INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The human immunodeficiency virus (HIV) Gag protein is important in many steps of the viral life cycle. The initial product of gag gene translation is a polyprotein which contains information necessary for assembly and release of virions (19, 37), into which viral RNA (24), env glycoprotein (14, 40), pol-encoded enzymes (29, 34), and other viral proteins are incorporated. Upon virion release from the cell surface, Gag polyproteins are cleaved by the viral protease into matrix, capsid, nucleocapsid, and other small proteins (20). Given the association of Gag polyprotein cleavage products with retroviral preintegration complexes (11), it is likely that Gag products are also important in early events following viral entry into the host cell prior to retroviral DNA integration. This is supported by the identification of a putative nuclear localization signal within the matrix protein (10, 16) as well as the observation that some gag mutations have no apparent effect on virion assembly yet significantly reduce virion infectivity (7, 30, 36).

Identification of cellular proteins necessary for the multiple functions of Gag is important in understanding the role Gag protein plays in the retroviral life cycle (9, 18). Cyclophilins, which are members of a large family of cellular proteins with multiple functions, have been shown to interact specifically with HIV-1 Gag (26). Furthermore, it has been demonstrated that human cyclophilin A (CyPA) is incorporated into HIV-1 virions, but not those of other primate immunodeficiency viruses, and that HIV-1 virion incorporation of CyPA is essential to its infectivity (15, 35). Both virion incorporation of CyPA and virion infectivity are disrupted in a dose-dependent fashion by cyclosporin A (CsA) and nonimmunosuppressive analogs of CsA which bind CyPA (3, 5, 15, 33, 35). HIV-1 capsid (CA) has been shown not only to mediate CyPA incorporation but also to confer sensitivity of virion infectivity to CsA (13, 35). Using HIV-simian immunodeficiency virus chimeras, it has been shown that a small, highly conserved, proline-rich segment of HIV-1 CA mediates CyPA incorporation into HIV-1 virions (15). Substitution of a single conserved proline within this segment by alanine (P222A) diminishes CyPA incorporation into HIV-1 virions and decreases virion infectivity (15).

Here we demonstrate that mutation of this conserved residue does not interfere with CyPA incorporation into HIV-1 virions or virus replication in all cell types. We also show that cells with high CyPA content support the replication of the HIV-1 P222A mutant.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cells and viruses. Plasmids pYKJR-CSF (21), pNL4-3 (2), and pMM4 (27) were used to produce stocks of HIV-1 strains JR-CSF, NL4-3, and HXB2, respectively. To introduce mutations into the pNL4-3 gag gene, a StuI-ApaI fragment (nucleotides 14173 to 2011) was ligated into pBluescript II KS() (Stratagene, La Jolla, Calif.) to produce the vector pKS-gag. Site-directed mutagenesis was performed with single-stranded phagemid DNA released from Escherichia coli transformed by pKS-gag and infected with M13K07. The following mutagenic oligonucleotides were used to produce proline-to-alanine mutations at Gag amino acid codons 217 and 222: 5'ATAGATTGCATGCCGTGCATGCAGGG3' and 5'CATGCAGGGGCAATTGCACCAGG3', respectively. A BssHII-SpeI fragment (nucleotides 712 to 1508) containing the respective mutation, confirmed by the dideoxy-chain termination method, was ligated into the proviral vector pNL4-3. The P222A mutation was then introduced into the pYKJR-CSF and pMM4 plasmids by subcloning of the fragment NarI-ApaI (nucleotides 639 to 2011) from the pNL4-3/P222A mutant. Virus replication was examined in CEM cells and in phytohemagglutinin (PHA)-stimulated peripheral blood lymphocytes (PBLs) infected with supernatants of COS-7 cells transfected with these proviral constructs as described previously (42). Supernatants from transfected COS-7 cells or infected CEM cells were overlayed onto two-step gradients (15 to 60% sucrose diluted in TN [10 mM Tris-HCl, pH 7.5, 100 mM NaCl]). Virus recovered from the interface of these step gradients was centrifuged overnight at 150,000 × g through 15 to 60% continuous sucrose gradients prepared with TN. Fractions were recovered from the continuous gradients as recently described (12). The p24 content of the sucrose gradient fractions and supernatants of transfected or infected cells was determined by p24 enzyme-linked immunosorbent assay (ELISA) (Abbott Laboratories, Abbott Park, Ill.).

