Mutations in the innate immune receptor NOD2 are the greatest single genetic risk factor for Crohn’s disease, yet the mechanisms by which NOD2 regulates intestinal homeostasis remain unclear. We used a CRISPR-generated zebrafish model to determine the impacts of NOD2 deficiency on intestinal health. In cellular, molecular, and transcriptomic studies, we uncovered substantial effects of NOD2 deficiency on epithelial and immune compartments, including deregulated expression of developmental pathways that establish and maintain the gut epithelium, and an unexpected increase in the expression of multiple estrogen-response genes. In functional assays, we uncovered a mechanistic link between estrogenic signals and NOD2-deficiency phenotypes, whereby exposure to estrogen alone replicated the effects of NOD2-deficiency, and treatment with the estrogen receptor modulator tamoxifen reverted the epithelial defects observed in nod2 mutants. Our findings identify a NOD2-estrogen regulatory axis that supports intestinal homeostasis and suggest that hormonal signaling may contribute to sex-specific aspects of Crohn’s disease.

NOD2 deficiency reduces intestinal length but does not alter microbiota composition in zebrafish

To investigate the role of NOD2 in intestinal homeostasis, we used CRISPR-Cas9 mutagenesis to generate nod2 mutant zebrafish that harbor a 4-base pair deletion in the first coding exon, resulting in a frameshift and premature stop codon that produces a severely truncated, non-functional NOD2 protein (Fig 1A). Adult nod2^−/− zebrafish were homozygous viable and showed modest drops in body weight and standard body length by six months (Fig 1B and 1C). In contrast, nod2^−/− adults exhibited an approximately 20% reduction in total intestinal length (Fig 1D). This reduction remained significant after normalization to either body length or body weight, indicating a genotype-specific decrease in relative gut size rather than a secondary effect of overall body scaling (Fig 1E and 1F). Histological analysis suggested architectural differences in the mutant intestine, including reduced epithelial fold height and altered mucosal organization (S1 Fig), features that are examined in more detail in later figures. Using a custom-made rabbit anti-zebrafish NOD2 antibody, we confirmed intestinal NOD2 protein expression in 4-month-old adult WT intestines and validated the absence of functional NOD2 protein in nod2^−/− mutants (S1 Fig).

As interactions between NOD2 and sex are minimally defined, we initially used mixed-sex in-crossed sibling cohorts to define genotype-dependent intestinal phenotypes. Where later experiments identified sex-dependent variation, follow-up analyses were performed with sex recorded and analyzed explicitly. To determine if NOD2 deficiency modifies microbiome composition in zebrafish, we performed a deep-sequencing analysis of the V4 region of the bacterial 16S rRNA gene from 6-month-old adult WT and nod2^−/− siblings. Under cohousing conditions, NOD2 loss did not have discernible effects on microbiota composition, representation, or diversity (Fig 1G and 1K). In contrast, when genotypes were maintained in separate tanks on the same recirculating water system under otherwise identical conditions, mutants showed comparable alpha diversity, but modest differences in beta diversity and relative taxonomic abundance relative to wildtype counterparts (Fig 1L and 1P). These findings mirror data from mice [15,33–35], and highlight the utility of fish for examining homeostatic roles of NOD2.

High-resolution profiling of NOD2-deficient intestines uncovers early immune and epithelial alterations

To better determine the impacts of NOD2 deficiency on gut homeostasis, we performed single-cell RNA sequencing on whole intestines dissected from 6-month-old co-housed, WT, and nod2^−/− siblings derived from a heterozygous in-cross (2 males and 3 females in each group), effectively minimizing artefactual impacts of sex, environment, and genetic variability. As zebrafish intestines are comparatively small but contain most cell types found in mammalian guts [23,27], this approach allowed us to characterize the consequences of NOD2 deficiency for all major intestinal cell types. In total, we generated high-quality transcriptomic data for 11,490 cells, that resolved into 13 distinct transcriptional clusters (Fig 2A and 2B) based on established markers from both zebrafish [23] and mammalian gut studies [27,36], including critical immune-regulatory epithelial, stromal, and hematopoietic lineages. Although nod2 mRNA levels were below detection thresholds, our data revealed broad expression of NOD2 pathway components and signatures of activation across epithelial and immune subsets, closely paralleling expression patterns reported in the human intestine (S2 Fig). Gene ontology analysis uncovered widespread transcriptional changes in nod2^−/− intestines, with impacts on processes related to growth, development, and survival across epithelial and immune compartments, as well as modest shifts in cell type proportions, including apparent decreases in goblet cell numbers and increases in leukocyte abundance (Fig 2C and 2D).

Given the central role of gut-associated leukocytes in CD [37,38] and the high expression of NOD2 in intestinal immune cells [39–43], we elected to characterize the effects of NOD2 deficiency on leukocyte transcriptional states in greater detail (Fig 2E–2H). Gut-associated leukocytes were identified based on transcriptional clustering and canonical immune marker gene expression (S3 Fig). We discovered notable impacts of NOD2 deficiency on expression of co-stimulatory and cytotoxic markers in T cells (Fig 2E); reduced expression of il22 and il7r in ILC3s (Fig 2F); and dysregulated NF-κB signaling in granulocytes (Fig 2H). In macrophages, expression of the CD-linked cytokine cxcl8b.1 was significantly downregulated, while fibrosis and autophagy-associated genes (cd9a, atg4c) were upregulated (Fig 2G). Additionally, nod2^−/− IECs exhibited altered expression of chemokines (e.g., cxcl8a.6, ccl20b, cxcr3.1) and cytokine signaling components (e.g., il2rga, il17rc, il34, S4A and S4B Fig), while stromal cells altered transcripts related to leukocyte activation and adhesion (e.g., cd74a, cd248a, S4A and S4B Fig), pointing to substantial, intestine-wide shifts in the immune landscape of NOD2-deficient adults.

To evaluate the in vivo consequences of these transcriptional shifts, we used immunohistochemistry to quantify the abundance of cells that express the pan-leukocyte Lcp1 marker in 18-month-old NOD2-deficient guts and WT controls (Fig 2I and 2J). Consistent with the widespread impacts of NOD2 deficiency on leukocyte transcriptional states, we observed a significant accumulation of Lcp1⁺ cells in aged nod2^−/− intestines relative to wildtype controls (Fig 2K), confirming that NOD2-deficienct intestines harbor a larger population of functionally dysregulated immune cells.

