Molluscum Contagiosum Virus Transcriptome in Abortively Infected Cultured Cells and a Human Skin Lesion

ABSTRACT

Molluscum contagiosum virus (MOCV), the only circulating human-specific poxvirus, has a worldwide distribution and causes benign skin lesions that may persist for months in young children and severe infections in immunosuppressed adults. Studies of MOCV are restricted by the lack of an efficient animal model or a cell culture replication system. We used next-generation sequencing to analyze and compare polyadenylated RNAs from abortive MOCV infections of several cell lines and a human skin lesion. Viral RNAs were detected for 14 days after MOCV infection of cultured cells; however, there was little change in the RNA species during this time and a similar pattern occurred in the presence of an inhibitor of protein synthesis, indicating a block preventing postreplicative gene expression. Moreover, a considerable number of MOCV RNAs mapped to homologs of orthopoxvirus early genes, but few did so to homologs of intermediate or late genes. The RNAs made during in vitro infections represent a subset of RNAs detected in human skin lesions which mapped to homologs of numerous postreplicative as well as early orthopoxvirus genes. Transfection experiments using fluorescent protein and luciferase reporters demonstrated that vaccinia virus recognized MOCV intermediate and late promoters, indicating similar gene regulation. The specific recognition of the intermediate promoter in MOCV-infected cells provided evidence for the synthesis of intermediate transcription factors, which are products of early genes, but not for late transcription factors. Transcriptome sequencing (RNA-seq) and reporter gene assays may be useful for testing engineered cell lines and conditions that ultimately could provide an in vitro replication system.

IMPORTANCE The inability to propagate molluscum contagiosum virus, which causes benign skin lesions in young children and more extensive infections in immunosuppressed adults, has constrained our understanding of the biology of this human-specific virus. In the present study, we characterized the RNAs synthesized in abortively infected cultured cells and a human skin lesion by next-generation sequencing. These studies provided an initial transcription map of the MOCV genome, suggested temporal regulation of gene expression, and indicated that the in vitro replication block occurs prior to intermediate and late gene expression. RNA-seq and reporter assays, as described here, may help to further evaluate MOCV gene expression and define conditions that could enable MOCV replication in vitro.

INTRODUCTION

Molluscum contagiosum virus (MOCV) is the sole member of the Molluscipoxvirus genus of the Chordopoxvirinae subfamily of the Poxviridae (1). Although many poxviruses cause zoonoses, variola virus (the causative agent of smallpox) and MOCV are the only known human-specific poxviruses (2, 3). MOCV has a worldwide distribution and commonly infects young healthy children, where it causes papular skin lesions that may persist for many months before spontaneous resolution (4). However, widespread disfiguring skin lesions may occur in individuals with immunodeficiencies. For the latter, the most successful therapy is treatment of the underlying immunodeficiency. Although several MOCV variants have been recognized by restriction endonuclease analysis and limited DNA sequencing, they produce indistinguishable lesions (4).

Knowledge of MOCV is limited because of the lack of either a cell culture system or useful animal model. The inoculation of primate cells with MOCV produces an abortive infection with cell rounding and related cytopathic effects (CPE) (5). However, the cells regain a more normal appearance after 48 h (6, 7). Evidence that MOCV gene expression is necessary for CPE was supported by the ability of inhibitors of RNA and protein synthesis to prevent this phenomenon (7, 8). Following infection, electron microscopy revealed MOCV cores within the cytoplasm, consistent with early gene expression, but the disassembly of the cores or assembly of new virus particles was not observed (7). More direct evidence for early gene expression in human fibroblasts was obtained by RNA-DNA hybridization and reverse transcription-PCR (RT-PCR) (7, 8). Although MOCV DNA replication was not detected, the transcription of postreplicative (PR) genes was reported (9). In contrast to the inability of MOCV to replicate in cell culture systems, some success has been obtained by infecting human foreskin tissue grafted into immune deficient mice (10, 11).

