APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Nov. 2011, p. 7647–7655 Vol. 77, No. 210099-2240/11/$12.00 doi:10.1128/AEM.05565-11Copyright © 2011, American Society for Microbiology. All Rights Reserved.

Geographical Location Determines the Population Structurein Phyllosphere Microbial Communities of a

Salt-Excreting Desert Tree�†Omri M. Finkel,1 Adrien Y. Burch,2 Steven E. Lindow,2

Anton F. Post,3 and Shimshon Belkin1*Institute of Life Sciences, Hebrew University of Jerusalem, Jerusalem, Israel1; Department of Plant and Microbial Biology,

University of California, Berkeley, Berkeley, California2; and Josephine Bay Paul Center forComparative Molecular Biology and Evolution, Marine Biology Laboratory,

Woods Hole, Massachusetts3

Received 25 May 2011/Accepted 3 September 2011

The leaf surfaces of Tamarix, a salt-secreting desert tree, harbor a diverse community of microbial epiphytes.This ecosystem presents a unique combination of ecological characteristics and imposes a set of extreme stressconditions. The composition of the microbial community along ecological gradients was studied from analysesof microbial richness and diversity in the phyllosphere of three Tamarix species in the Mediterranean and DeadSea regions in Israel and in two locations in the United States. Over 200,000 sequences of the 16S V6 and 18SV9 hypervariable regions revealed a diverse community, with 788 bacterial and 64 eukaryotic genera but onlyone archaeal genus. Both geographic location and tree species were determinants of microbial communitystructures, with the former being more dominant. Tree leaves of all three species in the Mediterranean regionwere dominated by Halomonas and Halobacteria, whereas trees from the Dead Sea area were dominated byActinomycetales and Bacillales. Our findings demonstrate that microbial phyllosphere communities on differentTamarix species are highly similar in the same locale, whereas trees of the same species that grow in differentclimatic regions host distinct microbial communities.

The leaf surface of terrestrial plants, the phyllosphere, pro-vides an extensive habitat for microorganisms. Covering anestimated surface area of 6.4 � 108 km2 and comprising themain interface between terrestrial biomass and the atmo-sphere, this environment harbors a substantial microbial pop-ulation, consisting of up to �1026 bacterial cells (22) as well aseukaryotes and archaea. The global phyllosphere is extremelydiverse, with multiple variables contributing to a dazzling mi-crobial diversity of (potentially) millions of bacterial species(20). Different plant species select for different bacterial con-sortia (20, 38, 39, 40), a feature attributed to the difference incomposition of phytochemicals found on the leaf surfaces (33,39). Another factor that contributes to variability is seasonality,as distinct successional patterns have been shown to occuramong microbial communities of certain plant species (16, 31).Variability in the spatial distribution of microbes has beenshown to exist at several scales, such as within the same tree oreven the same leaf (2, 17, 22). Intraspecies variability at largergeographical scales has recently become the focus of somestudies (18, 32) in phyllosphere ecology.

In a paper discussing the diversity of phyllosphere microbi-ota, Lambais et al. (20) raised the question of whether thesame tree species harbor similar communities in different lo-cations. In the present communication we provide some an-

swers to this question from a study of the biogeographicalpatterns in microbial community composition in the phyllo-sphere of the genus Tamarix. This extremely resilient tree isadapted to a wide range of water availability conditions, whichcontributes to its nearly global distribution as well as to itssuccess as an invading species in the New World (36). One ofthe adaptation mechanisms of Tamarix trees to a wide range ofsalinities is their ability to secrete solutes onto the leaf surface,rendering it extremely saline (36) and, in some species, alkaline(30, 37). Microorganisms living on this surface are exposed toa stress “cocktail” consisting of elevated and fluctuating salin-ity, periodic desiccation, moderately high temperatures, highlevels of UV radiation, and in some cases high alkalinity. De-spite these harsh conditions, a substantial epiphytic microbialpopulation thrives on the surface of Tamarix leaves (30).

The extreme conditions on Tamarix leaf surfaces suggestthat the resident microbial community is uniquely adapted tothese conditions, a feature that would be reflected in its geno-typic makeup. Coupled with the global distribution of Tamarixspp. over different climate zones, this genus makes an inter-esting candidate for the study of basic questions in the bioge-ography of epiphytic microorganisms. We have focused onthree Tamarix species: Tamarix aphylla, Tamarix nilotica, andTamarix tetragina. Leaf samples were taken from trees in fourdifferent locations in Israel as well as from two sites in theUnited States. 16S/18S rRNA hypervariable tag sequencing(35) was employed in order to analyze the microbial richnessand diversity of these leaf samples and determine their spatialvariability, representing a novel attempt to deeply analyze mi-crobial richness in the desert phyllosphere.

* Corresponding author. Mailing address: Institute of Life Sciences,the Hebrew University of Jerusalem, Jerusalem 91904, Israel. Phone:972 2 6584192. Fax: 972 2 6585559. E-mail: shimshon@vms.huji.ac.il.

† Supplemental material for this article may be found at http://aem.asm.org/.

� Published ahead of print on 16 September 2011.

7647

on Decem

ber 22, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

MATERIALS AND METHODS

Sampling. Fourteen leaf samples were collected from four locations in Israel(Table 1): by the Mediterranean Sea (Mediterranean), an arid site north of theDead Sea (Dead Sea arid), a site in the Negev Desert highlands (Negev), and anoasis by the Dead Sea (Dead Sea oasis). In addition, samples were collected fromtwo locations in the United States: a tree located in a cultivated garden on theisland of Martha’s Vineyard, MA (Martha’s Vineyard) and a site in Davis,northern California (California). Not all tree species were present in all locations(Table 1). In all cases, duplicate samples were collected from multiple trees.

