Supporting Information Valm et al. 10.1073/pnas.1101134108 SI Materials and Methods Microbial Cell Culture. Escherichia coli K12 (ATCC 10798) were grown in LB broth (Becton Dickinson) at 37 °C under aerobic conditions and harvested at an OD 600 of 0.5. Streptococcus san- guinis (ATCC 10556), Neisseria sicca (ATCC 9913), Eikenella corrodens (ATCC 23834), Actinomyces naeslundii (ATCC 19039), and Haemophilus influenzae (ATCC 10211) were grown in Brain Heart Infusion (BHI) broth (Becton Dickinson) at 37 °C in static culture in ambient air supplemented with 10% CO 2 . Rothia dentocariosa (ATCC 17931) and Gemella bergeriae (ATCC 700627) were grown in BHI broth in static culture at 37 °C under normal aerobic atmosphere. Fusobacterium nucleatum subsp. nucleatum (ATCC 25586), Veillonella atypica (ATCC 17744), Selenomonas flueggei (ATCC 43531), Capnocytophaga gingivalis (ATCC 33624), Porphyromonas gingivalis (ATCC 33277), Lepto- trichia buccalis (ATCC 14201), and Campylobacter showae (ATCC 51146) were grown in prereduced BHI supplemented with L-cysteine, hemin, and vitamin K 1 (Beckton Dickinson) at 37 °C in static culture in an anaerobic chamber (Coy Laboratory Products) under an atmosphere of 90% N 2 , 10% H 2 , except that the medium for V. atypica was supplemented with 1.5% sodium lactate (Fisher Scientific), and the 5-mL medium for C. showae was supplemented with 0.25 mL of an aqueous solution of 6% each of formate (MP Biomedicals) and fumarate (Acros Or- ganics). Prevotella nigrescens (ATCC 25845) was grown in pre- reduced BHI supplemented with L-cysteine, hemin, and vitamin K 1 (Becton Dickinson) at 37 °C in static culture in an anaerobic GasPak jar with anaerobic atmosphere generated by a GasPak EZ anaerobe container system sachet (Becton Dickinson). Oral microbes were removed from incubation when the cultures became turbid; average growth time was 2 d (range 1–6 d). Treponema denticola (ATCC 35405) was procured from ATCC as a live culture in liquid broth and was fixed immediately upon receipt. Cells in their growth medium were fixed by addition of an equal volume of 4% paraformaldehyde (Electron Microscopy Sciences) followed by incubation at room temperature for 1.5 h. Fixed cells were washed two or three times in PBS, and then the final pellet was resuspended in 1× PBS, to which an equal volume of 100% ethyl alcohol (EtOH) was added. Cells were stored at −20 °C for up to 12 mo before fluorescence in situ hybridization (FISH). Oligonucleotide Probe Design. For E. coli experiments, eight ver- sions of the Eub338 probe (5′-GCT GCC TCC CGT AGG AGT- 3′) (1) were custom synthesized (Invitrogen). The Eub338 probe targets a conserved region of the bacterial 16S rRNA which is predicted to be present in most bacteria (http://www.microbial- ecology.net/probebase) and was verified to be present in E. coli. Each version of the probe had a different fluorophore conju- gated to its 5′ end from the following repertoire: Pacific Blue, Pacific Orange, BODIPY-FL, Oregon Green 514, Alexa Fluor 532, Alexa Fluor 546, Rhodamine Red-X, and Texas Red-X. Oligonucleotide probes used to label oral microbes were designed to target the 16S rRNA and are listed in Table S1. The probes used were identified as specific genus-level probes for target organisms in the literature or were designed de novo using the probe design function of the ARB program (www.arb-home. de) with a database of 16S sequences from the Human Oral Microbiome Database (HOMD) (www.homd.org). HOMD 16S rRNA RefSeq file v.10 was downloaded on November 7, 2008. Probes generally were designed with the following three criteria: They must be 18–21 nucleotides in length, must have no nucle- otide mismatches to all named species in the target genus, and must have at least two central mismatches to all nontarget spe- cies in the HOMD. Exceptions to these criteria for each probe are listed in Table S1. In addition, when it was possible to design more than one probe for a target taxon that met the above cri- teria, the overall change in free energy, ΔG, for probe binding to target was calculated for the candidate probes and considered when choosing a suitable