|Home | About | Journals | Submit | Contact Us | Français|
Mounting evidence supports the concept that Merkel cell polyomavirus (MCV) is a causal factor underlying most cases of a highly lethal form of skin cancer known as Merkel cell carcinoma. To explore the possibility that polyomaviruses commonly infect healthy human skin, we developed an improved rolling circle amplification (RCA) technique to isolate circular DNA viral genomes from skin swab specimens. Complete MCV genomes were recovered from 14/35 (40%) healthy adults, providing the first full-length, apparently wild-type cloned genomes for this polyomavirus species. RCA analysis also revealed the existence of two previously unknown polyomavirus species that we name human polyomavirus-6 (HPyV6) and HPyV7. Biochemical experiments show that polyomavirus DNA is shed from the skin in the form of assembled virions. A pilot serological study indicates that infection or co-infection with the three skin-tropic polyomaviruses is very common. Thus, at least three polyomavirus species are constituents of the human skin microbiome.
The Polyomaviridae are a family of non-enveloped viruses that carry circular, double-stranded DNA genomes. The family is named for some members’ ability to induce various types of tumors in experimentally infected animals. For example, the polyomaviruses BKV and JCV, which persistently infect the urinary epithelia in a great majority of humans, can cause tumors in experimentally inoculated rodents (Corallini et al., 1978; Ohsumi et al., 1986). Although BKV and JCV have been indirectly associated with the development of various forms of human cancer, such as prostate cancer and colorectal cancer (respectively), conclusive proof of a causal relationship between BKV or JCV and human cancers has remained elusive (reviewed in (Abend et al., 2009; Maginnis and Atwood, 2009)). Two more recently-discovered human polyomaviruses, WUV and KIV (Allander et al., 2007; Gaynor et al., 2007), have also been shown to infect a majority of humans, but clear links between these two viruses and human disease, including cancer, have not so far been identified (reviewed in (Dalianis et al., 2009)).
The recent discovery of a fifth human polyomavirus associated with an unusual form of skin cancer called Merkel cell carcinoma (MCC) has rekindled research interest in the possibility that polyomaviruses cause cancer in humans ((Feng et al., 2008), reviewed in (Zur Hausen, 2009)). DNA sequences of the newly discovered virus, named Merkel cell polyomavirus (MCV), are present in about 80% of MCC tumor specimens. Furthermore, MCV genomes have been shown to be clonally integrated into the cellular DNA of some MCC tumors and their metastases. A majority of MCC tumors also display ongoing expression of the MCV large T antigen oncoprotein (Shuda et al., 2009). Taken together, the results strongly suggest a causal relationship between MCV and a majority of MCC cases.
Although serological evidence indicates that most adults have been immunologically exposed to MCV (Carter et al., 2009; Kean et al., 2009; Pastrana et al., 2009; Tolstov et al., 2009), the nature of MCV infection in healthy individuals remains unclear. Subgenomic fragments of MCV DNA have been detected in a variety of healthy specimen types, including skin, saliva, gut, and respiratory secretion samples (Feng et al., 2008; Loyo et al., 2009; Wieland et al., 2009). At present, only four full-length MCV genomes have been reported, each of which was cloned by PCR-based amplification of tumor-derived DNA (Feng et al., 2008; Katano et al., 2009). All four available genomes carry truncating mutations in the T antigen gene in a pattern typical of MCV sequences amplified from tumors (Shuda et al., 2008). The tumor-derived reference isolate, MCV-350, also encodes functional defects in its origin of replication and VP1 capsid protein gene (Kwun et al., 2009; Pastrana et al., 2009).
The extent to which the complete genomes of tumor-derived MCV strains are distinct from strains circulating among healthy individuals is not known. To further explore the tissue tropism and sequence diversity of MCV in individuals without MCC, we set out to capture full-length wild type (wt) MCV DNA shed from the skin of healthy volunteers. Detection of MCV DNA was facilitated by a method known as rolling circle amplification (RCA), a random-primed extension reaction that employs a high fidelity DNA polymerase from bacteriophage phi29 to selectively amplify circular DNA molecules, such as polyomavirus genomic DNA (reviewed in (Johne et al., 2009)).
Analysis of cloned RCA products revealed the presence of wild-type MCV genomes, as well as a variety of other circular dsDNA molecules, including sequences of a wide variety of human papillomavirus (HPV) species known to infect the skin. Sequencing of cloned RCA products also revealed the existence of two previously unknown polyomaviruses. The results draw a striking parallel between papillomaviruses and polyomaviruses, and reveal an intriguing new pair of polyomavirus target species for studies aimed at uncovering additional links between Polyomaviridae and human cancer or other diseases.
