Tristetraprolin specifically regulates the expression and alternative splicing of immune response genes in HeLa cells

To explore the molecular mechanism of TTP-mediated transcriptional regulation, RNA-seq experiments were performed. As shown in Fig. 1a and b, the efficacy of ZFP36 overexpression was assessed by RT-PCR and western blots, approximately 10 times. Four cDNA libraries were constructed and sequenced using the Illumina Xten platform for 2 × 150 paired-end reads per sample.

After removing the adaptor sequences and low quality sequencing reads, we obtained a total of 98 ± 14.8 million high quality reads from each sample (Table 1). When these reads were mapped onto the human GRCH38 genome using Tophat2, 89.22–92.68% were aligned and about 95.46–96.8% were uniquely aligned (Table 1). To compare the gene expression patterns between individuals, we reassessed the gene and transcript quantifications with Cufflinks [44]. The expression values in units of fragments per kilo base of exon model per million fragments mapped (FPKM) was calculated. RNA-seq yielded robust expression results for 28,837 genes (see Additional file 1: Table S1). Effective overexpression of ZFP36 was further confirmed in a parallel RNA-seq analysis (Fig. 1c). FPKM values for all 28,837 genes were used to calculate a correlation matrix based on the Pearson’s correlation coefficient.

Based on the high quality RNA-seq data obtained from the ZFP36-OE cells and the control, we used criteria of an absolute fold change ≥1.5 or ≤ 2/3, FDR < 0.05 with the edge R package [45] to identify genes that were potentially regulated by TTP at the transcriptional level. We identified 596 and 231 genes that were upregulated and downregulated, respectively. The details of the differentially expressed genes (DEGs) can be found in Additional file 2. A volcano plot was constructed to display the DEGs associated with ZFP36 overexpression (Fig. 1d). A heat map analysis of the DEG expression patterns in the RNA-seq samples displayed high consistency with the TTP-mediated transcription in both data sets (Fig. 2a). These results indicated that TTP extensively regulated gene expression.

Based on the cut-off criteria, the upregulated DEGs were enriched in 66 GO terms, and the downregulated DEGs in 4 GO terms (see Additional files 3 and 4 for details). Strikingly, in the biological process terms of the GO analysis, the upregulated genes in the ZFP36-OE cells were primarily associated with the innate immune response, including the “type I interferon-mediated signaling pathway,” “response to virus,” “defense response to virus,” and “interferon-gamma-mediated signaling pathway.” The TTP up-regulated genes included, IL23A, OASL, TRIM22, IFIT2, HERC5, IFIT3, IFI27, TNF, CSF2, IL6, CCL2, RELB, CXCL2, TRAF1, and DDX58. TTP overexpression also resulted in the upregulation of genes enriched in the “cytokine-mediated signaling pathway” and “positive regulation of the apoptotic process” (Fig. 2b; for details see Additional file 4). In contrast, the downregulated genes were primarily enriched regarding “synaptic transmission.” The enrichment of genes involved in “signal transduction” was also observed (Fig. 2c).

When the corrected p-value for the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were set at < 0.05; no pathways were enriched (Additional file 1: Figure S1; for details see Additional files 5 and 6). Surprisingly, when the adjusted p-value for the KEGG pathways were set at < 0.05, some genes were dysregulated in the TTP overexpressing samples annotated with KEGG categories and were mainly involved in the NF- κB signaling pathway (e.g., RELB, DDX58, CXCL2, IL1R1, and TRAF1) or chemokine signaling pathways (for details see Additional file 6).

We selected DEGs or non-DEGs involved in the immune response for further reverse transcription and quantitative real-time (qPCR) validation analysis. The qPCR results were highly consistent with the sequencing data (Fig. 2d and Table 2).

Since HeLa is a well-characterized HPV+ cervical cancer cell line [46, 47], it could be possible that the activated expression of the immune response genes by ZFP36 overexpression is due to an attempt to reactivate the expression of the HPV genome. To address this possibility, we detected the level of the HPV gene expression (E6, E7, and L1) by qRT-PCR analysis in ZFP36-overexpressed and control cells. None of the HPV genes displayed an increased expression upon ZFP36 overexpression (Fig. 2e). The specific PCR primer pairs that were designed are presented in Table 2 [48]. Moreover, we also mapped the RNA-seq reads in this study onto the HPV genome, which demonstrated that extremely few reads originated from HPV and supported the lack of HPV activation. These results confirmed the lack of HPV activation upon ZFP36 overexpression in HeLa cells.

