In all experiments, 10 to 12-week-old C57BL/6 Muc2^+/+ and Muc2^-/- littermates (equal numbers of male and female mice) were used. Wild-type mice were purchased from Charles River Laboratories International (St-Constant, Quebec, Canada) and Muc2^-/- mice obtained from Dr. Velcich, were bred in-house to generate littermate Muc2^+/- breeding pairs and backcrossed on a C57BL/6 background. Germ-free (GF) mice on a C57BL/6 background were purchased from the International Microbiome Centre at the University of Calgary. Mice were kept in sterilized, filter-top cages and maintained on standard food and water ad libitum in specific-pathogen-free (SPF) conditions. The Health Sciences Animal Care Committee from the University of Calgary examined the animal care and treatment protocol (AC18-0218) and approved the experimental procedures proposed and certifies with the applicant that the care and treatment of animals used were in accordance with the principles outlined in the most recent policies on the “Guide to the Care and Use of Experimental Animals” by The Canadian Council on Animal Care.

The human colon cancer cell line LS174T was maintained in Eagle’s Minimum Essential Medium (EMEM) supplemented with 10% heat-inactivated fetal bovine serum, 10 mM HEPES, and 10 mM penicillin and streptomycin (complete EMEM). Cells were maintained in a humidified incubator with 5% CO₂. Media was changed every 2 to 3 days and cells were passed once a week, as the cells became confluent. For protein extraction, LS174T cells were plated onto 6-well plates (2.5 x 10⁵ cells/well) in 3 mL of complete EMEM. For RNA extraction, LS174T cells were plated onto 12-well plates (1.5 x 10⁵ cells/well) in 2 mL of complete EMEM. 10 μM PMA was added to wells for 2 hours to induce MUC2 expression where indicated. Tunicamycin (5 μg/mL) was added to wells for 12 hours to interrupt N-glycosylation where indicated.

RNA from cells were extracted using the Omega Bio-Tek E.Z.N.A. Total RNA Kit, as per the manufacturer’s instructions. RNA from tissues were homogenized in Ribozol RNA Extraction Reagent (VWR) and then extracted as per the manufacturer’s instructions. For DSS-treated tissue, the DSS was removed as previously described (17). Briefly, RNA was incubated with 2 M lithium chloride on ice for 2 hours, followed by centrifugation at 14 000 x g for 30 min at 4 °C. The supernatant was removed and discarded, and the pellet of RNA was dissolved in 200 μL RNase-free water. The purification process was repeated once more. The pelleted RNA was resuspended in 200 μL RNase-free water and precipitated at -20°C for 30 min in 3 M sodium acetate and 100% ethanol. Once the RNA was isolated, RNA yield was measured using a NanoDrop1000 spectrophotometer. 500 ng of RNA was reverse transcribed using the qScript cDNA SuperMix (QuantaBio). qPCR was performed using primers the following primers and annealing temperatures: Human MUC2 (F: CAGCACCGATTGCTGAGTTG, R: GCTGGTCATCTCAATGGCAG, 56°), Human FCGBP (F: GGACCTCAAGAACACTGGCA, R: GAGGATGGAGACTGAAGCGG, 60°), Human GAPDH (F: TGATGACATCAAGAAGGTGGT-GAAG, R: TCCTTGGAGGCCATGTGGGCCAT, 60°), Human MATH1 (F: TGCACTTCTCGAC-TTTCGAGGACA, R: AACTTGCCTCATCCGAGTCACTGT, 63°), Human SPDEF (F: TTGCTA-CTCAAGCCCCACAG, R: CCGGATGATGCCCTTCTTGT, 58°), Human ATF4 (F: TGCTGTCT-GCCGGTTTAAGT, R: CTGCTGCCTCTAATACGCCA, Mouse Muc2 (F: GAAGCCAGATCCCG-AAACCA, R: CCAGCTTGTGGGTGAGGTAG, 60°), Mouse Fcgbp (F: GACAAAGCCTATCT-CTGCCGT, R: CTTGGGTAGCAGACAGGGAA, 58°), Mouse β-Actin (F: TGGGGTGTTGAAGG-TCTC, R: CTACAATGACTGCGTGTG, 58°). Mouse Math1 (F: AAAGGAGGCTGGCAG-CAA, R: TGGTTCAGCCCGTGCAT, 58°), Mouse Spdef (F: GACTCACACTCAAGGGGCAA, R: TCAG-AAGAGTCGTCCGTCCT, 58°). Reactions were done in a Corbett Rotor Gene 3000 system. Data analysis was done using the 2^−ΔΔCT method to interrogate fold changes in gene expression.

