Metabolic control of enteroendocrine cell fate through a redox state sensor CtBP

Abstract

Enteroendocrine (EE) cells monitor the intestinal nutrient composition and consequently control organismal physiology through hormonal signaling. In addition to the immediate effects on hormone secretion, nutrients influence EE cell abundance by affecting the determination and maintenance of cell fate. EE cells are known to import and respond to dietary sugars, but how the sugar-induced changes in the intracellular metabolic state are sensed to control the immediate and long-term responses of EE cells, remains poorly understood. We report that the NADH binding transcriptional cofactor C-terminal binding protein (CtBP) acts at the interface between nutrient sensing and fate regulation of Drosophila larval EE cells, thus controlling organismal energy metabolism and survival on a high sugar diet. CtBP dimerization in EE cells is regulated through the redox balance of nicotinamide cofactors controlled by glycolysis and pentose phosphate pathway, allowing EE cells sense their internal metabolic state in response to sugar catabolism. CtBP interacts with the EE cell fate determining transcription factor Prospero through a conserved binding motif and binds to genomic targets controlling EE cell fate and size, such as components of Notch and insulin/mTOR pathways. Collectively, our findings uncover a modality where changes in intracellular redox state serve as an instructive signal to control EE cell function to globally control organismal homeostasis.

Figure 1.. CtBP regulates larval energy metabolism and is necessary for sugar tolerance

(A) Schematic overview of the whole-body RNAi screen for sugar intolerance, focusing on proteins with Rossmann fold with predicted nicotinamide cofactor binding activity. (B) Genetic screen identifies several genes necessary for sugar tolerance, including C-terminal binding protein (CtBP). Pupariation index (PI), rewarding for rapid pupariation and high survival rate, was calculated for each condition. X-axis displays the genotype effect, i.e. log_₂ fold change (Log2FC) of PI between control and experimental genotype on low sugar diet (LSD). Y-axis displays the sugar intolerance, i.e. Log2FC of PI of each genotype on LSD vs. high sugar diet (HSD). (C) Whole-body knockdown of CtBP results in sugar intolerance, manifested as increased lethality and delayed larval development on HSD (20% sucrose). Pupariation rate data for control and CtBP RNAi larvae on HSD were compared using the log-rank test (χ² = 109.3, P < 0.0001). (D) Whole-body ctbp loss-of-function (CtBP⁰³⁴⁶³/CtBP^87De−10) decreases sugar tolerance compared to heterozygous control (+/CtBP⁰³⁴⁶). Survival data were compared using the log-rank test (χ² = 351.0, P < 0.0001). (E) ctbp-deficient (CtBP⁰³⁴⁶³/CtBP^87De−10) larvae exhibit a hunger-like behavior on a sugar-only diet (5% sucrose), quantified by the number of mouth hook movements per min. Two-tailed t-test, P = 0.0135. (F) ctbp mutant (CtBP⁰³⁴⁶³/CtBP^87De−10) larvae have higher starvation resistance compared to control (+/CtBP⁰³⁴⁶) (log-rank test, χ² = 50.48, P < 0.0001). ctbp-deficient (CtBP⁰³⁴⁶³/CtBP^87De−10) larvae have elevated levels of glucose (G) and fructose (H) in their hemolymph on sugar-only diet compared to control (+/CtBP⁰³⁴⁶) animals (two-tailed t-test, P = 0.0297 and P = 0.0012, respectively). (I) Loss of ctbp function (CtBP⁰³⁴⁶³/CtBP^87De−10) leads to hemolymph acidification (two-tailed t-test, P = 0.0135). This is accompanied by reduced accumulation of triacylglycerols (J) and decreased body mass (K) (two-tailed t-test, P = 0.0177 and P = 0.0403, respectively). (L) ctbp-deficient (CtBP⁰³⁴⁶³/CtBP^87De−10) larvae have reduced intestinal lipid content quantified in (M) as a lipid volume (mm³) per cell (two-tailed t-test, P = 0.0403). (N) The intestines of ctbp-deficient (CtBP⁰³⁴⁶³/CtBP^87De−10) larvae are significantly thinner than those of control larvae (+/CtBP⁰³⁴⁶) (two-tailed t-test, P = 0.0175).

