Human Bocavirus 1 NP1 acts as an ssDNA-binding protein to help AAV2 DNA replication and cooperates with RPA to regulate AAV2 capsid expression

ABSTRACT

Adeno-associated virus (AAV) requires co-infection with helper virus for efficient replication. We previously reported that Human Bocavirus 1 (HBoV1) genes, including NP1, NS2, and BocaSR, were critical for AAV2 replication. Here, we first demonstrate the essential roles of the NP1 protein in AAV2 DNA replication and protein expression. We show that NP1 binds to single-strand DNA (ssDNA) at least 30 nucleotides (nt) in length in a sequence-independent manner. Furthermore, NP1 colocalized with the BrdU-labeled AAV2 DNA replication center, and the loss of the ssDNA-binding ability of NP1 by site-directed mutation completely abolished AAV2 DNA replication. We used affinity-tagged NP1 protein to identify host cellular proteins associated with NP1 in cells cotransfected with the HBoV1 helper genes and AAV2 duplex genome. Of the identified proteins, we demonstrate that NP1 directly binds to the DBD-F domain of the RPA70 subunit with a high affinity through the residues 101–121. By reconstituting the heterotrimer protein RPA in vitro using gel filtration, we demonstrate that NP1 physically associates with RPA to form a heterologous complex characterized by typical fast-on/fast-off kinetics. Following a dominant-negative strategy, we found that NP1-RPA complex mainly plays a role in expressing AAV2 capsid protein by enhancing the transcriptional activity of the p40 promoter. Our study revealed a novel mechanism by which HBoV1 NP1 protein supports AAV2 DNA replication and capsid protein expression through its ssDNA-binding ability and direct interaction with RPA, respectively.

IMPORTANCE

Recombinant adeno-associated virus (rAAV) vectors have been extensively used in clinical gene therapy strategies. However, a limitation of these gene therapy strategies is the efficient production of the required vectors, as AAV alone is replication-deficient in the host cells. HBoV1 provides the simplest AAV2 helper genes consisting of NP1, NS2, and BocaSR. An important question regarding the helper function of HBoV1 is whether it provides any direct function that supports AAV2 DNA replication and protein expression. Also of interest is how HBoV1 interplays with potential host factors to constitute a permissive environment for AAV2 replication. Our studies revealed that the multifunctional protein NP1 plays important roles in AAV2 DNA replication via its sequence-independent ssDNA-binding ability and in regulating AAV2 capsid protein expression by physically interacting with host protein RPA. Our findings present theoretical guidance for the future application of the HBoV1 helper genes in the rAAV vector production.

INTRODUCTION

Recombinant adeno-associated virus (rAAV) vectors are promising for clinical gene therapy strategies due to their excellent in vivo safety profile (1). Several rAAV-based gene therapeutics have reached the market, such as Luxturna (for the treatment of inherited retinal disease) and Zolgensma (for the treatment of spinal muscular atrophy) (2). As a replication-deficient virus, the productive replication of AAV must be completed with the assistance of helper viruses, which include adenovirus (Ad) and herpes simplex virus 1 (HSV-1) (3). However, host factors are also required for full AAV replication (4–6). For example, replication factor C (RFC), proliferating cell nuclear antigen (PCNA), and DNA polymerase δ promote AAV genome replication in vitro (6) and colocalize with the AAV replication center in vivo (5). Less is known about how host factors interplay with the helper virus to regulate AAV replication.

AAV, belonging to the genus Dependovirus of the family Parvoviridae, possesses an ssDNA genome of approximately 4.8 kilobases (kb) in length, flanked by two hairpin-shaped inverted terminal repeats (ITRs) (3). The ITRs provide free 3′-OH primers to initiate AAV DNA replication by a single-strand displacement mechanism, forming the replicative intermediates duplex-DNA that are covalently closed at one or both ends (7). The presence of duplex-DNA intermediates immediately triggers subsequent replication and transcription of the AAV genome (8, 9). The AAV genome consists of two open reading frames (ORFs) (termed rep and cap) encoding eight viral proteins regulated by three different promoters (p5, p19, and p40) (10–15). The rep gene encodes Rep78, Rep68, Rep52, and Rep40 proteins translated from p5 and p19 promoters drived-mRNAs with alternative splicing (10, 16). The Rep78 and Rep68 proteins, which exhibit helicase and endonuclease activities, are capable of catalyzing site-specific and strand-specific endonucleolytic cleavage at the terminal resolution site (TRS) of the covalently closed terminal sequence (17–22). The resulting linear duplex DNA end is then unwound to reform a terminal hairpin, thus providing a new 3′-OH primer for strand displacement synthesis (7). Rep52 and Rep40 proteins are mainly responsible for packaging the newly synthesized plus and minus single-stranded genomes into the preformed capsids with equal efficiency (23). The AAV capsid has icosahedral symmetry and consists of the proteins VP1, VP2, and VP3, present in a molar ratio of 1:1:10 (24–26). The capsid proteins encoded by the cap gene are translated from mRNAs transcribed by the p40 promoter (11, 27), which is activated by the Rep78/68 protein (28). A recently identified assembly activating protein (AAP) is also translated by p40 promoter-driven mRNA. AAP is mainly localized in the nucleus and serves as a scaffold for the nuclear import, proper folding, and proportional assembly of capsid proteins (29, 30).

Human Bocavirus 1 (HBoV1), belonging to the genus Bocaparvovirus in the Parvoviridae family, was recently identified as a helper virus for AAV2 (31). The HBoV1 genome encodes five non-structural proteins (NS1, NS2, NS3, NS4, and NP1) and three structural proteins (VP1, VP2, and VP3) by alternative splicing and alternative polyadenylation of the pre-mRNA transcribed from the p5 promoter (32–34). Using an intragenic pol III promoter, the HBoV1 genome at nucleotides 5,199–5,388 also encodes a noncoding RNA, BocaSR (bocavirus transcribed small noncoding RNA) (35). Although the exact molecular mechanism is unknown, we previously reported that the co-expression of NP1, NS4, and BocaSR is essential for the DNA replication and progeny virus production in HEK293 cells transfected with AAV2 duplex genome (31). We also found that the NS2 protein can functionally replace NS4 in supporting AAV2 replication. Furthermore, in AAV2 virus-infected HEK293 cells, NS2, but not NS4, together with NP1 and BocaSR, supported AAV2 replication (31). However, expressing HBoV1 helper genes (NP1, NS2, and BocaSR) is not as efficient as Ad helper genes (E2A, E4, and VA) in promoting rAAV vector production in HEK293 cells, largely due to the inefficient expression of AAV Rep52 and capsid proteins (28). Interestingly, adding the Ad E2A gene to the above HBoV1 helper genes significantly increased rAAV DNA replication and rAAV vector production to levels higher than Ad helper genes alone (28).

The highly conserved NP1 protein of the Bocavirus genus is a multifunctional protein. It localizes to the nucleus and promotes pre-mRNA maturation to ensure proper synthesis of capsid proteins (33, 36–39). NP1 does this by suppressing the cleavage and polyadenylation of RNAs at the viral internal polyadenylation site (pA)p and by facilitating splicing of the upstream intron adjacent to (pA)p. The critical role of NP1 interaction with CPSF6 in the maturation of VP-encoding mRNAs was also revealed (40). The available findings also demonstrated that NP1 is localized at the HBoV1 DNA replication centers and plays an enhancement role in viral DNA replication (38, 41–43), although the detailed mechanisms are unclear. Recently, NP1 was shown to interact with host factors Ku70 and RPA70, which are proteins involved in DNA repair, and facilitate HBoV1 DNA replication (44). What host factors are involved in regulating AAV2 replication through interaction with the NP1 protein in the HBoV1 helper circumstance needs further investigation.

In this study, we aimed to delineate how the HBoV1 NP1 protein exerts its helper function in supporting AAV2 replication. We find that NP1 can bind ssDNA in a sequence-independent manner, indispensable for efficient viral DNA replication in AAV2 duplex genome transfected HEK293 cells or in AAV2-infected HEK293 cells. Besides, we perform an affinity pull-down assay to identify host proteins that interact with NP1. Among the identified proteins, we detailed the direct interaction between NP1 and the host heterotrimeric protein RPA and characterized the binding interfaces in both proteins. Moreover, we found that NP1 and RPA not only form a heterologous complex in vitro but also have discrete co-localization in vivo and are both recruited to the AAV2 replication center. Significantly, cellular functional analyses performed by over-expressing RPA70 to competitively deconstruct the NP1-RPA complex reveal that the NP1-RPA interaction does not contribute to the replication of AAV2 DNA, but directs the expression of the capsid proteins by ensuring the transcriptional activity of the p40 promoter.