Viral protein and cell lysate analysis. Sucrose gradient-purified viral stocks (see above) containing equivalent amounts of p24 antigen were precipitated with acetone, separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to nitrocellulose membranes as previously described (32). Cell lysates prepared as previously described (32) were used for Western blot analysis of cellular CyPA content. Serially diluted lysates of CEM cells or CyPA (generously provided by W. Sundquist, University of Utah) were included as quantitative standards. Detection of proteins was performed with rabbit antiserum against human CyPA (Affinity Bioreagents, Golden, Colo.) or a mixture of human monoclonal antibodies against p24 (71-31, 91-6, and 98-4.3) obtained from Susan Zolla-Pazner through the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases. Bound antibody was detected with horseradish peroxidase-conjugated secondary antibodies and a chemiluminescent detection system (New England Biolabs, Beverly, Mass.) (see Fig. 3 and 5) or alkaline phosphatase-conjugated secondary antibodies with a precipitation substrate (see Fig. 4) as previously described (31).

DNA PCR analysis. PCR detection of HIV-1 DNA synthesis and cellular -globin gene sequences was performed by previously described methods (22, 41). Dried gels were exposed to film to produce autoradiographs and were analyzed with an Ambis (San Diego, Calif.) radioanalytic imager. A standard curve to relate counts per minute to copies of the HIV-1 genome was generated by linear regression and used to quantitate the signals from experimental samples as previously described (23).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Some P222A mutants replicate in CEM cells and PBLs. To study the effect of the P222A mutation on viral growth and CyPA incorporation, virus stocks were prepared by collecting the supernatant of COS-7 cells transfected with pNL4-3 (2) or a proviral vector in which a P222A mutation was encoded, pNL4-3/P222A. Virus stocks containing equivalent p24 contents were then used to infect PBLs and CEM cells, and the p24 content in the culture supernatant was monitored. We chose to study HIV-1_NL4-3 and HIV-1_NL4-3/P222A initially, since HIV-1_NL4-3 contains open reading frames for all known HIV-1 accessory genes. HIV-1_NL4-3/P222A was found to replicate with somewhat slower kinetics but with similar virus production to that of wild-type HIV-1_NL4-3 in CEM cells (Fig. 1D) and also slower kinetics and at reduced yet significant levels in PBLs (Fig. 1A). The growth kinetics of mutant and wild-type virus stocks following three serial passages in CEM cells remained identical to those observed with the initial infection of CEM, making reversion or compensatory mutations unlikely reasons for the growth of the P222A virus in CEM cells (data not shown). To confirm this, we amplified HIV-1 DNA sequences from cells infected with stocks of HIV-1_NL4-3 and HIV-1_NL4-3/P222A, which had been passaged three times in CEM cells. This DNA was ligated into the pCR II TA cloning vector (Invitrogen, San Diego, Calif.), and nucleotides 1450 to 1581 were sequenced. One of 12 clones from cells infected with HIV-1_NL4-3/P222A contained a single point mutation 62 nucleotides downstream from the P222A mutation site. None of the P222A mutants had reverted or contained substitutions at the codon encoding 222A. Four clones from cells infected with HIV-1_NL4-3 were without mutations within the region sequenced (data not shown). The same P222A mutation was then introduced into the proviral vectors for JR-CSF (21) and HXB2 (27). While these mutants failed to replicate in PBLs (Fig. 1B and C), HIV-1_HXB2/P222A did replicate in CEM cells in a pattern similar to that seen following infection of CEM cells by HIV-1_NL4-3/P222A (Fig. 1D and E). Thus, while the P222A mutation has been shown to abrogate HIV-1 replication in Jurkat T cells (7, 15), we found that the observed phenotype was dependent on the cell type and, possibly, the virus strain.

