Introduction

Tuberculosis (TB) remains a significant global health challenge, with approximately 10.6 million new cases reported in 2022, leading to 1.3 million deaths worldwide ¹. Mycobacterium tuberculosis (M. tb) variant Mycobacterium bovis (M. bovis), a zoonotic pathogen, accounts for approximately 10% of human TB cases in regions with high prevalence of zoonotic diseases. The immune evasion mechanisms employed by M. tb pose challenges in understanding the host immune response and the progression of TB.

Recent studies have highlighted the interplay between chronic infections, such as TB, and cancer development ^{2, 3}. Emerging evidence suggests that the inflammatory environment induced by persistent infections can contribute to cellular transformations associated with epithelial-mesenchymal transition (EMT), a process linked to cancer metastasis ⁴. However, the specific mechanisms underlying this relationship, particularly concerning host factors involved in TB infection, remain poorly understood.

In this context, we focus on RNA-binding motif protein X-linked 2 (RBMX2), a host factor that has emerged as a potential modulator of both TB infection and cancer progression. In previous studies, we also reported that RBMX2 can regulate the retention of APAF-1 introns through alternative splicing, mediating apoptosis following M. bovis infection ⁵. Moreover, previous research also indicates that RBMX displays significant and contrasting effects depending on the tumor type, functioning either as an oncogene or a tumor suppressor. For instance, its overexpression has been associated with hepatocellular carcinoma ⁶ and T-cell lymphomas⁷, while a reduction in RBMX levels has been noted in Pancreatic ductal adenocarcinoma⁸, suggesting a role in tumorigenesis and metastasis. However, RBMX2 involvement in TB infection and its regulatory effects on EMT have not been thoroughly investigated.

This study aims to elucidate the role of RBMX2 in M. bovis infection and its impact on EMT in epithelial cells. We hypothesize that RBMX2 promotes the adhesion and invasion of M. bovis in host cells while facilitating the EMT process through the activation of the p65/MMP-9 signaling pathway. By investigating these interactions, we seek to provide insights into the potential dual role of RBMX2 in TB and lung cancer, highlighting its significance as a therapeutic target.

Results

Elevated expression of RBMX2 in Mycobacterium bovis-infected cells and tuberculous Lesions

M. bovis was used to infect the EBL cell library that was constructed using CRISPR/Cas9 technology in our laboratory at a multiplicity of infection (MOI) of 100:1. We assessed cell survival and conducted high-throughput sequencing on the viable cells to identify key host factors with potential anti-M. bovis capabilities. Notably, the RBMX2 knockout cells exhibited significant resistance to the infection.

Further analysis using AlphaFold multimer revealed that amino acid positions 56–134 of RBMX2 contain an RNA recognition sequence (Supplementary Fig. 1A). Additionally, we identified 611 RBM family sequences at the bovine protein exon level and categorized them into subfamilies based on motif similarity. RBMX2 was primarily classified within subfamily VII, which includes proteins such as EIF3G, RBM14, RBM42, RBMX44, RBM17, PUF60, SART3, and RBM25 (Supplementary Fig. 1B). A comparison of amino acid sequences indicated a high degree of conservation of RBMX2 across different species (Supplementary Table 1).

To further confirm RBMX2 expression in EBL cells following infection with M. bovis and M. bovis Bacillus Calmette–Guerin (BCG), we collected RNA samples at 24, 48, and 72 hours post-infection (hpi). Real-time quantitative polymerase chain reaction (RT-qPCR) results demonstrated that RBMX2 expression was significantly upregulated after both M. bovis and M. bovis BCG infections (Figs. 1A and B). Moreover, elevated RBMX2 levels were observed in bovine lung macrophages (BoMac cells), bovine lung alveolar primary cells, and human lung epithelial cells (A549 cells) following M. bovis infection (Figs. 1C-E).

Figure 1

Additionally, we investigated RBMX2 expression in clinical tissue samples from cows affected by M. bovis. Livers and lung samples with visible tuberculous lesions were obtained from a slaughterhouse. RT-qPCR analysis revealed that RBMX2 expression was upregulated in both the lungs and livers infected with M. bovis (Fig. 1F).

RBMX2 did not affect cell proliferation but inhibited epithelial cell survival during M. bovis infection

In generating RBMX2 monoclonal knockout (KO) cells, monoclonal cells were selected through limited dilution in 96-well plates. Sanger sequencing was conducted to identify different monoclonal cells with distinct knockout sites (Supplementary Fig. 2A).

The impact of RBMX2 knockout on EBL cells indicated that RBMX2 did not alter cell morphology, as evidenced by cytoskeleton staining using phalloidine (Supplementary Fig. 2B). Additionally, cellular proliferation was assessed using the EDU assay (Figs. 1G and H) and CCK-8 assay (Fig. 1I), along with cell cycle progression evaluated by Flow cytometry (Supplementary Fig. 2C).

To investigate the potential role of RBMX2 in resistance to M. bovis infection, we assessed the survival rate of EBL cells and H1299 cells at different hours post-infection using CCK-8 assay. The findings revealed an enhanced survival rate in both RBMX2 knockout polyclonal and monoclonal EBL cells following M. bovis infection (Figs. 1J and K). Furthermore, RBMX2-silenced H1299 cells exhibited a higher survival rate compared to H1299 ShNc cells after M. bovis infection (Fig. 1L). Notably, RBMX2 knockout in EBL cells and silencing in H1299 cells demonstrated significant improvements in cells survival at 96 and 120 hpi compared to Wild Type (WT) EBL cells and H1299 ShNC cells, as evidenced by crystal violet staining (Supplementary Fig. 2D).

RBMX2-regulated genes associated with cell tight junction and EMT-related pathways

In investigate the role of RBMX2 knockout in the process of M. bovis infection, we selected RBMX2 knockout EBL cells and WT EBL cells at 0 (2 hpi, recorded as 0 hpi), 24, and 48 hpi for RNA-Seq analysis. A total of 16,079 genes were detected, with a significance level of p < 0.05 and log2 (fold change) > 2 compared to WT EBL cells. Notably, 42 genes are significantly regulated at all three-time points (0, 24, and 48 hpi) (Fig. 2A). Among these, 11 genes were significantly upregulated, while 31 genes were significantly downregulated. A heat map illustrates the expression level of each gene across various samples, with each row representing a specific gene and each column representing the expression levels in a given sample (Fig. 2A).

Figure 2

The functional significance of these differentially expressed genes (DEGs) in EBL cells infected with RBMX2 knockout was further explored using GO and KEGG analyses. The most significantly affected biological processes included epithelial cell differentiation, cell adhesion and biological adhesion. The cellular components most affected were the extracellular region, extracellular region part, extracellular space, and plasma membrane part. The molecular functions primarily associated with activin receptor activity, type I (Fig. 2B).

The KEGG pathway entries revealed that these genes were linked to a various pathways, including ECM–receptor interaction, cGMP-PKG signaling pathway, PI3K-Akt signaling pathway, cytokine-cytokine receptor interaction, and vascular smooth muscle contraction (Fig. 2C). Proteomics also validated these phenotypes related to cell tight junction and EMT using 48hpi protein samples, highlighting pathways such as extracellular space, integrin binding, basement membrane, and glucose homeostasis via GO analysis, along with cell adhesion molecules, ECM-receptor interaction, and TGF-beta signaling pathway through KEGG analysis (Figs. 2D and E).

We also examined the impact of varying infection durations on RBMX2 knockout EBL cellular lines via GO analysis. At 0 hpi, genes were primarily related to the pathways of cell junctions, extracellular regions, and cell junction organization (Supplementary Fig. 3A). At 24 hpi, genes were mainly associated with pathways of the basement membrane, cell adhesion, integrin binding and cell migration (Supplementary Fig. 3B). By 48 hpi, genes were annotated into epithelial cell differentiation and were negatively regulated during epithelial cell proliferation (Supplementary Fig. 3C). This indicated that RBMX2 can regulate cellular connectivity throughout the stages of M. bovis infection.

