, one noteworthy limitation is that qPCR-based

approaches require prior knowledge of the specific target gene of interest. This inevitably results in

the fact that any qPCR-based method can not be used to analyze the sequences of unknown species

which are likely to account for the vast majority of the world's millions of species. The development

of “omic” approaches in recent years, however, can circumvent this problem by providing a PCR-

independent assessment of microbial diversity. Hence, combing qPCR technique with other

approaches, such as metagenomic and metatranscriptomic analysis, can enable researchers to gain a

more comprehensive and in-depth understanding of complex microbial communities.

2.3 Transcriptomics The development of next-generation sequencing techniques, allowing for the analysis of microbial

population at a large-scale, has brought forth novel applications, such as metatranscriptomics and

metaproteomics which revolutionize the study of complex microbial communities, such as biofilms.

The transcriptome is the complete collection of transcribed elements of the genome present in a cell

or tissue at a specific development stage or physiological condition. The difference in gene

expression patterns between biofilm and planktonic bacteria modes of growth has been well

established [148, 149]. Moreover, it is not surprising that the intricate interactions between species

in multispecies biofilms can also bring about major changes in gene expression compared to single

species biofilms.

All existing technologies that have been developed to deduce and quantify the transcriptome can be

attributed to either hybridization- or sequencing-based approaches. Hybridization-based approaches

are typically based on the custom-made microarrays or commercial high-density oligo microarrays

that are incubated with fluorescently labeled cDNA. Although microarray techniques are high

throughput and relatively inexpensive, these methods have several limitations, such as high

28

background levels caused by cross-hybridization, difficulty in detecting and quantifying low-

abundance species owing to the analog nature of the signal and challenges with comparing

expression levels between laboratories and across platforms. Furthermore, microarray analyses rely

on existing knowledge about genome sequence [150]. These limitations, however, have been

surmounted by sequencing-based approaches since the remarkable sequencing technology have

exploded onto the scene, offering dramatically lower per-base costs. Especially, the recently

developed next-generation sequencing techniques have opened new doors in the field of

transcriptomic analysis, prompting rapid emergence of RNA sequencing, termed RNA-Seq, which

directly determines the cDNA sequence from an organism of interest. The major steps involved in

bacterial transcriptome sequencing are depicted in Figure 2 [151]. In general, a population of RNA

is converted to cDNA library and then sequenced in a high-throughput manner to obtain short reads.

Thereafter, the resulting reads are assembled using either de novo or genome-guided approach to

produce a genome-scale transcription map that consists of both the transcriptional structure and/or

level of expression for each gene [152]. The output length of sequence reads and depth of coverage

vary depend on the different DNA sequencing platform used. Three major commercial next-

generation sequencing platforms: Roche's 454, Illumina's Solexa and Applied Biosystems' SOLiD

are commonly used worldwide.

Although RNA-Seq, as a novel field of research, is still under active development, it has clear

advantages over existing technologies. First, it allows the detection of transcripts in non-model

organisms with unknown genomic information. Given the tremendous diversity of uncultivated

microorganisms, future transcriptomic studies have the potential to identify new non-coding RNA

families. Second, it has a much larger dynamic range (spanning five orders of magnitude) compared

with DNA microarrays, and its high accuracy for quantifying expression levels has been verified by

quantitative PCR [153]. Third, it can reveal the precise location of transcription boundaries and

identify single-nucleotide polymorphism (SNPs) in the transcribed regions which make RNA-Seq

useful in complex transcriptomic studies. Overall, RNA-Seq is such a powerful and promising tool

that it enables the entire transcriptome to be analyzed in a very high-throughput and quantitative

manner and offers single-base resolution while concurrently, profiling gene expression levels at a

genome scale. By the use of RNA-Seq technology, the difference of gene expression between

mature P. aeruginosa biofilms and planktonic cells was firstly evaluated recently [154]. A set of

genes that were specifically regulated in biofilms were identified, including genes involved in type

three secretion, adaptation to microaerophilic growth and the production of extracellular matrix

components, indicating that biofilms are not just surface attached cells in stationary phase.

Moreover, the qualitative analysis of the RNA-Seq data revealed the enrichment of the 5'-ends of

the original transcripts, enabling an accurate prediction of transcriptional start sites (TSS). In spite

of many studies on expression profile of monospecies biofilms, metatranscriptomic analysis of

multispecies biofilm opens up the possibilities for assessing gene expression profiles of whole

microbial communities, facilitating the identification of genes of importance under different

environmental conditions. Such a study was conducted recently to study patterns of community

gene expression in a multispecies biofilm model composed of oral bacteria and periodontal

pathogens. The metatranscriptomic data demonstrated the changes in gene expression profiles of the

29

organisms present in the healthy community after the addition of periodontal pathogens to this

model and these changes can be accurately evaluated, focusing either on changes at the gene level

or treating the transcriptome of the community as a whole [155].

Figure 2 Strategies used in RNA-Seq experiments for bacterial transcriptomic analysis. (a) Outline of the

general steps involved in a typical RNA-Seq experiment. (b) Details of an RNA-Seq experiment used for

whole-transcriptome profiling of Burkholderia cenocepacia [156]. (c) Procedure used to identify sRNAs

associated with Hfq in Salmonella typhimurium [157]. (d) Differential RNA-Seq (dRNA-Seq) used to

identify putative transcriptional start sites in Helicobacter pylori [158].

Despite the appealing advantages described above, metatranscriptomic studies of microbial

assemblages in situ are still rare so far. This is due to several technological challenges associated

with the processing of RNA samples, including the recovery of high-quality mRNA, high demand

for computational power and biostatistical expertise and relatively high costs of more depth

sequencing for more complex transcriptome. The newly emerging metraproteomics analysis, which

aims at assessing the entire protein complement of environmental microbiota at a given time point

[159], has a huge potential to link the diversity and activity of microbial communities with their

impact on ecological functions. With the falling costs of sequencing, it can be expected that more

and more transcriptomics and proteomics will contribute to a deeper understanding of functional

dynamics of microbial communities and evolutionary processes. Thence, the query is proposed

recently about if it is the time to start an integrated omics approach to biofilms - “biofomics”. This

will then lead to the construction of a free on-line database where biofilm signatures are identified

and interrogated, including the ability of a microorganism to attach to surfaces, interact with its

neighbors and form biofilms, which therefore can provide comprehensive data sets about the overall

behavior of the microorganism or system [160]. Nevertheless, before commencing such an action, a

30

versatile, reliable and high-throughput biofilm growing device and appropriate methods for biofilm

analysis should be selected with the purpose of minimizing variations and the consequent need of

taking a higher number of replicates. However, whether such a device and analytical methods are

already fully developed, especially for multispecies biofilms, remains open to question.

31

3 Biofilms and protozoa

3.1 Protozoa Protozoa are unicellular, ubiquitous and colorless eukaryotes with size varying from a few to

hundreds of micrometers. Most protozoa are heterotrophic and generally feed on bacteria and other

smaller microorganisms, e.g. fungi and algae. A typical predator-prey interaction exhibits three

stages of feeding process: contact, capture and ingestion (phagocytosis). Motility plays an important

role for protozoa in capturing food and avoiding unfavorable environmental conditions. Based on

the types of locomotion, protozoa can be divided into three major groups: amoebae, flagellates and

ciliates. Amoebae produce pseudopodia for both locomotion and food-acquiring, while flagellates

possess one or more flagella and ciliates use numerous small cilia.

In agricultural soil, heterotrophic flagellates and naked amoebae predominate, with numbers

ranging from l0,000 to 1,00,000 per gram of arable soil and by forming cysts in their life cycle, they

can persist through adverse soil conditions, such as drought stress [31]. Protozoa are important for

soil nutrient cycling by feeding on bacteria and releasing excess nitrogen into the soil environment

which make them valuable in maintaining microbial equilibrium in the soil. Moreover, protozoa

graze different bacteria to different degrees depending on the characteristics of bacteria, including

cell size [161], cell surface properties [162], rate of motion [163], extent of biofilm formation [164]

as well as the nutritional and biochemical status [165]. There is also cumulative evidence that

secondary metabolites produced by bacteria may make them less susceptible to grazing [166-168].

Although amoebae grazing appears to be non-size-selective, flagellates and ciliates have preference

for medium-size bacterial cells [169, 170]. By altering bacterial size distribution, grazing is likely to

affect the bacterial community structure. In addition, another mechanism involved in the effect of

protozoa has on the bacterial community structure is the lower edibility of gram-positive bacteria

[171] due to the lower rate of digestion of cell wall [172]. However, this is not always the case as

many gram-negative bacteria are inedible while many gram-positive bacteria are adequate food

sources for protozoa [173, 174]. For instance, both Bacillus licheniformis and Pseudomonas

aeruginosa were found to produce toxic substance to the amoebae [175] and some other features of

bacterial cells like size, cell morphology and motility, are not related to gram status but have been

shown to influence feeding selectivity. Additionally, the bacterial community structure may also be

affected by non-selective grazing. For example, growth rates have been demonstrated to affect the

survival of bacteria in environments with intense protozoan predation [176, 177], as a fast-growing

species may survive at higher densities than the slow-growing organisms by compensating for cell

loss faster. Moreover, by decreasing bacterial numbers and releasing nutrient immobilized by soil

microorganisms, protozoa grazing may reduce interspecies competition for substrates [178]. There

is also report that shifts in bacterial community composition resulted from the enhancement of

grazing which disturbed the established balance between population-specific growth and mortality

rates of bacteria by stimulating viral activity [179].

3.2 Biofilms - the response of cell consortia to protozoan grazing Environmental bacteria, as an integral part of microbial food chain, are frequently confronted with

consuming protozoa, which is considered to be a major cause of bacterial death in most freshwater,

32

marine and moist soil habitats [180]. Given their pronounced effects on prey fitness, such as

reduced bacterial biomass and changes in both the composition and morphological structure,

bacteria have developed diverse strategies against protozoa predation. One of the remarkably

effective ways is the formation of cell clusters, e.g., biofilms, although it is not clear up to now that

whether enhanced microcolony formation under grazing pressure is a direct defense strategy or an

indirect stimulation as a result of nutrient recycling and/or chemical cues [181].

Generally, in natural environment, protozoan colonization of bacterial biofilms has three succession

stages. First, early surface colonizers, heterotrophic flagellates, colonize the surface within minutes

after exposure due to their high mobility and abundance in the environment. This is followed by

ciliates and then the later amoebae. The efficient grazing protection provided by microcolony

formation depends on the stage of biofilm development and the feeding mode of the grazer [181,

182], that is early biofilms are colonized mainly by generalists which feed on suspended and

attached bacteria, while later biofilm stages are colonized by more permanently attached specialists

feeding on surface-associated bacteria [183]. Clearly, biofilms provide a dramatically different

feeding environment for protozoa in contrast to planktonic cells. For example, introducing ciliate

grazers to biofilms formed by the yeast, Cryptococcus spp., could result in the 1.75 higher levels of

biofilm metabolism compared with non-grazed controls. Furthermore, the preferential grazing on

the noncellular biofilm matrix over cells embedding in the biofilm demonstrated that EPS could

serve as source of nutrients and energy for protists which benefits the biofilm not only from

physical protection against ingestion but also from the enhanced nutrient recycling [184]. In

addition, the close proximity of bacteria within biofilms can result in enhanced opportunities for

interactions such as horizontal gene transfer and quorum sensing (QS), which may induce multiple

anti-predator mechanisms that are expressed at different stages of biofilm development. Evidences

that QS is involved in regulating anti-predator biofilms have been reported in various

microorganisms. Matz and colleagues proved the key roles of microcolonies formation (Figure 3)

and inhibitor production induced by quorum sensing in the resistance of Pseudomonas aeruginosa

biofilms to protozoan grazing [19]. And the further study showed that inhibitor production of

mature P. aeruginosa biofilms is effective against a wider range of biofilm-feeding predators while

microcolony-mediated protection is only beneficial in the early stages of biofilm formation [182].

A recent study suggested that protozoan resistance of pathogen P. aeruginosa did not result from

activation of QS-regulated public goods, but from the larger and stronger biofilms and thus,

concluded that protist predation can favor cooperation within bacterial species [185]. While in

Serratia marcescens, QS-controlled, biofilm-specific differentiation of filaments and cell chains in

biofilms provides an efficient mechanism against protozoan grazing [21], Vibrio cholera can resist a

range of predators by QS-induced production of an unknown toxin [186], which has been proved to

be more important in grazing resistance of late biofilms compared with the protection provided by

EPS. In early river biofilms under semi-natural conditions, it was shown that the presence of

heterotrophic flagellates significantly reduced the abundance of single bacterial cells, but stimulated

the formation of bacterial microcolonies [187]. There is also report indicating that protozoa from

healthy activated sludge could initially disturb the biofilm development in flow cells but later

stimulate its growth [188]. This stimulation could be caused by release of nutrients or non-selective

33

protozoan grazing, which can result in total domination of faster-growing bacteria, whereas some

strains are able to even outgrow predators [189-192]. Evidence that biofilm-specific chemical

defenses against protozoan predators are a widespread phenomenon in a diverse set of marine

bacteria has been presented in a study comparing efficacy of chemical defenses in biofilms and

planktonic phases of growth [193]. Despite the intense grazing pressure faced by biofilms due to the

lack of mobility to escape protozoan grazing, the fact that bacteria live as biofilms in many

environments, could indicate that biofilms have a relatively high anti-predator fitness. And this

increased fitness is likely to be favored as a grazing resistance mechanism that has evolved during

the long history of close ecological associations between bacteria and protozoa.

(a) (b) (c)

Figure 3 The effects of flagellate Rhynchomonas nasuta grazing on Pseudomonas aeruginosa biofilms [9,

19]. (a) 3-day-old biofilms of P. aeruginosa PAO1 growing without flagellate. (b) 3-day-old biofilms of P.

aeruginosa PAO1 growing with flagellate. (c) 7-day-old biofilms of P. aeruginosa PAO1. Biofilms were

pregrown for 3 days before the addition of the flagellate. Scale bar = 50 µm.

Nevertheless, some flagellates, such as Cafeteria roenbergensi, Bodo saliens and Caecitellus

parvulus, as efficient bacterivores, are proved to be able to reduce the numbers of surface-deposited

bacteria just as other protozoa restrict numbers of suspended bacteria [194]. Protozoan grazing have

been reported to affect biofilm development by inducing fragmentation and sloughing [195, 196] as

well as changing exopolymers distributions and chemical natures of biofilms [197], hence, result in

the enhanced spatial and temporal heterogeneity within biofilm communities. Acanthamoeba

castellanii and Colpoda maupasi were proved to significantly influence the development and

population dynamics of mixed biofilm communities. A. castellanii, as a predominant biofilm grazer,

integrated in biofilms, whereas, C. maupasi reduced biofilm thickness by up to 60% as a result of

grazing and/or their movement causing sloughing, indicating biofilm growth may not provide total

protection against protozoa grazing [198]. Böhme A et al. [21] demonstrated the predation by

protozoa with different feeding modes and motility resulted in different morphological structures of

multispecies biofilms, including smaller microcolonies with lower maximal and basal layer

thickness, larger or mushroom-shaped microcolonies. Furthermore, protozoan grazing could

improve mass transfer of nutrients into biofilms, thus accelerate microbial growth. Studies on

feeding interactions of two contrasting ciliates with bacterial biofilms have shown that feeding

preferences for spatially separated Pseudomonas costantinii biofilms over Serratia plymuthica

34

biofilms can be initiated by the detection of dissolved chemical cues or contact-based detection of

bacterial attributes [199].

3.3 Protozoa and biofilms - reservoirs of pathogenic bacteria The response of bacteria living in a biofilm to the protozoa can change from defensive to

exploitative, making protozoa an environmental reservoir for bacterial pathogens. For instance, the

growth/survival of bacteria inside the amoebae could give rise to several facultative and obligate

pathogens, e.g. Listeria, Mycobacterium and Legionella [200]. What's more, intracellular bacterial

growth in protozoa or as a biofilm in soil could lead to distinct phenotypes. Legionella pneumophila

and Mycobacterium avium grown in amoebae were reported to be more invasive for macrophages

and/or epithelial cells than those grown in vitro [201, 202], moreover, L. pneumophila, replication

in amoebae could cause large morphological and biochemical changes and induce an antibiotic -

and biocide-resistant phenotype [203]. The fact that Legionellosis, in some cases, results from

inhalation of aerosols of biofilm- and amoebae-containing L. pneumophila, suggests that protozoa

probably have a crucial role in the maintenance and dissemination of human pathogenic bacteria in

the environment. Similar to the resistance mechanisms of Pseudomonas aeruginosa biofilms

against protozoan grazing, the resistance of biofilm infections to phagocyte-mediated host defense

also results from the presence of matrix-enclosed aggregates and the quorum sensing-induced

secretion of virulence factors [204, 205]. From a population dynamic-ecological perspective, the

control of the replication of the pathogens at the stage of nonspecific, constitutive host defenses is

analogous to that of a predator-prey system, with the pathogens being the prey and the phagocytic

cells the predators [206]. Future studies directed at unraveling the role of protozoan grazing in the

structural and functional stability of biofilms could contribute to a better understanding of the

environmental persistence and continuous evolution of bacterial pathogens.

35

4 Where are we going with biofilms? - In the context of microbial ecology Since ecological interaction is one of the main drivers for shaping microbial community, it is

obvious that understanding the mechanisms underlying microbial ecology of biofilm is the core of

research. As a scientific discipline, microbial ecology examines three aspects of a community [207]:

Structure – Which microorganisms are present?

Function – Which metabolic capabilities are available/expressed that can support adaptation of this

community to environmental conditions?

Interaction – How microorganisms interact with one another and with the surrounding environment?

These questions can be answered by the more recent and exciting application of the molecular-

biology tools, which have marked a major breakthrough in transition from reductionism to holism,

that is to elucidating the microbial community as an integral, coevolving system instead of

identifying independent individuals. Transcriptomics and proteomics have come on the scene after

genomics to allow a comprehensive study of the community function at the transcription and

translation levels. Furthermore, metabolomics, described as the comprehensive analysis of the

complete set of metabolites (metabolome) produced within each bacterial species and the entire

microflora [208], can potentially provide a more accurate snapshot of the actual physiological state

of the microorganisms [209, 210] and thus complement transcriptomics and proteomics to assess

genetic function as metabolic pathway fluxs [211]. Since one factor affecting the dynamics of

microbial community in an environmental niche is the metabolic activity from each member,

metabolomics play a crucial role in exploring the biofilm dynamics and interactions within the

microbial ecosystems. A good illustration of linking metabolomics with proteomics to assess

functional differentiation and interactions was provided by studying a dual-species biofilm

composed of Leptospirillum groups II and III. The findings demonstrated that strong metabolomic

segregation exhibited organism-specific correlation patterns which reflected the functional

differentiation of these two species, indicating that the evolutionary divergence had lessened

competition between co-existing microorganisms and allowed them to occupy distinct niches [212].

Moreover, from reconstruction of near-complete or partial recovery of genomes, metabolic network

could be deduced in a natural acidophilic biofilm which provided insights into community

interactions and functions [213]. Besides, metabolic cooperation that is achieved by synergistic

relationship between microorganisms or mutual exchange of metabolites within a community, can

be inferred by analyzing metabolomic data [214].

The emergence of various high-throughput technologies now allows more “meta-omics”, e.g.

metagenomics, metatranscriptomics and metametabolomics that aim to characterize complete

microbial ecosystems. Thus, eco-systems biology was proposed by combining these

microbiolomics to interrogate the diversity, function and ecology of the microbial community [214].

Biofilms, as the dominate lifestyle of microorganisms in nature, are undoubtedly ideal for

conducting these studies which will provide a more complete picture of biofilm community

behavior in the face of environmental stresses. However, the complementary utilization of different

“omics” methods also put forward the challenge in meta-data analysis as well as technical

challenges, including nucleic acids and protein extraction from environmental samples, mRNA

36

instability and low abundance of certain gene transcripts in total RNA. Especially, compared with

genomic, transcriptomic and proteomic analysis, the simultaneous determination of metabolome at

a given physiological state is extremely difficult due to the more variable products generated from

metabolic reactions [215]. Besides, this may be further complicated when confronted with

multispecies biofilms in the environment as a result of the structural and physiological complexity

and unevenness of the microbial community. Although the “meta-omics” technologies are expected

to revolutionize our understanding of the microbial community, due to the heterogeneity of

microniches within a biofilm, particular attention should also be paid to transcriptomic analysis at

single cell resolution in order to identify cell-to-cell spatiotemporal variations. Furthermore, prior to

sequencing, hypothesis-driven research should be taken into consideration as it may allow more

efficient identification and verification of predictions generated by the large-scale data [216].

In parallel, the high-resolution microscopic approaches also continue to shed light on the physical

and structural properties of biofilms which enable the in situ investigation of cell attachment,

organization and succession within microbial communities. Nevertheless, the three-dimensional,

computer-based visualization and quantification of microbial communities in natural ecosystems are

still challenges because of the insufficient discrimination of different species and the limited

oligonucleotide probes available for various species. When reporter genes are applied, genetic

manipulation is also involved for the microorganism being studied which appears to be practically

infeasible for unknown species in nature, resulting in the restricted application of these methods to

only single species or simple mixed-species population. This, thus, calls for combining multiple

technologies or even combining interdisciplinary research efforts in biofilm science. Moreover, the

need to standardize experimental approaches, which would allow for more reliable and comparable

results, is also increasingly appreciated.

More attention is being given to multispecies biofilms than ever before, as evidenced by a surge of

publications in this field since 2003. While some molecular mechanisms underlying intra- and inter-

species interactions between individual strains have been clarified, mostly are only applicable to

dual-species biofilms, which is notably in contrast to the composition of biofilms in environmental

habitats. The biofilms in natural environment are much more complex entities than what we have

examined in the laboratory owing to the structural and physiological heterogeneity as well as the

extensive and striking interrelations between their components. And the coordinated behaviors of

biofilm have led to the idea that biofilm acts more like an orchestrated multicellular organism than a

collection of organisms. This view has been supported by the observation that the altruistic self-

sacrifice of the majority of the population when exposed to adverse conditions, such as programmed

cell death [217], could ultimately enhance the survival of the whole community [113], which

violates the principles of Darwinian evolution. This can be explained by the statement that Darwin’s

evolutionary theory focused intensely on one level of existence (organic species) while failing to

conceive the existence of multispecies communities and their potential role in the origin of species

[21]. However, objections are also raised to this fascinating analogy, mainly because upon their

decay, the biofilm cells can turn back to planktonic form in response to environmental signals in

contrast to the irreversible differentiation of the cells in a multicellular organism. Additionally,

whether the biofilm-specific genes that hierarchically regulate transition through specific stages in

37

biofilm formation have been identified successfully, is still be questioned [116]. Hence, the term

“biofilm phenotype” is commonly used rather than “biofilm genotype”. From another point of view,

[218]“city of microbes” was put forward to describe the association of microorganisms in biofilms,

where inhabitants select location, limit settlements of too many bacteria, store energy in

exopolysaccharide, transfer genetic material horizontally and communicate with their neighbors.

When conditions in the “city” are less comfortable, inhabitants migration occurs [218, 219]. In the

ecological context, the perception of biofilms as microbial landscapes has been proposed by Battin

TJ et al. [220] to emphasize the spatially explicit dimension and unify numerous facets of biofilms,

such as biodiversity, dynamics and ecosystem function.

In summary, multispecies biofilm provides a valuable model for testing ecological and evolutionary

theories relating to interspecies interactions. By integrating ecological and evolutionary frameworks

with the promising molecular technologies mentioned in this thesis, a genetic and biochemical

understanding of the interactions between biofilm landscape components and the operating

ecological process is supposed to be achieved, and thus helps us to manage biofilms to provide

services to society.

38

5 References 1. Kolenbrander PE, Palmer RJ, Periasamy S, Jakubovics NS (2010) Oral multispecies biofilm

development and the key role of cell–cell distance. Nature Reviews Microbiology 8: 471-480.

2. Costerton JW, Cheng K, Geesey GG, Ladd TI, Nickel JC, Dasgupta M, Marrie TJ (1987) Bacterial

biofilms in nature and disease. Annual Reviews in Microbiology 41: 435-464.

3. Donlan RM, Costerton JW (2002) Biofilms: survival mechanisms of clinically relevant

microorganisms. Clinical microbiology reviews 15: 167-193.

4. Costerton J, Stewart PS, Greenberg E (1999) Bacterial biofilms: a common cause of persistent

infections. Science 284: 1318-1322.

5. Potera C (1996) Biofilms invade microbiology. Science 273: 1795-1797.

6. Davey ME, O'toole GA (2000) Microbial biofilms: from ecology to molecular genetics.

Microbiology and molecular biology reviews 64: 847-867.

7. Almås ÅR, Mulder J, Bakken LR (2005) Trace metal exposure of soil bacteria depends on their

position in the soil matrix. Environmental science & technology 39: 5927-5932.

8. McKew BA, Taylor JD, McGenity TJ, Underwood GJ (2010) Resistance and resilience of benthic

biofilm communities from a temperate saltmarsh to desiccation and rewetting. The ISME journal 5:

30-41.

9. Matz C, Kjelleberg S (2005) Off the hook–how bacteria survive protozoan grazing. Trends in

microbiology 13: 302-307.

10. Stewart PS, William Costerton J (2001) Antibiotic resistance of bacteria in biofilms. The Lancet 358:

135-138.

11. Jesaitis AJ, Franklin MJ, Berglund D, Sasaki M, Lord CI, Bleazard JB, Duffy JE, Beyenal H,

Lewandowski Z (2003) Compromised host defense on Pseudomonas aeruginosa biofilms:

characterization of neutrophil and biofilm interactions. The Journal of Immunology 171: 4329-4339.

12. Jaffe A, Chabbert YA, Semonin O (1982) Role of porin proteins OmpF and OmpC in the permeation

of beta-lactams. Antimicrobial agents and chemotherapy 22: 942-948.

13. Roberson EB, Firestone MK (1992) Relationship between desiccation and exopolysaccharide

production in a soil Pseudomonas sp. Applied and Environmental Microbiology 58: 1284-1291.

14. Harrison JJ, Turner RJ, Ceri H (2005) Persister cells, the biofilm matrix and tolerance to metal

cations in biofilm and planktonic Pseudomonas aeruginosa. Environmental Microbiology 7: 981-

994.

15. Vuong C, Voyich JM, Fischer ER, Braughton KR, Whitney AR, DeLeo FR, Otto M (2004)

Polysaccharide intercellular adhesin (PIA) protects Staphylococcus epidermidis against major

components of the human innate immune system. Cellular microbiology 6: 269-275.

16. Evans D, Allison D, Brown M, Gilbert P (1991) Susceptibility of Pseudomonas aeruginosa and

Escherichia coli biofilms towards ciprofloxacin: effect of specific growth rate. Journal of

Antimicrobial Chemotherapy 27: 177-184.

17. Hengge-Aronis R (1996) Regulation of gene expression during entry into stationary phase.

Escherichia coli and Salmonella: cellular and molecular biology, 2nd ed ASM Press, Washington,

DC: 1497-1512.

18. Bagge N, Hentzer M, Andersen JB, Ciofu O, Givskov M, Høiby N (2004) Dynamics and spatial

distribution of β-lactamase expression in Pseudomonas aeruginosa biofilms. Antimicrobial agents

and chemotherapy 48: 1168-1174.

19. Matz C, Bergfeld T, Rice SA, Kjelleberg S (2004) Microcolonies, quorum sensing and cytotoxicity

determine the survival of Pseudomonas aeruginosa biofilms exposed to protozoan grazing.

Environmental microbiology 6: 218-226.

20. Lewis K (2001) Riddle of biofilm resistance. Antimicrobial agents and chemotherapy 45: 999-1007.

21. Queck SY, Weitere M, Moreno AM, Rice SA, Kjelleberg S (2006) The role of quorum sensing

mediated developmental traits in the resistance of Serratia marcescens biofilms against protozoan

grazing. Environmental microbiology 8: 1017-1025.

22. Hall-Stoodley L, Costerton JW, Stoodley P (2004) Bacterial biofilms: from the natural environment

to infectious diseases. Nature Reviews Microbiology 2: 95-108.

39

23. Torsvik V, Goksøyr J, Daae FL (1990) High diversity in DNA of soil bacteria. Applied and

environmental microbiology 56: 782-787.

24. Gans J, Wolinsky M, Dunbar J (2005) Computational improvements reveal great bacterial diversity

and high metal toxicity in soil. Science 309: 1387-1390.

25. Nunan N, Wu K, Young IM, Crawford JW, Ritz K (2003) Spatial distribution of bacterial

communities and their relationships with the micro-architecture of soil. FEMS Microbiology

Ecology 44: 203-215.

26. Sutherland IW (2001) The biofilm matrix–an immobilized but dynamic microbial environment.

Trends in microbiology 9: 222-227.

27. Chang W-S, van de Mortel M, Nielsen L, de Guzman GN, Li X, Halverson LJ (2007) Alginate

production by Pseudomonas putida creates a hydrated microenvironment and contributes to biofilm

architecture and stress tolerance under water-limiting conditions. Journal of bacteriology 189: 8290-

8299.

28. Lee S-E, Kim S-H, Lee J-U, Chon H-T (2013) Technical development of geomicrobiological

fixation of heavy metals in soils and sediments by biosorption and bioreduction-case histories in

Korea. EGU General Assembly Conference Abstracts, vol. 15, pp. 2702.

29. Sandoval-Carrasco CA, Ahuatzi-Chacón D, Galíndez-Mayer J, Ruiz-Ordaz N, Juárez-Ramírez C,

Martínez-Jerónimo F (2013) Biodegradation of a mixture of the herbicides ametryn, and 2, 4-

dichlorophenoxyacetic acid (2, 4-D) in a compartmentalized biofilm reactor. Bioresource technology

145: 33-36.

30. Walker TS, Bais HP, Déziel E, Schweizer HP, Rahme LG, Fall R, Vivanco JM (2004) Pseudomonas

aeruginosa-plant root interactions. Pathogenicity, biofilm formation, and root exudation. Plant

physiology 134: 320-331.

31. Ekelund F, Rønn R (1994) Notes on protozoa in agricultural soil with emphasis on heterotrophic

flagellates and naked amoebae and their ecology. FEMS Microbiology Reviews 15: 321-353.

32. Marsh PD (2006) Dental plaque as a biofilm and a microbial community–implications for health and

disease. BMC Oral Health 6: S14.

33. Suppiger A, Schmid N, Aguilar C, Pessi G, Eberl L (2013) Two quorum sensing systems control

biofilm formation and virulence in members of the Burkholderia cepacia complex. Virulence 4: 19-

18.

34. Huber B, Riedel K, Hentzer M, Heydorn A, Gotschlich A, Givskov M, Molin S, Eberl L (2001) The

cep quorum-sensing system of Burkholderia cepacia H111 controls biofilm formation and swarming

motility. Microbiology 147: 2517-2528.

35. Schmid N, Pessi G, Deng Y, Aguilar C, Carlier AL, Grunau A, Omasits U, Zhang L-H, Ahrens CH,

Eberl L (2012) The AHL-and BDSF-dependent Quorum sensing systems control specific and

overlapping sets of genes in Burkholderia cenocepacia H111. PloS one 7: e49966.

36. Savage VJ, Chopra I, O'Neill AJ (2013) Staphylococcus aureus biofilms promote horizontal transfer

of antibiotic resistance. Antimicrobial agents and chemotherapy 57: 1968-1970.

37. Martínez JL (2008) Antibiotics and antibiotic resistance genes in natural environments. Science 321:

365-367.

38. Forsberg KJ, Reyes A, Wang B, Selleck EM, Sommer MO, Dantas G (2012) The shared antibiotic

resistome of soil bacteria and human pathogens. Science 337: 1107-1111.

39. Seneviratne G, Kecskés ML, Kennedy IR (2008) Biofilmed biofertilisers: novel inoculants for

efficient nutrient use in plants. In: Kennedy, I.R., Choudhury, A.T.M.A., Kecskés, M.L., Rose, M.T.

(Eds.), Efficient Nutrient Use in Rice Production in Vietnam Achieved Using Inoculant Biofertilisers.

Proceedings of a Project (SMCN/2002/073) Workshop Held in Hanoi, Vietnam, 12e13 October 2007.

ACIAR Proceedings No. 130, Canberra, pp. 126e130.

40. Bianciotto V, Bonfante P (2002) Arbuscular mycorrhizal fungi: a specialised niche for rhizospheric

and endocellular bacteria. Antonie van Leeuwenhoek 81: 365-371.

41. Seneviratne G, Zavahir J, Bandara W, Weerasekara M (2008) Fungal-bacterial biofilms: their

development for novel biotechnological applications. World Journal of Microbiology and

Biotechnology 24: 739-743.

40

42. Jayasinghearachchi H, Seneviratae G (2010) A Mushroom-fungus helps improve endophytic

colonization of tomato by Pseudomonas fluorescens through biofilm formation. Research Journal of

Microbiology 5: 689-695.

43. Jayasinghearachchi H, Seneviratne G (2004) A bradyrhizobial-Penicillium spp. biofilm with

nitrogenase activity improves N2 fixing symbiosis of soybean. Biology and fertility of soils 40: 432-

434.

44. LoVLEY DR, Klug MJ (1986) Model for the distribution of sulfate reduction and methanogenesis in

freshwater sediments. Geochimica et Cosmochimica Acta 50: 11-18.

45. Stewart PS, Franklin MJ (2008) Physiological heterogeneity in biofilms. Nature Reviews

Microbiology 6: 199-210.

46. Matz C (2011) Competition, Communication, Cooperation: Molecular Crosstalk in Multi-species

BiofilmsBiofilm Highlights. In: Flemming, H.-C., Wingender, J., Szewzyk, U. (Eds.), Biofilm

Highlight. Springer Series on Biofilms, vol. 5. Springer-Verlag Berlin, Heidelberg, pp. 29-40.

47. Rickard AH, Gilbert P, High NJ, Kolenbrander PE, Handley PS (2003) Bacterial coaggregation: an

integral process in the development of multi-species biofilms. Trends in microbiology 11: 94-100.

48. Kolenbrander PE, Andersen RN, Blehert DS, Egland PG, Foster JS, Palmer RJ (2002)

Communication among oral bacteria. Microbiology and molecular biology reviews 66: 486-505.

49. Klayman BJ, Volden PA, Stewart PS, Camper AK (2009) Escherichia coli O157: H7 requires

colonizing partner to adhere and persist in a capillary flow cell. Environmental science & technology

43: 2105-2111.

50. Bradshaw DJ, Marsh PD, Watson GK, Allison C (1998) Role of Fusobacterium nucleatum and

coaggregation in anaerobe survival in planktonic and biofilm oral microbial communities during

aeration. Infection and immunity 66: 4729-4732.

51. Gilbert P, Maira-Litran T, McBain AJ, Rickard AH, Whyte FW (2002) The physiology and

collective recalcitrance of microbial biofilm communities. Advances in microbial physiology 46:

203-256.

52. Adam B, Baillie GS, DOUGLAS LJ (2002) Mixed species biofilms of Candida albicans and

Staphylococcus epidermidis. Journal of medical microbiology 51: 344-349.

