Spatial organization of a model 15-member human gutmicrobiota established in gnotobiotic miceJessica L. Mark Welcha,1, Yuko Hasegawaa, Nathan P. McNultyb,c, Jeffrey I. Gordonb,c, and Gary G. Borisyd,1

aThe Josephine Bay Paul Center for Comparative Molecular Biology and Evolution, Marine Biological Laboratory, Woods Hole, MA 02543; bCenter forGenome Sciences and Systems Biology, Washington University School of Medicine, St. Louis, MO 63110; cCenter for Gut Microbiome and Nutrition Research,Washington University School of Medicine, St. Louis, MO 63110; and dDepartment of Microbiology, The Forsyth Institute, Cambridge, MA 02142

Contributed by Gary G. Borisy, September 12, 2017 (sent for review July 3, 2017; reviewed by Angela E. Douglas and Ruth E. Ley)

Knowledge of the spatial organization of the gut microbiota isimportant for understanding the physical andmolecular interactionsamong its members. These interactions are thought to influencemicrobial succession, community stability, syntrophic relationships,and resiliency in the face of perturbations. The complexity anddynamism of the gut microbiota pose considerable challenges forquantitative analysis of its spatial organization. Here, we illustratean approach for addressing this challenge, using (i) a model, defined15-member consortium of phylogenetically diverse, sequenced hu-man gut bacterial strains introduced into adult gnotobiotic mice feda polysaccharide-rich diet, and (ii) in situ hybridization and spectralimaging analysis methods that allow simultaneous detection ofmultiple bacterial strains at multiple spatial scales. Differences inthe binding affinities of strains for substrates such as mucus or foodparticles, combined with more rapid replication in a preferred mi-crohabitat, could, in principle, lead to localized clonally expandedaggregates composed of one or a few taxa. However, our resultsreveal a colonic community that is mixed at micrometer scales, withdistinct spatial distributions of some taxa relative to one another,notably at the border between the mucosa and the lumen. Our datasuggest that lumen and mucosa in the proximal colon should beconceptualized not as stratified compartments but as componentsof an incompletely mixed bioreactor. Employing the experimentalapproaches described should allow direct tests of whether and howspecified host and microbial factors influence the nature and func-tional contributions of “microscale” mixing to the dynamic opera-tions of the microbiota in health and disease.

gut microbial ecology | community biogeography | bacterial–bacterialinteractions | microbiome function | multiplex fluorescence imaging

The functions expressed by members of a microbial communityare impacted by their neighbors (1–4) and by physiological

features of their environment (5–9). A substantial body of theorysuggests that spatial structure has a powerful influence on theevolutionary stability of mutualistic relationships among mem-bers of a microbial community and between the community andits host (10–17). Depending on the details of the model or theexperimental system, cooperative interactions can be either sta-bilized or destabilized by spatial structure (reviewed in refs. 12,16, and 18). Recently, Coyte et al. used modeling to predict thatthe host would benefit from compartmentalizing microbial spe-cies in the gut to weaken interactions between species and pro-mote community stability (16).Microbes display diverse adherence properties and growth rates

that could contribute to spatially structured communities. In thegut, microbes can adhere to the epithelium and mucins (19–21);these components of the ecosystem are arranged nonrandomly inways that could lead to spatial structuring of adherent communitymembers (22, 23). Similarly, partially digested food particles in thelumen could serve as sites of attachment (24–28). Differentialreplication of a microbe based on its localization in the mucuslayer or the lumen (29) could itself generate a spatially structuredmicrobial consortium or could amplify differences established bydifferential adherence.

The lumen of the gut is considered to be a compartmentinhabited by a microbiota distinct from that of the mucus layer(23, 29–33). This view is largely based on studies that have com-pared mucosal samples to feces (30, 34–36). In contrast, studiesthat have directly compared mucosa-associated communities withadjacent luminal contents have generally shown only modest dif-ferences in the relative proportions of taxa (29, 33, 37–39), al-though exceptions occur. For example, Yasuda et al. (33) reportedthat Helicobacteraceae dominated the colonic mucosa in rhesusmacaques but were a minor constituent in the luminal community.The complexity and dynamism of the gut microbiota pose

considerable challenges for quantitative analysis of its spatial or-ganization. The most comprehensive analysis to date used gno-tobiotic mice colonized with a human fecal community andfluorescence in situ hybridization (FISH) with probes havingspecificities that ranged from phylum level to genus level (5).Other studies have used FISH, antibody staining, or labeling ofpolysaccharides to investigate the distribution of particular taxawithin the microbiota, or of the microbiota as a whole in thepresence of host perturbations (38, 40–48). Using microbes ge-netically engineered to express distinct combinations of two fluo-rescent proteins, Whitaker et al. (49) were able to discriminate sixengineered strains of Bacteroides in the gut of gnotobiotic mice.Adherence to available substrates such as mucus, food particles,

or other microbes, including to the polysaccharide-rich capsularstructures that some community members produce, may localizean organism to a preferred microhabitat but may only modestlyprolong its residence time in the gut. Interestingly, modelingstudies performed in experimental bioreactors, notably recently

Significance

Spatial structure is postulated to have a powerful influence onestablishing and sustaining the signaling and metabolic ex-changes that define relationships among members of the gutmicrobiota and host. However, information about gut communityspatial structure is limited. Simultaneous imaging of componentsof a 15-member model human gut bacterial community over arange of spatial scales in gnotobiotic mice revealed that the colonis better conceptualized as an incompletely mixed bioreactor,rather than having sharply stratified luminal and mucosal com-partments. Identifying host and microbial factors that constrainthe ability of community members to establish sizeable single oroligotaxon agglomerations should yield new insights about how“micro”-scale mixing defines community function.

Author contributions: J.L.M.W., Y.H., N.P.M., J.I.G., and G.G.B. designed research;J.L.M.W., Y.H., and N.P.M. performed research; J.L.M.W., Y.H., and G.G.B. analyzed data;and J.L.M.W., Y.H., N.P.M., J.I.G., and G.G.B. wrote the paper.

Reviewers: A.E.D., Cornell University; and R.E.L., Max Planck Institute.

The authors declare no conflict of interest.

Published under the PNAS license.1To whom correspondence may be addressed. Email: jmarkwelch@mbl.edu or gborisy@forsyth.org.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1711596114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1711596114 PNAS | Published online October 9, 2017 | E9105–E9114

MICRO

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020

developed gut-on-a-chip technology, have shown that peristalticmixing is a key factor in maintaining high bacterial densities,counteracting the tendency of flow to cause rapid depletion ofbacteria (50). The importance of “precise” spatial positioning ofmicrobiota members relative to their metabolic partners to thehealthy functioning of the microbiota is largely unknown. More-over, published studies have yet to examine the biogeography of adiverse microbiota within the colon at a species level, and in amanner where most members of a complex community could betargeted simultaneously without their prior genetic engineering toproduce reporter proteins. In the present report, we use a FISHapproach that allows simultaneous identification of many bacterialspecies (51, 52) to study the spatial organization of a defined 15-member community of sequenced and phylogenetically diversehuman gut-derived taxa installed in the guts of gnotobiotic mice.While this artificial community is simplified relative to natural gutcommunities, it was complex enough to allow for a variety of spatialdistributions and metabolic interactions between its members (4, 53,54). Moreover, our multiplexed imaging approach provided anopportunity to examine the degree of spatial organization of thisdefined community at multiple scales, ranging from hundreds ofmicrometers across the diameter of the gut, to a mesoscale of tensof micrometers, to just a few micrometers at the highest resolution.

