INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

An intact nef gene is important for efficient replication and pathogenicity of human and simian immunodeficiency viruses (HIV and SIV, respectively) (14, 31, 34, 63). Several in vitro activities of HIV and SIV Nef which may be relevant for viral pathogenicity have been described. Nef downregulates cell surface expression of CD4 (2, 6, 20, 46) and of class I major histocompatibility complex (MHC-I) molecules (12, 40, 56). Furthermore, Nef increases the infectivity of viral particles and enhances viral replication in primary lymphocytes (11, 16, 22, 37, 42, 47, 58).

Although the molecular basis of these effects has not been fully elucidated, several studies have shown that Nef can interfere with cellular signal transduction pathways (1, 4, 17, 24, 29, 43, 57) and interacts with a number of cellular proteins and kinases (reviewed in reference 54). It has been shown previously that the conserved P(xxP)₃ element in HIV type 1 (HIV-1) Nef is important for the association with a 62-kDa phosphoprotein, termed Nef-associated kinase (NAK), which may belong to the family of p21-activated protein kinases (5, 45, 48, 55). The PxxP motif of HIV-1 Nef also mediates the interaction with the SH3 domains of the Src tyrosine family kinases Hck, Lyn, and Fyn and the SH3 domain of VAV (3, 18, 25, 38, 39, 53). Since p21-activated protein kinases do not contain SH3 domains and are unlikely to interact directly with the PxxP sequence in Nef, it has been suggested that an SH3 domain-containing tyrosine kinase, p62, and HIV-1 Nef may form a protein complex (45). Recently, it has been suggested that HIV-1 Nef association with the  chain of the T-cell receptor complex activates NAK and leads to the induction of Fas-Fas ligand expression (64).

A P(xxP)_1-2 motif is also conserved in SIV and HIV-2 Nef. Similarly to HIV-1 Nef, mutations in the PxxP motif in SIV Nef can disrupt association with p62 (32, 37). Mutation of the PxxP motif in SIVmac Nef, however, does not impair tyrosine or serine phosphorylation of Nef and does not disrupt the interaction with the tyrosine kinase Src (37). Furthermore, we have previously found that SIVmac239 containing changes of P₁₀₄xxP₁₀₇ to A₁₀₄xxA₁₀₇ in Nef (AxxA-Nef) showed full pathogenic potential in rhesus macaques (37). A₁₀₇P reversion, which is sufficient to restore the Nef-p62 association, was observed in only a minor fraction of nef sequences derived from rapidly progressing animals at the time of AIDS-related death at 9 and 18 weeks postinfection (wpi) (37). In another study, however, such reversions came to predominate in the majority of macaques chronically infected with the SIVmac239 AxxA-Nef mutant (32).

Notably, the SIVmac Nef protein, unlike HIV-1 Nef, contains N-terminal Y₂₈xxL/Y₃₉xxS motifs, which resemble consensus sequences for SH2 binding domains (17). An SIVmac239 variant containing an additional YxxL motif in Nef, which generates an immunoreceptor tyrosine-based activation motif, causes extensive T-lymphocyte activation and acute disease in macaques (17, 43). It has also been postulated that the Y₂₈GRL/Y₃₉SQS motifs in SIV Nef may represent tyrosine-based endocytosis signals important for downregulation of CD4 (8, 49).

