Role of the SH3-Ligand Domain of Simian Immunodeficiency Virus Nef in Interaction with Nef-Associated Kinase and Simian AIDS in Rhesus Macaques

Abstract

The nef gene of the human and simian immunodeficiency viruses (HIV and SIV) is dispensable for viral replication in T-cell lines; however, it is essential for high virus loads and progression to simian AIDS (SAIDS) in SIV-infected adult rhesus macaques. Nef proteins from HIV type 1 (HIV-1), HIV-2, and SIV contain a proline-Xaa-Xaa-proline (PxxP) motif. The region of Nef with this motif is similar to the Src homology region 3 (SH3) ligand domain found in many cell signaling proteins. In virus-infected lymphoid cells, Nef interacts with a cellular serine/threonine kinase, designated Nef-associated kinase (NAK). In this study, analysis of viral clones containing point mutations in the nef gene of the pathogenic clone SIVmac239 revealed that several strictly conserved residues in the PxxP region were essential for Nef-NAK interaction. The results of this analysis of Nef mutations in in vitro kinase assays indicated that the PxxP region in SIV Nef was strikingly similar to the consensus sequence for SH3 ligand domains possessing the minus orientation. To test the significance of the PxxP motif of Nef for viral pathogenesis, each proline was mutated to an alanine to produce the viral clone SIVmac239-P¹⁰⁴A/P¹⁰⁷A. This clone, expressing Nef that does not associate with NAK, was inoculated into seven juvenile rhesus macaques. In vitro kinase assays were performed on virus recovered from each animal; the ability of Nef to associate with NAK was restored in five of these animals as early as 8 weeks after infection. Analysis of nef genes from these viruses revealed patterns of genotypic reversion in the mutated PxxP motif. These revertant genotypes, which included a second-site suppressor mutation, restored the ability of Nef to interact with NAK. Additionally, the proportion of revertant viruses increased progressively during the course of infection in these animals, and two of these animals developed fatal SAIDS. Taken together, these results demonstrated that in vivo selection for the ability of SIV Nef to associate with NAK was correlated with the induction of SAIDS. Accordingly, these studies implicate a role for the conserved SH3 ligand domain for Nef function in virally induced immunodeficiency.

The nef gene of primate lentiviruses (human immunodeficiency virus types 1 and 2 [HIV-1 and HIV-2] and simian immunodeficiency virus [SIV]) encodes a 27- to 35-kDa protein that is myristoylated at the N terminus and localized largely in cell membranes (8, 41, 48). This gene is dispensable for virus replication in vitro in cultures of CD4-positive T cells and macrophages. Kestler et al. have shown that expression of an intact SIV nef gene was essential for the maintenance of high viral loads and progression to simian AIDS in adult rhesus macaques (19). The importance of nef in the virus-host relationship was also highlighted by the observation that some long-term survivors (humans) of HIV-1 infection contain low levels of a virus with deletions in nef (9, 29). Nonetheless, in neonate macaques, the requirement of nef for pathogenesis can be overcome by inoculation with high doses of an SIV clone with a deletion in nef (4, 51). Thus, it appears that age is one host factor that influences the role of this viral gene in immunodeficiency disease.

Several functional properties have been ascribed to Nef of primate lentiviruses, including downregulation of the cell surface receptor CD4 and major histocompatibility complex (MHC) class I molecules on T cells, enhancement of virion infectivity, and modulation of T-cell activation (8, 41, 48). Nef was shown to exert inhibitory effects on the induction of transcription factors NF-κB and AP-1, interleukin-2, and interleukin-2 receptor alpha chain (37). Other reports described activation of T-cell proliferation by Nef, which correlated with increased virus production (1, 32). The effect of Nef on T-cell activation is most probably mediated through T-cell signaling pathways (1, 6, 47). An in vivo role for Nef in cell signaling has been investigated by experiments performed with SIV variants containing a nef allele with a signal sequence termed the immunoreceptor tyrosine-based activation motif (ITAM) (10, 28). The presence of an ITAM in the Nef of a clone of SIVmac239 enabled the virus to activate resting peripheral blood mononuclear cells (PBMC) and replicate at high levels and to produce acute fatal disease in adult macaques (10). These properties of the viral clone with an ITAM, in tissue culture cells and in animals, are similar to those of SIVpbj14, which is a variant virus that also contains this ITAM in Nef (12).

A number of cell signaling proteins, including tyrosine (Lck, Hck, Src, and Lyn) and serine/threonine kinases (protein kinase C-theta, p21-activated kinase [PAK]), have been reported to associate with Nef (reviewed in reference 41). However, the physiological relevance of the interaction of Nef with these various cell signaling proteins remains to be established. In our studies, cell extracts from HIV-1- and SIV-infected lymphoid cells were immunoprecipitated with anti-Nef antibody and the immunoprecipitates were subsequently incubated in an in vitro kinase reaction. This assay revealed two cellular proteins of 62 and 72 kDa (p62 and p72, respectively) that coimmunoprecipitated with Nef (43, 44). The kinase in these immunoprecipitates is designated Nef-associated kinase (NAK). Several lines of evidence have shown that p62 belongs to the PAK family of cellular serine kinases (27, 35, 45). However, the exact identity of p62, as well as that of p72, remains to be determined. Additional in vitro kinase assays of immunoprecipitates of infected cell extracts, performed with anti-PAK antibodies, demonstrated hyperphosphorylation of p72 in such immunoprecipitates; thus, Nef activates PAK (45). In vivo studies of an infectious SIV clone with a mutation in Nef, abrogating NAK activation, suggested that the interaction of Nef with NAK was important for viral pathogenesis in juvenile rhesus macaques (45).

One of the prominent structural features within Nef, with potential for interactions with cell signaling proteins, is a region with strong homology to a binding ligand for Src homology 3 (SH3) domains found in various tyrosine kinases and signaling adapter proteins (31, 36). The role of the SH3 ligand domain (the proline-Xaa-Xaa-proline [PxxP] motif), which is highly conserved in Nef of primate lentiviruses, was explored in vitro; this motif exhibited highly specific interactions with SH3 domains of the tyrosine kinases Hck and Lyn (15, 22, 33, 42). Mutational analysis revealed that the integrity of the PxxP motif in HIV-1 Nef was important for the enhancement of virion infectivity; in contrast, CD4 downregulation by Nef was apparently unaffected by mutations in the PxxP motif (42). The Nef-mediated enhancement of HIV-1 virion infectivity has been correlated with Nef-NAK interaction, which was also shown to depend on the integrity of the PxxP region of Nef (14, 16, 50). Nevertheless, a recent study examining the role of the SH3 ligand domain of SIV Nef in rhesus macaques infected with a viral clone with mutations in the PxxP motif (P¹⁰⁴ KVP¹⁰⁷) concluded that this highly conserved motif was not important for progression to simian AIDS (SAIDS) (20).

We examined the SH3 ligand domain of SIV Nef (containing the PxxP motif P¹⁰⁴KVP¹⁰⁷) in Nef-NAK interactions by analyzing SIV clones containing point mutations in this domain in lymphoid cell cultures. Several key amino acid residues in this domain of Nef were shown to be necessary for Nef-NAK interaction and PAK activation. These experiments defined a consensus sequence, which is very similar to the sequence shown to be important in SH3 domain/SH3 ligand interactions in HIV-1 Nef (15, 22, 33, 42). Additionally, the in vivo role of the SH3 ligand domain for pathogenesis was investigated by the inoculation of seven juvenile rhesus macaques with an SIVmac239 mutant, which encodes a Nef in which both prolines of the SH3 ligand domain were mutated to alanines. Reversion of the Nef mutations in animals inoculated with this mutant virus revealed strong selective pressure for restoration of the SH3 ligand domain. These studies also demonstrated that the ability of Nef to associate with NAK and activate PAK was correlated with progression to fatal SAIDS.

