Altered microRNA expression and pre mRNA splicing events reveal new mechanisms associated with early stage Mycobacterium avium subspecies paratuberculosis infection 1Scientific RepoRts | 6 24964 | DOI[.]
www.nature.com/scientificreports OPEN received: 04 September 2015 accepted: 08 April 2016 Published: 22 April 2016 Altered microRNA expression and pre-mRNA splicing events reveal new mechanisms associated with early stage Mycobacterium avium subspecies paratuberculosis infection Guanxiang Liang1, Nilusha Malmuthuge1, Yongjuan Guan2,3, Yuwei Ren1,4, Philip J. Griebel5,6 & Le Luo Guan1 The molecular regulatory mechanisms of host responses to Mycobacterium avium subsp paratuberculosis (MAP) infection during the early subclinical stage are still not clear In this study, surgically isolated ileal segments in newborn calves (n = 5) were used to establish in vivo MAP infection adjacent to an uninfected control intestinal compartment RNA-Seq was used to profile the whole transcriptome (mRNAs) and the microRNAome (miRNAs) of ileal tissues collected at one-month post-infection The most related function of the differentially expressed mRNAs between infected and uninfected tissues was “proliferation of endothelial cells”, indicating that MAP infection may lead to the over-proliferation of endothelial cells In addition, 46.2% of detected mRNAs displayed alternative splicing events The pre-mRNA of two genes related to macrophage maturation (monocyte to macrophage differentiation-associated) and lysosome function (adenosine deaminase) showed differential alternative splicing events, suggesting that specific changes in the pre-mRNA splicing sites may be a mechanism by which MAP escapes host immune responses Moreover, 9 miRNAs were differentially expressed after MAP infection The integrated analysis of microRNAome and transcriptome revealed that these miRNAs might regulate host responses to MAP infection, such as “proliferation of endothelial cells” (bta-miR-196 b), “bacteria recognition” (bta-miR-146 b), and “regulation of the inflammatory response” (bta-miR-146 b) Johne’ s disease (JD) is chronic granulomatous enteritis of ruminants caused by Mycobacterium avium subspecies paratuberculosis (MAP)1 Clinical symptoms of JD in cattle include persistent diarrhea, progressive weight loss, decreased production and death2 Although only 10–15% of MAP-infected cattle may develop clinical disease3,4, the carriers shed MAP into feces and milk5, which are the main sources of infection for other animals and possibly a zoonotic threat to humans6 To date, vaccines for JD are capable of controlling MAP shedding and clinical disease, but are not effective in preventing MAP infection7 In addition, animals infected by MAP usually undergo a long asymptomatic period and the diagnosis of MAP infection during the early subclinical stage remains challenging8,9 Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, AB, Canada 2UWA Institute of Agriculture and School of Animal Biology, University of Western Australia, Crawley, WA, Australia Department of Animal Biology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America 4Key Laboratory of Agricultural Animal Genetics, Breeding and Reproduction of Ministry of Education, Huazhong Agriculture University, Wuhan, Hubei, China 5Vaccine and Infectious Disease Organization, University of Saskatchewan, Saskatoon, SK, Canada 6School of Public Health, University of Saskatchewan, Saskatoon, SK, Canada Correspondence and requests for materials should be addressed to L.L.G (email: lguan@ualberta.ca) Scientific Reports | 6:24964 | DOI: 10.1038/srep24964 www.nature.com/scientificreports/ Figure 1. Quality control of RNA-Seq dataset (A) The distribution of genomic locations of RNA-Seq reads in control compartment (B) The distribution of genomic locations of RNA-Seq reads in infected compartment (C) The plot of RNA-Seq coverage of gene body (D) Saturation curve for gene number detection; X-axis number of the mapped reads; Y-axis - number of the expressed genes (FPM> 1) (E) Length distribution of small RNA reads (F) Cumulative frequency of detected miRNAs MAP infects the gastrointestinal tract primarily through the ileum or distal small intestine during the first few months after a calf is born10,11 Bacteria enter via M-cells overlying lymphoid follicles in the ileal Peyer’s patches (PPs), and establish a persistent infection