One possible explanation for why mutations of the PxxP motif in SIV Nef do not prevent disease progression in macaques is that SIV Nef, unlike HIV-1 Nef, interacts with both SH2 and SH3 domains to perform critical functions. In the present study, we assessed this possibility and further investigated the relevance of the Nef-NAK association for SIV replication in vivo. A Nef mutant containing combined changes in both the proline- and the tyrosine-based motifs (FFAA-Nef) associated with the cellular tyrosine kinase Src and was functional in most in vitro assay systems. The FFAA-Nef did not associate with p62, however, and showed a reduced ability to stimulate viral replication in primary lymphocytes. SIVmac239 containing the FFAA-Nef replicated to high levels in infected rhesus macaques. Concordant with previous studies (32, 37), we show that A₁₀₄xxA₁₀₇P₁₀₄xxP₁₀₇ reversions are not a prerequisite for efficient replication of SIVmac in rhesus macaques. The Y₂₈/Y₃₉F₂₈/F₃₉ substitutions did not revert, suggesting that these N-terminal tyrosines in SIV Nef are dispensable for efficient viral replication in vivo.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of SIVmac239 Nef mutants and expression plasmids. The open SIVmac239 wild-type nef (239wt); the AxxA-Nef mutant, SIVmac239 nef*, containing a stop signal at the 93rd codon; and NU, containing deletions of 515 bp in the nef-long terminal repeat region, have been previously described (26, 30, 37, 52). The 239wt nef was amplified using the mutagenic primer pSIV-YYFF (5'- GAAGATCTGCGACAGAGACTCTTGCGGGCGCGTGGGGAGACTTT TGGGAGACTCTTAGGAGAGGTGGAAGATGGATTCTCGCAATC-3'), which introduced two AT substitutions (underlined) at nucleotide positions 83 and 116 in nef, and primer SL17 (5'-GTCCCTGCTGTTTCAGCGAGTTTCC-3'). The PCR product was digested with BglII (boldface) and NcoI. Subsequently, gel-purified fragments encompassing bp 43 to 205 of the nef open reading frame, predicting changes of Y₂₈F and Y₃₉F (Fig. 1), were inserted into plasmids carrying the 239wt, AxxA, AxxP, and PxxA nef alleles. The mutant nef alleles were subsequently inserted into a modified pBR322 vector containing the full-length SIVmac239 proviral DNA as described previously (37). These mutants are hereinafter referred to as FF-, FFAA-, FFAP-, and FFPA-Nef variants. For protein expression, these nef open reading frames were amplified with primers SL1 and SL34 and cloned into the expression vector pFJ as described previously (17). Amplification with SL34 results in fusion of the AU-1 peptide tag to the C terminus of Nef. All PCR-derived inserts were sequenced to confirm that only the intended changes were present (Fig. 1).

Cells, virus stocks, and infectivity assays. COS, 293T, and CEMx174 cells were cultured as described previously (9). Rhesus peripheral blood mononuclear cells (rPBMC) were isolated as described previously (37), immediately infected with aliquots of the virus stocks containing 2 ng of p27, and kept in RPMI 1640 with 10% fetal calf serum (FCS). Residual virus was removed by washing the cells 16 to 18 h after infection. After 6 days postinfection, cells were stimulated with phytohemagglutinin (PHA; 2 µg/ml; Sigma) for 3 days and washed and maintained in RPMI 1640 with 20% FCS and 100 U of interleukin 2 (IL-2) per ml. Virus production was measured by reverse transcriptase assay (50). The herpesvirus saimiri-transformed T-cell line 221 (1) was maintained in the presence of 100 U of IL-2/ml (Boehringer, Heidelberg, Germany) and 20% FCS. Infections were performed in the presence of 50 U of IL-2/ml and 5% FCS. Virus infectivity was determined using sMAGI cells as described previously (10), except that no DEAE-dextran was added to the infections, and quantitated using the Galacto-Light Plus chemiluminescence reporter assay kit (Tropix, Bedford, Mass.), as recommended by the manufacturer.

Transfections. For virus production, 293T cells were transfected by the calcium phosphate method (15) with 10 µg of the proviral constructs. The medium was changed after overnight incubation, and virus was harvested 24 h later. The amount of p27 antigen in the plasma was determined by a commercial HIV-1-HIV-2 enzyme-linked immunosorbent assay (Immunogenetics, Zwijndrecht, Belgium). COS cells were transfected by the DEAE-dextran method (13).

Infection of rhesus macaques. Juvenile rhesus macaques of Indian origin were infected by intravenous inoculation of SIVmac239 FFAA-Nef, containing 10 ng of p27 produced by transfected 293T cells. Limiting-dilution analysis in CEMx174 cells showed that the virus stocks contained approximately 1,000 50% tissue culture infective doses per ng of p27 antigen. The animals were healthy and seronegative for SIV, D-type retroviruses, and simian T-cell leukemia virus type 1 at the time of infection. Sera and cells were collected at regular intervals, and serological, virological, and immunological analysis was performed as described previously (59-61). Urinary neopterin levels were determined and normalized for urinary creatine concentration as previously described (19, 36).

