Transcriptome responses of intestinal epithelial cells induced by membrane vesicles of Listeriamonocytogenes

Highlights

• •We report the transcriptome profiling of Caco-2 cells upon interaction with membrane vesicles of L. monocytogenes.
• •Exposure of MVs to intestinal epithelial cells extensively altered the host transcriptome.
• •Pathway analysis confirming that MVs appears as principally responsible for the induction of immune signaling pathways.
• •We identified several non-coding RNAs (ncRNAs), possibly involved in the regulation of early manipulation of the host gene expression, essential for the persistence of L. monocytogenes.
• •The findings have opened the way for more detailed studies on the roles of membrane vesicles in the host-pathogen interaction during L. monocytogenes infection.

Abstract

Membrane vesicles (MVs) serve as an essential virulence factor in several pathogenic bacteria. The release of MVs by Listeria monocytogenes is only recently recognized; still, the enigmatic role of MVs in pathogenesis is yet to be established. We report the transcriptome response of Caco-2 cells upon exposure to MVs and the L. monocytogenes that leads to observe the up-regulation of autophagy-related genes in the early phase of exposure to MVs. Transcription of inflammatory cytokines is to the peak at the fourth hour of exposure. An array of differentially expressed genes was associated with actin cytoskeleton rearrangement, autophagy, cell cycle arrest, and induction of oxidative stress. At a later time point, transcriptional programs are generated upon interaction with MVs to evade innate immune signals, by modulating the expression of anti-inflammatory genes. KEGG pathway analysis is palpably confirming that MVs appear principally responsible for the induction of immune signaling pathways. Besides, MVs induced the expression of cell cycle regulatory genes, likely responsible for the ability to prolong host cell survival, thus protecting the replicative niche for L. monocytogenes. Notably, we identified several non-coding RNAs (ncRNAs), possibly involved in the regulation of early manipulation of the host gene expression, essential for the persistence of L. monocytogenes.

Graphical abstract

Introduction

Membrane vesicles (MVs) are the bi-layered structure of 20 to 500 nm in size and consist of macromolecules, such as phospholipids, proteins, lipopolysaccharide (LPS), and nucleic acid (Pathirana et al., 2016). MVs are proteoliposomes shed by the outward bulges producing spherical buds involving periplasmic membrane and accumulating peptidoglycan fragments (Reviewed in Toyofuku et al., 2019). MVs are composed of a single membrane derived from the outer membrane (OM) and contain OM proteins, lipopolysaccharide (LPS), and other lipids, while the vesicle lumen mainly contains periplasmic proteins (Ellis et al., TN 2010). The release of MVs benefits the microbe, mediating microbial interactions with the human host and within the bacterial community (Bomberger et al., 2009; A Kulp and Kuehn, 2010). MV synthesis and release by bacteria is a universal and widespread phenomenon (Toyofuku et al., 2019).

Generally, bacteria release MVs by blebbing their membrane. MVs produced by Gram-negative bacteria have been extensively studied. However, MVs from Gram-positive bacteria were overlooked for decades under the impression that the Gram-positive bacteria lack an outer membrane and the peptidoglycan-rich cell wall rigidity would not allow vesicle blebbing (Lee et al., 2009). The MVs from Gram-negative bacteria are known as outer membrane vesicles (OMVs), whereas MVs from Gram-positive bacteria are called membrane vesicles (MVs) (Avila-Calderón et al., 2015). A few reports demonstrated the biogenesis of MVs in Gram-positive bacteria (Brown et al., 2015; Toyofuku et al., 2019). The secretion and release of MVs into the extracellular milieu appear similar in Gram-negative and Gram-positive bacteria (Deatheragea and Cooksona, 2012).

In recent years, several Gram-positive bacteria such as Staphylococcus aureus, Bacillus spp., Streptococcus spp., Clostridium perfringens, and Listeria monocytogenes have been found to produce MVs during in vitro culture and/or in vivo murine infection (Thay et al., 2013; Rivera et al., 2010; Resch et al., 2016; Jiang et al., 2014; Lee et al., 2013). In addition, MVs production also protects bacteria under stressful conditions by eliminating accumulated damaged DNAs and proteins (McBroom et al., AJ 2007).

Diverse functions ascribed to MVs include promoting virulence, biofilm formation, signal transduction, stress response, cytotoxicity, and host pathology (S Liao et al., 2014; Mondal et al., 2016; Karthikeyan et al., 2020). L. monocytogenes secretes MVs under physiological stress. The release of MVs could benefit the pathogen by enhancing the interactions with the host or within bacterial communities [Kulp A, Kuehn MJ A 2010; Bomberger et al., 2009]. Further, MVs contain various cargo molecules which can modulate host pathology (Jung et al.,2016; Bielaszewska et al., 2017; Finethy et al., 2017). Thus, MVs help pathogens to buffer the host immune response and to invade and establish the infection successfully [Jung et al., 2016; Bielaszewska et al., 2017; Finethy et al., 2017].

L. monocytogenes, the etiologic agent of listeriosis, remains a serious public health concern, mainly associated with the contaminated food coupled with a high mortality rate (Swaminathan et al., B 2007; Allerberger et al., 2010). The capability of L. monocytogenes to internalize the host by hijacking their immune-defense mechanism keeps this pathogen unique, possibly through the recruitment of MVs. L. monocytogenes use an array of virulence effectors that act in one or more steps during the cellular infection (Camejo et al., 2011).

The MVs from L. monocytogenes may serve as the cargo for delivering virulence factors to the host cells. Thus, MVs perform an essential function in bacterial pathogenesis without the direct interaction between pathogen and host cells (Bielaszewska et al., 2017; Elluri et al., 2014). We previously reported the proteome analysis of L. monocytogenes MTCC 1143 (Serotype 4b) (Karthikeyan et al., 2019). We previously demonstrated that MVs of L. monocytogenes serotype 4b could interact with Caco-2 cells and subsequently cause a cytopathogenic effect. In this study, we report the transcriptome responses of the Caco-2 cells upon exposure to MVs of L. monocytogenes serotype 1/2a and compared the responses to the Caco-2 cells infected with L. monocytogenes.

