Ratiometric Imaging of Extracellular pH in Bacterial Biofilms withC-SNARF-4

Sebastian Schlafer,a,b Javier E. Garcia,b Matilde Greve,a Merete K. Raarup,c Bente Nyvad,b Irene Digeb

iNANO Interdisciplinary Nanoscience Center, Science and Technology, Aarhus University, Aarhus, Denmarka; Department of Dentistry, Health, Aarhus University, Aarhus,Denmarkb; Stereology and Electron Microscopy Research Laboratory, Centre for Stochastic Geometry and Advanced Bioimaging, Health, Aarhus University, Aarhus,Denmarkc

pH in the extracellular matrix of bacterial biofilms is of central importance for microbial metabolism. Biofilms possess a com-plex three-dimensional architecture characterized by chemically different microenvironments in close proximity. For decades,pH measurements in biofilms have been limited to monitoring bulk pH with electrodes. Although pH microelectrodes with abetter spatial resolution have been developed, they do not permit the monitoring of horizontal pH gradients in biofilms in realtime. Quantitative fluorescence microscopy can overcome these problems, but none of the hitherto employed methods differen-tiated accurately between extracellular and intracellular microbial pH and visualized extracellular pH in all areas of the biofilms.Here, we developed a method to reliably monitor extracellular biofilm pH microscopically with the ratiometric pH-sensitive dyeC-SNARF-4, choosing dental biofilms as an example. Fluorescent emissions of C-SNARF-4 can be used to calculate extracellularpH irrespective of the dye concentration. We showed that at pH values of <6, C-SNARF-4 stained 15 bacterial species frequentlyisolated from dental biofilm and visualized the entire bacterial biomass in in vivo-grown dental biofilms with unknown speciescomposition. We then employed digital image analysis to remove the bacterial biomass from the microscopic images and ade-quately calculate extracellular pH values. As a proof of concept, we monitored the extracellular pH drop in in vivo-grown dentalbiofilms fermenting glucose. The combination of pH ratiometry with C-SNARF-4 and digital image analysis allows the accuratemonitoring of extracellular pH in bacterial biofilms in three dimensions in real time and represents a significant improvement topreviously employed methods of biofilm pH measurement.

Bacterial biofilms are involved in both beneficial and damagingprocesses. They are currently exploited in bioremediation, for

example, in wastewater treatment plants (1), and in the produc-tion of biofuels or bioelectricity in microbial fuel cells (2, 3).Moreover, there is a multitude of symbiotic relationships betweenmicrobial biofilms and higher organisms that contribute to phys-iological homeostasis (4, 5). The functioning of the human gut,for example, depends strongly on microbial processes occurringin enteric biofilms (6, 7). On the other hand, biofilms are involvedin food spoilage (8) and undesirable biofouling in the marine,industrial, and medical fields (9). They are the causative agents ofmany human diseases, such as wound and implant/device-relatedinfections (10, 11), endocarditis (12), cystic fibrosis (13), perio-dontitis (14), and dental caries (15).

Bacterial biofilms possess a complex three-dimensional archi-tecture (16). Reaction-diffusion interactions limit the penetrationof solutes into and out of the biofilm (17), which leads to steepgradients of nutrients and metabolites and the creation of hetero-geneous microenvironments that are colonized by organisms indifferent physiological states and with different biochemical re-quirements (18). Obligately anaerobic organisms, for example,can thrive in biofilms in overtly aerobic environments, such assupragingival tooth surfaces or seawater (19, 20). Studying thearchitecture and bacterial metabolism in biofilms is of crucial im-portance to understanding, exploiting, and controlling biofilms.

The three-dimensional mapping of pH in bacterial biofilmsmerits intensive research. Biofilm pH has a central influence onthe metabolic processes carried out in different areas of a biofilm.pH in microbial fuel cell biofilms, for example, has a significantimpact on electric current production (21). In dental biofilms, pHat the biofilm-tooth interface is the key virulence factor for the

development of caries lesions, one of the most prevalent diseasesof mankind. For decades, pH in dental biofilms was recorded withpH-sensitive electrodes that only allowed bulk pH measurements(22–24). The development of pH microelectrodes with tip diam-eters as small as 10 �m (25) has improved the spatial resolutionand made the recording of vertical pH gradients in dental biofilmspossible (26, 27). pH microelectrodes also are employed to mea-sure pH z profiles in wastewater treatment biofilms, microbial fuelcells, and in the context of biocorrosion (28–30). Still, the inser-tion of microelectrodes disturbs the biofilm mechanically, and it isnot possible to measure horizontal pH gradients in biofilms.

In recent years, various fluorescence microscopic methodshave been developed to overcome these problems and to mapbiofilm pH three dimensionally in real time. Two-photon excita-tion followed by time-gated fluorescence life-time imaging(FLIM) uses pH-dependent differences in the fluorescence life-time decay of, for example, carboxyfluorescein to determine local

Received 31 August 2014 Accepted 5 December 2014

Accepted manuscript posted online 12 December 2014

Citation Schlafer S, Garcia JE, Greve M, Raarup MK, Nyvad B, Dige I. 2015.Ratiometric imaging of extracellular pH in bacterial biofilms with C-SNARF-4.Appl Environ Microbiol 81:1267–1273. doi:10.1128/AEM.02831-14.

Editor: S.-J. Liu

Address correspondence to Sebastian Schlafer, sebastians@microbiology.au.dk.

Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.02831-14.

