Lactococcus lactis NCDO 2118 wild-type (L. lactis WT) strain was grown at 30°C in M17 medium (Difco) containing 0.5% glucose (GM17), without agitation, or in the same medium solidified with 1.5% agar for 18 h. The nucleotide sequence encoding the SlpB surface protein from Propionibacterium freudenreichii CIRM-BIA 129 (Pf 129) was obtained from the database of the National Center for Biotechnology Information (NCBI), deposited under accession number CDP48273.1 (CDS 5503..7173). The sequence was optimized for expression in Lactococcus lactis NCDO2118 bacteria in the OptimumGeneTM program (GenScript Corporation) and synthesized by GenScript Corporation (Piscataway, NJ, USA) and cloned into pUC57 vector. The optimized SlpB protein sequence was synthesized with the restriction sites NsiI-3′ and EcoRI-5 to cloning into the plasmid pXYSEC (chloramphenicol resistance). The NsiI/EcoRI digested and purified SlpB ORF and pXY:SEC fragments were ligated by T4 DNA ligase (Invitrogen) to obtain the pXYSEC:slpB plasmid, which was established by transformation in E. coli Top10 and selected with 10 μg/ml of chloramphenicol (Cm) in Luria Bertani Agar (Miyoshi et al., 2004). Routinely, the primers SlpB-Forward 5′-GAT​CCC​CCG​TCT​GAA​CGA​ACT​T-3′ and SlpB-Reverse 5′-CGA​CAT​CAT​TGA​ACA​TGC​TGA​AGA​GC-3′ were used for plasmid construction verification by PCR and agarose gel, and also for sequencing (PCR product size: 1,816 bp). Then, the optimized gene sequence for SlpB was subcloned into the pXYSEC plasmid and competent L. lactis NCDO 2118 bacteria was transformed by electroporation as previously described by Langella et al. (1993), and grown at 30°C in M17 medium (Difco) containing 0.5% glucose (GM17) without agitation containing 10 μg/ml of chloramphenicol. To confirm the final construction of pXYSEC:slpB, a DNA sequencing was performed by fluorochrome-labeled dideoxynucleotides method (BigDye Terminator v3.1 Cycle Sequencing, Applied Biosystems, USA). Recombinant L. lactis NCDO2118 strain was grown in Difco M17 broth, supplemented with either 0.5% glucose (GM17) or 1% xylose (XM17) and chloramphenicol (10 μg/ml) at 30°C without agitation. On the first day, single colonies of recombinant L. lactis NCDO2118 harboring pXYSEC:slpB (L. lactis-SlpB) were cultured in 5 ml of GM17. On the second day, the overnight culture was diluted 1:10,000 in XM17 to induce the expression of the slpB gene ORF. Proteins sample preparation from L. lactis wild type and recombinant L. lactis-SlpB cultures was performed as previously described (Miyoshi et al., 2004). To verify protein production, a Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS–PAGE) and Western blotting was done as previously described by Do Carmo et al., 2017.

Conventional female C57BL/6 mice of 8 weeks of age, obtained at the Universidade Federal de Minas Gerais (UFMG, Belo Horizonte, Brazil), were used in this work. They were housed in plastic cages in a room with controlled temperature (18°C–23°C), light cycle 14 h light/10 h dark, relative humidity (40%–60%), and ad-libitum access to food and water. All experimental procedures realized in this work were approved by the Ethics Committee on Animal Experimentation of the Universidade Federal de Minas Gerais (CEUA-UFMG, Brazil) by the protocol no. 148/2020.

L. lactis NCDO2118 WT (L. lactis) and L. lactis NCDO2118 pXYSEC: slpB (L. lactis-SlpB) strains were prepared daily for the animals, using intragastric gavage as a form of administration. Both strains were grown in M17 medium (Difco) added with glucose (0.5%), for the wild-type strain, and xylose (1%) + 10 μg/ml of Cm, for the L. lactis-SlpB strain. Bacteria cultures (L. lactis and L. lactis-SlpB), were incubated for 24 h at 30°C, and 1 ml of each culture was centrifuged at 3,500 rpm for 10 min and washed with PBS pH 7.4 twice to remove the antibiotic. Thus, each bacterial pellet corresponds to a daily dose with 5 × 10⁹ CFU/per dose of bacteria, which was then resuspended in 100 µl of PBS pH 7.4.

