Constructing a replication-competent clone of PARV4 in the context of AAV5 ITRs
We first cloned nucleotide nt 127-5268 of the PARV4 genome (GenBank accession no.: NC_007018) into vector pCR2.1, which encompasses a coding region with two large open reading frames (ORF) and a portion of the terminal repeats at the 5' and 3' ends (Jones et al., 2005
). This incomplete PARV4 genome lacks the terminal repeats at two ends, and therefore will not replicate in any cell culture system. To determine the transcriptional profile of PARV4 in a replication-competent system, we incorporated the incomplete PARV4 genome into the context of AAV5 ITRs, which resulted in plasmid p5TRPARV4. It replicated efficiently in 293 cells that expressed AAV5 Rep78 protein and necessary Ad5 gene products (i.e., E2, E4orf6 and VA RNA from the pHelper plasmid) (Guan et al., 2008
) (, lane 4). In contrast, the pSKPARV4 clone, which did not contain the AAV5 ITRs, did not replicate in 293 cells (, lane 8). Thus, we used the p5TRPARV4 construct for further analysis of PARV4 gene expression.
Replication of the construct p5TRPARV4 in 293 cells
5'/3' rapid amplification of cDNA ends (RACE) and reverse transcription (RT)-PCR identification of the transcription units of the PARV4 genome
We used total RNA extracted from 293 cells that had been co-transfected with p5TRPARV4, pAAV5Rep78, and pHelper to analyze PARV4 mRNA transcripts. The 5' RACE using the primer Rnt960 generated a predominant band of ~690 nts (, lane 1). Sequence analysis of this band revealed that mRNA transcripts spanning the region from the 5’ end to nt 960 initiated at nt 262. The 5' RACE with primers Rnt3840 and Rnt3440 produced major bands of ~700 nts and ~300 nts, and minor bands of ~600 nts and ~200 nts, respectively (, lanes 2&5, respectively). Sequence analysis of the major bands revealed the presence of a 5’ end of PARV4 mRNAs at nt 1950, and splicing sites of D2 donor at nt 2060, A2 acceptor at nt 2309, D3 donor at nt 2367, and A3 acceptor at nt 3329. In addition, sequence analysis of the minor bands confirmed the 5’ end of PARV4 mRNAs at nt 1950, and revealed the splicing sites of D3m donor at nt 2013 and A3 acceptor at nt 3329. Taken together, these results demonstrated the existence of two promoters at map unit 6 and 38, namely P6 and P38, which transcribed PARV4 mRNAs of ORFs at the left and right hands of the PARV4 genome, respectively. The 3' RACE with primer Fnt3518 produced two predominant bands of ~1,500 nts and ~300 nts (, lane 7). Sequence analysis of these two bands revealed that they terminated at nt 5149 and nt 3771, respectively, suggesting that PARV4 mRNAs are polyadenylated at either nt 5149 or nt 3771.
5’/3’ RACE and RT-PCR analyses of PARV4 mRNAs
More extensive RT-PCR analysis using different sets of primers followed by sequencing also showed the same results as those obtained by 5'/3' RACE, which confirmed the donor and acceptor sites for alternative splicing of PARV4 mRNAs generated from the P6 and P38 promoters (, lanes 3, 4, and 6). Sequence analysis of the DNA fragments amplified by PCR using primers Fnt283 and Rnt2060 revealed an intron that lies in the NS1-encoding region from the D1 donor (nt 418) to the A1 acceptor sites (nt 1630).
Taken together, these data provide the locations of all transcription units in the PARV4 genome by 5’/3’ RACE and RT-PCR, which are diagramed in . The sequence results of the amplified fragments are shown in detail in .
Sequencing results of the DNA bands obtained from 5'/3' RACE and RT-PCR
RNase protection assay (RPA) analysis of PARV4 mRNA transcripts
To determine the relative abundance of each PARV4 mRNA transcript, we used 8 anti-sense probes to protect individual transcript. A schematic diagram of the probes with their putatively protected bands and respective sizes is shown in .
Transcription mapping by RNase protection assays (RPAs)
Probe P1, which spanned the P6 promoter and D1 donor site, protected bands of 219 and 157 nts in sizes. These bands confirmed the mRNA initiation site at nt 262 and the D1 donor site at nt 418. Notably, the spliced band (157 nts) at the D1 donor site accounted for less than 5% of the total protected bands (, lane 2), suggesting that mRNA splicing at the D1 site is rather poor.