Although productive infection in vitro has not been demonstrated, genome sequencing provides a wealth of information (12, 13). Of 163 initially annotated open reading frames (ORFs) in the 190-kbp MOCV genome, 105 have homologs in well-studied orthopoxviruses. Moreover, the physical arrangement of the conserved genes within the MOCV and orthopoxvirus genomes is almost identical. The conservation of the DNA concatemer resolution sequence, the T5NT early transcription termination sequence, and A+T-rich putative promoter sequence in MOCV imply that the basic features of MOCV replication and gene expression are likely to be similar to those of other chordopoxviruses. A subset of novel MOCV genes is predicted to have roles in host interactions (14). The sequence information has enabled the functional analysis of some such genes by expressing them in uninfected cells or via recombinant vaccinia virus (VACV) (15, 16) (17–22).

A greater understanding of the biology of MOCV and the basis for its inability to replicate in cell culture could come from a transcriptional analysis. Next-generation deep RNA sequencing recently has delineated the VACV transcriptome (23), the prototype member of the Poxviridae. In the present study, we applied transcriptome sequencing (RNA-seq) to analyze the MOCV transcriptome during abortive infections and from a human skin lesion. Viral RNAs were detected for 14 days after MOCV infection of cultured cells; however, there was little change in the RNA species during this time, and they represented only a subset of RNAs isolated from a human skin lesion. Several lines of evidence indicated that the RNAs synthesized in cultured cells belong to the early class. First, there was a high correlation between RNAs made in the absence and presence of a protein synthesis inhibitor. Second, although there are MOCV homologs of all classes of VACV genes, most highly expressed MOCV RNAs made during in vitro infections corresponded to the early homologs. Third, the consensus promoter sequence for most expressed MOCV genes was similar to that of VACV early promoters. Finally, the synthesis of intermediate transcription factor proteins, which are products of early genes, in MOCV-infected cells was supported by transfection studies with reporter genes regulated by an MOCV intermediate promoter. Further RNA-seq and reporter gene analyses may help to evaluate and optimize conditions that enhance MOCV infection.

MATERIALS AND METHODS

Virus preparation.

Molluscum contagiosum was isolated from two patients in their early twenties with immunodeficiency diseases. MOCV was classified as subtype 1 based on RNA sequence matches to positions 148 to 290 of the reference genome sequence (accession number M74033.1) of the putative XL2 gene (24) and overall correspondence to the complete subtype 1 genome sequence (12). The patients had widely disseminated molluscum contagiosum over large areas of their bodies. Several lesions, which had been present for many months, were removed by curettage (scraping and excising lesions with a sharp curette) and stored at −80°C. For virus stock preparation, tissue plugs from the same individual were pooled and macerated with a Dounce homogenizer in 1 ml of 1 mM Tris-HCl, pH 9.0. The suspension was frozen and thawed three times, followed by three cycles of sonication. The virus preparation then was cleared by slow-speed centrifugation (300 × g for 10 min) to remove large debris, and the cleared supernatant was divided into aliquots and stored at −80°C. The virus concentration was estimated using a virus counter (ViroCyt 2100; Denver, CO) (25). For RNA extraction, tissue plugs from molluscum skin lesions were immersed in RNAlater (AM7021; Life Technologies), kept at 4°C, and processed within 72 h.

Electron microscopy.

Carbon-coated, Rh-flashed copper mesh grids (400 mesh; Ted Pella) with nitrocellulose supporting film were placed on 15-μl droplets of solution for 5 min before negative staining in 1% uranyl acetate. Virus particles were observed and photographed at 120 kV in a Tecnai 12.

Cell culture and standard infection protocol.

MRC-5 cells and Huh7.5.1 cells in essential modified Eagle medium (EMEM; Quality Biologicals) containing 10% fetal bovine serum, as well as Vero cells in the same medium containing 8% serum, were seeded into 6-well plates. Upon reaching near confluence (∼10⁶ cells), the medium was removed and the cells were overlaid with 1 ml of EMEM plus 2% serum containing 10 μl of virus stock estimated to give a multiplicity of 5 to 10 particles per cell. After adsorption for 2 h at 37°C and 5% CO₂, the medium was replaced with EMEM containing 10% serum and the incubation continued for up to 14 days.

RNA isolation and library preparation.

