Cellular and viral factors regulating Merkel cell polyomavirus replication

Abstract

Merkel cell polyomavirus (MCV), a previously unrecognized component of the human viral skin flora, was discovered as a mutated and clonally-integrated virus inserted into Merkel cell carcinoma (MCC) genomes. We reconstructed a replicating MCV clone (MCV-HF), and then mutated viral sites required for replication or interaction with cellular proteins to examine replication efficiency and viral gene expression. Three days after MCV-HF transfection into 293 cells, although replication is not robust, encapsidated viral DNA and protein can be readily isolated by density gradient centrifugation and typical ∼40 nm diameter polyomavirus virions are identified by electron microscopy. The virus has an orderly gene expression cascade during replication in which large T (LT) and 57kT proteins are first expressed by day 2, followed by expression of small T (sT) and VP1 proteins. VP1 and sT proteins are not detected, and spliced 57kT is markedly diminished, in the replication-defective virus suggesting that early gene splicing and late gene transcription may be dependent on viral DNA replication. MCV replication and encapsidation is increased by overexpression of MCV sT, consistent with sT being a limiting factor during virus replication. Mutation of the MCV LT vacuolar sorting protein hVam6p (Vps39) binding site also enhances MCV replication while exogenous hVam6p overexpression reduces MCV virion production by >90%. Although MCV-HF generates encapsidated wild-type MCV virions, we did not find conditions for persistent transmission to recipient cell lines suggesting that MCV has a highly restricted tropism. These studies identify and highlight the role of polyomavirus DNA replication in viral gene expression and show that viral sT and cellular hVam6p are important factors regulating MCV replication. MCV-HF is a molecular clone that can be readily manipulated to investigate factors affecting MCV replication.

Figure 1. MCV genome.

(A) Full-length of MCV genomes identified from 5 MCC tumors (MCV350, MCV339, MCV344, MCV349, MCV352), 1 MCC cell line (MCVMKL-1) and 1 PBMC sample (MCV85). T antigen ORFs are shown in blue arrows, VP ORFs in pink arrows. Numbers stand for positions in MCV genome. Black solid boxes indicate genomic deletions in MCV genome. (B) Phylogenetic tree of MCV genomes. The consensus genome (MCV-HF, JF813003) is located in the center of a tree including MCV350 (EU375803), MCV339 (EU375804), MCV344 (JF812999), MCV349 (JF813000), MCV352 (JF813001), MCVMKL-1 (FJ173815), MCV85 (JF813002) and other MCV sequences obtained from human skin and Kaposi's sarcoma . (C) The consensus MCV-HF genome can be linearized at BsrFI site (4,596 nt) and cloned for propagation in E. coli. Sites for mutations engineered into two MCV-HF genomes (MCV-Rep⁻ (C/A) and MCV-hVam6p⁻ (TG/GC)) are shown.

Figure 2. Coordinated viral gene expression during MCV replication in 293 cells.

One microgram of recircularized MCV genomes (wild-type MCV-HF or replication-defective MCV-Rep⁻) was transfected into 293 cells. Immunoblotting was performed to examine T antigen expression over 5 days (indicated by hollow arrows) and VP1 protein (indicated by solid arrow) using CM2B4 (LT, 57kT), CM8E6 (sT) and CM9B2 (VP1) antibodies, respectively. Alpha-tubulin detection was used as a protein loading control. LT protein is expressed equally at day 2 for both viruses but decreases for MCV-Rep⁻ on days 3–5. VP1 increases on days 3–5 only for MCV-HF, corresponding to viral DNA replication (Fig. 3D). Other early proteins are also diminished (57kT) or absent (sT) in the replication deficient MCV-Rep⁻.

Figure 3. Fractionation of viral capsid protein VP1 by Optiprep™ density gradient ultracentrifugation.