53. Boonchan S, Britz ML, Stanley GA (2000) Degradation and mineralization of high-molecular-

weight polycyclic aromatic hydrocarbons by defined fungal-bacterial cocultures. Applied and

Environmental Microbiology 66: 1007-1019.

54. Farabegoli G, Chiavola A, Rolle E (2008) Remediation of chlorophenol-and phenol-contaminated

groundwater by a sequencing batch biofilm reactor. Water Science and Technology 58: 295–301.

55. Flemming H-C, Wingender J (2010) The biofilm matrix. Nature Reviews Microbiology 8: 623-633.

56. Richard G, Burns, Richard P. Dick (2002) Microbes and Enzymes in Biofilms. Enzymes in the

Environment: Activity, Ecology, and Applications 86: 307.

57. Jakubovics NS, Gill SR, Iobst SE, Vickerman M, Kolenbrander PE (2008) Regulation of gene

expression in a mixed-genus community: stabilized arginine biosynthesis in Streptococcus gordonii

by coaggregation with Actinomyces naeslundii. Journal of bacteriology 190: 3646-3657.

58. Parsek MR, Greenberg E (2005) Sociomicrobiology: the connections between quorum sensing and

biofilms. Trends in microbiology 13: 27-33.

59. Häussler S, Parsek MR (2010) Biofilms 2009: new perspectives at the heart of surface-associated

microbial communities. Journal of bacteriology 192: 2941-2949.

60. Riedel K, Hentzer M, Geisenberger O, Huber B, Steidle A, Wu H, Høiby N, Givskov M, Molin S,

Eberl L (2001) N-acylhomoserine-lactone-mediated communication between Pseudomonas

aeruginosa and Burkholderia cepacia in mixed biofilms. Microbiology 147: 3249-3262.

61. Eberl L, Tümmler B (2004) Pseudomonas aeruginosa and Burkholderia cepacia in cystic fibrosis:

genome evolution, interactions and adaptation. International journal of medical microbiology 294:

123-131.

62. Leadbetter JR, Greenberg E (2000) Metabolism of acyl-homoserine lactone quorum-sensing signals

by Variovorax paradoxus. Journal of Bacteriology 182: 6921-6926.

41

63. Liu X, Jia J, Popat R, Ortori CA, Li J, Diggle SP, Gao K, Cámara M (2011) Characterisation of two

quorum sensing systems in the endophytic Serratia plymuthica strain G3: differential control of

motility and biofilm formation according to life-style. BMC microbiology 11: 26.

64. De Kievit T (2009) Quorum sensing in Pseudomonas aeruginosa biofilms. Environmental

microbiology 11: 279-288.

65. Suntharalingam P, Cvitkovitch DG (2005) Quorum sensing in streptococcal biofilm formation.

Trends in microbiology 13: 3-6.

66. Saito Y, Fujii R, Nakagawa KI, Kuramitsu H, Okuda K, Ishihara K (2008) Stimulation of

Fusobacterium nucleatum biofilm formation by Porphyromonas gingivalis. Oral microbiology and

immunology 23: 1-6.

67. Jang Y-J, Sim J, Jun H-K, Choi B-K (2013) Differential effect of autoinducer 2 of Fusobacterium

nucleatum on oral streptococci. Archives of oral biology 58: 1594-1602.

68. Schauder S, Bassler BL (2001) The languages of bacteria. Genes & Development 15: 1468-1480.

69. Reading NC, Sperandio V (2006) Quorum sensing: the many languages of bacteria. FEMS

microbiology letters 254: 1-11.

70. Deng Y, Wu Je, Eberl L, Zhang L-H (2010) Structural and functional characterization of diffusible

signal factor family quorum-sensing signals produced by members of the Burkholderia cepacia

complex. Applied and environmental microbiology 76: 4675-4683.

71. Davies DG, Marques CN (2009) A fatty acid messenger is responsible for inducing dispersion in

microbial biofilms. Journal of bacteriology 191: 1393-1403.

72. Ryan RP, Dow JM (2011) Communication with a growing family: diffusible signal factor (DSF)

signaling in bacteria. Trends in microbiology 19: 145-152.

73. Boon C, Deng Y, Wang L-H, He Y, Xu J-L, Fan Y, Pan SQ, Zhang L-H (2007) A novel DSF-like

signal from Burkholderia cenocepacia interferes with Candida albicans morphological transition.

The ISME journal 2: 27-36.

74. Ryan RP, Fouhy Y, Garcia BF, Watt SA, Niehaus K, Yang L, Tolker‐Nielsen T, Dow JM (2008)

Interspecies signalling via the Stenotrophomonas maltophilia diffusible signal factor influences

biofilm formation and polymyxin tolerance in Pseudomonas aeruginosa. Molecular microbiology 68:

75-86.

75. Keller L, Surette MG (2006) Communication in bacteria: an ecological and evolutionary perspective.

Nature Reviews Microbiology 4: 249-258.

76. West SA, Winzer K, Gardner A, Diggle SP (2012) Quorum sensing and the confusion about

diffusion. Trends in microbiology 20: 586-594.

77. Hense BA, Kuttler C, Müller J, Rothballer M, Hartmann A, Kreft J-U (2007) Does efficiency

sensing unify diffusion and quorum sensing? Nature Reviews Microbiology 5: 230-239.

78. Tischler AD, Camilli A (2004) Cyclic diguanylate (c‐di‐GMP) regulates Vibrio cholerae biofilm

formation. Molecular microbiology 53: 857-869.

79. Beyhan S, Odell LS, Yildiz FH (2008) Identification and characterization of cyclic diguanylate

signaling systems controlling rugosity in Vibrio cholerae. Journal of bacteriology 190: 7392-7405.

80. Waters CM, Lu W, Rabinowitz JD, Bassler BL (2008) Quorum sensing controls biofilm formation in

Vibrio cholerae through modulation of cyclic di-GMP levels and repression of vpsT. Journal of

bacteriology 190: 2527-2536.

81. Caiazza NC, Merritt JH, Brothers KM, O'Toole GA (2007) Inverse regulation of biofilm formation

and swarming motility by Pseudomonas aeruginosa PA14. Journal of bacteriology 189: 3603-3612.

82. Newell PD, Monds RD, O'Toole GA (2009) LapD is a bis-(3′, 5′)-cyclic dimeric GMP-binding

protein that regulates surface attachment by Pseudomonas fluorescens Pf0–1. Proceedings of the

National Academy of Sciences 106: 3461-3466.

83. Tribble GD, Rigney TW, Dao D-HV, Wong CT, Kerr JE, Taylor BE, Pacha S, Kaplan HB (2012)

Natural competence is a major mechanism for horizontal DNA transfer in the oral pathogen

Porphyromonas gingivalis. MBio 3.

84. Wang B, Chi B, Kuramitsu H (2002) Genetic exchange between Treponema denticola and

Streptococcus gordonii in biofilms. Oral microbiology and immunology 17: 108-112.

42

85. Case RJ, Labbate M, Kjelleberg S (2008) AHL-driven quorum-sensing circuits: their frequency and

function among the Proteobacteria. The ISME journal 2: 345-349.

86. Huang Y, Zeng Y, Yu Z, Zhang J, Feng H, Lin X (2013) In silico and experimental methods

revealed highly diverse bacteria with quorum sensing and aromatics biodegradation systems-a

potential broad application on bioremediation. Bioresource technology 148: 311-316.

87. Burmølle M, Bahl MI, Jensen LB, Sørensen SJ, Hansen LH (2008) Type 3 fimbriae, encoded by the

conjugative plasmid pOLA52, enhance biofilm formation and transfer frequencies in

Enterobacteriaceae strains. Microbiology 154: 187-195.

88. Ma H, Bryers JD (2013) Non-invasive determination of conjugative transfer of plasmids bearing

antibiotic-resistance genes in biofilm-bound bacteria: effects of substrate loading and antibiotic

selection. Applied microbiology and biotechnology 97: 317-328.

89. Ehrlich GD, Ahmed A, Earl J, Hiller NL, Costerton JW, Stoodley P, Post JC, DeMeo P, Hu FZ

(2010) The distributed genome hypothesis as a rubric for understanding evolution in situ during

chronic bacterial biofilm infectious processes. FEMS Immunology & Medical Microbiology 59:

269-279.

90. Ghigo J-M (2001) Natural conjugative plasmids induce bacterial biofilm development. Nature 412:

442-445.

91. Schwering M, Song J, Louie M, Turner RJ, Ceri H (2013) Multi-species biofilms defined from

drinking water microorganisms provide increased protection against chlorine disinfection.

Biofouling 29: 917-928.

92. Lee KWK, Periasamy S, Mukherjee M, Xie C, Kjelleberg S, Rice SA (2013) Biofilm development

and enhanced stress resistance of a model, mixed-species community biofilm. The ISME Journal 1-

14.

93. Kreth J, Merritt J, Shi W, Qi F (2005) Competition and coexistence between Streptococcus mutans

and Streptococcus sanguinis in the dental biofilm. Journal of bacteriology 187: 7193-7203.

94. Rao D, Webb JS, Kjelleberg S (2005) Competitive interactions in mixed-species biofilms containing

the marine bacterium Pseudoalteromonas tunicata. Applied and environmental microbiology 71:

1729-1736.

95. Costa R, Gomes N, Krögerrecklenfort E, Opelt K, Berg G, Smalla K (2007) Pseudomonas

community structure and antagonistic potential in the rhizosphere: insights gained by combining

phylogenetic and functional gene‐based analyses. Environmental microbiology 9: 2260-2273.

96. Tait K, Sutherland I (2002) Antagonistic interactions amongst bacteriocin‐producing enteric bacteria

in dual species biofilms. Journal of applied microbiology 93: 345-352.

97. Foster KR, Bell T (2012) Competition, not cooperation, dominates interactions among culturable

microbial species. Current Biology 22: 1845-1850.

98. Molina L, Constantinescu F, Michel L, Reimmann C, Duffy B, Défago G (2003) Degradation of

pathogen quorum‐sensing molecules by soil bacteria: a preventive and curative biological control

mechanism. FEMS Microbiology Ecology 45: 71-81.

99. Peleg AY, Tampakakis E, Fuchs BB, Eliopoulos GM, Moellering RC, Mylonakis E (2008)

Prokaryote–eukaryote interactions identified by using Caenorhabditis elegans. Proceedings of the

National Academy of Sciences 105: 14585-14590.

100. West SA, Diggle SP, Buckling A, Gardner A, Griffin AS (2007) The social lives of microbes. Annu

Rev Ecol Evol Syst 38: 53-77.

101. Jass J, Surman S, Walker JT (2003) Medical biofilms. In: J. Jass, S. Surman, J. Walker (Eds.), pp.

51–72.

102. Kuenen J, Robertson LA, van Gemerden H (1985) Microbial interactions among aerobic and

anaerobic sulfur-oxidizing bacteria. Adv Microb Ecol 8: 1-59.

103. Clark F (1969) Ecological associations among soil micro-organisms. In Soil Biology 125-161.

104. Park D (1960) Antagonism—the background to soil fungi. The ecology of soil fungi: 148-159.

105. Burmølle M, Ren D, Bjarnsholt T, Sørensen SJ (2013) Interactions in multispecies biofilms: do they

actually matter? Trends in microbiology. doi: 10.1016/j.tim.2013.12.004

106. Hansen SK, Rainey PB, Haagensen JA, Molin S (2007) Evolution of species interactions in a biofilm

community. Nature 445: 533-536.

43

107. Ren D, Madsen JS, de la Cruz-Perera CI, Bergmark L, Sørensen SJ, Burmølle M (2013) High-

Throughput Screening of Multispecies Biofilm Formation and Quantitative PCR-Based Assessment

of Individual Species Proportions, Useful for Exploring Interspecific Bacterial Interactions.

Microbial ecology. doi: 10.1007/s00248-013-0315-z

108. Mitri S, Foster K (2013) A Genotypic View of Social Interactions in Microbial Communities.

Annual review of genetics 47: 247-273.

109. Nadell CD, Xavier JB, Foster KR (2009) The sociobiology of biofilms. FEMS microbiology reviews

33: 206-224.

110. West SA, Griffin AS, Gardner A, Diggle SP (2006) Social evolution theory for microorganisms.

Nature Reviews Microbiology 4: 597-607.

111. Zhang W, Sileika T, Packman AI (2013) Effects of fluid flow conditions on interactions between

species in biofilms. FEMS microbiology ecology 2: 344-354.

112. Nadell CD, Foster KR, Xavier JB (2010) Emergence of spatial structure in cell groups and the

evolution of cooperation. PLoS computational biology 6: e1000716.

113. Kreft J-U (2004) Biofilms promote altruism. Microbiology 150: 2751-2760.

114. Kelvin LKW (2013) Role of cell-cell signaling, differentiation and fitness in multi-species biofilms.

Dissertation, Nanyang Technologial University.

115. Morris JJ, Lenski RE, Zinser ER (2012) The Black Queen Hypothesis: evolution of dependencies

through adaptive gene loss. MBio 3 doi: 10.1128/mBio.00036–12

116. Monds RD, O’Toole GA (2009) The developmental model of microbial biofilms: ten years of a

paradigm up for review. Trends in microbiology 17: 73-87.

117. Schink B (2002) Synergistic interactions in the microbial world. Antonie Van Leeuwenhoek 81:

257-261.

118. Bakke R, Trulear M, Robinson J, Characklis W (1984) Activity of Pseudomonas aeruginosa in

biofilms: steady state. Biotechnology and bioengineering 26: 1418-1424.

119. Nickel J, Wright J, Ruseska I, Marrie T, Whitfield C, Costerton J (1985) Antibiotic resistance of

Pseudomonas aeruginosa colonizing a urinary catheter in vitro. European journal of clinical

microbiology 4: 213-218.

120. Kharazmi A, Giwercman B, Høiby N (1999) Robbins device in biofilm research. Methods in

enzymology 310: 207-215.

121. Stepanović S, Vuković D, Dakić I, Savić B, Švabić-Vlahović M (2000) A modified microtiter-plate

test for quantification of staphylococcal biofilm formation. Journal of Microbiological Methods 40:

175-179.

122. Harrison J, Turner R, Ceri H (2005) High-throughput metal susceptibility testing of microbial

biofilms. BMC microbiology 5: 53.

123. Bragança S, Azevedo N, Simões LC, Keevil C, Vieira M (2006) Use of fluorescent in situ

hybridization for the visualization of Helicobacter pylori in real drinking water biofilms. Water

Science and Technology 55:387–393.

124. Chavant P, Gaillard-Martinie B, Talon R, Hébraud M, Bernardi T (2007) A new device for rapid

evaluation of biofilm formation potential by bacteria. Journal of microbiological methods 68: 605-

612.

125. Silagyi K, Kim S-H, Martin Lo Y, Wei C-i (2009) Production of biofilm and quorum sensing by

Escherichia coli O157: H7 and its transfer from contact surfaces to meat, poultry, ready-to-eat deli,

and produce products. Food microbiology 26: 514-519.

126. Min K, Rickard A (2009) Coaggregation by the freshwater bacterium Sphingomonas natatoria alters

dual-species biofilm formation. Applied and environmental microbiology 75: 3987-3997.

127. Ceri H, Olson M, Stremick C, Read R, Morck D, Buret A (1999) The Calgary Biofilm Device: new

technology for rapid determination of antibiotic susceptibilities of bacterial biofilms. Journal of

clinical microbiology 37: 1771-1776.

128. Harrison JJ, Ceri H, Yerly J, Stremick CA, Hu Y, Martinuzzi R, Turner RJ (2006) The use of

microscopy and three-dimensional visualization to evaluate the structure of microbial biofilms

cultivated in the Calgary Biofilm Device. Biological procedures online 8: 194-215.

44

129. Schubotz F, Meyer‐Dombard D, Bradley A, Fredricks H, Hinrichs KU, Shock E, Summons R (2013)

Spatial and temporal variability of biomarkers and microbial diversity reveal metabolic and

community flexibility in Streamer Biofilm Communities in the Lower Geyser Basin, Yellowstone

National Park. Geobiology 11: 549-569.

130. Zhu Y, Dashper SG, Chen Y-Y, Crawford S, Slakeski N, Reynolds EC (2013) Porphyromonas

gingivalis and Treponema denticola Synergistic Polymicrobial Biofilm Development. PloS one 8:

e71727.

131. Wagner M (2009) Single-cell ecophysiology of microbes as revealed by Raman microspectroscopy

or secondary ion mass spectrometry imaging. Annual review of microbiology 63: 411-429.

132. Benoit MR, Conant CG, Ionescu-Zanetti C, Schwartz M, Matin A (2010) New device for high-

throughput viability screening of flow biofilms. Applied and environmental microbiology 76: 4136-

4142.

133. Livak KJ, Flood S, Marmaro J, Giusti W, Deetz K (1995) Oligonucleotides with fluorescent dyes at

opposite ends provide a quenched probe system useful for detecting PCR product and nucleic acid

hybridization. Genome Research 4: 357-362.

134. Smith CJ, Nedwell DB, Dong LF, Osborn AM (2007) Diversity and abundance of nitrate reductase

genes (narG and napA), nitrite reductase genes (nirS and nrfA), and their transcripts in estuarine

sediments. Applied and environmental microbiology 73: 3612-3622.

135. Baldwin BR, Nakatsu CH, Nies L (2003) Detection and enumeration of aromatic oxygenase genes

by multiplex and real-time PCR. Applied and Environmental Microbiology 69: 3350-3358.

136. Kutyavin IV, Afonina IA, Mills A, Gorn VV, Lukhtanov EA, Belousov ES, Singer MJ, Walburger

DK, Lokhov SG, Gall AA (2000) 3′-minor groove binder-DNA probes increase sequence specificity

at PCR extension temperatures. Nucleic Acids Research 28: 655-661.

137. Maeda H, Fujimoto C, Haruki Y, Maeda T, Kokeguchi S, Petelin M, Arai H, Tanimoto I, Nishimura

F, Takashiba S (2003) Quantitative real‐time PCR using TaqMan and SYBR Green for

Actinobacillus actinomycetemcomitans, Porphyromonas gingivalis, Prevotella intermedia, tetQ gene

and total bacteria. FEMS Immunology & Medical Microbiology 39: 81-86.

138. Sha Q, Vattem DA, Forstner MR, Hahn D (2013) Quantifying Salmonella population dynamics in

water and biofilms. Microbial ecology 65: 60-67.

139. Zhang L, Fritsch M, Hammond L, Landreville R, Slatculescu C, Colavita A, Mah T-F (2013)

Identification of Genes Involved in Pseudomonas aeruginosa Biofilm-Specific Resistance to

Antibiotics. PloS one 8: e61625.

140. Pan Y, Breidt F (2007) Enumeration of viable Listeria monocytogenes cells by real-time PCR with

propidium monoazide and ethidium monoazide in the presence of dead cells. Applied and

environmental microbiology 73: 8028-8031.

141. Chen NT, Chang CW (2010) Rapid quantification of viable legionellae in water and biofilm using

ethidium monoazide coupled with real‐time quantitative PCR. Journal of applied microbiology 109:

623-634.

142. Àlvarez G, González M, Isabal S, Blanc V, León R (2013) Method to quantify live and dead cells in

multi-species oral biofilm by real-time PCR with propidium monoazide. AMB Express 3: 1-8.

143. Lenz AP, Williamson KS, Pitts B, Stewart PS, Franklin MJ (2008) Localized gene expression in

Pseudomonas aeruginosa biofilms. Applied and environmental microbiology 74: 4463-4471.

144. Martin-Laurent F, Philippot L, Hallet S, Chaussod R, Germon J, Soulas G, Catroux G (2001) DNA

extraction from soils: old bias for new microbial diversity analysis methods. Applied and

environmental microbiology 67: 2354-2359.

145. Love JL, Scholes P, Gilpin B, Savill M, Lin S, Samuel L (2006) Evaluation of uncertainty in

quantitative real-time PCR. Journal of microbiological methods 67: 349-356.

146. Smith CJ, Osborn AM (2009) Advantages and limitations of quantitative PCR (Q‐PCR)‐based

approaches in microbial ecology. FEMS microbiology ecology 67: 6-20.

147. Smith CJ, Nedwell DB, Dong LF, Osborn AM (2006) Evaluation of quantitative polymerase chain

reaction‐based approaches for determining gene copy and gene transcript numbers in environmental

samples. Environmental Microbiology 8: 804-815.

45

148. Lo A, Seers C, Boyce J, Dashper S, Slakeski N, Lissel JP, Reynolds E (2009) Comparative

transcriptomic analysis of Porphyromonas gingivalis biofilm and planktonic cells. BMC

microbiology 9: 18.

149. Shemesh M, Tam A, Steinberg D (2007) Differential gene expression profiling of Streptococcus

mutans cultured under biofilm and planktonic conditions. Microbiology 153: 1307-1317.

150. Shendure J (2008) The beginning of the end for microarrays? Nature Methods 5: 585-587.

151. Febrer M, McLay K, Caccamo M, Twomey KB, Ryan RP (2011) Advances in bacterial

transcriptome and transposon insertion-site profiling using second-generation sequencing. Trends in

biotechnology 29: 586-594.

152. Wang Z, Gerstein M, Snyder M (2009) RNA-Seq: a revolutionary tool for transcriptomics. Nature

Reviews Genetics 10: 57-63.

153. Nagalakshmi U, Wang Z, Waern K, Shou C, Raha D, Gerstein M, Snyder M (2008) The

transcriptional landscape of the yeast genome defined by RNA sequencing. Science 320: 1344-1349.

154. Dötsch A, Eckweiler D, Schniederjans M, Zimmermann A, Jensen V, Scharfe M, Geffers R,

Häussler S (2012) The Pseudomonas aeruginosa transcriptome in planktonic cultures and static

biofilms using RNA sequencing. PloS one 7: e31092.

155. Frias-Lopez J, Duran-Pinedo A (2012) Effect of periodontal pathogens on the metatranscriptome of

a healthy multispecies biofilm model. Journal of bacteriology 194: 2082-2095.

156. Yoder-Himes D, Chain P, Zhu Y, Wurtzel O, Rubin E, Tiedje JM, Sorek R (2009) Mapping the

Burkholderia cenocepacia niche response via high-throughput sequencing. Proceedings of the

National Academy of Sciences 106: 3976-3981.

157. Sittka A, Lucchini S, Papenfort K, Sharma CM, Rolle K, Binnewies TT, Hinton JC, Vogel J (2008)

Deep sequencing analysis of small noncoding RNA and mRNA targets of the global post-

transcriptional regulator, Hfq. PLoS genetics 4: e1000163.

158. Sharma CM, Hoffmann S, Darfeuille F, Reignier J, Findeiß S, Sittka A, Chabas S, Reiche K,

Hackermüller J, Reinhardt R (2010) The primary transcriptome of the major human pathogen

Helicobacter pylori. Nature 464: 250-255.

159. Wilmes P, Bond PL (2004) The application of two‐dimensional polyacrylamide gel electrophoresis

and downstream analyses to a mixed community of prokaryotic microorganisms. Environmental

Microbiology 6: 911-920.

160. Azevedo NF, Lopes SP, Keevil CW, Pereira MO, Vieira MJ (2009) Time to “go large” on biofilm

research: advantages of an omics approach. Biotechnology letters 31: 477-485.

161. Posch T, Jezbera J, Vrba J, Šimek K, Pernthaler J, Andreatta S, Sonntag B (2001) Size selective

feeding in Cyclidium glaucoma (Ciliophora, Scuticociliatida) and its effects on bacterial community

structure: a study from a continuous cultivation system. Microbial ecology 42: 217-227.

162. Monger BC, Landry MR, Brown SL (1999) Feeding selection of heterotrophic marine

nanoflagellates based on the surface hydrophobicity of their picoplankton prey. Limnology and

oceanography 44: 1917-1927.

163. Matz C, Jürgens K (2005) High motility reduces grazing mortality of planktonic bacteria. Applied

and environmental microbiology 71: 921-929.

164. Hahn M, Moore E, Höfle M (2000) Role of microcolony formation in the protistan grazing defense

of the aquatic bacterium Pseudomonas sp. MWH1. Microbial ecology 39: 175-185.

165. Shannon SP, Chrzanowski TH, Grover JP (2007) Prey food quality affects flagellate ingestion rates.

Microbial ecology 53: 66-73.

166. Pedersen AL, Nybroe O, Winding A, Ekelund F, Bjørnlund L (2009) Bacterial feeders, the nematode

Caenorhabditis elegans and the flagellate Cercomonas longicauda, have different effects on

outcome of competition among the Pseudomonas biocontrol strains CHA0 and DSS73. Microbial

ecology 57: 501-509.

167. Jousset A, Lara E, Wall LG, Valverde C (2006) Secondary metabolites help biocontrol strain

Pseudomonas fluorescens CHA0 to escape protozoan grazing. Applied and environmental

microbiology 72: 7083-7090.

46

168. Matz C, Deines P, Boenigk J, Arndt H, Eberl L, Kjelleberg S, Jürgens K (2004) Impact of violacein-

producing bacteria on survival and feeding of bacterivorous nanoflagellates. Applied and

environmental microbiology 70: 1593-1599.

169. Hahn MW, Höfle MG (2001) Grazing of protozoa and its effect on populations of aquatic bacteria.

FEMS Microbiology Ecology 35: 113-121.

170. Jürgens K, Güde H (1994) The potential importance of grazing-resistant bacteria in planktonic

systems. Marine ecology progress series Oldendorf 112: 169-188.

171. Rønn R, McCaig AE, Griffiths BS, Prosser JI (2002) Impact of protozoan grazing on bacterial

community structure in soil microcosms. Applied and Environmental Microbiology 68: 6094-6105.

172. Iriberri J, Azua I, Labirua‐Iturburu A, Artolozaga I, Barcina I (1994) Differential elimination of

enteric bacteria by protists in a freshwater system. Journal of Applied Microbiology 77: 476-483.

173. Weekers PH, Bodelier PL, Wijen JP, Vogels GD (1993) Effects of grazing by the free-living soil

amoebae Acanthamoeba castellanii, Acanthamoeba polyphaga, and Hartmannella vermiformis on

various bacteria. Applied and environmental microbiology 59: 2317-2319.

174. Singh B (1941) Selectivity in bacterial food by soil amoebae in pure mixed culture and in sterilized

soil. Annals of Applied Biology 28: 52-64.

175. Groscop JA, Brent MM (1964) The effects of selected strains of pigmented microorganisms on small

free-living amoebae. Canadian journal of microbiology 10: 579-584.

176. Gurijala KR, Alexander M (1990) Effect of growth rate and hydrophobicity on bacteria surviving

protozoan grazing. Applied and environmental microbiology 56: 1631-1635.

177. Sinclair JL, Alexander M (1989) Effect of protozoan predation on relative abundance of fast-and

slow-growing bacteria. Canadian Journal of microbiology 35: 578-582.

178. Petropoulos P, Gilbride KA (2005) Nitrification in activated sludge batch reactors is linked to

protozoan grazing of the bacterial population. Canadian journal of microbiology 51: 791-799.

179. Šimek K, Pernthaler J, Weinbauer MG, Horn k K, Dolan JR, Nedoma J, Ma n M, Amann R (2001)

Changes in bacterial community composition and dynamics and viral mortality rates associated with

enhanced flagellate grazing in a mesoeutrophic reservoir. Applied and environmental microbiology

67: 2723-2733.

180. Fenchel T (1987) Ecology of protozoa: the biology of free-living phagotrophic protists. . In: Brock

TD (ed) Brock/Springer series in contemporary bioscience. Science and Technology, Madison, WI.

181. Erken M, Farrenschon N, Speckmann S, Arndt H, Weitere M (2012) Quantification of individual

flagellate-bacteria interactions within semi-natural biofilms. Protist 163: 632-642.

182. Weitere M, Bergfeld T, Rice SA, Matz C, Kjelleberg S (2005) Grazing resistance of Pseudomonas

aeruginosa biofilms depends on type of protective mechanism, developmental stage and protozoan

feeding mode. Environmental microbiology 7: 1593-1601.

183. Arndt H, Schmidt-Denter K, Auer B, Weitere M (2003) Protozoans and biofilms. In Fossil and

Recent Biofilms. Krumbein, W.E., Paterson, D.M., and Zavarzin, G.A., (eds). Dordrecht: Kluwer

Academic Publishers, in press.

184. Joubert L-M, Wolfaardt G, Botha A (2006) Microbial exopolymers link predator and prey in a model

yeast biofilm system. Microbial ecology 52: 187-197.

185. Friman V-P, Diggle SP, Buckling A (2013) Protist predation can favour cooperation within bacterial

species. Biology letters 9: 20130548.

186. Sun S, Kjelleberg S, McDougald D (2013) Relative contributions of Vibrio polysaccharide and

quorum sensing to the resistance of Vibrio cholerae to predation by heterotrophic protists. PloS one

8: e56338.

187. Wey JK, Scherwass A, Norf H, Arndt H, Weitere M (2008) Effects of protozoan grazing within river

biofilms under semi-natural conditions. Aquatic Microbial Ecology 52: 283.

188. Rychert K, Neu TR (2010) Protozoan impact on bacterial biofilm formation. Biological Letters 47:

3-10.

189. Cole JJ (1999) Aquatic microbiology for ecosystem scientists: new and recycled paradigms in

ecological microbiology. Ecosystems 2: 215-225.

47

190. Pernthaler J, Posch T, Simek K, Vrba J, Amann R, Psenner R (1997) Contrasting bacterial strategies

to coexist with a flagellate predator in an experimental microbial assemblage. Applied and

Environmental Microbiology 63: 596-601.

191. Sieburth J, Davis P (1982) The role of heterotrophic nanoplankton in the grazing and nurturing of

planktonic bacteria in the Sargasso and Caribbean Seas. Ann Inst Oceanogr 58: 285-296.

192. Sherr BF, Sherr EB, Andrew TL, Fallon RD, Newell SY (1986) Trophic interactions between

heterotrophic protozoa and bacterioplankton in estuarine water analyzed with selective metabolic

inhibitors. Marine Ecology-Progress Series 32: 169-179.

193. Matz C, Webb JS, Schupp PJ, Phang SY, Penesyan A, Egan S, Steinberg P, Kjelleberg S (2008)

Marine biofilm bacteria evade eukaryotic predation by targeted chemical defense. PloS one 3: e2744.

194. Zubkov MV, Sleigh MA (2000) Comparison of growth efficiencies of protozoa growing on bacteria

deposited on surfaces and in suspension. Journal of Eukaryotic Microbiology 47: 62-69.

195. Jackson S, Jones E (1991) Interactions within biofilms: the disruption of biofilm structure by

protozoa. Kieler Meeresforsch 8: 264-268.

196. Pedersen K (1982) Factors regulating microbial biofilm development in a system with slowly

flowing seawater. Applied and environmental microbiology 44: 1196-1204.

197. Wolfaardt G, Lawrence J, Headley J, Robarts R, Caldwell D (1994) Microbial exopolymers provide

a mechanism for bioaccumulation of contaminants. Microbial ecology 27: 279-291.

198. Huws S, McBain A, Gilbert P (2005) Protozoan grazing and its impact upon population dynamics in

biofilm communities. Journal of applied microbiology 98: 238-244.

199. Dopheide A, Lear G, Stott R, Lewis G (2011) Preferential feeding by the ciliates Chilodonella and

Tetrahymena spp. and effects of these protozoa on bacterial biofilm structure and composition.

Applied and environmental microbiology 77: 4564-4572.

200. Brown MR, Barker J (1999) Unexplored reservoirs of pathogenic bacteria: protozoa and biofilms.

Trends in microbiology 7: 46-50.

201. Cirillo JD, Falkow S, Tompkins LS (1994) Growth of Legionella pneumophila in Acanthamoeba

castellanii enhances invasion. Infection and immunity 62: 3254-3261.

202. Cirillo JD, Falkow S, Tompkins LS, Bermudez LE (1997) Interaction of Mycobacterium avium with

environmental amoebae enhances virulence. Infection and Immunity 65: 3759-3767.

203. Barker J, Scaife H, Brown M (1995) Intraphagocytic growth induces an antibiotic-resistant

phenotype of Legionella pneumophila. Antimicrobial agents and chemotherapy 39: 2684-2688.

204. Singh PK, Schaefer AL, Parsek MR, Moninger TO, Welsh MJ, Greenberg E (2000) Quorum-sensing

signals indicate that cystic fibrosis lungs are infected with bacterial biofilms. Nature 407: 762-764.

205. Høiby N, Krogh Johansen H, Moser C, Song Z, Ciofu O, Kharazmi A (2001) Pseudomonas

aeruginosa and the in vitro and in vivo biofilm mode of growth. Microbes and Infection 3: 23-35.

206. Levin BR, Antia R (2001) Why we don't get sick: the within-host population dynamics of bacterial

infections. Science 292: 1112-1115.

207. Rittmann BE (2007) Where are we with biofilms now? Where are we going? Water science and

technology 55: 1-8.

208. Washio J, Mayanagi G, Takahashi N (2010) Challenge to Metabolomics of Oral Biofilm:—From

“What Are They?” to “What Are They Doing?”—. Journal of Oral Biosciences 52: 225-232.

209. Nicholson JK, Lindon JC (2008) Systems biology: metabonomics. Nature 455: 1054-1056.

210. Garcia DE, Baidoo EE, Benke PI, Pingitore F, Tang YJ, Villa S, Keasling JD (2008) Separation and

mass spectrometry in microbial metabolomics. Current opinion in microbiology 11: 233-239.

211. Hollywood K, Brison DR, Goodacre R (2006) Metabolomics: current technologies and future trends.

Proteomics 6: 4716-4723.

212. Wilmes P, Bowen BP, Thomas BC, Mueller RS, Denef VJ, VerBerkmoes NC, Hettich RL, Northen

TR, Banfield JF (2010) Metabolome-proteome differentiation coupled to microbial divergence.

MBio 1.

213. Tyson GW, Chapman J, Hugenholtz P, Allen EE, Ram RJ, Richardson PM, Solovyev VV, Rubin

EM, Rokhsar DS, Banfield JF (2004) Community structure and metabolism through reconstruction

of microbial genomes from the environment. Nature 428: 37-43.

48

214. Raes J, Bork P (2008) Molecular eco-systems biology: towards an understanding of community

function. Nature Reviews Microbiology 6: 693-699.

215. Ohashi Y, Hirayama A, Ishikawa T, Nakamura S, Shimizu K, Ueno Y, Tomita M, Soga T (2008)

Depiction of metabolome changes in histidine-starved Escherichia coli by CE-TOFMS. Molecular

BioSystems 4: 135-147.

216. Jansson JK, Prosser JI (2013) Microbiology: The life beneath our feet. Nature 494: 40-41.

217. Rice KC, Bayles KW (2003) Death's toolbox: examining the molecular components of bacterial

programmed cell death. Molecular microbiology 50: 729-738.

218. Watnick P, Kolter R (2000) Biofilm, city of microbes. Journal of Bacteriology 182: 2675-2679.

219. Nikolaev YA, Plakunov V (2007) Biofilm—“City of microbes” or an analogue of multicellular

organisms? Microbiology 76: 125-138.

220. Battin TJ, Sloan WT, Kjelleberg S, Daims H, Head IM, Curtis TP, Eberl L (2007) Microbial

landscapes: new paths to biofilm research. Nature Reviews Microbiology 5: 76-81.