Studies explaining the host cell response against the infection of L. monocytogenes are widely available (Witte et al., 2012). To our knowledge, no published data is available on the immune pathophysiology corresponding to the transcriptome of the host response to the content of MVs from L. monocytogenes. Understanding the role of MVs from L. monocytogenes is indispensable to realizing the capacity of this pathogen. Therefore, we report the host cell (Caco-2 cells) responses against the MVs of L. monocytogenes strain 10403S using RNA-Seq technology, a robust approach for studying global gene response while infection (Nie et al., 2012; Srikumar et al., 2015). RNA-Seq allows the simultaneous sequencing of millions of RNA transcripts, thus enabling the identification of global gene expression under different conditions (Morozova et al., O 2008; Wang et al., 2009). Hence, we have used RNA-Seq to map the changes in host transcriptome during MVs interaction globally. Differential gene expression analysis revealed the modulation of gene expression over time, while pathway and gene ontology analyses provided novel insights into the processes during the interaction.

We focused on host cell responses during the early phase of interaction, as this period is essential for the establishment of infection. We were able to recognize MVs-induced signaling pathways, dysregulation of the microtubule, and perturbed cytoskeletal network in the host cells. A set of novel small nucleolar RNAs (snoRNAs), micro-RNAs (miRNAs), and small Cajal body-specific RNAs (scaRNAs) were identified that might be considered as a specific biomarker of L. monocytogenes-induced infections. Also, pathways down-regulated in Caco-2 cells during the MVs interactions were mapped, which include the cell adhesion, cytoskeletal regulators, apoptosis, and extracellular matrix synthesis. This is the first transcriptome analysis of the dynamic interaction of MVs from L. monocytogenes with intestinal epithelium and provides new insights into the MVs-dependent pathogenesis of L. monocytogenes.

Materials and methods

Bacterial strains and culture growth

L. monocytogenes 10403S strain (Gifted by Daniel Portney, University of California, Berkeley, USA) was regularly grown on brain heart infusion agar (BHI) (Himedia; Mumbai India) at 37°C.

Isolation and purification of MVs

L. monocytogenes was inoculated into fresh BHI broth and incubated overnight at 37 °C. A 10 ml of the overnight culture was transferred to 1 L of medium and was allowed to grow until the optical density at 600 nm (OD 600) reached 1.0 at 37 °C with shaking. Cells were pelleted using centrifugation at 6000 × g for 20 min at 4 °C twice. The supernatant was collected and filtered through a 0.2 μm membrane filter (Millipore, Billerica, USA). The filtrate obtained was subjected to ultracentrifugation at 1,42,032 x g, for 2 h at 4 °C in Type 45 Ti in Beckman Ultracentrifuge. The MVs settled as the pellet was then resuspended in 100 µl of PBS. The MVs were further purified by sucrose density gradient centrifugation, as described earlier (Kulkarni et al., 2014). Briefly, layers of equal volumes of 70%, 60%, and 20% sucrose were added in polyallomer tubes from bottom to top, and the MVs suspension was added on the top of the layers. The tubes were ultracentrifuged in Beckman SW 60 Ti at 1,64,609 x g at 4 °C for 6 h. Different fractions were collected and diluted in 50 mM phosphate buffer (pH 7.4), and the presence of MVs was detected by dynamic light scattering (Nanoparticle analyzer, Horiba scientific, Z-100 obtain from Japan). The fractions containing MVs were pooled together and purified by ultracentrifugation again in Beckman Type 60 Ti for 90 min at 250,000 x g at 4 °C. The pellet thus obtained was reconstituted in 50 mM phosphate buffer (pH 7.4) and stored at −80 °C until further use.

Particle size characterization or dynamic light scattering (DLS)

Purified MVs were diluted with PBS to a final protein concentration of 0.06 µg/ml. The size distribution analysis was performed and recorded at 90˚angle with a laser of wavelength 632 nm. The data was analyzed by Horiba software, and the average hydrodynamic radius was obtained. The measurements were conducted at 25°C with 40 to 50 runs for each sample, and the average intensity weighted diameter was calculated. The average diameter was obtained for MVs isolated from three independent batches.

Exposure of L. monocytogenes and MVs to Caco-2 cells and RNA extraction

Growth and maintenance of the human colon adenocarcinoma cell line, Caco-2, were performed as described earlier (Karthikeyan et al., 2019). For bacterial infection, Caco-2 cells in antibiotic-free cell culture media were seeded in a 6-well plate (Hi-media) at a concentration of 1 × 10⁵ cells. Cells were then infected with three technical replicates of L. monocytogenes 10403S at a multiplicity of infection (MOI) of 10 and incubated at 37°C, 5% CO_2. After a 1 h infection period, infected monolayers were washed once with 1 ml sterile PBS, and gentamicin was added (50 µg/ml) at that point. The cells were further incubated for different time points (4 h and 8 h) at 37 °C, 5% CO₂.

Similarly, for MVs exposure, Caco-2 cells were treated with three technical replicates of MVs (10 µg) and incubated for 4 h and 8 h at 37 °C, 5% CO₂. Untreated cells were used as control. At each time points, cells were washed and scraped off gently. Total RNA was then extracted from Caco-2 cells using the Qiagen RNeasy mini kit according to the manufacturer's instructions. The concentration and quality of extracted RNA were assessed using an Agilent 2100 bioanalyzer and stored at − 80 °C.

Library preparation, RNA sequencing, and data analysis

Illumina NextSeq RNA Sample Prep Kit (version 3) was used according to the manufacturer's instructions for RNA-Seq sample preparation. Poly (A)-enriched cDNA libraries were generated using the Illumina NextSeq sample preparation kit (San Diego, CA) and checked for quality and quantity using the Bioanalyzer. Paired-end reads were obtained using the Illumina NexSeq 500 platform. Trimmomatic was used to remove any remaining Illumina adapter sequences from reads and to trim bases off the start or the end of a read when the quality score fell below a threshold (Bolger et al., 2014). Sequence quality metrics were assessed using FastQC. Reads were aligned independently to the Homo sapiens genome (GRCh38/hg38) obtained from the UCSC genome browser (http://genome.ucsc.edu) using HISAT (v 2.0.10) (Kim et al., 2015). After read mapping to hg38, SAMtools (http://www.htslib.org/) was used to filter bam files for uniquely mapped reads. The resulting BAM files were used for all further downstream analyses.

