Abstract
Lactobacillus displacement from the vaginal microbiome associates with adverse health outcomes and is linked to increased risk of preterm birth. Glycans mediate bacterial adhesion events involved in colonisation and infection. Using customised glycan microarrays, we establish glycan interaction profiles of vaginal bacteria implicated in reproductive health. Glycan binding signatures of the opportunistic pathogens Escherichia coli, Fusobacterium nucleatum and Streptococcus agalactiae to oligomannose N-glycans, galactose-terminating glycans and hyaluronic acid, respectively are highly distinct from Lactobacillus commensals. Binding to sulphated glycosaminoglycans by vaginal bacteria is pH dependent, as is binding to neutral and sialic acid-terminating glycans by F. nucleatum. Adhesion of Lactobacillus crispatus, Lactobacillus iners, Gardnerella vaginalis, S. agalactiae and F. nucleatum to vaginal epithelial cells is partially mediated by chondroitin sulphate. S. agalactiae binding to chondroitin sulphate C oligosaccharides is inhibited by L. crispatus. This study highlights glycans as mediators of vaginal bacterial binding events involved in reproductive health and disease.
Similar content being viewed by others
Introduction
The vaginal microbiome is an important mediator of reproductive health and is a risk modifier of adverse pregnancy outcomes such as preterm birth1,2,3. Loss of Lactobacillus species and overgrowth of gram-negative bacteria are microbiological hallmarks of the common vaginal condition, bacterial vaginosis (BV)4. In pregnancy, Lactobacillus crispatus dominance of the cervicovaginal niche is protective against preterm birth5,6,7, whereas Lactobacillus iners or a shift towards a high diversity composition is associated with increased risk of preterm birth8,9. This high diversity vaginal microbiota is often enriched in opportunistic, potentially pathogenic bacteria such as Gardnerella vaginalis, Streptococcus agalactiae, Sneathia and Prevotella species that promote reproductive tract inflammation8,10. L. crispatus prevents pathogen colonisation of the vaginal microbiome via several mechanisms including production of antimicrobial compounds and lactic acid, leading to low vaginal pH < 4.5 in most healthy asymptomatic women11,12.
Glycans are made up of sugars arranged to form molecular epitopes specifically recognised by antibodies, lectins and carbohydrate binding proteins13. They form the bulk of the glycocalyx and are the main component of mucus in human secretions, including the cervicovaginal fluid (CVF)14. Glycans constitute the first point of contact in many cell-cell interactions, including host-microbiota interactions8. Host glycans modulate dynamics of microbial colonisation, serving as attachment and adhesion sites as well as nutrient sources15,16,17,18. Microbes can modify host glycans to avoid immune surveillance19 during infection or induce immune tolerance through molecular mimicry20,21. There is increasing evidence that host glycan-bacteria interactions are important modifiers of microbiota composition in the lower reproductive tract, yet the knowledge on specific glycan structures involved is limited17,18,20,22. Glycan microarrays offer a powerful strategy to elucidate substrate specificity of glycan binding proteins, including those derived from microbes such as viruses, bacteria or parasites23,24,25,26.
Here, we report the development of customised glycan microarrays for the detection and characterisation of the glycan binding specificities of key bacterial species from the vaginal microbiota implicated in reproductive health, like L. crispatus, L. iners and G. vaginalis and opportunistic pathogens, including S. agalactiae, E. coli and F. nucleatum, whose presence is associated with dysbiosis and adverse pregnancy and neonatal outcomes15,27. Our results reveal distinct glycan binding profiles by the opportunistic pathogens studied here, when compared collectively to commensal species within the vaginal microbiota and highlight pH-dependent interactions with sulphated glycosaminoglycans (GAGs) such as chondroitin sulphate (CS), a key mediator of bacterial adherence to vaginal epithelial cells.
Results
Glycan library and microarray design for bacterial binding studies
A library of 187 sequence-defined glycan probes (Supplementary Data 1) was constructed, and these were included in six different glycan microarray sets (Supplementary Fig. 1) used for the study. This library was designed to encompass representative glycans and glyco-epitopes from the glycocalyx and displayed on mucins, which form the major structural component of CVF28 (Fig. 1). Glycans were converted into amino-terminating probes for their immobilisation on N-hydroxysuccinimide (NHS)-functionalised glass slides, producing covalent glycan microarrays. Mass spectrometry (MS) and, when necessary, NMR, were used for glycan probe characterization (Source Data File 1). As part of the microarray quality control (QC), the immobilisation and epitope integrity of the glycan probes used in this study were assessed using a collection of 33 sequence-specific glycan recognition systems (Supplementary Data 2), which exhibited the expected binding patterns with only a few exceptions (Source Data File 2) of additional or weak binding to related glycans not described before, and which are noted in the MIRAGE document under the Section of Glycan identification and quality control (Supplementary Data 3).
A Pie chart representing the glycan categories included in a customised sequence-defined glycan library for binding studies of vaginal microbiota. Numbers indicate the number of probes in main categories (Supplementary Data 1). Examples of glycans in each category are shown. PS polysaccharide, NS non-sulphated. B Diagram representing the methodology used for analysis of whole bacteria binding to sequence-defined glycans on microarrays. Glycans are covalently immobilised on NHS glass slides and binding performed with fluorescently labelled bacterial cultures of vaginal microbiota. C Bacterial collection with strain specification used for the study (Supplementary Data 4). * Strains purchased from ATCC or DMZ collections.
The well characterised binding of the pathogenic urinary tract infection (UTI)-associated strain E. coli C600 to oligomannose N-glycans was used to optimise conditions for bacterial binding29. Both fixed and live fluorescently labelled E. coli C600 bound specifically to oligomannose N-glycans on Microarray set 1 (Fig. 2A) with 69 glycan probes including N-, O-, blood group related, sialylated and N-acetylglucosamine (GlcNAc)-terminating glycans (Supplementary Fig. 2A). Consistent with previous findings of oligomannose binding by the type 1 pilus adhesin FimH of uropathogenic E. coli30, strong binding to the N-glycan Man5 was observed across several independent experiments (Supplementary Fig. 2B)29. Concanavalin A, a plant lectin with known high avidity for α-linked mannose31, used as control in the same assay, showed strong binding to oligomannose N-glycans and weak binding to the complex-type N-glycan NA2. Glycan binding of three other E. coli strains (E. coli 789, 900 and 901) isolated from the urine of UTI patients32 was also tested (Fig. 2B, C). E. coli 789 and 901 strains strongly bound oligomannose N-glycans. Binding by E. coli 900 to oligomannose N-glycans was comparatively weak. Electron microscopy (EM) analysis of the E. coli isolates revealed that E. coli 900 displays substantially shorter and fewer fimbriae compared with the three other strains tested (Fig. 2D).
A Printing layout of glycan Microarray set 1 indicating the ___location of the main glycan categories for the 69 sequence-defined glycan probes included in the microarray (Supplementary Data 1). B Bar charts showing the mean fluorescence intensities of bacterial binding signals to glycans printed on quadruplicated spots on the glycan Microarray set 1. Error bars represent the SD of quadruplicates for each glycan probe on the microarray (n = 1). Assay is representative of three independent experiments. Structures of bound glycans are shown. Fixed cells were used. Source data are provided as a Source Data File. C Top: Microarray images of fixed fluorescently labelled bacteria binding to glycans on Microarray set 1. Bottom: Heatmaps of relative binding strengths amongst the four E. coli strains tested to glycans on Microarray set 1. Colour code: Dark blue (0–1 %); Light blue (1–10 %); Yellow (10–30 %); Orange (30–70%); Red (70–100%). 100%, the maximum binding score observed across the four experiments. D Electron-microscopy images of whole bacterial cells of E. coli strains C600, 789, 900, and 901. Enlargements showing differences in fimbriae sizes. White bars in enlargements indicate 200 nm. Images are representative of two independent experiments.
Comparison of glycan binding profiles of vaginal commensals and pathogens
We next performed glycan binding profiling of key vaginal commensals such as L. crispatus and L. iners, and potentially pathogenic bacteria including G. vaginalis, S. agalactiae, E. coli and F. nucleatum. Binding experiments were conducted on a total of 22 strains, 17 of which were patient-derived isolates32 (Supplementary Data 4) at mild acidic (pH 4) and neutral pH (pH 7). This was to reflect the pH gradient within the female reproductive tract which ranges from pH <4.5 in the lower vagina, to pH 6.7-7.2 in the cervix and pH 7.5-8.0 in the upper uterine cavity11. We first examined glycan binding to a set of 90 neutral and sialylated probes on Microarray set 2 (Probes #1-90; Supplementary Data 1). Glycan binding footprints are summarised in Fig. 3, with representative strain-level binding for each species presented in Supplementary Fig. 3. Bacterial binding was not observed to any of the short glycan probes with an amino-propyl (-sp) linker (probes #91-110 on Microarray set 3; Supplementary Data 1), despite showing expected binding with lectins and antibodies in QC tests. Oligomannose N-glycan binding by fimbriated E. coli strains was highly specific, a feature not shared by other species tested (Fig. 3 and Supplementary Fig. 3). While almost all species showed strong binding profiles at pH 4 and pH 7 to a subset of neutral and sialylated glycans, F. nucleatum 23726 strongly bound fucosylated, sialylated and galactose (Gal)-terminating glycans only at neutral pH, emphasising the significance of pH for this bacterium (Supplementary Figs. 3B, 4 and Fig. 4). Mucin O-glycans (#1, #2, #3 and #5), Gal-terminating poly-N-acetyllactosamine (polyLacNAc) glycans (#27-29, #30-39), including glycans with internal blood group epitopes, and non-sulphated chondroitin oligosaccharides (#88-90) were specifically bound by F. nucleatum 23627 (Supplementary Fig. 3). This result was validated using on-array inhibition assays which showed a minimum 60% decrease in binding by F. nucleatum 23627 to all Gal- and N-acetylgalactosamine (GalNAc)-terminating glycans in the presence of free Gal, but not free glucose (Glc) (Fig. 4A). Strong binding by F. nucleatum 23726 to α(2-6) sialic acid-terminating glycans (#5, #13-15, #24 and 26) and several linear and branched glycans containing the blood group H type-1 epitope (#54, #79-82) was also observed. Interestingly, binding to the branched α(2-6) sialyl core-1 O-glycan (#5), the α(2-6) sialylated glycan LSTb (#24) and the blood group terminating glycan (#78), which also contain a Gal-terminating antenna, was inhibited by free Gal, whereas binding to linear α(2-6) sialic acid-capped polyLacNAc (#26) or sialyl N-glycans (#13-15), was not. Binding of F. nucleatum to epithelial cells has been shown to be mediated by the autotransporter and Gal-inhibitable adhesin, Fap227. Deletion of fap2 abrogated F. nucleatum 23726 binding to Gal-terminating glycans, blood group terminating glycans and O-glycans. However, significant binding remained to sialic acid-terminating glycans from fetuin (Fig. 4B).