Protein samples were taken from 6-well plates after various times and treatments. The lysates were extracted, washed with cold PBS, and subsequently lysed with 300 μL complete lysis buffer (1 mmol/L EDTA, 20 mmol/L Tris-HCl, 0.1% SDS, 100 mmol/L NaCl, 0.5% Triton X-100, phenylmethylsulfonyl fluoride, and a protease inhibitor cocktail; Sigma-Aldrich) on ice for 30 min. Samples were then centrifuged at 8000 x g for 10 min and the pelleted debris was removed. The remaining protein was quantified by a BCA protein assay and adjusted to equal concentrations of protein, and then resuspended into 5X Laemmli buffer and boiled for 5 min. The lysates were loaded into 7.5%, 10%, 12%, or 15% SDS polyacrylamide gels and then transferred onto a nitrocellulose membrane. After transfer, the membranes were then blocked in 5% milk powder in PBS and 0.1% Tween. The membranes were then probed with primary antibodies: MUC2 (in-house antibody against CsCl purified glycosylated MUC2 mucin (18) or Santa Cruz sc-515106 for the C-terminus), FCGBP (Cloud-Clone PAP389Hu01), and CLCA1 (Abcam 1808521). As a housekeeping marker, mouse monoclonal anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Calbiochem 6C5) was used. Corresponding HRP-conjugated secondary antibodies were diluted in 3% milk powder in PBS and 0.1% Tween and then visualized using ChemiLucent ECL detection (EMD Millipore) and imaged on a Bio-Rad ChemiDoc Imager. For samples prepared under non-reducing conditions, β-Mercaptoethanol and SDS were omitted from the 5X sample buffer.

1.5x10⁵ cells were seeded on sterile No. 1.5 glass coverslips and left to grow for 24 hours. Coverslips were washed twice with ice-cold PBS, then fixed with 4% paraformaldehyde for 15 min at 37°C, and then washed twice with PBS again. Cells were then permeabilized with PBS containing 0.35% Triton for 5 min at room temperature. After washing thrice with PBST (PBS with 0.1% Tween), coverslips were blocked for 1 hour with 5% donkey serum in PBST, washed twice, and incubated overnight with primary antibodies for MUC2 (sc-515106, Santa Cruz) or FCGBP (PAP389Hu01, Cloud-Clone) in a humidified chamber at 4°C. The next day, coverslips were washed with PBST thrice and incubated for 1 hour with fluorescent secondary antibodies and DAPI (Life Technologies). For migration and proliferation studies, cells were counterstained with Paxillin (05-417, Millipore) and the incorporation of EdU quantified as per the manufacturer’s instruction (C10637, Thermo Fisher), respectively. Coverslips were mounted with FluorSave (Calbriochem) and imaged using a Nikon A1R confocal laser scanning microscope or Nikon Ti Eclipse Widefield microscope. STED imaging used the same protocol; however, imaging was performed using an Aberrior STEDYCON microscope attached to a Nikon Ti Eclipse Widefield microscope. Quantification and image analysis were done in FIJI (ImageJ) to determine protein intensity and colocalization.

Mucin granules were prepared from confluent flasks of LS174T cells. Cells were washed with PBS, collected, and centrifuged at 300 x g for 5 min. The resulting pellet was resuspended in MES homogenization buffer (250 mM sucrose, 25 mM MES, 0.1 mM MgSO4, 2 mM EGTA, 0.1 mM PMSF) and homogenized in a Dounce homogenizer. The homogenate was then centrifuged at 300xg for 5 min and the supernatant was collected and re-centrifuged at 300xg for 10 min. The cleared supernatant was then centrifuged at 2100 x g for 15 min. The resulting pellet was then washed 3 times with PBS and the crude mucin granules were either left untreated or treated with either 1% CHAPS or 1% octyl-glucoside. For further purification, the crude granules were resuspended into 2 mL HEPES buffer (20 mM HEPES, 137 mM NaCl) and carefully placed on top of a sucrose gradient (2 mL fractions of each 2 M sucrose, 1.5 M sucrose, 1 M sucrose, and 0.5 M sucrose) and ultracentrifuge at 20 000xg overnight. After centrifugation, 1 mL fractions were isolated, washed three times in PBS, and prepared for analysis.

WT and MUC2 KO LS174T goblet-like cells were used for shotgun proteomics analysis. Protein samples were lysed with 1% SDS, 0.1 M EDTA in 200 nM HEPES (pH 8), and protease inhibitor tablets (Roche). Proteins were denatured with the addition of a final concentration of 10 mM DTT. Samples were alkylated by incubation with a final concentration of 15 mM iodoacetamide in the dark for 25 min at room temperature at a pH of 6.5. Next, to label peptide α- and ϵ-amines, samples were incubated for 18 h at 37°C with isotopically heavy [40 mM 13CD2O + 20 mM NaBH3CN (sodium cyanoborohydride)] or light labels [40 mM light formaldehyde (CH2O) + 20 mM NaBH3CN]. These were all final concentrations. Samples were subjected to C18 chromatography before being subjected to liquid chromatography and tandem mass spectrometry (LC-MS/MS).