Reads aligned to annotated genes were quantified with the feature-count program Subread using gene annotations from the UCSC genome browser (Y Liao et al., 2014). Uniquely mapped reads were subjected to the DEseq2-Bioconductor R package for identification and quantification of genes that were significantly differentially expressed between the conditions, following standard normalization procedures (Love et al., 2014). The values of the FPKM (fragment per kilobase max. transcript length per million mapped reads) were computed for each library from the raw read counts. Normalized counts were transformed to rlog values to create a heat map. The list of DESeq2 determined differentially expressed genes (DEGs) were filtered with a conservative absolute log2 fold change cut-off and corrected p-value (p<0.05). Heatmap was generated using Heatmapper (Babicki et al., 2016). Boxplot was generated using BoxPlotR (Spitzer et al., 2014). Lists of differentially expressed genes were further annotated with ConsensusPathDB (Kamburov et al., 2013), BionetDB (Mudunuri et al., 2009), and pathway information using the Kyoto Encyclopaedia of Genes and Genomes (KEGG, http://www.genome.jp/kegg/).

Gene ontology (GO) enrichment analysis

GO categories enriched in the DEGs lists were identified using the ConsensusPathDB and DAVID (Huang et al., 2009). For each comparison, up-regulated and down-regulated gene sets were subjected to ConsensusPathDB and DAVID. A p-value cut-off of 0.01 was used.

KEGG pathways and network analyses

KEGG pathway analysis using ConsensusPathDB-Homo sapiens was performed to identify signaling and metabolic pathways over-represented in the differentially expressed gene lists. For each KEGG pathway, a p-value was calculated using a hypergeometric test, with a cut-off of 0.01 to identify enriched KEGG pathways. Genes that were more than 2-fold change in infected cells relative to uninfected controls at each time point were used as input with up- and down-regulated genes considered separately.

Non-coding RNA analysis

Non-coding RNAs and microRNAs were predicted from lncRNome (http://genome.igib.res.in /lncRNome/), LNCipedia (www.lncipedia.org), and miRBase (www. mirbase.org) databases.

Results

Purification of MVs from L. monocytogenes strain 10403S

L. monocytogenes strain 10403S was grown in BHI broth at 37 °C, and MVs were isolated from culture supernatant by ultracentrifugation. The electron micrographs revealed that the MVs of L. monocytogenes are spherical in size. The mean hydrodynamic radius of vesicles was 192.3 nm in size, as determined by DLS. All further experiments were performed using these purified MVs.

Exposure of MVs to intestinal epithelial cells extensively altered the host transcriptome

Caco-2 cells were treated with MVs and incubated for 4 h and 8 h post-challenge in three biologically independent experiments at the same time point. RNA integrity of all the experiments was examined, and the data has been shown in supplementary Fig S1. After 4 h of incubation, a total of 1189 transcripts were significantly altered upon exposure to MVs of L. monocytogenes (>1.5- fold regulation, p-value<0.05, p-adj<0.05); of these, 669 were up-regulated and 520 were down-regulated. At 8 h, 989 genes were significantly altered. Of these, 360 were up-regulated while 629 were down-regulated (>1.5- fold regulation, p-value<0.05, p-adj<0.05). MA plot was generated for all the data sets and shown in Fig. 1. DEGs in treated cells compared to the untreated cells were identified by calculating values of FPKM. The top 50 up-regulated and down-regulated genes are shown in Supplementary Table 1. Gene set enrichment analysis using ConsensusDB and DAVID showed significantly enriched functional groups that were altered by MVs exposure.

Fig. 1.

The host transcriptional responses to the infection by L. monocytogenes and its MVs. (A) Bland-Altman (MA) plots showing the differential expression of genes in Caco-2 cells after 4 h of exposure with MVs compared to the control. (B) at 8 h exposure with MVs (C) at 4 h infection with L. monocytogenes (D) at 8 h infection with L. monocytogenes. The plots were generated by the DESeq R package.

Epithelial response to interaction with MVs

Primary response genes, which include immediate-early and delayed genes, play pivotal roles in a wide range of biological processes, including differentiation, proliferation, survival, stress, innate and adaptive immune responses, and glucose metabolism. At 4 h, many of the up-regulated genes belong to endocytosis, autophagy, actin-cytoskeleton rearrangements, cell cycle, and pro-inflammatory cytokines. Genes were categorized according to biological function and canonical pathways. The top GO terms enriched in each category with the p-values are shown in Fig. 2.

Fig. 2.

Gene ontology enriched terms for genes differentially regulated in Caco-2 cells upon exposure with MVs. (A) Pathways significantly regulated in intestinal epithelial Caco-2 cells exposure with MVs 4 h (B) Pathways and process significantly regulated in intestinal epithelial Caco-2 cells exposed to MVs at 8 h time point. Terms were sorted according to P-value.

The GO enrichment analysis in biological processes revealed the unique features of MVs interaction with host. The most enriched early up-regulated clusters include the endocytosis, defense response, and pro-inflammatory response. In contrast, the most enriched early down-regulated group is involved in the regulation of metabolic pathways. Gene expression of cellular pathways is noticeably altered by host cells exposure to MVs. The majority of the affected pathways are vital for innate immune responses (P < 0.05, tested by the Wilcox test, followed by the Benjamini and Hochberg multiple test correction), according to the KEGG pathway annotation (Fig. 3). Invading microbial pathogens are sensed by the host PRR, such as Toll-like receptors (TLR), NOD-like receptors (NLR), and RIG-I-like receptors (RLR), through binding of PRRs to their ligands called pathogen-associated molecular patterns (PAMPs), which leads to the activation of host immune responses to microbial infections. DEGs in TLR and NLR signaling pathways, as well as cytokine-cytokine receptor interaction and chemokine signaling pathways, were significantly enriched (Table 1). Most of the pathways associated with innate immune responses were up-regulated, whereas those involved in the basal cellular metabolic pathways were generally down-regulated.

Fig. 3.