For KEGG analysis, genes linked to the MAPK signaling pathway, chemical carcinogen-DNA adducts, and chemical carcinogen–receptor activation were observed at 0 hpi (Supplementary Fig. 3D). At 24 hpi, significant enrichment was found in the ECM–receptor interaction, PI3K–Akt signaling pathway, and focal adhesion (Supplementary Fig. 3E). Upon enrichment analysis at 48 hpi, significant enrichment was noted in the TGF-beta signaling pathway, transcriptional misregulation in cancer, microRNAs in cancer, small cell lung cancer, and p53 signaling pathway (Supplementary Fig. 3F).

To validate the RNA-seq data, we confirmed a random selection of 10 genes using RT-qPCR (Fig. 2F). The mRNA levels of the selected genes were consistent with the RNA-seq outcomes, including FRMD4B, MCTP1, HSPB8, ST14, and OMD, which are linked to scaffold proteins, cell adhesion, and epithelial barriers. Additionally, MCTP1, HSPB8, IL-24, 1L-7, GPX2, and NOS3 were associated with apoptosis, inflammation, and oxidative stress.

RBMX2 interferes with the integrity of tight junctions in epithelial cells caused by M. bovis

Through transcriptome sequencing, we identified 41 genes that underwent transcript changes following M. bovis infection. Of these, 22 genes were upregulated, while 19 genes were downregulated (Supplementary Figs. 4A and B). The downregulated genes were primarily associated with tight junction and leukocyte transendothelial migration (Fig. 3A). In contrast, the upregulated genes were mainly linked to immunity-related pathways, including the defense response to viruses and regulation of viral processes (Fig. 3B).

Figure 3

Based on these finding,we hypothesized that RBMX2 could damage the integrity of epithelial cell tight junction. To validate this prediction, we constructed a model of M. bovis-destroyed tight junctions in EBL cells. EBL cells were infected with M. bovis and we subsequently performed RT-qPCR and Western blot (WB) analyses to assess the expression of cell tight junction-related mRNAs (TJP1, OCLN, CLDN-5, and CLDN-7) and proteins (ZO-1, OCLN, and CLDN-5). We found that the expression levels of all mRNAs and proteins was significantly reduced following M. bovis infection compared to the uninfected coutrol (Figs. 3C and D).

Tight junctions are critical intercellular adhesion structures that define the permeability of barrier-forming epithelial cells ⁹. The ability of epithelial cells to maintain functional intercellular adhesion is essential for forming a tight epithelial protective barrier ¹⁰. To further investigate the effect of M. bovis on epithelial cell adhesion, we conducted a cell adhesion assay, revealing a significant decrease in the intercellular adhesion ratio (Fig. 3E). These results validated that M. bovis can compromise the tight junction of EBL cells.

To explore the impact of RBMX2 on the epithelial cells barrier, we infected RBMX2 knockout and WT EBL cells with M. bovis and assessed the expression of epithelial barrier-related mRNAs (TJP1, OCLN, and CLDN-5) and proteins (ZO-1, OCLN, and CLDN-5) using RT-qPCR and WB. The RBMX2 knockout EBL cells exhibited a significant upregulation of all three proteins compared to WT EBL cells (Figs. 3F and G). The cell adhesion assay also demonstrated that RBMX2 knockout EBL cells displayed a heightened adhesion ratio relative to WT EBL cells (Fig. 3H).

Tight junctions are essential component of the epithelial barrier and can be compromised by bacteria infections, contributing to inflammation ^{11, 12}. To investigate whether RBMX2 mediates intracellular inflammation by regulating the tightness of the epithelial barrier, we measured the mRNA levels of inflammatory factors (IL-1β, TNF, and IL-6) in RBMX2 knockout EBL cells post-infection. The results showed a decrease in the expression of these inflammatory factors in RBMX2 knockout EBL cells compared to WT EBL cells after M. bovis infection (Fig. 3I).

In summary, our findings indicate that RBMX2 affects the integrity of epithelial cell tight junctions, and knocking out RBMX2 enhances tight junction formation while decreasing the inflammatory response induced by M. bovis.

RBMX2 disrupted the epithelial barrier via activating p65

The MAPK pathway and NF-kappaB pathway play crucial roles in regulating tight junctions between epithelial cells ^13-16. To further elucidate the regulatory mechanism of RBMX2 in the epithelial barrier’s tight junctions, we found that the MAPK pathway was enriched in our transcriptome sequencing data, and the potential regulatory transcription factor p65 was predicted through the JASPAR database (Supplementary Table 2).

We assessed the phosphorylation levels of MAPK pathway proteins and p65 protein in RBMX2 knockout EBL cells, revealing that the phosphorylation of p65 and MAPK/p38/JNK was significantly reduced in these cells after M. bovis infection compared to WT EBL cells (Figs. 4A and B).

Figure 4

To determine whether RBMX2 specifically regulates the MAPK/p38/JNK proteins and p65 to promote tight junction disruption and decrease intercellular adhesion, RBMX2 knockout EBL cells were treated with PMA (p65 activator), Anisomycin (JNK activator), and ML141 (p38 activator) for 12 h prior to M. bovis infection. The optimal concentrations for these activators were determined through WB and CCK-8 cell viability assays, yielding 100 nm for PMA, 10 μm for Anisomysin, and 10 μm for ML141, all without significant effects on cell viability (Supplementary Figs. 5A, B, and C). WB results indicated that the p65 and MAPK/p38 pathways could impede the expression of tight junction proteins (ZO-1, OCLN, and CLDN-5) following M. bovis infection (Figs. 4C and D).

To examine whether RBMX2 regulates cell adhesion ratio via p65 and MAPK/p38/JNK pathways, we investigated the impact of p65, p38 and JNK activators on cell adhesion in RBMX2 knockout EBL cells. Cell adhesion assays showed that treatment with p65 activator significantly reduced cell adhesion ratio compared to the p38 and JNK activators (Figs. 4E and F).

To further substantiating the association between p65 and tight junction regulation in WT EBL cells after M. bovis infection, we used siRNA to inhibit p65. WB analysis demonstrated that siRNA significantly reduced p65 and p-p65 expression (Figs. 4G and H). WB and Immunofluorescence (IF) analyses revealed that suppressing p65 facilitated the expression of ZO-1, OCLN, and CLDN-5 after M. bovis infection (Figs. 4G, H and I). Additionally, silencing p65 in WT EBL cells inhibited the invasion of M. bovis into epithelial cells, as shown by plate counting assay (Fig. 4J).

Finally, to investigate the relationship between RBMX2-mediated p65 and the findings above, EBL cells were infected with M. bovis and M. bovis BCG. WB and high-content live imaging system outcomes demonstrated diminished nuclear translocation of the p65 protein in RBMX2 knockout EBL cells compared with WT EBL cells (Figs. 4K and L).

In summary, our findings validate that RBMX2 can regulate the phosphorylation and nuclear translocation of p65 protein, leading to the degradation of tight junction proteins in EBL cells infected with M. bovis.

RBMX2 promotes adhesion, invasion, and intracellular survival of pathogens

Disrupting the intercellular tight junction barrier enhances bacterial adhesion and invasion ^17,18. Based on these findings, we hypothesize that RBMX2 may facilitates the adhesion and invasion of M. bovis by degrading the tight junction proteins in EBL cells. To test this hypothesis, EBL cells were infected with M. bovis at an MOI of 20:1, plate count assays indicated a significant reduction in M. bovis adhesion in RBMX2-KO#17 and KO#23 EBL cells compared with WT EBL cells (Fig. 4M) at 15 min, 30 min, 1 h, and 2 h post-infection (Fig. 4N). Furthermore, the invasion of M. bovis in RBMX2-KO#17 and KO#23 EBL cells was decreased at 2, 4, and 6 h post-infection compared with WT EBL cells. Additionally, the intracellular survival of M. bovis in RBMX2-KO#17 and KO#23 EBL cells was lower at 24, 48, 72 and 96 hpi compared with WT EBL cells (Fig. 4O). Silencing RBMX2 in human lung epithelial cells (H1299) also led to a significant reduction in the adhesion, invasion, and intracellular survival of M. bovis (Figs. 4P-R). Thus, the host factor RBMX2 is critical for promoting the adhesion, invasion, and intracellular survival of M. bovis.