All MS experiments were carried out by the Southern Alberta Mass Spectrometry (SAMS) core facility at the University of Calgary, Canada as previously described (19). Analysis was performed on an Orbitrap Fusion Lumos Tribrid MS (Thermo Scientific) operated with Xcalibur (version 4.0.21.10) and coupled to a Thermo system. Tryptic peptides (2 μg) were loaded onto a C18 trap (75 μm x 2 cm; Acclaim PepMap 100, P/N 164946; ThermoScientific) at a flow rate of 2 μl/min of solvent A (0.1% formic acid and 3% acetonitrile in LC-MS grade water). Peptides were eluted using a 120 min gradient from 5 to 40% (5–28% in 105 min followed by an increase to 40% B in 15 min) of solvent B (0.1% formic acid in 80% LC-MS grade acetonitrile) at a flow rate of 0.3 μL/min and separated on a C18 analytical column (75 μm x 50 cm; PepMapRSLC C18; P/N ES803; Thermo Scientific). Peptides were then electrosprayed using 2.3 kV voltage into the ion transfer tube (300 °C) of the Orbitrap Lumos operating in positive mode. The Orbitrap first performed a full MS scan at a resolution of 120,000 FWHM to detect precursor ions with an m/z between 375 and 1575 and a +2 to +7 charge. The Orbitrap AGC (Auto Gain Control) and the maximum injection time were set at 4 x 105 and 50 ms, respectively. The Orbitrap was operated using the top speed mode with a 3 s cycle time for precursor selection. The most intense precursor ions presenting a peptidic isotopic profile and having an intensity threshold of at least 5000 were isolated using the quadrupole and fragmented with HCD (30% collision energy) in the ion routing multipole. The fragment ions (MS2) were analyzed in the ion trap at a rapid scan rate. The AGC and the maximum injection time were set at 1 x 104 and 35 ms, respectively, for the ion trap. Dynamic exclusion was enabled for 45 s to avoid the acquisition of the same precursor ion with a similar m/z (plus or minus 10 ppm).

Spectral data were matched to peptide sequences in the human UniProt protein database using the MASCOT (www.matrixscience.com). The cleavage site specificity was set to Trypsin/P for the proteomics data, with up to two missed cleavages allowed. After protein identification, the resulting hits were submitted to Metascape (20) for bioinformatic analysis to generate the gene ontology pathway analyses and transcription factor analysis presented. For the protein-protein interaction study, STRING v11 (https://string-db.org) software was used (21).

Mucin preparations from LS174T secretions were collected as previously described (22). Briefly, spent media from LS174T cells were dialyzed, lyophilized, resuspended in column buffer, and applied to a Sepharose 4B columns (50 by 2.5 cm; Bio-Rad Laboratories) that were equilibrated with column buffer at 4°C. The column was calibrated by using known standards: blue dextran (2000 kDA), thyroglobulin (669 kDa), bovine serum albumin (67 kDa), and Chymotrysinogen (25 kDa). Four mL fractions were collected and read on a DU800 Spectrophotometer (Beckman Coulter). Appropriate samples were pooled and dialyzed against deionized water at 4°C overnight, lyophilized, and resuspended into 100 µL of water. Protein concentration was determined through a BCA protein assay prior to preparing samples for Western blotting.

Colonic organoids were derived as previously described from Muc2^+/+ and Muc2^-/- littermates (23, 24). Briefly, the colons were excised, and crypts digested using a gentle cell dissociation buffer (StemCell Technologies) and approximately 100 crypts were embedded into GFR-matrigel (BD) domes. Organoids were fed every two days with advanced DMEM supplemented with L-WRN (ATCC CRL-3276) as well as N2, B27, GlutaMax, SB202190, Nicotinamide, N-acetylcysteine, A83-01, and mEGF. Colonoids were passaged once every 7-10 days with TrypeLE. For experiments, colonoids were grown for 7 days in complete media and then fed with L-WRN-conditioned media containing only N2, B27, Nicotinamide, and Glutamax for 3 days. Five µM DAPT was administered to the experimental media 24 hours prior to harvesting for experiments using colonoids skewed to a goblet cell phenotype.

Production and measurement of growth factors by WT and MUC2 KO LS174T cells were assessed using a Luminex multiplex assay (Eve Technologies, Calgary, AB, Canada). Wounds were generated in cells using a 2-well silicone insert with a defined 500 µm cell-free gap (IBIDI), and after 48 hours, the culture inserts were removed, and the supernatants were collected for day 0 measurements. Fresh media were added and removed on days 2 and 4. The supernatant was centrifuged at 10 000 x g for 10 min to remove cellular debris prior to the analysis.

The animals were treated with an antibiotic cocktail as previously described to decrease the bacterial load (25). Briefly, the mice were gavaged every 12 hours with an antibiotic cocktail. For the first 3 days, the cocktail contained amphotericin-B (1 mg/kg). Starting on day 4, the cocktail contained vancomycin (50 mg/kg), neomycin (100mg/kg), metronidazole (100 mg/kg), and amphotericin-B (1 mg/kg) for the next 14 days. Additionally, ampicillin (1 mg/mL) was added to the drinking water and consumed ad libitum. The animals were used 2 days following the antibiotic treatment in order for it to be cleared from systemic circulation.

Chemical colitis was induced as previously described (26) using DSS (MW: 36,000–50,000; MP Biochemicals, Santa Ana, CA) dissolved in tap water. The Muc2^+/+ animals received 3.5% for 5 days, while the Muc2^-/- mice received 1% for 3 days follow up by water ad libitum. The mice were weighed daily and sacrificed on days 4, 8, 12, and 15 post-DSS.

All experiments were repeated at least 3 independent times. All statistical analysis was done using GraphPad Prism 8. Statistical significance between two or more groups was analyzed with two-way ANOVA, whereas the Student’s t-test was used for comparison between 2 groups. Significance was assumed at P < 0.05. Error bars depict mean ± standard deviation (SD).

