Dimerization and template switching in the 5' untranslated region between various subtypes of human immunodeficiency virus type 1

Abstract

The human immunodeficiency virus type 1 (HIV-1) particle contains two identical RNA strands, each corresponding to the entire genome. The 5' untranslated region (UTR) of each RNA strand contains extensive secondary and tertiary structures that are instrumental in different steps of the viral replication cycle. We have characterized the 5' UTRs of nine different HIV-1 isolates representing subtypes A through G and, by comparing their homodimerization and heterodimerization potentials, found that complementarity between the palindromic sequences in the dimerization initiation site (DIS) hairpins is necessary and sufficient for in vitro dimerization of two subtype RNAs. The 5' UTR sequences were used to design donor and acceptor templates for a coupled in vitro dimerization-reverse transcription assay. We showed that template switching during reverse transcription is increased with a matching DIS palindrome and further stimulated proportional to the level of homology between the templates. The presence of the HIV-1 nucleocapsid protein NCp7 increased the template-switching efficiency for matching DIS palindromes twofold, whereas the recombination efficiency was increased sevenfold with a nonmatching palindrome. Since NCp7 did not effect the dimerization of nonmatching palindromes, we concluded that the protein most likely stimulates the strand transfer reaction. An analysis of the distribution of template-switching events revealed that it occurs throughout the 5' UTR. Together, these results demonstrate that the template switching of HIV-1 reverse transcriptase occurs frequently in vitro and that this process is facilitated mainly by template proximity and the level of homology.

FIG. 1.

(A) Alignment of the cloned leader sequences of eight circulating HIV-1 isolates obtained from eight infected individuals. The sequence of subtype B-HXB2 is included as a reference sequence. The subtype assignment is shown to the left (see Material and Methods), and the positions of the functional RNA elements are denoted as follows: TAR; PolyA, hairpin containing polyA site used in the 3′ LTR; PBS; DIS; SD, major SD; PSI. Conservation of alignment positions is indicated by box shading: black; 100%; gray, 80%; light gray, 60%. (B) Unrooted phylogenetic tree based on the alignment of the subtype sequences shown in panel A. The branch lengths were calculated by using the PHYLIP program (J. Felsenstein, 1993. PHYLIP [Phylogeny Inference Package] version 3.5c. Distributed by the author. Department of Genetics, University of Washington, Seattle).

FIG. 2.

Effect of Mg²⁺ ions on the stability of subtype dimers. (A) Dimerization of all subtype leader transcripts analyzed on native polyacrylamide gels containing 5 mM MgCl₂ (TBM₅) or 1 mM EDTA (TBE₁). The positions of the monomer and dimer species are indicated to the left of the gels by the letters M and D, respectively. (B) Quantification of dimerization efficiency (D = [number of dimers]/[number of dimers + number of monomers]) for all subtypes by a similar gel retardation assay of subtype homodimers as shown in panel A in the presence of decreasing amounts of MgCl₂ (5, 1, and 0.1 mM) and in the presence of EDTA.

FIG. 2.

Effect of Mg²⁺ ions on the stability of subtype dimers. (A) Dimerization of all subtype leader transcripts analyzed on native polyacrylamide gels containing 5 mM MgCl₂ (TBM₅) or 1 mM EDTA (TBE₁). The positions of the monomer and dimer species are indicated to the left of the gels by the letters M and D, respectively. (B) Quantification of dimerization efficiency (D = [number of dimers]/[number of dimers + number of monomers]) for all subtypes by a similar gel retardation assay of subtype homodimers as shown in panel A in the presence of decreasing amounts of MgCl₂ (5, 1, and 0.1 mM) and in the presence of EDTA.

FIG. 3.

Heterodimerization between subtypes is dependent exclusively on the DIS palindrome. (A) Heterodimerization between labeled transcripts of subtypes AC, B, or G (Donor; nucleotides 183 to 404) and longer unlabeled leader transcripts of subtypes B (Acceptor; nucleotides 1 to 466) or C2 (Acceptor; nucleotides 1 to 543) at a molar ratio of 1:1. (B) Heterodimerization between labeled transcripts of subtype AC (nucleotides 183 to 381) or subtypes B, C2, and D (Donor; nucleotides 183 to 404) with a 20-nucleotide 5′-tag sequence and longer unlabeled transcripts of subtypes AC, B, C1, C2, D, AE, F1, G, AG, F1′, and GC1 (Acceptor; nucleotides 1 to 355) at a molar ratio of 1:5. The percentage of heterodimers compared to that of homodimers was calculated as follows: H = 2 × h/(d + 2 × h) × 100%, where d and h are the intensities of the dimer and heterodimer, respectively. The DIS palindrome classes corresponding to donors AC, B, C2, and D are indicated in brackets. (C) Heterodimerization in the presence and absence of NCp7 by using a selected set of the same RNAs as shown in panel B. NCp7 was removed by phenol extraction prior to loading. The multimers that were formed only when using RNAs with matching DIS palindromes and in the presence of NCp7 are marked. The positions of the monomer, dimer, and heterodimer species are indicated with M, D, and H, respectively.