KEGG pathway enrichment analysis for genes differentially regulated in Caco-2 cells upon exposure with MVs. (A) Pathways significantly regulated in 4 h (B) Pathways significantly regulated during 8 h. Enrichment pathways were sorted based on q-value, and significance was assessed by the Wilcox test.

Table 1.

KEGG pathways enriched in host cells upon exposure with MVs.

KEGG term	Description	Genes	P-value
hsa04668	TNF signaling pathway	CXCL10, CFLAR, CXCL5, CXCL2, CXCL3, PIK3CA, CCL2, JUN, PTGS2, CREB3L1,MMP14, ICAM1, NFKBIA	0.00024414
hsa04621	NOD-like receptor signaling pathway	CXCL2,GBP1,GBP2,CCL2,JUN,BRCC3,NFKBIA,TICAM1,CXCL3,CXCL8,NFKBIB,GABARAPL1	0.000488
hsa04657	IL-17 signaling pathway	MMP13, CXCL10, CXCL5, CXCL2, CXCL3, CCL2. JUN, PTGS2, HSP90B1, MMP1, CXCL8, NFKBIA	0.000488
hsa04062	Chemokine signaling pathway	CXCL10, CXCL5, CXCL2, CXCL3, PIK3CA, CCL2, CXCL16, NFKB, IGNG5, CXCL8, NFKBIB, SOS1, SOS2	0.001709
hsa05164	Influenza A	TMPRSS13, DDX58, CXCL10, HSPA1A, TMPRSS4, PIK3CA, CCL2, JUN, NFKBIA, NXT1, TLR7, TICAM1, ICAM1, CXCL8, NFKBIB, HSPA2	0.002686
hsa04064	NF-κB signaling pathway	CSNK2A3, CFLAR, CXCL2, DDX58. PTGS2, NFKBIA, TICAM1, ICAM1, CXCL8	0.003906
hsa05160	Hepatitis C	DDX58, EGFR, PIK3CA, NFKBIA, CLDN, CLDN3, TICAM1, CXCL8, EIF3E, SOS1, SOS2	0.004883
hsa05200	Pathways in cancer	TXNRD1, MYC, TPR, MET, JUN, HSP90B1, IGF2, CALM1, EGFR, PIK3CA, RALBP1, NFKBIA, SHH, LAMC1, E2F1, LAMC2, MMP1, CCDC6, GNG5, IL2RG, PTGS2, PMAIP1, CXCL8, RBX1, CDK4, SOS1, SOS2	0.006998
hsa05134	Legionellosis	HSPA1A, CXCL2, CXCL3, HSPA2, C3, NFKBIA, CXCL8, EEF1A1, EEF1A2	0.007813
hsa04010	MAPK signaling pathway	MYC, EGFR, HSPA1A, MET, HSPA2, JUN, CACNG8, CACNG4, EREG, AREG, DUSP5, DUSP6, IGF2, NF1, SOS1, SOS2	0.009186
hsa04926	Relaxin signaling pathway	MMP13, GNG5,EGFR, PIK3CA, NFKBIA, JUN,CREB3L1, MMP1, SOS1, SOS2	0.009766
hsa00230	Purine metabolism	RRM1, APRT,POLR2J2, POLR2,GPOLE4, POLR3K, PRIM1, ATIC, POLR2C, POLR2D, GART, ENPP4, NTPCR, POLR1C, POLR2, IPOLR2J	0.005157
hsa04723	Retrograde endocannabinoid signaling	NDUFA9, NDUFB2, NDUFB7, NDUFS1,NDUFS3, PLCB4, NDUFA3, NDUFA4, MAPK1, NDUFA8	0.005859
hsa04060	Cytokine-cytokine receptor interaction	CCL2, EGFR, EDA, IL2RG, CXCL10, LIF TNFRSF11B, XCL2	0.007813
hsa03050	Proteasome	PSMB10, PSMC2, POMP, PSMA2, PSMA, PSMB1, PSMB4, PSMD14	0.007813
hsa00190	Oxidative phosphorylation	ATP5PB, ATP6V1A, COX6B2, NDUFA3, ATP6V0D1, ATP5PO, COX5B, COX6B1, COX7B, COX8A, UQCRQ, UQCRC1, NDUFA4, UQCRFS1, NDUFA8, UQCR11, NDUFA9, NDUFB2, NDUFB7, NDUFS1, NDUFS3, ATP5F1B	0.000256

Endocytosis and autophagy responses

The initial attachment and binding of MVs to epithelial surface triggered a dynamic host response as observed by GO enrichment of functional groups, including response to the endocytic pathway and autophagy. At 4 h time point, the CAV1 gene was highly up-regulated (55 fold), suggesting that MVs could have been internalized via caveolin-mediated endocytosis. The transcriptome signatures of host cells also revealed the modulations of cell-to-cell junction proteins, in particular, claudin 3 (CLDN3) and claudin 8 (CLDN8), which serves as the sealing component of tight junction forming a paracellular barrier, highlighting MVs invasion based on parasitosis. Transcripts involved in the autophagy process were also significantly up-regulated at 4 h time point. Importantly, GABARAPL1, FEZ1, MAPILC3B, RAB7A, ATP6V1G1 (GO:0,006,914) were significantly up-regulated. At 8 h time point, autophagy-related genes were not regulated considerably, suggesting that autophagy might be an early stage mechanism of host response during MVs interactions. It was observed that autophagy-related genes standstill, suggesting that autophagy might be a primary immune reaction of the host while MVs interaction.

