Harnessing gut microbes for glycan detection and quantification

Abstract

Glycans facilitate critical biological functions and control the mammalian gut microbiota composition by supplying differentially accessible nutrients to distinct microbial subsets. Therefore, identifying unique glycan substrates that support defined microbial populations could inform therapeutic avenues to treat diseases via modulation of the gut microbiota composition and metabolism. However, examining heterogeneous glycan mixtures for individual microbial substrates is hindered by glycan structural complexity and diversity, which presents substantial challenges to glycomics approaches. Fortuitously, gut microbes encode specialized sensor proteins that recognize unique glycan structures and in-turn activate predictable, specific, and dynamic transcriptional responses. Here, we harness this microbial machinery to indicate the presence and abundance of compositionally similar, yet structurally distinct glycans, using a transcriptional reporter we develop. We implement these tools to examine glycan mixtures, isolate target molecules for downstream characterization, and quantify the recovered products. We assert that this toolkit could dramatically enhance our understanding of the mammalian intestinal environment and identify host-microbial interactions critical for human health.

Fig. 1. Bt PUL transcription reflects target glycan abundance.

a A cartoon depicting prototypical glycan utilization by PUL-encoded gene products. SGBPs (green) specifically bind target glycans in the extracellular milieu, which are transported across the outer membrane (OM) by the SusCD complex (blue). Internalized glycans are depolymerized by glycosyl hydrolase (GH) and polysaccharide lyase (PL) activities (purple) generating ligands that convert PUL sensors (red) to their active form. PUL sensors are deactivated following removal of their ligands from the periplasm by GH activities or importation across the inner membrane (IM; orange). b Ligand-bound sensors direct increased PUL transcription by sequence-specific interactions with PUL promoters. These promoters can be introduced into pBolux to drive glycan-responsive bioluminescence. c Growth of wild-type Bt (black) or strains lacking the CS-inducible susC (ΔBT3332, blue), 3 CS-specific lyases (ΔBT3324 ΔBT3350 ΩBT4410, purple), CS-sensor (ΔBT3334, red), or a glucuronyl hydrolase (ΔBT3348, orange) were measured during anaerobic culture in minimal media containing 0.1% CS as a sole carbon source. Values are the mean of 8 biological replicates and error bars are SEM in color-matched shading. d The fold increase of the CS-inducible susC (BT3332) mRNA levels were measured by qPCR in either wild-type Bt (open bar) or a strain lacking the CS-sensor (red bar) following introduction of 0.2% CS into the culture media. The fold increase was calculated as the change in transcript levels from cultures before and 2 h after the introduction of CS. Values are the mean of 6 biological replicates, error bars are SEM, and P-values were computed using a two-tailed student’s t-test and ** indicates values <0.01 and ns > 0.05. e Growth of wild-type Bt (black) or strains lacking the levan-inducible susC (ΔBT1763, blue), 4 levanases (ΔBT1760-1759 ΔBT3082 ΩBT1765, purple), fructan sensor (ΔBT1754, red), or a putative inner-membrane fructose importer (ΩBT1758, orange) were measured during anaerobic culture in minimal media containing levan as a sole carbon source. Values are the mean of 8 biological replicates, error bars are SEM in color-matched shading. f The fold increase of levan-inducible susC (BT1763) mRNA levels were measured in either wild-type Bt (open bars) or a strain lacking the levan sensor (red bars) following the introduction of 0.2% levan in the culture media. The fold change was calculated as the change in transcript levels from cultures before and 2 h after levan introduction. Values are the mean of 6 biological replicates, error bars are SEM, and P-values were computed using a two-tailed student’s t-test and ** indicates values <0.01, * <0.05 and ns > 0.05. g The fold change in BT1763 mRNA levels were measured by qPCR in wild-type Bt as described in (d), 120 min following the introduction of mixtures of 0.2%, 0.02%, or 0.002% levan supplemented with galactose to 0.5% total carbohydrate. Values are the mean of 6 independent measurements, error bars are SEM, and P-values were calculated by 2-way ANOVA with Tukey’s honest significance test and *** represents values <0.001. Source data are provided as a Source Data file.

Fig. 2. Construction of a Bacteroides-optimized bioluminescent reporter.

a Schematic depicting the construction of a bioluminescent reporter that encodes the entire Pl lux cassette under control of the Bt rpoD promoter and rpiL* RBS (top); a Bacteroides-optimized lux cassette with rearranged luxA-E (indicated by the shaded regions) and Bt intergenic regions from a constitutively expressed Bt operon (BT1160-1155; middle); or pBolux which has BamHI and SpeI sites positioned upstream of the Bacteroides-optimized lux cassette (bottom) in the multi-copy plasmid pLYL01. b Relative luminescence (solid lines) or growth (dashed lines) from Bt strains harboring an empty vector (black) or plasmids containing either the lux operon from Pl (plux-Pl, blue) or the Bacteroides-optimized lux cassette (plux-Bt, red) expressed from the Bt rpoD promoter and rpiL* RBS were measured during growth in minimal media containing 0.5% galactose as the sole carbon source. c The relative luminescence (solid lines) or growth (dashed lines) of Bt strains harboring empty pBolux (black) or a plasmid with the Bt rpoD promoter cloned into the BamHI and SpeI sites (red) during growth in galactose as the sole carbon source. All values in b, c are the mean of 8 biological replicates and error bars are SEM in color-matched shading. Source data are provided as a Source Data file.