The mice were divided randomly into four main groups, each containing six animals per group (Supplementary Figure S1). Group 1 represented a healthy control group that received only water for drinking (control group). The mice from groups 2–4 (experimental groups) received DSS (36–50 kDa, MP Biomedicals, CAT 260110, LOT Q5756), as the only drinking source, prepared to a concentration of 1.7% in filtered drinking water and provided to the animals daily, during 7 days, according to the acute colitis model previously described (Wirtz et al., 2017). Animals from group 2 received only DSS solution (group DSS) and no treatment. Mice from groups 3 and 4 received the bacteria dose during the 7 days by gavage (all experimental days), together with DSS in drinking water. Precisely, mice from group 3 received intragastric doses (100 µl containing 5 × 10⁹ CFU) of L. lactis NCDO 2118 WT (group DSS + NCDO 2118 WT), and animals in group 4 received (100 µl containing 5 × 10⁹ CFU) of L. lactis NCDO 2118 pXYSEC:slpB (group DSS + NCDO2118 pXYSEC:slpB). The mice were euthanized on the seventh day. All in vivo experiments were done in biological triplicate.

During all experimental days, the water, food intake, and mice body weight were recorded daily. On the last experimental day, the disease activity index (DAI) was determined, as described by Cooper et al. (1993), attributing a score of the three major colitis clinical signs: weight loss, intensity of diarrhea, and presence of rectal bleeding.

A longitudinal abdominal incision was performed in all mice to access the intestine and colon, and then to be used in future analyses. The colon length of each mouse (measured from the cecum to rectum) were used to indicate the mean of each experimental group (cm). The colon distal part was collected, washed with PBS, and stored in segment rolls for histomorphological analysis. These rolls were immersed in formaldehyde solution (4%, v/v) for tissue fixation and, after that, they were embedded in paraffin. A section (4 µm) was placed on a glass slide and stained with Hematoxylin-Eosin (HE) (Marchal Bressenot et al., 2015). The sections were photographed (×20 magnification objective) using a digital camera (Spot Insight Color) coupled to an optical microscope (Olympus, BX-41, Japan). The histological inflammation score was determined by a pathologist using the score previously described by Wirtz et al. (2017). Consider: tissue damage (0: none; 1: isolated focal epithelial damage; 2: mucosal erosions and ulcerations; 3: extensive damage deep into the bowel wall) and lamina propria inflammatory cell infiltration (0: infrequent; 1: increased, some neutrophils; 2: the submucosal presence of inflammatory cell clusters; 3: transmural cell infiltrations). The total score ranging from 0 (no changes) to 6 (widespread cellular infiltrations and extensive tissue damage) were obtained by the sum of these two sub-scores (tissue damage and lamina propria inflammatory cell infiltration). Other cuts of the paraffinized colon samples were produced and stained by the periodic acid-Schiff (PAS) (Prisciandaro et al., 2011) in order to count the mucus-producing goblet cells. Ten random field images from each sample were made using the ×40 objective, and then, using ImageJ software (version 1.8.0) the intact goblet cells were counted. The total number of goblet cells was expressed as the number of cells per high-power field (HPF) (×40, 108.2 µm²).

Neutrophil infiltration levels in the colon tissue were assessed by measurement of myeloperoxidase activity (MPO), as previously described by Porto et al., 2019. For MPO quantification, a piece of colon tissue (100 mg) was homogenized proportionally in 1.9 ml/100 mg of PBS and centrifuged at 10,000 × g for 10 min. The pellet formed was lysed and centrifuged again. The pellet formed was resuspended proportionally in 1.9 ml/100 mg of 0.5% HTAB (hexadecyltrimethylammonium bromide) diluted in PBS. The suspension was submitted to freeze–thaw cycle (3x) using liquid nitrogen and then, centrifuged at 12,000 × g at 4°C, for 10 min. In order to perform the enzymatic assay, we added an equal amount of substrate (1.5 mM L⁻¹ of o-phenylenediamine and 6.6 mM L⁻¹ of H₂O₂ in 0.075 mM L⁻¹ of Tris–HCl pH 8.0) to the supernatant. To stop the enzymatic reaction, 50 μl of 1 M H₂SO₄ was added. The absorbance was read in a spectrophotometer (Spectramax M3, Molecular Devices, LLC, Sunnyvale, CA, USA), at 492 nm.

