RNA-binding protein PTBP1 mediates HSV-1 attachment and infection through regulation of heparan sulfate 3-O-sulfotransferase gene expression

Abstract

Background

The RNA-binding protein polypyrimidine tract-binding protein 1 (PTBP1), also known as heterogeneous nuclear ribonucleoprotein I (hnRNP I), mediates gene expression through splicing regulation. Its role in virus infection is undefined.

Results

We show that genetic ablation of PTBP1 renders cell resistant to herpes simplex virus 1 (HSV-1) infection. HSV-1 utilizes 3-O-sulfated heparan sulfate proteoglycans (HSPGs) for attachment and for infection of epithelial cells. We found that knockout of PTBP1 expression resulted in loss of HS3ST3A1 and HS3ST3B1, heparan sulfate glucosaminyl 3-O-sulfotransferase genes for 3-O-sulfation of the heparan sulfate (HS) chains of HSPGs. Each of the HS3ST3A1/HS3ST3B1 genes is composed of 2 exons separated by an extraordinarily long intron whose removal requires PTBP1-associated looping. We found that PTBP1 interacted with the intronic region of HS3ST3A1/HS3ST3B1 pre-mRNAs and modulated their processing to mRNA. The essential role of PTBP1 in functional HS3ST3A1 expression and in HSV-1 infection was demonstrated by ectopic re-expression in the knockout (ko) cells. In addition, we showed that targeting PTBP1 expression by microRNA mimics reduced disease symptoms in a mouse herpetic stromal keratitis (HSK) model.

Conclusions

The results demonstrate that PTBP1 mediates HSV-1 infection of epithelial cells through splicing regulation of HS3ST3A1/HS3ST3B1. These studies provide a new area for novel therapeutic strategies through splicing regulation.

Graphical abstract

By Sun et al. RNA-binding protein PTBP1 mediates HSV-1 attachment and infection through regulation of heparan sulfate 3-O-sulfotransferase gene expression.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13578-025-01495-7.

Background

pre-mRNA alternative splicing (AS) occurs in more than 90% of genes in the human genome and can affect protein functions by influencing enzymatic activity, protein locations or interactions [1, 2]. AS is carried out by spliceosome that can recognize splice signals and join alternative exons by simultaneously removing the much longer intronic sequences [3]. In vertebrates, where introns have a median length of approximately 1 kb, exon definition interactions predominate [4, 5]. Long-range RNA looping occurs so that distal regions of pre-mRNA transcripts are brought together by RNA-binding proteins (RBPs) that recognize sequences or structures at distal sites on the same RNA or by RNA-RNA complementarity [6].

PTBP1 is an RBP and functions as a major splicing regulator that can simultaneously recognize multiple binding sites in the same transcript with its RNA recognition motif-type domains (RRMs) [6–9]. The PTBP1-RNA interaction promotes PTBP1-mediated RNA loop formation that defines cassette exon skipping or intron removal [9, 10]. Analyses of the 3D map of PTBP1-RNA interaction uncovered that intronic RNA loops tend to facilitate asymmetrical intron removal, whereas loops spanning across cassette exon primarily repress splicing [9].

In a preliminary study, we observed loss of HS3ST3A1 and HS3ST3B1 expression from PTBP1-ko HeLa cells. The HS3ST3A1 and HS3ST3B1 genes, also known as 3OST3A1 and 3OST3B1, respectively, encode heparan sulfate-glucosaminyl 3-sulfotransferases (HS3STs) that catalyze the 3-O-sulfation of the heparan sulfate chains of HSPGs [11–13]. HS is a group of widely distributed polysaccharides on the cell surface and in the extracellular matrices. These bioactive polysaccharides provide docking sites for cytokines and ligands involved in diverse biological processes [14–17]. A wide range of pathogens, especially bacteria and viruses including SARS-CoV-2 and HIV-1 viruses, also use HSPGs as receptors or co-receptors to mediate their attachment to host cells [18–22]. Although the 3-O-sulfated HS represents only a small percentage of HS from biological sources, this subpopulation is closely associated with HS biological functions [23, 24]. For example, heparin, an HSPG approved by the FDA as an anticoagulant, contains a 3-O-sulfated glucosamine residue critical for antithrombin binding activity [25, 26]. Human herpes simplex virus 1 (HSV-1), one of the most common pathogens of human infection, employs 3-O-sulfated HS for initial attachment and infection [27–29]. Thus, understanding how PTBP1 regulates HS3ST3A1/B1 expression will inform the use of HS mimetics in anticoagulant and in antiviral agent development.

Each of the HS3ST3A1 and HS3ST3B1 genes is composed by two exons separated by a long intron (103.7 kb in HS3ST3A1 and 43.0 kb in HS3ST3B1, respectively). We speculated that the decreases in HS3ST3A1/B1 expression could be a result of ineffective removal of the long introns without PTBP1 participation. Whether PTBP1 mediates HS modification gene expression remains unknown.

In this study, we focused on PTBP1 regulation of HS3ST3A1/B1 expression and on HSV-1 infection and found that PTBP1 mediates HSV-1 infection through proper splicing of HS3ST3A1 and HS3ST3B1. Here, we report the identification of PTBP1 as a host factor mediating HSV-1 infection through splicing regulation of HS modification genes.

Materials and methods

Ethics statement

A protocol for the use and care of laboratory animals was approved by the IACUC of Nanjing University (D2103051). All investigations are subjected to regulations of Jiangsu Province Laboratory Animal License Management Measures and conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research (2021 edition).