Heatmap showing the relative binding intensities of live fluorescently labelled bacterial cultures of L. crispatus, L. iners, G. vaginalis, S. agalactiae, and fixed E. coli and F. nucleatum strains to 90 neutral and sialic acid terminating glycans at pH 4 and pH 7 on Microarray set 2 with glycans printed at high density (300 μM). Colour code: Dark blue (0–1%); Light blue (1–10%); Yellow (10–30%); Orange (30–70%); Red (70–100%). 100%, the maximum binding score observed across pH 4 and pH 7 for a given strain. Glycan probes are grouped according to their sequences into categories and annotated by the coloured panels defined at the bottom of the figure. Histograms of representative strains are shown in Supplementary Fig. 3. Grey cells correspond to glycan probe #65 that did not pass quality control in this set.
(Left panels) Heatmaps of relative binding strengths of fixed fluorescently labelled F. nucleatum 23726 to neutral and sialylated glycans at pH7 in the presence of 10 mM galactose, 10 mM glucose or buffer only. Source data are provided in Source Data File 5. (Right panels) Heatmaps of relative binding strengths of fixed fluorescently labelled wt, Δgalkt and Δgalkt Δfap2 F. nucleatum 23726 to neutral and sialylated glycans at pH7. Values on heatmap indicate the mean fluorescence intensities of quadruplicate spots on Microarray set 2a and 2b. Colour code for binding intensities: Dark blue (0–10%); Light blue (10–25%); Yellow (25–50%); Orange (50–70%); Red (70–100%). 100%, the maximum binding score observed across the three conditions tested. Glycan structures corresponding to selected bound glycan probes are shown. Glycans categories are coloured as follows: Yellow: Mucin core related; Pink: Sialylated; Orange: LacNAc terminating; Blue: Blood group A, B, H glycans; Peach: Lewis blood group related; Purple: Fucosylated branched glycans from human milk; Grey: Non-sulphated chondroitin sulphate. * (-104) Grey cells were flagged for artefact/background on slide.
Vaginal bacteria bind blood group A, B, H and Lewis active glycans as well as sialylated glycans
Previous studies have reported high levels of fucosylation and sialylation in the N-and O-glycans of CVF from pregnant and non-pregnant women28,33. Moreover, expression of specific blood group epitopes including A, B, H(O), Lewis a, Lewis x and Lewis y have been reported in CVF33. We observed binding by L. crispatus, L. iners, S. agalactiae, G. vaginalis and E. coli strains to blood group-terminating glycans with A, B, H and Lewis epitopes on linear type 1, type 2 or type 4 polyLacNAc chains or branched human milk oligosaccharides (HMOs) (Fig. 3 and Supplementary Fig. 3), but not to the Gal-terminating type 1 or type 2 backbones (#27-34). Binding to blood group glycans A, B, H, Lewis x and Lewis b (#42, #45, #47, #49, #55, #60, #71, #82) was consistent across all L. crispatus strains tested at both pH 4 and pH 7, whereas for L. iners, G. vaginalis and S. agalactiae, binding was both pH and strain dependent (Fig. 3), except for binding to the HMO MFLNH with a terminal blood group H epitope (#82) by G. vaginalis, which was observed for all strains tested. No binding was detected by L. iners 13335 at neutral pH (Supplementary Fig. 3), again highlighting species-dependent pH mediation of glycan binding events. At strain level, differences were also observed between clinically derived E. coli strains (789, 900, and 901) compared to the ATCC C600 strain. S. agalactiae did not bind to the linear LNFP-I probe #42 but did bind strongly to the bi-antennary HMO glycans (#82 and 83) containing the LNFP-I sequence (Fig. 3)34. All bacterial strains exhibited strong binding to α(2-3) or α(2-6) sialic acid-terminating polyLacNAc glycans apart from E. coli C600 (Fig. 3). Weak, yet consistent, binding to sialylated N-glycans containing a mixture of α(2-3) and α(2-6) sialyl linkages was observed for most L. crispatus strains tested and some strains of L. iners, G. vaginalis and S. agalactiae (Fig. 3 and Supplementary Fig. 3).
Vaginal bacteria show pH dependent binding to CS and heparin
We next investigated the capacity of vaginal bacteria to bind GAGs, which have been shown to be abundantly expressed on the cell surface and in the extracellular matrix of endometrial and cervico-vaginal epithelia35,36,37. To test bacterial binding to GAGs, Microarray set 4 containing 36 GAG oligosaccharides representing different chain lengths of chondroitin sulphate A (CSA), CSB (dermatan sulphate), CSC, keratan sulphate (KS), heparin (a highly sulfated form of HS), and hyaluronic acid (HA) was constructed (Supplementary Data 1). At pH 4, most strains bound CS, heparin and KS oligosaccharides, but comparatively weak or no binding was observed to the non-sulfated KS probes (Fig. 5A). Of the E. coli strains tested, only C600 and 901 demonstrated weak binding to oligosaccharides derived from CSs, heparin and KS. Binding to CSs and heparin was pH dependent for most bacterial strains tested. At pH 7, only F. nucleatum and some Lactobacillus strains bound strongly to CSA or heparin oligosaccharides, respectively (Fig. 5A). The pH-dependent binding of bacteria to CSs and heparin oligosaccharides was observed in both HEPES and acetate buffers (Supplementary Fig. 5A).
A Heatmap showing the relative binding intensities to 36 GAG oligosaccharides at pH4 and pH7 of fixed fluorescently labelled strains of E. coli and F. nucleatum, and live L. crispatus, L. iners, G. vaginalis and S. agalactiae. The average rank of the mean fluorescence intensities from at least two independent experiments on Microarray set 5 was calculated (Source Data file 6) and the resulting values coloured as follows: Dark blue (0–10 %); Light blue (10–25 %); Yellow (25–50 %); Orange (50–70%); Red (70–100%). 100%, the maximum binding score observed across pH4 and pH7 for a given strain. B Glycan microarray images of fluorescently labelled E. coli C600, F. nucleatum 23726, L. crispatus 1398, L. iners 13335, G. vaginalis 775 and S. agalactiae 776 binding to GAG polysaccharides at pH4 or pH7. Images are representative of three independent experiments. C Bar charts showing the mean fluorescence intensities of bacterial binding at pH4 or pH7 to GAG polysaccharides on Microarray set 6 printed at 0.5 mg/ml. Error bars represent SD of quadruplicate spots for each polysaccharide on the microarray (n = 1). Assay is representative of three independent experiments. Source data are provided in Source Data File 7.
Binding to HA was restricted to a few strains of S. agalactiae and G. vaginalis. HA binding by S. agalactiae has been previously reported38,39,40. We identified in the urogenital derived strains of S. agalactiae 49, 767 and 776 a putative hyaluronidase with an appended carbohydrate binding module (CBM), that shares significant homology to the CBM70 ___domain of the S. pneumoniae hyaluronidase that targets HA41,42. Consistent with microarray binding data of whole bacteria, Isothermal Titration Calorimetry (ITC) demonstrated that a 14-mer HA oligosaccharide is a substrate for the recombinant full-length S. agalactiae 767 hyaluronidase (Sa Hyal_767) (Supplementary Fig. 6A, B). No enzymatic reaction was observed with heparin 14-mer or the negative control LNnT tetrasaccharide (Supplementary Fig. 6C).
The interactions of CS and heparin oligosaccharides with L. crispatus 1398, L. iners 13335, G. vaginalis 775, S. agalactiae 776, E. coli C600 and F. nucleatum 23726 were further investigated with biotinylated CSA, CSB, CSC and heparin polysaccharides immobilised on NHS functionalized glass slides on Microarray set 5 (Supplementary Fig. 7A). Again, strong pH-dependent binding to all GAGs was observed for L. crispatus 1398, G. vaginalis 775, S. agalactiae 776 and F. nucleatum 23726; whereas L. iners only showed weak binding signal in HEPES buffer (Fig. 5B, C and Supplementary Fig. 5B). F. nucleatum 23726 gave strong binding to CSA, which is enriched in endometrial tissue and placenta43,44, and heparin polysaccharides at pH 7 (Supplementary Fig. 7B–D). E. coli C600 did not bind GAG polysaccharides in any of the conditions tested.
Binding of L. crispatus, L. iners, S. agalactiae and F. nucleatum to vaginal epithelial cells is partially mediated by CS
Using flow cytometry, we observed that cultured vaginal epithelial cells (VK2) demonstrate high surface expression levels of CSs and heparan sulphate (HS) but not HA or KS (Supplementary Fig. 8A). Treatment of VK2 cells with chondroitinase ABC resulted in a significant decrease in anti-CS56 binding (Supplementary Fig. 8B). Binding of vaginal epithelial cells to fluorescently labelled L. crispatus 1398, L. iners 13336, G. vaginalis 775, S. agalactiae 776 and F. nucleatum 23726 was observed at pH 4 and pH 7, however chondroitinase ABC treatment significantly reduced bacterial binding at pH 4 but not at pH 7 (Fig. 6A and Supplementary Fig. 8C). Consistent with this, pre-incubation of L. crispatus with CSC polysaccharide significantly reduced bacterial binding to VK2 cells at pH 4 (Fig. 6B). Treatment with heparinase to remove HS did not affect L. crispatus 1398 or G. vaginalis 775 binding (Supplementary Fig. 8D). As several studies have indicated that L. crispatus may inhibit S. agalactiae colonisation in the vaginal niche45,46,47,48, we next performed on array competition assays, where binding of S. agalactiae 776 to CSC 14-mer was assessed in the presence of increasing amounts of L. crispatus 1398. Binding of S. agalactiae 776 to CSC 14-mer decreased in a dose-dependent manner with increasing concentrations of L. crispatus (Fig. 6C), indicating that L. crispatus can compete with S. agalactiae for CSC binding.