Even though FCGBP has been referred to as a mucus-associated protein, no studies to date have reported if FCGBP and MUC2 mRNA expression are coordinately regulated or if the proteins are secreted together in colonic mucus constitutively, and in response to a mucus secretagogue. To address this possibility, we quantified the temporal expression of MUC2 and FCGBP mRNA in WT (27) and CRISPR/Cas9 MUC2 KO LS174T goblet-like cells (MUC2 KO) in culture (23). Following a lag phase from 2 to 6 days, MUC2 and FCGBP mRNA expression in WT cells significantly peaked on days 8 and 7, respectively ( Figure 1A ). In contrast, FCGBP mRNA expression in MUC2 KO cells showed no significant changes in expression between days 2 and 9. To determine if MUC2 and FCGBP mRNA expression correlated with protein production, Western blots were done every 2 days that revealed FCGBP and glycosylated MUC2 (Gly MUC2) protein expression in WT cells increased temporally with peak expression at day 7 consistent with mRNA expression ( Figures 1A, B , Supplementary Figure 1A ). In MUC2 KO cells, FCGBP protein expression similarly increased temporally with peak expression at day 7 ( Figure 1B ). To determine if FCGBP gene expression was inducible similar to MUC2 (28), cells were stimulated with the MUC2 secretagogue, phorbol 12-myristate 13-acetate (PMA) (29). Surprisingly, PMA significantly upregulated the expression of both MUC2 and FCGBP mRNA and protein production in WT goblet-like cells, whereas FCGBP mRNA expression and protein were unchanged in MUC2 KO cells ( Figures 1C, D , Supplementary Figure 1B ). These results demonstrate that both MUC2 and FCGBP associate with each other in WT goblet-like cells and are secreted together in response to a mucus secretagogue. This was in marked contrast to what was observed in MUC2 KO cells. To determine the proteins that were basally expressed in WT and MUC2 KO cells, the global proteome was quantified by quantitative shotgun proteomic analysis ( Figure 1E , Supplementary Table 1 ). WT cells were metabolically active and among the most abundant proteins that were upregulated were MUC2 mucin and heat shock protein beta-1, whereas in MUC2 KO cells, FCGBP and UDP-glucuronosyltransferases were highly upregulated ( Table 1 ). To elucidate insights into the function of the proteome, gene ontology enrichment analysis showed that WT cells were upregulated for genes associated with temperature stimulus, biosynthetic process, protein stabilization and translation, while MUC2 KO cells were highly upregulated for the E-cadherin stabilization pathway ( Figure 1F ). WT cells also showed an increase in biological processes associated with responses to stimulus and metabolic process ( Figure 1G ), which could include responses to mucus secretagogues such as PMA and the highly demanding metabolic process of producing and secreting MUC2 mucin. Using immunofluorescence confocal microscopy, WT goblet-like cells showed ~45% FCGBP that was colocalized with ~80% MUC2 mucin with both proteins abundantly present in the theca at the apical surface of the cell ( Figures 2A, B ). Surprisingly, another pool of FCGBP protein in WT was diffused throughout the cytoplasm not associated with MUC2 ( Figure 2A top left middle panel). Consistent with the proteome data that showed high FCGBP protein expression in MUC2 KO cells ( Table 1 ), FCGBP expression was significantly higher (~40%) and diffusely distributed throughout the cytoplasm in MUC2 KO as compared to WT cells ( Figures 2C, D , top right middle panel). These results demonstrate that FCGBP and MUC2 mucin are coordinately regulated in WT goblet-like cells and suggest that the presence of MUC2 influences the expression and secretion of FCGBP. In addition, a free cytoplasmic pool of FCGBP not associated with MUC2 mucin suggests that FCGBP may have diverse functions within the cell.

To verify if cultured human goblet-like cells show a similar expression profile for Fcgbp and MUC2 in vivo, colonic organoids were generated from Muc2^+/+ and Muc2^-/- littermates. To optimize goblet cell lineage, colonoids were pre-treated with DAPT, a NOTCH inhibitor that increases the number of goblet-like cells (30, 31). The Muc2^+/+ colonoids differentiated with DAPT showed a significant increase in Muc2 mRNA and protein expression but not for Fcgbp ( Figures 3A, B ). The Muc2^-/- colonoids treated with DAPT showed a slight but not significant increase in Fcgbp mRNA expression since these mice retain the presence of goblet-like cells in the absence of Muc2 (32) ( Figures 3A, B ).

To determine if there was a regional difference in basal expression of Muc2 and Fcgbp in the proximal and distal colon of mice, tissues were isolated and analyzed for mRNA and protein expression. We have previously shown that mature goblet cell populations are highest in the proximal colon (33). In accordance, Muc2 mRNA levels were markedly increased in the proximal but not the distal colon of Muc2^+/+ whereas, Fcgbp mRNA levels were similar in Muc2^+/+ and Muc2^-/- littermates ( Figure 3C ). Predictably, glycosylated Muc2 protein was highly expressed in the proximal as compared to the distal colon in Muc2^+/+ littermates ( Figure 3D ). While Fcgbp protein expression was high in the proximal colon of both genotypes, it was absent in the distal colon of Muc2^-/- littermates ( Figure 3D ). Proteomic analysis of proximal/mid colonic tissues revealed that Fcgbp and Clca1 were highly abundant in Muc2^-/- , whereas Muc2 and Fcgbp were abundant in Muc2^+/+ littermates ( Table 2 highlighted). These observations indicate that Muc2 and Fcgbp are coordinately regulated throughout the length of the colon with the highest expression in the proximal colon associated with high numbers of goblet cells as compared to the distal colon.