To define the effects of NOD2 deficiency on leukocyte accumulation within the intestine, we stratified Lcp1⁺ cells by anatomical position along the intestinal fold axis, separating villus tip, villus length (mid), basal, and stromal compartments (Fig 2L). This spatial analysis showed that the increase in Lcp1⁺ cells in nod2^−/− intestines was distributed across epithelial and stromal regions rather than confined to a single niche, consistent with broad immune compartment remodeling rather than a purely localized infiltrate.

NOD2 regulates intestinal epithelial progenitor dynamics

Alongside anticipated effects on immune-regulatory leukocytes, we noted dysregulated gene expression in NOD2-deficient epithelial lineages, including prominent effects on pathways linked to cell growth, development, and survival (Fig 2C). Given the role of the intestinal epithelium in preserving immune homeostasis, we asked if NOD2 modifies epithelial renewal. To answer this question, we focused on cellular features associated with tissue replenishment, starting with progenitor cell proliferation. Using our single-cell RNA sequencing data, we found that NOD2 deficiency significantly attenuated the expression of key growth regulators in the progenitor cell populations (Fig 3A), including critical elements of the Wnt/β-catenin (wnt7ba), TGFβ (tgif1), and growth (myca) pathways (Fig 3B), suggesting that NOD2 deficiency disrupts the molecular framework required for intestinal epithelial growth and renewal. Feature plots confirmed expression of the respective genes within progenitor populations, supporting their relevance to early epithelial renewal programs (Fig 3C and 3D).

To test putative roles for NOD2 in epithelial cell proliferation, we quantified incorporation of the S phase marker EdU in intestines of WT and nod2^−/− larvae at seven days post-fertilization. In this baseline comparison, we quantified EdU⁺ cells from whole-mount confocal images by counting EdU⁺ nuclei across the entire intestine, to obtain an unbiased whole-organ measure of proliferation. We discovered that NOD2 deficiency led to a 50% drop in the number of cycling cells (S5A and S5B Fig), confirming that NOD2 is required for epithelial proliferation in vivo.

To determine if this effect depends on microbial cues, we repeated the EdU labeling assay with seven-day post-fertilization axenic larvae alongside a conventionally reared control population. For these comparisons, we quantified EdU⁺ cells from whole-mount intestines by confocal imaging and counting EdU⁺ nuclei within a standardized intestinal region of interest (ROI), preserving tissue architecture and ensuring consistent quantification across rearing conditions. In agreement with a previous study [29], we found that wildtype larvae exhibited reduced proliferation under axenic conditions (Fig 3F and 3G). Furthermore, and consistent with data in S5 Fig, we observed a near 50% decline of EdU incorporation in conventionally reared nod2^−/− intestines relative to their wildtype counterparts. However, we also discovered that proliferation rates were equivalent between both genotypes in the absence of commensal microbes (Fig 3F and 3G), suggesting that NOD2 promotes epithelial proliferation in response to bacterial cues.

To directly test if the canonical NOD2 ligand, MDP, induces intestinal epithelial proliferation via NOD2, we used flow cytometry as an orthogonal, quantitative approach to complement whole-mount imaging and enable unbiased enumeration of cycling intestinal cells following acute ligand exposure. Using this method, we quantified EdU⁺ cells in dissociated intestines from axenic WT and NOD2-deficient larvae (7 dpf) treated with MDP. We found that MDP treatment more than doubled the number of EdU⁺ cells in axenic WT larvae, establishing it as a potent microbial cue for intestinal growth (Fig 3H). In contrast, the pro-growth effect of MDP was substantially impaired in nod2^−/− larvae with only a minor increase in EdU⁺ cells, demonstrating that the MDP-NOD2 axis links microbial sensing to ISC proliferation.

We then assessed proliferation in WT and NOD2-deficient 6-month-old adults using PCNA labeling, which marks cells in the S phase of the cell cycle. Here, we once more observed a significant reduction in proliferative PCNA⁺ cells in intestines of NOD2-deficient adults relative to WT controls (Fig 3I and 3J). Stratification by sex revealed that this reduction was evident in both males and females, with a stronger effect size observed in females (S6A and S6B Fig). These findings indicate that NOD2-dependent control of epithelial proliferation persists into adulthood and exhibits sex-dependent magnitude. Together, our results establish NOD2 as a regulator of epithelial renewal, linking microbial recognition to epithelial cell proliferation during developmental and adult life stages, and demonstrating an essential role for NOD2 in intestinal growth and homeostasis.

Loss of NOD2 disrupts epithelial homeostasis by altering cell death and differentiation

While NOD2 is an established microbial sensor, its role in epithelial cell dynamics is less well understood. Our transcriptomic analysis indicated that NOD2 deficiency alters the expression of genes involved in proliferation and cell death across all major secretory epithelial lineages, including goblet cells, enteroendocrine cells, and tuft cells (Figs 2C and 4A). As NOD2 dysfunction can impair secretory cell function [44], we determined if NOD2 is required for epithelial cell composition and survival. To characterize the impacts of NOD2 loss on epithelial cell survival and identity in the intestinal epithelial barrier, we quantified apoptotic and Anxa4⁺ cell populations in 6-month-old cohoused WT and nod2^−/− in-crossed siblings (Fig 4B–4D). In fish, the 2F11 antibody labels Anxa4⁺ goblet, tuft, and enteroendocrine cells, as well as recently discovered BEST4 cells. We found that nod2 mutants had a significant increase in TUNEL⁺ cells and a loss of Anxa4⁺ secretory cells relative to WT controls (Fig 4B–4D), indicating elevated apoptosis and loss of secretory cell populations in NOD2-deficient guts. Moreover, analysis by sex revealed that both male and female nod2^−/− intestines showed reductions in Anxa4⁺ secretory cells, an accumulation of Lcp1⁺ cells, and significantly more TUNEL+ cells compared to WT siblings (S7 Fig). These effects were more pronounced in nod2^−/− females, which displayed higher leukocyte infiltration and apoptotic burden relative to nod2^−/− males (S7 Fig), suggesting that NOD2 deficiency disrupts epithelial integrity in both sexes but may have a stronger effect on females.

Beyond cell survival, epithelial architecture depends on transcriptional patterning programs that establish the cellular heterogeneity required for intestinal function. Our differential gene expression analysis uncovered downregulated expression of key patterning genes in nod2^−/− intestines, including map2k4a, epb41l4b, camk2g1, pak1, and sox21a (S8A and S8B Fig), suggesting that NOD2 modifies genetic programs that regulate spatial identity and cell fate. To test roles for NOD2 in epithelial patterning, we quantified Anxa4⁺ secretory cells in intestines of WT and nod2^−/− larvae (7 dpf) raised under either conventional or axenic conditions. Under conventional conditions, nod2^−/− larvae exhibited a significant reduction in Anxa4⁺ cells compared to WT siblings (Fig 4E and 4F). This reduction was not observed in axenic nod2^−/− larvae, suggesting that commensal microbes typically engage NOD2 to promote maturation of secretory lineages in the gut. Supporting this, we discovered that adult nod2^−/− intestines also exhibited shorter villi and fewer goblet cells (Fig 4G and 4H), reflecting impaired secretory lineage differentiation, and mirroring patient data that associate NOD2 mutations with goblet cell loss [45,46].