A Flow cytometry-based analysis of live bacterial binding to VK2 cells treated or not with chondroitinase ABC. Bar charts show the mean fluorescence intensities of the bound bacteria at pH4 and pH7. Error bars represent SD of three independent experiments for each condition (n = 3). B Flow cytometry-based analysis of the binding of L. crispatus preincubated or not with CSC polysaccharide to VK2 cells. Bar charts show the mean fluorescence intensities of the bound bacteria at pH4 and pH7. Error bars represent SD of three independent experiments for each condition (n = 3). C Left: Bar charts show the mean fluorescence intensities of the binding at pH4 of fluorescently labelled S. agalactiae 776 and L. crispatus 1398 to a CSC 14-mer oligosaccharide in the presence of increasing amounts of L. crispatus 1398. Error bars represent SD of triplicates on the microarray for each glycan (n = 1). One experiment representative of two independent experiments is shown. Right: Representative images of S. agalactiae 776 cells binding to CSC 14-mer on the microarray in the presence of increasing amounts of L. crispatus 1398. *** p < 0.0001; *p < 0.039; ns: non-significant; 2-way ANOVA (A) and **p < 0.0067; ns: non-significant B). Source data are provided in the Source Data File 7.
Discussion
Using customised glycan microarrays, we unveiled a diverse range of glycan structures recognised by vaginal bacteria implicated in reproductive health and disease states. These include GAGs that decorate plasma membrane and extracellular matrix glycoproteins, neutral and sialylated glycans carried in glycolipids at the plasma membrane, and N- and O-glycans presented on membrane-bound and secreted glycoproteins like mucins, the major protein constituents of CVF. We identified glycans specifically recognised by the opportunistic pathogens E. coli, F. nucleatum and S. agalactiae, that are not bound by commensals, whereas other glycan binding activities were shared by both beneficial and potentially pathogenic bacteria. Further, we demonstrated that adhesion of bacteria to vaginal epithelial cells is partially glycan mediated.
The utility of our glycan microarray approach was first tested using four E. coli strains. E. coli strain 900 demonstrated reduced oligomannose glycan binding compared to the other strains tested. Oligomannose binding by E. coli is known to be mediated by the adhesin, FimH1, which is located at the tip of the fimbria30. Consistent with our microarray results, electron microscopy showed that E. coli strain 900 had comparably shorter and fewer fimbria, despite being cultured in similar conditions. The abnormally short and reduced number of cellular appendices in E. coli 900 is likely to be abrogating oligomannose binding in this strain. These results thus highlight the capacity of glycan microarray studies to effectively discern binding characteristics attributable to different bacterial strain morpho-phenotypes.
Glycan microarray technologies, since first introduced in early 200049,50, have revolutionized the molecular dissection of glycan-protein interactions in a miniaturised and high-throughput fashion. Significant advances have been made in applying glycan microarrays to analyze the glycan-binding specificities of microbial glycan-binding proteins, including adhesins (lectins) from viruses, fungi, and bacteria, bacterial toxins and CBMs, as well as whole virus particles. However, the application of this technology to whole bacterial cell binding has remained relatively limited. A few studies have utilised glycan array analysis to investigate host glycan binding by whole bacterial cells of pathogenic bacteria, including Campylobacter jejuni51, Group A Streptococcus52 and Neisseria meningitidis53,54. These studies have provided valuable insights into bacteria-glycan interactions, but also present several limitations. For example, many of these studies examine only a limited selection of bacterial species and strains, and the glycan libraries used are often narrow in scope, being restricted to commercial glycans or exclusively sialylated glycans. The conclusions drawn from these studies are often limited to monosaccharide specificity and less quantitative. To our knowledge, this is the first time that GAG oligosaccharides of different types and chain lengths have been interrogated for bacterial cell binding side by side with neutral and sialylated glycans of diverse types.
During assay condition optimisation, fucose-dependent blood group binding was detectable only at a high glycan printing concentration (300 µM), likely due to increased glycan density. In contrast, strong binding of anti-carbohydrate antibodies was observed even at lower concentrations (50–100 µM) on the same array (Supplementary Fig. 9). This suggests an overall weak binding avidity of bacterial cells to the blood group glycans presented on the array. Thus, for screening our vaginal microbiota collection, a printing concentration of 300 µM was used for Microarray set 2. With the exception of F. nucleatum, we observe that most strains bound to neutral glycans derivatised with the AEAB (N-(aminoethyl)-2-aminobenzamide) linker (Fig. 3), suggesting that presentation and the chemical environment around the glycan, are likely to be important for whole bacterial binding to glycans55,56. Again, this effect was not observed in antibody and plant lectin binding (Source Data File 2).
Bacteria express multiple glycan-binding proteins (GBPs), including lectins and CBMs, which can simultaneously mediate interactions with host glycans, and the subcellular localisation of these GBPs varies (e.g., secreted vs membrane-bound). This diversity likely explains our observation of bacteria binding to multiple glycan types with varying signal intensities. Our recent comparative genomics study revealed a wider repertoire of GBPs in vaginal pathogens and pathobionts compared to commensal species42. Except for a few glycans, the binding profiles of commensals L. crispatus and L. iners and potentially pathogenic bacteria like G. vaginalis and S. agalactiae, observed here were overlapping. This might reflect competition for similar glycans as described for GAG binding, but we cannot rule out that modification and utilisation of glycans, which involves a wide repertoire of additional enzymes involved in transport and processing of glycans inside the bacteria57,58, might also differ from commensals to potentially pathogenic bacteria. Further glycan-binding activities may also exist that were not captured by the approaches used in this study.
Our GAG array results highlighted strong binding of L. crispatus, L. iners, G. vaginalis, F. nucleatum and S. agalactiae to the sulphated GAGs, CSs and Heparin, which was largely pH dependent. Binding of E. coli to the same GAGs was comparatively weak. S. agalactiae and some strains of G. vaginalis also demonstrated strong binding to the non-sulphated GAG, HA. These findings likely reflect the differing capacity of vaginal bacteria to adapt to dynamic changes in pH and adhere to different glycan structures encountered within the reproductive tract35,37. For example, during pregnancy, HA expression is highly increased in the cervix, whereas under-sulphated CSA is enriched in the neutral or slightly basic environment of the placenta11,43,59. The GAG binding characteristics of F. nucleatum and S. agalactiae described here are consistent with their capacity to colonise both the lower and upper reproductive tract in pregnancy, where pH is higher than in the vagina, which could potentially lead to infection and increased risk of preterm birth and neonatal sepsis60,61,62.
Several studies have linked F. nucleatum colonisation of the cervicovaginal niche during pregnancy with increased risk of preterm birth and subsequent risk of early onset neonatal sepsis63,64. Amniotic fluid infection has also been previously linked to Fusobacteria colonisation of the vagina65. Our data provides evidence of direct and specific recognition by F. nucleatum of glycans terminating in Gal and GalNAc, such as the mucin core-1 O-glycan (Galb1-3GalNAc), non-sulphated chondroitin and several polyLacNAc glycans. Binding to Gal, GalNAc and blood group glycans, but not sialylated glycans was abolished in the fap2 deletion mutant. These results are consistent with previous studies using indirect methods like galactose or lectin competition/ inhibition of bacterial binding to cell lines and tissues, including placental epithelial cells, that reported F. nucleatum binds to Gal-terminating glycans (including the mucin core 1 O-glycan) through the autotransporter Fap227. Gal-terminating LacNAc chains targeted by Fusobacterium are present in O-glycans from CVF66, which may provide a potential explanation for Fusobacterium adhesion and colonisation of the vaginal niche65. Binding to sialic acid terminating N-glycans, also present in CVF, was observed in a fap2 deletion mutant suggesting that other adhesins mediate sialic acid binding by F. nucleatum67,68.
In the lower reproductive tract, a decrease in sialylation of poly-fucosylated N-glycans in cervicovaginal fluid is associated with the presence of high diversity CST-IV community state type33, which includes known sialic acid foragers like Gardnerella18,69. Here, using glycan arrays, we show that vaginal species including L. crispatus, L. iners, G. vaginalis and S. agalactiae have the capacity to bind several blood groups and sialic acid-terminating glycans (N- and O-glycans) present in CVF33. Notably, binding to these non-GAG glycans was weaker compared to the binding to GAG sequences. Detectable signals required higher glycan probe concentrations, suggesting a need for increased surface density. Interestingly, except for L. crispatus, condition and strain-specific variations in neutral and sialic acid glycan binding were observed. Subspecies and strain to strain variability of G. vaginalis have been shown to differ in terms of virulence factors expressed, biofilm formation capacity and/or variability on the mucosal immune responses elicited70,71,72. Such variations could underlie differing functional abilities, including glycan binding as observed in this study, with exposure to different lectins, capsule or cell wall components. Consistent with these findings, specific sub-species or strains of G. vaginalis have recently been implicated in increased risk of preterm birth5,73,74. Despite this, consistent binding to certain glycans like H type 1 containing human milk oligosaccharides was observed.
Phenotypic differences in host blood group glycan expression have been correlated to host disease susceptibility by human pathogens such as enterotoxigenic E. coli or Helicobacter pylori, which express adhesins targeting a specific blood group epitope75. In addition, blood group B individuals are more susceptible to E. coli and Group B Streptococcus infections76,77. Histo blood group antigen (HBGA) expression can be used in patient risk stratification78,79. In the gut, blood group glycans are used by commensal bacteria as carbon sources thus shaping microbiota composition, which has implications for chronic inflammatory bowel disease57,80,81. Probiotic lactic acid-producing bacteria like Levilactobacillus brevis and L. gasseri have also been reported to bind to blood group antigens that might help with colonisation82. Blood group glycan expression in human epithelial cells and secretions such as saliva, breast milk or mucus depends on the secretor gene (FUT2). The FUT2 α-1,2-fucosyltransferase enables synthesis of the H(O) antigen, which can be elongated by either a GalNAc-transferase or a galactosyl-transferase to produce A and B antigens, respectively. Our findings that vaginal pathogens and commensals bind to blood group glycans will inform follow up studies on blood group glycan binding and utilisation by vaginal microbiota to provide mechanistic insight into recent reports of increased preterm birth risk in non-secretor individuals harbouring pathogen enriched vaginal microbiomes83. Further studies investigating the relationship of microbiota composition, HBGA expression and blood group antigen binding and utilisation by bacteria in pathogen-driven disease, particularly in the context of reproductive tract infections, are also warranted.
Our study also demonstrated that L. crispatus competes for binding to specific glycan structures targeted by vaginal pathogens, such as S. agalactiae, in a pH dependent manner. This may provide a potential explanation for reduced S. agalactiae colonisation during pregnancy in women dominated by L. crispatus45,84. This has important clinical implications as S. agalactiae exposure from the maternal genital tract during delivery is the primary risk factor for early-onset GBS infection, which remains one of the most common infectious causes of mortality in neonates85,86. L. crispatus competition for glycan-binding sites in the cervicovaginal niche may also contribute to its capacity to dominate the niche2,87. Regulation of pH-dependent competition for glycan binding sites in the vaginal niche offers a strategy for the development of prebiotics designed to encourage vaginal colonisation of L. crispatus, to prevent pathogen colonisation and thus reducing the incidence of BV, preterm birth and associated neonatal complications88.