MATERIALS AND METHODS

Cells and antibodies.

CEMx174 cells, of the human lymphoid T/B-cell hybrid cell line, which is permissive to replication of various SIVmac strains, were provided by James Hoxie (University of Pennsylvania). Rabbit polyclonal antiserum was generated against SIV Nef expressed in Escherichia coli (recombinant Nef protein was provided by Casey Morrow, University of Alabama). An SIV Nef monoclonal antibody (17.2) was a generous gift from Kai Krohn (University of Tampere, Tampere, Finland); this antibody was raised against a synthetic peptide between amino acids 69 and 75 of SIVmac251 Nef and detects SIVmac239 Nef. Rabbit polyclonal anti-rat PAK-1 antibody was purchased from Santa Cruz Biotechnologies (Santa Cruz, Calif.); this antibody was raised against the N-terminal peptide (20 amino acids) of rat PAK-1 and detects human PAK-1.

Construction of SIV nef point mutants.

The SIVmac239nef+ clone was produced by mutating the premature stop codon in nef of the original pathogenic SIVmac239 clone (GenBank accession no. M33262) to a codon for glutamate, which is the most common codon in revertant viruses recovered from macaques infected with this latter viral clone (19). The nef gene and protein in SIVmac239nef+ are designated “prototype.”

The proviral genome of SIVmac239nef+ has been cloned in two halves, divided at the unique SphI restriction site (in the vpr gene at position 6707) for ease of handling of the proviral DNA (5). Both halves were inserted into the pGEM7 vector (Promega, Madison, Wis.); the 5′ half of the viral clone was designated pVP-1, and the 3′ half was designated pVP-2nef (36a). Mutants with point mutations in the PxxP region of Nef were generated, either by PCR or by oligonucleotide mutagenesis, to change codons of various amino acids to alanine codons (Fig. 1A). For PCR mutagenesis, the primers contained two convenient restriction sites, a BglII site (position 9375) at the 5′ end of the 5′ primer, and an AflIII site (position 9686) at the 3′ end of the 3′ primer. Thus, DNA fragments (311 bp) amplified with these primers contained BglII sites at their 5′ ends and AflIII sites at their 3′ ends. The 3′-end primer also contained the intended point mutations in the nef gene. After PCR amplification, the amplified DNA fragments were cloned directly into the pCRII vector as specified by the manufacturer TA cloning kit; Invitrogen, San Diego, Calif.). All mutations were confirmed by DNA sequencing. The mutant fragments between the BglII and AflIII sites were then used to replace the corresponding fragment in pVP-2(nef+).

FIG. 1.

In vitro kinase assay of SIVmac239 Nef mutants for Nef association with NAK and activation of PAK. (A) The conserved SH3 ligand domains from HIV-1 and SIV Nef are represented by amino acid sequences from the HIV-1-NL4-3 and SIVmac239 clones. SIV Nef mutants that affected NAK association and PAK activation are denoted by asterisks. For all mutants, the codon for the prototype amino acid was changed to the codon for alanine. (B and C) In vitro kinase assays were performed on anti-SIV Nef (B) and anti-PAK-1 (C) immunoprecipitates from extracts of uninfected CEMx174 cells and CEMx174 cells chronically infected with SIVmac239nef+, SIVmac239-RR/LL, SIVmac239-  P¹⁰⁴A/ P¹⁰⁷A, SIVmac239-R¹⁰³A, SIVmac239-P¹⁰⁴A, SIVmac239-K¹⁰⁵A, SIVmac239- V¹⁰⁶A, SIVmac239-P¹⁰⁷A, SIVmac239-L¹⁰⁸A, SIVmac239-R¹⁰⁹A, SIVmac239-F¹²²A, and SIVmac239Δnef. A total of 10⁷ cells were analyzed per lane. (D) The level of Nef expression in cell lines infected with prototype and mutant viruses was determined by immunoblot analysis with the anti-SIV Nef monoclonal antibody 17.2. The lanes are the same as in panels B and C. Immunoprecipitates were electrophoresed on a 12% polyacrylamide gel under denaturing conditions and transferred to PVDF membranes. Phosphorylation of proteins was visualized by autoradiography. Kinase and immunoblot assays were performed as described in Materials and Methods.

For mutagenesis by oligonucleotides, a DNA fragment containing part of the nef gene between the SacI (position 9487) and NdeI (position 10008) sites in pVP-2(nef+), was subcloned into a derivative of pUC19 that lacked the AflIII restriction site. A unique ClaI site was then engineered at position 9600, without resulting in amino acid changes in Nef; this clone was designated pIK3. The PxxP region of Nef was encoded between the new ClaI site and the unique AflIII site in the nef gene. Mutant oligonucleotides between ClaI and AflIII sites (86 bp) were designed to contain the intended point mutations in the region encoding PxxP. The double-stranded oligonucleotides representing the mutant DNA fragments were used to replace the corresponding prototype ClaI-AflIII fragment in pIK3. All the mutations were confirmed by DNA sequencing. Finally, the SacI-NdeI fragment from pIK3, containing the point mutation(s) in the region encoding PxxP, replaced the corresponding fragment in pVP-2(nef+). Complete nef mutant proviruses were obtained by joining pVP-1 and the various mutant pVP-2 plasmids at the SphI site.

Production of virus stocks from proviral clones.

Plasmids containing full-length SIVmac239nef+ proviral DNA and mutant proviral DNA were transfected into CEMx174 cells. Transfections were performed in duplicate by electroporation as previously described (49). Briefly, exponentially growing cells were resuspended in serum-free RPMI medium at a concentration of 10⁷ cells/ml. Plasmid DNA (5 μg) was mixed with 0.4 ml of this cell suspension, and electroporation was performed in a 0.4-cm cuvette at 960 μF capacitance and 200 V by using a gene pulser (Bio-Rad, Richmond, Calif.). Electroporated cells were cultured and microscopically monitored at daily intervals for virus production by the appearance of cytopathic effects such as multinucleated syncytia and giant cells. After 5 to 8 days of culture, cell-free virus stocks were obtained by removal of cells by centrifugation (2,500 × g for 5 min) and removal of cell debris from the supernatant by filtration through 0.45-μm-pore-size filters. Virus stocks were stored frozen at −70°C in 1-ml aliquots. Titers of virus stocks were determined in microtiter plates containing CEMx174 cells; the 50% tissue culture infective dose (TCID₅₀) was calculated by using the end-point dilution method of Reed and Muench (38).

Chronically infected cell lines.

CEMx174 cells chronically infected with SIVmac239nef+ and mutant viruses were derived as outgrowths of acute infection with the viruses as previously described (43). These cell lines were stored frozen at 5 × 10⁶ cells/ml at −135°C in RPMI containing 20% fetal calf serum and 10% dimethyl sulfoxide.

In vitro kinase assays and anti-Nef immunoblots.

In vitro kinase assays were performed, as described previously (43), on immunoprecipitates obtained from lysates of virally infected CEMx174 cells by using either rabbit anti-SIV Nef antibody or rabbit anti-rat PAK-1 antibody. For anti-Nef immunoblots, proteins were transferred from sodium dodecyl sulfate-polyacrylamide gels to polyvinylidene difluoride (PVDF) membranes (Bio-Rad). After three washes with Blotto buffer (5% nonfat dry milk in phosphate-buffered saline [pH 7.4]), the PVDF membranes were incubated with anti-SIV Nef monoclonal antibody (antibody 17.2). The membranes were washed three times with wash buffer (250 mM NaCl, 50 mM Tris [pH 7.4], 0.1% Tween 20) and incubated with goat anti-mouse secondary antibody conjugated to alkaline phosphatase (Southern Biotechnology Associates, Birmingham, Ala.). After being washed three times with wash buffer and once with 100 mM Tris (pH 9.5), the blots were developed with the BCIP/NBT detection kit (Vector Laboratories, Burlingame, Calif.).