The extent of eosinophil infiltration into the tissues was assessed by measuring eosinophil peroxidase (EPO) activity, as previously described by Vieira et al. (2009). For EPO quantification, a piece of colon tissue (100 mg) was homogenized proportionally in 1.9 ml/100 mg of PBS and centrifuged at 10,000 × g for 10 min 4°C. The precipitate was subjected to hypotonic lysis, where 0.9 ml of a solution containing 0.2% NaCl was added prior to the addition of an equal volume of solution containing 1.6% NaCl and 5% glucose. The samples were again homogenized and centrifuged (10,000 × g, at 4°C, for 10 min). The supernatant was discarded, and the pellet was resuspended in 1.9 ml of 0.5% HTAB (hexadecyltrimethylammonium bromide) diluted in PBS. After three cycles of freeze–thaw in liquid nitrogen, the samples were centrifuged at 4°C, 10,000 g for 10 min. To test the enzyme activity, the obtained supernatant was mixed with a substrate (1:1) containing 1.5 mmol/L of o-phenylenediamine, 6.6 mmol/L of H₂O₂, and 0.075 mmol/L of Tris-HCl pH 8. After 30 min the reaction was stopped with 50 μl of 1 M H₂SO₄. The absorption was measured in a spectrophotometer (Spectramax M3, Molecular Devices, LLC, Sunnyvale, CA, United States) at 492 nm.

The secretory immunoglobulin A (sIgA) of the intestinal lavage was determined by ELISA, according to Cordeiro et al., 2021. For the quantification of the samples was used a 96 well-plates (Nunc-Immuno Plates, MaxiSorp) coated with anti-IgA antibodies (Southern Biotechnology, Birmingham, AL, United States) and incubated overnight. After the incubation, the plates were washed in saline-Tween (saline with 0.05% of Tween-20—SIGMA Chemical Co.) and blocked with 200 µl of PBS-casein (0.05%) for 1 h, at room temperature. After that, the intestinal lavage contents were added, and the plate was serially diluted (1:100) and incubated at room temperature for 1 h. Plates were washed with saline-Tween and then, biotin-conjugated anti-mouse IgA antibodies were added (Southern Biotechnology) (1: 10,000 in PBS-casein). Plates were incubated for 1 h at 37°C and then, biotinylated monoclonal antibodies anti-IgA (BD Bioscience) were added and incubated for 1 h at room temperature. Following this, peroxidase-labeled streptavidin (Southern Biotechnology) was added. Plates were washed in saline-Tween and incubated again with 100 µl of orthophenylenediamine (OPD) (Sigma, St. Louis, MO, USA) and H₂O₂ (0.04%), for 1 h, at room temperature. To stop the reaction, 20 µl/well of 2N H₂SO₄ was added. Absorbance reading was performed on Bio-Rad Model 450 Microplate Reader, at 492 nm. The results of total sIgA were measured, according to the standard curve, in a concentration of sIgA (ng) per ml of intestinal fluid.

In order to obtain the quantitative gene expression in colon fragments, the methodology was carried out according to Do Carmo et al. (2020). Fragments of 1 cm of the colon were collected. Total RNA was isolated using PureLink RNA Mini Kit (Thermo Fisher Scientific) according to the protocol of the manufacturer. Afterward, DNase I (Invitrogen; Waltham, MA, USA) was used to digest residual genomic DNA of samples, and then Turbo DNA-free Kit (Ambion; Austin, TX, USA) was used for DNA removal following the protocol of the manufacturer. RNA quality was checked by agarose gel and NanoDrop^® ND-1000 (260/230 ratio). To obtain the samples cDNA the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems; Foster City, CA, United States) was used. Quantitative PCR (qPCR) was determined using iTaq universal SYBR green supermix (Biorad; Hercules, CA, United States) and gene specific primers (Table 1), for Mucin 2 (muc-2), Zonula occludens 1 (zo-1), zonula occludens 2 (zo-2), Claudin-1 (cln-1), Claudin-5 (cln-5), Occludin (ocln), inducible nitric oxide synthase (inos), peroxisome proliferator-activated receptor-gamma (pparg), and cytokine genes for interleukin-10 (il-10), il-17, as well as housekeeping genes encoding β-actin (actβ) and GAPDH (gapdh). The amplification cycles were performed as described: 95°C for 30 s, and 40 cycles of 95°C for 15 s and 60°C for 30 s on ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). Results were expressed as a fold-change of expression levels, using the mean and standard deviations of target expression (2^−ΔΔCt).