Fig. 3. Construction of a glycan-responsive reporter in Bt.

a Relative luminescence from wild-type Bt strains harboring pBolux (black) or a plasmid including the promoter region preceding the CS-inducible susC gene (P-BT3332, pink) following the introduction of CS as the sole carbon source normalized by the relative luminescence of identical cultures supplied galactose. Values are the mean of 12 biological replicates and error is SEM in color-matched shading. b Relative luminescence from wild-type Bt (black) or strains lacking a CS-inducible susC (ΔBT3332, blue), 3 CS-specific lyases (ΔBT3324 ΔBT3350 ΩBT4410, purple), CS-sensor (ΔBT3334, red), or a glucuronyl hydrolase (ΔBT3348, orange) harboring P-BT3332 following the introduction of an equal mixture of CS and galactose normalized to measurements from identical strains supplied galactose alone. c Relative luminescence from wild-type Bt (black) or strains lacking 3 CS-specific lyases (ΔBT3324 ΔBT3350 ΩBT4410, purple) or the CS-sensor (ΔBT3334, red) harboring P-BT3332 following the introduction of a mixture of unsulfated CS disaccharide (di0S) and galactose normalized to measurements from identical strains supplied galactose alone. Values are the mean of 6 biological replicates and error is SEM in color-matched shading. d Relative luminescence from wild-type Bt (solid lines) or strains lacking the CS-sensor (ΔBT3334, dashed lines) harboring P-BT3332 following the introduction of a mixture of galactose and either hyaluronic acid (HA, green) or heparin (blue) and galactose normalized to measurements from identical strains supplied galactose alone. For b, d, values are the mean of 8 biological replicates and error is SEM in color-matched shading. Source data are provided as a Source Data file.

Fig. 4. A levan-responsive reporter reveals multiple levanases coordinate fructan utilization in Bt.

a Relative luminescence from wild-type Bt harboring pBolux (black) or a plasmid including the promoter region preceding the levan-inducible susC gene (P-BT1763, pink) following the introduction of levan as the sole carbon source and normalized by the relative luminescence of identical cultures supplied galactose. Values are the mean of 12 biological replicates and error is SEM in color-matched shading. b, c Relative luminescence from wild-type Bt (black) or strains lacking the levan-inducible susC (ΔBT1763, blue), 4 levan-specific hydrolases (ΔBT1760-1759 ΔBT3082 ΩBT1765, purple), fructan sensor (ΔBT1754, red), or a putative inner-membrane fructose transporter (ΩBT1758, orange) harboring P-BT1763 were measured following the introduction of an equal mixture of galactose and (b) levan or (c) fructose and normalized by the relative luminescence of identical cultures supplied galactose alone. d Growth of wild-type Bt (black) or strains lacking the levan-inducible susC (ΔBT1763, blue), 4 levan-specific hydrolases (ΔBT1760-1759 ΔBT3082 ΩBT1765, purple), fructan sensor (ΔBT1754, red), or a putative inner-membrane fructose transporter (ΩBT1758, orange) were measured during anaerobic culture in minimal media containing fructose as a sole carbon source. e Relative luminescence of wild-type Bt or strains lacking all other levanases except BT1760 (ΔBT1759 ΔBT3082 ΩBT1765, pink), BT1759 (ΔBT1760 ΔBT3082 ΩBT1765, teal), BT3082 (ΔBT1760-59 ΩBT1765, lavender), or BT1765 (ΔBT1760-59 ΔBT3082, purple) harboring the levan-responsive reporter following the introduction of a mixture of levan and galactose normalized with measurements from identical cultures supplied galactose alone. f Relative luminescence of wild-type Bt or strains lacking BT1760 (pink), BT1759 (teal), BT3082 (lavender), or BT1765 (purple) harboring P-BT1763 following the introduction of a mixture of levan and galactose normalized with measurements from identical cultures supplied galactose alone. For b–f, values are the mean of 8 biological replicates and error is SEM in color-matched shading. Source data are provided as a Source Data file.

Fig. 5. Species-specific responses enable PUL reporters to distinguish between compositionally identical yet structurally distinct glycans.

a Growth of wild-type Bt (gray) and Bo (black) or a strain lacking the Bo inulin sensor (ΔBACOVA_04496, red) were measured during anaerobic culture in minimal media containing inulin as a sole carbon source. b Relative luminescence from wild-type Bo harboring pBolux (black) or a plasmid including the promoter region preceding the inulin-inducible susC gene (P-BACOVA_04505, pink) were measured following the introduction of inulin as the sole carbon source and normalized by the relative luminescence from identical cultures supplied galactose. Values are the mean of 12 biological replicates and error is SEM in color-matched shading. Relative luminescence from wild-type Bo harboring P-BACOVA_04505 (black) or an isogenic strain lacking the Bo inulin sensor (ΔBACOVA_04496, red) were measured following the introduction of an equal mixture of galactose and (c) inulin or (d) fructose and normalized by the relative luminescence of identical cultures supplied galactose alone. For panels a, c, d, values are the mean of 8 biological replicates and error is SEM in color-matched shading. Source data are provided as a Source Data file.