ResultsStrategy for Imaging the Distribution of Microbes in the GnotobioticColon. The experimental design and workflow of our strategy areshown in Fig. 1. Adult germ-free mice were colonized with aconsortium of 15 human gut-derived bacterial species (Table S1),and gut segments were fixed and embedded in a glycol meth-acrylate (GMA) resin. The GMA remained in place not onlyduring sectioning but also during subsequent in situ hybridizationand imaging steps, preserving the 3D structure of the gut contents.The proportional representation of members of this community,whose genome sequences had been defined, was initially estimatedby shotgun sequencing of DNA prepared from fecal samples (seeTable S2 for the results of Community Profiling by Sequencing;COPRO-Seq).

To construct a probe set that would provide information on thedistribution of all 15 taxa simultaneously, we initially employed acombinatorial labeling and spectral imaging strategy where eachtaxon was labeled with a unique binary combination chosen fromamong six fluorophores (51, 52). However, we found that because ofthe high microbial density in the colon, adjacent bacterial cells fre-quently overlapped one another in the same pixel of the image,despite confocal optical sectioning. This overlap resulted in ambig-uous combinations of binary signals. Therefore, we reverted to astrategy of labeling each microbe with a single fluorophore. To vi-sualize the distribution of each taxon, we hybridized some sectionswith mixtures of oligonucleotides targeting six or seven taxa in-dividually (e.g., probe sets 1 and 2 in Table S1). To evaluate thedistribution of the entire community, we used a mixture (probe set 3)comprising three sets of oligonucleotides: (i) six probes, each la-beled with a different fluorophore, targeting six species (Bacteroidescellulosilyticus, Bacteroides caccae, Parabacteroides distasonis,Ruminococcus torques, Clostridium scindens, and Collinsella aero-faciens); (ii) four probes, all labeled with Alexa 488, targeting fourmoderately abundant members of Bacteroides (Bacteroides the-taiotaomicron, Bacteroides vulgatus, Bacteroides ovatus, Bacteroidesuniformis); and (iii) five probes, all labeled with Alexa 532, tar-geting five low abundance Firmicutes (Eubacterium rectale, Clos-tridium spiroforme, Faecalibacterium prausnitzii, Ruminococcusobeum, and Dorea longicatena) (see Methods and Fig. S1 for de-tails of probe design and for validation of their taxon specificities).In addition, we labeled all bacteria with the universal bacterialprobe Eub338 conjugated to Alexa 514 and used DAPI to labelhost and bacterial DNA. In some experiments, wheat germ ag-glutinin (WGA), which binds to N-acetyl-D-glucosamine and sialicacid, or an antibody to mouse colonic mucin (MCM) was includedto localize mucus. Finally, we made use of endogenous auto-fluorescence to image food particles and host tissue.When gut samples are imaged at a resolution sufficient to dis-

tinguish individual microbial cells, the field of view is necessarilylimited (∼100 × 100 μm) and the depth of field is shallow (∼1 μm),so that ∼10,000 cubic micrometers of material are visualized in asingle image. The density of microbes in the small intestine has

Merge

Individualunmixed images

BinarizedImage

Watershedimage

Fixation,methacrylateembedding

False-color overlay of unmixedimages for visual inspection

Unmixed images foreach fluorophore

Encase inagarose

Oralgavage

5 µ m thickcross sections

Abundanceheat map

Proximalcolon

Linearunmixing

FISH and tilescan imaging

Segmentedtaxon-assignedimage for cell

counts

Colonicsegment

Germ-freemouse

Germ-freemouse

Gnotobioticmouse

Gnotobioticmouse

Fig. 1. Experimental design and workflow. Germ-free mice were gavaged with a 15-member consortium of human gut bacterial strains and killed 14 d later.Segments of proximal colon were encased in agarose, fixed with formaldehyde, embedded in glycol methacrylate (GMA) resin, sectioned, and hybridized withfluorescent probes. The resin remained in place during hybridization and imaging steps, preserving 3D spatial structure. Gut cross-sections were imaged as atile scan of multiple fields of view to image entire sections at high resolution. Each field of view was imaged by excitation with six laser lines sequentially andprocessed by linear unmixing to create separate images for each of nine fluorophores and for autofluorescence from host tissue and ingested food particles.Each image was then segmented into binary images to allow for an automated count of cells in each square of an 8 × 8 grid for each field of view. The resultsare displayed as a heat map of cell abundance. The unmixed fluorophore channels were also false-colored and overlaid to produce the final unmixed image.

E9106 | www.pnas.org/cgi/doi/10.1073/pnas.1711596114 Mark Welch et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020

been reported to be 104 to 107 cells per mL of luminal contents,whereas in the colon and feces it is orders of magnitude greater (asmuch as 1011–1012 cells per g; refs. 55 and 56). This translates to1,000 microbes per field of view in the colon and feces, but thisnumber decreases 10,000- to 1 million-fold in the jejunum andileum. The FISH probe that reacted with all 15 members of thecommunity (Eub338) confirmed that the distal small intestine ofthese gnotobiotic mice was sparsely colonized (Fig. S2A), whereasthe colon was densely colonized (Fig. S2B; compare with the ab-sence of microbes in the ileum and colon of germ-free controls inFig. S2 C and D). Because of its significantly higher microbialdensity, we chose to focus our multiscale multiplex imaging anal-ysis on the proximal colon.

Characterizing the Distribution of Bacteria in the Gut Is a MultiscaleProblem. To quantitatively evaluate the overall distribution ofbacteria in the proximal colon, we tile-scanned five nearlycomplete cross-sections and a total of 372 microscopic fieldsfrom this region in two mice and represented the data as heatmaps of bacterial cell abundance (see Fig. 1 for a summary of theapproach). As demonstrated by the universal Eub338 probe,bacteria were not distributed homogeneously across cross-sections; rather, they were concentrated at the border betweenthe lumen and mucosa, as well as in patches in the interior of thelumen (Fig. 2A). These regions of highest density contained∼50 cells per 100 μm2. Substantially lower densities occurred inthe remainder of the lumen, consisting primarily of regions inand surrounding large autofluorescent particles that we interpretas remnants of food (Fig. 2A).

