Interactions between host and intestinal crypt-resided biofilms are controlled by epithelial fucosylation

•Intestinal crypt is an important niche for in vivo biofilm formation during colitis•Epithelial fucosylation restrains crypt-resided pathogenic and commensal biofilms•Liberated fucose restricts biofilms by inhibiting biofilm-related gene expression•Fucose represses biofilm formation in vivo and alleviates experimental colitis As highly organized consortia of bacteria, biofilms have long been implicated in aggravating inflammation. However, our understanding regarding in vivo host-biofilm interactions in the complex tissue environments remains limited. Here, we show a unique pattern of crypt occupation by mucus-associated biofilms during the early stage of colitis, which is genetically dependent on bacterial biofilm-forming capacity and restricted by host epithelial α1,2-fucosylation. α1,2-Fucosylation deficiency leads to markedly augmented crypt occupation by biofilms originated from pathogenic Salmonella Typhimurium or indigenous Escherichia coli, resulting in exacerbated intestinal inflammation. Mechanistically, α1,2-fucosylation-mediated restriction of biofilms relies on interactions between bacteria and liberated fucose from biofilm-occupied mucus. Fucose represses biofilm formation and biofilm-related genes in vitro and in vivo. Finally, fucose administration ameliorates experimental colitis, suggesting therapeutic potential of fucose for biofilm-related disorders. This work illustrates host-biofilm interactions during gut inflammation and identifies fucosylation as a physiological strategy for restraining biofilm formation. As highly organized consortia of bacteria, biofilms have long been implicated in aggravating inflammation. However, our understanding regarding in vivo host-biofilm interactions in the complex tissue environments remains limited. Here, we show a unique pattern of crypt occupation by mucus-associated biofilms during the early stage of colitis, which is genetically dependent on bacterial biofilm-forming capacity and restricted by host epithelial α1,2-fucosylation. α1,2-Fucosylation deficiency leads to markedly augmented crypt occupation by biofilms originated from pathogenic Salmonella Typhimurium or indigenous Escherichia coli, resulting in exacerbated intestinal inflammation. Mechanistically, α1,2-fucosylation-mediated restriction of biofilms relies on interactions between bacteria and liberated fucose from biofilm-occupied mucus. Fucose represses biofilm formation and biofilm-related genes in vitro and in vivo. Finally, fucose administration ameliorates experimental colitis, suggesting therapeutic potential of fucose for biofilm-related disorders. This work illustrates host-biofilm interactions during gut inflammation and identifies fucosylation as a physiological strategy for restraining biofilm formation. Biofilms consist of microbial communities embedded in extracellular matrix with highly organized structures, protecting the encased microbes from adverse environmental conditions.1Lynch A.S. Robertson G.T. Bacterial and fungal biofilm infections.Annu. Rev. Med. 2008; 59: 415-428https://doi.org/10.1146/annurev.med.59.110106.132000Google Scholar,2Domingue J.C. Drewes J.L. Merlo C.A. Housseau F. Sears C.L. Host responses to mucosal biofilms in the lung and gut.Mucosal Immunol. 2020; 13: 413-422https://doi.org/10.1038/s41385-020-0270-1Google Scholar Biofilm-forming microorganisms are phylogenetically diverse, yet share common features, such as adherence and adaptation to biotic surfaces or man-made materials, formation of living cellular clusters encased in extracellular polymeric substance matrix, and exhibition of viscoelasticity properties.3Koo H. Allan R.N. Howlin R.P. Stoodley P. Hall-Stoodley L. Targeting microbial biofilms: current and prospective therapeutic strategies.Nat. Rev. Microbiol. 2017; 15: 740-755https://doi.org/10.1038/nrmicro.2017.99Google Scholar,4Parsek M.R. Singh P.K. Bacterial biofilms: an emerging link to disease pathogenesis.Annu. Rev. Microbiol. 2003; 57: 677-701https://doi.org/10.1146/annurev.micro.57.030502.090720Google Scholar,5Rumbaugh K.P. Sauer K. Biofilm dispersion.Nat. Rev. Microbiol. 