Probe P3, which spanned the A1 acceptor site of the first intron, protected bands of 320 and 131 nts in sizes (, lane 10). The protected band of 131 nts confirmed the position of the A1 acceptor site at nt 1630. Similar to what observed at the D1 donor site, only a few PARV4 mRNAs were spliced at the A1 acceptor site, suggesting the weakness of the splicing capability of the first intron spanning D1-A1 sites.
Probe 4, which spanned the P38 promoter and D3m donor site, protected bands of 377, 330, 111 and 64 nts in sizes (, lane 3). Probe 5, which spanned the P38 promoter, D3m and D2 donor sites, protected bands of 240, 211, 140, 111 and 64 nts in sizes (, lane 4). These bands confirmed the locations of the P38 promoter as well as the D3m and D2 donor sites at nt 1950, 2013 and 2060, respectively. Only ~5% of mRNAs generated from the P38 promoter were spliced at the D3m donor site, which resulted in a band of 64 nts. The majority of the P38-generated mRNAs were spliced at the D2 site, which produced a band of 111 nts.
Probe P6, which spanned the A2 acceptor and the D3 donor sites, protected bands of 233, 120, and 59 nts in sizes. These bands confirmed the presence of the PARV4 mRNAs spliced at the A2 acceptor site (nt 2309) and the D3 donor site (nt 2367) (, lane 5). A putative band of 172 nts, which refers to the transcripts that are spliced at the A2 site but not at the D3 donor site, was not detected, suggesting that the PARV4 mRNAs that were spliced at the A2 site were spliced concurrently at the D3 donor site,
Probe P7, which spanned the A3 acceptor site, protected two bands of 232 and 112 nts in sizes, respectively. Thus, the A3 acceptor site was mapped to nt 3329 (, lane 6). The ratio of unspliced vs. spliced PARV4 mRNA transcripts across this region was ~1:4.
Probe P8 was used to confirm the internal poly A signal (pA)p site (, lane 7). No band of ~170 nts in size was apparent, which suggests that the (pA)p site at nt 3771 detected by 3’ RACE is not significantly used for processing PARV4 pre-mRNA. Probe P9 protected a major band at ~183 nts, which was produced when PARV4 mRNAs were terminated at the (pA)d site, confirming the predominant polyadenylation site of PARV4 mRNA transcripts at nt 5149 (, lane 8).
Northern blot analysis of PARV4 mRNA transcripts
Hybridization of PARV4 mRNAs with the Cap probe (nt 3,841-5,200) detected four major bands (, lane 3), which were putatively NS1a-encoding R1a mRNA (~4.0 kb), NS1b-encoding R1b mRNA (~5.0 kb), VP1-encoding R3 mRNA (~3.2 kb), and VP2-encoding R4 and R5 mRNAs (~2.0 kb). The putative NS2-encoding R2 mRNA (~2.4 kb) had a much lower expression level and was not clearly visualized in either lane 2, where the NS probe (nt 283-560) was supposed to detect the R2 mRNA, or in lane 3, where the Cap probe was used, in . The NS probe was included to detect all transcripts generated from the P6 promoter; however, this probe only revealed a single major band, the R1a mRNA at ~4.0 kb. There were two smear areas above and below the R1a band, which potentially correspond to the R1b mRNA at ~5.0 kb and R2 (R2a+R2b) mRNAs at ~2.4 kb, respectively (, lane 2).
Analysis with the PARV4 genome probe (NSCap) showed an overall RNA profile generated from transfection of p5TRPARV4 (, lane 4). Five major species of mRNA transcripts were observed, which were consistent in size with those detected when using the Cap probe and predicted to encode NS1 (R1a and R1b), VP1 (R3), and VP2 (R4 and R5). The R4/R5 mRNAs were the most abundant transcripts and constituted ~80% of the total viral mRNAs.
Accordingly, we generated a genetic map of PARV4 as a summary of the results obtained from 5’/3’ RACE, RT-PCR, RPA, and Northern blot analyses (). The five major PARV4 mRNA transcripts and two minor transcripts (R2a and R2b) are depicted with their respective sizes, abundance by percentage, and encoded proteins.
Expression strategy of PARV4 proteins
We next examined protein expression of PARV4 by transfecting. To determine whether the NS1a or NS1b protein is expressed after transfection of p5TRPARV4 and pSKPARV4 in 293 cells, we expressed NS1a and NS1b individually as size controls (, lanes 5&6). We found that NS1a was clearly detected, but not NS1b, in both p5TRPARV4- and pSKPARV4-transfected cells (, lanes 2&4). The bands of both NS1a were slightly lower than the one as the size control, which are likely due to the uneven migrations of NS1a in each lane (). Although the predicted size of the NS1a protein is 66 kDa (), it appeared larger than 72 kDa (~80 kDa) in the blot, which was likely due to a modification of the NS1 protein (e.g., phosphorylation) (Nuesch et al., 1998
Western blot analysis of PARV4 proteins
The anti-N-NS1 antibody used for Western blot was produced against the N-terminus (aa 27-44) of the NS1 protein (), which is potentially able to detect both NS1 and NS2 proteins. However, we did not detect specific bands around 7 kDa, which should have corresponded to the putative NS2 protein, using a high percentage (15%) PAGE gel (data not shown). These results indicated that the NS2 protein is likely expressed at a very low level.