In summary, this is the first study using glycan microarrays to test whole vaginal bacteria-glycan binding events with a structurally diverse collection of glycans. Our study describes a methodological approach by which customised glycan arrays can be used to identify complex bacterial binding events of biological and clinical significance. The identification of sequence-defined glycan structures specifically bound by vaginal pathogens and commensals represents an important step towards the design and development of strategies aimed to prevent or modulate bacterial binding in the vaginal niche. Although the bacteria examined in our study are associated with adverse pregnancy outcomes, these species are also more broadly implicated in other areas of reproductive tract health and disease, including bacterial vaginosis, risk of STI acquisition and HPV infection and progression of cervical cancer89,90,91. Glycomimetics, synthetic compounds mimicking native glycans, could be introduced to the cervicovaginal mucosa to block or compete for pathogen binding sites or alternatively, to encourage and support adhesion and engraftment of live biotherapeutics (probiotics). In addition, the sequence-defined glycan binding profiles of various bacteria provide valuable information for further detailed studies, such as interaction studies at the protein level or in vivo experiments. While our current study focused on the vaginal niche, the methods described can be readily adapted for investigations of colonisation and infection in other sites, including the gut, skin and respiratory tract, where collective evaluation of bacterial recognition of glycan binding can inform microbiota ecology dynamics and modulating studies.
Methods
Glycan materials: amino-terminating glycan probes
AEAB-terminating92 high-mannose N-glycan probes #7-#12 were purchased from NatGlycan (Atlanta, Georgia, USA). Aminopropyl-terminating probes #6, #91-110, #170, #174-#177, and #187, synthesised chemically, were purchased from GlycoNZ (Auckland, New Zealand). The following amino-terminating glycans were obtained as gifts from different resources or collaborators: Chemically synthesised aminopropyl (Sp)-terminating probes #155 and #156 were from the Core D of the US Consortium for Functional Glycomics; probes #165 and #169 were from Prof. Nicolai Bovin (Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences), probe #18293 from Prof Chunxia Li (Ocean University of China, Qingdao, China), and GalNAc β-linked to Ser (#163) and Thr (#164)93 were from Ulrika Westerlind (Umeå University). The glycine (Gly)-terminating glycan probes #79-#81, #84, #151-#154, #157-159, #160-162, and #183-186 are N-glycosides and were prepared by conjugation of the respective reducing sugars with glycine as described94, and purified further by HPLC. All amino-terminating glycans were analyzed by electrospray (ESI) or matrix-assisted laser desorption/ionisation (MALDI) mass spectrometry (MS) (Source Data File 1) with hexose-containing glycans quantified by micro-scale hexose assay (see below) before microarray printing.
Preparation and analysis of amino-terminating glycan probes
Deprotection of Fmoc-protected and per-acetylated serine- and threonine-linked mucin cores
Probes #1-#3 and #166-#168 were prepared by deprotection of fluorenylmethoxycarbonyl (Fmoc)- and per-acetyl Ser- and Thr-linked mucin cores purchased from Sussex Research Laboratories (Ottawa, Canada). In brief, 1 N NaOCH3 in methanol ( ~ pH 10-11) was added to the solution of the protected sample (1 mg in 0.5 ml methanol). The solution was stirred for 3 h at room temperature and the pH was adjusted to pH 4-5 using acetic acid. The sample solution was dried under a N2 stream and re-dissolved in 0.05 M ammonium acetate before purification by gel filtration on a Superdex Peptide column (Pharmacia, Uppsala, Sweden) with elution by 0.05 M ammonium acetate. The collected fraction was freeze-dried, and ammonium acetate removed by repeated co-evaporation with water. The deprotected samples were analyzed by MS and quantified by micro-scale hexose assay (see below) before use.
Extended and Ser-linked mucin core 3 O-glycans
Enzymatically prepared Ser-terminating probes #178-#180 were from Prof Kelly Moremen (Complex Carbohydrate Research Center, University of Georgia) and were purified by gel filtration chromatography using Superdex Peptide column before use.
Ser-terminating probe #181 was chemically synthesised and as a gift from Prof Dr Akihiro Imamura of Gifu University (Yanagido, Japan). These Ser-linked and extended mucin core 3 O-glycans were analyzed by ESI-MS (Fig. 1 in Source Data File 1) and quantified by hexose assay before microarray printing.
Preparation of amino-terminating probes from azido-terminating glycans by click chemistry
Probes #19-#22, #27, #28, #30-39, #41, #43, #44, #48, #50, #51, #54, #56, #57, #61-#63, #65-#67, #69, #70, #72-#77 were prepared from chemoenzymatically synthesised 3-azidopropyl glycans95 after conversion into amino-terminating glycans. Conversion of the azido- into amino-terminating glycans was performed as described previously96 with the following modifications. To the freeze-dried mixture of azido-glycans (50 nmol) and triethylene glycol 2-aminoethyl propargyl ether (50 nmol, Sigma-Aldrich, Gillingham, England) was added pre-mixed 200 ml solution of t-butyl alcohol/water (2:1, v/v) containing CuSO4 (30 nmol), sodium ascorbate (30 nmol) and tris-benzyltriazolylmethyl amine (TBTA, 60 nmol). The mixture was incubated at room temperature for 1 h with occasional vortexing. The solvent was removed under reduced pressure and the residue purified by gel filtration on a Superdex Peptide column as described above. Further purification was carried out by HPLC on a XBridge amide column (Waters, Wilmslow, England) with a binary solvent system of A: acetonitrile/H2O 70:30 and B: acetonitrile /H2O 20:80, with both containing 10 mM KH2PO4 (pH 3.0). The elution was carried out by gradient 5-25%B in 45 min at a flow rate of 1 ml/min and detection by UV 206 nm. The collected fractions were desalted by gel filtration on a Superdex Peptide column as described above. The converted propyl triazole-PEG4-amine (PTPA)-terminating probes were purified by HPLC (Fig. 2 in Source Data File 1) for representative HPLC profiles) and analyzed by MALDI-MS (Table 1 in Source Data File 1 for representative mass spectra). Quantitation was by micro-scale hexose assay (see below) before microarray printing.
Preparation of AEAB-terminating probes from reducing glycans by reductive-amination
Neutral glycan (#29, #40, #42, #45-#47, #49, #52, #53, #55, #58-#60, #64, #68, #71 and #82) and sialylated probes (#16, #17, #23-#25), purchased respectively from Elicityl (Crolles, France) or Dextra Laboratories (Reading, England), were prepared from reducing sugars whereas sulfated KS oligosaccharide probes #126-#135, were prepared from chemically synthesised reducing sugars gifted by Tokyo Chemical Industry (Tokyo, Japan). The sialylated probes #18 and #26, were prepared from the enzymatic synthesised reducing sugars (Dr Apoorva Srivastava, Department of Chemistry, Imperial College London). Preparation of the fluorescent reagent AEAB and conjugation with reducing sugars by reductive-amination were performed as previously described92. For purification of the conjugation product, a Sep-Pak Aminopropyl cartridge (Waters, Wilmslow, England) was used with elution by acetonitrile /H2O for the neutral sugars and acetonitrile /0.05 M ammonium acetate for the sialylated sugars. Further purification was carried out by HPLC using Amide column (XBridge, Waters). The gradient elution was acetonitrile /H2O for the neutral and acetonitrile /0.05 M ammonium acetate for the sialylated sugars at a flow rate of 1 ml/min. with detection by UV 330 nm. The detection and probe quantitation were performed using by UV 330 nm as described92.
Glycosaminoglycan oligosaccharide-AEAB probes
Oligosaccharide fractions from heparin97, chondroitin sulphate A, B and C98, and keratan sulphate98 were prepared by partial digestion with heparin lyase I (Sigma), chondroitinase ABC (Sigma), and keratanase II, respectively, as previously described. Hyaluronic acid (HA) fragments were prepared in two different ways. HA lyase (EC 4.2.2.1; from Streptomyces hyalurolyticus; Sigma) was used to obtain unsaturated ΔUA-terminating oligosaccharides and HA hydrolase (EC 3.2.1.35; from bovine testes; Sigma) was used to prepare the unmodified HA oligosaccharides99. After conversion into AEAB probes, the conjugation products of KS DP10 and 14, (#136-#138), HA DP4, 8, 11-14 (#139-#146), and CS DP6, 10 and 14 (#111-#122) and non-sulphated chondroitin DP10, 12 and 14 (#88-#90), and the conjugation products of heparin DP6, 10 and 14 (#123-#125) were purified using an amino-cartridge (Sep-Pak Aminopropyl, Waters) with elution by sodium acetate (0.05 M). Additional purification was carried out by a short column of strong anion exchange (HiTrap Q, GE Healthcare, Amersham, England) with elution by a gradient of NaCl. The collected fractions were desalted on a Superdex Peptide column with elution by ammonium acetate (0.05 M).
Serine- and threonine-linked sialyl mucin core 1 probes
Mono- and disialyl core 1 probes #4, #5, #171-#173 with either Ser or Thr were isolated from human urine as described100. Further purification was carried out by HPLC with XBridge amide column (Waters, Wilmslow, England) with a binary solvent system of A: acetonitrile/H2O 70:30 and B: acetonitrile/H2O 20:80 (both containing 7.5 mM KH2PO4). The elution was carried out by a gradient 15-20%B in 40 min at a flow rate of 1 ml/min and detection by UV 206 nm.
Human milk oligosaccharide AEAB probes
AEAB probes #78 and #83-#87 were prepared from DP6-8 and DP10 fractions of HMOs as described101. Amide and PGC columns were used for separation and purification. Their sequences were determined by negative-ion ESI-CID-MS/MS98.
Complex-type N-glycan AEAB probes from fetuin N-glycans
Fully sialylated bi- and tri-antennary N-glycan AEAB probes #13-#15, were prepared from fetuin N-glycans released by PNGase F (New England Biolabs, Ipswich, MA) and isolated from a short column of Bio-Gel P6 as described102. The total released N-glycans were conjugated to AEAB and the access AEAB reagent was removed by NH2-cartridge. The N-glycan-AEAB fraction was fractionated by HPLC using a Hypersil APS-2 column (Fisher Scientific, Loughborough, England) with a gradient by acetonitrile/H2O containing ammonium acetate. Detection was by UV at 330 nm and the flow rate was at 1 ml/min.