To directly assess the impact of NOD2 deficiency on goblet cell-associated mucus output, we performed whole-mount Alcian blue staining of intestinal mucus in wild-type and nod2 mutant larvae (7 dpf). We performed the quantification using a fixed, pre-defined color threshold applied uniformly across all images with blinded analysis to prevent observer bias. Because Alcian blue labels mucins rather than goblet cells themselves, this assay reports mucin-positive area and not goblet cell number or density. We found that nod2 mutants displayed a significant reduction in Alcian blue⁺ area under homeostatic conditions (Fig 4I and 4J), indicating reduced goblet cell function. Treatment with dextran sulfate sodium (DSS), a colitogenic polysaccharide that induces epithelial damage and inflammation in vertebrates, led to an expansion of the Alcian blue⁺ mucin signal in both genotypes, with stained area approximately doubling in each. While the relative increase was similar, mucin-positive area remained lower in nod2^−/− larvae, suggesting that epithelial stress can partially rescue goblet cell function through NOD2-independent mechanisms, though the response remains blunted in the absence of NOD2.

Combined, our observations point to a critical role for NOD2 in maintaining epithelial homeostasis, but raise a key mechanistic question: what is the underlying mechanism that contributes to the epithelial phenotypes observed in nod2^−/− mutants?

NOD2 deficiency enhances estrogen signaling in the intestinal epithelium

To investigate the molecular pathways responsible for epithelial disruption in nod2^−/− intestines, we examined our transcriptomic data for signatures of impaired signaling in NOD2-deficient adult intestines and identified enrichment of estrogen response pathways as the most prominently affected (Fig 5A and 5B). The zebrafish genome encodes esr1, and esr2, orthologs of mammalian ERα and ERβ, respectively. In fish, a genome duplication event led to evolution of esr2a and esr2b paralogs from the ancestral esr2 gene. We observed enriched expression of esr2b specifically in IECs (Fig 5C), and widespread upregulation of estrogen-response genes, including esr2b and multiple vitellogenins (vtg1-vtg7), across epithelial lineages in nod2^−/− intestines compared to their wildtype sibling controls (Fig 5D and 5E). Notably, estrogen-response genes were also elevated in immune and stromal subsets of nod2^−/− intestines (Fig 5F and 5G), suggesting that NOD2 deficiency broadly impacts estrogenic signaling across the intestinal cellular landscape. To validate the estrogen-related transcriptional changes observed in our scRNA-seq data, we performed qPCR for vtg1, vtg2, and vtg4 in 6-month-old co-housed WT and nod2^−/− zebrafish intestines, separated by sex (Fig 5H and 5I). We found that, in males, vtg1, vtg2, and vtg4 levels remained unchanged between WT and nod2^−/− (Fig 5H). In contrast, nod2^−/− females exhibited significant increases in vtg gene expression compared to WT controls (Fig 5I), indicating that NOD2 deficiency enhances estrogen-responsive gene expression in a sex-specific manner.

Estrogen signaling modulates intestinal epithelial proliferation, goblet cell abundance, and leukocyte recruitment through a NOD2-dependent mechanism

Given the links between estrogen signaling and epithelial cell proliferation and survival [47–49], we asked if estrogen and NOD2 interact to control intestinal cell fates. To this end, we analyzed epithelial function in WT and NOD2-deficient larvae (7 dpf) that we treated with 17β-estradiol (E2), or the selective estrogen receptor modulator, tamoxifen, which acts as a potent estrogen receptor antagonist in larval zebrafish [50,51]. As larvae are estrogen-sensitive hermaphrodites [52], they provide a unique opportunity to examine sex-independent effects of estrogen on gut physiology. To test if interactions between estrogen and NOD2 modify intestinal renewal, we first quantified intestinal epithelial proliferation in wild-type and nod2^−/− larvae that we treated with E2, or the vehicle control DMSO, for 24 hours. We discovered that E2 treatment reduced the number of proliferative cells by over 50% in wild-type larvae but had negligible impacts on proliferation in NOD2-deficient counterparts (Fig 6A and 6B). Notably, pre-treatment with tamoxifen had the opposing effect. Tamoxifen exposure significantly elevated intestinal proliferation rates in NOD2-deficient larvae without discernible impacts on wildtype controls (Fig 6C and 6D).

Looking at goblet cell function, we detected similar interactions between NOD2 and estrogen. Exposure to E2 significantly reduced the Alcian blue⁺ area of wildtype larval intestines (7 dpf) but had no effect on Alcian blue levels in nod2^−/− intestines. (Fig 6E and 6F), suggesting that epithelial responses to estrogen are maximally engaged in the absence of NOD2. Finally, we found that tamoxifen treatment nearly doubled the intestinal Alcian blue-positive area of nod2^−/− intestines, with no discernible effects on WT siblings (Fig 6G and 6H).

To determine if estrogen-NOD2 interactions also influence the intestinal accumulation of gut-associated phagocytes, we analyzed macrophage populations in intestines of Tg(mpeg1.1:EGFP) larvae that we treated with E2 or tamoxifen for 24 or 6 hours, respectively (7 dpf). Supporting effects of E2 on mucosal defenses, we discovered that E2 treatment more than doubled the number of gut-associated macrophages in wildtype larvae, matching levels found after DSS exposure, whereas tamoxifen exposure significantly reduced macrophage accumulation (Fig 6I–6L). In sum, our work shows that treatment of wildtype larvae with E2 replicates the proliferative and barrier defects noted in NOD2-deficient larvae, while tamoxifen exposure partially rescues barrier defects in nod2^−/− intestines, indicating a role for the NOD2-estrogen axis in the maintenance of intestinal epithelial homeostasis.

Estrogen signaling impairs epithelial recovery in a NOD2-dependent manner

Finally, to assess the consequences of estrogen-NOD2 interactions during colitogenic challenges, we monitored survival rates of wild-type and NOD2-deficient larvae that we exposed to DSS. Kaplan–Meier survival analysis revealed a significant reduction in survival for nod2^−/− larvae following DSS exposure compared to WT controls (Fig 7A), consistent with previous work [31] showing increased sensitivity of NOD2-deficient zebrafish to epithelial injury.