As gut microbiota are known to affect goblet cell lineage and Muc2 mucin production and function (25, 26, 34–36), we quantified the expression of Muc2 and Fcgbp in specific-pathogen-free (SPF) animals and those that were either untreated or given broad-spectrum antibiotics (Abx) to reduce microbiota diversity and abundance (25, 37) as well as germ-free (GF) mice. SPF mice treated with Abx showed a significantly decreased Muc2 mRNA expression and a greater reduction in Fcgbp (38–40) ( Figure 3E ). GF mice had very low basal Muc2 whereas Fcgbp mRNA expression was high and showed lower Muc2 protein expression in Abx-treated mice and even lower levels in GF mice. Fcgbp protein expression was similarly decreased in Abx-treated SPF mice whereas basal levels in GF mice were high ( Figure 3F ). Fcgbp was not decreased in GF mice as we hypothesized given that the mucus layer was noticeably thinner and less defined in GF mice (39, 41). As antibiotic treatment decreased Fcgbp in SPF mice, we cannot rule out the possibility that the antibiotics themselves may influence the expression of Fcgbp.

As MUC2 and FCGBP are highly colocalized in cultured goblet-like cells ( Figure 2A ), this suggests that perhaps both proteins are biosynthesized and packaged together in mucin granules. To determine if these proteins were tightly bound to each other or are spatially separated within mucin granules, we isolated and purified MUC2 mucin granules. To do this, log-phase goblet-like cells (days 5-6) were homogenized, and following a series of differential centrifugation, the enriched mucin granules were isolated using sucrose gradients ( Figure 4A ). The purified granules enriched for FCGBP and the glycosylated MUC2 were partitioned in fraction 9, with a density of ~1.16 – 1.19 g/ml sucrose, suggesting that both proteins were localized with each other. By electron microscopy, purified mucin granules were distinct and irregular shapes, heterogenous in diameter and size (24, 42), and contained a lipid bilayer ( Figure 4B ). Removal of the lipid bilayer with weak detergents, CHAPS, and octyl-glucoside surprisingly increased immunoreactivity of both FCGBP and MUC2 ( Figure 4C ), clearly indicating that both these proteins were localized and shielded within mucin granules. However, the major question still remained: how do FCGBP and MUC2 biochemically interact with each other via glycan-glycan and/or protein-protein interaction within mucin granules? Recent studies have shown that FCGBP is not covalently bound to MUC2 in secreted mucus (16). As interchain disulfide bonds involve several highly conserved, cysteine-rich protein domains that are usually N- but not O-glycosylated (15), we hypothesized that FCGBP interacts via N-linked glycan-glycan interaction with MUC2 C-terminal interchain bonds. Thus, to confirm the importance of N-linked glycans in MUC2 and FCGBP, goblet-like cells were treated with the N-linked glycosylation inhibitor tunicamycin (18) that showed a shift in the migration pattern of the apoproteins on agarose and SDS-PAGE gels cumulating with a significant decrease in total glycosylated MUC2 mucin ( Figure 4D ). To determine if FCGBP and MUC2 were dissociated from each other in mucin granules in the goblet cell theca (as in Figure 2A in WT cells), cells were imaged basally and after incubation with tunicamycin. As predicted, FCGBP and MUC2 were significantly dissociated with each other following treatment with tunicamycin and the proteins were dispersed throughout the cytoplasm with no discernable goblet cell theca as compared to untreated controls ( Supplementary Figures 2A, B ). Whole cells lysates from tunicamycin treated cells showed shifts in FCGBP (agarose gel) and MUC2 on both agarose and SDS-PAGE acrylamide gels ( Supplementary Figure 2C ). These results demonstrate that FCGBP and MUC2 probably interacted together via N-linked glycans in mucin granules ( Figure 4D ) and dissociated from the goblet cell theca and dispersed in the cell cytoplasm ( Supplementary Figure 2A ). However, we cannot exclude the possibility that N-linked glycans were also necessary for proper intracellular compartmentalization/sorting of FCGBP and MUC2.

Based on the Western blot data showing that granules are enriched for FCGBP and MUC2 mucin ( Figure 4C ), microscopy was used to visualize the colocalization of the proteins. Using widefield microscopy, ~97% of the granules were positive for MUC2 (red) and FCGBP (green) using fluorescent-labeled antibody-tagged proteins ( Figures 4E, F ). As mucin granules are small (0.5 – 2 µm) (42), the resolution of widefield microscopy was not sufficient to discern the spatial organization of the antibody-tagged proteins. To resolve this, we used stimulated emission depletion (STED) super-resolution microscopy to visualize and image the proteins in granules with a high resolution of approximately 30-50 nm (43) vs the resolution of confocal microscopy. STED microscopy revealed that MUC2 (red) and FCGBP (green) proteins were randomly distributed within a granule, were heterogeneous, and not all granules had the same localization of each protein ( Figures 4G, H ). However, in most granules, MUC2 and FCGBP proteins were spatially colocalized with each other (53% and 58%, respectively; Figure 4I ). These results suggest that both MUC2 and FCGBP are highly colocalized in heterogenous subpopulations of mucin granules.

