TIGAR regulates intestinal mucus barrier integrity by inhibiting lactylation of G6PD/6PGD in ulcerative colitis

Abstract

Oxidative stress and metabolic dysregulation in goblet cells are pivotal in ulcerative colitis (UC) pathogenesis. TIGAR promotes the synthesis of NADPH and contributes to mitigate oxidative stress, but how it regulates NADPH production and affects UC remains unclear. Here we demonstrate that TIGAR inhibits lactylation of the key NADPH-synthesizing enzymes G6PD (at K432) and 6PGD (at K38), thereby preserving their enzymatic activities by promoting G6PD homodimer formation and 6PGD binding to NADP⁺. In male UC mice, persistently low TIGAR expression elevates lactate levels, promoting the lactylation of G6PD and 6PGD and impairing their function. This process suppresses NADPH synthesis, exacerbating goblet cell oxidative stress. The resulting decline in Trx1 reductase activity induces S-nitrosylation of the mucin-processing enzyme AGR2, thereby inhibiting mature MUC2 production and compromising the intestinal mucus barrier. Our findings elucidate a mechanistic pathway through which TIGAR maintains cellular redox homeostasis, presenting it as a potential therapeutic target for UC.

This study reveals that reduced TIGAR expression in ulcerative colitis promotes G6PD/6PGD lactylation, inhibiting NADPH production, and promotes Trx1 and AGR2 S-nitrosylation, impairing mature MUC2 synthesis and damaging the intestinal mucus barrier.

Introduction

Ulcerative colitis (UC) is a chronic, nonspecific inflammatory bowel disease (IBD) of unknown etiology, primarily affecting the distal colonic mucosa and submucosa. Characterized by a prolonged and often recurrent course, UC significantly impacts patients’ quality of life and can lead to colitis-associated cancer^1,2. UC is associated with various factors, including the intestinal environment, immune response, inflammation, genetics, diet, and gut microbiota, etc^3,4. Recently, the relationship between the intestinal mucus barrier and UC has garnered increasing attention^5,6. Intestinal mucus serves as the first line of defense against the invasion of harmful factors and plays a crucial role in maintaining intestinal homeostasis^7,8. Current research primarily focuses on the effects of UC on intestinal mucus synthesis, mucus constituents, and the glycosylation modifications of mucins^9,10. However, whether UC affects the critical issue of converting mucin precursors to their mature forms remains unclear.

In the endoplasmic reticulum, the protein modification enzyme anterior gradient protein 2 (AGR2) plays a crucial role in the maturation of MUC2^11,12. Both basic and clinical studies have confirmed that abnormal expression or dysfunction of AGR2 can impair mucins synthesis and modification, which are essential factors in the pathogenesis of UC^11,12. However, there are no reports on whether UC affects the enzymatic activity of AGR2. The activity of AGR2 depends on its redox balance, and oxidative stress can reduce AGR2’s activity, thereby impeding mucin synthesis and modification¹³. Oxidative stress in goblet cells is widely recognized as a significant contributor to mucin modification disorders and the onset of UC, making antioxidative strategies crucial for UC therapy^14,15. Additionally, abnormal cellular metabolism constitutes another key factor in UC. Studies have shown that intestinal epithelial cells in animal models and patients with UC exhibit a distinctive metabolic pattern characterized by reduced oxygen consumption, increased glucose utilization, and elevated lactate production¹⁶. This suggests that UC may induce metabolic reprogramming in intestinal epithelial cells, shifting glucose metabolism from oxidative phosphorylation to glycolysis¹⁷. It is now established that lactate can interfere with cellular metabolism and induce the lactylation of certain metabolic enzymes¹⁸. Recently, the lactylation of non-histone proteins has emerged as a focal point of research in this field^19,20. In this context, the TP53-induced glycolysis and apoptosis regulator (TIGAR) has gained attention for its regulatory effects on glucose metabolism and oxidative stress^21,22.

TIGAR exhibits phosphatase activity, catalyzing the dephosphorylation of fructose-2,6-bisphosphate to generate fructose-6-phosphate, which can be converted into glucose-6-phosphate. The latter then enters the pentose phosphate pathway (PPP) to synthesize reduced coenzyme II (NADPH)²³. NADPH is one of the most crucial reducing agents in the body, providing hydrogen essential for maintaining the stability of various redox systems, including the glutathione and thioredoxin systems²⁴. In living organisms, more than 60% of NADPH is synthesized via the PPP, where glucose-6-phosphate dehydrogenase (G6PD) and 6-phosphogluconate dehydrogenase (6PGD) serve as the rate-limiting enzymes regulating NADPH production^25,26. TIGAR’s role in promoting glucose entry into the PPP and facilitating NADPH synthesis has been widely recognized. However, the relationship between TIGAR and these two rate-limiting enzymes remains underexplored. Only one study has demonstrated that TIGAR can promote the synthesis of G6PD, but did not involve its enzymatic activity²⁷. Furthermore, no investigation has been reported into whether TIGAR influences 6PGD activity. Given TIGAR’s roles in inhibiting glycolysis and reducing lactate synthesis, we hypothesize that TIGAR may regulate the activities of these two enzymes through lactylation modification, thereby impacting NADPH synthesis.

Our findings revealed that TIGAR significantly inhibits the lactylation of both G6PD and 6PGD, and identified the lactylation sites at lysine residues 432 and 38 in their respective amino acid sequences. Low expression of TIGAR was found to be a key factor contributing to the enhanced lactylation of G6PD and 6PGD, resulting in decreased NADPH synthesis, induced S-nitrosylation of AGR2, impaired MUC2 maturation, and subsequent damage to the intestinal mucus barrier in male UC mice. This study establishes TIGAR as a critical molecule that influences MUC2 modification and mucus integrity by affecting cellular redox homeostasis and glucose metabolism. These findings provide experimental evidence supporting that TIGAR is a potential therapeutic target for UC.

Results

Intestinal TIGAR deletion aggravates colitis via oxidative stress in UC mice

The symptoms of DSS-induced colitis in mice closely resemble those of human ulcerative colitis²⁸. Following 7 days of DSS induction, the HE staining results demonstrated that the crypt interstice was markedly widened (Supplementary Fig. 1a), and the crypt height was significantly reduced (Supplementary Fig. 1c). Correspondingly, the UC mice showed more severe disruption of the intestinal mucosal epithelium barrier accompanied by increased inflammatory cell infiltration (Supplementary Fig. 1b). AB-PAS staining indicated reduced neutral and acidic mucus synthesis (Supplementary Fig. 1a, d, g). We utilized Cy3-conjugated EUB338 probes, which specifically target bacterial DNA, and the MUC2-N antibody, which recognizes mature MUC2, to evaluate the damage to the mucus barrier in UC mice²⁹. Both the proximal and distal colons of UC mice exhibited damage to the inner mucous layer (Supplementary Fig. 1e, f), thinning the outer mucous layer, and bacteria migrated to the inner mucous layer (Supplementary Fig. 1e, f). Additionally, we observed that TIGAR expression began to decrease from day 5 and showed a more pronounced reduction by day 7 (Supplementary Fig. 1h). Therefore, our results suggest that reduced TIGAR expression in colon tissue may play a significant role in the pathogenesis of ulcerative colitis.

To investigate the relationship between TIGAR and UC, we generated mice with conditional knockout of TIGAR in the intestinal tract (TIGAR^f/fVil1-Cre) by crossing TIGAR^f/f mice with Vil1-Cre mice, enabling conditional TIGAR knockout in the gut system (Fig. 1j), and verified by western blot (Fig. 1l). Following the same procedure of DSS administration, the TIGAR^f/fVil1-Cre mice exhibited symptoms including mucopurulent1 bloody stool, highly distorted crypts, increased intercryptal spaces, and decreased intestinal mucus synthesis (Fig. 1a, b). These changes appeared two days earlier and were associated with more severe pathological damage in TIGAR^f/fVil1-Cre mice compared to TIGAR^f/f mice (Fig. 1a), suggesting that TIGAR deletion accelerated the onset and progression of UC.

Fig. 1. Intestinal TIGAR deletion aggravates colitis via oxidative stress in UC mice.

a, b TIGAR^f/f and TIGAR^f/fVil1-Cre mice were administered 2.5% DSS for 5 and 7 days, and the distal colon tissue slices were obtained. Colon H&E staining (left) and PAS/Alcian blue staining (right). Scale bar: 100 μm or 50 μm (a). Immunostaining was performed with an anti-MUC2-C (green) antibody, and FISH was performed with a bacterial 16S rRNA gene probe (red). The white vertical lines indicate the thickness of the mucus. Scale bar: 50 μm (b). c–i TIGAR^f/f and TIGAR^f/fVil1-Cre mice were fed with 2.5% DSS for 5 days, GSH level (c), GSH/GSSG ratio (d), NADPH level (e), NADPH/NADP⁺ ratio (f), Trx1 reductase activity (g), TrxR activity (h), the concentration of NO (i) were detected in colonic tissue lysates (n = 5 biological replicates). j Protocol for creating gut-specific TIGAR knockout mice (TIGAR^f/fVil1-Cre). k Immunoblot of Trx1 and TrxR levels in TIGAR^f/f and TIGAR^f/fVil1-Cre mice colonic tissue. l, m The AGR2, CHOP, GRP78, and TIGAR (l), MUC2 precursor and mature MUC2 (m) levels in colon tissue lysates were detected by western blotting. n–q HT-29 CL.16E cells were cultured for 12 h in the presence or absence of TNF-α (100 ng/ml). Fluorescence intensity was measured by flow cytometry using DCFH-DA dye to detect ROS content change (n). The DAF-AMDA fluorescent probe was used to detect changes in NO content (o). A MitoSox probe was used to assess MitoROS changes by flow cytometry (p). A dihydroethidium (DHE) probe was used to detect changes in superoxide (q). n = 6 biological replicates. Data presented are representative of three independent experiments and shown as mean ± SD, and p-values were determined by Mann-Whitney U test for c–f, two-sided unpaired Student’s t-test for (g–i), and one-way ANOVA with Tukey’s HSD post hoc test for (n–q).