For the quantification of cytokines, the samples were weighed, and 50 mg of colon tissue was homogenized in 1 ml of PBS solution containing Tween-20 (0.05%) (Sigma-Aldrich, St. Louis, MO, USA), Phenylmethylsulfonyl fluoride (PMSF) 0.1 mM (Sigma-Aldrich, St. Louis, MO, USA), 0.1 mM benzethonium chloride (Sigma-Aldrich, St. Louis, MO, USA), 10 mM EDTA (Synth, São Paulo, São Paulo, Brazil), and aprotinin A 20 KIU (Sigma-Aldrich, St. Louis, MO, USA). The homogenized samples were then centrifuged at 3,000 × g for 10 min at 4°C, and the supernatants were collected to perform the enzyme-linked immunosorbent assay (ELISA). Plates were coated with purified monoclonal antibodies reactive with cytokines IL-1β, IL-10, IL-12, p70, IL-17, TGFβ1, and TNF-α (B&D Systems, Inc., USA), overnight at 4°C. Then, plate wells were washed, supernatants were added, and the plates were again incubated overnight at 4°C. On the third day, biotinylated monoclonal antibodies against cytokines (R&D Systems, Inc., USA) were added to the plates and incubated for 2 h at room temperature. Color was developed at room temperature with 100 μl/well of orthophenylenediamine (1 mg/m) and 0.04% (v/v) H₂O₂ substrate in sodium citrate buffer. The reaction was stopped by the addition of 20 μl/well of 2N H₂SO₄. The absorbance was measured at 492 nm using a Microplate Reader Model 680 (BIO-RAD).

Data were analyzed using one-way ANOVA followed by Tukey’s post-test and performed in GraphPad Prism version 9.1 for Windows (GraphPad Software, San Diego, CA, USA). The experimental assays were performed in triplicate, and the results were expressed as mean ± standard deviation. Asterisks demonstrated in all figures represent the significant differences between the experimental groups and were indicated as follows: *p< 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Expression of the P. freudenreichii SlpB protein by L. lactis NCDO 2118 was first verified by Western blotting (Supplementary Figure S2). We then investigated the ability of such expression to enhance the probiotic properties of the L. lactis in the context of DSS-colitis. First, the liquid and food consumption, as well as the weight of the animals, were monitored during the seven experimental days. In the DSS group, significant -weight loss was observed (p < 0.001 and p < 0.0001, respectively) on the sixth and seventh days (Figures 1A, B). However, at the end of the 7 days, both treatments with L. lactis-SlpB and the control group L. lactis WT were able to limit such weight loss of the animals (p < 0.001 and p < 0.01, respectively), when compared with the DSS group. Concerning liquid and food intake, no significant change was observed between experimental groups (Supplementary Figure S3).

Regarding disease activity index (DAI) analysis (Figures 2A, B), as expected, the DSS significantly increased the score (5.85 ± 3.18) in the disease control group (DSS Control), when compared with the healthy group (Control) on the sixth and seventh days (p < 0.05 and p < 0.0001, respectively). At the end of 7 days, L. lactis-SlpB administration was shown to mitigate the signs of clinical colitis, based on DAI score (3.40 ± 1.67), when compared with the DSS control (p < 0.01). Treatment with L. lactis WT strain failed, by contrast, to reduce DAI score in this DSS mice model (6.20 ± 1.30). Additionally, L. lactis-SlpB administration significantly limited the colon length shortening (p < 0.05) caused by the DSS administration, when compared with the DSS group (Figure 2C). It was more effective than L. lactis WT, which failed to do so (Figure 2C).

Histological analysis revealed that consumption of L. lactis-SlpB was able to mitigate the colon damage caused by DSS administration (Figure 2D). Precisely, it preserved colon morphological structure, reduced inflammatory cell infiltration in the lamina propria, submucosa, and muscular layer. Furthermore, animals that received L. lactis WT showed a slight decrease in mucosal damage, when compared with DSS control, but some ulcerations and a large infiltration of inflammatory cells were still observed. These results become evident through the analysis of the histopathological score, shown in Figure 2E, where, among the groups that consumed the DSS solution, the group L. lactis-SlpB showed significant differences in the histopathological score (3.14 ± 0.37), when compared with the group treated with L. lactis WT (4.00 ± 0.5, p < 0.05) and the DSS control group (5.0 ± 0.63, p < 0.0001). As expected, DSS-colitis induction resulted in a substantial decrease in goblet cells number in the DSS control group (67.08 ± 17.85 goblet cell/hpf). A significant increase in the number of goblet cells (Figures 3A, B) was exclusively observed in the group treated with L. lactis-SlpB (99.90 ± 20.51 goblet cell/hpf), when compared with the DSS control group (p < 0, 05). It was, however, not enough to re-establish the levels of the control group (133.1 ± 14.62 goblet cell/hpf, p < 0.05). In addition, L. lactis WT strain was not able to significantly increase the number of goblet cells (compared with the DSS-control group), and there was no statistical difference between L. lactis-SlpB and L. lactis WT strains (90.20 ± 6.64 goblet cell/hpf). A decrease in crypt depth (Figure 3C) was observed in the DSS Control group (191.1 μm ± 45.4, p < 0.05), compared with the Control group (232.3 μm ± 41.73). However, there was no statistical difference in the depth of the Crypts of Lieberkühn, between the treated groups L. lactis WT (212.5 μm ± 28.91) and L. lactis-SlpB (209.7 μm ± 46.30) with the control group (232.3 μm ± 41.73) and DSS group (191.1 μm ± 45.4).