Copyright © 2015, American Society for Microbiology. All Rights Reserved.

doi:10.1128/AEM.02831-14

February 2015 Volume 81 Number 4 aem.asm.org 1267Applied and Environmental Microbiology

on April 30, 2019 by guest

http://aem.asm

.org/D

ownloaded from

pH in a biofilm sample (31). pH-sensitive nanoparticles compris-ing both pH-sensitive and pH-insensitive dye molecules havebeen developed to quantify pH in biofilms by comparing the flu-orescent emission of both dyes (32). Yet another approach is theuse of ratiometric dyes, such as C-SNARF-4, that display differentfluorescent emission spectra depending on the state of protona-tion of the dye. Calculating the fluorescent emission ratio at twodifferent wavelengths allows the determination of local pH irre-spective of dye concentration and compartmentalization (33, 34).

All fluorescence microscopic approaches have to face twomajor challenges. Although none of the methodologies rely ona homogeneous probe concentration for pH calculations, thefluorescent dye still has to penetrate the biofilm and bind in allextracellular areas with a sufficient concentration to allow forquantification. While small dye molecules readily diffuse intothe biofilm, penetration proved to be a considerable problemfor pH-sensitive nanoparticles. Hidalgo et al. reported thatstaining of the biofilm was not possible with nanoparticles of 70nm and 30 nm in diameter, and 10-nm-sized particles mostlyadhered to bacterial cells, which made it impossible to recordpH in cell-free areas of the biofilm (32).

Additionally, it is crucial to differentiate between extracellularand intracellular pH in the biofilms. Intracellular pH will differconsiderably from extracellular pH due to bacterial homeostasis,and in the case of dental caries, only extracellular pH will affect theunderlying tooth. Both C-SNARF-4 and carboxyfluorescein pen-etrate bacterial cells at low pH. In previous studies by Vroom et al.(31), Hunter and Beveridge (34), and Franks et al. (35), no effortwas undertaken to differentiate between the intra- and extracellu-lar compartments; consequently, intracellular fluorescence con-tributed to the recorded pH values. Moreover, Hunter and Bever-idge and Franks et al. employed genetically modified organismsexpressing the fluorescent proteins green fluorescent protein(GFP) or mCherry to stain the biofilms. While the use of addi-tional fluorescent stains to visualize bacteria in biofilms wouldpermit us to differentiate between intra- and extracellular com-partments, the application of fluorescent dyes other than the pHmarker itself brings about the risk of fluorescent contaminationand false measurements in the extracellular space.

In the present study, we therefore investigate the use of C-SNARF-4 in a dual function, as an intracellular bacterial stain andan extracellular pH marker. Since dental plaque is a classic exam-ple of an acid-producing biofilm, we investigated the staining abil-ities of C-SNARF-4 in the range of pH 4.0 to 8.0 on 15 differentbacterial species frequently isolated from dental biofilm. More-over, we used C-SNARF-4 to stain in vivo-grown dental biofilmsof unknown composition and to monitor extracellular pH in thesebiofilms at the biofilm-substratum interface.

MATERIALS AND METHODSBacterial strains. Actinomyces naeslundii AK6, Actinomyces viscosus(CCUG 33710), Enterococcus faecalis (DSM 20478T), Lactobacillus para-casei subsp. paracasei (DSM 20020), Streptococcus gordonii (ATCC10558T), Streptococcus mitis SK24, Streptococcus mutans (DSM 20523T),Streptococcus oralis (NCTC 7864T), and Streptococcus sanguinis (ATCC10556T) were cultivated aerobically on blood agar (SSI, Copenhagen,Denmark) and transferred to Todd Hewitt broth (THB; Roth, Karlsruhe,Germany) at 35°C until mid- to late exponential phase prior to experi-mental use. Bifidobacterium dentium (DSM 20436T), Fusobacterium nu-cleatum subsp. polymorphum (ATCC 10953T), Neisseria subflava (DSM17610T), Porphyromonas gingivalis (ATCC 33277T), Prevotella intermedia

(CCUG 24041T), and Veillonella parvula (CCUG 5123T) were cultivatedanaerobically on modified Columbia blood agar (36) and transferred to aliquid plaque medium (37) at 36.5°C until mid- to late exponential phasebefore experimental use.

Actinomyces naeslundii AK6 and Streptococcus mitis SK24 were kindlyprovided by M. Kilian, Department of Biomedicine, Aarhus University,Denmark.

Visualization of bacteria in planktonic culture with C-SNARF-4.Optical-bottom 96-well plates (Sigma-Aldrich, Brøndby, Denmark) werecoated with a thin film of porcine gelatin (type A; 0.2% [wt/vol] in Milli-Qwater; Sigma-Aldrich, Brøndby, Denmark) and dried overnight. Bacteriawere washed twice in 0.9% sterile NaCl and adjusted to an optical densityof 0.4 (550 nm). Bacterial suspensions were mixed with HEPES buffersolutions (50 mM, adjusted to pH 4.0 to 8.0 in steps of 0.5 pH units) at aratio of 1:2 and set to settle in the 96-well plates for 1 h. All wells werewashed twice with HEPES buffer of the right pH to remove nonadherentbacterial cells. Thereafter, bacteria were stained with C-SNARF-4 (LifeTechnologies, Nærum, Denmark) for 30 min and imaged with a confocalmicroscope (Zeiss LSM 510 META; Jena, Germany). The microscope wasequipped with a 63� water immersion objective with a 1.2 numericalaperture (Plan Apochromat). A 543-nm laser line (250 to 300 �W) wasused to excite C-SNARF-4, and fluorescence emission was monitoredsimultaneously within 576- to 608-nm (green) and 629- to 661-nm (red)intervals (META detector).