2020; 18: 571-586https://doi.org/10.1038/s41579-020-0385-0Google Scholar Biofilm formation initiates a well-coordinated program including alterations in gene expression, intercellular communication, and community-level communication.6Liu J. Prindle A. Humphries J. Gabalda-Sagarra M. Asally M. Lee D.y.D. Ly S. Garcia-Ojalvo J. Süel G.M. Metabolic co-dependence gives rise to collective oscillations within biofilms.Nature. 2015; 523: 550-554https://doi.org/10.1038/nature14660Google Scholar,7Liu J. Martinez-Corral R. Prindle A. Lee D.Y.D. Larkin J. Gabalda-Sagarra M. Garcia-Ojalvo J. Süel G.M. Coupling between distant biofilms and emergence of nutrient time-sharing.Science. 2017; 356: 638-642https://doi.org/10.1126/science.aah4204Google Scholar For instance, in Escherichia coli, genes controlling curli fibers and type I fimbriae production such as csgD and fimB, motility regulation such as motA and fliA and quorum sensing such as luxS and mqsR are involved in biofilm formation and development.8Beloin C. Roux A. Ghigo J.M. Escherichia coli biofilms.Curr. Top. Microbiol. Immunol. 2008; 322: 249-289https://doi.org/10.1007/978-3-540-75418-3_12Google Scholar One characteristic property of the pathogenic biofilm lifestyle is the ability to survive in the presence of antimicrobial agents and host immunity, which contributes to drug resistance and persistent and recurrent infections.9Lebeaux D. Ghigo J.M. Beloin C. Biofilm-related infections: bridging the gap between clinical management and fundamental aspects of recalcitrance toward antibiotics.Microbiol. Mol. Biol. Rev. 2014; 78: 510-543https://doi.org/10.1128/mmbr.00013-14Google Scholar However, to date, there is scarce knowledge regarding how host and biofilm interact in complex tissue microenvironments and how the host prevents and overcomes potentially detrimental biofilm formation in vivo. Thus, investigations on host-biofilm interactions represent one fundamental yet understudied aspect of biofilm biology and would greatly advance our understanding of biofilm pathogenicity in relevant disease conditions. In the intestine, emerging evidence has linked colonic biofilms to the pathogenesis of inflammatory bowel diseases (IBDs) and colorectal cancers.10Li S. Konstantinov S.R. Smits R. Peppelenbosch M.P. Bacterial Biofilms in Colorectal Cancer Initiation and Progression.Trends Mol. Med. 2017; 23: 18-30https://doi.org/10.1016/j.molmed.2016.11.004Google Scholar,11Motta J.P. Wallace J.L. Buret A.G. Deraison C. Vergnolle N. Gastrointestinal biofilms in health and disease.Nat. Rev. Gastroenterol. Hepatol. 2021; 18: 314-334https://doi.org/10.1038/s41575-020-00397-yGoogle Scholar In colorectal cancer patients, biofilms have been observed located in the inner mucus layer above the epithelial surface, which may promote inflammatory responses and aggravating inflammation-associated oncogenesis.2Domingue J.C. Drewes J.L. Merlo C.A. Housseau F. Sears C.L. Host responses to mucosal biofilms in the lung and gut.Mucosal Immunol. 2020; 13: 413-422https://doi.org/10.1038/s41385-020-0270-1Google Scholar These observations in patients have sparked interest in long-neglected biofilms in intestinal diseases and raise a number of fundamental questions regarding intestine-associated biofilms. First, it has been conventionally conceived that bacterium-intrinsic alterations at the individual or population levels drive biofilm formation.12Knippel R.J. Drewes J.L. Sears C.L. The Cancer Microbiome: Recent Highlights and Knowledge Gaps.Cancer Discov. 2021; 11: 2378-2395https://doi.org/10.1158/2159-8290.cd-21-0324Google Scholar Given the close proximity of biofilms with intestinal tissues, it is unclear whether host factors are involved in facilitating or limiting biofilm formation. Second, as the intestinal tissues uniquely consist of well-organized crypt structures, the extents of biofilm-epithelium interactions are not well characterized and it is plausible that biofilms make extensive contacts with intestinal tissues in addition to being confined to the epithelial surface. Third, the contributions of pathogenic versus commensal microorganisms to intestinal biofilms are poorly understood. We are merely at the dawn of appreciating the importance and unveiling the key regulators of intestinal biofilms, and many questions of great significance and urgency, including the