Suboptimal Nucleotides in the Infectious, Pathogenic Simian Immunodeficiency Virus Clone SIVmac239

Abstract

We analyzed virus sequences in two monkeys infected with SIVmac239 and two monkeys infected with SHIVnef that maintained high, persisting viral loads. Sequence changes were observed consistently at four loci in all four animals: a single nucleotide change in the Lys-tRNA primer binding site in the 5′ long terminal repeat; two nucleotide changes that resulted in two amino acid changes in the pol gene product; and a single nucleotide change in the region of the simian immunodeficiency virus genome where the rev and env genes overlap, resulting in changes in the predicted amino acid sequences of both gene products. None of these mutations were seen in short-term cultures of CEM×174 cells infected with SIVmac239 or SHIVnef. At all four positions in all four animals, the new sequences represented consensus sequences for primate lentiviruses, whereas the inoculum sequences at these four loci have either never been or rarely been reported outside of SIVmac239. Thus, although cloned SIVmac239 is consistently pathogenic and consistently induces high viral load set points, it is clearly less than optimal at these four nucleotide positions.

SIV and SHIVnef infections.

Experimental infection of rhesus monkeys with cloned SIVmac239 has been useful for studying the pathogenesis of AIDS. The utility of this system stems in large part from the consistent behavior of this strain in experimentally infected rhesus monkeys. More than 20 monkeys have been infected experimentally with SIVmac239 at the New England Regional Primate Research Center. SIVmac239-infected rhesus monkeys have displayed consistent viral loads both at peak height ([20 ± 15] × 10⁶ RNA copies per ml of plasma [n = 10]) and at set point ([5.5 ± 5.0] × 10⁶ RNA copies per ml of plasma [n = 10]) and have progressed to AIDS in a time frame that is suitable for laboratory investigation (8). The defined sequence of the molecularly cloned virus (17) has provided numerous opportunities for detailed studies that would be difficult or impossible otherwise. We have taken advantage of this system to engineer a SHIVnef recombinant by exchanging human immunodeficiency virus type 1 (HIV-1) nef sequences for simian immunodeficiency virus (SIV) nef sequences in the cloned SIVmac239 backbone (1). We have demonstrated previously that this recombinant is pathogenic in a majority of experimentally infected rhesus macaques. Here, we report on two juvenile rhesus macaques that were inoculated intravenously with SHIVnef and two that were inoculated intravenously with SIVmac239. The stocks used for monkey inoculation were generated by DEAE dextran-mediated transfection of the cell line CEM×174 (16) and were harvested at or near the peak of virus production (day 11 for SHIVnef and day 12 for SIVmac239). Whole blood samples were obtained from these animals at regular intervals, and plasma was purified. Virion-associated SIV RNA in plasma was then quantitated as previously described (19). All four animals displayed high, persisting SIV RNA concentrations throughout the course of infection (Fig. 1). Three of the four monkeys have progressed to AIDS and death, 40 to 96 weeks postinoculation.

FIG. 1.

SIV RNA levels, per milliliter of plasma, at the indicated time postinoculation for monkeys infected with SHIVnef (SHIV) or SIVmac239 (239). The dashed line indicates the threshold sensitivity of the assay, 300 copy eq/ml.

Determination of SIV sequences isolated from infected macaques.

We examined sequence changes from the parental virus in four infected monkeys. In order to acquire SIV template DNA for PCR, we purified cellular DNA from 5 × 10⁶ peripheral blood mononuclear cells by a previously described saturated-NaCl precipitation technique (2). One microgram of the purified cellular DNA was used as a template for PCR amplification of SIV sequences in three overlapping fragments of approximately 3.5 kbp with SIV-specific primers. The resulting fragments derived from PCR were column purified (Qiagen, Santa Clarita, Calif.), and the entire SIV sequence was then determined using universal and reverse primers, as well as SIV-specific primers, with an ABI 377 DNA sequencer (Perkin-Elmer Cetus).