49

50

6 Manuscripts

Manuscript 1

High-Throughput Screening of Multispecies Biofilm Formation and Quantitative

PCR-Based Assessment of Individual Species Proportions, Useful for Exploring

Interspecific Bacterial Interactions.

51

METHODS

High-Throughput Screening of Multispecies BiofilmFormation and Quantitative PCR-Based Assessmentof Individual Species Proportions, Useful for ExploringInterspecific Bacterial Interactions

Dawei Ren & Jonas Stenløkke Madsen &

Claudia I de la Cruz-Perera & Lasse Bergmark &

Søren J. Sørensen & Mette Burmølle

Received: 26 April 2013 /Accepted: 14 October 2013# Springer Science+Business Media New York 2013

Abstract Multispecies biofilms are predominant in almost allnatural environments, where myriads of resident microorgan-isms interact with each other in both synergistic and antago-nistic manners. The interspecies interactions among differentbacteria are, despite the ubiquity of these communities, stillpoorly understood. Here, we report a rapid, reproducible andsensitive approach for quantitative screening of biofilm for-mation by bacteria when cultivated as mono- and multispeciesbiofilms, based on the Nunc-TSP lid system and crystal violetstaining. The relative proportion of the individual species in afour-species biofilm was assessed using quantitative PCRbased on SYBR Green I fluorescence with specific primers.The results indicated strong synergistic interactions in a four-species biofilm model community with a more than 3-foldincrease in biofilm formation and demonstrated the strongdominance of two strains, Xanthomonas retroflexus andPaenibacillus amylolyticus . The developed approach can beused as a standard procedure for evaluating interspecies inter-actions in defined microbial communities. This will be ofsignificant value in the quantitative study of the microbialcomposition of multispecies biofilms both in natural

environments and infectious diseases to increase our under-standing of the mechanisms that underlie cooperation, com-petition and fitness of individual species in mixed-speciesbiofilms.

Introduction

Biofilms are defined as polymeric matrix-enclosed bacterialcommunities associated with surfaces or interfaces [1]. Theyare considered the dominant lifestyle of bacteria both in envi-ronmental ecosystems and human hosts, and typically com-prise a large number of different bacteria living together [2].Soil is an example of an environment that contains a largenumber of very diverse bacteria and numerous available sur-faces for multispecies biofilm formation [3].

Coresidence of diverse bacteria in multispecies biofilms islikely to catalyse complex interactions, resulting in increased ordecreased biofilm biomass [4–7], which may in turn affect theoverall function of the biofilm community. As example, Pseu-domonas aeruginosa PAO1 and Burkholderia sp . NK8 showedenhanced biofilm formation in a dual-species biofilm, directlybenefitting bioremediation potential, as chlorobenzoates weremore efficiently degraded [8] and similar biofilm synergies wereobserved in drinking water systems [7]. Despite an increase inbiofilm-related studies over the past decades, most of these havefocused on monospecies biofilms or specific ecological nichessuch as mixed-species oral biofilms. Our current knowledgeregarding the prevalence, physiology and complexity of multi-species biofilm is still incomplete, partly due to the lack ofreproducible screening methods.

Two measurements are of particular importance when ex-ploring interactions in biofilms; namely, the biomass (or

Electronic supplementary material The online version of this article(doi:10.1007/s00248-013-0315-z) contains supplementary material,which is available to authorized users.

D. Ren : J. S. Madsen : C. I. de la Cruz-Perera : L. Bergmark :S. J. Sørensen :M. BurmølleSection of Microbiology, Department of Biology,University of Copenhagen, Copenhagen, Denmark

M. Burmølle (*)Section of Microbiology, University of Copenhagen,Universitetsparken 15, bygn 1, 2100 Copenhagen Ø, Denmarke-mail: burmolle@bio.ku.dk

Microb EcolDOI 10.1007/s00248-013-0315-z

productivity) of the total biofilm and of the individual strains.Measurements of the overall biomass can, when comparedwith the amount that each of the residing species are able toproduce as monospecies biofilms, be used to distinguish theeffect of interactions on the extent of biofilm that is formed, asto whether this is synergistic, neutral or antagonistic. Beingable to differentiate the success of each different bacterialmember of the biofilm can furthermore resolve interactionsas being mutual, commensal or parasitic, which is fundamen-tal for exploring and understanding the selective forces oper-ating in and shaping multispecies communities.

Several methods and protocols have been developed forstudying biofilms. Flow cells combined with confocal scan-ning laser microscopy [9] is the most favoured tool and has theadvantage of enabling one to obtain quantitative informationon both the overall biomass and of the individual strains ifcombined with correct labelling techniques (e.g. fluorescencein situ hybridization, FISH) but is unfortunately not suited forhigh-throughput comparative screening studies. Furthermore,identification of specific strains in multispecies biofilms bythis approach is highly labour-intensive and requires experthandling in order to avoid pitfalls such as uneven staining dueto the limited probe penetration into biofilms, artefacts causedby hybridization and dehydration procedures in FISH [10, 11].Microtiter plates are suitable for performing biofilm quantifi-cation, usually based on crystal violet retention, owing to theirhigh-throughput screening capability and the simplicity ofprotocols [12]. However, the reproducibility of this assay isproblematic, especially when multispecies consortia areanalysed. In addition, molecular analysis of biofilm cellsrequires efficient and complete disruption of the cells followedby extraction of the target molecules, which still needs to beimproved in multispecies biofilms.

In this study, we report an easily applicable and reproduc-ible approach for consistently quantifying multispecies bio-film formation and to evaluate interactions, in regard to theoverall biofilm formation and relative proportions of individ-ual species.We present an optimized DNA extraction protocoland SYBR Green-based quantitative PCR (qPCR) assay forthe selective, rapid and sensitive detection of four species inmultispecies biofilm. The reproducibility and broad applica-bility of this specific detection procedure makes this methoduseful for most types of defined biofilms, not limited to thesoil isolates used in this study.

Materials and Methods

Soil Isolates and Culture Conditions

The bacterial strains used in this study (Table 1) were obtainedfrom agricultural soil as described previously [13]. From thetotal strain pool isolated by de la Cruz-Perera et al. [13], we

selected seven strains based on growth compatibility (seebelow). Two of the selected strains (6 and 7) were not de-scribed by de la Cruz-Perera et al., but these were isolated andidentified by procedures identical to those referred to above.

Optimization of Growth Media

To determine the optimal growth conditions and evaluate thebiofilm-forming capabilities of the seven selected soil isolates,each strain was grown individually inMinimal Medium (basalmedium 500 mL: NaH2PO4·H2O 0.5 g, K2HPO4·3H2O2.125 g, NH4Cl 1.0 g, pH 7.2; trace metals 500 mL:nitrilotriacetic acid 0.0615 g, MgSO4·7H2O 0.1 g, FeSO4·7H2O 0.006 g, ZnSO4 ·7H2O 0.0015 g, MnSO4 ·H2O0.0015 g, pH 7.0) supplemented with 0.2 % D(+)-Glucose,nutrient-low R2B medium (yeast extract 0.5 g, proteose pep-tone 0.5 g, casamino acids 0.5 g, glucose 0.5 g, soluble starch0.5 g, sodium pyruvate 0.3 g, K2HPO4 0.3 g, MgSO4·7H2O0.05 g, in 1 L distilled water) and nutrient-rich TSB medium(tryptic soy broth, Merck KgaA, Germany). When solidmedium was needed, 1.5 % (wt/vol) agar was added.

The strains were subcultured from frozen glycerol stocksonto tryptic soy agar (TSA) plates for 48 h at 24 °C, and thencolonies were transferred onto Minimal Medium agar platescontaining 0.2 % glucose, R2B agar plates or TSA plates.Colonies from solid type media were inoculated into 5 mLliquid media of the same type and incubated with shaking(250 rpm) at 24 °C overnight.

Cultivation of Biofilms

Both 96-well cell culture plates (cat. no. 655 180, Greiner Bio-One, Germany) and Nunc-TSP lid system (cat. no. 445497,Thermo Scientific, Denmark), which comprises a 96-wellplate lid with pegs that extend into each well, were used tocultivate biofilms. The seven selected strains were screened

Table 1 Identification of the seven soil isolates by 16S rRNA analysis

Strain no.a GenBank accession no. Closest relativeb

1 JQ890536 Pseudomonas lutea

2 JQ890538 Stenotrophomonas rhizophila

3 JQ890537 Xanthomonas retroflexus

4 JQ890542 Ochrobactrum rhizosphaerae

5 JQ890539 Microbacterium oxydans

6 JQ890541 Arthrobacter nitroguajacolicus

7 JQ890540 Paenibacillus amylolyticus

a The numbers 1 to 7 were designated to the seven strains forsimplificationb The sequences had 98 to 100 % base identity to the closest relative inGenBank

D. Ren et al.

for biofilm formation as single species and in four-speciescombinations as described below.

The colonies (grown on agar plates for 24 h) were inocu-lated into 5 mL of TSB medium and incubated overnight at24 °C with shaking (250 rpm). One hundred to 600 μL ofthese stationary phase bacterial cultures were transferred tofresh TSB medium and grown until an optical density at600 nm (OD600) of ~1.0 was reached. The cell suspensionswere then adjusted to an OD600 of 0.15 by dilution in TSBmedium. A total of 160 μL of monospecies or four mixedspecies (40 μL of each species) exponentially growing cul-tures were added to each well. To some wells, the samevolume of fresh TSB medium was added to obtain a back-ground value, which was subtracted from values obtainedfrom the wells containing cells. The plates were sealed withParafilm and incubated with shaking (200 rpm) at 24 °C for24 h. The biofilm assays were performed three times ondifferent days (biological replicates) with four technical repli-cates every time.

Quantification of Biofilm Formation by Use of CVand TTC

A modified version of the crystal violet (CV) method fordetection of biofilms using 96-well cell culture plate [12]was applied as previously described. Additionally, the assaywas further modified for quantifying biofilms formed on pegsof the Nunc-TSP lid system, previously referred to as theCalgary method [14]. After 24 h incubation, in order to washoff planktonic cells, the peg lid was transferred successively tothree microtiter plates containing 200 μL phosphate bufferedsaline per well, followed by staining of the biofilms formed onthe pegs with 180 μL of an aqueous 1 % (w /v ) CV solution.After 20 min of staining, the lid was rinsed again three timesand then placed into a new microtiter plate with 200 μL of96 % ethanol in each well. After allowing 30 min for the stainto dissolve into the ethanol, the absorbance of CV at 590 nmwas determined by using an EL 340 BioKinetics reader(BioTek Instruments,Winooski, Vt.). The CV-ethanol suspen-sion was diluted with 96 % ethanol when the OD590 wasabove 1.1.

Parallel with the CV-based biofilm assays described above,an alternative method, based on 2, 3, 5-triphenyltetreazoliumchloride (TTC) reduction, was also implemented and evaluat-ed. After 24-h incubation, the pegs with attached cells wererinsed three times as described above. Thereafter, the peg lidwas mounted on a new microtiter plate with 200 μL freshmedium containing 0.01 % TTC. The plate was then sealedwith Parafilm, wrapped in foil and incubated with shaking(200 rpm) for another 24 h. In order to resuspend the formedformazan, the peg lid was transferred to a new microtiter platecontaining 200 μL of 96 % ethanol per well. Finally, theabsorbance was measured at 490 nm.

The statistical analyses were conducted using ANOVA test(SPSS version 17.0 for Windows). A p value of ≤0.05 wasregarded as a statistically significant difference.

DNA Extraction from Four-Species Biofilms

The 24-h biofilms formed on peg lids were rinsed three timeswith PBS to remove non-adherent bacteria as described above.Then pegs were broken from the lid from the back (i.e. withoutdirect contact to the biofilm-covered part of the peg) usingsterile forceps and transferred into Lysing Matrix E tubes(provided by the FastDNA™ SPIN Kit for soil). Each pegwas placed in one tube. Aliquots of 882 µl of sodium phosphatebuffer was added to each tube and biofilms were disruptedfrom pegs by bead beating using the Savant FastPrepFP120 for 40s at setting 6.0. The pegs were removed andstained with crystal violet to verify that after bead beating,all the cells were detached from the pegs. Ninety-eightmicrolitres of lysozyme solution (20 mg/ml), dissolved insodium phosphate buffer was added, and the samples wereincubated for 1 h at 37 °C. Next, 122 μL of MT buffer wasadded to each sample, followed by bead beating twice for40s at setting 6.0 with cooling on ice during the short timeinterval. The complete cell lysis was visually confirmed withan optical microscope. Genomic DNA was then extractedusing FastDNA™ SPIN Kit for soil (Qbiogene, Illkirch,France) according to the manufacturer's instructions and quan-tified with a Qubit fluorometer (Invitrogen, Carlsbad, CA,USA).

DNA Sequence Analysis and Design of the Species-SpecificPrimers

Multiple alignment of sequences from the four strains 2, 3, 5and 7 was done using the DNAMAN software (version 7,Lynnon corporation). The species-specific primer pairs forSYBR Green qPCR were designed manually based on thevariable regions of 16S rRNA genes (Table 2) according to theguidelines set by Primer Express (version 3.0) from AppliedBiosystems with an approximate maximum amplicon size of300 bp. The obtained best fitting primers were checked byDNAMAN to avoid hairpins or primer–dimer formations. Thespecificity of each pair of primers was checked against theother three, non-target strains using conventional PCRs asfollows. Reaction mixtures of 25 μL contained 16 μL H2O,1 μL genomic DNA, 5 μL 5× Phusion HF Buffer, 0.5 μL10 mM dNTPs, 1 μL 10 μM of each primer, and 0.5 μLPhusion DNA polymerase (Phusion high-fidelity PCR kit;Finnzymes, Espoo, Finland). Amplifications were performedwith the following cycling protocol: 5 min at 95 °C, followedby 30 cycles of 30s at 94 °C, 30s at 61 °C/62 °C, 20s at 72 °Cand a final elongation step of 5 min at 72 °C in DNA EngineDyad Peltier Thermal Cycler (MJ Research Inc.). The

High-Throughput Screening and Species Abundance Analysis in Multispecies Biofilms

amplified products were separated by 1.0 % agarose gelelectrophoresis, stained with ethidium bromide andphotographed under UV illumination.

Plasmid Standards Used for Absolute Quantification

A plasmid standard containing the target region was preparedfor each primer pair as follows. The 16S rRNA gene fragmentswere amplified by conventional PCR using the correspondingprimers as mentioned above. The products were cloned usingthe TOPO TA cloning kit (Invitrogen, Carlsbad, CA, USA).Plasmids were isolated with QIAprep Spin Miniprep Kit(Qiagen Gmbh, Hilden, Germany). The qualities were evaluat-ed by agarose gel electrophoresis and concentrations weremeasured by Qubit Fluorometer (Invitrogen, Carlsbad, CA,USA). 16S rRNA gene copy numbers were calculated assum-ing that the average molecular mass of a double-stranded DNAmolecule is 660 g/mol. Four standard curves were generatedusing triplicate 5-fold dilutions of plasmid DNAs, and thecorresponding slope was used to calculate PCR amplificationefficiency (E) according to the equation of E =10(−1 slope). 16SrRNA gene copy numbers in unknown samples were thendetermined by interpolation from the standard curve using theirrespective threshold cycle (Ct) values. The Ct value representsthe number of PCR amplification cycles needed to producefluorescence intensity above a pre-defined threshold.

Quantitative PCR Based on SYBR Green I Fluorescence

For each sample, four separate qPCR reactions were per-formed with the Mx3000PTM Real-time PCR System (Strata-gene, USA) using each pair of species-specific primers. Eachof the reaction components per 20 μL were 7 μL H2O, 10 μL2× Brilliant III Ultra-Fast SYBR® Green QPCR Master Mix(Agilent Technologies), 1 μL 10 μM of each primer and 1 μLgenomic DNA (either standard or sample). The PCRprogrammes were carried out as follows: 95 °C for 10 min,40 cycles of 95 °C for 30s, 59–62 °C for 30s and 72 °C for

20s/30s, followed by a standard melting/dissociation curvesegment. Each sample was run in triplicates and a negativecontrol was included in every run.

Results

Isolates and Media Optimization

The seven selected strains, representing common soil bacteria,were all able to grow on Minimal Medium supplemented with0.2 % D(+)-Glucose (data not shown). These strains wereselected for further media optimization and multispecies bio-film formation. The strain identities and the accession numbersof the 16S rRNA gene sequences, used in this study for design-ing specific primers (strain 2, 3, 5 and 7), are shown in Table 1.

All seven strains showed slower growth on Minimal Me-dium+Glucose (0.2 %) and R2A plates than on TSA platesand took much more time to reach exponential growth phasein liquid media at 24 °C (data not shown). Since all sevenstrains grew well on TSA/TSB, this medium was chosen forfurther cultivation of mono- and four-species biofilms.

Assay for Biofilm Formation

Monospecies biofilm formation was assayed both with andwithout the peg lid system using both the CVretention and theTTC reduction assays (Fig. 1a–c). To evaluate the reproduc-ibility within technical replicates and between individual ex-perimental series (biological replicates), all biofilm assayswere performed in four replicates and conducted three indi-vidual times on different days under the same conditions.

Compared with the 96-well cell culture plate, the Nunc-TSP lid system gave less variation between replicates andbetween individual days, which were directly reflected in thebiofilm formation of strains 1, 2 and 4. These three strains,which were identified as poor biofilm formers in the Nunc-TSP lid system (Fig. 1b), showed significantly different

Table 2 Species-specific primers based on 16S rRNA gene sequences

Strain no. Identity Primer position Sequence (5′–3′) Length (bp) Product size (bp)

2 S . rhizophila FP 159 GCCTTGCGCGGATAGATG 18 240

RP 399 CGGGTATTAGCCGACTGCTT 20

3 X . retroflexus FP 135 GCCTTGCGCGATTGAATG 18 252

RP 387 CCGTCATCCCAACCAGGTATT 21

5 M . oxydans FP 935 TCAACTCTTTGGACACTCGTAAACA 25 213

RP 1148 CATGCGTGAAGCCCAAGAC 19

7 P. amylolyticus FP 800 GATACCCTTGGTGCCGAAGTT 21 145

RP 945 CGGTCAGAGGGATGTCAAGAC 21

The numbers in the primer position show the positions of the target sites of each species

FP Forward primer, RP Reverse primer

D. Ren et al.

abilities for biofilm formation from day to day (P <0.05) whenusing the 96-well cell culture plate (Fig. 1a). Another differ-ence between the assays conduced in the two types of micro-titer plates was the amount of biofilm formed by strain 6,which displayed biofilm formation in the 96-well cell cultureplate, but was incapable of attaching to the pegs. The Nunc-TSP lid system could not only provide more reproducibleresults, but it also removed the possibility that aggregationmay be linked to sedimentation of the microorganisms in testwells [15], which might have been the case for strain 6(subpanels a vs. b of Fig. 1). Based on these characteristics,the Nunc-TSP lid system was chosen as a better device forhigh-throughput screening for synergistic interactions in mul-tispecies biofilm formation.

In addition to the CV method, which quantifies total at-tached material, we also used the TTC method, which evalu-ates cell activity. Respiring cells reduce the colourless TTCsolution to the red insoluble formazan, which can be dissolvedin 96 % ethanol and measured spectrophotometrically. Thedata shown in Fig. 1c represents the TTC absorbance at490 nm. Low day-to-day variation was observed, but thelow output signals resulted in less distinct difference betweenbiofilm formers and non-formers. In addition, some absor-bance measurements were close to the resolution limit of theBioreader. Therefore, the TTC reduction assay was not ap-plied for the following mixed-species biofilm formation.

The results obtained by the Nunc-TSP lid system showedthat only strain 3 was able to establish biofilm on its own; CVvalues obtained for this strain were more than 10-fold higherthan for any of the other strains (Fig. 1b). Based on theseresults, strain 3 was identified as being a good biofilm former,whereas the remaining six strains were characterized as poorbiofilm formers.

Biofilm Formation by Single-Species and Four-SpeciesConsortia

The seven selected strains were screened for biofilm forma-tion, by using the Nunc-TSP lid system, as single species andin all possible combinations of four, in order to identify a four-species model consortium, suitable for studies of interspeciesinteractions. As shown in Fig. 2, the combination of strains 2,3, 5 and 7 gave no significant variation between replicates and

between individual days, indicating the high reproducibility ofthe assay. Obviously, with strain 3 as exception, the strainsshowed weak ability to form biofilm. However, when the fourisolates coexisted in the biofilm, the biofilm biomass in-creased by >300 % compared to the single-species biofilms,

(a)

Isolates

1 2 3 4 5 6 7

Bio

film

form

atio

n (O

D 5

90)

0

1

2

3

4

5

day1 day2 day3

(b)

Isolates1 2 3 4 5 6 7

Bio

film

form

atio

n (O

D 5

90)

0.2

3.0

5.0

7.0

0.0

0.4

4.0

6.0 day1 day2 day3

(c)

Isolates

1 2 3 4 5 6 7

Bio

film

form

atio

n (O

D 5

90)

0.00

0.05

0.10

0.15

0.20

0.25

day1 day2 day3

�Fig. 1 Biofilm formation in 96-well cell culture plate (a) and Nunc-TSPlid system (b , c) by the seven isolates. After 24 h of incubation, thebiofilm formation was quantified by staining with crystal violet followedby absorbance measurements at 590 nm (a , b) or the presence of reducedTTC was quantified by measuring the absorbance at 490 (c). Threeparallel bars represent means ± standard error for four replicates onthree different days representing three biological replicates (day 1, day2 and day 3). 1 , Pseudomonas lutea; 2 , Stenotrophomonas rhizophila ; 3 ,X . retroflexus ; 4 , Ochrobactrum rhizosphaerae ; 5 , Microbacteriumoxydans; 6 , Arthrobacter nitroguajacolicus ; 7 , P. amylolyticus

High-Throughput Screening and Species Abundance Analysis in Multispecies Biofilms

indicating a strong synergy in this multispecies biofilm. Thisfour-species consortium was therefore selected as a multispe-cies biofilm model for further studies of the species dynamics.

Quantification of Four Strains by Quantitative PCRin Multispecies Biofilm

DNAwas successfully extracted from the biofilms by excisionof entire pegs followed by cell disruption. The biofilms werecompletely removed from the pegs (confirmed by CV stain-ing, data not shown), and complete cell lysis was confirmedby microscopy. The amount of DNA extracted frommultispe-cies biofilms composed of strains 2, 3, 5 and 7 was 1.9–2.1 μg/peg.

Based on the variable regions of 16S rRNA genes, fourpairs of specific primers were designed for SYBR GreenqPCR (Table 2). For simplification and to enable quantifica-tion based on SYBR Green, the primers were designed forapplication in four separate reactions for each sample. Thespecificity of the primers was confirmed by conventionalPCRs, verifying that primers were strictly species specific.The high linearity of the Ct values plotted in the standardcurves was verified by the correlation coefficient (RSq) valuesof >0.99. The amplification efficiencies (E) ranged from 80 to90 % (Supplementary Fig. 1). No detectable peaks that wereassociated with primer–dimer or other non-specific PCR prod-ucts were observed in the melting curves (data not shown),and only the single bands of the expected size amplicon ineach qPCR assay were detectable by agarose gel electropho-resis (data not shown). The standard plasmid DNA used for

the standard curves ranged from 3×106 to 3×107 copies/μLand were used in 5-fold dilution series. The DNA extractedfrom biofilms was diluted appropriately to ensure that all theunknown samples were within the range of the standardcurves. 16S rRNA gene copy number of each species wascalculated and is shown in Table 3. As the exact copy numbersof the 16S rRNA gene per cell in the four species are currentlyunknown, the cell numbers were estimated based on otherspecies in the same genus [16–19].

Discussion

Multispecies biofilms are prevalent in almost all environ-ments. A pressing need, therefore, exists to better the under-standing of the social interactions and selective forces thatdrive bacterial communities in multispecies biofilm. Duringthe past few decades, simultaneous staining of multiple spe-cies by FISH has been widely used in combination withconfocal laser scanning microscopy (CLSM) for species dif-ferentiation in oral biofilms [20–22]. This approach is, how-ever, not easily applicable to many different isolates fromvarious environments and it is only partly quantitative. In thisstudy, we have developed an approach to consistently quantifybiofilm formation, which enables high-throughput screeningof the prevalence of synergistic interactions and assessment ofthe proportions of individual bacterial species in biofilms. Thegeneral procedure of this approach is illustrated in Fig. 3which can be used as a standard procedure for evaluating

Isolates2 3 5 7 2357

Bio

film

for

mat

ion

(OD

590

)

0

3

6

9

12

15

18

21

24

day 1day 2day 3

Fig. 2 Biofilms formed by single-species and four-species communities(2 , Stenotrophomonas rhizophila; 3 , X . retroflexus; 5 ,M . oxydans ; and7 , P. amylolyticus) in Nunc-TSP lid system. After 24 h of incubation, thebiofilm formation was quantified by staining with crystal violet. Threeparallel bars represent the average absorbance at OD 590 nm for fourreplicates on three different days representing three biological replicates(day 1, day 2 and day 3). The bars represent means ± standard error forfour replicates

Table 3 The copy numbers of 16S rRNA gene and the estimated cellnumbers derived from four separate qPCRs

Strainno.

Identity 16S rRNAgene copies(per μL)

Estimated copynumber of the16S rRNA gene(per cell)

Estimatednumber ofcells (per μL)

2 S . rhizophila 7.91E+006 4a 1.98E+006

3 X . retroflexus 1.98E+009 2b 9.90E+008

5 M . oxydans 8.89E+006 2c 4.45E+006

7 P. amylolyticus 9.95E+008 12d 8.29E+007

a The copy number was estimated to be 4, as Stenotrophomonasmaltophilia is known to have four copies [17]b The copy number of the 16S rRNA gene in X . retroflexus is estimated tobe 2, as Xanthomonas axonopodis , Xanthomonas campestris ,Xanthomonas oryzae, Xanthomonas citri, Xanthomonas albilineans areknown to have two copies, respectively [18]c The copy number of the 16S rRNA gene in M . oxydans is estimated tobe 2, as Microbacterium testaceum is known to have two copies [19]d The copy number of the 16S rRNA gene in P. amylolyticus is estimatedto be 12, as other Paenibacillus species are known to have an average 12copies [16]

D. Ren et al.

interspecies interaction in multispecies biofilms. The right-hand part of this figure illustrates the parallel strain identifica-tion and primer design. Clearly, this protocol is applicable formultispecies biofilms composed of a defined number ofknown species. Other approaches, including metagenomicanalysis, are more suitable when interactions in more complexcommunities are explored. However, the standard methodpresented in this study could be much useful in some naturalsettings where species diversity is more restricted e.g. chronicinfections or when limited key species are the research focus.

Multispecies Biofilm Assay

Mixed species may facilitate synergistic or antagonistic inter-actions between consortia members in biofilms. The co-culturing of four strains in this study provided an easy wayto detect changes in biofilm formation from mono- to multi-species biofilms and allowed high-throughput screening ofmany isolates with high reproducibility. Biofilms have beenproposed to exist in a fine balance between competition andcooperation [23], which can be tipped by various types ofinfluences such as surrounding environmental factors andquorum sensing-dependent gene regulation. Inconsistent re-sults in biofilm formation assays have previous been reported[24–27] due to different substrates, media, inoculum size andculture conditions, and the inconsistency is most likely to beworsened by the highly heterogeneous nature of multispeciesbiofilms. In this study, the three independent repetitive runs

and four technical replicates per run were performed to assessthe reproducibility of the quantitative analysis.

Poor reproducibility of the microtiter plate assay (reflectedby higher standard errors (Fig. 1a)) can be caused by pipet-ting uncertainties, pieces of biofilm detaching, etc. In theNunc-TSP lid system, the number of pipetting steps is dra-matically reduced, which decreases the handling-inducedvariability. This was best reflected by the lower standarderrors, indicative of relatively uniform biofilms throughoutthe four replicate wells (Fig.1b, c and Fig. 2).

TTC reduction, as a simple colorimetric method, has beenwidely used to evaluate the cell activity in plant tissue [28],fungal spores, yeast and bacteria cultures [29]. The low TTCvalues observed in this study were probably due to therelatively low metabolic activities of cells and the enhancedlevels of non-biological material within mature biofilm com-pared to the CV assay where the extracellular matrix and allbacterial cells are stained. The combined use of CV and therespiratory indicator CTC (5-cyano-2,3-ditolyl tetrazo-lium chloride) in high-throughput biofilm assays was pre-viously reported by Pitts et al., who also found cells that hadlost metabolic activity could still contribute to the totalamount of biomass [30]. Therefore, in this study, TTC wasless suitable for quantitative biofilm measurements of micro-bial biomass.

DNA Extraction and qPCR Analysis of Species Distribution

Comparative studies of gene and protein expression ofbiofilm-associated cells have proven that these are significant-ly different from planktonically grown cells [31, 32]. Theprotocol optimized in the present study successfully detachedand lysed both the Gram-negative and Gram-positive cells andis likely to become useful in many applications of DNA-basedbiofilm research. With the appropriate modification, the pro-tocol is additionally suitable for RNA and protein extractions.

qPCR has been successfully applied for quantifying bacterialabundance in plaque biofilm [33], faucet biofilm [34] andbiofilms in wounds [35] due to its speed, sensitivity and spec-ificity. The targets of qPCR can be species-specific genes or16S ribosome RNA genes [36, 37]. 16S rRNA gene proved tobe an excellent target for both quantitative broad-range PCR[38] and group-/species-specific PCR [39, 40] in complexcommunities. The qPCR assay offers several advantages overFISH: low risk of contamination by amplified products, thesimplicity and rapidity of data analysis and low detection limit.Using the real-time TaqMan assay, Price et al. [41] were able toexamine specific bacterial populations in biofilms grown fromhuman saliva and their susceptibility to chlorhexidine. Ren et al.reported a consistent function-related bacterial distribution inanode biofilms using both FISH and qPCR analyses [42].

In the present study, we describe a specific qPCR assay toexamine the population dynamics in multispecies biofilms.

Strain isolation

Media optimization

Mono- and multi-species biofilm cultivation

Target selection

DNA extraction

Quantitative PCR

DNA extraction

16S rRNA gene sequencing

Species-specific primer design

Fig. 3 General procedure for high-throughput screening and evaluationof interactions in multispecies biofilms. The right-hand part illustratesthe parallel strain identification and primer design for use in qPCR

High-Throughput Screening and Species Abundance Analysis in Multispecies Biofilms

The significant synergistic interaction observed in a biofilmconsisting of four soil bacteria make this consortium a pow-erful model to study development and interactions in multi-species biofilms. For this work, 16S rRNA gene was targetedfor SYBR Green assay profiling of the four strains run inseparate reactions. The hypervariable regions interspersedwith the conserved regions make 16S rRNA gene an attractivetarget for both universal and specific primers. In addition, thepublic databases of 16S rRNA gene sequences are easilyaccessible, including GenBank, Greengenes and RibosomalDatabase Project, which are valuable for bacterial identifica-tion and investigation in microbial ecology and evolution.TaqMan and SYBR Green qPCR are two frequently usedassays. Despite of the significantly high specificity of the detec-tion with TaqMan probe, SYBRGreen qPCR is widely used dueto the low cost and the ease in designing primers and optimizingassays. Maeda et al. [43] have reported that there are no signif-icant differences between the TaqMan and SYBR Green chem-istry in their specificity and sensitivity. However, the effect ofSYBR Green qPCR is also limited in terms of the number ofspecies it is practical to analyse. In such cases, the use of TaqManwould be preferred in a complex community to enable morespecies to be analysed, as it opens up for multiplex qPCR, whichonly requires one specific probe per target (instead of two spe-cific primers) and reduces the amount of samples being run.

Concluding Remarks

Overall, we present a sensitive and reliable high-throughputmethod to investigate the interspecific interactions betweenbacterial isolates with the static co-culture assay followed by aqPCR assay that uses species-specific primers to measure the16S rRNA gene numbers of each species in multispeciesbiofilms. By quantification and comparisons of the biomassof each species when grown alone and in the multispeciesbiofilm, understanding of the interspecific interactions (coop-erative, mutualistic, competitive) that operate in the multispe-cies biofilm, is obtainable. To our knowledge, this presentedapproach applied in screenings for overall synergism andantagonism within multispecies biofilms composed of soilisolates is firstly reported in this study. As the ubiquity ofbiofilm formation is receiving increasingly more attentionamong scientists, this developed method will be a valuabletool for studying the social interactions and selective forcesthat drive complex bacterial communities in natural environ-ments as well as infectious diseases.

Acknowledgments This research was supported by funding to MetteBurmølle by The Danish Council for Independent Research, Technologyand Production, ref no: 09-090701, to Søren Sørensen by The DanishCouncil for Independent Research, Natural Sciences and the Danish Inno-vation Consortium, SiB, ref no: 11804520 to Jonas Stenløkke Madsen.

