Role of the membrane-associated accessory protein (MAAP) in adeno-associated virus (AAV) infection

ABSTRACT

Adeno-associated viruses (AAVs) package a single-stranded (ss) DNA genome of 4.7 kb in their capsid of ~20 nm in diameter. AAV replication requires co-infection of a helper virus, such as adenovirus. During the optimization of recombinant AAV production, a small viral nonstructural protein, membrane-associated accessory protein (MAAP), was identified. However, the function of the MAAP in the context of AAV infection remains unknown. Here, we investigated the expression strategy and function of the MAAP during infection of both AAV2 and AAV5 in human embryonic kidney (HEK)293 cells. We found that AAV2 MAAP2 and AAV5 MAAP5 are expressed from the capsid gene (cap)-transcribing mRNA spliced from the donor to the second splice site that encodes VP2 and VP3. Thus, this AAV cap gene transcribes a multicistronic mRNA that can be translated to four viral proteins, MAAP, VP2, AAP, and VP3 in order. In AAV2 infection, MAAP2 predominantly localized in the cytoplasm, alongside the capsid, near the nuclear and plasma membranes, but a fraction of MAAP2 exhibited nuclear localization. In AAV5 infection, MAAP5 revealed a distinct pattern, predominantly localizing within the nucleus. In the cells infected with an MAAP knockout mutant of AAV2 or AAV5, both viral DNA replication and virus replication increased, whereas virus egress decreased, and the decrease in virus egress can be restored by providing MAAP in trans. In summary, MAAP, a novel AAV nonstructural protein translated from a multicistronic viral cap mRNA, not only facilitates cellular egress of AAV but also likely negatively affects viral DNA replication during infection.

IMPORTANCE

Recombinant adeno-associated virus (rAAV) has been used as a gene delivery vector in clinical gene therapy. In current gene therapies employing rAAV, a high dose of the vector is required. Consequently, there is a high demand for efficient and high-purity vector production systems. In this study, we demonstrated that membrane-associated accessory protein (MAAP), a small viral nonstructural protein, is translated from the same viral mRNA transcript encoding VP2 and VP3. In AAV-infected cells, apart from its prevalent expression in the cytoplasm with localization near the plasma and nuclear membranes, the MAAP also exhibits notable localization within the nucleus. During AAV infection, MAAP expression increases the cellular egress of progeny virions and decreases viral DNA replication and progeny virion production. Thus, the choice of MAAP expression has pros and cons during AAV infection, which could provide a guide to rAAV production.

INTRODUCTION

Adeno-associated viruses (AAVs) are small, non-enveloped viruses with a linear 4.7-kb single-stranded DNA (ssDNA) genome packaged into a non-enveloped icosahedral (T = 1) capsid of ~20 nm in diameter (1). AAVs belong to the genus Dependoparvovirus of the family Parvoviridae (2). Recombinant (r) AAVs have been widely used as vectors in human gene therapy because they are naturally nonpathogenic to humans and have a broad tropism (3–5). The efficacy, persistence, and safety of rAAV vectors have been well-demonstrated in human clinical trials and clinically approved treatments (4). Thirteen natural human and nonhuman primate AAV serotypes (AAV1-13) and variants of more than one hundred have been identified or engineered (5). They have been used in in vitro gene transfer experiments and have had their tropisms expanded for tissue- and organ-specific gene deliveries (6–12). Currently, five AAV-based gene therapy medicines, Luxturna, Zolgensma, Hemgenix, Elevidys, and Roctavian, have been approved by the US Food and Drug Administration (FDA) for the treatment of spinal muscular atrophy, retinal dystrophy, hemophilia B, Duchenne muscular dystrophy, and hemophilia A, respectively (13–17). Today, novel AAV capsids are being developed to fulfill the demand for efficient, convenient, and optimized vectors in gene therapy. Although a significant amount of research has been carried out in vector engineering to generate new capsid variants, there is still a high demand to elucidate the biology of AAVs in depth for high yield production of effective rAAV vectors for clinical applications.

While Luxturna, Zolgensma, and Elevidys use AAV2, AAV9, and AAVrh74 capsid, respectively, to deliver the transgene, both Hemgenix and Roctavian use the AAV5 capsid. Based on the sequences of the large capsid protein VP1, naturally occurring primate AAVs have been phylogenetically grouped as Clades A–H, with AAV5 as the most distinct one (18, 19). The genus Dependoparvovirus has been classified into Dependoparvovirus A (AAV1–4 and 6–13) and Dependoparvovirus B (AAV5, bovine AAV, and goat AAV) (20). The gene expression strategy used by AAVs in Dependoparvovirus B is different from that employed by AAVs in Dependoparvovirus A (21–25).

The AAV2 genome comprises two open-reading frames (ORFs) for replication proteins (Rep) and capsid proteins (VP/Cap), respectively, between two inverted terminal repeats (ITR) of ~145 nucleotides (26, 27). The rep genes, which are transcribed from P5 and P19 promoters, respectively, encode the large Rep78/68 and small Rep52/40 proteins that are required for viral DNA replication, viral mRNA transcription and processing, and viral genome packaging (1, 21, 28). AAV2 P40 promoter-transcribed precursor (pre-)mRNA is alternatively spliced from the donor (D) at nt 1,906 to two acceptor (A) sites, A1 at nt 2,201 and A2 at nt 2,228, at a ratio of ~1:10 (22, 29). The P40-transcribed mRNA spliced from D to A1 (P40-SplA1) encodes VP1, and the P40-transcribed mRNA that is spliced from D to A2 (P40-SplA2) encodes VP2 and VP3, which have a stoichiometry of 1:1:10 (30, 31). The cap gene also encodes two auxiliary small viral nonstructural proteins, the assembly-activating protein (AAP) (32) and the recently identified membrane-associated accessory protein (MAAP) (33). It is known that the MAAP is expressed in the VP1-coding sequence from a + 1 frameshifted ORF at a noncanonical start codon CUG and has 119 amino acids (aa) and a molecular weight of ~14 kDa (33, 34). The transcription profile of AAV5 is different from that of AAV2 (23). First, AAV5 P7 and P19 promoters transcribe mRNAs that encode Rep78 and Rep52 only, respectively, which are polyadenylated at the proximal polyadenylation site that lies within the AAV5 intron. Second, the AAV5 P41 promoter, which transcribes VP-encoding mRNAs, is active per se without the requirement of Rep78 to activate (35, 36).