Leaves were collected between 11:00 a.m. and 2:00 p.m. from different parts ofeach tree, at random, into sterile paper envelopes; the samples were returned tothe laboratory and processed within 2 to 5 h of sampling. The leaves were placedinside 50-ml sterile plastic test tubes (Falcon) and immediately immersed insterile phosphate-buffered saline (PBS) medium (4 g of leaf/40 ml of PBS buffer,pH 7.4). Bacteria were dislodged from the leaves using a sonication tub (Tran-sistor/ultrasonic T7; L&R Manufacturing Co.) for 2 min at medium intensity andby vortexing six times for 10 s each time at 5-min intervals. The leaf wash (LW)was separated from the leaf debris by decanting and kept for analysis. Chloro-phyll content of the leaves was measured as described by Lorenzen (23).

DNA extraction. Leaf washes were filtered on a 0.22-�m-pore-size membranefilter (Millipore) which was subjected to total community DNA extraction, usinga Soil Microbial DNA extraction kit (Zymo Research, Orange, CA).

16S/18S rRNA tag pyrosequencing. Details of the 16S/18S rRNA tag pyrose-quencing method have been described elsewhere (1, 9, 14, 35). In short, DNAsamples were PCR amplified (30 cycles) using three sets of primers (Table 2)flanking either the bacterial and archaeal V6 hypervariable region or the eu-karyotic V9 hypervariable region in the small-subunit rRNA gene. Each ribo-somal primer set is flanked by pyrosequencing linker sequences and by a 5-nu-cleotide key for sample identification among mixed amplicon libraries. Anequimolar mix from each sample was used to prepare pyrosequencing beads viaemulsion PCR. Beads were loaded onto a picotiter plate (estimated at�1,000,000 beads/plate), and pyrosequencing was performed using a GenomeSequencer FLX System (Roche).

To decrease error rates, the FLX system output was passed through severalquality filters additional to the ones built into the system (13). The followingreads were omitted from microbial community analyses: reads with ambiguousnucleotides, reads shorter than 50 nucleotides, reads lacking the sample key

and/or primer sequence at either end, reads with erroneous primer sequences ateither end, and reads that could not be unambiguously assigned to a sample.

The remaining sequences were assigned the taxonomic classification of themost similar reference sequences in the V6 reference database (V6 RefDB) (9),based on a global alignment of the query sequence against the reference se-quences in the V6 RefDB (Global Alignment for Sequence Taxonomy [GAST])(14). In cases where a tag was equidistant to multiple reference sequences, it wasclassified to the level of the most resolved taxon shared by at least two-thirds ofthe reference sequences nearest to that tag.

All 14 samples were sequenced using bacterial primers, and five samples weresequenced using eukaryotic primers. Only two samples (the two MediterraneanT. nilotica trees) were sequenced using an archaeal primer set (Table 2) as thesewere the only samples from which an archaeal 16S rRNA gene PCR product wasobtained.

To estimate the species richness of samples independently of taxonomic as-signment, all of the tag sequences from a single sample or from categorizedgroups of samples were aligned using a pairwise alignment method and clusteredusing a 2% single-linkage preclustering methodology followed by average-link-age clustering based on pairwise alignments (15), a clustering method designedto minimize operational taxonomic unit (OTU) inflation caused by sequencingerrors. Rarefaction curves, the Chao1 richness estimator, an abundance-basedcoverage estimator (ACE) (7), and the Shannon diversity index (H�) were cal-culated using mothur (34). Comparisons between community dissimilarity andenvironmental conditions were carried out using environmentally fitted nonmet-ric multidimensional scaling (NMDS) plots (27) and partial Mantel correlations(21), both using the Vegan package in R (28). Mantel tests were used sincevalues in dissimilarity matrices are not independent. Partial Mantel tests wereused in order to isolate the effect of correlated parameters from one another. Pvalues were calculated using 100,000 permutations on rows and columns ofdissimilarity matrices.

Chemical analysis of the leaf environment. Leaf wash for chemical analysiswas prepared as described above, replacing PBS with double-distilled water(DDW). Electrical conductivity (EC) was measured using a conductivity meter(S30 Seveneasy Conductivity; Mettler, Toledo, OH). The correlation betweenEC values (mS/cm) and the concentration of Na� ions in the leaf wash (mg/liter)was established using an inductively coupled plasma optic emission spectrometer([ICP/OES] Optima 3000; Perkin Elmer, Waltham, MA) and was found to be

TABLE 1. A list of sample collection sites, their location, sampling dates, and the number of each Tamarix species sampled at each location

Collection site Coordinates Sampling date(mo/day/yr)

No. of samples by species

T. nilotica T. aphylla T. tetragina

Dead Sea arid 31°51�16.56�N, 35°31�53.16�E 7/23/09 2 1Dead Sea oasis 31°42�46.57�N, 35°27�13.77�E 7/23/09 1Mediterranean 32°33�33.97�N, 34°54�30.61�E 7/23/09 2 2 2Negev 30°52�40.49�N, 34°47�11.84�E 7/23/09 1Martha’s Vineyard 41°27�34.70�N, 70°33�27.85�W 8/16/09 1a

California 38°32�20.72�N, 121°45�57.86�W 8/28/09 2

a Tentative species identification.