Our initial study goal was to isolate full-length, wt MCV genomic DNA from swabs drawn across the surface of human skin. The skin of the forehead was chosen based on a recent report by Wieland and colleagues showing that short MCV PCR products can be amplified from this easily sampled skin surface (Wieland et al., 2009). DNA was extracted from the skin swab specimens, and then subjected to random hexamer-primed RCA.
Under ideal circumstances, RCA produces a long polymer of tandem repeats of any circular template present in the reaction mixture. RCA reactions performed on DNA extracted from skin swabs were analyzed by digesting the finished reaction with restriction enzymes, such as BamHI or EcoRI, that would be expected to cut known MCV isolates to unit length. Agarose gel analysis of an initial set of samples showed a smoothly distributed smear of products, suggesting that the majority of RCA products were derived from random fragments of linear cellular DNA, as opposed to discrete circular DNA templates (data not shown). To overcome this problem, we incorporated a pre-processing step in which the extracted DNA was digested with exodeoxyribonuclease V, an enzyme that digests linear DNA molecules but spares double-stranded circular DNA molecules. This pre-processing step was augmented by the inclusion of a restriction enzyme that we reasoned would be unlikely to digest MCV DNA. Like other polyomavirus genomes, available MCV sequences are relatively deficient in CpG dinucleotide motifs (White et al., 2009). The restriction enzyme NotI, which contains two CpG dinucleotides in its eight base pair recognition motif, was therefore used for the pre-digestion step. The recognition sequence of the restriction enzyme SalI, which contains one CpG motif, also tends not to occur in polyomavirus genomes, so this enzyme was chosen for the pre-treatment step in repeat sampling experiments (see below).
Swab specimens were collected from a total of 35 study volunteers. Sixteen of the volunteers were sampled on two occasions, roughly three months apart. Modified RCA analysis of swab samples revealed discrete restriction fragment banding patterns for most individuals. An example of a restriction fragment analysis of skin swab RCA samples is shown in Figure S1.
It is well established that most humans chronically shed the DNA of a variety of skin-tropic human papillomavirus (HPV) species from the surface of their skin (Antonsson et al., 2003). Thus, it was not surprising to find that a majority of cloned RCA restriction fragments exhibited varying degrees of homology to human papillomaviruses of the skin-tropic genera beta and gamma (Figure S1 and data not shown). The complete genome of a previously unknown Gammapapillomavirus species was cloned from RCAs of subject 3. The clone was submitted to the Human Papillomavirus Reference Laboratory (Heidelberg, Germany) and has been designated HPV type 127.
In addition to HPV-related fragments, a variety of bacterial plasmid-like sequences were observed. One 5 kb plasmid-like sequence was cloned from RCAs of subject 26. The plasmid-like sequence was named pMobRep based on the presence of ORFs encoding proteins with roughly 50% similarity to bacterial MobA, MobC and RepA proteins. Overall, approximately 10% of cloned RCA restriction fragments did not have clear homologs in BLAST or BLASTX searches.
RCA samples from study subjects 17 and 30 displayed prominent ~5 kb BamHI or EcoRI restriction fragments. Sequencing of the cloned restriction fragments revealed them to be complete MCV genomes (isolates R17a, R17b and R30a). High-fidelity PCR, with RCA-amplified material as the template, was used to clone complete MCV genomes from an additional twelve study subjects. A cumulative total of 14/35 (40%) of study subjects yielded complete MCV genomes at one or both time points (Figure 1).
As expected, each of the MCV isolates encodes full-length open reading frames for predicted MCV proteins, including the large T antigen. The one exception is isolate 09b, in which a one basepair insertion results in truncation of the C-terminal 15 amino acids of the predicted small t antigen protein. Each of the clones encodes an identical consensus T antigen peptide sequence, with the exception of clone 16b, which differs from the consensus by only two amino acids. The protein non-coding control regions (NCCRs) of the MCV clones contain a small number of single nucleotide substitutions and short insertions or deletions. Interestingly, the NCCR of isolate 16b encodes a 25 basepair tandem repeat resulting in duplication of a GGGNGGRR sequence motif that has previously been shown to enhance transcription and in vitro replication in JCV, SV40 and a variety of other virus types (Martin et al., 1985). An identical 25 basepair tandem repeat exists in the previously reported tumor-derived MCV clone TKS (Katano et al., 2009).