The MAAP has been predicted to be expressed not only in Dependoparvovirus A but also in Dependoparvovirus B AAVs (33, 37). However, which mRNA is responsible for expression of AAV MAAP is still unknown. The MAAP is predicted to be mostly structurally disordered and to contain two short protein-binding domains (38). Structural predictions show six discernable regions: (1) a short, disordered N-terminus (aa 1–15); (2) a short, hydrophobic stretch containing at least one cysteine (C) (aa 16–22) can form a β-strand; (3) a central, T-/S-rich region predicted disordered (aa 23–73), rich in charged amino acids; (4) a region devoid of a predicted secondary structure (aa 74–83), is ordered; (5) a disordered region that has the potential to form an α-helix (aa 84–94); (6) a C-terminal amphipathic α-helix (aa 95–116) is predicted to bind membranes (38). The C-terminus has been experimentally proven to play an essential role in membrane anchoring and promoting extracellular secretion of rAAV vectors (37). In contrast to the AAP (32, 39–41), the function of the MAAP is not well-studied. During rAAV production, AAV8 MAAP (MAAP8) has been observed in the promotion of secreting produced vectors by interacting with extracellular vesicles (37). In cells infected with adenovirus (Ad), transfection of a full-length AAV2 genome clone demonstrated that MAAP2 functions as an accelerator for AAV2 replication and likely in viral genome packaging (34). However, the exact roles of the MAAP during AAV infection and its expression strategy remain elusive, and nothing is known about the expression of AAV5 MAAP (MAAP5) and its function.

In this report, we first studied the expression strategy of both MAAP2 and MAAP5, representative of Dependoparvovirus A and B, respectively, by transfection of cap-expressing constructs in human embryonic kidney (HEK)293 cells. We then investigated cellular localization and functions of both MAAP2 and MAAP5 during infection of AAV2 and AAV5, respectively, in HEK293 cells transfected with the pHelper plasmid that expressed Ad helper genes, E2a, E4orf6, VA, and L4-22K/33K (42), a nonlytic AAV infection system without lysis of the infected cells. This nonlytic infection system allows observing progeny virion egress without cell lysis by Ad.

RESULTS

The expression strategy of AAV2 MAAP (MAAP2)

To answer a fundamental question on AAV gene expression, which P40-transcribed mRNA expresses MAAP2 and AAP2, we constructed a CMV (immediate early, IE) promoter-driven AAV2 cap gene expression cassette, pCMVAAV2Cap, in which the AAV2 P40 promoter was replaced by the CMV IE promoter because the P40 promoter requires AAV2 Rep68/78 in trans and ITR or P5 promoter in cis to become activated (28) (Fig. 1A). To observe the expression only from P40-SplA1and P40-SplA2 mRNAs (Fig. 2A), we removed the D-A1 and D-A2 introns, respectively, in pCMVAAV2Cap to construct two cDNA plasmids: pCMVAAV2Cap^D/A1 and pCMVAAV2Cap^D/A2 (Fig. 1A). Transfection of pCMVAAV2Cap in HEK293 cells expressed five viral proteins, MAAP, AAP, VP1, VP2, and VP3 (Fig. 1B, lane 2). When the intron D-A1 was removed, pCMVAAV2Cap^D/A1 expressed strong bands of VP1 and AAP but weak bands of MAAP, VP2, and VP3 (Fig. 1B, lane 3). In the cells transfected with pCMVAAV2Cap^D/A2, VP1 was not expressed; VP2 and VP3 were expressed at a ratio of ~1:5, similar to pCMVAAV2Cap-transfected cells; and both MAAP and AAP were expressed at a similar or a higher level than that in the pCMVAAV2Cap-transfected cells (Fig. 1B, lane 4). Reverse transcription (RT)-quantitative (q) PCR confirmed that pCMVAAV2Cap^D/A1 only expressed D/A1 cap mRNA and pCMVAAV2Cap^D/A2 only expressed D/A2 cap mRNA (Fig. 1C), as determined using D/A1- and D/A2-specific probes, respectively, in RT-qPCR (Table S1).

Fig 1.

MAAP2 expression strategy. (A) Schematic diagram of the AAV2 cap-expressing constructs. The parent plasmid pCMVAAV2Cap, two cap-encoding cDNA constructs (pCMVAAV2Cap^D/A1 and pCMVAAV2Cap^D/A2), and two synthetic intron (Sintron) replacement constructs (pCMVAAV2Cap^{D^A1} and pCMVAAV2Cap^{D^A2}) are depicted with transcription and translation units at the nucleotide numbers of the AAV2 genome (GenBank accession no.: AF043303). Open-reading frames (ORFs) located in the cap gene are outlined in colors. (B&D) Western blotting. HEK293 cells were transfected with plasmids, as indicated. At 2 days post-transfection (dpt), cells were collected and analyzed with Western blotting using anti-MAAP2, anti-AAP2, and anti-VP for expression of MAAP2, AAP2, and VP1, VP2, and VP3 as indicated. β-actin is shown as a loading control. (C&E) mRNA quantification by RT-qPCR. HEK293 cells were transfected with plasmids as indicated. At 2 dpt, cells were collected for total RNA isolation. cDNAs were generated by reverse transcription and quantified by multiplex qPCR using D-A1 or D-A2 splicing mRNA-specific probes and a β-actin-specific probe (as an internal control). The left bar in panels C&E shows D-A1 mRNA of the cDNA or D-A1-spliced mRNA of the original intron (C) or Sintron (E) that were detected by a D-A1 splicing-specific probe. The right bars in panels C&E show D-A2 mRNA of the cDNA or D-A2 spliced mRNA of the genuine intron (C) or Sintron (E) that were detected by a D-A2 splicing-specific probe. Relative gene expressions are interpreted using the 2^-∆∆CT method. UD means undetected. (F) Western blotting. HEK293 cells were transfected with an AAVRep2Cap2 (R2C2) plasmid and its derivatives as indicated plasmids, together with pHelper. At 2 dpt, cells were collected and analyzed with Western blotting using anti-MAAP2, anti-AAP2, anti-VP, and anti-Rep for expression of AAV2 MAAP2, AAP2, VP1, VP2, VP3, and Rep proteins, as indicated. β-actin is shown as a loading control.

Fig 2.

Cap gene expression map of AAV. (A) AAV2 cap gene. The AAV2 cap gene is depicted. AAV2 P40 promote- transcribed mRNAs include P40 unspliced mRNA and P40 alternatively spliced mRNAs at D and A1 or A2, designated as P40-SplA1 or P40-SplA2 mRNA, respectively. As the P40-SplA2 is the predominant P40-transcribed mRNA during AAV2 and Ad5 co-infection (splicing at A1 vs A2 is at a ratio of >1:10), P40-SplA1 mRNA largely expresses VP1 at nt 2,203 (AUG) and minor AAP2 at nt 2,729 (CUG), and P40-SplA2 mRNA expresses MAAP, VP2, AAP, and VP3 at initiation codons of nt 2,282 (CUG), nt 2,614 (ACG), nt 2,729 (CUG), and nt 2,809 (AUG), respectively. (B) AAV5 cap gene. The AAV5 cap gene is depicted. The AAV5 P41 promoter transcribes a pre-mRNA that is unspliced (P41-unspl) or alternatively spliced mRNAs at D and A1 or A2, which are designated as P41-SplA1 or P41-SplA2 mRNA, respectively. P41-SplA2 is the abundant P41-transcribed mRNA during infection, expressing MAAP, VP2, AAP, and VP3 in order at initiation codons of nt 2,232 (ATT), nt 2,615 (ACG), nt 2,715 (CUG), and nt 2,783 (AUG), respectively. P41-SplA1 mRNA expresses VP1 at initiation codons of nt 2,207 (AUG) and very minor AAP5 at nt 2,715 (CUG). The MAAP and AAP are translated from the +1 frame-shifted ORF. The nucleotide numbers of the AAV2 and AAV5 genomes refer to GenBank accession #AF043303 and #AF085716, respectively.