in submucosal macrophages1 Many studies have focused on the mechanisms of MAP infection by characterizing host innate and adaptive immune responses in vitro (macrogphage cell line) and in vivo (ileal tissue) during the subclinical period, revealing a pronounced effect on immune cells, the systemic immune system, and the mucosal immune system12,13 Recently, gene expression changes in MAP-infected macrophages and whole blood of MAP-infected calves have been reported14,15 However, little is known about transcriptome alterations and the molecular mechanisms regulating the host response to MAP at the site of infection during the subclinical stage of disease A previous study reported the existence of MAP in ileal tissues and MAP-specific immune responses (such as interferon gamma responses) in calves one month after MAP infection, suggesting that a persistent infection was established within one month post-infection16 Moreover, the post-transcriptional regulation by microRNAs (miRNAs) and alternative splicing can play a role in host responses to pathogenic bacteria17,18 Thus, we hypothesized that the regulatory mechanisms of miRNAs and alternative splicing of pre-mRNAs may be associated with host responses during persistent MAP infection This study used an in vivo model to localize MAP infection to the terminal small intestine and studied gene expression and post-transcriptional regulation (miRNA expression and pre-mRNA splicing) at the site of infection one-month post-infection These transcriptional and post-transcriptional changes provide new insights into the mechanisms by which MAP effectively evades host immune responses and establishes a persistent infection Results RNA-Seq profiling of MAP-infected and control compartments of ileum tissues. The surgically isolated ileum in calves (10–14 days old) was subdivided into two compartments; MAP-infected (infected) and non-infected (control) Intestinal tissues from each ileal compartment, including the PPs, were collected from all animals (n = 5) at one month post-infection (age = 40–44 days) and prepared for transcriptome profiling using RNA-Seq A total of 300,676,396 paired-end sequence reads were obtained from the 10 samples On average, ~89% of these reads were mapped to the bovine reference genome (UMD3.1, Supplementary Table S1) The quality of RNA-Seq data was evaluated via the genomic regions of reads, the RNA-Seq 3′ /5′ bias and the sequencing depth Approximately 78% of the reads were derived from exonic and intronic regions, gene upstream and downstream regions, whereas 22% were derived from intergenic regions (Fig. 1A,B) In addition, the coverage of reads along each transcript revealed no noticeable 3′ /5′ bias, confirming the acceptable quality of sequencing data (Fig. 1C) The number of detected genes increased along with increasing sequencing reads and the gene number eventually plateaud, revealing that the most expressed genes were detected by RNA-Seq (Fig. 1D) Scientific Reports | 6:24964 | DOI: 10.1038/srep24964 www.nature.com/scientificreports/ Figure 2. Profiling of mRNAs expression, alternative splicing events and miRNAs expression (A) PCA plot of mRNAs expression (B) PCA plot of alternative splicing events (C) PCA plot of miRNAs expression (D) Categories of alternative splicing events (E) Number of different alternative splicing events (F) Hierarchical cluster of mRNAs expression (G) Hierarchical cluster of miRNAs expression The red and green frames indicate two clusters based on miRNAs expression A total of 17,789,335 small RNA reads were obtained from the same 10 samples (Supplementary Table S1) and 58.94% of the total reads were 21 nt and 22 nt in length (Fig. 1E) The reads (15,207,875) with 18–25 nt lengths were used for further miRNA expression analysis Among these reads, 14,062,510 reads were mapped to a known miRNA database (miRBase version 21), which resulted in 3,802 reads representing 55 putative novel miRNAs The remaining 1,137,761 reads that were not identified as miRNAs belonged to other small noncoding RNAs (tRNA, snoRNA, snRNA and others) Cumulative frequency analysis revealed that the 20 most highly expressed miRNAs accounted for approximately 90% of the sequenced reads, representing the majority of expressed miRNAs (Fig. 1F) Transcriptome, alternative splicing events, and miRNA expression changes after MAP infection. An average of 14,444 ± 187 (mean ± SD) and 14,430 ± 87 genes were identified (fragments per million mapped fragments (FPM) > 1) in control and infected ileal compartments, respectively In total, 13,046 genes were commonly expressed in all 10 samples Functions of the 3,000 most highly expressed genes were related to “metabolic process” and “protein synthesis” (Supplementary Table S2) Principal component analysis (PCA) and hierarchical cluster analysis showed no clear separation between infected and control samples (Fig. 2A,F) The transcriptome profile of animal #5 control compartment was an outlier from all control samples of other animals (Fig. 2A), thus this animal was subsequently removed from further analysis An average of 20,036 ± 3,163 alternative splicing events were detected in the control compartment tissues, whereas 21,005 ± 870 were detected in infected tissues There were no significant differences between the two groups in terms of the number of alternative splicing events However, the number of alternative splicing events differed greatly between control and infected compartments (14,598 vs 19,661) of animal #5, which was also an outlier (Fig. 2B) These differences led us to exclude animal #5 from further pre-mRNA splicing analyses The remaining four animals revealed 26,045 alternative splicing events (5,903 alternative acceptor, 3,535 alternative donor, 1,262 alternative first exon, 324 alternative last exon, 4,344 cassette, 705 coordinate cassette, 9,918 intron retention, and 54 mutually exclusive events) from 6,695 genes expressed in the ileal transcriptome (46.2% of total expressed genes) (Fig. 2D,E) Scientific Reports | 6:24964 | DOI: 10.1038/srep24964 www.nature.com/scientificreports/ Fold change Regulation Fold change Regulation Fold change PNMA6A 0.27 D ACE2 1.63 U Regulation ABCC2 1.82 DEFB 0.40 D KIFC3 1.64 U U SPSB4 1.82 EIF1AY 0.46 D EFNA5 U 1.65 U ARHGEF4 1.87 RPL36A 0.48 D U FAM46B 1.66 U SEMA6B 1.88 SAA3 0.52 U D RHOBTB3 1.67 U PDK4 1.89 Bta-mir-425 U 0.55 D MB21D2 1.67 U ENPP3 1.90 U CFAP46 0.55 D COL4A1 1.68 U PCK1 1.92 U BEST2 0.56 D VIP 1.70 U G3 × 7I8 1.94 U GPR113 0.57 D TINAGL1 1.70 U FRMD3 1.95 U PTPRO 0.58 D DNAJA4 1.70 U CDH13 1.97 U Bta-mir-421 0.59 D DDAH1 1.70 U ATP8B1 1.97 U FCER1A 0.61 D WFS1 1.72 U PNCK 1.97 U MPP3 0.62 D KLK10 1.72 U TMEM132C 1.99 U MYBPH 0.65 D IYD 1.73 U ZCCHC12 2.01 U MAFF 1.52 U TRIM40 1.73 U PABPC5 2.01 U RND3 1.52 U CAV2 1.73 U TRPM6 2.02 U ERI1 1.52 U ME3 1.74 U IGFBP2 2.03 U SYT4 1.55 U CPQ 1.74 U HSPA6 2.03 U PECI 1.55 U HOXD9 1.75 U SAMD4A 2.06 U NOS3 1.57 U DUSP26 1.75 U LOC101906113 2.10 U MLF1 1.57 U TMEM56 1.76 U DPF3 2.15 U ART5 1.58 U AGPAT9 1.76 U LEKR1 2.64 U COL4A2 1.58 U E1BLE0 1.78 U PDE7B 2.65 U ITGA1 1.59 U EXOC3L2 1.79 U GRP 2.84 U ADSSL1 1.62 U CCDC70 1.79 U DYTN 3.36 U PRSS23 1.62 U ANGPTL4 1.82 U PAK7 4.41 U SLC16A13 1.63 U GJB3 1.82 U PSAPL1 7.03 U Gene Symbol Gene Symbol Gene Symbol Table 1. The differentially expressed genes between MAP-infected and control tissues detected by RNASeq Notes: “D” indicates downregulated genes in MAP-infected vs control tissues; “U” indicates upregulated genes in MAP-infected vs control tissues; Fold changes were defined as ratios of arithmetic means of FPM when compare MAP-infected vs control tissues The significances were declared at fold change > 1.5 and P 1) was 375 ± 26 in infected compartments and 375 ± 25 in control compartments with 280 miRNAs commonly expressed in all 10 samples The most highly expressed miRNA was miR-143 Although PCA did not show a clear separation (Fig. 2C), hierarchical cluster displayed a separation between the miRNA expression patterns of control and infected tissues (Fig. 2G) Unlike mRNA profile and alternative splicing events, the control compartment of animal #5 was not an outlier (Fig. 2C); however, to be consistent, this animal was excluded from further miRNAs differential expression analyses Altered expression of mRNAs in MAP-infected ileal compartments. The analysis of differentially expressed (DE) mRNAs between infected (n = 4) and control compartments (n = 4) using edgeR19 revealed 81 DE genes (14 downregulated and 67 upregulated in infected tissues vs control tissues; P 1.5, Table 1) PCA plots revealed a clear clustering of these DE genes based on the ileal compartments (infected vs control) (Fig. 3A) The most relevant function of DE genes estimated by the Ingenuity Pathway Analysis (IPA) was “proliferation of endothelial cells”, which showed an activated trend in infected tissues (z-score = 1.01) (Fig. 3B) Moreover, the IPA revealed that the DE genes were significantly related to “glucose metabolism disorder” and downregulated “proliferation of muscle cells” (z-score = −1.67) (Fig. 3B) To evaluate the impact of MAP infection on host innate and adaptive immune responses, genes related to innate (793 ± 6, annotated by GO: 0045087) or adaptive immune responses (220 ± 3, annotated by GO: 0002250) were selected Although some genes in certain individual animals revealed fold changes more than 1.5 (Fig. 3C,D), neither innate nor adaptive immune-related genes expressions showed statistical differences between two compartments (fold change > 1.5, P 1.5 A1, A2, A3 and A4 represent Animal #1, Animal #2, Animal #3, Animal #4, respectively (D) Fold change of adaptive immune-related genes when comparing infected vs control tissues with data presented as described above A1, A2, A3 and A4 represent Animal #1, Animal #2, Animal #3, Animal #4, respectively significantly different alternative splicing events between control and infected compartments of all four animals (Supplementary Table S3) Among them, an alternative first exon event was detected (adjusted P 10%) in the mRNA of monocyte to macrophage differentiation-associated (MMD) (Fig. 4A) In MAP-infected tissues, the expression (fragments per kilobase of exon per million mapped reads (FPKM)) of exon 1A (the first annotated exon in bovine genome UMD 3.1) decreased among all animals (log2FPKM in MAP-infected vs log2FPKM in control: 6.0 vs 3.8 in Animal #1; 5.6 vs 4.3 in Animal #2; 5.8 vs 4.6 in Animal #3; 6.0 vs 4.2 in Animal #4) (Fig. 4A), while the expression of exon 1B (the alternatively spliced form for the first exon) increased in all animals (log2FPKM in MAP-infected vs log2FPKM in control: 1.3 vs 3.6 in Animal #1; 1.5 vs 3.2 in Animal #2; 1.7 vs 3.1 in Animal #3; 1.0 vs 3.4 in Animal #4) when compared to control tissues (Fig. 4A) In addition, an intron retention event was detected (adjusted P 10%) in adenosine deaminase (ADA) transcript (Fig. 4B) A higher expression of intron (intron region between exon and exon of the ADA mRNA) were detected in MAP-infected tissues among all animals, compared to that of control tissues (log2FPKM in MAP-infected vs log2FPKM in control: 1.7 vs 0.8 in Animal #1; 1.2 vs 2.4 in Animal #2; 0.7 vs 2.0 in Animal #3; 0.5 vs 1.9 in Animal #4) (Fig. 4B) Subsequently, further protein sequence analysis (translate RNA sequence to protein sequence) on the above alternative spliced forms of MMD and ADA mRNAs in infected tissues revealed the potential introduction of stop codons, when the detected alternative splicing events happen (Supplementary Table S4) The multiplex reverse transcription quantitative PCR (RT-qPCR) was further performed to verify the differences between MAP-infected and control tissues in alternative spliced forms of above two genes The primers and probes were designed to target the identical isoforms that were detected by RNA-Seq Using the primers and probes designed; only two isoforms were detected for both genes (data not shown) The ratio between the expression of isoform (exon 1A + exon 2) and isoform (exon 1B + exon 2) for MMD decreased significantly (P = 0.0031, paired t-test) in infected tissues (0.21 ± 0.05) when compared to control tissues (0.10 ± 0.05) (Fig. 4C) Similarly, the ratio between the expression of isoform (exon 4 + exon 5) and Scientific Reports | 6:24964 | DOI: 10.1038/srep24964 www.nature.com/scientificreports/ Figure 4. Analysis of alternative splicing events (A) The alternative first exon event of MMD The figure shows the genomic location of this event, and Exon 1A and Exon 1B were alternatively spliced as the first exon of MMD mRNA The Y-axis represents the normalized expression level (log2FPKM) of Exon 1A or Exon 1B (B) The intron retention event for ADA The figure showed the genomic location of this event and Intron was alternatively spliced The Y-axis represented the normalized expression level (log2FPKM) of Exon 4, Exon or Intron RT-qPCR validation of alternative splicing changes for MMD (C) and ADA (D) Primers and probes were designed based on each event Y-axis of the bar plot represented the ratio between the expressions of two isoforms: isoform / isoform *significant difference at P