In addtion, pathway activators were employed to investigate the relationship between pathway activation and the regulation of M. bovis adhesion and invasion by RBMX2. The results showed that applying a p65 activator to RBMX2 knockout EBL cells significantly impaired the tight junction function of the epithelial barrier, thereby enhancing M. bovis adhesion and invasion in EBL cells by regulating the phosphorylation and nucleation of p65 (Supplementary Figs. 6A and B). Hence, RBMX2 promotes M. bovis adhesion and invasion in EBL cells by regulating the phosphorylation and nucleation of p65

To further explore the influence of RBMX2 on adhesion and invasion in the context of different virulence Mycobacterial species, we utilized attenuated virulent M. bovis BCG and Mycobacterium smegmatis (M. smegmatis) to infect EBL cells. The plate count assays revealed a significant reduction in the adhesion of both M. bovis BCG and M. smegmatis in RBMX2 knockout EBL cells at various time points post-infection compared to WT EBL cells (Supplementary Figs. 6C and D). Additionally, the invasion of M. bovis BCG and M. smegmatis was markedly diminished at different hpi in RBMX2 knockout EBL cells (Supplementary Figs. 6E and F). Consequently, the virulence of Mycobacterium does not appear to affect the capacity of RBMX2 to promote the adhesion and invasion of M. bovis BCG and M. smegmatis in EBL cells.

Furthermore, we examined the role of RBMX2 in regulating other bacteria associated with bovine pneumonia by infecting EBL cells with Salmonella and Escherichia coli (E. coli). Our findings indicated that RBMX2 knockout EBL cells exhibited a degree of inhibition in bacterial adhesion and invasion across all tested bacterial species compared to WT EBL cells (Supplementary Figs. 6G-J).

In summary, our experiments provide compelling evidence that RBMX2 broadly promotes the adhesion and invasion of pathogens associated with bovine pneumonia, highlighting its potential as a target for developing disease-resistant cattle.

RBMX2 is highly expressed in lung cancer and regulates cancer-related metabolites

Previous studies have demonstrated a link between bacterial infection and the initiation of EMT ^19-21. Prolonged infection with M. tb or M. bovis induces oxidative stress, activates Toll-like receptors (TLR), and elicits the release of inflammatory cytokines ^22-24. Consequently, this phenomenon creates a favorable microenvironment for tumor development, progression, and dissemination ^{25, 26}. Epidemiological investigations have established a correlation between TB and the occurrence of lung cancer ^27-30. However, the precise cellular mechanism underlying this association remain unclear.

In our research, we conducted a comprehensive analysis of gene families, revealing a remarkable degree of RBMX2 conservation among bovine, monkey, and human sources (Fig. 5A). To further elucidate the involvement of RBMX2 in cancer, we utilized the TIMER2.0 database to assess its expression patterns across various cancer types within The Cancer Genome Atlas (TCGA). Our findings indicated a notable upregulation of RBMX2 in lung cancer, specifically in lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) (Fig. 5B).

Figure 5

To investigate the elevated expression of RBMX2 in lung cancer further, we measured its expression in both lung cancer epithelial cell lines and normal lung epithelial cell lines using RT-qPCR. The results demonstrated that RBMX2 expression in lung cancer epithelial cell lines (NCIH1299, NCIH460, and CALU1) was significantly higher than that in normal lung epithelial cells (BEAS-2B) (Fig. 5C). Additionally, immunofluorescence (IF) analysis revealed that RBMX2 expression levels were also higher in LUAD and LUSC tumor tissues compared to adjacent non-cancerous lung tissues (Figs. 5D-F).

Furthermore, we investigate whether RBMX2 regulates specific EMT-related metabolites to mediate the EMT process in EBL cells following M. bovis infection. Our metabolomic analysis of EBL cell samples at 48 hpi revealed that RBMX2 knockout primarily enriched pathways related to nucleotide metabolism, biosynthesis of cofactors, biosynthesis of nucleotide sugars, pentose and glucuronate interconversions, vascular smooth muscle contraction, amino sugar, and nucleotide sugar metabolism, chemokine signaling pathway, aldosterone synthesis and secretion, and cGMP-PKG signaling pathway. Thes pathways are associated with tumor cells proliferation, migration, and invasion (Fig. 5G). Differential metabolites are related to tumor metabolism, mainly including 6-glucuronic acid estriol, adenosine, uridine 5 ‘-triphosphate, and 5’ - deoxy-5 ’- (methylthio) (Figure 5H).

RBMX2 promotes the transformation of EBL cells from epithelial cells to mesenchymal cells

To investigate the association between RBMX2 and EMT induced by M. bovis, we intially infected EBL cells directly with M. bovis. This infection did not effectively induce the expression of mesenchymal cell marker proteins (Supplementary Fig. 7A). Previous literature indicates that M. bovis infection in macrophages, along with its secreted proteins, can stimulate the production of cytokines such as TNF, IL-6, and TGF-β, all of which promote EMT process ^31-34.

To further explore the impact of M. bovis infection on EMT progression in EBL cells, we constructed a coculture model using M. bovis-infected BoMac cells (Fig. 6A and Supplementary Fig. 7B). In this model, we identified EMT-related cytokines, including IL-6 and TNF, in M. bovis-infected BoMac cells, revealing a significant increase compared to uninfected BoMac and EBL cells (Fig. 6B). Notably, RBMX2 expression was upregulated in EBL cells within this coculture model following M. bovis infection (Fig. 6C).

Figure 6

We then examined the morphology changes in epithelial cells within the coculture model using electron microscopy. The results demonstrated that most cells transitioned from circular to an elongated morphology (Fig. 6D). Staining of the EBL cells cytoskeleton with ghost pen cyclic peptide highlighted significant morphological alterations, with EBL cells transforming from spherical to spindle-shaped (Supplementary Fig. 7C).

Additionally, we assessed the expression of EMT-related mRNAs and proteins in EBL cells after coculture with M. bovis-infected BoMac cells. This analysis reveald a significant downregulation of the epithelial cell marker protein E-cadherin and a notable upregulation of mesenchymal markers N-cadherin and MMP-9, as determined by RT-qPCR and WB (Fig. 6E, Supplementary Fig. 7D). The coculture with, M. bovis-infected BoMac cells enhanced the migratory and invasive properties of EBL cells, as demonstrated by Transwell assay (Figs. 6F and G).

To confirm the impact of RBMX2 knockout on the EMT process in the coculture model, we observed that RBMX2 knockout cells exhibited significant upregulation of the epithelial cell marker protein E-cadherin and downregulation of the mesenchymal cell marker N-cadherin and MMP-9 compared to WT EBL cells, as shown by RT-qPCR and WB (Fig. 6H and Supplementary Fig. 7E). Wound healing and Transwell assays demonstrated that the migration and invasion rates of RBMX2 knockout cells were markedly lower than those of WT EBL cells (Figs. 6I, J, and K).

Finally, we silenced RBMX2 in the human lung cancer epithelial cell line H1299, which expresses high levels of RBMX2, to assess the effect on EMT-related proteins, as well as invasion and migration ability following M. bovis infection. The results indicated that RBMX2 significantly inhibited the EMT process in H1299 cells post-infection (Supplementary Figs. 8A and B).

