Contribution of glutamine residues in the helix 4-5 loop to capsid-capsid interactions in simian immunodeficiency virus of macaques

Abstract

Following retrovirus entry, the viral capsid (CA) disassembles into its component capsid proteins. The rate of this uncoating process, which is regulated by CA-CA interactions and by the association of the capsid with host cell factors like cyclophilin A (CypA), can influence the efficiency of reverse transcription. Inspection of the CA sequences of lentiviruses reveals that several species of simian immunodeficiency viruses (SIVs) have lost the glycine-proline motif in the helix 4-5 loop important for CypA binding; instead, the helix 4-5 loop in these SIVs exhibits an increase in the number of glutamine residues. In this study, we investigated the role of these glutamine residues in SIVmac239 replication. Changes in these residues, particularly glutamine 89 and glutamine 92, resulted in a decreased efficiency of core condensation, decreased stability of the capsids in infected cells, and blocks to reverse transcription. In some cases, coexpression of two different CA mutants produced chimeric virions that exhibited higher infectivity than either parental mutant virus. For this complementation of infectivity, glutamine 89 was apparently required on one of the complementing pair of mutants and glutamine 92 on the other. Modeling suggests that glutamines 89 and 92 are located on the distal face of hexameric capsid spokes and thus are well positioned to contribute to interhexamer interactions. Requirements to evade host restriction factors like TRIMCyp may drive some SIV lineages to evolve means other than CypA binding to stabilize the capsid. One solution used by several SIV strains consists of glutamine-based bonding.

Importance: The retroviral capsid is an assembly of individual capsid proteins that surrounds the viral RNA. After a retrovirus enters a cell, the capsid must disassemble, or uncoat, at a proper rate. The interactions among capsid proteins contribute to this rate of uncoating. We found that some simian immunodeficiency viruses use arrays of glutamine residues, which can form hydrogen bonds efficiently, to keep their capsids stable. This strategy may allow these viruses to forego the use of capsid-stabilizing factors from the host cell, some of which have antiviral activity.

FIG 1

Alignment of primate immunodeficiency virus sequences from the region surrounding the helix 4-5 loop. Eighty complete CA sequences from the HIV Sequence Compendium were aligned using ClustalW and ordered using PHYLIP. Two groups of helix 4-5 loop sequences are apparent: group GP loops, containing a potential cyclophilin A binding motif GP (highlighted in orange), and group Q loops, which are rich in glutamine residues (highlighted in yellow). The former group contains HIV-1, HIV-2, SIVcpz, and some other SIV strains, the last consisting of SIVmac239 and several SIV lineages.

FIG 2

Expression and processing of SIVmac239 helix 4-5 loop CA mutants. (A) Site-directed mutagenesis was used to alter, singly or in combination, each of four glutamine residues in the SIVmac239 helix 4-5 loop to glycines (represented by asterisks). The glutamine residues (Q85, Q89, Q90, and Q92) are numbered using SIVmac239 residue positions, according to current convention (57). The CA mutants are shown in numerical order in the column on the left; in the column on the right, the mutants are organized according to the number of glutamine residues altered. (B) The processing of the Gag precursor polyprotein was examined by metabolically labeling transfected 293T cells producing wild-type (WT) and mutant SIVmac239 virions. Secreted viruses were collected, purified over a 20% sucrose column, and run on a 12% bis-Tris gel. Loading was normalized by the amount of ³⁵S label. The domain organization of the SIVmac239 p55 Gag precursor polyprotein is illustrated beneath the gel.

FIG 3

Infection efficiency of mutant viruses. (A and B) VSV-G pseudotypes of the indicated wild-type and mutant SIVmac239 viruses were manufactured in HEK293T cells, normalized for RT activity, and serially diluted before infection of Cf2Th cells (A) and HeLa cells (B). The GFP-positive cells from five independent infections were counted and normalized to the values observed for 50,000 RT units of the wild-type (WT) virus. The infection efficiency of each mutant virus was classified as near wild type (green), intermediate (orange), or severely defective (red). (C) SIVmac239 CA mutants grouped according to infection efficiency. (D) Infection efficiency inversely correlates with the number of altered glutamines. Mutants with two altered glutamines exhibit a wide range of phenotypes. The bar indicates the median for each category. Third-order polynomial regression (r²) was 0.989.

FIG 4

Capsid morphology of the SIVmac239 mutants. (A) Viruses lacking envelope glycoproteins were manufactured, purified, and stained for EM imaging, as detailed in Materials and Methods. The morphology of the virion cores was analyzed. If a virion was sectioned to reveal a conical core in a particle ∼100 nm in diameter, it was scored as conical. If the virion section revealed a well-defined core that was not conical, it was scored as round. If the sectioned virion contained no defined core, or a core of improper size, it was scored as “other.” Examples of conical, round, and other cores for WT and selected mutant virions are shown. (B) Distribution of core morphologies for WT and mutant virions. For WT virions, n = 166; between 162 and 297 virion particles were scored for each mutant. (C) Inverse correlation between the percentage of virion particles scored as other and the replication efficiency of the mutants in Cf2Th target cells (for an input of 25,000 RT units of virus). Spearman r_S = −0.6063; P = 0.01285 (two-tailed test).