Cells, virus, and reagents

HEK293T (ATCC, CRL-3216, human embryonic kidney cells), Vero (ATCC, CCL-81, African green monkey kidney epithelial cells), HeLa (ATCC, CCL-2, human cervical carcinoma), and N2a (ATCC, CCL-131, mouse Neuro-2a neuroblastoma) cells were cultured in DMEM (Life Technologies) plus 10% fetal bovine serum (FBS) and penicillin/streptomycin (Life Technologies) at 37 °C in an incubator with 95% humidified air and 5% CO₂. The cells were tested routinely for Mycoplasma contamination with a PCR detection kit (Beyotime, Shanghai). PTBP1- and HS3ST3A1-ko cells were generated in the lab using the CRISPR/Cas9 system. A CRISPR design tool (available at crispr.mit.edu) was used to select single-guide RNA (sgRNA) targeting PTBP1 (accession NM_002819.5, 5’-CTTCATCGAGATGAACACGG-3’) or HS3ST3A1 (accession NM_006042.3, 5’-ACCCTGTCCGGCCCCGTCGT-3’). To generate PTBP1- or HS3ST3A1-ko cells, HeLa or 293 T cells were transfected with the sgRNA and GenCrispr Cas9-C-NLS Nuclease (Z03385, GenScript, Nanjing) using Lipofectamine 2000 (Life Technologies). Individual clones were selected and validated by immunoblotting analysis for absence of the corresponding protein and by DNA sequencing. A Cell Counting Kit-8 (C0038, Beyotime) was used for assessment of cell viability by following the manufacturer’s instructions.

HSV-1 strain F was obtained from ATCC (Manassas, USA). The antibodies against PTBP1 (Abcam, ab133734), HS3ST3A1 (SinoBiological, 209,208-T08), GAPDH (Abcam, ab9484), HS3ST3A1 (Santa Cruz, sc390024, mouse specific), the HSV-1 ICP0 (Santa Cruz, sc53070), ICP27 (Santa Cruz, sc69806), and gB (Santa Cruz, sc56987) were obtained from commercial sources. Acyclovir (ACV, A126073) was purchased from Aladdin Scientific (Shanghai).

For virus titrations, serially-diluted HSV-1 was allowed to adsorb to Vero cells at 37 ℃ for 1 h. The cells were then overlaid with a medium containing carboxymethyl cellulose (CMC) and incubated at 37 °C in a humidified incubator with 5% CO₂ until well-defined plaques were visible. After fixation, the cells were stained with crystal violet for enumeration for virus titers [30, 31]. The titers were expressed as plaque-forming units per milliliter (PFU/ml).

Preparation of bulk RNA library for RNA-sequencing (RNA-seq)

For bulk RNA-seq library preparation, the parental and PTBP1-ko HeLa cells were collected and total RNA was extracted using TRIzol reagent to generate mRNA-seq library using KAPA stranded mRNA library preparation kit (KK8421, KAPA Biosystems). Samples were prepared for bulk RNA-seq libraries preparation and were sequenced using 150 bp paired-end Illumina NovaSeq 6000 platform by Annoroad Gene Technology Company (Beijing). The data have been deposited at the NCBI (accession number of GSE297152).

For bulk paired-end RNA-seq raw data, Skewer (v0.2.2) was used to trim adaptor and low-quality bases (q < 20). After quality control by FastQC (v0.11.5), trimmed sequencing reads were mapped to hg38 reference genome by using STAR (2.5.2b) with parameters (–twopassMode Basic –outSAMstrandField intronMotif –alignSJstitchMismatchNmax 5 -1 5 5). Subsequently, the StringTie (v2.2.1) was used to assign mapped reads to genomic features with hg38 Ref Seq gene annotation downloaded from UCSC table browser and normalized the count reads to FPKM. The edgeR/DESeq2 (v1.38.1) was used for analyzing differentially expressed genes with a cut-off of |Log₂FC|≥ 1 and p ≤ 0.05. The genes with altered transcription post PTBP1 ko were imported into the DAVID database (https://davidbioinformatics.nih.gov/) to perform GO and the KEGG pathway enrichment analysis. The top GO enrichments and KEGG pathways showing the highest counts based on a significance threshold of p ≤ 0.05 were selected.

PTBP1-RNA profiling using publicly available data

To analyze PTBP1-RNA interaction, raw data files were retrieved from the Gene Expression Omnibus (GEO) database. These included GSE57278, GSE114720 (both 293 T cells), GSE137925 (K562 cells), and GSE210583 (HeLa cells). Peak calling with MACS2 (v2.2.6) was performed to identify regions of PTBP1 interaction (https://github.com/jsh58/MACS). The Integrative Genomics Viewer (IGV) was used for visualization of PTBP1 RNA-binding events.

RNA immunoprecipitation (RIP) assay

The RIP assay was performed by following a reported protocol [33]. The parental and PTBP1-ko 293 T cells in 10-cm culture plates were collected with a cell scrapper into 1 ml of a polysome lysis buffer containing 10 mM HEPES (pH 7.4), 100 mM KCl, 0.5% NP-40, 0.1% SDS, 0.5 mM DTT, 20 U/ml RNaseOUT (Invitrogen), and a cocktail of Roche EDTA-free protease inhibitors. After incubation on ice for 5 min with occasional vortexing, cell lysates were collected by centrifugation and used for immunoprecipitation by incubating with BeyoMag Protein A/G magnetic beads (Beyotime, P2108) that were pre-bound with anti-PTBP1 or a control IgG for 3 h at 4 °C. The beads were then washed with a wash buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM MgCl₂, 1 mM DTT) for five times, followed by resuspension in 150 μl digestion buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM MgCl₂, 0.05% NP-40, 1% SDS) for proteinase K (1.2 mg/ml) digestion 55 °C for 30 min. After deactivation by heating at 65 °C for 5 min, the samples were then treated with 15 μg/ml DNase I (Beyotime, D7073) at 37 °C for 10 min to remove potential DNA contamination. Total RNA in these samples was collected by ethanol precipitation after phenol–chloroform extraction and was used for RT-PCR studies. The primers are listed in Table S1.

RT-PCR and real time semi-quantitative PCR (RT-qPCR)

Gene expression was determined by reverse transcription qPCR (RT-qPCR). Total RNA from cells or mouse tissues was extracted using TRIzol reagent. For reverse transcription, total RNA (1 μg) was first treated with gDNA wiper (Vazyme) to remove genomic DNA contamination, then reverse transcribed into cDNA using HiScript III SuperMix (Vazyme, #R323) and used for analysis by RT-PCR or RT-qPCR. The primers are listed in Table S1.