RESULTS

HBoV1 NP1 protein can bind to ssDNA and colocalize with the AAV2 replication center

It has been reported that HBoV1 NP1 protein localized in BrdU-labeled HBoV1 replication center, where the ssDNA was accumulated. However, NP1 did not interact with NS proteins (45), suggesting that NP1 may bind with ssDNA directly or interact with cellular proteins recruited to the viral replication center. We wondered whether the same is true for NP1 during AAV2 replication. Thus, we sought to purify the NP1 protein, combined with in vitro DNA pull-down assay, to probe the ability of NP1 to bind to ssDNA or double-strand DNA (dsDNA). By removing the N-terminal disordered dual nuclear localization signal (NLS) sequence (36) (Fig. 1A), the resulting His-tagged NP1ΔNLS construct was successfully purified and then subjected to a heparin column to eliminate any possible contaminating nucleic acids, which was monitored by the absorbance ratio between 260 nm and 280 nm. The gel filtration result showed that the purified protein was eluted with an estimated molecular weight of 162 kDa (Fig. 1B), indicating NP1ΔNLS forms a high-order oligomeric state in the solution. More importantly, although NP1ΔNLS did not enter the nucleus after expression in HEK293 cells but retained in the cytoplasm (Fig. S1A), it could be translocated into the nucleus upon co-expression with NS2 and BocaSR to initiate AAV2 replication (Fig. S1B). Western blot, Southern blot, and Q-PCR results demonstrated that NP1ΔNLS did not differ significantly from the full-length NP1 protein’s ability to promote AAV2 protein expression, DNA replication, and progeny virus production (Fig. S1C through F). Using in vitro pull-down assays, we found that only ssDNA but not dsDNA interacted with NP1ΔNLS (Fig. 1C), representing that NP1 can bind to ssDNA. Given that the used oligonucleotides were generated with random sequences, this result also suggests that NP1 binds to ssDNA in a sequence-independent manner. Notably, the length of ssDNA shorter than 30 nt did not exhibit binding to NP1ΔNLS, whereas the length of ssDNA over 30 nt also did not significantly affect its affinity to the protein (Fig. 1C).

Fig 1.

Confirmation of HBoV1 NP1 binding to ssDNA and colocalizing with the AAV2 replication center. (A) A schematic picture of NP1 and NP1ΔNLS mutant. The canonical NLS (cNLS, key sites highlighted in red font) and noncanonical NLS (ncNLS) in amino acids 7–50 of NP1 were highlighted in a blue square. For expression and purification of NP1, the residues 1–50 were truncated and added with a C-terminal 6 × His tag for purification. (B) Analytical gel filtration of purified NP1ΔNLS. The eluted fractions were subjected to SDS-PAGE analysis. The elution volume marker was shown on the top of the gel. The top-right corner presents a calibration curve for gel filtration using proteins of known sizes. The elution volume of NP1ΔNLS projected on the curve allows the determination of the corresponding molecular weights. (C) In vitro DNA pull-down assay, biotin-labeled dsDNA or ssDNA with different lengths was incubated with equal amounts of purified NP1ΔNLS protein. NP1ΔNLS bound with biotinylated DNAs were pelleted down using streptavidin beads, boiled resolved in SDS-PAGE gel, and stained with Coomassie blue (pull-down panel). Streptavidin beads without any DNA were used as negative control (pull-down panel). Input represents equal amounts of NP1ΔNLS used in each pull-down assay. (D) Representive IF images on HEK293 cells transfected with plasmids labeled on the left. At 48 h post-transfection, cells were labeled by incorporating BrdU into the newly synthesized viral ssDNA. The BrdU incorporated ssDNA were detected with anti-BrdU antibody and visualized using a secondary antibody conjugated with AF-594. The Myc-tagged proteins were probed with anti-Myc antibodies and visualized using a secondary antibody conjugated with AF-488. DAPI was used to identify nuclei. Scale bar, 10 µm.

To detect whether NP1 colocalizes with the AAV2 replication center, we transfected HEK293 cells with pIAAV2 plasmid in the presence or absence of HBoV1 helper genes. After pulse-labeling the cells with 5-Bromodeoxyuridinc (BrdU), we observed that larger sites of BrdU localization appeared, which colocalized with accumulated NP1 protein (Fig. 1D), suggesting that NP1 can bind the AAV2 genome. Notably, NP1ΔNLS also localized to BrdU-labeled AAV2 DNA replication sites (Fig. S1B), indicating that the NLS is dispensable for NP1 binding to the viral genome. The above data demonstrate that NP1 has a sequence-independent ssDNA binding ability and can localize to the AAV2 genome replication sites without the NLS.

ssDNA binding ability is essential for NP1 to support AAV2 DNA replication

As the minimal region of NP1 responsible for ssDNA-binding includes residues from 90 to 219 (NP1Δ90) (Fig. 2A, lane 3 vs lanes 7 and 8, and Table S1), and the R163Aand M165A double mutated variant of NP1ΔNLS lost the interaction with ssDNA (Fig. 2B, lane 5 vs lane 4, Fig. S2; Table S1), we wondered whether NP1 utilizes its ssDNA-binding ability to play a role in AAV2 genome replication. In NS2, BocaSR, and AAV2 duplex DNA transfected HEK293 cells, we compared co-transfections with wild-type NP1 to those with R163A and M165A mutant and found that the double mutated NP1 variant failed to initiate AAV2 DNA replication (Fig. 2C, lane 4 vs lane 3, and Fig. 2D). The expression of AAV2 capsid protein was simultaneously decreased in R163A and M165A mutant expressing cells (Fig. 2E, lane 4 vs lane 3), resulting from the less efficient AAV2 DNA replication. Consistently, the number of progeny virions detected by Q-PCR technology was also significantly lower in HEK293 cells transfected with R163A and M165 mutant than in HEK293 cells transfected with wild-type NP1 (Fig. 2F). Notably, the expression of AAV2 Rep proteins was not markedly affected (Fig. 2E), suggesting that the impaired replication of the AAV2 genome was indeed caused by the loss of the ability of R163A and M165A mutant to bind to ssDNA. Parallelly, the same experiments were conducted with AAV2-infected HEK293 cells, and similar results were observed. As shown in Fig. 2G through I, compared with wild-type NP1, the R163A and M165A mutant failed to support AAV2 DNA replication and capsid protein expression. The number of progeny virions produced by co-transfection with R163A and M165A mutant was lower than that produced by co-transfection with wild-type NP1 in HEK293 cells, which was at a level similar to the negative control (Fig. 2J). These results demonstrated that the ssDNA-binding ability of NP1 protein is indispensable for supporting efficient AAV2 DNA replication following transfection and during AAV2 infection.

Fig 2.

NP1 helps AAV2 replication depending on its ssDNA-binding ability. (A) Performing in vitro DNA pull-down assays to determine the minimal sequence for NP1 binding ssDNA. Streptavidin beads were pre-immobilized with 5′-biotin labeled ssDNA in 50 nt length, resuspended in buffers containing purified NP1 truncations, and rotated at 4°C for 2 h. Shown are Coomassie blue stained SDS-PAGE gels of input and bound samples. Streptavidin beads alone were loaded as a negative control. (B) Biotin-labeled dsDNA and ssDNA pull-down of NP1ΔNLS wild type or mutant. Shown are Coomassie blue stained SDS-PAGE gels of input and bound samples. Streptavidin beads alone were loaded as the negative control. (C–F) HEK293 cells were cotransfected with pIAAV2, NS2, BocaSR, and NP1 or its mutant. At 48 h post-transfection, the cells were harvested and subjected to Southern blot, Western blot, and real-time PCR analysis. (C) Southern blot analysis. Hirt DNAs were extracted, treated with DpnI, blotted on Nylon membrane, and hybridized with a DIG-labeled AAV2 probe. As indicated next to the image, the detected bands are dRF (dimer replicative form) DNA, mRF (monomer replicative form) DNA, and ssDNA (AAV2 single-strand genome). The 7.6 kb and 4.3 kb AAV2 DNA size markers and DpnI-digested DNA are also labeled. (D) The levels of mRF DNA were quantified with ImageJ software and the quantities obtained from two independent experiments are presented as relative levels to the control (lane 2, panel C). P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. **, P < 0.01. (E) Western blotting. Immunoblots were probed with anti-AAV2 Rep, anti-AAV2 VP, anti-HBoV1 NS4, and anti-Myc antibodies. β-actin was reprobed as a loading control. The detected proteins are indicated next to the images. (F) Quantification of progeny AAV2. Progeny AAV2 was quantified by real-time PCR. The virus production levels were shown as DNase-resistant particles (DRP) per cell. Error bars show standard deviations, which were obtained from three independent experiments. P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. ****, P < 0.0001. (G–J) The HEK293 cells were firstly infected with AAV2 and then transfected with NS2, BocaSR, NP1, or NP1 mutant at 8 h post-infection. At 48 h post-transfection, the cells were harvested for Southern blot, Western blot, and real-time PCR analysis. (G) Southern blot analysis. The Hirt DNA samples without DpnI-digesting were blotted on Nylon membrane, and hybridized with a DIG-labeled AAV2 probe. As indicated next to the image, the detected bands are dRF DNA, mRF DNA, ssDNA, and AAV2 DNA size markers. (H) The intensities of mRF DNA were quantified with ImageJ and the quantities obtained from two independent experiments are presented as relative levels to the control (lane 2, panel G). P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. **, P < 0.01. (I) Western blot analysis. AAV2 proteins were probed with anti-Rep, anti-VP antibodies, and HBoV1 proteins were probed with anti-NS4, anti-Myc antibodies. β-actin was subjected as a loading control. (J) Quantification of progeny AAV2. Progeny AAV2 was quantified by real-time PCR. The virus production levels were shown as DNase-resistant particles (DRP) per cell. Error bars show standard deviations, which were obtained from three independent experiments. P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. ****, P < 0.0001.