We next asked if estrogen modulates intestinal resilience in a NOD2-dependent manner. To this end, we treated WT and nod2^−/− larvae with E2 prior for 24 hours to a DSS challenge. As anticipated, we found that nod2^−/− larvae exhibited poor survival regardless of E2 treatment status, and we found that wildtype larvae effectively survived an acute DSS challenge, confirming a role for NOD2 in regulating host responses to a pro-inflammatory challenge (Fig 7B). Strikingly, we discovered that E2-treated WT larvae phenocopied the heightened mortality observed in nod2^−/− animals, indicating that estrogen impairs epithelial recovery in a NOD2-dependent mechanism (Fig 7B).

Finally, we tested if tamoxifen pretreatment restores resilience to DSS exposure in nod2 mutant larvae. Remarkably, tamoxifen significantly improved survival in nod2^−/− larvae (Fig 7C), indicating that attenuation of estrogen signaling rescues NOD2-associated defects in epithelial recovery. However, we also found, unexpectedly, that WT larvae displayed enhanced sensitivity to DSS following pre-treatment with tamoxifen. We speculate that this opposing response suggests that the effects of estrogen receptor blockade on intestinal integrity depend on the underlying inflammatory context and basal NOD2 activity, rather than acting as a uniform protective mechanism.

In sum, our work shows that loss of NOD2 significantly elevates E2 signaling in the intestine, that E2 exposure is sufficient to recapitulate several key NOD2-deficiency phenotypes in zebrafish larvae, and that pharmacological attenuation of E2 signaling restores epithelial homeostasis in NOD2-deficient fish, indicating that estrogen and NOD2 oppose each other to regulate intestinal homeostasis.

Zebrafish strains and maintenance

All zebrafish experiments were performed at the University of Alberta using protocols approved by the University’s Animal Care & Use Committees, Biosciences and Health Sciences (#3032), operating under the guidelines of the Canadian Council of Animal Care. Wild-type (TL), nod2^−/− knockout zebrafish and Tg(mep1.1:EGFP) (4–9 months-old; approx. equal sex distribution) were reared at 29°C in tank water (Instant Ocean Sea Salt dissolved in reverse osmosis water maintained at a conductivity of 1,000 μS and pH buffered to pH 7.0 with sodium bicarbonate) under a 14-hour/10-hour light/dark cycle using standard zebrafish husbandry protocols. Feeding regimens were age-specific. From 6 to 14 days post-fertilization (dpf), larvae were fed GEMMA Micro 75 twice daily and live rotifers once daily. From 15 to 28 dpf, larvae received a 1:1 mixture of GEMMA Micro 75 and GEMMA Micro 150 twice daily, along with live rotifers once daily. Juvenile fish (29 dpf to 13 weeks) were fed GEMMA Micro 300 twice daily and live rotifers once daily. Adult zebrafish were maintained on a once-daily diet of GEMMA Micro 300 and live rotifers. For larval analysis, breeding tanks were set up overnight with 1 male and 1 female separated by a divider until morning. Fish were bred for 1 hour, then embryos were collected, rinsed gently with facility water, and transferred to a petri dish (fisherbrand) with 30 mL embryo media (EM) and 30 embryos per dish. Embryos were raised at 29 °C under a 14-hour/ 10-hour light/ dark cycle until 7 dpf.

Where applicable, adult experiments used either mixed-sex cohorts or sex-stratified groups as indicated in the figure legends. Larval experiments were performed at stages prior to sexual differentiation. Experiments were performed in both larval and adult zebrafish to match assay requirements to biological context. Larval models were used where optical transparency, whole-mount imaging, germ-free derivation, and short-term pharmacologic manipulation are required. Adult models were used to assess mature intestinal architecture, long-term epithelial homeostasis, microbiome composition, and sex-dependent phenotypes. Developmental stage is specified for each experiment in the figure legends and indicated by icons in figure panels.

Generation and validation of the nod2-mutant zebrafish line

crRNA:tracrRNA duplex and gRNA:Cas9 RNP complexes were prepared with the nod2 crRNA sequence: TTGGACCTGCTACTTGCTC. Generation and validation of the nod2-mutant zebrafish line was performed as previously described [31], with modifications detailed below. Briefly, Wild-type TL zebrafish embryos were microinjected at the 1–4-cell stage with 1 nL of the CRISPR-Cas9 injection mix per embryo, using a reduced injection volume compared to the original protocol. At 5 days post-fertilization (dpf), larvae were collected for genomic DNA extraction. The genomic region encompassing the CRISPR target site was amplified by PCR, purified, and subjected to Sanger sequencing to assess mutation efficiency. Embryos with evidence of CRISPR-Cas9-induced mutations were raised to adulthood and outcrossed to wild-type TL zebrafish. Genomic DNA was extracted from F1 larvae at 5 dpf and screened for germline transmission of nod2 mutations using high-resolution melting (HRM) analysis. Clutches positive for mutations were raised to adulthood, and heterozygous carriers were identified by HRM, PCR, and Sanger sequencing. Founder fish harboring the same mutation were selected and in-crossed, and F2 embryos were sequenced to confirm the presence of the intended mutation. The resultant mutant line carried a 4-base pair deletion (CTTG) in exon 2 of nod2, resulting in a frameshift and truncated, non-functional protein.

Fish weight, body length, and gut length measurement

Adult zebrafish were anesthetized in Tricaine (MS-222) until complete loss of touch response and then blotted gently on absorbent paper to remove excess water. Body weight was recorded immediately using a precision analytical balance (±0.1 mg sensitivity), with each individual weighed prior to any dissection to avoid dehydration artifacts. Standard body length was measured from the tip of the snout to the base of the caudal fin from lateral images acquired under a dissection microscope using an adjacent metric ruler for calibration.

For gut length measurements, fish were euthanized, and the gastrointestinal tract was carefully dissected under a stereomicroscope using fine forceps. Surrounding tissue was removed without stretching the tissue. The intestine was then laid flat in phosphate-buffered saline (PBS) in a glass Petri dish and imaged with a ruler under a dissection scope. Total gut length was quantified using ImageJ, measuring the longest continuous distance from anterior to distal end. All measurements were performed blinded to genotype, and at least three technical measurements per intestine were averaged to account for handling variability. For size-normalized analyses, gut length was divided by body length or body weight, as indicated in the figure panels.