Micro-scale hexose quantitation
The micro-scale dot orcinol assay on TLC plates was performed as previously described103. In brief, hexose-containing probes, especially those not AEAB conjugated, were dissolved in H2O (at approx. 0.5-1.0 mg/ml), and 1 μl was spotted onto a silica gel TLC plate. Standard solutions (Galactose in H2O) were also applied alongside the glycan probes. The plate was dried and sprayed with orcinol/H2SO4 staining reagent and heated in an oven at 105 oC to develop the colour. Quantitation was carried out by scanning at 550 nm on a CAMAG TLC scanner (Omicron Research, Hungerford, England).
MS analysis of amino-terminating glycan probes
MS analysis of most amino-terminating probes was carried out by an electrospray mass spectrometer on a Waters (Manchester, UK) Q-TOF-type mass spectrometer SYNAPT-G2 in either positive- or negative ion mode. Cone voltage was at 50 eV or 80 eV and capillary voltage at 3 kV. Source temperature was 80 oC and the desolvation temperature 150 oC. A scan rate of 1.5 sec/scan was used and the acquired spectra were summed for presentation. For analysis, glycan probes were dissolved in H2O typically at a concentration of 10–20 pmol/μl, of which 1 μl was injected via an HPLC injector. Solvent acetonitrile/2 mM ammonium bicarbonate 1:1 was delivered by a HPLC pump (Waters) at a flow rate of 10 μl/min. Selected probes were also analyzed by MALDI-MS on an Axima MALDI Resonance mass spectrometer with a QIT-TOF configuration (Shimadzu). A nitrogen laser was used to irradiate samples at 337 nm, with an average of 100-200 shots accumulated. A matrix solution (0.5 μl) of 2,5-dihydroxybenzoic acid (20 mg/ml) in a mixture of methanol/water (1:1) was deposited on the sample target before application of the sample solution (0.5 μl).
Biotinylated GAG polysaccharide probes
Biotinylated CSA, CSB, CSC and heparin (#147-#150, respectively) were prepared and purified essentially as described104. The level of biotin introduced was one for every 50 carboxyl groups105. In brief, 50 mg of GAG polysaccharide were dissolved in 0.1 M MES buffer (pH 5.5) before addition of 109 ml of 25 mM biotin LC-hydrazide (Pierce, Tattenhall, UK) solution in Me2SO and 130 ml of 25 mg/ml 1-ethyl-2-(2-dimethylamino propyl) carbodiimide solution in 0.1 M MES buffer. The reaction mixtures were stirred at room temperature for 24 h and dialysed extensively against deionized water. A short Sephadex G-10 column (1.6 × 30 cm) was used to remove remaining biotin and other reagents. Quantitation of GAG polysaccharides and their modified forms was carried out by carbazole assay106.
Bacterial culture and labelling
Bacterial isolates used in this study were commercially sourced (ATCC and DSMZ) or were kindly gifted by the Lawley and Wolfe lab (Supplementary Data 4). Isolates from the Wolfe lab were originally obtained from non-pregnant women who provided verbal and written consent following Loyola University Medical Centre Institutional Review Board approval as detailed32. The identity of the bacterial species used was confirmed using sanger sequencing targeting the 16S rRNA gene and MALDI_TOF.
Bacterial Culture
E. coli isolates were streaked on LB agar (Sigma, Lennox recipe) and incubated overnight at 37 °C. LB broth was inoculated from cultures on plates and incubated overnight in a shaking incubator at 37 °C and 200 rpm. Anaerobic bacteria were cultured in the following media: L. crispatus, MRS agar and broth media; L. iners and G. vaginalis, Columbia (Difco) + 5% v/v defibrinated horse blood (Oxoid) agar and Vaginal Microbe Media (VMM, Patent protected); S. agalactiae, BHI (Oxoid) agar and broth media; and F. nucleatum, Columbia (Difco) + 5% v/v Defibrinated horse blood (Oxoid) agar and BHI (Oxoid) broth supplemented with 5 g/l yeast extract (Sigma Aldrich), 0.005 g/l L-cysteine hydrochloride monohydrate (Merck), 0.01 g/l D-(+)-cellobiose (Sigma Aldrich) and 0.01 g/l D-(+)-maltose (0.01 g/l) media. Agar (1.2% w/v) was added to the media to prepare agar plates. Isolates of anaerobes were streaked on agar plates and incubated overnight at 37 °C in an anaerobic chamber (10% CO2, 10% hydrogen, 80% nitrogen and 70% humidity). A total of 10 ml of degassed broth was inoculated and incubated overnight in anaerobic conditions at 37 °C. The starter broth was used to inoculate the required volume of fresh media and was incubated as described above until the broth reached an OD600 ≈ 1.0.
Fluorescent labelling of cultured live bacteria
For staining of live bacteria, 2 ml aliquots of culture were centrifuged for 10 minutes, and the resulting pellet washed twice with 1 ml of cold sterile Hanks Balanced Saline solution (HBSS) (Gibco). 1 × 108 bacterial cells were stained with 10 μM CellTrace™ CFSE (Molecular Probes) or 10 μM CellTrace™ Far Red (Molecular Probes) in 250 uL of PBS and incubated for 1 h in a rotating wheel at room temperature protected from light. Bacteria were centrifuged, the pellet was washed twice with cold sterile HBSS and resuspended in 1 ml of cold sterile HBSS. Stained bacteria were kept on ice or at 4 °C and protected from light until further use (within 2 h for live bacteria).
Fluorescent labelling of fixed bacterial cultures
E. coli and F. nucleatum cultures were centrifuged for 10 minutes to pellet bacteria and cells were washed twice in cold sterile PBS. Bacteria were fixed with 4% formaldehyde in PBS in a rotating wheel at room temperature for 15 min. E. coli or F. nucleatum fixed cells were centrifuged and washed twice in cold sterile PBS, the cell pellet was resuspended in HBSS (250 uL for 1 × 108 bacterial cells) before staining with 10 μM SYTO™59 (Invitrogen) or 10 μM SYTO™ BC Green (Invitrogen) respectively, for 30 min in a rotating wheel at room temperature protected from light. Bacteria were centrifuged, the pellet was washed twice with sterile cold PBS and resuspended in the desired volume of cold sterile PBS. Stained bacteria were kept on ice or at 4 °C and protected from light until further use (ideally no more than 1 week).
All centrifugations were done at 5000 x g for 10 min in a refrigerated centrifuge set at 4 °C, except for E. coli where 3000 x g were used to avoid fimbriae shedding. HBSS media and buffers required for staining and fixing procedures of anaerobe bacterial cultures were degassed before use. Wide-bore tips were used while handling bacterial cultures to avoid fimbriae shedding. To ensure the bacterial cultures were not contaminated during culturing, we used chromogenic agar (CHROMagar™) and light microscopy to visualise homogeneous and distinct cell morphology.
Glycan microarray preparation and analyses
The generation of covalent sequence-defined glycan arrays and the GAG polysaccharide microarray on NHS-functionalised glass slides (Nexterion H) was performed essentially as described before49,98. Five oligosaccharide microarray sets and one GAG polysaccharide array set were generated (Supplementary Fig. 1 and Supplementary Data 1). Details of the preparation of the microarrays and the methods used for binding assays and data analysis are in accordance with MIRAGE (Minimum Information Required for A Glycomics Experiment) guidelines for reporting of glycan microarray-based data107 (Supplementary Data 3).
Binding analyses with anti-carbohydrate antibodies, plant lectins and other glycan binding proteins
Details of the antibodies, lectins and glycan binding proteins used are summarised in Supplementary Data 2. Microarray analyses were performed at ambient temperature essentially as previously described108. No blocking was used for overlay assays on covalent glycan array glass slides. Pre-complexation of hSigLec15-Fc with biotinylated anti-human IgG (1:1 by weight) was performed at 4 °C for 30 min before adding it to the microarray.
Bacterial binding analyses
Microarray analyses of fluorescently labelled bacteria (see above) was performed as follows. Fluorescently labelled bacterial cultures at OD = 1 were resuspended in binding buffer (HBS: 10 mM HEPES at pH 7.0 or pH 4, 150 mM NaCl, supplemented with 5 mM CaCl2 or Acetate buffer: 10 mM sodium acetate, at pH 7.0 or pH 4, 150 mM NaCl, supplemented with 5 mM CaCl2) and were used for overlays on glycan microarrays. 100 µL of bacterial suspension was applied to the incubation chamber with the microarrays, incubated for 1 h at room temperature under mild agitation on an oscillating platform and washed three times with binding buffer, followed by two washes with HPLC grade water to remove salts from the array. Slides were dried under a mild nitrogen flow and scanned for quantitation as described below. HBS at pH4 was prepared immediately before use by adjusting the pH of HBS with HCl. Acetate buffer pH4 was prepared immediately before use by adjusting the pH with Acetic acid. The pH of the bacterial suspensions was monitored before and after overlays using pH strips.
Microarray data analysis
Images of fluorescently bound antibodies, lectins, glycan binding proteins and bacteria were acquired with a GenePix 4300 A scanner from Molecular Devices, quantified using GenePix software and processed with CarbArryART109. Bar charts and heatmaps were generated using excel. RANK quantitation is described in MIRAGE (Supplementary Data 3) and performed as previously described.
Glycan binding on-array inhibition assays
For on-array inhibition studies of F. nucleatum 23726 glycan binding, fluorescently labelled cultures of bacteria were incubated with 10 mM D-galactose, 10 mM D-glucose, or HEPES buffer (HBS: 10 mM HEPES at pH 7.5, 150 mM NaCl, and 5 mM CaCl2), for 15 min at room temperature before performing bacterial overlay on glycan microarrays.
Bacterial competition assays on array
For on-array competition studies between L. crispatus and S. agalactiae, 50 μl of 0.2 × 108, 0.4 × 108, 0.8 × 108, or 1.6 × 108 cells/ml of CellTrace™ CFSE labelled L. crispatus cultures were added to 50 μl of 0.4 × 108 cells/ml CellTrace™ Far Red fluorescently labelled cultures of S. agalactiae (see staining procedures described above and MIRAGE (Supplementary Data 3) before performing bacterial overlay on glycan microarrays.