Next, we were interested in determining whether NP1Δ90 is as efficient as the full-length NP1 to help AAV2 DNA replication. Unexpectedly, the results showed that the GFP-tagged NP1Δ90 (Fig. 3A; Table S1) variant only slightly stimulated AAV2 DNA replication (Fig. 3B, lane 5 vs lanes 2 and 3, and Fig. 3C), which is likely due to the insufficient accumulation of this variant in the nucleus. Thus, we chose to retain the NLS sequence with GFP-NP1Δ90 to enable the resulting variant (GFP-NP1Δ51–90, Fig. 3A; Table S1) to distribute in the nucleus. As expected, the GFP-NP1Δ51–90 variant rescued AAV2 DNA replication completely (Fig. 3B, lane 4 vs lane 3, and Fig. 3C). However, in the absence of residues from 51 to 90, neither GFP-NP1Δ51–90 nor GFP-NP1Δ90 efficiently complement AAV2 capsid proteins expression, compared with wild-type NP1 (Fig. 3D, lanes 4 and 5 vs lane 3). We, therefore, concluded that the ssDNA-binding region is sufficient for AAV2 DNA replication when NP1 is imported into the nucleus. In support of this notion, expression of deletion mutants of NP1 (Fig. 3A; Table S1, Myc-NP1Δ60–70, Myc-NP1Δ70–80, and Myc-NP1Δ80–90) did not abrogate the helper function for AAV2 DNA replication (Fig. 3E, lanes 4–6 vs lane 3, and Fig. 3F), although lower expression levels of capsid proteins were observed (Fig. 3G, lanes 4–6 vs lane 3).

Fig 3.

Expression of the ssDNA-binding sequence of NP1 in the nucleus can support AAV2 genome replication. (A) Schematic illustration of GFP-tagged or Myc-tagged full-length NP1 and its truncations. The deleted residues were represented in gray dashed lines. (B and E) HEK293 cells were transfected with plasmids as indicated. Mock, without transfection. The Hirt DNA samples extracted from lysed cells were DpnI digested at 37° for 5 h and analyzed by Southern blotting with a digoxigenin-labeled AAV2 probe. dRF, mRF, ssDNA, and AAV2 DNA size markers are indicated. (C and F) The levels of mRF DNA were quantified with ImageJ and the quantities obtained from two independent experiments are presented as relative levels to the control. Lane 2 in panel (B) and lane 2 in panel (E) are controls for panels (C and F), respectively. P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. **, P < 0.01; ***, P < 0.001; n.s., no significance. (D and G) HEK293 cells were transfected with plasmids as indicated. Mock, without transfection. The lysates from transfected cells were analyzed by Western blotting using anti-AAV2 Rep, anti-AAV2 VP, anti-HBoV1 NS4, anti-GFP, and anti-Myc antibodies. The identities of the detected bands are indicated at the right of the blots. β-actin was reprobed as a loading control.

In summary, these results demonstrate that the ssDNA-binding ability is necessary and sufficient for NP1 to promote AAV2 DNA replication, while the residues 51–90 of NP1 play a role in expressing AAV2 capsid proteins.

Identification of host cellular factors interacting with NP1

We then asked whether host factors interact with NP1 to facilitate AAV2 replication. To this end, HEK293 cells cotransfected with pIAAV2, FLAG-NP1, NS2, and BocaSR plasmids were harvested and immunoprecipitated (IP) with anti-FLAG antibody or control IgG antibody. The precipitated samples were checked for protein composition using SDS-PAGE electrophoresis and stained with Coomassie blue, which revealed five protein bands in the anti-FLAG-NP1 group unique from those in the control IgG group (Fig. 4; Fig. S1 to S5, lane 3 vs lane 2). The results of the mass spectrum analysis of these five protein bands identified several proteins that potentially interact with FLAG-NP1 (Fig. 4B; Table S2). Interestingly, two subunits of the host heterotrimeric protein RPA, RPA70, and RPA32, were included with high confidence in the reservoir of interacting proteins (Fig. 4B), suggesting that AAV2 replication is linked to NP1-RPA interaction.

Fig 4.

Identification of host cellular proteins interacting with NP1. (A) FLAG-NP1 pull-down assay in the circumstance of AAV2 replication. HEK293 cells were cotransfected with pIAAV2, FLAG-NP1, NS2, and BocaSR expressing plasmids, as shown. At 48 h post-transfection, 90% of the cell lysates were immunoprecipitated with an anti-FLAG or control IgG, followed by separating on SDS-PAGE and staining with Coomassie blue. The bands (S1–S5) uniquely stained in the sample immunoprecipitated with anti-FLAG but not in the control IgG sample are denoted with stars and indicated next to the image. 10% of the lysed cells were loaded as input. (B) Putative NP1-interacting proteins that were identified by IP-mass spectrometry. Proteins identified in each band (S1–S5) pulled down with anti-Flag antibody are shown with protein names. (C) Schematic diagram of subunits RPA70, RPA32, and RPA14 that form the RPA complex. Domains are labeled and shown in colored rectangles. Protein interaction sites are lined in black, ssDNA-binding sites in cyan, and the trimerization core in orange-red. (D) Co-IP to determine the interaction between NP1 and RPA subunits. The plasmids used to transfect HEK293 cells are indicated. At 48 h post-transfection, the cells were lysed, mock-treated, or individually treated with different nucleases (DNase, RNase, and Benzonase) to eliminate possible nucleic acids mediated interactions. The treated lysate was immunoprecipitated with anti-Myc or control IgG antibodies. Western blot analysis of the immunoprecipitated materials was carried out using anti-RPA70, anti-RPA32, and anti-Myc antibodies. IgG-L represented the light chains of anti-Myc or normal IgG.

RPA is an abundant, highly conserved protein in eukaryotic cells (46). As shown in Fig. 4C, RPA, consisting of RPA70, RPA32, and RPA14 subunits, binds to ssDNA in subnanomolar affinity through four oligonucleotide/olig-osaccharide-binding (OB) folds that are referred to as DNA-binding domains (DBD) A, B, C, and D (47–50). By contrast, the DBD-F domain of RPA70 and the winged helix (WH) domain of RPA32 mainly mediate protein-protein interactions, which regulate the in vivo function of RPA in a complex manner (51). The DBD-E of the RPA14 does interacts with DBD-C and DBD-D to form the RPA heterotrimer (52). To examine the interaction of NP1 and RPA, we performed a Co-IP assay for Myc-NP1 and endogenous RPA70, RPA32, and RPA14. As shown in Fig. 4D, the Myc-tagged NP1 simultaneously precipitated the three RPA subunits and the addition of nucleases (DNase, RNase, or Benzonase) into the binding/wash buffer did not affect the interactions, suggesting that their associations are not dependent on the presence of nucleic acids (Fig. 4D, lane 3 vs lanes 4–6). These results demonstrated that host cellular protein RPA specifically interacts with NP1.

NP1 physically binds to the DBD-F domain of subunit RPA70 and forms a heterologous complex with RPA in vitro

Next, we tested if NP1 directly interacts with RPA subunit using GST pull-down assays and Bio-layer interferometry (BLI) technology. While the His-tagged NP1ΔNLS and GST-tagged RPA32/RPA14 subunit were successfully purified from Escherichia coli (E.coli) cells, the full-length RPA70 protein failed to be expressed in bacteria, and we thus divided its sequence into three different fragments (named RPA70-F, RPA70-AB, and RPA70-C) based on the defined functional domains (Fig. 4C). Besides, we also prepared the GST protein and tested its NP1ΔNLS-binding property.

The results of GST pull-down assays shown in Fig. 5A indicated that neither GST-RPA32 nor GST-RPA14 can retain NP1ΔNLS, which is in agreement with the measurement results of BLI (Fig. 6A). Importantly, although the association of GST-RPA70-AB or GST-RPA70-C with NP1ΔNLS was not observed (Fig. 5B and 6B), GST-RPA70-F associated with NP1ΔNLS (Fig. 5B) in typical slow-on/slow-off kinetics (K_on = 4.95 × 10⁵ M⁻¹ s⁻¹, K_off = 1.99 × 10⁻³ s⁻¹), and the equilibrium dissociation constant (K_D) were determined to be 4.02 nM (Fig. 6C and E). In addition, the fact that NP1ΔNLS coeluted with different RPA70-F truncations (Table S3) from the Superdex-200 column further supports our result (Fig. S3). Therefore, these experiments suggest that NP1 could physically interact with RPA70 by binding to the DBD-F domain. To define the region of NP1 that is associated with GST-RPA70-F, we used additional GST pull-down assays and examined the binding of GST-RPA70-F to a series of NP1 fragments (Table S3). By comparing with the GST control (Fig. 5C), these pull-down assays indicated that amino acids 101–121 are critical for NP1 to bind to GST-RPA70-F (Fig. 5D and E).

Fig 5.