Viral genomic DNA was extracted from murine trigeminal ganglion as reported [32]. Viral DNA in trigeminal ganglion was quantified by measuring the viral TK DNA on a QuantStudio 5 real-time PCR system (Applied Biosystems) using ChamQ Universal SYBR qPCR Master Mix (Vazyme, #Q711). The 2^−ΔΔCT method was used for calculation. The copy numbers of viral genome were estimated using a fragment of known amount TK DNA as a standard for the calculation.

RNA decay analysis

To evaluate whether PTBP1 affects the stability of HS3ST3A1 and HS3ST3B1 mRNA, an RNA decay assay was performed using actinomycin D (ActD; Sigma, SBR00013) to inhibit host transcription. The parental and PTBP1-ko HeLa cells were treated with 5 μg/ml ActD for varying times and then used for total RNA extraction. HS3ST3A1 and HS3ST3B1 transcript levels in the samples were quantitatively measured by RT-qPCR using GAPDH as an internal control for normalization. The percentage of the mRNA remaining at each time point was calculated using untreated control (0 h) as 100%. The half-lives of the mRNA were extrapolated.

Subcellular fractionation for RNA distribution analysis

The nuclear and cytoplasmic fractionation was performed by following a reported method [34]. The parental and PTBP1-ko 293 T cells (1 × 10⁷ cells/tube) were resuspended in hypotonic buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0.15% Triton X-100, and a cocktail of EDTA-free protease inhibitors). After incubation on ice for 8 min with occasional pipetting, the samples were centrifuged to separate the cytoplasmic fraction from the nuclear pellet. Total RNA in these fractions was extracted using TRIzol reagent and used for detection of pre-mRNA and mature mRNA by RT-qPCR. The primers used for HS3ST3A1 detection are listed in Table S1.

Plasmid construction and transfection

pCMV3-PTBP1 (NM_002819.5, HG17890-UT) and pCMV3-HS3ST3A1 (NM_006042.3, HG22150-UT) were purchased from SinoBiological. The PTBP1 C23S mutant was synthesized by GenScript. For transfection studies, the plasmids were transfected using Lipofectamine 2000. The cells were used for assays at 48 h post transfection.

Immunoblotting

For immunoblotting, total proteins (10 μg) were resolved by SDS-PAGE. The proteins were then transferred to polyvinylidene difluoride (PVDF) membranes (Millipore), blocked with 5% fat-free milk in TBST for 1 h and incubated with a primary antibody at 4 ℃ for overnight. The signals were detected by incubating with an appropriate secondary antibody (HRP-conjugated) followed by ECL reagent (Beyotime).

HSV-1 binding assay

To assess the impact of PTBP1 ko on HSV-1 binding, the parental and PTBP1-ko HeLa cells were detached and washed three times with ice-cold serum-free DMEM (SFM). Aliquots of cells (2 × 10⁶ cells in 100 μl SFM/tube) were incubated with HSV-1 at approximately 10 PFU/cell on ice for an hour. At the end of the incubation, the cells were washed with SFM for 3 times to remove unbound virus. Cell-attached virus was detected by immunoblotting with an anti-gB antibody and by qPCR to quantitatively measure viral genomic DNA after cell lysis. All steps were performed on ice or kept at 4 °C to ensure that only surface-bound virus was detected.

Mouse herpetic stromal keratitis (HSK) model for miRNA mimics with agomiR-124

The miR-124-3p and agomiR-124 that mimics mature miR-124-3p and a control agomiR-NC were purchased from GenePharma (Shanghai, China). The effect of miR-124-3p and agomiR-124 on PTBP1 expression was confirmed by knockdown of PTBP1 expression in cell cultures or in mice.

To knockdown PTBP1 expression in animals, Balb/C inbred female mice of 6–8 weeks old (Beijing Vital River Laboratory Animal Technology Co., Ltd) were anesthetized by IP injection of ketamine (100 mg/kg) and xylazine (5 mg/kg). Intrastromal injection of agomir was performed with deeply anesthetized mice by following reported protocols [35, 36]. In brief, a small intrastromal pocket was carefully created in the mid-peripheral cornea using a 29-gauge needle. The 33-gauge needle was then inserted toward the central cornea, and the agomiR-124 or agomiR-NC (at 500 μM in normal saline, approximately 2 μl in volume) was delivered into the corneal stroma under a stereoscopic microscope (Olympus, SZ61). The effect of agomiR-124 on PTBP1 as well as HS3ST3A1 expression was monitored by RT-qPCR. At 48 h post agomiR-124 treatment, PTBP1 and HS3ST3A1 expression was reduced by ~ 80% and ~ 50%, respectively.

To test the therapeutic potential, the mice (n = 3) were infected at 48 h post agomir treatment by corneal scarification with HSV-1 (5 × 10⁴ PFU in 5 μl PBS) by following a reported protocol [37]. In parallel experiment, a group of antiviral drug acyclovir-treated mice (ACV, 50 mg/kg, ip, daily) was included as a control for antiviral treatment. Mice were monitored daily, and ocular swabs were collected on day 3 and day 5 post infection. The mice were killed on day 7 post infection and used for the detection of viral genomic DNA in the trigeminal ganglia by qPCR.

Statistical analysis

Statistical analysis was conducted with SPSS 17.0 software package. All data are presented as mean ± SD. For comparisons between two groups, an unpaired two-tailed Student’s t-test was used and one-way ANOVA for multiple group comparisons. Differences were considered using * p ≤ 0.05, ** p ≤ 0.01, and *** p ≤ 0.001. ns stands for no significance.

Results

Loss of HS3ST3A1 and HS3ST3B1 expression from PTBP1 deletion

PTBP1 is a well characterized splicing regulator and influences pre-mRNA splicing and RNA metabolism. To investigate the consequence of PTBP1 deletion on host gene expression, we generated PTBP1-ko HeLa cells (Fig. 1A) and carried out RNA-seq analysis to identify globally altered transcripts. As expected, PTBP1-ko cells had significantly altered patterns of gene expression (Fig. 1B). We identified 993 significantly up-regulated genes along with 683 down-regulated genes from PTBP1-ko HeLa cells (Fig. 1C). The impacted genes were highly concentrated in several signaling pathways, including proteoglycan processing and binding, MAPK signaling, system development, cell differentiation, and HSPG modifications (Fig. 1D and E).