TRIzol reagent (15596-018; Ambion) was used for RNA purification by following the manufacturer's protocol. RNA was extracted with phenol-chloroform-isoamyl alcohol (1:1:25), ethanol precipitated, and suspended in water. Polyadenylated RNA was captured on solid-phase magnetic beads (61006; Life Technologies), and RNA libraries were prepared using a ScriptSeq v2 kit (SSV21106; Epicentre). For RNA extracted from cells infected at 31°C, an rRNA removal step was included by following the manufacturer's protocol (MRZH116; Epicentre) before poly(A) selection.

RNA sequencing.

An Illumina GA IIx system was used to sequence the libraries made from MRC-5 cells infected at 37°C and a tissue lesion, and an Illumina HiSeq system was used to sequence the libraries made from MRC-5, Vero, and Huh7.5.1 cells infected at 31°C. Reads were trimmed of adaptor sequences and poor-quality sequences, and the remaining poor-quality reads were filtered out. The trimmed reads were mapped to MOCV's genome using the Bowtie 2 alignment program. Default settings were used, except for allowing two alignments per read for the terminal repeats. Gene counts were determined by the number of MOCV reads mapped to an ORF in the corresponding strand. Read counts were normalized to total read counts and ORF length across all samples.

Transcriptome analysis.

MOCV tilings (WIG format files) of viral reads were imported into Mochiview (26) to generate whole-genome transcriptome maps as previously described (23). The read counts were normalized and organized by time point, and each ORF was associated with its predicted expression class. The results were converted into annotation files using an in-house script and imported into Geneious 6.0 software. R software was used for data analysis, including statistics and the generation of bar plots, histograms, and correlation coefficients. An arbitrary threshold set on the median for each time point and the presence of the early promoter consensus were used to generate the MOCV expression map.

Plasmids and transfection assays.

DNA constructs containing a green fluorescent protein (GFP) or luciferase reporter ORF under the control of MOCV promoters MC044, MC095, and MC069 were engineered by overlapping PCR. The promoter sequences were GCTCTTTTTGTATTTAAAGCTAGCGAGGAGAAAGTAAATAGCG, CCTTTGGTGCAGATCTCGCGAATAATAA, and TTCCTCGTGTGGGTCTTAATTGTGGCACTTTAA, respectively. The constructs were inserted into a pCR-bluntII plasmid backbone (K2800-20; Invitrogen). Expression was determined in MRC-5, HUH7.5.1, Vero, BHK-21, and BS-C-1 cells infected with VACV or ectromelia virus. At 2 h after infection, the reporter plasmids were transfected using Opti-MEM medium (31985-062; GIBCO) with Lipofectamine 2000 (11668-500; Invitrogen) according to the manufacturer's protocol. Luciferase expression was measured using a luminometer (Berthold Sirius luminometer), while GFP expression was evaluated by fluorescence microscopy.

Ethics.

The Institutional Review Board of the National Institute of Allergy and Infectious Diseases, NIH, approved this research, and the patients provided written informed consent.

Microarray accession number.

The cDNA sequences were deposited at the National Center for Biotechnology Information Sequence Read Archive (SRA) under SRA accession number SRP069070.

RESULTS

Characterization of MOCV stocks.

Because MOCV has not been grown in cell culture, virus stocks were prepared from human skin lesions (27). The presence of typical poxvirus particles in the supernatant following the clarification of tissue homogenates by low-speed centrifugation was confirmed by electron microscopy (Fig. 1A), and the numbers of virions were quantified with a Virocyt 2100, a dedicated flow cytometer that scores virus particles (vp) on the basis of coincidence of detected DNA and protein. MOCV stocks typically contained ∼1 × 10⁹ vp/ml.

FIG 1.

Electron microscopy and CPE of MOCV stocks. (A) Electron microscopy. An MOCV preparation was deposited on a grid and negatively stained with 1% uranyl acetate. A cluster of virus particles is shown on the left and a single virion at higher magnification on the right. Magnifications are indicated at the bottom. (B) Multiplicity-dependent CPE. Nearly confluent MRC-5 cells were infected with MOCV at a multiplicity of 0 to 40 virus particles (vp) per cell at 37°C. Microscopic images were made at 8 h after infection. (C) Protein synthesis requirement for cytopathic effect. Nearly confluent MRC-5 cells were untreated or treated with 300 μg/ml of CHX for 30 min and then infected with MOCV at a multiplicity of approximately 10 vp/cell as described for panel B. Microscopic images of cells incubated for 4 h at 37°C or 15 h at 31°C are shown.