Initially, cells of S. mitis were incubated with different concentrationsof C-SNARF-4 (20 �M, 30 �M, and 50 �M) at pH 4 to 8 to determine theideal concentration for bacterial visualization. While 20 �M yielded suf-ficient contrast between cells and background, a concentration of 50 �Mwas judged to give the best cell/background ratio (see Fig. S1 in the sup-plemental material) and was used in all subsequent experiments. For allbacterial strains and pH values, images were acquired in three differentmicroscopic fields of view, and the x-y positions were marked in the mi-croscope software. All samples then were counterstained with BacLight(Invitrogen, Taastrup, Denmark) according to the manufacturer’s in-structions, and identical microscopic fields of view were imaged to checkif all cells had been visualized by C-SNARF-4. BacLight was excited with488-nm and 543-nm laser lines, and fluorescent emission was detectedwith the META detector set to 500 to 554 nm and 554 to 608 nm, respec-tively. Both channels in BacLight images were pseudocolored in red andboth channels in C-SNARF-4 images in green, and the correspondingC-SNARF-4 and BacLight images were merged in Photoshop (Adobe, SanJose, CA). All experiments were repeated on another day.

In situ biofilm growth. Dental biofilms were grown in situ on custom-made nonfluorescent glass slabs (4 by 4 by 1 mm; Menzel, Braunschweig,Germany) with a surface roughness of 1,200 grit. Glass slabs weremounted slightly recessed on the buccal flanges of an individually de-signed lower jaw splint worn by a volunteer. The in situ model is describedin more detail in Dige et al. (38). The splint was worn for periods of 48 hand was removed from the mouth only for oral hygiene and during intakeof food or liquids other than water. The experimental protocol was ap-proved by the Ethics Committee of Aarhus County (M-20100032).

Visualization of bacteria in dental biofilms with C-SNARF-4. Afterbiofilm growth, the glass slabs were removed carefully from the splint.Biofilms (15 to 35 �m thick) were set to equilibrate for 30 min in custom-made wells filled with HEPES buffer (50 �M; pH 4.0 to 8.0 in steps of 0.5pH units) and C-SNARF-4 (50 �M). For each biofilm, images were ac-quired directly above the biofilm-glass interface in three different micro-scopic fields of view, and the x-y positions were marked in the microscopesoftware. Following counterstaining with BacLight, identical microscopicfields of view were imaged to check if the entire biofilm had been visual-ized by C-SNARF-4. Experiments were performed in duplicate. To test thevisualization of thicker biofilms with C-SNARF-4, dental biofilm wassampled with sterile curettes, transferred to custom-made glass slabs, andincubated overnight in tryptic soy broth (Scharlau, Barcelona, Spain) at35°C. Thereafter, biofilms were incubated for 45 min with sterile saliva

Schlafer et al.

1268 aem.asm.org February 2015 Volume 81 Number 4Applied and Environmental Microbiology

on April 30, 2019 by guest

http://aem.asm

.org/D

ownloaded from

containing 0.4% (wt/vol) glucose and 50 �M C-SNARF-4. z-stacks span-ning the entire height of the biofilms were acquired with an interslicedistance of 3 �m. The experiments were performed in duplicate.

Calibration of C-SNARF-4. For confocal microscopic calibration, 100�l of HEPES buffer solution (50 mM; adjusted to pH 4.5 to 8.5 in steps of0.1 pH units), containing C-SNARF-4 at a concentration of 20 �M, wereimaged in custom-made wells. For image acquisition, a Zeiss LSM 510META (Jena, Germany) with a 40�/1.2-numeric-aperture water immer-sion objective (C-Apochromat) was used. The probe was excited with a543-nm laser line (250 to 300 �W), and fluorescence emission was mon-itored simultaneously within 576- to 608-nm (green) and 629- to 661-nm(red) intervals (META detector), with the pinhole set to 2 Airy units(2-�m optical slice thickness). Images were 364 by 364 pixels (141 by 141�m2) in size and were acquired with a pixel dwell time of 18 �s, lineaverage of 2, 0.4 �m/pixel (zoom 1), and 12-bit intensity resolution. Theprocedure was repeated three times and was performed at 37°C with an XLincubator (PeCON, Erbach, Germany). A measurement was performedon unstained HEPES buffer for background subtraction and showedvalues similar to those for images of the stained buffer solution ac-quired with the 543-nm laser switched off. Therefore, the latter pro-cedure was performed for every third pH value for background sub-traction.

For ratio calculation, regions of 100 by 100 pixels were defined withineach image, and the averages and standard deviations were determinedusing the LSM acquisition software. Subsequently, the ratio R, standarddeviations, sR, and standard errors of means, SR, were calculated for eachpH value as previously described (39). The resulting values of R wereplotted in SigmaPlot 10 (Systat Software, Inc., San Jose, CA, USA) andfitted to the following function:

pH � ln� 1.61

R � 0.0937� 1� · 0.397 � 6.12 (1)

The calibration data and fitted curve are shown in Fig. S2 in the sup-plemental material.

Biofilm pH imaging. In situ-grown biofilms were placed in custom-made wells filled with C-SNARF-4 (50 �M) and salivary solution pre-pared according to the method of De Jong et al. (40). For biofilm pHimaging, the microscope was kept at 37°C with an XL incubator(PeCon, Erbach, Germany). Glucose was added to a concentration of0.4% (wt/vol), and in each biofilm, one microscopic field of view wasimaged every 30 s for 5 min immediately after the addition of glucose.Images were acquired at the bottom of the biofilm, just above thebiofilm-glass interface, since pH in the bottommost layer of dentalbiofilms has the biggest influence on the caries process. For back-ground subtraction, images were acquired with the 543-nm laserswitched off. Thereafter, biofilms were counterstained with BacLightand the same fields of view were imaged to verify that the entire bac-terial biomass was visualized with C-SNARF-4. Six replicate biofilmswere studied.