NP1 forms a heterologous complex with RPA in vitro by directly binding to the DBD-F domain of subunit RPA70. (A and B) GST pull-down assays were performed with GST-RPA32, GST-RPA14 GST-RPA70-F, GST-RPA70-AB, GST-RPA70-C, or GST, and purified NP1ΔNLS. Coomassie blue stained SDS-PAGE gels of purified proteins used (left) and bound samples (right) are shown. (C and D) GST pull-down assays were performed with GST (C) or GST-RPA70-F (D) and purified NP1 truncations. Coomassie blue stained SDS-PAGE gels of purified proteins used (left) and bound samples (right) are shown. (E) The amount of NP1 truncations retained was expressed relative to the amount of GST-RPA70-F in the bound sample and then normalized to the amount of NP1ΔNLS. All values are presented as mean ± SD, derived from three independent experiments. (F) Superposed gel filtration chromatograms of RPA and NP1ΔNLS-RPA complex show the respective elution volumes. The right panel shows SDS-PAGE gels of peaks from gel filtration.

Fig 6.

The binding profiles of interactions between GST-tagged proteins and purified NP1ΔNLS or RPA via BLI. (A) NP1ΔNLS binds to immobilized GST-RPA32, GST-RPA14, or GST. (B) NP1ΔNLS binds to immobilized GST-RPA70-F/AB/C or GST. (C) A series of diluted NP1ΔNLS binds to GST-RPA70-F immobilized on the sensor. The concentration used for binding evaluation is listed. (D) RPA reconstituted in vitro binds to immobilized GST-NP1ΔNLS. The concentration used for binding evaluation is listed. (E) Kinetic binding parameters of protein interactions measured using BLI assays. The K_D values were calculated using a 1:1 binding model.

We also examined whether NP1ΔNLS interacts with the RPA heterotrimer by gel filtration chromatography. The RPA protein was expressed and purified as previously described (52). As shown in Fig. 5F, a stable heterologous complex composed of NP1ΔNLS and RPA was present in the gel filtration, indicating that NP1 directly interacts with RPA. To further support this conclusion, we determined the real-time binding kinetics between NP1ΔNLS and RPA via BLI. The observed profile revealed typical fast-on/fast-off binding characteristics (K_on = 6.19 × 10⁵ M⁻¹ s⁻¹, K_off = 2.31 × 10⁻¹ s⁻¹), and the affinity (K_D) of the association was calculated to be 373 nM (Fig. 6D and E). In conclusion, these biochemical experiments indicate that NP1 interacts with the DBD-F domain of RPA70 through a small fragment involving residues 101–121 and further associates with RPA via the RPA70 subunit to form a heterologous complex.

NP1-RPA complex located at the AAV2 replication center

Since NP1 directly interacts with RPA in vitro, we set out to study whether NP1 colocalizes with RPA in cells. First, HEK293 cells transiently expressing Myc-tagged NP1 were used to examine the subcellular localization of NP1 and endogenous RPA by confocal microscopy. The NP1 was detected with an antibody against the Myc-tag, while the trimeric RPA was imaged by immunostaining with an antibody against the RPA70 subunit. We observed that NP1 and RPA were distributed throughout the nucleus and had a clear co-localization (Fig. 7A). Next, we monitored the co-localization of NP1 and RPA in HEK293 cells during the productive replication of AAV2. The results showed that NP1 largely colocalizes with RPA at discrete sites within the nucleus (Fig. 7B), suggesting that viral replication could alter the distribution pattern of NP1 and RPA in the nucleus. Finally, given that NP1 is localized in the AAV2 replication center (Fig. 1D; Fig. S1B) and that RPA has been reported to be a component of AAV replication center formed by transfection of minimal helper genes from Ad or HSV-1 (5, 53), we investigated the co-localization of RPA with the AAV2 replication centers in which HBoV1 genes supply helper activity. We employed the Rep78 protein to represent the viral replication centers. In pIAAV2, we inserted a FLAG-tag after the first methionine of the Rep78 and used an anti-FLAG antibody to detect the localization of Rep78 in the cells. Similar to NP1, we observed a specific pattern of co-localization of the anti-RPA70 staining with the anti-FLAG antibody staining in AAV2 dsDNA transfected cells (Fig. 7C), indicating that RPA localizes within the viral replication centers maintained by the HBoV1 helper system.

Fig 7.

NP1 colocalizes with RPA in vivo, and both recruit to the AAV2 replication center. (A and B) Representive IF images on HEK293 cells transfected with plasmids expressing proteins as indicated. The Myc-tagged NP1 protein was detected with an anti-Myc antibody and visualized using an AF-488 conjugated donkey anti-mouse secondary antibody. The endogenous expressed RPA was probed with an anti-RPA70 antibody and visualized using an AF-594 conjugated donkey anti-rabbit secondary antibody. DAPI was used to identify nuclei. Scale bar, 10 µm. (C) Representive IF images on HEK293 cells transfected with plasmids expressing proteins as indicated. The Myc-tagged NP1 protein or endogenous expressed RPA protein was detected with an anti-Myc antibody or anti-RPA70 antibody and visualized using AF-594 conjugated donkey anti-mouse or donkey anti-rabbit secondary antibodies. The FLAG-tagged Rep78 protein, representing the AAV2 replication center, was probed with an anti-FLAG antibody and visualized using an AF-488 conjugated donkey anti-mouse secondary antibody. DAPI was used to identify nuclei. Scale bar, 10 µm.

Collectively, we confirmed that NP1 and RPA not only bind together to form a stable complex in vitro but also have significant co-localization in vivo, and both are recruited into the viral replication center during AAV2 productive replication, supporting the idea that NP1-RPA complex may play a role in the AAV2 life cycle.

AAV2 capsid protein levels were downregulated upon over-expression of the RPA70 subunit

To further prove the function of the interaction between NP1 and RPA in AAV2 replication, we constructed an NP1 mutant (NP1Δ101–121) in which the DBD-F binding region (residues from 101 to 121, Fig. 5D) was deleted and looked at its impact on viral replication. The results showed that the number of progeny virions remarkably decreased when NP1Δ101–121 was present, compared to wild-type NP1 (Fig. 8A). Western blot analysis revealed that VP1 to VP3 proteins were expressed in wild-type NP1-expressing HEK293 cells but not in NP1Δ101–121-expressing or control HEK293 cells, whereas Rep78 and Rep52 proteins were not affected (Fig. 8B). Southern blotting of low-molecular-weight DNA (Hirt DNA) samples extracted from transfected HEK293 cells showed that NP1Δ101–121 did not rescue the level of the replicative form DNA of AAV2 to that in cells supplemented with full-length NP1 (Fig. 8C and D). Accordingly, substituting NP1 with NP1Δ101–121 in cells to deconstruct the NP1-RPA complex may disrupt AAV2 DNA replication, leading to reduced expression of capsid proteins and lower production of viral particles. Paradoxically, we could not demonstrate whether removing the DBD-F bound region would affect the ssDNA binding capacity of NP1 and, thus, the replication of the AAV2 genome.

Fig 8.

A deletion mutant of NP1 that does not interact with RPA70-F in vitro loses its function in promoting AAV2 replication in vivo. (A) Quantification of progeny AAV2. At 48 h post-transfection, the cells were collected, lysed, and quantified for DRP by real-time PCR. The virus production levels are shown as DRP per cell. Error bars show standard deviations, which were obtained from three independent experiments. P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. ***, P < 0.001. (B) Comparison of AAV2 Rep and capsid proteins in HEK293 cells transfected with NP1 wild-type or deletion mutant (NP1Δ101–121) plasmid with NS2, BocaSR, and pIAAV2. Cells were lysed and subjected to Western blotting after 48 h transfection. (C) Southern blot analysis of AAV2 genome replication using DpnI-digested Hirt DNA extracted from cells transfected with NP1 wild-type or deletion mutant (NP1Δ101–121). Cotransfected plasmids are shown, and dRF, mRF, ssDNA, and AAV2 DNA size markers are indicated. (D) The levels of mRF DNA were quantified with ImageJ and the quantities obtained from two independent experiments are presented as relative levels to the control (lane 2, panel C). P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. ****, P < 0.0001.

Consequently, we attempted to apply a dominant-negative strategy to induce the over-expression of RPA70 in HEK293 cells to look for an influence on AAV2 replication. However, the confocal microscopy revealed that the recombinant ^FLAGRPA70 protein was retained in the cytoplasm rather than colocalized with NP1 in the nucleus (Fig. 9A). To address this issue, we added an NLS tail to the ^FLAGRPA70 coding sequence to facilitate the translocation of ^FLAGRPA70 into the nucleus. The resulting ^FLAGRPA70-NLS overlapped well with NP1 within the nucleus (Fig. 9B). By cotransfecting plasmids expressing ^FLAGRPA70-NLS with NP1 at 1:1, 1:3, and 1:5 ratios into HEK293 cells, we observed that the levels of both AAV2 viral release and capsid proteins were gradually decreased compared to the non-^FLAGRPA70-NLS expressing cells (Fig. 9C and D, lane 3 vs lanes 4–6). Astonishingly, the results of Southern blot analysis indicate that the viral DNA replication was unaffected (Fig. 9E and F). More importantly, similar results were also obtained in AAV2-infected HEK293 cells, with less release of progeny virions, declined expression levels of capsid proteins, and intact DNA replication (Fig. 9H through K). Meanwhile, we noted a reduction in newly synthesized ssDNA when ^FLAGRPA70-NLS was increasingly expressed either in AAV2 duplex genome transfected HEK293 cells or in AAV2-infected HEK293 cells (Fig. 9E, G, J and L), most probably due to the shortage of viral capsids. As a control, ^FLAGRPA32-NLS showed strong nuclear co-localization with NP1 (Fig. S4A), but increasing its expression did not affect AAV2 protein expression and DNA replication (Fig. S4B through I), confirming that the potential negative effects of ^FLAGRPA70-NLS over-expression did not cause the results shown in Fig. 9. These results strongly demonstrated that the function of the NP1-RPA interaction is to regulate the expression of AAV2 capsid protein, but not to engage in viral DNA replication.