Our initial experiments were performed with MRC-5 cells, which are diploid fibroblasts derived from human fetal lung (28) that have been used to make vaccines, including one for smallpox. In order to make a preliminary assessment of the biological activity of the MOCV preparations and the correspondence between the numbers of particles and infectious units, MRC-5 cells were incubated with serial virus dilutions at 37°C and examined 8 h after infection by light microscopy. Between multiplicities of 4 and 40 PFU per cell, the entire monolayer exhibited CPE (Fig. 1B). At a multiplicity of 10 vp per cell, CPE occurred as early as 4 h at 37°C, and virtually the entire monolayer was affected by 15 h, even at 31°C (Fig. 1C). Both temperatures were used in later experiments. The CPE was prevented by pretreatment of the cells with cycloheximide (CHX) (Fig. 1C) or by UV irradiation of the inoculum (not shown), confirming the importance of MOCV gene expression for this phenomenon. As previously observed, the CPE was transient and the monolayers recovered within 72 h (not shown).

Analysis of MOCV RNA from MRC-5 cells and a human skin lesion.

Total polyadenylated RNA was isolated from MOCV-infected MRC-5 cells incubated at 37°C for 0, 4, 8, 12, and 120 h in the first experiment. An additional sample was obtained from cells that had been infected for 12 h in the presence of CHX in order to limit expression to RNA made in the absence of de novo protein synthesis. In a second experiment, RNA was isolated at 12 and 120 h to overlap the previous analysis and also at 7, 9, and 14 days. For comparison, polyadenylated RNA also was isolated from a human MOCV lesion. RNA-seq libraries were prepared and analyzed on an Illumina sequencer. The proportion of total reads that mapped to the MOCV genome increased from about 0.1% at 4 h to about 1% at 120 h and maintained a similar percentage for an additional 5 days (see Table S1 in the supplemental material).

The normalized read counts corresponding to each annotated MOCV ORF (GenBank accession number U60315.1) at each time point are presented in Table S2 in the supplemental material. The numbers of reads associated with each ORF varied by over 3 logs, from undetectable to >25,000, with 67 ORFs having read counts greater than 100 at two or more time points (see Table S2). The number of reads and their alignment with the MOCV genome are shown in Fig. 2. Counts above and below the line represent transcripts corresponding to sequences on the plus and minus strands directed to the right and left, respectively. The majority of transcripts mapped near the ends of the genome, with those on the left predominantly on the minus strand and those on the right predominantly on the plus strand, a pattern similar to that occurring at early times after VACV infection (23). Except for peak heights, the patterns of MOCV RNAs at all times looked remarkably similar to each other as well as to the pattern obtained from MOCV-infected MRC-5 cells at 12 h with CHX (Fig. 2). This impression was verified by calculating coefficients of determination (r²). For the first experiment the r² values for the 4-h sample compared to the 8-, 12-, and 120-h samples were 0.93, 0.88, and 0.75, respectively (Fig. 3A). The declining r² values with time indicate changes in the amounts of individual RNAs that could be due to differential RNA stability or synthesis. The r² values for each time point versus the 12-h CHX time varied from 0.91 to 0.7 (Fig. 3A). In the second experiment, the r² values of the 120-h sample versus the 7-, 9-, and 14-day samples were 0.97, 0.96, and 0.96, respectively, indicating little or no change in RNA populations during this time period (Fig. 3B).

FIG 2.

Alignment of RNA read counts along the MOCV genome. (Left) Nearly confluent MRC-5 cells were infected with approximately 10 vp/cell of MOCV at 37°C and harvested at 4 to 120 h postinfection (hpi). An additional monolayer was treated with CHX, infected with MOCV, and harvested at 12 h. Polyadenylated RNA was isolated and sequencing was performed as described in Materials and Methods. Read counts per nucleotide (vertical axis) were plotted along the plus (upper) and minus (lower) strands of the MOCV genome (horizontal axis). (Right) In an independent experiment, MRC-5 cells were infected as described for cells examined on the left, except that they were harvested at 12 h (not shown), 120 h, 7 days, 9 days, and 14 days postinfection (dpi). In addition, RNA was isolated from a human skin lesion and sequenced as described above.