Consumption of L. lactis WT and/or L. lactis-SlpB significantly decreased the amount of colon enzyme activity of MPO (57.27 ± 48.43 and 17.50 ± 6.65, respectively) (Figure 4A), with statistically significant differences for both treatments, when compared with the DSS control group (232.7 ± 115.9, p < 0.0001). The same scenario was repeated when the EPO enzyme activity was quantified (Figure 4B), where both strains, L. lactis WT and L. lactis-SlpB, proved effective to reduce EPO levels (0.07 ± 0.05 and 0.04 ± 0.04, respectively), with statistically significant differences, compared with the DSS Control group (0.28 ± 0.10, p < 0.0001). In addition, the results shown in Supplementary Figure S4 demonstrate high levels of secretory IgA (sIgA) in the inflamed control group DSS (91.30 μg/ml ± 40.93). However, no statistical differences between the groups L. lactis WT (63.54 μg/ml ± 17.17) and L. lactis-slpB (59.44 μg/ml ± 26.85 and control group (62.81 μg/ml ± 28.17) was observed.

In the context of DSS-induced colitis, consumption of L. lactis-slpB increased significantly (p < 0.001) the colonic mRNA expression levels (Figure 5) of muc-2 gene and epithelial barrier genes zo-1, cln-1, and cln-5 (1.75 ± 0.87; 2.65 ± 0.72; 1.46 ± 0.54; 1.56 ± 0.35, respectively), when compared with the DSS Control group (0.60 ± 0.21; 0.97 ± 0.49; 0.41 ± 0.25; 0.90 ± 0.31, respectively). Interestingly, no difference in the expression levels of the zo-2 gene was found between the experimental groups.

The increase in the inos gene expression levels triggered by DSS administration in the inflammatory control group (8.31 ± 4.66) were controlled by the administration of L. lactis-SlpB strain (1.39 ± 1.08, p < 0.01) (Figure 6A). On the other hand, mRNA levels of pparγ were decreased in the DSS Control group (0.59 ± 0.20) and in the L. lactis WT group (0.58 ± 0.30), but the L. lactis-SlpB administration restored significant levels of pparγ colonic expression (1.07 ± 0.54, p < 0.01) (Figure 6B). Regarding the expression of genes encoding pro and anti-inflammatory cytokines (Figures 6C, D), L. lactis-SlpB group showed significantly reduced levels of il-17 gene expression (0.39 ± 0.27, p < 0.01), compared with the DSS control group (1.83 ± 1.03). Finally, the gene expression of anti-inflammatory cytokine il-10 was reduced in the DSS Control group (0.31 ± 0.11), but was restored in the group that received the L. lactis-SlpB strain (1.25 ± 0.55, p < 0.001). However, there were no statistical differences in the colonic expression levels of il-17 and il-10 cytokines, between the groups receiving L. lactis WT or L. lactis-SlpB.

Levels of colonic pro-inflammatory cytokines IL-17, IL-12, and TNF-α (13.66 ± 2.285, 20.17 ± 1.36; 17.41 ± 6.01, respectively) (Figures 7A–C) were increased in the DSS control group. In contrast, L. lactis-SlpB group showed significantly reduced levels of these cytokines, TNF-α, IL-17, and IL-12 (19.98 ± 7.04, p < 0.05; 10.51 ± 2.20, p < 0.05; 14.20 ± 1.20, p < 0.05, respectively), compared with the DSS control group. In addition, an increase in the colonic levels of IL-10 and TGF-β cytokines (Figures 7D, E) was observed in the group treated with L. lactis-SlpB (59.10 ± 23.14; 468.70 ± 261.60, p < 0.05, respectively) compared with the DSS Control group (31.96 ± 4.91; 183.80 ± 63.11, respectively). However, L. lactis-SlpB failed to decrease levels of Il-1β induced by DSS (Figure 7F).