In conclusion, we successfully established a coculture model involving M. bovis-infected BoMac cells that induce EMT in EBL cells, thereby demonstrating that the host factor RBMX2 effectively promotes the EMT process in this context.

RBMX2 facilitates the EMT process via the p65/MMP-9 pathway

EMT enhances the migration and invasion capabilities of tumor cells ^35-37. Both p65 and MAPK/p38/JNK pathways have been shown to regulate the EMT process through various mechanisms. ^38-41. To elucidate the precise regulatory role of RBMX2 in the EMT process within the coculture model of EBL cells, we assessed the MAPK pathway and p65 protein levels via WB. Our findings revealed a significant reduction in the phosphorylation of p65 and MAPK/p38/JNK in RBMX2 knockout cells infected with M. bovis, compared to WT EBL cells (Supplementary Fig. 8C).

To determine the specific pathways regulating the EMT phenotype of p65 and MAPK/p38/JNK in the coculture model, RBMX2 knockout EBL cells were treated with PMA, Anisomysin, and ML141. We observed that p65 downregulates the expression level of E-cadherin while simultaneously upregulating MMP-9 expression (Fig. 7A). The MAPK/p38 pathway inhibits E-cadherin expression while promoting N-cadherin expression. Conversely, the MAPK/JNK pathway impedes N-cadherin expression while facilitate the expression of E-cadherin and MMP-9. Furthermore, we comprehensively investigated the correlation between pathway activation and the migration and invasion of phenotypes, demonstrating that the activation of the p65 pathway enhances the migratory and invasive capabilities of EBL cells, as evidenced by the Transwell assay (Figs. 7B and C) and wound-healing assay (Figs. 7D and E).

Figure 7

Invasion and migration are critical stages in tumor development, influenced by p65-dependent factors such as matrix metalloproteinases, urokinase fibrinogen activators, and interleukin-8 ^42-46. To further substantiate the correlation between the p65 protein and the regulation of the EMT process in EBL cells post-M. bovis infection, we employed siRNA to inhibit p65 expression in WT EBL cells. This suppression was found to inhibit MMP-9 expression, as demonstrated by WB (Figs. 7F and G).

To verify the regulatory mechanism between RBMX2 and p65, we first identified multiple amino acid residues in the RBMX2 promoter that form strong interactions through docking (Fig. 7H). Additionally, we queried the RBMX2 promoter region for binding sites of the transcription factor p65 using JASPAR data, subsequently validating this interaction using a dual luciferase reporter plasmid system (Fig. 7I) and Chip-PCR (Fig. 7J). Notably, RBMX2 and p65 exhibited multiple amino acid residues forming strong bonds in protein docking, yielding a docking score of 1978.643 (Fig. 7K).

Moreover, MMP-9 is known to induce EMT in epithelial cells, enhancing their invasiveness. The regulatory mechanisms involving p65 and MMP-9 have been documented in several studies ^{47, 48}. In our research, we identified multiple binding sites in the promoter sequences of p65 and MMP-9 through molecular docking (Fig. 7L) and confirmed the biding site between MMP-9 promoter and p65 protein via Chip-PCR (Fig. 7M). Using protein docking, we validated the relationship between bovine p65 protein and MMP-9, with a binding score of 1784.378 (Fig. 7N).

In summary, RBMX2 facilitates the EMT process of EBL cells following M. bovis infection by activating the p65/MMP-9 pathway.

Discussion

The resurgence of M. bovis-associated TB presents a substantial global challenge, affecting both livestock and human populations. Notably, M. bovis exhibits over 99% nucleotide similarity with M. tb ⁴⁹. The pathogenesis of TB is complex process influenced by a confluence of bacterial, host, and environmental factors. Both M. tb and M. bovis have evolved sophisticated strategies to evade host immune responses, enabling their long-term intracellular persistence. It is estimated that approximately one-third of the global population harbors latent TB infections ^{50, 51}.

The successful establishment of infection by mycobacteria is contingent upon their ability to circumvent early innate immune defenses, employing both transcriptional and post-transcriptional regulatory strategies within host macrophages ^52-56. Despite notable advancements in the field, unraveling the mechanisms of immune evasion and understanding the dynamics of latent infections caused by M. tb and M. bovis remains a substantial challenge. The cellular immune response to these pathogens is inherently complex, further compounded by the microdiversity of mycobacteria within individual hosts and the variability of immune responses among different individuals ^{54, 57}.

Lung epithelial cells serve as the primary physical barrier against infection and play a pivotal role in the innate immune response. These cells not only function as a frontline defense but also recruit and activate antigen-presenting cells, such as macrophages, to initiate adaptive immune responses against M. bovis infection ^{58, 59}. Given that alveolar epithelial cells can act as reservoirs for M. bovis, their role in the infection process is particularly significant. Additionally, activated macrophages and neutrophils enhance the bactericidal effects of alveolar epithelial cells, further contributing to the host’s defense mechanisms ^{60, 61}.

In our study, we utilized a CRISPR-Cas9 mutant library developed in our laboratory to investigate potential host factors influencing M. bovis infection. Notably, we identified RBMX2 as a protein with significant anti-M. bovis infection capabilities. The RBMX2 protein, characterized by an RNA recognition motif within its 56-134 amino acid residue region, is implicated in mRNA splicing through spliceosomes and is emerging as a potential molecular marker for sperm activity.⁶². The downregulation of RBMX2 in the X chromosome of lung telocytes suggests its involvement in cellular immunity ⁶³.

Further exploration of RBM genes in cattle revealed that EIF3G, RBM14, RBM42, RBMX44, RBM17, PUF60, SART3, and RBM25 belong to the same subfamily as RBMX2. Previous studies have demonstrated that genes within this subfamily can regulate the proliferation and lifecycle of cancer cells. For instance, EIF3G modulates the mTOR signaling pathway, inhibiting the proliferation and metastasis of bladder cancer cells ⁶⁴. Similarly, RBM14 has been linked to the reprogramming of glycolysis in lung cancer, acting as a novel epigenetically activated oncogene ⁶⁰. Despite these insights, the specific function of RBMX2 in the context of M. bovis infection and its potential role in cancer pathogenesis remain largely unexplored.

Our findings reveal that RBMX2 is upregulated in TB-infected cells and tissues, demonstrating the capacity to inhibit M. bovis invasion. Transcriptomic analyses indicate that RBMX2 may disrupt tight junctions within epithelial cells and promote epithelial-mesenchymal transition (EMT) following M. bovis infection.

Our findings reveal that RBMX2 is upregulated in TB-infected cells and tissues, demonstrating its capacity to inhibit M. bovis invasion. Transcriptomic analyses suggest that RBMX2 may disrupt tight junctions within epithelial cells and promote EMT following M. bovis infection.

Epithelial cells serve as more than passive barriers; they actively participate in innate immunity by regulating cytokine secretion and maintaining barrier integrity ⁶¹. Disruptions in epithelial tight junctions, often exacerbated by pro-inflammatory stimuli from pathogenic bacteria, can facilitate bacterial translocation and subsequent infection ^{61 65, 66}

In our experiments, we observed that the disruption of the epithelial barrier facilitated M. bovis adhesion and invasion. Conversely, the knockout of RBMX2 stabilized the epithelial barrier, attenuating M. bovis invasion and intracellular survival. This stabilization also mitigated downstream innate immune responses, reducing cellular inflammation, ROS production, and apoptosis.

The loss of tight junction integrity is a precursor to EMT, a process increasingly recognized for its role in cancer progression ^{20, 65-67}. Recent studies have established a link between bacterial infections and EMT induction, particularly in the context of gastric adenocarcinomas ^{68, 69}. Epidemiological evidence suggests that TB may serve as a risk factor for lung cancer ^{70, 71}; however, the underlying cellular mechanisms remain elusive. Chronic TB infection has been implicated in lung carcinogenesis, with reports indicating that BCG vaccination can enhance the survival of tumor cells under inflammatory conditions ^{72 73}.