Fig 9.

AAV2 capsid protein levels were downregulated in HEK293 cells over-expressing the RPA70 subunit. (A and B) Representive IF images on HEK293 cells transfected with plasmids expressing proteins as indicated. The FLAG-tagged RPA70 or RPA70-NLS protein was detected with an anti-FLAG antibody and visualized using an AF-594 conjugated donkey anti-mouse secondary antibody. The NP1 protein was localized with the GFP tag. DAPI was used to identify nuclei. Scale bar, 10 µm. (C–E) HEK293 cells were cotransfected with pIAAV2 and different combinations of Bocavirus Helper genes along with increasing amounts of ^FLAGRPA70-NLS plasmid in the full Bocavirus Helper groups. At 48 h post-transfection, the cells were harvested and subjected to real-time PCR, western blot, and Southern blot to analyze AAV2 progeny virus production, protein expression, and genome replication. (C) Quantification of progeny AAV2. The virus production levels are shown as DRP per cell. Error bars show standard deviations, which were obtained from three independent experiments. P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. ****, P < 0.0001. (D) Western blotting. AAV2 proteins were probed with anti-Rep and anti-VP antibodies; HBoV1 NS2 was probed with an anti-NS4 antibody; NP1 was probed with an anti-Myc tag antibody; ^FLAGRPA70-NLS was probed with an anti-FLAG tag antibody. β-actin was reprobed as a loading control. (E) Southern blot analysis. The Hirt DNA was extracted, treated with DpnI, and subjected to Southern blotting with a DIG-labeled AAV2 probe. dRF DNA, mRF DNA, ssDNA, DpnI-digested DNA, and AAV2 DNA size markers are indicated. (F) The intensities of mRF DNA (lanes 3–6, panel E) were quantified with ImageJ and the quantities obtained from two independent experiments are presented as relative levels to the control (lane 3, panel E). P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. n.s., no significance. (G) The levels of viral ssDNA detected in panel (E), lanes 3–6 were quantified using ImageJ and values obtained from two independent experiments are presented as relative levels to the control (lane 3, panel E), which was set to 100%. *, P < 0.05; **, P < 0.01. (H–J) HEK293 cells were firstly infected with AAV2, 8 h post-infection, the cells were transfected with different combinations of Bocavirus Helper genes along with increasing amounts of ^FLAGRPA70-NLS plasmid in the full Bocavirus Helper groups. At 48 h post-transfection, the cells were harvested and subjected to real-time PCR, western blot, and Southern blot to analyze AAV2 progeny virus production, protein expression, and genome replication. (H) Quantification of progeny AAV2. The virus production levels are shown as DRP per cell. Error bars show standard deviations, which were obtained from three independent experiments. P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. ****, P < 0.0001. (I) Western blot analysis. AAV2 proteins were probed with anti-Rep and anti-VP antibodies; HBoV1 NS2 was probed with an anti-NS4 antibody; NP1 was probed with an anti-Myc tag antibody; ^FLAGRPA70-NLS was probed with an anti-FLAG tag antibody. β-actin was subjected as a loading control. (J) Southern blotting for AAV2 DNA replication. The Hirt DNA samples were extracted and directly analyzed by Southern blotting without DpnI-digesting. The blot was hybridized with DIG-labeled AAV2 probe. dRF DNA, mRF DNA, ssDNA, and AAV2 DNA size markers are indicated. (K) The intensities of mRF DNA (lanes 3–6, panel J) were quantified with ImageJ and the quantities obtained from two independent experiments are presented as relative levels to the control (lane 3, panel J). P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. n.s., no significance. (L) The levels of viral ssDNA detected in panel (J), lanes 3–6 were quantified using ImageJ and values obtained from two independent experiments are presented as relative levels to the control (lane 3, panel J), which was set to 100%. *, P < 0.05; **, P < 0.01.

Over-expression of RPA70 reduces the AAV2 p40 transcriptional activity

Next, we aimed to investigate the underlying mechanisms involved in the regulatory role of the NP1-RPA complex in expressing AAV2 capsid protein. Considering that NP1 has been reported to be required for the splicing of HBoV1 cap gene mRNAs, thereby facilitating the expression of HBoV1 capsid protein (33, 39). We first examined whether the NP1-RPA complex is implicated in the alternative splicing process of the AAV2 cap gene. For this purpose, a plasmid called pRep-GFP was constructed where the AAV2 rep genes were expressed under their native p5 and p19 promoters. Still, the cap gene was replaced by the GFP ORF, so that the intron acceptor A2 site was omitted, while the p40 promoter remained in situ (Fig. 10A). However, although the alternative splice sites in the cap gene were lost with the replacement GFP coding sequence, increasing expression of ^FLAGRPA70-NLS still dramatically reduced the levels of GFP proteins (Fig. 10B). We further quantified the mRNA of GFP in the condition of ^FLAGRPA70-NLS over-expression. As shown in Fig. 10C, ^FLAGRPA70-NLS suppressed GFP expression at the mRNA level. These results indicate that the NP1-RPA complex does not regulate AAV2 capsid protein levels via the alternative splicing process.

Fig 10.

Over-expression of RPA70 subunit inhibited the AAV2 p40 promoter in HEK293 cells. (A, D, and F) Plasmids pRep-GFP (A), pRep-GFP-Fluc (D), pAAV-GFP-Fluc (F), or pRep (F) are schematically illustrated. (B) Over-expressing ^FLAGRPA70-NLS reduces the level of GFP protein but not AAV2 Rep proteins. HEK293 cells were cotransfected with plasmids as indicated and lysed after 48 h transfection, followed by Western blotting using appropriate antibodies. (C) Over-expressed ^FLAGRPA70-NLS inhibited GFP mRNA transcription. HEK293 cells were transfected with plasmids as indicated. At 48 h post-transfection, the cells were harvested and total RNAs were purified, converted to cDNA, quantified by real-time PCR and normalized to GAPDH. P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. ***, P < 0.001; ****, P < 0.0001. (E) Over-expressed ^FLAGRPA70-NLS inhibited p40 promoter-driven Luciferase activity. HEK293 cells were cotransfected with pRep-Fluc, different combinations of Bocavirus Helper genes, and increasing amounts of ^FLAGRPA70-NLS in the full Bocavirus Helper groups. At 48 h post-transfection, cells were harvested for luciferase activity measurement. P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. ****, P < 0.0001. (G) Southern blot analysis of rAAV2 DNA replication. HEK293 cells were cotransfected pAAV-GFP-Fluc, pRep, and different combination of Bocavirus Helper genes along with increasing amounts of ^FLAGRPA70-NLS. The Hirt DNAs were extracted, treated with DpnI, and were subjected to Southern blot analysis with a DIG-labeled GFP probe. dRF DNA, mRF DNA, DpnI-digested DNA, and AAV2 DNA size markers are indicated. (H) The levels of mRF DNA were quantified with ImageJ and the quantities obtained from two independent experiments are presented as relative levels to the control (lane 3, panel G). P values were determined by ordinary one-way ANOVA with Tukey’s multiple-comparisons test. n.s., no significance. (I) Western blotting. HEK293 cells were cotransfected with plasmids as the same as in Fig. 9G. Cells were collected, lysed, and subjected to Western blot analysis using appropriate antibodies.

Since the transcription of the AAV2 cap gene is under the control of the p40 promoter (11, 27), we subsequently sought to perform a luciferase reporter assay to evaluate whether the transcriptional activity of the p40 promoter was compromised in HEK293 cells expressing ^FLAGRPA70-NLS. We constructed a plasmid pRep-Fluc (Fig. 10D), in which the luciferase ORF immediately followed the AAV2 rep genes and was under the control of the p40 promoter, like the cap gene. As expected, specific repression of luciferase activity was observed in ^FLAGRPA70-NLS-expression cells compared with the control cells that did not express ^FLAGRPA70-NLS (Fig. 10E), indicating that the p40 promoter was poorly transactivated. Taken all these results together, we concluded that the NP1-RPA complex regulates AAV2 capsid protein expression by governing the p40 promoter transcriptional activity.