Next, we replaced the D-A1 and D-A2 introns with a strong synthetic chimeric intron (Sintron) in pCMVAAV2Cap to observe viral protein expression from P40-SplA1and P40-SplA2 mRNAs that underwent the mRNA splicing process. Similarly, the MAAP was detected at a relatively high level in HEK293 cells transfected with pCMVAAV2Cap^{D^A2}, compared to that with pCMVAAV2Cap^{D^A1}, whereas AAP expression was detected at an opposite level (Fig. 1D, lanes 4 vs 3). VP2 and VP3 were detected in cells transfected with either plasmid, whereas VP1 was only detected in pCMVAAV2Cap^{D^A1}-transfected cells (Fig. 1D, lane 3). RT-qPCR confirmed pCMVAAV2Cap^{D^A1} only expressed D^A1 cap mRNA and pCMVAAV2Cap^{D^A2} only expressed D^A2 cap mRNA (Fig. 1E).

To confirm the cap gene expression in the AAV2 genome, based on an AAV2 rep- and cap-expressing plasmid (R2C2), we generated D/A1 and D/A2 cDNA and D^A1 and D^A2 Sintron-replaced R2C2 plasmids. R2C2 has the two ITRs removed. We observed that both R2C2^D/A1 and R2C2^{D^A1} only expressed VP1 and a very minor band of AAP2 (Fig. 1F, lanes 3 and 5), and both R2C2^D/A2 and R2C2^{D^A2} expressed abundant AAP2, VP2, VP3, and MAAP2 (Fig. 1F, lanes 4 and 6). The parent R2C2 expressed AAP2, VP1-3, and MAAP from the cap gene, as well as Rep proteins (Fig. 1F, lane 2). Because the intron was removed or replaced, these R2C2 mutants only expressed the smaller Rep68 and Rep40 (Fig. 1F, lanes 3–6).

Considering the P40 mRNAs are spliced at A1 vs A2 at a ratio of ~1:10 during AAV2 infection in the presence of Ad5 (29), we concluded that the P40-SplA2 mRNA is the dominant mRNA transcript to translate MAAP, AAP, VP2 and VP3, but only the P40-SplA1 mRNA expressed VP1. The expressions of AAV2 MAAP2, VP2, and VP3 from P40-SplA1 mRNA generated from the CMV-driven cap gene are likely due to the strong CMV promoter, which was confirmed in the AAV5 cap gene expression from the AAV5 native P41 promoter (Fig. 3). The AAV2 cap gene expression strategy is summarized in Fig. 2A.

Fig 3.

MAAP5 expression strategy. (A) Schematic diagram of the AAV5 cap-expressing constructs. Wild-type (WT) plasmid pP41AAV5Cap, two cap-encoding cDNA constructs (pP41AAV5Cap^D/A1 and pP41AAV5Cap^D/A2), and AUU-mutated cDNA construct (pP41AAV5Cap^D/A2∆ATT) are depicted with transcription and translation units at the nucleotide numbers of the AAV5 genome (GenBank accession #AF085716). Open-reading frames (ORFs) located in the cap gene are outlined in colors. (B&C) Western blotting. HEK293 cells were transfected with the p41AAV5Cap-based plasmids (B) and AAV5RepCap (R5C5)-based plasmids (C) as indicated. At 2 dpt, cells were collected and analyzed with Western blotting using anti-MAAP5, anti-AAP5, anti-VP, and anti-Rep for expression of MAAP5, AAP5, VP1-3, and Rep proteins, respectively, as indicated. (D&E) Determination of the start codon of MAAP5. (D) A strategy of mutating potential MAAP5 start codons is depicted. Numbers indicate mutated nucleotides. (E) Western blotting. HEK293 cells were transfected with plasmids as indicated. At 2 dpt, cells were collected and analyzed with Western blotting using anti-Flag and anti-MAAP5. β-actin is shown as a loading control.

The expression strategy of AAV5 MAAP (MAAP5)

Concurrently, the MAAP5 transcription profile was also investigated in the context of the genuine AAV5 P41 promoter-driven AAV5 cap gene. We employed pP41AAV5Cap, in which the P41 promoter is active independent of any cis or trans factors (35) (Fig. 3A). Two plasmids, pP41AAV5Cap^D/A1 and pP41AAV5Cap^D/A2, were constructed by removing D-A1 and D-A2 introns of AAV5, respectively, to examine the expressions of cap gene mRNAs (Fig. 3A). MAAP5 expression was observed obviously when D-A2 was removed in transfected HEK293 cells, but not when D-A1 was removed (Fig. 3B, MAAP5, lane 4 vs lane 3). AAP5 expression exhibited at variable levels from the three constructs (Fig. 3B, AAP5, lanes 2–4). The expression pattern of AAV5 VP1–3 from the two mutants was similar to that of AAV2 counterparts (Fig. 1B).

To confirm the cap gene expression in the AAV5 genome, we generated two cDNA constructs from an AAV5 rep and cap-expressing plasmid (R5C5) that has the two ITRs removed. We observed R5C5^D/A1 expressed largely VP1 and a weak band of AAP5 (Fig. 3C, lane 3) and R5C5^D/A2 expressed abundant AAP5, VP2, VP3, and MAAP5 (Fig. 3C, lane 4). The parent R5C5 expressed AAP2, VP1–3, and MAAP from the cap gene, as well as Rep proteins (Fig. 3C, lane 2). Since the intron was removed or replaced, the R5C5 mutants only expressed smaller Rep68 and Rep40 (Fig. 3C, lanes 3–5).

To determine the start codon of MAAP5, we mutated several potential start codons referring the MAAP2 amino acid sequence in the context of pP41AAV5MAAP^Flag (Fig. 3D). Among all mutants with potential start codons, we only did not detect MAAP5 expression in cells transfected with pP41AAV5MAAP^Flag(ATT∆) (Fig. 3E), suggesting that MAAP5 was translated from a noncanonical start codon, AUU, which was immediately next to the A2 splice site (Fig. 2B). To ensure that the blockage of MAAP5 expression was independent of any splicing issue that may be caused by mutation in the splicing region, we mutated the potential start codon in the constructs, pP41AAV5Cap^D/A2∆ATT and R5C5^D/A2∆ATT, in which D-A2 intron was already removed. These two mutants led to no decrease in the expression of other proteins from D-A2 cDNA-transcribed mRNA while abolishing MAAP5 expression (Fig. 3B and C, lane 5).