Clones 10b, 16b and TKS were isolated from individuals born in Asia. The fact that all three clones are divergent from the main cluster of MCV isolates derived from individuals born predominantly in North America and Europe raises the possibility that different MCV strains are prevalent in geographically distinct human populations. This would be reminiscent of the situation for BKV and JCV, whose worldwide sequence divergence patterns have been shown to resemble “out of Africa” models of prehistoric human migration (Yogo et al., 2004; Zhong et al., 2009).
Analysis of the cloned MCV genomes revealed that each subject who tested positive at both time points shed identical or nearly identical MCV sequences at both time points (Figure 1). This suggests that each subject is chronically infected with a specific MCV isolate. The result also indicates that there was not extensive cross-contamination among subject samples.
A confounding unknown for serological studies of MCV exposure has been the possibility that distinct MCV genotypes or serotypes might be disproportionately associated with MCC tumors. The fact that the four available full-length tumor-derived MCV clones are interspersed among the apparently wild-type clones reported here (Figure 1) argues against the possibility that particular MCV genotypes are over-represented in tumors.
Pairwise alignments of the VP1 proteins encoded by the new MCV isolates against all available MCV VP1 protein sequences in GenBank revealed a minimum identity of 98.6%. Among human polyomaviruses, only BKV is known to have subspecies that are serologically distinct (Knowles et al., 1989). The VP1 protein divergence for MCV isolates is comparable to the divergence seen within BKV serogroups (Figure S1). The results argue against the existence of distinct serotypes among available MCV isolates.
Sequencing of a 1.6 kb BamHI fragment from the initial RCA analysis of subject 7 revealed limited homology to the VP2 protein of WU polyomavirus in a BLASTX search. The full genomic sequence of the previously unknown virus species was amplified by PCR and cloned. The novel virus is the sixth fully sequenced polyomavirus species of human origin (reviewed in (zur Hausen, 2008)). We therefore suggest the name “human polyomavirus-6” (HPyV6).
To search for additional polyomaviruses similar to HPyV6, we performed PCR using degenerate PCR primers targeting conserved portions of various polyomavirus genomes. Since it did not appear to be possible to design “universal” primers capable of binding a broad range of polyomavirus species, we designed partially degenerate primers targeting sequences conserved between specific pairs of polyomaviruses (see Supplemental Experimental Procedures). A primer pair designed to bind both HPyV6 and WUV successfully amplified a unique polyomavirus-like sequence. The complete genome of this second novel polyomavirus was amplified by PCR and was found to be 68% identical to HPyV6 at the nucleotide level. We named the new species HPyV7. The degenerate primer sets did not detect any additional polyomavirus species beyond MCV, HPyV6 or HPyV7.
Complete HPyV6 and HPyV7 genomes were cloned from a total of 5/35 or 4/35 individuals, respectively. As with MCV, repeat sampling showed that individual subjects continued to shed a similar or identical HPyV6 or 7 sequence at both time points (data not shown), again suggesting a chronic clonal infection.
Phylogenetic comparison of the complete genomes of HPyV6 and 7 to other polyomavirus species shows that the two new viruses occupy a distinct clade (Figure 2). The two new viruses also occupy a longer common branch with WUV and KIV. This apparent affiliation is attributable primarily to the extreme divergence of the late regions (VP1 or VP2 capsid genes) of these four species from all other polyomaviruses (Figure S2). In contrast to the late regions, the early regions of HPyV6 and 7 are quite distinct from those of WUV and KIV (data not shown). A possible explanation for this observation could be that the WU/KI/6/7 cluster arose after an ancient recombination event that joined the early and late regions of two distantly related polyomavirus species.
Productive viral infection is associated with the generation of progeny virions. To determine whether the polyomavirus DNA detected in the skin swab specimens was shed in the form of assembled virions, skin swabs contributed by nine volunteers were extracted using nondenaturing conditions. A broad-spectrum endonuclease was added to the extract with the goal of digesting any DNA not protected within a virion. The digested extracts were subjected to an Optiprep step gradient ultracentrifugation technique designed to separate virions from free DNA and other extract components (Buck et al., 2004).