Inflammatory pathways

Gene set enrichment analyses (GSEA) revealed the up-regulation of innate immune response genes, including inflammation, inflammasome signaling (encompassing Nod-like receptor and TLR7), and cytokine signaling (inflammation and cell death IL-17 signaling pathway (hsa04657)). The top up-regulated genes include C—C motif chemokine ligand 2 (CCL2) (FC-59.7), C-X-C motif chemokine ligand 5 (CXCL5), C-X-C motif chemokine ligand 2 (CXCL2), C-X-C motif chemokine ligand 3 (CXCL3), C-X-C motif chemokine ligand 8 (CXCL8), C-X-C motif chemokine ligand 16 (CXCL16), and C-X-C motif chemokine ligand 10 (CXCL10). Similarly, the transcript levels of genes encoding pro-inflammatory cytokines peaked at 4 h. The up-regulated genes enriched for the KEGG-pathways includes 'cytokine-cytokine interaction pathway' (hsa04060), NF-kappa B signaling pathway (hsa04064) 'TNF signaling pathway' (hsa04668) and IL-17 signaling pathways (Table 1). Expression of the pattern recognition receptors, TLRs, plays an essential role in the activation of the host immune responses. TLR7 gene was found to be up-regulated explicitly at 4 h time point. At 8 h, up-regulated genes associated with inhibition of complement system activation (CD55, complement decay-accelerating factor), EGFR, CCL2, and XCL2 (GO:0,006,954) were also observed. The tumor necrosis factor receptor type1 death domain (TRADD) was down-regulated, and several cytokine genes were not significantly regulated as compared to the early response.

Response to oxidative stress and metabolic process induced by MVs

The induction of reactive oxygen species (ROS) by host cells represents the first line of defense against intracellular pathogens. At 4 h, the transcriptome of Caco-2 cells showed the induction of thioredoxin reductase 1 (TXNRD1), EGFR, MMP14, CFLAR, NDUFA12, ATP2A2 (GO:0,006,979). On the other hand, genes encoded for oxidative phosphorylation were found to be down-regulated (GO:0,006,119) ATP5PO, NDUFB2, COX5B, NDUFS1, UQCRQ, NDUFS3, NDUFA4, NDUFA8.

Differential expression of genes involved in signaling pathways

Differential expression of several genes involved in signaling pathways (P < 0.05, Wilcox test), according to the KEGG annotation was observed. Notably, the TNF signaling pathway, NOD-like receptor signaling pathway, IL-17 signaling pathway, NF-κB signaling pathway, mitogen-activated protein kinase (MAPK) signaling pathway, relaxin signaling pathway, and PI3K-Akt signaling pathway were observed (Table 1).

Non-coding RNAs

RNA sequencing of Caco-2 cells exposure to MVs revealed several lincRNAs, microRNAs, snoRNAs, and snRNAs that were significantly regulated (Fig 4). We also observed 157 host snoRNAs that were differentially regulated (Supplementary Table S2). These snoRNAs were categorized into 53 H/ACA box type (SNORD) and 104 H/ACA box type (SNORA) groups. Both types of snoRNAs were significantly down-regulated upon interaction with MVs. In addition, long intergenic non-protein coding RNA (lincRNA) (n = 10), microRNA (miRNA) (n = 6), antisense RNA (n = 16), small nuclear RNA (snRNA) (n = 13), small Cajal body-specific RNA (SCARNA1) (n = 8) were significantly regulated during 4 h interaction. Similarly, lincRNAs (n = 4), microRNA (n = 1), snoRNA (n = 64), and snRNA (n = 8) were significantly regulated at 8 h.

Fig. 4.

Regulation of non-coding RNAs in Caco-2 cells exposure with MVs. (A) Types of differentially regulated non-coding RNAs upon infection with MVs 4 h (B) Differentially regulated non-coding RNAs in 8 h.

Similar host responses induced by L. monocytogenes and its MVs

When compared to the untreated cells, 2888 genes (1375 were up-regulated, and 1531 were down-regulated) were differentially expressed in Caco-2 cells infected with L. monocytogenes during 4 h post-time point, whereas 1189 (669 were up-regulated, and 520 were down-regulated) were differentially expressed in Caco-2 cells exposed with MVs (Fig. 5A). Similarly, during the 8 h time point revealed, a total of 2216 genes were significantly regulated. Of these, 727 were up-regulated, and 1489 were down-regulated (>1.5- fold regulation, p-value<0.05, p-adj<0.05), whereas 989 genes were significantly regulated in MVs-exposed cells. Of these, 360 were up-regulated while 629 were down-regulated (>1.5- fold regulation, P-value<0.05, P-adj<0.05) (Fig. 5B). Overall, 560 genes were commonly regulated in both sets during the 4 h time point, whereas 523 genes were commonly regulated in both sets during the 8 h time point (Fig. 5A-B). Furthermore, we performed pathway analysis with the distinct subset of genes modulated by both treatments. The results revealed that the genes regulated in Caco-2 cells by both interactions of MVs and L. monocytogenes showed enrichment of common GO terms. Analysis of the differentially expressed genes commonly altered by the exposure of L. monocytogenes and it's MVs (Supplementary Table S3) revealed the enrichment of genes involved in the inflammatory response, cellular response to tumor necrosis factor (TNF), immune response, apoptotic cell signaling, ER stress, and others. The common DEGs could be categorized into several canonical pathways, such as TNFsignaling, TLR signaling pathway, nuclear Factor-kB (NF-κB), chemokine signaling, and cytokine-cytokine receptor interaction (Fig. 6A). In conclusion, these results suggest that both L. monocytogenes and its MVs could modulate the expression of a similar set of genes and pathways in the host cell (Caco-2 cells).

Fig. 5.

Comparison of the gene expression pattern in Caco-2 cells infected with L. monocytogenes and its MVs. Venn diagram depicting the number and distribution of common differentially expressed genes in 4 h (A) and 8 h (B).

Fig. 6.

Heatmap showing differential gene expression in Caco-2 cells at 4 h of infection with L. monocytogenes and its MVs. Gene ontology (GO) group of the innate immune response (A), mitochondrial (B), apoptosis (C), and non-coding RNAs (D).

Dissimilar host responses to L. monocytogenes and its MVs

At 4 h time point, several genes associated with mitochondrial-mediated transport, apoptosis, immune response, and miRNAs were differentially regulated by L. monocytogenes as well as its MVs. For instance, mitochondrial-mediated transport genes such as MRPL22, MRPS6, MPZL2, NDUFS3, and COX5B were up-regulated upon exposure to MVs, whereas these genes were significantly down-regulated in response to L. monocytogenes (Fig. 6B). Similarly, innate immune response and apoptotic related genes such as CXCL2, CXCL8, ICAM1, MUC13, C13, NECTIN2, GATA6, ARF3, TIMP3, NDUF3, MSX2, NFKBIA, and RPS27A were differentially regulated (Fig. 6A & C). Also, several lincRNAs, microRNAs, snoRNAs, and snRNAs were significantly down-regulated in response to MVs as compared to L. monocytogenes (Fig. 6D).