TTP negatively regulates immune response genes by binding to the ARE elements located at the 3′-UTR of the mRNAs of these target genes, which leads to target mRNA degradation [10]. We hypothesized that the upregulated gene expression in HeLa cells should not reflect its binding to AU-rich elements in the 3′-UTRs of immune response genes. Consistent with this hypothesis, we found that apart from IL-23A, the level of all reported TTP target gene expression involving the binding of the mRNA AU-rich elements was not substantially altered (Fig. 3a). To correlate TTP-regulated gene expression with its 3′-UTR binding activity, we analyzed the previously published iCLIP data [28] and obtained mouse macrophage mRNAs containing TTP binding peaks at the 3′-UTR regions. We then calculated the percentage of TTP-regulated genes in HeLa cells whose mouse homologues contain a TTP binding peak in the 3′-UTR of mRNA in mouse macrophages. The statistical analysis showed that TTP-mediated down-regulated genes contained more 3′-UTR binding sites in mice, consistent with the view that TTP may downregulate gene expression in HeLa cells via its canonical 3′-UTR binding activity (Fig. 3b). To further test whether the TTP-regulated genes in HeLa cells are related to the AU-rich elements in the 3′-UTRs, we performed a motif search analysis, which showed that the frequency of AU-rich elements were substantially higher in the TTP-mediated downregulated genes (Fig. 3c). Taken together, these results confirmed that TTP may downregulate the expression of some genes in HeLa cells via its known 3′-UTR binding activity. Moreover, these findings are consistent with our hypothesis that TTP might upregulate gene expression at the transcriptional level, which may be distinct from the TTP-mRNA binding activity.

One key objective of this study was to gain an insight into the role of TTP in the regulation of alternative splicing. To use transcriptome data to explore the TTP-dependent AS events in HeLa cells, we assessed the data quality using an analysis of alternative splicing. Among the 83.3 M ± 14.2 M uniquely mapped reads from ZFP36-OE and control HeLa cells, 40.64–41.82% were junction reads (Table 1). When compared to the reference genome annotation, we found 68.66% annotated exons (252,220 out of 367,321 annotated exons), 166,260 annotated splice junctions, and 211,482 novel splice junctions by using the Tophat2 pipeline.

To investigate the global changes in the AS profiles in response to ZPF36 overexpression, AS events were analyzed from the RNA-seq dataset using ABLas software [49]. We detected 21,681 known alternative splicing events (ASEs) in the model gene that we termed the reference genome, and 49,136 novel ASEs, excluding intron retention events (Additional file 1: Table S2).

We applied a stringent cut-off p-value of ≤0.05, changed T-value ≥0.2 (See Methods) to identify regulated alternative splicing events (RASEs) with a high confidence, which resulted in 497 RASEs (the complete RASEs can be found in Additional files 7 and 8). TTP-regulated AS events included the Cassette exon (69)/exon skipping (86 ES), alternative 5′ splice site (127 A5SS), and alternative 3′ splice site (137 A5SS) (Fig. 4a). These data suggested that TTP globally regulated alternative splicing events in HeLa cells. To analyze whether the increase in AS events could be attributed to altered transcription [50], we overlapped the genes whose level of expression and alternative splicing were both regulated by TTP, and identified seven such genes: GARNL3, PVRL4, ABCA5, TTC21A, HES4, LRRC37A11P, and SLX1B-SULT1A4 (Fig. 4b). This finding indicated that transcriptional regulation and alternative splicing might be partially coupled.

It was further revealed that the genes regulated by TTP-mediated alternative splicing were highly enriched for “positive regulation of I-kappaB kinase/NF-kappaB,” “TRIF-dependent toll-like receptor signaling pathway,” “MyD88-independent toll-like receptor signaling pathway,” “toll-like receptor 3 signaling pathway,” and “regulation of transcription, DNA-dependent” (GO biological process terms, Fig. 4c). Enriched KEGG pathways (p < 0.05) included those involved in “Renal cell carcinoma,” “Notch signaling pathway,” and “TNF signaling pathway” (Fig. 4c).

To validate the alternative splicing events identified by the RNA-Seq data, 16 potential alternative splicing events were analyzed by q-PCR. PCR primer pairs (Table 2) were designed to amplify the long and short splicing isoforms in the same reaction. Out of 16 tested events, 13 alternative splicing events validated by q-PCR were in agreement with the RNA-Seq results. The 13 validated splicing events were located in the following genes (TRIM38, IRF3, TLR4, LTBR, CASP8, MEF2A, MAP2K7, ATF2, SLC44A2, CARHSP1, CREM, TEAD2, and MYCBP2) (Fig. 5 and Additional file 1: Figure S2).

Primer pairs used for Hot Fusion were designed by CE Design V1.04 with gene-specific sequences, in conjunction with the portion of the vector pIRES-hrGFP-1a sequences, each with a 17 bp–30 bp overlap.