Biofilm pH analysis. In order to exclusively determine extracellularpH in the biofilms, the program daime (digital image analysis in microbialecology, v.2.0) was employed to remove the entire bacterial biomass fromall pictures (41). Red- and green-channel C-SNARF-4 biofilm imageswere exported separately as TIF files from the microscope software (LSMImage Examiner; Zeiss, Jena, Germany). In daime, green-channelC-SNARF-4 images were segmented with individually chosen brightnessthresholds using the “Automatic segmentation” “Custom threshold”function. During segmentation, bacterial cells were recognized as ob-jects. The object layer of the segmented images then was transferred tothe corresponding red-channel C-SNARF-4 images. Thereafter, theobject editor function was used to reject and delete all objects in bothgreen- and red-channel images, leaving only the extracellular matrix inthe images. In ImageJ (http://rsb.info.nih.gov/ij; v.1.47), backgroundfluorescence was subtracted, the mean filter (radius, 1 pixel) was em-ployed to compensate for detector noise, and the ratio between fluo-

rescence intensities detected with the green (576 to 608 nm) and red(629 to 661 nm) filter sets was determined for every extracellular pixel.Average ratios and standard deviations were calculated for each field ofview and converted to pH values according to equation 1.

FIG 1 Cells of different bacterial species in HEPES buffer were imaged with aconfocal microscope after staining with C-SNARF-4 (A, C, E, G, and I) andcounterstaining with BacLight (B, D, F, H, and J). (A and B) F. nucleatumsubsp. polymorphum, kept at pH 5.0. (C and D) L. paracasei subsp. paracasei,pH 4.5. (E and F) P. gingivalis, pH 5.0. (G and H) B. dentium, pH 4.0. (I and J)S. sanguinis, pH 4.5. At low pH (4.0 to 5.5), all cells of all organisms were reliablystained by C-SNARF-4 compared to the positive-control stain. Both viable andmembrane-compromised cells (red in the BacLight image) were targeted by C-SNARF-4. Bars, 20 �m.

Ratiometric pH Imaging in Biofilms with C-SNARF-4

February 2015 Volume 81 Number 4 aem.asm.org 1269Applied and Environmental Microbiology

on April 30, 2019 by guest

http://aem.asm

.org/D

ownloaded from

RESULTSStaining of bacterial species with C-SNARF-4. All 15 bacterialspecies included in the study were successfully stained with C-SNARF-4 at low pH (4 to 5.5). Confocal microscopic imagingshowed that the dye was up-concentrated in the cells, which madethem clearly distinguishable from the buffer background. Stainingwith 20 �M C-SNARF-4 yielded sufficient contrast between cellsand the background, but a concentration of 50 �M was judged togive the best cell/background ratio (see Fig. S1 in the supplementalmaterial). Counterstaining with BacLight and superimposition ofthe corresponding C-SNARF-4 and BacLight images showed thatall cells took up C-SNARF-4 and that both viable and membrane-compromised cells were visualized. Figure 1 shows selected bacte-rial species stained with C-SNARF-4 and the correspondingBacLight images (see Fig. S3 in the supplemental material for theother investigated bacterial species).

At pH values above 6.0, the contrast between bacterial cellsstained with C-SNARF-4 and the background decreased, and cellswere more difficult to identify (Fig. 2 depicts examples). Table 1lists the pH ranges suitable for the visualization of each organism.

Staining of dental biofilms with C-SNARF-4. In vivo-growndental biofilms typically reached a thickness of 15 to 35 �m at theend of the 48-h growth phase. C-SNARF-4 was able to visualizebacteria in all layers of these dental biofilms kept at low pH (4.0 to5.5). Counterstaining with BacLight proved that the entire bacte-rial biomass was visualized with C-SNARF-4 (Fig. 3A to D). Asseen for planktonic cells, the contrast between cells and extracel-

lular matrix decreased at higher pH, and the bacterial biomass wasmore difficult to identify (Fig. 3E and F). In thicker dental biofilms(up to 120 �m) exposed to 0.4% (wt/vol) glucose, a sharp contrastbetween bacterial cells and extracellular matrix was observed up toa depth of 75 �m (see Fig. S4 in the supplemental material). Indeeper layers of the biofilm, the contrast decreased, and differen-tiation between extra- and intracellular compartments becamedifficult.

Calculation of extracellular pH in dental biofilms stainedwith C-SNARF-4. When C-SNARF-4-stained biofilms were incu-bated with 0.4% (wt/vol) glucose, the bacteria became visiblewithin a very short time (�1 min). Digital image analysis withdaime reliably identified and removed the bacterial biomass in allbiofilm images, leaving only the extracellular compartment (seeFig. S5 in the supplemental material). Ratiometric pH analysis ofthe C-SNARF-4-stained biofilms incubated with glucose revealedthat the bacterial biomass was visible already when the average pHin the microscopic field of view was well above pH 6. Figure 4shows a biofilm with rapid acid production, where the average pHdrops to 5.31 within 5 min (Fig. 4A to E), and a biofilm withmoderate acid production, where average pH drops to 6.26 within5 min (Fig. 4F to J). In both biofilms, the bacterial biomass isdistinguishable from the first image (Fig. 4A and F), as shown byBacLight counterstaining (Fig. 4B and G). The average extracellu-lar pHs at the time point the first images were taken were 6.49 (Fig.4C) and 6.75 (Fig. 4H), respectively.