Oxidative stress induced by DSS damages the mucus barrier³⁰. TIGAR is an important glucose-regulating enzyme that promotes NADPH synthesis and inhibits oxidative stress³¹. To evaluate TIGAR’s impact on the colonic mucus barrier, we measured oxidative stress markers (GSH, NADPH, NO) and activities of Trx1 and TrxR in colon tissues of TIGAR^f/fVil1-Cre mice treated with DSS. Compared to TIGAR^f/f mice, GSH content, GSH/GSSG ratio, NADPH content, NADPH/NADP⁺ ratio, and Trx1 reductase activity were significantly reduced (Fig. 1c–g), while NO levels were elevated (Fig. 1i). TrxR1 activity and protein levels of Trx1 and TrxR1 remained unchanged (Fig. 1h, k). Oxidative stress can induce the misfolding of ER proteins³², and GRP78 and CHOP levels were significantly elevated in colon tissues of TIGAR^f/fVil1-Cre mice (Fig. 1l). ER stress also caused MUC2 precursor misfolding, leading to their accumulation and reduced mature MUC2 synthesis (Fig. 1m). We established HT29 CL.16E cells with stable TIGAR knockout using CRISPR/Cas9 (Supplementary Fig. 2i). In vitro experiments mirrored the findings in TIGAR^f/fVil1-Cre mice (Supplementary Fig. 2a–i). TIGAR knockout in HT29 CL.16E cells increased ROS, NO, mitochondrial ROS, and superoxide levels (Fig. 1n–q). These results suggest that TIGAR deletion exacerbates intestinal oxidative stress and disrupts MUC2 maturation in UC mice.

TIGAR deficiency accumulate 6-phosphogluconate and lactate by reprograming the PPP and glycolysis

TIGAR has long been considered to primarily function as a fructose-2,6-bisphosphatase, which suppresses glycolysis and subsequently promotes the flux of the PPP to generate NADPH to reduce intracellular ROS²¹. However, there remains a lack of understanding regarding the specific mechanisms by which TIGAR regulates the distribution of metabolic flux between glycolysis and the PPP. For this reason, we first conducted untargeted metabolomics to investigate the static changes of metabolites. We found a dramatic increase in 6-phosphogluconate (6-PG) levels in TIGAR-knockout HT29 CL.16E cells (Fig. 2a), which was identified by the mzVault library (Fig. 2b) and comparing the retention time of the standard. Then, we also observed a consistent increase in 6-PG in TIGAR-knockout cells (HT29 CL.16E and HIEC-6) and colon tissues from TIGAR^f/fVil1-Cre mice, with DSS induction further exacerbating the elevation of 6-PG (Fig. 2c). At the same time, our research revealed elevated lactate levels in both TIGAR-knockout cell lines and colon tissues from TIGAR^f/fVil1-Cre mice and DSS-induced TIGAR^f/fVil1-Cre mice (Fig. 2d). The KEGG enrichment analysis of metabolite revealed that the PPP and glycolysis were both affected pathways following TIGAR-knockout in HT29 CL.16E cells (Supplementary Fig. 3a). These results suggest that 6-PG and lactate, key metabolites in the PPP and glycolysis, may have increased synthesis or suppressed catabolism without TIGAR. Furthermore, DSS induction appears to exacerbate this metabolic imbalance. To investigate the specific mechanisms underlying the reprogramming of glucose metabolism following TIAGR depletion, we need to perform isotope tracing analysis. First, we must ascertain whether the absence of TIGAR affects glucose uptake in cells or animals. Notably, our results indicated that the absence of TIGAR does not influence cellular glucose uptake (Supplementary Fig. 3b, c). The glucose oxidation rate appears to remain comparable between TIGAR^f/fVil1-Cre and TIGAR^f/f mice, as inferred from analyses of oxygen consumption and respiratory quotient (RQ) (Supplementary Fig. 3d–f)^33,34.

Fig. 2. TIGAR deficiency accumulated 6-phosphogluconate and lactate by reprogramming the PPP and glycolysis.

a Volcano plot displaying metabolites differentially expressed upon TIGAR knockout in HT29 CL.16E cells (FC > 2, and two-sided Student’s t-test, P < 0.05). b Comparison of the mass spectra in negative ion mode of the unknown with m/z = 275.0175 Da (top) and 6-PG from the mzVault library (bottom). c, d Untargeted metabolomics revealed relative amounts of 6PG and lactate from different samples: HT29 CL.16E cells (n = 3 biological replicates); HIEC-6 cells (n = 6 biological replicates). the colon tissue (n = 3 biological replicates). e Schematic diagram of a multi-timepoint stable isotope tracing experimental protocol. f The fractional contribution of ¹³C₆-glucose to intermediates of the glycolysis and PPP in HT29 CL.16E cells upon sgTIGAR compared to control after 2 h of ¹³C₆-glucose pretreatment (n = 4 biological replicates). g, h Dynamic ¹³C₆-glucose labeling of HT29 CL.16E cells showing the fraction (left) and the mean of relative amount (right) of ¹³C incorporation into 6-PG (g) and lactate (h). ‘M + n’ equals the molecular mass plus the number of incorporated heavy carbons (n = 4 biological replicates). i Schematic representation of glycolysis and PPP, RNA-seq data of colon from the TIGAR^f/f and TIGAR^f/fVil1-Cre mice (n = 3 biological replicates). j The fractional contribution of ¹³C₆-glucose to intermediates of the glycolysis and PPP in the colon from TIGAR^f/fVil1-Cre mice compared to control after 1 h of ¹³C₆-glucose pretreatment (n = 4 biological replicates). The first and second data sets in panel (c), as well as F1,6 P, DHAP, and 3-PG in panel (f), and DHAP, 3-PG, PEP, lactate, and R5P in panel (j), were compared using two-sided unpaired Welch’s t-test. 6PG in panel (j) was analyzed using the Mann–Whitney U test. One-way ANOVA with Tukey’s HSD post hoc test was applied to the fourth data set in panel c and the fourth data set in panel (d). Data in panels g (left) and h (left) were analyzed by two-way repeated-measures ANOVA. All other comparisons were performed using a two-sided Student’s t-test.

Then, we employed U-¹³C-glucose for stable isotope tracing to analyze metabolic flux changes within the glycolysis and PPP (Fig. 2e). Following incubation with U-¹³C-glucose 2 h, we observed a reduced carbon flux through the PPP and increased carbon flux through glycolysis upon TIGAR-knockout in HT29 CL.16E cells, with the fractions of ¹³C-labeled glucose 6-phosphate (G6P), fructose-6-phosphate (F6P), 2-phosphoglycerate (2-PG), pyruvate (PYR), and lactate accumulated in TIGAR-deficient. In contrast, the fractions of ¹³C-labeled 6-PG and ribose 5-phosphate (R5P) in the PPP pathway were remarkably decreased (Fig. 2f). Next, we observed the temporal patterns of isotope incorporation into metabolites multiple times points in cells. The isotope exchange rate for 6-PG was lower in TIGAR-deficient cells compared to the control group. At the same time, the total 6-PG (labeled and unlabeled) levels remained consistently higher in TIGAR-deficient cells (Fig. 2g). This suggested that both the production and consumption of 6-PG were inhibited in TIGAR-deficient cells, with the inhibition of the consumption pathway being more pronounced. In contrast, lactate exhibited a different pattern. While the total lactate levels were higher in TIGAR-deficient cells compared to controls, the isotope exchange rate of lactate was also elevated, indicating that the production pathways of lactate are activated in TIGAR-deficient cells (Fig. 2h). To confirm the consistency of this metabolic reprogramming pattern caused by TIGAR deficiency in intestinal epithelial cells, we validated the findings in HIEC-6 cells and the colons of the TIGAR^f/fVil1-Cre mice. The results were consistent with those observed in HT-29 cells (Fig. 2j, Supplementary Fig. 3g–o).