probe (2). For two selected taxa, a genus-level probe could not be designed that met the above criteria, and so a family-level probe was designed instead. These probes targeted all oral microbes in the families Pasteurellaceae and Neisseriaceae. For Combinatorial Labeling and Spectral Imaging (CLASI)-FISH analyses, two versions of each of the 15 taxon-specific probes were synthesized. Both versions of each oligonucleotide probe were conjugated to one of two different fluorophores from the following repertoire of six, to give 15 unique binary fluorophore combinations: Alexa Fluor 488, Alexa Fluor 514, Alexa Fluor 555, Alexa Fluor 594, Alexa Fluor 647, and Rhodamine Red-X. FISH. Fixed E. coli cells were suspended in hybridization buffer [0.9 M NaCl, 0.02 M Tris (pH 7.5), 0.01% SDS, and 20% “Hi- Di” grade formamide (Applied Biosystems)]. To each of the 28 tubes, first one and then a second fluorophore-labeled probe was added to give a final probe concentration of 5 ng/μL for each probe. Tubes were incubated at 46 °C for 4 h. Cells were washed in wash 1 buffer [0.9 M NaCl, 0.02 M Tris (pH 7.5), 0.01% SDS, 20% formamide] for 15 min at 48 °C and then in wash 2 buffer [0.9 M NaCl, 0.02 M Tris (pH 7.5), 0.01% SDS] for 15 min at 48 °C. Cells were resuspended in 0.025 M NaCl + 0.02 M Tris (pH 7.5). Five microliters of each of the cell suspensions from each tube was pipetted into a single tube to give a mixture which then was spotted on an Ultrastick slide (Thermofisher). The slide was placed in a humid chamber for 60 min at room temperature to allow cells to settle and then was rinsed very briefly in ice-cold EtOH and allowed to air dry. The dried specimen was mounted in Vectashield antifade solution (Vector Laboratories). FISH on cultured oral microbes and dispersed dental plaque was performed in Eppendorf tubes. Fixed cultured cells were suspended in hybridization buffer and then were combined into one tube. An aliquot of fixed dental plaque was suspended in hybridization buffer. To these tubes, all 30 of the taxon-specific probes were added to give a final probe concentration of 2 μM for each probe. In addition, an aliquot of Eub338 probe conjugated to Pacific Blue was added to the microbe mixture tubes and the dispersed plaque tubes at a final concentration of 2 μM. In both the mixtures of laboratory-grown microbes and dispersed dental plaque samples, cells of the genera Streptococcus, Gemella, and Leptotrichia were considerably brighter than all other cells. Therefore, to reduce the overall dynamic range in recorded spectral images, three unlabeled oligonucleotides were added to all oral microbe FISH reaction tubes to compete with fluo- rophore-conjugated probes for target sequences. These probes were the STR405 probe at 2 μM, the GEM572 probe at 1 μM, and the LEP568 probe at 0.5 μM. The tubes were incubated at 46 °C for 18 h. Cells were washed as above; then 40 μL of cell suspension was pipetted onto custom Excel Adhesion slides (Thermofisher), and the slides were placed in a humid chamber for 60 min at room temperature. The slides were dehydrated in an ethanol series, mounted in ProLong Gold antifade reagent (Invitrogen), and allowed to cure for 72 h at room temperature in the dark before being imaged. Valm et al. www.pnas.org/cgi/content/short/1101134108 1 of 11
11
Embed
Supporting Information - Proceedings of the National ... · PDF fileSupporting Information ... of the image sets using reference spectra appropriate for that ... The linear unmixing
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Supporting InformationValm et al. 10.1073/pnas.1101134108SI Materials and MethodsMicrobial Cell Culture. Escherichia coli K12 (ATCC 10798) weregrown in LB broth (Becton Dickinson) at 37 °C under aerobicconditions and harvested at an OD600 of 0.5. Streptococcus san-guinis (ATCC 10556), Neisseria sicca (ATCC 9913), Eikenellacorrodens (ATCC 23834), Actinomyces naeslundii (ATCC 19039),and Haemophilus influenzae (ATCC 10211) were grown in BrainHeart Infusion (BHI) broth (Becton Dickinson) at 37 °C in staticculture in ambient air supplemented with 10% CO2. Rothiadentocariosa (ATCC 17931) and