To gain insights into the functional interactions between FCGBP and MUC2 and other putative granule proteins, purified mucin granules were electrophoresed by SDS-PAGE under reducing and non-reducing conditions, and the global proteome of the different molecular weight fractions was resolved using quantitative shotgun proteomic analysis ( Supplementary Table 2 , Figures 5A, B ). Reducing and non-reducing conditions were used to maximize the localization of the MAPs that could interact with MUC2 by disulfide bonding. Proteins of interest related to MUC2 mucin, MAPs, and proteins involved in mucus exocytosis were found in all 4 fractions in reducing and non-reducing gels ( Figure 5B , Table 3 ). Predictably, MUC2 was abundant in fractions 1 and 2 (stacking gel; SG) in reducing and non-reducing gels but was also found throughout the other 2 fractions in the resolving gel. AGR2, which plays a role in MUC2 assembly, folding, and trafficking, was located throughout the gel under non-reducing conditions. Similarly, the SNARE proteins, STXB1, STX3, and SNP23 which are associated with MUC2 mucin exocytosis (24, 45), were located throughout the gel. Importantly, FCGBP was present in fractions 1-3 under reducing conditions and in fractions 1 and 2 under non-reducing conditions. The higher abundance of FCGBP and MUC2 in fractions 1 and 2 (stacking gel) under non-reducing conditions suggests that the interaction between these proteins was easily disrupted by reducing agents. Importantly, of all the MAPs, only FCGBP and AGR2 were identified in purified granules by proteomic analysis. These findings highlight the abundance of FCGBP/MUC2 packaged within mucin granules.

To gain insights into the functions of the mucin granule proteome, the data from reducing ( Figures 5C, G ) and non-reducing ( Supplementary Figures 3A, D ) gels were analyzed for pathway and gene ontology enrichment and protein network using meta-analysis (20) (metascape.org). All 4 fractions showed enrichment in proteins that were involved in high-metabolic activities such as metabolism of lipids, cellular response to stress, glucose metabolism, and vesicle-mediated transport ( Figure 5C ). More specifically, these proteins were associated with metabolic and cellular processes as well as biological processes related to immune system processes and signaling ( Figure 5D ). The gene ontology pathway data demonstrated that the mucin granule proteins are complex, consisting of proteins that are related to a range of processes and functions associated with high metabolic activity. In addition to the gene ontology pathway analysis, the transcription factor for MUC2, SP1 (46), was highly associated with mucin granule proteins ( Figure 5E ). The distribution of the proteins analyzed through a Venn diagram showed overlap of proteins between each of the 4 fractions ( Figure 5F ). This was consistent with the meta-analysis findings that showed enrichment of proteins in all 4 fractions involved in metabolic and cellular processes and transcription factors. Importantly, MUC2 was present in all 4 fractions under both reducing and non-reducing conditions, whereas FCGBP was present in fractions 1-3 in reducing and fractions 1-2 in non-reducing gels.

To elucidate the functional interactions between MUC2 and FCGBP and other MAPs, we used STRING-db v11 analysis database (21) (https://string-db.org). Strikingly, only FCGBP showed multiple interactions with the other MAPs including CLCA1, ZG16, and TFF3. ARG2, CLCA1, ZG16, and KLK1 did not interact with each other ( Supplementary Figure 3D ). Other protein interactions that were identified included the transcription factor NPAS4, adhesion molecules (PCDH17 and CDHR5) and the complement-associated proteins C1QC ( Figure 5G ). The MUC2 protein interactions were predominantly with MUC12, MUC4, MUC16, MUC7, MUC1 and MUC17 ( Figure 5G ). Additionally, numerous polypeptide N-acetyl-galactosaminyltransferases (GALNT14, GALNT4, GALNT2, and GALNT7) known to be associated with mucin glycosylation were identified (39) ( Table 3 ). Importantly, no MAPs other than FCGBP interacted with MUC2 by STRING analysis. Other proteins of interest identified from the proteomic data ( Table 3 ) include CDHR5, an important adhesion molecule between microvilli and enterocytes of the intestine (47). Additionally, multiple N-acetylgalactosaminyltransferases were identified (GALNT1, GALNT2, GALNT3, GALNT4, GALNT5, GALNT6, GALNT7, GALNT10, GALNT12, GALNT13 and B3GALNT2). This family of glycosyltransferases is important in the O-glycosylation of mucins such as MUC2 (48). N-Acetylglucosaminyltransferases were also found including MGAT1, MGAT2, MGT4A, and MGT5A. This subgroup of acetyl-glucosaminyltransferases is associated with N-linked glycans in organelles throughout the body (49, 50). Of particular interest were proteins that had previously been identified in secreted mucus from mouse and human studies and determined to be closely related to Muc2 through proteomic analysis (11, 51). Indeed, several proteins that were identified in secreted mucus throughout the GI tract were identified in the mucin granules ( Table 3 ) including ERP29, TAGL2, S100a6, PIGR, RNH, PSMD2, ELAVL1, CTSD, DCN, FAM3D, GPA33, LGALS3, and LGALS3BP. While no functional significance has been ascribed to these proteins within mucin granules or secreted mucus, it was of interest to observe the overlap of proteins within granules and in secreted MUC2 mucus. Taken together, the proteomic data highlight the complex nature of the mucin granules and demonstrate that the contents contain numerous proteins known to be associated with secreted MUC2 mucus.