Four changes were consistently observed in the SIV sequences from all four animals (Table 1): a single nucleotide change in the Lys-tRNA primer binding site in the 5′ long terminal repeat (LTR) (Table 1); two nucleotide changes that resulted in two amino acid changes in the pol gene product, one of which was in reverse transcriptase and one of which was in integrase (Table 1); and a single nucleotide change in the region of the SIV genome where the rev and env genes overlap, resulting in a change in the amino acid sequences of both gene products (Table 1). The tat gene also overlaps the rev and env genes in this region, but the nucleotide change in this region did not affect its amino acid sequence (Table 1). Previous analyses of virus stocks at the Lys-tRNA primer binding site, integrase, and rev/env loci revealed no evidence of these reversions (5, 10; L. Denekamp, R. C. Desrosiers, and L. Alexander, unpublished data).

TABLE 1.

Changes in cloned SIVmac239 sequences of rhesus macaques

Gene or genetic elementa	Nucleotide in:		Amino acid inb:
	SIVmac239	Monkey isolate	SIVmac239	Monkey isolate
5′ LTR (PBS)	T	C	NA	NA
pol (RT)	C	T	Ser	Leu
pol (Int)	C	T	Ala	Val
rev	A	G	Lys	Arg
tat	A	G	Lys	Lys
env	A	G	Arg	Gly

^aPBS, Lys-tRNA primer binding site; RT, reverse transcriptase; Int, integrase. 

^bNA, not applicable (i.e., these LTR sequences do not encode amino acids). 

Alignment of primate lentivirus sequences.

We aligned SIVmac239 to lentivirus sequences contained in the Los Alamos Sequence Database (B. Korber, HIV-1 sequence database posting, 1999). At the four loci that were consistently altered in monkeys infected with this cloned virus, the sequences in HIV-1, HIV-2, and SIV are all highly conserved (Fig. 2). However, the cloned SIVmac239 sequences are unique at all four of these loci (Fig. 2). The changes that were detected at these loci in monkey-passaged virus resulted in consensus sequences being introduced into the SIV genome (Fig. 2).

FIG. 2.

Alignment of primate lentivirus sequences. SIVmac239 sequences are aligned with previously observed SIV, HIV-1, and HIV-2 sequences (B. Korber, HIV-1 sequence database posting, 1999). SIVmac239 sequences are shown on the top line of each panel and are used as the basis of comparison with the other sequences. The numbers above the top lines indicate the amino acid or nucleotide position of the depicted gene or genetic element, respectively. A dot indicates homology with SIVmac239 at a particular locus; a dash indicates that a nucleotide or amino acid is not contained in a particular sequence. Positions at which cloned SIVmac239 sequences are suboptimal are depicted in white on black. (A) 5′ LTR sequences surrounding the primer binding site (shaded). (B and C) Amino acid sequences of the reverse transcriptase and integrase subunits of Pol. (D) Amino acid sequences of Rev. The nucleotide change that resulted in an amino acid change in Rev also resulted in an amino acid change in Env (Table 1). However, since primate lentivirus Env sequences are not conserved at this locus, an alignment of these sequences was not possible.

Implications of the observed sequence changes.

SIVmac239 has been a consistently pathogenic strain in rhesus monkeys and thus has been used extensively as a model for HIV-1 infection of humans (1, 3, 4, 6–9, 11, 12–15, 18, 20, 23). The median time frame for progression to and death from AIDS in macaques infected with SIVmac239 at the New England Regional Primate Research Center is approximately 13 months. Despite its uniform pathogenicity and uniformly high viral load set points (10⁵ to 10⁷ RNA copies per ml of plasma), the SIVmac239 clone is clearly less than optimal at the four positions described here. Our data also suggest that a SIVmac239 derivative strain that contains all four changes in sequence might exhibit enhanced viral replication in comparison to parental SIVmac239. This enhancement could lead to an accelerated onset of AIDS in experimental monkey infections and could potentially facilitate a more consistent outcome.

HIV-1 sequences from recently infected individuals are relatively homogeneous despite the fact that virus harbored by the donor is heterogeneous (21, 22, 24, 25). This observation has suggested that HIV-1 infection results from very few and perhaps even a single infectious particle. It seems reasonable that many, or even most, human infections with HIV-1 may initiate similarly, with suboptimal sequences at specific loci, as we describe here. In some cases, suboptimal sequence polymorphisms may be difficult to revert (2). Depending on how many suboptimal polymorphisms are transmitted, the ease of their reversion, and the severity of their attenuating effects, they might influence the ability of virus to replicate efficiently and the rate of disease progression. Our results suggest that the presence of as many as four clearly suboptimal nucleotides can still result in consistently high viral loads and disease progression.