ones listed above, await answers. However, due to the lack of reliable animal models for studying intestinal biofilms, genetic dissections of bacterial and host regulators and establishment of causal connections for biofilms in intestinal inflammation in vivo remain challenging. In a healthy host, bacteria might encounter great difficulties in stably establishing pathogenic biofilms in the gut due to the robust host defense mechanisms such as epithelial barrier function. The barrier functions of intestinal epithelial cells (IECs) are the first line of host defense against microbial invasion of tissues.13Gallo R.L. Hooper L.V. Epithelial antimicrobial defence of the skin and intestine.Nat. Rev. Immunol. 2012; 12: 503-516https://doi.org/10.1038/nri3228Google Scholar,14Peterson L.W. Artis D. Intestinal epithelial cells: regulators of barrier function and immune homeostasis.Nat. Rev. Immunol. 2014; 14: 141-153Google Scholar Defects in IEC barrier functions often result in increased susceptibility to infections and commensal bacterial invasions, which are associated with disrupted homeostasis and IBDs.15Ramanan D. Cadwell K. Intrinsic Defense Mechanisms of the Intestinal Epithelium.Cell Host Microbe. 2016; 19: 434-441https://doi.org/10.1016/j.chom.2016.03.003Google Scholar,16Caruso R. Lo B.C. Núñez G. Host-microbiota interactions in inflammatory bowel disease.Nat. Rev. Immunol. 2020; 20: 411-426https://doi.org/10.1038/s41577-019-0268-7Google Scholar Heavily glycosylated mucus secreted from goblet cells covers the mucosal surfaces, which constitute a key component of the host defense system.17Kudelka M.R. Stowell S.R. Cummings R.D. Neish A.S. Intestinal epithelial glycosylation in homeostasis and gut microbiota interactions in IBD.Nat. Rev. Gastroenterol. Hepatol. 2020; 17: 597-617https://doi.org/10.1038/s41575-020-0331-7Google Scholar,18Johansson M.E.V. Hansson G.C. Immunological aspects of intestinal mucus and mucins.Nat. Rev. Immunol. 2016; 16: 639-649https://doi.org/10.1038/nri.2016.88Google Scholar α1,2-Fucosylation, a form of O-linked glycosylation of IECs, is catalyzed by fucosyltransferase 1 (Fut1) and Fut2, which mediate the addition of L-fucose via an α1-2 linkage to the terminal β-D-galactose residues of mucosal glycans.19Goto Y. Uematsu S. Kiyono H. Epithelial glycosylation in gut homeostasis and inflammation.Nat. Immunol. 2016; 17: 1244-1251https://doi.org/10.1038/ni.3587Google Scholar α1,2-Fucose decorates mucosal glycans on the apical surface of epithelium, which provides the unique opportunity for α1,2-fucose to contact and interact with luminal microbes. Polymorphisms in the FUT2 gene are associated with increased risks of various human disorders, including bacterial infections and IBDs.19Goto Y. Uematsu S. Kiyono H. Epithelial glycosylation in gut homeostasis and inflammation.Nat. Immunol. 2016; 17: 1244-1251https://doi.org/10.1038/ni.3587Google Scholar Animal studies have also demonstrated that α1,2-fucosylation protects mice against infections by Salmonella Typhimurium or Citrobacter rodentium.20Pickard J.M. Maurice C.F. Kinnebrew M.A. Abt M.C. Schenten D. Golovkina T.V. Bogatyrev S.R. Ismagilov R.F. Pamer E.G. Turnbaugh P.J. Chervonsky A.V. Rapid fucosylation of intestinal epithelium sustains host-commensal symbiosis in sickness.Nature. 2014; 514: 638-641https://doi.org/10.1038/nature13823Google Scholar,21Goto Y. Obata T. Kunisawa J. Sato S. Ivanov I.I. Lamichhane A. Takeyama N. Kamioka M. Sakamoto M. Matsuki T. et al.Innate lymphoid cells regulate intestinal epithelial cell glycosylation.Science. 2014; 3451254009https://doi.org/10.1126/science.1254009Google Scholar,22Pham T.A.N. Clare S. Goulding D. Arasteh J.M. Stares M.D. Browne H.P. Keane J.A. Page A.J. Kumasaka N. Kane L. et al.Epithelial IL-22RA1-mediated fucosylation promotes intestinal colonization resistance to an opportunistic pathogen.Cell Host Microbe. 