INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Retrovirus assembly involves precursor polyproteins encoded by the gag, pol, and env genes. The Gag polyprotein plays a central role in virion assembly and is itself sufficient to produce noninfectious virus-like particles. The Pol polyprotein is translated as a Gag-Pol fusion protein that is produced by either a ribosomal frameshift or a read-through of the termination codon. In several retrovirus subfamilies, including Moloney murine leukemia virus (MLV) and human immunodeficiency virus (HIV), Pol consists of protease (PR), reverse transcriptase (RT), and integrase (IN). Thus, assembly of the viral enzymes occurs in the context of a large Gag-Pol precursor polyprotein. During assembly of MLV and HIV virions, the dimerization of PR induces proteolysis and subsequently cleavage of the Gag and Gag-Pol precursors into their mature protein products. This results in a rearrangement of the cleaved Gag and Pol polypeptides within the virion, which leads to condensation of the virus core, a process that is essential for formation of infectious virions (for reviews, see references 27 and 29).

The retroviral IN protein plays a central role in the formation of proviral DNA. After the virus enters a target cell and reverse transcription of the RNA genome is completed, IN catalyzes integration of the nascent viral cDNA into the host cell genome. From extensive in vitro analysis, significant progress has been made in understanding the structure of the IN protein, the contribution of IN subdomains to enzyme activity, and the mechanism of the integration reaction. Using recombinant IN and oligonucleotides that represent the viral DNA ends, the in vitro integration reaction proceeds in two steps: IN removes two nucleotides from the 3' terminus of the linear DNA molecule (3'-end processing), which is then joined to the target DNA (strand transfer). Through amino acid sequence alignment and in vitro activity studies of wild-type and mutant IN, residues have been identified that are conserved among retroviral IN proteins, including those comprising a zinc finger-like motif (HHCC) located in the N terminus and a D,D(35)E motif required for polynucleotide transfer. Mutations in the D,D(35)E catalytic triad abolish strand transfer activity in vitro, indicating that this motif represents the catalytic site of IN. Mutations in the HHCC and D,D(35)E motifs of HIV type 1 (HIV-1) IN disrupt both 3'-end processing and strand transfer in vitro. The carboxyl terminus of IN is least conserved and possesses nonspecific DNA binding properties (for a review, see reference 4). The C terminus of the HIV-1 IN protein has been associated with nuclear import of the preintegration complex (13).

In vivo analyses of HIV-1 IN structure and function have suggested that the IN protein may play important nonenzymatic roles during the early stages of the virus life cycle (10). However, these in vivo studies with infectious virus (proviral DNA) have been complicated since mutations in IN affect the Gag-Pol precursor protein and often cause defects during the late stages of the virus life cycle. Numerous studies have reported IN mutations that cause defects in virion assembly, production, maturation, protein composition, or nuclear import of the preintegration complex (2, 3, 5, 6, 11, 22, 26). Some HIV-1 IN mutations impair viral DNA synthesis without an apparent effect on late-stage events (11, 15, 19). Similarly, studies of the IN of Ty3, a retrovirus-like element of Saccharomyces cerevisiae, identified nonconserved IN mutations that reduced the level of replicated DNA, despite normal levels of exogenous RT activity and capsid maturation (21).

To specifically analyze the function of the mature IN protein during early stages of the virus life cycle, we developed a trans-complementation approach. This enables the packaging of functional Pol proteins (RT and IN) into virions by their expression in trans, independently of the Gag-Pol precursor (17, 18, 31, 32). By expressing IN as a fusion partner of virus protein R (Vpr), fully functional IN can be incorporated into wild-type or IN mutant virions in trans (17, 32). Complementation studies indicated that the trans-IN protein rescued the defects in DNA synthesis caused by certain HIV-1 IN mutations, including those in the highly conserved HHCC motif (31). Complementation of this defect did not require the trans-IN protein to be enzymatically (integration) active, confirming that the IN protein itself augments DNA synthesis in vivo.