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Effect of Gag P222A mutation on HIV-1 replication in PBLs and CEM cells. Following infection of 2 × 106 CEM cells or PHA-stimulated PBLs with HIV-1 virus stock containing 20 ng of p24, viral replication was monitored by measuring the amount of p24 in the viral culture supernatant (ordinate) as a function of days postinfection (abscissa). CEM cells were split 1:3 at 72 hpi and every 48 h thereafter. , wild type; , P222A mutant.

HIV-1 DNA synthesis is decreased following infection by P222A mutants. DNA PCR was then used to determine if the apparent reduced infectivity of P222A mutants in PHA-stimulated PBLs results from defects in early events in the retroviral life cycle. HIV-1_NL4-3 and HIV-1_NL4-3/P222A virus stocks containing equivalent amounts of p24 were used to infect PBLs and CEM cells, and the HIV-1 DNA produced at 8 and 19 h postinfection (hpi) was analyzed by quantitative PCR (Fig. 2A). Approximately fivefold less HIV-1 DNA was present at these time points following infection of CEM by HIV-1_NL4-3/P222A than following infection of CEM by HIV-1_NL4-3 (Fig. 2A). About 10-fold less HIV-1 DNA was synthesized following infection of PBL by P222A mutant virus than following infection of PBLs by wild-type HIV-1 (Fig. 2A). Similarly, virus stocks containing equivalent p24 contents were used to infect PBLs and CEM cells, and the HIV-1 DNA present was analyzed at 24 and 72 hpi by quantitative PCR (Fig. 2B). HIV-1_NL4-3 produced abundant HIV-1 DNA by 24 hpi, with an increased amount by 72 hpi, consistent with spread, in both PBLs and CEM cells. HIV-1_NL4-3/P222A produced a similar pattern only in CEM cells (Fig. 2B). Thus, the amount of HIV-1 DNA produced following infection parallels the pattern of growth seen for mutant virus, which replicated only in CEM cells, and wild-type virus, which replicated in both cell types.

View larger version (56K): [in this window] [in a new window]   FIG. 2.   PCR analysis of HIV-1 DNA in HIV-1NL4-3- or HIV-1NL4-3/P222A-infected CEM cells or PHA-stimulated PBLs. CEM cells or PHA-stimulated PBLs were infected with HIV-1NL4-3 or HIV-1NL4-3/P222A (20 ng of p24/2 × 106), and cells were harvested at 8 and 19 hpi (A) or 24 and 72 hpi (B). DNA was purified and subjected to PCR analysis with the R/U5 (AA55/667), LTR (long terminal repeat)/gag (M661/667), or -globin (LA1/LA2) primers (42). LV1 and LV2, live virus stock infections done in duplicate; HI, infections done with heat-inactivated virus stock (42). HIV-1 DNA corresponds to the number of copies per lane, while -globin DNA is given as micrograms per lane.

CyPA content of P222A mutant virions produced in some cell types is only moderately reduced. It has been shown that impaired CyPA incorporation into HIV-1 virions reduces virion infectivity (15, 35). Hence, having found near wild-type levels of replication of P222A mutants in CEM cells, we next evaluated the CyPA content of mutant P222A virions and wild-type NL4-3 virions following infection of CEM cells. CEM cells were infected with wild-type NL4-3, P217A, and P222A HIV-1 viral stocks obtained by collecting supernatants of COS-7 cells transfected with the proviral constructs pNL4-3 (2), pNL4-3/P217A, and pNL4-3/P222A as previously described (42). The cell culture medium was changed, and the supernatant was collected every 24 h. Virus was purified, by sucrose gradient centrifugation, from the CEM supernatants collected at the time of peak virus production determined by p24 ELISA (72 hpi for NL4-3 and 120 hpi for P217A and P222A). Western blot analysis showed that virus samples contained equal amounts of viral protein (Fig. 3A). Supernatant of mock-infected CEM cells (Fig. 3, lane 1) was also analyzed in a similar manner to demonstrate that CyPA-containing vesicles were not produced under these conditions (4, 28). Mutant P217A virions, previously shown to contain amounts of CyPA similar to those in wild-type virions (15), were analyzed as an additional control (Fig. 3, lane 4). Mutant P222A virions produced following infection of CEM cells had only moderately decreased CyPA content compared with wild-type and P217A virions from CEM cells (Fig. 3B). Similarly, P222A virions produced by transfection of COS-7 cells had less than a threefold reduction of CyPA content compared with wild-type virions produced in the same manner (Fig. 4B). This difference is somewhat less than the approximately fivefold difference seen when the CyPA contents of P222A mutant and wild-type virions produced by transfection of 293T cells were compared (7). Thus, P222A mutant HIV-1 virions from COS-7 cells may contain 1.5 to 2 times as much CyPA as P222A mutant HIV-1 virions from 293T cells. While this difference is not large, it is possible that slight alterations of the CA-CyPA stoichiometry of HIV-1 may affect virion infectivity significantly (25).