To further investigate the underlying mechanism of this metabolic reprogramming induced by TIGAR deficiency, we performed RNA sequencing on the colons of the TIGAR^f/f and TIGAR^f/fVil1-Cre mice. RNA sequencing analysis demonstrated that there were 1261 differentially expressed genes (p-adjust<0.05, FC > 1.2). KEGG pathway enrichment analysis indicated that these differentially expressed genes were mainly enriched in pathways metabolism, human diseases, cellular processes, etc (Supplementary Fig. 4a). Given that TIGAR is a gene involved in the regulation of cellular metabolism, we analyzed the differentially expressed genes within the metabolism module, which consisted of a total of 107 genes. The absence of TIGAR was found to upregulate the expression of multiple glycolysis-related genes, including Hk2, Pkm, and Ldha (Supplementary Fig. 4b). Gene set enrichment analysis (GSEA) revealed substantial alterations in glycolysis and gluconeogenesis pathways associated with TIGAR deficiency (Supplementary Fig. 4c, d). An increase in lactate flux can be attributed to the upregulation of metabolic enzymes involved in glycolysis at the RNA level (Supplementary Fig. 4b). However, the RNA levels of metabolic enzymes upstream and downstream of 6-PG in RNA-seq showed no significant changes with TIGAR deficiency, including G6PD and PGD (Supplementary Fig. 4b). Additionally, no significant differences were observed in the protein levels of G6PD and 6PGD (Supplementary Fig. 4e–g). Therefore, we speculate that the accumulation of 6-PG may be related to changes in G6PD and 6PGD enzymatic activity.

TIGAR deficiency exacerbates oxidative stress by promoting the lactylation of G6PD and 6PGD

We next investigated the activities of G6PD and 6PGD, and their enzymatic activities were significantly reduced in TIGAR^f/fVil1-Cre mice (Supplementary Fig. 5a, b). Specifically, the dimerization of G6PD was markedly decreased in TIGAR knockout cells and mice (Supplementary Fig. 5c, d), whereas the dimerization of 6PGD remained stable (Supplementary Fig. 5e). These findings suggested that TIGAR regulated the activity of these two enzymes, potentially through distinct mechanisms, thereby impacting NADPH synthesis. Our research revealed elevated lactate levels in both colon tissues from UC mice and TIGAR^f/fVil1-Cre mice and in TNFα-induced HT29 CL.16E and HIEC-6 cells (Fig. 2d). Given lactate’s role as a substrate for protein lactylation³⁵, we hypothesized that TIGAR knockout induces lactylation of G6PD and 6PGD. Pan-Kla levels were elevated in UC mice (Fig. 3a, b), TNFα-induced cells (Fig. 3c and Supplementary Fig. 5h–j), TIGAR^f/fVil1-Cre mice colon tissues (Fig. 3a, b), and TIGAR-knockout HT29 CL.16E cells (Fig. 3c and Supplementary Fig. 5j). Western blot analysis showed significantly elevated lactylation of G6PD and 6PGD in DSS-induced mouse colon tissues (Fig. 3d), with similar results in cell experiments (Fig. 3e).

Fig. 3. TIGAR deficiency exacerbates oxidative stress by promoting the lactation of G6PD and 6PGD.

a, b The changes in pan-lactylation in the colonic lysates were detected by western blot. c TIGAR knocked out HT29CL.16E cells were treated with TNF-α (100 ng/ml) for 12 h. Pan-lactylation changes were analyzed by Western blot. d The lactylation of 6PGD and G6PD was detected using pan-anti-lactyllysine antibody. e, f The lactylation levels of G6PD and 6PGD were analyzed by immunoprecipitation (e), the level of a dimer of G6PD was detected by western blotting (f). g Cartoon representation of G6PD structure. Dual G6PD monomers are stacked into a dimer. h Cartoon representation of the crystal structure of G6PD bound to NADP⁺. i Representative image of the structure of human 6PGD bound to NADP⁺ made using Pymol database. j The lactylation of G6PD were detected using a pan-anti-lactyllysine antibody in cells cultured with or without lactate (10 μM) for 12 h. k Cibacron Blue PD were used to detect the binding of G6PD WT and mutants with NADP⁺. l The dimer of G6PD was investigated using a flag antibody in HT29 CL.16E cells. m The activity of G6PD was detected in cells cultured with or without lactate. n = 5 biological replicates. n The activity of G6PD was assessed in cells cultured with or without TNF-α and FX11 (20 μM, 24 h). n = 5 biological replicates. o up: The lactylation levels of 6PGD were detected in HT29 CL.16E cells. down: Cibacron Blue PD were used to assess the binding of 6PGD WT and mutants with NADP⁺. p The dimer of 6PGD was investigated using a flag antibody in HT29 CL.16E cells. q The activity of G6PD was detected in HT29 CL.16E cells. n = 5 biological replicates. r The activity of G6PD was assessed in cells cultured with or without lactate and FX11. n = 5 biological replicates. s, t The G6PD or 6PGD knockdown HT29 CL.16E cells with re-expression of G6PD (s) or 6PGD (t) mutations were used to detected the mature MUC2 by western blot. Data presented are representative of three independent experiments and shown as mean ± SD, p-values were determined by one-way ANOVA with Tukey’s HSD post hoc test for (m, n, q, r).

To investigate the effect of lactylation on mucus synthesis in G6PD and 6PGD, we conducted proteomics analysis in HT-29 cells. LC-MS revealed lactylation sites at G6PD K432 and 6PGD K38 (Supplementary Fig. 5f, g). G6PD and 6PGD activity depends on dimerization and NADP⁺ binding. TIGAR-knockout reduced G6PD dimer formation in HT29 CL.16E cells (Fig. 3f). Molecular docking simulations predicted lactylation effects on their activity. G6PD K432 was near the dimer interface, forming a hydrogen bond with M411 of the opposing monomer, suggesting that lactylation may disrupt dimer formation (Fig. 3g). Previous studies reported that lactylation at K45 hinders G6PD dimer formation and possibly affects NADP⁺ binding³⁶. However, our results showed that K45 was close to the NADP⁺ binding domain, likely affecting NADP⁺ binding rather than dimerization (Fig. 3h). 6PGD K38, located in the catalytic NADP⁺ binding domain, forms a hydrogen bond with NADP⁺, indicating that its lactylation may affect NADP⁺ binding (Fig. 3i). Thus, the lactylation of 6PGD K38 may affect the binding to NADP⁺. We explored the impact of lactylation on dimer formation and NADP⁺ binding by mutating specific sites. WT and G6PD mutant alleles were expressed in G6PD knockout HT29 CL.16E cells using the same sense mutation. Mutations at K45 and K432 inhibited G6PD lactylation (Fig. 3j), K45 primarily affected NADP⁺ binding (Fig. 3k). While K432R impacted dimer formation (Fig. 3l). This was due to their locations: K432 was in the dimerization domain (Fig. 3g), and K45 was in the NADP⁺-binding domain (Fig. 3h). Moreover, lactate was found to inhibit the activity of G6PD (Fig. 3m). However, when the lactylation site K432 was mutated, the enzymatic activity was restored (Fig. 3m). We utilized FX11 to inhibit LDH activity and observed that FX11 was capable of restoring G6PD activity (Fig. 3n). Mutating K38 of 6PGD also inhibited its lactylation (Fig. 3o) and hindered NADP⁺ binding without affecting dimer formation (Fig. 3o, p). Similarly, treatment with lactate inhibited the activity of 6PGD, however, when the lactylation site K38 was mutated, the enzymatic activity was restored (Fig. 3q). Following the inhibition of LDH, the activity of 6PGD was restored (Fig. 3r). Inhibiting lactylation of G6PD and 6PGD promotes MUC2 maturation (Fig. 3s, t). Subsequently, we generated G6PD or 6PGD-deficient HT29 CL.16E cells using the CRISPR-Cas9 method and found that G6PD or 6PGD-deficiency increased ROS, NO, and superoxide (Supplementary Fig. 5k–p). Re-expressing wild type or mutant G6PD/6PGD decreased these levels induced by TIGAR knockout (Supplementary Fig. 5q–s), with the mutants showing a more pronounced effect (Supplementary Fig. 5q–s).

TIGAR deletion leads to AGR2 S-nitrosylation and inhibits MUC2 maturation

Oxidative stress reduces intestinal mucus synthesis, and AGR2 plays a crucial role in the synthesis and modification of MUC2¹¹. The synthesis of mature MUC2 was reduced in the colon tissue of UC mice (Supplementary Fig. 6a). We utilized the MUC2-VNTR antibody, which targets immature MUC2²⁹, to investigate its interaction and cellular co-localization with AGR2. The interaction between AGR2 and the MUC2 precursor has been validated through immunoprecipitation (Supplementary Fig. 6b, c), and their co-localization has been confirmed by immunofluorescence analysis of cell and colon tissue (Supplementary Fig. 6d–f). We then generated AGR2-knockout HT29 CL.16E cells and measured precursor MUC2 and mature MUC2 levels. The results demonstrated that AGR2 knockout led to an accumulation of precursor MUC2, whereas the level of mature MUC2 was reduced (Supplementary Fig. 6g). Our previous studies have confirmed that AGR2 S-glutathione modification influences mucus processing and synthesis¹³. We measured total protein S-glutathione levels in the colon tissue of UC mice. Although total protein S-glutathione showed a slight increase, this change was not statistically significant (Fig. 4a). This indicates that other posttranslational modifications may affect the function of AGR2 in UC mice.