Inoculation of rhesus macaques.

All the animals in the study were colony-bred juvenile rhesus macaques (Macaca mulatta) free of simian retrovirus (SRV) type D, SIV, and simian T-lymphotropic virus; these animals were housed at the California Regional Primate Research Center at the University of California, Davis. Seven macaques were intravenously inoculated with 1,000 TCID₅₀ of mutant virus carrying the proline-to-alanine mutation in the two prolines within the SH3 ligand domain of Nef (SIVmac239-P¹⁰⁴A/P¹⁰⁷A). Two additional animals each received an intravenous injection of 1,000 TCID₅₀ of the pathogenic clone SIVmac239nef+. At fixed intervals, blood samples were drawn from the inoculated animals for virus isolation, complete blood cell count, CD4 and CD8 cell count (by flow cytometry), and determination of anti-SIV antibody levels (with the HIV-2 antibody kit [Genetic Systems, Shasta, Minn.]). Virus load measurements were made on plasma samples obtained from EDTA-anticoagulated blood by a branched-chain DNA (bDNA) assay developed for quantitating SIV RNA copies in plasma (P. Dailey and J. Booth, Chiron Diagnostics, Chiron Reference Testing Laboratory, Emeryville, Calif.). Complete physical examinations were performed on the animals at regular intervals. The animals were also monitored for body weight and clinical signs of disease. Opportunistic infections were diagnosed by standard microbiological techniques performed on clinical samples at the Clinical Microbiology Laboratory, California Regional Primate Research Center. Macaques that became seriously ill and were nonresponsive to therapeutic interventions (e.g., enhanced diets and antibiotics) were humanely euthanized with a barbiturate overdose. Some macaques were tested for SRV at necropsy by sensitive PCR amplification of DNA from PBMC and lymph nodes with SRV primers (24) and by immunoblot analysis of plasma for antibodies to whole virus (23).

PCR amplification and sequencing of nef genes from infected animals.

DNA was isolated, using the Qiagen blood kit as specified by the manufacturer (Qiagen Inc., Chatsworth, Calif.), from either 400 μl of whole blood or 10⁷ PBMC obtained from infected animals. The DNA thus obtained was used as the template for amplification of nef by nested PCR. Two PCR amplifications were performed per time point. In the first round (30 cycles) of PCR, the 5′ and 3′ primers 5′- CCAGAGGCTCTCTGCGACCCTAC and 5′-AGAGGGCTTTAAGCAAGCAAGCGTG, respectively, were used for nef amplification. For the second round (30 cycles), the 5′-primer remained the same while the 3′- primer was 5′-GCCTCTCCGCAGAGCGACTGAATAC. PCR amplifications were performed in a DNA thermal cycler (Perkin-Elmer, Foster City, Calif.) with Taq polymerase (Perkin-Elmer). The final product (992 bp), containing the full-length nef gene, was directly cloned into the pCRII vector, and the DNA was sequenced on both strands.

Isolation of virus from PBMC of infected macaques.

Virus isolations were performed as previously described by coculturing PBMC from infected animals with CEMx174 cells (30). For in vitro kinase assays, acutely infected CEMx174 cells, produced by cocultivation with PBMC from macaques infected with SIVmac239-P¹⁰⁴A/P¹⁰⁷A, were lysed as described previously (45). The lysates were analyzed in the in vitro kinase assays as described above to test for the restoration of the ability of Nef to associate with NAK.

Histological examination of tissues from infected rhesus macaques.

To assess histopathologic changes during the course of infection, axillary and inguinal lymph nodes were obtained by percutaneous biopsy under ketamine hydrochloride anesthesia. Lymph node biopsy specimens and all tissues collected at necropsy were fixed in 10% buffered formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin for microscopic examination.

RESULTS

Analysis of SIV Nef point mutants in in vitro kinase assays.

To examine the role of the SH3 ligand domain of Nef in in vitro kinase assays, which measure the association of Nef with NAK (the cellular serine kinase) and activation of PAK in lymphoid cells, several viruses were constructed with point mutations (i.e., alanine substitutions) spanning the highly conserved PxxP region of SIVmac239 Nef. This region is R¹⁰³ P¹⁰⁴ K¹⁰⁵V¹⁰⁶P¹⁰⁷L¹⁰⁸R¹⁰⁹, as illustrated in Fig. 1A. One additional point mutant (also an alanine substitution) was in a highly conserved phenylalanine residue (F¹²²), located 12 residues C-terminal to the PxxP motif in Nef. This latter mutant was constructed because of the predicted importance of the counterpart residue F⁹⁰ of HIV-1 Nef in the interaction with the SH3 domain of the tyrosine kinase Hck (22). The in vitro kinase analysis of anti-Nef immunoprecipitates of extracts from chronically infected cell lines revealed that several but not all residues in the PxxP region of SIVmac239 Nef were important in its association with NAK. The double proline mutation, P¹⁰⁴A/P¹⁰⁷A, abrogated the ability of Nef to associate with NAK (Fig. 1B, lane 4). Of the two prolines (P¹⁰⁴ and P¹⁰⁷), P¹⁰⁷ was critical for the association of Nef with NAK (lanes 6 and 9). Four other residues in the PxxP region, V¹⁰⁶, L¹⁰⁸, R¹⁰⁹, and F¹²², were also essential for Nef association with NAK (lanes 8, 10, 11, and 12). Mutations in residues R¹⁰³ and K¹⁰⁵ did not have significant effects on Nef association with NAK in chronically infected lymphoid cells (lanes 5 and 7).

The ability of these SIV Nef mutants to activate PAK, as indicated by hyperphosphorylation of p72 (45), was also investigated. These kinase assays were performed on anti-PAK-1 immunoprecipitates from CEMx174 cells chronically infected with all of the Nef mutant viruses described above. The same Nef mutants which were defective in association with NAK (P¹⁰⁴A/P¹⁰⁷A, V¹⁰⁶A, P¹⁰⁷A, L¹⁰⁸A, R¹⁰⁹A and F¹²²A [Fig. 1B]) were also defective in activation of PAK (Fig. 1C, lanes 4 and 8 through 12). This finding was consistent with our previously reported results showing that Nef mutants with mutations in the double-arginine motif (R¹³⁷R¹³⁸) were defective in both NAK association and activation of PAK (Fig. 1B and C, lanes 3) (45). Within the SH3 ligand domain of SIV Nef, the residue P¹⁰⁴ could be altered to A without affecting association or activation of PAK in the in vitro kinase assays (Fig. 1B and C, lanes 6). Because prolines in SH3 ligand domains of cellular proteins contribute to SH3 binding through hydrophobic interactions (25), it is possible that in the absence of P¹⁰⁴ in Nef, some other neighboring hydrophobic residue(s) (e.g., V¹⁰²) contributes to the interaction with NAK. Immunoblot analysis demonstrated that in CEMx174 cell lines chronically infected with prototype SIVmac239 or various Nef mutant viruses, all of the mutant Nef proteins were expressed at levels similar to those in the cells infected with prototype Nef (Fig. 1D). Taken together, the analysis of these SIV mutants in the NAK association assay (on immunoprecipitates prepared with anti-Nef antibody) and the PAK activation assay (on immunoprecipitates prepared with anti-PAK-1 antibody) demonstrate the importance of V¹⁰⁶P¹⁰⁷L¹⁰⁸R¹⁰⁹, and F¹²² in the SH3 ligand domain of Nef.