2014; 16: 504-516https://doi.org/10.1016/j.chom.2014.08.017Google Scholar However, how epithelial α1,2-fucosylation protects the host against intestinal microbes is still unclear, and whether the interactions between host and invasive microorganisms controlled by mucosal α1,2-fucosylation involve biofilm formation remains unknown. In order to study host-biofilm interactions in a well-defined system, we established animal models of intestinal biofilms that rendered reliable detections of biofilms in situ in intestinal tissue environments feasible and enabled genetic manipulations of the bacterial as well as host components during the process of biofilm formation. Leveraging on the in vivo models, we investigated host-biofilm interactions in intestinal inflammation induced by pathogens and symbionts. Interestingly, we identified mucus-associated biofilm formation as a previously unrecognized mechanism for bacterial occupation of the intestinal crypt and depicted host-biofilm interactions mediated by epithelial α1,2-fucosylation. Our results establish a key role for fucosylation in host defense against biofilm-mediated occupation of tissues by pathogens and opportunistically commensal bacteria and provide insights into the pathogenic mechanisms underlying the contributions of mucosal biofilms to intestinal diseases. To determine how the interactions between host and invasive microorganisms are controlled by mucosal α1,2-fucosylation, we first analyzed the Fut2 gene expression and found it to be predominantly restricted to gastrointestinal mucosal tissues (Figures S1A and S1B). We then characterized tissue distributions of fucosylation signals in situ by immunostaining with the α1,2-fucose-recognizing lectin Ulex europaeus agglutinin-1 (UEA-1), which showed that α1,2-fucose was enriched in the granules of goblet cells, mucus in the crypts, and the inner mucus layer in the large intestine of wild-type (WT) mice (Figures S1C and S1D). In contrast, other types of glycans did not exhibit similar tissue localization patterns as α1,2-fucose (Figures S1E–S1H), prompting us to investigate the specific role of large-intestinal epithelial α1,2-fucosylation in host-microbiota interaction. For the following functional studies, Fut2-deficient (Fut2−/−) and Fut2-sufficient (Fut2+/−) cohoused littermates were used, whereby ablation of Fut2 led to efficient deletion of the Fut2 gene and loss of α1,2-fucose on goblet cells and mucus (Figures S2A and S2B). At steady state, Fut2−/− mice did not develop spontaneous colitis, as shown by intact tissue architecture, normal weights, and unaltered expression of several immune-related genes (Figures S2C–S2G). Intriguingly, immunofluorescence staining of Muc2 and bacteria showed increased bacterial aggregation penetrating into the inner mucus layer of colonic epithelia in Fut2−/− mice compared to littermate controls, whereas there were no differences in the thickness of Muc2-positive mucus layers between these two genotypes of mice (Figures 1A and 1B ). In line with the observations of bacterial penetration, detection of bacteria adjacent to the colonic epithelial surfaces by fluorescence in situ hybridization (FISH) demonstrated decreased extents of bacterial exclusion from colonic epithelial surfaces in Fut2−/− mice (Figures 1C and 1D). Moreover, decreased segregation of bacteria was further corroborated by periodic acid-Schiff (PAS)/hematoxylin staining, which demonstrated that Fut2−/− mice displayed thinner bacterium-free mucus layers on the colonic epithelia than Fut2+/− littermates (Figures 1E and 1F). Thus, these results indicate that commensal bacteria are apt to aggregate in the inner mucus of large intestine at steady state upon loss of epithelial α1,2-fucosylation. Commensal invasion of inner mucus and epithelium prior to the onset of inflammation has been observed in murine colitis models.23Johansson M.E.V. Gustafsson J.K. Sjöberg K.E. Petersson J. Holm L. Sjövall H. Hansson G.C. Bacteria penetrate the inner mucus layer before inflammation in the dextran sulfate colitis model.PLoS One. 2010; 5e12238https://doi.org/10.1371/journal.pone.0012238Google Scholar,24Johansson M.E.V. Gustafsson J.K. Holmén-Larsson J. Jabbar K.S. Xia L. Xu H. Ghishan F.K. Carvalho F.A. Gewirtz A.T. Sjövall H. Hansson G.C. Bacteria penetrate the normally impenetrable inner colon mucus layer in both murine colitis models and patients with ulcerative colitis.Gut. 2014; 63: 281-291https://doi.org/10.1136/gutjnl-2012-303207Google Scholar To determine whether epithelial α1,2-fucosylation is required for protection from commensal bacterial invasion in dextran sulfate sodium (DSS)-induced colitis, Fut2−/− and