TABLE 2. Primers used in this study

Primer name 5�–3� sequencea Description (reference)b

967F GCCTCCCTCGCGCCATCAG-XXXXX-CNACGCGAAGAACCTTANCCAACGCGAAAAACCTTACCCAACGCGCAGAACCTTACCATACGCGARGAACCTTACCCTAACCGANGAACCTYACC

454 Adapter A-recognition key-bacterial V6forward primer (35)

1046R GCCTTGCCAGCCCGCTCAG-CGACAGCCATGCANCACCTCGACAACCATGCANCACCTCGACGGCCATGCANCACCTCGACGACCATGCANCACCT

454 Adapter B-bacterial V6 reverse primer (35)

958F GCCTCCCTCGCGCCATCAG-XXXXX-AATTGGANTCAACGCCGG 454 Adapter A-recognition key-archaeal V6forward primer (35)

1048R GCCTTGCCAGCCCGCTCAG-CGRCGGCCATGCACCWC 454 Adapter B-archaeal V6 reverse primer (35)1380F GCCTCCCTCGCGCCATCAG-XXXXX-CCCTGCCHTTTGTACACAC 454 Adapter A-recognition key-eukaryotic V9

forward primer (1)1510R GCCTTGCCAGCCCGCTCAG-CCTTCYGCAGGTTCACCTAC 454 Adapter B-eukaryotic V9 reverse primer (1)

a The 454 adapter sequences are indicated in italics. The boldface X residues represent a recognition key for sample identification.b Primer sequence components are described in respective order.

7648 FINKEL ET AL. APPL. ENVIRON. MICROBIOL.

on Decem

ber 22, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

linear. pH was determined using a pH meter equipped with a combination glasselectrode (model 420, Orion; ThermoOrion). Total dissolved organic carbon(DOC) was determined using a Formacs total organic carbon high-temperaturecombustion analyzer (Skalar Analytical B. V. Breda, The Netherlands), followingthe removal of inorganic carbon by lowering the pH to �2.0.

Nucleotide sequence accession number. The sequences determined in thisstudy have been deposited in the NCBI Short Read Archive under accessionnumber SRA026088.2.

RESULTS AND DISCUSSION

Chemical composition of the leaf surface deposits. Resultsof chemical analysis of leaf washes (LWs) are presented inTable 3. Electrical conductivity (EC), a proxy for salinity, wassignificantly higher in the Dead Sea arid samples than at othersampling sites (t 0.016, P 0.01). This may be attributed tothe high July-August temperatures in the Dead Sea area, av-eraging a daily maximum of 39°C (compared with 31 to 32°C inthe Negev and the Mediterranean coastal plain [19]), coupled

with low water availability at the site. Also, a significant differ-ence in pH values appears to exist among tree species (one-wayanalysis of variance [ANOVA], P � 105). Leaf washes fromT. aphylla trees were alkaline (mean pH, 9.5) while leaf washesfrom T. nilotica and T. tetragina were neutral (mean pH, 7.15).The concentration of dissolved organic carbon (DOC) on theleaf surface was high in all cases, varying between 1 and 4 mgof C/g of leaf. For comparison, the sugar concentration ontomato leaves was determined at 1.55 �g/g (26). Organic car-bon concentrations reported in this study, as well as thosepreviously published for T. aphylla (4, 30), are thus approxi-mately 3 orders of magnitude higher. In addition to providingrich carbon and energy sources for the phyllosphere bacteria, itis likely that these compounds also aid in desiccation andosmotic stress responses.

Sequence recovery and microbial abundance. Our sequenc-ing effort yielded 158,980 bacterial V6 sequences, �60 bp inlength, from 14 samples, averaging 11,355 sequences per sam-ple. Of these sequences, 31,547 were identified as chloroplastand mitochondrial 16S rRNA derived from the host tree (Ta-bles 4 and 5) and were removed from the data set. In theeukaryotic V9 data sets (Table 2), 14,294 out of the 48,673sequences were identified as plant 18S rRNA in a total of fivesamples.

The distribution of organelle-derived sequences was not ran-dom. The mean ratio between bacteria-derived and organelle-derived amplicons of Dead Sea arid trees is 0.31 � 0.17, com-pared with 292 � 496 for Mediterranean trees. A similar trendappears in the eukaryotic data set: the mean ratio betweeneukaryote-derived amplicons and plant-derived amplicons ofDead Sea arid trees is 0.80 � 0.17, compared with 184 � 224for Mediterranean trees. Some of this variation can be ex-plained by differences in chloroplast content between sites(170 � 49 mg of chlorophyll/g of leaves in Mediterranean treesand 70 � 16 mg of chlorophyll/g of leaves in Dead Sea trees).However, this difference cannot explain the approximately1,000-fold difference observed in the ratio between bacteria-derived amplicons and chloroplast-derived amplicons. A pos-

TABLE 3. EC, pH, and DOC content of leaf washesa

Sample (site, species, tree no.)b EC(mS/cm) pH DOCc

(mg of C/g)

Mediterranean, T. tetragina 1 13.43 6.54 1.8Mediterranean, T. tetragina 2 6.9 7.26 2.17Mediterranean, T. nilotica 1 7.9 7.26 3.67Mediterranean, T. nilotica 2 7.83 7.59 3.45Mediterranean, T. aphylla 1 3.18 9.74 1.46Mediterranean, T. aphylla 2 4.37 9.94 1.5Dead Sea arid, T. aphylla 13.83 9.1 2.58Dead Sea arid, T. nilotica 1 10.68 7.35 1.13Dead Sea arid, T. nilotica 2 15.95 7.28 2.87Dead Sea oasis, T. nilotica 6.16 6.88 2.79Negev, T. aphylla 4.37 9.73 3.04California, T. aphylla 1 3.72 9.74 1.7California, T. aphylla 2 2.14 8.97 0.95Martha’s Vineyard, Tamarix sp. 1.2 7.05 3.5

a Leaf washes were performed with 1 g of leaf/10 ml of H2O. EC, electricalconductivity; DOC, dissolved organic carbon.

b Numbers indicate that two trees of the same species were sampled at thesame location.

c DOC values are normalized per gram of leaf.