Several studies have analyzed MLV IN function in vivo by introducing mutations into the IN coding region of full-length proviral DNA. Nearly all of these mutations caused a severe loss of virus infectivity. Similarly to HIV-1, some mutations were also found to impair late-stage events of the virus life cycle (virion production or proteolytic processing of Gag and Pol), while others appeared to affect only early events (23-25). Southern blot analysis of unintegrated cDNA did not reveal a notable impairment in viral DNA synthesis (8, 9, 24, 25). These findings suggested that IN mutant viruses exhibited reduced infectivity as a result of a defect in the integration of proviral DNA. Since Southern blot analysis of unintegrated DNA measures only a portion of the total viral DNA synthesized, we reexamined the effects of IN mutations on MLV DNA synthesis in infected cells using sensitive, semiquantitative PCR-based methods. Our results indicate that complete deletion or C-terminal truncation of IN results in a reproducible reduction in viral cDNA synthesis. To rule out the possibility that this phenotype was due to negative effects of the mutations on a late-stage event of the virus life cycle, we used MLV Gag-IN fusion proteins to package IN into MLV virions in trans. This analysis demonstrated that trans-IN protein (derived from Gag-IN) rescued defects in both viral DNA synthesis and infectivity, confirming that the mature MLV IN protein itself augments reverse transcription in infected cells. This finding is consistent with those for HIV-1 and suggests a role of the IN protein in viral DNA synthesis that may be conserved among all retroviruses.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and antibodies. 293T and HeLa cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U of penicillin, and 0.1 mg of streptomycin/ml. Rabbit polyclonal antisera against the MLV Pol, p30^Gag, and RT proteins were generous gifts from Monica Roth, Alan Rein, and Stephen Goff, respectively.

Expression plasmids. Our analysis of IN was performed using the MLV pkat vector system described previously (12). IN mutations were introduced into the pkat-gag-pol packaging plasmid. To facilitate cloning of the various IN mutants, pkat-gag-pol was first modified by introducing a unique MluI site at the end of the IN coding region (Fig. 1A). To abolish the catalytic activity of IN, the pkat-IN^D184A plasmid was constructed. Partially overlapping 5' SalI-ScaI and 3' ScaI-MluI DNA fragments were PCR amplified from pkat-gag-pol. The internal sense primer replaced alanine with an aspartic acid residue at position 184 (D184A). The ScaI site was introduced by design of the internal primers and did not change the amino acid sequence. The pkat-IN^H61A plasmid was constructed to disrupt the zinc finger domain of IN. Partially overlapping 5' SalI-XhoI and 3' XhoI-MluI DNA fragments were PCR amplified from pkat-gag-pol. The internal antisense primer replaced alanine with a histidine residue at position 61 (H61A). The XhoI site was introduced via primer design without changing the amino acid sequence. The pkat-IN³⁴ plasmid deleted 34 amino acids from the carboxyl terminus of IN. It was constructed by ligating an SalI-MluI DNA fragment into pkat-gag-pol. The 3' mutagenic primer deleted 102 bp of IN sequence and introduced a TAA translational stop codon at amino acid position 374. The pkat-IN plasmid deleted the entire IN protein. It was constructed by ligating a PCR-amplified, RT-containing SalI-MluI DNA fragment into SalI-MluI-cut pkat-gag-pol. The 3' antisense primer introduced an MluI site and included a TAA translational stop codon after the last codon of RT. The pkat-RT-IN plasmid deleted the 3' end of Pol, including most of RT (301 residues) and all of the IN coding region. It was constructed by cutting pkat-gag-pol with SalI and MluI, filling in with Klenow enzyme, and then religating with T4 DNA ligase.

View larger version (26K): [in this window] [in a new window]   FIG. 1.   DNA construction of MLV packaging, vector, and trans-IN expression plasmids. (A) Packaging constructs. The IN coding region of the wild-type pkat-gag-pol plasmid was mutated as illustrated, generating the D184A, H61A, 34, and IN mutants. The pkat-RT-IN mutation deletes all of IN and the 3' half of RT. SD, splice donor sequence. (B) Gene transfer (vector) plasmids. The effects of IN mutations on infectivity and integration were analyzed using the pkat-GFP and pkat-GFP-puro gene transfer vectors. In both vectors, the GFP reporter is under the control of the CMV promoter. Using an IRES, the pkat-GFP-puro vector enables GFP and puromycin to be expressed from the same mRNA. (C) trans-IN expression plasmids. The pkat-gagNCIN, pkat-gagNCIN34, and pkat-gagNCIND184A plasmids fused gag (minus the NC domain) in frame with wild-type and mutant (34 and D184A) IN. At the gag-IN junction, 42 bp of RT sequence was included (darkened bar) to preserve the natural protease cleavage site at the N terminus of IN.