Determination of reversion frequency. SIV sequences spanning the entire nef gene were amplified from rPBMC DNA with a nested-PCR approach or from DNA isolated from positive PBMC-CEMx174 bulk cocultivation by one round of amplification as described previously (33, 37). Viral plasma RNA was isolated with the QIAamp RNA kit (Qiagen, Basel, Switzerland), reverse transcribed with Superscript reverse transcriptase (GibcoBRL, Eggenstein, Germany), and subjected to a standard nested-PCR approach. PCR fragments were sequenced directly or following subcloning into the pCRII vector (Invitrogen Corp., San Diego, Calif.). Sequencing was performed with the PRISM sequencing kit (Perkin-Elmer, Foster City, Calif.) and with an automated Applied Biosystems 373 DNA sequencer following the protocols of the manufacturers. Reversions were quantitated by comparison to the standard curves as described previously (35, 37).

Biochemical analysis of Nef. Transfected COS cells or infected CEMx174 cells were lysed, and cleared cell lysates were used for immunoprecipitation and in vitro protein kinase assays as described previously (17, 37). Expression of Nef proteins in whole cellular lysates was analyzed by immunoblotting (37) as described by the manufacturer of the enhanced chemiluminescence system (ECL; Amersham, Chicago, Ill.). Dose-response analysis of the effect of Nef on CD4 and MHC-I cell surface expression was performed with Jurkat T cells expressing high levels of CD4, as described previously (23, 46).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

In vitro replication and infectivity. The localization of the mutations introduced into SIVmac239 Nef is shown in Fig. 1. The tyrosines at position 28 and 39 in SIV Nef were replaced with phenylalanines because such mutations disrupt tyrosine-based sorting signals and are known to decrease the affinity of these signals for adapter complexes by 2 orders of magnitude (7) and to minimize effects on secondary structure. In an experiment using high-titered virus stocks derived from transfected 293T cells, an intact nef gene enhanced SIV infectivity in sMAGI cells about eightfold (Fig. 2A). The mutations in the tyrosine and proline residues in SIV Nef did not significantly reduce virion infectivity.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Schematic presentation of the mutations introduced into the SIVmac239 Nef. Numbers specify the amino acid position in Nef. PPT, polypurine tract. Indicated are the mutations in the N-terminal tyrosine residues and in the PxxP motif.

View larger version (22K): [in this window] [in a new window]   FIG. 2.   Infectivity and replication of SIVmac nef variants. (A) sMAGI cells were infected in triplicate with five different 293T cell-derived virus stocks containing 100 ng of p27. Infectivity is shown relative to that of 239wt virus. (B) Replication kinetics of SIVmac239 nef variants in primary rhesus lymphocytes. rPBMC were infected immediately after isolation and stimulated with PHA at day 6 postinfection. Similar results were obtained with PBMC from three other rhesus macaques. (C) Replication of the indicated Nef mutants in 221 cells in the presence of 50 U of IL-2/ml and 5% FCS. Infections, cell culture, and quantitation of infectivity were performed as described in Materials and Methods. Reverse transcriptase (RT) activity was determined using a phosphorimager. P.S.L., photostimulated light emission.

The Y₂₈F, Y₃₉F, P₁₀₄A, and P₁₀₇A mutations in SIV Nef do not affect the ability to downregulate CD4 and MHC-I. The relative abilities of the 239wt and mutant Nef proteins to downregulate CD4 and MHC-I surface expression were assayed in dose-response experiments in human CD4⁺ Jurkat T cells using a transient-transfection assay. It has been previously shown that individual mutations in the N-terminal tyrosine residues or in the PxxP motif did not significantly affect the ability of SIVmac Nef to downmodulate cell surface expression of CD4 or MHC-I molecules (37, 41). As shown in Fig. 3, combining these mutations on the same Nef molecule also had only a marginal effect on the ability of Nef to downregulate CD4 and MHC-I molecules.