FIG 3.

Correlation between the numbers of reads that map to individual ORFs from different samples. (A) Comparison of reads from 4 to 120 h. The times at which RNAs were obtained in the absence of CHX or at 12 h in the presence of CHX are indicated. Each dot represents a comparison of read counts that mapped to an individual ORF. The coefficient of determination (r²) is indicated with the number of asterisks being proportional to significance. (B) Comparison of reads from 120 h to 14 days and skin lesion. The format was the same as that described for panel A.

In contrast to the data from in vitro infections, the read counts of the RNAs obtained from the skin lesion were more broadly distributed along the genome, suggesting that they represent transcripts from all promoter classes (Fig. 2). The r² values calculated by comparing the in vitro and in vivo RNAs were only about 0.1 (Fig. 3B), confirming the large difference in the regulation of gene expression in vitro and in vivo. The positive correlation between the MOCV RNAs made in the absence and presence of CHX suggested that the in vitro subset consisted mainly of transcripts from early genes.

Transcription of MOCV homologs of VACV genes.

Of the 163 annotated MOCV ORFs, 103 have VACV homologs (13). The experimentally determined promoter classifications of the VACV homologs (29) of the corresponding MOCV genes are shown in Table S2 in the supplemental material. Histograms of the MOCV read counts corresponding to the early, intermediate, and late VACV homologs are shown in Fig. 4. We first considered the RNAs corresponding to genes expressed by cells infected with MOCV in the presence of CHX, which are early by definition. In that case, the median value of the read counts that were associated with the MOCV homologs of VACV early genes was more than a log higher than median values of homologs of intermediate and late genes (Fig. 4). This result suggested similarity in gene regulation between MOCV and VACV. For in vitro infections in the absence of CHX, early homologs also had higher read counts than the intermediate and late homologs (Fig. 4), consistent with the relatively constant pattern of total MOCV RNAs over time (Fig. 2). In contrast, many intermediate and late mRNA homologs from skin lesions had similar and even higher abundance than the early homologs (Fig. 4).

FIG 4.

Histograms of MOCV read counts corresponding to the early, intermediate, and late VACV homologs. Data from Table S2 in the supplemental material were plotted for the samples indicated. The horizontal axis shows the number of reads on a log scale. The vertical axis (density) represents the fraction of ORFs.

The close packing of ORFs and the uneven read counts across an ORF due to differences in efficiency of PCR make it difficult to distinguish between readthrough due to incomplete transcription termination from an adjacent gene and that due to differences in the regulation of individual genes in VACV and MOCV. We suspect that the relatively low read counts associated with homologs of intermediate and late mRNAs in MRC-5 cells (Fig. 4) are due mainly to readthrough. However, a few of the VACV homologs transcribed abundantly in MRC-5 cells corresponded to intermediate and late genes (see Table S2 in the supplemental material), which could indicate differences in the regulation of those genes in VACV and MOCV.

Senkevich and coworkers had noted that the putative promoter sequences associated with the MOCV ORFs closely resembled their VACV homologs, suggesting similar regulation (13). A promoter consensus sequence (Fig. 5A), using the 100-bp regions upstream of 75 MOCV ORFs that were expressed in MRC-5 cells, closely resembled the VACV early promoter consensus (30, 31). By identifying ORFs with counts above the median from MRC-5 cells and the presence of the early consensus motif, a map of highly expressed early genes was generated (Fig. 5B).

FIG 5.

Early promoter consensus sequence and expression map. (A) The 100-bp sequences upstream of 75 ORFs expressed in MRC-5 cells were analyzed using the MEME program. The size of the letter is proportional to the frequency of occurrence of that nucleotide. (B) MOCV in vitro expression map. The ORFs that were colored yellow were expressed above the median in MRC-5 cells and contained the early promoter motif indicated in panel A. The remaining ORFs were colored green.

RNA synthesized at lower temperature and in additional cell lines.

