Hierarchical glycolytic pathways control the carbohydrate utilization regulator in human gut Bacteroides

Abstract

Human dietary choices control the gut microbiome. Industrialized populations consume abundant amounts of glucose and fructose, resulting in microbe-dependent intestinal disorders. Simple sugars inhibit the carbohydrate utilization regulator (Cur), a transcription factor in members of the prominent gut bacterial phylum, Bacteroidetes. Cur controls products necessary for carbohydrate utilization, host immunomodulation, and intestinal colonization. Here, we demonstrate how simple sugars decrease Cur activity in the mammalian gut. Our findings in two Bacteroides species show that ATP-dependent fructose-1,6-bisphosphate (FBP) synthesis is necessary for glucose or fructose to inhibit Cur, but dispensable for growth because of an essential pyrophosphate (PPi)-dependent enzyme. Furthermore, we show that ATP-dependent FBP synthesis is required to regulate Cur in the gut but does not contribute to fitness when cur is absent, indicating PPi is sufficient to drive glycolysis in these bacteria. Our findings reveal how sugar-rich diets inhibit Cur, thereby disrupting Bacteroides fitness and diminishing products that are beneficial to the host.

Fig. 1. A bioluminescent transcriptional reporter of Cur activity.

a, b Bioluminescence from a wild-type Bt strain harboring P-fusA2 (black) or P-∆22 bp (blue) cultured in media containing (a) fucose or (b) galNAc as the sole carbon source normalized to measurements collected from isogenic strains harboring a promoter-less pBolux plasmid. c Normalized bioluminescence from a wild-type Bt strain harboring P-fusA2 (black) or P-∆22 bp (blue) and a ∆cur Bt strain harboring P-fusA2 (red) cultured in fructose as a sole carbon source. d Normalized bioluminescence from a wild-type Bt strain harboring P-fusA2 cultured in PMOG as the sole carbon source (black) or in combination with 0.1% (orange), 0.2% (purple), or 0.5% (green) fructose. For (a–d), n = 8 biological replicates, error is SEM in color matched shading.

Fig. 2. Fructose and glucose require phosphorylation for Cur inhibition.

a Normalized bioluminescence from ∆sensor^PUL22 harboring P-fusA2 cultured in media containing PMOG (black), PMOG and fructose (red), or PMOG and glucose (blue). b Fold change of fusA2 transcript amounts from wild-type, ∆sensor^PUL22, and ∆frk^PUL22 cultured in media containing PMOG as the sole carbon source 10 minutes following the addition of 0.2% glucose (left) or 60 minutes after 0.2% fructose addition (right). c Fold change of frk^PUL22 transcripts from wild-type or ∆sensor^PUL22 60 minutes following the addition of 0.2% fructose to cells cultured in media containing PMOG. d Fructokinase activity of purified Frk^PUL22 protein. e Growth of wild-type or ∆frk^PUL22 in minimal media containing fructose as the sole carbon source. f, g Normalized bioluminescence from (f) ∆frk^PUL22 or (g) ∆glk harboring P-fusA2 cultured in media containing PMOG (black) as the sole carbon source or equal mixtures of PMOG and fructose (red) or glucose (blue). h Fold change of fusA2 (left) and chb^PUL80 (right) transcript amounts from ∆glk cultured in media containing PMOG as the sole carbon source 10 minutes following 0.2% glucose addition or 60 min following fructose addition. For (a, e–g) n = 8 biological replicates; error is SEM in color matched shading. For (b, c, h), n = 6 biological replicates; error is SEM. For (d), n = 4 technical replicates; error is SEM. P-values were calculated by 2-way ANOVA with Fisher’s LSD test and * represents values < 0.05, ** <0.01, *** <0.001. Exact P-values are provided in the Source Data file.

Fig. 3. Distinct enzyme classes synthesize FBP in Bt.

a Schematic of the glycolytic pathway in Bt. b, c Enzyme kinetics of purified PfkA (green circles), PfkB (purple squares), PfkC (white triangles), or Pfp (black diamonds) in reactions containing either (b) ATP or (c) PPi as a phosphoryl donor. d PfkA activity in the presence of increasing PPi amounts. e PPi amounts in whole cell lysates of wild-type Bt grown in fructose or PMOG as the sole carbon source. For (a) abbreviations are as follows: Glu glucose, Fru fructose, G6P glucose-6P, F6P fructose-6P, FBP fructose bisphosphate, PEP phosphoenolpyruvate, ATP adenosine triphosphate, PPi pyrophosphate, Pfk phosphofructokinase, Pfp phosphofructose phosphotransferase, Pgi phosphoglucoseisomerase, Glk glucokinase, Frk fructokinase. For (b–d) n = 4 technical replicates, error is SEM. For (e) n = 4 biological replicates, error is SEM; P-values were calculated by 1-way ANOVA with Fisher’s LSD test and *** represents values < 0.001. Exact P-values are provided in the Source Data file.