Fig. 4. TIGAR deletion leads to AGR2 S-nitrosylation and inhibits MUC2 maturation.

a–d C57BL/6 N mice were administered 2.5% DSS for 5 or 7 days. Western blotting was used to detect the changes in S-glutathionylation of proteins in colon tissues (a), the concentration of NO, n = 5 mice (b), and biotin switch assay was performed to investigate the changes in S-nitrosylation of proteins in the colon tissues after 5 (c) or 7 (d) days DSS feeding (protein sample not treated with biotin-HPDH was used as a negative control to demonstrate that S-nitrosylation proteins were successfully labeled, the same method was employed for S-nitrosylation detection henceforth). e Assessment of S-nitrosylation in frozen colon tissue using biotin derivatization and analysis by confocal microscopy. Scale bar: 50 μm. f, g The cysteine thiol level of AGR2 in colon tissue of mice after 5 (f) or 7 (g) days of DSS administration. h The changes of SNO-AGR2 in the colon tissue of mice were detected by biotin switch assay after 7 days of DSS feeding. i, j The tissue lysate was incubated with MUC2-VTNR or AGR2 antibody conjugated to Thermo Scientific Pierce Protein A/G Magnetic Beads for 24 hours at 4 °C with gentle rotation. Subsequently, western blot analysis was performed using anti-MUC2-VTNR and anti-AGR2 antibodies. k, l AGR2 wild-type, and C81S mutant were expressed in HT29 CL.16E cell, and the alterations in the precursor (k) and mature (l) MUC2 were analyzed using western blot. m, n TIGAR^f/f (WT) and TIGAR^f/fVil1-Cre (KO) mice were administered 2.5% DSS for 5 days, after which the levels of total protein S-nitrosylation (m) and AGR2 S-nitrosylation (n) in colonic tissues were assessed using the biotin switch assay. Data presented are representative of three independent experiments and shown as mean ± SD, and p-values were determined by two-sided unpaired Student’s t-test for (b).

Previous studies have shown that changes in intracellular NO levels due to oxidative stress are linked to protein S-nitrosylation³⁷. Given the significant increase in NO levels in UC mice (Fig. 4b), we hypothesized and confirmed elevated S-nitrosylation of proteins in the colon tissues of UC mice (Fig. 4c–e). Additionally, AGR2 sulfhydryl groups were significantly reduced (Fig. 4f, g), while AGR2 S-nitrosylation was increased (Fig. 4h). AGR2 facilitates MUC2 maturation by forming disulfide bonds with MUC2 precursors via its Cys81 residue^11,38. Our results show that increased AGR2 S-nitrosylation in UC mice reduce its binding affinity to MUC2, as confirmed by immunoprecipitation (Fig. 4i, j). Mutation at Cys81 led to the accumulation of MUC2 precursors and a decrease in mature mucin synthesis (Fig. 4k, l). We also observed increased S-nitrosylation of proteins and AGR2 in TIGAR^f/fVil1-Cre mice (Fig. 4m, n). These findings indicate that AGR2 Cys81 S-nitrosylation inhibits its interaction with MUC2 precursors, disrupting MUC2 maturation. TIGAR promotes MUC2 maturation by preventing AGR2 S-nitrosylation and maintaining the mucus barrier.

TIGAR deletion facilitates the Trx1-mediated S-nitrosylation of AGR2

Trx1 is a key intracellular oxidoreductase involved in denitrosylation and transnitrosylation, with its function regulated by the intracellular redox state³⁹. In this study, we observed increased S-nitrosylation of Trx1 in the colon tissue of UC mice (Supplementary Fig. 7a), accompanied by reduced reductase activity and denitrosylation function (Supplementary Fig. 7b). Similar results were found in TNFα-induced HT29 CL.16E cells (Supplementary Fig. 7c, d). Overexpression of Trx1 increased SNO-protein levels, while knockdown decreased them (Supplementary Fig. 7e), indicating Trx1 primarily acts as an oxidase under inflammatory conditions. The redox status of Cys32 and Cys35 in Trx1 modulates its ability to S-nitrosylate or denitrosylate target proteins^39,40. TIGAR depletion increased Trx1 S-nitrosylation (Supplementary Fig. 7f–h), and redox electrophoresis showed disulfide bond formation in Trx1 Cys32-Cys35 and Cys62-Cys69 upon TNF-α treatment (Supplementary Fig. 7i), confirming loss of reductase activity. Trx1 was nitrosylated at Cys73, crucial for transnitrosylation⁴¹. Mass spectrometry confirmed S-nitrosylation at Cys73, a conserved site in mammals (Supplementary Fig. 7j, k). The NADPH regulates the activity of Trx1 oxidoreductase⁴². NADPH supplementation in TIGAR knockout cells reduced protein and Trx1 S-nitrosylation (Supplementary Fig. 7l, m), indicating TIGAR maintains Trx1 reductase activity by enhancing NADPH synthesis.

Trx1 transfers SNO to target proteins after S-nitrosylation at Cys73³⁹. The primary sequence motif analysis on peptides containing identified S-nitrosylation sites revealed that proteins harboring CXXXXA, CXXI, and EC motifs were more susceptible to S-nitrosylation, with Cys81 of AGR2 located within these motifs (Supplementary Fig. 8a–c). Trx1-mediated transnitrosylation depends on target protein interaction^39,41. Immunoprecipitation and immunofluorescence confirmed Trx1-AGR2 interaction (Fig. 5a, b). Trx1 knockdown reduced AGR2 S-nitrosylation (Fig. 5c and Supplementary Fig. 8d), while overexpression increased it (Fig. 5d and Supplementary Fig. 8e), suggesting Trx1 transfers nitroso to AGR2 in inflammation. Using plasmids with synonymous mutations, we first attenuated endogenous Trx1 in HT29 CL.16E cells using shRNA and subsequently expressed Flag-Trx1 (Supplementary Fig. 8f). Following Trx1 attenuation, both SNO-protein levels and AGR2 S-nitrosylation decreased (Supplementary Fig. 8g, h). In contrast, these S-nitrosylation levels increased upon Trx1 re-expression (Supplementary Fig. 8g, h). We found that mutating Cys32 or Cys35 reduced SNO-protein (Fig. 5e, g), SNO-Trx1 (Fig. 5h), and SNO-AGR2 (Fig. 5f, i), while simultaneous mutation of Cys32 and Cys35 increased their S-nitrosylation (Fig. 5e–i), and reduced the synthesis of MUC2 (Fig. 5i). Mutating Cys73 significantly reduced their S-nitrosylation levels (Fig. 5g–i) and increased the synthesis of MUC2 (Fig. 5i), indicating Cys32 and Cys35 are involved in denitrosylation, while Cys73 is critical for transnitrosylation. Next, we found that NADPH addition after TIGAR deletion or Trx1 transnitrosylation inhibition with PX12 reduced AGR2 S-nitrosylation and increased the synthesis of MUC2 (Fig. 5j). Exogenous NADPH had no effect on AGR2 S-nitrosylation and the synthesis of MUC2 after Trx1 knockdown or simultaneous mutation of Cys32 and Cys35 (Fig. 5k, l), indicating NADPH modulates AGR2 S-nitrosylation by altering Trx1 reductase activity. In summary, TIGAR reduces AGR2 S-nitrosylation by promoting NADPH synthesis, thereby maintaining Trx1 reductase activity and ultimately enhancing mucus synthesis.

Fig. 5. TIGAR deletion facilitates the Trx1-mediated S-nitrosylation of AGR2.

a The interaction between AGR2 and Trx1 was investigated through immunoprecipitation assays utilizing antibodies specific to AGR2 and Trx1 in HT29 CL.16E cells. b Immunofluorescence detected the co-localization of AGR2 and Trx1 in HT29 CL.16E cells. Scale bar: 10 μm. c, d The S-nitrosylation of AGR2 was examined in Trx1-knockdown (c) or Trx1-overexpressing (d) HT29 CL.16E cells following TNFα treatment (100 ng/mL, 24 h, the same method was employed for induction of the inflammatory state in cells henceforth). A protein sample not treated with biotin-HPDH was used as a negative control to demonstrate that S-nitrosylation proteins were successfully labeled (the same method was employed for S-nitrosylation detection henceforth). e, f Flag-Trx1, Flag-Trx1C32S, Flag-Trx1C35S, and Flag-Trx1C32/35S (simultaneous mutation on Cysteine 32 and 35) were expressed in HT29 CL.16E cells with endogenous Trx1 knocked down. After treatment with TNFα, the total protein S-nitrosylation (e) and AGR2 S-nitrosylation (f) were detected by biotin switch assay. g–i The mutants of Trx1 were expressed in HT29 CL.16E cells with endogenous Trx1 knocked down, and the SNO-protein (g), SNO-Trx1 (h), SNO-AGR2 and MUC2 (i) were detected. j The SNO-AGR2 and MUC2 were detected in TIGAR-knockout HT29 CL.16E cells after treatment with NADPH (20 mM, 12 h) and PX12 (10 μM, 12 h) under inflammatory conditions. k The SNO-AGR2 and MUC2 was observed in HT29 CL.16E cells following NADPH treatment after Trx1 knockdown. l The mutant of Trx1 was expressed in HT29 CL.16E cells following the knockdown of endogenous Trx1. Subsequently, the SNO-AGR2 and MUC2 were assessed and stimulated as indicated for 12 h. The data presented are representative of three independent experiments.