Clinical outcome of macaques infected with SIVmac239- P¹⁰⁴A/P¹⁰⁷A.

All the rhesus macaques for this study were juvenile animals, ranging in age from 27 to 49 months at the time of inoculation. Seven macaques received a cell-free preparation of SIVmac239-P¹⁰⁴A/P¹⁰⁷A at 1,000 TCID₅₀ by the intravenous route. To serve as comparisons, two macaques were inoculated with the pathogenic molecular clone SIVmac239nef+. All the animals were monitored for virus load, antiviral antibodies, and hematological and clinical signs of immunodeficiency.

At the beginning of this in vivo study of nef function, a group of four juvenile macaques were inoculated with SIVmac239- P¹⁰⁴A/P¹⁰⁷A. Two macaques in this group (Mmu 25905 and Mmu 27659) showed high levels of viral RNA in plasma in the first 2 to 4 weeks of infection and a decline of about 3 orders of magnitude by 12 weeks after inoculation (Fig. 2A). Thereafter, the virus load rose in Mmu 27659 and remained above the detection limit in Mmu 25905. During the course of infection, both animals exhibited clinical signs consistent with progression to SAIDS, including a decline in CD4 T-cell counts to levels below the lower limit for the reference range for healthy uninfected animals (i.e., below 500 CD4 T cells/μl) (Fig. 3A and Table 1). Necropsy performed on Mmu 25905, which was euthanized 32 weeks after infection, revealed depletion in several lymphoid organs, the presence of Mycobacterium avium, and other opportunistic bacterial infections in segments of the gastrointestinal tract (Table 1). Mmu 27659, necropsied at 46 weeks, showed lymphoid hyperplasia and lymphoid-cell depletion; another prominent feature was a gastritis associated with the opportunistic agent Helicobacter (Table 1). Thus, both animals exhibited several signs of SAIDS. Virus recovered form both of these animals at several time points during the course of infection and at necropsy showed phenotypic and genotypic reversions in Nef (see below and Table 2).

FIG. 2.

Virus load measured by bDNA assay in plasma from infected macaques (expressed as viral RNA copies per milliliter of plasma). Solid lines indicate animals that were euthanized within the study, and the stars show the time at which each animal was euthanized for necropsy. Dashed lines indicate animals that were alive at the conclusion of the study. The lower limit of detection for the bDNA assay is 10,000 copies of viral RNA per ml. (A) Virus load in plasma in seven macaques infected by the intravenous route with the mutant SIVmac239-P¹⁰⁴A/P¹⁰⁷A. (B) Virus load in plasma in control macaques infected with virus containing full-length nef, SIVmac239nef+.

FIG. 3.

CD4 cell count in peripheral blood of infected macaques. Lymphocyte subsets in peripheral blood were analyzed by flow cytometry, in a FACScan (Becton Dickinson, Mountain View, Calif.) with marker antibodies recognizing CD4 T cells (OKT4), CD8 T cells (Leu2A), CD2 T cells (Leu5b), and CD19 B cells (Leu16). Solid lines indicate animals that were euthanized within the study, and the stars show the time at which each animal was euthanized for necropsy. Dashed lines indicate animals that were alive at the conclusion of the study. (A) CD4 cell count in seven macaques infected with SIVmac239-P¹⁰⁴A/P¹⁰⁷A. (B) CD4 cell count in control macaques infected with virus containing full-length nef, SIVmac239nef+.

TABLE 1.

Clinical and pathological findings in juvenile macaques infected with SIVmac239-P¹⁰⁴A/P¹⁰⁷A and SIVmac239nef⁺

Virus and animala	Phenotype reversionb	Sequence of Nefc	Clinical signs and premortem laboratory data	Postmortem findings
SIVmac239-P104A/P107A
Mmu 25905	Yes (20 wk p.i.d)	AKVP	Lymphadenopathy and splenomegaly at 4 to 20 wk p.i., Staphylococcus rhinitis at 2 wk p.i., lymphoid hyperplasia at 8 wk p.i.	Necropsy at 32 wk p.i., lymphoid hyperplasia to lymphoid depletion, multiorgan bacterial infections (gastritis [Helicobacter], colitis [spirochetosis]), Mycobacterium avium
Mmu 27659	Yes (12 wk p.i.)	PRVA, PKVP	Marked weight loss (10%), partial anorexia, diarrhea, jaundice at 46 wk p.i., lymphoid hyperplasia at 32 wk p.i., neutrophilic leukocytosis and CD4 T-cell lymphocytopenia at 42–46 wk p.i., rectal culture positive for Balantidium coli	Necropsy at 46 wk p.i.; lymphoid hyperplasia to lymphoid depletion; thymic atrophy and wasting; jejunal villus blunting and atrophy; lymphocytic, plasmacytic, eosinophilic enteritis and colitis; bacterial gastritis (Helicobacter); lymphocytic, plasmacytic, eosinophilic cholangiohepatitis; chronic peritonitis
Mmu 26785	No	AKVA	Severe weight loss (15%), anorexia, diarrhea, severe hematologic changes (microcytic hypochromic anemia, hypoproteinemia at 4–19 wk p.i., neutropenia 2–4 wk p.i., marked eosinophilia 4–12 wk p.i., thrombocytopenia)	Necropsy at 19 wk p.i.; peripheral and visceral lymph node depletion; multifocal cytomegalovirus neuritis and pneumonitis with intranuclear inclusion bodies and multinucleated giant cells; lymphocytic and eosinophilic colitis with bacterial colonization; histiocytic, lymphocytic enteritis with villus atrophy; membranous glomerulonephritis
Mmu 27626	No	AKVA	Thrombocytopenia at 61–93 wk p.i., monocytosis at 61–93 wk p.i., normal weight gain	Necropsy at 93 wk p.i., hepatitis with marked liver pathology, widespread lymphoid hyperplasia, no evidence of lymphoid depletion, moderate multiorgan vasculitis, B-cell lymphoma and ulcer in ileum, no evidence of opportunistic infections
Mmu 27879	Yes (12 wk p.i.)	AKVP	Persistent lymphadenopathy, splenomegaly from 8–53 wk p.i.; hyperproteinemia at 24– 42 wk p.i.; thrombocytopenia at 24–53 wk p.i.; persistent bacterial rhinitis at 16– 53 wk p.i.; decreased CD4/CD8 cell ratio	Alive 85 wk p.i.
Mmu 28000	Yes (12 wk p.i.)	AKVP	Transient lymphadenopathy during primary viremia at 3 wk p.i.	Alive 85 wk p.i.
Mmu 26902	Yes (8 wk p.i.)	AKVP	Transient lymphadenopathy during primary viremia at 3 wk p.i.; persistent lymphadenopathy, splenomegaly from 8–34 wk p.i.; thrombocytopenia at 20–77 wk p.i.; decreased CD4/CD8 cell ratio	Alive 77 wk p.i.
SIVmac239nef+
Mmu 26084	NAe	Prototype (PKVP)	Weight loss, diarrhea, lymphadenopathy, thrombocytopenia, splenomegaly, pneumonia	Necropsy at 51 wk p.i.; lymphoid depletion, thymic atrophy, Pneumocystis carinii pneumonia; encephalitis
Mmu 27098	NA	Prototype (PKVP)	Weight loss, diarrhea, lymphadenopathy, persistent bacterial rhinitis	Necropsy at 87 wk p.i., lymphoid hyperplasia to depletion, gastrointestinal giardiasis and cryptosporidiosis

^aAge at inoculation: Mmu 25905, 48 months; Mmu 27659, 27 months; Mmu 26785, 35 months; Mmu 27626, 27 months; Mmu 27879, 36 months; Mmu 28000, 29 months; Mmu 26902, 49 months; Mmu 26873, 31 months; Mmu 26939, 30 months; Mmu 26084, 39 months; Mmu 27098, 26 months. 