Heat maps of the individual bacterial taxa showed a broadlysimilar pattern of distribution at the scale of tens to hundreds ofmicrometers (Fig. 2B). This similarity was evaluated quantita-tively by line-scan analysis of transects running from the edge ofthe mucosa into the lumen (Fig. 2C). In these scans, the transectsextended far enough into the lumen to determine the width ofthe high density microbial zone near the mucosa but not so far asto intersect patches of high luminal concentration. Data werecollected for 245 transects, and abundances were expressed asthe fraction of bacterial cells present in discrete segments of thetransect. When averaged across the 245 transects, the abundancepattern for each taxon was similar, peaking close to the mucosaand then falling off sharply to a low level at distances greaterthan 75 μm into the lumen.We focused our analysis on four “microhabitats”: (i) densely

colonized regions adjacent to the colonic mucosa, (ii) densepatches of microbes within the lumen of the colon, (iii) sparselycolonized areas within the lumen, and (iv) sparsely colonizedcrypts of Lieberkühn. All of the abundant taxa were detectablein each of these microhabitats. In some images, large auto-fluorescent particles of food pressed close to the mucosa so thatthe microbe-rich region was only 10–20 μm thick (Fig. 3A). Otherimages show a wider, densely colonized region at the edge of themucosa (Fig. 3B). The lumen contained densely colonized re-gions bordered by irregularly shaped autofluorescent food par-ticles of variable size and brightness (Fig. 3C). A diversemicrobial community was thinly arrayed on some of these par-ticles and colonized the edges of cavities within them (Fig. 3 Aand C). This latter observation is consistent with the notion that

A B

0

5

10

15

20

Highestdensity

Lowestdensity

500 µm

50

30

20

10

Taxo

n re

lativ

eab

unda

nce

Distance from edge of proximal colonic mucosa (µm)

B. caccae

C. scindensR. torques

C B. cellulosilyticus4 Bacteroides

C. aerofaciens

-57 -38 -19 0

Edge ofmucosa

19 38 57 76 95 114 133 152 171 190 209 228 247 266 285 304 323 342 361 380

40

0

7

6

5

4

3

2

1

0

16

2

141210864

0

1618

2

141210864

0

910

12

876543

0

3.5

3

2.5

2

1.5

1

0.5

0

30

20

10

40

0

20

10

15

5

0

B. cellulosilyticus 4 Bacteroides

5 FirmicutesC. scindensP. distasonisB. caccae

C. aerofaciens R. torques

1 FOV =8 x 8 grid

Mucosa

MucosaLumen

500 µm

Fig. 2. Microbes are most abundant near host tissue and in patches in the lumen. Microbial density in each region of the cross-section is shown as a heat map(A) representing the number of cells hybridizing to the Eub338 probe and ranging from zero (dark blue) to 158 cells (red) per 19 × 19 μm grid square. The heatmap is overlaid on a tile scan (Inset) of the section showing autofluorescence from host tissue and ingested food particles. (B) The density of individual taxa orgroups of taxa shows a similar distribution. (C) A line-scan analysis of cell abundance along transects perpendicular to host tissue illustrates that taxa havesimilar distributions at this scale. Values shown are the percent of cells observed in each quantum of the transect, and are the mean of 245 transects from atotal of five intestinal cross-sections from two mice. Transects consisted of 24 adjacent grid squares placed so that the fifth grid square contained the firstbacterial cell labeled with the universal probe Eub338. Thus, the first four grid squares (76 μm) of each transect line would be empty unless occupied with abacterial cell hybridizing to a specific probe but not to Eub338.

Mark Welch et al. PNAS | Published online October 9, 2017 | E9107

MICRO

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020

food particles can serve as platforms for attachment of bacterialtaxa (including potential syntrophic partners). Crypts were colo-nized by a sparse, taxonomically mixed community (Fig. 3D),suggesting that they were colonized from the lumen and did not

serve as microhabitats for any select groups within the 15-membermodel human gut microbiota.The overall composition of the microbial community differed only

modestly from microhabitat to microhabitat within a cross-section

A B

B. cellulosilyticus 4 Bacteroides 4 Bacteroides B. caccae C. aerofaciens

B. cellulosilyticus B. theta

autofluorescence

B. thetaB. vulgatus C. aerofaciens R. torquesautofluorescence

C. aerofaciens C. scindens R. torquesautofluorescenceautofluorescence

B. cellulosilyticus 4 Bacteroides 4 Bacteroides B. caccae C. aerofaciens C. aerofaciens C. scindens R. torquesautofluorescenceautofluorescence

B. cellulosilyticus B. theta

autofluorescence

B. thetaB. vulgatus C. aerofaciens R. torquesautofluorescence

C D

.

B. cellulosilyticus B. theta

autofluorescence

B. thetaB. vulgatus B. ovatusR. torquesautofluorescenceDAPIDAPI

5 μm

20 μm

Fig. 3. Colonization patterns in distinct microhabitats in the colon. Tiled images show the distribution of microbes relative to host tissue and large autofluorescentfood particles. Images shown are representative of the region proximal to the epithelium (A and B), the region distal to the epithelium (lumen; C), and crypts (D). Whiteboxes show the positions of higher magnification views (Lower) where individual bacterial cells are visible; low-magnification image in C shows the image location inthe lumen. Microbes were spatially mixed at micrometer scales in all microhabitats. Legend in A also applies to C. Scale bars in D apply throughout the figure.

E9108 | www.pnas.org/cgi/doi/10.1073/pnas.1711596114 Mark Welch et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020

when measured at a mesoscale (i.e., on the order of 100 μm). Toquantify these differences, we compared, from a single cross-section, 11 fields of view (each 152 × 152 μm) containing pri-marily the dense community and 12 fields containing almostexclusively the sparser, food particle-associated community. Thesix Bacteroides species collectively dominated in all microhabi-tats; for example, B. cellulosilyticus made up 43 ± 4% (mean ±SD) of the cells in the densely colonized fields and 46 ± 4% inthe sparsely colonized fields (Table S2). C. aerofaciens was mostabundant in the densely colonized areas, where it made up 11 ±1% of bacteria cells versus 6 ± 2% within the sparsely colonizedareas. Of the two members of Clostridia targeted with specificprobes, C. scindens comprised 8 ± 2–4% of the community inboth microhabitats, while R. torques made up 3 ± 1% in densefields and 2 ± 1% in sparsely colonized regions of the lumen.Thus, modest differences in overall microbial community com-position in mucus-rich and food-rich regions were detectable.Summing FISH results across all of the sections yielded relativeabundances that in many cases were comparable to resultsobtained from COPRO-Seq analysis of fecal DNA (Table S2).Communities observed at 10-μm scales consisted of a mixture

of diverse species, with the relative proportions of species varyingfrom region to region. The autofluorescent particles we observedranged in size from distinctively shaped objects several hundredsof micrometers long (Fig. 2) down to small “blobs” only a fewmicrometers in diameter. While the larger objects were generallyonly thinly colonized, smaller particles were present in denselycolonized patches, with the exception of some patches near themucosa (Fig. 3).

Quantitative Analysis of Spatial Organization at Micrometer Scales.To assess whether the spatial arrangement of microbes at mi-crometer scales differed relative to the visible landmarks of theepithelial border and large food particles, we carried out ananalysis of spatial arrangement using the method of linear

dipoles as implemented in DAIME (57). Linear dipole analysiscalculates a pair cross-correlation function for two categories ofobjects by estimating the probability, normalized to their densityin the image, that the ends of a line segment of a given length willtouch both of them.Linear dipole analysis demonstrated a tendency of microbes to

be in low abundance near crypts and invaginations in the mu-cosal epithelium (Fig. 4). This result is not surprising becausethese regions are generally occupied by a dense mucus layerresistant to bacterial penetration (58). Bordering this largelymicrobe-free zone there was dense colonization of microbes. Forpurposes of analysis, we marked this border of dense coloniza-tion with a hand-drawn line and calculated the spatial cross-correlation between the microbes and this line (Fig. 4). The re-sults revealed a concentration of microbes within ∼30 μm of thisline (Fig. 4). By contrast, and counterintuitively, microbes wereunderrepresented within 5–10 micrometers of large food parti-cles (Fig. 4). It is possible that this underrepresentation was duein part to a failure to detect the edges of food particles accu-rately, as food particles were detected by virtue of their endog-enous fluorescence rather than by specific staining. However, theimages also give no visual evidence of any clustering of microbesclose to these autofluorescent particles.