As depicted in Figure 3, qPCR signals for MCV, HPyV6 and HPyV7 were observed in fractions 2–6 of the swab extract gradient. These fractions had densities of 1.29, 1.22, 1.19. 1.16 and 1.07 g/ml, respectively. This is a relatively broad migration profile compared to recombinant MCV VP1/VP2 virus-like particles (VLPs), which typically show an apparent density focused around 1.21 g/ml (Buck et al., 2004; Pastrana et al., 2009; Tolstov et al., 2009). A possible explanation for the broad migration profile is that the viral DNA is protected within virions that are heterogeneously complexed with relatively buoyant compounds, such as skin-derived lipids or cosmetics. This concept is consistent with recent observations showing that recombinant MCV VP1 capsomers can bind buoyant glycosphingolipids in the setting of sucrose gradients (Erickson et al., 2009).
The swab extract gradient contained a total of roughly two million MCV genome equivalents and roughly 5,000 HPyV6 genome equivalents. Cycle sequencing of the faint PCR products observed in fractions 3, 5 and 6 with the HPyV7-specific primer set confirmed an appropriate HPyV7 sequence. However, the overall signal for this species was too low to allow a confident interpretation. Gradient analyses of three independent sets of swab specimens showed similar results (data not shown).
As a positive control for the gradient analysis, we generated modest yields of MCV virions by transfecting 293TT cells with recombinant MCV isolate R17a genomic DNA. The MCV virions produced in transfected 293TT cells showed a focused migration profile, with a strong peak in fraction 3.
No qPCR signal was detected in a control gradient in which free MCV isolate R17a plasmid DNA was spiked into a mock extract containing endonuclease (data not shown). Free MCV DNA that was not subjected to nuclease treatment remained predominantly in the upper portion of the ultracentrifuge gradient (Figure 3), consistent with the greater buoyancy of non-encapsidated DNA in Optiprep.
The data suggest that the high density nuclease-resistant DNA in the skin swab extracts and transfected cell lysates is in the form of assembled virions.To confirm this explanation, we performed an experiment in which monoclonal antibodies specific for the VP1 major capsid protein of MCV (Pastrana et al., manuscript submitted) were used to strip MCV DNA out of Optiprep fraction material. The results, shown in the bottom left panel of Figure 3, confirm that a majority of the MCV DNA in the swab extract and 293TT-produced virion gradients is associated with VP1 protein. Taken together, the results clearly demonstrate that MCV DNA is shed from the skin in the form of assembled virions. The high density of a majority of the nuclease-resistant HPyV6 (and perhaps HPyV7) DNA in Optiprep gradients suggests that these viruses are also shed in the form of virions.
Polyomavirus VP1 proteins typically have the ability to spontaneously self-assemble into virus-like particles (VLPs) when expressed as recombinant proteins. Enzyme-linked immunosorbent assays (ELISAs) employing recombinant polyomavirus VLPs have been widely used to detect virus-specific antibody responses aroused during natural infection. To develop ELISAs targeting HPyV6 and HPyV7, the VP1 protein of each virus species was expressed in transfected 293TT cells and purified by Optiprep gradient ultracentrifugation. Although the HPyV6 and HPyV7 VP1 proteins migrated to the core fractions of the gradient where VLPs are typically found, electron micrographs revealed irregular particles smaller than the 45–50 nm diameters expected for full-size polyomavirus VLPs (Figure S3). The VP1 subviral particles were of sufficient purity for use in ELISAs. Full-size murine polyomavirus (MPyV) or MCV-based VLPs were produced in 293TT cells for use in negative and positive control ELISAs, respectively (Tolstov et al., 2009).
A set of 95 serum samples that had previously been tested for reactivity against MCV (Pastrana et al., 2009; Tolstov et al., 2009) were re-tested in separate ELISAs against VP1 particles based on MPyV, MCV, HPyV6 or HPyV7 (Figure 4). Each serum sample showed a distinct pattern of reactivity to each of the three human polyomavirus VP1 proteins, suggesting that each ELISA is polyomavirus species-specific. The 58/95 (61%) rate of seropositivity for MCV is similar to previous observations for this group of sera (set “Commercial Donor” reported in (Tolstov et al., 2009)). Reactivity against HPyV6 was also very common, with 66/95 (69%) subjects scoring seropositive. The seropositivity rate for HPyV7 was lower, with 33/95 (35%) subjects scoring seropositive. The frequency of double seropositivity for pairs of virus species was similar to frequencies predicted by the products of individual frequencies (chart in Figure 4). This suggests that serological exposure to each cutaneous polyomavirus species is an independent variable.