References

1. Costerton JW, Lewandowski Z, Caldwell DE, Korber DR, Lappin-Scott HM (1995) Microbial biofilms. Annual Reviews inMicrobiology 49(1):711–745

2. Hall-Stoodley L, Costerton JW, Stoodley P (2004) Bacterial biofilms:from the natural environment to infectious diseases. Nat RevMicrobiol 2(2):95–108

3. Burmølle M, Kjøller A, Sørensen SJ (2012) An invisible workforce:biofilms in the soil. In: Lear G, Lewis G (eds) Microbial biofilms:current research and applications. Caister Academic press, Norfolk, UK

4. Burmølle M, Webb JS, Rao D, Hansen LH, Sørensen SJ, KjellebergS (2006) Enhanced biofilm formation and increased resistance toantimicrobial agents and bacterial invasion are caused by synergisticinteractions in multispecies biofilms. Appl Environ Microb 72(6):3916–3923

5. Pathak AK, Sharma S, Shrivastva P (2012) Multi-species biofilm ofCandida albicans and non-Candida albicans Candida species onacrylic substrate. Journal of Applied Oral Science 20(1):70–75

6. Wen ZT, Yates D, Ahn SJ, Burne RA (2010) Biofilm formation andvirulence expression by Streptococcus mutans are altered whengrown in dual-species model. BMC microbiology 10(1):111

7. Schwering M, Song J, Louie M, Turner RJ, Ceri H (2013) Multi-species biofilms defined from drinking water microorganisms pro-vide increased protection against chlorine disinfection. Biofouling(ahead-of-print):1-12

8. Yoshida S, Ogawa N, Fujii T, Tsushima S (2009) Enhanced biofilmformation and 3-chlorobenzoate degrading activity by the bacterialconsortium of Burkholderia sp. NK8 and Pseudomonas aeruginosaPAO1. J Appl Microbiol 106(3):790–800

9. Tolker-Nielsen T, Brinch UC, Ragas PC, Andersen JB, Jacobsen CS,Molin S (2000) Development and dynamics of pseudomonas sp.Biofilms. J Bacteriol 182(22):6482–6489

10. Evans RP, Nelson CL, Bowen WR, Kleve MG, Hickmon SG (1998)Visualization of bacterial glycocalyx with a scanning electron micro-scope. Clin Orthop Relat Res 347:243

11. Morck D, Raybould T, Acres S, Babiuk L, Nelligan J, Costerton J(1987) Electron microscopic description of glycocalyx and fimbriaeon the surface of Pasteurella haemolytica-A1. Can J Vet Res 51(1):83

12. O'Toole GA, Kolter R (1998) Initiation of biofilm formation inPseudomonas fluorescens WCS365 proceeds via multiple, conver-gent signalling pathways: a genetic analysis. Mol Microbiol 28(3):449–461

13. de la Cruz-Perera CI, Ren D, Blanchet M, Dendooven L, Marsch R,Sørensen SJ, Burmølle M (2013) The ability of soil bacteria toreceive the conjugative IncP1 plasmid, pKJK10, is different in amixed community compared to single strains. FEMS MicrobiolLett 338(1):95–100

14. Ceri H, OlsonM, Stremick C, Read R,Morck D, Buret A (1999) TheCalgary biofilm device: new technology for rapid determination ofantibiotic susceptibilities of bacterial biofilms. J Clin Microbiol37(6):1771–1776

15. Harrison JJ, Stremick CA, Turner RJ, Allan ND, Olson ME, Ceri H(2010) Microtiter susceptibility testing of microbes growing on peglids: a miniaturized biofilm model for high-throughput screening.Nature protocols 5(7):1236–1254

16. Shin SH, Kim S, Kim JY, Song HY, Cho SJ, Lee K-I, Lim HK, ParkN-J, Hwang IT, Yang K-S (2012) Genome sequence of Paenibacillusterrae HPL-003, a xylanase-producing bacterium isolated from soilfound in forest residue. J Bacteriol 194(5):1266–1266

17. Lira F, Hernández A, Belda E, Sánchez MB, Moya A, Silva FJ,Martínez JL (2012) Whole-genome sequence of Stenotrophomonasmaltophilia D457, a clinical isolate and a model strain. J Bacteriol194(13):3563–3564

D. Ren et al.

18. da Silva AR, Ferro JA, Reinach F, Farah C, Furlan L, Quaggio R,Monteiro-Vitorello C, Van Sluys M, Almeida N, Alves L (2002)Comparison of the genomes of two Xanthomonas pathogens withdiffering host specificities. Nature 417(6887):459–463

19. Morohoshi T, Wang W-Z, Someya N, Ikeda T (2011) Genomesequence of Microbacterium testaceum StLB037, an N -acylhomoserine lactone-degrading bacterium isolated from potatoleaves. J Bacteriol 193(8):2072–2073

20. Al-Ahmad A, Wunder A, Auschill TM, Follo M, Braun G, HellwigE, Arweiler NB (2007) The in vivo dynamics of Streptococcus spp.,Actinomyces naeslundii , Fusobacterium nucleatum and Veillonellaspp. in dental plaque biofilm as analysed by five-colour multiplexfluorescence in situ hybridization. J Med Microbiol 56(5):681–687

21. Dige I, Nilsson H, Kilian M, Nyvad B (2007) In situ identification ofstreptococci and other bacteria in initial dental biofilm by confocallaser scanningmicroscopy and fluorescence in situ hybridization. EurJ Oral Sci 115(6):459–467

22. Thurnheer T, Gmür R, Guggenheim B (2004) Multiplex FISH anal-ysis of a six-species bacterial biofilm. J Microbiol Meth 56(1):37–47

23. Nadell CD, Xavier JB, Foster KR (2009) The sociobiology ofbiofilms. FEMS Microbiol Rev 33(1):206–224

24. Bechet M, Blondeau R (2003) Factors associated with the adherenceand biofilm formation byAeromonas caviae on glass surfaces. J ApplMicrobiol 94(6):1072–1078

25. Dag I, Kiraz N, Yasemin O (2010) Evaluation of different detectionmethods of biofilm formation in clinical Candida isolates. Afr JMicrobiol Res 4(24):2763–2768

26. Nett JE, Cain MT, Crawford K, Andes DR (2011) Optimizing aCandida biofilm microtiter plate model for measurement of antifun-gal susceptibility by tetrazolium salt assay. J Clin Microbiol 49(4):1426–1433

27. Zara S, Gross MK, Zara G, Budroni M, Bakalinsky AT (2010)Ethanol-independent biofilm formation by a flor wine yeast strainof Saccharomyces cerevisiae . Appl Environ Microb 76(12):4089–4091

28. Gupta S, Kaur S, Sehgal S, Sharma A, Chhuneja P, Bains NS (2010)Genotypic variation for cellular thermotolerance in Aegilops tauschiiCoss., the D genome progenitor of wheat. Euphytica 175(3):373–381

29. Savaroglu F, Filik I, Oztopcu V, Kabadere S, Ilhan S, Uyar R (2011)Determination of antimicrobial and antiproliferative activities of theaquatic moss Fontinalis antipyretica Hedw. Turk J Biol 35:361–369

30. Pitts B, Hamilton MA, Zelver N, Stewart PS (2003) A microtiter-plate screening method for biofilm disinfection and removal. JMicrobiol Meth 54(2):269–276

31. Moreno-Paz M, Gómez MJ, Arcas A, Parro V (2010) Environmentaltranscriptome analysis reveals physiological differences betweenbiofilm and planktonic modes of life of the iron oxidizing bacteria

Leptospirillum spp. in their natural microbial community. BMCgenomics 11(1):404

32. Resch A, Leicht S, Saric M, Pásztor L, Jakob A, Götz F, Nordheim A(2006) Comparative proteome analysis of Staphylococcus aureusbiofilm and planktonic cells and correlation with transcriptome pro-filing. Proteomics 6(6):1867–1877

33. Kirakodu S, Govindaswami M, Novak M, Ebersole J, Novak K(2008) Optimizing qPCR for the quantification of periodontal path-ogens in a complex plaque biofilm. The open dentistry journal 2:49

34. Beumer A, King D, Donohue M, Mistry J, Covert T, Pfaller S (2010)Detection of Mycobacterium avium subsp. paratuberculosis indrinking water and biofilms by quantitative PCR. Appl EnvironMicrob 76(21):7367–7370

35. Dalton T, Dowd SE, Wolcott RD, Sun Y, Watters C, Griswold JA,Rumbaugh KP (2011) An in vivo polymicrobial biofilm woundinfection model to study interspecies interactions. PloS one 6(11):e27317

36. Suzuki N, Nakano Y, Yoshida A, Yamashita Y, Kiyoura Y (2004)Real-time TaqMan PCR for quantifying oral bacteria during biofilmformation. J Clin Microbiol 42(8):3827–3830

37. Tsushima I, Kindaichi T, Okabe S (2007) Quantification of anaerobicammonium-oxidizing bacteria in enrichment cultures by real-timePCR. Water Res 41(4):785–794

38. Nadkarni MA, Martin FE, Jacques NA, Hunter N (2002)Determination of bacterial load by real-time PCR using a broad-range (universal) probe and primers set. Microbiology 148(1):257–266

39. Matsuki T, Watanabe K, Tanaka R, Fukuda M, Oyaizu H (1999)Distribution of bifidobacterial species in human intestinal microfloraexamined with 16S rRNA-gene-targeted species-specific primers.Appl Environ Microb 65(10):4506–4512

40. Matsuki T, Watanabe K, Fujimoto J, Takada T, Tanaka R (2004) Useof 16S rRNA gene-targeted group-specific primers for real-time PCRanalysis of predominant bacteria in human feces. Appl EnvironMicrobiol 70(12):7220–7228

41. Price R, Viscount H, Stanley M, Leung KP (2007) Targeted profilingof oral bacteria in human saliva and in vitro biofilms with quantitativereal-time PCR. Biofouling 23(3):203–213

42. Ren Z, Steinberg L, Regan J (2008) Electricity production andmicrobial biofilm characterization in cellulose-fed microbial fuelcells. Water Sci Technol 58(3):617

43. Maeda H, Fujimoto C, Haruki Y, Maeda T, Kokeguchi S, Petelin M,Arai H, Tanimoto I, Nishimura F, Takashiba S (2006) Quantitativereal-time PCR using TaqMan and SYBR Green for Actinobacillusactinomycetemcomitans , Porphyromonas gingivalis , Prevotellaintermedia , tetQ gene and total bacteria. FEMS Immunology &Medical Microbiology 39(1):81–86

High-Throughput Screening and Species Abundance Analysis in Multispecies Biofilms

Manuscript 2

High prevalence of biofilm synergy among bacterial soil isolates in co-cultures

indicates bacterial interspecific cooperation.

High prevalence of biofilm synergy among bacterial soil isolates in co-cultures

indicates bacterial interspecific cooperation

Dawei Ren, Jonas Stenløkke Madsen, Søren J. Sørensen*, Mette Burmølle*

Section of Microbiology, Department of Biology, University of Copenhagen

*Corresponding authors

Postal address: Universitetsparken 15, Bygn. 1, 2100 København Ø, Denmark

e-mail Mette Burmølle: burmolle@bio.ku.dk e-mail Søren J. Sørensen: sjs@bio.ku.dk

Running title: High prevalence of synergy in multispecies biofilms

Key words: Cooperation / Multi-species biofilm / Social evolution / Soil bacteria / Synergy

Abstract

Biofilms that form on roots, litter and soil particles typically contain multiple bacterial species.

Currently, little is known about the interactions transpiring in these multi-species biofilms as only

few studies have been based on environmental isolates. In this study, we assessed the prevalence of

synergism and antagonism in biofilm formation among seven soil-isolates, which were co-cultured

in combinations of four species. In 63% of the four-species biofilms, an increase of biofilm biomass

was observed when compared to those formed by single-species. In contrast, only 6% of the

multispecies biofilms displayed a decreasing trend, thus demonstrating a high prevalence of

synergism in multispecies biofilm formation. One four-species consortium, composed of

Stenotrophomonas rhizophila, Xanthomonas retroflexus, Microbacterium oxydans and

Paenibacillus amylolyticus, showed particularly strong biofilm synergy and was selected for further

studies. X. retroflexus was the only strain, out of the four, capable of forming abundant biofilm on

its own, under the in vitro conditions provided. In accordance, strain specific quantitative PCRs

revealed that X. retroflexus was highly dominant in the four-species consortium (> 97% of total

biofilm cell number). Although the three other strains were present in low abundances, they were all

indispensable for the strong synergistic effect to occur in the four-species biofilms. Moreover,

absolute cell numbers of each strain were significantly enhanced when comparing multi- to single-

species biofilms. This indicated that all the individual strains benefited from joining the multi-

species community. Our results show a high prevalence of synergy in biofilm formation in

multispecies consortia isolated from a natural bacterial habitat and suggest that interspecific

cooperation occurred.

Introduction

Biofilms in environmental systems are comprised of

a large number of different bacteria living together

(Hall-Stoodley et al., 2004). Soil, for example, are

systems that typically are high in bacterial numbers,

diversity and readily available surfaces, which

suggest a good setting for multispecies biofilm

formation (Burmølle et al., 2012). The members of

multispecies biofilms may influence each other

antagonistically; e.g. through resource competition

or production of inhibitory compounds (Rao et al.,

2005), or synergistically; via mechanisms such as

syntrophy, biofilm induction or improved resistance

in biofilms (Burmølle et al., 2013).

We have previously observed strong synergy

among four epiphytic isolates from a marine

environment. When comparing single species

biofilms with four-species biofilms, enhanced

biofilm biomass was observed in the latter, in

addition to higher resistance to antibacterial agents

and better protection from bacterial invasion

(Burmølle et al., 2006). Diaz et al., likewise,

revealed a synergistic partnership between Candida

albicans and streptococci where C. albicans

promoted the ability of streptococci to form

biofilms on abiotic surfaces or on the surface of an

oral mucosa analogue (Diaz et al., 2012).

Additionally, several studies have shown that some

species, which were unable to form a biofilm alone,

could promote the biomass of mixed-species

biofilms (Filoche et al., 2004, Klayman et al., 2009,

Sharma et al., 2005, Yamada et al., 2005). In a

recent study, Lee et al., (2013) addressed the

protective effect of the resistant species in a three-

species biofilm. These resistant species protected

the more sensitive ones from inhibitory compounds

and the overall species ratio remained constant,

implying that resistance mechanisms may serve as

public goods in multispecies biofilms (Lee et al.,

2013).

Multi-species biofilms play an essential role in

maintaining the ecological balance in soil

(Burmølle et al., 2012) and when compared with

single-species biofilms and planktonic counterparts,

multispecies biofilm formation seem to promote

certain advantages, including increased resistance to

antibacterial compounds, enhanced protection from

desiccation and protozoan predation as well as high

rate of horizontal gene transfer (Davey and O'toole

2000, Jefferson 2004, Sørensen et al., 2005).

Therefore a pressing need exists for more research

directed towards understanding of the social

interactions and selective forces that drive bacterial

biofilm communities. Specifically, the net effects

and gains/losses regarding biofilm biomass and

protection level of each strain in a mixed

community must be evaluated in order to

characterize the underlying interactions as being

either cooperative or competitive.

In this study we assessed the prevalence of

synergism in four-species biofilm formation of

different soil isolates. Using quantitative PCR

(qPCR), we determined the presence and

progression dynamics of the individual strains at

different stages of biofilm development in a

selected four-species biofilm community. We

demonstrate a high prevalence of synergistic effects

in multispecies biofilm, indicative of cooperative

forces shaping these communities.

Materials and methods

Bacterial strains used in this study

Seven agricultural, bacterial isolates, previously

identified and characterized, were used in this study:

1) Pseudomonas lutea, 2) Stenotrophomonas

rhizophila, 3) Xanthomonas retroflexus, 4)

Ochrobactrum rhizosphaerae, 5) Microbacterium

oxydans, 6) Arthrobacter nitroguajacolicus and 7)

Paenibacillus amylolyticus (de la Cruz Perera et al.,

2013, Ren et al., 2013).

Biofilm quantification by use of crystal violet

assay

Biofilm formation was assayed and quantified as

previously described (Ren et al., 2013). Briefly,

exponential phase cultures of the seven selected

strains were adjusted to an OD 600 of 0.15 in TSB

(Tryptic Soy Broth, Merck KgaA, Germany)

medium and then inoculated into Nunc-TSP plate

(Cat. No. 445497, Thermo Scientific, Denmark).

The inoculum volumes were 160 μL for

monospecies biofilms and 40 μL for each species in

four-species biofilms. After 24-hour incubation at

24°C with shaking (200 rpm), biofilm formation

was quantified by a modified crystal violet assay

based on the Calgary device.

The four-species consortium composed of strains 2,

3, 5 and 7 was examined for synergy in single-

species and all possible combinations of two-,

three- and four-species biofilms. The inoculum

volumes of each strain were equivalent and added

up to a total of 160 μL. Biofilm assays were

performed as described above.

The biofilm experiments were repeated three times

on three independent days with four replicates every

time. The statistical analyses were conducted using

ANOVA test (SPSS version 17.0 for Windows). P

values < 0.05 were regarded as statistically

significant.

Strain specific qPCR on biofilm and planktonic

fractions

Cell numbers of mono- and co-cultures of strain 2,

3, 5 and 7 were assessed by qPCR as follows. The

four-species biofilms attached on the pegs and

planktonic cells in wells were collected at six time

points (4h, 8h, 12h, 16h, 20h and 24h after co-

inoculation). In addition, both single-species

biofilms and associated planktonic fractions were

sampled at 24 h. Three replicates were prepared at

each time point. Cells were lysed by lysozyme

digestion and bead beating, followed by DNA

extraction using FastDNA™ SPIN Kit for soil

(Qbiogene, Illkirch, France) and species specific

qPCRs as previously reported (Ren et al., 2013).

The PCR programs were adjusted as follows: 95°C

for 2 min, 40 cycles of 95°C for 15s, 64°C for 20s

and 72°C for 20s, followed by a standard

melting/dissociation curve segment. Each sample

was run in duplicate wells and a no template control

was included in every run. As the exact copy

numbers of the 16S rRNA gene per cell in the four

species are currently unknown, the cell numbers

were calculated based on other species in the same

genus as previously described (Ren et al., 2013).

Results

Biofilm formation by single-species and four-

species consortia

The prevalence of synergistic and antagonistic

effects in biofilm formation among the seven soil

isolates was examined by co-culturing all possible

combinations of four strains. A total of 35 different

four-species consortia were screened for biofilm

formation by using the Nunc-TSP lid system and

quantified by the modified CV staining method,

previously demonstrated to be suitable and

reproducible (Ren et al., 2013).

Whether synergistic or antagonistic effects

dominated in the individual four-species biofilm

were assessed by relating the measured absorbance

of the multi-species biofilm (Abs590 MS) to that of

the best single-species biofilm former present in the

relevant combination (Abs590 BS) as follows:

(Abs590 MS - St dev) > (Abs590 BS + St dev) =

Synergism

(Abs590 MS + St dev) < (Abs590 BS - St dev) =

Antagonism

This is based on the assumptions that in the absence

of interactions 1) the cell density of single- and

multi-species biofilms are equal so neither more nor

less biofilm is expected to be formed by multiple

species compared to single species with similar

nutrient availability unless interactions causing

synergistic or antagonistic effects occur and 2) the

best biofilm former dominates the biofilm. Figure 1

shows data from a representative dataset; data from

the three biological replicates are presented as

Supplementary material (Table S1).

Figure 1 Four-species biofilm formation of seven soil isolates. The observed data points (dark grey dots) were

collected by quantifying biofilm formation of all four-species combinations using the crystal violet assay. Error bars

represent standard deviations of four replicates. Light grey bars indicate the amount of monospecies biofilm produced

by the best biofilm former present in the combination. Data points (incl. standard deviations) above grey areas indicate

synergistic effects in four species biofilms and data points (incl. standard deviations) below grey areas indicate

antagonism (see text for further details). Strains: 1- P. lutea, 2- S. rhizophila, 3- X. retroflexus, 4- O. rhizosphaerae, 5-

M. oxydans, 6- A. nitroguajacolicus, 7- P. amylolyticus.

A total of 22 four-species combinations, which

accounted for 63% of all combinations (35),

showed synergy in biofilm formation, while only 6%

(2 of 35) of the four-species combinations showed

antagonistic effects owing to the diminished

biomass compared with single-species biofilms

formation. Furthermore, 10 of the 22 combinations

showing synergy were composed only of poor

biofilm-forming strains, especially combinations 1-

2-5-7 and 1-2-6-7, in which biofilm biomass had

increased by more than 5-fold (P < 0.05). This may

indicate that interspecific interactions had led to

cooperative biofilm formation by these strains that

were unable to form biofilm when grown

individually. In total, 31% (11 of 35) of all

combinations were categorized as having no

significant change in biomass as (Abs590 MS - St

dev) < (Abs590 BS + St dev) and (Abs590 MS + St

dev) > (Abs590 BS - St dev).

Figure 2 Biofilms formed by four isolates 2-

Stenotrophomonas rhizophila, 3- Xanthomonas

retroflexus, 5- Microbacterium oxydans and 7-

Paenibacillus amylolyticus when equal aliquots of the

diluted cultures were incubated in co-cultures of two,

three and four isolates. Assays for the detection of

synergistic effects were performed three times

(Experiment 1, 2 and 3) with 4 replicates each time.

Error bars represent ± SEM of four replicates.

Synergism in biofilm formation among S.

rhizophila, X. retroflexus, M. oxydans and P.

amylolyticus

An example of strong synergy was found in strain

combination 2-3-5-7 (S. rhizophila, X. retroflexus,

M. oxydans and P. amylolyticus, respectively,

Figure 1 and Supplementary Table S1). To

determine whether all these four strains contributed

to the enhanced biomass, each strain was grown as

single-species biofilm and in all possible

combinations as two, three and four-species biofilm

and analyzed for synergy in biofilm formation as

described above (Figure 2).

Two- and three-species biofilms did, with the

exception of combinations 3-5 (P = 0.026) and 3-5-

7 (P = 0.001), not differ significantly (P > 0.05)

compared with the amount of biofilm produced by

the single species. Strain 7, P. amylolyticus, which

grew much slower in TSB medium than the other

three strains (data not shown), did not produce any

biofilm alone, but as strain 2, it stimulated biofilm

formation of the other three species. When the four

isolates were co-cultured, the biomass increased by

more than 4-fold compared to that of single-species

biofilms and 2-5 fold compared to that of three-

species biofilms. This indicates that each of these

four strains was indispensable to induce strong

synergy, and interestingly, this applied both to the

good biofilm former, strain 3 X. retroflexus as well

as the three other strains that did not produce

biofilm in monocultures.

Figure 3 The estimated cell numbers per peg/well of each strain in four-species biofilms and associated planktonic

fractions of S. rhizophila, X. retroflexus, M. oxydans and P. amylolyticus at six time points (4 h, 8 h, 12 h, 16 h, 20 h

and 24 h after co-incubation) based on qPCR. MP – Multispecies planktonic cells, MB – Multispecies biofilms. Each

point represents the mean of three replicates, with vertical lines representing ± standard deviations.

Cell numbers in mono- and co-cultures

Specific qPCRs were performed in order to evaluate

the total cell numbers of each strain in both biofilms

and planktonic communities at different stages of

multi-species biofilm development. Additionally,

the cell numbers in single-species biofilms and

planktonic cells at 24 h were measured. The DNA

samples from different time points were diluted

appropriately in order to yield results within the

dynamic range of the standard curves, which were

generated using duplicate 10-fold dilutions of

plasmid DNAs (Ren et al., 2013). R-Square (RSq)

values, based on threshold cycles of the standard

curves, ranged from 0.988 to 1 and application

efficiencies (E) ranged from 84.7 - 98.0 %

(Supplementary material, Table S2).

The cell numbers of each species in the

multispecies biofilm (peg) and in the associated

planktonic fraction (well) are shown in Figure 3. In

general, a marked increase in cell numbers was

observed from 4h to 12h for all of the four strains in

the biofilm fraction. After 12 hours, only the cell

numbers of strain 3 increased continuously, whereas

those of the other three species in the multi-species

biofilm remained constant or decreased. Cell

numbers of all four species in the planktonic

fraction increased during the first 12-16 hours,

hereafter planktonic growth leveled off, indicating

nutrient depletion and transition to a steady state

condition.

Strains

S. rhizophila X. retroflexus M. oxydans P. amylolyticus

Cell

num

bers

(lo

g1

0)

0

2

4

6

8

10

Multispecies biofilms

Single-species biofilms

Figure 4 The estimated cell numbers for each strain (S. rhizophila, X. retroflexus, M. oxydans and P. amylolyticus) in

four-species and single-species biofilms at 24h. Bars represent means ± standard deviation for three replicates.

The cell numbers of each strain in the multi-

species biofilm and in the single-species biofilms

at 24h are shown in Figure 4. For all of the four

species, cell numbers were significantly higher

(P < 0.05) in four-species biofilms compared to

single-species biofilms, indicative of individual

fitness gains when joining the multispecies

biofilm, when interpreting enhanced cell

numbers in biofilms as fitness gain.

The ratio of total cell numbers at 24h of the four

strains S. rhizophila, X. retroflexus, M. oxydans

and P. amylolyticus was 4:900:9:15 in

multispecies biofilms (Table 1). Thus, there is a

strong dominance of X. retroflexus in the four-

species biofilm, as this strain constitutes > 97%

of the total cell number. It is worth noticing that,

in contrast to biofilms (Figure 4), planktonic cell

numbers of the four species, except for P.

amylolyticus, are lower in the four-species co-

culture compared to single-species planktonic

biomass (Table 1). The summarized cell

numbers from the biofilm and the planktonic

fraction from a specific well are not

representative of the cell number of that well

because there are other surfaces available for

bacterial attachment in the wells besides the pegs.

Discussion

High prevalence of synergism in biofilm

formation

Single-species biofilms are rare in natural

environments, especially in agricultural soil where

micro-communities exposed to plenty of organic

matter have the potential to develop into

multispecies biofilms with high bacterial density

and diversity (Burmølle et al., 2010, Narisawa et al.,

2008, Rodríguez and Bishop 2007). Such conditions

in bacterial habitats are likely to facilitate the

development of intricate relationships between

different species. While many previous studies have

focused on interspecies interactions within oral

microbial communities (Kuramitsu et al., 2007,

Palmer Jr et al., 2001, Saito et al., 2008, Sharma et

al., 2005, Wang and Kuramitsu 2005), research on

multispecies biofilms composed of soil bacteria is

still in its infancy. In this study, seven soil bacteria,

isolated from one micro-habitat, were screened in

order to assess the prevalence of interactions

leading to synergism in biofilm formation of four-

species co-cultures.

Table 1 Absolute cell numbers per peg/well measured by quantitative PCRs at 24h in multispecies planktonic cells,

single-species planktonic cells, multispecies biofilms and single-species biofilms. Each value is the mean of three

replicates.

Strains Multispecies

planktonic cells

Single-species

planktonic cells

Multispecies

biofilms

Single-species

biofilms

S. rhizophila 2.33E+07 5.79E+07 1.35E+07 9.14E+03

X. retroflexus 3.38E+08 5.46E+08 2.97E+09 4.92E+07

M. oxydans 6.42E+06 1.58E+08 2.93E+07 4.73E+04

P. amylolyticus 2.81E+07 1.33E+07 4.78E+07 2.10E+03

In total, 63% of four-species biofilms showed

synergistic effects in biofilm formation while only 6%

displayed antagonistic activity. This is in agreement

with the assumption that the driving force in

bacterial community development is the self-

organization and cooperation rather than

competition of individual microorganisms (Daniels

et al., 2004, Davies et al., 1998, Parsek and

Greenberg 2005). The recently presented “Black

Queen Hypothesis”, explains how multispecies

cooperation can evolve (Morris et al., 2012). This

hypothesis considers cooperation in complex

bacterial communities as being a consequence of

species adapting to the presence of each other. In

order to enhance own fitness, species delete vital

functions or pathways that are provided by the

surrounding bacteria. This leads to a dependency

and is therefore an irreversible commitment to

living in close association with other species, which

may often require development of more complex

systems to ensure that the coexistence is maintained.

When applying the Black Queen Hypothesis on the

results of the present study, the strains may prefer

living in the multispecies biofilm in order to keep

vital partners in close contact, or the ability of

biofilm formation could be the function that is lost

by some of the species.

The synergistic interactions were also observed

when four poor biofilm formers were co-cultured

(Figure 1), which verified the previous findings

wherein the biofilm forming ability of individual

bacteria is not necessarily an indicator of their

potential in multispecies biofilms (Bharathi et al.,

2011, Burmølle et al., 2006). This shifting of

biofilm pattern from weak to moderate or strong

can be the result of metabolic interactions (Møller

et al., 1998), enhanced coaggregation (Rickard et

al., 2003), organized spatial distribution (Skillman

et al., 1998) and/or facilitated initial surface

attachment (Klayman et al., 2009, Simões et al.,

2008), e.g. bridging bacteria may facilitate the

association of other species that do not coaggregate

directly with each other. Thus, species that do not

form biofilms as single strains may benefit from the

advantages associated to biofilm formation,

including enhanced protection from outside stress

and expanded niche availability, by engaging in

multispecies communities.

In contrast to our results, Foster et al., recently

showed that the great majority of interactions in

pairwise species combinations of bacterial isolates

from tree-hole rainwater pools were net negative

and very few strong higher-order positive effects

arose from combinations composed of more than

two species (Foster and Bell 2012). The two studies

differ with respect to bacterial habitats targeted for

isolation (soil vs. tree-holes), productivity

parameter assessed (biofilm formation vs. CO2

production) and different definitions of

synergism/cooperation, which may explain the

observed differences. The tree-hole species tend to

use similar resources which may be the key factor

that lead to competition among microbes (Lawrence

et al., 2012), while in nutrient rich agricultural soil,

bacteria are likely to be tightly associated and by

their collective metabolic activities, individual

bacterial consortia stabilize their environment and

fertilize the soil. The key to whether bacterial

species compete or cooperate may lie in their

possibility of long-term co-adaptation and degree of

niche overlap. We are currently investigating the

significance of these parameters for the net

interactions of bacterial communities.

Enhanced biofilm formation in co-cultures

When exploring strain dynamics of the selected

four-species community, the strongest synergy was

induced only when all the four strains S. rhizophila,

X. retroflexus, M. oxydans and P. amylolyticus were

co-cultivated (Figure 2). This demonstrates that

each strain played an important role in the enhanced

biofilm formation, regardless of the capability of

single-species biofilm formation. Divergence in

resource use may be one of the reasons that lead to

increased productivity of the entire community. By

adapting to consume different resources, and to

metabolize waste products produced by other

species, the consortium collectively decompose

substrates in the medium more effectively. In the

present study, the synergistic enhancement of

biomass of the four co-cultured species was only

observed when grown as a biofilm; Planktonic co-

culturing did not lead to changes in overall biomass

(data not shown). This indicates that a structured

environment is highly important for the synergy to

occur. Thus co-metabolism (syntrophy) may

explain the observed synergy, but only when

coupled to a structured environment that enables

tight cell-cell associations optimizing

product/substrate availability. Similar observations

have previously been reported (Hansen et al., 2007,

Stewart et al., 1997).

qPCR results showed a strong dominance (> 97%)

of X. retroflexus in the four-species biofilm, which

is consistent with the results from the crystal violet

assay where X. retroflexus was the only good

biofilm former out of the four strains. Despite its

monospecies biofilm forming abilities, cell numbers

of strain X. retroflexus were enhanced

approximately 60-fold when co-cultured in biofilms

with the other strains. The remaining three strains

constituted < 3% of the total cell number of the

four-species biofilms, however they all showed

enhanced cell numbers in the multispecies biofilm

by over three orders of magnitude when compared

to cell numbers in monospecies biofilms. Thus, S.

rhizophila, M. oxydans and P. amylolyticus,

although present in low abundances and inability of

monospecies biofilm formation, stimulate of X.

retroflexus and obtain higher cell numbers when

present in the multispecies biofilm. This significant

change in capacity of biofilm formation may be

explained by the fact that, species evolving in

communities may have higher growth rates when

assayed in the presence of other species (Lawrence

et al., 2012), especially in structured communities

(Hansen et al., 2007, Stewart et al., 1997). P.

amylolyticus, of which cell numbers was most

strongly enhanced, has previously been reported to

produce antibacterial agents with broad spectrum

activity against both Gram-negative and Gram-

positive bacteria (DeCrescenzo Henriksen et al.,

2007). However, in this study, this strain appeared

to stimulate rather than inhibit the growth of other

three strains. Further studies to identify the changes

in gene expression patterns in this four-species

biofilm will lead to a deeper understanding of the

underlying molecular mechanisms of synergism in

bacterial communities.

Synergism in biofilm formation indicate

cooperation

Several observations in this study indicate that

bacteria increase fitness from joining multispecies

biofilms. If this fitness advantage applies to all of

the species present, the underlying interaction is

categorized as being cooperative (West et al., 2007).

Four such observations are discussed here. (i) We

observed a high prevalence of biofilm synergy in

four-species biofilms, strongly dominating over

antagonistic effects. This implies selection for

living in multispecies communities, indirectly

indicating that bacteria may enhance fitness when

joining multispecies biofilms. The fitness advantage

may be caused by growth promotion that enhances

bacterial biomass and thereby the direct fitness, but

also the advantages gained from the biofilm-

associated bacterial protection may be of major

significance in natural environments with high

levels of stress. (ii) Each of the four species in the

selected community was indispensable for the

synergistic effect observed on biofilm biovolume.

These vital interdependencies may evolve over

many years under continuous selective pressures,

whereby only fitness enhancing relationships are

favored. (iii) The cell numbers of all four species

were higher in multi-species biofilms when

compared to biofilms composed of single species.

Thus, when focusing on biofilm-associated growth,

there is a direct gain in fitness for all strains from

joining the four-species biofilm. (iv) The observed

synergistic effect only applied to the four-species

biofilm; cell numbers in the planktonic fraction

decreased for three out of the four strains. Based on

this, there seem to be selection for or forces driving

bacteria to form multi-species biofilms.

Alternatively, metabolic cooperation occurs in the

biofilm, whereby the growth rates of biofilm-

resident strains are enhanced.

In conclusion, the observations presented in this

study points in the direction of multispecies

biofilms as being a favorable bacterial habitat; they

gain protection and opportunities of engaging in

mutually beneficial cooperative interactions.

Acknowledgements

This study was partly funded by grants to SJS and MB

from the Danish Innovation Consortium SiB, ref no:

11804520, and The Danish Council for Independent

Research; ref no: DFF-1335- 00071 and ref no: DFF-

1323-00235 (SIMICOM).

References

Bharathi P, Bhowmick P, Shekar M, Karunasagar I

(2011). Biofilm formation by pure and mixed culture of

Lactobacillus isolates on polystyrene surface in varying

nutrient conditions. Biotechnol. Bioinf. Bioeng. 1: 93-98.

Burmølle M, Webb JS, Rao D, Hansen LH, Sørensen SJ,

Kjelleberg S (2006). Enhanced biofilm formation and

increased resistance to antimicrobial agents and bacterial

invasion are caused by synergistic interactions in

multispecies biofilms. Applied and environmental

microbiology 72: 3916-3923.

Burmølle M, Thomsen TR, Fazli M, Dige I, Christensen

L, Homøe P et al., (2010). Biofilms in chronic infections

– a matter of opportunity–monospecies biofilms in

multispecies infections. FEMS Immunology & Medical

Microbiology 59: 324-336.

Burmølle M, Kjøller A, Sørensen SJ (2012). An

Invisible Workforce: Biofilms in the Soil. Microbial

Biofilms: Current Research and Applications: 61.

Burmølle M, Ren D, Bjarnsholt T, Sørensen SJ (2013).

Interactions in multispecies biofilms: do they actually

matter? Trends in Microbiology, In Press: DOI:

10.1016/j.tim.2013.12.004

Daniels R, Vanderleyden J, Michiels J (2004). Quorum

sensing and swarming migration in bacteria. FEMS

microbiology reviews 28: 261-289.

Davey ME, O'toole GA (2000). Microbial biofilms: from

ecology to molecular genetics. Microbiology and

molecular biology reviews 64: 847-867.

Davies DG, Parsek MR, Pearson JP, Iglewski BH,

Costerton J, Greenberg E (1998). The involvement of

cell-to-cell signals in the development of a bacterial

biofilm. Science 280: 295-298.

de la Cruz Perera CI, Ren D, Blanchet M, Dendooven L,

Marsch R, Sørensen SJ et al., (2013). The ability of soil

bacteria to receive the conjugative IncP1 plasmid,

pKJK10, is different in a mixed community compared to

single strains. FEMS Microbiology Letters 338: 95-100.

DeCrescenzo Henriksen E, Phillips D, Peterson J (2007).

Polymyxin E production by P. amylolyticus. Letters in

applied microbiology 45: 491-496.

Diaz PI, Xie Z, Sobue T, Thompson A, Biyikoglu B,

Ricker A et al., (2012). Synergistic interaction between

Candida albicans and commensal oral streptococci in a

novel in vitro mucosal model. Infection and immunity 80:

620-632.

Filoche S, Anderson S, Sissons C (2004). Biofilm

growth of Lactobacillus species is promoted by

Actinomyces species and Streptococcus mutans. Oral

microbiology and immunology 19: 322-326.

Foster KR, Bell T (2012). Competition, not cooperation,

dominates interactions among culturable microbial

species. Current Biology 22: 1845-1850.

Hall-Stoodley L, Costerton JW, Stoodley P (2004).

Bacterial biofilms: from the natural environment to

infectious diseases. Nature Reviews Microbiology 2: 95-

108.