Fut2+/− littermates were administered 4% DSS in the drinking water for 5 days. Fut2−/− mice exhibited profound weight loss and increased severity of intestinal inflammation with high mortality, while Fut2+/− mice were mildly affected (Figures 1G–1J). In addition, Fut2−/− mice showed more STAT3-activated IECs at day 10 than Fut2+/− littermates (Figures 1K and 1L), indicative of enhanced epithelial immune activation.25Pickert G. Neufert C. Leppkes M. Zheng Y. Wittkopf N. Warntjen M. Lehr H.A. Hirth S. Weigmann B. Wirtz S. et al.STAT3 links IL-22 signaling in intestinal epithelial cells to mucosal wound healing.J. Exp. Med. 2009; 206: 1465-1472https://doi.org/10.1084/jem.20082683Google Scholar,26Taniguchi K. Wu L.W. Grivennikov S.I. de Jong P.R. Lian I. Yu F.X. Wang K. Ho S.B. Boland B.S. Chang J.T. et al.A gp130-Src-YAP module links inflammation to epithelial regeneration.Nature. 2015; 519: 57-62https://doi.org/10.1038/nature14228Google Scholar,27Fichtner-Feigl S. Kesselring R. Strober W. Chronic inflammation and the development of malignancy in the GI tract.Trends Immunol. 2015; 36: 451-459https://doi.org/10.1016/j.it.2015.06.007Google Scholar Next, we assessed the occupation of mucosal tissues by commensal bacteria at the early stage of colitis via analyzing the levels of their genomic DNA relative to host DNA in the mucosal tissues (Figures 1M and S2H). Notably, among various bacterial species, tissue-associated indigenous E. coli was consistently increased in both cecum and colon in Fut2−/− mice compared to Fut2+/− mice (Figure 1M), which was corroborated by immunofluorescence examination of cecum sections at the early stage of DSS-induced colitis (Figure 1N). On the contrary, there were no significant differences in the fecal E. coli numbers between Fut2−/− and Fut2+/− mice (Figures S2I and S2J). Moreover, compared with their littermate controls, Fut2−/− mice showed decreased segregation of indigenous E. coli from colonic epithelial surfaces at the recovery stage of DSS-induced colitis (Figures 1O and 1P). Collectively, these results demonstrate that epithelial α1,2-fucosylation facilitates host defense against commensal bacterial occupation of mucus and ameliorates experimental colitis. Following the above observations with commensal bacteria, we next assessed whether the segregation of pathogens from large-intestinal mucus barrier was also controlled by α1,2-fucosylation in the S. Typhimurium infection-induced colitis model (Figure 2A, left).28Barthel M. Hapfelmeier S. Quintanilla-Martínez L. Kremer M. Rohde M. Hogardt M. Pfeffer K. Rüssmann H. Hardt W.D. Pretreatment of mice with streptomycin provides a Salmonella enterica serovar Typhimurium colitis model that allows analysis of both pathogen and host.Infect. Immun. 2003; 71: 2839-2858https://doi.org/10.1128/iai.71.5.2839-2858.2003Google Scholar Despite the fact that the luminal S. Typhimurium burdens were comparable between Fut2−/− and Fut2+/− littermates (Figure 2B), Fut2−/− mice showed increased S. Typhimurium occupation of mucus and the cecal crypt during early infection compared to Fut2+/− littermates, as assessed by staining of S. Typhimurium and mucus (Figures 2C and 2D). In addition, Fut2−/− mice displayed decreased extents of S. Typhimurium exclusion from colonic epithelial surfaces, as indicated by evident S. Typhimurium occupation of inner mucus and tissues compared to their littermate controls (Figures 2E and 2F). Intriguingly, most of the S. Typhimurium had reached the bottom of cecal crypts in Fut2−/− mice during early infection (Figure 2G). Occupation of mucus and crypts by bacteria correlated with the subsequently observed exacerbated intestinal damage and increased proliferation of the epithelium in Fut2−/− mice (Figures S3A–S3J). Although cecum epithelium tips are a hotspot for S. Typhimurium infection,29Furter M. Sellin M.E. Hansson G.C. Hardt W.D. Mucus Architecture and Near-Surface Swimming Affect Distinct Salmonella Typhimurium Infection Patterns along the Murine Intestinal Tract.Cell Rep. 2019; 27: 2665-2678.e3https://doi.org/10.1016/j.celrep.2019.04.106Google Scholar loss of α1,2-fucosylation did not influence the bacterial burden of cecum epithelium tips (Figures S3K and S3L), which may be attributed to the lack of mucus coverage on the