MAAP5 is expressed from P41-SplA2 mRNA with an unusual non-AUG start codon AUU, whereas AAP5 is expressed abundantly from P41-SplA2 mRNA and weakly from P41-SplA1mRNA. Importantly, MAAP5 was found to be 17 aa longer than MAAP2. The AAV5 cap gene expression strategy is summarized in Fig. 2B.

MAAP2 localizes mostly in the cytoplasm but with some in the nucleus, whereas MAAP5 largely localizes in the nucleus

Previously, overexpression of the MAAP in transfected cells showed that it was associated with the plasma membrane and appeared as punctuated patterns throughout the cell (33, 37). We determined the subcellular localization of MAAP2 and MAAP5 in nonlytic virus infection. HEK293 cells were infected with AAV2 or AAV5 and were co-transfected with pHelper. As the expression of AAV Rep proteins induces cell cycle arrest and apoptosis (43, 44), to ensure a nonlytic infection, we analyzed infected cells at 2 days post-infection (dpi). AAV-infected cells were co-immunostained with anti-AAV2 intact capsid (A20) and anti-MAAP2 antibodies or with anti-AAV5 intact capsid (ADK5b) and anti-MAAP5 antibodies. The specificity of both anti-MAAP antibodies was confirmed by the absence of signals observed in mock-infected cells (Fig. 4 and 5B, Mock), as well as the colocalization of anti-MAAP and anti-Flag staining in cells expressing Flag-tagged MAAP (Fig. 4 and 5C).

Fig 4.

MAAP2 predominantly localizes in the cytoplasm, with some presence observed in the nucleus during AAV2 infection. HEK293 cells were infected with wtAAV2 or mock-infected followed by pHelper transfection. At 2 dpt or dpi, infected cells were co-immunostained for MAAP2 and AAV2 intact capsid and observed under a Leica STED microscope with a 100 x objective lens (A) or for Z-stack at a 14.27 µm vertically scanned area (B). Representative Z-stack images of a collection of 42 layers and produced Z-stack images are presented. The colors of confocal images correspond to blue for DAPI, red for MAAP, and green for AAV intact capsids. Size bar = 5 µm. Representative confocal images are shown. (C) HEK293 cells were transfected with pCI-MAAP2^Flag or mock-transfected. At 2 dpt, transfected cells were co-immunostained using anti-MAAP2 and anti-Flag and observed under a Leica STED microscope with a 100 x objective lens. The colors of confocal images correspond to blue for DAPI, green for MAAP2, and red for Flag. Size bar = 5 µm. Representative confocal images are shown.

Fig 5.

MAAP5 localizes both in the cytoplasm and nucleus during AAV5 infection. HEK293 cells were infected with wtAAV5 or mock-infected followed by pHelper transfection. At 2 dpt or dpi, infected cells were co-immunostained with MAAP5 and AAV5 intact capsid and observed under a Leica STED microscope with a 100 x objective lens (A) or for Z-stack at a 10.91 µm vertically scanned area (B). Representative Z-stack images of a collection of 50 layers and produced Z-stack images are presented. The confocal images are color-coded, with DAPI represented in blue, MAAP5 in red, and intact AAV5 particles in green. Size bar = 5 µm. Representative confocal images are shown. (C) HEK293 cells were transfected with p41AAV5MAAP5^Flag or mock-transfected. At 2 dpt, transfected cells were co-immunostained with anti-MAAP5 and anti-Flag and observed under a Leica STED microscope with a 100 x objective lens. The colors of confocal images correspond to blue for DAPI, green for MAAP5, and red for Flag. Size bar = 5 µm. Representative confocal images are shown.

While the MAAP2 localized abundantly in the cytoplasm near the nuclear and plasma membranes, we also observed that it weakly localized in the nucleus (Fig. 4A). However, the nuclear localization of MAAP5 was much more obvious (Fig. 5A). The Z-stack images of confocal microscopy clearly showed both MAAP2 and MAAP5 were detectable in the nucleus, where most of the capsids localized (Fig. 4B and 5B ). We reasoned that the produced capsids in the nuclei trafficked to the cytoplasm and egressed outside of the cells rapidly, so fewer capsid signals were observed in the cytoplasm than in the nucleus. Thus, we confirmed that both MAAP2 and MAAP5 localize not only in the cytoplasm but also in the nucleus, suggesting that the MAAP has a function in the nucleus.

MAAP knockout does not affect the expression of other viral proteins

To examine the function of MAAP2 in nonlytic AAV2 infection, based on an AAV2 infectious clone SSV9 (45), we constructed a mutant infectious clone without the expression of a full-length MAAP (SSV9^∆MAAP) by introducing an early stop codon at th19th aa (leucine) that does not alter any amino acid in the VP1 ORF (Fig. 6A), which led to a complete knockout of MAAP2 in SSV9^∆MAAP-transfected HEK293 cells, while other viral proteins were expressed at similar levels (Fig. 6B). Similarly, an AAV5 infectious clone, pAAV5 (23), was introduced with an early stop codon at the 36th aa (leucine) without affecting VP1 ORF to construct pAAV5^∆MAAP (Fig. 6D), which led to a complete knockout of MAAP5 in pAAV5^∆MAAP-transfected HEK293 cells, while other viral proteins were expressed at similar levels (Fig. 6E). We next transfected SSV9^∆MAAP and pAAV5^∆MAAP into HEK293 cells, followed by purification of MAAP knockout mutant viruses, AAV2^∆MAAP and AAV5^∆MAAP. The infection of AAV2^∆MAAP or AAV5^∆MAAP demonstrated that only the MAAP was ablated, while other viral proteins were largely unaffected (Fig. 6C and F). The slightly increased AAP2 expression in SSV9^∆MAAP-transfected cells or AAV2^∆MAAP-infected cells was likely due to the knockout of the MAAP ORF.

Fig 6.

Construction of MAAP knockout AAV infectious clone and mutant virus infections. (A) Construction of an MAAP2 knockout mutant infectious clone of AAV2 (SSV9^∆MAAP). MAAP2 is translated from a + 1 frame-shifted ORF in the VP-encoding mRNAs started at nt 2,282 and ended at nt 2,641 with a noncanonical start codon (CUG). The 19th amino acid (leucine/L) of the MAAP2 was mutated to a stop codon (UAG) in AAV2 infectious clone SSV9, resulting in SSV9^∆MAAP. (B&C) Western blotting of AAV2 proteins. (B) Transfection. HEK293 cells were transfected with SSV9 and SSV9^ΔMAAP, followed by pHelper transfection. (C) Infection. HEK293 cells were infected with wtAAV2 or AAV2^ΔMAAP, followed by pHelper transfection. At 2 dpt or dpi, cells were harvested and immunoprobed for AAV2 proteins: VP and Rep proteins, AAP2, and MAAP2. β-actin is shown as a loading control. (D) Construction of an MAAP5 knockout mutant infectious clone of AAV5 (pAAV5^∆MAAP). MAAP5 is translated from a + 1 frame-shifted ORF on the cap-encoding mRNAs started at nt 2,232 and ended at nt 2,639 with a noncanonical start codon (AUU). The 36th amino acid (leucine/L) of the MAAP5 was mutated to a stop codon (UAG) in an AAV5 infectious clone pAAV5, resulting in pAAV5^∆MAAP. (E&F) Western blotting of AAV5 proteins. HEK293 cells were transfected with pAAV5 or pAAV5^ΔMAAP (E) or infected with AAV5 or AAV5^ΔMAAP (F), followed by pHelper transfection. At 2 dpt or dpi, cells were harvested and immunoprobed for AAV5 proteins: VP and Rep proteins, AAP5, and MAAP5. β-actin is shown as a loading control.