Mammalian cell culture
VK2/E6E7 vaginal epithelial cells (ATCC- CRL-2616) were cultured and maintained in Keratinocyte SFM 1X (Gibco) containing Calcium Chloride 0.4 mM and 1% w/v penicillin– streptomycin (Sigma Aldrich). Cells were cultured at 37 °C and 5% CO2 in a humidified incubator and were free of mycoplasma, which was routinely tested for using the MycoStrip Detection Kit (InvivoGen).
Chondroitinase and heparin lyase III digestion of VK2 cells
The VK2/E6E7 cells were gently scrapped and resuspended in PBS at a concentration of 8.105 cells/ml. Chondroitinase ABC (E.C. 4.2.2.4, Sigma) and heparin lyase III (E.C. 4.2.2.8, recombinant, IBEX Technologies, Montreal, Canada) were added at a concentration of 0,005 U/ml and 0,001 U/ml respectively and the cells were incubated at 37 °C for 2 h. The cells were centrifuged 5 min at 300 x g and resuspended in PBS at the appropriate pH for antibody or bacteria binding analyses.
Assessment of VK2-glycan and bacteria binding by flow cytometry
Sub-confluent cultures of VK2/E6E7 cells were gently scrapped and resuspended in PBS. Single cell suspensions were stained with anti-chondroitin sulphate (CS56, Abcam), anti-keratan sulphate (MZ15, a present from Fiona Watt, Imperial, London), anti-heparan sulphate (clone 10E4, Amsbio) or biotinylated hyaluronic acid binding protein (Millipore), followed by Alexa Fluor 488-conjugated anti-mIgG (ThermoFisher) or Alexa Fluor 647-conjugated anti-mIgM antibodies (ThermoFisher) or streptavidin PE (Phycoerythrin) conjugate (Miltenyibiotech), in PBS for 30 min at 4 °C. Cells were analyzed using a FACSCalibur flow cytometer (Becton Dickinson). Intact VK2 cells were gated based on their FFS and SSC, excluding cells with extreme low or high FSC and SSC values, see example in Supplementary Fig. 8). Flow cytometry data was analyzed using CellQuest software (BD Biosciences).
For bacterial binding assays, VK2/E6E7 cells were added to bacteria suspensions to obtain a ratio of bacterial particles to cells or MOI (Multiplicity of Infection)=20 for all bacterial strains except for G. vaginalis at pH 4 where MOI = 100 was used. Cell mixtures were gently stirred for 1 h at 4 °C. Where appropriate, fluorescently labelled bacteria were incubated in PBS containing polysaccharides at 5 mg/ml for 30 min before binding assays. Cells were analyzed using a FACSCalibur flow cytometer (Becton Dickinson) and analyzed using the CellQuest software (BD Biosciences).
Electron microscopy
Overnight cultures of E. coli isolates were imaged by negative stain. For staining, 10 µl of culture was deposited onto glow-discharged 300 mesh carbon-coated copper grids (Agar Scientific), and left for 2 min. The grid was washed twice with 10 µl of water. The sample was stained with 0.2% (w/v) or 0.04% (w/v) phosphotungstic acid for 7 sec, after which the stain was blotted off. The grid was imaged on a Tecnai 12 Spirit transmission electron microscope (FEI) operating at 120 kV and with a 2 K eagle camera (FEI).
SaHyal_767 protein expression and purification
The gene WP_000403395.1 encoding a putative hyaluronate lyase enzyme from S. agalactiae 767 was analyzed using InterProScan. The designed construct Sa_Hyal_767 includes both the carbohydrate-binding and catalytic ___domain (residues 106-1081). Tyr578 was mutated to Phe to reduce enzymatic activity on glycan binding assays and was codon optimised for E.coli expression. The construct was synthesised and cloned between the BamHI/XhoI sites in the multiple cloning site of the expression vector pET28a(+) (GenScript) (Supplementary Fig. 10). The resulting recombinant protein has a hexa-histidine-tag at the N-terminal. Protein expression was carried out in E. coli strain BL21(DE3) induced by adding 1 mM isopropyl-β-d-thiogalactopyranoside, overnight at 19 °C. IMAC purification was performed with a imidazole gradient. Fractions containing the purified recombinant proteins were buffer exchanged to 50 mM HEPES, pH 7.5, 100 mM NaCl, 5 mM CaCl2 and 0.02% sodium azide. The final yield of purified protein was approximately 8 mg/L of culture.
Isothermal calorimetry (ITC) assays
The theoretical basis of kinetic measurements by ITC is described elsewhere110. Briefly, in a single injection ITC experiment, the total heat measured is proportional to the apparent enthalpy (ΔHapp), and the number of moles of product generated. The reaction rate can be related to the amount of heat generated over time. The affinity for the substrate (Km) and the turnover rate (kcat) values of the enzyme can be obtained by fitting the rate of the reaction versus the substrate concentration plots with a Michaelis-Menten curve. Experiments were performed on MicroCal PEAQ-ITC calorimeter (Malvern). One single 38 μl injection was performed with a speed of 0.5 μL s–1. The instrument was set to high-feedback mode, reference power 5 μcal sec-1, stirring speed of 650 rpm, and an experimental temperature of 25 °C. A 600 second pre-injection delay was applied for baseline stabilisation after equilibration. Apparent enthalpy of the reaction, heat rate (dQ/dt), and Michaelis–Menten plots were generated using MicroCal PEAQ-ITC Analysis Software (Malvern).
Sa_Hyal_767 activity against HA 14-mer, heparin 14-mer, and LNT was tested at pH 7.5. The enzyme was dialysed overnight into a buffer either containing 50 mM HEPES at pH 7.5, 150 mM NaCl, and 5 mM CaCl2. The lyophilised glycans were dissolved in the same corresponding buffer. A Sa_Hyal_767 solution at 50 nM was placed in the sample cell, and a 0.5 mM glycan solution was loaded in the injection syringe. Substrate autohydrolysis and heat of dilution phenomena were assessed by replacing the enzyme solution in the calorimetric cell with reaction buffer and carrying out the measurements with identical experimental settings.
Statistical analysis
In glycan microarray binding analyses each assay was performed with four technical replicates and results in graphs presented as mean +/− SD of quadruplicates (Source Data Files 2–7). Bacterial binding to polysaccharide GAG arrays was performed in three independent assays. Results in graphs are presented as mean +/− SD of quadruplicates on the microarray of representative experiments or as mean +/− SD of three independent experiments. Statistical significance of data from independent experiments was assessed using an unpaired two-tailed Student’s t test. Microscopy images are representative and similar results were observed in different fields of view with two independent biological replicates. Mammalian cell-based assays were performed with at least three biological replicates. Results in graphs are presented as mean +/− SD. Statistical significance was assessed using 2-way ANOVA.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
Source data and glycan microarray data are provided with this paper as Source Data Files. Original glycan microarray image files and gpr files are available upon request. Glycan microarray processed data is provided in Source Data Files 2-6. Glycan microarray metadata is summarised in Supplementary Data 3 with MIRAGE information. There are no restrictions on the data availability. Source data are provided with this paper.
References
Golob, J. et al. Microbiome Preterm Birth DREAM Challenge: Crowdsourcing machine learning approaches to advance preterm birth research. Cell Rep. Med. 5, 101350 (2024).
Huang, C. et al. Meta-analysis reveals the vaginal microbiome is a better predictor of earlier than later preterm birth. BMC Biol. 21, 199 (2023).
Gudnadottir, U. et al. The vaginal microbiome and the risk of preterm birth: a systematic review and network meta-analysis. Sci. Rep. 12, 7926 (2022).
van de Wijgert, J. & Jespers, V. The global health impact of vaginal dysbiosis. Res Microbiol 168, 859–864 (2017).
Callahan, B. et al. Replication and refinement of a vaginal microbial signature of preterm birth in two racially distinct cohorts of US women. Proc. Natl. Acad. Sci. USA 114, 9966–9971 (2017).
Fettweis, J. et al. The vaginal microbiome and preterm birth. Nat. Med. 25, 1012–1021 (2019).
Kindinger, L. M. et al. The interaction between vaginal microbiota, cervical length, and vaginal progesterone treatment for preterm birth risk. Microbiome 5, 6 (2017).
Chan, D. et al. Microbial-driven preterm labour involves crosstalk between the innate and adaptive immune response. Nat. Commun. 13, 975 (2022).
Kindinger, L. M. et al. Relationship between vaginal microbial dysbiosis, inflammation, and pregnancy outcomes in cervical cerclage. Sci. Transl. Med 8, 350ra102 (2016).
Decout, A. et al. Lactobacillus crispatus S-layer proteins modulate innate immune response and inflammation in the lower female reproductive tract. Nat. Commun. 15, 10879 (2024).
Lykke, M. et al. Vaginal, cervical and uterine ph in women with normal and abnormal vaginal microbiota. Pathog. (Basel, Switz.) 10, 90 (2021).
Ravel, J. et al. Vaginal microbiome of reproductive-age women. Proc. Natl. Acad. Sci. USA 108, 4680–4687 (2011).
Varki, A. & Gagneux, P. Biological functions of glycans. Essent. Glycobiol. 7, 3–49 (2015).
Vagios, S. & Mitchell, C. Mutual Preservation: A review of interactions between cervicovaginal mucus and microbiota. Front. Cell. Infect. Microbiol. 11, 676114 (2021).
Burcham, L. et al. Role of MUC5B during Group B streptococcal vaginal colonization. mBio 13, e0003922 (2022).
Takagi, J. et al. Mucin O-glycans are natural inhibitors of candida albicans pathogenicity. Nat. Chem. Biol. 18, 762–773 (2022).
Jenkins, D. et al. Bacterial amylases enable glycogen degradation by the vaginal microbiome. Nat. Microbiol. 8, 1641–1652 (2023).
Agarwal, K. et al. Glycan cross-feeding supports mutualism between Fusobacterium and the vaginal microbiota. PLoS Biol. 18, e3000788 (2020).
Kolar, S. et al. Group B streptococcus evades host immunity by degrading hyaluronan. Cell Host Microbe 18, 694–704 (2015).
Morrill, S. et al. Gardnerella vaginolysin potentiates glycan molecular mimicry by Neisseria gonorrhoeae. J. Infect. Dis. 228, 1610–1620 (2023).
de Jong, H., Wösten, M. & Wennekes, T. Sweet impersonators: Molecular mimicry of host glycans by bacteria. Glycobiology 32, 11–22 (2022).
Navarro, S. et al. Glycogen availability and pH variation in a medium simulating vaginal fluid influence the growth of vaginal Lactobacillus species and gardnerella vaginalis. BMC Microbiol. 23, 186 (2023).
