Adeno-associated virus DNA replication is induced by genes that are essential for HSV-1 DNA synthesis

Adeno-associated virus (AAV) DNA replication is not detectable unless cells are coinfected with a helper adenovirus (Ad) or herpesvirus or unless AAV infection is carried out in certain established cell lines that have been treated with various metabolic inhibitors or uv irradiation. In helper-dependent infections, it has been shown that AAV DNA synthesis depends on one or more early Ad genes, whereas little is known concerning any herpesvirus gene that promotes AAV DNA synthesis. In this study we tested the ability of four cloned Xbal fragments of herpes simplex virus type 1 (HSV-1) DNA to induce AAV DNA synthesis in Vero cells. Cotransfections, which were carried out with pAV1 (an infectious AAV2 plasmid), revealed that AAV DNA synthesis could be optimally induced by three of these clones (C,D, and F) plus a clone of the HSV-1 ICP4 (IE 175) gene. ICP4, an immediate early gene, was presumably required to activate expression of other HSV genes. To help identify the additionally needed HSV genes, we tested Xbal C,D, and F subclones that contain genes previously found necessary for origin-dependent HSV DNA synthesis and found that at least five of these genes (UL 5, 8, 9, 29, and 30) contributed to the induction of AAV DNA synthesis. In contrast to their absolute requirement for HSV DNA synthesis, none of these genes were strictly necessary for AAV DNA replication. Because they are all known to specify proteins that are directly involved in HSV DNA synthesis, our results suggest that some or all of their products also may directly participate in the replication of AAV DNA.     

Studies on the defectiveness of adeno-associated virus (AAV). I. Effects of phosphonoacetic acid and 2-deoxy-D-glucose on the replication of AAV.

Adeno-associated viruses (AAV) are defective parvoviruses which require the presence of a helper virus, either an adenovirus or a herpesvirus, to initiate their replication cycle. The mechanisms by which these helper viruses can promote AAV macromolecular synthesis have been investigated in systems where the replication of the helper virus was altered. In the first system, phosphonoacetic acid (PAA), a specific inhibitor of herpesvirus-coded DNA polymerase, was used in AAV-herpes simplex virus (HSV) coinfections to determine what effect this drug would have on the replication of AAV. It was found that in the presence of increasing concentrations of PAA, the synthesis of AAV DNA, structural proteins, and immunofluorescent (IF) capsid antigens was inhibited. However, when an adenovirus helper was used in place of HSV, this inhibition was not seen. It was also found that the addition of the drug 5 hr after infection was still effective in inhibiting AAV capsid antigen synthesis completely. This finding indicates that the PAA-sensitive event required by AAV occurs relatively late in the HSV cycle. Restoration of HSV DNA polymerase activity by reversal of a PAA block was not sufficient for initiating AAV replication. De novo protein synthesis was required also. In the second system, the effects of 2-deoxy-d-glucose (2DG), an inhibitor of protein glycosylation, on AAV replication were also examined. In AAV coinfections with HSV, increasing concentrations of 2DG inhibited the production of AAV IF capsid antigens. Conversely, production of AAV intracellular proteins was enhanced with increasing 2DG concentrations. Under these conditions the pattern of IF staining for the major AAV polypeptide was normal. Thus 2DG appears to interfere with the assembly of AAV proteins into a capsid configuration. In experiments with an adenovirus helper, 2DG was found to inhibit initiation of AAV replication through early adenovirus functions.     

Ubiquitous human adeno-associated virus type 2 autonomously replicates in differentiating keratinocytes of a normal skin model

Since its discovery in 1966, adeno-associated virus type 2 (AAV) has been described as a helper-dependent parvovirus. However, in this study we demonstrate that AAV undergoes its complete life cycle, devoid of helper viruses or genotoxic agents, in the organotypic epithelial raft tissue culture system, a model of normal skin. AAV progeny production directly correlated with epithelial differentiation, as nondifferentiating keratinocytes were defective for this activity. Large nuclear virus arrays of particles of approximately 26 nm (parvovirus size) were observed in the granular layers of the raft epithelium by electron microscopy. Additionally, dosage-dependent histologic changes, some of which might be interpreted as cytopathology, were induced in the AAV-infected epithelial tissues. These data suggest a new biological model for AAV; that is, AAV is an epithelial-tropic autonomous parvovirus that can alter normal squamous differentiation.     

Adeno-associated virus autointerference.