Micrometer-Scale Analysis Reveals Differences in the Organization ofSpecific Taxa with Respect to the Mucosa and One Another. Turningfrom the distribution of bacteria overall to the distribution ofindividual taxa within the community, we detected an un-derrepresentation of R. torques in the dense community at theborder between the mucosa and the lumen. Bacteroides wereabundantly represented throughout this microbe-dense region,but R. torques exhibited marked scarcity in the region closest tothe mucosa (Fig. 5). To quantify the distribution of taxa relativeto the edge of the mucosa, we marked the limit of dense mi-crobial colonization with a 1-μm-thick line (Fig. 5) and used

100 µm100 µm

00.5

11.5

22.5

33.5

0.4 4.6 8.813.0

17.221.4

25.629.8

34.038.2

42.446.5

50.754.9

59.163.3

67.571.7

75.980.1

84.388.5

92.796.9

Pai

r cor

rela

tion

Microns

to food edgeto host edgeto edge of microbial colonization

Fig. 4. Spatial analysis of bacteria relative to visible landmarks. A tiled image (Upper Left) was segmented to identify bacterial cells and to outline the edgeof large food particles, host tissue, and the edge of dense colonization (Upper Right; orange, green, and purple lines). Spatial correlation analysis was carriedout using the method of linear dipoles (57), which calculates the pair cross-correlation function as the probability that two categories of object are located ata given distance from one another, normalized to their density in the image. This analysis revealed that bacteria tend to localize within 30 μm of the markededge of colonization, but are underrepresented within 20 μm of the epithelium and within 5–10 μm of large food particles. Interiors of food particles and hosttissue (Upper Right, magenta) were excluded from the analysis. Dotted lines indicate 95% confidence intervals generated by dividing the image into 10 radialsectors for analysis. Food particles and host tissue were identified by autofluorescence with 405-nm excitation, and their edges were outlined by eroding theimage by seven pixels (∼1 μm) using FIJI. The edge of dense colonization was outlined by hand.

Mark Welch et al. PNAS | Published online October 9, 2017 | E9109

MICRO

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020

linear dipole analysis to calculate the cross-correlation betweeneach taxon and the line. The use of the line, rather than themucosa itself, was necessary because of crypts and invaginationsin the mucosa that resulted in variable distances between the

epithelial border and microbial community, whereas the edge ofmicrobial colonization itself was relatively smooth. The resultsconfirmed the visual observation that R. torques is frequently,although not universally, underrepresented in the microbial

0

0.5

1

1.5

2

0.2 2.3 4.4 6.5 8.6 10.7

12.8

14.9 17 19

.121

.223

.324

.8

B. cellulosilyticus4 BacteroidesC. aerofaciensB. caccaeC. scindensR. torques

B. cellulosilyticusto lineR. torquesto line

0

0.2

0.4

0.6

0.8

1

1.2

0.2 1.9 3.6 5.2 6.9 8.6 10.3

12.0

13.6

15.3

17.0

18.7

20.3

22.0

23.7

24.5

Pai

r cor

rela

tion

Pai

r cor

rela

tion

Microns

Microns10 µm

Fig. 5. Distinctive community organization in the 10 μm closest to the mucosa. The microbial community near the mucosal epithelium is abundantlypopulated by B. cellulosilyticus, but R. torques is underrepresented in a narrow band close to the mucosa. The white boxed area in Inset shown in Upper Leftdenotes the field shown at higher magnification in Upper Left and Lower Left. (Upper Left) A representative image near the mucosa showing all taxa and a1-μm-thick line representing the edge of microbial colonization. (Upper Right) Pair cross-correlation (PCC) analysis showing the probability of detecting a cellat each distance from the line, normalized to the density of cells in the image. Analysis was carried out using the method of linear dipoles as implemented inDAIME (57). (Lower Left) The B. cellulosilyticus and R. torques channels are shown separately for clarity and to demonstrate the rarity of R. torques in the 5- to10-μm zone at the edge of the microbe-dense region. (Lower Right) Results of PCC analysis depicted as the mean of all 11 images from two sections in which a100 μm length and 40 μm width of mucosal border was visible. Confidence intervals of 95% are shown for these two bacterial strains.

B. cellulosilyticus B. thetaiotaomicronB. vulgatus C. aerofaciens R. torques

R2 = 0.193

0

10B. v

ulga

tus

cells

per g

rid s

quar

e

R. t

orqu

esce

llspe

r grid

squ

are

20

30

40

0 20 40 60 80 100 0 20 40 60 80 100

B. vulgatus vs. B. cellulosilyticus R. torques vs. B. cellulosilyticus

R2 = 0.155

-4

0

4

8

12

16

B. cellulosilyticus cells per grid squareB. cellulosilyticus cells per grid square

Fig. 6. Distinctive organization of microbes relative to one another. Abundance of each taxon was tabulated within 1,572 grid squares measuring 19 × 19 μm (cf.individual squares in the heat map in Fig. 2) from a section hybridized with the species-specific probe set 1. Scatter plots of individual taxa show that the abundance ofB. cellulosilyticus and B. vulgatus are positively correlated (Upper Left) while the abundance of B. cellulosilyticus and R. torques are negatively correlated (Upper Right).Scatter plots include only those grid squares that contain a high density of bacterial cells (at least 50% of the maximum density). An image of such a densely populatedregion (Lower) shows that B. cellulosilyticus and B. vulgatus are abundant in the same region of the image, while the abundance of R. torques is highest where abundanceof the Bacteroides is low. These spatial relationships are consistent across mice, as demonstrated by analysis of both mice with the comprehensive probe set 3 (Fig. S4).

E9110 | www.pnas.org/cgi/doi/10.1073/pnas.1711596114 Mark Welch et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020

community positioned in the 5- to 10-μm region closest to hostcells (Fig. 5).Differences in the distribution of individual taxa relative to one

another were also evident at micrometer scales. Images were di-vided into grid squares of 19 × 19 μm, and bacterial cells withineach grid square were tallied. Scatter plots comparing the abun-dance of pairs of taxa within these grid squares showed a linearregression with a positive slope (Fig. S3), likely reflecting thepresence of a mixed microbial community in regions of both highand low microbial abundance. One possibility is that the primarydriver of correlations in abundances among taxa is simply theoverall abundance of bacterial cells in any given 19 × 19 μm gridsquare. However, analyzing only “densely populated” grid squares(defined as those containing at least half the maximum number ofbacterial cells) revealed differences in taxon distribution (Fig. 6and Figs. S3 and S4). For example, abundance of the mostprominent taxon, B. cellulosilyticus, was positively correlated withthat of B. vulgatus but negatively correlated with R. torques.Among the possible explanations for these observed distributionaldifferences are cooperative (attractive) versus competitive (re-pulsive) interactions between taxa, or an indirect effect of bindingor proliferation of taxa in distinct microenvironments.To further investigate factors underlying differences in taxon

density and distribution, we imaged the distribution of bacteria,mucus, and food in additional cross-sections using the Eub338

bacterial probe in conjunction with WGA and a mouse colonicmucin antibody (anti-MCM) to localize mucus (Fig. 7 and Figs.S5 and S6). As before, bacteria were detected in high abundancein regions of the lumen as well as close to the mucosa (Fig. 7 Band C). WGA and anti-MCM showed similar staining patterns toone another (Fig. S5), and a qualitative inspection of the imagesrevealed dense concentrations of mucus in regions correspondingto areas of high bacterial density, as well as in goblet cells (Fig. 7B and C and Fig. S6). Large autofluorescent particles, by con-trast, generally occupied regions in which bacteria were notabundant, although small autofluorescent particles were mixedwith mucus and bacteria even in regions of highest bacterialdensity (Fig. 7D). These results are consistent with the notionthat dense concentrations of mucus support high bacterial den-sity, both in the loose mucus layer at the mucosal border and inthe interior of the lumen.

