Wild-type C57BL/6 mice that were 6 to 7 weeks old, derived from a C57BL/6 background (Jackson Laboratory, Bar Harbor, ME, USA), were kept under controlled conditions of a temperature of 25°C and a 12-h light/dark cycle in autoclaved microisolator filtered cages, with sterile bedding. Mice were fed with irradiated control diet (Research Diets, D09051102) or an autoclavable rodent breeder diet (LabDiet, St. Louis, MO, USA; 50I3) with ad libitum access to food and sterilized water. Before the infection, the mice were free of E. coli, confirmed by fecal samples cultured on MacConkey agar plates.

The EAEC strains used in this study were the wild-type EAEC 042 and a derivative mutant for AAF/II (ΔAAF/II), which was provided by Dr. James Nataro Laboratory, University of Virginia (Charlottesville, VA, USA). E. coli strains HS and 3030 (carrying afa gene) reference strains plus DAEC-MXR strain, isolated from a 1-year-old Mexican child with diarrhea, were included as controls.

Before inoculation, the intestinal pH of C57BL/6 mice was neutralized with NaHCO₃ (0.35 M); then after 15 min, mice were orally inoculated via gavage needle with 9.5 × 10⁹, 1 × 10¹⁰, or 5 × 10⁹ colony-forming unit (CFU) of EAEC 042 or control strains in 100 μl of vehicle buffer (saline solution 0.8%), as previously described by Bernal-Reynaga et al. (2013). In addition, another set of mice was orally inoculated with the vehicle buffer.

Every day, body weight was measured and fecal samples were collected until the day mice were sacrificed by cervical dislocation under chloroform inhalation. Briefly, collected fecal samples were weighed and resuspended in phosphate-buffered saline (PBS); then logarithmic 10⁻¹–10⁻³ serial dilutions were prepared, plated onto MacConkey agar plates, and incubated at 37°C for 24 h Bernal-Reynaga et al. (2013). The next day, colonies were counted, and their concentration was determined and expressed in CFU/g of feces. In order to verify that the recovered bacteria were EAEC 042, a multiplex PCR targeting aggR, aatA, and aap genes was performed on the isolated colonies (Cerna et al., 2003).

At 7 days post-infection, ileum and colonic mucosal tissues were collected, weighed, rinsed with PBS, and then frozen. Tissue DNA was extracted with QuickGene SP kit DNA (FUJIFILM, Tokyo, Japan). The tissue bacterial load was determined by quantitative real-time PCR, screening for aaiC gene (F:5′ ATTGTCCTCAGGCATTTCAC 3′ and R:5′ ACGACACCCCTGATAAACAA 3′) (Panchalingam et al., 2012) in a StepOne Thermo Fisher (Waltham, MA, USA) thermocycle with the following conditions: 50°C 2 min, 95°C 5 min (1 cycle); 95°C 45 s, 58°C 20 s, 72°C 20 s (38 cycles). The melting curve was set at 95°C 15 s, 68°C 1 min, and 85°C 15 s.

Excised ileum and colonic mucosal tissue specimens were fixed with 100% ethanol or Carnoy’s solution (60% ethanol, 30% chloroform, and 10% acetic acid), embedded in paraffin, and cut into 5-µm sections. Sections were stained with H&E and periodic acid-Schiff (PAS) and by immunofluorescence and immunochemistry staining. Mucosa and villus histological evaluations were performed in H&E-stained tissues, and sections stained with PAS were used to evaluate mucus production. Images were processed using ImageJ 2.0.0-rc-65/1.52i software.

Fixed tissues were deparaffinized/dehydrated, and the slides were then treated for antigen retrieval. Briefly, a staining dish containing citrated buffer (0.01 M, pH 6) was introduced into a preheated water bath until the water reached a temperature of 95°C–100°C, then the slides were immersed in the staining dish, and the lid was placed loosely, allowing the slides to incubate for 20 min. Finally, slides were removed and left at room temperature.

For immunofluorescence, slides were incubated for 2 h with PBS containing 5% bovine serum albumin (BSA) to block unspecific binding. Immunolabeling with primary antibodies was performed overnight at 4°C followed by washing and incubation with species-specific fluorescently labeled secondary antibodies conjugated with Alexa 488 (A11008; Invitrogen, Carlsbad, CA, USA) and Alexa 568 (A11077; Invitrogen, CA, USA). The following primary antibodies were used: Alexa 568-labeled phalloidin, anti-ZO-1 (61-7300; Invitrogen, CA, USA), anti-β-catenin ((H-102): sc-7199; Santa Cruz Biotechnology, Texas, USA), anti-MUC1 (ab15481; Abcam, Boston, MA, USA), and anti-EAEC 042 ΔAAF/II (home produced).