View larger version (16K): [in this window] [in a new window]   FIG. 3.   The N-terminal tyrosine residues and the PxxP motif in SIV Nef are dispensable for CD4 and MHC-I downregulation. Human CD4+ Jurkat T cells were transiently transfected with the indicated amounts of plasmids expressing the 239wt-, AxxA-, FF-, or FFAA-Nef, and cell surface expression of CD4 (A) and MHC-I (B) was determined as described in Materials and Methods. The values on the y axis give the levels of CD4 and MHC-I expression, represented by the peak channel number of red or green fluorescence on CD20+ cells (23, 46).

P₁₀₇A disrupts FFAA-Nef association with NAK activity. COS cells were transfected with expression constructs for 239wt or mutant Nef proteins, and in vitro kinase assays were performed on Nef immune complexes. Five phosphorylated proteins with apparent molecular masses of 34, 62, 80, 90, and 110 to 130 kDa (Fig. 4A, lane 2) were detected with 239wt Nef. A similar pattern was observed for the FF-Nef mutant (Fig. 4, lane 3). In agreement with previous studies (32, 37), coprecipitation and/or phosphorylation of these cellular proteins was abolished for AxxA-Nef (Fig. 4, lane 4). Analysis of mutants containing changes in single proline residues revealed that the FFAP-Nef showed a pattern of phosphorylation of proteins similar to that of 239wt Nef, whereas the FFPA-Nef itself was impaired in the association with cellular phosphoproteins (Fig. 4A, lanes 5 and 6). All mutant Nef proteins were expressed at comparable levels as verified by immunoblotting on whole-cell lysates (Fig. 4A, lower panel). The results show that, of the four mutations in FFAA-Nef, only P₁₀₇A is critical for the association with the serine/threonine kinase that phosphorylates these substrates. However, all mutant Nef proteins were phosphorylated in this immune complex kinase assay (Fig. 4A, lanes 3 to 6).

View larger version (58K): [in this window] [in a new window]   FIG. 4.   Kinase association and phosphorylation of mutant SIV Nef proteins. (A) N-terminal tyrosine residues are not critical for p62 association of SIV Nef. COS cells were transfected with pFJ vector control (lane 1) or pFJ constructs for 239wt-Nef (lane 2), FF-Nef (lane 3), FFAA-Nef (lane 4), FFPA-Nef (lane 5), or FFAP-Nef (lane 6). Nef immunocomplexes were precipitated and subjected to in vitro kinase assays as described previously (17). Labeled proteins were detected by autoradiography (2 h). Nef expression was detected by immunoblotting in 15 µg of whole cellular protein as shown in the lower panel using anti-AU1 antibody. (B) Y28 and Y39 of SIVmac239 Nef are not required for tyrosine phosphorylation and Src kinase association. COS cells were transfected with 5 µg of pFJ expression constructs for Src (Src), 239wt, and FFAA Nef as indicated. Nef immune complexes were precipitated with anti-AU1 antibodies, and immunoblotted proteins were detected with anti-Src antibodies (lanes 1 to 6). Src expression was verified by immunoblotting in 15 µg of whole cellular protein as shown in the lower panel using anti-Src antibodies (lanes 7 to 12). Molecular size markers (in kilodaltons) are indicated on the left of each panel. The positions of Nef, Src, heavy chains (HC), and light chains (LC) are indicated. IP, immunoprecipitation; IB, immunoblot.

Tyrosine phosphorylation and Src association. Expression constructs for 239wt and FFAA-Nef were transfected into COS cells together with pFJ expression constructs for the tyrosine kinase Src. Nef immune complexes were purified, and tyrosine phosphorylation was detected by immunoblotting. Both 239wt and FFAA-Nef showed tyrosine phosphorylation and coprecipitation of a 56-kDa tyrosine-phosphorylated protein (Fig. 4B). To verify that the tyrosine-phosphorylated 56-kDa protein in Nef immune complexes is indeed Src, the membrane was reprobed with an anti-Src antibody. Src was coprecipitated together with 239-Nef and FFAA-Nef (Fig. 4B, lower panel). Thus, these tyrosine and proline residues in SIV-Nef are not required for Src tyrosine kinase association or Nef tyrosine phosphorylation.