View larger version (26K): [in this window] [in a new window]   FIG. 3.   CyPA content of HIV-1NL4-3, HIV-1NL4-3/P217A, and HIV- 1NL4-3/P222A virions. Sucrose gradient-purified HIV-1NL4-3, HIV-1NL4-3/P217A, and HIV-1NL4-3/P222A virions containing 230 ng of p24 collected from CEM cells infected with virus stocks from supernatants of COS-7 cell transfections were acetone precipitated, separated by SDS-PAGE, and transferred to nitrocellulose. Membranes were labelled with a mixture of human monoclonal antibodies against capsid (CA) (A) or rabbit polyclonal antibody against CyPA (B). Bound antibody was detected with horseradish peroxidase-conjugated secondary antibody and a chemoluminescent detection system. Serially diluted human CyPA was used for standards. M, molecular mass standards; WT, wild type.

View larger version (19K): [in this window] [in a new window]   FIG. 4.   CyPA content of HIV-1NL4-3 and HIV-1NL4-3/P222A virions produced by COS-7 cells. Different amounts (measured by p24 ELISA) of sucrose gradient-purified HIV-1NL4-3 and HIV-1NL4-3/P222A virions from COS-7 cell transfections were precipitated by centrifugation, and their proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes. Membranes were labelled with human monoclonal antibody against CA (A) or rabbit polyclonal antibody against human CyPA (B). Bound antibody was detected with alkaline phosphatase-conjugated secondary antibody with a precipitation substrate. Serially diluted CEM cells were analyzed similarly as standards. M, molecular mass standards; WT, wild type.

Cells with higher CyPA content support the replication of mutant virus. The CyPA contents of several cell lines were next evaluated in order to determine if the CyPA contents of susceptible cell lines might influence the ability of HIV- 1_NL4-3/P222A to incorporate CyPA and replicate. The CyPA content of COS-7 cells was similar to that of CEM cells and was approximately 4-fold greater than that of Jurkat cells and approximately 10-fold greater than that of PBLs (Fig. 5A), while the total cellular protein content of these cell types, determined by the Bradford dye-binding procedure (Bio-Rad Laboratories, Hercules, Calif.) in duplicate was found to vary by less than 16% (data not shown). Infection of CEM cells with the P222A mutant resulted in a yield of virus similar to that from infection of these cells with wild-type HIV-1 (Fig. 1D and E). On the other hand, cells with lower CyPA content, such as Jurkat cells and PBLs, either failed to support replication of the P222A mutant (4, 11) (Fig. 1B and C) or supported only attenuated replication of this mutant (Fig. 1A). Thus, the ability of P222A mutants to replicate in different cell types correlated with cellular CyPA content. Likewise, infection of CEM cells with the P217A mutant, which fails to replicate in Jurkat cells (15) or PBLs, resulted in a yield of virus similar to that from infection of CEM cells with wild-type HIV-1 (data not shown). P217 is one of only nine CA residues which make a series of favorable contacts with the CyPA active site (17). Furthermore, the binding affinity of P217A CA is threefold lower than that of wild-type HIV-1 (39). Thus, while CyPA incorporation of the P217A mutant is similar to that of wild-type HIV-1 (15) (Fig. 3), it is possible that cellular CyPA content also influenced the ability of the P217A mutant to replicate.