Preparations were mounted in a fluorescent mounting medium with DAPI (Vectashield) and examined using a laser scanning confocal imaging system (Leica Biosystems, Wetzlar, Germany; SP8, confocal) at the National Laboratory of Experimental Services (LANSE), CINVESTAV-IPN. Images were processed using ImageJ 2.0.0-rc-65/1.52i software.

From pool stools of the same three mice collected 1 day before EAEC 042 infection and 21 days after infection, DNA was extracted using a QIAamp DNA stool kit (Qiagen, Hilden, Germany) following the manufacturer’s protocol. However, the last column elution was done using 200 μl of nuclease-free water. DNA was quantified using a NanoDrop2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and DNA integrity was verified by PCR assay as previously described by Simon et al. (2014), using primers that amplify a 107-bp fragment of the 16S rRNA gene of E. coli (Simon et al., 2014).

DNA obtained from the stool samples was used for the amplification of the 16S rRNA gene with the primers targeting regions V3–V5 (357-F: 5′-CTCCTACGGGAGGCAGCAG-3′ and CD-R: 5′-CTTGTGCGGGCCCCCGTCAATTC-3′). Sequencing libraries were prepared following the 16S Metagenomic Sequencing Libraries Preparation guide and were sequenced on the Illumina-MiSeq platform at the CINVESTAV-Langebio Laboratory to generate tagged paired-end reads. The obtained sequences have been submitted to the NCBI database under accession number SRR11485509-11485628. Raw fastq files were analyzed by Quantitative Insights Into Microbial Ecology (QIIME2-2019.10) (Bolyen et al., 2019) and processed using the Deblur algorithm to denoise and infer exact amplicon sequence variants (ASVs). The curated ASVs were aligned and annotated by the naive Bayes classifier using the SILVA 132-99nb database.

In order to calculate alpha diversity, the complete ASVs count table was rarefied to 71,364 sequences per sample. Microbiota alpha diversity was quantified as community richness by the Shannon index, while microbiome beta diversity was estimated using samples distances by Bray–Curtis. A two-dimensional scatter plot was generated using non-metric multidimensional scaling analysis (NMDS). The bacterial taxonomic distributions of sample communities were visualized. Taxa abundance was estimated by linear discriminant analysis effect size (LEfSe) (Segata et al., 2011).

Predicted microbial metabolic functions were determined by Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) as described by Douglas et al. (2020). The 16S rRNA metagenome dataset was used to predict metabolic functions using the Kyoto Encyclopedia of Genes and Genomes (KEGG) and the ortholog classification database at hierarchy level 3 pathways.

Results are expressed as mean ± SE. The statistical significance of differences between groups was determined with a t-test or ANOVA test, as appropriate, by using GraphPad Prism 7.2 software. Microbiome alpha diversity comparison was done by the Wilcoxon test. Taxa analysis was performed by LEfSe using R package microeco (Segata et al., 2011). Metabolic pathway comparison was determined by ANOVA test using STAMP v2.1.3 (Statistical Analysis of Taxonomic and Function software) (Parks et al., 2014). p-Values <0.05 were considered significant.

The effect of EAEC 042-infection on the gut barrier was investigated by evaluating the expression of ZO-1 tight junctions and β-catenin adherence junction proteins, on the ileum and colon, after 7 days post-challenge and uninfected mice (control group) by immunofluorescence. As illustrated in Figure 5 , ZO-1 localization remained unaltered in both ileum and colon tissues, suggesting that tight junctions are not affected at this stage of EAEC 042 infection and confirming that tEAEC, which is not an invasive pathogen, does not disrupt the epithelial barrier integrity as do S. enterica serovar Typhimurium and Shigella flexneri, both invasive bacteria (Tafazoli et al., 2003; Meza-segura et al., 2021). On the other hand, EAEC 042 infection influenced β-catenin expression in the ileum since it was observed not only in the latero-basal cell region, adherence junction, and control tissues but also in the apical cell region ( Figure 6 ). In contrast, in colon-infected tissues, β-catenin expression in the latero-basal cell region was diminished in comparison with control colon tissues; however, discrete β-catenin expression was observed in the apical cell region as well ( Figure 6 ). Overall, our results reveal that at 7 days post-EAEC 042 infection, β-catenin expression is differential in ileum and colon tissues. So far, it is not clear if EAEC 042 infection and pathogenesis in humans take place in the ileum or colon, but a significant effect on β-catenin expression observed on ileum tissues suggests that at least in the EAEC 042 murine model, infection impacts ileal tissues more than the colon.