Collectively, we successfully rescued MAAP knockout AAV mutants, AAV2^∆MAAP and AAV5^∆MAAP.

AAV DNA replication increases in the absence of the MAAP

We next examined the viral DNA replications and mRNA expressions of the MAAP mutants. To this end, we infected HEK293 cells with wtAAV2 and AAV2^∆MAAPor wtAAV5 and AAV5^∆MAAP, with co-transfection of pHelper, and complemented MAAP2 in AAV2^∆MAAP-infected cells and MAAP5 in AAV5^∆MAAP-infected cells by transfection of pCI-MAAP2 and pCI-MAAP5, respectively. pCI-MAAP2 uses a noncanonical start codon (CUG), while pCI-MAAP5 initiates at AUG, which ensured an expression level similar to that during wtAAV2 or wtAAV5 infection. At 2 dpi, we harvested infected cells for lower molecular weight (Hirt) DNA extraction and Southern blotting. MAAP ablation in AAV2^∆MAAP- and AAV5^∆MAAP-infected cells caused a significant increase in the intermediates of viral DNA, mRF (monomeric replicative form) and dRF (double replicative form), and ssDNA (Fig. 7A, lanes 4 vs 3; Fig. 8A, lanes 4 vs 3; and Fig. 7B and Fig. 8B). Although there were no statistical significances, MAAP complementation mostly rescued the level of replicative viral DNA in AAV2^∆MAAP- or AAV5^∆MAAP-infected cells to that in wtAAV2-infected cells (Fig. 7A, lanes 5 vs 3 and Fig. 7B) or wtAAV5-infected cells (Fig. 8A, lanes 5 vs 3 and Fig. 8B) to certain degrees. Northern blotting showed a similar pattern of viral mRNAs among total RNAs extracted from wtAAV2-, AAV2^∆MAAP-, and MAAP2-complemented AAV2^∆MAAP-infected cells (Fig. 7C). As expected, viral proteins (except for MAAP) were expressed at similar levels in AAV2^∆MAAP-infected cells as in wtAAV2-infected cells and when the MAAP was expressed in trans in AAV2^∆MAAP-infected cells (Fig. 7D). Similarly, in AAV5^∆MAAP-infected cells, viral mRNAs and proteins were expressed at levels similar to those in wtAAV5-infected cells and at the similar levels were restored when the MAAP was provided in trans (Fig. 8C and D).

Fig 7.

MAAP2 knockout increases viral DNA replication in AAV2^∆MAAP-infected cells. HEK293 cells were infected with wtAAV2 and AAV2^ΔMAAP or mock-infected, followed by transfection with pHelper and pCI-empty (─) or pCI-MAAP2 (+) as indicated. At 2 dpt or dpi, the cells were harvested for analyses as follows. (A&B) Viral DNA detection. (A) Southern blotting. Hirt DNA was extracted from the infected cells, digested with Dpn I, and subjected to electrophoresis and blotted using an undigested SSV9 probe. The AAV2 double-stranded (ds) DNA genome excised from SSV9 was loaded as a size marker (Lane 1). dRF-double replicative form, mRF-monomeric replicative form, ssDNA-single stranded DNA, and Dpn I-digested bands are indicated. (B) mRF quantification. mRF data were normalized to Dpn I-digested bands that serve as loading controls. Relative folds of intensity are shown. Means and standard deviations were calculated based on three independent experiments. *, P < 0.05; ***, P < 0.001; and ns, no significant difference. (C) Northern blot analysis. Total RNA was extracted from AAV2-infected cells using TRIzol reagent. Ten micrograms RNA of each sample was separated on 1% agarose gel, ethidium bromide-stained, and visualized under a UV light for the 28S and 18S ribosomal (r) RNAs, as well as the ssRNA ladder. The gel was then transferred onto a nitrocellulose membrane, followed by hybridization with an undigested SSV9 probe. P5, P19, and P40 mRNAs correspond to mRNAs transcribed from P5, P19, and P40 promoters, respectively. (D) Western blot analysis. Infected cells were lysed and immunoprobed for AAV2 proteins: VP and Rep proteins, MAAP2 and AAP2, as indicated. β-actin is shown as a loading control.

Fig 8.

MAAP5 knockout increases viral DNA replication in AAV5^∆MAAP-infected cells. HEK293 cells were infected with wtAAV5 and AAV5^ΔMAAP, respectively, or mock-infected, followed by transfection with pHelper and pCI-empty (─) or pCI-MAAP5 (+) as indicated. At 2 dpt or dpi, the cells were harvested for analyses as follows. (A&B) Detection of viral DNA. (A) Southern blotting. Hirt DNA was extracted from the infected cells, digested with Dpn I, and subjected to electrophoresis and blotted using a probe made of the AAV5 dsDNA genome excised from pAAV5, which was loaded as a size marker as well (Lane 1). dRF: double replicative form; mRF: monomeric replicative form; ssDNA: single-stranded DNA; MitoDNA: mitochondrial DNA are indicated. (B) mRF quantification. mRF data were normalized to mitochondrial DNA bands that serve as loading controls. Relative folds of intensity are shown. Means and standard deviations were calculated based on three independent experiments. *, P < 0.05; and ns, no significant difference. (C) Northern blot analysis. Total RNA was extracted from AAV5-infected cells using TRIzol reagent.Ten micrograms RNA of each sample was separated on 1% agarose gel, ethidium bromide-stained, and visualized under a UV light for the 28S and 18S rRNAs, as well as the ssRNA ladder. The gel was then transferred onto a nitrocellulose membrane, followed by hybridization with a cap and rep probes, which were digested from pAAV5. P7, P19, and P41 mRNAs correspond to mRNAs transcribed from P7, P19, and P41 promoters, respectively. (D) Western blot analysis. Infected cells were lysed and immunoblotted for AAV5 VP and Rep proteins, MAAP5 and AAP5, as indicated. β-actin is shown as a loading control.

All these data indicated that ablation of MAAP2 and MAAP5 significantly increases viral DNA replication (by 5.4-fold and 2.1-fold, respectively), but not much in the expression of viral mRNAs and proteins.