Clinical findings in macaques infected with the SIVmac239 FFAA-Nef variant. Three rhesus macaques (Mm7999, Mm8154, and Mm8324) were inoculated intravenously with the SIVmac239 FFAA-Nef variant. All animals showed a moderate transient anemia by 2 wpi and developed high titers of SIV-specific antibodies (Fig. 5A). At three time points prior to infection, the absolute number of CD4⁺ cells ranged from 1,228 to 2,665/mm³ in these juvenile rhesus macaques (2,168 ± 303/mm³; n = 9). All animals showed a marked reduction of CD4⁺ T cells during acute infection (day 0, 2,111 ± 390/mm³; 2 wpi, 789 ± 235/mm³; 4 wpi, 678 ± 295/mm³) (Fig. 5B). The decline was most apparent for the CD4⁺/CD29⁺ memory T-cell subset (day 0, 226 ± 47/mm³; 2 wpi, 34 ± 3/mm³; 4 wpi, 48 ± 12/mm³). In agreement with the declining CD4⁺ cell counts, the average T4/T8 ratio also declined from 1.38 ± 0.19 to 0.65 ± 0.20 within the first 2 weeks after inoculation. All animals became chronically infected and maintained CD4⁺ lymphocyte counts of >300/mm³ during 68 (Mm7999), 39 (Mm8154), and 98 (Mm824) weeks of follow-up (Fig. 5B). Mm7999 showed a mild (8 wpi) to moderate (16 wpi) lymphadenopathy and severe splenomegaly by 24 wpi. Mm7999 was euthanatized at 68 wpi because of a neoplasia at the right eye. Postmortem examination revealed an obstructive infiltratively growing malignant B-cell lymphoma at the right orbit and a moderate to severe hyperplasia to depletion of the lymphatic organs. The second animal, Mm8154, developed a moderate to marked lymphadenopathy by 12 wpi and a severe splenomegaly by 24 wpi. Mm8154 became anemic by 36 wpi, as indicated by decreased hemoglobin values (101 g/liter versus a median prevalue of 129 ± 5 g/liter; n = 5), a reduced hematocrit (29.3% versus a median prevalue of [38.6 ± 1.5]%; n = 5), and a reduced number of erythrocytes (4.5 Tera/liter versus a prevalue of 5.4 ± 0.1 Tera/liter; n = 5). The animal lost approximately 20% of its body weight during the last 7 weeks of life and had to be euthanatized at 39 wpi because of weakness. Histopathologic examination at necropsy revealed lymphoid hyperplasia and a moderate interstitial pneumonia. The remaining animal, Mm8324, developed a mild (16 wpi) to moderate (40 wpi) lymphadenopathy and a moderate splenomegaly. Multiple bouts of diarrhea were observed by 60 wpi and were induced by infections with Campylobacter strains and protozoans like Balantidium and Giardia strains. The animal was euthanatized at 98 wpi because of a tumor in the nasal tract. Histopathologic examination at necropsy identified the neoplasia as a highly malignant B-cell lymphoma. Furthermore, severe follicular hyperplasia of the lymph nodes and spleen and lymphohistiocytic infiltrates with follicular morphology in multiple other organs including brain, liver, kidney, bladder, skin, muscle, and pancreas were observed. Thus, all animals developed signs and symptoms of immunodeficiency during the course of infection.

View larger version (14K): [in this window] [in a new window]   FIG. 5.   Humoral immune response and CD4+ T-cell counts in infected macaques. The figure shows SIV enzyme-linked immunosorbent assay antibody titers (A) and absolute number of CD4+ T cells in peripheral blood (B) in three macaques, Mm7999 (), Mm8154 (), and Mm8324 (), infected with the FFAA-Nef mutant. Mm7999 had to be euthanatized at 68 wpi, and Mm8154 had to be euthanatized at 39 wpi, while Mm8324 was still alive after 98 weeks of follow-up.