FIG. 3.

Heterodimerization between subtypes is dependent exclusively on the DIS palindrome. (A) Heterodimerization between labeled transcripts of subtypes AC, B, or G (Donor; nucleotides 183 to 404) and longer unlabeled leader transcripts of subtypes B (Acceptor; nucleotides 1 to 466) or C2 (Acceptor; nucleotides 1 to 543) at a molar ratio of 1:1. (B) Heterodimerization between labeled transcripts of subtype AC (nucleotides 183 to 381) or subtypes B, C2, and D (Donor; nucleotides 183 to 404) with a 20-nucleotide 5′-tag sequence and longer unlabeled transcripts of subtypes AC, B, C1, C2, D, AE, F1, G, AG, F1′, and GC1 (Acceptor; nucleotides 1 to 355) at a molar ratio of 1:5. The percentage of heterodimers compared to that of homodimers was calculated as follows: H = 2 × h/(d + 2 × h) × 100%, where d and h are the intensities of the dimer and heterodimer, respectively. The DIS palindrome classes corresponding to donors AC, B, C2, and D are indicated in brackets. (C) Heterodimerization in the presence and absence of NCp7 by using a selected set of the same RNAs as shown in panel B. NCp7 was removed by phenol extraction prior to loading. The multimers that were formed only when using RNAs with matching DIS palindromes and in the presence of NCp7 are marked. The positions of the monomer, dimer, and heterodimer species are indicated with M, D, and H, respectively.

FIG. 3.

Heterodimerization between subtypes is dependent exclusively on the DIS palindrome. (A) Heterodimerization between labeled transcripts of subtypes AC, B, or G (Donor; nucleotides 183 to 404) and longer unlabeled leader transcripts of subtypes B (Acceptor; nucleotides 1 to 466) or C2 (Acceptor; nucleotides 1 to 543) at a molar ratio of 1:1. (B) Heterodimerization between labeled transcripts of subtype AC (nucleotides 183 to 381) or subtypes B, C2, and D (Donor; nucleotides 183 to 404) with a 20-nucleotide 5′-tag sequence and longer unlabeled transcripts of subtypes AC, B, C1, C2, D, AE, F1, G, AG, F1′, and GC1 (Acceptor; nucleotides 1 to 355) at a molar ratio of 1:5. The percentage of heterodimers compared to that of homodimers was calculated as follows: H = 2 × h/(d + 2 × h) × 100%, where d and h are the intensities of the dimer and heterodimer, respectively. The DIS palindrome classes corresponding to donors AC, B, C2, and D are indicated in brackets. (C) Heterodimerization in the presence and absence of NCp7 by using a selected set of the same RNAs as shown in panel B. NCp7 was removed by phenol extraction prior to loading. The multimers that were formed only when using RNAs with matching DIS palindromes and in the presence of NCp7 are marked. The positions of the monomer, dimer, and heterodimer species are indicated with M, D, and H, respectively.

FIG. 4.

Template-switching efficiency. (A) Experimental setup showing the donor and acceptor transcripts drawn to scale with unique tags at their 5′ end, the region of homology (shaded box), and the RT primer. (B) Template-switching efficiency T for all combinations of AC, B, C2, and D (Donor; nucleotides 183 to 444) and AC, B, C1, C2, D, AE, F1, G, AG, F1′, and GC1 (Acceptor; nucleotides 1 to 355) at a molar ratio of 1:5. Only elongated products are shown from the primer extension gels. F indicates the full-length donor product; S indicates primers that have switched to the acceptor template. The template-switching efficiency (T) was calculated as s/(f + s) × 100%, where s is the intensity of the template-switching product in the primer extension reaction and f is the intensity of the band corresponding to the full-length product of the donor. (C) Template-switching efficiencies for selected combinations of the donor to the acceptor under different conditions. The donor and acceptor templates were incubated under dimerization conditions for 5 and 40 min as indicated or treated with NCp7 in NC buffer for 15 min prior to reverse transcription (NCp7).

FIG. 5.

Distribution of template-switching events for different subtype combinations and conditions. (A) Illustration of the experimental setup with indication of regions of identity (shaded boxes) and phylogenetic markers (black lines) between subtypes B and D. The transcripts, primers, regions of identity, and markers are drawn to scale. (B) Distribution of template-switching events. Prior to RT transcription, the RNA substrates were dimerized for either 5 or 40 min or incubated for 15 min in the presence of NCp7. Analysis of recombinant clones was performed, and the identified template-switching events are summarized in Table 1. The y axis indicates the frequency of template switching normalized to the number of nucleotides within each region of identity. Standard error bars are estimated as Δp_i = p_i[(n_i)^−1/2 + (N)^−1/2], where N is the total number of recombinants cloned and sequenced in each independent experiment and n_i is the number of recombinants occurring in the interval i. All regions of identity less than 5 nucleotides long are omitted from the figure but are shown in Table 1. Three independent experiments were performed for the donor D-to-acceptor B combination (5 and 40 min) with similar distribution. All three experiments were included in the data analysis. Only single template-switching events were included. n.d., not determined.