MAAP expression decreases total AAV production and increases virus egress from the infected cells

To examine the progeny virus production in wtAAV2 and AAV2^∆MAAP-infected or wtAAV5 and AAV5^∆MAAP-infected cells, we co-transfected the infected cells with pHelper and pCI-empty or pCI-MAAP2/5 as indicated. Expression of AAV Rep proteins induces cell cycle arrest and apoptosis (43, 44). To ensure a nonlytic infection, at 2 dpi, cells and the media were collected for virus extraction and quantification using an AAV2 rep-specific probe and a AAV2 cap-specific probe, respectively, for AAV2 or an AAV5 rep-specific probe for AAV5. AAV2^∆MAAP infection showed 4.6-fold and 2.8-fold increases in progeny virions in the cell pellet and total virus yield, respectively, compared to wtAAV2 infection (Fig. 9A and C). Notably, the number of released viruses into the media showed a 52-fold decrease in AAV2^∆MAAP-infected cells (Fig. 9A and C). The released virus number was decreased by 4.9-fold in AAV5^∆MAAP-infected cells, whereas total virus yield increased by ~1.5-fold (Fig. 10A). Surprisingly, the trans-complementation of the MAAP resulted in a successful restoration of MAAP function (Fig. 9B, D and 10B)

Fig 9.

MAAP2 expression decreases total AAV2 production and increases virus egress from the infected cells. HEK293 cells were infected with wtAAV2 or mutant virus AAV2^ΔMAAP followed by co-transfection of pHelper with equal molar pCI-empty (─) or pCI-MAAP2 (+), as indicated. At 2 dpt or dpi, infected cells and media were harvested for subsequent experiments. (A&C) Virus production. Progeny viruses were extracted from the cell pellet and medium, respectively, and quantified by qPCR using an AAV2 rep-specific probe (A) and an AAV2 cap-specific probe (C). Related bars represent virus genomic copies (vgc) in the pellet, medium, and total (pellet plus medium), respectively. (B&D) The ratio of the progeny viruses resided in-cells to in-media. Bars indicate progeny virus yield ratios in-cells vs in-media based on the results determined by qPCR using an AAV2 rep-specific probe (B) and an AAV2 cap-specific probe (D), respectively. Means and standard deviations were calculated based on three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001.

Fig 10.

MAAP5 expression decreases total AAV5 production and increases virus egress from the infected cells. HEK293 cells were infected with wtAAV5 or mutant virus AAV5^ΔMAAP followed by co-transfection of pHelper with pCI-empty (─) or pCI-MAAP5 (+), as indicated. At 2 dpt or dpi, infected cells and media were harvested for subsequent experiments. (A) Virus production. Progeny viruses were extracted from the cell pellet and medium, respectively, and quantified by qPCR using an AAV5 rep-specific probe. Related bars represent vgc in the pellet, medium and total (pellet plus medium), respectively. (B) The ratio of the progeny viruses resided in-cells to in-media. Bars indicate progeny virus yield ratios in-cells vs in-media based on the results determined by qPCR using an AAV5 rep-specific probe. Means and standard deviations were calculated based on three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001.

Collectively, these data showed that both MAAP2 and MAAP5 decrease total AAV production and facilitate virus egress out of the infected cells during AAV infection.

DISCUSSION

In this study, we showed that both MAAP2 and MAAP5 are expressed from the D-A2-spliced cap mRNA. During nonlytic AAV2 infection, MAAP2 is mostly expressed in the cytoplasm with nuclear localization to some extent. In addition, AAV^∆MAAP replicated more efficiently than wtAAV, which was consistent with their viral DNA replication. MAAP5 is expressed from a noncanonical start codon AUU at nt 2,232, which extends 17 aa of the N-terminus of MAAP5. Importantly, MAAP5 abundantly localizes in the nucleus, highlighting the function of the MAAP in viral DNA replication, in addition to its role in virus egress out of the infected cells.

AAV uses a quadcistronic cap mRNA to translate viral proteins

Although the AAP was identified over 10 years ago and the MAAP 5 years ago, it was unknown which cap mRNA encodes the AAP and MAAP. This study clearly demonstrated that AAV2 P40-SplA2 mRNA encodes MAAP, VP2, AAP and VP3 in order, and AAV2 P40-SplA1 mRNA encodes VP1, as well as AAP, but weakly. Since the different antibodies were used to probe the MAAP, AAP, and VP, the exact expression levels of individual proteins remain unknown. The noncanonical translation initiation is controlled by the scanning mechanism of eukaryotic translation, in which the ribosome binds the 5’-end of mRNA through the interaction of its m7G cap with initiation factor 4F (eIF4F) and initiates translation from the start codon closest to the cap. If the start codon is weak or noncanonical, only a part of the ribosome translates the ORF, and the remainder of the ribosome resumes scanning of the mRNA for the next start codon (46–48). For AAV2 P40-SplA2 mRNA, the noncanonical start codons, CUG for MAAP, ACG for VP2, and CUG for AAP, located upstream of the AUG start codon of VP3 confer AAV2 P40-SplA2 mRNA to encode four proteins (Fig. 2A). Although AAV5 P41-SplA3 mRNA also encodes four proteins, MAAP, VP2, AAP, and VP3, in order (Fig. 2B), MAAP5 was expressed from an unusual non-AUG codon AUU, located immediately after the A2 splice acceptor. AAV5 P41-SplA1 mRNA encodes VP1 and also AAP5 but weakly.

Most eukaryotic mRNAs are monocistronic; however, polycistronic mRNAs exist and mostly only translate to two to three proteins, including some viral mRNAs (49). We previously revealed the R2 mRNA of Aleutian mink disease parvovirus (AMDV) is a tricistronic mRNA that is translated to a nonstructural protein NS2 and two structural proteins, VP1 and VP2 (50, 51). Therefore, the nature of the AAV cap gene (SplA2) mRNA as a quadcistronic mRNA is unique among eukaryotic gene translation with the most known cistrons encoded in eukaryotic cells. Further investigation underlying the regulation of both in cis and in trans is warranted.

MAAP likely negatively regulates AAV DNA replication and virus production

MAAP ablation increases the total production of progeny viruses, which is likely related to the increase in viral DNA replication. The basic-amino-acid-rich (BR) presence of the MAAP at the C-terminus may function as a nuclear localization (NLS) (34) and lead to nuclear localization of the MAAP. It is interesting that the nuclear expression of MAAP5 was much more obvious than that of MAAP2, which perhaps is related to the extended 17 aa at the N-terminus. The decrease in viral DNA replication might be a consequence of the MAAP localized in the nucleus. Without MAAP expression, fewer progeny virions are egressed out of the infected cells, which may result in more virions stuck in the nucleus and thus more viral DNA templates for replication. However, others have observed that MAAP ablation or early termination decreased AAV2 production in an AAV2 infectious clone transfected cells infected with Ad (34). We believe that this discrepant observation is probably attributable to the lytic infection system employed by them. As the MAAP is not colocalized with the viral DNA replication center (data not shown), we hypothesize that the MAAP plays an indirect role in negative regulation of viral DNA replication. Lately, the screening of an MAAP expression library has revealed mutants that exhibit a gradual enhancement in the overall AAV titer (52), thereby substantiating the role of the MAAP in AAV replication.