Our analyses suggested that there is a block in postreplicative transcription under in vitro infection conditions. Since MOCV replicates in the skin, we considered that virus replication is temperature sensitive at 37°C. Another possibility was that MOCV is specifically host range restricted in MRC-5 cells. To investigate these possibilities, RNA-seq data were obtained by infecting human MRC-5, human Huh7.5.1 cells, and monkey Vero cells at 31°C for 7 days. Huh7.5.1 cells are an epithelium-like hepatoma cell line that has a defect in the Rig-1 pathway (32) and has been used for difficult-to-propagate viruses, such as hepatitis C (33) and host range-restricted VACV (34). Vero epithelial cells, derived from an African green monkey kidney, are interferon deficient and commonly used for virus propagation (35). The read counts are listed in Table S3 in the supplemental material and are plotted along the MOCV genome in Fig. 6. The RNA-seq pattern in MRC-5 cells infected at 31°C (Fig. 6) closely resembled the pattern obtained at 37°C (Fig. 2). Although there were differences particularly in the RNA peak heights from Huh7.5.1 cells compared to those from Vero and MRC-5 cells, the overall patterns were similar (Fig. 6) and distinct from that of skin lesions (Fig. 2). Therefore, neither lowering the temperature nor using other cell lines overcame the restriction for postreplicative MOCV gene expression.

FIG 6.

RNA-seq patterns in multiple cell lines infected at 31°C. MRC-5, Huh7.5.1, and Vero cells were infected with approximately 10 vp per cell and incubated at 31°C for 7 days. At the end of this time, RNA was extracted and analyzed as described in the legend to Fig. 2.

MOCV promoter recognition in VACV- and MOCV-infected cells.

The expression of intermediate and late VACV genes depends on the synthesis of specific transcription factors and the replication of the genome. The intermediate transcription factors are expressed early in infection prior to viral DNA replication, whereas the late transcription factors are synthesized postreplicatively. This difference provides a way to distinguish between VACV intermediate and late promoters. Transfected reporter genes regulated by an intermediate promoter, but not an exclusively late promoter, will be expressed in cells infected in the presence of a DNA replication inhibitor (36, 37). Early genes are transcribed within virus particles, and early promoters have low activity in transfection assays (38) and were not tested. We synthesized the putative promoter sequences present upstream of MC044, MC069, and MC095 and inserted them adjacent to the ORF of GFP. MOCV promoter activities were confirmed by transfecting the above-described plasmids into BS-C-1 cells infected with VACV or ectromelia virus in the presence or absence of AraC. Fluorescent cells, indicating GFP expression, occurred with all three promoters, suggesting they belong to the intermediate or late class (Fig. 7A). However, only MCO44 promoter activity was prominent in the presence of AraC, indicating that VACV and ectromelia virus recognized that MOCV sequence as an intermediate promoter. To obtain more quantitative data, the same promoters were cloned adjacent to the firefly luciferase gene and transfected into BS-C-1 cells infected with VACV in the presence or absence of AraC. Again, all three promoters functioned in the absence of AraC, but only MC044 functioned as a strong intermediate promoter in the presence of the drug (Fig. 7B).

FIG 7.

Recognition of MOCV promoters by VACV and MOCV. (A) VACV recognition of MOCV intermediate and late promoters determined by GFP synthesis. DNA sequences upstream of ORFs MC044, MC095, and MC069 were cloned into a plasmid adjacent to the enhanced GFP ORF. The plasmids were transfected into BS-C-1 cells that were mock infected or infected with 10 PFU/cell of VACV in the presence or absence of AraC and incubated for 18 h. Fluorescence microscopic images are shown. (B) VACV recognition of MOCV intermediate and late promoter activities quantified by luciferase activity. The same promoter sequences used for panel A were cloned upstream of the firefly luciferase ORF and transfected into mock- or VACV-infected cells in triplicate as described for panel A. After 2 days, luciferase activity was determined. Luciferase activity was plotted as relative light units (RLU) per second. Standard deviation bars are shown. Asterisks indicate P values of <0.05. (C) Synthesis of intermediate transcription factors in MOCV-infected cells determined by luciferase expression. Triplicate cultures of Huh7.5.1 and MRC-5 cells were mock infected (−) or infected (+) with approximately 10 vp per cell of MOCV and then transfected with the same plasmids used for panel B. After 120 h, luciferase activity was measured. Standard deviation bars are shown. (D) Cells were infected and transfected at 37°C or 31°C as described for panel C. After 7 days, luciferase activity was measured. Standard deviation bars and P values are shown.