F-primer: agcccgggcggatccgaattcATGGATCTGACTGCCATCTACG

R-primer: gtcatccttgtagtcctcgagCTCAGAAACAGAGAGGCGATTG

The human HeLa cell line was purchased from the Institute of Biochemistry and Cell Biology (Chinese Academy of Sciences, Shanghai, China). The cells were seeded into a petri dish (100 mm) at a density of 1 × 10⁵ cells/mL per well, and cultured at 37 °C with 5% CO₂ in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% fetal bovine serum (FBS) (Hyclone), penicillin (100 U/mL), and streptomycin (100 g/mL). The transfection of HeLa cells with a ZFP36-overexpressing plasmid was performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. The transfected cells were harvested after 48 h.

GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was used as a control. cDNA synthesis was performed using standard procedures and real time PCR was performed on the Bio-Rad S1000 with Bestar SYBR Green RT-PCR Master Mix (DBI Bioscience). The concentration of each transcript was then normalized to the level of GAPDH mRNA using the 2- ΔΔCT method [60]. Comparisons were performed with a paired Student’s t-test by using GraphPad Prism software (San Diego, CA).

Total RNA was extracted using TRIzol (Ambion). The RNA was further purified with two phenol-chloroform treatments and RQ1 DNase (Promega) was administered to remove the DNA. The quality and quantity of the purified RNA was assessed by measuring the absorbance at 260 nm/280 nm (A260/A280) using a Smartspec Plus (BioRad). The integrity of the RNA was further verified by 1.5% agarose gel electrophoresis.

For each sample, 1 μg of the total RNA was used for RNA-seq library preparation via VAHTS Stranded mRNA-seq Library Prep Kit (Vazyme). Polyadenylated mRNA was purified and fragmented, and then converted into double stranded cDNA. After the step of end repair and A tailing, the DNAs were ligated to VAHTS RNA Adapters (Vazyme). The purified ligation products corresponding to 200–500 bps were digested with heat-labile UDG, and the single stranded cDNA was amplified, purified, quantified, and stored at − 80 °C before sequencing.

For high-throughput sequencing, the libraries were prepared following the manufacturer’s instructions and applied to the Illumina HiSeq X Ten system for 150 nt paired-end sequencing.

In this study, to elucidate the validity of the RNA-seq data, qPCR was performed for some selected DEGs, and normalized to the reference gene, GAPDH. Information regarding the primers is presented in Table 2. The same RNA samples for RNA-seq were used for qPCR. The PCR conditions consist of denaturing at 95 °C for 10 min, 40 cycles of denaturing at 95 °C for 15 s, followed by annealing and extension at 60 °C for 1 min. PCR amplifications were performed in triplicate for each sample.

Moreover, a qPCR assay was used to analyze the alternative splicing events in HeLa cells. The primers used to detect the pre-mRNA splicing are presented in Table 2. To detect one of the alternative isoforms, one primer was designed in the alternative exon, and an opposing primer was designed in a constitutive exon. To detect the other alternative isoform, a boundary-spanning primer for the sequence encompassing the exon-exon junction with the opposing primer in a constitutive exon was used.

The adaptors and low quality bases were trimmed from raw sequencing reads by using FASTX-Toolkit (Version 0.0.13). The filtered reads were then aligned to the GRCh38 genome by TopHat2 [61], allowing four mismatches. Uniquely aligned reads were selected to calculate the fragments per kilobase per million mapped reads (FPKM) that represents the expression levels of genes [44]. The software edgeR [45] was utilized to screen for DEGs, based on fold change (FC ≥ 1.5) and false discovery rate (FDR < 0.05). To predict the gene function and calculate the functional category distribution frequency, enriched KEGG pathway and Gene Ontology (GO) terms were identified by using KOBAS 2.0 server [62] and in house tool, respectively. Hypergeometric test and Benjamini-Hochberg FDR controlling procedure were used to define the enriched significance of each pathway (corrected p-value < 0.05). The alternative splicing events (ASEs) and regulated alternative splicing events (RASEs) between the samples were defined and quantified using the ABLas pipeline as described previously [49]. After detecting the ASEs in each RNA-seq sample, Fisher’s exact test was used to calculate significant p-values, with the alternative reads and model reads of the samples as input data, respectively. The change ratio of alternatively spliced reads and constitutively spliced reads between the compared samples was defined as the RASE ratio. We set p-value < 0.05 and RASE ratio > 0.2 as the thresholds for RASE detection.

To explore the mRNA binding profile of TTP, we obtained and analyzed mouse macrophage mRNAs containing TTP binding peaks at the 3′-UTR regions from published data (GSE63466).