Fig. 1.

Global transcription profile identifies PTBP1 as a regulator of HS3ST3A1 and B1 expression. A PTBP1 knockout using CRISPR/Cas9 by targeting PTBP1 exon 5. The ko effect was confirmed by immunoblotting for absence of PTBP1 expression. B Heatmap depicting altered genes in the parental (WT) and PTBP1-ko HeLa cells (n = 3). C Volcano plot of RNA-seq analysis for altered genes in the parental (WT) and PTBP1-ko HeLa cells, p ≤ 0.05 and |log₂FC|≥ 1. D Gene Ontology analyses of altered genes in PTBP1-ko cells. Top 10 enriched GO terms are highlighted for differentially expressed genes based on their biological processes, cellular components, and molecular functions. E KEGG enrichment analysis of altered genes in PTBP1-ko HeLa cells. F Heatmaps of 3OST family genes in the parental (WT) and PTBP1-ko HeLa cells. G Schematic illustration of enzymes involved in heparan sulfate chain biosynthesis and modifications. Xyl (Xylose), Gal (Galactose), GlcA (Glucuronic acid), GlcNAc (N-acetylglucosamine), IdoA (Iduronic acid). H Heatmaps of HS biosynthesis gene expression in the parental (WT) and PTBP1-ko HeLa cells. I Analysis of 3OST family genes in the parental (WT) and PTBP1-ko HeLa cells by RT-qPCR. The experiments were performed 3 times independently. ns, no significance, *** p ≤ 0.001. J Validation of PTBP1 and HS3ST3A1 expression in the parental (WT) and PTBP1-ko HeLa cells by immunofluorescence. GAPDH was used as an internal control. The experiment was performed 3 times independently. K Effects of PTBP1 ko on HS3ST3 mRNA stability. The parental and PTBP1-ko HeLa cells were treated with ActD for times as indicated. HS3ST3A1/B1 mRNA decay was assessed by RT-qPCR using GAPDH as the internal control for normalization. Data are represented as a log10 plot of the percentage of mRNA remaining versus mock-treated control. Data are shown as mean ± SD of triplicate measurement

Of particular interest, we observed that PTBP1 ko resulted in loss of HS3ST3A1 and HS3ST3B1 (Fig. 1F), genes encoding 3OST3A1 and 3OST3B1, respectively, for HS chain 3-O-sulfation [14, 38]. PTBP1 ko did not significantly affect other genes involved in the biosynthesis of HS chains (Fig. 1G and H) nor other members of the 3OST family genes as was validated by RT-qPCR (Fig. 1I). The loss of HS3ST3A1 expression in the PTBP1-ko cells was confirmed by immunoblotting (Fig. 1J).

PTBP1 plays a critical role in splicing regulation as well as in RNA maturation and stability [39]. To distinguish whether the reduced expression of HS3ST3A1 and HS3ST3B1 in PTBP1-ko cells was associated with decreased mRNA stability or defects in RNA processing, we performed an RNA decay assay by treating the cells with actinomycin D (ActD). The levels of HS3ST3A1 and HS3ST3B1 mRNA in ActD-treated and mock-treated samples were quantified at different time points post-treatment. PTBP1 ko did not significantly affect HS3ST3A1 and HS3ST3B1 mRNA stability since no significant difference in mRNA stability was detected between the parental and PTBP1-ko HeLa cells (Fig. 1K).

The human brain tissues are known to have low levels of PTBP1 expression due to tissue specific microRNA repression [40–42]. Analysis of the Genotype-Tissue Expression (GTEx) Consortium (http://www.gtexportal.org/home/) atlas of genetic regulatory effects across human tissues uncovered low expression of both HS3ST3A1 and HS3ST3B1 alongside low PTBP1 expression in the brain tissues (Fig. S1). These results indicated that reduced HS3ST3A1 and HS3ST3B1 expression arises naturally in low PTBP1 expression settings.

Loss of PTBP1 expression causes cells resistant to HSV-1 infection

HSV-1 uses 3-O-sulfated HSPGs for cell surface attachment and infection of epithelial cells [28]. Thus, reduced expression of HS3ST3A1/B1 could potentially impair cell susceptibility to HSV-1 infection due to reduction in cell attachment. In this regard, we performed cell attachment assay by incubating HSV-1 virus with the parental and PTBP1-ko HeLa cells. Consistent with their role in supporting HSV-1 infection [28, 43], loss of HS3ST3A1/B1 expression significantly impaired HSV-1 attachment to the PTBP1-ko cells (Fig. 2A).

Fig. 2.

Depletion of PTBP1 expression reduces HSV-1 binding and infection due to loss of 3OST3. A Depletion of PTBP1 expression on HSV-1 cell binding. Left panel: schematic of the viral binding assay. Middle panel: qPCR analysis of viral genomic DNA in the parental (WT) and PTBP1-ko HeLa cells. Right panel: immunoblotting of HSV-1 gB protein. GAPDH was used as a loading control. B and C PTBP1 ko on HSV-1 infection in HeLa cells. The cells were infected with HSV-1 for 24 h. HSV-1 ICP0 and ICP27 production (B) and HSV-1 titers in culture medium (C) were detected. D and E. PTBP1 ko on HSV-1 infection in 293 T cells. The cells were infected with HSV-1 for 24 h. HSV-1 ICP0 and ICP27 production (D) and HSV-1 titers in culture medium (E) were detected. F Re-expression of PTBP1 on HSV-1 infection. PTBP1-ko 293 T cells were transfected for ectopic expression of PTBP1 for 48 h. The cells were then infected with HSV-1 for 24 h. The expression of PTBP1 and HSV-1 ICP proteins were examined by immunoblotting. Vec, empty vector. G HS3ST3A1 ko using CRISPR/Cas9 by targeting exon 1 of HS3ST3A1. H Expression of 3OST family genes in 293 T cells determined by RT-qPCR using GAPDH as an internal control for normalization. The gene expression was presented relative to HS3ST3A1. I and J Genetic ablation of HS3ST3A1 on HSV-1 infection in 293 T cells. HSV-1 infection was determined by immunoblotting (I) and by plaque forming assay (J). K Re-expression of HS3ST3A1 restores cell susceptibility to HSV-1 infection. HS3ST3A1-ko 293 T cells were transfected with pCMV3-HS3ST3A1 for 48 h and then infected with HSV-1 infection for 24 h. The HS3ST3A1 and viral gene ICP27 expression were examined by immunoblotting. L and M Expression of HS3ST3A1 in PTBP1-ko cells restores cell susceptibility to HSV-1 infection. HSV-1 infection was determined by immunoblotting (L) and by RT-qPCR (M). N Overexpression of PTBP1 in HS3ST3A1-ko cells on HSV-1 infection. HS3ST3A1-ko 293 T cells were transfected with a PTBP1 expression plasmid or empty vector, followed by HSV-1 infection. PTBP1 and viral protein ICP0 expression was assessed by immunoblotting. The experiments were performed 3 times independently. Data are shown as mean ± SD. ns, no significance, ** p ≤ 0.01, *** p ≤ 0.001