FIG 2 When kept in HEPES buffer at higher pH (6.0 to 7.0), the contrast between C-SNARF-4-stained bacteria and the background decreases and cells becomedifficult to identify. (A) F. nucleatum subsp. polymorphum, pH 6.0. (B) V. parvula, pH 6.0. (C) A. viscosus, pH 6.0. (D) S. oralis, pH 6.5. Bars, 20 �m.

TABLE 1 Visualization of bacterial species kept in HEPES buffer at different pH values with C-SNARF-4

Strain

Visualization at pHa:

4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0

A. naeslundii AK 6 � � � (�) (�) � � � �A. viscosus CCUG 33710 � � � � (�) � � � �B. dentium DSM 20436 � � � (�) � � � � �E. faecalis DSM 20478 � � � � � � � � �F. nucleatum ATCC 10953 � � � (�) � � � � �L. paracasei DSM 20020 � � � (�) � � � � �N. subflava DSM 17610 � � � � (�) � � � �P. gingivalis ATCC 33277 � � � � � � � � �P. intermedia CCUG 24041 � � � � � � � � �S. gordonii ATCC 10558 � � � (�) � � � � �S. mitis SK24 � � � (�) � � � � �S. mutans DSM 20523 � � � (�) � � � � �S. oralis NCTC 7864 � � � � � � � � �S. sanguinis ATCC 10556 � � � � � � � � �V. parvula CCUG 5123 � � � � (�) � � � �a � indicates that a strain was visualized well at a given pH. (�) indicates that a strain was visualized but with low contrast compared to the extracellular background. � indicatesthat a strain could not be distinguished from the extracellular background at a given pH.

Schlafer et al.

1270 aem.asm.org February 2015 Volume 81 Number 4Applied and Environmental Microbiology

on April 30, 2019 by guest

http://aem.asm

.org/D

ownloaded from

DISCUSSION

pH in biofilms is an ecological determinant for bacterial metabo-lism. In dental biofilm, it is the key virulence factor for the devel-opment of caries lesions, and only fluorescence microscopic tech-niques offer the possibility to monitor biofilm pH in differentmicroenvironments in real time. Fluorescence life-time imaging(FLIM), pH-sensitive nanoparticles, and pH-sensitive ratiometricdyes have been employed to measure biofilm pH microscopically.Although none of these techniques rely on a uniform distributionof the fluorescent dye in the biofilms for pH calculations, the dyeconcentration has to be high enough in all areas of the biofilm toallow for quantification. This proved to be a problem for pH-sensitive nanoparticles, which mainly bound to bacterial surfacesand left channels and voids filled with extracellular matrix un-stained (32). In dental biofilms, pH in these areas is of crucialimportance for the outcome of caries. The present study provesthat the ratiometric dye C-SNARF-4 penetrates well into young invivo-grown multispecies biofilms and allows the recording of pHin all areas of interest (Fig. 4).

Extracellular pH and intracellular bacterial pH differ consider-ably in biofilms. Quantitative microscopic techniques should be

able to differentiate reliably between these compartments, butnone of the hitherto published reports have addressed this prob-lem sufficiently. Vroom et al. employed carboxyfluorescein forFLIM and claimed that the charged carboxy group prevents thedye from penetrating the cell (31). With a pKa of �6.5, however,most of the molecule is protonated and, consequently, unchargedat low pH. In the published confocal microscopic images, bacterialcells are clearly visible in carboxyfluorescein-stained biofilms,which indicates that the dye was up-concentrated in the cells. As aconsequence, the displayed pH values represent a mixture of pHin intra- and extracellular compartments. Likewise, Hunterand Beveridge, who were the first to use C-SNARF-4 in a bac-terial biofilm, report that Pseudomonas aeruginosa cells internal-ized the dye and were discernible from the background (34).However, no measures were taken to remove intracellular areasfrom the microscopic images before pH calculation. Franks etal. employed C-SNARF-4 in current-generating Geobacter sul-furreducens biofilms (35), using the same approach as Hunterand Beveridge. Without differentiating between extracellularand intracellular compartments, pH was calculated for squareareas in the analyzed microscopic fields of view, yielding anaverage of intra- and extracellular pH.

In order to truly map extracellular pH in a biofilm, it is criticalto visualize the bacterial biomass in the biofilm and exclude itfrom pH calculations. Using a second, non-pH-sensitive dye tostain bacteria and later remove them via digital image analysiswould induce the risk of fluorescent contamination of the extra-cellular space, which might compromise pH calculations. Hunterand Beveridge used green fluorescent protein (GFP) to visualize P.aeruginosa, and there is an overlap between the fluorescent emis-sions of C-SNARF-4 and GFP. The fluorescent emission ofmCherry, used by Franks et al. to visualize G. sulfurreducens cells,peaks at 611 nm and shows considerable overlapping of the red-spectrum emission of C-SNARF-4. Leakage of these dyes into theextracellular compartment compromises fluorescence intensityratios calculated for pH measurement.

Therefore, in the present study, we investigated the use of C-SNARF-4 in a double function, as both a pH-sensitive ratiometricdye and a bacterial stain. We chose 15 different bacterial strainscommonly isolated from supragingival dental biofilm and ob-served that C-SNARF-4 stained both viable and membrane-compromised cells at low pH (4.0 to 5.5) (Fig. 1; see also Fig. S3in the supplemental material). Moreover, we showed that C-SNARF-4 stained the entire bacterial biomass in in vivo-growndental biofilms of unknown bacterial composition (Fig. 3A toD and 4). This suggests that C-SNARF-4 serves as a universalbacterial stain.