As MUC2 and FCGBP were shown to be abundantly expressed, packaged, and localized in mucin granules, we next investigated how these proteins interact and partition in secreted native polymeric mucus gels. To determine if FCGBP and MUC2 were covalently attached to each other (15), secreted mucus from LS174T spend culture media was purified under native conditions (Tris-HCL buffer, pH 8.0) using Sepharose 4B (S4B) column chromatography as previously described (52). High molecular weight MUC2 mucin eluted in the void volume (V₀) ( Figure 6A , fractions 29–36) whereas other non-mucin proteins eluted at different molecular weights fractions designated F1, F2, and F3 (color coded). To visualize the proteins, S4B protein fractions separated by SDS-PAGE ( Figure 6B ) under reducing [left panels, 4% stacking gel (SG) and 12% resolving gel (RG)] and non-reducing [right panel, 4% stacking gel (SG) and 7.5% resolving gel (RG)] conditions revealed that the majority of FCGBP and MUC2 were bound to each other (red boxed area). Under non-reducing conditions, FCGBP was present with MUC2 in the stacking gel (SG; right panels above the red arrowheads), and under reducing conditions, FCGBP separated into several protein fractions with two major bands at 70 and 26 KDa (left panels). Under reducing conditions, most of the 70 KDa FCGBP was unbound with MUC2 and was present in all fractions but was abundant in the low molecular weight F2 fractions (Fractions 70-77). Under non-reducing conditions, the majority of FCGBP and MUC2 were present together in the stacking gel suggesting a role for disulfide bonds between these proteins in secreted mucus. However, STRING-db analysis revealed no protein-protein interactions between FCGBP and MUC2. Instead, FCGBP showed direct interaction with other MAPs including CLCA1 ( Figure 5G , stippled circle), and indeed, blots probed for CLCA1 showed its presence with FCGBP in V₀ under reducing and non-reducing conditions ( Figure 6B , third panel). To elucidate if MUC2 and FCGBP were covalently bound, V₀ MUC2 mucin from S4B was purified using cesium chloride (CsCl) density gradient ultracentrifugation that separates covalently from non-covalently bound proteins (52–54). Quantification of the CsCl density gradient fractions ( Figure 6C ) revealed that FCGBP partitioned exclusively in the low-density fraction 4 (1.31 g/mL) as a 26 KDa protein whereas, MUC2 was present in the high-density fractions 7-9 (1.48-1.58 g/mL) and remained in the stacking gel (SG, Figure 6D ). These findings show that FCGBP and MUC2 are not covalently bound in secreted mucus, contrary to previous speculations (15). Based on this unique interaction, FCGBP-MUC2 binding is probably destabilized in the gut by different pH levels and reducing agents similar to the conditions that occur in acute/chronic inflammation in infectious and inflammatory bowel disease (55, 56).

To quantify how MUC2 and FCGBP in mucin granules partition and merge with each other when secreted constitutively from LS174T cells to form mucus strands, STED immunofluorescence microscopy was used to follow the granule contents ( Figure 7 ). Cells that were actively secreting mucus were probed with fluorescent-labeled antibodies for MUC2 (red) and FCGBP (green) and a fluorescent stain for DAPI (blue). Newly secreted mucus strands from the plume (near the cell surface, nuclei stain blue) to the tail were clearly visible that showed the spatial distribution of FCGBP and MUC2 ( Figures 7A, B ). Regionally, the mucus strand showed that FCGBP ( Figure 7A , green) was predominant in the mid-body, whereas MUC2 (red) was highest at the cell junction with lower quantities of FCGBP ( Figures 7A, C ). At the tail region, the granule proteins lost their structure with granule coalescence to form red mucus strands. The fate of FCGBP and MUC2 in older mucus strands ( Figure 7D ) lost their organization and contained abundant aggregates of MUC2 and FCGBP that were either loosely attached to the cell body or were being pinched off the cell body (as shown). These data demonstrate that FCGBP and MUC2 mucin are secreted together and play a critical role in the spatial organization of the mucus strand.

As both MUC2 and FCGBP were packaged in mucin granules and were secreted together to form the mucus layer, it was of interest to determine how these proteins interacted in wounded cells and in restitution. To do this, WT and MUC2 KO goblet-like cells were plated into 2-well silicone inserts with a defined 50 μm cell-free gap. After the cells became confluent (2 days), the insert was removed, and the cells were allowed to migrate to close the wound gap (open white arrows). Strikingly, he wounds were completely healed by day 2 in MUC2 KO but took until day 6 to heal in WT cells ( Figures 8A, B ). To visualize the partitioning of MUC2 and/or FCGBP during restitution, the cells grown without an insert acted as basal controls and following the removal of the insert, the cell fronts were imaged using confocal microscopy on days 1, 3, and 5. Basally, both MUC2 and FCGBP were colocalized in WT cells, whereas FCGBP was diffusely present throughout the cytoplasm in MUC2 KO cells ( Figure 8C , quantification in Supplementary Figure 4A ). Unexpectedly, once wounded, MUC2 and FCGBP were both highly polarized at the wound margin of WT cells and followed the leading edge of the cells until wound closure at day 6. In contrast, FCGBP unbound to MUC2 in MUC2 KO cells did not partition to the leading edge of the wound, suggesting that cytoplasmic FCGBP might play a role in restitution in the absence of MUC2 that healed within 2 days. The evidence for this was that the expression of MUC2 and FCGBP mRNA were significantly decreased in WT cells during restitution, whereas FCGBP mRNA expression remained high and unaltered in MUC2 KO cells ( Figure 8D ). In both cell types, the stress response gene ATF4 decreased significantly in wounded cells but returned to control levels upon wound closure only in MUC2 KO cells on days 5-7 ( Figure 8D ). Wound healing was mediated primarily by cell migration as the secretory cell lineage transcription factors MATH1 and SPDEF were significantly downregulated post wound ( Supplementary Figure 4B ).