We have analyzed an autointerference phenomenon exhibited by adeno-associated virus type 2 (AAV) when grown in KB cells coinfected with adenovirus type 2 as the helper. Infectious AAV particles that banded at 1.41 g/cm³ in CsCl were purified by three cycles of centrifuging in CsCl equilibrium gradients. When cells were infected with an increasing multiplicity of these AAV particles there was a corresponding decrease in production of infectious progeny AAV. There was also an AAV multiplicity-dependent inhibition of production of infectious adenovirus and inhibition of Ad DNA replication. The viral DNA in the Hirt supernatant fraction extracted from cells infected with different multiplicities of AAV was analyzed in neutral sucrose gradients. At low multiplicities of infection with AAV, the main AAV DNA species synthesized was the mature 14.5 S (standard) viral genome. In higher multiplicity infections with AAV increasing amounts of aberrant 10 S AAV DNA molecules accumulated and the proportion of 14.5 S AAV DNA decreased. Restriction endonuclease cleavage showed that the 10 S DNA was enriched for the left- or right-hand terminal regions of the AAV genome. These molecules may be analogous to the previously characterized aberrant DNA molecules found in light-density AAV particles. Thus, the AAV autointerference is correlated with production of the aberrant deleted AAV genomes.     

Structural polypeptides of adenovirus-associated virus top component.

Low density adenovirus-associated virus (AAV) from upper AAV-adenovirus bands on isopycnic CsCl gradients were purified by sedimentation through sucrose gradients and by immunoprecipitation of filtered particles. These particles appeared to be “empty” capsids because of their uptake of negative stain and their lack of isotopically labeled DNA. Analysis of these particles by electrophoresis on sodium dodecyl sulfate-polyacrylamide gels demonstrated three polypeptides. These polypeptides showed identical molecular weights and the same relative concentrations as the structural polypeptides of marker complete AAV virions when compared by this technique. These findings indicate that the apparently “empty” AAV capsid does not lack any of the three polypeptides found in the complete virus particle, suggesting that none of them are core proteins.     

Viral DNA synthesis in vitro with nuclei isolated from adeno-associated virus type 1-infected cells

Electron microscope autoradiograms revealed that grains (viral DNA) were found in the nuclei of cells coinfected with adeno-associated virus type 1 (AAV1) and a temperature-sensitive mutant of human adenovirus type 31 (H31tsA13) at the nonpermissive temperature (40°). Nuclei isolated from cells coinfected with AAV1 and H31tsA13 at 40° were active in AAV1-DNA synthesis in vitro. Alkaline sucrose gradient analysis of the in vitro products showed that DNA synthesized with nuclei isolated from coinfected cells sedimented faster than marker AAV1-DNA, although DNA synthesized with nuclei isolated from cells coinfected in the presence of cycloheximide cosedimented with the marker DNA. This fast-sedimenting DNA shifted to the position of the marker DNA after treatment with papain, trypsin, and sodium dodecyl sulfate followed by extraction with phenol. This observation suggests the involvement of a protein in the formation of fast-sedimenting viral DNA.     

Analysis of proteins, helper dependence, and seroepidemiology of a new human parvovirus

A new type of defective parvovirus, tentatively designated as adeno-associated virus type 5 (AAV-5), is characterized as far as its proteins, its helper dependence, and its seroepidemiology are concerned. The protein analysis of AAV-5 in polyacrylamide gels demonstrated the presence of three structural polypeptides, corresponding to VP 1, VP 2, and VP 3 of other AAV types. The preparation of monoclonal antibodies against AAV-5 permitted the analysis of viral structural antigen expression by using adenovirus type 12 (Ad 12) or several herpes group viruses as helper viruses, respectively. AAV-5-infected cell cultures coinfected with either Ad 12, Herpes simplex virus (HSV), Cytomegalovirus (CMV), or Varicella Zoster virus (VZV) efficiently synthesize AAV-5 specific antigens. Epstein-Barr virus (EBV) and Herpesvirus saimiri, in contrast, provide only a very weak helper activity for AAV 5 antigen expression. The development of a specific ELISA test permitted screening of human sera for antibodies to AAV-5. Forty-five percent of 926 sera from all age groups and approximately 60% of the adult population reveal antibodies to structural components of this virus. The seroepidemiology differs from that reported for other AAV serotypes. Highest average titers against AAV-5 are observed in the age group between 15 and 20 years. Sera from patients with cervical carcinoma revealed average titers of antibodies well below those of age-matched control groups. Attempts to find higher antibody levels against AAV-5 in specific human diseases failed thus far.     