Replication of the SIVmac239 FFAA-Nef variant in vivo. An intact nef gene provides a strong replicative advantage during the acute phase of SIV infection. In macaques infected with nef-open SIVmac239, the peak levels of p27 plasma viremia were about 64-fold higher than those in animals inoculated with nef-deletion SIV (Fig. 6A; Table 1). Similar differences were observed in the peak viral RNA load (Fig. 6B; Table 1). On average, the peak levels of p27 plasma viremia and of viral RNA in the three animals that received the FFAA-Nef variant were not significantly lower than for 239wt infection (Table 1). Together with the two animals previously analyzed (37), we have investigated five rhesus macaques infected with SIVmac239 containing the P107A substitution in Nef that disrupts p62 association. Altogether, the peak levels of both p27 viremia and viral RNA did not differ significantly from those observed in animals inoculated with pathogenic SIVmac239 (Fig. 6A and B; Table 1). In agreement with comparable efficiencies of viral replication, the levels of urinary neopterin, an indicator for immune activation (19), were approximately 10-fold elevated during acute infection with 239wt and the FFAA- or AxxA-Nef SIV variants. In contrast, only a threefold increase was observed after infection with attenuated nef-deletion forms (Fig. 6C; Table 1). Thus, mutations that disrupt Nef-NAK association did not attenuate SIVmac replication in rhesus macaques during the acute phase of infection.

View larger version (18K): [in this window] [in a new window]   FIG. 6.   Peak levels of viral load in rhesus macaques acutely infected with SIVmac239 Nef variants. The figure shows maximum levels of p27 plasma antigenemia (A), viral RNA (B), and urinary neopterin (C) in five animals infected with the FFAA-Nef () or AxxA-Nef () mutant (37). For comparison, values obtained from 4 (A and C) and 10 (B) macaques infected with NU and from 8 to 12 animals infected with 239wt are indicated. Peak levels of p27 plasma antigenemia and of viral RNA were always observed at 2 wpi. The neopterin/creatine ratio is expressed for each animal as the fold increase over the mean ratio determined prior to infection. Mean values were obtained from at least three samples taken within a 5-day interval and peaked between 5 and 10 or 11 and 15 days postinfection.

View larger version (23K): [in this window] [in a new window]   FIG. 7.   Replication of the SIVmac239 FFAA-Nef variant in vivo. The figure shows numbers of infectious cells per 1 million PBMC (A) and viral RNA loads (B) in Mm7999 (), Mm8154 (), and Mm8324 (), infected with the FFAA-Nef mutant. The detection limit for viral RNA is approximately 40 copies/ml (61). For comparison, average values obtained from 4 (A) and 10 (B) macaques infected with SIVmac239 NU () and from 8 to 12 animals infected with wild-type SIVmac239 () are indicated. The grey bars give standard deviations. Parameters were determined as described in Materials and Methods.

High viral load precedes reversions in SIVmac239 FFAA-Nef. To investigate if the ability of Nef to coprecipitate p62 was restored, virus was reisolated by cocultivation of PBMC with CEMx174 cells, and in vitro kinase assays were performed with positive bulk cocultivations. Significant levels of phenotypic reversions were first observed at 16 wpi in Mm8324 and at 20 wpi in Mm7999 and Mm8154 (Fig. 8 and data not shown). Notably, at 12 wpi the cell-associated viral load in these animals was comparable to that with 239wt infection (Table 1). Genotypic reversions were detected by direct sequence analysis of nef-spanning PCR fragments amplified directly from PBMC or from positive bulk cocultivations and by sequencing of single clones. As summarized in Table 2, changes in codon 107 were first observed at 16 wpi in Mm7999 and Mm8324 and at 20 wpi in Mm8154. Concordant with the faint signal at 62 kDa in in vitro kinase assays (Fig. 8), about 5% of the virus population obtained from Mm7999 at 16 wpi contained reversions (Table 2). Several additional alterations in nef were detected (Table 2). However, analysis of single nef alleles derived from the three infected animals confirmed that only those containing the A₁₀₇P reversion were positive for the Nef-NAK interaction (data not shown). A similar selective pressure seems to exist for P₁₀₄, which has no effect on Nef-NAK association, since in Mm8154 A₁₀₄P reversion was already detected by 8 wpi (Table 2). In Mm8324, a complex pattern of changes to proline residues at positions 101, 104, and 107 was observed (Table 2 and data not shown). At 16 wpi, two of six nef alleles predicted the original mutant SxxAxxA sequence, two contained PxxAxxP, and the remaining two contained PxxAxxA and SxxAxxP. At 20 wpi, four of six nef alleles predicted PxxAxxP and two predicted SxxPxxA. Nine of 10 nef alleles obtained after 20 wpi predicted a PxxPxxP motif. In contrast to the reversions in the mutated PxxP motif, no changes to N-terminal tyrosine residues Y₂₈ and Y₃₉ were observed in the three animals (Table 2). Unexpectedly, however, a F₃₉S change was first observed at 8 wpi in Mm7999 and came to predominate. Our results indicate that there is little if any selective pressure for tyrosine residues at positions 28 and 39 in SIV Nef.