Generating single-cell suspensions for single-cell RNA-sequencing

Single-cell suspensions were prepared from adult zebrafish intestines as previously described [23], with minor modifications. Briefly, five adult WT TL and five adult nod2^−/− cohoused in-crossed siblings with equal sex distribution (2 males:3 females) were selected for tissue collection. Intestines were dissected, minced, and digested in a dissociation cocktail containing fresh collagenase A (1 mg/mL), 40 μg/mL Proteinase K, and 10× Trypsin (0.25% final). Following digestion, cell suspensions were washed, filtered, and live cells isolated using OptiPrep Density Gradient Medium according to OptiPrep Application Sheet C13. Viable cells were counted and processed for single-cell RNA sequencing as described [23]. Cell suspensions were then diluted to a concentration of 1,200 cell/μL and immediately run through the 10× Genomics Chromium Controller with Chromium Single Cell 3′ Library & Gel Bead Kit v3.1. Library QC and sequencing was performed by Novogene using the Illumina HiSeq platform.

Processing and analysis of single-cell RNA-seq data

Single-cell RNA-seq data were processed and analyzed as previously described [23], with minor modifications. Briefly, Cell Ranger v7.0.1 (10× Genomics) was used to demultiplex raw sequencing files and align reads to the zebrafish reference genome (Ensembl GRCz11.96). Cell Ranger output matrices were analyzed using the Seurat R package (v5.0.1) in RStudio. Cells with fewer than 200 unique molecular identifiers (UMIs), greater than 2,500 UMIs, or greater than 25% mitochondrial reads were removed. Seurat was used to normalize expression values and perform clustering (resolution 0.8, 50 principal components), with optimal PCs determined by JackStraw scores and elbow plots. Marker genes were identified using the “FindMarkers” function, and clusters were annotated using known zebrafish cell type markers or mammalian orthologs.

16S rRNA gene sequencing

Briefly, four adult WT TL and four nod2^−/− 6-month-old in-crossed siblings with equal sex distribution were selected for microbiome analysis. For the co-housing condition, WT and nod2^−/− adults were maintained together in the same recirculating tank for ≥8 weeks prior to sampling to ensure shared environmental microbiota. For the single-housing condition, WT and nod2^−/− adults were maintained in separate tanks on the same recirculating water system under identical water quality, feeding schedule, and density for the same duration; despite shared system water, tank-level separation can still permit divergence in gut microbial communities. All fish were fasted prior to dissections. Zebrafish intestines were dissected with surrounding organs removed. Individual guts were collected and homogenized in 300 μL of Powerbead Solution. The Qiagen DNeasy UltraClean Microbial Kit (Cat No. 12224–250) was used to extract microbial DNA. To assess the intestinal bacterial community composition, the V4 variable region of the 16s rRNA gene encompassed by the 515 forward primer and 806 reverse primer was sequenced. Quality control and sequencing were performed by Novogene Corporation using illumina Novaseq Platform PE250. Sequences were processed with QIIME2-2024.10. The DADA2 pipeline was used to join paired-end reads, remove chimeric sequences, and to generate the feature table used to resolve amplicon sequence variants using default parameters. DADA2 denoising resulted in 694,045 reads for co-housed and 814,688 reads for single-housed. Amplicon sequence variants (ASVs) represented by fewer than 200 reads across all samples were removed. A naïve Bayes classifier trained on SILVA138 99% full-length reference sequences was used to assign taxonomy. Taxonomy assignments were verified using NCBI blast. The sequence table was then filtered to exclude any sequences that were unassigned, not assigned past phylum level, or assigned as eukaryota, resulting in 575,576 sequences (co-housed) or 673,162 sequences (single-housed) corresponding to 25 unique features.

RNA extraction and cDNA synthesis

Total RNA was extracted from zebrafish intestinal tissue using TRIzol Reagent (Thermo Fisher Scientific) following manufacturer protocols with minor modifications. Intestinal samples were collected from five WT TL males, five WT TL females, five nod2^−/− males, and five nod2^−/− females, all derived from cohoused in-crossed sibling groups. Fish were fasted prior to dissection. Individual intestines were isolated, cleared of surrounding tissue, and homogenized in 250 μL of TRIzol using a motorized pestle. Homogenates were brought to 1 mL total volume with additional TRIzol and stored at −70 °C until further processing. For phase separation, 400 μL of chloroform was added, followed by 15 s vortexing and a 3-minute incubation at room temperature. The aqueous phase was collected, and chloroform extraction was repeated at a 1:1 ratio. RNA was precipitated by adding 50 μL of 3 M sodium acetate (pH 5.2) and 1 mL of 95% ethanol, followed by overnight incubation at 4 °C. RNA was pelleted by centrifugation (12,000g, 10 min, 4 °C), washed twice in 0.5 mL 75% ethanol (7,500g, 10 min, 4 °C), and resuspended in 50 μL of nuclease-free water. RNA concentrations were determined using a NanoDrop spectrophotometer and diluted to a standardized concentration for downstream cDNA synthesis. cDNA was synthesized from up to 3 μg of total RNA using the qScript cDNA SuperMix (Quantabio, no. 95048-100) in a total reaction volume of 20 μL. Reactions were incubated at 25 °C for 5 min, 42 °C for 30 min, and 85 °C for 5 min, followed by a 4 °C hold.

Quantitative RT-qPCR

Quantitative PCR was performed using PerfeCTa SYBR Green SuperMix (Quantabio, no. 95054-500) on a One-Step real-time PCR thermocycler. Reactions were carried out in a 10 μL volume containing 2.5 μL of cDNA and 1.6 μM of each forward and reverse primer. Plates were centrifuged at 200g for 1 min prior to thermocycling. The qPCR program consisted of an initial denaturation at 95 °C for 3 min, followed by 40 amplification cycles of 95 °C for 15 s and 60 °C for 45 s, with melt curve analysis included at the end of the run. All reactions were performed in technical triplicates. Cycle threshold (Ct) values were averaged across replicates, and relative gene expression was calculated using the ΔΔCt method, with ef1α used as the endogenous reference gene.

Zebrafish RT-qPCR primer sequences

The following sequences were used (given in order from the 5′ to 3′ end): vtg1 qPCR forward, GCCAAAAAGCTGGGTAAACA; vtg1 qPCR reverse, AGTTCCGTCTGGATTGATGG; vtg2 qPCR forward, TGCCGCATGAAACTTGAATCT; vtg2 qPCR reverse, GTTCTTACTGGTGCACAGCC; vtg4 qPCR forward, TCCAGACGGTACTTTCACCA; vtg4 qPCR reverse, CTGACAGTTCTGCATCAACACA; ef1a qPCR forward, GCATACACTAAGAAGATCGGC; ef1a qPCR reverse, TCTTCCATCCCTTGAACCAG.