Gemella bergeriae (ATCC700627) were grown in BHI broth in static culture at 37 °C undernormal aerobic atmosphere. Fusobacterium nucleatum subsp.nucleatum (ATCC 25586), Veillonella atypica (ATCC 17744),Selenomonas flueggei (ATCC 43531), Capnocytophaga gingivalis(ATCC 33624), Porphyromonas gingivalis (ATCC 33277), Lepto-trichia buccalis (ATCC 14201), and Campylobacter showae(ATCC 51146) were grown in prereduced BHI supplementedwith L-cysteine, hemin, and vitamin K1 (Beckton Dickinson) at37 °C in static culture in an anaerobic chamber (Coy LaboratoryProducts) under an atmosphere of 90% N2, 10% H2, except thatthe medium for V. atypica was supplemented with 1.5% sodiumlactate (Fisher Scientific), and the 5-mL medium for C. showaewas supplemented with 0.25 mL of an aqueous solution of 6%each of formate (MP Biomedicals) and fumarate (Acros Or-ganics). Prevotella nigrescens (ATCC 25845) was grown in pre-reduced BHI supplemented with L-cysteine, hemin, and vitaminK1 (Becton Dickinson) at 37 °C in static culture in an anaerobicGasPak jar with anaerobic atmosphere generated by a GasPakEZ anaerobe container system sachet (Becton Dickinson).Oral microbes were removed from incubation when the culturesbecame turbid; average growth time was 2 d (range 1–6 d).Treponema denticola (ATCC 35405) was procured from ATCCas a live culture in liquid broth and was fixed immediatelyupon receipt.Cells in their growth medium were fixed by addition of an equal
volume of 4% paraformaldehyde (Electron Microscopy Sciences)followed by incubation at room temperature for 1.5 h. Fixed cellswere washed two or three times in PBS, and then the final pelletwas resuspended in 1× PBS, to which an equal volume of 100%ethyl alcohol (EtOH) was added. Cells were stored at −20 °C forup to 12 mo before fluorescence in situ hybridization (FISH).
Oligonucleotide Probe Design. For E. coli experiments, eight ver-sions of the Eub338 probe (5′-GCT GCC TCC CGT AGG AGT-3′) (1) were custom synthesized (Invitrogen). The Eub338 probetargets a conserved region of the bacterial 16S rRNA which ispredicted to be present in most bacteria (http://www.microbial-ecology.net/probebase) and was verified to be present in E. coli.Each version of the probe had a different fluorophore conju-gated to its 5′ end from the following repertoire: Pacific Blue,Pacific Orange, BODIPY-FL, Oregon Green 514, Alexa Fluor532, Alexa Fluor 546, Rhodamine Red-X, and Texas Red-X.Oligonucleotide probes used to label oral microbes were
designed to target the 16S rRNA and are listed in Table S1. Theprobes used were identified as specific genus-level probes fortarget organisms in the literature or were designed de novo usingthe probe design function of the ARB program (www.arb-home.de) with a database of 16S sequences from the Human OralMicrobiome Database (HOMD) (www.homd.org). HOMD 16SrRNA RefSeq file v.10 was downloaded on November 7, 2008.Probes generally were designed with the following three criteria:They must be 18–21 nucleotides in length, must have no nucle-
otide mismatches to all named species in the target genus, andmust have at least two central mismatches to all nontarget spe-cies in the HOMD. Exceptions to these criteria for each probeare listed in Table S1. In addition, when it was possible to designmore than one probe for a target taxon that met the above cri-teria, the overall change in free energy, ΔG, for probe binding totarget was calculated for the candidate probes and consideredwhen choosing a suitable probe (2). For two selected taxa, agenus-level probe could not be designed that met the abovecriteria, and so a family-level probe was designed instead. Theseprobes targeted all oral microbes in the families Pasteurellaceaeand Neisseriaceae. For Combinatorial Labeling and SpectralImaging (CLASI)-FISH analyses, two versions of each of the 15taxon-specific probes were synthesized. Both versions of eacholigonucleotide probe were conjugated to one of two differentfluorophores from the following repertoire of six, to give 15unique binary fluorophore combinations: Alexa Fluor 488, AlexaFluor 514, Alexa Fluor 555, Alexa Fluor 594, Alexa Fluor 647,and Rhodamine Red-X.