Using the anti-C-VP2 antibody raised against the C-terminus (aa 500-552) of the VP2 protein, we detected a major band at a molecular weight of ~65 kDa, which corresponded to the VP2 protein, in both p5TRPARV4 and pSKPARV4-transfected cells (, lanes 2&4). Notably, VP1, which has a predicted molecular weight of 101 kDa, was not detected. However, we observed a minor band around 90 kDa above the VP2 band, which could be the VP1 protein translated from the second AUG (nt 2531) or the third AUG (nt 2678) on the VP1-encoding R3 mRNA (). However, it is also possible that the band of ~90 kDa was a cleaved product of the VP1 protein of 101 kDa, since the capsid proteins of other parvoviruses have been reported to be cleaved (Farr et al., 2006
; Cheng et al., 2009
Both the PARV4 NS1a and NS1b proteins induce cell cycle arrest at G2/M phase
As shown in other genera of parvoviruses, the large nonstructural protein NS1 or Rep78 protein is able to induce cell cycle arrest (Chen and Qiu, 2010
), which is also a hallmark of B19V infection of primary erythroid progenitor cells (Wan et al., 2010
). Since PARV4 was previously detected in human bone marrow (Manning et al., 2007
; Longhi et al., 2007
), we examined the effect of PARV4 NS1 on the cell cycle progression of CD34+
HSCs. We first performed a sequence alignment among PARV4 NS1, B19V NS1, and AAV2 Rep78 proteins (). Highly conserved regions were identified in the helicase motifs (the four sites of Walker boxes, A, B, B’ and C sites (Jindal et al., 1994
; Walker et al., 1997
; Momoeda et al., 1994
)). Therefore, we introduced mutations in boxes A and B of both NS1a and NS1b to explore the potential role of the helicase motifs of NS1 in disturbing cell cycle progression.
The helicase motifs of the PARV4 NS1 are responsible for G2/M arrest induced by PARV4 NS1 protein
We transduced lentiviruses in CD34+
HSCs in order to express NS1a and NS1b. At 48 hrs post-transduction, we analyzed the cell cycle progression using flow cytometry. As shown in , an obvious G2/M arrest (p
< 0.01) was observed in both NS1a and NS1b-expressing CD34+
HSCs compared to RFP-expressing cells. Notably, both NS1a(mA+B) and NS1b(mA+B), which bear the mutations in boxes A+B, significantly rescued the G2/M arrest (p
< 0.01, ). The expression levels of NS1a, NS1b and their mutants in transduced cells were quantified at similar levels as shown with mean fluorescence intensity in . These results demonstrated that the PARV4 NS1a/b proteins are capable of inducing a G2/M arrest in NS1-expressing cells. Thus, these data reveal a remarkable feature of the PARV4 NS1, which is similar to that seen with B19V infection of primary erythroid progenitor cells (Wan et al., 2010
) and clearly involves the predicted helicase motifs.
The PARV4 VP1 unique region does not exhibit phospholipase A2-like activity in vitro
The VP1 unique region (VP1u) of parvoviruses functions like PLA2
with conserved motifs, which is critical for virus infection (Zadori et al., 2001
). We first aligned the PARV4 VP1u with that of B19V, and identified conserved PLA2
motifs (). To examine the PLA2
-like activity of the PARV4 VP1u, three forms of VP1u (1M-VP1u, 52M-VP1u, and 101M-VP1u) were purified (), which originated from three different methionines of the putative VP1u. This approach was taken because we had not determined the start AUG for translation of PARV4 VP1. Purified B19V VP1u and GST proteins were used as positive and negative controls, respectively. As seen in , both B19V VP1u and bee venom exhibited significant PLA2
-like activity (Dorsch et al., 2002
); however, none of the PARV4 VP1u proteins possessed PLA2
-like function in vitro
, implying that either the PARV4 VP1u requires a structural conformation for PLA2
-like activity or PLA2
-like activity is not required for PARV4 infection.
The PARV4 VP1 unique region does not exhibit phospholipase A2-like activity in vitro