At 8 h, innate immune response genes such as NFKBIA, HIST1H2BG, HIST1H2BE, ID2 were differentially regulated (Fig. 7A). Similarly, the cytoskeleton and microtubule-associated genes were differentially regulated (Fig. 7B). Few cell cycle-related genes such as MCM6, MYC, and CCNB1 were up-regulated in response to MVs exposure and down-regulated with L. monocytogenes (Fig. 7C). Likewise, mitochondrial-mediated transport genes and several lincRNAs, microRNAs, snoRNAs, and snRNAs, were differentially regulated (Fig. 7D-E).

Fig. 7.

Heatmap showing differential gene expression in Caco-2 cells at 8 h of infection with L. monocytogenes and its MVs. Gene ontology (GO) group of the innate immune response (A), actin-cytoskeleton (B), cell cycle (C), mitochondrial (D), and non-coding RNAs (E).

Overall, 597 genes were differentially expressed in Caco-2 cells in response to MVs exposure but not with the L. monocytogenes alone. These genes showed positive and negative regulation of inflammatory response, endocytosis, and positive regulation of apoptosis. MVs-regulated genes could be categorized into several pathways, such as PI3k-Akt signaling pathway, MAPK signaling pathway, NOD-like receptor signaling pathway, cAMP signaling pathway, TNF, and NF-κB signaling.

MVs modulated pro-survival pathways of the host cell (Table 2)

Table 2.

List of pro-survival pathway genes significantly expressed in host cells upon exposure with MVs.

Related GO term	Description	Gene name	Fold change	P-value
GO:0,032,956	Actin cytoskeleton rearrangement	ARF6 ARHGAP28 ARHGAP40 OPHN1 TMSB4X TRIOBP CAPZB TPM3 TACSTD2 MICAL3	2.3 11.4 17.8 41.7 0.47 17.4 12.8 2.3 1.7 23.6	0.002523 0.00474 0.000297 0.000301 0.003534 0.003334 0.045622 0.041 0.039126 0.005866
GO:0,006,914	Autophagy	GABARAPL1 MAP1LC3B RAB7A FEZ1 LRSAM1 TICAM1 TRIM5 PIK3CA	11.3474 2.1 9.1 29.4 10.092 3.54946 23.5817 2.34366	0.000154 0.004651 3.45E-08 8.21E-06 0.001255 0.00028 0.001425 0.005209
GO:0,045,087	The immune evasion signaling process	CCL2 CD55 RPS27A SAMHD1 XCL2 TRIM15 MSRB1 RBCK1	41.1639 20.2412 0.363462 0.430985 5.22208 40.5199 0.374091 45.03	0.001415 0.032958 0.028371 0.013921 0.020581 0.003208 0.004295 0.000856
GO:0,007,049 GO:0,000,077 GO:2,001,022	Regulation of cell cycle and DNA damage response	SYF2 TRIAP1 NABP1 NUDC NUP43 MAD2L1 MRE11 CDCA5 MNS1 MRNIP EGFR SMCHD1	14.1682 0.3893 0.257413 0.100511 8.52728 0.398981 0.285434 0.38159 10.2083 7.66825 7.72826 6.4	9.93E-09 0.000309 1.95E-05 0.005435 0.000821 0.000124 0.005237 0.00069 3.21E-07 0.001941 1.09E-11 4.25E-06

• (i)Actin-cytoskeleton rearrangements in host cells

Several genes associated with actin cytoskeletal rearrangements were up- and down-regulated. Enrichment analysis suggested that MVs induce the disruption and remodeling of the cytoskeleton. In particular, actin filament organization (GO:0,007,015), regulation of actin filament (GO:0,030,832), regulation of actin polymerization or depolymerization (GO:0,008,064) were within the enrichment category by GO analysis. Subsets of genes associated with actin cytoskeleton rearrangements were consistently up-regulated by MVs interaction (Fig. 8). Importantly, the process related to actin cytoskeleton rearrangement was the most enriched within the up-regulated genes. In particular, ARHGAP28, ARHGAP40, ARF, TRIOBP, and OPHN1 were up-regulated at 4 h, and CAPZB, TPM3, TRIOBP, TACSTD2, TMSB4X, and MICAL3 were up-regulated at 8 h time period respectively.

• (i)Autophagy and xenophagy response

Fig. 8.

Differentially expressed genes belong to actin-cytoskeleton rearrangement (GO:0,032,956) in response to MVs exposure. Left panel (A) ARF (B) ARHGAP28 (C) OPHN1(D) WASF2. Red notch indicates uninfected control and green indicate test. Right panel. Differentially expressed genes belong to autophagy (GO:0,006,914) in response to MVs exposure. (A) GABARAPL1 (B) FEZ1 (C) MAP1LC3B (D) RAB7A. The red notch box indicates uninfected control and green indicate test.

GO term analysis revealed that autophagy-related genes were significantly up-regulated during the early response of MVs and down-regulated by L. monocytogenes. Autophagy and xenophagy (macroautophagy) related genes, including GABARAPL1, FEZ1, MAP1LC3B, and RAB7, were significantly up-regulated by MVs in early response (Fig. 8). At 8 h, only the GABARAPL1 gene was found to be differentially expressed, whereas other autophagy genes were not significantly regulated, suggesting a dramatic remodeling of autophagosome formation following MVs interaction. Thus, MVs might be involved in inhibition or avoid autophagy of the host cells to favor the intracellular survival of L. monocytogenes.

• (i)Immune evasion signals

At 4 h, many cytokine-related genes were significantly up-regulated (Fig. 9 ). However, only a few cytokines CCL2, XCL2 and CXCL10 were up-regulated considerably at 8 h. Overexpression of TRIM15 and RBCK1 was observed at 8 h. The up-regulation of RBCK1 can negatively regulate TAB2/3 and TNF induced NF-κB activation. Also, overexpression of RBCK1 can inhibit the inflammatory signaling cascade.

• (i)Regulation of cell cycle and DNA damage response

Fig. 9.