Since EAEC 042 AAF/II plays an important role in pathogenesis, its role during infection was evaluated using EAEC 042 wild-type and EAEC 042 ΔAAF/II mutant strains and antibodies increased against the mutant strain that also recognizes 042 wild-type bacteria. Since growth shortfalls were observed among infected mice at the beginning of the infection ( Figure 1 ), ileal tissues were evaluated for the bacterial presence and β-catenin expression at this stage of infection. As illustrated in Figure 7 , at 4 days post-infection, the wild-type EAEC 042 strain seems to form a biofilm on the apical surface of the ileal villi, as is shown with the red fluorescent dye. In contrast, no biofilm formation was observed on the ileum of EAEC 042 ΔAAF/II-infected mice, and only very few bacteria were observed on the ileum surface ( Figure 7 ). Biofilm formation by strain 042 has been demonstrated in the neonatal mouse model as well (Roche et al., 2010). At 4 days post-challenge, the β-catenin expression on the ileum of EAEC 042 mice was latero-basal and apical, like the expression observed at 7 days post-infection ( Figure 6 ), in contrast with the animals infected with the EAEC 042 ΔAAF/II where β-catenin expression was similar to that observed on the ileum of control mice ( Figure 7 ). In 2017, Boll et al. proposed that MUC1, epithelial transmembrane mucin, is the cell receptor for EAEC 042 strain and that AAF/II mediates this interaction on T84 human colonocytes, facilitating bacterial adhesion (Boll et al., 2017). Therefore, using the newly developed mouse model, the expression and distribution of MUC1 on ileum tissues of EAEC 042, EAEC 042 ΔAAF/II-infected mice, and control mice at 3 days post-infection were evaluated. As illustrated in Figure 8 , once more, a biofilm formation was observed on the apical surface of the ileal villi of mice infected with EAEC 042 and none on the ileum of EAEC 042 ΔAAF/II-infected mice, as previously observed on day 4. It seems that in comparison with control and EAEC 042 ΔAAF/II mice, MUC1 is highly expressed on the surface of the ileal tissues of EAEC 042-infected mice. Notably, on the ileal tissues of EAEC 042-infected mice, we observed colocalization of MUC1 and EAEC 042 bacteria (yellow dots), but not on the ileum of mice infected with the AAF/II mutant.

Together, our results suggest that EAEC 042 strain forms a biofilm on the apical surface of ileal villi of infected mice, and it seems to be dependent on the AAF/II. So far, in vitro experiments have demonstrated the importance of EAEC 042 AAF/II in bacteria adherence assays to T84 cells (Gonyar et al., 2020). Furthermore, to the best of our knowledge, it is reported for the first time that the 042 EAEC strain induces the expression of β-catenin on the apical surface of ileum enterocytes in vivo, which may be a result of the biofilm and/or a result of AAF/II fimbria interaction with a surface molecule. Due to the colocalization between EAEC 042 strain and MUC1, it seems that MUC1 may as well be the AAF/II receptor in mice, as previously shown in in vitro experiments with human T84 cells (Boll et al., 2017). Furthermore, EAEC 042 infection enhances MUC1 expression in human fetal small intestinal xenografts in an AAF/II-dependent manner (Boll et al., 2017).

On the other hand, the expression of β-catenin on the apical surface of ileum enterocytes, only among animals infected with the EAEC 042 wild type and not in animals infected with EAEC 042 AAF/II-mutant, suggests that potentially AAF/II fimbria directly or indirectly participates in β-catenin expression and localization in vivo. It has been reported that the cytoplasmic tail of MUC1 competes with E-cadherin for binding to β-catenin in human colon carcinoma HCT116/V cells (Ren et al., 2002). Furthermore, it has been reported that in human AGS gastric adenocarcinoma cells infected with Helicobacter pylori, MUC1 formed a co-immune precipitation (co-IP) complex with β-catenin and the CagA virulence factor, where MUC1 overexpression reduced CagA/β-catenin co-IP (Guang et al., 2012). In addition, in the absence of MUC1 overexpression in AGS cells, H. pylori infection increased the nuclear level of β-catenin, whereas MUC1 overexpression decreased bacteria-driven β-catenin nuclear localization (Guang et al., 2012). Therefore, it is plausible that in EAEC 042 infection in mice, AAF/II fimbria binds to MUC1 present on the enterocyte surface, inducing β-catenin expression, which binds to the cytoplasmic tail of MUC1 enterocytes at the apical surface. On the other hand, it has been demonstrated that β-catenin is a transcriptional factor that has been proven to induce the expression of IL-8 genes in H. pylori-infected AGS cells (Guang et al., 2012; Ma and Hottiger, 2016). It has been reported that EAEC 042 induces in vitro production of IL-8, and its polymorphisms are associated with different clinical outcomes among EAEC-infected patients and subjects with EAEC-associated diarrhea; individuals with AA IL-8 genotype at position 251 produce greater concentrations of fecal IL-8 than those with the AT or TT genotypes (Steiners et al., 1998; Jiang et al., 2003).