Hansen SK, Rainey PB, Haagensen JA, Molin S (2007).

Evolution of species interactions in a biofilm community.

Nature 445: 533-536.

Jefferson KK (2004). What drives bacteria to produce a

biofilm? FEMS microbiology letters 236: 163-173.

Klayman BJ, Volden PA, Stewart PS, Camper AK

(2009). Escherichia coli O157: H7 requires colonizing

partner to adhere and persist in a capillary flow cell.

Environmental science & technology 43: 2105-2111.

Kuramitsu HK, He X, Lux R, Anderson MH, Shi W

(2007). Interspecies interactions within oral microbial

communities. Microbiology and molecular biology

reviews 71: 653-670.

Lawrence D, Fiegna F, Behrends V, Bundy JG,

Phillimore AB, Bell T et al., (2012). Species interactions

alter evolutionary responses to a novel environment.

PLoS biology 10: e1001330.

Lee KWK, Periasamy S, Mukherjee M, Xie C,

Kjelleberg S, Rice SA (2013). Biofilm development and

enhanced stress resistance of a model, mixed-species

community biofilm. The ISME journal 1-14.

Morris JJ, Lenski RE, Zinser ER (2012). The Black

Queen Hypothesis: evolution of dependencies through

adaptive gene loss. MBio 3: DOI: 10.1128/mBio.00036–

12

Møller S, Sternberg C, Andersen JB, Christensen BB,

Ramos JL, Givskov M et al., (1998). In situ gene

expression in mixed-culture biofilms: evidence of

metabolic interactions between community members.

Applied and environmental microbiology 64: 721-732.

Narisawa N, Haruta S, Arai H, Ishii M, Igarashi Y

(2008). Coexistence of antibiotic producing and

antibiotic-sensitive bacteria in biofilms is mediated by

resistant bacteria. Applied and environmental

microbiology 74: 3887-3894.

Palmer Jr RJ, Kazmerzak K, Hansen MC, Kolenbrander

PE (2001). Mutualism versus independence: strategies of

mixed-species oral biofilms in vitro using saliva as the

sole nutrient source. Infection and immunity 69: 5794-

5804.

Parsek MR, Greenberg E (2005). Sociomicrobiology:

the connections between quorum sensing and biofilms.

Trends in microbiology 13: 27-33.

Rao D, Webb JS, Kjelleberg S (2005). Competitive

interactions in mixed-species biofilms containing the

marine bacterium Pseudoalteromonas tunicata. Applied

and environmental microbiology 71: 1729-1736.

Ren D, Madsen J, Cruz-Perera C, Bergmark L, Sørensen

S, Burmølle M (2013). High-Throughput Screening of

Multispecies Biofilm Formation and Quantitative PCR-

Based Assessment of Individual Species Proportions,

Useful for Exploring Interspecific Bacterial Interactions.

Microbial ecology: DOI: 10.1007/s00248-013-0315-z

Rickard AH, Gilbert P, High NJ, Kolenbrander PE,

Handley PS (2003). Bacterial coaggregation: an integral

process in the development of multi-species biofilms.

Trends in microbiology 11: 94-100.

Rodríguez SJ, Bishop PL (2007). Three-dimensional

quantification of soil biofilms using image analysis.

Environmental Engineering Science 24: 96-103.

Saito Y, Fujii R, Nakagawa KI, Kuramitsu H, Okuda K,

Ishihara K (2008). Stimulation of Fusobacterium

nucleatum biofilm formation by Porphyromonas

gingivalis. Oral microbiology and immunology 23: 1-6.

Sharma 420 A, Inagaki S, Sigurdson W, Kuramitsu H

(2005). Synergy between Tannerella forsythia and

Fusobacterium nucleatum in biofilm formation. Oral

microbiology and immunology 20: 39-42.

Simões LC, Simoes M, Vieira MJ (2008). Intergeneric

coaggregation among drinking water bacteria: evidence

of a role for Acinetobacter calcoaceticus as a bridging

bacterium. Applied and environmental microbiology 74:

1259-1263.

Skillman LC, Sutherland IW, Jones MV, Goulsbra A

(1998). Green fluorescent protein as a novel species-

specific marker in enteric dual-species biofilms.

Microbiology 144: 2095-2101.

Stewart PS, Camper AK, Handran S, Huang C-T,

Warnecke M (1997). Spatial distribution and coexistence

of Klebsiella pneumoniae and Pseudomonas aeruginosa

in biofilms. Microbial ecology 33: 2-10.

Sørensen SJ, Bailey M, Hansen LH, Kroer N, Wuertz S

(2005). Studying plasmid horizontal transfer in situ: a

critical review. Nature Reviews Microbiology 3: 700-

710.

Wang BY, Kuramitsu HK (2005). Interactions between

oral bacteria: inhibition of Streptococcus mutans

bacteriocin production by Streptococcus gordonii.

Applied and environmental microbiology 71: 354-362.

West SA, Diggle SP, Buckling A, Gardner A, Griffin AS

(2007). The social lives of microbes. Annual Review of

Ecology, Evolution, and Systematics 38: 53-77.

Yamada M, Ikegami A, Kuramitsu HK (2005).

Synergistic biofilm formation by Treponema denticola

and Porphyromonas gingivalis. FEMS microbiology

letters 250: 271-277.

Supplementary Table 1

Biofilms formed by single species and combinations (four strains). 40μL of each of the four diluted cultures

was inoculated in each well. The plate was incubated for 24 h followed by crystal violet staining and

absorbance measurements at 590 nm. The ratio of (Abs 590 multispecies biofilm – St dev) / (Abs 590 best

strain + St dev) or (Abs 590 multispecies biofilm + St dev) / (Abs 590 best strain - St dev) was calculated

and the average value from three experiments is presented as the “Fold change”.

Strains Biofilm formation (Abs 590)

a

Fold change Experiment-1 Experiment-2 Experiment-3

1 c 0.143 ± 0.038

b 0.180 ± 0.106 0.256 ± 0.036 -

2 0.126 ± 0.011 0.077 ± 0.004 0.077 ± 0.016 -

3 3.769 ± 0.589 6.606 ± 1.171 6.245 ± 0.502 -

4 0.381 ± 0.009 0.190 ± 0.037 0.302 ± 0.011 -

5 0.157 ± 0.009 0.094 ± 0.048 0.166 ± 0.018 -

6 0.055 ± 0.014 0.049 ± 0.028 0.149 ± 0.073 -

7 0.004 ± 0.003 0.017 ± 0.004 0.033 ± 0.010 -

1234 3.763 ± 0.289 4.649 ± 0.435 3.777 ± 0.144 0.910

1245 0.245 ± 0.021 0.281 ± 0.037 0.247 ± 0.030 Nd

1256 0.593 ± 0.084 0.992 ± 0.490 0.390 ± 0.061 1.767

1267 0.668 ± 0.176 1.895 ± 0.320 2.564 ± 0.589 5.325

1235 4.642 ± 0.689 10.889 ± 0.905 6.050 ± 1.544 N

1236 13.734 ± 0.433 12.697 ± 2.195 13.790 ± 1.572 1.910

1237 6.290 ± 1.043 6.937 ± 1.727 6.142 ± 0.742 N

1246 0.633 ± 0.241 1.972 ± 0.621 1.978 ± 0.093 3.001

1247 0.882 ± 0.049 0.987 ± 0.253 1.840 ± 0.511 3.113

1257 1.146 ± 0.207 1.947 ± 0.464 1.782 ± 0.187 5.292

1345 5.246 ± 0.806 10.877 ± 2.125 5.198 ± 0.961 N

1356 8.454 ± 1.014 11.587 ± 0.575 9.646 ± 0.526 1.460

1367 9.950 ± 1.818 11.033 ± 0.339 10.226 ± 1.822 1.442

1346 10.528 ± 0.676 12.201 ± 1.132 12.650 ± 1.069 1.721

1347 13.542 ± 0.905 12.526 ± 1.317 9.806 ± 0.983 1.730

1357 13.078 ± 1.068 13.601 ± 0.622 11.502 ± 0.738 1.893

1456 0.168 ± 0.049 0.103 ± 0.013 0.268 ± 0.005 0.741

1467 1.053 ± 0.298 0.707 ± 0.175 0.839 ± 0.394 1.862

1457 1.257 ± 0.155 0.676 ± 0.173 1.169 ± 0.334 2.624

1567 0.876 ± 0.143 0.498 ± 0.220 1.375 ± 0.694 2.229

2345 3.452 ± 0.655 6.246 ± 0.559 3.674 ± 0.091 N

2356 7.292 ± 1.099 7.646 ± 1.411 6.153 ± 1.172 N

2367 9.224 ± 0.588 9.102 ± 1.934 9.813 ± 1.833 1.260

2346 7.436 ± 1.394 6.762 ± 0.894 7.757 ± 1.833 N

2347 14.468 ± 1.888 15.250 ± 0.894 13.813 ± 0.588 2.127

2357 21.067 ± 1.335 22.885 ± 2.816 19.045 ± 2.158 3.002

2456 0.330 ± 0.040 1.023 ± 0.084 0.514 ± 0.083 1.785

2467 0.200 ± 0.070 0.526 ± 0.136 0.413 ± 0.119 N

2457 0.310 ± 0.060 0.281 ± 0.070 0.471 ± 0.115 N

2567 0.794 ± 0.480 1.028 ± 0.467 1.502 ± 0.444 3.929

3456 7.020 ± 0.180 6.542 ± 0.370 7.325 ± 1.049 1.021

3467 8.340 ± 1.117 6.710 ± 1.350 6.533 ± 0.973 N

3457 8.908 ± 2.148 9.374 ± 1.759 11.777 ± 2.374 1.259

3567 11.436 ± 1.773 11.006 ± 1.709 10.053 ± 2.061 1.427

4567 0.129 ± 0.034 0.085 ± 0.063 0.399 ± 0.154 N a Assays for the detection of biofilm formation were performed three times (Experiment 1, 2 and 3).

b Values represent the means ± standard deviation of four replicates in each experiment.

c 1- Pseudomonas lutea, 2- Stenotrophomonas rhizophila, 3- Xanthomonas retroflexus, 4- Ochrobactrum rhizosphaerae,

5- Microbacterium oxydans, 6- Arthrobacter nitroguajacolicus, 7- Paenibacillus amylolyticus. d Combinations which showed (Abs 590 multispecies biofilm – St dev) < (Abs 590 best strain + St dev) and (Abs 590

multispecies biofilm + St dev) > (Abs 590 best strain - St dev) were identified as no significant change in biomass.

Supplementary Table 2

Standard curves used to measure the cell numbers of strains S. rhizophila, X. retroflexus, M. oxydans and P.

amylolyticus in mono- and multi-species biofilms and planktonic cell fractions.

Species RSq Application

efficiency(E)

Estimated copy number of

the 16S rRNA gene (per

cell)a

S. rhizophila 1.000 87.1% 4

X. retroflexus 0.999 94.3% 2

M. oxydans 0.988 98.0% 2

P. amylolyticus 0.999 84.7% 12 a The copy numbers of the 16S rRNA gene were estimated according to other species in the same genus (Ren et al.

2013).

Manuscript 3

Metatranscriptome analysis of multispecies biofilms indicates strain- and community-

dependent changes in gene expression

Metatranscriptome analysis of multispecies biofilms indicates strain- and

community- dependent changes in gene expression

Lea Benedicte Skov Hansen+, Dawei Ren

+, Søren J. Sørensen*, Mette Burmølle*

+Shared first authorship

Section of Microbiology, Department of Biology, University of Copenhagen

*Corresponding authors

Postal address: Universitetsparken 15, Bygn. 1, 2100 København Ø, Denmark

e-mail Mette Burmølle: burmolle@bio.ku.dk e-mail Søren J. Sørensen: sjs@bio.ku.dk

Key words: Metatranscriptome / Multispecies biofilm / Soil bacteria / Synergistic interactions

Abstract

It has gradually become more apparent that bacteria often exist in naturally formed multispecies

biofilms. Within these biofilms interspecies interaction seems to play an important role in shaping

the function and activities of these dynamic communities. However, little is known about the effect

of interspecies interaction on a gene expression level in these multispecies biofilms. This study

represents a comparative gene expression analysis of the Xanthomonas retroflexus transcriptomes

when grown in a single-species biofilm and in dual- and four-species biofilms with

Stenotrophomonas rhizophila, Microbacterium oxydans and Paenibacillus amylolyticus. The result

revealed a species specific response to coexistence in the dual-species biofilm, where the overall

strongest change in expression profile was observed in the dual-species biofilms of X. retroflexus

and P. amylolyticus. Furthermore, a distinct expression pattern was detected in the four-species

biofilm including changes in expression of genes, which were not observed differentially expressed

in any of the dual-species biofilms. This non-linear response in the four-species biofilm indicates a

complex interaction pattern at the gene regulation level when increasing the number of species in

the co-cultures. This is in correspondence with the previously described emergent behaviour of

enhanced biofilm formation by this specific four-species consortia isolated from soil [1]. 70 genes

were found differentially expressed when co-culturing X. retroflexus with other species, which

include genes involved in amino acid metabolism, membrane bound efflux system and MazE/MazF

toxin-antitoxin system, suggesting the enhanced resistance of multispecies biofilms.

Introduction

It is now acknowledged that biofilms, as the

dominant microbial lifestyle in nature, are

composed of multiple species, where extensive

interactions between different species are

bound to play a crucial role in shaping these

dynamic communities [2]. In general,

interspecies interactions, either synergistic or

antagonistic, involve physical contact,

metabolic communication, quorum sensing and

genetic exchange [3]. The synergistic effect of

living as a mixed community, which often

exhibits enhanced biomass compared with

either single species grown alone, is implicated

in the development of several beneficial

phenotypes, including promoted cell attachment

due to coaggregation [4], co-metabolism where

one species takes advantage of the metabolite

produced by a neighbouring species [5], and

cross-species protection against antimicrobials

or host immune responses [6, 7]. The

synergistic interactions in multispecies biofilms

are considered to be an effective strategy in

degrading organic matter [8, 9] in nature and

evading host clearance in chronic infections

[10], indicating the profound impact

interactions have both on major ecological

processes and on clinical settings.

While the reductionist approach has advanced

biofilm research by analysing single-species

biofilms or individual species in a complex

microbial community, interactions between

different organisms within multispecies

biofilms as a key area for providing insight into

evolutionary ecology and developing new

therapeutics, is still in its infancy. The rise of

metatranscriptomic analysis to study the global

gene expression profiling of the whole

microbial communities, has been greatly

accelerated by RNA-Seq in the past few years,

and thus opens new possibilities for clarifying

the functional interaction networks within

multispecies microbial communities. This new

technology has been recently applied to an oral

biofilm model [11], where cell-cell interactions

are believed to play integral roles in the

development of biofilm architecture and

pathogenicity. The dramatic changes in gene

expression profiles of this multispecies biofilm

model in the presence of periodontal pathogens

were accurately assessed. Moreover, small

noncoding RNAs with important gene

regulatory roles were identified, showing

promise of metatranscriptomic analysis in

complex microbial communities [11].

The synergism in a four-species biofilm

composed of Stenotrophomonas rhizophila,

Xanthomonas retroflexus, Microbacterium

oxydans and Paenibacillus amylolyticus, was

presented in our previous studies with increased

biofilm formation over 3-fold relative to single-

species biofilms [1]. As the only good biofilm

former, the prevalence (> 97% of total biofilm

cell number) of strain X. retroflexus was

demonstrated by species specific quantitative

PCR [12]. Moreover, despite of the low

abundance of three other strains, they were all

indispensable for the strong synergy that

occurred in the four-species biofilm which

make this consortium a powerful model to

study intricate interactions in multispecies

biofilms.

In the present study, the gene expression profile

of X. retroflexus in a single-species biofilm was

compared to its expression profiles in dual-

species biofilms with S. rhizophila, M.oxydans

or P. amylolyticus as well as in a four-species

biofilm. This represents, to our knowledge, the

first evaluation of gene expression changes due

to the interspecies interactions in a multispecies

biofilm model composed of soil isolates. We

utilized an RNA-Seq-based metatranscriptomic

approach, and found a significant effect of the

interspecies interactions on the gene expression.

Materials and methods

Bacterial strains and growth conditions

The four species used in this study were

Stenotrophomonas rhizophila (JQ890538),

Xanthomonas retroflexus (JQ890537),

Microbacterium oxydans (JQ890539), and

Paenibacillus amylolyticus (JQ890540) [1].

Strains were streaked from frozen glycerol

stocks onto TSA (Tryptic Soy Agar) plates and

incubated for 48h at 24°C. Hereafter, the

freshly isolated colonies were transferred into 5

mL of TSB (Tryptic Soy Broth) and incubated

with shaking (250 rpm) at 24°C overnight.

Biofilm cultivation

Biofilms for gene expression analysis by RNA-

sequencing were grown in six-well polystyrene

plates (Greiner Bio-One, Germany). Overnight

cultures of each strain were subcultured to

exponential phase and adjusted to an OD600nm

of 0.15 in fresh TSB, as described by Ren et al.

2013 [1]. A total of 4 mL of single-species,

dual-species or four-species diluted cultures

(admixtures of each species at equal cell

density) were inoculated in each well and

allowed to grow for 24h at 24°C. Single- and

dual-species biofilms were prepared in

triplicate wells, while four-species biofilm were

prepared in five replicate wells.

DNA and RNA extraction

Isolation of genomic DNA from overnight

cultures of four bacteria was performed as

previously described using FastDNA™ SPIN

Kit for soil (Qbiogene, Illkirch, France) [1].

For RNA extraction, planktonic cells from

biofilm-containing wells were gently removed,

wells were rinsed three times using phosphate

buffer solution (PBS) and then the biofilms

were scraped with sterile pipette tips into 1 mL

of PBS. The biofilm suspensions were

immediately centrifuged and resuspended in

100 μL of PBS followed by addition of 500 μL

of RNAlater® (Ambion). The RNAlater-

preserved samples were kept at 4 °C overnight,

after which RNAlater was removed by adding

600 μL cold PBS and centrifuging at 8000 rpm

for 5 min at 4 °C. The pellet was resuspended

in 200 μL of lysozyme solution (20 μg/mL) and

incubated at room temperature for 10 min, with

vortexing for 10s every 2 min. After 700 μL of

buffer RLT was added and vortexed for 10s, the

obtained solution was transferred into bead-

beating tube (provided by the FastDNA™

SPIN Kit for soil) and bead beaten using the

Savant FastPrep FP120 for 30s at setting 5.0,

followed by centrifugation at 13000 rpm for 8

min at 4 °C. 0.85 mL of supernatant was

transferred to new microcentrifuge tube. 0.47

mL of 100% ethanol was added and mixed by

pipetting. 0.7mL of lysate was transferred on to

RNeasy mini spin column. Total RNA was then

further purified from each biofilm sample using

the RNeasy mini kit (Qiagen) according to

manufacturer’s instructions.

Genomic DNA removal was conducted

according to the instructions in DNAfree™ Kit

(Ambion), except the incubation time was

extended to 1 hour. The complete removal of

DNA was confirmed by RT-qPCR using 20 µL

SYBR Green reactions on Mx 3000 (Stratagene,

Cedar Creek, Texas). The qPCR targeted the

16S rRNA gene by the eubacterial primers

EUB338 and EUB518 [13]. All qPCR reactions

were run in technical duplicates and contained

10 µL of Brilliant III SYBR Green QPCR

Master Mix (Stratagene, Cedar Creek, Texas), 1

µL forward primer (final concentration 385

nM), 1 µL reverse primer (final concentration

385 nM), 1 µL template DNA (100-fold diluted

to avoid PCR inhibitors in accordance to Bustin

SA et al. [14], and 7 µL ddH2O. The program

was modified from Bergmark L et al. [15],

combining the annealing and extension step:

95 °C for 3 min followed by 40 cycles of 95 °C

for 10 s, 60 °C for 20 s, and a final dissociation

curve. The standard curve for bacteria was

made from a pure culture of Pseudomonas

putida kt2440 [16, 17]. mRNA was

subsequently isolated using the MicroExpress

Bacterial mRNA Purification Kit (Ambion)

following the manufacturer’s instructions.

Sequencing

The genomic DNA was prepared for

sequencing using the NEBNext Quick DNA

Sample Prep. Master Mix 2 (New England

BioLabs Inc., Ipswitch, MA, USA). DNA was

amplified by PCR and sequenced as 250 bp

paired end libraries on an Illumina MiSeq

(Illumina, San Diego, CA, USA). The RNA

transcripts were prepared using ScriptSeq™ v2

RNA-Seq Library Preparation Kit (Epicentre

Biotechnologies, Madison, WI, USA) and

sequenced as 50 bp single reads on an Illumina

HiSeq 2000 (Illumina, San Diego, CA, USA).

Data analysis

The quality filtering of the genomic and

metatranscriptomic libraries was carried out in

similar ways. Only reads approved by the

CASAVA v1.8.2 (Illumina, San Diego, CA,

USA) were included and adaptor remnants

were removed. End nucleotides with a quality

Phred score below 20 were filtered. After the

end trim process, sequences shorter than 50

nucleotides (nt) or 45 nt were removed for the

genomic and metatranscriptome libraries,

respectively. Reads were discarded if they

displayed a mean Phred score below 15 or a

mean Phred score below 10 in a sliding window

of 10 nt for the genomic libraries and a mean

Phred score below 15 across a sliding window

of 5 nt for the metatranscriptomes. The quality

filtering was carried out using Biopieces.

Velvet v.1.2.07 was used to assemble the

genomic data [18] and Prodigal v2.50 was used

for gene calling [19]. A gene catalogue was

created including the genes from all four

genomes.

All metatranscriptome libraries were mapped to

the four genomes using Bowtie2 v.2.0.5 [20]

and number nt mapping to genes in the gene

catalogue was calculated, which yielded an

expression matrix. The 1000 most expressed

genes from X. retroflexus were extracted from

the matrix and the Bray-Curtis dissimilarity

metric was calculated between samples and

visualized in a NMDS plot. EdgeR was utilized

to infer any statistical significant differential

expression of the top 1000 genes [21]. The

average number of nt mapping to a gene within

biofilm types was calculated and the logarithm

to base 2 fold changes (log FC) was calculated

between single-species biofilm and the four

different multispecies biofilms. Genes showing

log FCs between -3 and 3 were excluded and

infinite values were treated according to the

expression levels of the gene in the single-

species biofilm. The remaining genes were

annotated using protein Blast and the non-

redundant protein database maintained by

National Center for Biotechnology Information

(NCBI - 2014-01-09), furthermore, conserved

Pfam domains were located using the Pfam web

based batch search tool and signal peptides

were identified using the signalP 4.1 server [22-

24].

Results

Sequencing statistics

In order to conduct a comparative gene

expression analysis of X. retroflexus (Xr) grown

in single-species, dual-species and four-species

biofilms with S. rhizophila (Sr), M. oxydans

(Mo) and P. amylolyticus (Pa), the genomes of

these four species were sequenced to create a

mapping template for the RNA-Seq transcripts.

The raw output from the 250 bp paired end

Illumina MiSeq DNA sequencing were quality

filtered and between 53.47% and 68.83% nt

remained, which yielded between 210 Mb and

1,475 Mb for assembly. The largest estimated

genome size of 6.3 Mb and the highest number

of coding sequence (CDS) belonged to Pa

(Table S1). The smallest genome size of 2.5

Mb and lowest number of CDS belonged to Sr,

however the assembly statistics for Sr indicated

a poor assembly of the genome with a low N50

and high number of contigs (Table S1).

The transcriptome and metatranscriptome

libraries were sequenced, quality filtered and

mapped against the genome templates. After

quality filtering the RNA transcripts, 75.82% nt

remained (Table S2). The filtered reads were

mapped against all four genomes and 92.39%

nt could be remapped, however only 2.46%

mapped against CDSs (Table S2). The

remaining 97.54% nt were mainly from rRNAs

and few mapped to intergenic regions. In

general, a higher percentage of the RNA

transcripts mapped to CDSs from biofilms

containing Pa (2.32% to 4.86%) (Table S2).

Expression activity

The proportion of transcripts from the four

genomes was estimated to assess the relative

species specific expression activity in each of

the metatranscriptomes from the dual- and four-

species biofilms (Figure 1). A noise level of

2.20% was detected, which represents the

average percentage of transcripts that mapped

unspecifically to conserved regions in the

genomes. The abundances of Sr and Mo were

close to the detection limit, making their true

expression activities difficult to assess. Pa

transcripts dominated the transcriptomes with

more than 50% and showed the highest

percentage in the dual-species biofilms (Xr+Pa).

A decrease in abundance could be observed for

Pa in the four-species biofilm while Xr showed

a relative increase compared to their

abundances in the dual-species biofilms,

however this was not statistically significant

(Student’s t-test, p-values > 0.05).

Comparative expression analysis of the

single-, dual-, and four-species biofilms

In order to investigate gene expression changes

in Xr induced by interspecies interactions, Xr

expression profiles of dual- and four-species

biofilms were compared to that of the Xr

single-species biofilm. For this purpose, all

transcripts from Xr were extracted from the

metatranscriptomes and expression profiles of

Xr in each sample were obtained. To infer the

correlation between the sample replicates, a

pearson correlation matrix was created

including all 17 samples and the replicates

showed good correlations with values between

0.85 and 0.98 (Table S3). The high rRNA

content in the samples resulted in low sequence

coverage of the CDSs and higher deviation for

low abundant transcripts. Hence, only the 1000

highest expressed genes in Xr were used for

further analysis, in order to minimize errors

caused by the low genome coverage and low

abundant transcripts (Table S1).

Figure 1 The species specific transcript abundance

in dual- and four-species biofilms. The broken line

represents the average percentage of transcript that

map unspecifically to conserved regions in the

genome and represents the detection limit (2.20%).

A) The relative abundance of transcripts that

mapped against the Xr genome, in the Xr+Sr,

Xr+Mo, Xr+Pa and Xr+Sr+Mo+Pa biofilms. B) The

relative abundance of transcripts that mapped

against the Sr genome in the Xr+Sr and

Xr+Sr+Mo+Pa biofilms. C) The relative abundance

of transcripts that mapped against the Mo genome

in the Xr+Mo and Xr+Sr+Mo+Pa biofilms. D) The

relative abundance of transcripts that mapped

against the Pa genome in the Xr+Pa and

Xr+Sr+Mo+Pa biofilms.

After the samples were normalized to the Xr

transcript library sizes and transformed using

the natural logarithm, a similarity analysis of

the biofilms was conducted based on the

expression of the top 1000 genes. The Bray-

Curtis dissimilarity metric was applied and the

results are reflected in an NMDS plot (Figure

2). Three clusters appeared in the plot. The first

cluster consisted of the single-species biofilms

of Xr and the dual-species biofilms of Xr+Sr

and Xr+Mo. Within this cluster the Xr and

Xr+Mo replicates created separate sub-clusters,

however the replicates of Xr+Sr showed larger

deviation and did not cluster together. The

second cluster consisted of the three replicates

of Xr+Pa in close proximity of each other and

the third tight cluster consisted of the five

replicates of the four-species biofilms,

Xr+Sr+Mo+Pa.

The 1000 highest expressed genes in Xr were

tested for significantly different expression

across the five sample groups using statistics

based on empirical Bayes methods [25]. This

analysis resulted in 587 differentially expressed

genes (p-values < 0.05 with a Benjamini and

Hochberg correction). To detect the active

genes in the transition from growing in a

single-species biofilm to dual- and four-species

biofilms, the ratio of the average number of nt

mapping to a specific gene in the single-species

biofilms vs. each of the four different

multispecies biofilms was calculated. A total of

70 genes had an logarithm to base 2 fold

change (log FC) above 3 or below -3, which

correspond to a 8 times up- or down-regulation.

The log FCs of the 70 genes are displayed in a

heatmap (Figure 3) and the result correlated

well with the observations from the NMDS plot.

Many expression ratios for Xr+Sr and Xr+Mo

were close to zero, which indicated a closer

relation to the single-species biofilms. However,

two genes (Xr21 and Xr22) showed an up-

regulation in both dual-species biofilms (Xr+Sr

and Xr+Mo) and four genes (Xr23, Xr24, Xr25

and Xr26) were down-regulated in the Xr+Sr

biofilms. The Xr+Pa and Xr+Sr+Mo+Pa

biofilms clustered together and most genes

seemed to be up- and down-regulated in both

biofilm types. However, a substantial number

of genes were regulated in the four-species

biofilms only. However, 3 genes (Xr3, Xr4 and

Xr27) displayed unique regulation in the Xr+Pa

biofilms.

Figure 2 NMDS plot of the Bray-Curtis

dissimilarity measure between the Xanthomonas

retroflexus (Xr) expression profiles of the 1000

most abundant transcripts, when grown in dual- and

four-species biofilms with Stenotrophomonas

rhizophila (Sr), Microbacterium oxydans (Mo) and

Paenibacillus amylolyticus (Pa).

Functional analysis of the differentially

expressed genes

The 70 genes with highly responding

expression profiles were annotated using Blast

against the non-redundant database at NCBI

and a search for conserved domains was

conducted and signal peptides were detected

[22-24]. These annotation results can be

observed in Table S4. Some genes annotated to

functions outside the cytoplasm but did not

contain a signal peptide, however some genes

were truncated and the signal peptide could be

missing.

The differentially expressed genes were divided

into functional groups, where the most

prominent were membrane proteins, genes

involved in regulation, and amino acid

metabolisms. Table 1 lists genes encoding

membrane proteins and many of these can be

assigned into subgroups of potential efflux

systems (Xr14, Xr18, Xr38 and Xr66), receptors

(Xr19, Xr27 and Xr67), transport proteins (Xr1,

Xr5, Xr10, Xr31 and Xr41), and chemotaxis

proteins (Xr7 and Xr34). Genes involved in

efflux systems and receptors displayed a

consistent up-regulation in the Xr+Pa and the

four-species biofilms compared to the single-

species biofilms. However, Xr27 only showed

an up-regulation in the Xr+Pa biofilms (Figure

3). The transport and chemotaxis subgroups

showed more variable expressions with both

up- and down-regulations in the Xr+Pa and

four-species biofilms, except for Xr31 and Xr34,

which were up-regulated in the four-species

biofilms only.

Few of the membrane protein-encoding genes

annotated to more specific functions. A domain

for CsgG (PF03783.9) was identified in Xr44,

which was up-regulated in both the Xr+Pa and

four-species biofilms. The CgsG domain is

found in an outer membrane lipoprotein and it

is thought to be limiting factor in the assembly

of curli, which are adhesive fibres [26]. The

Xr2 gene encodes a domain (PF08813.6),

which is connected to bacterial integrins and

this gene was down-regulated in the Xr+Pa and

four-species biofilms. Integrins are also

connected to cell adhesion [27]. Furthermore, a

domain found in an outer membrane lysozyme

type-c inhibitor (PF09864.4) was located in the

Xr30 gene. This gene was up-regulated in the

four-species biofilms, indicating resistance

toward lysozyme [28].

Table 1 Differential expressed genes of Xanthomonas retroflexus (Xr) encoding for membrane proteins

ID name e-val Pfam e-val sig-

Pa

Xr+Sr Xr+Mo Xr+Pa Xr+Sr+

Mo+Pa

Xr1 major facilitator superfamily

transmembrane nitrite

extrusion

0 PF07690.11 4.7E-13 N -1.33 -0.90 -3.35 -4.11

Xr 2 hypothetical protein

SMD_0239

2.00E-151 PF08813.6 5.4E-42 Y -1.05 -1.43 -3.30 -2.81

Xr5 transport-associated protein 1.00E-66 PF04972.12 5.1E-17 Y -0.19 -0.02 -3.57 -3.14

Xr7 methyl-accepting

chemotaxis protein

3.00E-99 PF13426.1 2.5E-16 N -0.45 -0.63 -8.60 -5.31

Xr10 membrane protein 7.00E-112 PF08212.7 3.7E-47 N 1.10 0.50 -3.14 -5.82

Xr14 membrane protein 0 PF02321.13 3.9E-35 Y 0.76 1.80 2.63 3.41

Xr18 RND family efflux

transporter MFP subunit

0 PF00529.15 1.6E-59 Y 0.84 2.29 4.76 4.92

Xr 19 TonB-dependent receptor 0 PF00593.19 1.6E-24 Y 1.88 2.10 4.21 4.83

Xr 27 TonB-dependent

siderophore receptor

0 PF00593.19 4.4E-23 Y 1.10 0.34 3.07 0.30

Xr30 hypothetical protein 1.00E-131 PF09864.4 3.9E-13 Y 0.24 -0.07 0.18 3.28

Xr31 TolA protein 0 PF13103.1 1.6E-10 N 0.04 -2.01 1.08 3.56

Xr32 membrane protein 9.00E-41 - - N 0.00 -2.22 1.72 3.36

Xr34 methyl-accepting

chemotaxis protein

0 PF00015.16,

PF00672.20

2.6E-55,

7.2E-11

N -1.16 -1.62 1.71 3.24

Xr38 efflux transporter, RND

family, MFP subunit

0 PF12700.2 1.7E-30 N -1.50 0.19 3.33 4.99

Xr41 Vitamin B12 transporter

btuB

0 PF00593.19 2.2E-28 N -0.45 2.08 2.67 4.24

Xr44 hypothetical protein, partial 0 PF03783.9 1.9E-13 N 0.38 1.51 3.05 3.90

Xr66 multidrug transporter 1.00E-68 PF00893.14 9.3E-21 N -0.66 -0.03 2.96 3.52

Xr67 TonB-denpendent receptor 0 PF00593.19 1.8E-27 Y -1.19 -0.91 2.12 3.60

aIdentified signal peptides in the gene: yes (Y) or no (N).

Table 2 Differential expressed genes of Xanthomonas retroflexus (Xr) encoding functions in regulation

ID name e-val Pfam e-val sig-

Pa

Xr+Sr Xr+Mo Xr+Pa Xr+Sr+

Mo+Pa

Xr4 hypothetical protein 2.00E-41 PF05532.7 6.1E-20 N -0.95 -0.27 -3.10 -1.26

Xr12 AbrB family transcriptional

regulator

2.00E-39 PF04014.13 5E-09 N 2.62 0.10 6.83 7.47

Xr17 hypothetical protein

SMD_2382

6.00E-64 PF02604.14 2.2E-09 N 1.73 1.45 4.27 3.58

Xr20 leucyl aminopeptidase 0 PF00883.16 2.1E-115 N 2.40 1.40 3.64 4.54

Xr22 HTH-type transcriptional

regulator betI

4.00E-98 PF00440.18 1.9E-11 N 3.75 3.34 0.27 -0.85

Xr39 hypothetical protein Smlt2244 1.00E-84 PF01850.16 1.3E-08 N -1.83 -0.74 4.28 5.00

Xr58 ArsR family transcriptional

regulator

0 PF08241.7,

PF01022.15

1.1E-22,

2.5E-14

N -1.89 -0.28 1.81 5.24

Xr63 NusA antitermination factor 0 PF08529.6,

PF13184.1,

PF14520.1,

PF00575.18

2.4E-41,

5.5E-27,

9.5E-10,

2.4E-09

N -0.27 0.49 1.75 3.42

Xr65 DEAD/DEAH box helicase 0 PF00270.24 2.4E-46 N -0.86 0.43 3.43 3.07

Xr70 ArsR family transcriptional

regulator

2.00E-65 PF12840.2 2.7E-19 N -0.50 0.10 1.65 3.95

aIdentified signal peptides in the gene: yes (Y) or no (N).