In particular, to further verify that the NP1-RPA interaction was only for the AAV2 capsid protein and not for the viral DNA, we separately assessed viral DNA replication by constructing a rAAV2 plasmid pAAV-GFP-Fluc harboring GFP and luciferase expression cassettes flanked by two AAV2 ITRs (Fig. 10F). The rAAV2 plasmid was cotransfected with AAV2 rep and HBoV1 helper genes along with varying amounts of ^FLAGRPA70-NLS into HEK293 cells, and the rAAV2 DNA replication was examined by Southern blotting. The results showed that co-expression of NS2 and BocaSR did not initiate rAAV2 genome replication (Fig. 10G, lane 2), although it stimulated AAV2 Rep 78 and Rep 52 expression (Fig. 10I, lane 2). Adding NP1 to make an entire HBoV1 helper efficiently promoted rAAV2 DNA replication (Fig. 10G, lane 3). Unsurprisingly, increased ^FLAGRPA70-NLS did not impede rAAV2 DNA replication (Fig. 10G, lane 3 vs lanes 4–6, and Fig. 10H). Parallelly, ^FLAGRPA70-NLS over-expression did not affect AAV2 Rep proteins and NP1, NS2 protein expression (Fig. 10I, lane 3 vs lanes 4–6). Therefore, these results demonstrated that NP1-RPA interaction is not required for replicating the AAV2 genome.

DISCUSSION

In this study, we precisely dissected the role of NP1 in supporting AAV2 replication. We demonstrated that NP1, through its ssDNA binding ability and interaction with the host factor RPA, controls AAV2 DNA replication and capsid protein expression (Fig. 11). Our study provides a unique example of how a small viral non-structural protein facilitates the multifaceted regulation of AAV2 replication.

Fig 11.

Model showing how NP1 and NP1-RPA complex implicated in AAV2 replication. In the initial stage of AAV2 replication, the presence of the HBoV1 NP1 protein ensures efficient AAV2 DNA replication by the single-strand displacement mechanism. Upon the formation of the duplex-DNA intermediates, the rapidly assembled NP1-RPA complex is enriched to guide the activation of the p40 promoter to ensure the synthesis of the capsid protein.

Previous studies have found that the important role of the NP1 protein in viral DNA replication is conserved in MVC, BPV1, and HBoV1 (34). In HBoV1, NP1 is thought to colocalize with NS proteins (NS1 to NS4) within the viral replication centers and to be directly involved in viral DNA replication at OriR (45). However, the direct interaction between NP1 and NS proteins was not detected, and the underlying mechanism for the recruitment of NP1 remains largely unknown. Here, together with the results that NP1 has the ssDNA-binding ability and punctually colocalizes with the AAV2 DNA in the nucleus, we infer that NP1 is capable of direct binding to the AAV2 genome and protecting the single-stranded template DNA or the single-stranded progeny DNA from random nicking or from binding by other proteins that might interfere with DNA replication, and more studies are required to clarify this.

Although AAV does not encode its ssDBP, the role of helper virus-originated ssDBP in AAV DNA replication, such as E2A and ICP8 encoded by Ad and HSV-1, respectively, has been reported (5, 53–59). E2A and ICP8 overlap with the AAV Rep protein within the viral replication center (53, 57, 60). In an in vitro assay, adding Ad E2A or HSV-1 ICP8 enhances the processive replication of AAV2 DNA (53, 54). Moreover, either E2A or ICP8 can physically interact with the purified AAV Rep78/68 protein to increase the binding of Rep78/68 to the AAV replication origin site and stimulate the endonuclease activity of Rep78/68 (53), thereby improving their cleavage efficiency at the TRS sites and promoting AAV DNA replication in vitro. However, several groups have argued that E2A is more important for AAV rep and cap gene expression because mutations in the E2A gene, including gene deletion, have a modest effect on DNA replication (58, 59). Here, we have directly linked the ssDNA-binding ability of NP1 with the sufficient replication of AAV2 DNA through biochemical and cellular functional experiments. Our data indicate that the ssDNA-binding activity of NP1 is indispensable for AAV2 DNA replication and expression of the ssDNA-binding patch of NP1 in the nucleus can fully rescue the viral genome replication. Intriguingly, while the residues 51–90 are not responsible for the AAV2 DNA replication, they are associated with capsid protein expression (Fig. 3D and F), suggesting a multifunctional role of NP1 in AAV2 replication. Our results also raised some interesting questions about this HBoV1-encoded ssDBP: First, does NP1 bind the ITRs, the only cis-acting elements required for AAV genome replication? Second, should NP1, like E2A and ICP8, interact with Rep proteins to promote AAV DNA replication? Answering these questions will be interesting directions for future studies.

We demonstrate that NP1 binds directly to the DBD-F domain of RPA70, which contributes little to the binding of RPA70 to ssDNA, but is more engaged in the protein-protein interaction with other factors that regulate in vivo function of RPA (51). Combined with the results of the co-elution of NP1 with RPA from the gel filtration column, we conclude that NP1 is physically associated with RPA by binding to the subunit RPA70 DBD-F domain. RPA32 and RPA14 subunits are not involved in the interaction between NP1 and RPA since no obvious retention or binding of NP1 was observed in the GST pull-down and BLI assay. However, we cannot exclude the possibility that NP1 interacts with RPA32 or RPA14, as the conformational change of RPA upon NP1 binding may bring these two subunits to a more suitable position that interacts with NP1.

RPA is a central ssDBP in eukaryotic cells and has been implicated in diverse roles ranging from DNA replication, repair, and recombination to the regulation of gene transcription (61–65). The requirement for RPA in AAV DNA replication has been controversial. RPA is localized to the AAV replication center and is shown in vitro to interact physically with Rep78/68 protein to enhance binding and nicking at replication origin sites (53). Moreover, in an established in vitro AAV DNA replication system using crude extracts from the uninfected cells, the addition of RPA can support the processive replication of AAV DNA (54, 66). In a later study, however, the authors purified the components of the in vitro replication reaction and found that RPA was no longer required (6). We observed that both NP1 and RPA are recruited to the AAV replication center. Notably, NP1 binds RPA with high affinity, potentially promoting efficient recruitment. The co-localization of Rep78 protein with NP1 and RPA suggests that the interaction between the NP1-RPA complex and Rep78 is possible and requires further validation. Significantly, we found that over-expression of ^FLAGRPA70-NLS, which competes with endogenous RPA for binding to NP1, attenuated the transcriptional activity of the p40 promoter and resulted in decreased levels of capsid proteins, but viral DNA replication was not abolished, strongly suggesting that RPA is dispensable for AAV2 genome replication. However, we do not think that RPA could promote AAV2 DNA replication without NP1, as no AAV2 DNA replicative forms were detected when HEK293 cells were transfected with NS2 and BocaSR.

The fact that the NP1-RPA complex affects the transcription of the AAV2 cap gene is unexpected but reasonable. First, while NP1 plays a pivotal role in viral DNA replication (38, 41–44), this multifunctional protein has been reported to process viral pre-mRNA, regulate HBoV1 NS proteins expression and interact with host factor CPSF6 to facilitate (pA)p site-reading to enable full-length VP-encoding transcripts (33, 37–40). Second, accumulating evidence suggests that RPA is an important player in transcription. RPA70 binds to ssDNA generated by the transcription machinery, such as R loops and transcription bubbles, and can interact with RNA polymerase II and the activation domains of the transcription factors (61, 63, 64, 67–69). The involvement of RPA70 in the transcriptional activation of the heat shock factor protein and BRCA1 has been reported (67, 70). In contrast, the transcriptional repression role of RPA70 in metallothionein-IIA, endothelial nitric oxide synthase, and ARE-NRE-containing genes has also been revealed (63, 71, 72). The essential role of RPA70 in the chromatin-remodeling-based regulation of lipid oxidation-related gene transcription was recently demonstrated (65). Although the detailed role of the NP1-RPA complex in regulating the activity of the p40 promoter was not further explored here, we speculate that it may be related to the Rep78 protein. First, the colocalization of NP1 or RPA with Rep 78 in vivo and the results that Rep78 was immunoprecipitated simultaneously with RPA by Myc-NP1 (data not shown) strongly suggest that the NP1-RPA complex may further associated with Rep78 to form a triplet complex. Since Rep78 can activate the p40 promoter, we think that the interaction of NP1-RPA with Rep78 may enhance this activation, that is, RPA stimulates the ATP-dependent helicase activity of Rep78 to unwind the replicative intermediate duplex-DNA to form a regional R-loop-like ssDNA, which is immediately stabilized by the ssDNA binding protein NP1, RPA or NP1-RPA complex, for the subsequent use of transcription. However, this hypothesis remains to be demonstrated.

Our work expands the current understanding of the role of HBoV1 NP1 protein in AAV DNA replication. We discovered the critical role of the NP1-RPA complex in controlling capsid protein synthesis, and these results pave the way for developing a more efficient rAAV production system.

MATERIALS AND METHODS

Plasmid construction

AAV2 related plasmids

pIAAV2, which contains a full-length AAV2 genome, was constructed in our lab. The pAAV-MCS vector containing the AAV2 ITR (Agilent Technologies) was digested with NotI restriction enzyme, treated with alkaline phosphatase, and followed by gel purification. The pAAV-RC2 (Agilent Technologies) plasmid was used as the template to amplify AAV2 Rep and Cap genes. The complete P5 promoter sequence and AAV2 poly(A) sequence, plus the junction sequence between poly(A) and ITR, were synthesized by Sangon Biotech. The three DNA elements were assembled by overlapping PCR and cloned into a digested pAAV-MCS vector. The plasmid integrity was verified by Sanger sequencing and SmaI enzyme digestion. pIAAV2^FLAG was constructed by inserting sequence (DYKDDDD) into the vector’s backbone via in-fusion cloning.

pRep-GFP, which remains the rep ORF and p5, p19, p40 promoters, and AAV2 poly(A) signal sequence was constructed via a two-step cloning strategy. First, cap ORF in the pIAAV2 plasmid was replaced with GFP coding sequence to construct pAAV2-GFP plasmid. Second, the p5-Rep-GFP-Poly(A) segment was PCR amplified and cloned into a pSK vector to create pRep-GFP.

pRep-Fluc, which remains the rep ORF and p5, p19, p40 promoters and AAV2 polyA signal, was constructed by replacing the GFP ORF in the pRep-GFP with luciferase coding sequence via in-fusion cloning.

pRep and pAAV-GFP-Fluc were constructed as described previously (28).