It is well established that most humans become infected with BKV and JCV polyomaviruses during childhood. Both of these polyomavirus species can chronically infect the stratified epithelial surfaces of the lower urinary tract, where virions are periodically shed into the urine (Hogan et al., 1980; Boldorini et al., 2005; Singh et al., 2006). In the current work, we show that virions of a separate group of polyomavirus species are shed from the surface of a different stratified epithelial surface, the skin. Our data suggest that nearly all adults are persistently infected with one or more of the cutaneous polyomavirus species.
Although known human polyomaviruses appear not to cause overt clinical symptoms in a great majority of infected individuals, they can cause disease in subjects with impaired immune function. For example, BKV can induce nephropathy in kidney transplant recipients and JCV causes progressive multifocal leukoencephalopathy, a fatal brain disease that sometimes strikes immunocompromised individuals (reviewed in (Jiang et al., 2009)). Similarly, the observation that risk of MCC increases dramatically in immunocompromised subjects provided an initial line of evidence supporting the concept that MCC might be caused by an infectious agent (Engels et al., 2002; Feng et al., 2008). By extension, it seems likely that diseases that might be caused by HPyV6 or HPyV7 would occur at greater frequency in immunocompromised subjects. Past work establishing causal links between polyomaviruses and human disease suggests that possible disease associations could be explored through PCR or in situ detection of the viral DNA in lesion samples (reviewed in (Buck and Lowy, 2009)), or by immunohistochemical detection of T antigen protein expression in tissue specimens (Shuda et al., 2009). Quantitative serological analysis using the VP1 ELISA reported here might also be used to identify possible links between high titer polyomavirus-specific seroreactivity and candidate disease states (Carter et al., 2009; Pastrana et al., 2009; Tolstov et al., 2009).
Several of the MCV isolates reported in this work bear striking similarity to MCV clones isolated from MCC tumors, with one pair (16b versus TKS) differing by only three nucleotides in the entire 5.4 kb MCV genome. Analysis of VP1 peptide sequences similarly suggests that all available MCV isolates occupy a single serotype (Figure S1). Wild-type and tumor-derived MCV isolates also appear to carry relatively consistent non-coding control regions (NCCRs). This contrasts with the complex insertions and deletions that can sometimes be found in the NCCRs of BKV and JCV isolates derived from diseased tissues (reviewed in (White et al., 2009)). Together, the results argue against the possibility that MCC tumors are caused by a subset of MCV types distinct from those that commonly infect healthy individuals.
The chronic shed of polyomavirus virions from the surface of the skin is strikingly reminiscent of the Papillomaviridae, a different family of non-enveloped circular dsDNA viruses that replicate exclusively in keratinocytes of the epidermis and are transmitted environmentally or via interpersonal contact. Although the presence of MCV in MCC tumors indicates that the virus can enter Merkel cells or their epidermal precursors (Van Keymeulen et al., 2009), the fact that Merkel cells make up fewer than 1% of cells in the epidermis strongly suggests that more common epidermal cells, such as keratinocytes or melanocytes, are involved in the active production of virions.
For papillomaviruses, the production of virions is closely tied to keratinocyte differentiation (reviewed in (Doorbar, 2005)). Consequently, papillomaviruses cannot be propagated in standard monolayer cell cultures. Although cultured cell monolayers transfected with recombinant MCV genomes can produce virions (Figure 3), the overall production rate is very low, averaging only about one virion per cell. Cells transfected with HPyV6 or HPyV7 genomes appear to yield similarly meager numbers of virions (data not shown). The results raise the possibility that the life cycles of the cutaneous polyomaviruses are tied to epidermal differentiation.
The family Polyomaviridae currently contains only one genus, Polyomavirus. We are not aware of an exact standard for what constitutes a distinct polyomavirus species and the assignment of species names is informally governed by a confusing set of at least four different conventions. This contrasts sharply with the situation for the family Papillomaviridae, which has recently benefited from the development of a precise set of classification rules (Bernard et al., 2010; De Villiers et al., 2004). Given the remarkable range of biological similarities between the two families – including their apparent capacity to occasionally undergo inter-familial recombination (Woolford et al., 2007) – we feel that the application of a papillomavirus-like taxonomic system to the polyomaviruses could be of great utility. Based on our understanding of the system, we would suggest that HPyV6 and HPyV7 be considered two distinct species that occupy a novel genus-like clade. Since HPyV6 and HPyV7 are the fifth genus-like clade of polyomaviruses known to contain a human-tropic species, we would suggest the name “Epsilonpolyomavirus” for their clade. While we are in favor of the idea that genus designations might ultimately be superimposed on existing polyomavirus species, it is our view that reassigning “HPyV#” species names to the five previously-discovered human polyomaviruses would be undesirable and would violate the ICTV’s Code of Virus Classification, sections 2.1 and 3.9 <www.ictvonline.org>.