Figure 2.. CtBP acts in enteroendocrine cells to control organismal metabolism

(A) ctbp deficiency (CtBP⁰³⁴⁶³/CtBP^87De−10) in Drosophila larvae results in a reduction on relative number of EE cells on HSD (30% sucrose), quantified in (B). The data were compared with unpaired t-test, P = 0.0012. (C) Upper panel: The distribution of Dh31-, Tk-, and AstA-expressing EE cells in Drosophila larvae: a schematic overview. Lower panel: Whole-body ctbp loss-of-function (CtBP⁰³⁴⁶³/CtBP^87De−10) results in a loss of Dh31-, Tk-, and AstA-expressing EE cells on HSD. (D) Quantification of C. Two-tailed t-tests, *P < 0.01; **P = 0.0032. (E) Targeted loss of CtBP in EE cells (Pros-GAL4>CtBP RNAi) leads to reduced relative number of EE cells in larvae on HSD. Two-tailed t-test, P = 0.0046. (F) EE cell specific loss of CtBP function (Pros-GAL4 > CtBP^RNAi) on HSD reduces EE cell size (Mann-Whitney U test, P = 0.00313). (G) Loss of CtBP function in EE cells (Pros-GAL4>CtBP RNAi) leads to sugar intolerance of Drosophila larvae. The pupariation rate of larvae fed on HSD (20% sucrose) were compared with the log-rank test (χ² = 88.6, P < 0.0001). (H) Intestinal lipid levels are reduced in larvae with EE-specific CtBP deficiency (Pros-GAL4>CtBP RNAi), quantified as relative Oil Red O intensity in (I), unpaired t-test, ****P < 0.0001. (J) Loss of Dh31, Tk and AstA function in EE cells (Pros-GAL4> RNAi) leads to sugar intolerance of Drosophila larvae. The pupariation rate of larvae fed on HSD (20% sucrose) were compared using a log-rank test with the adjustment for multiple comparisons (χ² = 59.4, P < 0.0001). (K) Intestinal lipid levels are reduced in larvae with EE-specific Dh31-deficiency (Pros-GAL4>Dh31 RNAi), quantified as relative Oil Red O intensity in (L), unpaired t-test, ****P < 0.0001.

Figure 3.. Nicotinamide redox cofactors control EE cell function

(A) The ratiometric SoNar and iNap sensors were expressed by Pros-GAL4^ts driver to monitor the NAD⁺/NADH and NADPH/NADP⁺ ratios in EE cells. GAL4 expression was activated 16 h prior EE cell redox state measurements. EE cells of HSD fed larvae showed elevated reduction of NAD⁺ and NADP⁺ in EE cells when compared to those of 2-deoxyglucose (2-DG)-fed animals. The data were analyzed using the Mann-Whitney test, with ****P < 0.0001. (B) Feeding larvae with nicotinamide riboside (NR), a precursor of nicotinamide adenine dinucleotide, increases the proportion of EE cells (Pros⁺) in the Drosophila intestine. Data were analyzed using an unpaired t-test (****P < 0.0001). NR feeding also increases the number of Dh31-positive cells, with statistical significance determined by unpaired t-tests (*P = 0.0149; ***P = 0.0008). (C) EE cell specific knockdown (Pros-GAL4) of Naprt, a critical enzyme in NAD biosynthesis, leads to sugar intolerance. Pupariation rates of control and experimental genotype on HSD were compared using the log-rank test (χ² = 64.2, P < 0.0001). (D) EE cell specific knockdown (Pros-GAL4) of Naprt reduces relative EE cell numbers in the Drosophila intestine as well as (E) number of Dh31, AstA and Tk positive cells. The data were analyzed using unpaired t-test, *P < 0.03; ****P < 0.0001. (F) Transgenes expressing the bacterial enzymes LbNOX and TPNOX, which oxidize NADH and NADPH in EE cells, respectively, impair larval development on HSD. Pupariation rate is analyzed using the log-rank test, revealing significant effects for both LbNOX (χ² = 43.8, P < 0.0001) and TPNOX (χ² = 54.1, P < 0.0001).

