We generated a replicating MCV molecular clone that can be manipulated to assess effects on virus DNA replication and encapsidation. As previously seen in origin-replication studies 
, introduction of the MCV350 strain point mutation into the MCV-HF replication origin abolishes the clone's ability to replicate. We also confirm that MCV sT protein expression together with LT expression is required for optimal MCV replication, a feature of viral genome replication that MCV shares with the human JC polyomavirus 
Fully-encapsidated MCV virions from MCV-HF were isolated in fraction 4 at 1.24 g/ml on iodixanol (Optiprep™) gradients. Evidence that these are fully-encapsidated viruses include isolation of nuclease-protected DNA and ultra-high molecular weight aggregates of VP1 protein specific to this fraction. In the 1970s, hyperosmolar CsCl isopycnic gradients were used to isolate polyomaviruses having higher apparent buoyant densities (e.g., 1.34 g/ml 
). It is now well-established that hyperosmolar CsCl gradients overestimate the densities of large macromolecular structures, such as viruses, since CsCl gradients dehydrate virions, replacing water with heavy salts, which reduces viability and artifactually increases the buoyant density. JC virus has a buoyant density of 1.20 g/ml on linear sucrose gradients and 1.35 g/ml in CsCl gradients 
. Similarly, goose hemorrhagic polyomavirus virion has a 1.20 g/ml density in sucrose and an apparent 1.34–1.35 g/ml density with CsCl gradients 
. MCV DNA has been isolated from skin at 1.22 g/ml density using iodixanol gradients, which is in agreement with our findings 
. The encapsidated MCV we identify in fraction 4 using uranyl acetate negative staining have the same size (~40 nm) as similarly prepared MCV VLP but are smaller than MCV VLP (55–58 nm) visualized using phosphotungstic acid staining 
MCV, like other polyomaviruses, has a strictly regulated order of viral gene expression that leads to encapsidation of the viral genome. MCV LT and 57kT proteins are expressed early after MCV-HF transfection. Subsequently sT and VP1 (and presumably other virion protein components) are detectably expressed, leading to generation of self-assembling viral particles. Our study reveals an unexpected and interesting complexity for MCV gene expression. MCV LT, sT and 57kT are generated from alternatively-spliced, overlapping genes driven by the same early promoter. MCV-Rep−
expression for LT is initially identical to MCV-HF, indicating that the early promoter is intact, but other early proteins are diminished (57kT) or absent (sT) suggesting the possibility that viral DNA replication also regulates early TAg splicing. Late VP1 protein expression also is not detected using the MCV-Rep−
genome indicating that it is likely the early-to-late promoter switch also depends on viral DNA replication. This is similar to mouse polyomavirus, a close relative to MCV in the murine polyomavirus clade 
. Well-established late lytic expression and packaging cascades for other viruses, such as herpesviruses, are also dependent on active viral DNA replication 
. Although our results are consistent with DNA replication-dependent viral transcription, MCV-Rep−
is mutated with a single pentamer sequence (PS) 7/8 substitution in the MCV origin and so we cannot exclude the possibility that this point mutation also affects late promoter activity in addition to genome replication. Additional studies are needed to determine if DNA replication also regulates MCV splicing patterns.
One key factor determining MCV-HF virion production is the abundance of sT protein, which our data suggests is in turn increased in a positive feedback loop during MCV replication. When sT is coexpressed with the transfected MCV-HF genome, there is 5–6 fold increased virus production. For mouse polyomavirus, sT can signal to AP-1 and PEA3 factors to promote viral replication and late gene expression 
. Based on SV40 studies, the most prominent role of sT involves inhibition or retargeting of substrates for the major cellular phosphatase, PP2A 
. We previously identified MCV sT as an important accessory factor for efficient MCV DNA replication 
, but it is unknown whether this is due to PP2A or 4E-BP1 targeting, or due to targeting of our other cellular factors 
. The Southern blot in reveals that MCV DNA replication is sparse and most of the replicated DNA is fragmented. This suggests that MCV genome replication, at least under conditions of plasmid transfection, is inefficient and small increases in successful MCV genome replication may have a large impact on virion replication and encapsidation. Whether the same is true during natural infection will only be determined through development of a successful MCV transmission system.
Our findings that hVam6p inhibits MCV production are unexpected and, to a degree, paradoxical. MCV LT retains a conserved hVam6p-binding domain that represses virion production. Other polyomaviruses also encode gene sequences that repress viral replication 
, which might reflect the need for these small DNA viruses to suppress virus replication to sustain a chronic, persistent infection without virion production (e.g., latency). Although the effects of hVam6p on MCV replication are clear, the mechanism by which hVam6p suppresses replication is not. Our origin replication studies in the presence and absence of hVam6p overexpression, together with LT.W209A mutations in LT coding region, show no direct effect of hVam6p on initiation of viral DNA replication. Additional studies are also needed to investigate this effect, however, since Southern blotting reveals amplified viral DNA replication for MCV-hVam6p−
virus compared to MCV-HF, suggesting hVam6p might be inhibitory to viral DNA replication in the context of the full viral genome.
It is noteworthy that overexpression of this vacuolar sorting protein has a profound antiviral effect on MCV replication. hVam6p possesses citron and clathrin homology domains, the latter being involved in MCV LT binding 
, that are important to its functions in the HOPS-CORVET complex as an accessory factor for endosomal fusion 
. In yeast, the hVam6p homolog also has been reported to act as a guanine nucleotide exchange factor for Gtr1 that contributes to TORC1 activation 
. An isoform of hVam6p, TRAP-1-like protein (TLP) regulates the balance between Smad2 and Smad3 in TGF-ß signaling 
. Previous studies did not find either mTOR activation or TGF-ß signaling, however, to be appreciably altered by MCV LT expression 
. In our current study, even low levels of hVam6p transfection reduce MCV-HF virus production to levels similar to the replication-deficient clone demonstrating that it is a potent factor restricting MCV replication. For the MCV-hVam6p−
virus, hVam6p inhibits replication in a dose-dependent manner, which may be due to either the W209A substitution incompletely disabling the hVam6p-binding site or hVam6p acting in other steps of MCV replication beyond those depending on LT. If hVam6p plays a role in inhibiting egress of MCV, this might represent a novel component of innate immunity. Tetherin, for example, is a recently-discovered innate immune component that prevents enveloped viral budding from cells 
. Detailed analysis of hVam6p's role in antiviral responses is beyond our current study but development of a replicating MCV clone that can be genetically manipulated provides a critical reagent for use in these follow-on investigations.
Secondary MCV transmission was not detected in our study suggesting that MCV may have a tissue tropism that is not easily modeled in undifferentiated tissue culture. MCV resembles some other small DNA viruses in this way, such as JCV as well as other human tumor viruses including human papillomaviruses (HPV), hepatitis B and C viruses and KSHV. Poor MCV transmissibility could be due in part to low virus yields from our molecular clone system that might be improved with other MCV strains or cell culture settings. Cloning MCV-HF, however, provides a useful tool for these testing conditions to optimize virus yield that may ultimately allow a laboratory transmission model.