Mehta, A., Heimburg-Molinaro, J. & Cummings, R. Tools for generating and analyzing glycan microarray data. Beilstein J. Org. Chem. 16, 2260–2271 (2020).
Stappers, M. et al. Recognition of DHN-melanin by a C-type lectin receptor is required for immunity to Aspergillus. Nature 555, 382–386 (2018).
Shahin, M. et al. Characterization of a glycan-binding complex of minor pilins completes the analysis of Streptococcus sanguinis type 4 pili subunits. Proc. Natl. Acad. Sci. USA 120, e2216237120 (2023).
Liu, Y. et al. Altered receptor specificity and cell tropism of D222G hemagglutinin mutants isolated from fatal cases of pandemic A(H1N1) 2009 influenza virus. J. Virol. 84, 12069–12074 (2010).
Parhi, L. et al. Placental colonization by fusobacterium nucleatum is mediated by binding of the Fap2 lectin to placentally displayed Gal-GalNAc. Cell Rep. 38, 110537 (2022).
Andersch-Björkman, Y., Thomsson, K., Holmén Larsson, J., Ekerhovd, E. & Hansson, G. Large scale identification of proteins, mucins, and their O-glycosylation in the endocervical mucus during the menstrual cycle. Mol. Cell. Proteom.: MCP 6, 708–716 (2007).
Rosenstein, I. et al. New type of adhesive specificity revealed by oligosaccharide probes in Escherichia coli from patients with urinary tract infection. Lancet (Lond., Engl.) 2, 1327–1330 (1988).
Spaulding, C. et al. Selective depletion of uropathogenic E. coli from the gut by a FimH antagonist. Nature 546, 528–532 (2017).
Heimburg-Molinaro, J., Song, X., Smith, D. & Cummings, R. Preparation and analysis of glycan microarrays. Current protocols in protein science Chapter 12 (2011).
Thomas-White, K. et al. Culturing of female bladder bacteria reveals an interconnected urogenital microbiota. Nat. Commun. 9, 1557 (2018).
Wu, G. et al. N-glycosylation of cervicovaginal fluid reflects microbial community, immune activity, and pregnancy status. Sci. Rep. 12, 16948 (2022).
Lin, A. et al. Human milk oligosaccharides inhibit growth of group B Streptococcus. J. Biol. Chem. 292 (2017).
Nasciutti, L. et al. Distribution of chondroitin sulfate in human endometrium. Micron (Oxf., Engl.: 1993) 37, 544–550 (2006).
Mahendroo, M. Cervical hyaluronan biology in pregnancy, parturition and preterm birth. Matrix Biol.: J. Int. Soc. Matrix Biol. 78-79, 24–31 (2019).
Akgul, Y., Holt, R., Mummert, M., Word, A. & Mahendroo, M. Dynamic changes in cervical glycosaminoglycan composition during normal pregnancy and preterm birth. Endocrinology 153, 3493–3503 (2012).
Li, S. & Jedrzejas, M. Hyaluronan binding and degradation by Streptococcus agalactiae hyaluronate lyase. J. Biol. Chem. 276, 41407–41416 (2001).
Mello, L., De Groot, B., Li, S. & Jedrzejas, M. Structure and flexibility of Streptococcus agalactiae hyaluronate lyase complex with its substrate. Insights into the mechanism of processive degradation of hyaluronan. J. Biol. Chem. 277, 36678–36688 (2002).
Baron, M., Bolduc, G., Goldberg, M., Aupérin, T. & Madoff, L. Alpha C protein of group B Streptococcus binds host cell surface glycosaminoglycan and enters cells by an actin-dependent mechanism. J Biol. Chem. 279 (2004).
Suits, M. et al. Conformational analysis of the streptococcus pneumoniae hyaluronate lyase and characterization of its hyaluronan-specific carbohydrate-binding module. J. Biol. Chem. 289, 27264–27277 (2014).
Bonnardel, F. et al. Proteome-wide prediction of bacterial carbohydrate-binding proteins as a tool for understanding commensal and pathogen colonisation of the vaginal microbiome. NPJ biofilms microbiomes 7, 49 (2021).
Beeson, J. et al. Adhesion of plasmodium falciparum-infected erythrocytes to hyaluronic acid in placental malaria. Nat. Med. 6, 86–90 (2000).
Chai, W., Beeson, J. & Lawson, A. The structural motif in chondroitin sulfate for adhesion of Plasmodium falciparum-infected erythrocytes comprises disaccharide units of 4-O-sulfated and non-sulfated N-acetylgalactosamine linked to glucuronic acid. J. Biol. Chem. 277, 22438–22446 (2002).
Starc, M., Lučovnik, M., Eržen, Vrlič, P. & Jeverica, S. Protective Effect of Lactobacillus crispatus against Vaginal Colonization with Group B Streptococci in the Third Trimester of Pregnancy. Pathog. (Basel, Switz.) 11, 980 (2022).
Maidment, T. et al. Group B Streptococcus vaginal colonisation throughout pregnancy is associated with decreased Lactobacillus crispatus and increased Lactobacillus iners abundance in the vaginal microbial community. Front. Cell. Infect. Microbiol. 14 (2024).
Marziali, G., Foschi, C., Parolin, C., Vitali, B. & Marangoni, A. In-vitro effect of vaginal lactobacilli against group B Streptococcus. Microbial pathogenesis 136 (2019).
He, Y. et al. Evaluation of the Inhibitory Effects of Lactobacillus gasseri and Lactobacillus crispatus on the Adhesion of Seven Common Lower Genital Tract Infection-Causing Pathogens to Vaginal Epithelial Cells. Front. Med. 7 (2020).
Blixt, O. et al. Printed covalent glycan array for ligand profiling of diverse glycan binding proteins. Proc. Natl. Acad. Sci. USA 101, 17033–17038 (2004).
Fukui, S., Feizi, T., Galustian, C., Lawson, A. & Chai, W. Oligosaccharide microarrays for high-throughput detection and specificity assignments of carbohydrate-protein interactions. Nat. Biotech. 20 (2002).
Day, C. et al. Differential carbohydrate recognition by Campylobacter jejuni strain 11168: influences of temperature and growth conditions. PloS one 4, e4927 (2009).
De Oliveira, D. et al. Blood Group Antigen Recognition via the Group A Streptococcal M Protein Mediates Host Colonization. mBio 8, e02237–02216 (2017).
Mubaiwa, T. et al. The glycointeractome of serogroup B Neisseria meningitidis strain MC58. Sci. Rep. 7 (2017).
Deng, L. et al. Oral streptococci utilize a Siglec-like ___domain of serine-rich repeat adhesins to preferentially target platelet sialoglycans in human blood. PLoS pathogens 10 (2014).
Feizi, T. Angling for recognition. Curr. Biol.: CB 2, 185–187 (1992).
Grant, O., Smith, H., Firsova, D., Fadda, E. & Woods, R. Presentation, presentation, presentation! Molecular-level insight into linker effects on glycan array screening data. Glycobiology 24, 17–25 (2014).
Zhernakova, D. et al. Host genetic regulation of human gut microbial structural variation. Nature 625, 813–821 (2024).
Poole, J., Day, C. J., von Itzstein, M., Paton, J. C. & Jennings, M. P. Glycointeractions in bacterial pathogenesis. Nat. Rev. Microbiol. 16, 440–452 (2018).
Vornhagen, J. et al. Bacterial Hyaluronidase Promotes Ascending GBS Infection and Preterm Birth. mBio 7, e00781–00716 (2016).
Han, Y. Fusobacterium nucleatum: a commensal-turned pathogen. Curr. Opin. Microbiol. 23, 141–147 (2015).
Gonçalves, B. et al. Group B streptococcus infection during pregnancy and infancy: estimates of regional and global burden. Lancet Glob. health 10, e807–e819 (2022).
Bianchi-Jassir, F. et al. Preterm birth associated with group b streptococcus maternal colonization worldwide: systematic review and meta-analyses. Clin. Infect. Dis.: Off. Publ. Infect. Dis. Soc. Am. 65, S133–S142 (2017).
Brown, R. et al. Vaginal dysbiosis increases risk of preterm fetal membrane rupture, neonatal sepsis and is exacerbated by erythromycin. BMC Med. 16, 9 (2018).
Wang, X. et al. Comparative microbial analysis of paired amniotic fluid and cord blood from pregnancies complicated by preterm birth and early-onset neonatal sepsis. PloS one 8 (2013).
Hitti, J. et al. Vaginal indicators of amniotic fluid infection in preterm labor. Obstet. Gynecol. 97, 211–219 (2001).
Wu, G. et al. Glycomics of cervicovaginal fluid from women at risk of preterm birth reveals immuno-regulatory epitopes that are hallmarks of cancer and viral glycosylation. Sci Rep. 14, 20813 (2024).
Shaniztki, B., Ganeshkumar, N. & Weiss, E. Characterization of a novel N-acetylneuraminic acid-specific fusobacterium nucleatum PK1594 adhesin. Oral. Microbiol. Immunol. 13, 47–50 (1998).
Umaña, A. et al. Utilizing whole fusobacterium genomes to identify, correct, and characterize potential virulence protein families. J. Bacteriol. 201, e00273–00219 (2019).
Robinson, L., Schwebke, J., Lewis, W. & Lewis, A. Identification and characterization of NanH2 and NanH3, enzymes responsible for sialidase activity in the vaginal bacterium Gardnerella vaginalis. J. Biol. Chem. 294, 5230–5245 (2019).
Shvartsman, E., Hill, J., Sandstrom, P. & MacDonald, K. Gardnerella revisited: species heterogeneity, virulence factors, mucosal immune responses, and contributions to bacterial vaginosis. Infection and immunity 91 (2023).
Vaneechoutte, M. et al. Emended description of Gardnerella vaginalis and description of Gardnerella leopoldii sp. nov., Gardnerella piotii sp. nov. and Gardnerella swidsinskii sp. nov., with delineation of 13 genomic species within the genus Gardnerella. Int. J. Syst. Evol. Microbiol. 69, 10.1099/ijsem.0.003200 (2019).
Munch, M. et al. Gardnerella species and their association with bacterial vaginosis. J. Infect. Dis. 230 (2024).
Pace, R. et al. Complex species and strain ecology of the vaginal microbiome from pregnancy to postpartum and association with preterm birth. Med. 2, 1027–1049 (2021).
Bulavaitė, A., Maier, T. & Pleckaityte, M. Discrimination of gardnerella species by combining MALDI-TOF protein profile, chaperonin cpn60 sequences, and phenotypic characteristics. Pathogens 10, 277 (2021).