We next determined whether luciferase was expressed from the plasmids in cells infected with MOCV, which would occur only if transcription factor RNAs were synthesized and translated. The highest level of luciferase activity occurred with the MC044 promoter in Huh7.5.1 cells (Fig. 7C). The same promoter gave considerably less activity in MRC-5 cells, and the activity with the other promoters was not much above background. The above-described experiment was performed with cells kept at 37°C. More luciferase was detected when the Huh7.5.1 and MRC-5 cells were incubated at 31°C than at 37°C (Fig. 7D). The activity with the intermediate MC044 promoter provided evidence that the intermediate transcription factors were synthesized in cells infected with MOCV. Furthermore, the specificity for MC044 compared to that for MC095 and MC069 was consistent with the replication block occurring prior to intermediate or late gene expression. The higher activity with Huh7.5.1 cells than MRC5 cells could represent cell-specific differences in MOCV gene expression, mRNA stability, or the stability of MOCV proteins or luciferase.

DISCUSSION

The regulation of MOCV transcription was predicted to be similar to that of orthopoxviruses based on a comparison of their genome sequences (13). The present study, however, provides the first comprehensive experimental analysis of the MOCV transcriptome in cells infected in vitro and a naturally infected human skin lesion. The method closely followed that used for the analysis of the VACV transcriptome in cultured human cells (23). There were several notable differences between the data obtained with VACV and MOCV. First, for VACV the percentage of viral polyadenylated RNA amounted to 25 to 55% of the total within 4 h, whereas in MOCV-infected cells the percentage remained approximately 1%, even though all of the cells appeared to be infected based on the CPE. Second, there was a transition between the early pattern and the postreplicative pattern between 2 and 4 h after VACV infection, whereas the initial pattern persisted in MOCV-infected cells for 14 days and represented a subset of RNAs from a skin lesion. Several lines of evidence indicated that the persistent MOCV pattern corresponds to the early stage of VACV transcription. First, there was a high correlation between RNAs made in the absence and presence of a protein synthesis inhibitor, which prevents the transcription of intermediate and late genes. Second, although there are MOCV homologs of all classes of VACV genes, most MOCV RNAs made in vitro corresponded to the early homologs. Third, the consensus promoter sequence of the expressed MOCV genes was similar to that of VACV early promoters. Further, the synthesis of intermediate transcription factor proteins, products of early genes, in MOCV-infected cells was supported by transfection studies with reporter genes regulated by an MOCV intermediate promoter. There is some ambiguity in analyzing individual genes because of transcriptional readthrough, which is characteristic of orthopoxvirus RNA synthesis and likely is occurring with MOCV. Bugert and coworkers (9) reported the detection of RNAs by RT-PCR corresponding to MC095R and MC106L, which are homologous to VACV postreplicative genes. We also found RNAs corresponding to MC095R; however, they were detected in the presence of CHX, suggesting that they resulted either from readthrough of early MC094R RNA or early promoter activity of MC095. We also detected RNAs corresponding to MC106L, in the presence of CHX but not in it absence, which could be derived by readthrough from early MC107L. Taken together, we believe that the evidence for a block in MOCV postreplicative gene expression in cultured cells is overwhelming and consistent with the inability of others (9, 10) to demonstrate MOCV genome replication.

The CPE following infection with MOCV is similar to the cell rounding seen early after infection with VACV. In both cases, CPE is prevented by inhibitors of RNA and protein synthesis as well as UV irradiation of virus particles (7, 39), suggesting that early viral gene expression is involved. However, no noncytopathic VACV mutant has been described, and the putative protein(s) involved is unknown. The recovery of MOCV-infected cells from severe CPE and the continued isolation of RNA for 14 days, although previously documented (9), are remarkable. The latter likely is due to the persistence of virus cores in MOCV-infected cells (7), since there is evidence that intact cores synthesize early RNAs in VACV-infected cells (40).