(A) Twelve fractions were collected from highest to lowest density, and analyzed by immunoblotting with CM9B2 antibody to detect VP1 capsid protein. Assembled 45 kDa VP1 protein is isolated in fraction 4. Unassembled, free VP1 protein is present in Fractions 9 and 10. The positive control (+) is virus-like particle (VLP) prepared from 293TT cells by MCV VP1 and VP2 transfection. (B) Typical 40 nm diameter icosahedral Merkel cell polyomavirus particles present in fraction 4 (upper panel) and VP1/VP2-containing MCV virus-like particles for comparison (bottom panel). (C) Nuclease-resistant MCV DNA in various gradient fractions quantitated by real time PCR. Highest levels of encapsidated DNA are present in fraction 4, corresponding to the fraction having MCV virions. (D) Time-course for MCV virion production after transfection of 1 µg replication competent (MCV-HF) or incompetent (MCV-Rep⁻) genomes into 293 cells was determined on lysed cells by quantitative PCR after nuclease treatment. Genome replication and packaging of MCV-HF is evident by day 3.

Figure 4. MCV genome replication.

Southern blot (right panel) for MCV for MCV-HF (lane 1), MCV-Rep⁻ (lane 2) and MCV-hVam6p⁻ (lane 3) viruses four days after transfection of 1 µg circular genomic DNA into 293 cells. Panel on left shows the ethidium bromide-stained gel prior to transfer indicating equal DNA loading. Bands for the full-length 5.4 kb MCV genome are present as DpnI-resistant bands in MCV-HF and MCV-hVam6p⁻ viruses (lanes 1 and 3) but not in the replication deficient MCV-Rep⁻ virus (lane 2). The replication efficiency was measured by the ratio between the DpnI-resistant 5.4 kb band and the DpnI-sensitive band. The MCV-hVam6p⁻ virus generates ∼2-fold more full length genome compared to wild-type MCV-HF virus. Replicated viral DNAs also show the presence of extensive subgenomic fragments.

Figure 5. Quantitative PCR for MCV virion production for MCV-HF, MCV-Rep⁻ and MCV-hVam6p⁻ viruses.

One microgram MCV clone DNAs were transfected into 293 cells stably transduced to express MCV sT or LT proteins (not shown). DNA was extracted and treated with benzonase and RNase to discriminate packaged viral DNA. The nuclease-resistant MCV genome was precipitated and measured by quantitative PCR after proteinase K treatment. Cellular sT expression increases virion production for both MCV-HF and MCV-hVam6p⁻ viruses. Comparison of MCV-HF and MCV-hVam6p⁻ shows that loss of the hVam6p binding site also increases virus production. Coexpression of sT and mutation of the hVam6p binding site in the MCV genome are additive in virion production compared to MCV-HF without sT coexpression.

Figure 6. Comparison of viral packaging for MCV-HF, MCV-Rep⁻ and MCV-hVam6p⁻ viruses.

Optiprep™ density gradient fractions from wild type (MCV-HF) and mutant viruses (MCV-Rep⁻, MCV-hVam6p⁻) generated from transfected 293 cells were used for Western blotting. Dilutions of MCV virus-like particles (VLP) provide a marker for the relative abundance of VP1 protein in each fraction. Assembled MCV-hVam6p⁻ virus VP1 expression is ∼10 fold increased in fraction 4 compared to MCV-HF.

Figure 7. Effect of hVam6p coexpression on MCV virion production.

293 cells cotransfected with hVam6p and MCV genomes at day 4 as measured by nuclease-resistant DNA by quantitative PCR. Circularized viral plasmids (1 µg) together with varying amounts of hVam6p expression plasmid were simultaneously transfected during this experiment.

Figure 8. Effects of hVam6p on in vitro MCV origin replication.

293 cells were transfected with plasmids containing MCV origin and equal amounts of either wild type of genomic T antigen (TAg), LT cDNA, or the corresponding constructs containing the hVam6p-binding site mutations (TAg.W209A and LT.W209A). Origin replication was assessed through Southern blotting by comparing the ratio of DpnI-resistant (replicated) to DpnI-sensitive (unreplicated) DNA. For each condition, replication in the absence or presence of simultaneously cotransfected hVam6p expression plasmid was determined. Expression of the genomic TAg containing both sT and LT showed increased replication of the MCV origin compared to the LT cDNA regardless of hVam6p coexpression. Neither mutation of the hVam6p binding site on LT nor coexpression of exogenous hVam6p significantly altered MCV origin replication.