Moreover, M. tb-infected THP-1 cells have been shown to induce EMT in lung adenocarcinoma epithelial cells ⁷⁴. Chronic infection with M. tb is associated with oxidative stress and inflammatory cytokine production, fostering an environment conducive to tumor progression ^{75, 76}. Our analysis of RBMX2 across various cancers revealed increased expression levels in lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC), suggesting a conserved role in tumor biology across species.

In light of these findings, we constructed a model of M. bovis infection in EBL cells to investigate EMT induction. Initial results indicated that M. bovis alone did not induce EMT; however, a co-culture model incorporating M. bovis-infected BoMac cells successfully induced EMT in EBL cells. Notably, the knockout of RBMX2 in this context inhibited EMT, suggesting that RBMX2 may elevate the risk of lung cancer through EMT induction following M. bovis infection.

In conclusion, our study identifies RBMX2 as a novel host factor influencing susceptibility to M. bovis infection, playing a crucial role in both TB pathogenesis and the EMT process associated with lung cancer. Targeting RBMX2 may present a promising avenue for the prevention and treatment of TB in both humans and animals. Additionally, the effective modulation of RBMX2 could potentially mitigate the incidence of TB-associated EMT and its implications for lung cancer development in the near future.

Conclusion

Our findings demonstrate that RBMX2 significantly facilitates the invasion of M. bovis by promoting the activation of the p65 protein. This activation undermines the integrity of the epithelial cell barrier and promotes the EMT of epithelial cells through the p65/MMP-9 signaling pathway. These insights elucidate the critical role of RBMX2 in M. bovis pathogenesis and highlight its potential as a therapeutic target for mitigating infection-related complications.

Methods

Patients and lung tissues

This study was approved by the Ethics Committee of Inner Mongolia Autonomous Region People’s Hospital, and written informed consent was obtained from all participating patients (Approval Number: No. 2020021). A total of 35 specimens of lung cancer tissue, along with adjacent normal lung tissue, were collected. We randomly selected the above specimens, including age and gender. All patients involved in this study had not received any medication, chemotherapy, or radiation therapy prior to surgical resection. The samples were subsequently prepared for immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) assays.

Bovine clinical tissues

This project has received approval from the Experimental Animal Ethics Center of Huazhong Agricultural University (Approval Number: HZAUCA-2024-0014). Following examination by researchers at the National Animal Tuberculosis Professional Laboratory (Wuhan), bovine tuberculosis tissue samples were collected from the lungs and livers of slaughtered cattle, with each specimen measuring at least 5 cm × 5 cm × 5 cm and specifically including tissues exhibiting tuberculous lesions.

Cell lines

EBL and 293T RRID:CVCL_0063 cells were generously provided by M. Heller from Friedrich-Loeffler-Institute. These cells were cultured in heat-inactivated 10% fetal bovine serum supplemented with Dulbecco’s modified Eagle medium (DMEM, Gibco, USA) at 37°C and 5% CO₂⁷⁷. BoMac cells were kindly provided by Judith R. Stabel from the Johne’s Disease Research Project at the United States Department of Agriculture in Ames, Iowa, and were maintained according to previously established protocols ⁷⁸. BoMac cells were grown in heat-inactivated 10% fetal bovine serum supplemented with Roswell Park Memorial Institute 1640 (RPMI 1640, Gibco, USA) at 37°C and 5% CO₂. Bovine lung alveolar primary cells were isolated in our laboratory and cultured in heat-inactivated 10% fetal bovine serum supplemented with DMEM (Gibco, USA) at 37°C and 5% CO₂. BEAS-2B RRID:CVCL_0168, NCIH1299, NCIH460 RRID:CVCL_0459, and CALU1 epithelial cells were all provided by Professor Shi Lei from Lanzhou University, and all culture procedures were conducted according to standardized protocols. We conduct regular testing on all cell lines to eliminate contamination from mycoplasma and black fungus.

Bacterium

M. bovis (ATCC:19210), originally isolated from a cow, is maintained in this laboratory. M. smegmatis mc²155 (NC_008596.1) and M. bovis BCG-Pasteur (ATCC:35734) were generously provided by Professor Luiz Bermudez from Oregon State University. All strains were cultured in a Middlebrook 7H9 broth (BD, MD, USA) supplemented with 0.5% glycerol (Sigma, MO, USA), 10% oleic acid–albumin–dextrose–catalase (OADC, BD, USA) and 0.05% Tween 80 (Sigma, MO, USA) or on Middlebrook 7H11 agar plates (BD, MD, USA) containing 0.5% glycerol (Sigma, MO, USA) and 10% OADC (BD).

Prior to infection, the optical densities of bacterial cultures at 600 nm (OD600) were adjusted to the required MOI using the standard turbidimetric card. The cultures were then centrifuged at 3,000 g for 10 min. The precipitated bacteria were resuspended in medium and dispersed using an insulin syringe. Subsequently, 50 μL of 10-fold serially diluted bacterial suspension was plated onto Middlebrook 7H11 agar to determine the number of viable bacteria (colony-forming units, CFUs). All experiments involving M. bovis were conducted in strict accordance with the biosafety protocols established for the Animal Biosafety Level 3 Laboratory of the National Key Laboratory of Agricultural Microbiology at Huazhong Agricultural University.

E. coli (ATCC:25922) was donated by Professors Zhou Rui and Wang Xiangru of Huazhong Agricultural University, while Salmonella (ATCC:14028) has been preserved and passed down through generations in our laboratory. These strains were resuscitated, and single colonies were purified, cultured, and grown in Luria–Bertani broth.

Generation of the RBMX2-KO EBL cells

The small guide RNA (sgRNA) sequence targeting the bovine RBMX2 gene (5’-GAATGAGCGTGAGGTCGAAC-3’) was cloned into the pKLV2-U6gRNA5 (BbsI)-PGKpuro2ABFP (#67991), kindly provided by Professor Zhao Shuhong of Huazhong Agricultural University, to construct the recombined lentivirus. EBL cells were then infected with either the RBMX2 pKLV2- U6gRNA5(BbsI)-PGKpuro2ABFP (#67991) lentivirus or the empty vector pKLV2-U6gRNA5(BbsI)- PGKpuro2ABFP (#67991) lentivirus as a negative control. After 48 to 60 hpi, puromycin (2.0 mg/mL) was added to select the positive clones. Finally, the monoclonal cells obtained through limiting dilution were expanded, and the knockout of RBMX2 was confirmed via PCR assay (Table S1).

Table S1

Extraction of total RNA and RT-qPCR

Cold phosphate-buffered saline (PBS, HyClone, China) was used to wash the cells three times, after which 1 mL of Trizol (Invitrogen, USA) was added per well to lyse the cells. The lysate was collected in EP tubes, and 200 μL of chloroform was subsequently added. The mixture was vortexed for 30 seconds and then centrifuged at 12,000 rpm for 10 minutes at 4°C. Following centrifugation, 500 μL of the supernatant was transferred to a new EP tube. To this supernatant, 500 μL of isopropanol was added and mixed gently by inversion. The mixture was allowed to stand for 10 minutes at 4°C before being centrifuged again at 12,000 rpm for 15 minutes at 4°C. The supernatant was discarded, revealing the RNA pellet. Next, the RNA pellet was washed with 1 mL of 75% ethanol and centrifuged at 7,500 rpm for 5 minutes at 4°C. The supernatant was removed, and the RNA pellet was air-dried for 15 minutes. Subsequently, 20 μL of DEPC-treated water was added, and the mixture was incubated at 58°C in a water bath for 10 minutes to dissolve the RNA. Purified RNA was then obtained.

To assess RNA purity, the OD260/OD280 ratio was measured using a NanoDrop ND-1000 instrument (Agilent Inc., USA), with values between 1.8 and 2.0 indicating acceptable purity. RNA integrity and potential contamination with genomic DNA were evaluated using denaturing agarose gel electrophoresis. The samples were stored at -80°C for later analysis.