View larger version (78K): [in this window] [in a new window]   FIG. 8.   Phenotypic reversions of the FFAA-Nef mutant in vivo. In vitro kinase assays were performed using positive CEMx174-PBMC bulk cocultivations obtained from Mm7999 at the indicated wpi. Nef and p27 expression levels in the lysates of infected cells were verified by immunoblotting using rabbit anti-Nef antiserum or a monoclonal anti p27 antibody (27). The right panel shows results obtained from CEMx174 cells infected with 239nef*, 239wt, and uninfected (uninf.) cells. Analogous experiments were performed with reisolates recovered from the remaining two animals infected with the FFAA-Nef mutant. Nef-NAK association was restored by 20 wpi in Mm8154 and by 16 wpi in Mm8324. Molecular masses in kilodaltons are shown at left.

View larger version (43K): [in this window] [in a new window]   FIG. 9.   Detection of reversions in codons 104 and 107 of the nef open reading frame in plasma RNA. Viral plasma RNA was isolated, reverse transcribed, and subjected to nested PCR. DNA fragments were sequenced directly, and the percentage of reversions (Rev) was estimated from standard curves as described previously (37). Only the first positions of codons 104 and 107 are shown, since nucleotide substitutions in the third position do not change the amino acid sequence. Nucleotides are shown in black (G) and grey (C). Numbers below the curves indicate the percentages of GCCC (AP) reversions.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, changes in the mutated PxxP motif that restored the Nef-NAK association were consistently observed between 12 and 20 wpi. These observations are similar to those made by Khan et al. (32), who found reversion of AxxA to AxxP in four out of seven infected macaques and reversion to PxxP in one animal. In the remaining two animals studied by Khan et al. and in our initial study, reversions were not observed (32, 37). Therefore, these three studies consistently indicate that the selective pressure for prolines at positions 104 and 107 in SIV Nef is weak, because (i) reversions occurred slowly, compared to inactivating point mutations in nef (31); (ii) they did not precede the development of a high viral load; and (iii) several macaques progressed to fatal disease in the virtual absence of viruses containing reversions of the AxxA mutations. The present study also demonstrates that mutations in the N-terminal tyrosine residues in SIV Nef, which were postulated to be important for the interaction of 239 Nef with adapter protein complexes (49), have little effect on Nef function in vivo.

The initial studies on the role of Nef in SIV pathogenesis showed that single point mutations that disrupt the nef open reading frame revert within 1 to 2 weeks after inoculation and multiple mutations or deletions that impair important aspects of Nef function result in an attenuated phenotype prior to reversion (9, 31, 62). Altogether, 12 rhesus macaques have been infected with SIVmac239 variants carrying nef alleles predicting a 107PA (CCCGCC) change that disrupts Nef association with NAK (32, 37). Four of these animals died in the virtual absence of revertant viruses (32, 37). Reversions prior to AIDS-related death were observed in the remaining eight animals. It has been suggested that the putative SH3-ligand domain plays an important role in SIV-induced immunodeficiency, because reversions were observed in most infected animals (32). However, this finding is reminiscent of the results from previous studies of vpr. Point mutations in the SIVmac239 vpr initiation codon reverted in three of five infected rhesus macaques within 4 to 8 weeks after inoculation (36). While the initial conclusion from those experiments was that vpr is important for both SIV replication in vivo and disease progression, subsequent studies revealed that a vpr deletion mutant replicated with kinetics similar to that of wild-type SIVmac239 and induced AIDS and death in infected macaques (21, 28). The finding that reversions in the mutated SIVmac Nef PxxP motif do not precede the development of a high viral load and occur only at late time points and only in a subset of animals that develop fatal disease provides evidence that these reversions are not required for efficient replication or AIDS progression.