FISH. Fixed E. coli cells were suspended in hybridization buffer[0.9 M NaCl, 0.02 M Tris (pH 7.5), 0.01% SDS, and 20% “Hi-Di” grade formamide (Applied Biosystems)]. To each of the 28tubes, first one and then a second fluorophore-labeled probe wasadded to give a final probe concentration of 5 ng/μL for eachprobe. Tubes were incubated at 46 °C for 4 h. Cells were washedin wash 1 buffer [0.9 M NaCl, 0.02 M Tris (pH 7.5), 0.01% SDS,20% formamide] for 15 min at 48 °C and then in wash 2 buffer[0.9 M NaCl, 0.02 M Tris (pH 7.5), 0.01% SDS] for 15 min at48 °C. Cells were resuspended in 0.025 M NaCl + 0.02 M Tris(pH 7.5). Five microliters of each of the cell suspensions fromeach tube was pipetted into a single tube to give a mixture whichthen was spotted on an Ultrastick slide (Thermofisher). The slidewas placed in a humid chamber for 60 min at room temperatureto allow cells to settle and then was rinsed very briefly in ice-coldEtOH and allowed to air dry. The dried specimen was mountedin Vectashield antifade solution (Vector Laboratories).FISH on cultured oral microbes and dispersed dental plaque
was performed in Eppendorf tubes. Fixed cultured cells weresuspended in hybridization buffer and then were combined intoone tube. An aliquot of fixed dental plaque was suspended inhybridization buffer. To these tubes, all 30 of the taxon-specificprobes were added to give a final probe concentration of 2 μM foreach probe. In addition, an aliquot of Eub338 probe conjugatedto Pacific Blue was added to the microbe mixture tubes and thedispersed plaque tubes at a final concentration of 2 μM. In boththe mixtures of laboratory-grown microbes and dispersed dentalplaque samples, cells of the genera Streptococcus, Gemella, andLeptotrichia were considerably brighter than all other cells.Therefore, to reduce the overall dynamic range in recordedspectral images, three unlabeled oligonucleotides were added toall oral microbe FISH reaction tubes to compete with fluo-rophore-conjugated probes for target sequences. These probeswere the STR405 probe at 2 μM, the GEM572 probe at 1 μM,and the LEP568 probe at 0.5 μM. The tubes were incubated at46 °C for 18 h. Cells were washed as above; then 40 μL of cellsuspension was pipetted onto custom Excel Adhesion slides(Thermofisher), and the slides were placed in a humid chamberfor 60 min at room temperature. The slides were dehydrated inan ethanol series, mounted in ProLong Gold antifade reagent(Invitrogen), and allowed to cure for 72 h at room temperature inthe dark before being imaged.
Valm et al. www.pnas.org/cgi/content/short/1101134108 1 of 11
Empirical Probe Specificity Testing. All oligonucleotide probes firstwere evaluated for specificity in silico by evaluating all probesequences against all 16S sequences in the HOMD, using thecriteria described in the text. All 15 probes then were testedempirically for specificity in control experiments against all 14nontarget organisms and their one target organism used in thisstudy. For probe testing, each probe was synthesized and con-jugated to Alexa Fluor 488 (Invitrogen). As a positive control, allcells were labeled simultaneously with the Eub338 probe con-jugated to Rhodamine Red-X. Eub338 labeled all cells, but notaxon-specific probe showed significant cross-hybridization tounintended targets under the hybridization conditions used.
Image Acquisition. Spectral images were acquired with eithera Nikon C1si (Nikon Instruments) or a Zeiss 710 (Carl Zeiss)laser scanning confocal microscope equipped with a 32-channelmultianode spectral detector and three or five laser lines, re-spectively. Images were acquired with a 20×/0.8 NA objective lensor a 100×/1.4 NA objective (Nikon) or with a 63×/1.4 NA ob-jective (Zeiss) with 5-nm (Nikon) or 9.7-nm (Zeiss) channelwidths, meaning that each of the available 32 anodes on thespectral detector collected light that corresponded to either a 5-or 9.7-nm bandwidth of the visible spectrum. Spectral imageacquisitions were made for each field of view in order of de-scending excitation wavelength, 561 nm, 488 nm, and 405 nm(Nikon) or 633 nm, 594 nm, 561 nm, and 488 nm (Zeiss). Allspectral images were acquired as the line average of three (Ni-kon) or four (Zeiss) scans. In addition, for experiments donewith the Zeiss instrument, a standard fluorescence image wasacquired with 405-nm excitation as the line average of 16 ac-quisitions for each field of view for collection of the Eub338-Pacific Blue signal.