Reverse transcription of the RNA samples was performed using HiScript III RT SuperMix for qPCR (+gDNA wiper, Vazyme, China). Four microliters of 4× gDNA wiper mix were added to 1 μg of RNA, followed by the addition of 16 μL of RNase-free ddH2O. The mixture was incubated at 42°C for 2 minutes to eliminate genomic DNA. Reverse transcription was subsequently carried out by adding 4 μL of 5× HiScript III qRT SuperMix, with the mixture incubated at 37°C for 15 minutes and then at 85°C for 5 seconds to synthesize cDNA.

cDNA expression across different sample groups was quantified using AceQ qPCR SYBR Green Master Mix (Vazyme, China) in a ViiA7 real-time PCR machine (Applied Biosystems Inc., USA). The final volume of each real-time PCR reaction was 20 μL, comprising 10 μL of 2× AceQ qPCR SYBR Green Master Mix, 0.4 μL of upstream primer (10 μM), 0.4 μL of downstream primer (10 μM), 0.4 μL of 50× ROX reference dye 2, 3 μL of cDNA template, and 5.8 μL of ddH₂O. The PCR conditions were as follows: 95°C for 5 minutes (1 cycle), 95°C for 10 seconds (40 cycles), and 60°C for 30 seconds (40 cycles). The primer sequences for RT-qPCR are provided in Table S2.

Table S2

Western blot

RIPA reagent (Sigma, USA), supplemented with protease inhibitors and phosphatase inhibitors (Roche, China), was added to the cell samples, and the cells were lysed on ice for 30 minutes. The supernatant was collected by centrifugation at 12,000 rpm and 4°C for 10 minutes. Protein concentrations of each sample were determined using a BCA kit (Beyotime, China), and the proteins were adjusted to equal concentrations.

The protein samples were mixed with 5× protein loading buffer and boiled for 10 minutes. Proteins were then separated using SDS-PAGE on 10% polyacrylamide gels at 100 V for 90 minutes, followed by transfer to polyvinylidene difluoride (PVDF) membranes (Millipore, Germany) at 100 V for 70 minutes. The PVDF membrane was incubated in 5% skimmed milk for 4 hours to block non-specific binding. The membrane was washed three times with Tris-buffered saline containing 0.15% Tween-20 (TBST) for 5 minutes each.

Subsequently, the membrane was incubated with a primary antibody at 4°C for 12 hours, followed by a secondary antibody at room temperature for 1 hour. The following primary antibodies were utilized: ZO-1 (Bioss, China), OCLN (Proteintech, China), CLADN-5 (Proteintech, China), MMP-9 (Proteintech, China), E-cadherin (Proteintech, China), N-cadherin (Bioss, China), LaminA/C (Proteintech, China), β-actin (Bioss, China), p65 (CST, USA), p-p65 (CST, USA), and a MAPK-related antibody (CST, USA). The following secondary antibodies were used: anti-mouse HRP secondary antibody and anti-rabbit HRP (Invitrogen, USA).

Phylogenetic tree construction and gene structure and function predictive analysis of the bovine RBM family

The bovine RBM protein sequences were converted into FASTA format files and imported into MEGA-X for comparative analysis. The neighbor-joining method was employed to construct the phylogenetic tree of the CsGRF protein. The iTOL website (https://itol.embl.de/) was utilized to enhance the visualization of the CsGRF phylogenetic tree. Additionally, the MEME Suite database (http://meme-suite.org/) was used for the online analysis of conserved motifs within the RBM gene family.

Adhesion assay

EBL cells were infected with different live bacteria at an MOI of 20 for durations of 15 minutes, 30 minutes, 1 hour, and 2 hours at 4°C. Following infection, the cells were washed three times with phosphate-buffered saline (PBS; Gibco, USA) to remove any extracellular bacteria. After thorough washing, cell lysis was achieved using 0.1% Triton X-100 (Beyotime, China). The resulting cell lysates were plated onto 7H11 or LB agar plates and incubated for several days to facilitate colony-forming unit (CFU) enumeration.

Invasion assay

EBL cells were infected with various live bacteria at an MOI of 20 for durations of 2, 4, and 6 hours at 37°C. Following infection, gentamycin (Procell, China) was added at a concentration of 100 μg/mL for 2 hours to eliminate any extracellular bacteria. The infected cells were then washed three times with phosphate-buffered saline (PBS) (Gibco, USA) to remove remaining extracellular bacteria. After sufficient washing, the cells were lysed using 0.1% Triton X-100 (Beyotime, China). The resulting cell lysates were subsequently plated onto 7H11 or LB agar plates and incubated for several days to allow for colony-forming unit (CFU) enumeration.

Intracellular survival assay

EBL cells were infected with various live bacteria at an MOI of 20 for 6 hours at 37°C. Following infection, gentamycin (Procell, China) was added at a concentration of 100 μg/mL to eliminate extracellular bacteria for 2 hours. The infected cells were then washed three times with phosphate-buffered saline (PBS) (Gibco, USA) to remove any remaining extracellular bacteria, which is referred to as the 0-hour time point.

Subsequently, the infected cells were cultured in complete medium supplemented with Triton X-100 (0.1%) (Beyotime, China) at various time points: 0, 24, 48, and 72 hours hpi. Cell lysates were plated onto 7H11 or LB agar plates and allowed to incubate for several days to facilitate colony-forming unit (CFU) counting.

RNA-Seq, proteomics and metabolomics

Total RNA was isolated using Trizol Reagent (Invitrogen Life Technologies), and the concentration, quality, and integrity were determined using a NanoDrop spectrophotometer (Thermo Scientific). Three micrograms of RNA were used as input material for RNA sample preparation. Sequencing libraries were generated according to the following steps: mRNA was purified from total RNA using poly-T oligo-attached magnetic beads. Fragmentation was performed using divalent cations under elevated temperature in an Illumina proprietary fragmentation buffer. First-strand cDNA was synthesized using random oligonucleotides and Super Script II. Second-strand cDNA synthesis was subsequently performed using DNA Polymerase I and RNase H. The remaining overhangs were converted into blunt ends via exonuclease/polymerase activities, and the enzymes were removed. After adenylation of the 3′ ends of the DNA fragments, Illumina PE adapter oligonucleotides were ligated for hybridization. To select cDNA fragments of the preferred length of 400–500 bp, the library fragments were purified using the AMPure XP system (Beckman Coulter, Beverly, CA, USA). DNA fragments with ligated adaptor molecules at both ends were selectively enriched using the Illumina PCR Primer Cocktail in a 15-cycle PCR reaction. The products were purified (AMPure XP system) and quantified using the Agilent high-sensitivity DNA assay on a Bioanalyzer 2100 system (Agilent). The sequencing library was then sequenced on the NovaSeq 6000 platform (Illumina) by Shanghai Personal Biotechnology Co., Ltd.

RNA sequencing services were provided by Personal Biotechnology Co., Ltd. Shanghai, China. The data were analyzed using the free online platform Personalbio GenesCloud (https://www.genescloud.cn). Proteomics and metabolomics were performed by METWARE.

Differential expression analysis of the transcriptome

HTSeq (0.9.1) RRID:SCR_005514 statistical methods were employed to compare the read count values for each gene, which served as the original expression measurements, while FPKM was utilized for normalization. Subsequently, differential gene expression was analyzed using DESeq (1.30.0) RRID:SCR_000154 under the following screening criteria: a fold change of |log2FoldChange| > 2 and a significant P-value < 0.05. We employed the R language Pheatmap (1.0.8) software package to conduct bidirectional clustering analysis of all differentially expressed genes across samples. A heatmap was generated based on the expression levels of the same gene in various samples and the expression patterns of different genes within the same sample, utilizing the Euclidean method for distance calculation and the complete linkage method for clustering. Data analysis was performed using the free online platform Personalbio GenesCloud (https://www.genescloud.cn).