We also performed an infection assay to demonstrate the effect of PTBP1 ko on HSV-1 infection. HeLa cells were permissible to HSV-1 infection as was demonstrated by expression of HSV-1 infected cell proteins (ICP0 and ICP27) and by production of infectious virus (Fig. 2B and C). Genetic ablation of PTBP1 blocked HSV-1 infection. In general, we detected close to 3 log reduction of virus titers in the PTBP1-ko cells compared to that of the parental HeLa cells. Together, these results demonstrated that loss of PTBP1 expression in HeLa cells caused cell resistant to HSV-1 infection through reduction in virus attachment.

The essential role of PTBP1 expression in HSV-1 infection was confirmed by re-expression of PTBP1 protein in PTBP1-ko cells. For this purpose, we generated PTBP1-ko 293 T cells since these cells have a significantly higher transfection efficiency. Similar to what was observed with HeLa cells, PTBP1-ko 293 T cells had marked reduction in viral protein expression and in HSV-1 infection (Fig. 2D and E). Fittingly, we found that ectopic expression of PTBP1 by transient transfection restored cell infectivity to HSV-1 infection (Fig. 2F). Thus, HSV-1 infection required PTBP1 expression.

Finally, we performed two additional assays to demonstrate that PTBP1-mediated HSV-1 infection was through HS3ST3A1/B1 expression. Firstly, we chose to knockout HS3ST3A1 from 293 T cells (Fig. 2G) and used the cells for the HSV-1 infection assay since the cells mainly expressed HS3ST3A1 (Fig. 2H). The 293 T cells (WT) were susceptible to HSV-1 infection. Knockout of HS3ST3A1 resulted in significant reduction of HSV-1 ICP27 expression (Fig. 2I) and infectious HSV-1 production (Fig. 2J). Ectopic re-expression of HS3ST3A1 in these ko cells rescued HSV-1 infection (Fig. 2K), confirming that HSV-1 productive infection required HS3ST3A1 expression. Secondly, we tested if expression of HS3ST3A1 in the PTBP1-ko 293 T cells would rescue HSV-1 infection. Consistent with its role in HSV-1 infection, HS3ST3A1 expression in PTBP1-ko cells allowed HSV-1 infection (Fig. 2L and M), independent of PTBP1 expression; whereas overexpression of PTBP1 in the HS3ST3A1-ko cells failed to rescue HSV-1 infection in these cells (Fig. 2N). These results thus demonstrated that PTBP1 mediated HSV-1 infection likely through HS3ST3A1/B1 expression regulation.

Requirement of PTBP1 for HS3ST3A1/B1 pre-mRNA processing

As an RNA-binding protein, PTBP1 modulates multiple aspects of RNA metabolism through interactions with pre-mRNAs. Analysis of PTBP1-RNA interactions from publicly available datasets revealed abundant interactions of PTBP1 with the pre-mRNA of HS3ST3A1 and HS3ST3B1 (Fig. 3A and S2). Notably, the association of PTBP1 with pre-mRNA was mainly confined within the intron region likely due to their overwhelming length (Fig. 3A). To confirm a direct interaction between PTBP1 and the introns of HS3ST3A1/HS3ST3B1 pre-mRNA, we performed a RIP assay and detected specific interactions of PTBP1 with HS3ST3A1/HS3ST3B1 pre-mRNA using gene specific primers targeting the intronic regions (Fig. 3B).

Fig. 3.

Requirement of PTBP1 for HS3ST3A1/B1 pre-mRNA processing. A Profiling of PTBP1 binding to HS3ST3A1 and HS3ST3B1 pre-mRNA. The publicly available data (GSE57278) was retrieved and used to obtain RNA binding profiles of HS3ST3A1 and HS3ST3B1 pre-mRNA association with PTBP1 in 293 T cells. B Validation of PTBP1 and HS3ST3A1/B1 pre-mRNA interactions by RNA-immunoprecipitation assay. RIP was performed using 293 T cells with an anti-PTBP1 antibody or IgG as a control. Two sets of primers (L and R) targeting the corresponding intronic region of HS3ST3A1 or B1 were used to detect immunoprecipitation enrichment by qPCR. C RT-PCR validation of HS3ST3A1/B1 pre-mRNA processing in parental HeLa (wt HeLa) and PTBP1-ko cells. Diagram showing the pre-mRNA of HS3ST3A1 and HS3ST3B1 and primers used for the detection of pre-mRNA and mRNA by RT-PCR. Total RNA was first treated with gDNA wiper to remove contamination from gDNA, followed by reverse transcription. The expression of corresponding genes was detected by PCR amplification (+ RT) using non-transcriptase treated samples (-RT) as controls. Primer combinations spanning the 2 exons (F1/R2, F2/R2, and F2/R3) were included to demonstrate lack of gene splicing events in the PTBP1-ko cells. The experiment was performed 3 times independently. GAPDH was used as a loading control. D Diagram showing the corresponding sequences of WT PTBP1 and the C23S mutant. E Detection of PTBP1 dimer formation by immunoblotting. PTBP1-ko 293 T cells were transfected with WT or PTBP1 C23S mutant for ectopic re-expression of PTBP1. The cell lysates were separated by SDS-PAGE under reducing condition or non-denaturing native condition. GAPDH was used as a loading control. The experiment was performed 3 times. F Re-expression of WT PTBP1, but not C23S mutant, reinstates proper HS3ST3A1/B1 expression detected by RT-PCR. G and H Ectopic expression of WT but not PTBP1 C23S mutant rescues infectivity to HSV-1 infection. PTBP1-ko 293 T cells were transfected with WT or PTBP1 C23S mutant for 48 h. PTBP1 expression and HSV-1 ICP27 production in the parental 293 T (WT) and ko cells were detected by immunoblotting using GAPDH as a loading control (G). Similarly, viral gene expression was confirmed by RT-qPCR (H). The experiments were performed 3 times independently with representative gels shown. Data are presented as mean ± SD. ns, no significance, * p ≤ 0.05, *** p ≤ 0.001