Overall, our studies identified the AAV VP-coding mRNAs are multicistronic mRNAs using translation noncanonical initiation codons. While we verified the role of the MAAP in the cellular egress of both AAV2 and AAV5 in their nonlytic virus infection systems, we found that the MAAP localizes not only in the cytoplasm but also in the nucleus. The MAAP plays a negative role in viral DNA replication and thereafter progeny virus production during nonlytic AAV infection.

MATERIALS AND METHODS

Cells and culture

Human embryonic kidney HEK293 cells (293AAV) were obtained from Cell Biolabs, Inc. San Diego, CA. The cells were maintained in Dulbecco’s modified Eagle medium (DMEM) (#SH30022; Cytiva Life Science, Marlborough, MA) supplemented with 10% fetal bovine serum (FBS; #F0926, MilliporeSigma, St. Louis, MO) and 100 units of penicillin–streptomycin.

Plasmid construction

AAV cap-expressing plasmids

pCMVAAV2Cap is a CMV (immediate early, IE) promoter-driven AAV2 cap gene expression plasmid described previously (53). It was used as a parental plasmid to generate two cDNA constructs of AAV2, pCMVAAV2Cap^D/A2 and pCMVAAV2Cap^D/A2, and two synthetic chimeric intron (Sintron)-replaced constructs, pCMVAAV2Cap^{D^A1} and pCMVAAV2Cap^{D^A2}, using HiFi DNA assembly (NEB, #E2621) and synthesized DNA at Twist Bioscience (South San Francisco, CA). The Sintron is the chimeric intron taken from pCI (Promega, Madison, WI). The pP41AAV5Cap is an AAV5 P41 promoter-driven AAV5 cap gene expression plasmid in which P41 is a genuine AAV5 promoter (35). It was used to generate cDNA constructs of AAV5, pP41AAV5Cap^D/A1, pP41AAV5Cap^D/A2, and pP41AAV5Cap^D/A2∆ATT. pP41AAV5MAAP^Flag was made by fusing a Flag tag at the C-terminus of MAAP ORF and was used as a parent plasmid to generate these mutants, pP41AAV5 MAAP^∆VP1ATG, pP41AAV5MAAP^∆TTG, pP41AAV5MAAP^∆ATT, pP41AAV5MAAP^∆GTT&TGT, pP41AAV5MAAP^∆GTG, and pP41AAV5MAAP^∆CTG by using HiFi DNA assembly.

AAV rep and cap gene-expressing plasmids

AAV2 RepCap (R2C2) plasmid contain the AAV2 genome nt 144–4,492, and AAV5 RepCap (R5C5) contains the AAV5 genome nt 185–4,448, which have been described previously (23, 28). R2C2^D/A1 and R2C2^D/A2 were constructed by deletion of the introns D-A1 and D-A2, respectively, in R2C2, and R2C2^{D^A1} and R2C2^D/A2 were constructed by replacement of the introns D-A1 and D-A2, respectively, with the Sintron. R5C5^D/A1 and R5C5^D/A2 were constructed by deletion of the introns D-A1 and D-A2, respectively, and R5C5^D/A2∆ATT was constructed by mutation of the MAAP initiation codon ATT to AGG in R5C5^D/A2.

AAV infectious clone and mutants

SSV9 (pSub201), an AAV2 infectious clone, containing a full-length AAV2 genome (GenBank accession #AF043303) (45) was a kind gift from Dr. R. J. Samulski at the University of North Carolina, Chapel Hill (28). The SSV9^∆MAAP plasmid was generated by introducing a single-nucleotide change in the 19th amino acid leucine using HiFi DNA assembly, which results in a stop codon in MAAP2 ORF. pAAV5 is an AAV5 infectious clone (pAV5) that contains a full-length AAV5 genome (GenBank accession #AF085716) (23). The pAAV5^∆MAAP plasmid was generated by introducing a single-nucleotide change in the 36th amino acid leucine using HiFi DNA assembly, which resulted in a stop codon in MAAP5 ORF. The pHelper plasmid was obtained from Agilent Technologies, Inc., Santa Clara, CA (#240071).

MAAP expressing plasmids

The CMV IE promoter-driven pCI mammalian expression vector was purchased (Promega, Madison, WI). CUG-started MAAP2 ORF and AUG-started MAAP5 were cloned into the parental plasmid to generate pCI-MAAP2 and pCI-MAAP5, respectively. pCI-MAAP2^Flag was constructed by addition of a Flag tag to the C-terminus of MAAP2 ORF in pCI-MAAP2.

Wild-type (wt)AAV and MAAP knockout mutant production

wtAAV2 or AAV2^∆MAAP were produced as described previously (54) with modifications. SSV9, SSV9^∆MAAP, pAAV5, and pAAV5^∆MAAP plasmids were transfected in HEK293 cells with pHelper at a 1:1 molar ratio using PEI Max (#24765, Polysciences) at a 1:3 (DNA:PEI Max) ratio. At 2 days post-transfection, the cells were harvested, lysed, and treated with deoxycholate (10%) and DNase I (4 mg/mL). The lysate was clarified by the addition of CsCl, followed by CsCl ultracentrifugation for two rounds. Finally, purified viruses were dialyzed against PBS buffer, and the titers (viral genome copies, vgc, per mL) were quantified by quantitative real-time (qPCR) using rep- or cap-specific probe after treatment with benzonase (#71205–3, MilliporeSigma).

Virus infection

HEK293 cells were seeded on culture plates/dishes based on following experiments 1 day before infection. On the day of infection, cells were infected with the virus at a multiplicity of infection (MOI) of 10,000 vgc/cell of relevant viruses in 0.25 vol of a proper culturing medium at room temperature on a rocking platform. After an hour of incubation, the virus-containing medium was removed, and cells were washed twice with DPBS. The cells were then replaced with fresh media followed by transfection of indicated plasmids using PEI Max at a ratio of 1:3 (DNA:PEI Max). At 2 dpi, cells or medium were collected and treated as indicated in the figure legends.

Quantification of progeny virions in cells and in the media

HEK293 cells of one well of a 6-well plate were infected with wtAAV2/5 and AAV2/5^∆MAAP, followed by transfection of 1.5 µg of pHelper with an equal molar of pCI-MAAP2/5 or pCI-empty. The cells were harvested at 2 dpi, and the media were collected in separate tubes. The cells were resuspended in 500 µL of 10 mM Tris-HCl (pH 8.0). Resuspended cells and 500 µL of 2 mL media were then lysed and treated with deoxycholate (10%) and DNase I (4 mg/mL). DNA was extracted from the crude lysate using the Quick-DNA//RNA Pathogen kit (#R1042, Zymo Research) and quantified by qPCR using the rep- or cap-specific probe for AAV2 or AAV5 after treatment with Benzonase (#71205–3, MilliporeSigma).