NC

NC

34

NC

NC

184A

NC

Virus infectivity and integration assays. Culture supernatants from 293T cells were collected 72 h after transfection, clarified by low-speed centrifugation (1,000 × g, 10 min), and filtered through 0.45-µm-pore-size sterile filters. HeLa cells grown in 24-well plates were infected with serial fivefold dilutions of supernatant in DMEM containing 1% FBS and 10 µg of DEAE-dextran/ml at 37°C for 4 h. The medium was then replaced with fresh DMEM containing 10% FBS. Two days later, GFP-positive (green) cell colonies were counted using a fluorescence microscope. Each GFP-positive colony was counted as 1 virus infectious unit. For the analysis of proviral integration, DMEM containing puromycin (4 µg/ml) was added to the infected HeLa cells and the number of resistant, GFP-positive cell colonies was counted after 14 days of selection.

Semiquantitative analysis of viral DNA synthesis. The PCR technique used to monitor the synthesis of viral DNA in infected cells was similar to that described earlier (31). Briefly, transfection-derived virus was filtered through 0.45-µm-pore-size filters and incubated with DNase I (4 µg/ml; Worthington Inc.) at 37°C for 1 h to minimize DNA contamination. One milliliter of each virus stock (normalized for RT activity or the density of capsid protein [p30] on Western blotting) was used to infect 1 million HeLa cells. The cells were washed twice with DMEM by low-speed centrifugation after 4 h and lysed after 18 h. Total DNA was extracted by organic methods, resuspended in 100 µl of distilled water, and treated with the DpnI restriction endonuclease to digest bacterially derived plasmid DNA. Four hundred nanograms of each DNA extract was subjected to 25 rounds of PCR amplification using primers designed to detect late (R-gag [sense nucleotides 1 to 22, 5'-GCGCCAGTCCTCCGATTGACTG-3', and antisense nucleotides 626 to 602, 5'-GCCCATATTCTCAGACAAATACAGAAA-3']) products of reverse transcription. PCR products were separated on 1.5% agarose gels and stained with ethidium bromide. The relative amount of amplified DNA was determined by comparison to known standards. As standards, fivefold serial dilutions of vector DNA ranging from 50 to 31,250 copies were analyzed in parallel.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of IN mutant virions. Our analysis of MLV IN protein function was performed in infected cells using the pkat MLV vector system described previously (12). Mutations were introduced into the IN sequence of the pkat-gag-pol packaging plasmid. These included mutations that disrupted the catalytic triad (D184A) or the N-terminal HHCC motif (H61A) and those that deleted either a portion of the C terminus (34) or the entire protein (IN) (Fig. 1A). Virions were generated by individually transfecting 293T cell cultures with the mutant packaging constructs, the pkat-GFP gene transfer vector, and the pkat-env plasmid. Wild-type virions were produced using the pkat-gag-pol packaging construct as a control. Virions collected from the culture supernatants by ultracentrifugation were analyzed by immunoblotting. The mature form (46 kDa) of the D184A and H61A mutant IN proteins was detected using anti-Pol antiserum (Fig. 2A). The 34 mutant was detected migrating with a lower molecular weight, consistent with the 34-amino-acid truncation. No IN protein was detected for the IN mutant. Virions were also analyzed using anti-Gag antibody specific for the MLV p30 capsid protein. The D184A and H61A mutants were identical to the wild type with respect to the amount and proteolytic processing of the p65 Gag precursor protein. In the case of 34 mutant virions, slightly elevated amounts of incompletely processed Gag were detected (Fig. 2B). The RT-IN mutant (included as a control) also exhibited increased levels of unprocessed and incompletely processed Gag protein. In comparing the signal intensities of p30, there did not appear to be a significant change in virion production among the different IN mutants. In using anti-RT antiserum to probe replica blots, no differences were detected between wild-type and IN mutant virions in either the amount or processing of the 80-kDa RT protein (Fig. 2C). A truncated RT protein (~56 kDa) was detected in the RT-IN mutant, consistent with the 301-amino-acid residue deletion. Culture supernatants were also analyzed for RT activity, and except for the RT-IN mutant, all of the IN mutants exhibited levels similar to that of wild type (Table 1). Taken together, these results demonstrated that there was not a severe defect in the assembly, release, and maturation of the IN mutant virions.