PTBP1 association with intronic RNA tends to induce RNA loops facilitating asymmetric intron removal [6, 9]. To demonstrate if loss of HS3ST3A1 and HS3ST3B1 expression was resulted from ineffective pre-mRNA processing without PTBP1 expression, we adopted an RT-PCR approach by measuring mRNA production in these cells (Fig. 3C). We designed 2 pairs of primers located within individual exons and another pair that would bridge the two exons for mRNA detection (Table S1). We detected the existence of individual exons from the parental and PTBP1-ko HeLa cells with expected sizes (Fig. 3C, Table S1). However, we only detected mRNA from the parental HeLa cells, but not the PTBP1-ko cells by RT-PCR (Fig. 3C), using the primer pair for mRNA detection. The band of PCR amplification was specific since combination of primers spanning the 2 exons failed to detect the presence of processed mRNA from the ko cells (Fig. 3C). Thus, proper HS3ST3A1 and HS3ST3B1 pre-mRNA processing required PTBP1 expression.

The conclusion was substantiated by ectopic expression of PTBP1 in the ko cells. PTBP1 regulates gene splicing which requires PTBP1 dimerization [7, 44]. Thus, PTBP1-ko 293 T cells were transfected with a plasmid for WT PTBP1 expression (Fig. 3D). As a control, we also included a dimerization defective PTBP1 mutant (C23S) for the assay [9]. We detected PTBP1 dimer formation in WT PTBP1 transfected, but not the C23S mutant transfected, cells under non-denature condition (Fig. 3E). As expected, ectopic re-expression of WT PTBP1, but not the C23S mutant, resulted in HS3ST3A1/B1 expression in PTBP1-ko 293 T cells (Fig. 3F). Critically, re-expression of WT PTBP1 reconditioned the ko cells to HSV-1 infection (Fig. 3G and H), while transfection with the C23S mutant did not translate to HSV-1 infection in PTBP1-ko cells (Fig. 3G and H).

The defect in HS3ST3A1/B1 pre-mRNA processing in PTBP1-ko cells was also demonstrated by reduced mRNA in the cytoplasmic fractions. For this purpose, the parental 293 T and PTBP1-ko cells were fractionated into cytoplasmic and nuclear fractions. The ratios of pre-mRNA and mature mRNA in these fractions were determined by RT-qPCR with appropriate primers. In PTBP1-ko cells, mature mRNA from the cytoplasmic fraction was markedly reduced compared to those in the parental cells. For comparison, nuclear pre-mRNA levels were significantly elevated in the ko cells, indicating a defect in pre-mRNA splicing in the absence of PTBP1 expression (Fig. S3).

Unlike in human cells, very few HSV-1 genes are known to be spliced, although the same pre-mRNA processing machinery is shared. We selected those HSV-1 genes that have been reported to undergo splicing [45, 46] and checked whether PTBP1-ko also affected their processing (Fig. S4A). Although an overall reduction of viral gene transcripts was detected from the ko cells, which was consistent with reduced viral binding (Fig. 2A), we found PTBP1 ko, and similarly HS3ST3A1 ko, did not affect the splicing patterns of ICP0, ICP22 and UL5 genes compared to those from HSV-1-infected parental 293 T cells (Fig. S4B and C). Thus, loss of PTBP1 or HS3ST3A1 expression did not measurably alter HSV-1 pre-mRNA splicing, rather, these host factors primarily affected infection by limiting viral attachment and entry.

Suppression of PTBP1 expression by microRNA mimicry with agomiR-124 ameliorates herpetic stromal keratitis in mice

Having established the role of PTBP1 in HS3ST3 expression and in HSV-1 infection, we explored a therapeutic potential of PTBP1 suppression against HSV-1 infection using a murine HSK model. The miR-124 has been a well characterized microRNA against PTBP1 expression [40, 41, 47]. We thus performed experiments to validate the effect of miR-124-3p on PTBP1 expression in cell cultures. Transfection with miR-124-3p, but not a control microRNA (miR-NC), significantly reduced PTBP1 expression in both human and mouse cells (Fig. S5A) as well as HSV-1 infection (Fig. S5B and S5C). Importantly, the effect was dependent on PTBP1 expression since reintroduction of PTBP1 by transient transfection with a plasmid encoding PTBP1 reversed the inhibitory effect of miR-124 on HSV-1 infection (Fig. S5D).

For animal studies, we applied microRNA mimics using agomiR-124, a form of chemically modified miRNA agonist suitable for in vivo applications [48, 49]. Since the mouse and human miR-124 as well as its target of human and mouse PTBP1 are identical, we thus constructed an agomiR-124 based on the miR-124-3p sequence information. The effect of agomiR-124 on PTBP1 expression was first tested in mice by intrastromally injection with agomiR-124 (Fig. 4A). Treatment with agomiR-124, but not control agomiR-NC, resulted in significant reduction of PTBP1 and HS3ST3A1 expression (Fig. 4B). At 48 h post the injection, PTBP1 and HS3ST3A1 were detected at approximately 20% and 50% of untreated controls (Fig. 4B), indicating PTBP1-targeting agomiR-124 suppressed PTBP1 and HS3ST3A1 expression both in cell cultures and in animals.