DiscussionWe investigated the spatial arrangement of members of a de-fined artificial 15-member human gut microbiota in the proximalcolon of gnotobiotic mice at macroscale, mesoscale, and micro-meter scale. At a macrolevel scale of hundreds of micrometers,bacterial cell density was heterogeneous throughout the colon butwas highest in regions rich in mucus, both in a layer near the ep-ithelium and in patches in the lumen. Compositional “patchiness”

B

C

D

i

i

ii

ii

iii

iii

iv

iv

a b c

A Bacteria (Eub338-Rhodamine Red X)Mucus (Wheat Germ Agglutinin-Alexa488)Autofluorescence

a

b

c

Bacteria Mucus Food & Tissue Overlay

Fig. 7. Dense bacterial aggregations occupy regions rich in mucus. (A) Cross-section of colon highlighting location of panels shown at higher magnificationbelow. (B) Bacterial density (i) is heterogeneous in the lumen. Staining with fluorophore-labeled WGA (ii) shows high density of mucus in areas of the lumenthat contain abundant bacteria. Large autofluorescent food particles (iii) occupy areas of the lumen in which bacterial density is low. (iv) Overlay of i–iii.(C) Bacterial density (i) is also high in a narrow zone located at the edge of the mucosa. WGA staining (ii) shows a high density of mucus in this zone.Autofluorescent food particles (iii) are located within micrometers of the mucosa. (D) High-magnification views showing regions of the lumen with abundantmucus (a), large food particles (b), and a large food particle pressed close to the mucosa (c). The cross-section shown is adjacent to the one presented in Fig. 2.(Scale bars: A, 200 μm; B and C, 50 μm; D, 10 μm.)

Mark Welch et al. PNAS | Published online October 9, 2017 | E9111

MICRO

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020

was observed at a mesoscale (tens of micrometers), while at mi-crometer scales, each microhabitat was generally occupied by acomplex community of microbes intermingled with one another.This extensive mixing at micrometer scales, the absence of largemicrocolonies associated with the mucosa or food particles, andthe broad similarity of luminal and mucosa-adjacent communitieslead us to view the lumen and mucosa in the proximal colon not asstratified compartments but as components of a partially mixedbioreactor. By “bioreactor,” we mean an environment constructedto harbor microbes and harness their metabolism of availablenutrients, where the degree of spatial homogenization of com-munity members is a manifestation of a complex dynamic involvingmany factors including flux of mucus and epithelial cells into thelumen, the affinities of community members for these host con-stituents and food particles, cooperative/competitive (attractive/repulsive) interactions between microorganisms, and peristalticflow of bulk material through the lumen.One might have expected a priori that differences between taxa

in their binding affinities for substrates such as mucus or foodparticles, combined with more rapid replication in a preferredmicrohabitat, would lead to localized clonal expansion. Theresulting communities would have spatial structure in the form ofmicrocolonies and a distinctive community composition within themucus layer compared with luminal contents. Instead, the spatialdistribution of both substrates and microbes suggests a dynamicmodel for the proximal colon in which mixing and dispersal by hostfactors tend to homogenize the community. Such mixing could alsoexplain why large clusters with strongly distinctive taxonomiccomposition were not observed between the mucus layer and thelumen. The turnover time for mucus is ∼6 h for mucus in gobletcells and as little as a single hour for the inner mucus layer in thedistal colon (59). The replication time of two common gut sym-bionts, B. thetaiotaomicron and Escherichia coli, has been estimatedat 3 h in the mucus layer and 3–8 h in colonic contents (29). Thus,these and other gut bacterial taxa may carry out only a few roundsof cell division within the mucus layer before being shed along withmucus into the lumen. Adhesion to the epithelium could supportpersistence of microbes for a longer time, but epithelial cells arecontinuously discarded; with the exception of Paneth cells, theother three mouse gut epithelial lineages turn over every 3–5 d(60–64). Microbes in the lumen also have limited opportunity forreplication before being expelled, as the contents of the gut tra-verse the mouse intestine in 4–6 h (65–67). We hypothesize thatthrough rapid turnover and mixing of the mucus layer, the epi-thelium, and the gut contents as a whole, the mammalian host actsto diminish the ability of the microbiota to establish spatiallysegregated communities or sizeable single-taxon agglomerationsboth in the lumen and adjacent to the mucosa. Whether thephysiology of the microbes themselves also fosters mixing is aquestion raised by the extensive micrometer-scale taxonomicintermingling in densely populated mucus-rich regions of our im-ages. The bacteria comprising the artificial human gut microbiotacharacterized in this study are nonflagellated. Diffusion is unlikelyto be a significant force within the viscous gel of the mucus layer(29), although there is little information about how foraging ofmucus glycans by resident microbes affects mucus viscosity. Re-gardless of the means by which mixing is achieved, its effect is todiminish micrometer-scale spatial structure in the community.The microbial assemblage we employed was simple enough so

most of the abundant taxa could be simultaneously visualized withmultilabel FISH and complex enough to permit a variety of pos-sible taxon–taxon interactions and spatial distributions. Inevitably,our findings are dictated by the microbes that we chose to createthis synthetic community. We do not know whether the consor-tium of 15 taxa studied display a spatial organization that is rep-resentative of the native mouse gut microbiota. It is possible thatthis low diversity consortium of human gut-derived bacterialstrains is less likely to form spatially structured communities than

the native mouse gut microbiota. Moreover, members of a morecomplex community of human gut microbes, representing lineagesfrom Bacteria and other domains of life (e.g., methanogenicarchaeons and eukaryotes) could demonstrate more pronounceddifferences in their replication rates and stronger spatial associa-tions with one another or with food particles, host cells, mucus, orother features of the gut habitat.The observed spatial arrangement differs dramatically from the

highly ordered and more clustered arrangements visible in humandental plaque (68). These dissimilarities could reflect a number offactors including differences in flow rates in the two ecosystemsand the absence of a surface in the gut to which microbes canstably adhere. In the mouth, continuous rapid flow of saliva en-sures that adherence to a surface, either directly or indirectly viabinding to other adherent microbes, is critical for remaining in thehabitat. Further, chemical communication in salivary flow is mosteffective at distances on the order of micrometers (1) so precisepositioning relative to metabolic partners is critical.Deeper understanding of spatial relationships in the gut should