A 110-kDa nuclear shuttle protein, nucleolin, specifically binds to adeno-associated virus type 2 (AAV-2) capsid.

A 110-kDa protein was copurified with adeno-associated virus type 2 (AAV-2) virions after CsCl density gradient isopycnic centrifugation. Amino acid sequence of peptides derived from this protein after tryptic digestion, monoclonal antibody production, and Western blot analysis showed that the copurified protein was the major nucleolar phosphoprotein, human nucleolin. Virus overlay assays demonstrated that AAV-2 capsid specifically bound to the human nucleolin, and immunoprecipitation studies confirmed the in vitro binding of nucleolin and intact AAV-2 capsids but not denatured viral proteins. Double-immunofluorescence staining of infected cells showed that AAV capsid and nucleolin were colocalized in both cytoplasm and nucleus. In addition, when cytoplasmic and nuclear fractions were extracted from AAV-infected KB cells at different time points postinfection, immunoprecipitation data and Western blotting showed that AAV capsid formation and nucleolin interact specifically and share their subcellular localization in infected cells. With the known functions of nucleolin in the synthesis of rRNA and ribosome assembly, binding to single-stranded DNA, and acting as a shuttle between cytoplasm and nucleolus, our data showing that AAV-2 capsid binds specifically to nucleolin both in vitro and in vivo suggest a key role of nucleolin in AAV-2 replication, particularly in capsid assembly.     

Adeno-associated virus and development of cervical neoplasia.

Evidence from several sources has suggested that adeno-associated virus (AAV) infection might protect against cervical cancer, in part, by interfering with human papillomavirus (HPV)-induced tumorigenesis. Detection of AAV type 2 (AAV-2) DNA in cervical tissues has been reported. However, there have been few in vivo studies of women with cervical HPV infection or neoplasia, and these have reported inconsistent results. Therefore, we used polymerase chain reaction (PCR) assays targeted to the AAV-2 rep and cap genes to test tissue specimens from women in an epidemiological study of cervical neoplasia in Jamaica. We tested 105 women with low-grade cervical intraepithelial neoplasia (CIN-1), 92 women with CIN-3/carcinoma in situ or invasive cancer (CIN-3/CA), and 94 normal subjects. PCR amplification of human beta-globin DNA was found in almost all cervical specimens, indicating that these materials were adequate for PCR testing. The prevalence of HPV DNA, determined by HPV L1 consensus primer PCR was, as expected, strongly associated with presence and grade of neoplasia. Each of the AAV PCR assays detected as few as 10 copies of the virus genome. However, none of the 291 cervical specimens from Jamaican subjects tested positive for AAV DNA. Negative AAV PCR results were also obtained in tests of cervical samples from 79 university students in the United States. Exposure to AAV was assessed further by serology. Using a whole virus AAV-2 sandwich enzyme-linked immunosorbent assay, we found no relationship between AAV antibodies and presence or grade of neoplasia in either the Jamaican study subjects or women enrolled in a U.S. cervical cancer case (n = 74) -control (n = 77) study. Overall, the data provide no evidence that AAV infection plays a role in cervical tumorigenesis or that AAV commonly infects cervical epithelial cells.     

Dissociation of carcinogen-induced SV40 DNA-amplification and amplification of AAV DNA in a Chinese hamster cell line

Time kinetics of AAV-5 DNA amplification induced by MNNG in cells of a Chinese hamster line (CO631) and a Syrian hamster line (Elona) were compared to kinetics of SV40 DNA amplification in these cells. AAV-5 DNA amplification starts in both lines about 8 hr following treatment of AAV-infected cells with MNNG. SV40 DNA amplification is induced only in CO631 cells. The time it becomes detectable varies but in no case was before 23 hr after MNNG treatment. In CO631 cells a second round of AAV DNA amplification takes place. It starts at times paralleling the MNNG-induced synthesis of SV40 DNA but appears to be independent of genotoxic treatment. Elona cells amplify neither SV40 DNA after exposure to chemical carcinogens nor AAV-5 DNA in untreated cells. In both lines combined treatment with MNNG and AAV-5 resulted in marked cytopathogenic changes and cell death. In the test systems used here viral DNA amplification did not lead to synthesis of infectious virus, thus, the viral cycle remained abortive.     