Image Analysis. Spectral images were processed first with NikonEZc1 or Zeiss Zen software. Linear unmixing was applied to eachof the image sets using reference spectra appropriate for thatacquisition. The reference spectra used for unmixing each exci-tation included all the fluorophores in the experiment that wereexcited by that laser line excitation wavelength (even if they wereexcited more efficiently by another wavelength used in the ac-quisition) and background. After unmixing of all image sets takenusing different excitation wavelengths, some fluorophores wererepresented by unmixed channels from more than one excitation.For these fluorophores, the channel taken with the less favorableexcitation wavelength was discarded from the analysis because itwas redundant and was of lower signal-to-noise quality than theacquisition with the more favorable excitation wavelength. Forexample, the 561-nm excitation image sets generated unmixedchannels for Alexa Fluor 532, Alexa Fluor 546, Rhodamine Red-X, and Texas Red-X, whereas the 488-nm excitation acquisitiongenerated unmixed channels for Bodipy-Fl, Oregon Green 514,Alexa Fluor 532, Alexa Fluor 546, and Rhodamine Red-X. Boththese two excitation wavelengths excited Alexa Fluor 532, AlexaFluor 546, and Rhodamine Red-X, but the more favorable ex-citation for Alexa Fluor 532 was 488 nm, and the more favorableexcitation for Alexa Fluor 546 and Rhodamine Red-X was 561nm. Consequently, the Alexa Fluor 532 channel was discardedfrom analysis of the 561-nm excitation, and the other three flu-orophore channels were kept. The same comparative procedurewas used throughout spectral imaging analysis. All referencespectra were measured in separate singly-labeled E. coli imagedusing the same acquisition conditions as the CLASI-FISH im-ages. The linear unmixing operation resulted in image stacks inTIFF format; these stacks were opened in ImageJ (3) or Math-ematica (Wolfram Research) for further analysis. The pixel-based images were converted to a particle basis by segmentingcells from background and from each other. The average in-tensity for each cell (averaged over all pixels in each particle) in
each of the fluorophore channels then was measured, and thetwo highest fluorophores were identified to assign the cell itslabel type. For every particle, if the intensity of the third highestfluorophore was >60% of the second highest fluorophore, thecell was declared ambiguous. Ambiguity is assumed to resultfrom the presence of cells of different taxa that touch in theplane of focus but are not separated by the segmentation pro-cedure described above or from cells of different taxa over-lapping in the z-dimension within the plane of focus of theconfocal microscope image.
Image Presentation. E. coli and laboratory-grown oral microbeartificial mixture images are presented in the text as raw spectralimages in which the three or five separate excitation acquisitionswere unmixed using microscope software, either Nikon EZc1 forE. coli or Zeiss Zen for oral microbes, and each of the fluo-rophore channels was pseudocolored. The channels then weremerged using the logical operator “OR” to create the raw spectralimage, in which the visible color is the summation of the in-dividual channel pseudocolors. The binary assigned images werecreated in Mathematica. After cells were segmented from back-ground and particle binary combination was identified for eachcell, the cells were pseudocolored in 1 of 28 or 15 different colorscorresponding to its binary label type or taxon identity.