TIGAR maintains Trx1 and AGR2 activity through G6PD and 6PGD

Our results reveal an important role of G6PD and 6PGD in TIGAR-triggered oxidative stress; the NADPH/NADP⁺ ratio and GSH/GSSG ratio were declined by knockdown of G6PD and 6PGD (Supplementary Fig. 9a–d). We next investigated whether G6PD and 6PGD regulate the activity of Trx1 and AGR2. There was a substantial reduction in the reductase activity of Trx1 after G6PD and 6PGD knockout in HT29 CL.16E cells (Fig. 6a, b), accompanied by an over-oxidation (Fig. 6c). Furthermore, the S-nitrosylation of total protein, Trx1, and AGR2 was significantly elevated in G6PD or 6PGD-deficient HT29 CL.16E cells (Fig. 6d–h and Supplementary Fig. 9e, f), but decreased after activating G6PD (Supplementary Fig. 9e, f). Moreover, the S-nitrosylation of Trx1 was reduced in G6PD or 6PGD knockout cells cultured in the presence of NADPH (Fig. 6i, j). These results confirmed that G6PD and 6PGD were crucial enzymes for promoting NADPH synthesis and maintaining the reductase activity of Trx1.

Fig. 6. TIGAR maintains Trx1 and AGR2 activity through G6PD and 6PGD.

a, b Changes of Trx1 activity after G6PD or 6PGD knockout in HT29 CL.16E cells. n = 5 biological replicates. c The redox potential of Trx1 was determined by redox western analysis. Three bands were separated from top to bottom by natural non-reducing gel electrophoresis. These correspond to the completely oxidized form (Ox2), the mixture of reduced and oxidized forms (Ox1), and the completely reduced form (R). Due to the difference in the number of negative charges, the completely reduced Trx1 treated with a reducing agent (DTT) moves faster than the oxidized Trx1 treated with an oxidizing agent (diamine). d Changes of total protein S-nitrosylation after G6PD or 6PGD knockout in HT29 CL.16E cells. e–h The levels of Trx1 and AGR2 S-nitrosylation in G6PD or 6PGD depletion HT29 CL.16E cells were detected by biotin switch assay after exposing with or without TNFα (100 ng/mL) for 12 h. i, j the changes of S-nitrosylation of Trx1 in G6PD or 6PGD knockout cells cultured in the presence of NADPH. k–n TIGAR was re-expressed in cells with TIGAR knocked out, and the cells were treated with G6PDi-1 (10 μm) and 6PGD-IN-1 (10 μm) for 12 h. Trx1 (k, m) and AGR2 (l, n) S-nitrosylation levels were detected. o–r TIGAR was co-expressed with WT and mutant of G6PD or 6PGD, and the S-nitrosylation of Trx1 and AGR2 was detected. Data presented are representative of three independent experiments and shown as mean ± SD, and p-values were determined by one-way ANOVA with Tukey’s HSD post hoc test for (a, b).

To further substantiate that TIGAR sustains Trx1 reductase activity through 6PGD and G6PD, we initially performed a knockout of endogenous TIGAR in HT29 CL.16E cells, followed by the re-expression of TIGAR, which resulted in a decline of Trx1 and AGR2 S-nitrosylation (Fig. 6k–n). However, upon inhibiting G6PD or 6PGD, the S-nitrosylation levels of Trx1 and AGR2 were markedly elevated (Fig. 6k–n). We further verified the effect of G6PD and 6PGD lactylation on the S-nitrosylation of Trx1 and AGR2. Restriction of their lactylation can reduce the S-nitrosylation of Trx1 and AGR2 and restore Trx1 activity impaired by TIGAR knockout (Fig. 6o–r and Supplementary Fig. 9g, h), while also preventing MUC2 synthesis disorder (Supplementary Fig. 9i, j). These findings indicated that TIGAR preserves the activities of Trx1 and AGR2 by mitigating G6PD and 6PGD lactylation, thereby maintaining the synthesis of mature MUC2.

Mutant G6PD and 6PGD in TIGAR-KO mice alleviate DSS-induced colitis

To validate the lactylation sites of G6PD (K432) and 6PGD (K38) identified in this study and to assess the literature-reported G6PD (K45) site³⁶, we generated K432R and K45R mutations in G6PD and K38R in 6PGD, introduced these mutants into TIGAR-knockout mice, and induced colitis with DSS four weeks later. The K432 and K38 mutations effectively prevented lactylation of G6PD and 6PGD, respectively (Fig. 7g, h), and significantly restored their enzymatic activities in TIGAR-knockout mice (Fig. 7e, f). In contrast, the K45 mutation had no significant effect on G6PD activity (Fig. 7e). Further investigation revealed that either the K432 or K38 mutations significantly reduced S-nitrosylation levels of Trx1 and AGR2, and markedly enhanced mature MUC2 synthesis (Fig. 7g, h). Morphological analysis showed that mutation of G6PD at K432 or 6PGD at K38 ameliorated colitis-associated intestinal damage, enhanced mucus production, preserved mucosal layer integrity (Fig. 7a–d), and markedly reduced bacterial load in the inner mucus layer (Fig. 7c, d).

Fig. 7. Expression of WT and mutant G6PD and 6PGD in TIGAR-KO mice alleviates DSS-induced colitis.

G6PD and 6PGD AAV were injected into TIGAR^f/fVil1-Cre mice, four weeks later, induced colitis with DSS for 5 days, and distal colon were obtained from each group mice. a, b Colonic H&E staining (left) and PAS/Alcian blue staining (right) following the overexpression of both WT and mutant forms of G6PD (WT, K45R and K432R) and 6PGD (WT and K38R) in TIGAR-KO mice, scale bar: 100 μm or 50 μm. c, d Immunostaining of colon sections was performed using an anti-MUC2-C (green) antibody and FISH (bacterial 16S rRNA gene probe, red), scale bar: 50 μm. e, f The activities of G6PD and 6PGD in colon tissues following the overexpression of wild-type and mutant forms of G6PD (WT, K45R, and K432R) and 6PGD (WT and K38R) in TIGAR-KO mice, n = 5 biological replicates. g, h The nitrosylation levels of TRX1 and AGR2, as well as the expression of mature MUC2, G6PD, and 6PGD, were assessed by western blot following the overexpression of wild-type and mutant forms of G6PD and 6PGD in TIGAR-KO mice. i Schematic illustration of TIGAR maintains the intestinal mucus barrier by inhibiting lactate production, reducing lactylation of G6PD and 6PGD, promoting NADPH synthesis, inhibiting nitrosylation of TRX1 and AGR2, and facilitating MUC2 maturation. Data presented are representative of three independent experiments and shown as mean ± SD, and p-values were determined by one-way ANOVA with Tukey’s HSD post hoc test for (e, f).

Another finding of this study is that even unmutated wild-type G6PD or 6PGD exhibited some biological activity upon expression, but the magnitude of their effect was substantially lower than that of the K432R or K38R mutants, respectively. In contrast to the control (wild-type), the G6PD K45R mutant displayed no marked alterations in lactylation modification, enzymatic activity, and biological functions (Fig. 7a–h). Collectively, these findings indicate that although K45 is a site of G6PD lactylation, as supported by both published evidence³⁶ and our cellular experiments, it has a comparatively modest impact on G6PD’s post-translational modification, activity, and antioxidant capacity relative to K432.

Discussion

This study preliminarily elucidates TIGAR’s pivotal role and the underlying mechanisms in sustaining the intestinal mucus barrier in UC mice by promoting NADPH synthesis and inhibiting oxidative stress. We discover that the low TIGAR expression in UC mice is associated with increased lactylation modifications of two key rate-limiting enzymes involved in NADPH synthesis, namely G6PD, and 6PGD. Additionally, we identify the critical modification sites on these enzymes and explore how these modifications influence the enzymatic activity. Moreover, our findings reveal that S-nitrosylation of AGR2, a key enzyme in MUC2 maturation, significantly undermines the integrity of the intestinal mucus barrier in UC mice. We propose that the reduced TIGAR expression in UC mice leads to diminished NADPH synthesis, which enhances the trans-nitrosylation activity of Trx1, ultimately promoting AGR2 S-nitrosylation. This cascade of events is a primary factor disrupting MUC2 maturation and compromising the integrity of the intestinal mucus barrier.

Over a decade ago, researchers identified that abnormal expression of TIGAR was closely associated with the onset and progression of UC, suggesting that it may serve as a potential therapeutic target for this condition⁴³. However, the causal relationship between TIGAR and UC remains uncertain. In this study, we employed a conditional knockout mouse model to establish a causal link between TIGAR and the onset of UC. Following the knockout of TIGAR, pathological changes in the colon associated with UC manifested earlier, and the severity of colon damage was significantly exacerbated (Fig. 1a). Further investigations revealed that TIGAR enhances NADPH synthesis through two distinct mechanisms, thereby maintaining the stability of the intestinal mucosal barrier in UC mice. First, by promoting increased glucose entry into the PPP, thus providing precursors for NADPH synthesis (Fig. 2f, j); and second, by inhibiting glycolysis and reducing lactate production (Fig. 2d), which suppresses lactylation modifications of G6PD and 6PGD (Fig. 3d, e), thereby preserving their enzymatic activity.

This study identified a lactylation modification site at lysine 432 (K432) on G6PD through mass spectrometry (MS), marking a departure from the sole previous finding in this field, which indicated that the lactylation site on mouse G6PD is at lysine 45 (K45). The authors believe that lactylation at K45 could inhibit the formation of G6PD homodimers, subsequently reducing its activity³⁶. Accordingly, we individually mutated both sites and confirmed that K432 and K45 are lactylation sites on G6PD, but mutations at K432 exert a more pronounced effect on G6PD lactylation (Fig. 3j). Furthermore, lactylation at K432 affects G6PD dimer formation, while K45 influences G6PD’s binding to NADP⁺ without impacting its dimer formation (Fig. 3k, l). Structural biology analyses reveal that K45 of G6PD is located within its NADP⁺ binding domain (Fig. 3h), whereas K432 is situated in the region responsible for dimer formation (Fig. 3g). Therefore, our findings align more closely with the functional characteristics of the protein domains in G6PD and are supported by relevant literature⁴⁴.