GO and KEGG enrichment analysis

We mapped all genes to terms in the Gene Ontology (GO) database and calculated the number of differentially enriched genes for each term. GO enrichment analysis was conducted on the differentially expressed genes using the topGO package, where P-values were determined using the hypergeometric distribution method. A threshold of P-value < 0.05 was set to identify significant enrichment, allowing us to pinpoint GO terms associated with significantly enriched differential genes and to elucidate the primary biological functions they perform. Additionally, we employed ClusterProfiler (3.4.4) software to conduct KEGG pathway enrichment analysis for the differential genes, concentrating on pathways with a P-value < 0.05. The data analysis was carried out using the free online platform Personalbio GenesCloud (https://www.genescloud.cn).

Flow cytometry assay

Cells were washed with cold PBS and then centrifuged at 1,500 rpm for 5 minutes; the supernatant was discarded. The cells were fixed in 70% ethanol for 12 hours. Following fixation, the cells were centrifuged again at 1,500 rpm for 5 minutes, and the supernatant was removed. The cells were washed once with cold PBS.

For staining, a Cell Cycle and Apoptosis Analysis Kit (Beyotime, China) was utilized. The staining mixture was prepared with a total volume of 535 μL, comprising 500 μL of staining buffer, 25 μL of PI dye, and 10 μL of RNase A. The samples were incubated with the staining mixture at 37°C for 30 minutes prior to detection.

Cell viability assay

EBL cells were infected with M. bovis at an MOI of 100. Cell viability was assessed at 48, 72, 96, 120, and 144 hpi using the CCK-8 assay (Dojindo, Kumamoto, Japan).

At each time point, fresh medium containing 10% CCK-8 was added to a 96-well cell plate and incubated for 60 minutes at 37°C. The cell viability was then determined by measuring the absorbance at 450 nm. The percentage of cell viability was calculated using the following formula:

cell viability (%) = [A (infection group) − A (blank group)] / [A (infection group) A (blank group)] × 100

Wound-healing assay

Approximately 2×10⁵ treated cells were added to a 6-well plate after EBL cells were cocultured with BoMac cells for 48 hours following M. bovis infection and subsequently digested with 1× trypsin. Once the cells were evenly spread, a cross was drawn along a ruler using a 200 μL pipette tip to create a wound.

The cells were washed three times with PBS (Gibco, USA) to remove any detached cells, and serum-free medium was added to each well. Images were taken at 0, 12, and 24 hours of culture, and the healing rate was calculated using ImageJ software.

Cell adhesion assay

To assess the level of cell adhesion between RBMX2 knockout and wild-type (WT) EBL cells after M. bovis infection, a cell adhesion assay was performed using a cell adhesion detection kit (Bestbio, China) according to the manufacturer’s protocol. Matrigel (20 mg/L) was added to a 96-well plate at 100 μL per well. Subsequently, 3 × 10⁴ cells were added to the adhesion plate and incubated in a CO2 incubator for 6 hours. After incubation, the wells were washed twice with PBS, and the adhered cells were stained with 10 μL of cell stain solution A for 2 hours at 37°C. The level of cell adhesion was quantified by measuring the absorbance at 560 nm.

Transwell assay

5×10⁴ treated cells and 200 μL of serum-free medium were added to each upper chamber of a Transwell (Corning, USA), while 550 μL of complete medium was added to the lower chamber. The cells were incubated for 24 hours.

The following day, the upper chamber was removed, and 550 μL of 4% paraformaldehyde (Beyotime, China) was added for 30 minutes. After fixation, 550 μL of crystal violet dye (Beyotime, China) was added to each well for another 30 minutes. The upper chamber was then washed with PBS for 30 seconds and dried upside down with a cotton swab.

After drying, the membrane of the chamber was removed and placed on a glass slide, followed by the addition of a neutral gum seal. Three random fields of view were selected under the microscope for imaging, and the cell count was performed using ImageJ RRID:SCR_003070 software.

Crystal violet staining assay

EBL cells were infected with M. bovis at an MOI of 100. The surviving cells on six-well plates were then fixed with 4% paraformaldehyde (Beyotime, China). After fixation, the cells were stained with 0.1% crystal violet staining solution (Beyotime, China) and rinsed five times with PBS (Gibco, USA). The plates were placed in a 37°C oven for 6 hours before being photographed.

M. bovis infection of BoMac and coculture with EBL cells

BoMac cells were seeded into the upper chamber of a 0.4 μm pore size Transwell (Corning, USA) at a density of 3×10⁴ cells per insert in RPMI 1640 medium supplemented with 10% FBS. The following day, the cells were infected with M. bovis at an MOI of 5 for 4 hours. After infection, the cells were washed three times with 1× warm PBS and treated with Gentamycin (100 μg/mL) for 2 hours to eliminate any extracellular bacilli.

The infected and uninfected BoMac cells were then incubated for 24 hours to mitigate the effects of M. bovis infection. Afterward, the cells underwent an additional 24-hour incubation period in fresh RPMI 1640 medium before coculturing with EBL cells.

In parallel, EBL cells were seeded at a density of 2×10⁵ cells per well onto 12-well plates containing DMEM supplemented with 10% FBS. The following day, culture inserts containing either M. bovis-infected or uninfected BoMac cells were introduced into the wells of the 12-well plates containing EBL cells, and the coculture was incubated for up to 24 and 48 hours in serum-free DMEM.

Small interfering RNA (siRNA) transfection

EBL cells were transfected with p65 siRNA (Tsingke, China) using jetPRIME (Polyplus, France) following the manufacturer’s instructions. Scrambled siRNA (Tsingke, China) served as a negative control.

Cells were seeded in 12-well plates and cultured at 37°C in a 5% CO₂ atmosphere. To prepare the transfection mixture, 0.8 μg of jetPRIME was incubated with 1 μL of siRNA for 10 minutes. This mixture was then added to the cells and incubated for 12 hours.

Following the incubation, RNA extraction was performed, and the efficiency of transfection was calculated.

The siRNA sequences used were as follows:

Cattle p65-1 siRNA, sense 5′-GCAGUUUGAUACCGAUGAA (dT)(dT)-3′;

Cattle p65-1 siRNA, antisense 5′-UUCAUCGGUAUCAAACUGC (dT)(dT)-3′;

Cattle p65-2 siRNA, sense 5′-GGACGUACGAGACCUUCAA (dT)(dT)-3′;

Cattle p65-2 siRNA, antisense 5′-UUGAAGGUCUCGUACGUCC (dT)(dT)-3′.

P65 nuclear translocation assay

For the p65 nuclear translocation assay, RBMX2 knockout and wild-type cell lines were transfected with the pCMV-EGFP-p65 plasmid. Following transfection, bovine lung epithelial cells were infected with M. bovis.

After infection, the cell nuclei were stained with Hoechst dye. The entry of p65 into the nucleus was analyzed at different time points using a confocal high-resolution cell imaging analysis system from PerkinElmer Life and Analytical Sciences Ltd. (Britain). This imaging technique allowed for the assessment of p65 translocation into the nucleus in response to infection.

CHIP-PCR

Formaldehyde cross-linking and ultrasonic fragmentation of cells were performed. One plate of cells was removed, and the volume of the culture medium was measured. 37% formaldehyde was added to achieve a final concentration of 1%, and the plates were incubated at 37°C for 10 minutes. Glycine (2.5 M) was then added to the plates at a final concentration of 125 mM to terminate cross-linking. After mixing, the culture medium was left at room temperature for 5 minutes, and the cells were cleaned three times with cold PBS. The cells were scraped off with PBS, centrifuged at 2,000g for 5 minutes, and the supernatant was removed. IP lysis solution containing a protease inhibitor was added to lyse the cells (the amount of lysis solution depending on the amount of cell precipitation), and the mixture was fully lysed on ice for 30 minutes, with the cells being repeatedly blown with a gun or shaken on a vortex mixer to ensure complete lysis.