Input Cell Counting. Inputs of fixed microbial cell populations usedto create mixtures were counted using a semiautomated cell-counting procedure. E. coli were imaged in phase contrast on aZeiss Axioskop 2 equipped with 20×/0.5 NA phase objective(Carl Zeiss). An equal volume of cell suspension from each ofthe 28 different populations of FISH-labeled E. coli was pipettedinto one chamber of a Bright-Line hemocytometer (AmericanOptical). Cells were allowed to settle on the slide for 20 minbefore counting. Three 40-μm squares on each side of the centralcounting region were chosen randomly for counting, for a totalof six measurements. Images of each counting region were ac-quired with a Zeiss Axiocam. Images were imported into ImageJ,and cells were segmented from background and counted. Cellcounts are the mean of six measurements.An equal volume of fixed microbe populations of each of the 15
taxa used in the mixed oral species experiment were spotted ontoGold Seal Ultrastick slides for counting (Thermo Fisher). Cellswere allowed to settle on coated slides for 60 min in a humidchamber. Slides were rinsed briefly 2× in 100% EtOH and thenwere air dried. Slides were mounted in ProLong Gold antifademedium (Invitrogen). Images of five separate fields of view ac-quired with a 63×/1.4 NA objective were imported into ImageJfor automated counting as described for E. coli. Cell counts re-ported are the mean of five measurements. Correlation value, r,was calculated as the Pearson’s correlation coefficient betweenthe mean values of input into the artificial mixture and meanvalues of output as measured with CLASI-FISH.
Spatial Analysis. A computer program to analyze the spatial dis-tribution of microbes in semidispersed dental plaque was de-veloped on the Mathematica platform. Briefly, taxon-identifiedcells in segmented plaque images were subjected to two rounds ofimage dilation. Features were dilated two pixels, corresponding toa nondiagonal distance in the imaged sample of 264 nm. Thedifference between the number of noncontiguous features innondilated and dilated images was used as an estimate of thenumber of cells that touch. Thus, in this analysis, two cells touch,by definition, if they are within four pixels, or 528 nm, of eachother. This distance, which is approximately twice the Rayleighresolution limit in the acquired images, is assumed to be slightlygreater than the error in segmentation that might artificiallyreduce the apparent size of a microbe in the binary image usingthe segmentation procedure described above. This analysis was
Valm et al. www.pnas.org/cgi/content/short/1101134108 2 of 11
applied sequentially to all possible binary subsets of the 15-taxonsegmented images, including the 15 self-touching subsets, to give105 + 15 = 120 independent measurements of inter- and in-traspecies touching occurrences in each field of view acquired. Inaddition, we compared the results from our spatial analysisprogram with results obtained using the cluster analysis functionof DAIME (4), and concluded that similar results were obtainedusing both methods.
Image Modeling. In 15 separate experiments, aliquots of fixedhuman dental plaque were labeled with 1 of the 15 genus- orfamily-specific FISH probes. Images of these labeled cells wereacquired and processed using the same protocol used for CLASI-FISH–labeled plaque. In ImageJ, features from segmented im-ages were fit to ellipses to determine the length of the major andminor axes for every cell in the images. These data were im-ported into Mathematica where the mean lengths of the majorand minor axes for cells of each taxon were calculated and usedto create model ellipsoid cells. Model images were constructedwith the mean number of each type of cell observed in 24 fieldsof view of CLASI-FISH–labeled plaque. A random-numbergenerator was used to define the x,y center locations of eachmodel cell in the image as well as the orientations of each cell. Intotal, 100 model images were constructed. Model images thenwere subjected to the same spatial analysis program as images ofCLASI-FISH–labeled plaque.
Spatial Plot. We analyzed the spatial interrelationships of mi-crobes in partially dispersed plaque by assessing the frequencywith which cells of the same taxon or cells of different taxatouched each other. Two meaningful frequencies exist: the frac-tion of cells of type A that touch type B, and the fraction of cells oftype B that touch type A. We refer to the lower-abundance taxonfrom any pair as the “target cell” and higher-abundance taxon asthe “base cell.” To determine if the number of touching events
observed in images of partially dispersed plaque occurred moreoften than would be expected from random associations of themicrobes present, the 225 associations in each of 24 fields of viewcalculated for each of the target and base cells were comparedwith the frequencies of percent associations in model images ofrandomly placed cells with a one-tailed Student t test. For thisanalysis, the null hypothesis was set to 0. For all populations withP values ≤0.05, the percent association in random images wassubtracted from the percent association in observed images togive the final frequency of percent association (Tables S2–S5).The lower-left triangular matrix of these associations, not in-cluding the main diagonal, was used to construct a graph of allintergenus associations observed in dispersed dental plaque. Forevery intergenus association there are two measures of percent,the percent of the first genus that was found to associate with thesecond, and the percent of the second genus that was found toassociate with the first. By taking only the lower-left triangularmatrix, we include in the spatial analysis plot every possiblepercent association only once, namely the percent of the lower-abundance genus that associates with the higher-abundance ge-nus. The spatial analysis plot was created in Mathematica. Thearea of each circle reflects the relative abundance of each genusor family observed in all 24 fields of view of semidispersed dentalplaque, by the formula
radius of circle ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiabundance
π
r:
The positions of each circle in the plot were determined using theSpring Electrical Embedding option in the Mathematica Graph-Plot function. A line connecting two circles indicates that cellsof the lower-abundance taxon were found to associate with cellsof the higher-abundance taxon to which it is connected withP value ≥0.05 and with frequency ≥3%.