These differences arise from the disparity in mutation strategies in G6PD. Our study adopts arginine substitution for lysine (K45R) rather than the previously reported alanine substitution (K45A). Both lysine and arginine have a free amino group, and their charges and structures are highly similar; however, there are significant differences in the side chain properties of alanine and lysine⁴⁵. Consequently, the K45A mutation reported in the literature not only affects the lactylation of G6PD, but it may also impact its dimerization. Furthermore, we examined the effects of K432 and K45 mutations on G6PD lactylation, enzymatic activity, and mucus barrier integrity in vivo. The K432 mutation markedly reduced G6PD lactylation, enhanced its enzymatic activity, increased mature mucin production, and alleviated UC-associated mucus barrier disruption (Fig. 7a, c, e, g). In contrast, the K45R mutation had no significant effect. Our findings identify K432 as the primary lactylation site regulating G6PD activity, refining the mechanistic understanding of how lactylation modulates its function.

A significant breakthrough of this study is the identification of another critical link in the regulation of NADPH synthesis by TIGAR, specifically the second rate-limiting enzyme, 6PGD, in the oxidative phase of the PPP. To date, no similar reports have been published. We initially identified a lactylation modification site at lysine 38 (K38) through LC-MS analysis. Subsequently, the effect of K38 lactylation on 6PGD function was analyzed using molecular docking and mutation. Our study demonstrated that K38 resides within the NADP⁺ binding domain, and its lactylation inhibits 6PGD binding to NADP⁺, blocks NADPH production (Fig. 3k). In vivo studies demonstrated that the K38 mutation markedly suppresses lactylation of 6PGD in UC mice, enhanced its enzymatic activity, leading to increased mature mucin production and alleviation of UC induced mucus barrier disruption (Fig. 7b, d, f, h).

Intriguingly, the discovery of 6PGD’s role was quite serendipitous. Our initial focus was on the effects of TIGAR knockout on G6PD, but we unexpectedly observed a significant accumulation of 6-phosphogluconate (Fig. 2c, g). In the oxidative phase of the PPP, G6PD and 6PGD are interlinked metabolic enzymes; the former catalyzes the conversion of glucose-6-phosphate to 6-phosphogluconate, while the latter catalyzes the transformation of 6-phosphogluconate into ribulose-5-phosphate⁴⁶. Following TIGAR knockout, both enzymes underwent lactylation modifications, ultimately resulting in decreased activity of these two enzymes. The substantial accumulation of 6-phosphogluconate indicates that the impact of TIGAR on 6PGD is more pronounced. While the role of TIGAR in promoting NADPH synthesis is well recognized, previous studies have primarily focused on its regulation of G6PD^26,47, neglecting its influence on 6PGD. The findings from this study not only contribute to a deeper understanding of how TIGAR regulates glucose metabolism and maintains redox balance but also provide compelling evidence for considering TIGAR as a potential target for UC therapy.

It is now established that the impairment of mucin maturation is one of the key factors leading to the disruption of the intestinal mucus barrier^6,48. This study found that the cysteine 81 (Cys81) of AGR2, a key enzyme regulating MUC2 maturation in goblet cells of UC mice, underwent S-nitrosylation, hindering the binding of AGR2 to MUC2 precursor (Fig. 4h–j and Supplementary Fig. 6b, c). This modification leads to the accumulation of immature mucins in goblet cells, exacerbates ERS, and impairs mucin maturation, ultimately compromising the integrity of the intestinal mucus barrier (Supplementary Fig. 6a, e). Further studies found that after knocking out TIAGR, the degree of S-nitrosylation of AGR2 was more evident. However, this modification was not directly regulated by NADPH but rather by Trx1 (Fig. 5j, k). Trx1 is recognized for its dual role as both a reductase and an oxidase, exhibiting capabilities for denitrosylation or trans-nitrosylation that depend on the oxidative state of its environment and the NADPH levels^40,49. Cellular experiments show that under TNF stimulation, oxidative stress occurs in the cells, leading to the formation of disulfide bonds between Cys32 and Cys35 in Trx1, inhibiting its ability to bind to S-nitrosothiol (SNO) groups and resulting in a loss of denitrosylation function (Fig. 5e). Under such conditions, Trx1 can transfer the SNO group bound to its Cys73 to AGR2 (Fig. 5i and Supplementary Fig. 8d), facilitating its trans-nitrosylation effect. Knocking out TIGAR makes the above reaction more pronounced. Consequently, the levels of cellular oxidative stress and the extent of NADPH depletion determine whether Trx1 performs denitrosylation or transnitrosylation. In summary, this study provides evidence that UC induces S-nitrosylation of AGR2 and illustrates the specific process by which TIGAR regulates the activity of AGR2 through Trx1.

This study confirms that the impaired conversion of immature MUC2 to its mature form in the goblet cells of UC mice is a crucial factor contributing to the damage of the intestinal mucus barrier. The persistent low expression of TIGAR in UC mice promotes the lactylation of G6PD and 6PGD, inhibits NADPH synthesis, induces the over-oxidation of Trx1, and leads to the S-nitrosylation of AGR2. These processes obstruct the synthesis of mature MUC2 and impair the integrity of the intestinal mucus barrier. This work provides a theoretical foundation for considering TIGAR as a potential therapeutic target for UC treatment. Although this study thoroughly explores the mechanisms by which TIGAR promotes mucin maturation and maintains the mucus barrier, it remains to be determined whether TIGAR can regulate subsequent processing events of mucins, such as glycosylation and sialylation modification of mucins. This issue will be the focus of our future research.

Methods

Cell lines and culture

The 293 T cell line (ATCC, CRL-3216, fetal) was maintained in our laboratory and was cultured in DMEM medium (Gibco, Carlsbad, CA, USA) supplemented with 10% FBS (Gibco, Carlsbad, CA, USA) at 37°C in a 5% CO₂ incubator. The HIEC-6 cell line (CRL-3266, fetal) was purchased from ATCC and cultured in DMEM medium (Gibco, Carlsbad, CA, USA) supplemented with 5% FBS (Gibco, Carlsbad, CA, USA), 10 mM HEPES, 5 mg/ml EGF, 100 U/ml penicillin/streptomycin, 4 mM Glutamax at 37°C in a 5% CO₂ incubator.

Sodium butyrate was utilized to differentiate HT29 cells into mature goblet cells, thereby obtaining mucus-secreting HT29 CL.16E cells⁵⁰. Briefly, the Human colon cancer cell line HT29 (ATCC, HTB-38, female) was cultured in a 5 mmol/L sodium butyrate solution for 9 days. Subsequently, the cells were passaged and maintained in a medium supplemented with sodium butyrate for 14 days to generate the HT-29 Cl.16E cell line. Over time, these cells differentiate into goblet cell phenotypes characterized by their ability to produce substantial amounts of mucus. The HT-29 Cl.16E cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂. The culture medium consisted of DMEM supplemented with 10% FBS, 100 U/mL penicillin/streptomycin, and 1 mM sodium pyruvate.

Animals and experimental models

Healthy male C57BL/6 N mice (22–25 g, 8–10 weeks of age) were obtained from the Third Military Medical University Laboratory Animal Center. The animals were housed under SPF conditions in individually ventilated cages, with a 12-h light/dark cycle and ad libitum access to food and water. Before the experiment, the mice were acclimatized to a standardized diet for one week. All procedures adhered to institutional guidelines and regulations and the experimental protocol was approved by the Experimental Animal Welfare and Ethics Committee of the Third Military Medical University (Approval number: AMUWEC20210636).

TIGAR^f/f mice have two lox sites on both sides of exon 2 in the TIGAR locus. Floxed mice were crossed with Vil1-Cre mice to generate tissue-specific TIGAR knockout mice. Mice were administered orally with 2.5% DSS dissolved in sterile distilled water for 7 days to establish the UC model. The mice were randomly assigned to the Cont, UC (DSS), TIGAR^f/f, DSS+TIGAR^f/f, and DSS+TIGAR^f/fVil1-Cre groups. Following the experimental period, the mice were euthanized, and the colon tissues were collected for subsequent analysis.

The adeno-associated viruses (AAVs) used in the overexpression of G6PD (WT, K45R and K432R) and 6PGD (WT and K38R) were manufactured by Cyagen Biosciences Inc (Shanghai, China). 2 × 10¹¹ genomes of the AAV9 vector for G6PD (WT, K45R and K432R) and 6PGD (WT and K38R) overexpression and control viruses were injected intraperitoneally into mice. Four weeks later, DSS was used to induce colitis for 5 days.

The animals used in this study were male. Unless otherwise specified, all references to animals in the text refer to males.