TABLE 4. The number of bacterial tags obtained for each sample and the available level of taxonomic resolution

Sample (site, species, tree no.)aNo. of reads Percent assigned by taxonomic rankb

Total Bacterial Organelle-derived DNA Order Family Genus Species

Dead Sea arid, T. nilotica 1 9,981 3,080 6,911 92.9 81.5 63.7 2.4Dead Sea arid, T. nilotica 2 11,675 1,216 10,459 91.3 80 58.3 2.8Dead Sea arid, T. aphylla 1 10,134 2,685 7,449 94.5 84.4 68.3 2Mediterranean, T. nilotica 1 12,149 12,029 120 98.9 94.4 92.3 0.1Mediterranean, T. nilotica 2 12,801 10,733 2,068 95.6 92 81.5 0.1Mediterranean, T. tetragina 1 14,114 13,794 320 94.7 94.2 66.6 0.1Mediterranean, T. tetragina 2 13,250 13,185 65 96.1 91 71.1 0Mediterranean, T. aphylla 1 10,366 10,358 8 99.9 99.1 98.5 0.1Mediterranean, T. aphylla 2 11,304 11,198 106 99.8 98.1 97.3 0.1Dead Sea oasis, T. nilotica 10,033 6,646 3,387 96.5 88.7 77.8 6.5Negev, T. aphylla 10,202 10,162 40 99.4 91.3 64.6 0.2Martha’s Vineyard, Tamarix sp. 10,123 9,859 264 98.5 84.9 65.1 0.7California, T. aphylla 1 10,649 9,843 806 78.6 70.4 51.6 1.2California, T. aphylla 2 12,199 11,910 289 43.3 41.8 9.7 0.2

Total 158,980 126,688 32,292 91.2 86.1 70.3 1.2

a Numbers indicate that two trees of the same species were sampled at the same location.b Percent taxonomic rank calculations do not include sequences recognized as plant organelle sequences.

VOL. 77, 2011 GEOGRAPHY DETERMINES PHYLLOSPHERE MICROBIAL DIVERSITY 7649

on Decem

ber 22, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

sible explanation is that this ratio is caused mainly by a vari-ability of 2 to 3 orders of magnitude in microbial biomassbetween the Dead Sea and Mediterranean sites. This hypoth-esis is supported by counts of CFU on LB agar plates contain-ing 5% NaCl. Mediterranean trees harbored 2 � 105 to 6 � 105

CFU per gram of leaves, while Dead Sea trees harbored only4 � 103 CFU/g.

It should be noted that the taxonomic resolution of theeukaryotic hypervariable region V9 was distinctly lower thanthat of the bacterial V6 region: only 7.3% of eukaryotic V9could be assigned by GAST to taxonomic ranks below order. Incomparison, 86.0% of bacterial V6 sequences could be as-signed to family, and �70% could be assigned at the genuslevel.

Archaeal amplicons were obtained only from the two Med-iterranean T. nilotica samples, yielding 13,281 and 14,107 V6sequences. Over 99% of the sequences were identified as mem-bers of the family Halobacteriaceae, and 92% of these wereassigned to three known genera within this family (see TableS1 in the supplemental material).

Species richness and alpha diversity. Bacterial species rich-ness across all samples was estimated by the ACE richnessestimator (7) to be 5,265 when operational taxonomic units(OTUs) were clustered with a 6% dissimilarity cutoff, and8,754 at a 3% dissimilarity cutoff (see Fig. S1 in the supple-mental material). Taxonomic assignments of the reads identi-fied 788 genera. Interestingly, using simple average-neighborclustering, the clustering method we have used prior to “iron-ing out the wrinkles” (15), resulted in an estimated 9,322OTUs with a 3% dissimilarity cutoff, indicating that OTUinflation due to sequencing errors was less that 10%.

Categorizing samples according to tree species (Fig. 1A) andgeographical location (Fig. 1B) clearly indicates that bacterialspecies richness and diversity correlated with geographic loca-tion rather than with Tamarix species. The ACE richness esti-mate for the Dead Sea arid samples was 1,963 OTUs (mean,1,239 OTUs per tree) at 94% similarity, almost twice as high asthe ACE value for Mediterranean samples (1,115 OTUs;mean, 500 OTUs per tree). The ACE values for T. niloticafrom all locations compared with those for T. aphylla from alllocations were similar: 2,018 for T. nilotica (mean, 865 OTUsper tree) and 1,593 for T. aphylla (mean, 581 OTUs per tree).Shannon diversity indices displayed a similar pattern, withDead Sea samples being about twice as diverse as Mediterra-nean samples (Fig. 1).