1. Amann RI, et al. (1990) Combination of 16S rRNA-targeted oligonucleotide probes withflow cytometry for analyzing mixed microbial populations. Appl Environ Microbiol 56:1919–1925.
2. Yilmaz LS, Noguera DR (2004) Mechanistic approach to the problem of hybridizationefficiency in fluorescent in situ hybridization. Appl Environ Microbiol 70:7126–7139.
3. RasbandWS, Image J (2010) (US National Institutes of Health; Bethesda, MD). Availableat http://rsb.info.nih.gov/ij. Accessed July 29, 2010.
4. Daims H, Lücker S, Wagner M (2006) daime, a novel image analysis program formicrobial ecology and biofilm research. Environ Microbiol 8:200–213.
Valm et al. www.pnas.org/cgi/content/short/1101134108 3 of 11
Fig. S1. Pure populations of 28 different binary-labeled E. coli are correctly identified in their label type. The same labeled cells used to create the mixturedescribed in the text were spotted separately on microscope slides for imaging. Displayed here are pseudocolored images of each of the 28 pure populations ofbinary-labeled E. coli. (A) Spectral images after linear unmixing in which the color in every pixel corresponds to the merge of all eight fluorophore channels. (B)Binary label-type images, in which the same fields as in Awere segmented from background and then pseudocolored 1 of 28 colors representing each of the 28binary-label combinations. Each panel is a detail from a field of view acquired with a 20×/0.8 NA objective. (Scale bar: 50 μm.)
Valm et al. www.pnas.org/cgi/content/short/1101134108 4 of 11
Fig. S2. Quantitative analysis of pure populations of 28 different label types of E. coli. Histograms of number of cells in a field of view on the y axis plottedagainst intensity for all eight fluorophores used to label E. coli on the x axis. Histograms were generated after particle analysis performed on low-magnificationimages of pure populations of E. coli acquired and analyzed with settings identical to those used for the mixture described in the text. Most combinations showexactly two fluorophores with mean intensities much higher than the other six fluorophores used in the experiment, resulting in accurate binary label as-signment.
Valm et al. www.pnas.org/cgi/content/short/1101134108 5 of 11
Fig. S3. Relative abundance of each of the 15 probed taxa in dispersed human dental plaque. This dispersed plaque sample from a single donor is dominated bygenera that are known to colonize early andmostly on the supragingival surfaces of teeth or are known to be abundant in saliva, consistent with the dental plaquesampling protocol used. Error bars represent the SD in the measurement of the mean percent abundance from 24 fields of view from two separate FISH assays.
Fig. S4. Model cells and image to determine frequency of random associations. (A) Representative image details from 15 separate FISH-labeled samples ofsemidispersed human dental plaque. All cells are labeled with a taxon-specific probe conjugated to Alexa Fluor 488. (B) Images of labeled cells were processedusing the image-segmentation protocol used with CLASI-FISH–labeled plaque. Ellipses were fit to every cell, and the mean lengths of the major and minor axeswere computed and used to design model cells. (Scale bars in A and B: 5 μm.) (C) Model images of randomly placed cells were constructed and subjected to thesame spatial analysis as images of semidispersed plaque to determine the frequency of random associations. (Scale bar: 25 μm.)
Valm et al. www.pnas.org/cgi/content/short/1101134108 6 of 11
Table S1. Details of probes used in 15-taxon oral microbe mixture experiments
Target genus orfamily
Probename Probe sequence (5′–3′) Ref. Notes on specificity against HOMD sequences
Streptococcus STR405 TAGCCGTCCCTTTCTGGT 1 No mismatches to 54 representatives of Streptococcus inHOMD v10. One central mismatch to Streptococcus sp. OralTaxon 064. One noncentral mismatch to Lactococcus lactis OralTaxon 804.