Generating germ-free zebrafish

Fish embryos were made germ-free as described in [78]. A clutch of embryos was collected then washed and split into two cohorts. The CV cohort was kept in sterile EM, while the GF cohort was kept in sterile EM supplemented with ampicillin (100 μg/mL), kanamycin (5 μg/mL), amphotericin B (250 ng/mL), and gentamicin (50 μg/mL). Embryos were washed every 2 hours with EM or EM plus antibiotics for CV and GF cohorts, respectively. Once at 50% epiboly, the GF cohort was successively washed three times in EM, then 2 min in 0.1% polyvinylpyrrolidone-iodine (PVP-I) in EM, followed by three EM washes, then a 20-minute incubation with 0.003% sodium hypochlorite (bleach) in EM. Embryos were washed three more times then transferred into tissue culture flasks with sterile EM. The CV cohort received the same number and duration of washes, using EM in lieu of dilute PVP-I or bleach. All work was performed in a biosafety cabinet sterilized first with 10% bleach, followed by 70% ethanol. We tested for bacterial contamination in GF flasks at 4 days post-fertilization, according to established protocol. EM was collected from CV and GF culture flasks to test for bacteria by plating on TSA. Parental tank water and sterile filtered water were used as a positive and negative control, respectively, where bacteria were positively identified in parental tank water and confirmed absent from sterile water. CV and GF flasks with bacteria present or absent, respectively, were used for subsequent analysis.

Flow cytometry of intestinal proliferation in zebrafish larvae

To quantify intestinal epithelial proliferation, zebrafish larvae were immersed in 100 μg/mL 5-ethynyl-2′-deoxyuridine (EdU; 10 mM stock in DMSO) for 24 hours in embryo medium containing 1% DMSO, with or without muramyl dipeptide (MDP) at 1 μg/mL. A 10 mg/ml working stock of MDP (Sigma-Aldrich A9519) was diluted to 1 μg/ml for all treatments. For MDP treatments, axenic larvae were exposed to MDP for 16 hours prior to EdU labeling. Following treatment, intestines were dissected from 15 larvae per condition in ice-cold PBS and subjected to enzymatic dissociation in a cocktail containing 1 mg/mL collagenase, 0.25% trypsin, and 40 μg/mL proteinase K at 37 °C for 30 min with intermittent gentle trituration. Dissociated cells were filtered through a 40 μm strainer, pelleted by centrifugation (300g, 10 min, 4 °C), and washed with PBS. Cells were fixed in 4% paraformaldehyde for 15 min at room temperature, washed with FACS buffer (PBS without calcium or magnesium supplemented with 2% heat-inactivated calf serum), and permeabilized in saponin buffer (0.1% saponin in FACS buffer) for 10 min. EdU detection was performed using the Click-iT EdU Alexa Fluor 488 Imaging Kit according to the manufacturer’s instructions. After EdU labeling, cells were stained with DRAQ5 (1:1,000) for nuclear visualization. Flow cytometry was conducted on the Attune NxT cytometer, and EdU⁺ proliferating cells were quantified based on fluorescence intensity using appropriate controls for gating.

Estrogen and tamoxifen treatment of zebrafish larvae

A working stock of 20 mg/mL 17β-estradiol (E2; Sigma) was prepared in DMSO and diluted to a final concentration of 25 ng/µL in 1xEM for all treatments. No more than 20 zebrafish larvae at 5 dpf were placed in each petri dish and incubated in 25 ng/µL E2 or control DMSO for 24 hours. Larvae were then collected for downstream applications, including immunofluorescence and whole-mount Alcian blue staining. Similarly, a 20 mg/mL stock of tamoxifen (Sigma), dissolved in DMSO, was diluted to a final concentration of 1.5 ng/µL in 1 × EM. Larvae (5 dpf; ≤20 per dish) were incubated in 1.5 ng/µL tamoxifen for 6 hours prior to collection for EdU labeling and Alcian blue staining. For DSS survival assays, larvae were pretreated with E2 for 24 hours or tamoxifen for 6 hours before initiating DSS exposure.

DSS treatment of zebrafish larvae

DSS treatments were performed as previously described [32], with 0.5% DSS (Sigma, cat. no. 42867) used for all experiments and treatments initiated at 5 dpf. Briefly, no more than 20 zebrafish larvae were placed in each petri dish and incubated in freshly prepared 0.5% DSS dissolved in 1× EM for 24 hours. Larvae were collected for further experimental use after removal of DSS. For survival assays, larvae were continuously exposed to 0.5% DSS in 1× EM beginning at 5 dpf. Media was replaced daily by transferring larvae into fresh DSS-containing petri dishes. Survival was monitored over time and quantified by Kaplan–Meier analysis, using 20 larvae per genotype per condition.

Whole-mount Alcian blue staining and quantification

This protocol was adapted and modified from [32]. To visualize goblet cells, zebrafish larvae (7 dpf) were fixed in 4% PFA at 4 °C for 48 hours, rinsed in 1× PBS for 5 min, and incubated in acidic ethanol (70% ethanol containing 1% concentrated HCl) for 10 min. Alcian blue stains secreted acidic mucins rather than goblet cells directly; therefore, this assay quantifies mucin-positive area and does not measure goblet cell number or density. Larvae were stained in 0.1% Alcian blue (prepared in 3% acetic acid, pH 2.5) for 3 hours at room temperature, washed three times in acidic ethanol, and incubated overnight in acidic ethanol to reduce background staining. Larvae were mounted in 0.7% low-melting-point agarose and imaged under identical illumination and exposure settings. For quantification, images were analyzed in Fiji (ImageJ) using blinded analysis. A region of interest (ROI) corresponding to the mid and distal intestinal segments, defined anteriorly by the hepatic bend and posteriorly by the cloaca, was outlined while excluding extra-intestinal tissues. Images were converted to RGB, and a single fixed color threshold for Alcian blue signal was established using control images and then applied uniformly to all samples within the experiment without per-image adjustment. Threshold values were held constant across genotypes and treatments to avoid segmentation bias. Alcian blue-positive regions were segmented using this fixed threshold, and the Alcian blue-positive area and total ROI area were measured using the Analyze Particles tool. Data are reported as percent Alcian blue-positive intestinal area, calculated as (AB⁺ area/total intestinal ROI area) × 100, reflecting the proportion of the intestinal surface occupied by goblet cell-associated mucin staining. Each dot in quantification plots represents a single larval intestine.

