Varicella-zoster virus VLT-ORF63 fusion transcript induces broad viral gene expression during reactivation from neuronal latency

Abstract

Varicella-zoster virus (VZV) establishes lifelong neuronal latency in most humans world-wide, reactivating in one-third to cause herpes zoster and occasionally chronic pain. How VZV establishes, maintains and reactivates from latency is largely unknown. VZV transcription during latency is restricted to the latency-associated transcript (VLT) and RNA 63 (encoding ORF63) in naturally VZV-infected human trigeminal ganglia (TG). While significantly more abundant, VLT levels positively correlated with RNA 63 suggesting co-regulated transcription during latency. Here, we identify VLT-ORF63 fusion transcripts and confirm VLT-ORF63, but not RNA 63, expression in human TG neurons. During in vitro latency, VLT is transcribed, whereas VLT-ORF63 expression is induced by reactivation stimuli. One isoform of VLT-ORF63, encoding a fusion protein combining VLT and ORF63 proteins, induces broad viral gene transcription. Collectively, our findings show that VZV expresses a unique set of VLT-ORF63 transcripts, potentially involved in the transition from latency to lytic VZV infection.

Fig. 1. Transcription profile across the VLT and ORF63 loci during lytic VZV infection of human epithelial cells and sensory neurons.

a Coverage plots denoting dRNA-Seq (ARPE-19 epithelial cells, dark blue) and cDNA-Seq (human iPSC-derived sensory neurons [HSN], teal) data aligned to the top strand of the VZV genome. dRNA-Seq data is representative of two independently sequenced biological replicates with RNA extracted from lytically VZV-infected ARPE-19 cells at 4 days post infection (4 dpi). cDNA-Seq data was generated from a pool of biological replicates collected from lytically VZV-infected HSN at 14 dpi. Schematics of the major transcripts from VLT and ORF63 loci are shown in following colors: lytic VLT isoform (_lytVLT, red), canonical RNA 63-1 (63, green), lytic VLT-ORF63 isoforms (_lytVLT63-1/2/3, purple) and latent VLT isoform (VLT, orange). Additional canonical VZV transcription units present are shown in grey, with ORF numbers indicated. Wide boxes indicate canonical CDS domains and while thin boxes indicate UTRs, respectively. Y-axis values indicate the maximum read depth of that track. b Analysis of VLT, RNA 63 and VLT-ORF63 isoform expression by RT-qPCR analysis using the same two independent experiments in ARPE-19 cells (dark blue) and HSN (teal) for long-read sequencing. The primer locations used for RT-qPCR analysis detecting transcripts from VLT to ORF63 loci are depicted in Supplementary Fig. 1. Source data are provided as a Source Data file. ORF; open reading frame, CDS; coding sequence, UTR; untranslated region.

Fig. 2. Transcription profile across the VLT and ORF63 loci in latently VZV-infected human trigeminal ganglia.

a Detection of VLT, canonical RNA 63-1 and VLT63 isoforms by RT-qPCR analysis in latently VZV-infected human trigeminal ganglia (TG) (n = 4; post-mortem interval 4–4.5 h). Data on individual TG samples are shown as unique symbols. Source data are provided as a Source Data file. b Detection of VZV RNA 63 and VLT RNA by multiplex fluorescent in situ hybridization (ISH) on human TG (upper two panels) and human herpes zoster skin biopsy (bottom panel). Representative images are shown for n = 7 human TG and n = 2 zoster skin biopsies. Asterisks indicate autofluorescent lipofuscin granules in neurons. Scale bar: 10 µm. Arrowheads indicate RNA 63 (red) and/or VLT (green) ISH signal. Right panels: enlargements of area indicated by white box. c Putative transcription start sites (TSS) of VLT (row 1), VLT63 (row 2) in human TGs and VLT in latently VZV-infected human iPSC-derived sensory neurons (HSN) (row 3), as determined by 5′-RACE analysis. Schematic top shows major latent VLT and VLT-ORF63 transcript isoforms and location of primers used for 5′-RACE analysis. Bottom: VLT sequence of VLT exon 1, as previously determined by RNA-seq on human TGs. Flanking regions are shown with arrows indicating putative TSS by 5′-RACE analysis.