^bPhenotype reversion: Nef is positive for association with NAK (weeks postinfection at which NAK-positive phenotype was detected). 

^cSequence of Nef based on data in Table 2; the prototype sequence of SIV Nef is P¹⁰⁴K¹⁰⁵V¹⁰⁶P¹⁰⁷. 

^dp.i., postinoculation. 

^eNA, not applicable. 

TABLE 2.

Genotypic reversions in Nef of viruses recovered from juvenile rhesus macaques infected with SIVmac239-P¹⁰⁴A/P¹⁰⁷A

Animal	Time (wk) postinfectiona	No. of clones	Amino acid sequence from 104 to 107b
Mmu 25905	20	16	AKVP
	30	15	AKVP
Mmu 27659	12	3	PRVA
		11	AKVA
	24	1	PKVP
		9	PRVA
	42	6	PKVP
		6	PRVA
	46	7	PKVP
		3	PRVA
Mmu 26785	19	10	AKVA
Mmu 27626	93	4	AKVA
		5	AKVT
Mmu 27879	12	16	AKVP
		4	AKVA
	24	19	AKVP
Mmu 28000	12	15	AKVP
		9	AKVA
	17	8	AKVP
Mmu 26902	8	17	AKVP
	12	26	AKVP

^aTime of virus recovery. 

^bDNA sequence analysis was conducted on PCR amplified nef clones from viruses recovered from macaques infected with the mutant virus. The genotypic reversions are presented as the deduced amino acid sequence. The prototype Nef sequence is P¹⁰⁴K¹⁰⁵V¹⁰⁶P¹⁰⁷. 

The third macaque, Mmu 26785, in this group of four animals inoculated with SIVmac239-P¹⁰⁴A/P¹⁰⁷A, showed very high levels of viral RNA in plasma throughout the course of infection (Fig. 2A). Interestingly, the CD4 T-cell counts in this animal remained in the normal range (Fig. 3A). An unusual characteristic in Mmu 26785, not commonly found in rhesus macaques infected with various pathogenic strains of SIV, was a lack of detectable anti-SIV antibodies in plasma (see below). Additionally, very high levels of eosinophils were detected in the peripheral blood of Mmu 26785. The percentages of leukocytes identified as eosinophils in this animal preinoculation and during the course of infection are as follows: preinoculation, 8%; 4 weeks, 60%; 8 weeks, 41%; 12 weeks, 21%; and 19 weeks, 1%. Because of severe weight loss and untreatable diarrhea, this animal was euthanized at 19 weeks after virus inoculation. SRV was not detected at necropsy by PCR amplification and immunoblot analysis. Histopathologic analysis at necropsy revealed lymphoid cell depletion, infection with cytomegalovirus, and inflammation in the gastrointestinal tract (Table 1). Virus recovered from Mmu 26785, both during the course of infection and at necropsy did not show reversion in Nef (see below and Table 2).

The remaining animal in this group of four macaques inoculated with SIVmac239-P¹⁰⁴A/P¹⁰⁷A, Mmu 27626, showed an initial high virus load at 2 weeks followed by a decline of about 2 orders of magnitude for the remainder of the infection (Fig. 2A). CD4 T-cell counts were in the normal range until about 69 weeks, when levels fell below the normal range but remained at 160 to 500 cells/μl until necropsy at 93 weeks (Fig. 3A). Histopathologic analysis at necropsy of Mmu 27626 revealed widespread lymphoid hyperplasia with no evidence of lymphoid cell depletion (Table 1). No signs of opportunistic infection were evident. SRV was not detected at necropsy by PCR amplification and immunoblot analysis. A prominent feature encountered in the postmortem examination was an ulcer caused by an extranodal lymphosarcoma in the ileum; immunohistochemical analysis demonstrated that the cells in this tumor reacted predominantly with antibody to a B-cell marker (CD 20) (data not shown). The virus recovered at necropsy from Mmu 27626 tested negative in the NAK assay (data not shown). Analysis of nef gene sequences of virus at necropsy revealed a mixture including the mutant sequence A¹⁰⁴KVA¹⁰⁷ and the novel sequence A¹⁰⁴KVT¹⁰⁷ (see below and Table 2).

Three additional juvenile macaques (Mmu 26902, Mmu 27879, and Mmu 28000) were infected with SIVmac239- P¹⁰⁴A/P¹⁰⁷A about 1 year after inoculation of the above-mentioned macaques, to provide a larger number of animals for this in vivo genetic analysis of SIV nef function. After the initial peak of viremia, virus levels declined 1 to 2 orders of magnitude in Mmu 26902 and Mmu 27879. In Mmu 28000, virus levels in plasma declined to below the level of detection (Fig. 2A). In these three animals, the number of CD4 T cells remained in the normal range during the course of infection, except for that in Mmu 26902, which showed a decline to below 500 CD4 T cells per μl in peripheral blood at 61 weeks after infection (Fig. 3A). Interestingly, the two animals with the higher virus load, Mmu 27879 and Mmu 26902, displayed signs consistent with progression to SAIDS, including persistent generalized lymphadenopathy, decline of the CD4/CD8 T-cell ratio, and thrombocytopenia (Table 1). The third macaque, Mmu 28000, exhibited no abnormal hematolgical findings (Fig. 3A; Table 1). Levels of anti-SIV antibody titers in plasma were in the high range in Mmu 27879 and in the moderate range in Mmu 26902 and Mmu 28000 (see below). The nef gene of the viruses from these three animals reverted to the NAK-positive phenotype and genotype (see below and Table 2). The time taken to develop fatal SAIDS after infection with pathogenic SIVmac239 is variable (19). Long observation periods are warranted because one adult rhesus macaque infected with pathogenic SIVmac239 required 5 years to develop fatal SAIDS (29a). Accordingly, Mmu 27879, Mmu 26902, and Mmu 28000 will continue to be monitored for virological, immunological, and clinical parameters.

Two control macaques infected with SIVmac239nef+ (Mmu 26084 and Mmu 27098) displayed a high peak of viral RNA levels in plasma 2 weeks postinfection followed by a decline of 1 to 2 orders of magnitude during the remainder of the infection (Fig. 2B). These two animals showed several signs of SAIDS both during the course of infection and at necropsy (Table 1; Fig. 3B).

Antiviral immune responses in infected macaques.

Plasma samples from the macaques in this study were tested in a quantitative enzyme-linked immunosorbent assay which measured antibody levels against the whole virus. By 8 weeks postinfection, all animals except Mmu 26785 had antibody levels in plasma greater than 1/3,200; in general, analysis of plasma samples collected 24 weeks postinfection and at later time points showed that these antibody levels continued to rise. The following titers for anti-SIV antibodies were measured in the 24-week plasma samples: Mmu 25905, 1/51,200; Mmu 27659, 1/102,400; Mmu 27626, 1/102,400; Mmu 27879, 1/409,600; Mmu 28000, 1/51,200; and Mmu 26902, 1/204,800. In striking contrast, antibody titers in plasma samples collected from Mmu 26785 at 8, 12, and 19 weeks were below the detection limit of this assay (<1/100). Other investigators have also reported undetectable antiviral antibody levels in adult rhesus macaques exhibiting rapid disease progression after infection with pathogenic strains of SIV (17, 53).