Immunofluorescence and fluorescence microscopy

Adult zebrafish intestines were fixed in 4% PFA (diluted in 1xPBS) overnight at 4 °C. Intestines were washed twice in 1× PBS. Intestinal segments were then cryoprotected in 15% (w/v) sucrose/PBS at room temperature until sunk followed by sinking in 30% (w/v) sucrose/PBS (overnight at 4 °C). Intestines were mounted in Optimal Cutting Temperature Embedding Medium, then frozen on dry ice. Seven-micrometer cryosections were collected on Superfrost Plus slides. After allowing sections to adhere onto slides, slides were immersed in PBS for 20 min at room temperature. Tissue was blocked for 1 hour at room temperature in 3% (w/v) BSA dissolved in PBST (1× PBS + 0.2% (v/v) Tween 20). Primary antibodies were diluted in blocking buffer. Sections were incubated in primary antibody solution overnight in a humid chamber at 4 °C. The primary antibodies used in this paper was mouse anti-zebrafish gut secretory cell epitopes antibody [FIS 2F11/2] 1/500 (abcam 71286), and chicken polyclonal anti-GFP (ThermoFisher Scientific Cat# PA1-9533). Excess primary antibody was washed by immersing slides in fresh PBST for 1.5 hours. Sections were incubated in secondary antibody solution (prepared in blocking buffer) for 1 hour at room temperature in a humid chamber, protected from light. To stain apoptotic cells, secondary solution was removed then sections incubated in TUNEL reaction mixture following the manufacturer’s suggestion (Roche 12156792910). Lastly, nuclei were stained with Hoechst (Molecular Probes Cat# H-3569) diluted 1:2,000 in PBST for 10 min protected from light. Slides were washed in PBST by brief immersion followed by a 30-min incubation in fresh PBST protected from light. Slides were mounted in Fluoromount Aqueous Mounting Medium.

Zebrafish larvae (6–7 dpf) were incubated in embryo medium with 1% DMSO and 5 mM EdU for 8 hours at 29 °C. Fish intestines were dissected in PBS and fixed in 4% paraformaldehyde in PBS overnight at 4 °C. Guts were washed three times with PBSTx (0.75% Triton X-100 in 1× PBS), blocked for 1 hour in PBSTx + 3% BSA at room temperature, and stained with primary antibodies in blocking buffer overnight at 4 °C. Guts were washed with PBSTx and then stained for 1 hour at room temperature with secondary antibodies and nuclear stain (1/1,000 Hoechst 33258), followed by rinse in PBSTx. EdU detection was performed by incubating guts in Click-iT reaction cocktail for 30 min at room temperature. Guts were washed in PBSTx followed by extra washing in PBS. Whole larval intestines were mounted on microscope slides using Flouromount and visualized.

EdU proliferation quantification

For imaging-based proliferation assays, whole-mount intestines were imaged by confocal microscopy using identical acquisition settings across experimental groups. EdU⁺ nuclei were quantified either across the entire intestine (baseline genotype comparisons) or within a standardized intestinal ROI spanning the mid-intestine (rearing-condition comparisons), depending on experimental design. Nuclei were identified by DNA staining and EdU⁺ cells were counted using blinded analysis. Percent EdU⁺ values were calculated relative to total epithelial nuclei within the quantified region. Flow cytometry-based EdU quantification was performed on dissociated intestines where indicated to enable unbiased enumeration following acute ligand stimulation.

Quantification of Anxa4⁺, Lcp1⁺, and TUNEL⁺ cells

Quantification of Anxa4⁺ secretory cells, Lcp1⁺ leukocytes, and TUNEL⁺ apoptotic cells was performed on sagittal intestinal sections using Fiji (ImageJ). Individual villi were manually outlined using the DNA (nuclear) channel to define the epithelial region of interest. Within each villus ROI, Anxa4⁺ , Lcp1⁺ , and TUNEL⁺ cells were detected using a fixed threshold applied uniformly across all samples and counted manually using the cell counter tool. Total epithelial cell number per villus was determined by counting all DNA⁺ nuclei within the same ROI. To account for differences in villus size and cellular density between sections and fish, data were expressed as the percentage of marker-positive cells per villus, calculated as (number of Anxa4⁺ , Lcp1⁺ , or TUNEL⁺ cells ÷ total number of DNA⁺ nuclei within the ROI) × 100. Multiple villi were analyzed per fish to capture intra-intestinal variability. For statistical analysis, measurements from individual villi were averaged per fish, and each fish was treated as a single biological replicate (n = 5 per group), thereby avoiding pseudoreplication. Where indicated, individual villus measurements are displayed to illustrate within-sample variability.

Immunohistochemistry

Adult zebrafish intestines were fixed at 4 °C in BT fixative: 4% PFA, 0.15 mM CaCl₂, 4% Sucrose in 0.1 M phosphate buffer (pH 7.4). Intestinal segments were processed for paraffin embedding and 5 μm sections collected on Superfrost Plus slides. Tissue was deparaffinized, rehydrated, then boiled in 10 mM sodium citrate pH 6, 0.05% Tween 20 for 20 min at 98 °C to unmask antigen. After cooling sections to room temperature for 30 min, sections were incubated in 3% hydrogen peroxide for 10 min, washed twice in dH₂O, and once in PBSt (PBS + 0.5% Triton X-100). Sections were then blocked for 1 hour at room temperature in 10% (v/v) normal goat serum (NGS)/PBSt followed by overnight incubation in primary antibody solution (prepared in blocking buffer) in a humid chamber at 4 °C. The primary antibodies used in this paper were: mouse anti-PCNA 1:5,000 (Sigma P8825), rabbit anti-Lcp1 1:1,000 (GeneTex GTX124420) and a custom anti-zebrafish NOD2 antibody (1:1,000, Biologics International Corp (BIC)). Slides were washed three times in PBSt and tissue incubated in SignalStain Boost Detection Reagent (HRP, Rabbit or Mouse) for 30 min at room temperature. Colorimetric detection was done by incubating in SignalStain DAB Chromogen solution for 5 min. Sections were counterstained with Alcian blue (Aldrich 19,978-8) for 3 min, counterstained in quarter-strength Hematoxylin Gill III (Leica Ca.# 3801542) for 30 s, dehydrated in ethanol, cleared with toluene and mounted in DPX.

Microscopy

Histological samples were imaged on a ZEISS AXIO A1 compound light microscope with a SeBaCam 5.1MP camera. Confocal images were captured on an Olympus IX-81 microscope fitted with a CSU-X1 spinning disk confocal scan-head, Hamamatsu EMCCD (C9100-13) camera and operated with PerkinElmer’s Volocity. Confocal z stack image processing was done in Fiji.