be gained as methods are developed for quantifying the distribu-tion of nutrients and metabolic products over different spatialscales. Gnotobiotic animal models harboring defined consortia ofmicrobes and fed diets of known composition and physical (e.g.,particulate) properties represent a starting point for these types ofstudies. For example, follow-up experiments in which gnotobioticmice are fed purified dietary fibers of defined composition and sizecould provide an opportunity to identify taxa that exhibit re-producible patterns of cooccurrence or spatial association with thepopulation of food particles represented in their colons. In addi-tion, the relative effects of propulsive contractions, nonpropulsivemixing, and/or the rate of renewal of the mucus layer can be ex-plored using gnotobiotic animals with mutations that affect theirgut motility and/or the composition of their mucus (e.g., ref. 67).These genetic manipulations can also be applied to gnotobioticzebrafish where the transparency of the organism can support invivo imaging of microbial communities (69, 70). This latter featureavoids a potentially confounding variable; namely, that mesoscaleand microscale spatial structure may be disrupted in unknown waysduring the processing of gut tissues taken from euthanized animals.The resulting datasets should enable modeling both in silico and inex vivo experimental bioreactors, including “gut-on-a-chip” sys-tems (71). Together, these efforts should provide informationabout forces that promote or retard mixing at microscale levels.Put another way, studies of microscale mixing provide an oppor-tunity to determine how this parameter is related to the niches(“jobs”) of community members (both symbionts and pathogens),the expressed functional properties of a microbiota (including itsresiliency to perturbations), and whether changes in this spatialfeature are a reflection of, or causally related to, various typesof dysbioses.

Materials and MethodsCollection of Samples from Gnotobiotic Mice and Preparation for Imaging. Allexperiments involving mice were performed using protocols approved by theAnimal Studies Committee of the Washington University School of Medicine.Two 8-wk-old, male, germ-free C57BL/6J micewere gavagedwith a 15-memberbacterial consortium prepared and administered using procedures described inearlier reports (4, 54), and then maintained in a gnotobiotic isolator under astrict light cycle (lights on at 0600 hours and off at 1800 hours). Two additionalmice were gavaged with a two-member bacterial community composed ofB. thetaioatomicron VPI-5482 and E. rectale ATCC 33656, and two control micewere maintained as germ-free. All animals were fed a sterilized, low-fat, plantpolysaccharide-rich chow (Product 7378000; B&K Universal) ad libitum. All micewere killed 14 d after gavage of their bacterial consortium.

Thedistal fourth of the small intestine and theproximal thirdof the colonweresnap-frozen in optimal cutting temperature compound and stored at −80 °C.These frozen segments were cut into 5- to 10-mm-long pieces, and molten0.5% agarose was pipetted onto the two cut ends of each piece. Samples werethen fixed in 2% paraformaldehyde in PBS for 12 h at 4 °C, then washed,

E9112 | www.pnas.org/cgi/doi/10.1073/pnas.1711596114 Mark Welch et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020

resubmerged in molten 0.5% agarose, dehydrated in acetone for 1 h, infil-trated with Technovit 8100 glycol methacrylate resin with several changes over12 h at 4 °C, and were then transferred to embedding solution where theywere allowed to solidify for 12 h at 4 °C. Embedded samples were sectionedwith a Sorvall JB-4 microtome (Dupont Instruments) to 5–10 μm thicknessand subjected to fluorescent labeling experiments. We employed para-formaldehyde rather than Carnoy’s fixative because paraformaldehyde iscommonly used for microbial FISH (72, 73). Moreover, our direct comparison ofthe two methods revealed similar preservation of mucus and other spatiallandmarks in methacrylate-embedded colonic segments fixed with eitherCarnoy’s solution or paraformaldehyde (74).

Probe Design. Candidate probes identified using the “probe design” functionof the ARB software package (75) were further analyzed by calculating theoverall free energy of hybridization (ΔG0

overall; ref. 76) for each. A set of 16SrRNA-directed oligonucleotides was selected that were predicted to betaxon-specific at the same hybridization stringency (arbitrarily chosen to be20% formamide). Probes were synthesized (Invitrogen) with a 5′ fluo-rophore and their specificity evaluated by hybridization to pure cultures oftarget and nontarget bacterial strains.

FISH Analysis of Colonic Sections. FISH was carried out at 46 °C for 6 h in 0.9 MNaCl, 0.02 M Tris pH 7.5, 0.01% SDS, 20% HiDi formamide (Applied Bio-systems), and 2 μM of each probe. Sections were then washed twice for10 min each at 48 °C in wash buffer (0.215 M NaCl, 0.02 M Tris pH 7.5, 5 mMEDTA), incubated in DAPI in PBS at room temperature for 15 min in the dark,washed in PBS, dehydrated through a series of ethanol washes [3 min eachin 50%, 80%, and 96% (vol/vol) ethanol], mounted in ProLong Gold antifadereagent (catalog no. P36934; Invitrogen) with a No. 1.5 coverslip, and placedin the dark at room temperature for at least 24 h before imaging. FISH onpure cultures was carried out in the same way, but without DAPI staining.

Lectin and Mucin Staining. Labeling of methacrylate sections with WGA wascarried out after the FISH hybridization and washing steps. Slides were in-cubated in 40 μg/mL WGA conjugated to Alexa Fluor 488 (catalog no.

W11261; Invitrogen) in PBS at room temperature for 15 min in the dark.Slides were washed briefly in PBS, dipped in distilled H2O, dehydratedthrough a series of ethanol washes and mounted as above. Slides to be la-beled with anti-mouse colonic mucin (anti-MCM, a gift of Dr. Ingrid B. Renes,Erasmus MC-Josephine Nefkens Institute, Rotterdam, The Netherlands) weretreated with blocking buffer (2% goat serum; 1% BSA; 0.2% Triton X-100;0.05% Tween 20) for 1 h at room temperature, incubated with anti-MCMdiluted 1:50 in blocking buffer for 12 h at 4 °C, rinsed three times for 3 mineach in PBS, incubated with blocking buffer at room temperature for 1 h,incubated with a 1:1,000 dilution of Alexa Fluor 633 goat anti-rabbit IgG(catalog no. A21070; Invitrogen) in blocking buffer for 1 h, rinsed threetimes for 3 min each in PBS, incubated with WGA, washed, and mountedas above.

Image Acquisition and Linear Unmixing. Spectral image stacks were acquiredusing a Zeiss LSM 710 confocal microscope equipped with a QUASAR spectraldetector and a Plan-Apochromat 63×, 1.4 N.A. objective lens or a Zeiss LSM780 confocal microscope equipped with a GaAsP spectral detector and aPlan-Apochromat 40×, 1.4 N.A. objective. Excitation wavelengths of 633 nm,594 nm, 561 nm, 514 nm, 488 nm, and 405 nm were employed sequentially.Linear unmixing was carried out using ZEN software (Carl Zeiss) or using acustom algorithm on the Mathematica platform (Wolfram Research) usingthe concatenated reference spectra shown in Fig. S1. Details of the unmixingprocedure and figure preparation are given in SI Materials and Methods.

Spatial Arrangement Analysis. Images of each taxon were segmented in FIJIusing the IsoData or RenyiEntropy global thresholding algorithm (77) andwere then imported into DAIME version 2.1 for analysis of spatial arrange-ment using linear dipoles.

ACKNOWLEDGMENTS. We thank David O’Donnell and Maria Karlsson forhelp with gnotobiotic animal husbandry; Louie Kerr for microscopy support;and Christopher Rieken, Blair Rossetti, and Liping Xun for outstanding techni-cal support. This work was supported in part by NIH Grants DE 022586,DK30292, DK70977, and DK78669.