Table 3 Differential expressed genes of Xanthomonas retroflexus (Xr) encoding functions in amino acid

metabolisms

ID name e-val Pfam e-val sig-

Pa

Xr+Sr Xr+Mo Xr+Pa Xr+Sr+

Mo+Pa

Xr23 methylcrotonoyl-CoA

carboxylase

0 PF02786.12,

PF00289.17,

PF00364.17

6.5E-73,

1.8E-38,

3.4E-15

N -3.85 -1.03 0.43 2.02

Xr28 aromatic amino acid

aminotransferase

0 PF00155.16 2.8E-80 N 1.32 -3.09 1.91 2.67

Xr37 aminotransferase 0 PF00733.16 2E-58 N 0.00 0.58 4.05 4.19

Xr49 branched-chain amino acid

aminotransferase

0 - -

N 0.15 0.43 1.19 3.52

Xr51 glycine system protein H 2.00E-83 PF01597.14 2.2E-45 N 0.34 0.87 1.93 3.20

Xr56 biotin synthase 0 PF06968.8,

PF04055.16

1.5E-28,

2E-15

N -3.33 0.68 2.93 3.05

aIdentified signal peptides in the gene: yes (Y) or no (N).

Figure 3 Heatmap displaying the logarithmic (to

base 2) fold change (log FC) of the expression of

Xanthomonas retroflexus (Xr) genes between the

single-species biofilm and each of the four different

multispecies biofilms. All genes were among the

1000 most abundant Xr transcripts. They displayed

a logarithmic fold change above 3 (equivalent to 8

times difference in expression levels) compared to

the single-species biofilm and were tested

significantly different (p-values < 0.05 with

Benjamini Hochberg correction). A euclidean

clustering of both the samples and the genes are

displayed as similarity trees on the x- and y-axes,

respectively.

Table 2 summarizes all differentially expressed

genes that encode regulatory functions. Five

genes are involved in transcription regulation

(Xr22, Xr58, Xr63, Xr65 and Xr70) and these

genes were mainly up-regulated in the four-

species biofilms compared to the single-species

biofilm. An exception was Xr65, which was

also up-regulated in the Xr+Po biofilms and

Xr22, which was one of the two genes up-

regulated in the Xr+Sr and Xr+Mo biofilms

(Figure 3). Toxin-antitoxin systems were also

present among the differentially expressed

genes and these systems have been shown to

play a role in regulation [29]. An antitoxin

domain (PF02604.14) was identified in Xr17.

Furthermore, Xr12 encodes the AbrB/MazE

antitoxin and Xr39 contains a PIN domain

(PF01850.16) that is often present in toxins

[29]. Xr12 and Xr39 are consecutive genes in

the genome of Xr, and this indicates the

presence of a complete toxin-antitoxin system

in the differentially expressed genes [29]. All

three genes seemed to be up-regulated in both

the Xr+Pa and four-species biofilms and

particular Xr12 showed a high shift in

regulation level (Figure 3).

Genes participating in other regulation

mechanisms were also identified (Table 2). Xr4

was down-regulated only in the Xr+Pa biofilms;

this gene encodes a hypothetical protein with a

CsbD domain that has been connected to stress

response [30]. Leucyl aminopeptidase, involved

in the Glutathione metabolism, is encoded by

Xr20 and up-regulation both in the Xr+Pa and

the four-species biofilms was observed [31].

Glutathione is known to participate in

responses to environmental factors [32].

Table 3 lists genes involved in amino acid

metabolism. Branched chain amino acid

aminotransferase and methylcrotonoyl-CoA

carboxylase (Xr49 and Xr23) are both known

for their participation in Leucine, Valine and

Isoleucine metabolisms and were up-regulated

in the four-species biofilms. Furthermore, X23

displayed a down-regulation in the Xr+Sr

biofilm. Biotin synthase, encoded by Xr56,

participates in the production of biotin; a

cofactor for methylcrotonoyl-CoA and limiting

to amino acid decomposition [33]. This gene

was up-regulated in both the Xr+Pa and four-

species biofilms. Additionally, Xr28 encodes an

aromatic amino acid aminotransferase, which

catalyses the last step in Tyrosine and

Phenylalanine synthesis. Xr28 was slightly

down-regulated in the Xr+Mo biofilms and up-

regulated in both the Xr+Pa and four-species

biofilms. The aminotransferase (Xr37) contains

an AsnB domain, representing an alternative

pathway for Asparagine synthesis from

Aspartate in the absence of AsnA and this gene

was up-regulated in both the Xr+Pa and four-

species biofilms. Xr51 encodes a glycine

cleavage system protein H, which participates

in the catabolism of Glycine and was up-

regulated in the four-species biofilms.

Besides genes coding for membrane proteins or

participating in regulation and amino acid

metabolisms, four differentially expressed

genes were annotated to peptidoglycan systems.

Three of these genes (Xr24, Xr60 and Xr64)

were up-regulated in the four-species biofilm,

indicating an increased peptidoglycan

production, furthermore, Xr24 was down-

regulated in the Xr+Sr biofilms. The fourth

gene (Xr8), which was down-regulated in the

Xr+Pa and four-species biofilms, encodes a

domain (PF03734.9) often found in an L,D-

transpeptidase that has been shown to provide

an alternative pathway for peptidoglycan

synthesis [34]. Furthermore, the gene (Xr15)

encoding a cell division protein (FtsK), which

plays an important role in cell division, seemed

to be up-regulated in both the Xr+Pa and four-

species biofilm.

In general, many genes involved in cell defense

systems have been identified as being

differentially expressed, including the

membrane bound efflux systems involved in

multidrug exclusion and the lysozyme type-c

inhibition (Table 1). In connection to this, a

beta-lactamase domain was identified in Xr52,

which was up-regulated in the four-species

biofilm. Additionally, an up-regulation of Xr29

in the four-species biofilm was observed, which

encodes a DNA processing protein (DprA) that

is involved in DNA recombination.

Discussion

This study demonstrates the effect of

interspecies interactions in multispecies

biofilms on the gene expression in X.

retroflexus (Xr). Comparisons of the expression

profiles of X. retroflexus grown in single- and

dual-species biofilms with S. rhizophila (Sr), M.

oxydans (Mo) and P. amylolyticus (Pa),

revealed shifts in gene expression, however,

these changes seemed to be species dependent.

Through further investigations of the four-

species biofilms, indication of a non-linear

response to multiple species was identified,

suggesting complex and dependent interaction

patterns. This is in good correspondence with

the previously described emergent behaviour of

enhanced biofilm formation by this four-species

consortium [12]. The genes identified in the

differential expression analysis revealed

changes in expression of regulatory elements,

membrane proteins and genes involved in

amino acid metabolism. Only few studies exist

that investigated the effect of coexistence in

multispecies biofilms on a gene expression

level and the results presented here contributes

to further understanding of the interaction

mechanisms in multispecies biofilms [11, 35,

36].

Pa displayed high gene expression activity

when coexisting with Xr in biofilms. Over 50%

of the mRNA transcripts in the

metatranscriptomes of the Xr+Pa and four-

species biofilms annotated back to the genome

of Pa, indicating high activity levels. Our

previous study showed that Pa did not display

good biofilm production abilities [1] and the

biofilm production decreased in the dual-

species biofilm Xr+Pa compared to the single-

species biofilm of Xr according to Ren et al., in

review [12]. However, the cell abundance of Pa

showed a significantly increase in four-species

biofilm which is consistent with the high

activity level of Pa in the present study. But,

we should note that the relative abundance of

mRNA can vary between 1-5% of the total

cellular RNA and the percentage of transcripts

mapping to CDSs increased in samples

containing Pa, which indicated a higher

mRNA/rRNA ratio in Pa compared to Xr, Sr

and Mo [37]. An increased mRNA/rRNA ratio

would yield more mRNAs per cell and this

could introduce a bias in the assessment of the

relative activity. Furthermore, Pa displayed the

largest estimated genome size and this could

also affect the proportion of transcripts from Pa.

The proportion of transcripts annotating back to

Sr and Mo was close to the detection limit and

indicated a low activity level. The low activity

level in the four-species biofilms correlated

with the low cell copy numbers found by the

qPCR analysis in our previous study [12].

Coexistence of different species in biofilms

seemed to alter the gene expression, however,

the response was species specific [11, 35]. The

presence of Pa clearly induced a change in the

gene expression profiles of Xr, compared to the

profiles observed in the single-species biofilm

and was responsible for the induced regulation

of the majority of the 70 highly differentially

expressed genes. It did show high activity

levels in the biofilms and this could be causing

the response observed in the Xr expression. The

induced gene regulation in the dual-species

biofilm of Xr+Sr and Xr+Mo was less evident.

However, the Xr+Mo replicates did create a

cluster, which separated from the single-species

biofilm and some genes were differently

regulated in their presence. This was consistent

with the biofilm formation of Xr+Mo which

significantly increased compared with the

single-species biofilm of Xr, while the Xr+Sr

biofilms did not show any significantly increase

in biomass [12]. Transcripts from Sr and Mo

constituted a smaller proportion of the

metatranscriptomes and showed low abundance

in the previous study of the four-species biofilm

composition [12]. The relative low abundance

and activity could explain the low effect of the

presence of Sr and Mo on the Xr gene

expression.

Interestingly, when inferring the four-species

biofilms, a unique expression pattern for Xr

emerged, with regulations uniquely observed

for this combination. The five biological

replicates of the four-species biofilms created a

tight cluster that separated from the other

biofilm samples (Figure 2) and several genes

displayed expression shifts, which were not

observed in any of the dual-species biofilms

(Figure 3). Hence, the co-occurrence of Sr, Mo

and Pa induced a different expression pattern in

Xr, compared to the gene expression in the

dual-species biofilms, which is consistent with

our previous results that each strain is

indispensable for the synergistic interactions in

this four-species biofilm [12]. This observation

was not evident in a gene expression study of

Saccharomyces cerevisiae exposed to multiple

physiological transitions [38]. They showed

that when exposed to different simultaneous

stresses, the cellular response approximated the

sum of the responses for each individual stress

and indicated a linear response to multiple

physiological changes [38]. The current result

differs from these observations and suggests a

non-linear response to multiple species,

indicating that gene expression is regulated by a

complex interspecies interaction network.

The functional annotation of the 70

significantly differentially expressed genes

revealed expression changes of regulatory

elements in the presence of Pa [11]. Genes

coding for transcription factors and toxin-

antitoxin system displayed an increase in

expression in Xr. Especially the potential

MazE/MazF toxin-antitoxin system was up-

regulated and has been reported to be a general

regulatory element and involved in biofilm

formation [39]. Also, the MazE/MazF system

has been connected with programmed cell

death and lysis, which was widely reported in

Xanthomonas [40, 41]. Furthermore, genes

encoding transcriptional regulators show

unique expression levels in the four-species

biofilm only. These changes in expression of

regulatory elements possibly control the

phenotypic adaptation of the cells to the

community composition in the biofilms [42].

The regulation of genes encoding membrane

proteins with roles in transport, efflux systems,

receptors and chemotaxis indicated that many

of the phenotypic changes induced by

coexistence in biofilm were in relation to the

surrounding microenvironment [11]. Some of

the identified transporters could potentially play

a role in secreting extracellular polymeric

substances (EPS). Furthermore, the up-

regulation of the potential cell division protein,

the potential participant in curli production and

chemotaxis systems in the presence of Pa could

also indicate higher growth rate, adhesion and

specific growth direction. An increase in EPS

excretion and directional growth could be a

competitive response to coexistence and an

advance for Xr [43]. A further indication of

enhanced growth was the observed up-

regulation of peptidoglycan production in the

four-species biofilm only. All these

observations are correlated with previous

quantitative PCR (qPCR) measurements which

demonstrated the increase in cell number of Xr

when co-cultured with other three species [12].

Up-regulation of potential defense system like

efflux pump was evident in the presence of Pa

and a potential lysozyme type-c inhibitor

together with a potential beta-lactamase were

up-regulated in the four-species biofilm. The

up-regulation of these genes indicates an

upgrade in defense mechanisms against toxins

in this multispecies biofilm [44], which could

positively affect the total resistance of the

biofilm. One of the main emergent behaviour

of bacterial biofilm is enhanced antibiotic

resistance. This is usually thought to be due to

lack of diffusion and low growth rate [45, 46].

However our results may indicate that this is

also due to specific induction of resistance trait.

The expression changes in amino acid

metabolisms represented a potential

interspecies cooperation in the multispecies

biofilms [36]. The aromatic amino acid

aminotransferase displayed an up-regulation in

the presence of Pa. Through further inspection,

the function of this gene is not immediately

present in the genome of Pa, hence, an

enhanced production and export of tyrosine and

phenylalanine could constitute an example of

interspecies cooperation. Furthermore, the

catabolism of valine, leucine, isoleucine and

glycine seemed to be up-regulated in the four-

species biofilm and this could indicate an

increased exchange of amino acids or

polypeptides. The interspecies interaction could

be further investigated by investigating the

expression profiles of Sr, Mo and Pa.

The current comparative gene expression study

of X. retroflexus (Xr) grown in single-, dual-

and four-species biofilms with S. rhizophila

(Sr), M. oxydans (Mo) and P. amylolyticus (Pa)

displayed phenotypic adaptation according to

the species composition in the biofilms. By

inferring the differentially expressed genes, the

adaptations seemed to respond to both positive

and negative relations. It can be hypothesized

that interspecies interactions in biofilms are

balances between cooperation and competition.

To further investigate the effects of interspecies

interactions, the phenotypic changes in

knockout mutants of candidate genes identified

in the comparative gene expression analysis

might further elucidate their functional effect

on interspecies interaction in multispecies

biofilms.

References

1. Ren D, Madsen JS, de la Cruz-Perera CI,

Bergmark L, Sørensen SJ, Burmølle M (2013)

High-Throughput Screening of Multispecies

Biofilm Formation and Quantitative PCR-Based

Assessment of Individual Species Proportions,

Useful for Exploring Interspecific Bacterial

Interactions. Microbial ecology: 1-9.

2. Hall-Stoodley L, Costerton JW, Stoodley P

(2004) Bacterial biofilms: from the natural

environment to infectious diseases. Nature

Reviews Microbiology 2: 95-108.

3. Burmølle M, Ren D, Bjarnsholt T, Sørensen SJ

(2014) Interactions in multispecies biofilms: do

they actually matter? Trends in microbiology.

doi: 10.1016/j.tim.2013.12.004

4. Vornhagen J, Stevens M, McCormick DW,

Dowd SE, Eisenberg JN, Boles BR, Rickard AH

(2013) Coaggregation occurs amongst bacteria

within and between biofilms in domestic

showerheads. Biofouling 29: 53-68.

5. Farabegoli G, Chiavola A, Rolle E (2008)

Remediation of chlorophenol-and phenol-

contaminated groundwater by a sequencing

batch biofilm reactor.

6. Schwering M, Song J, Louie M, Turner RJ, Ceri

H (2013) Multi-species biofilms defined from

drinking water microorganisms provide

increased protection against chlorine

disinfection. Biofouling 29: 917-928.

7. Høiby N, Bjarnsholt T, Givskov M, Molin S,

Ciofu O (2010) Antibiotic resistance of bacterial

biofilms. International journal of antimicrobial

agents 35: 322-332.

8. Seneviratne G, Zavahir J, Bandara W,

Weerasekara M (2008) Fungal-bacterial

biofilms: their development for novel

biotechnological applications. World Journal of

Microbiology and Biotechnology 24: 739-743.

9. LoVLEY DR, Klug MJ (1986) Model for the

distribution of sulfate reduction and

methanogenesis in freshwater sediments.

Geochimica et Cosmochimica Acta 50: 11-18.

10. Harrison F (2007) Microbial ecology of the

cystic fibrosis lung. Microbiology 153: 917-923.

11. Frias-Lopez J, Duran-Pinedo A (2012) Effect of

periodontal pathogens on the metatranscriptome

of a healthy multispecies biofilm model. Journal

of bacteriology 194: 2082-2095.

12. Ren D, Madsen JS, Sørensen SJ, Burmølle M

(2013) High prevalence of biofilm synergy

among bacterial soil isolates in co-cultures

indicates bacterial interspecific cooperation.

The ISME journal. Under review.

13. Fierer N, Jackson JA, Vilgalys R, Jackson RB

(2005) Assessment of soil microbial community

structure by use of taxon-specific quantitative

PCR assays. Applied and Environmental

Microbiology 71: 4117-4120.

14. Bustin SA, Benes V, Garson JA, Hellemans J,

Huggett J, Kubista M, Mueller R, Nolan T,

Pfaffl MW, Shipley GL (2009) The MIQE

guidelines: minimum information for

publication of quantitative real-time PCR

experiments. Clinical chemistry 55: 611-622.

15. Bergmark L, Poulsen PHB, Al‐Soud WA,

Norman A, Hansen LH, Sørensen SJ (2012)

Assessment of the specificity of Burkholderia

and Pseudomonas qPCR assays for detection of

these genera in soil using 454 pyrosequencing.

FEMS microbiology letters 333: 77-84.

16. Park J-W, Crowley DE (2005) Normalization of

soil DNA extraction for accurate quantification

of target genes by real-time PCR and DGGE.

BioTechniques 38: 579-586.

17. VanGuilder HD, Vrana KE, Freeman WM

(2008) Twenty-five years of quantitative PCR

for gene expression analysis. Biotechniques 44:

619.

18. Zerbino DR, Birney E (2008) Velvet:

algorithms for de novo short read assembly

using de Bruijn graphs. Genome research 18:

821-829.

19. Hyatt D, Chen G-L, Locascio PF, Land ML,

Larimer FW, Hauser LJ (2010) Prodigal:

prokaryotic gene recognition and translation

initiation site identification. BMC

bioinformatics 11: 119-119.

20. Langmead B, Salzberg SL (2012) Fast gapped-

read alignment with Bowtie 2. Nature methods

9: 357-359.

21. Robinson MD, McCarthy DJ, Smyth GK (2010)

edgeR: a Bioconductor package for differential

expression analysis of digital gene expression

data. Bioinformatics (Oxford, England) 26: 139-

140.

22. Altschup SF, Gish W, Pennsylvania T, Park U

(1990) Basic Local Alignment Search Tool. J

mol biol 215: 403-410.

23. Punta M, Coggill PC, Eberhardt RY, Mistry J,

Tate J, Boursnell C, Pang N, Forslund K, Ceric

G, Clements J, Heger A, Holm L, Sonnhammer

ELL, Eddy SR, Bateman A, Finn RD (2012)

The Pfam protein families database. Nucleic

acids research 40: D290-301.

24. Petersen TN, Brunak S, von Heijne G, Nielsen

H (2011) SignalP 4.0: discriminating signal

peptides from transmembrane regions. Nat Meth

8: 785-786.

25. McCarthy DJ, Chen Y, Smyth GK (2012)

Differential expression analysis of multifactor

RNA-Seq experiments with respect to

biological variation. Nucleic acids research 40:

4288-4297.

26. Loferer H, Hammar M, Normark S (1997)

Availability of the fibre subunit CsgA and the

nucleator protein CsgB during assembly of

fibronectin-binding curli is limited by the

intracellular concentration of the novel

lipoprotein CsgG. Molecular microbiology 26:

11-23.

27. Kelly G, Prasannan S, Daniell S, Fleming K,

Frankel G, Dougan G, Connerton I, Matthews S

(1999) Structure of the cell-adhesion fragment

of intimin from enteropathogenic Escherichia

coli. Nature structural biology 6: 313-318.

28. Callewaert L, Aertsen A, Deckers D,

Vanoirbeek KGa, Vanderkelen L, Van

Herreweghe JM, Masschalck B, Nakimbugwe D,

Robben J, Michiels CW (2008) A new family of

lysozyme inhibitors contributing to lysozyme

tolerance in gram-negative bacteria. PLoS

pathogens 4: e1000019-e1000019.

29. Anantharaman V, Aravind L (2003) New

connections in the prokaryotic toxin-antitoxin

network: relationship with the eukaryotic

nonsense-mediated RNA decay system.

Genome biology 4: R81-R81.

30. Prágai Z, Harwood CR (2002) Regulatory

interactions between the Pho and sigma σB-

dependent general stress regulons of Bacillus

subtilis. Microbiology (Reading, England) 148:

1593-1602.

31. Cappiello M, Lazzarotti A, Buono F, Scaloni A,

D'Ambrosio C, Amodeo P, Méndez BL, Pelosi

P, Del Corso A, Mura U (2004) New role for

leucyl aminopeptidase in glutathione turnover.

The Biochemical journal 378: 35-44.

32. Riccillo PM, Muglia CI, Bruijn FJD, Roe AJ,

Booth IR, Aguilar OM, Bruijn FJDE, Booth

IANR (2000) Glutathione Is Involved in

Environmental Stress Responses in Rhizobium

tropici, Including Acid Tolerance Glutathione Is

Involved in Environmental Stress Responses in

Rhizobium tropici, Including Acid Tolerance.

doi: 10.1128/JB.182.6.1748-1753.2000.

33. Wood HG, Barden RE (1977) Biotin enzymes.

Ann. Rev. Bioehem. 46: 385-413.

34. Biarrotte-Sorin S, Hugonnet J-E, Delfosse V,

Mainardi J-L, Gutmann L, Arthur M, Mayer C

(2006) Crystal structure of a novel beta-lactam-

insensitive peptidoglycan transpeptidase.

Journal of molecular biology 359: 533-538.

35. Wen ZT, Yates D, Ahn S-J, Burne Ra (2010)

Biofilm formation and virulence expression by

Streptococcus mutans are altered when grown in

dual-species model. BMC microbiology 10:

111-111.

36. Jakubovics NS, Gill SR, Iobst SE, Vickerman

MM, Kolenbrander PE (2008) Regulation of

gene expression in a mixed-genus community:

stabilized arginine biosynthesis in Streptococcus

gordonii by coaggregation with Actinomyces

naeslundii. Journal of bacteriology 190: 3646-

3657.

37. Niedhardt FC, Umbarger HE (1996) Chemical

composition of Escherichia coli. In: Curtiss, R,

III, Ingraham, JL, Lin, ECC, Low, KB,

Magasanik, B, Regnikoff, WS (eds.). American

Society of Microbiology, Washington D. C., pp.

13-16.

38. Gasch aP, Spellman PT, Kao CM, Carmel-Harel

O, Eisen MB, Storz G, Botstein D, Brown PO

(2000) Genomic expression programs in the

response of yeast cells to environmental

changes. Molecular biology of the cell 11:

4241-4257.

39. Wang X, Wood TK (2011) Toxin-antitoxin

systems influence biofilm and persister cell

formation and the general stress response.

Applied and environmental microbiology 77:

5577-5583.

40. Engelberg-Kulka H, Hazan R, Amitai S (2005)

mazEF: a chromosomal toxin-antitoxin module

that triggers programmed cell death in bacteria.

Journal of cell science 118: 4327-4332.

41. Rice KC, Bayles KW (2008) Molecular control

of bacterial death and lysis. Microbiology and

molecular biology reviews: MMBR 72: 85-109,

table of contents

42. López-Maury L, Marguerat S, Bähler J (2008)

Tuning gene expression to changing

environments: from rapid responses to

evolutionary adaptation. Nature reviews

Genetics 9: 583-593.

43. Xavier JB, Foster KR (2007) Cooperation and

conflict in microbial biofilms. Proceedings of

the National Academy of Sciences of the United

States of America 104: 876-881.

44. Yang L, Liu Y, Wu H, Hóiby N, Molin S, Song

Z-j (2011) Current understanding of multi-

species biofilms. International journal of oral

science 3: 74-81.

45. Harrison JJ, Turner RJ, Ceri H (2005) Persister

cells, the biofilm matrix and tolerance to metal

cations in biofilm and planktonic Pseudomonas

aeruginosa. Environmental Microbiology 7:

981-994.

46. Evans D, Allison D, Brown M, Gilbert P (1991)

Susceptibility of Pseudomonas aeruginosa and

Escherichia coli biofilms towards ciprofloxacin:

effect of specific growth rate. Journal of

Antimicrobial Chemotherapy 27: 177-184.

Supplementary tables and figures

Table S1 Genome assembly statistics

Strain Trim (Mb) N50a Max. contig

length (CL)

Min.

CL

Mean CL contigs Genome

size (Mb)

CDSb

Xanthomonas

retroflexus (Xr)

441.58 107,385 410,416 984 49,439 94 4.6 4,205

Stenotrophomonas

rhizophila (Sr)

210.48 5351 28,238 1,577 4,957 501 2.5 2,607

Microbacterium

Oxydans (Mo)

531.54 1,156,050 1,214,131 540 307,568 13 4 3,799

Paenibacillus

amylolyticus (Pa)

1,475.01 36,803 168,316 508 18,823 335 6.3 5,584

aN50: 50% of the entire assembly is contained in contigs or scaffolds. bCoding sequences.

Table S2 Metatranscriptome statistics

Sample Raw (Mb)* Trim (Mb)** % Mapped

(Mb)***

% CDS

(Mb)****

%

Xr a# 440.14 345.11 78.41% 337.03 97.66% 3.88 1.15%

Xr b 551.73 432.69 78.42% 418.96 96.83% 4.20 1.00%

Xr c 641.36 512.55 79.92% 498.34 97.23% 5.34 1.07%

Xr+Sr a 720.76 524.87 72.82% 505.63 96.33% 8.02 1.59%

Xr+Sr b 881.80 654.41 74.21% 631.82 96.55% 15.59 2.47%

Xr+Sr c 805.52 573.65 71.21% 549.14 95.73% 5.03 0.92%

Xr+Mo a 778.30 519.54 66.75% 492.34 94.76% 7.07 1.44%

Xr+Mo b 1,129.83 807.85 71.50% 773.56 95.76% 6.01 0.78%

Xr+Mo c 593.42 413.70 69.71% 395.37 95.57% 3.76 0.95%

Xr+Pa a 1,058.75 797.95 75.37% 712.08 89.24% 34.40 4.83%

Xr+Pa b 998.45 690.20 69.13% 611.14 88.54% 23.82 3.90%

Xr+Pa c 951.04 736.32 77.42% 671.19 91.16% 28.66 4.27%

Xr+Sr+Mo+Pa a 1,753.43 1,392.47 79.41% 1236.95 88.83% 39.89 3.23%

Xr+Sr+Mo+Pa b 1,202.92 969.95 80.63% 847.00 87.32% 24.61 2.91%

Xr+Sr+Mo+Pa c 980.82 785.00 80.03% 724.87 92.34% 21.70 2.99%

Xr+Sr+Mo+Pa d 1,324.05 1,044.41 78.88% 935.45 89.57% 23.82 2.55%

Xr+Sr+Mo+Pa e 895.13 708.44 79.14% 661.58 93.38% 15.33 2.32%

*The number of mega bases from the raw output of the sequencer. **Mega bases remaining after the quality filtering process. ***Mega bases mapped to the genomes. ****Mega bases maped to protein coding sequences.

#3 replicates for Xr, Xr+Sr, Xr+Mo and Xr+Pa biofilms; 5 replicates for Xr+Sr+Mo+Pa biofilm.

Table S3: Pearson correlation between the samples

Xr a Xr b Xr c Xr+Sr

a

Xr+Sr

b

Xr+Sr

c

Xr+Mo

a

Xr+Mo

b

Xr+Mo

c

Xr+Pa

a

Xr+Pa

b

Xr+Pa

c

Xr+Sr

+Mo+Pa a

Xr+Sr

+Mo+Pa b

Xr+Sr

+Mo+Pa c

Xr+Sr

+Mo+Pa d

Xr+Sr

+Mo+Pa e

Xr a# 1.00 0.95 0.95 0.88 0.91 0.87 0.96 0.92 0.97 0.89 0.89 0.89 0.25 0.30 0.24 0.29 0.27

Xr b 0.95 1.00 0.97 0.87 0.85 0.92 0.93 0.93 0.92 0.87 0.87 0.85 0.33 0.38 0.32 0.37 0.35

Xr c 0.95 0.97 1.00 0.92 0.89 0.95 0.93 0.92 0.93 0.84 0.86 0.84 0.32 0.36 0.31 0.36 0.32

Xr+Sr a 0.88 0.87 0.92 1.00 0.95 0.92 0.85 0.91 0.88 0.72 0.74 0.73 0.23 0.28 0.23 0.27 0.25

Xr+Sr b 0.91 0.85 0.89 0.95 1.00 0.85 0.89 0.87 0.93 0.77 0.80 0.77 0.15 0.21 0.15 0.19 0.19

Xr+Sr c 0.87 0.92 0.95 0.92 0.85 1.00 0.87 0.91 0.88 0.80 0.80 0.79 0.42 0.44 0.41 0.46 0.39

Xr+Mo a 0.96 0.93 0.93 0.85 0.89 0.87 1.00 0.91 0.98 0.89 0.90 0.85 0.30 0.34 0.28 0.33 0.31

Xr+Mo b 0.92 0.93 0.92 0.91 0.87 0.91 0.91 1.00 0.90 0.80 0.80 0.75 0.32 0.36 0.31 0.36 0.33

Xr+Mo c 0.97 0.92 0.93 0.88 0.93 0.88 0.98 0.90 1.00 0.90 0.91 0.88 0.26 0.31 0.25 0.30 0.28

Xr+Pa a 0.89 0.87 0.84 0.72 0.77 0.80 0.89 0.80 0.90 1.00 0.98 0.96 0.44 0.46 0.42 0.46 0.42

Xr+Pa b 0.89 0.87 0.86 0.74 0.80 0.80 0.90 0.80 0.91 0.98 1.00 0.95 0.40 0.43 0.39 0.42 0.39

Xr+Pa c 0.89 0.85 0.84 0.73 0.77 0.79 0.85 0.75 0.88 0.96 0.95 1.00 0.43 0.44 0.42 0.45 0.40

Xr+Sr+Mo+Pa a 0.25 0.33 0.32 0.23 0.15 0.42 0.30 0.32 0.26 0.44 0.40 0.43 1.00 0.92 0.98 0.98 0.87

Xr+Sr+Mo+Pa b 0.30 0.38 0.36 0.28 0.21 0.44 0.34 0.36 0.31 0.46 0.43 0.44 0.92 1.00 0.93 0.96 0.98

Xr+Sr+Mo+Pa c 0.24 0.32 0.31 0.23 0.15 0.41 0.28 0.31 0.25 0.42 0.39 0.42 0.98 0.93 1.00 0.98 0.91

Xr+Sr+Mo+Pa d 0.29 0.37 0.36 0.27 0.19 0.46 0.33 0.36 0.30 0.46 0.42 0.45 0.98 0.96 0.98 1.00 0.93

Xr+Sr+Mo+Pa e 0.27 0.35 0.32 0.25 0.19 0.39 0.31 0.33 0.28 0.42 0.39 0.40 0.87 0.98 0.91 0.93 1.00

#3 replicates for Xr, Xr+Sr, Xr+Mo and Xr+Pa biofilms; 5 replicates for Xr+Sr+Mo+Pa biofilm.