Fig. 3. Protein coding potential of VLT-ORF63 fusion transcripts.

a Schematic presentation of VLT and VLT-ORF63 isoform transcripts with predicted encoded proteins. Red boxes indicate exons of lytic isoforms of VLT and VLT-ORF63 transcripts, orange boxes indicate exons of latent isoforms of VLT and VLT-ORF63 transcripts, and a black box indicates a part of RNA 63-1 5′-UTR in canonical RNA 63-1 transcript. The end of VLT-ORF63 transcripts indicates stop codon for ORF63 CDS. Black horizontal lines indicate location of encoded open reading frames (ORFs). The blue square indicates the first start codon (ATG), the black triangle indicates first stop codon within the same frame as the start codon, the green diamond indicates start codon of canonical ORF63, and grey circles indicate downstream ATG codons within the same frame as the start codon. Pointed rectangles show translated protein from corresponding ORFs with a black box indicating the 24 amino acid linker peptides translated from a part of RNA 63-1 5′-UTR in the canonical RNA 63-1 transcript. UTR; untranslated region, CDS; coding sequence. Confocal microscopic images of ARPE-19 cells b transfected with CS-CA-VZV plasmids (48 h post transfection), c lytically infected with VZV (4 dpi), and d herpes zoster skin lesions. b, c Cells were stained with anti-pORF63 mAb (green) and anti-pORF63 pAb (red) for CS-CA-ORF63, and anti-pVLT-ORF63 pAb (green), anti-pVLT pAb (red) and anti-pORF63 mAb (blue) for CS-CA-VLT-ORF63-1 and CS-CA-VLT-ORF63-2 and lytic VZV infection. Nuclei were stained with DAPI (cyan) and images are representative of results from two independent experiments. Magnification; x600 and x3 digital zoom with 5 µm scale bars (except 3rd row in b, and 3rd and 4th rows in c) and x600 and x2 digital zoom with 10 µm scale bars (3rd row in b, and 1st and 2nd rows in c). mAb; monoclonal antibody, pAb; polyclonal antibody. d Nuclei were stained with DAPI (blue) and images are representative for two independent stainings performed on one control and two herpes zoster skin biopsies. Magnification: x200 with 50 µm scale bars (1st and 2nd rows) and x200 and x3 digital zoom with 20 µm scale bars (3rd row).

Fig. 4. Effect of anisomycin treatment on VZV transcription in latently VZV-infected sensory neurons in vitro.

a RT-qPCR analysis for transcription across VLT and ORF63 loci in latently VZV-infected human iPSC-derived sensory neurons (HSN) cultures (n = 3). Data on individual HSN culture experiments are shown as unique symbols. b, c Latently VZV-infected HSN cultures were depleted of neurotrophic factors (NGF [nerve growth factor], GDNF [glial-derived neurotrophic factor], BDNF [brain derived neurotrophic factor] and NT-3 [neurotrophin-3]) and treated with anti-NGF antibody (Ab) for 14 days. In total, n = 40 independent cultures were subjected to reactivation stimuli. b Representative examples of a HSN cultures showing complete reactivation (2/40) and early/abortive reactivation (38/40) by infectious focus forming assay after transferring HSN onto ARPE-19 cells. c Representative examples of viral gene expression in cultures showing complete reactivation (n = 1 shown; white circle) and early/abortive reactivation (n = 5 shown; colored triangles). d, e HSN cultures were treated with d anisomycin (n = 6), or e DMSO as solvent control (n = 3) at both the somal and axonal compartment for 1 h, washed twice and cultured for 7 days before RT-qPCR analysis. Data on individual HSN culture experiments are shown as unique symbols and/or colors. Only VLT exon 1–2 is shown as representative of VLT. Source data are provided as a Source Data file (a–e).

Fig. 5. Effect of ectopic VLT-ORF63 expression on VZV gene expression in latently VZV-infected sensory neurons in vitro.

At 14 days after establishment of VZV latency in human iPSC-derived sensory neurons (HSN), following VZV genes were transduced by replication incompetent lentivirus vectors: empty vector (white), ORF63 (grey), VLT63-1 (purple) and VLT63-2 (cyan). Transduced HSN were cultured for 14 days (n = 4 replicates/vector) and subjected to RT-qPCR analysis. Technical duplicates were utilized per sample and all the biologically independent data is shown as unique symbols for a endogenous or transduced VZV genes, b IE genes, c E genes, and d L genes. Source data are provided as a Source Data file.