Southern blotting

Hirt DNA preparation and Southern blotting were essentially performed as described previously (55). Briefly, at 2 dpi, harvested cells were processed for Hirt DNA extraction followed by Dpn I digestion. Digested DNA samples and a size marker (an AAV2 dsDNA genome digested from SSV9 for AAV2 and an AAV5 dsDNA genome digested from pAAV5 for AAV5) were run in 1% agarose gel overnight. Next, resolved DNAs on the gel were transferred to a nitrocellulose membrane (#1212590, GVS) and hybridized with an [^α-32P] dCTP-labeled SSV9 (undigested) or AAV5 dsDNA genome probe (digested from pAAV5). Phosphoscreen-captured hybridization signals were scanned and visualized on an Amersham Typhoon Biomolecular Imager (Cytiva). mRF, Dpn I-digested bands and Mito-DNA bands were quantified using ImageJ (56).

RNA isolation and Northern blotting

At 2 dpi, infected HEK293 cells were harvested, and total RNA from indicated samples was extracted using TRIzol (#15596026, Invitrogen) according to the manufacturer’s instructions.

Northern blotting was performed as reported previously (50, 57). Briefly, 10 µg RNA for each sample was run on a formaldehyde-containing agarose (1%) gel overnight. The gel was then stained with ethidium bromide for visualization of the 28 s and 18 s rRNA, as well as ssRNA ladder (NEB, # N0362S). Resolved RNAs were then transferred to a nitrocellulose membrane and hybridized with an [^α-32P] dCTP-labeled SSV9 (undigested)- or an AAV5 rep or AAV5 cap gene probe. Phosphoscreen-captured hybridization signals were scanned and visualized on an Amersham Typhoon Biomolecular Imager (Cytiva).

Western blotting

At 2 dpi or post-transfection, the cells were harvested, lysed, and separated in Tris–Glycine precast gel (NuSep Inc, #NN10-816, #NB10-816, #NB10-420) with a protein ladder (#P008, GoldBio). After transferring the proteins on a polyvinylidene difluoride (PVDF) membrane (#IPVH00010, MilliporeSigma), the membrane (blot) was blocked in 5% non-fat dry milk containing TBS-T for 1 hour. The blot was then incubated with an antibody diluted in 1% non-fat dry milk containing TBS-T overnight. After washing, the blot was incubated with an infrared dye-conjugated secondary antibody for 1 hour and visualized on an Odyssey imaging system (LI-COR Biotechnology, Lincoln, NE).

Reverse transcription and qPCR quantification

At 2 dpt, cells were harvested, and total RNA was extracted. The RNA samples were treated with RNase-free DNase (#M6101, Promega) according to the manufacturer’s instructions. DNase-treated total RNA (2 µg) was reverse-transcribed to generate the first-strand cDNA using Moloney Murine Leukemia Virus Reverse Transcriptase (M-MLV RT) (#M1705, Promega) according to the manufacturer’s instructions.

The generated first-strand cDNAs were then quantified by multiplex qPCR, as described previously (55, 58, 59). Briefly, D-A1- or D-A2-spliced mRNA-specific probes (FAM-labeled) were used for quantification, and a β-actin-specific probe (JOE-labeled) was employed as an internal control. The sequences of D-A1- or D-A2-spliced mRNA- and β-actin-specific probes and primers are presented in Table S1. Relative gene expressions were analyzed with the 2^–∆∆Ct method (60). The 2^–∆∆Ct method is an interpretation of relative quantification to the untreated control as a fold change of a target gene expression that was already normalized to the internal control. Briefly, the internal control β-actin mRNA was used to normalize the mRNAs from the reverse transcription in one qPCR reaction. The RNA sample generated from pCMVAAV2Cap-transfected cells was employed as a reference.

Immunostaining and confocal microscopy

At 2 dpi, cells were harvested and cytospun on slides. The cells were then fixed (4% PFA) for 15 minutes and permeabilized (0.5% Triton-X-100) for 5 minutes. Then, the cells were incubated with the first antibody for 1 hour and the second antibody for another hour, after blocking for half an hour with 2% BSA. Between each treatment step, cells were washed with PBS for 5 minutes. Prepared slides were observed and imaged under a Leica TCS SP8 STED 3 × Super Resolution Microscope and processed with LAS X Life Science Microscope Software (Leica).

Antibodies

First antibodies

The anti-MAAP2 antibody was produced by immunization of purified GST-fused MAAP2 in rats according to a reported protocol (61, 62). One anti-AAP2 antibody was a kind gift from Dr. Dirk Grimm (63), and another one was produced by immunization of an AAP2 peptide (Cys-RSTSSRTSSARRIKDASRR)-conjugated KLH (keyhole limpet hemocyanin) in the rabbit at Biomatik (Wilmington, DE). Anti-MAAP5 and anti-AAP5 were produced by immunization of an MAAP5 peptide (Cys-GETSERQSFRPRKGFSN) and an AAP5 peptide (GPRDACRPSLRRSLRCRS-Cys)-conjugated KLH (keyhole limpet hemocyanin), respectively, in the rabbit at Biomatik (Wilmington, DE). Anti-VP (#03–61084 for Western blotting of both AAV2 and AAV5), anti-AAV2 intact capsid (#03–65255), anti-AAV5 intact capsid (#03–610149), and anti-Rep (#03–61069 for Western blotting of both AAV2 and AAV5) were purchased from ARP (Waltham, MA). Anti-Flag (#200–301-B13) was purchased from Rockland (Limerick, PA).

Secondary antibodies

Anti-Rat DyLight 800 (#SA5-10024) was purchased from Invitrogen. Anti-Rabbit DyLight 800 (#5151S) and anti-mouse DyLight 800 (#5257S) were purchased from Cell Signaling (Danvers, MA). Alexa Fluor 488-conjugated anti-Rat (#A48262), Alexa Fluor 488-conjugated anti-Mouse (#A11001), Alexa Fluor 555-conjugated anti-Mouse (#A32773), Alexa Fluor 594-conjugated anti-Rat (#A48271) and Alexa Fluor 594-conjugated anti-Rabbit (#A32754) were purchased from Invitrogen (Waltham, MA).

Statistical analysis

GraphPad Prism 9 (GraphPad Software) was used for statistical analysis. Means and standard deviations (SD) were calculated based on three independent experiments. Statistical significances (P-value) were determined by Student’s t test (n.s., no statistically significant difference; *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001).

DATA AVAILABILITY

All data needed to evaluate the conclusions are present in the paper and the supplemental material.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jvi.00633-24.

Table S1. jvi.00633-24-s0001.docx.

Sequences of qPCR primer-probe sets.

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.