Polymorphism of endogenous murine leukemia viruses revealed by isoelectric focusing in polyacrylamide gels.

The synthesis of adeno-associated virus (AAV2) RNA in KB3 cells coinfected with adenovirus type 2 as a helper has been studied. Previous studies revealed a discrete 20 S AAV RNA species which was present in the nucleus and polysomes and a second, heterogenous population of smaller AAV RNA molecules (4 to 18 S) present only in the nucleus and nonpolysomal regions of the cytoplasm. In the present study, the AAV genome sequence representation and polyriboadenylate [poly(A)] sequences in AAV RNA were correlated with the size of the RNA and its cellular distribution. Most of the 20 S AAV RNA contained a poly(A) sequence of 200 nucleotides in length, whereas the heterogenous 4 to 18 S AAV RNA contained little or no poly(A). Both the poly(A)(+) and the poly(A)(?) RNA, as well as RNA isolated from the cell nucleus, cytoplasm, or polysomes, contained the same set of AAV RNA sequences complementary to 70 to 75% of the AAV DNA minus strand. These results indicate that a single polyadenylated AAV mRNA species is synthesized and that this represents most, if not all, of the entire portion of the AAV genome that is stably transcribed.     

Temporal acceleration of the human papillomavirus life cycle by adeno-associated virus (AAV) type 2 superinfection in natural host tissue

Epidemiologically, certain human papillomaviruses are positively associated with cervical cancer, while adeno-associated viruses (AAV-2) are negatively associated with this same cancer. Both HPV and AAV productively replicate in differentiating keratinocytes of the skin and interact with each other. However, AAV has a relatively fast life cycle, generating infectious progeny by the third to fourth day of an organotypic epithelial raft culture. In contrast, HPV is slow, generating infectious progeny only after 10-12 days. As earlier studies indicated that these two skin-tropic virus types significantly affect each other's life cycle, we investigated if the temporal kinetics of the slow HPV life cycle was affected by the fast AAV in raft cultures. Here it is shown that the presence of AAV-2 at a variety of multiplicities of infection (m.o.i.) resulted in early onset HPV-31b DNA replication. Using plasmids which each expressed only one of the four rep proteins, an enhancement affect was seen for all four rep proteins of AAV, with Rep40 having the highest activity. Furthermore, AAV (m.o.i. of 5) also resulted in a temporally accelerated production of HPV infectious units, seen as early as Day 4, with high levels of viral progeny being produced by Day 6.5. Like earlier studies at Day 12, histological differences were seen at Day 6.5 between AAV-infected and mock-infected HPV/rafts. These data suggest that under specific conditions the AAV rep trans-factors can positively regulate HPV gene expression in addition to the usual negative regulation that has been consistently observed by the rep proteins. These data also suggest that AAV has a significant effect upon the temporal kinetics of the HPV life cycle in natural host tissue. However, it is unclear if or how this AAV-induced fast HPV life cycle mechanistically correlates with lower rates of HPV-associated cervical disease.     

Mechanisms of adeno-associated virus genome encapsidation

The defective parvovirus, adeno-associated virus (AAV), is under close scrutiny as a human gene therapy vector. AAV's non-pathogenic character, reliance on helper virus co-infection for replication and wide tissue tropism, make it an appealing vector system. The virus' simplicity and ability to generate high titer vector preparations have contributed to its wide spread use in the gene therapy community. The single stranded AAV DNA genome is encased in a 20-25 nm diameter, icosahedral protein capsid. Assembly of AAV occurs in two distinct phases. First, the three capsid proteins, VP1-3, are rapidly synthesized and assembled into an empty virion in the nucleus. In the second, rate-limiting phase, single-strand genomic DNA is inserted into pre-formed capsids. Our rudimentary knowledge of these two phases comes from radioactive labeling pulse-chase experiments, cellular fractionation and immunocytological analysis of infected cells. Although the overall pattern of virus assembly and encapsidation is known, the biochemical mechanisms involved in these processes are not understood. Elucidation of the processes of capsid assembly and encapsidation may lead to improved vector production. While all of the parvoviruses share the characteristic icosahedral particle, differences in their surface topologies dictate different receptor binding and tissue tropism. Based on the analysis of the molecular structures of the parvoviruses and capsid mutagenesis studies, investigators have manipulated the capsid to change tissue tropism and to target different cell types, thus expanding the targeting potential of AAV vectors.