Fig. 4.

Suppression of PTBP1 by microRNA mimics with agomiR-124 remedies herpetic stromal keratitis. A Illustration of PTBP1 knockdown by agomiR-124 on gene expression and on HSV-1 infection. B RT-qPCR analysis of mouse Ptbp1 and Hs3st3a1 expression. The corneal samples were collected at 48 h post agomiR-124 or agomiR-NC injection (n = 3). Mouse Gapdh was used as an internal control. C and D Infectious virus in tears (n = 3) collected on day 3 (C) and day 5 (D) post inoculation. An antiviral drug group was included by daily treatment with 50 mg/kg acyclovir (ACV; ip, qd). E qPCR analysis of HSV-1 genomic DNA in TG collected on day 7 post agomiR-124 or agomiR-NC treatment. Viral genomic DNA was detected with a fragment of the TK DNA as a standard. Genome equivalent was calculated and used to represent genome copy number. The immunoblotting and qPCR experiments were performed 3 times independently, while virus titers were from one representative experiment using triplicate samples. Data are mean ± SD. ns, no significance, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001

Next, we infected the mice by corneal scarification with HSV-1 (5 × 10⁴ PFU per eye). Virus shedding in the tears was determined on day 3 and 5 post HSV-1 infection. Compared to mock- or agomiR-NC-treated group, mice in the agomiR-124 and the antiviral drug ACV-treated groups had reduced virus shedding in the tears (Fig. 4C and D). HSV-1 infection by corneal inoculation leads to neuroinvasion which can be measured by detection of viral DNA in the trigeminal ganglion (TG) [50]. Thus, the animals were sacrificed on day 7 after infection and HSV-1 genomic DNA in the TG was determined by qPCR. Compared with mock or agomiR-NC treated group, mice in agomiR-124 or ACV treated groups had significant reduction of viral DNA (Fig. 4E), indicating suppression of PTBP1 expression alleviated HSK symptoms and HSV-1 infection.

Discussion

PTBP1 is a member of the heterogeneous nuclear ribonucleoproteins that contribute to multiple aspects of RNA metabolism including alternative splicing, mRNA stabilization, and transcriptional and translational regulation. Several hnRNP proteins are involved in HSV-1 infection by regulating innate immune response or virus release [51–53]. In this study, we provide evidence demonstrating PTBP1 modulating HSV-1 infection through proper gene splicing of HSV-1 attachment receptors. Each of the HS3ST3A1 and B1 pre-mRNA contains an extraordinarily long intron that would require RNA looping for proper splicing. We showed that PTBP1 interacted with the intronic region of the HS3ST3A1 and B1 pre-mRNA, while genetic ablation of PTBP1 expression resulted in loss of HS3ST3A1 and B1 expression. In contrast to human cells, very few HSV-1 genes are known to be spliced, although the same pre-mRNA processing machinery is shared. We found PTBP1 ko did not markedly affect viral gene splicing, instead PTBP1 deletion prior to HSV-1 infection mainly affected host factors involved in HSV-1 cell attachment. Thus, loss of PTBP1 caused cell resistance to HSV-1 infection through reduced virus binding. Further studies are required to determine whether a looping mechanism is involved in PTBP1-mediated HS3ST3A1/B1 intron removal. Nonetheless, this study demonstrates that PTBP1 regulates HSV-1 infection through splicing regulation of attachment related genes.

Alternative splicing of pre-mRNA is a crucial aspect of gene regulation and has a pivotal role in tissue development and differentiation, and in key cellular pathways. It is carried out by spliceosome that can recognize splice signals and join alternative exons by simultaneously removing intronic sequences in pre-mRNA [2, 3]. Therefore, the relative abundance of alternatively spliced isoforms (alternative splicing outcomes) is critically influenced by the splice site strength, cis-regulatory sequences in the pre-mRNA, the expression levels of trans-acting factors including RNA-binding proteins and splicing factors [3]. In addition to the canonical splicing signals, different RNA-binding proteins (such as Nova, PTBP1, hnRNPA1, and SRSF1) can influence splicing through interactions with unique auxiliary sequences within the introns and exons [9, 54, 55]. As an RNA-binding protein, PTBP1 can either repress or activate splicing through loop formations, where loops in individual introns preferentially promote cassette exon splicing by accelerating asymmetric intron removal, while those spanning across cassette exon primarily repress splicing [7, 9]. Each of the HS3ST3A1/B1 genes contains a super long intron that shows abundant interactions with PTBP1, to potentially facilitate asymmetrical intron removal. Thus, PTBP1 mediates HSV-1 infection through regulation of HS3ST3A1/B1 expression.

The effect of PTBP1 on HSV-1 infection can be at multiple stages. As an RNA binding protein, PTBP1 has the ability to shuttle between the nuclear and cytosolic compartment, potentially for viral RNA transport [56, 57]. HSV-1 multiplies in fibroblast and epithelial cells for lytic infection, while it establishes viral latency in neuronal like cells. The PTB family consists of three members in mammalian genomes, with PTBP1 expressed in most cell types, PTBP2 (also known as nPTB) exclusively found in the nervous system, and PTBP3 (also known as ROD1) predominately detected in immune cells [58]. PTBP1 depletion leads to upregulation of PTBP2 through splicing regulation of the PTBP2 gene, causing fibroblast cell differentiation to neuronal-like cells [40, 59]. Thus, PTBP1 may function as a restriction factor modulating HSV-1 latency and lytic switch. We focused on a condition by knocking out PTBP1 expression prior to HSV-1 infection. Although we did not detect distinguishable alteration of HSV-1 gene splicing patterns in the parental and PTBP1-ko cells, this does not exclude the possibility that PTBP1, a critical co-factor of splicing regulation, may be involved in viral gene splicing as observed by Zheng and colleagues [60, 61]. In addition, viral proteins may also be involved in host gene splicing regulation. For example, HSV-1 ICP27 promotes cotranscriptional cellular pre-mRNA polyadenylation as well as aberrant pre-mRNA splicing of host genes, contributing to virus-induced host shutoff required for efficient viral growth [46, 62, 63]. Thus, PTBP1 can utilize mechanisms to control and regulate HSV-1 infection and host cellular responses.