Figure 4.. CtBP dimerization is regulated by metabolic pathways affecting NAD(P)H redox balance

(A) Schematic presentation of the BiFC system to analyze CtBP homodimerization by the activity of metabolic pathways or enzymes influencing the redox balance of nicotinamide cofactors. GAL4 expression (Pros-GAL4^ts) was activated 16 h prior measurement for all BiFC experiments. (B) CtBP dimerization as measured by BiFC reporters expressed in EE cells (Pros-GAL4^ts>CtBP-CC^OE:CtBP-VN^OE), EE cells marked by anti-Pros antibodies. (C) CtBP dimerization (number of BiFC positive cells within EE cell population) is inhibited by LbNOX and TpNOX overexpression, which oxidize NADH and NADPH, respectively. The data are analyzed using ordinary one-way ANOVA, ****P < 0.0001. ^nsindicates no significant difference in the studied parameter. (D) CtBP homodimerization in EE cells (BiFC) is regulated by sugar catabolism and nicotinamide metabolism. A diet containing 2-deoxyglucose (DG) was used to inhibit glucose catabolism. Diets with NR and AG1, a small molecule activator of the enzyme glucose-6-phosphate dehydrogenase (G6PD), were used to modulate the metabolic state. The data are analyzed using one-way ANOVA with Kruskal-Wallis correction for not normal distribution, with *P=0.0143, ***P = 0.0002, ****P < 0.0001. ^nsindicates no significant difference in the studied parameter. (E) Glucose catabolism through glycolysis and the pentose phosphate pathway activates CtBP homodimerization in the EE cells. The enzymes of glycolysis (phosphofructokinase, pfk) and the pentose phosphate pathway (G6PD, Zw) were depleted in EE cells. The data are analyzed using ordinary one-way ANOVA, with *P=0.0264, ****P < 0.0001. (F) Depletion of pfk and Zw in larval EE cells leads to sugar intolerance. Larval development on HSD (30% sucrose) was fully impaired with pfk knockdown (Pros-GAL4>) (χ² = 128.0, P < 0.0001) and modestly delayed by Zw knockdown (χ² = 12.1, P = 0.0004), as determined by the log-rank test. (G) Knockdown of pfk and Zw (Pros-GAL4>) in larvae fed on moderate (MSD, 5% sucrose) or high sugar diet (HSD, 30% sucrose) reduces the proportion of EE cells in the larval intestine. These genetic manipulations also decrease populations of Dh31-positive (H) and Tk-positive (I) cells (unpaired t-test, with **P < 0.01, ***P < 0.001).