Previous estimates of microbial richness of the phyllosphere

of a single tree species, obtained by Sanger sequencing of 16Sclone libraries, were 1 to 2 orders of magnitude lower (8, 20,30) but could not serve for a valid comparison of richness withthe study presented here due to the difference in samplingdepths between sequencing methods. Recently, pyrosequenc-ing was applied in the study of phyllosphere microbial diversityamong a variety of tree species, albeit with a lower samplingdepth, reaching 600 to 1,500 reads per sample (32). While ararefaction analysis was not presented, the authors reportedthat they identified 164 to 424 OTUs with a 3% dissimilaritycutoff among the first 750 reads. In comparison, Tamarix leafsamples yielded lower species richness, between 71 to 286OTUs, for an equivalent sampling depth. This difference maybe partially attributed to the targeting of a different hypervari-able region (10), as well as to differences in clustering methods.While it is tempting to link the lower Tamarix species richnessto the extreme abiotic conditions on its leaves, evidence pointsotherwise: the Dead Sea sites, which had the highest temper-ature and salinity among the samples in this study, also dis-played the highest taxonomic richness.

Species richness of eukaryotes (see Table S2 in the supple-mental material) was an order of magnitude lower than that ofbacteria. In contrast to bacteria, geographic location did notseem to affect eukaryotic species richness: ACE richness esti-mates for Dead Sea and Mediterranean trees were similar (170and 176 OTUs, respectively).

Species diversity. Beta diversity, a measure of the propor-tion of unique taxa between different samples, was calculatedaccording to Yue and Clayton (41), providing a weighted dis-tance measure that accounts for the relative abundance of taxain each sample, and the results were clustered using the Kitschalgorithm. The use of a weighted algorithm is necessary in suchan analysis due to the variety of sample sizes. The same anal-ysis, performed using other weighted distance measures suchas Morisita-Horn (12) and an unweighted one (6), yielded verysimilar results (data not shown). Two related non-Tamarixsamples were added to the analysis: a sample from the gut ofa T. nilotica phyllosphere insect, Oxyrrachis versicolor, collectedfrom a tree in an oasis by the Dead Sea (O. M. Finkel et al.,unpublished data), and a sample taken from Dead Sea waterduring a 1992 algal bloom (5). To evaluate the taxonomicdepth of differences and similarities between samples, theseanalyses were performed at three levels: order, family, andgenus. Very similar patterns were seen for the different taxonlevels. The order-level analysis is presented in Fig. 2, while thatof family and genus can be found in the supplemental materials

TABLE 5. The number of eukaryotic tags obtained for each sample and the available level of taxonomic resolution

Sample (site, species,tree no.)b

No. of reads Percent assigned by taxonomic ranka

Total Eukaryotic Plant DNA Metazoan DNA Order Family Genus Species

Dead Sea arid, T. nilotica 1 6,979 2,624 3,638 717 61.1 10.9 10.9 10Dead Sea arid, T. aphylla 1 2,615 1,056 1,554 5 57.7 12.2 12.2 9Mediterranean, T. nilotica 1 15,515 14,362 568 585 98.9 2.9 2.9 2.8Mediterranean, T. aphylla 1 13,395 11,874 39 1,482 82.8 4.2 4.2 4.1Dead Sea oasis, T. nilotica 10,169 517 8,495 1,157 49.9 6.3 6.3 4.2

Total 48,673 34,379 14,294 789.2 70.1 7.3 7.3 6

a Percent taxonomic rank calculations do not include sequences recognized as plant genomic sequences or as metazoan sequences.b Numbers indicate that two trees of the same species were sampled at the same location.

7650 FINKEL ET AL. APPL. ENVIRON. MICROBIOL.

on Decem

ber 22, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

(see Fig. S2 and S3). Analysis at a species resolution wouldyield little information as most reads could not be assigned atthis level.

Two main clusters emerged for all three taxonomic levels: (i)a Mediterranean cluster that includes phyllosphere communi-ties from the Mediterranean and Negev samples, with theMartha’s Vineyard sample forming an outgroup within thiscluster, and (ii) a Dead Sea cluster encompassing all samples

from sites by the Dead Sea, with the sample designated Cali-fornia 1 and the O. versicolor sample forming outgroups. WhileCalifornia 2 and Dead Sea water were also included in thiscluster, their shared branch lengths were so short that theyshould each be considered an autonomous cluster. The Med-iterranean cluster appears rather uniform, characterized bysmall differences among its members. The Dead Sea cluster ischaracterized by greater branch lengths, coinciding with in-

FIG. 1. Rarefaction curves comparing bacterial species richness between T. nilotica and T. aphylla (A) and between Dead Sea and Mediter-ranean sampling sites (B). OTUs are categorized according to a 6% dissimilarity cutoff. Chao and ACE richness estimates and Shannon diversityindex are listed in the inset tables.

VOL. 77, 2011 GEOGRAPHY DETERMINES PHYLLOSPHERE MICROBIAL DIVERSITY 7651

on Decem

ber 22, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

creasing dissimilarity with increasing geographical distance. Allthree Dead Sea arid samples clustered together while all othersamples each formed their own clusters, with the Dead Seaoasis sample as the nearest relative to the Dead Sea aridcluster. Surprisingly, the Dead Sea water sample and, to alesser extent, the O. versicolor sample, did not form outgroupsto the entire data set although their shared branch lengths withother members of their cluster are considerably shorter thantheir unique branch lengths.