Differentially expressed pro-inflammatory response genes in response to MVs exposure. The top up-regulated genes were those encoding the C—C motif chemokine ligand 2(CCL2), C-X-C motif chemokine ligand 5(CXCL5), C-X-C motif chemokine ligand 2(CXCL2), C-X-C motif chemokine ligand 3(CXCL3), C-X-C motif chemokine ligand 8(CXCL8), C-X-C motif chemokine ligand 16(CXCL16), C-X-C motif chemokine ligand 10(CXCL10). The red notch box indicates uninfected control and green indicate test.

MVs differentially regulate cell cycle-associated genes in host cells such as SYF2, TRIAP1, NABP1, NUDC, NUP43, and MAD2L1 (Fig. 10). The checkpoint gene MRE11 was significantly down-regulated (Fig. 10A) during the early phase of exposure, suggesting a positive regulation of the cell cycle. Also, the DNA damage response genes such as MRNIP, MNS1, EGFR, and SMCHD1 were significantly up-regulated during MVs exposure, suggesting a dramatic response to DNA damage.

Fig 10.

Differentially expressed genes belong to the regulation of cell cycle and DNA damage response (GO:0,007,049), (GO:0,000,077), GO:2,001,022: (A) MRE1, (B) SYSF2 (C) CDCA5 (D) TRIAP1 (E) NABP1 (F) MAD2L1 (G) MRNIP, (H) MNS. Red notch indicates uninfected control and green indicate test.

Discussions

The secretion of MVs is a fundamental process for both Gram-positive and Gram-negative bacteria. These MVs are primarily aiding for the transfer of DNA, proteins, cell-to-cell signaling, biofilm formation, stress response, delivery of toxin to the host cell and host-cell interactions (Bielaszewska et al., 2017). MVs of Gram-negative bacteria and their roles in virulence are well characterized (Mondal et al., 2016; Jung et al., 2016; Bielaszewska et al., 2017). Recent data suggest that vesicles of Gram-positive bacteria are shown to be associated with virulence. MVs derived from Gram-positive bacteria may play a pivotal role in bacterial pathogenesis, similar to MVs from Gram-negative bacteria (Mondal et al., 2016; Jung et al., 2016; Bielaszewska et al., 2017; Liu et al., 2018).

Although numerous studies provide evidence for the production of MVs by Gram-positive bacterium, the knowledge is still scarce on the functional characterization of MVs of L. monocytogenes (Jiang et al., 2014; Kamburov et al., 2013; Karthikeyan et al., 2020). In the previous study, we demonstrated that L. monocytogenes secrete biologically active MVs and could interact with Caco-2 cells. Also, we reported the proteome of L. monocytogenes serotype 4b strain. In this study, we report the purification MVs from a L. monocytogenes serotype 1/2a strain. In both strains, MVs were spherical in shape with sizes ranging from 200 nm to 300 nm in diameter. Major virulence factors such as LLO, inlB, PI-PLCA, autolysin, flagellin were observed MVs of both strains. We had previously shown that MVs could interact with Caco-2 cells and cause a cytopathogenic effect. Thus, we hypothesized that the transcriptome in the Caco-2 cells could be thoroughly influenced when exposed to the MVs of L. monocytogenes. As an extension, we studied the transcriptome responses of host cells (Caco-2 cells) upon exposure to the MVs of L. monocytogenes and compared the transcriptome responses of Caco-2 cells infected with L. monocytogenes cells. There is a vast knowledge gap on how MVs regulate the expression of gene transcription in the host during pathogenesis. To address this issue, we providently performed transcriptome profiling to establish evidence for understanding the transcriptional response of the host upon exposure to MVs and the whole bacterial cells. In this study, we emphasized the modulation of gene expression in Caco-2 cells upon exposure to the MVs of L. monocytogenes as the exact mechanisms underlying pathogenesis.

We used a human epithelial cell line Caco-2 cells since it has been successfully employed as an infection model to investigate the pathogenesis of L. monocytogenes and other bacteria. Recently, our group characterized the interaction of MVs with Caco-2 cells and demonstrated that MVs were internalized via actin-mediated endocytosis (Karthikeyan et al., 2019). The GO enrichment analysis in biological processes revealed the distinctive features of MVs interaction. The majority of enriched clusters comprise transcripts participating in endocytosis, defense response, and pro-inflammatory responses. These transcripts were significantly up-regulated during the early phase of exposure with MVs. In contrast, the most enriched early down-regulated groups were involved in the regulation of metabolic pathways.

The significant events occurring during the interplay between MVs and Caco-2 cells were identified in our transcriptome analysis. We found that MVs substantially regulates the expression of genes associated with the actin cytoskeleton network, as highlighted by the GO enrichment analysis. Loss of cytoskeletal integrity and increased epithelial permeability are usual hallmarks of inflammation. Cytoskeletal rearrangement or loss of integrity will lead to numerous beneficial events for pathogens to survive and disseminate into host cells (Colonne et al., 2016). Further, autophagy-related genes were significantly up-regulated in response to the early interaction with MVs, and not expressed during the late phase of exposure. Thus, molecular evidence is aligning with the previous report that MVs of L. monocytogenes could involve modulating autophagy during interaction (Vdovikova et al., 2017).

Recognition of infectious agents by the host cells results in the modulation of transcriptional programs to tackle the infection (Asrat et al., 2015). However, the activation of innate immunity by PRRs in response to infection with L. monocytogenes is still poorly understood. Also, the pattern recognition receptor that mediates the recognition of MVs is still unknown. In this study, we identified TLR7 implicated in pattern recognition was up-regulated upon interaction with MVs. Generally, the TLR family recognizes various categories of pathogen-associated molecules. Peptidoglycan from L. monocytogenes is recognized by TLR2 (Takeda et al., 2003), and major virulence factor listeriolysin O is a ligand for TLR4 (Jin et al., 2004). Recently, TLR10 was also regulated in macrophage as well as epithelial cells upon infection with L. monocytogenes (Regan et al., 2013). In this study, we observed that TLR7 was up-regulated in response to MVs of L. monocytogenes. However, possible ligand from the MVs is unknown. The murine TLR7 protein (human orthologue, TLR7/8) is shown to respond with viral single-stranded RNAs (ssRNAs), as well as to Streptococcal bacterial RNAs in dendritic cells (Eigenbrod et al., 2015; Diebold 2008). Possibly, RNAs packaged in the MVs could activate host signaling cascades via TLR7.