In vivo genotypic and phenotypic reversions in nef.

To investigate potential reversions in the nef sequence (genotypic reversion) and Nef function (phenotypic reversion), virus was isolated at various time points from PBMC obtained from macaques infected with the mutant clone SIVmac239- P¹⁰⁴A/P¹⁰⁷A. In these recovered viruses, DNA from the infected cells was used for PCR amplification to determine the sequence of nef genes and Nef function was analyzed in in vitro kinase assays.

Reversions in the nef sequence were first detected at the following time points: Mmu 26902 at 8 weeks postinoculation; Mmu 27659, Mmu 27879, and Mmu 28000 at 12 weeks; and Mmu 25905 at 20 weeks (Table 2). The virus isolated from this group of macaques, except Mmu 27659, displayed a change of A¹⁰⁷ to P. This pattern of change, which produced the sequence A¹⁰⁴KVP¹⁰⁷, was a revertant phenotype, as demonstrated in the previous analysis of mutants with point mutations in the PxxP region in the in vitro kinase assay (Fig. 1B and C); this assay revealed that the genotype A¹⁰⁴KVP¹⁰⁷ was positive for NAK association and PAK activation. In most of these animals, the proportion of nef clones with the sequence positive for Nef-NAK interaction increased at subsequent time points (Table 2). As an example of phenotypic reversion for the A¹⁰⁴KVP¹⁰⁷ genotype, the results of an in vitro kinase assay are shown for virus isolated from Mmu 25905 at 20 and 30 weeks postinoculation; strong phosphorylation of both p62 and p72 was observed (Fig. 4A, lanes 4 and 5). Taken together, these findings demonstrated significant selection pressure in vivo for the ability of Nef to associate with NAK.

FIG. 4.

Phenotypic reversions of the SIVmac239-P¹⁰⁴A/P¹⁰⁷A mutant in macaques. In vitro kinase assays were performed on anti-Nef immunoprecipitates from uninfected CEMx174 cells and from CEMx174 cells infected with prototype virus and viruses recovered from several macaques infected with the mutant virus. (A) Control kinase assays were done on anti-SIV Nef immunoprecipitates from extracts of uninfected CEMx174 cells and on extracts of CEMx174 cells chronically infected with SIVmac239nef+ or SIVmac239-P¹⁰⁴A/P¹⁰⁷A. In vitro kinase assay results are shown for anti-Nef immunoprecipitates from extracts of CEMx174 cells acutely infected with virus isolated from Mmu 25905 at weeks 20 (20wk) and 30, with virus from Mmu 27659 at weeks 8 and 24, and with virus from Mmu 27626 at week 24 postinoculation. (B) In vitro kinase assay performed on virus isolated from Mmu 27626 and Mmu 27659 at 42 weeks (42wk) postinoculation, analyzed in the same manner as in panel A with controls: in vitro kinase analysis of anti-SIV Nef immunoprecipitates from extracts of uninfected CEMx174 cells and CEMx174 cells chronically infected with SIVmac239nef+. (C) Association of Nef with NAK in the mutant virus, SIVmac239-K¹⁰⁵R/P¹⁰⁷A; the mutation in nef in this virus was constructed on the basis of the second-site revertant sequence in virus isolated from Mmu 27659 (Table 2). Results of in vitro kinase analysis of anti-SIV Nef immunoprecipitates from uninfected CEMx174 cells and from CEMx174 cells chronically infected with SIVmac239nef+ and SIVmac239-K¹⁰⁵R/P¹⁰⁷A are shown. A total of 10⁷ cells were analyzed per lane. The kinase assays were performed as described in Materials and Methods. Immunoprecipitates were electrophoresed on a 12% polyacrylamide gel under denaturing conditions, and phosphorylation of proteins was visualized by autoradiography.

The pattern of genotypic reversions observed in the virus isolated from Mmu 27659 was more complex than the pattern in the viruses from the other animals. At 12 weeks after inoculation of Mmu 27659, 3 of 10 nef clones displayed reversion at position 104 (changing A¹⁰⁴ to P) whereas alanine at position 107 (A¹⁰⁷) remained unchanged (Table 2). Based on the analysis of point mutants in the in vitro kinase assay, this change of A¹⁰⁴ to P alone would not result in association of Nef with NAK (Fig. 1B and C). However, all three of the Nef clones containing the A¹⁰⁴ to P reversion also contained a second-site mutation at position 105, changing K¹⁰⁵ to R (Table 2). To determine whether Nef with this revertant sequence was capable of association with NAK, this sequence, P¹⁰⁴R¹⁰⁵VA¹⁰⁷, was introduced into the SIVmac239nef+ clone to produce the mutant clone designated SIVmac239-K¹⁰⁵R/P¹⁰⁷A. This mutant virus revealed a low but positive level of Nef association with NAK (Fig. 4C, lane 3). The level of Nef-NAK interaction for this mutant Nef was similar to the level detected in virus isolated from Mmu 27659 at 24 weeks postinoculation (Fig. 4A, lane 7), at which time point the mutant sequence P¹⁰⁴RVA¹⁰⁷ predominated (Table 2). Additionally, the nef clones in Mmu 27659 exhibited further evolution during the course of infection. Importantly, at 42 weeks, 6 of 12 nef clones displayed reversion to the prototype sequence (P¹⁰⁴KVP¹⁰⁷) whereas the other 6 nef clones still contained the P¹⁰⁴R¹⁰⁵VA¹⁰⁷ sequence (Table 2); both of these genotypes are positive for the NAK association phenotype. This increase in the proportion of virus with the P¹⁰⁴KVP¹⁰⁷ genotype was reflected in the striking increase in the phosphorylation of p62 and p72 in the in vitro kinase assay performed on virus isolated at the same time point (Fig. 4B, lane 4). The proportion of nef clones containing the prototype sequence (P¹⁰⁴KVP¹⁰⁷) increased from 50% at 42 weeks to 70% at 46 weeks (Table 2). Taken together, the pattern of reversions in Mmu 27659 not only reinforced the significance of in vivo selection of Nef with the ability to associate with NAK but also highlighted the importance of Nef residue P¹⁰⁴ in the SH3 ligand domain.

Analysis of virus recovered from the animal displaying very rapid disease progression, Mmu 26785, showed that the mutant nef gene did not revert from the mutant sequence either during the course of infection or at necropsy at 19 weeks postinoculation (Table 2). Thus, the lack of genotypic revertants was consistent with the NAK-negative phenotype of virus recovered from this animal at necropsy (data not shown).

Virus in Mmu 27626 did not exhibit reversion in the nef sequence (Table 2) or in the ability of Nef to associate with NAK as analyzed at 24 weeks (Fig. 4A, lane 8) and 42 weeks (Fig. 4B, lane 3) postinfection. At necropsy at 93 weeks, nef clones from Mmu 27626 exhibited the mutant sequence A¹⁰⁴KVA¹⁰⁷ as well as a variant sequence, A¹⁰⁴KVT¹⁰⁷ (Table 2). Nef produced by virus recovered from this animal at necropsy did not associate with NAK in the in vitro kinase assay (data not shown).

DISCUSSION

Structural features of the SH3 ligand domain in Nef.