Mutations in the PxxP motif in SIV Nef reduced the ability of Nef to stimulate viral replication in rPBMC. Surprisingly, the reduced replicative capacity of the SIVmac239 FFAA-Nef variant in vitro was not associated with a significant delay in replication kinetics or a markedly reduced viral load in acutely infected rhesus macaques. One possible explanation for why revertants are not seen early in infection is that full activity of Nef to stimulate SIV replication in rPBMC may be less important during the acute phase of infection, when a relatively high proportion of T lymphocytes might be already activated. Immune cell activation during acute infection is evident in increased levels of neopterin in serum (Table 1), and it has been shown previously that this serologic activation marker correlates with cell surface activation markers (51). Consistent with this hypothesis, few or no reversions in the mutated PxxP motif were observed in three rapidly progressing macaques, which showed high levels of immune activation until AIDS-related death (32, 37). Furthermore, in chronically infected macaques changes in the mutated PxxP motif were consistently observed between 8 and 20 wpi and not during the acute phase of infection when maximum levels of viral replication are observed. It is unclear whether this selection is for the restoration of the Nef-p62 association or for an increased ability of Nef to stimulate viral replication. We feel that the second is more likely to be important, because mutations in P₁₀₄ do also revert although they do not affect Nef-p62 association. The P(xxP)₃ motifs in HIV-1 Nef likely have higher functional relevance than the P(xxP)_1-2 motifs in SIV Nef based on accumulated data (23, 29, 32, 37, 44, 53). Consistent with this, we show that, in contrast to HIV-1 Nef (23, 44), the PxxP motif in SIV Nef is not required for MHC-I downmodulation.

The N-terminal YxxL/YxxS motifs, which resemble consensus sequences for SH2 binding domains, are present in SIV but not in HIV-1 Nef. These tyrosine residues were mutated to investigate their importance for SIV Nef function, particularly in the absence of an intact putative SH3 domain. We found that the FFAA-Nef was active in CD4 and MHC-I downregulation and able to enhance virion infectivity. The mutant Nef also associated with the tyrosine kinase Src and was phosphorylated on tyrosine residues. Our results indicate that these tyrosine and proline residues in the context of SIVmac239 Nef are not required for important interactions with SH2 or SH3 domains. This is in contrast to HIV-1 Nef, for which a strong interaction between the PxxP motif and the SH3 domain of Src family kinases has been demonstrated previously (3, 25, 38, 39). In addition to possibly being involved in SH2 interactions, it has been reported that tyrosines Y₂₈ and Y₃₉ may have a role in the association of SIVmac239 Nef with µ subunits of clathrin-adapter complexes that mediate protein sorting from the plasma membrane and the Golgi apparatus (8, 49). However, it has also been shown that these tyrosines are not important for the ability of SIV Nef to downregulate CD4 and MHC-I molecules, and the interactions of 239-Nef with adapter protein complexes that are important for CD4 downregulation are mediated by other sequences in the N-terminal region of the SIV Nef molecule (41). In the present study, to disrupt the putative SH2 domain binding function of tyrosines Y₂₈ and Y₃₉ while minimizing the effects of these mutations on protein folding, these amino acid residues were mutated to phenylalanines, and as expected, these mutations did not affect the ability of Nef to downregulate CD4 and MHC-I. No reversions to tyrosines were detected in the infected animals throughout the observation period of up to 90 weeks, thus confirming that tyrosines Y₂₈ and Y₃₉ do not have important roles, either as putative SH2 binding sites or as µ adaptin binding sites, for 239-Nef function or SIV virulence.

In conclusion, our results suggest that mutations in both the tyrosine residues and the putative SH3 ligand domain do not disrupt major aspects of SIV Nef function. We confirm previous results (32, 37) showing that the selective pressure for proline residues at positions 104 and 107 in SIV Nef is weak. More importantly, however, the viral load prior to reversion is high, suggesting that both the putative SH3-ligand domain and the N-terminal tyrosine residues in SIVmac239 Nef do not have major roles for SIV virulence.