HSV-1 is a common virus that causes a wide range of diseases, including herpes labialis, herpes simplex encephalitis, and herpetic stromal keratitis [64]. HSV remains as a common cause of corneal disease and is a leading infectious cause of corneal blindness among developed nations. The estimated global incidence of herpetic keratitis is over 1.5 million and new monocular visual impairment cases can be 40,000 each year [65–67]. HSV-1 can infect virtually all cell types in vitro. HSV-1 infects epithelial cells and infection begins with the attachment of virus particles to susceptible cells. The initial step of cell attachment or binding during HSV-1 entry is mediated by envelope gB and gC. The binding is followed by fusion between virus envelope and cell membrane during which HSV-1 gD interacts with a modified form of HS known as 3-O-sulfated heparan sulfate [27, 28]. The 3OST enzyme governs an important modification of 3-O-sulfation on HS chains. An important reason lies in its ability to exploit HS for attachment to cells. 3-O-sulfated HS is sufficient to mediate HSV-1 attachment and fusion in primary corneal fibroblasts, where nectin-1 and HVEM are not expressed [27, 68]. Currently, there is no cure or successful vaccine to prevent latent or recurrent HSV-1 infections, posing a significant clinical challenge to managing HSK and preventing vision loss and representing a significant global burden of disease [67, 69]. We found that perturbation of PTBP1 expression displayed therapeutic potential against ocular HSV-1 infection using a murine HSK model. Thus, repression of 3OST expression by microRNA mimics represents a novel strategy to remedy HSV-1 infection.

Heparan sulfate proteoglycans are abundant cell-surface molecules that consist of a protein core to which heparan sulfate glycosaminoglycan chains are attached. Virtually every cell type in metazoan organisms produces heparan sulfate. The sulfation pattern in HS determines its biological function [70, 71]. The IdoA or GlcA and glucosamine residues in HS carry sulfo groups at different OH positions, including 2-O-sulfation of IdoA or GlcA residue and 6-O-sulfation, N-sulfation, and 3-O-sulfation of the glucosamine residues. A 3-O-sulfate group on the glucosamine sugar unit is linked to anticoagulant activity, to growth factor receptor signaling, and to HSV-1 an infection [26, 72, 73]. Despite the fact that the 3-O-sulfate modification is increasingly being linked to much broader biological functions, including tissue formation, brain symmetry, and neuronal growth [70, 74, 75], the regulation of HS3STs remains incompletely characterized. The 3-O-sulfation step is catalyzed by seven isoenzymes, the expression of which is spatiotemporally regulated in a broad range of tissues [24, 75]. A number of HS3ST isozymes have been implicated in HSV-1 entry to cells for lytic infection [27]. HSV-1 can establish viral latency in neuronal cells that is subjected to recurrent reactivation. Clinical evidence links recurrent HSV-1 reactivation to neuronal inflammation and the development of Alzheimer’s disease. The brain tissues are known to have low HS3ST3A1/B1 expression, while Alzheimer's disease brain has increased 3-O-sulfated heparan sulfate [75]. Whether this aids HSV-1 infection and spreading remains uninvestigated.

Conclusions

In summary, we show here that the host RNA splicing regulator PTBP1 modulates HSV-1 infection through splicing regulation of genes such as HS3ST3A1 and HS3ST3B1 of HS chain modification. Mechanistically, PTBP1 binds to the intron regions of HS3ST3A1 and HS3ST3B1 pre-mRNA at distinct locations, potentially inducing RNA looping for efficient removal of super long intron. HSV-1, a common pathogen of human infection, utilizes 3-O-sulfated heparan sulfate for infection of epithelial cells. Loss of HS3ST3A1/B1 expression by PTBP1 gene disruption or suppression with agomir markedly reduced HSV-1 infection. Many pathogens exploit HSPGs for attachment and for cell entry. Targeting the HS biosynthesis pathway or HS mimetics has been proposed as an effective approach against virus infection [19, 76, 77]. The characterization of PTBP1 in HS biosynthesis and in virus infection provides a strategy for antiviral agent discovery.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ACV

Acyclovir

ActD

Actinomycin D

AS

Alternative splicing

gD

HSV-1 glycoprotein D

hnRNP I

Heterogeneous nuclear ribonucleoprotein I

HS

Heparan sulfate

HSPG

Heparan sulfate proteoglycan

HS3ST

Heparan sulfate-glucosaminyl 3-sulfotransferase

HSK

Herpetic stromal keratitis

HSV

Herpes simplex virus

ICP

Infected cell protein

PFU

Plaque-forming unit

PTBP1

Polypyrimidine tract-binding protein-1

RIP

RNA immunoprecipitation

TG

Trigeminal ganglion

Author contributions

Conceptualization, EL, PC and WS; data curation and formal analysis, EL, WS, MZ, RW, JY, RL, and ATR; investigation, validation and methodology, WS, RL, MZ, RW; resources, JY, and XL; funding acquisition, PC and EL; supervision, EL; writing, WS and EL prepared original draft; editing, EL, WS, and ATR. All authors have reviewed and approved the current version for submission.

Funding

This work was supported by grants from the National Key R&D Program of China (2023YFC2308200 to PC) and from NSFC (81871636 to EL). ATR is on a scholarship provided from Chinese Scholarship Council.

Data availability

The RNA-seq data generated in this study were submitted to the GEO repository with an accession number of GSE297152 and will be publicly accessible upon acceptance of the paper. The basically data used to generate the figures and supporting analyses are included within the article and its supplementary files.

Declarations

Ethics approval and consent to participate

A protocol for the use and care of laboratory animals was approved by the IACUC of Nanjing University (D2103051). All investigations conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research (2021 edition).

Consent for publication

Not applicable.

Competing interests

The authors declare no conflict of interest.

Others to disclose

Authors disclose that no generative AI or AI-assisted technologies was used in the preparation of the manuscript.