An understanding of the SH3 ligand domain of Nef is built on our knowledge of the molecular basis of SH3-domain/SH3-ligand interactions (reviewed in reference 36) as well as of structural models of HIV-1 Nef (15, 22). From the analysis of cellular proteins that bind SH3 domains, the consensus sequence PxxPLR was identified. The prolines in this motif make direct contacts with aromatic residues, and the arginine makes ionic or salt bridges with acidic residues within SH3 domains (2). The salient features of the SH3 ligand motif found in Nef are as follows: (i) this motif is present in all isolates of HIV-1, HIV-2, and SIV; (ii) the PxVPLR sequence is conserved in HIV-1, HIV-2, and SIV (34, 46); and (iii) this motif exists in the “minus” orientation as determined by the location of the critical arginine residue on the C-terminal side of the prolines (25). The in vitro kinase analysis, presented in Fig. 1B and C, clearly demonstrated that the strictly conserved amino acid residues V¹⁰⁶, P¹⁰⁷, L¹⁰⁸, and R¹⁰⁹ of SIVmac239 Nef were critical for both association of Nef with NAK and Nef-mediated PAK activation. X-ray crystallographic studies of HIV-1 Nef complexed with the mutant SH3 domain of Fyn revealed that three amino acid residues in HIV-1 Nef, V⁷⁴, P⁷⁵, and R⁷⁷ (which correspond to V¹⁰⁶, P¹⁰⁷, and R¹⁰⁹, respectively, of SIV Nef), make key contacts with conserved residues within the SH3 domain (15, 22).

The specificity and affinity of binding of the SH3 domain of Hck to HIV-1 Nef depend on a tertiary interaction between a structure in Hck known as the RT loop, and a hydrophobic region in Nef that is relatively distant from the PxxP motif, involving the F⁹⁰ residue of HIV-1 Nef (21, 22). The RT loop is poorly conserved among different SH3 domains and therefore is likely to be involved in imparting specificity in other SH3-domain/SH3-ligand interactions as well. F⁹⁰ is strictly conserved in Nef of HIV-1, HIV-2, and SIV (34, 46). Accordingly, we constructed and tested a viral clone with a mutation in the corresponding residue, F¹²², in SIV Nef. This mutant was defective both in association of Nef with NAK (Fig. 1B) and in activation of PAK by Nef (Fig. 1C). Thus, the requirement for F¹²² for these in vitro SIV Nef functions supports a role for this amino acid residue in binding specificity as well.

Taken together, our findings, which demonstrate the importance of the residues V¹⁰⁶, P¹⁰⁷, L¹⁰⁸, R¹⁰⁹, and F¹²² of SIV Nef in the in vitro kinase assays, indicated that the SH3 ligand domain of Nef was similar to the SH3 ligand domains in cellular proteins involved in cell activation. Both prolines in the SIV Nef PxxP motif were not necessary, since the Nef proteins with the sequences A¹⁰⁴KVP¹⁰⁷ and P¹⁰⁴RVA¹⁰⁷ interact with NAK and activate PAK (Fig. 1B and 4C). Interestingly, a recent structural study of Src showed that an internal SH3 ligand domain in this cellular tyrosine kinase contains a single proline (52). A speculation is that the SH3 ligand domain of Nef may mediate the interaction with a unique member of the PAK family that contains an SH3 domain or, alternatively, it could function through an adapter molecule possessing an SH3 domain. Interestingly, the adapter molecule Nck, which contains three SH3 domains and one SH2 domain, has been shown to associate with PAK (13, 26). Further investigations are required to determine whether Nef interacts with NAK via Nck or via some other adapter molecule containing an SH3 domain. Additionally, a combination of structural and functional studies are needed to determine whether the NAK-negative phenotype of Nef mutants is due to the inability of mutant protein (i) to bind NAK or (ii) to bind but not activate NAK.

Nef reversion and disease with the SIVmac239-P¹⁰⁴A/P¹⁰⁷A mutant.

To test the importance of the SH3 ligand domain of Nef for the virus-host relationship, seven juvenile rhesus macaques were inoculated with the nef mutant virus SIVmac239- P¹⁰⁴A/P¹⁰⁷A and tested for virological and clinical parameters as well as for changes (i.e., reversions) in the nef mutations. Five of seven macaques showed reversion in the PxxP motif and restoration of the NAK-positive phenotype (summarized in Table 2); this pattern of reversion supports the importance of the SH3 ligand domain in Nef for SIV infection in vivo. Two macaques, Mmu 25905 and Mmu 27659, developed hematological abnormalities during the course of infection and progressed to fatal SAIDS (Table 1). All nef clones from Mmu 25905 contained A¹⁰⁴KVP¹⁰⁷ (Table 2), which is encoded by a genotype that is positive in the NAK assay (Fig. 4A). In Mmu 27659, the pattern of reversion demonstrated greater complexity during the course of infection. At 12 weeks, Nef from this animal contained the sequence P¹⁰⁴R¹⁰⁵V¹⁰⁶A¹⁰⁷ (Table 2); the second-site mutation, converting K¹⁰⁵ to R¹⁰⁵ and accompanied by reversion of A¹⁰⁴ to P¹⁰⁴, partially restored the Nef- NAK interaction (Fig. 4A). Interestingly, the viral load in Mmu 27659, which had shown a decrease after the acute phase of infection, began to increase at 12 weeks postinoculation (Fig. 2A). At subsequent time points, nef clones from this animal exhibited the prototype pattern P¹⁰⁴KVP¹⁰⁷ (Table 2), which fully restored Nef-NAK interaction (Fig. 4C). Importantly, the proportion of clones with the prototype Nef sequence progressively increased from 10% at 20 weeks to 70% at 46 weeks. This genetic evidence indicates that not only was there a selection for Nef with full ability to interact with NAK but also both prolines were important for disease progression in this animal.

The outcome of infection of Mmu 26785 with the SIVmac239- P¹⁰⁴A/P¹⁰⁷A mutant was strikingly different from that of the five animals showing reversions in nef. Mmu 26785 displayed very high virus loads, no detectable anti-SIV antibodies, no detectable reversions in Nef, and a rapid disease course (Table 1). Other investigators have also described a pattern of rapid disease progression with no detectable seroconversion in adult rhesus macaques infected with pathogenic strains of SIV (17, 53). It is possible that either a host genetic factor (e.g., MHC-1 genotype) or a cofactor (e.g., an unidentified infectious agent) produced a situation in Mmu 26785 in which the host immune response to SIV was compromised and a high virus load and rapid disease progression ensued. Unlike all the other macaques in our study, eosinophil levels in this animal were elevated at the time of virus inoculation and became very high during the course of infection (see Results); eosinophilia has been associated with parasite infections, allergy, and increased morbidity in HIV-1-infected individuals (7). Thus, it is possible that Mmu 26785 harbored an undetected parasite cofactor which could elicit disease by a virus with a mutation in nef. Interestingly, in another study, two of two adult macaques infected with the SIVmac239-P¹⁰⁴A/P¹⁰⁷A mutant clone also exhibited a pattern of rapid disease progression; these animals died 9 and 18 weeks after infection and showed weak antibody responses (20). It is difficult to draw reliable conclusions about the in vivo importance of viral gene functions based on studies with very few animals (i.e., only Mmu 26785 in our study and only two animals in the previous study [20]). Additionally, rapid disease progression, without seroconversion and without a chronic phase, in infected macaques is not a model for AIDS caused by HIV infection in humans (3, 40).