Fusobacterium FUS714 GGCTTCCCCATCGGCATT * No mismatches to all 19 representatives of Fusobacterium.Leptotrichia LEP568 GCCTAGATGCCCTTTATG * No mismatches to all 20 representatives of Leptotrichia. One
Veillonella VEI488 CCGTGGCTTTCTATTCCG 2 No mismatches to all 12 representatives of Veillonella.Neisseriaceae NEI1030 CCTGTGTTACGGCTCCCG * No mismatches to 24 representatives of family Neisseriaceae.
One mismatch to Kingella kingae,Neisseria sp. oral taxon 523, Eikenella sp. oral taxon 11,Kingella sp. Oral taxon 12,Neisseria sp. oral taxon 16, Kingella sp. oral taxon 459, E. coli.
Porphyromonas PGI350 CCTCACGCCTTACGACGG * No mismatches to nine representatives of Porphyromonas.One mismatch to P. endodontalis, Porphyromonas sp.oral taxon 395, and P. asaccharolytica.
Capnocytophaga CAP371 TCAGTCTTCCGACCATTG 3 No mismatches to all 19 representatives of Capnocytophaga andSphaerocytophaga S3 sp. oral taxon 337, 1 central mismatchto Bacteroidetes [G-6], oral taxon 516.
Pasteurellaceae PAS111 TCCCAAGCATTACTCACC * No mismatches to 16 representatives of family Pasteurellaceae.One mismatch to Cardiobacterium valvulum, Acinetobactersp. C1 oral taxon 408, Acinetobacter baumannii.
Gemella GEM572 TAAACCACCTGCGCGCGCTT * No mismatches to all five representatives of Gemella. One mismatchto Lactococcus lactis, Bacillus fusiformis, Bacillus anthracis.
Selenomonas SEL60 TCATTCGCTCCGTTCGAC * No mismatches to all 17 representatives of Selenomonas.Prevotella PRV392 GCACGCTACTTGGCTGG 4 No mismatches to 68 representatives of Prevotella. One mismatch
Actinomyces ACT476 ATCCAGCTACCGTCAACC 5 No mismatches to 11 representatives of Actinomyces. Onemismatch to Actinomyces gerencseriae and A. oricola.
Rothia ROT491 TAGCCGGCGCTTTCTCTG * No mismatches to all three representatives of Rothia.One mismatch to Microbacterium sp. oral taxon 185 andMicrobacterium sp. oral taxon 186.
Campylobacter CAM1021 ATTTCTGCAAGCAGACACTC * No mismatches to all eight representatives of Campylobacter.Treponema TRP684 TCTACAGATTCCACCCCTAC * No mismatches to 39 representatives of Treponema. One
mismatch to 17 other representatives of Treponema.
All probes were designed to be specific for their target genus or family within the context of the 619 species represented in the Human Oral MicrobiomeDatabase (HOMD). Probes have no mismatches to all representatives of the target genus or family and have two or more central mismatches to all other taxa inthe database. Exceptions to these criteria are listed in the table.* Probe designed in this study.
1. Paster BJ, Bartoszyk IM, Dewhirst FE (1998) Identification of oral streptococci using PCR-based, reverse-capture, checkerboard hybridization. Methods Cell Sci 20:223–231.2. Chalmers NI, Palmer RJ Jr., Cisar JO, Kolenbrander PE (2008) Characterization of a Streptococcus sp.-Veillonella sp. community micromanipulated from dental plaque. J Bacteriol 190:
8145–8154.3. Zijnge V, et al. (2010) Oral biofilm architecture on natural teeth. PLoS ONE 5:39321.4. Diaz PI, et al. (2006) Molecular characterization of subject-specific oral microflora during initial colonization of enamel. Appl Environ Microbiol 72:2837–2848.5. Gmür R, Lüthi-Schaller H (2007) A combined immunofluorescence and fluorescent in situ hybridization assay for single cell analyses of dental plaque microorganisms. J Microbiol
Methods 69:402–405.
Valm et al. www.pnas.org/cgi/content/short/1101134108 7 of 11