Longitudinal analysis of intra-host simian immunodeficiency virus recombination in varied tissues of the rhesus macaque model for neuroAIDS

Abstract

Human immunodeficiency virus intra-host recombination has never been studied in vivo both during early infection and throughout disease progression. The CD8-depleted rhesus macaque model of neuroAIDS was used to investigate the impact of recombination from early infection up to the onset of neuropathology in animals inoculated with a simian immunodeficiency virus (SIV) swarm. Several lymphoid and non-lymphoid tissues were collected longitudinally at 21 days post-infection (p.i.), 61 days p.i. and necropsy (75-118 days p.i.) from four macaques that developed SIV-encephalitis or meningitis, as well as from two animals euthanized at 21 days p.i. The number of recombinant sequences and breakpoints in different tissues and over time from each primate were compared. Breakpoint locations were mapped onto predicted RNA and protein secondary structures. Recombinants were found at each time point and in each primate as early as 21 days p.i. No association was found between recombination rates and specific tissue of origin. Several identical breakpoints were identified in sequences derived from different tissues in the same primate and among different primates. Breakpoints predominantly mapped to unpaired nucleotides or pseudoknots in RNA secondary structures, and proximal to glycosylation sites and cysteine residues in protein sequences, suggesting selective advantage in the emergence of specific recombinant sequences. Results indicate that recombinant sequences can become fixed very early after infection with a heterogeneous viral swarm. Features of RNA and protein secondary structure appear to play a role in driving the production of recombinants and their selection in the rapid disease model of neuroAIDS.

Fig. 1.

Putative SIV recombination breakpoints in infected macaques. The top axis represents the >1200 nucleotides incorporating the gp120 domain in SIV with black bars showing the approximate location of the beginning and end of conserved (C1, C2 and C3) and variable (V1, V2, V3 and V4) domains in our alignments. The three large blue boxes spanning all axes highlight the variable domains in the SIVmac251 viral swarm and subsequent macaques (D01–D06). Each plot uses coloured squares to display the estimated location of breakpoints found in sequence alignments using gard for each tissue collected at different time points. Each colour corresponds to a tissue: red, non-lymphoid; blue, lymphoid. A transparent black box incorporating several coloured squares indicates that multiple tissue alignments contained identical breakpoints.

Fig. 2.

Per cent recombination in SIV C1–C2 and V3–V4 domains. Primates D01–D06 are shown on the x-axis. Per cent recombination for C1–C2 and V3–V4 is shown on the y-axis.

Fig. 3.

Predominant secondary SIV RNA structures covering persistent recombinant breakpoints. Structures were generated using the BibiServ pknotsrg server. Persistent breakpoints identified in macaques are indicated by black arrows. Blue and red arrows indicate nucleotides that will encode cysteine residues and glycosylation motifs in proteins, respectively. (a) gp120 V2 substructure spanning SMM239 coordinates 7036–7158. (b) Two equally predominant gp120 C2 substructures spanning SMM239 coordinates 7234–7401; variable nucleotides between the two structures are circled in red. (c) gp120 V3 substructure spanning SMM239 coordinates 7563–7639. Paired nucleic acids are shown in blue and unpaired nucleic acids are shown in yellow. The MFE for each structure is given at the bottom.

Fig. 4.

V1 MFE RNA structures in a single SIV-infected animal. (a) Four sequences spanning the SIV V1 domain are shown with nucleotides that encode the putative pseudoknot highlighted in red. Beneath the nucleotide sequence is the structure mask for the RNA structures shown in (b). For the structure mask: ‘[…]’ denotes bonded RNA positions, ‘.’. denotes an unpaired position and ‘{…}’ denotes the locations of potential pseudoknots. (b) Structures are labelled as described in Fig. 3. Nucleotides involved in pseudoknot bonding are shown in bright yellow.

Fig. 5.

Evolutionary conserved breakpoints in SIV gp120. The secondary structure of the SIV gp120 protein was kindly provided by Brian T. Foley at the Los Alamos HIV Data Bank. A similar version was published previously (Hoxie, 1991). Putative glycosylation sites are indicated by black dots. The black line in the V1 domain represents the length-variable region removed prior to analysis. Red crosses indicate identical recombination sites that were found among different primates. Blue crosses represent identical recombination sites found with individual primates.

Fig. 6.

Model for pseudoknot-driven copy-choice recombination and generation of indels. Copy-choice recombination models state that recombination is due to reverse transcription slowing down at certain structural regions of RNA that can lead to strand switching. We propose that pseudoknots, along with other structural features, can also slow down reverse transcription. Additionally, pseudoknots may result in slipping during reverse transcription, thus leading to insertions or deletions, especially in the variable regions of gp120.