HBoV1 gene-related plasmids

PCR-amplified NP1 is cloned into pcDNA3.1-FLAG, pcDNA3.1-Myc, or pEGFP-C2. Mutations or deletions are generated following the site-mutagenesis PCR protocol as described (73).

The plasmid expressing NS2 protein or the noncoding BocaSR was constructed as previously described (31).

Cellular gene-related plasmids

The DNAs coding for RPA70 and RPA32 were amplified from cDNAs generated by reverse-transcription of the RNA samples extracted from HEK293 cells and inserted into pcDNA3.1-FLAG vector via the BamHI and XhoI sites. The NLS sequence (KRPAATKKAGQAKKKK) was added to the C-terminal of FLAG-tagged RPA70 or RPA32 by in-fusion cloning.

Bacterial expression plasmids

The vector pET-21B or pGEX-6p-1 was employed to express His-tagged or GST-tagged NP1ΔNLS protein. The constructs used to express NP1ΔNLS mutants were generated following the site-mutagenesis PCR protocol as described (73).

Antibodies

The anti-HBoV1 NS4 antibody was homemade and has been described previously (74). The purchased antibodies used in this study were listed: mouse anti-AAV2 Rep (#03-65169, American Research Products), mouse anti-AAV2 VP (#03-65158, American Research Products), rabbit anti-BrdU (#600-40-1-C29, ROCKLAND), mouse anti-FLAG-tag (#F1804, Sigma-Aldrich), mouse anti-β-actin (#A00702-100, GenScript), mouse anti-GFP (#66002-1-lg, Proteintech), rabbit anti-RPA70 (#ab79398, Abcam), rabbit anti-RPA32 (#10412-1-AP, Proteintech), HRP-conjugated anti-rabbit IgG (#A00098, GenScript), HRP-conjugated anti-mouse IgG (#115-035-174, Jackson ImmunoResearch), mouse anti-Myc-tag (#SC-40, Santa Cruz), Donkey anti-mouse AF594 (#715-585-151, Jackson ImmunoResearch), Donkey anti-rabbit AF594 (#711-585-152, Jackson ImmunoResearch), and Donkey anti-mouse AF488 (#715-545-151, Jackson ImmunoResearch).

Cell culture and plasmid transfection

HEK293 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (#SH30022.01, Hyclone) supplemented with 10% fetal bovine serum (#SV30208.02, Hyclone) at 37°C under a 5% CO₂ atmosphere. Cultured cells were transfected using the PEIMAX reagent (#24765-1, Polysciences, Warrington, PA) following the manufacturer’s instructions. An amount of 2 µg plasmids/well was used for the six-well plate, and 5 µg of plasmids was transfected for each 60 mm dish.

AAV2 virus production and infection

We produced AAV2 as described (31). Briefly, HEK293 cells were transfected with pIAAV2 plus Ad pHelper plasmids, using PEIMAX reagent. The cells were collected at 72 h post-transfection, lysed, and treated with excess DNase I. The clarified cell lysates were subjected to cesium chloride gradient ultracentrifugation. Virus-enriched fractions were collected, dialyzed against phosphate-buffered saline (PBS) (pH 7.4), and quantified by real-time PCR. The titers were determined as DRP (DNase-resistant particle) per microliters. The final virus stocks were kept in a −80°C freezer. HEK293 cells were infected with AAV2 at an MOI of 3000 DRP/cell. The virus (AAV2) was diluted in DMEM and incubated with the cells at 37°C for 4 h, and the cells were then washed once with PBS and fed fresh medium, followed by transfection at 8 h post-infection.

Quantification of AAV2 production by real-time PCR

HEK293 cells seeded into a six-well plate were transfected with pIAAV2 or infected with AAV2 in the presence of HBoV1 helper genes. The cells were lysed with fresh sodium deoxycholate lysis buffer [25 mM Tris-HCl (pH 8.0), 150 mM NaCl, 2 mM MgCl₂, and 0.5% sodium deoxycholate]. The cell lysate was treated with Benzonase nuclease (250 U/mL) for 2 h at 37°C to digest free nucleic acids. The reaction was stopped by adding EDTA to 10 mM, and the mixture was further digested with protease K for viral DNA extraction using a DNeasy blood and tissue kit (Qiagen), according to the manufacturer’s instructions. Viral genome numbers were determined by real-time PCR (forward primer 5′-TGGGTGGATTCGGACTTAAAC-3′, reverse primer 5′-CTGTGTGATGAAGGAAGCAAAC-3′) as DRP and normalized to each cell.

Protein expression and purification

The C-terminal 6 × His-tagged NP1ΔNLS was transformed into E.coli strain BL21 (DE3) and growth at 37°C in 2 × YT medium supplemented with antibiotic (ampicillin 100 µg/mL). When OD_{600 nm} reached 0.6–1.0, the cultures were induced with 0.1 mM isopropyl-β-d-thiogalactoside (IPTG) at 15°C for 16 h. The cells were harvested and lysed in buffer A containing 25 mM Tris-HCl (pH 8.0), 500 mM NaCl. The cell lysate supernatant, after centrifugation, was applied to a nickel-affinity column (Ni-NTA; GE Healthcare) and sequentially washed using buffer A plus 25, 50, and 100 mM imidazole. Eluted proteins were assessed using the heparin column (GE Healthcare) for removing nuclear acids from the targeted protein. The purity of NP1ΔNLS was further polished by gel filtration chromatography using a Superdex-200 column (GE Healthcare) in 10 mM Tris-HCl (pH 8.0) buffer containing 100 mM NaCl and 3 mM DTT. The peak fractions were pooled and assessed using SDS-PAGE gel electrophoresis and frozen in liquid nitrogen. The mutated or truncated NP1 proteins were expressed and purified similarly to the native NP1ΔNLS.

GST-tagged proteins were all expressed in E. coli BL21 (DE3) and induced with 0.1 mM IPTG at 16°C for 16 h. These proteins were captured using Glutathione Sepharose beads and eluted off the beads with buffer B containing 25 mM Tris-HCl (pH 8.0), 150 mM NaCl, 3 mM DTT plus 15 mM reduced glutathione (GSH). The GST-RPA70-F truncations were expressed and captured similarly, except they are on-column precision protease cleavaged to remove the GST tag and eluted in buffer B without GSH. All obtained proteins were extensively dialyzed against the PBS buffer and concentrated for use.

The RPA heterotrimer protein was expressed with minor changes as described (52). Briefly, RPA14 was subcloned into pGEX-6p-1 plasmid via BamHI/XhoI sites. RPA70 and RPA32 were inserted into the pACYCDuet-1 plasmid with BamHI/HindIII restriction sites and NdeI/XhoI restriction sites, respectively. The resulting construct contained RPA70 in MCS1 and RPA32 in MCS2. The plasmids expressing all three subunits of RPA were cotransformed into E. coli strain C43 and induced with 0.1 mM IPTG at 16°C for 16 h. The cells were harvested and lysed in buffer A supplemented with 1 mM phenylmethylsulphonyl fluoride (PMSF, Sangon). The cell lysate containing GST-RPA14 and His-RPA70-RPA32 was loaded onto the Ni-NTA column, followed by extensive washing, and eluted in buffer A containing 300 mM imidazole. The eluted proteins were incubated with precision protease at 20°C for 5 h and further purified using Glutathione Sepharose beads to remove the cleavaged GST tag. The preceding purification procedures of RPA are the same as NP1ΔNLS.

Analytical gel filtration of NP1ΔNLS

Purified NP1ΔNLS protein was adjusted to a 1 mg/mL concentration and manually injected into a Superdex-200 column equilibrated with 10 mM Tris-HCl (pH 8.0), 100 mM NaCl, and 3 mM DTT. The elution volume of NP1ΔNLS was assessed by subjecting standard proteins thyroglobulin (bovine) (670 kD), γ-globulin (bovine) (158 kD), ovalbumin (chicken) (44 kD), myoglobin (horse) (17 kD), and vitamin B12 (1.35 kD) to the same Superdex-200 column.

In vitro DNA pull-down assay

The 5′-biotin-labeled ssDNAs were synthesized and purchased from Sangon Biotech for the DNA pull-down experiments. To generate dsDNAs, equal moles of the complementary DNA oligonucleotides were annealed by boiling for 5 min and slowly cooling to room temperature. For each binding reaction, 200 µL of 1 µM biotin-DNA was incubated with Streptavidin-coated agarose beads (Genscript) at room temperature for 20 min with gentle rotation and washed thoroughly to eliminate unbound DNA. Subsequently, the beads immobilized with ssDNA or dsDNA were incubated with 10 µM of purified NP1ΔNLS wild type or variants in the DNA binding buffer [12 mM HEPEs-Na (pH 7.9), 4 mM Tris-HCl (pH 7.9), 150 mM NaCl, 60 mM KCl, 1 mM EDTA, 1 mM DTT, 12% glycerol, and 0.01% NP-40] at 4°C for 2 h and washed thoroughly in the binding buffer before subjected to SDS-PAGE analyses.