Another outcome of infection with the SIVmac239- P¹⁰⁴A/P¹⁰⁷A mutant was exemplified by Mmu 27626. This animal was sacrificed at 93 weeks because of severe hepatitis. Necropsy also revealed evidence of lymphoid hyperplasia; however, no lymphoid-cell depletion was noted in peripheral, inguinal, and mesenteric lymph nodes or in splenic or gut-associated lymphoid tissue (Table 1). Additionally, this animal did not show severe depletion of CD4 T cells in peripheral blood (Fig. 3A), did not exhibit weight loss, and did not present evidence of an opportunistic infection. The virus load in Mmu 27626 was relatively low throughout the course of infection. The virus recovered at necropsy was negative in the NAK assay (data not shown); the biological significance of the A¹⁰⁴KVT¹⁰⁷ pattern, in several nef clones obtained at necropsy, remains to be determined. Taken together, clinical and virological findings in Mmu 27626 do not support a diagnosis of typical SAIDS.

Level of revertant nef+ virus required to cause disease.

Lang et al. showed nef gene revertants in two adult rhesus macaques infected with the SIVmac239-P¹⁰⁴A/P¹⁰⁷A mutant (20). The total virus load in both animals was very high, and the level of genotypic revertants, restoring P¹⁰⁴ and P¹⁰⁷, was 5 to 10% of the level of total virus in each animal. Furthermore, this previous study measured reversion only at P¹⁰⁴ and P¹⁰⁷. Our study demonstrated that phenotypic reversion can also occur when a second-site suppressor mutation (underlined) is generated, i.e., P¹⁰⁴R¹⁰⁵VA¹⁰⁷ (Fig. 1B and 4C). Importantly, the ability of Nef to interact with NAK was not analyzed in virus recovered from the mutant-infected animals in the study by Lang et al. (20).

To test the ability of the NAK assay to detect a small proportion of NAK-positive virus in a mixture containing mutant (i.e., NAK-negative) virus, we performed a reconstruction experiment. Kinase assays were performed on extracts from mixtures of cells containing increasing proportions of the prototype SIVmac239nef+ cell line relative to the mutant SIVmac239-P¹⁰⁴A/P¹⁰⁷A cell line. In this experiment, Nef-NAK association was detected down to the 2% level of prototype to mutant virus-infected cell mixture (data not shown). Because the virus loads in the study by Lang et al. were very high, it is possible that the small proportion of revertants in their animals was sufficient to cause disease (20). Thus, a reinterpretation of the data in the previous paper is that the PxxP motif in Nef is important for SIV pathogenesis.

The dose of the virus used to establish infection can be another important variable for understanding the virus-host relationship and disease progression (39, 51). Our study used an inoculum of 1,000 TCID₅₀ of SIVmac239-P¹⁰⁴A/P¹⁰⁷A, whereas the study by Lang et al. used 10,000 TCID₅₀ of this mutant virus (39, 51). It is possible that the rapid disease course was influenced by the infecting dose of mutant virus as well as by host factors. Although the dose range required for transmission under natural conditions is not known for primate lentiviruses, the lower virus inoculum in the macaque model in our study is probably more physiologically relevant for elucidating the importance of viral gene functions in vivo. Nonetheless, the relationship of viral dose to outcome of infection (i.e., progression to SAIDS) remains to be determined for juvenile and adult rhesus macaques infected with either prototype or mutant SIV clones.

Reversion frequency and implications for Nef structure and function.

The time course of the appearance of viruses with reversions in the nef gene appears to be different in this study of the SIVmac239-P¹⁰⁴A/P¹⁰⁷A mutant and our previous study, which analyzed a viral clone (SIVmac239-RR/LL) with point mutations in two arginine residues (R¹³⁷R¹³⁸) in the highly conserved central domain of Nef (45). In two animals infected with SIVmac239RR/LL, the majority of nef clones had reverted to produce a functional Nef at 4 weeks after infection. In the present study, we detected revertant nef clones at 8 weeks in one animal (Mmu 26902) and at 12 weeks in three others (Mmu 27659, Mmu 27879, and Mmu 26902). It is possible that the nef mutations in the diarginine motif R¹³⁷R¹³⁸ cause a more pronounced effect on Nef structure and function than do the mutations in the PxxP motif; thus, selective pressures for restoring Nef function might be greater for SIVmac239-RR/LL. A clearer understanding of the roles of the domains of Nef in various functions ascribed to this viral protein (i.e., downregulation of CD4 and MHC class I antigens, augmentation of virion infectivity and viral replication, association with cellular kinases) is required to determine whether there is a hierarchy of Nef functions and whether each function is subject to different selection pressures.

Conclusions.

The analysis of SIV Nef mutants in the in vitro kinase assays, measuring Nef-NAK association and PAK activation, demonstrated the importance of V¹⁰⁶P¹⁰⁷L¹⁰⁸R¹⁰⁹ and F¹²² in the PxxP region of Nef; this sequence conforms to the consensus sequence requirements of SH3 ligand domains with “minus” orientation (2, 11). Additionally, our in vivo study highlighted the importance of the SH3 ligand domain in Nef by demonstrating reversion of the PxxP Nef mutant virus (i.e., SIVmac239-P¹⁰⁴A/P¹⁰⁷A) and disease progression in juvenile rhesus macaques. The animals in our study exhibited diverse outcomes with respect to pathogenesis and reversion of the mutations in the Nef SH3 ligand domain. Such different patterns in the virus-host relationship could occur because rhesus macaques are outbred animals and therefore differ in genes controlling antiviral immune responses (e.g., MHC class I genes) (18, 54). It is also possible that some captive macaques in different primate facilities contain an unidentified infectious agent which serves as a cofactor for pathogenesis in animals infected with either prototype SIV or certain SIV mutants. Accordingly, for pathogenesis studies to examine viral gene functions in outbred animals, it is essential to use a large number of animals to obtain an accurate assessment of viral gene function in vivo.

Sequence homology information also strongly supports the idea that the SH3 ligand domain of Nef plays a significant role in the virus-host relationship. HIV-1, HIV-2, and SIV Nef contain the SH3 ligand domain with several strictly conserved residues in addition to the two prolines; these residues are P¹⁰⁴xV¹⁰⁶P¹⁰⁷L¹⁰⁸R¹⁰⁹ and F¹²² in SIV Nef (34, 46). Such a high level of conservation clearly indicates an important functional role in vivo for this highly conserved feature of Nef proteins. Also, the structural and biochemical studies, which demonstrated the interaction of the SH3 ligand domain of Nef with the SH3 domain of certain tyrosine kinases, strongly support a role for the PxxP motif in Nef for cell activation (15, 22, 33, 42); several lines of evidence show that Nef influences one or more components of cell signaling (reviewed in reference 41). In vitro studies support a model for Nef in which this viral protein is multifunctional; Nef modulates cell activation, enhances virion infectivity and viral replication, and downregulates CD4 cell surface antigen and MHC class I antigens (8, 41, 48). An effect of Nef on a cell signaling molecule(s), via an interaction through the SH3 ligand domain, could influence one or more of these functions attributed to Nef. Although our data show a linkage between Nef-NAK interaction and SAIDS, these findings do not exclude the possibility that Nef interaction with other cellular proteins also contributes to disease progression. Further analyses are required to fully define cellular components interacting with Nef and to assess the relative importance of each of these functions for viral pathogenesis.

ADDENDUM IN PROOF

Two juvenile rhesus macaques (Mmu 28870 and Mmu 28790) were inoculated intravenously with 1,000 TCID₅₀ of virus recovered from Mmu 27626 at necropsy at 93 weeks postinoculation. After 8 weeks of infeciton, Mmu 28870 exhibited a high virus load that was comparable to that of prototype virus SIVmac239nef+. Analysis for NAK revealed that Nef reverted to a kinase-positive phenotype in the virus from Mmu 28870. In the other macaque infected with necropsy virus from Mmu 27626, virus load has remained low and virus from this animal has not reverted to a kinase-positive phenotype. Both animals are being monitored for clinical signs.