View larger version (59K): [in this window] [in a new window]   FIG. 2.   Immunoblot analysis of IN mutant virions. Cultures of 293T cells were separately transfected with 3 µg each of the different mutant packaging constructs plus 3 µg of the pkat-GFP gene transfer vector and 1 µg of the pkat-env plasmid. Six milliliters of each culture supernatant was collected after 72 h, and virions were pelleted by ultracentrifugation (125,000 × g for 2 h) using a Beckman SW-41 rotor. Pellets were solubilized in lysis buffer and analyzed by immunoblotting as described previously (33). Replica blots were probed separately with anti-Pol (A), anti-CA (B), and anti-RT (C) antibodies. WT, wild type.

IN mutations impair viral DNA synthesis. Our earlier studies demonstrated that the HIV-1 IN protein promotes viral cDNA synthesis independently of its enzymatic function (31). The results in Table 1 demonstrated that the infectivity of the IN and 34 mutants was consistently about 10-fold less than that of the integration-defective IN^D184A mutant. Since unintegrated cDNA of the MLV vector can express GFP (Table 2), this result suggested that the IN mutant was more severely impaired in DNA synthesis. To directly analyze viral DNA products of reverse transcription, HeLa cells were infected with normalized amounts of wild-type and IN mutant virions. Total DNA was extracted 18 h later, and cDNA was analyzed by semiquantitative PCR as described earlier (31). Approximately 10- to 20-fold less of the late (R-gag) DNA product was detected in cells infected with the IN and IN³⁴ mutants compared with wild type (Fig. 3). Identical results were obtained using primers that detected intermediate products of viral DNA synthesis (data not shown).

View larger version (16K): [in this window] [in a new window]   FIG. 3.   Mutations in IN impair DNA synthesis. Wild-type (WT) and IN mutant (IN, D184A, H61A, and 34) pkat-gag-pol were transfected into cultures of 293T cells together with pkat-GFP and pkat-env. After 72 h, the culture supernatants were filtered through 0.45-µm-pore-size filters and aliquoted. One aliquot set was analyzed for RT activity, and another set was analyzed by immunoblotting using anti-p30 antibody. The sample volumes of the third aliquot set were normalized based on RT activity or, in the case of the pkat-RT-IN control, the relative densities of the p30 protein. After treatment with RNase-free DNase (4 µg/ml), the virus stocks were placed on cultures of HeLa cells at 37°C. The cells were washed twice with DMEM by low-speed centrifugation after 4 h and lysed after 18 h. Total DNA was extracted, and 400 ng of each extract was analyzed using PCR to detect late (R-gag) viral DNA products of reverse transcription. The amplified products were resolved on 1.5% agarose gels and stained with ethidium bromide. As standards, serial fivefold dilutions, ranging from 50 to 13,250 copies of the pkat-GFP plasmid, were analyzed in parallel under identical conditions. The data shown are from a representative experiment that was repeated three times, each time with independent transfection-derived virus preparations. MW, molecular weight markers.

34

trans-IN protein complements DNA synthesis. It has been well documented that various mutations in HIV-1 IN impair viral DNA synthesis (11, 15, 19). However, it was not until recently that this phenotype was associated with a direct effect of the mature IN protein on the initiation of reverse transcription (31). Elucidation of this novel IN function was made possible by incorporating IN protein into HIV-1 virions in trans, as a fusion partner of Vpr. This enabled the function of the trans-IN protein to be studied against the background of IN mutant provirus, uncoupling the function of the mature IN protein during early events of the virus life cycle from defects in the Gag-Pol precursor protein and its effects on late events such as assembly. To examine the role of the MLV IN protein in reverse transcription, the IN coding region was fused with gag. This construction (designated pkat-gag^NCIN) deleted the NC sequence from gag and fused IN in frame at the 3' end of CA (Fig. 1C). This design was meant to minimize the effects (both positive and negative) of coexpressing wild-type Gag together with the pkat-gag-pol packaging construct (28). To confirm that the trans-IN protein was assembled into virions, the pkat-gag^NCIN expression plasmid was cotransfected into 293T cells with the IN packaging plasmid, the pkat-GFP-puro gene transfer vector, and the pkat-env plasmid. Progeny virions were shown to contain trans-IN protein that comigrated with wild type (Fig. 4, compare lanes 1 and 3). This result indicated that trans-IN was efficiently incorporated and subsequently liberated by proteolytic cleavage. Similarly, mature trans-IN protein was detected in virions when coexpressed with the IN³⁴ mutant packaging construct (lanes 4 and 5).