1. Kolenbrander PE, Palmer RJ, Jr, Periasamy S, Jakubovics NS (2010) Oral multispeciesbiofilm development and the key role of cell-cell distance. Nat Rev Microbiol 8:471–480.

2. Fischbach MA, Sonnenburg JL (2011) Eating for two: How metabolism establishesinterspecies interactions in the gut. Cell Host Microbe 10:336–347.

3. Traxler MF, Watrous JD, Alexandrov T, Dorrestein PC, Kolter R (2013) Interspeciesinteractions stimulate diversification of the Streptomyces coelicolor secreted metab-olome. MBio 4:e00459-13.

4. Wu M, et al. (2015) Genetic determinants of in vivo fitness and diet responsiveness inmultiple human gut Bacteroides. Science 350:aac5992.

5. Earle KA, et al. (2015) Quantitative imaging of gut microbiota spatial organization.Cell Host Microbe 18:478–488.

6. Abreu NAA, Taga ME (2016) Decoding molecular interactions in microbial commu-nities. FEMS Microbiol Rev 40:648–663.

7. Donaldson GP, Lee SM, Mazmanian SK (2016) Gut biogeography of the bacterialmicrobiota. Nat Rev Microbiol 14:20–32.

8. Stacy A, McNally L, Darch SE, Brown SP, Whiteley M (2016) The biogeography ofpolymicrobial infection. Nat Rev Microbiol 14:93–105.

9. Proctor DM, Relman DA (2017) The landscape ecology and microbiota of the humannose, mouth, and throat. Cell Host Microbe 21:421–432.

10. Nadell CD, Xavier JB, Foster KR (2009) The sociobiology of biofilms. FEMS MicrobiolRev 33:206–224.

11. Mitri S, Xavier JB, Foster KR (2011) Social evolution in multispecies biofilms. Proc NatlAcad Sci USA 108:10839–10846.

12. Estrela S, Brown SP (2013) Metabolic and demographic feedbacks shape the emergentspatial structure and function of microbial communities. PLOS Comput Biol 9:e1003398.

13. Stein RR, et al. (2013) Ecological modeling from time-series inference: Insight intodynamics and stability of intestinal microbiota. PLOS Comput Biol 9:e1003388.

14. Kovács AT (2014) Impact of spatial distribution on the development of mutualism inmicrobes. Front Microbiol 5:649.

15. Lambert G, Vyawahare S, Austin RH (2014) Bacteria and game theory: The rise and fallof cooperation in spatially heterogeneous environments. Interface Focus 4:20140029.

16. Coyte KZ, Schluter J, Foster KR (2015) The ecology of the microbiome: Networks,competition, and stability. Science 350:663–666.

17. Peña J, Wu B, Traulsen A (2016) Ordering structured populations in multiplayer co-operation games. J R Soc Interface 13:20150881.

18. Momeni B, Waite AJ, Shou W (2013) Spatial self-organization favors heterotypic co-operation over cheating. Elife 2:e00960.

19. Derrien M, et al. (2010) Mucin-bacterial interactions in the human oral cavity anddigestive tract. Gut Microbes 1:254–268.

20. Naughton JA, et al. (2013) Divergent mechanisms of interaction of Helicobacter pyloriand Campylobacter jejuni with mucus and mucins. Infect Immun 81:2838–2850.

21. Etzold S, Juge N (2014) Structural insights into bacterial recognition of intestinal

mucins. Curr Opin Struct Biol 28:23–31.22. Gunning AP, et al. (2013) Mining the “glycocode”–Exploring the spatial distribution

of glycans in gastrointestinal mucin using force spectroscopy. FASEB J 27:2342–2354.23. McLoughlin K, Schluter J, Rakoff-Nahoum S, Smith AL, Foster KR (2016) Host selection

of microbiota via differential adhesion. Cell Host Microbe 19:550–559.24. Shipman JA, Berleman JE, Salyers AA (2000) Characterization of four outer membrane

proteins involved in binding starch to the cell surface of Bacteroides thetaiotaomi-

cron. J Bacteriol 182:5365–5372.25. Sonnenburg JL, et al. (2005) Glycan foraging in vivo by an intestine-adapted bacterial

symbiont. Science 307:1955–1959.26. Macfarlane S, Macfarlane GT (2006) Composition and metabolic activities of bacterial

biofilms colonizing food residues in the human gut. Appl Environ Microbiol 72:

6204–6211.27. Walker AW, et al. (2008) The species composition of the human intestinal microbiota

differs between particle-associated and liquid phase communities. Environ Microbiol

10:3275–3283.28. Nagara Y, Takada T, Nagata Y, Kado S, Kushiro A (2017) Microscale spatial analysis

provides evidence for adhesive monopolization of dietary nutrients by specific in-

testinal bacteria. PLoS One 12:e0175497.29. Li H, et al. (2015) The outer mucus layer hosts a distinct intestinal microbial niche. Nat

Commun 6:8292.30. Stearns JC, et al. (2011) Bacterial biogeography of the human digestive tract. Sci Rep

1:170.31. Schluter J, Foster KR (2012) The evolution of mutualism in gut microbiota via host

epithelial selection. PLoS Biol 10:e1001424.32. Sommer F, Bäckhed F (2013) The gut microbiota–Masters of host development and

physiology. Nat Rev Microbiol 11:227–238.33. Yasuda K, et al. (2015) Biogeography of the intestinal mucosal and lumenal micro-

biome in the rhesus macaque. Cell Host Microbe 17:385–391.34. Zoetendal EG, et al. (2002) Mucosa-associated bacteria in the human gastrointestinal

tract are uniformly distributed along the colon and differ from the community re-

covered from feces. Appl Environ Microbiol 68:3401–3407.35. Morgan XC, et al. (2012) Dysfunction of the intestinal microbiome in inflammatory

bowel disease and treatment. Genome Biol 13:R79.36. Albenberg L, et al. (2014) Correlation between intraluminal oxygen gradient and

radial partitioning of intestinal microbiota. Gastroenterology 147:1055–1063.e8.37. Nava GM, Friedrichsen HJ, Stappenbeck TS (2011) Spatial organization of intestinal

microbiota in the mouse ascending colon. ISME J 5:627–638.38. Pédron T, et al. (2012) A crypt-specific core microbiota resides in the mouse colon.

MBio 3:e00116-12.

Mark Welch et al. PNAS | Published online October 9, 2017 | E9113

MICRO

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020

39. Malmuthuge N, Griebel PJ, Guan L (2014) Taxonomic identification of commensalbacteria associated with the mucosa and digesta throughout the gastrointestinal

tracts of preweaned calves. Appl Environ Microbiol 80:2021–2028.40. Schuppler M, Lötzsch K, Waidmann M, Autenrieth IB (2004) An abundance of Es-

cherichia coli is harbored by the mucosa-associated bacterial flora of interleukin-2-

deficient mice. Infect Immun 72:1983–1990.41. Chan V, et al. (2005) Visualization of Helicobacter species within the murine cecal

mucosa using specific fluorescence in situ hybridization. Helicobacter 10:114–124.42. Swidsinski A, Weber J, Loening-Baucke V, Hale LP, Lochs H (2005) Spatial organization

and composition of the mucosal flora in patients with inflammatory bowel disease.J Clin Microbiol 43:3380–3389.