Overall, 597 genes were differentially expressed in Caco-2 cells in response to MVs exposure but not with the L. monocytogenes alone. These genes showed positive and negative regulation of inflammatory response, endocytosis, and positive regulation of apoptosis. These differences in host cell expression might be due to several reasons, and the exact mechanism remains elusive. MVs serve as cargo for delivering virulence factors to host cells to promote bacteria pathogenicity and induce an inflammatory response. In our previous study, the global proteomic analysis of L. monocytogenes and its MVs illustrated a variety of common proteins overlapped. Many virulence factors were enriched in MVs, including the toxin protein of L. monocytogenes listeriolysin O, inlB, PI-PLCA, which further emphasize differences in the expression of genes in the bacterial cell and MVs.

Concurrently, LLO is highly enriched in MVs, which is a primary determinant of L. monocytogenes pathogenesis. LLO induces apoptosis, and high concentration may cause rapid lysis of host cells (Carrero et al., J.A. 2008; Seveau, 2014). Stimulation of cells with LLO alone induces significant modifications in host cells (Hamon et al., 2012). Similarly, a previous report from Vdokova et al. (2017) demonstrated that the LLO of MVs derived from L. monocytogenes is essential for host autophagy and thus contributes to the survival of bacteria within the host cells. Similarly, virulence proteins enriched MVs could trigger host cell modification irrespective of expression from whole bacteria. However, further research is required to determine if MVs derived from L. monocytogenes also induce significant differences in various cell lines compared to the whole bacterium. The L. monocytogenes-derived MVs are equipped with the cargo required to interact with host cells like the pathogen”.

Numerous genes involved in the innate and adaptive immune responses were regulated in response to MVs interaction. The component of the innate immune response such as CCL2, CXCL8, CXCL5, CXCL16 was among the most up-regulated genes, highlighting the recruitment of T cells and macrophages, a phenomenon associated with inflammation (Sokol et al., 2015). However, at 8 h, these genes were not significantly regulated, possibly to avoid excessive inflammation. Modulation of immune signals in the host has been demonstrated as an essential strategy developed by successful pathogens to subvert host responses, switching the immune responses into a hypo-responsive state (Gogos et al., 2000). Regulation of gene expression associated with different signaling cascades such as TLR, TNFR, NF-κB, and Akt signaling pathway were also identified in Caco-2 cells upon interaction with MVs, indicating that the MVs may modulate innate immune system components at various levels to evade and subvert host responses. Previous studies have demonstrated that the innate immune regulatory genes were activated in response to L. monocytogenes infection in Caco-2 cells (Baldwin et al., 2003). A recent study reported that several cytokine and chemokine genes were extensively up-regulated in the intestine of gnotobiotic mice in response to WT-L. monocytogenes infection, (Lecuit et al., 2007). Our study strongly demonstrated that MVs from L. monocytogenes could trigger or activate immune signaling in host cells.

A strategy developed by the intracellular pathogens to establish a niche during infection is the ability to control host cell damage and cell cycle to their advantage. MVs promoted increased expression of cell cycle regulatory genes such as cell cyclins and CDK, which may enhance host cell survival of L. monocytogenes. Parallel to it, down-regulation of DNA damage response gene MRE11 was also observed. These results corroborate with previous reports demonstrating that listeriolysin O could degrade mre11 and promote bacterial replication (Samba-Louaka et al., 2014). Thus, the listeriolysin O enriched in MVs may interact with host cells to promote dysregulation of DNA damage response. Similarly, the up-regulation of cell cycle-related genes in response to MVs interaction could also be a strategy to promote host cell survival, in turn, to retain a replicative niche for L. monocytogenes.

Another important observation in this study is the regulation of lincRNAs, microRNAs, snoRNAs, and snRNAs upon interaction with MVs. The survival of intracellular pathogens in host cells depends on the modulation of several cellular functions by regulating the host gene expression (Mumtaz et al., 2017). Manipulation of the host transcriptome serves as a selective advantage for the intracellular life cycle of the pathogen. Our findings suggest that MVs not only appear to interfere with cellular functions but promoting changes in several transcriptional and posttranscriptional regulatory elements that may extensively impact host cell functions during the later stage of infection. One such example is the differential expression of several non-coding RNAs, which participate in the regulatory roles in physiological and pathological responses (Kaakoush et al., 2015). Down-regulation of several ncRNAs upon interaction with MVs also raises interesting questions on the role of this class of RNAs in MVs interactions. Several reports on ncRNAs suggested that the regulatory RNAs used by intracellular pathogens may aid in surviving and evade immune responses (Kaakoush et al., 2015; Bayer-Santos et al., 2017). Altogether, the peptides/proteins captured inside the MVs of L. monocytogenes are the prime responsible factors for the modulation of immune responses of the host.

In conclusion, our study, employing the transcriptome profiling of Caco-2 cells upon interaction with MVs of L. monocytogenes. The RNA-Seq data are confirming the dynamic changes in host immune cellular pathways during the interaction and providing cues for understanding the MVs-mediated host response. Further, MVs induced the expression of cell cycle regulatory genes, which may result in the ability to prolong the host cell survival, thus may enrich the replicative niche for L. monocytogenes. Remarkably, we identified several non-coding RNAs (ncRNAs) that are significantly down-regulated, suggesting that an early manipulation of the host gene expression induced by MVs. The findings have opened the way for more detailed studies to understand the exclusive roles of MVs in the host-pathogen interaction.

CRediT authorship contribution statement

Raman Karthikeyan: Conceptualization, Methodology, Formal analysis, Validation, Writing – original draft. Pratapa Gayathri: Methodology, Formal analysis. Subbiah Ramasamy: Methodology, Formal analysis. Vemparthan Suvekbala: Formal analysis, Writing – review & editing. Medicharla V. Jagannadham: Conceptualization, Writing – review & editing. Jeyaprakash Rajendhran: Conceptualization, Investigation, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

• Data will be made available on request.