Stable knockout, knockdown and overexpressing cell lines

Stable knockout cell lines of TIGAR, Trx1, and G6PD were generated utilizing CRISPR/Cas9 technology. The double-stranded oligonucleotide complementary to the target sequence was cloned into the lentiCRISPRv2 vector and co-transfected with the packaging plasmid into HEK293 cells. Subsequently, a 48 h viral supernatant was collected to infect cells with polybrene. The infected cells were treated with a specific concentration of puromycin for 7 days, and the particular gRNA sequence information was presented in Supplementary Table 1.

The shRNA sequence was cloned into the pLKO.1-Puro vector and co-transfected with packaging plasmids to construct stable knockdown cell lines. The viral supernatant was collected 48 h post-transfection for cell infection, and infected cells were selected with puromycin for 7 days. Specific shRNA sequence information is provided in Supplementary Table 1.

Mammalian expression plasmids for Flag-, TIGAR, Trx1, AGR2, G6PD, 6PGD, and their mutants were constructed using standard molecular biology techniques. The WT or mutant sequences were cloned into the pLV4ltr-Puro-CMV vector. The shRNA approach was used to construct stable overexpressed cell lines.

Mucin detection

HT29 Cl.16E cells or colonic tissues were lysed using a lysis buffer containing 1 mmol/L EDTA, 10 mmol/L Tris-HCl, 10% SDS, 5 mmol/L sodium chloride, and 20% Triton X-100, supplemented with 5 mmol/L phenylmethylsulfonyl fluoride (PMSF) and a protease inhibitor cocktail (Sigma-Aldrich). The lysates were then subjected to compound AgPAGE separation and transferred to a nitrocellulose membrane. The membrane was probed with two MUC2 precursor antibodies: MUC2-VNTR (Invitrogen, MA5-12345) and MUC2-VNTR (Novus, NBP2-25221), capable of detecting immature MUC2. The amount of mature MUC2 was detected using a MUC2-N antibody (Abcam, ab90007) in colonic tissue and HT29 CL.16E cells.

H&E staining, AB-PAS staining, and Histological scoring

Intestinal slices with a thickness of 5 μm were fixed overnight in 4% Carnoy’s solution and then paraffin-embedded. H&E staining was performed by staining the sections with hematoxylin and dehydrating them in graded alcohols and xylene. For AB-PAS staining, dehydrated paraffin sections were stained with periodic acid, incubated with Schiff’s reagent, and then counterstained with hematoxylin before dehydration. The histological score for colonic epithelial injury was determined as follows: crypt structure was scored from normal (0) to severe crypt distortion with complete crypt loss (3); the degree of inflammatory cell infiltration was scored from normal (0) to dense inflammatory infiltration (3); the score for muscle thickening ranged from normal (0) to evident apparent muscle thickening (3); depletion of goblet cells was scored from absent (0) to present (1); crypt abscesses were scored from absent (0) to show (1).

To highlight the images of the mucosal layer changes, we selected different mouse samples for staining and statistical analysis of crypt height changes. The statistics for crypt height were derived from five mice per group. For each mouse, we measured the lengths of six consecutive crypts and calculated their average. Each point in the figure represents the average length of six consecutive crypts in one section of each mouse.

Determination of ROS, NO levels, and related redox couples

Cellular ROS levels were quantified using H2DCFDA (MCE, HY-D0940). Briefly, a 10 mM stock solution of H2DCFDA was prepared by dissolving it in DMSO and subsequently diluted for working concentrations. Adherent cells were incubated with PBS containing 5 μM H2DCFDA at 37°C for 30 min and harvested with 0.05% trypsin-EDTA. The cells were then resuspended in a fresh culture medium and immediately analyzed using NovoCyte flow cytometry (excited by a 488 nm laser) to measure fluorescence intensity. The MitoSox probe and DHE probe were used to detect mitochondrial ROS and superoxide, and the incubation method was consistent with that for ROS detection.

The concentration of NO was detected using a commercial kit (Beyotime, S0023). Briefly, cells and tissue samples were lysed with the NO detection-specific lysis buffer (Beyotime, S3090), and the reaction solution was prepared according to the instructions. After thorough mixing, the samples were incubated at room temperature for 10 min. Then, the OD value at 540 nm was measured, and the NO concentration was calculated based on the standard curve. The DAF-FM DA fluorescent probe (Beyotime, S0020S) was used to detect the intracellular NO level. Briefly, a staining solution with a final concentration of 5 μM DAF-FM DA was prepared. The cell culture medium was removed and added the staining working solution. The cells were incubated in a 37 °C cell culture incubator for 20 min. The cells were washed three times with PBS to thoroughly remove the residuary DAF-FM DA. Fluorescence images were collected using a laser confocal microscope.

NADPH/NADP⁺ ratios in colon tissues and cells were measured using commercial kits (Beyotime, S0180S). The cell or tissue is first lysed using NADPH/NADP⁺ extractive buffers. The supernatant was then collected for analysis after centrifugation at 12000 × g and 4 °C for 5 min. Generate standard curves according to instructions. The concentration of NADPH in the sample was determined after the decomposition of NADP⁺ in a water bath at 60 °C for 30 min. The ratio of NADPH to NADP⁺ is calculated as follows: NADPH/NADP⁺ = (NADPtotal- NADP⁺)/NADP⁺. The ratio of GSH/GSSG in colon tissues and cells was detected using commercial kits (Beyotime, S0053). Standard curves were generated for GSH and GSSG concentrations.

Trx1 redox state monitoring and determination of Trx1, TrxR activity

To investigate the redox status of Trx1 in colon tissues and cultured cells, protein samples were prepared by precipitation with trichloroacetic acid (Macklin, T885181). The samples were labeled using the -SulfoBiotics Protein Redox State Monitoring Kit Plus (Dojindo, SB12). Cell and colon tissue extracts were subjected to SDS-PAGE and irradiated with ultraviolet light on a radiator for 10 min to eliminate protein shifters. Proteins were then electrophoretically transferred to polyvinylidene fluoride membranes. The membranes were blocked with 5% skim milk, incubated overnight with an anti-Trx1 antibody, and subsequently incubated with a secondary antibody. Signal strength was quantified using an ECL luminescence system. The activities of Trx1 and TrxR were quantified using the Thioredoxin Fluorometric Activity Assay Kit (Cayman, 500228) and Oxidized Thioredoxin Reductase(TrxR) Activity Assay Kit (Solarbio, BC1155), respectively, according to the manufacturer’s protocols.

Immunofluorescence

Cells were fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.5% Triton X-100 for 15 min, and blocked with 5% normal serum for 1 hour at room temperature. Primary antibody was added and incubated overnight at 4 °C. After washing with TBST, fluorescent secondary antibody was applied and incubated at 37 °C for 1 h. DAPI staining was performed in the dark for 10 min, and specimens were sealed with anti-fluorescence quencher solution before imaging under a fluorescence microscope.

Western blot

The cell or tissue protein was collected, and its concentration was determined using the BCA method. Subsequently, the protein was transferred to a PVDF membrane through SDS-PAGE electrophoresis. The membrane was then blocked with 5% skim milk powder and incubated overnight at 4°C with the appropriate dilution of primary antibody working solution. Following this, the membrane was washed for 10 min each with TBST buffer, then incubated at room temperature for 90 min with the secondary antibody working solution. Subsequently, the film should be washed three times with TBST buffer and observed using an ECL luminescent solution. Antibodies: anti-TIGAR (Santa Cruz, sc-166290), anti-Trx1 (Abcam, ab273877), anti-Trx1 (ABclonal, A4024), anti-TrxR (Abcam, ab124954), anti-AGR2 (Abcam, ab76473), anti-G6PD (Abcam, ab993), anti-6PGD (Invitrogen, PA5-80894), Anti-L-Lactyl Lysine (PTM Bio, PTM-1401RM), anti-MUC2 (Invitrogen, MA5-12345), anti-MUC2 (Novus, NBP2-25221), anti-MUC2 (Abcam, ab90007), anti-Flag (protein tech, 20543-1-AP), anti-Flag (protein tech, 66008-4-Ig), anti-CHOP (Abcam, ab11419), anti-GRP78 (Abcam, ab108615), anti-GAPDH (protein tech, 60004-1-Ig), anti-Tubulin (protein tech, 66031-1-Ig).

Co‑immunoprecipitation (Co‑IP)

The co-immunoprecipitation assay used corresponding antibodies with the Pierce™ Classic Magnetic IP/Co-IP Kit (Thermo Fisher Scientific, 88804). In Brief, the cells were lysed on ice using an IP buffer containing a protease inhibitor, followed by centrifugation at 14000 x g for 20 min at 4 °C. The resulting supernatant was collected. Incubation with 5 μg of antibody was carried out overnight in a shaker at 4 °C, followed by the addition of 25 μl of clean Protein A/G magnetic beads. The mixture was then shaken for 4 h at 4 °C, and the magnetic beads were collected. Then, he collected magnetic beads, which were washed with IP buffer and wash solution buffer, and finally eluted with 100 μl of eluent to obtain proteins for western blot analysis.

Biotin switch assay for detection of S-nitrosylation

The S-nitrosylation modification of proteins was detected with minor improvements⁵¹. Cells were lysed using HENS buffer (250 mM HEPES, 4 mM EDTA, 0.1 mM neocuproine, 1% SDS), and equal protein amounts were incubated with MMTS at 50 °C for 20 min to block free cysteine thiols. After acetone precipitation and centrifugation, excess MMTS was removed using 70% acetone. The precipitate was resuspended in HENS buffer, mixed with biotin-HPDP and sodium ascorbate, and incubated for 1 hour in the dark. Following acetone precipitation and washing, the sample was neutralized and incubated with streptomycin at 4°C for 12–18 h. After low-speed centrifugation, the sample was washed, eluted, and analyzed by SDS-PAGE.