View larger version (31K): [in this window] [in a new window]   FIG. 5.   CyPA content of different cell types. Proteins from 3 × 104 or 6 × 103 COS-7, CEM, and Jurkat cells and PBLs (A) or 3 × 104 or 1.5 × 104 CEM, H9, and Molt-4 cells and PBLs (B) were separated by SDS-PAGE and transferred to nitrocellulose membranes. Membranes were decorated with rabbit polyclonal antibody against human CyPA. Horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin was used as a secondary antibody. Serially diluted human CyPA was used for standards. M, molecular mass standards.

Cellular CyPA content influences virion incorporation of CyPA. As noted previously, HIV-1 virions from cells with high CyPA content produced infectious P222A mutant virions with only moderately reduced CyPA content compared with that of wild-type HIV-1 virions (Fig. 3B and 4B), suggesting cellular CyPA content influenced virion incorporation of CyPA. During the preparation of this article, other investigators reported that H9 cells chronically infected with HIV-1_IIIB produced HIV-1 virions with increased CyPA content, higher infectivity, and greater resistance to drugs which bind CyPA than virions derived from Molt-4 cells chronically infected with the same virus (8). Therefore, we evaluated the CyPA content of H9 and Molt-4 cells. H9 cells were found to have approximately threefold-greater CyPA content than Molt-4 cells (Fig. 5B), while their total cellular protein content differed by less than 30% (data not shown). Hence, the CyPA content of HIV-1 virions produced by these cells is proportional to the CyPA content of the cells from which the virions are derived.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

It has been shown that CyPA is incorporated into HIV-1 virions via interaction with the Gag polyprotein (15). In addition, inhibition of CyPA incorporation into HIV-1 virions by CsA or a point mutation in the Gag polyprotein, P222A, interferes with virion infectivity (15, 35). However, the experiments described here demonstrated that the replication of P222A mutant virions was, at most, minimally impaired in cells with higher CyPA content.

We found that HIV-1 DNA synthesis was decreased following infection by the P222A mutant in both CEM cells and PBLs. This is similar to previous findings of a correlation between the disruption of incorporation of CyPA into HIV-1 virions and the failure to synthesize HIV-1 DNA in acutely infected cells (7). We also found that the relative decrease in HIV-1 DNA synthesis following infection by the P222A mutant compared with that of the wild-type virus in CEM cells and PBLs paralleled the replication kinetics of these virus strains, being more pronounced in PBLs than in CEM cells. Moreover, this difference also paralleled the relative CyPA concentrations of these cells, having been more pronounced in PBLs, which have lower CyPA content, than in CEM cells.

An interesting finding which appears to support a role for CyPA content of cells as an important determinant of the replication kinetics of HIV-1 is the production of isolates which are not only resistant to CsA but also fail to grow in the absence of CsA (1). These CsA-resistant/dependent mutants have been shown to have one of two point mutations, A224E or G226D. Both mutations result in the introduction of a negatively charged amino acid flanking one of three conserved prolines within the portion of Gag previously shown to be necessary for CyPA binding and incorporation into HIV-1 virions (7, 15, 35). While the P222A mutation markedly decreases CyPA binding by HIV-1 Gag (15, 35), the A224E and G226D mutations do not appear to alter significantly the CyPA-binding properties of Gag protein (6). This suggests that while the interaction of CyPA with HIV-1 Gag may be important for the replication of some HIV-1 isolates, it may be deleterious for others.