Fig. 4. ATP-dependent FBP production is required for Cur inhibition but dispensable for growth.

a Growth of wild-type (black), ∆pfkA (green), ∆pfkB (purple), or ∆pfkAB (blue) in minimal media containing fructose as a sole carbon source. b FBP synthesis measured from whole cell lysates of wild-type, ∆pfkA, ∆pfkB, or ∆pfkAB. c Log₂ fold change of steady-state glycolytic intermediates in ∆pfkA, ∆pfkB, or ∆pfkAB in glucose, fructose, PMOG, PMOG with glucose, or PMOG with fructose. d Bioluminescence from wild-type (black), ∆pfkA (green), ∆pfkB (purple), ∆pfkAB (blue), or ∆pfkA ∆cur (red) harboring P-fusA2 in minimal media containing equal amounts of PMOG and fructose. e Fold change of fusA2, chb^PUL80, and fucI transcript amounts relative to wild-type Bt from ∆pfkA, ∆pfkB, or ∆pfkAB cultured in minimal media containing fructose as a sole carbon source. f fusA2 transcript amounts in wild-type, ∆pfkA, ∆cur, and ∆pfkA ∆cur harboring empty vector (nbu) or complementing plasmids grown in glucose as the sole carbon source. For (c), abbreviations are as follows: G1P glucose-1P, G6P glucose-6P, FBP fructose bisphosphate, DHAP dihydroxyacetone-P, G3P 3-phosphoglycerate, G2P 2-phosphoglycerate, PEP phosphoenolpyruvate, PYR pyruvate, ATP adenosine triphosphate, ADP adenosine diphosphate, NAD nicotinamide adenine dinucleotide. For (a, d) n = 8 biological replicates; error is SEM in color matched shading. For (b) n = 6 biological replicates; error is SEM. For (e, f) n = 6 biological replicates; error is SEM. For (b, e) P-values were calculated by 2-way ANOVA with Fisher’s LSD test and * represents values < 0.05, ** <0.01, *** <0.001. For (f) P-values were calculated by 1-way ANOVA with Fisher’s LSD test and *** represents values < 0.001. Exact P-values are provided in the Source Data file.

Fig. 5. ATP-dependent FBP production is required for intestinal fitness by controlling Cur.

a Competitive fitness of ∆pfkA co-introduced into germ-free mice with equal amounts of wild-type Bt and fed a standard polysaccharide rich chow (black squares, n = 10 biological replicates) or a sugar-rich chow (pink squares, n = 8 biological replicates). b Competitive fitness ∆pfkA ∆cur co-introduced into germ-free mice with equal amounts of ∆cur and fed a standard polysaccharide rich chow (black circles, n = 10 biological replicates) or a sugar-rich chow (pink circles, n = 4 biological replicates). For all panels, error bars are SEM. P-values were calculated using 2-way ANOVA with Bonferroni correction and * represents values < 0.05, ** <0.01, *** <0.001. Exact P-values are provided in the Source Data file.

Fig. 6. Bf-pfkA is required for glucose- and fructose-mediated Cur inhibition and ATP-dependent FBP synthesis in B. fragilis.

a, b Bioluminescence from wild-type Bf (black), ∆Bf-pfkA (green), ∆Bf-cur (red), or ∆Bf-pfkA ∆Bf-cur (blue) harboring P-Bf-fusA2 were cultured in media containing a mixture of equal amounts of PMOG and (a) glucose or (b) fructose. c FBP synthetic rates measured from whole cell lysates of wild-type or ∆Bf-pfkA supplied ATP as a phosphoryl donor. For (a, b) n = 8 biological replicates; error is SEM in color matched shading. For (c) n = 5 biological replicates; error is SEM. P-values were calculated using an unpaired t-test and *** represents values < 0.001. Exact P-values are provided in the Source Data file.