43. Ahmed S, et al. (2007) Mucosa-associated bacterial diversity in relation to human

terminal ileum and colonic biopsy samples. Appl Environ Microbiol 73:7435–7442.44. Vaishnava S, et al. (2011) The antibacterial lectin RegIIIgamma promotes the spatial

segregation of microbiota and host in the intestine. Science 334:255–258.45. Cullender TC, et al. (2013) Innate and adaptive immunity interact to quench micro-

biome flagellar motility in the gut. Cell Host Microbe 14:571–581.46. Lee SM, et al. (2013) Bacterial colonization factors control specificity and stability of

the gut microbiota. Nature 501:426–429.47. Geva-Zatorsky N, et al. (2015) In vivo imaging and tracking of host-microbiota in-

teractions via metabolic labeling of gut anaerobic bacteria. Nat Med 21:1091–1100.48. Okumura R, et al. (2016) Lypd8 promotes the segregation of flagellated microbiota

and colonic epithelia. Nature 532:117–121.49. Whitaker WR, Shepherd ES, Sonnenburg JL (2017) Tunable expression tools enable

single-cell strain distinction in the gut microbiome. Cell 169:538–546.e12.50. Cremer J, et al. (2016) Effect of flow and peristaltic mixing on bacterial growth in a

gut-like channel. Proc Natl Acad Sci USA 113:11414–11419.51. Valm AM, et al. (2011) Systems-level analysis of microbial community organization

through combinatorial labeling and spectral imaging. Proc Natl Acad Sci USA 108:4152–4157.

52. Valm AM, Mark Welch JL, Borisy GG (2012) CLASI-FISH: Principles of combinatorial

labeling and spectral imaging. Syst Appl Microbiol 35:496–502.53. McNulty NP, et al. (2011) The impact of a consortium of fermented milk strains on the

gut microbiome of gnotobiotic mice and monozygotic twins. Sci Transl Med 3:106ra106.

54. McNulty NP, et al. (2013) Effects of diet on resource utilization by a model human gut

microbiota containing Bacteroides cellulosilyticus WH2, a symbiont with an extensiveglycobiome. PLoS Biol 11:e1001637.

55. Cummings JH, et al. (1990) The amount and composition of large bowel contents inman. Gastroenterology 98:A408.

56. Sarma-Rupavtarm RB, Ge Z, Schauer DB, Fox JG, Polz MF (2004) Spatial distribution

and stability of the eight microbial species of the altered schaedler flora in the mousegastrointestinal tract. Appl Environ Microbiol 70:2791–2800.

57. Daims H, Lücker S, Wagner M (2006) Daime, a novel image analysis program formicrobial ecology and biofilm research. Environ Microbiol 8:200–213.

58. Johansson MEV, et al. (2008) The inner of the two Muc2 mucin-dependent mucus

layers in colon is devoid of bacteria. Proc Natl Acad Sci USA 105:15064–15069.59. Johansson MEV (2012) Fast renewal of the distal colonic mucus layers by the surface

goblet cells as measured by in vivo labeling of mucin glycoproteins. PLoS One 7:e41009.

60. Cheng H, Leblond CP (1974) Origin, differentiation and renewal of the four mainepithelial cell types in the mouse small intestine. I. Columnar cell. Am J Anat 141:461–479.

61. Cheng H (1974) Origin, differentiation and renewal of the four main epithelial celltypes in the mouse small intestine. II. Mucous cells. Am J Anat 141:481–501.

62. Cheng H, Leblond CP (1974) Origin, differentiation and renewal of the four mainepithelial cell types in the mouse small intestine. III. Entero-endocrine cells. Am J Anat141:503–519.

63. Cheng H (1974) Origin, differentiation and renewal of the four main epithelial celltypes in the mouse small intestine. IV. Paneth cells. Am J Anat 141:521–535.

64. Cheng H, Leblond CP (1974) Origin, differentiation and renewal of the four mainepithelial cell types in the mouse small intestine. V. Unitarian theory of the origin ofthe four epithelial cell types. Am J Anat 141:537–561.

65. Degen LP, Phillips SF (1996) Variability of gastrointestinal transit in healthy womenand men. Gut 39:299–305.

66. Padmanabhan P, Grosse J, Asad ABMA, Radda GK, Golay X (2013) Gastrointestinaltransit measurements in mice with 99mTc-DTPA-labeled activated charcoal usingNanoSPECT-CT. EJNMMI Res 3:60.

67. Dey N, et al. (2015) Regulators of gut motility revealed by a gnotobiotic model ofdiet-microbiome interactions related to travel. Cell 163:95–107.

68. Mark Welch JL, Rossetti BJ, Rieken CW, Dewhirst FE, Borisy GG (2016) Biogeographyof a human oral microbiome at the micron scale. Proc Natl Acad Sci USA 113:E791–E800.

69. Rawls JF, Samuel BS, Gordon JI (2004) Gnotobiotic zebrafish reveal evolutionarilyconserved responses to the gut microbiota. Proc Natl Acad Sci USA 101:4596–4601.

70. Melancon E, et al. (2017) Best practices for germ-free derivation and gnotobioticzebrafish husbandry. Methods Cell Biol 138:61–100.

71. Kim HJ, Li H, Collins JJ, Ingber DE (2016) Contributions of microbiome and mechanicaldeformation to intestinal bacterial overgrowth and inflammation in a human gut-on-a-chip. Proc Natl Acad Sci USA 113:E7–E15.

72. Moter A, Göbel UB (2000) Fluorescence in situ hybridization (FISH) for direct visuali-zation of microorganisms. J Microbiol Methods 41:85–112.

73. Amann R, Fuchs BM (2008) Single-cell identification in microbial communities byimproved fluorescence in situ hybridization techniques. Nat Rev Microbiol 6:339–348.

74. Hasegawa Y, Mark Welch JL, Rossetti BJ, Borisy GG (2017) Preservation of three-dimensional spatial structure in the gut microbiome. bioRxiv, 10.1101/175224.

75. Ludwig W, et al. (2004) ARB: A software environment for sequence data. NucleicAcids Res 32:1363–1371.

76. Yilmaz LS, Noguera DR (2004) Mechanistic approach to the problem of hybridizationefficiency in fluorescent in situ hybridization. Appl Environ Microbiol 70:7126–7139.

77. Schindelin J, et al. (2012) Fiji: An open-source platform for biological-image analysis.Nat Methods 9:676–682.

78. Valm AM, Oldenbourg R, Borisy GG (2016) Multiplexed spectral imaging of 120 dif-ferent fluorescent labels. PLoS One 11:e0158495.

79. Bernsen J (1986) Dynamic thresholding of grey-level images. Proceedings of theEighth International Conference on Pattern Recognition (IEEE Computer Society, LosAlamitos, CA), pp 1251–1255.

80. Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of imageanalysis. Nat Methods 9:671–675.

81. Amann RI, et al. (1990) Combination of 16S rRNA-targeted oligonuleotide probeswith flow cytometry for analyzing mixed microbial populations. Appl EnvironMicrobiol 56:1919–1925.

E9114 | www.pnas.org/cgi/doi/10.1073/pnas.1711596114 Mark Welch et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

8, 2

020