To test which environmental or geographical parameterscorrelated with community dissimilarity, an NMDS plot wasdrawn for the Israeli samples only (Fig. 3). As the samplingsites were in different climatic regions, the changes in salinity,mean daily temperature, and latitude appear to have the samedirection. In order to isolate the effect of each of these vari-ables, partial Mantel tests were performed. Mean daily tem-perature (R 0.87, P 0.00045) and geographic location(R 0.404, P 0.0125) were both found to be significantlycorrelated with bacterial community dissimilarity, indicatingthat while the majority of the difference between the Dead Seacluster and the Mediterranean cluster appears to be due toclimatic factors, geographic isolation plays a significant role aswell.

The differences between sampling sites have manifestedthemselves in both the richness and diversity of microbial spe-cies from all three domains of life. Such variability within andbetween samples should be taken into account in attempting toestimate the global diversity of phyllosphere bacteria since it isnot based on interspecies variability alone. In a recent com-munication, Knief et al. (18) also showed that geographic lo-cation may play a more important role than host species indetermining microbial epibiont composition. However, other

recent evidence has shown that this is not necessarily a generalrule. Pinus ponderosa leaves in the United States and Australiahosted bacterial communities very similar to each other com-pared with sympatric trees of different species within the samegenus, the exact opposite of the trend we have seen for Tama-rix (32).

Figure 4 displays the main taxonomic orders comprising the

FIG. 2. Heat map and Kitsch tree imaging the level of similarity in bacterial order composition between samples as calculated by theYue-Clayton similarity index.

FIG. 3. Nonmetric multidimensional scaling (NMDS) of samplesfrom four sites in Israel. Relative positions of points represent com-munity similarity between sites. Vector direction represents the aver-age direction of environmental gradients. Vector length is proportionalto the magnitude of correlation between the environmental parameterand bacterial community composition. Only environmental variableswith P � 0.1 are shown. P values are based on 10,000 permutations.

7652 FINKEL ET AL. APPL. ENVIRON. MICROBIOL.

on Decem

ber 22, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

different samples, highlighting differences between the twogeographical/climatic clusters. A more detailed list of the iden-tified population components is presented in supplementalTable S3. The dominant bacterial order at the Mediterraneansite was Oceanospirillales, in which the genus Halomonas wasthe most abundant, comprising over 90% of the bacterial pop-ulation on T. aphylla trees and 50 to 60% of the population onT. nilotica and T. tetragina. This genus is also dominant in thetwo other samples that clustered together with the Mediterra-nean samples, Negev and Martha’s Vineyard. The majority ofinterspecies differences can be seen within proteobacteria (Fig.5): T. nilotica and T. tetragina support an abundance of anotherHalomonadaceae genus, Chromohalobacter, as well as Salini-sphaera (Gammaproteobacteria) and Rhodobacterales (Alpha-proteobacteria). The latter two genera are both nearly absenton T. aphylla.

While the Dead Sea cluster is characterized by greater di-versity than the Mediterranean cluster, both within and be-tween samples, most samples appear to be dominated by mem-bers of the two Gram-positive orders, Actinomycetales andBacillales. It should be noted that the four most dominantbacterial orders found in the Dead Sea proper—Bacillales(95% of the data set), Actinomycetales, Rhizobiales, and Burk-holderiales—are all relatively abundant on Tamarix trees fromthe same region (Fig. 4).

The absence of Halomonas, the most dominant genus onMediterranean, Negev, and Martha’s Vineyard samples, fromthe Dead Sea and California samples is intriguing. From aphenotypic standpoint, the abundant presence of Halomonasin most samples is not surprising. Most known members of this

genus are motile, can grow in up to 20% salinity, and are ableto utilize glycerol and trehalose (25), found in abundance onTamarix leaves (4). The one trait that is exclusively shared byall Dead Sea samples is temperature, which is considerablyhigher than in the other sampling sites, averaging summer dailymaxima of 39°C. It is possible that these temperatures or an-

FIG. 4. Diversity of taxonomic orders within and between samples, as assigned by GAST (14). Orders that comprised less than 1% of each dataset are binned together. Legend entries are sorted according to descending abundance and include taxonomic information as, in respective order,phylum, class, and order. NA, not assigned. The entire data set including rare taxa at the finest taxonomic resolution achieved is available in TableS3 in the supplemental material.

FIG. 5. Diversity of Proteobacteria within and between Mediterra-nean samples. Taxonomic information is indicated, in respective order,as class, order, family, and genus. NA, not assigned.

VOL. 77, 2011 GEOGRAPHY DETERMINES PHYLLOSPHERE MICROBIAL DIVERSITY 7653

on Decem

ber 22, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

other related factor (such as water availability) inhibits growthof Halomonas in this area. The sites with abundant Halomonasall have relatively high air humidity in common, and thus thesalt crystals on their leaves are more likely to absorb moisture.Indeed, recent results indicate that Halomonas isolates andculture-independent Halomonas 16S rRNA sequences werefound in samples from the Dead Sea area taken the followingwinter, when temperatures were lower and humidity higher(19) (data not shown). However, it has been reported (25) thatseveral strains of Halomonas can grow at up to 45°C. Furtherinvestigation of local Halomonas isolates is required in order totest the environmental variables that affect their growth.

An analysis performed on unicellular eukaryotic sequencesrevealed a similar dependence of their abundance on the en-vironment to that of bacteria, as samples from different Tama-rix species and the same environment clustered together (Fig.6). The dominant order in all of these samples is the fungalAscomycota. Within this order, families were distributed ac-cording to geographic location: Lecanoromycetes and Trichoco-maceae were found in all three Dead Sea samples but wereabsent in the Mediterranean. In contrast, Onygenales werepresent in samples from the Mediterranean and the Dead Seaoasis but were absent from the Dead Sea arid sample (seeTable S2 in the supplemental material).