Western blot analyses

After 48 h transfection, the cells were washed once with PBS and lysed with Hirt lysis buffer [10 mM Tris-HCl (pH 8.0), 10 mM EDTA, and 0.6% SDS] for 15 min. 10% of the cell lysates were resolved on 12% SDS-PAGE gels, transferred onto a polyvinylidene difluoride (PVDF) membrane, blocked with 5% non-fat milk, and probed with indicated primary and secondary antibodies accordingly. The remaining 90% of cell lysates were immediately used for Hirt DNA extraction.

Hirt DNA extraction and southern blotting

Southern blotting was used to analyze AAV2 DNA replication level and replication forms such as duplex DNA or ssDNA. (I) Hirt DNA extraction. The Hirt DNA was extracted from pIAAV2 transfected, or AAV2-infected HEK293 cells in the presence of different HBoV1 helper gene set combinations. Briefly, the cells cultured in six-well plates were washed once with PBS and lysed directly by adding 500 µL of Hirt lysis buffer [10 mM Tris-HCl (pH 8.0), 0.6% SDS, and 10 mM EDTA], 90% of the cell lysates were pass through 26 G needle several times to partial shear the chromosome DNA, followed by adding 5 M NaCl to a final concentration of 1.66 M. The mixture was incubated in an ice bath overnight before being cleared by centrifugation at 17,000 × g for 20 min. The supernatants were collected and treated with protease K at a final concentration of 1 mg/mL for 1 h. Hirt DNA was purified with a DNA extraction kit (Qiagen) according to the manufacturer’s instructions. DpnI was used to digest input DNA before Southern blotting was performed for Hirt DNA extracted from transfected cells. (II) Southern blotting. Southern blotting was performed according to our previously reported methods with minor modifications. Briefly, the Hirt DNA samples were digested with DpnI, and resolved on a 1% agarose gel. After denaturation and neutralization, the DNA samples in the gel were blotted onto a positively charged nylon membrane (Roche) and probed with a DIG-labeled probe using the DIG-High Prime DNA Labeling and Detection Starter Kit II (Roche), followed by the manufacturers' instructions. The template for the AAV2 probe was the EcoRI/XbaI-digested 4.3 kb of AAV2, which contains the AAV2 Rep- and VP-encoding sequence. For the detection of rAAV replication, a GFP gene fragment was used as a probe template. Hybridization signals were captured by exposure to X-ray film scanned with a professional X-ray scanner.

In vitro GST pull-down assay

For the pull-down assays, the mixture containing 20 µg of GST or GST-tagged protein and 200 µg of bait proteins was mixed in 500 µL of pull-down buffer [25 mM Tris-HCl (pH 8.0), 200 mM NaCl, 0.5% NP-40, 10% Glycerol, 1 mM EDTA, and 1 mM DTT] and rotated at 4°C for 2 h. The protein complex was captured with Glutathione Sepharose beads. After extensive washing, bound proteins were eluted and separated by SDS-PAGE on a 12% gel and visualized by Coomassie blue staining. Band intensity quantification was performed using the program ImageJ (75).

BLI assay

Protein-protein interactions were measured using an Octet R8 instrument with GST biosensors (ForteBio) at 30°C. For the affinity measurements, all the GST-tagged proteins were prepared in the PBST buffer (consisting of 1 × PBS and 0.02% Tween 20) and immobilized on the GST biosensor for the shift value reached at 1 nm. After a short baseline (120 s), samples containing NP1ΔNLS or RPA protein were twofold serially diluted in concentrations and were allowed to associate for 180 s, followed by dissociation for an additional 300 s. Octet data analysis software was used to fit the data to a 1:1 binding model to extract the association and dissociation rates. The K_D was calculated using the ratio K_off/K_on.

IP and mass-spectrum analysis

HEK293 cells were cotransfected with pIAAV2 plasmid, FLAG-NP1, NS2, and BocaSR expressing plasmids. At 48 h post-transfection, the cells were collected and lysed with NP-40 lysis buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.5% NP-40, 10% glycerol, and 1 mM EDTA, supplemented with protease and phosphatase inhibitor cocktails] on ice for 20 min. After centrifugation at 15,000 × g for 15 min, 10% of the clarified cell lysates were saved for input. The remains were equally distributed into two tubes, and anti-FLAG or control IgG antibody was added individually. The two mixtures were rotated at 4°C for 4 h, then incubated with Protein A/G beads for 2 h to capture the antigen-antibody complex. The extensively washed beads were resuspended in 2 × SDS loading buffer by boiling at 95°C for 10 min before being subjected to 10% SDS-PAGE analysis. The unique bands in the anti-FLAG group, visualized by Coomassie blue staining, were excised and sent for mass-spectrum analysis at APTBIO, Shanghai. Briefly, the gel pieces were first digested with trypsin and then analyzed using a Q Exactive mass spectrometer coupled to Easy-nLC 1000 (Thermo Fisher Scientific). Finally, the acquired data are analyzed with MASCOT software to obtain qualitative identification information for the target protein peptide molecules.

Co-IP assay

The Co-IP assays were conducted similarly to the above section. Briefly, cell monolayers were washed with PBS and incubated on ice with NP-40 lysis buffer. For each sample, 500 µL protein lysate was incubated with 1 µg antigen-specific antibody or normal IgG control antibody and 30 µL protein A/G agarose beads at 4°C for 4 h. The beads were washed three times with 1 mL of lysis buffer, and then the precipitate proteins were eluted with SDS loading buffer resolved in SDS-PAGE gel, followed by detection by Western blotting. An HRP-labeled light chain-specific antibody was used for detection to eliminate the interference of antibody-heavy chains.

Immunofluorescence (IF) analysis

HEK293 cells seeded onto circular coverslips were transfected with indicated plasmids. The cells were washed once with room temperature PBS before being fixed in 4% paraformaldehyde (PFA) for 20 min. The fixed cells were washed once with 30 mM glycine to neutralize excess PFA, followed by PBS washing and permeabilization with 0.1% Triton-100 in PBS. At room temperature, the permeabilized cells were blocked for 60 min in a blocking buffer (3% BSA in PBS). The cells were then incubated with diluted primary antibodies (mouse anti-Myc tag, mouse anti-FLAG tag, or rabbit anti-RPA70) overnight at 4°C. After 2 h incubation at room temperature with appropriate secondary antibodies (Donkey anti-mouse AF488, Donkey anti-rabbit AF594, or Donkey anti-mouse AF594), the cells were washed with PBST and stained using DAPI for 5 min. Coverslips were mounted on a glass slide with mounting medium and imaged using a confocal microscope system (NIKON-A1R/STORM).

For co-visualization of BrdU-labeled AAV2 replication center with NP1 or NP1ΔNLS, the cells transfected with pIAAV2 and HBoV1 helper gene sets were pulse-labeled with BrdU at a final concentration of 30 µM for 30 min. As described above, the cells were fixed, neutralized, permeabilized, and blocked. The cells were co-stained with antibodies against BrdU and against Myc-tag, followed by staining with appropriate secondary antibodies.

RNA extraction, reverse transcription, and real-time PCR

Total RNAs were isolated from the cells using the TRIzol Reagent (TaKaRa) following the manufacturer’s protocols. cDNA was then synthesized with M-MLV reverse transcriptase (Promega) and random hexamer primers (Vazyme). The newly synthesized cDNAs were used as the templates for amplifying a highly specific nucleotide region of target genes in quantitative PCR (qPCR) assay. Relative qPCR was performed on an ABI 7500 real-time PCR system (Applied Biosystems) using a SYBR Green Real-Time PCR Master Mix (TaKaRa) Dye. Each reaction was performed in triplicates. The mRNA levels of a housekeeping gene GAPDH were quantified as an internal control. The difference in gene expression was calculated based on 2^−ΔΔCT values. The primers 5′-agcagcacgacttcttcaagtcc-3′ and 5′-tgtagttgtactccagcttgtgc-3′ were used for GFP; primer pairs 5′-ctgggctacactgagcacc-3′ and 5′-aagtggtcgttgagggcaatg-3′ were used for GAPDH.

Luciferase reporter assay

The cells were cotransfected pRep-Fluc and plasmids expressing Myc-NP1, NS2, BocaSR, and ^FLAGRPA70-NLS. The Myc-NP1 and ^FLAGRPA70-NLS plasmids were transfected at indicated ratios. At 48 h post-transfection, the cells were lysed for measuring Firefly luciferase activities using the Luciferase Reporter Assay System (Promega), according to the manufacturer’s instructions.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 8.0; error bars show means and standard deviations. P values of statistical significance were determined by using one-way ANOVA analysis. ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; and *, P < 0.05 were regarded as statistically significant and n.s. indicated statistically not significant.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jvi.01515-23.

Supplemental figures and tables. jvi.01515-23-s0001.docx.

Figures S1 to S4; Tables S1 to S3.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.