View larger version (57K): [in this window] [in a new window]   FIG. 4.   Incorporation of trans-IN into IN mutant virions. Wild-type (WT) and mutant (IN and IN34) pkat-gag-pol plasmids were separately transfected into 293T cells by themselves (lanes 1, 2, and 4, respectively) or cotransfected with 1 µg of the pkatNCIN trans-IN expression plasmid (lanes 3 and 5). The 293T cell cultures were also transfected with pkat-GFP and pkat-env. After 72 h, the culture supernatants were passaged through 0.45-µm-pore-size filters, pelleted, and examined by immunoblot analysis using anti-Pol antibody.

34

NC

NC

34

NC

184A

34

34

View larger version (18K): [in this window] [in a new window]   FIG. 5.   trans-IN protein rescues viral DNA synthesis. Wild-type (WT) and mutant (IN and IN34) pkat-gag-pol plasmids were separately transfected into 293T cells by themselves (lanes 1, 2, and 5, respectively) or with the pkatNCIN, pkatNCIN34, and pkatNCIND184A trans-IN expression plasmids. The culture supernatants were collected 72 h later, filtered through 0.45-µm-pore-size filters, and aliquoted. Each aliquot was analyzed exactly as described for Fig. 3. The data shown are from a representative experiment that was repeated three times, each time with independent transfection-derived virus preparations. The rightmost lane contains molecular weight markers.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Several studies have examined the effects of MLV IN mutations on different steps in the virus life cycle and demonstrated that many of these mutations dramatically decreased virus infectivity. In using Southern blot analysis to analyze the unintegrated viral DNA following virus infection, IN mutations were not found to cause a defect in viral DNA synthesis, suggesting that reverse transcription occurred normally and that the observed defect in virus infectivity was due to an impairment in integration (8, 24, 25). Using sensitive PCR-based methods to analyze total viral DNA synthesis, our data show that certain IN mutations (IN and IN³⁴) decrease viral DNA synthesis 10- to 20-fold. One plausible explanation for our different results is that Southern blot analysis of unintegrated DNA (Hirt supernatants) measures only a portion of the viral DNA that is synthesized. Retroviral DNA exists in three forms in infected cells: integrated, unintegrated linear, and unintegrated circular. Certain IN mutations may change the relative amounts of these three DNA forms without affecting the total amount of viral DNA synthesized. In the case of HIV-1, mutations that affect integration may result in an accumulation of unintegrated circular DNA (10). Therefore, using PCR to directly measure total viral DNA provides a more sensitive and quantitative means to analyze the effect of IN mutations on viral DNA synthesis. Our results showing that certain IN mutations impair viral DNA synthesis are completely consistent with our analysis of virus infectivity. With a highly sensitive assay capable of detecting GFP expression from unintegrated DNA (Tables 1 and 2), the IN and IN³⁴ mutant viruses were shown to be consistently 10-fold-less infectious than the IN^D184A mutant virus. Taken together, our findings demonstrated, for the first time, that certain MLV IN mutations impair viral DNA synthesis in vivo.