Quantification and statistical analysis

Statistical analyses were performed using R. The statistical tests, group statistics, and P values used in each experiment are described in the respective figure legends.

S1 Fig. NOD2 protein expression in adult WT and nod2^−/− intestines.

A–D) Representative sagittal sections of 4-month-old cohoused WT and nod2^−/− adult intestines immunostained with a custom anti-zebrafish NOD2 antibody. Images were taken at 20× (A, C) and 40× (B, D) magnification. n = 5 fish per genotype (mixed sex). Scale bars = 50 μm. Icon indicates the developmental stage at which the experiment was performed.

https://doi.org/10.1371/journal.pbio.3003766.s001

(TIF)

S2 Fig. Single-cell expression of nod2 across epithelial, stromal, and immune lineages in WT adults.

A, B) tSNE plot (left) of single-cell RNA-seq from WT adult epithelial and stromal cells (n = 3,613), annotated by cell type. Heatmap (right) shows average expression of nod2 and NOD2-pathway genes across epithelial and stromal clusters. C, D) tSNE plot (left) of single-cell RNA-seq from WT adult intestinal immune cells (n = 823), annotated by lineage. Heatmap (right) shows average expression of nod2 and pathway-associated genes across immune subsets. The data underlying the graph in Figure are available in S11 Data.

https://doi.org/10.1371/journal.pbio.3003766.s002

(TIF)

S3 Fig. Identification of intestinal immune cell populations in WT and nod2^−/− intestines.

A) tSNE visualization of leukocytes from WT (left) and nod2^−/− (right) zebrafish intestines (6-month-old, cohoused), colored by immune cell type based on transcriptional clustering and marker gene expression. Major immune subsets identified include T cells, ILC3s, ILC2s, macrophages, granulocytes, B cells, and dendritic cells (DCs). B) Dotplot representation of the expression levels of established markers in the indicated cell types from integrated data set. The data underlying the graph in Figure are available in S12 Data.

https://doi.org/10.1371/journal.pbio.3003766.s003

(TIF)

S4 Fig. NOD2 deficiency enhances leukocyte-recruitment gene signatures in epithelial and stromal compartments.

A) tSNE plots of intestinal epithelial cells (IECs; blue) and stromal cells (pink) from 6-month-old, cohoused WT and nod2^−/− zebrafish. B) Heatmaps show differential expression of immune cell recruitment-associated genes in IECs and stromal cells. (For all genes adjusted P-value ≤ 0.05). The data underlying the graph in Figure are available in S13 Data.

https://doi.org/10.1371/journal.pbio.3003766.s004

(TIF)

S5 Fig. NOD2 is required for epithelial proliferation in the developing zebrafish gut.

A) Immunofluorescence images of whole-mount intestines from conventionally reared WT and nod2^−/− larvae (7 dpf), shown in rostral-to-caudal orientation from the base of the intestinal bulb to the cloaca. EdU⁺ cells (magenta) and nuclei (blue, DNA) are shown in merged false-colored images. White arrowheads highlight regions where nod2^−/− intestines lacking EdU⁺ cells. Scale bars = 150 μm. B) Quantification of total EdU⁺ cells per gut in WT and nod2^−/− larvae (7 dpf), determined from whole-mount confocal images by counting EdU⁺ nuclei across the entire intestinal epithelium. Each point represents an individual larva; nod2^−/− intestines exhibit a significant reduction in proliferating cells. P-values were calculated using Mann–Whitney U test. Icon indicates the developmental stage at which the experiment was performed. The data underlying the graph in Figure are available in S14 Data.

https://doi.org/10.1371/journal.pbio.3003766.s005

(TIF)

S6 Fig. Loss of NOD2 decreases intestinal epithelial proliferation with sex-specific effects.

A) Representative sagittal sections of posterior intestine from 6-month-old WT and nod2^−/− zebrafish, separated by sex. Sections were stained for proliferating cell nuclear antigen (PCNA; brown) and counterstained with Alcian blue to visualize goblet cells (cyan). Images are shown in rostral-to-caudal orientation and were collected from comparable posterior intestinal regions using identical imaging settings. Scale bars = 20 μm. B) Quantification of epithelial proliferation shown as the percentage of PCNA⁺ epithelial cells per villus in WT and nod2^−/− intestines, analyzed separately in males and females. Multiple regions of interest were analyzed per fish; measurements were averaged per fish, and each point represents an individual measurement obtained from multiple images per fish intestine (n = 5 zebrafish per genotype). P-values were calculated using Mann–Whitney U test or Kruskal–Wallis test. Icon indicates the developmental stage at which the experiment was performed. The data underlying the graph in Figure are available in S15 Data.

https://doi.org/10.1371/journal.pbio.3003766.s006

(TIF)

S7 Fig. NOD2 deficiency alters epithelial and immune cell populations and increases apoptosis in a sex-dependent manner.

A) Representative sagittal sections of WT and nod2^−/− intestines (6-month-old, cohoused adults) stained for DNA (nuclei), secretory cells (2F11/Anxa4), pan-leukocyte marker (Lcp1), and apoptotic cells (TUNEL). Merged panels show combined staining with outlined villus structures. Images are shown in rostral-to-caudal orientation. B–D) Quantification of % Anxa4⁺ cells (B), % Lcp1⁺ cells (C), and % TUNEL+ cells (D) from multiple sections of WT and nod2^−/− intestines. Multiple regions of interest were analyzed per fish; measurements were averaged per fish, and each point represents an individual measurement obtained from multiple images per fish intestine (n = 5 zebrafish per genotype). Scale bars = 50 μm. P-values were calculated using Mann–Whitney U test or Kruskal–Wallis test. Icon indicates the developmental stage at which the experiment was performed. The data underlying the graphs in Figure are available in S16 Data.

https://doi.org/10.1371/journal.pbio.3003766.s007

(TIF)

S8 Fig. Loss of NOD2 impairs expression of pattern specification genes in the intestinal epithelium.

A) Heatmap illustration of representative pattern specification processes gene expression in nod2^−/− versus WT adult intestinal epithelium (for all genes, adjusted p-value ≤ 0.05). B) Feature plots showing relative expression of select patterning genes (bambia, camk2g1, pak1, sox21a) in WT and nod2^−/− intestinal epithelium. Expression is reduced across enterocyte (EC) and lower right epithelial (LRE) clusters in nod2^−/− intestines, indicating broad loss of epithelial patterning cues. The data underlying the graphs in Figure are available in S17 Data.

https://doi.org/10.1371/journal.pbio.3003766.s008

(TIF)