cecum epithelium tips.29Furter M. Sellin M.E. Hansson G.C. Hardt W.D. Mucus Architecture and Near-Surface Swimming Affect Distinct Salmonella Typhimurium Infection Patterns along the Murine Intestinal Tract.Cell Rep. 2019; 27: 2665-2678.e3https://doi.org/10.1016/j.celrep.2019.04.106Google Scholar Notably, α1,2-fucosylation was not required for protection against S. Typhimurium infection in the small-intestine-dependent typhoid fever model (Figures S3M and S3N), suggesting large-intestine-specific effects of α1,2-fucosylation. In order to exclude the streptomycin-specific influence on the infection-induced colitis model, we next pretreated the mice with another antibiotic, vancomycin, prior to oral infection with S. Typhimurium (Figure 2A, right). Likewise, α1,2-fucosylation protected the host from S. Typhimurium-induced colitis and systemic infection in the vancomycin-pretreated model (Figures S3O–S3V). Importantly, Fut2−/− mice also displayed increased S. Typhimurium occupation of mucus and crypts in the cecum and colon at the early stage of colitis (Figures 2H–2K). Together, these results indicate that epithelial fucosylation deficiency leads to increased S. Typhimurium occupation of mucus and crypts at the early stage of colitis. Next, we sought to delineate the mechanisms by which α1,2-fucosylation limited bacterial crypt occupation. Under high-resolution confocal microscopy, S. Typhimurium mostly existed as single or planktonic cells negative for 4′,6-diamidino-2-phenylindole (DAPI) in the crypts of Fut2+/− animals during early infection (Figures 3A–3C , upper panels). In contrast, in the cecal crypts of Fut2−/− mice, high-cell-density S. Typhimurium aggregates were readily detectable, which were strongly co-stained with Muc2 and DAPI, indicative of extracellular bacterial DNA, and thus presented typical features of biofilms (Figures 3A–3C, bottom panels).3Koo H. Allan R.N. Howlin R.P. Stoodley P. Hall-Stoodley L. Targeting microbial biofilms: current and prospective therapeutic strategies.Nat. Rev. Microbiol. 2017; 15: 740-755https://doi.org/10.1038/nrmicro.2017.99Google Scholar Quantification of biofilm-containing crypts showed that loss of α1,2-fucosylation resulted in a significant increase of invasive mucus-associated S. Typhimurium biofilms in the cecal crypts early during infection (Figure 3D). Next, scanning electron microscopy (SEM) was used to better visualize S. Typhimurium biofilms. Compared to Fut2+/− littermates, Fut2−/− mice demonstrated significantly increased IEC-adhering bacterial aggregates adhering to the cecal epithelial surfaces (Figures 3E and 3F). Importantly, bacterial aggregates from Fut2+/− mice mostly showed smooth cell walls with minimal matrix attachment, yet the bacterial aggregates in Fut2−/− mice were encased in the extracellular matrix surrounding bacterial cells (Figure 3E), representing typical biofilm morphology. In addition, transmission electron microscopy (TEM) of cecal mucosa also revealed apparent bacterial aggregates surrounded by extracellular matrix in Fut2−/− mice, whereas Fut2+/− mice harbored bacteria as single cells with minimal surrounding matrix (Figure 3G). Intriguingly, the matrix of bacterial biofilms adhered to and filled in the microvilli of cecal IECs in Fut2−/− mice (Figure 3G, inset 3). In conjunction with the observations that S. Typhimurium biofilms were tightly associated with secreted Muc2 and goblet cells in the crypts (Figure 3C), these TEM data implicated strong adherence to IECs by mucus-associated S. Typhimurium biofilms in the crypts of Fut2−/− mice early during infection. Interestingly, the S. Typhimurium biofilms contained certain areas with strong Syto9 signals and weak Salmonella signals (Figure S3W), which suggested that the S. Typhimurium biofilms might contain other symbiotic bacteria. Furthermore, in contrast to 18 hours post infection (hpi), we did not observe biofilm formation in the crypt from both genotypes of mice at 24 hpi (Figure 3H), and we hypothesized that bacteria might have detached from the biofilm matrix and dispersed to infect host cells according to the current notion of the biofilm dispersal stage.5Rumbaugh K.P. Sauer K. Biofilm dispersion.Nat. Rev. Microbiol. 