Following image acquisition, we used the digital image analysissoftware daime to determine and remove the bacterial biomassfrom the microscopic images (see Fig. S5 in the supplementalmaterial). Thereafter, fluorescent ratios could be calculated exclu-sively for the extracellular compartment and converted to pH val-ues (Fig. 4). In principle, digital image analysis might well be em-ployed to remove the extracellular space in the biofilm images andcalculate fluorescent ratios inside bacterial cells. However, so farthe calibration of C-SNARF-4 has been performed only for theextracellular space, using characteristic biofilm matrix compo-nents (34) and unstained biofilms as controls (39). IntracellularpH recordings would require a thorough calibration of the dyeprior to quantitative interpretation of the fluorescence ratios (42).

FIG 3 In vivo-grown dental biofilms, kept in HEPES buffer at pH 4.0 (Aand B), pH 5.5 (C and D), and pH 7.0 (E and F), were stained with C-SNARF-4 (A, C, and E) and counterstained with BacLight (B, D, and F). Inthe range between pH 4.0 and pH 6.0, the entire bacterial biomass is reli-ably stained with C-SNARF-4 compared to the positive-control stain. Bothviable and membrane-compromised cells are stained with C-SNARF-4. AtpH values above 6.0, the contrast between C-SNARF-4-stained bacteria inthe biofilms and the background decreases, and the bacterial biomass be-comes hard to identify. Bars, 20 �m.

Ratiometric pH Imaging in Biofilms with C-SNARF-4

February 2015 Volume 81 Number 4 aem.asm.org 1271Applied and Environmental Microbiology

on April 30, 2019 by guest

http://aem.asm

.org/D

ownloaded from

At higher pH (5.5 to 8.0), the staining properties of C-SNARF-4 decreased, and it became difficult to distinguish be-tween bacterial cells and the extracellular space, both in pure cul-tures (Fig. 2) and in dental biofilms (Fig. 3E and F). C-SNARF-4has a pKa of �6.4, and only at low pH are the majority of the dyemolecules protonated and uncharged. Only the uncharged mole-cule penetrates the bacterial membrane, binds to intracellularstructures, and is up-concentrated in the bacteria. Therefore, ra-tiometric pH analysis with C-SNARF-4 is limited to biofilms inacidic environments.

Interestingly, cells in C-SNARF-4-stained dental biofilms in-cubated with glucose became visible immediately after exposureto glucose. The calculation of average pH in the imaged micro-scopic fields of view showed that the biofilms were clearly distin-guishable when the average pH was still well above 6.0 (i.e., 6.75)(Fig. 4). The most likely explanation for this observation is thatbacterial acid production lowered the pH in the immediate sur-roundings of the cell, so that C-SNARF-4 was up-concentrated inthe cells, while the average pH in the microscopic field of view stillwas high. This increases the applicability of C-SNARF-4 for mon-itoring extracellular pH in acid-producing biofilms.

Compared to FLIM and the use of pH-sensitive nanoparticles,ratiometric imaging with the pH-sensitive fluorescent probe C-SNARF-4 is cheaper and requires less technical equipment. Theprobe readily penetrates bacterial biofilms and equilibratesquickly, permitting the monitoring of biofilm pH immediatelyafter growth. This might be of particular importance for pH anal-ysis of in vivo-grown samples. pH ratiometry with C-SNARF-4 islimited by the use of confocal microscopy for image acquisition, asthicker biofilms cannot be imaged in their entirety with a confocalmicroscope due to the absorption of excitation energy and thescattering of photons by the specimen. In thick dental biofilms,reliable differentiation between extra- and intracellular fluores-

cence was possible up to a depth of ca. 75 �m (see Fig. S4 in thesupplemental material). To record pH z profiles in thicker bio-films, microelectrodes remain the method of choice.

We proved here that C-SNARF-4 can be used as both a bacte-rial stain and a pH-sensitive ratiometric probe. The combinationof confocal microscopic imaging and digital image analysis per-mits the removal of the bacterial biomass from biofilm images andmonitoring of extracellular pH microenvironments in real time inbacterial biofilms.

ACKNOWLEDGMENTS

This work was funded within the Danish National Advanced TechnologyFoundation platform ProSURF 016-2006-2.

We thank Rikke Meyer and Duncan Sutherland (iNANO, Science andTechnology, Aarhus University, Denmark) for fruitful discussions andMette Nikolajsen and Lene Grønkjær for excellent technical support. Mo-gens Kilian (Department of Biomedicine, Aarhus University, Denmark) isacknowledged for the kind donation of bacterial strains.

REFERENCES1. Edwards SJ, Kjellerup BV. 2013. Applications of biofilms in bioremedia-

tion and biotransformation of persistent organic pollutants, pharmaceu-ticals/personal care products, and heavy metals. Appl Microbiol Biotech-nol 97:9909 –9921. http://dx.doi.org/10.1007/s00253-013-5216-z.

2. Halan B, Buehler K, Schmid A. 2012. Biofilms as living catalysts incontinuous chemical syntheses. Trends Biotechnol 30:453– 465. http://dx.doi.org/10.1016/j.tibtech.2012.05.003.

3. Deng H, Chen Z, Zhao F. 2012. Energy from plants and microorganisms:progress in plant-microbial fuel cells. ChemSusChem 5:1006 –1011. http://dx.doi.org/10.1002/cssc.201100257.

4. Bogino PC, Oliva ML, Sorroche FG, Giordano W. 2013. The role ofbacterial biofilms and surface components in plant-bacterial associations.Int J Mol Sci 14:15838 –15859. http://dx.doi.org/10.3390/ijms140815838.

