No abstract

Incorporation of the envelope (Env) glycoprotein into budding virions is a key step in the replication cycle of lentiviruses. Previously, we provided genetic and biochemical evidence indicating that Env packaging into simian immunodeficiency virus (SIV) particles is mediated by the association of the Env cytoplasmic domain (CD) with the matrix (MA) domain of Gag. In this study, we developed an in vitro binding assay that, based on recombinant proteins expressed in bacteria, allowed us to demonstrate the physical interaction between the SIV Env CD and the MA in the absence of other viral or cellular proteins. We show that this association is blocked by mutations in each of the interacting domains that have been reported to interfere in vivo with the incorporation of Env into SIV virions. Moreover, we determined that the binding of SIV MA to the Env CD is saturable with a dissociation constant of 7x10(-7) M. Interestingly, the SIV MA is capable of specifically interacting in vitro with the human immunodeficiency virus type 1 Env CD, but not with that of the distantly related feline immunodeficiency virus. Our results strongly support the notion that the association between the SIV MA and Env CD plays a central role in the process of SIV Env incorporation into Gag-made particles.

Simian immunodeficiency viruses (SIVs) have an envelope (Env) glycoprotein with an unusually long cytoplasmic domain of 164 amino acids. In this article, we have characterized a series of SIV Env truncation mutants in which the cytoplasmic domain was progressively shortened from its carboxyl terminus by 20 amino acids. Expression by means of the vaccinia virus system showed that all of the SIV Env mutants were expressed and processed into the surface and transmembrane (TM) subunits. When the ability of the Env mutants to associate with SIV Gag particles was examined, we found that deletion of 20 to 80 residues from the carboxyl terminus of the SIV TM cytoplasmic tail abrogated the incorporation of the Env glycoprotein into particles. By contrast, further truncation of the SIV TM protein by 100 to 140 amino acids restored the ability of the Env protein to associate with Gag particles. Interestingly, mutants bearing a 44- or 24-amino acid cytoplasmic domain were incorporated at levels significantly higher than those of the wild-type Env. Single-cycle infectivity assays showed that Env mutants bearing cytoplasmic tails of 144 to 64 amino acids were highly inefficient at mediating virus entry. By contrast, truncation of the cytoplasmic domain to 44 or 24 amino acids drastically enhanced virus infectivity with respect to that conferred by the full-length Env protein. Our results demonstrate that small variations in the length of the SIV Env cytoplasmic domain dramatically influence Env-mediated viral functions.

Recombinant strains of replication-competent rhesus monkey rhadinovirus (RRV) were constructed in which strong promoter/enhancer elements were used to drive expression of simian immunodeficiency virus (SIV) Env or Gag or a Rev-Tat-Nef fusion protein. Cultured rhesus monkey fibroblasts infected with each recombinant strain were shown to express the expected protein. Three RRV-negative and two RRV-positive rhesus monkeys were inoculated intravenously with a mixture of these three recombinant RRVs. Expression of SIV Gag was readily detected in lymph node biopsy specimens taken at 3 weeks postimmunization. Impressive anti-SIV cellular immune responses were elicited on the basis of major histocompatibility complex (MHC) tetramer staining and gamma interferon enzyme-linked immunospot (ELISPOT) assays. Responses were much greater in magnitude in the monkeys that were initially RRV negative but were still readily detected in the two monkeys that were naturally infected with RRV at the time of immunization. By 3 weeks postimmunization, responses measured by MHC tetramer staining in the two Mamu-A*01 ؉ RRV-negative monkeys reached 9.3% and 13.1% of all CD8 ؉ T cells in peripheral blood to the Gag CM9 epitope and 2.3% and 7.3% of all CD8 ؉ T cells in peripheral blood to the Tat SL8 epitope. Virus-specific CD8 ؉ T cell responses persisted at high levels up to the time of challenge at 18 weeks postimmunization, and responding cells maintained an effector memory phenotype. Despite the ability of the RRVenv recombinant to express high levels of Env in cultured cells, and despite the appearance of strong anti-RRV antibody responses in immunized monkeys, anti-Env antibody responses were below our ability to detect them. Immunized monkeys, together with three unimmunized controls, were challenged intravenously with 10 monkey infectious doses of SIVmac239. All five immunized monkeys and all three controls became infected with SIV, but peak viral loads were 1.2 to 3.0 log 10 units lower and chronic-phase viral loads were 1.0 to 3.0 log 10 units lower in immunized animals than the geometric mean of unimmunized controls. These differences were statistically significant. Anti-Env antibody responses following challenge indicated an anamnestic response in the vaccinated monkeys. These findings further demonstrate the potential of recombinant herpesviruses as preventive vaccines for AIDS. We hypothesize that this live, replication-competent, persistent herpesvirus vector could match, or come close to matching, live attenuated strains of SIV in the degree of protection if the difficulty with elicitation of anti-Env antibody responses can be overcome.