After sonication, the mixture was centrifuged at 12,000 rpm at 4°C for 10 minutes to remove insoluble substances and collect the supernatant. 90 μL of the input was retained, and the rest was stored at −80°C. To confirm the presence of the target protein in the sample, 40 μL of the ultrasonic crushing product was taken, mixed with 10 μL of L5 reduced protein loading buffer, and heated for denaturation before performing Western blot detection.

For the remaining 50 μL of the product, 5 μL of Protease K and 2 μL of 5M NaCl (final NaCl concentration of 0.2M) were added and incubated at 55°C overnight for cross-linking. After cross-linking, the nucleic acid concentration was measured, and a portion of the sample was taken for PCR amplification, followed by agarose gel electrophoresis to detect the effectiveness of ultrasonic fragmentation and confirm the presence of the target DNA. After verifying the Input result, 100 μL of the ultrasonic crushing products was frozen at −80°C.

Next, 900 μL of ChIP Diffusion Buffer containing 1 mM PMSF and 20 μL of 50% of 1× PIC was added, along with an additional 60 μL of Protein A+G Agarose/Salmon Sperm DNA. The mixture was stirred at 4°C for 1 hour, allowed to stand at 4°C for 10 minutes to precipitate, and then centrifuged at 4000 rpm for 5 minutes. The sample was divided into two 1.5 mL EP tubes; the target protein IP antibody was added to one tube, while IgG (1 μg of the corresponding species) was added to the other. The samples were shaken overnight at 4°C for precipitation and cleaning of immune complexes.

After the overnight incubation, 200 μL of Protein A+G Agarose/Salmon Sperm DNA was added to each tube, shaken at 4°C for 2 hours, allowed to stand at 4°C for 10 minutes, and then centrifuged at 4,000 rpm for 1 minute. The supernatant was removed, and 8 μL of 5M NaCl and 20 μL of Protein K were added for overnight cross-linking at 55°C. Using databases like JASPAR to predict transcription factor binding sites, primers were designed and synthesized based on these binding sites. RT-PCR was employed to verify the binding, and after amplification, the products were taken for gel electrophoresis to confirm the correct fragment size (Table S3).

Table S3

Dual-luciferase reporter assay

293T cells in the logarithmic growth phase were adjusted to a density of 5 × 10⁴ cells/mL and inoculated into 48-well cell culture plates, with 300 cells per well. Each concentration gradient was represented by three replicate wells. On the second day after plating, when the cells reached a density of approximately 60-70%, transfection was performed with the following experimental groups: pGL4.10 RRID:Addgene_72684 basic+pRL TK+Over NC, pGL4.10 basic+pRL TK+Over p65, pGL4.10-RBMX2-WT+pRL TK+Over NC, pGL4.10-RBMX2-WT+pRL TK+Over p65, pGL4.10-RBMX2-MUT+pRL TK+Over NC, pGL4.10-RBMX2-MUT+pRL TK+Over p65. Transfection complexes were prepared, and dual-luciferase detection was carried out 48 hours post-transfection.

For cell lysis, the cell culture medium was removed, and the cells were gently rinsed twice with PBS. Then, 50 μL of 1× PLB lysis solution was added to each well, and the plates were placed on a shaking table at room temperature for 15 minutes to ensure complete lysis.

From each sample, 20 μL was taken for measurement. To assess Firefly luciferase activity, 100 μL of Firefly luciferase detection reagent (LAR II reagent, Promega, USA) was added, mixed well, and the relative light unit (RLU1) was measured. After this, 100 μL of Sea Kidney luciferase detection reagent (1× Stop&Glo® Reagent, Promega, USA) was added, mixed thoroughly, and the relative light unit (RLU2) was measured.

The activation level of the target reporter genes was compared between different samples by calculating the ratio of RLU1 from the Firefly luciferase assay to RLU2 from the Sea Kidney luciferase assay.

Protein docking

Molecular docking simulations were conducted to predict the formation of stable complexes between proteins RBMX2 or MMP9 with p65. The aligned sequences of RBMX2, MMP9, and p65 proteins were sourced from the UniProt RRID:SCR_002380 database. Their three-dimensional structures were predicted using AlphaFold and refined by constructing the structures in Avogadro. The structures were then optimized using the MMFF94 force field and exported in PDB format for further optimization in Gauss09.

The docking of the proteins was performed using Hdock, where each protein was treated separately as the receptor and ligand to assess their interactions ⁷⁹. The resulting docking affinities were annotated to provide insight into the binding interactions. Finally, PyMOL RRID:SCR_000305 was utilized to visualize the binding interaction geometries, allowing for a detailed examination of the molecular interactions between the proteins.

Statistical analysis

Statistical analysis was performed on all assays conducted in triplicate, with data expressed as the mean ± standard error of the mean. Each experiment was independently repeated three times. GraphPad Prism 7.0 RRID:SCR_002798 (La Jolla) was utilized for statistical analysis, employing a two-tailed unpaired t-test with Welch’s correction for comparisons between two groups. For comparisons among multiple groups, one-way or two-way ANOVA was applied, followed by the LSD test. Statistical significance was indicated at four levels: not significant (ns presents p > 0.05), * presents p < 0.05, ** presents p < 0.01, and *** presents p < 0.001.

Supplementary figures and tables

Supplementary Fig 1

Supplementary Fig 2

Supplementary Fig 3

Supplementary Fig 4

Supplementary Fig 5

Supplementary Fig 6

Supplementary Fig 7

Supplementary Fig 8

Supplementary Table 1

Acknowledgements

We would like to thank the National Key Laboratory of Agricultural Microbiology Core Facility for assistance in high-throughput microscopy, and we are grateful to Zhe Hu for his support of the sample preparation. This work was supported by the Major projects of agricultural biological breeding in China (2023ZD0405003), the National Natural Science Foundation of China (32072942), China Agriculture Research System of MOF and MARA (CARS-37), National Natural Science Foundation of China (81960772) and Key Project of Inner Mongolia Medical University (YKD2022ZD016).

Additional information

Author Contributions

YC and AG contributed to the conception and design of the study. CW, YP and HY carried out the experiment and wrote sections of the manuscript and made equal contribution. YJ, AK, ZY, KZ, and SX were involved in bacterial and cell research studies. CW, PY, HY, SX, LB, LZ, and CH performed the statistical analysis. CW, AG, and YC wrote the manuscript with all authors providing feedback. All authors contributed to manuscript revision, proof-reading, and approval of the submitted version.

Abbreviations

• RBMX2: RNA-binding motif protein X-linked 2

• EBL: embryo bovine lung cells

• BoMac: bovine macrophage cells

• A549: human pulmonary alveolar epithelial cells

• TB: tuberculosis

• M. bovis: Mycobacterium bovis

• M. tb: Mycobacterium tuberculosis

• BCG: Bacillus Calmette–Guerin

• WOAH: World Organisation for Animal Health

• hpi: hours postinfection

• M. smegmatis: Mycobacterium smegmatis

• E. coli: Escherichia coli

• p65: RELA proto-oncogene, NF-kB subunit

• MAPK: mitogen-activated protein kinase

• EMT: epithelial–mesenchymal transition

• PARP: poly ADP-ribose polymerase

• ZO-1: tight junction protein 1

• CLDN-5: claudin 5

• OCLN: occludin

• MMP-9: matrix metallopeptidase 9

• SP-A: surfactant protein A

• GO: Gene Ontology

• KEGG: Kyoto Encyclopedia of Genes and Genomes

• WB: Western blot

• RT-qPCR: real-time quantitative polymerase chain reaction

• IF: immunofluorescence

• ECMy: extracellular matrix

• TCGA: the cancer genome atlas