We considered two possibilities to explain this finding. First, IN mutation may change the structure of the Gag-Pol precursor and affect late stages of the viral life cycle (such as viral assembly), so as to impair later events including reverse transcription. Second, in the context of uncoating and formation of the reverse transcription complex the mature IN protein itself may augment reverse transcription. Previously, using a trans-complementation approach to package the HIV-1 IN protein independently of the Gag-Pol precursor we demonstrated that the mature HIV-1 IN protein is required for efficient initiation of viral DNA synthesis in infected cells (31). Here, we addressed whether the defect in MLV DNA synthesis was due to a change in Gag-Pol or the mature IN protein by trans-complementing IN mutants with wild-type and mutant Gag-IN fusion proteins. Our results show that the MLV trans-IN restored viral DNA synthesis to normal levels when coassembled with IN mutant-containing Gag-Pol precursor. Importantly, neither the Gag-IN³⁴ mutant nor trans-Gag precursor protein (without the IN fusion) rescued this defect. Moreover, Gag^NCIN rescued DNA synthesis at least as well as did a full-length Gag-IN fusion protein (data not shown), ruling out the possibility that NC protein derived from Gag-IN has a positive effect on DNA synthesis (1, 14, 16). Rather, these results indicate that the mature trans-IN protein itself was responsible for the restoration of viral DNA synthesis.

Recent studies of HIV-1 IN demonstrated three distinct phenotypes for IN mutant viruses: those that affect both viral DNA synthesis and DNA integration, those that affect only DNA synthesis, and those that affect only integration (15, 19). Our analysis of MLV indicates that the IN³⁴ mutant impaired both integration and viral DNA synthesis, which is consistent with our analysis of HIV-1 IN (31). Disruption of the HIV-1 IN HHCC motif affects both integration and viral DNA synthesis (15, 19, 31). In contrast, our analysis of MLV IN indicates that this motif may not be required for efficient viral DNA synthesis. This result was quite interesting, since the IN HHCC motif is conserved among all retroviruses and retroelements. Future studies will examine the genetic and structural basis for this difference. Our earlier findings for HIV-1 and now for MLV that catalytically defective (e.g., D116A and D184A, respectively) trans-IN complements DNA synthesis indicate a nonenzymatic, structure-based role of IN in viral DNA synthesis. Our results may suggest that this IN function is conserved among retroviruses. Since IN is essential for integrating the viral cDNA into the host cell genome, we suggest that retroviruses may have evolved a mechanism whereby IN acts to trigger viral DNA synthesis to ensure its association with the preintegration complex once reverse transcription is completed. Our findings and those of others (Monica Roth, personal communication) that retroviral IN and RT physically interact may help support this notion (31).

Viral vector systems are often used to study viral gene function since they can mimic that of the parental virus. MLV has been widely used as a vector for gene transfer because of its relatively simple genome organization and ability to infect and integrate DNA into the host cell genome (20). Our analysis of MLV IN was focused on early events of the virus life cycle. The MLV vector preserves all of the cis-acting elements necessary for efficient encapsidation of the viral RNA genome, reverse transcription, and integration. Our choice to utilize the MLV vector, rather than an infectious viral clone, was based, at least in part, on our previous analysis of HIV-1 assembly and IN function. Analysis of HIV-1 integration-defective provirus using HeLa-CD4--galactosidase (-Gal) indicator cells indicated that viral proteins could be expressed from unintegrated DNA. Integration-defective virus induced -Gal expression at a frequency of approximately 10 to 15% compared with wild-type virus. CD4-positive HeLa cells containing the LTR--Gal reporter can detect infection of integration-defective HIV-1 (2, 11, 30). This enabled us to subdivide the early steps of the HIV-1 virus life cycle and specifically analyze the effects of IN mutations on viral DNA synthesis. This was confirmed by our results using PCR methods to analyze unintegrated viral cDNA, which paralleled those using the CD4-HeLa--Gal indicator cells (31). In this study, we used the MLV-based vector containing an internal CMV promoter to drive GFP expression as a reporter for DNA synthesis-infectivity. Our results show that unintegrated vector DNA, derived from integration-defective virus, can exist transiently and express GFP (Table 1). By analysis of infected cells for transient gene expression (GFP-positive cell colonies) and proviral DNA integration (puromycin resistance), the frequency of GFP-positive cells was 3 orders of magnitude higher than that of integration (Table 2). Thus, the MLV vector system helped us to uncouple the integration function of IN from its nonenzymatic function. In combination with trans-IN complementation, these approaches may offer new opportunities to study both the enzymatic and nonenzymatic functions of MLV IN in the context of a replicating virus.