2020; 18: 571-586https://doi.org/10.1038/s41579-020-0385-0Google Scholar Consistent with our hypothesis, Fut2−/− mice showed an increased number of S. Typhimurium-positive expulsed enterocytes released from the cecum compared to Fut2+/− mice at 24 hpi (Figures 3H and 3I). Altogether, these data demonstrated that α1,2-fucosylation deficiency led to crypt occupation by mucus-associated S. Typhimurium biofilms at the early stage of infection, implying that epithelial α1,2-fucosylation acts as an endogenous host mechanism to limit crypt-invading biofilm formation (Figure 3J). Based on the results from S. Typhimurium infection, we next asked whether the opportunistically commensal bacteria such as E. coli could form mucus-associated biofilms to confer intestinal crypt occupation and whether α1,2-fucosylation could restrain the mucus-associated commensal biofilms by utilizing a vancomycin-induced E. coli overgrowth model (Figures 4A and S4A). Upon vancomycin pre-treatment, Fut2−/− mice exhibited enhanced susceptibility to DSS-induced colitis shown by profound weight loss, compromised survival, and severe pathological changes compared with Fut2+/− littermates (Figures S4B–S4F). During the early stage of colitis, tissue-associated E. coli was specifically enriched in Fut2−/− mice, whereas there were no significant differences in the fecal E. coli numbers between Fut2−/− and Fut2+/− mice (Figures S4G–S4I). Importantly, Fut2 deficiency led to a significantly increased incidence of invasive Muc2-associated E. coli biofilms in the crypts of cecum and colon at the early stage of DSS-induced colitis as indicated by high-cell-density E. coli aggregates co-localized with Muc2 and bacterial DNA (Figures 4B–4E). Moreover, TEM images revealed matrix-encased invasive bacterial aggregates adhering to the microvilli of cecal IECs in Fut2−/− mice, while, in Fut2+/− mice, invasive bacteria existed as single cells without adhering to the IEC microvilli (Figure 4F). Next, we wished to determine whether biofilm-forming capacity contributed to the occupation of crypts by E. coli. Genetically deleting the csgD gene in E. coli impaired biofilm formation without compromising bacterial growth in vitro (Figures 4G and 4H), as expected.8Beloin C. Roux A. Ghigo J.M. Escherichia coli biofilms.Curr. Top. Microbiol. Immunol. 2008; 322: 249-289https://doi.org/10.1007/978-3-540-75418-3_12Google Scholar When inoculated in vivo with WT E. coli in a 1:1 mixture into Fut2−/− mice, ΔcsgD E. coli exhibited markedly reduced biofilm-forming capacity in the crypts (Figures 4I–4K), genetically corroborating a critical role of biofilms in crypt occupation. Although many E. coli strains express curli at lower temperatures,30Serra D.O. Richter A.M. Hengge R. Cellulose as an architectural element in spatially structured Escherichia coli biofilms.J. Bacteriol. 2013; 195: 5540-5554https://doi.org/10.1128/jb.00946-13Google Scholar we consistently observed low-level expression of curli in WT E. coli (BW25113) in vitro at 37°C, with the ΔcsgD E. coli strain serving as a negative control for indication of specificity (Figure S4J). Importantly, high levels of curli protein (csgA) production were observed in crypt-residing E. coli biofilms instead of the luminal population of E. coli (Figure 4L), further supporting that the intestinal crypt was an important niche for in vivo biofilm formation. Similar to S. Typhimurium, E. coli biofilms also contained a small number of other symbiotic bacteria (Figure S4K). This observation suggests that fucosylation might inhibit the occupation of crypts by other symbiotic bacteria, which was supported by the increased susceptibility of Fut2−/− mice to DSS-induced colitis even in the absence of E. coli (Figures S4L–S4N). Collectively, these data suggest that indigenous E. coli occupation of crypts requires mucus-associated biofilm formation, which is restrained by epithelial α1,2-fucosylation (Figure 4M). We next sought to investigate the mechanisms underlying the control of invasive mucus-associated biofilms by epithelial α1,2-fucosylation. Loss of α1,2-fucosylation minimally affected the colon luminal commensal compositions and mucus secretion by goblet cells (Figures S5A–S5J). Upon close examination of the small number of commensal E. coli biofilms detected in the Fut2-sufficient mice, we found that