To gain mechanistic insights into why MUC2 KO cells healed within 2 days, secreted growth factors were assessed by a multiplex assay that showed significantly high levels of epidermal growth factor (EGF) in MUC2 KO, but not in WT cells; other growth factors (HGF and PLGF) did not play a role in wound healing ( Figure 9A ). To specifically determine if the altered rates of restitution were caused by differences in cell migration, proliferation, or both, wound closure was quantified using confocal microscopy on the wounded cells on days 1, 2, and 3. Cell migration was quantified by staining for the focal adhesion protein paxillin (57) and proliferation was measured by 5-ethynyl-2’-deoxyuridin (58) (EdU) incorporation. As predicted, there was a significant increase in paxillin and EdU incorporation within 2 days that was pronounced by day 3 upon wound closure in MUC2 KO as compared to WT cells ( Figures 9B, C ). Cell migration and proliferation are the key cellular processes required for wound closure and healing (59). In WT cells, the presence of FCGBP bound to MUC2 at the wound margin associated with significantly less EGF growth factor secretion, cell migration, and proliferation strongly suggests that cytoplasmic-free FCGBP in MUC2 KO cells played a significant role in wound healing.

To assess if the phenomenon we observed in human LS174T goblet-like cells was true of more complex and relevant diseases, the expression of MUC2/FCGBP was determined from human biopsy samples taken from healthy controls and patients with active ulcerative colitis (UC) and UC in remission (UC-R). In active UC patients, MUC2 and FCGBP were significantly upregulated 25-fold and 3-fold, respectively ( Supplementary Figure 4C ), and in UC patients in remission (UC-R), MUC2 and FCGBP expression were similar to healthy controls. These findings suggest that MUC2 and FCGBP are highly regulated during active disease and restitution. Taken together, these data suggest that MUC2 and FCGBP are coordinately altered during epithelial wounding/restitution and in the absence of MUC2, cytoplasmic FCGBP accelerates wound healing.

To determine if Muc2 and Fcgbp are coordinately regulated in an animal model of UC, Muc2^+/+ and Muc2^-/- littermates were given DSS in their drinking water to induce colitis and proximal colonic tissues were collected for RT-PCR and Western blot analysis from days 0 to 15 post-DSS as previously described (26, 60). For consistency, we used the same batch and dosage of DSS as our previous study (26). Weight loss was used as an indicator of disease activity associated with scores for stool consistency, blood in stool, and overall animal appearance in both genotypes. Histologically, tissue damage was associated with a thickened muscle layer, loss of crypt architecture, and cellular infiltration throughout the length of the colon, with most damage occurring distally. In both genotypes, colitis was associated with maximal weight loss from days 4-8 (acute disease) before returning gradually to basal weight by day 17 (healed lesions) post-DSS ( Figure 10A ). In Muc2^+/+ littermates, Muc2 expression was significantly decreased as early as day 4 post-DSS (induction phase) and remained below basal levels throughout to day 15 ( Figure 10B ) as shown previously (26, 61). In marked contrast, Fcgbp mRNA expression increased significantly by day 4 and remained elevated up to 15 days post-DSS ( Figure 10B ). In Muc2^-/- littermates, even though Fcgbp mRNA expression was significantly decreased as early as 4 days post-DSS and remained low up to day 15, colonic protein steadily decreased and was undetectable on days 12-15 post-DSS. Fcgbp protein expression in Muc2^+/+ was undetectable as early as day 4 post-DSS but returned to basal levels by day 15 post-DSS during restitution ( Figures 10C, D ). To determine if goblet cell lineage affected the expression of Muc2 and Fcgbp during acute disease and restitution, we quantified the expression of the secretory cell lineage Math1 and it was significantly downregulated as early as day 4 and remained low up to day 15 post-DSS ( Figure 10E ). As Math1 affects all secretory lineage, we analyzed Spdef expression, the transcription factor critical for terminal goblet cell differentiation and observed a steady increase in Spdef expression during restitution especially on days 12 and 15. As DSS can adversely affect shifts in microbiota abundance (25, 26) and microbiota can influence Muc2 and Fcgbp mRNA and protein expression ( Figure 3F ) that are critical in wound healing, these findings are not comparable to what was observed in cultured LS174T cells. Nonetheless, these results suggest that the restoration of Fcgbp in Muc2^+/+ littermates on day 15 perhaps played an overlooked role in mucosal healing.