Table S4 70 significant differentially expressed genes with a logFC above 3

ID Acc. name e-val Pfams e-val sig-P Xr+Sr Xr+Mo Xr+Pa Xr+Sr

+Mo+

Pa

Xr1 YP_001972537 major facilitator superfamily

transmembrane nitrite extrusion

0.00E+00 PF07690.11 4.7E-13 N -1.33 -0.90 -3.35 -4.11

Xr2 YP_006183023 hypothetical protein SMD_0239

2.00E-151 PF08813.6 5.4E-42 Y -1.05 -1.43 -3.30 -2.81

Xr3 YP_004794138 hypothetical protein 4.00E-99 PF09537.5 4.6E-29 N -0.96 0.04 -3.13 -1.78

Xr4 WP_019661831 hypothetical protein 2.00E-41 PF05532.7 6.1E-20 N -0.95 -0.27 -3.10 -1.26

Xr5 WP_005419358 transport-associated protein 1.00E-66 PF04972.12 5.1E-17 Y -0.19 -0.02 -3.57 -3.14

Xr6 WP_019184261 hypothetical protein 5.00E-110 PF05974.7 1.1E-49 N 0.01 -0.12 -3.28 -2.49

Xr7 YP_001972524 methyl-accepting chemotaxis

protein

3.00E-99 PF13426.1 2.5E-16 N -0.45 -0.63 -8.60 -5.31

Xr8 CCP12806 hypothetical protein SMSKK35_4264

0 PF03734.9, PF01471.13

3.2E-17, 6.6E-10

Y -0.13 -0.07 -3.91 -8.60

Xr9 WP_019659358 pyridoxamine 5'-phosphate

oxidase

5.00E-94 PF01243.15 0.00017 N 0.16 0.33 -4.24 -5.19

Xr10 WP_021201924 membrane protein 7.00E-112 PF08212.7 3.7E-47 N 1.10 0.50 -3.14 -5.82

Xr11 WP_010484413 glyoxalase 1.00E-73 PF12681.2 7.6E-8 N 0.26 -0.24 -3.06 -5.25

Xr12 YP_002028224 AbrB family transcriptional regulator

2.00E-39 PF04014.13 5.0E-9 N 2.62 0.10 6.83 7.47

Xr13 WP_021203640 azurin 2.00E-100 PF00127.15 1.1E-19 Y -0.13 3.07 4.46 6.72

Xr14 WP_008266809 membrane protein 0.00E+00 PF02321.13 3.9E-35 Y 0.76 1.80 2.63 3.41

Xr15 YP_002028320 cell division protein FtsK 0.00E+00 PF01580.13 1.2E-65 N 1.55 1.91 2.83 3.80

Xr16 YP_001973969 30S ribosomal protein S9 1.00E-86 PF00380.14 6.3E-47 N 1.08 0.76 3.53 3.55

Xr17 YP_006185095 hypothetical protein

SMD_2382

6.00E-64 PF02604.14 2.2E-09 N 1.73 1.45 4.27 3.58

Xr18 YP_002029850 RND family efflux transporter MFP subunit

0.00E+00 PF00529.15 1.6E-59 Y 0.84 2.29 4.76 4.92

Xr19 WP_019662033 TonB-dependent receptor 0.00E+00 PF00593.19 1.6E-24 Y 1.88 2.10 4.21 4.83

Xr20 YP_002026924 leucyl aminopeptidase 0 PF00883.16 2,1E-115 N 2,40 1,40 3,64 4,54

Xr21 YP_002029333 phospholipase

D/transphosphatidylase

0,00E+00 PF13091.1 1,3E-15 N 3,44 3,62 0,44 -0,75

Xr22 WP_006397171 HTH-type transcriptional

regulator betI

4,00E-98 PF00440.18 1,9E-11 N 3,75 3,34 0,27 -0,85

Xr23 YP_004790752 methylcrotonoyl-CoA

carboxylase

0 PF02786.12 6.5E-73,

1.8E-38,

3.4E-15

N -3,85 -1,03 0,43 2,02

Xr24 YP_002027511 integral membrane protein MviN

0,00E+00 PF03023.9 3,2E-122 N -3,96 -0,31 -0,42 2,00

Xr25 YP_006186711 porphobilinogen synthase 0,00E+00 PF00490.16 1,6E-146 N -4,03 -2,32 0,62 2,51

Xr26 YP_004791464 PepSY-associated TM helix domain-containing protein

0,00E+00 PF13703.1 6,5E-19 N -3,17 -1,56 1,36 2,95

Xr27 YP_004791875 TonB-dependent siderophore

receptor

0,00E+00 PF00593.19 4,4E-23 Y 1,10 0,34 3,07 0,30

Xr28 YP_002026409 aromatic amino acid

aminotransferase

0,00E+00 PF00155.16 2,8E-80 N 1,32 -3,09 1,91 2,67

Xr29 WP_019660782 DNA processing protein DprA 0,00E+00 PF02481.10 2,3E-72 N -0,34 -0,82 0,35 3,23

Xr30 WP_006363899 hypothetical protein 1,00E-131 PF09864.4 3,9E-13 Y 0,24 -0,07 0,18 3,28

Xr31 WP_005414275 TolA protein 0,00E+00 PF13103.1 1,6E-10 N 0,04 -2,01 1,08 3,56

Xr32 WP_017354885 membrane protein 9,00E-41 - - N inf -2,22 1,72 3,36

Xr33 WP_005415399 4-diphosphocytidyl-2-C-

methyl-D-erythritol kinase

4,00E-175 PF00288.21 9,5E-10 N -0,68 -1,57 2,15 3,96

Xr34 WP_005418317 methyl-accepting chemotaxis

protein

0 PF00015.16,

PF00672.20

2.6E-55,

7.2E-11

N -1,16 -1,62 1,71 3,24

Xr35 NP_639162 30S ribosomal protein S21 7,00E-41 PF01165.15 2,7E-24 N -0,14 -0,38 3,15 6,35

Xr36 WP_004149880 methionyl aminopeptidase 4,00E-175 PF00557.19 4,3E-47 N inf -0,45 3,93 5,07

Xr37 WP_021203383 aminotransferase 0 PF00733.16 2E-58 N inf 0,58 4,05 4,19

Xr38 WP_005414536 efflux transporter, RND family,

MFP subunit

0,00E+00 PF12700.2 1,7E-30 N -1,50 0,19 3,33 4,99

Xr39 YP_001972041 hypothetical protein Smlt2244 1,00E-84 PF01850.16 0,00000001

3

N -1,83 -0,74 4,28 5,00

Xr40 WP_006402337 cytochrome bd-I oxidase subunit I

0,00E+00 PF01654.12 1,5E-159 N -1,75 -0,85 3,37 4,45

Xr41 CCP09567 Vitamin B12 transporter btuB 0 PF00593.19 2,2E-28 N -0,45 2,08 2,67 4,24

Xr42 YP_002029700 inorganic pyrophosphatase 3,00E-125 PF00719.14 5,4E-59 N 0,62 1,80 3,45 3,84

Xr43 YP_002029231 hypothetical protein Smal_2848 0 - - N 0,65 1,52 2,68 4,55

Xr44 WP_005415097 hypothetical protein, partial 0,00E+00 PF03783.9 1,9E-13 N 0,38 1,51 3,05 3,90

Xr45 WP_019661298 cytidylate kinase 2,00E-152 PF02224.13 1,9E-52 N 0,21 1,32 2,77 3,94

Xr46 YP_001973281 angiotensin-converting peptidyl

dipeptidase protein

0 PF01401.13 1,5E-110 Y 0,37 1,54 1,87 3,02

Xr47 YP_002030341 hypothetical protein Smal_3959 0,00E+00 PF05960.6 1,7E-159 Y 0,56 1,49 1,64 3,12

Xr48 CCP09956 NAD synthetase 0,00E+00 PF02540.12,

PF00795.17

1.7E-69,

4E-16

Y 0,75 1,58 1,72 3,24

Xr49 WP_004144913 branched-chain amino acid

aminotransferase

0,00E+00 - - N 0,15 0,43 1,19 3,52

Xr50 YP_006184128 macrophage infectivity potentiator

1,00E-162 PF01346.13 1,4E-29 Y 0,12 0,79 2,18 3,25

Xr51 YP_002029454 glycine cleavage system protein

H

2,00E-83 PF01597.14 2,2E-45 N 0,34 0,87 1,93 3,20

Xr52 WP_010340581 hypothetical protein 4,00E-159 PF00753.22 0,000013 N 0,66 0,91 1,81 3,25

Xr53 WP_017356008 30S ribosomal protein S12 6,00E-83 PF00164.20 1E-53 N 0,86 0,55 2,10 4,81

Xr54 WP_004142671 phosphoribosylformylglycinamidine synthase

0,00E+00 PF13507.1, PF02769.17

5.7E-100, 5.1E-28

N 1,14 0,25 2,01 3,12

Xr55 YP_002029200 hypothetical protein Smal_2817 8,00E-132 PF02576.12 1,8E-41 N 0,34 0,07 2,78 3,67

Xr56 YP_006186679 biotin synthase 0,00E+00 PF06968.8,

PF04055.16

1.5E-28,

2E-15

N -3,33 0,68 2,93 3,05

Xr57 YP_003017703 sulfatase 0,00E+00 PF00884.18 1,2E-47 N -2,51 -0,96 2,62 3,02

Xr58 YP_002028993 ArsR family transcriptional

regulator

0,00E+00 PF08241.7,

PF01022.15

1.1E-22,

2.5E-14

N -1,89 -0,28 1,81 5,24

Xr59 CCP13965 3-hydroxydecanoyl-(acyl

carrier protein) dehydratase

3,00E-120 PF07977.8 1,8E-47 N -2,32 -0,17 2,27 5,04

Xr60 YP_001970648 D-alanine--D-alanine ligase 0,00E+00 PF07478.8 3,5E-53 N -1,63 0,66 1,93 3,79

Xr61 WP_010481806 alkyl hydroperoxide reductase

subunit F

0,00E+00 PF07992.9,

PF01063.14, PF13192.1

6.7E-32,

5.3E-29, 6.9E-08

N -0,42 1,02 1,48 3,27

Xr62 YP_001970812 30S ribosomal protein S4 3,00E-150 PF00163.14 4,2E-25 N -0,88 0,57 1,79 3,28

Xr63 YP_004793386 NusA antitermination factor 0,00E+00 PF08529.6,

PF13184.1,

PF14520.1, PF00575.18

2.4E-41,

5.5E-27,

9.5E-10, 2.4E-09

N -0,27 0,49 1,75 3,42

Xr64 YP_002029828 serine-type D-Ala-D-Ala

carboxypeptidase

0,00E+00 PF00768.15,

PF07943.8

2.6E-75,

9.2E-24

N -0,33 0,40 2,04 3,18

Xr65 YP_002026868 DEAD/DEAH box helicase 0 PF00270.24 2,4E-46 N -0,86 0,43 3,43 3,07

Xr66 WP_006450325 multidrug transporter 1,00E-68 PF00893.14 9,3E-21 N -0,66 -0,03 2,96 3,52

Xr67 WP_005408347 TonB-denpendent receptor 0,00E+00 PF00593.19 1,8E-27 Y -1,19 -0,91 2,12 3,60

Xr68 YP_002029000 FKBP-type peptidylprolyl isomerase

0,00E+00 PF01346.13 8,3E-26 Y -1,12 -0,44 2,43 3,53

Xr69 WP_005408268 trigger factor 0,00E+00 PF05697.8,

PF00254.23

2.6E-38,

1.2E-04

N -0,76 -0,47 1,63 3,47

Xr70 YP_002027450 ArsR family transcriptional

regulator

2,00E-65 PF12840.2 2,7E-19 N -0,50 0,10 1,65 3,95

Manuscript 4

Effects of grazing by flagellate Neocercomonas jutlandica on mono- and multi-

species biofilms

Effects of grazing by flagellate Neocercomonas jutlandica on mono- and multi-

species biofilms

Dawei Rena, Flemming Ekelund

b, Søren J. Sørensen

a,*, Mette Burmølle

a,*

aSection of Microbiology, Department of Biology, University of Copenhagen

bSection of Terrestrial Ecology, Department of Biology, University of Copenhagen

*Corresponding authors

Postal address: Universitetsparken 15, Bygn. 1, 2100 København Ø, Denmark

e-mail Mette Burmølle: burmolle@bio.ku.dk e-mail Søren J. Sørensen: sjs@bio.ku.dk

Key words: Biofilm / protozoan grazing / synergism / predator-prey interaction

Abstract

Protozoan grazing is considered to be a major regulating factor for bacterial populations in

agricultural soil. However, many soil inhabiting bacteria form biofilms and little is known about the

impact of predation as well as predator-bacterial prey interactions on biofilm dynamics. Here, we

used species specific quantitative PCR, to test effects of the heterotrophic flagellate Neocercomonas

jutlandica on biofilm formation. We worked with both a monospecies biofilm formed by

Xanthomonas retroflexus and a multispecies biofilm formed by soil isolates Stenotrophomonas

rhizophila, Xanthomonas retroflexus, Microbacterium oxydans and Paenibacillus amylolyticus. The

presence of N. jutlandica co-cultured with X. retroflexus increased the abundances of X. retroflexus

in both mono- and multi-species biofilms. While, N. jutlandica co-cultured with the mixture of four

bacteria could reduce the abundance of each species significantly in the four-species biofilm. These

conflicting observations indicate that the performance of protozoan grazing greatly relies on

predator-prey interactions, which probably induces protozoa with different grazing abilities.

Additionally, the synergistic interactions in this multispecies biofilm did not afford more protection

against predation compared with X. retroflexus biofilm. The constant ratio of cell numbers among X.

retroflexus, M. oxydans and P. amylolyticus was observed regardless of protozoan grazing,

suggesting these three species were spatially arranged in integrated communities rather than

individual microcolonies. However, these conclusions are based on the assumption that this

flagellate predator prefers surface attached cells. The experimental design needs to be optimized

and further studies should be conducted to verify these preliminary results. Furthermore, by

incorporating other types of protozoa, a more comprehensive assessment of protozoa-biofilm

interactions can be achieved.

Introduction

Biofilms, i.e. microbial aggregates associated

with surfaces, are ubiquitous in nature [1].

Typically, they contain multiple species of

bacteria, as well as archaea, fungi, algae and

protozoa. Protozoa occur ubiquitously and are

an important regulating factor for bacteria

growth and survival in nature [2]. The

dominant groups of protozoa in soil are

normally naked amoebae and heterotrophic

flagellates (HF) [3]. Heterotrophic flagellates

which utilize long flagella for propulsion are

the pioneer colonizer of the surfaces and

bacterial biofilms due to their high mobility and

abundance [4]. In agricultural soil,

heterotrophic flagellates are able to graze

bacteria in small pores because of their small

size and flexible cells and thus, play an

important role in the control of bacterial

communities as well as mineralization and

nutrient cycling by releasing the nitrogen taken

by bacteria [2].

The evidences that protozoan grazing could

shape biofilm morphology and spatial

arrangement of bacterial cells, such as

stimulating microcolonies formation, reducing

maximal and basal layer thickness and altering

mass transfer of nutrients, have been widely

reported [5, 6]. A generally held opinion is that

by growing in biofilms, bacteria can be offered

a protective niche against protozoan grazing,

owing to the physical protection by secreted

polymeric matrix [7] and enhanced

coordination by horizontal gene transfer and

quorum sensing [6, 8]. However, the counter

argument is also presented by the reports that

some protozoa showed marked preferences for

attached and aggregated bacteria, such as the

naked amoebae and many of the common

heterotrophic flagellates, as demonstrated for

Rhyncomonas nasuta and species of the genus

Bodo [9, 10]. Moreover, some pathogens, such

as Legionella, could multiply and become more

virulent and resistant after ingested by protozoa

in biofilms [11, 12], indicating the crucial role

of protozoa and biofilms in the evolution of

virulence factors in pathogenic bacteria.

Despite the feeding interactions between

protozoa and planktonic bacteria are well

understood [13-15], only a handful of studies

have attempted to assess the grazing impact on

biofilms, especially multispecies biofilms,

where interspecies interactions play an

important role in determining the structure,

function and dynamics of biofilms and

probably are involved in the defense

mechanism of bacterial biofilms against

protozoan grazing [16, 17].

Here, we combine species specific quantitative

PCR and a multispecies biofilm model, which

consists of four soil isolates and shows

synergism in biofilm formation. Then, the

effects of grazing by a flagellate on the

population dynamics of this multispecies

biofilm and a single-species biofilm were

quantified and compared, which could provide

valuable information about the impact of

interspecies interactions on the biofilm

resilience to grazing pressure in soil.

Materials and methods

Organisms and growth conditions

The bacterial strains Stenotrophomonas

rhizophila (JQ890538), Xanthomonas

retroflexus (JQ890537), Microbacterium

oxydans (JQ890539), and Paenibacillus

amylolyticus (JQ890540) were isolated from

agricultural soil as described previously [18].

Bacteria were routinely grown from frozen

glycerol stocks on Tryptic Soy Agar plates for

48h at 24°C. Thereafter, colonies were

transferred into full strength TSB (Tryptic Soy

Broth), incubated with shaking (250 rpm) at

24°C overnight for biofilm growth assay, or

into one-tenth strength TSB, incubated with

shaking (250 rpm) at 24°C for 48 h and then

used as bacterial prey for protozoan predators.

Neocercomonas jutlandica was originally

isolated from soil sampled at Research Center

Foulum (Jutland, Denmark), where it was the

most abundant flagellate species occurring in

microtiter plates used for enumeration of soil

protozoa [19]. Since then it has been kept in the

culture collection at Section for Terrestrial

Ecology, and has been used in several

experiments. Its growth kinetics was

determined in [20], and its susceptibility to the

fungicide propiconazole was assessed in [21]

where it was named Cercomonas crassicauda.

Ekelund F et al. [22] sequenced it and analysed

it phylogenetically, and hence gave it the name

Neocercomonas jutlandica. Finally, Pedersen

AL et al. [23, 24] showed that N. jutlandica

was among the more tolerant protozoa when

exposed to potentially toxic Pseudomonas spp.

We notice that N. jutlandica is synonymous

with Cercomonas jutlandica [25].

The derived protozoan cultures growing on

specific strains were prepared as follows. The

pure bacterial 48 h-cultures and the equal-

volume mixed culture of these four strains were

diluted 20-fold in weak phosphate buffer

(Neff ’s modified amoeba saline) [26], then the

stepwise dilution technique [27, 28] was

performed to provide protozoan cultures on

each strain separately and on the mixed bacteria.

In short, 100 μL protozoan cultures were

repeatedly transferred to 10 mL bacterial

cultures in cell culture flasks (Nunc A/S,

Roskilde, Denmark, # 156367, 25 cm3)

produced as described above and incubated in

darkness, at 9 °C, until late exponential phase.

More reproducible results could be obtained

from this approach than using a fixed cell

number as food source for protozoa when

making a standard comparison between cultures

according to Pedersen AL et al. [24].

Flagellate growth on bacterial cultures

We determined growth rates and peak

abundance of N. jutlandica on the four bacteria

separately and as a mixture (equal-volume

mixed) in 96-well microtiter plates (Cat. No.

655 180, Greiner Bio-One, Germany). For each

of the four bacteria and a mixture, 125 μL of

the 20-fold bacterial dilution and 25 μL

protozoan cultures (prepared as mentioned

above) were mixed and inoculated into each

well. Each particular combination of bacteria

and protozoa was set up in four replicates. The

wells containing only 150 μL bacteria were

blank controls. The batch cultures were then

left at 9 °C in darkness.

We used an inverted microscope (Olympus

CK30, 200 × magnification, phase contrast) to

quantify the protozoa. For each of the five

bacterial treatments, we counted the protozoa in

the four replicate wells every day from day 0 to

day 8, and on days 10, 12, 14, 16, 19 and 21,

where the cultures had approached the

stationary phase. At each counting, protozoa

spread five microscopic fields were counted in

each well. The protozoan numbers in each well

were calculated and plotted against times (day)

in SigmaPlot 12.5. The nonlinear regression

analyses were performed to fit these points to a

Sigmoidal, 3 Parameter equation f = a/(1+exp(-

(x-x0)/b)), where a, b and x0 are constants that

are determined by the nonlinear regression. The

intrinsic growth rate (r) defined as slope of the

sigmoid curve in the point of inflexion (r =

a/4b), and peak abundance (p) defined as the

limit of f when t →∞ (p = a) were calculated

as previous reported [29].

Quantitative PCR detection of the effects of

flagellate on both mono- and multi-species

biofilms

The biofilm formation assay was conducted in

96-well microtiter plates (Cat. No. 655 180,

Greiner Bio-One, Germany) as described

previously [30]. As displayed in Figure 1, two

experiments were conducted for assessing the

effects of the flagellate N. jutlandica on biofilm

development. Since X. retroflexus was

categorized as the only good biofilm former

among these four soil isolates in a previous

study [30], flagellate grazing on the single-

species biofilm formed by this strain was

performed to compare the abilities of flagellate

N. jutlandica to grazing monospecies and

multispecies biofilms.

(a)

(b)

Figure 1 Experimental setup for assessing the grazing effects of flagellate N. jutlandica on both single-

species (a) and multispecies biofilms (b). In short, 24-h biofilms in the wells were introduced with bacteria-

protozoan cultures or only bacterial cultures, followed by removal of planktonic cells and DNA extraction

X. retroflexus biofilms

X. retroflexus culture

after day-1, 2 and 3

X. retroflexus + derived protozoa

after day 1, 2 and 3

DNA extraction from biofilms

qPCR to determine cell numbers of X.

retroflexus

24 h

Four-speices biofilms

X. retroflexus culture

X. retroflexus + derived protozoa

after day 1, 2 and 3 DNA extraction from biofilms

qPCR to determine cell numbers of X.

retroflexus

four-speices bacterial culture

after day-3

bacterial culture + derived protozoa

after day 3 DNA extraction from biofilms

qPCR to determine cell numbers of four strains

24 h

from biofilms after 1, 2 and 3 days/3 days incubation. Thereafter, SYBR Green qPCRs were performed to

measure the cell numbers of only strain X. retroflexus or all the four strains.

In short, overnight bacterial cultures were

subcultured, grown to exponential phase in

TSB and then adjusted to give an OD600 of

0.15. Aliquots of 150 μL of diluted X.

retroflexus culture were then added to each well

to form single-species biofilms. For

multispecies biofilm study, equal volumes of

the four isolate cultures were thoroughly mixed

and 150 μL was added to each well. After 24 h

incubation, the biofilms were rinsed three times

with weak phosphate buffer to remove

planktonic cells and hereafter, the bacteria-

flagellate cultures prepared as in flagellate

growth experiment, that is 125 μL of 20-fold

diluted bacterial cultures and 25 μL protozoan

cultures in the flasks, was added to each well.

To the wells with pre-established single-species

biofilms, only single-species bacterial culture

and derived protozoan culture were added. To

the wells containing multispecies biofilms,

single-species or four-species mixed bacterial

cultures and respectively derived protozoan

cultures were added. In parallel, 150 μL of

single-species or mixed-species bacterial

cultures without protozoa was also inoculated

into some wells as controls. For all the

experiments investigating the effects of

protozoa feeding on biofilms, the 5-day-old

protozoan cultures in the flasks at 9 °C were

used to ensure that N. jutlandica was in the

exponential growth phase when inoculated in

the wells. Three replicate wells were prepared

for each treatment. The plates were then sealed

with Parafilm and incubated with shaking (100

rpm) at 24°C for 3 days.

The planktonic cells were removed after 1, 2

and 3 days/3 days, followed by biofilms were

scratched with 200 μL plastic pipette tips in

150 μL PBS. The wells were then stained with

crystal violet to verify that all the biofilm cells

were detached. Multispecies biofilm samples

before exposed to protozoan grazing were also

collected using the same procedure. Bacterial

DNA was extracted using FastDNA™ SPIN

Kit for soil (Qbiogene, Illkirch, France)

according to the method described in [30].

Three replicates of DNA samples extracted

from biofilms incubated with/without protozoa

were quantified by SYBR Green qPCR using

standard curves generated by serial 10-fold

dilutions of plasmid DNAs [30]. The cell

numbers of only strain X. retroflexus or all the

four strains were measured using species

specific primers and thermal profile setup

reported in [30]. All samples were run in

duplicate and a no template control was

included in every run.

Statistical analysis

One way ANOVA test (SPSS version 17.0 for

Windows) was conducted to evaluate the

effects of the flagellate predation on population

dynamics in both single-species and

multispecies biofilms. P values < 0.05 were

reported as statistically significant.

Results

N. jutlandica growth curves

The growth of N. jutlandica followed a similar

pattern on all the bacterial cultures, with an

exponential phase followed by a stationary

phase (Figure 2), which demonstrates that all

four strains are suitable as food for this

protozoan. Non-linear regression was used to

describe the growth pattern instead of simple

linear regression that is only applicable to

exponential growth phase. The data points

fitted very well to the sigmoid model, with Rsqr

values in the range of 0.9464 to 0.9786. Except

for P. amylolyticus, N. jutlandica growing on

the other bacterial cultures all could reach

stationary phase after 10 days. The flagellate

growth rates and peak abundances were

calculated and shown in Table 1. When feeding

on the four pure cultures, N. jutlandica in co-

culture with S. rhizophila yielded the highest

growth rates and cell numbers, indicating the

feeding preference of N. jutlandica for S.

rhizophila over the other three species.

However, when preying on the mixed bacterial

culture, N. jutlandica showed more rapid

growth even though the peak abundance was

compromised.S. rhizophila

Time (day)

0 5 10 15 20 25

Pro

tozo

an n

umber

s (log

10)

2.0

2.5

3.0

3.5

4.0

4.5

X. retroflexus

Time (day)

0 5 10 15 20 25

Pro

tozo

an n

umber

s (log

10)

2.0

2.5

3.0

3.5

4.0

4.5

M. oxydans

Time (day)

0 5 10 15 20 25

Pro

tozo

an n

umber

s (log 1

0)

2.0

2.5

3.0

3.5

4.0

4.5

P. amylolyticus

Time (day)

0 5 10 15 20 25

Pro

tozo

an n

umber

s (log 1

0)

2.0

2.5

3.0

3.5

4.0

4.5

Mixed strains

Time (day)

0 5 10 15 20 25

Pro

tozo

an n

umber

s (log

10)

2.0

2.5

3.0

3.5

4.0

4.5

Figure 2 N. jutlandica growth curves for the four

replicate batch cultures. The points indicate the

actual numbers of flagellate plotted against time of

day. The solid line represents the simple sigmoid

curve when data were fitted to the equation f =

a/(1+exp(-(x-x0)/b)).

Table 1 Growth rates (r) and peak abundances (p) of N. jutlandica consuming different planktonic bacteria.

S. rhizophila X. retroflexus M. oxydans P. amylolyticus Mixed strains

Growth rate

(day)a

0.96 1.17 0.97 1.70 0.89

Peak abundance

(cell number)b

1.74E+04 1.15E+04 1.16E+04 8.69E+03 1.41E+04

a Growth rate (r) is presented as the days it takes the protozoan to double in size (cell number).

b Peak abundance (p) is presented as the cell numbers of protozoan when t →∞, that is p = a.

Impact of N. jutlandica grazing on biofilms

under static condition

R-Square (RSq) values, based on threshold

cycles of the standard curves, ranged from

0.988 to 0.998 and application efficiencies (E)

ranged from 83.5-100.0%. The log 10 of

bacterial cell numbers were presented in Figure

3 and used in significance analysis.

As can be seen in Figure 3 (a and b), the results

obtained by qPCR showed that the presence of

flagellate N. jutlandica, which was co-cultured

with strain X. retroflexus, could increase the

cell numbers of this strain in both single- and

multi-species biofilms. The significant

enhancement was observed after 1 (P = 0.003)

and 2 (P = 0.008) days of incubation with N.

jutlandica in single-species biofilms, while this

difference gradually narrowed until day 3 (P =

0.787). In multispecies biofilms, this increase in

cell numbers of strain X. retroflexus was not

significantly different during the three days

between grazing and non-grazing biofilms (P >

0.6).

However, this bacterial growth-promoting

effect of N. jutlandica had changed (Fig. 3c)

when this flagellate was co-cultured with four-

species bacterial culture and then introduced

into multispecies biofilms, as evidenced by the

reductions in cell numbers of all the four strains

after 3 days’ predation. Moreover, these

reductions were significant for S. rhizophila (P

= 0.017), M. oxydans (P = 0.03) and P.

amylolyticus (P = 0.036), while a marginally

significant difference was revealed for X.

retroflexus (P = 0.06). These contrary

observations suggest that the grazing impact of

the flagellate N. jutlandica also depends on

their bacterial prey that co-cultured with before

inoculated into pre-established biofilms. It is

worth noting that the ratios of cell numbers of

the three strains X. retroflexus, M. oxydans and

P. amylolyticus remained almost constant (180:

24: 1) regardless of grazing or non-grazing

treatments, indicating the possible existence of

these three species as mixed microcolonies

within the multispecies biofilm. A greater

reduction was seen in the abundance of S.

rhizophila cells under predation pressure by N.

jutlandica, which is consistent with the prey

preference of this flagellate over strain S.

rhizophila derived from planktonic cultures. As

shown in Figure 3c, even without protozoan

grazing, the majority of bacterial cells still

dispersed from this multispecies biofilm after 3

days (P < 0.007) probably due to the low-

nutrient availability.

We notice that the added protozoa

unfortunately, apparently contained a

contaminant, which we determined to be

Pseudomonas sp. from its sequence of 16S

rRNA gene. Such contaminations are very

difficult to avoid when working with protozoa.

It never made up more than 20% of the

bacterial numbers; thus we do not consider it to

have had significant impact on the outcome of

the experiments.

(a) (b)

Time

day 1 day 2 day 3

Ce

ll n

um

be

rs o

f X

. re

tro

fle

xu

s (

log

10)

0

2

4

6

8

10

Without protozoan

With protozoan

(c)

Strains

S. rhizophila X. retroflexus M. oxydans P. amylolyticus

Ce

ll n

um

be

rs (

log

10)

0

2

4

6

8

10

0 day w ithout protozoan

3 day w ithout protozoan

3 day w ith protozoan

Figure 3 Effects of grazing by N. jutlandica upon the population dynamics of both single-species (a) and

multispecies biofilms (b and c). (a) X. retroflexus culture with/without protozoan feeding on X. retroflexus

was introduced into wells with pre-established X. retroflexus biofilm. Absolute cell numbers from biofilms

were measured using qPCR after 1, 2 and 3 days of treatment. (b) X. retroflexus culture with/without

protozoan feeding on X. retroflexus was introduced into wells with pre-established four-species biofilm and

X. retroflexus cell numbers from biofilms were measured using qPCR after 1, 2 and 3 days of treatment. (c)

Four-species culture with/without protozoan feeding on mixed strains was introduced into wells with pre-

established four-species biofilm. Absolute cell numbers of each species were measured using qPCR only

after 3 days of treatment. Bars represent means ± standard deviation for three replicates.

Time

day 1 day 2 day 3

Ce

ll n

um

be

r o

f X

. re

tro

fle

xu

s (

log

10)

0

2

4

6

8

Without protozoan

With protozoan

Discussion

Bacterial biofilms and protozoa are prevalent in

natural environments. Protozoa shape the

microbial communities, influence nutrient

availability and act as reservisors and vectors of

pathogenic bacteria [7, 31, 32]. Although, the

underlying mechanisms have been partly

elucidated from limited researches focusing on

single-species models [6, 33, 34], there is still a

deficiency of studies regarding the shaping

impact of protozoa on mixed-species biofilms,

which represent the predominate lifestyle in

most ecosystems. In a previous study, a strong

synergy was observed in a biofilm composed of

four soil isolates, which could contribute to

three-fold more biofilm biomass than the

single-species biofilms. Moreover, by

employing quantitative PCR, interspecific

cooperation and the dominance of X.

retroflexus in this microbial community had

been further demonstrated [30], indicating the

usefulness of qPCR assay in monitoring the

population dynamics. Here we combined this

powerful tool with the developed multispecies

biofilm model to evaluate the role of synergistic

interactions played in the impact of protozoan

grazing on the mixed-species biofilms.

Bacterial features, including cell size, surface

properties, speed of movement and ability of

biofilm formation [14, 35-37] could affect their

vulnerability towards grazers. Additionally,

some gram-positive bacteria showed lower

edibility than gram-negative bacteria due to

their thick and rigid cell walls [38].The larger

size (0.7 to 0.9 by 3.0 to 5.0 μm) compared

with other three species, motility by flagella

and gram status of strain P. amylolyticus [39]

may presumably explain the lower growth rate

and abundance of N. jutlandica when feeding

on this strain grown in planktonic culture.

Although high biofilm-forming capacities were

demonstrated in X. retroflexus and in co-culture

of the four species previously [30], these were

not observed under the oligotrophic condition

in the present study (data not shown), which

could be largely responsible for their

vulnerability towards N. jutlandica. The highest

growth rate of flagellate was found in mixed

culture of these four species which may result

from the enhanced prey density due to the

synergistic growth effect among these bacteria.

This synergistic interaction was observed in

biofilm formation with higher nutrient

concentration previously [30], thus, it requires

further validation as bacterial cultures were

diluted 20-fold in non-nutritious buffer (Neff ’s

modified amoeba saline) in the present study.

In order to maintain the vitality of N. jutlandica,

diluted bacterial cultures were inoculated

together with protozoan into the pre-established

biofilms. Even so, the added bacteria were

unlikely to result in the additional biofilm

formation, as confirmed by their inability to

produce any single- or multi-species biofilm in

the low-nutrient medium with or without the

presence of N. jutlandica (data not shown).

Hence, the changes in biofilm biomass after

addition of protozoa compared with addition of

diluted bacteria cultures only should result

solely from the influence of protozoan grazing

on biofilms.

A recent study has shown that while early-stage

biofilms are mainly regulated by random

attachment, mature biofilms seem to be largely

regulated by interactions within the bacterial

communities and with the abiotic environment,

whereas grazing by heterotrophic flagellates

may only has a detectable but limited effect

[17]. However, in this study, we demonstrate

that biofilm dynamics could be greatly affected

by both the abiotic condition (e.g. nutrient

starvation) and flagellate grazing. Previous

studies from Wey JK et al. [17] have shown

that within semi-natural river biofilms, while

some bacterial species were especially

susceptible to grazing by HF, others could

benefit from the resulting decreased

competition. This was not observed in the

present study, as the cell numbers of all species

in the presence of grazer decreased

significantly (Figure 3c). Interestingly,

although the cell number of strain S. rhizophila

showed reduction to a greater extent, an

approximate 3-fold reduction in cell numbers

was shown in all other three species, indicating

that these three species probably exist as

integrated communities rather than form

individual microcolonies.

The evidences that grazing pressure is

positively correlated with the formation of cell

clusters have been supported by the results

obtained both from monospecies laboratory

biofilms [6, 40] and from natural/semi-natural

multispecies biofilms [32, 41]. This could result

from an active defense mechanism [42] or a

passive process that the movement of HF and

their flagella drive the bacterial cells to the

substratum [17]. On the contrary, the

conclusion that total protection against

predation could not be fully met by growing as

a biofilm was also widely reported, as

exemplified by enhanced sloughing [43] and/or

markedly altered population dynamics [44] in

the presence of grazers. In this study, we aimed

to test the hypothesis that multispecies biofilms

should be more grazing-resistant compared

with monospecies biofilms owing to the

synergisitc interactions between community

members. However, conflicting results were

found. The increased bacterial abundance in the

presence of protozoa co-cultured with X.

retroflexus was reflected in both mono- and

multi-species biofilms. However, when used

the same flagellate but co-cultured with the

mixed bacteria, the significant reductions of

cell numbers were observed for all the four

bacteria. This can be explained by the higher

growth rate of N. jutlandica when feeding on

mixed bacteria compared with feeding on a

pure culture of X. retroflexus. These different

prey preferences suggest that this flagellate

probably has evolved better adaption to the

mixed bacteria which results in higher density

and vitality and thus higher ‘offensiveness’.

When co-cultured only with strain X.

retroflexus, this flagellate might become less

active grazer or the synergistic interactions

between biofilm members could enhance the

resistance of this multispecies biofilms against

protozoan grazing. However, the former

explanation seems more reasonable as in

single-species biofilms absent of interspecies

interactions, the decrease in cell numbers of X.

retroflexus was not observed either.

It is very important to note that we only have

been looking at the surface attached bacterial

cells. The conclusions obtained in this study are

based on the assumption that the chosen

flagellate have a preferencial grazing on surface

attached cells. However if this assumption is

wrong then we may have a situation where

majority of cells in planktonic phase will be

grazed and only surface attached cells will be

protected. In this situation we expect that the

monospecies cultures of the three bacteria

unable to produce biofilm by themselves will

be heavily reduced by grazing and that they

actually benefit greatly by being together in the

four-species consortium where biofilms

formation will provide them protections against

protozoan grazing. In order to test this we need

to conduct further experiments where we also

count bacterial cells in planktonic phase both in

mono cultures +/- protozoan grazing and in

four-species co-cultures. Thus, from the present

results, we can not support or reject this

hypothesis, and further experiments are needed

before it can be concluded which of the above

listed scenarios is supported.

Besides, it is not within the scope of this study

to predict whether the flagellate-prey

interactions obtained here, namely, the effects

of flagellate N. jutlandica on biofilms, could be

extrapolated to other protozoa. In future studies,

incorporation of other protozoa would certainly

be interesting; moreover, experimental

approach and axenizing procedure need

optimization to make this setup a valuable tool

for research on protozoa-biofilm interactions.

References

1. Donlan RM, Costerton JW (2002) Biofilms:

survival mechanisms of clinically relevant

microorganisms. Clinical microbiology reviews

15: 167-193.

2. Ekelund F, Rønn R (1994) Notes on protozoa in

agricultural soil with emphasis on heterotrophic

flagellates and naked amoebae and their ecology.

FEMS Microbiology Reviews 15: 321-353.

3. Ekelund F, Rønn R, Griffiths BS (2001)

Quantitative estimation of flagellate community

structure and diversity in soil samples. Protist

152: 301-314.

4. Arndt H, Schmidt-Denter K, Auer B, Weitere M

(2003) Protozoans and biofilms. Fossil and

recent biofilms 1: 161-179.

5. Böhme A, Risse‐Buhl U, Küsel K (2009)

Protists with different feeding modes change

biofilm morphology. FEMS microbiology

ecology 69: 158-169.

6. Matz C, Bergfeld T, Rice SA, Kjelleberg S

(2004) Microcolonies, quorum sensing and

cytotoxicity determine the survival of

Pseudomonas aeruginosa biofilms exposed to

protozoan grazing. Environmental microbiology

6: 218-226.

7. Joubert L-M, Wolfaardt G, Botha A (2006)

Microbial exopolymers link predator and prey

in a model yeast biofilm system. Microbial

ecology 52: 187-197.

8. Queck SY, Weitere M, Moreno AM, Rice SA,

Kjelleberg S (2006) The role of quorum sensing

mediated developmental traits in the resistance

of Serratia marcescens biofilms against

protozoan grazing. Environmental microbiology

8: 1017-1025.

9. Caron DA (1987) Grazing of attached bacteria

by heterotrophic microflagellates. Microbial

Ecology 13: 203-218.