HIV-1 virions whose CyPA incorporation is impaired by either CsA or the P222A Gag mutation are biochemically and morphologically indistinguishable from wild-type HIV-1 virions (6). In addition, in in vitro reactions, CyPA-deficient HIV-1 virions reverse transcribe their endogenous RNA templates similarly to wild-type HIV-1 virions (7). This suggests that CyPA is probably not crucial for virion assembly. However, HIV-1 DNA synthesis was significantly impaired following infection by CyPA-deficient HIV-1 virions (Fig. 2). Thus, it appears that the Gag-CyPA interaction is important for an early event in the virus life cycle, prior to initiation of reverse transcription (7). Since HIV-1 virions pseudotyped with amphotrophic murine leukemia virus Env proteins are rendered noninfectious and exhibit markedly decreased DNA synthesis when they contain the P222A Gag mutation or are produced in the presence of CsA, CyPA is likely important to HIV-1 at a point following receptor binding and membrane fusion (7). The most likely step between membrane fusion and reverse transcription in which CyPA and CA interactions are important appears to be virion uncoating (7). Recent analysis of the crystal structure of the N terminus of CA bound to human CyPA has led to the suggestion that CyPA may play a role in virion uncoating. It appears that CyPA binding of CA may be necessary for the destabilization of CA-CA interactions, thereby facilitating core disassembly (17). Thus, the stoichiometry of CA-CyPA (approximately 2,000:200) may be important for HIV-1 virion infectivity (25). Too little CyPA may render HIV-1 replication defective by allowing highly stable CA-CA interactions, thereby interfering with virion uncoating. This may be the mechanism by which drugs or mutations which interfere with CyPA incorporation into virions reduce HIV-1 virion infectivity. The P222A mutant has been shown to be defective in its ability to incorporate CyPA (15). In addition, P222A mutant Gag has been found to have a reduced affinity for CyPA compared with wild-type Gag (39). Hence, it is likely that a higher cellular CyPA concentration may be able to compensate for the decreased affinity of P222A mutant Gag for CyPA, increasing CyPA incorporation into P222A virions. This is supported by our finding that CEM cells, in which the P222A mutant replicated, contained higher levels of CyPA than those cell types (Jurkat cells and PBLs) in which this mutant failed to replicate. However, we found that the higher CyPA content of these cells was only partially able to restore the incorporation of CyPA into P222A virions. This may explain the slightly reduced rate of replication and decreased HIV-1 DNA synthesis observed following infection by the P222A mutant compared with that found following infection by wild-type virus even in cells with higher CyPA content. In contrast, CsA-resistant/dependent mutants contain Gag mutations (A224E or G226D) which likely destabilize CA-CA interactions, thereby facilitating virion uncoating in the absence of CyPA (25). However, these mutations do not significantly alter the CyPA-binding properties of Gag or virion incorporation of CyPA (6). Thus, these mutants may require CsA in order to decrease virion incorporation of CyPA, which might otherwise further destabilize CA-CA interactions, ultimately interfering with virus replication. Consistent with this model, double mutants (P222A and A224E), whose virion incorporation of CyPA is impaired, replicate in the absence of CsA (6). In addition, CsA-resistant/dependent mutants have recently been found to replicate in the absence of CsA in Jurkat cells which had decreased CyPA content compared with those cell types (CEM) in which these mutants require CsA to replicate (6). Finally, it has been shown that the cell type in which HIV-1_IIIB is produced determines the virion CyPA content, infectivity, and susceptibility to CsA (8). Analysis of the CyPA content of the chronically infected cells from which HIV-1 virions were harvested revealed that cells with greater CyPA content yield virions with higher CyPA content than cells with less CyPA (Fig. 5B).

Exposure of HIV-1 to CsA or nonimmunosuppressive CsA analogs results in marked inhibition of viral replication in cell culture (3, 5, 15, 33, 35). However, mutants are readily produced which are not only resistant to these drugs, but also fail to replicate in their absence in cell culture (1, 6). It is therefore conceivable that nonimmunosuppressive cyclosporin analogs, perhaps used cyclically, might be useful adjunctive agents for HIV-1 infection. However, our studies suggest HIV-1 may be eliminated from some cell types by CsA analogs, but virus replication could persist in other cells with higher CyPA content. This may be important in vivo, since multiple cell types are infected.