Arnolds, K. L., Martin, C. & Lozupone, C. Blood type and the microbiome- untangling a complex relationship with lessons from pathogens. Curr. Opin. Microbiol. 56, 59–66 (2020).
Abegaz, S. Human ABO Blood groups and their associations with different diseases. BioMed Res. Int. 2021, 6629060 (2021).
Regan, J., Chao, S. & James, L. Maternal ABO blood group type B: a risk factor in the developement of neonatal group B streptococcal disease. Pediatrics 62, 505–509 (1978).
Chanzu, N., Mwanda, W., Oyugi, J. & Anzala, O. Mucosal Blood Group Antigen Expression Profiles and HIV Infections: A Study among Female Sex Workers in Kenya. PloS one 10 (2015).
AbdelMassih, A. et al. The potential use of ABO blood group system for risk stratification of COVID-19. Med Hypotheses 145, 110343 (2020).
Lopera-Maya, E. et al. Effect of host genetics on the gut microbiome in 7,738 participants of the Dutch Microbiome Project. Nat. Genet. 54, 143–151 (2022).
Kononova, S., Litvinova, E., Vakhitov, T., Skalinskaya, M. & Sitkin, S. Acceptive Immunity: The role of fucosylated glycans in human host-microbiome interactions. Int. J. Mol. Sci. 22, 3854 (2021).
Uchida, H. et al. Lactic acid bacteria (LAB) bind to human B- or H-antigens expressed on intestinal mucosa. Biosci., Biotechnol., Biochem. 70, 3073–3076 (2006).
Kundu, S. et al. Secretor status is a modifier of vaginal microbiota-associated preterm birth risk. Microb. Genom. 10, 001323 (2024).
McCoy, J. et al. Cervicovaginal microbial-immune state and group B streptococcus colonization in pregnancy. Am. J. Perinatol. 41, e2539–e2546 (2024).
Verani, J., McGee, L. & Schrag, S. Prevention of perinatal group B streptococcal disease-revised guidelines from CDC, 2010. Mmwr. Recommendations Rep.: Morbidity Mortal. Wkly. report. Recommendations Rep. 59, 1–36 (2010).
Ahmadzia, H. & Heine, R. Diagnosis and management of group B streptococcus in pregnancy. Obstet. Gynecol. Clin. North Am. 41, 629–647 (2014).
Bayar, E., Bennett, P., Chan, D., Sykes, L. & MacIntyre, D. The pregnancy microbiome and preterm birth. Semin. Immunopathol. 42, 487–499 (2020).
Bayar, E. et al. Safety, tolerability, and acceptability of lactobacillus crispatus CTV-05 (LACTIN-V) in pregnant women at high-risk of preterm birth. Beneficial Microbes 14, 45–56 (2023).
Ravel, J. et al. Daily temporal dynamics of vaginal microbiota before, during and after episodes of bacterial vaginosis. Microbiome 1, 29 (2013).
Mitra, A. et al. The vaginal microbiota associates with the regression of untreated cervical intraepithelial neoplasia 2 lesions. Nat. Commun. 11, 1999 (2020).
Brotman, R. Vaginal microbiome and sexually transmitted infections: an epidemiologic perspective. J. Clinic. Invest. 121, 4610–4617 (2011).
Song, X. et al. Novel fluorescent glycan microarray strategy reveals ligands for galectins. Chem. Biol. 16, 36–47 (2009).
Li, C. et al. Noncovalent microarrays from synthetic amino-terminating glycans: Implications in expanding glycan microarray diversity and platform comparison. Glycobiology 31, 931–946 (2021).
Likhosherstov, L. M., Novikova, O. S., Malysheva, N. N. & Piskarev, V. E. Synthesis of mono- and di-α-l-fucosylated 2-acetamido-2-deoxy-N-glycyl-β-d-glucopyranosylamines modeling N-glycoprotein carbohydrate-peptide bond region based on 2-acetamido-N-(N-tert-butyloxycarbonylglycyl)-2-deoxy-β-d-glucopyranosylamine. Russian Chem. Bull. 64, 1445–1450 (2016).
Ye, J. et al. Reprogramming the enzymatic assembly line for site-specific fucosylation. Nat. Catal. 2, 514–522 (2019).
Gao, T. et al. Chemoenzymatic synthesis of O-mannose glycans containing sulfated or nonsulfated HNK-1 epitope. J. Am. Chem. Soc. 141, 19351–19359 (2019).
Chai, W., Luo, J., Lim, C. & Lawson, A. Characterization of heparin oligosaccharide mixtures as ammonium salts using electrospray mass spectrometry. Anal. Chem. 70, 2060–2066 (1998).
Wu, N. et al. Glycan markers of human stem cells assigned with beam search arrays. Mol. Cell. Proteom.: MCP 18, 1981–2002 (2019).
Chai, W., Beeson, J., Kogelberg, H., Brown, G. & Lawson, A. Inhibition of adhesion of Plasmodium falciparum-infected erythrocytes by structurally defined hyaluronic acid dodecasaccharides. Infect. Immun. 69, 420–425 (2001).
Parkkinen, J. & Finne, J. Isolation and structural characterization of five major sialyloligosaccharides and a sialylglycopeptide from normal human urine. Eur. J. Biochem. 136, 355–361 (1983).
Sun, S. et al. Toward automated identification of glycan branching patterns using multistage mass spectrometry with intelligent precursor selection. Anal. Chem. 90, 14412–14422 (2018).
Yu, G. et al. Effect and limitation of excess ammonium on the release of O-glycans in reducing forms from glycoproteins under mild alkaline conditions for glycomic and functional analysis. Anal. Chem. 82, 9534–9542 (2010).
Chai, W., Stoll, M., Galustian, C., Lawson, A. & Feizi, T. Neoglycolipid technology: deciphering information content of glycome. Methods Enzymol. 362, 160–195 (2003).
Beeson, J. et al. Structural basis for binding of Plasmodium falciparum erythrocyte membrane protein 1 to chondroitin sulfate and placental tissue and the influence of protein polymorphisms on binding specificity. J. Biol. Chem. 282, 22426–22436 (2007).
Mahoney, D., Blundell, C. & Day, A. Mapping the hyaluronan-binding site on the link module from human tumor necrosis factor-stimulated gene-6 by site-directed mutagenesis. J. Biol. Chem. 276, 22764–22771 (2001).
Bitter, T. & Muir, H. A modified uronic acid carbazole reaction. Anal. Biochem. 4, 330–334 (1962).
Liu, Y. et al. The minimum information required for a glycomics experiment (MIRAGE) project: improving the standards for reporting glycan microarray-based data. Glycobiology 27, 280–284 (2017).
Liu, Y. et al. Neoglycolipid-based oligosaccharide microarray system: preparation of NGLs and their noncovalent immobilization on nitrocellulose-coated glass slides for microarray analyses. Methods Mol. Biol. (Clifton, N. J.) 808, 117–136 (2012).
Akune, Y. et al. CarbArrayART: a new software tool for carbohydrate microarray data storage, processing, presentation, and reporting. Glycobiology 32, 552–555 (2022).
Hansen, L., Transtrum, M., Quinn, C. & Demarse, N. Enzyme-catalyzed and binding reaction kinetics determined by titration calorimetry. Biochim. Biophys. Acta. 1860, 957–966 (2016).
Acknowledgements
This work has been funded by the March of Dimes European Prematurity Research Centre grant 22-FY18-82 awarded to Imperial College London (ICL). The glycan microarray studies were performed in the ICL Carbohydrate Microarray Facility supported by the Wellcome Trust Biomedical Resource Grants (WT099197/Z/12/Z, 108430/Z/15/Z and 218304/Z/19/Z). The sequence-defined glycan microarrays contain oligosaccharides provided by collaborators including CFG Core D, Nicolai Bovin, Chunxia Li, Akihiro Imamura, Ulrika Westerlind, Kelly Moremen, and Apoorva Srivastava whom we thank. J.R.M. and the Division of Digestive Diseases at ICL receive financial and infrastructure support from the NIHR Imperial Biomedical Research Centre (BRC) based at ICL and ICL Healthcare NHS Trust. L.S. and B.G.M. are supported by Imperial Health Charity in partnership with the Parasol Foundation and the Rosetrees Trust. P.R.B. and L.S. receive infrastructure support from the NIHR Imperial Biomedical Research Centre (BRC) based at ICL and IC Healthcare NHS Trust. Y.L., A.S.P. and B.A.P., thank the GLYCOTwinning project HORIZON-WIDERA-2021-101079417 funded by European the Commission and a DL-57/2016 Program Contract to B.A.P. We would like to acknowledge the Centre for Biomolecular Spectroscopy at King’s College London that is funded by the Wellcome Trust and the British Heart Foundation (WT202767/Z/16/Z and IG/16/2/32273). V.P. receives financial support of the Ministry of Science and Higher Education of the Russian Federation (Contract No. 075-00277-24-00). The authors thank Prof Barbara Mulloy for her help in NMR data interpretation and Dr. Daniel Slade for the delta galkt and delta galkt fap2 F. nucleatum 23726 mutants.
Author information
Authors and Affiliations
Contributions
V.T.O., Y.L. and D.A.M. designed the work and wrote the manuscript with input from all authors. V.T.O. performed the bacterial microarray assays, the microarray quality control with help from L. R. and analysed the data. W.C. and Y.Z. prepared the glycan probes, A.D.M. printed the glycan microarrays and with V.T.O. established the optimal conditions for whole bacteria binding used in this manuscript. V.T.O. and B.G.M. set up and adapted the bacterial staining protocols. L.A.R. prepared the bacterial samples under J.R.M. supervision. A.C.D. designed and performed the flow cytometry based bacterial binding to VEC assays. A.S.P. and B.A.P. produced the full-length hyaluronidase and G.D.N performed the ITC assays. B.G.M. and Y.S.L. expanded the initial bacterial collection received from the Wolfe and Lawley labs. H.C. and V.P. produced some of the glycans used for this work. B.G.M. and H.A. performed and analysed the electron microscopy with help from T.R.D.C. Y.A. helped with annotation of glycan probes on in house databases. L.S., P.R.B., J.R.M. and T.F. provided useful insight and valuable input for the scientific work and experimental design.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks the anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Tajadura-Ortega, V., Chai, W., Roberts, L.A. et al. Identification and characterisation of vaginal bacteria-glycan interactions implicated in reproductive tract health and pregnancy outcomes. Nat Commun 16, 5207 (2025). https://doi.org/10.1038/s41467-025-60404-1
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-025-60404-1