Fig. 6. PUL reporters reflect dose-dependent transcription.

a Relative luminescence from a wild-type Bt strain harboring P-BT3332 following the introduction of twofold serial dilutions of 0.4% CS containing galactose to a total carbohydrate content of 0.5% and normalized to identical cultures supplied galactose alone. b Relative luminescence from a wild-type Bt strain harboring P-BT1763 following the introduction of twofold serial dilutions of 0.4% levan containing galactose to a total carbohydrate content of 0.5% and normalized to identical cultures supplied galactose alone. For a and b, values are the mean of 12 biological replicates and error is SEM in color-matched shading. c The AUC of response curves measured from wild-type Bt (black, see a) or a mutant lacking the glucuronyl hydrolase (Δgh, orange, see f) harboring P-BT3332 supplied twofold dilutions of CS to each strain and normalized by identical cultures supplied galactose alone. d The AUC of response curves measured from wild-type Bt (black, see b) or a mutant lacking a putative inner-membrane transporter (Ωimporter, orange; see Supplementary Fig. 9f) harboring P-BT1763 supplied twofold dilutions of levan to each strain and normalized by identical cultures supplied galactose alone. For c, d, values are the mean of 12 biological replicates, error bars are standard deviation, and P-values were computed by two-way ANOVA with Dunnett correction and *** indicates values <0.001, ** <0.01, * <0.05, and ns > 0.05. e The AUC of response curves measured from wild-type Bt strains harboring either P-BT3332 (open blue squares) supplied mixtures containing twofold serial dilutions of levan or P-BT1763 (open red circles) supplied twofold serial dilutions of CS with galactose to a total carbohydrate content of 0.5% and normalized by identical cultures supplied galactose alone. Values are the mean of 6 biological replicates, error bars represent standard deviation, and P-values were computed using two-way ANOVA with Tukey’s honest significance test and ns indicates values >0.05. f Relative luminescence from a gh-deficient Bt strain harboring P-BT3332 following the introduction of twofold serial dilutions of 0.4% CS containing galactose to a final carbohydrate content of 0.5% and normalized to identical cultures supplied galactose alone. Values are the mean of 12 biological replicates and error is SEM in color-matched shading. Source data are provided as a Source Data file.

Fig. 7. PUL reporters facilitate SGBP-mediated target glycan isolation.

Purified BT3330, BSA, or BT1761 was analyzed by affinity-PAGE in the presence of (a) 0.1% CS, (b) no glycan or (c) 0.1% levan. Images are representative of 3 independent experiments. The fold difference in AUC from wild-type Bt strains harboring either P-BT1763 (gray bars) or P-BT3332 (open bars) supplied fractions eluted from immobilized (d) BT3330 or (e) BT1761 in combination with 0.2% galactose and normalized against cultures supplied galactose alone. Normalized AUC values were divided by responses from identical strains supplied equivalent fractions eluted from columns containing control lysates prepared from cells harboring an empty vector (pT7-7; see Supplementary Fig. 10a, b). Values are the mean of 4 biological replicates, error bars are SEM, and P-values were computed using a two-tailed student’s t-test and *** indicate values <0.001, ** <0.01, * <0.05, and ns > 0.05. Relative luminescence from wild-type Bt or strains lacking BT3334 (red) or BT2826 (blue) harboring either (f) P-BT3332 or (g) P-BT2818 supplied an equal mixture of PMOG and galactose and normalized to identical cultures supplied galactose alone. Values are the mean of 8 biological replicates and error bars are SEM in color-matched shading. h Relative luminesecence from wild-type Bt strains harboring pBolux (black), P-BT3332 (green), or P-BT2818 (blue) supplied 0.4% galactose and a 1:10 dilution of pooled, concentrated material co-purifying with BT3330 pre-incubated with PMOG and normalized to identical cultures supplied galactose alone. Values are the mean of 4 biological replicates and error bars are SEM in color-matched shading. i The AUC from a wild-type Bt strain harboring P-BT3332 supplied 0.4% galactose containing either 0.1% PMOG (blue) or a 1:10 dilution of pooled, concentrated material co-purifying with BT3330 (red) pre-incubated with PMOG with and normalized to identical cultures supplied galactose alone. Measurements were collected alongside an identical strain supplied twofold serial dilutions of CS with galactose to a total carbohydrate content of 0.5% and normalized to identical cultures supplied galactose alone. Linear regression models (gray line) were computed in Prism, and sample concentrations were estimated with the derived equations. Values are the mean of 2 independent measurements, error bars are standard deviation, and the dashed gray lines are the 95% confidence interval. Source data are provided as a Source Data file.