Untargeted metabolomics analysis

For cellular samples, 1 mL of pre-chilled (at -80°C) 80% (v/v) methanol of MS grade was added to the culture dish, and cell debris was scraped off using a spatula. The cells were disrupted using a low-temperature ultrasonic disrupter for 60 seconds at -20°C, followed by a pause of 30 seconds, with this cycle repeated three times. 25 mg of colon tissue was weighed and mixed with 750 μL of pre-chilled 80% methanol for mouse colon samples. A low-temperature tissue grinder was utilized to homogenize the tissue for 60 seconds at -20°C, with a 30-second pause, repeated three times. The resulting homogenates from cells and colons were transferred and incubated at -20°C for 60 min. Subsequently, samples were centrifuged at 18,410 x g for 10 min at 4°C. The supernatant was collected after centrifugation, dried under nitrogen gas, reconstituted, and centrifuged again to obtain the supernatant for subsequent analysis. A 20 μL aliquot from each sample was combined to prepare a quality control (QC) mixture for mass spectrometry analyses.

An Orbitrap Exploris 120 Mass Spectrometer (Thermo Fisher Scientific) was used to analyze the hydrophilic metabolites extracted from cells and colon tissues. The column temperature was 45 °C. In negative ion mode, mobile phase A (5% ACN, 10 mM ammonium bicarbonate, pH 9) and mobile phase B (95% ACN, 10 mM ammonium bicarbonate, pH 9) were eluted by a gradient according to proportional mixing on a BEH Amide HILIC column (2.1 × 150 mm, 1.7 μm, Waters). The linear gradient was as follows: 0 min, 99% B; 0.1 min, 99% B; 6 min, 30% B; 6.5 min, 99% B; 10 min, 99% B. The spray voltage was 2.5 kV for negative ion mode. The capillary temperature was 350 °C for the negative ion mode. The mass range (m/z) was 70–1,050 for negative ion mode.

The raw files of untargeted metabolomics were searched using Compound Discoverer 3.3 (Thermo Fisher Scientific), which is based on the mzCloud and mzVault databases. The mass tolerance for precursor ions was set at 10 ppm, while the mass tolerance for fragment ions was defined as 15 ppm. For quantification purposes, a retention time shift of up to 0.25 min was permitted. For detailed methods, please refer to the Supplementary Information.

Isotope tracing analysis

Cells were cultured in a glucose-free medium supplemented with 10 mM D-glucose (MCE, HY-B0389) and U-¹³C-labeled glucose (MCE, HY-B0389A). Samples were collected at 1, 2, 4, 8, and 24 hours post-treatment. For metabolite extraction, 1 mL of pre-chilled 80% (v/v) methanol was added to each culture dish, and subsequent sample processing was performed according to established protocols for untargeted metabolomics.

In the in vivo experiment, mice were fasted for 12 h before treatment. 100 μL 1 M Unlabeled glucose and U-¹³C-labeled glucose were administered intravenously via the tail vein at 15-minute intervals. After the final injection, the mice were euthanized, and colon tissue samples were harvested for further analysis.

Isotope tracing analysis was performed using an Orbitrap Exploris 120 Mass Spectrometer (Thermo Fisher Scientific) with a column temperature maintained at 30 °C. In negative ion mode, gradient elution was conducted at a flow rate of 0.5 mL/min on a Premier ZHILIC column (2.1 × 100 mm, 1.7 μm, Waters) using mobile phase A (5% ACN, 15 mM ammonium acetate, pH 9) and mobile phase B (95% ACN, 15 mM ammonium acetate, pH 9). The linear gradient was as follows: 0 min, 90% B; 5 min, 65% B; 6 min, 65% B; 8 min, 90% B; 10 min, 90% B. The ion source used was H-ESI, operating in negative ion mode at 3 kV, with a sheath gas of 60 Arb, an auxiliary gas of 20 Arb, and a sweep gas of 2 Arb. The ion transfer tube temperature was maintained at 380°C, while the ion source evaporation temperature was set at 350°C. The collected data were analyzed using Compound Discoverer 3.3 software following the isotope-tracing analysis workflow. For detailed methods, please refer to the Supplementary Information.

Pan antibody-based PTM enrichment and LC-MS/MS analysis

The HT29 CL.16E cells were exposed to lactic acid for 24 h to enhance the Kla-modified peptides. The TNFα-treated HT29 CL.16E cells with Trx1 overexpression were incubated for 24h to improve S-nitrosylated peptides. The tryptic digest was resuspended in IP buffer (1 mM EDTA, 50 mM Tris-HCl, 100 mM NaCl, 0.5% NP-40, pH 8.0) and subsequently incubated with pre-washed Anti-L-Lactyl Lysine Antibody Conjugated Agarose Beads (PTM Bio, PTM-1404) or Anti-TMT Antibody beads (Thermo Scientific, 90076) at 4°C for 12 hours. The bound peptides were eluted using 0.1% trifluoroacetic acid (TFA). After vacuum drying, the eluted peptides were desalted using Millipore C18 ZipTips prior to LC-MS/MS analysis. LC-MS/MS analysis was supported by Jingjie PTM BioLabs. For detailed methods, please refer to the Supplementary Information.

RNA-Seq

Total RNA was extracted from colon tissues of three wild-type (WT) mice and three TIGAR knockout (KO) mice using TRIzolTM reagent (Invitrogen). The purity and concentration of RNA were assessed using the Agilent 2100 Bioanalyzer system (Agilent Technologies). The mRNA was isolated from total RNA using the Dynabeads® mRNA Purification Kit (Invitrogen). The mRNA was fragmented using a fragmentation buffer, and a cDNA library was constructed using the fragmented mRNA as a template. Sequencing was performed on the Illumina NovaSeq6000 platform at the Beijing Genomics Institute, and the data were aligned to the UCSC mm10 reference genome using TopHat.

Determination of G6PD/6PGD activity and the G6PD/6PGD dimer

G6PD activity was measured using a commercial Kit (Beyotime, S0189). In brief, 1 × 10⁶ cells or 20 mg of tissue samples were collected and lysed with G6PD extraction solution at 4 °C. The protein concentration was normalized and quantified using a BCA assay kit (Beyotime, P0010). 50 μL of the protein sample and 50 μL of the G6PD test solution were added to each well of a 96-well plate. Following a 10-minute incubation at 37 °C in the dark, the OD value was measured at 450 nm using a microplate reader. The 6PGD activity was measured using the 6-Phosphogluconate dehydrogenase Activity Assay Kit (Solarbio, BC2105) according to the manufacturer’s protocols.

G6PD or 6PGD dimer was detected using the DSS crosslinking method. Briefly, the culture medium was aspirated, and cells were washed twice with PBS. Cells were subsequently scraped in PBS containing an EDTA-free protease inhibitor cocktail. DSS was added to the samples, followed by cross-linking through rotation at 37 °C for 30 min. An equal volume of 1 M Tris (pH 7.5) was then added to achieve a final concentration of 20 mM, and the mixture was incubated at room temperature for 15 min. Cell lysis was performed using M-PER buffer and centrifugation at 16,000 × g for 10 min at 4°C. The supernatant was collected for subsequent Western blot analysis.

Statistical analysis

All statistical analyses were conducted using GraphPad Prism (v 8.0.1, GraphPad Inc.) or R v.4.1.1. Unless otherwise stated, results are reported as mean ± standard deviation (SD). Between-group comparisons were conducted using the unpaired Student’s t-test, unpaired t-test with Welch’s correction, or the Mann–Whitney U test, depending on the data’s normality and the equality of variances across groups. Multigroup comparisons were analyzed through one-way analysis of variance (ANOVA) with Tukey’s honestly significant difference (HSD) post hoc test or two-way repeated-measures ANOVA. All hypothesis tests were two-tailed, and a predetermined significance level of α = 0.05 was used as the threshold for statistical significance across all analyses.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

X.P. conceived the work and supervised the overall experiments. D.W., Y.W., L.X., and S.J.F. participated in relevant animal experiments and obtained specimens and data. D.W., S.S., P.Y.Z., X.L.Z., Y.W., T.Z., Z.H.Z., Y.Y., Q.Y.H., X.Y.L., Q.C. and C.Y.L. obtained the test data. D.W., S.S., P.Y.Z. performed bioinformatic analyses. X.P., P.Y.Z., D.W., and S.S. performed manuscript writing, reviewing, and editing. All authors discussed the results and commented on the manuscript.

Peer review

Peer review information

Nature Communications thanks Pin Wang, Wei Wei, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The proteomic data generated in this study have been deposited in the ProteomeXchange database under accession code PXD069158 and PXD069246. The RNA sequence data generated in this study have been deposited in the SRA database with the dataset identifier PRJNA1330206. The metabolomics project has been uploaded to Metabolomics Workbench and assigned the digital object identifier (DOI): 10.21228/M8VG27. All other study data are included in the article and/or Supplementary Information. Any additional information is available upon request to the corresponding author (Xi Peng, pxlrmm@tmmu.edu.cn). Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-70263-z.