In an effort to overcome the block in MOCV replication, we tried reducing the temperature to 31°C in order to mimic the lower temperature in the skin, the site of MOCV replication. However, the RNA pattern in MRC-5 cells remained similar to that obtained at 37°C. We also tried other cell lines, including Huh7.5.1 cells, which are deficient in the Rig1 pathway (32), and Vero cells, which are interferon deficient (35). In neither case were we able to substantially change the pattern of mRNAs.

We took a second approach to analyze the transcription program of MOCV and to provide evidence for the translation of early mRNAs. Studies with VACV have shown that the intermediate transcription factors are products of early genes (37). Therefore, reporter genes regulated by intermediate promoters can be expressed when transfected into cells even in the absence of viral DNA replication. Using this strategy, we showed that the DNA sequence upstream of the MC044L ORF was recognized by VACV as an intermediate promoter. Furthermore, a reporter gene regulated by the MC044L promoter was expressed in transfected MOCV cells, demonstrating that the transcription factors and likely RNA polymerase subunits, capping enzymes, and other proteins required for intermediate transcription in the case of VACV were made. Activity was higher in Huh7.5.1 cells than MRC-5 cells and at 31°C than at 37°C. Whether the low temperature directly improves MOCV gene expression, the stability of the reporter proteins, or the condition of the cells is uncertain. Sherwani et al. (41) reported GFP expression regulated by a synthetic VACV early/late promoter (42) in MOCV-infected cells. It would be interesting to know whether this promoter works in the presence of AraC or CHX, i.e., has intermediate promoter activity.

Why MOCV transcription does not progress beyond the early phase is uncertain. MOCV encodes highly conserved homologs of the orthopoxvirus genes that encode RNA polymerase subunits, transcription factors, and mRNA modification enzymes. As mentioned above, de novo synthesis of viral DNA has not been demonstrated following in vitro MOCV infection, and such a block would account for the restriction in intermediate and late transcription. Transcripts of MOCV homologs of genes directly involved in VACV DNA synthesis, including E9 DNA polymerase (MC039L), D4 uracil DNA glycosylase (MC093R), D5 helicase/primase (MC095R), and A20 processivity factor (MC026R), were detected (see Table S3 in the supplemental material), although we cannot determine whether the proteins all were made in sufficient amounts. Notably missing from MOCV, however, are orthopoxvirus genes that may be required in resting cells, including those encoding DNA ligase, ribonucleotide reductase subunits, glutaredoxin, thymidine kinase, thymidylate kinase, and deoxyuridine triphosphatase. While the latter proteins may restrict the sites of MOCV replication in vivo, cultured cells are actively dividing, making their absence less likely to be important. Wiebe and Traktman (43) pointed out that the absence in MOCV of the gene encoding the B1 kinase, which phosphorylates the cell protein known as barrier to autointegration factor, or BAF, could be important in determining cell tropism. BAF apparently acts by compacting viral DNA and preventing replication and subsequent transcription (44, 45). It would be interesting to determine whether the replication of MOCV is enhanced in BAF knockout cells.

The majority of orthopoxvirus genes that modulate the host response to infection also are missing from MOCV, although MOCV contains numerous unique genes, some of which may have analogous roles. Noticeably absent from MOCV is the E3 double-stranded RNA (dsRNA) binding protein or other protein with a similar dsRNA binding motif (E. Koonin, personal communication). The deletion of E3 from VACV results in a severe host range restriction involving the activation of RNase L and protein kinase R, and phosphorylation of eukaryotic translation initiation factor 2-α and the inhibition of both viral and host protein synthesis. It would be interesting to determine whether the activation of double-stranded RNA pathways occurs in MOCV-infected cells and if replication is enhanced in human cell lines in which these pathways are inactivated. An analysis of the effect of MOCV infection on the host transcriptome also would be enlightening. Although some changes were noted here, the present study was not designed to obtain statistically significant data on individual host mRNAs.

The present analysis of the MOCV transcriptome represents a starting point for further studies. The determination of RNA start sites and ribosome profiling, as has been done for VACV (31, 46), would provide a more precise transcription map. RNA-seq and the reporter gene assays described here may be useful for testing other cell lines and conditions that ultimately could provide an in vitro replication system.