5. Visick KL, Ruby EG. 2006. Vibrio fischeri and its host: it takes two totango. Curr Opin Microbiol 9:632– 638. http://dx.doi.org/10.1016/j.mib.2006.10.001.

FIG 4 pH was monitored ratiometrically with C-SNARF-4 in two in vivo-grown dental biofilms kept in sterile saliva (A to E and F to J) for 5 min after exposureto 0.4% (wt/vol) glucose. Images in panels A and F were acquired right after the addition of glucose. In both cases, the bacterial biomass is clearly distinguishable,as verified by BacLight staining (B and G). (C and H) Initial C-SNARF-4 images after removal of the bacterial biomass and ratiometric pH calculation. Forvisualization, false colors were assigned to each pH. Average pH in the images is 6.49 (C) and 6.75 (H). (D and I) The same fields of view as those shown in panelsC and H 5 min after acquisition of the first images. (E and J) The bacterial biomass was removed, pH was calculated, and false colors were applied. (E) In the firstbiofilm, the average pH in the microscopic field of view dropped to 5.31 within 5 min. (J) In the second biofilm, acid production was lower and pH was 6.26 after5 min. Bars, 20 �m.

Schlafer et al.

1272 aem.asm.org February 2015 Volume 81 Number 4Applied and Environmental Microbiology

on April 30, 2019 by guest

http://aem.asm

.org/D

ownloaded from

6. Van Wey AS, Cookson AL, Roy NC, McNabb WC, Soboleva TK,Shorten PR. 2011. Bacterial biofilms associated with food particles in thehuman large bowel. Mol Nutr Food Res 55:969 –978. http://dx.doi.org/10.1002/mnfr.201000589.

7. Taylor CJ, Mahenthiralingam E. 2006. Functional foods and paediatricgastro-intestinal health and disease. Ann Trop Paediatr 26:79 – 86. http://dx.doi.org/10.1179/146532806X107403.

8. Zottola EA, Sasahara KC. 1994. Microbial biofilms in the food processingindustry–should they be a concern? Int J Food Microbiol 23:125–148.http://dx.doi.org/10.1016/0168-1605(94)90047-7.

9. Mittelman MW. 1998. Structure and functional characteristics of bacte-rial biofilms in fluid processing operations. J Dairy Sci 81:2760 –2764.http://dx.doi.org/10.3168/jds.S0022-0302(98)75833-3.

10. Khardori N, Yassien M. 1995. Biofilms in device-related infections. J IndMicrobiol 15:141–147. http://dx.doi.org/10.1007/BF01569817.

11. Scali C, Kunimoto B. 2013. An update on chronic wounds and the role ofbiofilms. J Cutan Med Surg 17:371–376.

12. Ferro JM, Fonseca AC. 2014. Infective endocarditis. Handb Clin Neurol119:75–91. http://dx.doi.org/10.1016/B978-0-7020-4086-3.00007-2.

13. Hoiby N, Ciofu O, Bjarnsholt T. 2010. Pseudomonas aeruginosa bio-films in cystic fibrosis. Future Microbiol 5:1663–1674. http://dx.doi.org/10.2217/fmb.10.125.

14. Darveau RP. 2010. Periodontitis: a polymicrobial disruption of hosthomeostasis. Nat Rev Microbiol 8:481– 490. http://dx.doi.org/10.1038/nrmicro2337.

15. Takahashi N, Nyvad B. 2008. Caries ecology revisited: microbial dynam-ics and the caries process. Caries Res 42:409 – 418. http://dx.doi.org/10.1159/000159604.

16. Donlan RM. 2002. Biofilms: microbial life on surfaces. Emerg Infect Dis8:881– 890. http://dx.doi.org/10.3201/eid0809.020063.

17. Stewart PS. 2003. Diffusion in biofilms. J Bacteriol 185:1485–1491. http://dx.doi.org/10.1128/JB.185.5.1485-1491.2003.

18. Flemming HC, Wingender J. 2010. The biofilm matrix. Nat Rev Micro-biol 8:623– 633. http://dx.doi.org/10.1038/nrmicro2415.

19. Finnegan L, Garcia-Melgares M, Gmerek T, Huddleston WR, Palmer A,Robertson A, Shapiro S, Unkles SE. 2011. A survey of culturable aerobicand anaerobic marine bacteria in de novo biofilm formation on naturalsubstrates in St. Andrews Bay, Scotland. Antonie Van Leeuwenhoek 100:399 – 404. http://dx.doi.org/10.1007/s10482-011-9594-x.

20. Li J, Helmerhorst EJ, Leone CW, Troxler RF, Yaskell T, Haffajee AD,Socransky SS, Oppenheim FG. 2004. Identification of early microbialcolonizers in human dental biofilm. J Appl Microbiol 97:1311–1318. http://dx.doi.org/10.1111/j.1365-2672.2004.02420.x.

21. Torres CI, Kato MA, Rittmann BE. 2008. Proton transport inside thebiofilm limits electrical current generation by anode-respiring bacteria.Biotechnol Bioeng 100:872– 881. http://dx.doi.org/10.1002/bit.21821.

22. Stephan RM, Miller BF. 1943. A quantitative method for evaluatingphysical and chemical agents which modify production of acids in bacte-rial plaques on human teeth. J Dent Res 22:45–51. http://dx.doi.org/10.1177/00220345430220010601.

23. Schachtele CF, Jensen ME. 1982. Comparison of methods for monitoringchanges in the pH of human dental plaque. J Dent Res 61:1117–1125.http://dx.doi.org/10.1177/00220345820610100201.

24. Lingstrom P, Imfeld T, Birkhed D. 1993. Comparison of three differentmethods for measurement of plaque-pH in humans after consumption ofsoft bread and potato chips. J Dent Res 72:865– 870. http://dx.doi.org/10.1177/00220345930720050601.