An important question that remains unanswered is whatfraction of the community structure can be attributed to localfeatures, as opposed to limitations on dispersal of the mi-crobes. In other words, what role does geographic distance playin the microbial community shifts between locations (11, 24,29)? The depth of 16S tag sequencing and the rare biospherethat it reveals have partially answered that question. Sequencesassigned to an unidentified Halomonas species dominate theMediterranean, Negev, and Martha’s Vineyard samples butare also found in small numbers in Dead Sea and Californiasamples. This observation excludes migration barriers as thefactor responsible for the differences observed. An oppositeexample of a site-restricted taxon is much more difficult toidentify. Blattabacteriaceae, a little studied family of Flavobac-

teriales, are found in relatively small numbers (16 to 47 reads,�0.4%) in all samples from the Dead Sea area but are com-pletely absent from any of the other samples, despite the factthat sampling depth in the Dead Sea area was considerablylower. This might suggest that this bacterial family does notmigrate as easily as others do. To more carefully determine theextent to which the differences between communities observedin this study can be attributed to local environmental featuresas opposed to limitations on dispersal, the role that geographicdistance plays in the community shift observed between loca-tions should be further addressed (3). This can be done bysampling trees in sites that are as similar as possible in theirbiotic and abiotic surroundings but are geographically distantfrom each other. A further insight as to how “island-like” thephyllosphere environment actually is can be gained from com-paring microbial communities on Tamarix to those of sympa-tric species of different genera, to those of the soil in which thetrees grow, and to the air that connects them.

ACKNOWLEDGMENTS

Research was supported in part by the U.S.-Israel Binational Sci-ence Foundation grant number 2006324 to S.B. and S.E.L. O.M.F. andS.B. are indebted to the Gruss-Lipper Family Foundation for summerresearch fellowships (2009 and 2010) at the Marine Biology Labora-tory (Woods Hole, MA) that supported pyrosequencing and data anal-ysis. Earlier support of the Tamarix phyllosphere biodiversity study bythe Bridging the Rift Foundation is also gratefully acknowledged.

Our gratitude goes to the staff and scientists of the Josephine BayPaul Center in Comparative Molecular Biology and Evolution at theMarine Biology Laboratory for their help during and after the researchperiod.

REFERENCES

1. Amaral-Zettler, L. A., E. A. McCliment, H. W. Ducklow, and S. M. Huse.2009. A method for studying protistan diversity using massively parallelsequencing of V9 hypervariable regions of small-subunit ribosomal RNAgenes. PLoS One 4:e6372.

2. Andrews, J. H., and R. F. Harris. 2000. The ecology and biogeography ofmicroorganisms of plant surfaces. Annu. Rev. Phytopathol. 38:145–180.

3. Andrews, J. H., L. L. Kinkel, F. M. Berbee, and E. V. Nordheim. 1987. Fungi,leaves, and the theory of island biogeography. Microb. Ecol. 14:277–290.

4. Belkin, S., and N. Qvit-Raz. 2010. Life on a leaf: bacterial epiphytes of asalt-excreting desert tree, p. 393–406 In J. Seckbach and M. Grube (ed.),Symbioses and stress. Springer, Dordrecht, Netherlands.

5. Bodaker, I., et al. 2010. Comparative community genomics in the Dead Sea:an increasingly extreme environment. ISME J. 4:399–407.

6. Bray, J. R., and J. T. Curtis. 1957. An ordination of the upland forestcommunities of souther Wisconsin. Ecol. Monogr. 27:325–349.

7. Chao, A., and S. M. Lee. 1992. Estimating the number of classes via samplecoverage. J. Am. Stat. Assoc. 87:210–217.

8. Delmotte, N., et al. 2009. Community proteogenomics reveals insights intothe physiology of phyllosphere bacteria. Proc. Natl. Acad. Sci. U. S. A.106:16428–16433.

9. Dethlefsen, L., S. Huse, M. L. Sogin, and D. A. Relman. 2008. The pervasiveeffects of an antibiotic on the human gut microbiota, as revealed by deep 16SrRNA Sequencing. PLoS Biol. 6:e280.

10. Engelbrektson, A., et al. 2010. Experimental factors affecting PCR-basedestimates of microbial species richness and evenness. ISME J. 4:642–647.

11. Green, J., and B. J. M. Bohannan. 2006. Spatial scaling of microbial biodi-versity. Trends Ecol. Evol. 21:501–507.

12. Horn, H. S. 1977. Measurement of “overlap” in comparative ecologicalstudies. Am. Nat. 914:419–423.

13. Huse, S. M., J. A. Huber, H. G. Morrison, M. L. Sogin, and D. Mark Welch.2007. Accuracy and quality of massively parallel DNA pyrosequencing. Ge-nome Biol. 8:R143. http://genomebiology.com/content/8/7/R143.

14. Huse, S. M., et al. 2008. Exploring microbial diversity and taxonomy usingSSU rRNA hypervariable tag sequencing. PLoS Genet. 4:e1000255.

15. Huse, S. M., D. M. Welch, H. G. Morrison, and M. L. Sogin. 2010. Ironingout the wrinkles in the rare biosphere through improved OTU clustering.Environ. Microbiol. 12:1889–1898.