FIG 5

Fate-of-capsid assay. (A) Cf2Th cells were infected by wild-type and mutant viruses, as described in Materials and Methods. At 8 and 16 h after infection, cells were disrupted and the amount of pelletable CA protein in cytosolic lysates was measured. Results from three independent infections were normalized to those obtained for the wild-type virus at 8 h after infection. Significant differences in the amount of pelletable CA protein between WT and mutant viruses are indicated by single asterisks for a P value of <0.05 and double asterisks for a P value of <0.01. (B) Correlation between the amount of pelletable CA protein at 8 h after infection and the infectivity of the viruses in Cf2Th cells. Spearman r_S = 0.8557; P = 0.000024 (two-tailed test).

FIG 6

Viral cDNA synthesis. (A) Three independent wells of HeLa cells were infected with wild-type and mutant GFP-expressing viruses at an MOI of 0.5. Total DNA was extracted from the cells at 0, 2, 6, 12, 24, and 48 h following infection. Cells were assessed by FACS for GFP expression at 48 h postinfection. (B) One hundred nanograms of total DNA was used in quantitative real-time PCR assays. The early primer set measures the amount of plus-strand strong-stop DNA. The symbols for the mutants are colored according to replication efficiency, as in Fig. 3C. The correlation between the level of early reverse transcripts at 10 h postinfection and virus infection efficiency is shown. Spearman r_S = 0.6343; P = 0.008313 (two-tailed test). (D) The late primer set measures the amount of completely reverse-transcribed genomes.

FIG 7

Production of 2-LTR circles of viral DNA. (A) HeLa cells infected with wild-type and mutant GFP-expressing viruses were used as a source of DNA for quantitative real-time PCR assays, as described in the legend to Fig. 6. The 2-LTR primer set measures the amount of viral 2-LTR junctions, autointegrants, and minigenomes present in the cell. (B) Correlation between the level of 2-LTR DNA products at 25 h after infection and viral infection efficiency. Spearman r_S = 0.6523; P = 0.00617 (two-tailed test). (C) The percentage of canonical 2-LTR junctions, compared with the standard, amplified at 12, 24, and 48 h after infection is shown. All viruses that completed reverse transcription produced canonical junctions.

FIG 8

Complementation analysis. (A) In a pilot experiment, chimeric viruses were produced by transfecting the indicated ratios of wild type to Mut 15 virus DNA into HEK293T cells. The VSV-G-pseudotyped hybrid viruses were used to infect Cf2Th cells. GFP was measured in the cells 48 h later. (B) Three independent 1:1 transfections of the indicated mutants or of mutant and wild-type integrase-deficient pSIVmac239ΔnefΔenvEGFP DNA into HEK293T cells were performed to make hybrid viruses, and 12,500 cpm (RT units) of each virus preparation was used to infect Cf2Th cells. The relative infection efficiency of the mutants is proportionate to the diameter of the inner circle. Standard deviations, where large enough to be visible, are represented as an exterior ring. (C) The infection efficiencies are depicted (as in panel B) for hybrid viruses composed of 1:1 ratios of CA mutants with intermediate or severely defective infection efficiencies.

FIG 9

Infectivity complementation over a range of mutant ratios. (A) Mut 10 DNA was cotransfected at the indicated ratios with DNA for Mut 5, 6, 9, 13, and 14 into HEK293T cells. Hybrid viruses were used to infect Cf2Th cells and 48 h later, GFP expression in the target cells was measured. (B) The complementation of Mut 14 and Mut 10, 5, 13, and 6 was examined, as described for panel A. (C) The complementation of Mut 5 and Mut 10, 14, and 9 was examined, as described for panel A. (D) Analysis of the complementing pairs suggests that having Q89 available on one of the capsids and Q92 on the other leads to complementation. Mutant pairs that rely upon a single mutant to provide both Q89 and Q92, or pairs that lack either one, do not complement.

FIG 10

Structural explanation for the SIVmac239 CA mutant phenotypes. The amino acid sequence of the SIVmac239 CA protein was threaded into the crystal structure of the HIV-2 capsid (2WLV) (34). This structure was then aligned with models of the HIV-1 hexamer (3H4E [A] and 3J34 [B] [7, 48]) using the Matchmaker tool in Chimera (47). SIVmac239 hexamers were assembled into a lattice by alignment with the HIV-1 capsid models, matching critical interactions between the CA C-terminal domains. In panel A, the interactions of the capsid hexamers occur in a plane orthogonal to the axis of 3-fold pseudosymmetry. In panel B, the capsid model exhibits curvature, as might be seen in the mature assembled capsid. Both models are shown from two perspectives to emphasize the differing hexamer relationships. Glutamine residues 89 (yellow) and 92 (orange) could contribute to interhexamer interactions around the 3-fold pseudosymmetry axis that stabilize the capsid lattice. Such interactions may be particularly important during initial nucleation of the capsid assembly in the virion, as in panel A.

FIG 11

Phylogenetic tree of lentivirus gag genes. The phylogenetic tree of gag gene sequences of the indicated lentiviruses, adapted from reference , is shown. Group GP viruses, which contain a glycine-proline pair in the CA helix 4-5 loop, are colored green. Group Q viruses, with a glutamine-rich helix 4-5 loop, are colored red. The SIVgsn helix 4-5 loop cannot be assigned to either group.