The discoveries of MCV, WUV and KIV were all facilitated by the use of high-throughput shotgun sequencing (Allander et al., 2007; Feng et al., 2008; Gaynor et al., 2007). Since the initial goal of the current study was to capture full-length MCV clones, we elected to rely on traditional plasmid-based cloning techniques. These methods severely limited the throughput of our analysis, and at least a third of the restriction fragments visualized by agarose gel electrophoresis could not readily be cloned. The improved random-primed RCA method reported here should be adaptable for use with modern high-throughput sequencing methods, potentially allowing the future development of a more comprehensive catalog of polyomaviruses and other circular microbial DNA molecules commonly present on various human skin and mucosal surfaces.
Employees of the NCI’s Laboratory of Cellular Oncology were recruited for the study under NIH protocol 09-C-N219. Each volunteer was informed about the study and invited to consent to self-sampling by drawing a sterile wooden cotton-tipped swab across the skin of their forehead, hairline and eyebrows.
DNA was extracted from swab specimens using QIAquick PCR purification columns (Qiagen). The purified DNA was digested with NotI or SalI-HF (NEB) and Plasmid Safe (exonuclease V, Epicentre) then ethanol precipitated. The pelleted DNA was redissolved and amplified using an Illustra TempliPhi RCA kit (GE Healthcare). Additional details are given in Supplemental Experimental Procedures and at our laboratory website <http://home.ccr.cancer.gov/LCO/>.
PCR primers and cycling parameters are given in Supplemental Experimental Procedures.
GenBank accession numbers for the full-length viral genomes reported in this work are HM011538-HM011570. Sequence analysis was performed using MacVector 11 software. Neighbor-joining phylogenetic trees for nucleotide sequences were constructed from ClustalW alignments using Kimura’s two-parameter distance method. Protein analyses were aligned by Gonnet series and trees were constructed using uncorrected (“p”) without bootstrapping. Trees were displayed using FigTree 1.2 software <http://tree.bio.ed.ac.uk/software/figtree/>. Accession numbers for polyomavirus sequences used to construct phylogenetic trees are provided in Supplemental Experimental Procedures.
Virions and VLPs were extracted from sets of skin swabs from 9–11 individuals or lysates of roughly 2 × 108 293TT cells transfected for 1–2 weeks with reconstituted MCV isolate R17a genomic DNA. VLPs and subviral particles were produced by transfecting 293TT cells with VP1 expression plasmids pwP, pwM (Tolstov et al., 2009), p6VP1 or p7VP1 carrying, respectively, codon-modified versions of MPyV, MCV, HPyV6 or HPyV7 VP1. In all instances, the extraction buffer was composed of Dulbecco’s PBS supplemented with Triton X-100 and Benzonase Endonuclease (Sigma). A previously-reported Optiprep velocity/density ultracentrifuge gradient method was used to purify virions and VP1 particles away from other extract components (Buck and Thompson, 2007). Additional details are given in Supplemental Experimental Procedures.
Samples of pooled Optiprep gradient fractions were diluted in phosphate buffer then treated with a mixture of MCV VP1-specific IgM mAbs MV32 and MV62 (Pastrana et al, manuscript in preparation). An anti-sarcomeric actin IgM mAb (Invitrogen) was used as an isotype control. Antibodies were stripped out of solution using anti-IgM magnetic beads (Dynal, Invitrogen). The supernatant was digested with proteinase K, then subjected to ethanol precipitation and analyzed by MCV-specific qPCR. Additional details are given in Supplemental Experimental Procedures.
This research was funded by the Intramural Research Program of the NIH, with support from the Center for Cancer Research and the Director’s Innovation Award (National Cancer Institute). We are grateful to John Schiller and Doug Lowy for critical review of the manuscript and to Ted Pierson for useful discussions. We thank Ethel-Michele de Villiers and Len Norkin for guidance on the issue of polyomavirus species naming. We are grateful to Dr. de Villiers for re-sequencing and archiving HPV127 on behalf of the Human Polyomavirus Reference Laboratory.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.