16. Jackson, E. F., H. L. Echlin, and C. R. Jackson. 2006. Changes in thephyllosphere community of the resurrection fern, Polypodium polypodioides,associated with rainfall and wetting. FEMS Microbiol. Ecol. 58:236–246.

FIG. 6. Heat map and Kitsch tree imaging the level of similarity ineukaryotic order composition between samples as calculated by theYue-Clayton similarity index.

7654 FINKEL ET AL. APPL. ENVIRON. MICROBIOL.

on Decem

ber 22, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

17. Kinkel, L. L., M. Wilson, and S. E. Lindow. 1995. Effect of sampling scale onthe assessment of epiphytic bacterial populations. Microb. Ecol. 29:283–297.

18. Knief, C., A. Ramette, L. Frances, C. Alonso-Blanco, and J. A. Vorholt. 2010.Site and plant species are important determinants of the Methylobacteriumcommunity composition in the plant phyllosphere. ISME J. 4:719–728.

19. Kurtzman, D., and R. Kadmon. 1999. Mapping of temperature variables inIsrael: a comparison of different interpolation methods. Climate Res. 13:33–43.

20. Lambais, M. R., D. E. Crowley, J. C. Cury, R. C. Bull, and R. R. Rodrigues.2006. Bacterial diversity in tree canopies of the Atlantic forest. Science312:1917.

21. Legendre, P., and L. Legendre. 1998. Numerical ecology, 2nd Engl. ed.Elsevier, Amsterdam, the Netherlands.

22. Lindow, S. E., and M. T. Brandl. 2003. Microbiology of the phyllosphere.Appl. Environ. Microbiol. 69:1875–1883.

23. Lorenzen, C. J. 1967. Determination of chlorophyll and pheo-pigments:spectrophotometric equations. Limnol. Oceanogr. 12:343–346.

24. Martiny, J. B. H., et al. 2006. Microbial biogeography: putting microorgan-isms on the map. Nat. Rev. Microbiol. 4:102–112.

25. Mata, J. A., Martinez-Canovas, J. E. Quesada, and V. Bejar. 2002. A de-tailed phenotypic characterisation of the type strains of Halomonas species.Syst. Appl. Microbiol. 25:360–375.

26. Mercier, J., and S. E. Lindow. 2000. Role of leaf surface sugars in coloni-zation of plants by bacterial epiphytes. Appl. Environ. Microbiol. 66:369–374.

27. Minchin, P. R. 1987. An evaluation of relative robustness of techniques forecological ordinations. Vegetatio 71:145–156.

28. Oksanen, J., et al. 2008. Vegan: community ecology package. R packageversion 1.15-1. R Foundation for Statistical Computing, Vienna, Austria.http://vegan.r-forge.r-project.org/.

29. Prosser, J. I., et al. 2007. The role of ecological theory in microbial ecology.Nat. Rev. Microbiol. 5:384–392.

30. Qvit-Raz, N., E. Jurkevitch, and S. Belkin. 2008. Drop-size soda lakes:

transient microbial habitats on a salt-secreting desert tree. Genetics 178:1615–1622.

31. Redford, A. J., and N. Fierer. 2009. Bacterial succession on the leaf surface:a novel system for studying successional dynamics. Microb. Ecol. 58:189–198.

32. Redford, A. J., R. M. Bowers, R. Knight, Y. Linhart, and N. Fierer. 2010. Theecology of the phyllosphere: geographic and phylogenetic variability in thedistribution of bacteria on tree leaves. Environ. Microbiol. 12:2885–2893.

33. Ruppel, S., A. Krumbein, and M. Schreiner. 2008. Composition of thephyllospheric microbial populations on vegetable plants with different glu-cosinolate and carotenoid compositions. Microb. Ecol. 56:364–372.

34. Schloss, P. D., et al. 2009. Introducing mothur: open source, platform-independent, community-supported software for describing and comparingmicrobial communities. Appl. Environ. Microbiol. 75:7537–7541.

35. Sogin, M. L., et al. 2006. Microbial diversity in the deep sea and the under-explored “rare biosphere.” Proc. Natl. Acad. Sci. U. S. A. 103:12115–12120.

36. Waisel, Y. 1961. Ecological studies on Tamarix aphylla. L. Karst. Plant Soil13:356–364.

37. Waisel, Y. 1991. The glands of Tamarix aphylla: a system for salt secretion orfor carbon concentration? Physiol. Plant. 83:506–510.

38. Whipps, J. M., P. Hand, D. Pink, and G. D. Bending. 2008. Phyllospheremicrobiology with special reference to diversity and plant genotype. J. Appl.Microbiol. 105:1744–1755.

39. Yadav, R. K. P., E. M. Papatheodorou, K. Karamanoli, H. I. A. Constan-tinidou, and D. Vokou. 2008. Abundance and diversity of the phyllospherebacterial communities of Mediterranean perennial plants that differ in leafchemistry. Chemoecology 18:217–226.

40. Yang, C. H., D. E. Crowley, J. Borneman, and N. T. Keen. 2001. Microbialphyllosphere populations are more complex than previously realized. Proc.Natl. Acad. Sci. U. S. A. 98:3889–3894.

41. Yue, J. C., and M. K. Clayton. 2005. A similarity measure based on speciesproportions. Commun. Stat. Theor. Methods 34:2123–2132.

VOL. 77, 2011 GEOGRAPHY DETERMINES PHYLLOSPHERE MICROBIAL DIVERSITY 7655

on Decem

ber 22, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from