10. Sibbald MJ, Albright LJ (1988) Aggregated

and free bacteria as food sources for

heterotrophic microflagellates. Applied and

environmental microbiology 54: 613-616.

11. Taylor M, Ross K, Bentham R (2009)

Legionella, protozoa, and biofilms: interactions

within complex microbial systems. Microbial

ecology 58: 538-547.

12. Rowe M, Grant I (2006) Mycobacterium avium

ssp. paratuberculosis and its potential survival

tactics. Letters in applied microbiology 42: 305-

311.

13. Jürgens K, Matz C (2002) Predation as a

shaping force for the phenotypic and genotypic

composition of planktonic bacteria. Antonie van

Leeuwenhoek 81: 413-434.

14. Matz C, Jürgens K (2005) High motility reduces

grazing mortality of planktonic bacteria.

Applied and environmental microbiology 71:

921-929.

15. Roberts EC, Legrand C, Steinke M, Wootton

EC (2011) Mechanisms underlying chemical

interactions between predatory planktonic

protists and their prey. Journal of plankton

research 33: 833-841.

16. Matz C (2011) Competition, Communication,

Cooperation: Molecular Crosstalk in Multi-

species Biofilms. Biofilm Highlights. Springer,

pp. 29-40.

17. Wey JK, Jürgens K, Weitere M (2012) Seasonal

and successional influences on bacterial

community composition exceed that of

protozoan grazing in river biofilms. Applied and

environmental microbiology 78: 2013-2024.

18. de la Cruz‐Perera CI, Ren D, Blanchet M,

Dendooven L, Marsch R, Sørensen SJ,

Burmølle M (2013) The ability of soil bacteria

to receive the conjugative IncP1 plasmid,

pKJK10, is different in a mixed community

compared to single strains. FEMS microbiology

letters 338: 95-100.

19. Rønn R, Ekelund F, Christensen S (1995)

Optimizing soil extract and broth media for

MPN enumeration of naked amoebae and

heterotrophic flagellates in soil. Pedobiologia 39:

10-19.

20. Ekelund F (1996) Growth kinetics of five

common heterotrophic soil flagellates.

European journal of soil biology 32: 15-24.

21. Ekelund F, Westergaard K, Søe D (2000) The

toxicity of the fungicide propiconazole to soil

flagellates. Biology and fertility of soils 31: 70-

77.

22. Ekelund F, Daugbjerg N, Fredslund L (2004)

Phylogeny of Heteromita, Cercomonas and

Thaumatomonas based on SSU rDNA

sequences, including the description of

Neocercomonas jutlandica sp. nov., gen. nov.

European journal of protistology 40: 119-135.

23. Pedersen AL, Ekelund F, Johansen A, Winding

A (2010) Interaction of bacteria-feeding soil

flagellates and Pseudomonas spp. Biology and

Fertility of Soils 46: 151-158.

24. Pedersen AL, Winding A, Altenburger A,

Ekelund F (2011) Protozoan growth rates on

secondary‐metabolite‐producing Pseudomonas

spp. correlate with high‐level protozoan

taxonomy. FEMS Microbiology Letters 316:

16-22.

25. Karpov SA, Bass D, Mylnikov AP, Cavalier-

Smith T (2006) Molecular Phylogeny of

Cercomonadidae and Kinetid Patterns of

Cercomonas and Eocercomonas gen. nov.

(Cercomonadida, Cercozoa). Protist 157: 125-

158.

26. Page FC (1988) A New Key to Freshwater and

Soil Gymnamoebae: With Instructions for

Culture pp. 91-97. Freshwater Biological

Association Cumbria

27. Pelegri S, Christaki U, Dolan J, Rassoulzadegan

F (1999) Particulate and dissolved organic

carbon production by the heterotrophic

nanoflagellate Pteridomonas danica Patterson

and Fenchel. Microbial ecology 37: 276-284.

28. Mohapatra B, Fukami K (2004) Comparison of

the numerical grazing response of two marine

heterotrophic nanoflagellates fed with different

bacteria. Journal of Sea Research 52: 99-107.

29. Koch TA, Ekelund F (2005) Strains of the

Heterotrophic Flagellate Bodo designis from

Different Environments Vary Considerably with

Respect to Salinity Preference and SSU rRNA

Gene Composition. Protist 156: 97-112.

30. Ren D, Madsen JS, de la Cruz-Perera CI,

Bergmark L, Sørensen SJ, Burmølle M (2013)

High-Throughput Screening of Multispecies

Biofilm Formation and Quantitative PCR-Based

Assessment of Individual Species Proportions,

Useful for Exploring Interspecific Bacterial

Interactions. Microbial ecology. doi:

10.1007/s00248-013-0315-z

31. Brown MR, Barker J (1999) Unexplored

reservoirs of pathogenic bacteria: protozoa and

biofilms. Trends in microbiology 7: 46-50.

32. Wey JK, Scherwass A, Norf H, Arndt H,

Weitere M (2008) Effects of protozoan grazing

within river biofilms under semi-natural

conditions. Aquatic Microbial Ecology 52: 283.

33. Dopheide A, Lear G, Stott R, Lewis G (2011)

Preferential feeding by the ciliates Chilodonella

and Tetrahymena spp. and effects of these

protozoa on bacterial biofilm structure and

composition. Applied and environmental

microbiology 77: 4564-4572.

34. Friman V-P, Diggle SP, Buckling A (2013)

Protist predation can favour cooperation within

bacterial species. Biology letters 9: 20130548.

35. Posch T, Jezbera J, Vrba J, Šimek K, Pernthaler

J, Andreatta S, Sonntag B (2001) Size selective

feeding in Cyclidium glaucoma (Ciliophora,

Scuticociliatida) and its effects on bacterial

community structure: a study from a continuous

cultivation system. Microbial ecology 42: 217-

227.

36. Monger BC, Landry MR, Brown SL (1999)

Feeding selection of heterotrophic marine

nanoflagellates based on the surface

hydrophobicity of their picoplankton prey.

Limnology and oceanography 44: 1917-1927.

37. Hahn M, Moore E, Höfle M (2000) Role of

microcolony formation in the protistan grazing

defense of the aquatic bacterium Pseudomonas

sp. MWH1. Microbial ecology 39: 175-185.

38. Rønn R, McCaig AE, Griffiths BS, Prosser JI

(2002) Impact of protozoan grazing on bacterial

community structure in soil microcosms.

Applied and Environmental Microbiology 68:

6094-6105.

39. Shida O, Takagi H, Kadowaki K, Nakamura LK,

Komagata K (1997) Emended Description of

Paenibacillus amylolyticus and Description of

Paenibacillus illinoisensis sp. nov. and

Paenibacillus chibensis sp. nov. International

journal of systematic bacteriology 47: 299-306.

40. Matz C, McDougald D, Moreno AM, Yung PY,

Yildiz FH, Kjelleberg S (2005) Biofilm

formation and phenotypic variation enhance

predation-driven persistence of Vibrio cholerae.

Proceedings of the National Academy of

Sciences of the United States of America 102:

16819-16824.

41. Rychert K, Neu TR (2010) Protozoan impact on

bacterial biofilm formation. Biological Letters

47: 3-10.

42. Matz C, Kjelleberg S (2005) Off the hook–how

bacteria survive protozoan grazing. Trends in

microbiology 13: 302-307.

43. Jackson S, Jones E (1991) Interactions within

biofilms: the disruption of biofilm structure by

protozoa. Kieler Meeresforsch 8: 264-268.

44. Huws S, McBain A, Gilbert P (2005) Protozoan

grazing and its impact upon population

dynamics in biofilm communities. Journal of

applied microbiology 98: 238-244.

Manuscript 5

The ability of soil bacteria to receive the conjugative IncP1 plasmid, pKJK10, is

different in a mixed community compared to single strains

R E S EA RCH L E T T E R

The ability of soil bacteria to receive the conjugative IncP1plasmid, pKJK10, is different in a mixed community compared

to single strains

Claudia I. de la Cruz-Perera1,2, Dawei Ren1, Marine Blanchet1,3, Luc Dendooven2, Rodolfo Marsch2,Søren J. Sørensen1 & Mette Burmølle1

1Department of Biology, University of Copenhagen, Copenhagen, Denmark; 2Laboratory of Soil Ecology, Department of Biotechnology and

Bioengineering, Cinvestav, Mexico City, Mexico; and 3Laboratoire d’Oc�eanographie Microbienne, Observatoire Oc�eanologique de Banyuls,

Banyuls-sur-mer, France

Correspondence: Mette Burmølle, Section

of Microbiology, University of Copenhagen,

Universitetsparken 15, bygn 1,

2100 Copenhagen Ø, Denmark.

Tel.: +45 40220069;

e-mail: burmolle@bio.ku.dk

Received 18 October 2012; accepted 25

October 2012. Final version published online

22 November 2012.

DOI: 10.1111/1574-6968.12036

Editor: Paolina Garbeva

Keywords

horizontal gene transfer; bacterial

community; conjugation; host range;

soil bacteria.

Abstract

Horizontal gene transfer by conjugation is common among bacterial popula-

tions in soil. It is well known that the host range of plasmids depends on

several factors, including the identity of the plasmid host cell. In the present

study, however, we demonstrate that the composition of the recipient commu-

nity is also determining for the dissemination of a conjugative plasmid. We

isolated 15 different bacterial strains from soil and assessed the conjugation

frequencies of the IncP1 plasmid, pKJK10, by flow cytometry, from two differ-

ent donors, Escherichia coli and Pseudomonas putida, to either 15 different bac-

terial strains or to the mixed community composed of all the 15 strains. We

detected transfer of pKJK10 from P. putida to Stenotrophomonas rhizophila in a

diparental mating, but no transfer was observed to the mixed community. In

contrast, for E. coli, transfer was observed only to the mixed community, where

Ochrobactrum rhizosphaerae was identified as the dominating plasmid recipient.

Our results indicate that the presence of a bacterial community impacts the

plasmid permissiveness by affecting the ability of strains to receive the conjuga-

tive plasmid.

Introduction

Horizontal gene transfer (HTG) is a driving force in

bacterial evolution as it allows bacteria to rapidly acquire

complex new traits. Plasmids are one of the key vectors of

HTG, enabling genetic exchange between bacterial cells

across species and domain barriers (Poole, 2009; Boto,

2010), and they very often encode genes that confer adap-

tive traits to their host, such as antibiotic resistance, bio-

degradation pathways and virulence (de la Cruz & Davies,

2000). Transfer of these traits by conjugation requires the

donor and the recipient cells to be in direct contact.

Different abiotic and biotic factors affect the range of

conjugal exchange of genetic material between environ-

mental bacteria, such as nutrient availability, spatial archi-

tecture of the bacterial community, plasmid donor and

recipient relatedness and plasmid host type (van Elsas &

Bailey, 2002; De Gelder et al., 2005; Sørensen et al., 2005;

Seoane et al., 2011). The fraction of the cells in a com-

munity capable of receiving and maintaining conjugative

plasmids is highly dependent on several of these factors

and has been described as the plasmid permissiveness

(Musovic, 2010).

It has been shown that conjugative plasmids express

factors that favor the establishment of planktonic bacteria

in biofilm communities, thereby increasing the chances

for horizontal gene transmission (Ghigo, 2001; Reisner

et al., 2006; Madsen et al., 2012). Complex interspecies

communities facilitate synergistic interactions between

populations, affecting the function, stability and flexibility

of the community (James et al., 1995; Burmølle et al.,

2006).

In the present work, HTG by conjugation between sin-

gle populations and microbial communities isolated from

soil were investigated. The plasmid transfer frequencies

and the identities of the recipients of the plasmid, when

FEMS Microbiol Lett 338 (2013) 95–100 ª 2012 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

MIC

ROBI

OLO

GY

LET

TER

S

hosted by different donors, were compared. The bacterial

population was analyzed based on fluorescence properties

and sorted by flow cytometry (FCM) to detect and quan-

tify the plasmid transfer to the individual isolates and the

mixed community (Muller & Nebe-von-Caron, 2010).

Sequencing of the 16S rRNA gene from sorted transcon-

jugant cells was used to evaluate the host range of the

plasmid when a mixed microbial community was used as

recipient.

Materials and methods

Soil and leaf sampling

Soil samples were collected from an agricultural field in

T�astrup, Denmark, in the late summer of 2009. Soil was

sampled from the 5- to 10-cm layer. The soil water con-

tent upon sampling was 14.2%, and the water holding

capacity (WHC) was 60%. The soil was classified as sandy

loam with pH 7.2. Leaves of baby maize seedlings were

used for bacterial isolation. The seedlings were grown for

2 weeks in T�astrup soil before harvesting.

Bacterial strains, plasmids, and growth media

Escherichia coli CSH26::lacIq and Pseudomonas putida

KT2440::lacIq1, carrying pKJK10, a conjugative, green

fluorescent protein (GFP) tagged IncP1 plasmid, origi-

nally isolated from soil (Sengeløv et al., 2001; Bahl et al.,

2007) were used as donor strains. These strains were cul-

tured in Luria Bertani (LB) broth supplemented with

kanamycin monosulfate (50 mg mL�1); 1.5% (w/v) agar

was added when solid medium was needed. The recipient

strains (see below) were cultured in Tryptic Soy Broth

medium (TSB; 17 g peptone from casein, 3 g peptone

from soymeal, 2.5 g D(+)-Glucose, 5 g NaCl, 2.5 g

K2HPO4 in 1 L distilled water, pH 7.3).

Isolation and identification of recipient strains

from leaves incubated in soil

A 15 mg sample of a baby maize leaf was placed in 5 g

T�astrup soil adjusted to 40% WHC and incubated in tripli-

cate at room temperature for 17 days. After 7, 12, and

17 days, the leaves were picked up from the soil, washed

with PBS (8 g NaCl, 0.2 g KCl, 1.44 g Na2HPO4, and

KH2PO4, adjusted to 1 L distilled water and pH 7.4),

placed in a microfuge tube, added 1 mL PBS and vortexed

for 1 min. DNA was extracted from the cell suspension as

described below. Dilutions to 10�6 were made and 100 lLwere plated in triplicate onto Tryptic Soy Agar (TSA; Dif-

co) 10% supplemented with cycloheximide (50 mg mL�1)

and incubated at 25 °C for 2–5 days. Sixteen colonies from

each triplicate looking phenotypically different were

isolated and purified for DNA extraction. A denaturation

gradient gel electrophoresis (DGGE) analysis was

performed with 16S rRNA gene PCR fragments.

From the total of 48 strains from day 7, 15 morpholog-

ically different strains were selected for the use as recipi-

ents. The strains were grown overnight (ON) in 5 mL

TSB, the DNA was extracted using ‘Genomic Mini for

Universal Genomic DNA Isolation Kit’ (A&A Biotechnol-

ogy) and the 16S rRNA gene sequences were amplified

with primers 27F and 1492R (Lane, 1991) for identifica-

tion. The PCR mixture contained 0.5 lL DNA, 1XPhusion

GC buffer, 0.2 mM dNTP mixture, 1 U Phusion Hot

Start DNA Polymerase (FinnzymesOy, Espoo, Finland)

and 0.5 lM of each primer (TAG Copenhagen A/S,

Denmark). The final volume was adjusted with DNA-free

water to 50 lL. Amplification was as follow: initial dena-

turation at 98 °C for 30 s, followed by 35 cycles at 98 °Cfor 10 s, at 55 °C for 30 s and at 72 °C for 45 s. A final

primer extension reaction was performed at 72 °C for

6 min. The resulting sequence (1480 bp) was compared

with reference sequences by BLAST search (Altschul et al.,

1997) and aligned with them using CLUSTALX 1.7 program

(Thompson et al., 1997). Maximum-likelihood analyses

were performed using PhyML (Guindon & Gascuel,

2003). MODELTEST 3.06 (Posada, 2008) was used to select

appropriate models of sequence evolution by the Akaike

Information Criterion. The confidence at each node was

assessed by 500 bootstrap replicates. Similarities among

sequences were calculated using the MatGAT v.2.01 soft-

ware (Campanella et al., 2003). Taxonomic assignment

was carried out based on the Rosell�o-Mora and Aman

criteria (Rossell�o-Mora & Amann, 2001).

DGGE

The cells from the leaves-PBS solution and from the 48- to

15-strain pools were lysed by bead beating followed by

DNA extraction as specified above. The DNA was used for

a 16S rRNA gene PCR as described above and 1 lL of the

product was used as a template for a new PCR using inter-

nal primers with a GC clamp 341F and 518R (Muyzer

et al., 1993) and a polymerization step at 72 °C for 20 s.

This PCR product was loaded onto the DGGE gel, contain-

ing a denaturation gradient of 30–70% acrylamide, and

an electrophoresis was run in a Dcode system (Biorad) at

60 °C and 70 V for 17 h. The gel was stained with SYBR-

Gold (Invitrogene) in the dark for 45 min.

Filter mating assays

Prior to filter matings, the donor strains were grown in

5 mL LB broth at 250 r.p.m. at 30 °C (P. putida) and

ª 2012 Federation of European Microbiological Societies FEMS Microbiol Lett 338 (2013) 95–100Published by Blackwell Publishing Ltd. All rights reserved

96 C.I. de la Cruz-Perera et al.

37 °C (E. coli) for 18 h. These ON cell cultures were then

diluted 1 : 10 in fresh LB medium and grown under sim-

ilar conditions for three more hours to reach exponential

growth phase (OD600 � 0.6). The cells were then recol-

lected, washed twice, and resuspended in sterile PBS. The

recipient strains were cultured similarly in TSB at 25 °C.The lack of background fluorescence of the donor and

recipient strains was verified in the flow cytometer (see

specifications below) prior to their use in the filter mating

assay.

For the single-strain mating experiments, 10 lL of

donor and recipient, respectively, were spotted onto

0.2 lm nitrocellulose filters in triplicate, mixed, placed on

TSA and R2A (yeast extract 0.5 g, proteose peptone 0.5 g,

casamino acids 0.5 g, glucose 0.5 g, soluble starch 0.5 g,

sodium pyruvate 0.3 g, K2HPO4 0.3 g, MgSO4�7H2O

0.05 g, agar 15 g in 1 L distilled water) plates and incu-

bated at 25 °C for 20 h. The cells were then harvested from

the filter followed by resuspension in 1 mL PBS, and FCM

analysis as specified below. For the microbial community,

we spotted 5 lL of each isolate (OD600 � 0.3–0.7) and

75 lL of donor strain (either P. putida or E. coli, prepared

as described above) onto the filter, incubated and analyzed

by FCM at the same conditions as for the single-strain

matings. Controls with only donors or recipients were

included.

Quantification and sorting by FCM

Flow cytometric enumeration of cells was carried out

with a FACScalibur flow cytometer (Becton Dickinson,

San Jose, CA) equipped with a 15 mW argon laser

(488 nm). The following settings and voltages were used

during analysis: forward scatter = E01, side scatter

(SSC) = 350, and the fluorescent detectors FL1 (530/

30 nm), FL2 (585/42 nm), FL3 (650/30 nm) were set at

510 V. A threshold was set on the SSC, and no compen-

sation was used. All parameters were on logarithmic

mode. Samples were run at the ‘low’ flow rate setting for

1 min.

All the samples were diluted in PBS before flow

enumeration to assure optimal bacterial counts to

2000 events s�1. In part of the sample (100 lL), gfp-

expression was induced by incubation in LB with 1 mM

of isopropyl-b-D-1-thiogalactopyranoside (IPTG, SIGMA)

for 3 h at 30 °C (P. putida) and 37 °C (E. coli) to deter-

mine the number of donor cells (Musovic et al., 2006).

To isolate and identify recipients from the E. coli-com-

munity mating, one subsample of each replicate of the

cell extract was diluted to 1000 events s�1 to flow-sorted

(MoFlo; DAKO) at a flow rate of 400–1000 events s�1,

with an optimal setting of the sheath pressure of ca.

60 psi and drop drive frequency to ca. 95 kHz, using a

70-lm CytoNozzle tip on an enrichment sort option of

single-mode per single drop envelope. Dilutions up to

10�3 were made from approximately 70 000 cells of each

replicate, and 100 lL of each dilution were plated on

TSA plates supplemented with kanamycin, streptomycin

(100 mg mL�1) and tetracycline (20 mg mL�1) and incu-

bated at 25 °C for 2–5 days. Four green colonies of each

replicate were selected for DNA extraction and identified

by sequencing after the amplification of the 16S rRNA

gene as described above.

Data analysis was carried out with the CELLQUEST soft-

ware package. Two polygonal gates were defined in bivar-

iate FL1 vs. FL2 to count for green cells and in bivariate

SSC vs. FL2 density plot as a double check.

Statistical analysis

All microcosmic experiments were carried out in tripli-

cate. Standard deviations were calculated with Excel

(Microsoft®). A Student’s t-test was applied and probabil-

ities less than 0.05 were considered significant.

Results and discussion

Isolation of the bacterial community

Bacterial strains established on the leaves embedded in

soil were isolated to obtain a highly diverse bacterial

community with the capability of attachment. We used

maize leaves as the plants grow fast with no specific

requirements. The diversity of the established community

over time was followed by DGGE analysis. DGGE is a

simple and fast method to screen and compare the diver-

sity of a bacterial community, well suited for this study.

The number of bands in a DGGE lane reflects the degree

of bacterial diversity and lanes from the same gels can be

compared to explore changes in diversity (Muyzer et al.,

1993). Based on the number of bands associated to the

sampling days 7, 12, and 17 (Fig. 1, lane 1, 4, and 7,

respectively), a highly diverse community was observed

from day 7 and onwards. This was confirmed by compar-

ing colony morphology of the 48 isolated strains from the

different sampling points (data not shown). Due to this,

and the fact that the leaves were at this time point highly

decomposed (data not shown), the day 7-samples were

chosen for strain isolation.

To select a manageable, yet still diverse, subcommunity

from all of the isolated strains of the day 7 sampling, the

colony morphology of all the 48 isolates (from the three

replicates) was visually compared. Fifteen isolates appear-

ing morphologically different were chosen, and their

DGGE profile was compared with that of the 48 isolates

(Fig. 1, lane 2, and 3). Based on the number of bands, a

FEMS Microbiol Lett 338 (2013) 95–100 ª 2012 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

Gene transfer in mixed and single species communities 97

low number of strains were lost when the subcommunity

was selected, but most of the initial diversity was repre-

sented in the selected community (lane 3). From the 16S

rRNA gene sequencing analysis (Table 1), the isolates

were identified as typical soil bacteria, mostly gram nega-

tives, with the Pseudomonas and Chryseobacterium genera

as the most abundant. Based on this, the strains of the

selected subcommunity were considered as well-suited

potential recipients of the pKJK10 plasmid.

Transfer of pKJK10 to the individual soil

isolates and the mixed community

The donor strains used in this study encode the lacIq1

repressor gene from the chromosome, repressing GFP

expression from pKJK10 when present in these donor

strains, as the lac promoter regulates GFP expression in

this plasmid. Due to the lack of the lacIq1 repressor in the

soil isolates, GFP will be expressed if the plasmid is trans-

ferred into these cells. This system thus allows enumera-

tion of transconjugants and donors by direct sample

analysis and after IPTG induction, respectively (Sørensen

et al., 2003). Detection by FCM has several advantages in

such approaches because enumeration of transconjugant

cells is based solely on fluorescence markers. There is

therefore no need for only including strains with specific

antibiotic resistance profiles in the recipient community,

and the strains do not need to be capable of expressing

the resistance traits encoded by the plasmid to be charac-

terized as transconjugants.

The transfer frequency of the conjugative plasmid from

the two different donors to the soil isolates was calculated

as a transconjugant/donor ratio. No green cells were

observed in the negative controls with only the recipient

strains present (data not shown), indicating that none of

the soil isolates produced auto-fluorescence and that

green cells represented plasmid transfer events.

When E. coli was used as donor, no transfer of pKJK10

was detected to any of the individual 15 soil isolates, but

P. putida was observed to transfer pKJK10 to Stenotropho-

monas rhizophila. The plasmid transfer frequency from

P. putida to S. rhizophila was higher when the filters were

placed on TSA medium (1.07 � 3.05 9 10�1) compared

with R2A medium (0.33 � 2.32 9 10�2, Table 2), sup-

porting the fact that the metabolic state of the cells may

in some cases influence conjugation frequencies (van

Elsas & Bailey, 2002). These results reflect the fact that

the host range of plasmids depends on the identity of the

donor strain (De Gelder et al., 2005).

Day 7 12 17

Lane 1 2 3 4 5 6 7 8 9

Fig. 1. Denaturing gradient gel electrophoresis (DGGE) analysis of

the diversity of the bacterial communities isolated from maize leaves

after 7, 12, and 17 days of incubation in soil. The DGGE gel shows

the PCR amplified products of the 16S rRNA genes of the total

bacterial consortia present on the leaves on the sampling days 7, 12,

and 17 (lane 1, 4, and 7, respectively). The 16S rRNA gene profiles of

a total of 48 cultured strains from each day are presented in lane 2

(day 7), lane 5 (day 12), and lane 8 (day 17). Of these 48, 15

morphologically different strains were selected to constitute

representative communities (lanes 3, 6, and 9) of sampling days 7,

12, and 17, respectively. The DGGE analysis indicated that most

bacterial diversity was preserved when reducing the community from

48 to 15 strains (by comparing lane 2–3, 5–6, and 8–9). The 15-strain

community from sampling day 7 was used for gene transfer analysis.

Table 1. Identification of the isolated strains by 16S rRNA gene

analysis

Strain name Similarity (%)*

Flavobacterium psychrolimnae 97.8

Pseudomonas lutea 98.1

Pseudomonas brassicacearum 99.6

Pseudomonas fluorescens 99.7

Ochrobactrum rhizosphaerae 100

Chryseobacterium soldanellicol 98.2

Chryseobacterium letacus 98.5

Sphingobacteriaceae 93.7

Xanthomonas retroflexus 99.6

Micrococcaceae 94.2

Chryseobacterium ginsengisoli 99.2

Stenotrophomonas rhizophila 99.6

Microbacterium oxydans 100

Ensifer adherens 98.9

Janthinobacterium lividum 99.7

*Similarity from sequenced 16S rRNA genes calculated with the

MATGAT v.2.01 software (see text for more details).

ª 2012 Federation of European Microbiological Societies FEMS Microbiol Lett 338 (2013) 95–100Published by Blackwell Publishing Ltd. All rights reserved

98 C.I. de la Cruz-Perera et al.

In contrast to the results observed when transferring

pKJK10 to individual isolates, no plasmid transfer events

were observed from P. putida to the mixed community

consisting of the same 15 strains applied individually

above. Transconjugants were, however, obtained when

applying E. coli as donor of pKJK10. The green fluores-

cent transconjugant cells were sorted by FACS and

cultured on TSA agar plates. By sequence analysis of the

16S rRNA gene from four colonies from each replicate,

the selected transconjugants were shown all to be identi-

cal and identified as Ochrobactrum rhizosphaerae. This

does not exclude the possibility that other isolates may

also have received the plasmid, but it does show that

O. rhizosphaerae in fact did so and that it was the most

dominant strain among the plasmid recipients. Interest-

ingly, O. rhizosphaerae was not able to receive the plas-

mid in the individual mating experiment, indicating that

the plasmid permissibility does not only depend on the

abilities of the plasmid, host and recipient strains, but

also on the surrounding microbial community, which

may reduce or enhance plasmid transfer. Both of these

scenarios were observed in this study; transfer of pKJK10

from P. putida to S. rhizophila was observed in diparental

mating experiments, but not in a mixed community, pos-

sibly caused by reduced survival/competition ability of

the strains or by the fact that the donor and this specific

recipient populations had less opportunity for interaction

in the mixed community. In contrast, the presence of a

mixed community induced pKJK10 transfer from E. coli

to O. rhizosphaerae, which may be due to altered physical

cell–cell interaction or the presence of one or several

intermediate plasmid host(s). These ‘plasmid step-stones’

may facilitate plasmid transfer from E. coli to O. rhizosp-

haerae, but are unable to establish and stabilize the plas-

mid in their own population. Because it was not possible

to isolate the strains individually after growth in the com-

munity, the fraction of O. rhizosphaerae herein could not

be determined; It is possible that O. rhizosphaerae is the

dominating strain in the consortium or the most meta-

bolically active, explaining its enhanced abilities as plas-

mid recipient. Regardless of this strain being dominant or

representing a minor population of the community, it is

still intriguing that no plasmid transfer was observed in

the dual-strains mating from E. coli to O. rhizosphaerae.

The results of this study indicate that the surrounding

bacterial community strongly impacts the plasmid host

range, which needs to be considered when analyzing

potential plasmid dissemination in natural environments

in association to risk assessment. Plasmid mediated traits,

including antibiotic resistance and virulence, may spread

to natural bacterial populations in situ, in spite of

an apparent narrow host range detected in simple,

dual-strain-mating experiments.

Acknowledgements

This research was supported by funding to Søren Sørensen

by The Danish Council for Independent Research (Natural

Sciences), The Danish Council for Independent Research

(Technology and Production) (ref no: 09-090701, Mette

Burmølle) and the Department of Biotechnology and

Bioengineering (Cinvestav, Mexico). Claudia I. de La

Cruz-Perera received grant-aided support from ‘Consejo-

Nacional de Ciencia y Tecnologia’ (CONACyT, Mexico)

scholarship 166878.

References

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z,

Miller W & Lipman DJ (1997) Gapped BLAST and

PSI-BLAST: a new generation of protein database search

programs. Nucleic Acids Res 25: 3389–3402.Bahl MI, Hansen LH, Goesmann A & Sørensen SJ (2007) The

multiple antibiotic resistance IncP-1 plasmid pKJK5 isolated

from a soil environment is phylogenetically divergent from

members of the previously established alpha, beta and delta

sub-groups. Plasmid 58: 31–43.Boto L (2010) Horizontal gene transfer in evolution: facts and

challenges. Proc Biol Sci 277: 819–827.Burmølle M, Webb JS, Rao D, Hansen LH, Sørensen SJ &

Kjelleberg S (2006) Enhanced biofilm formation and

increased resistance towards antimicrobial agents and

bacterial invasion are caused by synergistic interactions in

multispecies biofilms. Appl Environ Microbiol 72:

3916–3923.Campanella JJ, Bitincka L & Smalley J (2003) MatGAT: an

application that generates similarity/identity matrices using

protein or DNA sequences. BMC Bioinformatics 4: 29.

De Gelder L, Vandecasteele FP, Brown CJ, Forney LJ & Top

EM (2005) Plasmid donor affects host range of promiscuous

IncP-1beta plasmid pB10 in an activated-sludge microbial

community. Appl Environ Microbiol 71: 5309–5317.

Table 2. Transfer frequencies of pKJK10 by different donors to

individual recipients and mixed communities*

Donor Recipient Frequency

E. coli CSH26::lacIq S. rhizophila –†

Mixed community 1.56 � 3.43 9 10�1

P. putida KT2440::lacIq S. rhizophila 1.07 � 3.05 9 10�1

0.33 � 2.32 9 10�2‡

Mixed community –†

*All values are transconjugants/donor ratios � standard deviation of

triplicate experiments.†No transconjugants were obtained on the filter (below detection

limit).‡Number of transconjugants obtained when placing the filters on R2A

plates.

FEMS Microbiol Lett 338 (2013) 95–100 ª 2012 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

Gene transfer in mixed and single species communities 99

de la Cruz F & Davies J (2000) Horizontal gene transfer and

the origin of species: lessons from bacteria. Trends Microbiol

8: 128–133.Ghigo JM (2001) Natural conjugative plasmids induce

bacterial biofilm development. Nature 412: 442–445.Guindon S & Gascuel O (2003) A simple, fast, and accurate

algorithm to estimate large phylogenies by maximum

likelihood. Syst Biol 52: 696–704.James GA, Beaudette L & Costerton JW (1995) Interspecies

bacterial interactions in biofilms. J Ind Microbiol 15:

257–262.Lane DJ (1991) 16S/23S rRNA sequencing. Nucleic Acid

Techniques in Bacterial Systematics (Stackebrandt E &

Goodfellow M, eds.), pp. 115–174. John Wiley & Sons, Inc.,

New York.

Madsen JS, Burmølle M, Hansen LH & Sørensen SJ (2012)

The interconnection between biofilm formation and

horizontal gene transfer. FEMS Immunol Med Microbiol 65:

183–195.Muller S & Nebe-von-Caron G (2010) Functional single-cell

analyses: flow cytometry and cell sorting of microbial

populations and communities. FEMS Microbiol Rev 34:

554–587.Musovic S, Oregaard G, Kroer N & Sørensen SJ (2006)

Cultivation-independent examination of horizontal transfer

and host range of an IncP-1 plasmid among Gram-positive

and Gram-negative bacteria indigenous to the barley

rhizosphere. Appl Environ Microbiol 72: 6687–6692.Musovic S, Dechesne A, Sørensen J & Smets BF (2010) Novel

assay to assess permissiveness of a soil microbial community

toward receipt of mobile genetic elements. Appl Environ

Microbiol 76: 4813–4818.

Muyzer G, de Waal EC & Uitterlinden AG (1993) Profiling of

complex microbial populations by denaturing gradient gel

electrophoresis analysis of polymerase chain reaction-

amplified genes coding for 16S rRNA. Appl Environ

Microbiol 59: 695–700.Poole AM (2009) Horizontal gene transfer and the earliest

stages of the evolution of life. Res Microbiol 160: 473–480.Posada D (2008) jModelTest: phylogenetic model averaging.

Mol Biol Evol 25: 1253–1256.Reisner A, Holler BM, Molin S & Zechner EL (2006)

Synergistic effects in mixed Escherichia coli biofilms:

conjugative plasmid transfer drives biofilm expansion.

J Bacteriol 188: 3582–3588.Rossell�o-Mora R & Amann R (2001) The species concept for

prokaryotes. FEMS Microbiol Rev 25: 39–67.Sengeløv G, Kristensen KJ, Sørensen AH, Kroer N & Sørensen

SJ (2001) Effect of genomic location on horizontal transfer

of a recombinant gene cassette between Pseudomonas strains

in the rhizosphere and spermosphere of barley seedlings.

Curr Microbiol 42: 160–167.Seoane J, Yankelevich T, Dechesne A, Merkey B, Sternberg C

& Smets BF (2011) An individual-based approach to explain

plasmid invasion in bacterial populations. FEMS Microbiol

Ecol 75: 17–27.Sørensen SJ, Sørensen AH, Hansen LH, Oregaard G & Veal D

(2003) Direct detection and quantification of horizontal

gene transfer by using flow cytometry and gfp as a reporter

gene. Curr Microbiol 47: 129–133.Sørensen SJ, Bailey M, Hansen LH, Kroer N & Wuertz S

(2005) Studying plasmid horizontal transfer in situ: a critical

review. Nat Rev Microbiol 3: 700–710.Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F &

Higgins DG (1997) The CLUSTAL_X windows interface:

flexible strategies for multiple sequence alignment aided by

quality analysis tools. Nucleic Acids Res 25: 4876–4882.van Elsas JD & Bailey MJ (2002) The ecology of transfer of

mobile genetic elements. FEMS Microbiol Ecol 42: 187–197.

ª 2012 Federation of European Microbiological Societies FEMS Microbiol Lett 338 (2013) 95–100Published by Blackwell Publishing Ltd. All rights reserved

100 C.I. de la Cruz-Perera et al.