25. Revsbech NP. 2005. Analysis of microbial communities with electro-chemical microsensors and microscale biosensors. Methods Enzymol 397:147–166. http://dx.doi.org/10.1016/S0076-6879(05)97009-2.

26. von Ohle C, Gieseke A, Nistico L, Decker EM, DeBeer D, Stoodley P.2010. Real-time microsensor measurement of local metabolic activities inex vivo dental biofilms exposed to sucrose and treated with chlorhexidine.

Appl Environ Microbiol 76:2326 –2334. http://dx.doi.org/10.1128/AEM.02090-09.

27. Zaura E, Buijs MJ, ten Cate JM. 2002. The effects of the solubility ofartificial fissures on plaque pH. J Dent Res 81:567–571. http://dx.doi.org/10.1177/154405910208100813.

28. Wang Z, Deng H, Chen L, Xiao Y, Zhao F. 2013. In situ measurementsof dissolved oxygen, pH and redox potential of biocathode microenviron-ments using microelectrodes. Bioresour Technol 132:387–390. http://dx.doi.org/10.1016/j.biortech.2012.11.026.

29. Xiao Y, Wu S, Yang ZH, Wang ZJ, Yan CZ, Zhao F. 2013. In situprobing the effect of potentials on the microenvironment of heterotrophicdenitrification biofilm with microelectrodes. Chemosphere 93:1295–1300. http://dx.doi.org/10.1016/j.chemosphere.2013.06.065.

30. Miyanaga K, Terashi R, Kawai H, Unno H, Tanji Y. 2007. Biocidaleffect of cathodic protection on bacterial viability in biofilm attachedto carbon steel. Biotechnol Bioeng 97:850 – 857. http://dx.doi.org/10.1002/bit.21278.

31. Vroom JM, De Grauw KJ, Gerritsen HC, Bradshaw DJ, Marsh PD,Watson GK, Birmingham JJ, Allison C. 1999. Depth penetration anddetection of pH gradients in biofilms by two-photon excitation micros-copy. Appl Environ Microbiol 65:3502–3511.

32. Hidalgo G, Burns A, Herz E, Hay AG, Houston PL, Wiesner U, LionLW. 2009. Functional tomographic fluorescence imaging of pH microen-vironments in microbial biofilms by use of silica nanoparticle sensors.Appl Environ Microbiol 75:7426 –7435. http://dx.doi.org/10.1128/AEM.01220-09.

33. Marcotte N, Brouwer AM. 2005. Carboxy SNARF-4F as a fluorescent pHprobe for ensemble and fluorescence correlation spectroscopies. J PhysChem B 109:11819 –11828. http://dx.doi.org/10.1021/jp0510138.

34. Hunter RC, Beveridge TJ. 2005. Application of a pH-sensitive fluoro-probe (C-SNARF-4) for pH microenvironment analysis in Pseudomonasaeruginosa biofilms. Appl Environ Microbiol 71:2501–2510. http://dx.doi.org/10.1128/AEM.71.5.2501-2510.2005.

35. Franks AE, Nevin KP, Jia HF, Izallalen M, Woodard TL, Lovley DR.2009. Novel strategy for three-dimensional real-time imaging of microbialfuel cell communities: monitoring the inhibitory effects of proton accu-mulation within the anode biofilm. Energy Environ Sci 2:113–119. http://dx.doi.org/10.1039/b816445b.

36. Hunt DE, Jones JV, Dowell VR, Jr. 1986. Selective medium for theisolation of Bacteroides gingivalis. J Clin Microbiol 23:441– 445.

37. Jensen SB, Loe H, Schiott CR, Theliade E. 1968. Experimental gingivitisin man. 4. Vancomycin induced changes in bacterial plaque compositionas related to development of gingival inflammation. J Periodont Res3:284 –293.

38. Dige I, Nilsson H, Kilian M, Nyvad B. 2007. In situ identification ofstreptococci and other bacteria in initial dental biofilm by confocal laserscanning microscopy and fluorescence in situ hybridization. Eur J Oral Sci115:459 – 467. http://dx.doi.org/10.1111/j.1600-0722.2007.00494.x.

39. Schlafer S, Raarup MK, Meyer RL, Sutherland DS, Dige I, NyengaardJR, Nyvad B. 2011. pH landscapes in a novel five-species model of earlydental biofilm. PLoS One 6:e25299. http://dx.doi.org/10.1371/journal.pone.0025299.

40. de Jong MH, van der Hoeven JS, Hvan OS, Olijve JH. 1984. Growth oforal Streptococcus species and Actinomyces viscosus in human saliva.Appl Environ Microbiol 47:901–904.

41. Daims H, Lucker S, Wagner M. 2006. daime, a novel image analysisprogram for microbial ecology and biofilm research. Environ Microbiol8:200 –213. http://dx.doi.org/10.1111/j.1462-2920.2005.00880.x.

42. Martinez KA, Kitko RD, Mershon JP, Adcox HE, Malek KA, BerkmenMB, Slonczewski JL. 2012. Cytoplasmic pH response to acid stress inindividual cells of Escherichia coli and Bacillus subtilis observed by fluo-rescence ratio imaging microscopy. Appl Environ Microbiol 78:3706 –3714. http://dx.doi.org/10.1128/AEM.00354-12.

Ratiometric pH Imaging in Biofilms with C-SNARF-4

February 2015 Volume 81 Number 4 aem.asm.org 1273Applied and Environmental Microbiology

on April 30, 2019 by guest

http://aem.asm

.org/D

ownloaded from