Figure 5.. CtBP genomic targets regulate EE cell size and fate

(A) Schematic overview presenting the principle of targeted DamID used to identify the genome-wide target loci of chromatin-binding proteins. For determining its genomic binding, CtBP was fused with DNA adenine methyltransferase (Dam) and expressed in EE cells. (B) CtBP genomic targets include regulators of fate and growth regulating signaling pathways. CNET plot showing gene overlaps among selected pathways of CtBP targets identified by TaDa. Blue-colored nodes represent CtBP target genes within the pathway network. (C) Genetic screen of selected CtBP targets depleted in EE cells (Pros-GAL4>RNAi) reveals regulators of EE cell size and sugar tolerance. The plot shows the effect of the knockdown on EE cell size (X-axis) and sugar intolerance (Y-axis), calculated as the Log2FC of pupariation index of each genotype on LSD vs. HSD (30% sucrose). (D) CtBP binding to the genomic locus encoding Insulin-like receptor (InR), identified by TaDa. (E) Knockdown of InR (Pros-GAL4>RNAi) reduces EE cell size (unpaired t-test, ****P < 0.0001) and (F) delays larval development (log-rank test, χ² = 39.3, P < 0.0001) on HSD (30% sucrose). (G) Transcriptional regulators Split ends (spen) and (H) Mastermind (mam) are direct targets of CtBP. (I) Depletion of spen in EE cells (Pros-Gal4>RNAi) results in sugar intolerance. Pupariation rate of larvae with EE-specific spen knockdown and its control was analyzed using the log-rank test (χ² = 80.6, P < 0.0001). (J) Depletion of spen in EE cells (Pros-Gal4>RNAi) decreases the proportion of EE cells in Drosophila intestine, unpaired t-test (****P < 0.0001). (K) Knockdown of mam in EE cells (Pros-Gal4>RNAi) causes sugar intolerance (log-rank test, χ² = 64.0, P < 0.0001), but (L) does not affect the proportion of EE cells in larval intestine, unpaired t-test, ^ns indicates P > 0.05. (M) Knockdown of spen in EE cells (Pros-Gal4>RNAi) decreases Dh31 cell numbers and increases Tk cell numbers, lowering the ratio of Dh31/Tk-positive cells (*P = 0.03, **P = 0.0016). (N) Knockdown of mam in EE cells (Pros-Gal4>RNAi) increases Dh31 cell numbers (***P = 0.001) with no significant effect (^ns) on Tk-positive EE populations, increasing the ratio of Dh31/Tk-positive cells (***P = 0.001). The data are analyzed with unpaired t-test.

Figure 6.. CtBP interacts with Prospero in EE cells

(A) CtBP shares >30% of its genomic targets with Pros in EE cells. Comparison of genomic target loci of CtBP and Pros, identified by EE cell-specific TaDa. (B) Common targets of CtBP and Prospero include many regulators of the mTOR and Notch pathways. CNET plot showing gene overlaps in the mTOR and Notch pathways among CtBP and Pros targets identified by TaDa. Blue-colored nodes represent CtBP-specific targets, while purple-colored nodes indicate genes co-regulated by CtBP and Pros. (C) Conserved substrate-binding cleft mediates interaction between Drosophila CtBP (dCtBP) and Prospero. AlphaFold2 multimer model of dCtBP (teal) bound to the Prospero ALSLV motif (residues 1284–1288, orange) is shown superimposed with crystal structures of human CtBP1 (hCtBP1, grey) bound to a canonical PIDLS motif (beige, PDB ID: 1HL3; Nardini et al., 2003), and human CtBP2 (hCtBP2, cyan) bound to a non-canonical ALDLS motif from RAI2 (PDB ID: 8ATI, cyan, Goradia et al., 2024). The zoomed-out view displays the structural alignment of substrate-binding domains across dCtBP, hCtBP1, and hCtBP2, highlighting motif engagement within the conserved PxDLS-binding cleft. The zoomed-in view emphasizes the structural conservation of the CtBP binding pocket and the similar configurations adopted by ALSLV, PIDLS, and ALDLS motifs, despite sequence differences. A computational estimate of the change in free binding energy (ΔΔG) identifies a phenylalanine-to-alanine mutation at position 53 in dCtBP (CtBPF53A) as a key substitution predicted to significantly reduce binding (see also Supplementary Figure 3C). Proteins are shown in cartoon representation, interacting residues as sticks, and NAD as spheres. (D) Co-immunoprecipitation of ectopically expressed FLAG-CtBP (CtBP^WT) and HA-Pros (Pros^WT) in Drosophila S2 cells confirms the ability of CtBP to interact with Pros. The CtBP-Pros interaction is significantly reduced by the CtBP F53A mutation, consistent with structural predictions. (E) Quantification of the data is displayed as a percentage of Co-IP relative to CtBP^WT and normalized to the input (unpaired t-test *P < 0.03). (F) CtBP interacts with Pros in EE cells as revealed by the Bimolecular Fluorescence Complementation (BiFC) assay. (G) CtBP-Pros interaction is significantly inhibited by 2-DG feeding, but in contrast to CtBP-CtBP interaction, it remains unaffected by NR feeding. Data are analyzed using the one-way ANOVA with Kruskal-Wallis correction for not normal distribution, with *P < 0.02. ns indicates no significant difference in the studied parameter.