Using a computational approach recently we identified a sequence embedded in the U3-R region of HIV-1 RNA that is highly complementary to human tRNA3Lys
(Piekna-Przybylska et al., 2010
). By reconstitution in vitro
, we showed that segments of this complementary sequence called motif 9 and segment 1 in U3 promote minus strand transfer. These sequences were proposed to bind the priming tRNA so as to circularize the RNA genome to bring the two R regions into proximity for transfer. Our experiences with modeling strand transfer made it clear that many properties of the participating proteins and RNA template structures collaborate to produce the highest transfer efficiency. To simulate HIV-1 minus strand transfer, our systems in vitro
contained RT, NC, human tRNA3Lys
, and two RNA templates. The donor RNA template for transfer represented the 5’ end of the HIV-1 RNA genome. The acceptor template represented the 3’ end. NC protein and the RT RNase H activity have been shown to be critical for efficient strand transfer, and so were necessary components. However, RNA systems were modified to exclude RNA structural features that promote transfer, other than motif 9 and segment 1. This was done to improve the ability of our assays to measure the effects of those specific sequences. The 3’ end of the RNA template was shortened to eliminate the stimulating effect of the invasion-driven mechanism of transfer and allow only the terminal transfer (Chen et al., 2003b
; Song et al., 2009
). The RNA template representing the 5’ end of HIV-1 RNA included only the first 199nt, as sequences beyond the PBS in the 3’ direction also stimulated transfer (Piekna-Przybylska et al., 2010
; Song et al., 2009
). Alteration or deletion of motif 9 and segment 1 reduced the efficiency of minus strand transfer from about 70% to just about 10%, supporting their proposed role in transfer (Piekna-Przybylska et al., 2010
Our current goal was to determine whether motif 9 and segment 1 had the expected biological function. The approach was to set up a cell culture assay that could detect the efficiency of minus strand transfer in viral systems with normal or altered motif 9 and segment 1. At the outset we realized that the system in vivo could not be as sensitive as the assay in vitro, since in vivo we could not eliminate all of the additional RNA structural factors that promote transfer.
Our approach was to generate three HIV-1 mutant viruses with altered sequences in motif 9 and segment 1. In transfected cells, these mutations do not affect the synthesis and stability of viral RNA, and normal level of virions are produced. However, if the U3 mutations affect the efficiency of minus strand DNA transfer, they would influence expression of the next generation of viruses after first round of reverse transcription in infected cells. The sequence of the promoter for synthesis of the RNA genome is also inherited in the U3 region at the 3’ end of the viral RNA, and overlaps with the tRNA gene-like sequence. Alteration of the promoter and related sequences will result in mutation of binding sites for transcription factor Sp1 and Sp3 (Pereira et al., 2000
). By using the single cycle of recombinant virus we were able to exclude these issues and limit the effect of mutations in U3 solely to the minus strand transfer reaction.
The major implication of our results is that tRNA-like U3 sequences have an influence on minus strand DNA transfer in the cell. While the qualitative interpretation of the results seems clear, the quantitative outcomes deserve comment. The observed reductions of 9% to 26% in DNA transfer are generally much smaller than the differences observed in vitro. However, as we mentioned above, our reconstituted system eliminated other RNA structural factors contributing to minus strand transfer. In addition, unknown factors may also help to circularize the template RNA to facilitate transfer. If the interactions of motif 9 and segment 1 with the tRNA were part of a group of circularizing mechanisms, the effect of the two segments would be moderate in terms of percentage change. However, since the minus strand transfer step is likely to affect the overall growth kinetics of the virus, there could be substantial evolutionary pressure supporting any mechanism that even moderately improves transfer efficiency.
The results showed that alteration of either motif 9 or segment 1 reduced transfer. Surprisingly, alteration of both had a lesser than expected effect. Unfortunately, the effects of any mutation can be pleiotropic. For example, strand transfer has been shown to be influenced by the folding structure of the acceptor template, since some structures can interfere with annealing to the (-)ssDNA. Folding analysis of the effects of our mutations does not predict substantial structural changes, but it is not possible to rule out subtle differences among the wild-type and mutant acceptor folding structures.
What other factors could bring the 5’ and 3’ ends of the viral RNA template into proximity? It was shown in vitro
that the invasion-driven mechanism of minus strand transfer would still support efficient transfer when branch migration of the acceptor template from the invasion site to the terminus transfer site was not allowed (Song et al., 2008
). We interpreted this result to mean that interaction of the acceptor RNA at the invasion site held the donor and acceptor RNAs together. This means that the invasion mechanism itself is a mechanism for template circularization, although differing from the U3 interactions in that it occurs after some (-)ssDNA synthesis. A specific interaction was also demonstrated in vitro
between the R element at the 3’ end of HIV-1 and a gag
sequence, leading to circularization of the genome (Gee et al., 2006
; Ooms et al., 2007
) (). Analysis in vitro
revealed that these interactions enhance minus strand transfer when a DNA primer is used to start reverse transcription (Beerens and Kjems, 2010
). As the R element is involved in minus strand transfer the interactions with gag
during genome circularization are disrupted by (-)ssDNA synthesis. However, the interactions of tRNA3Lys
with sequences in U3 could still maintain circularization until reverse transcription reaches the U3 region after transfer.
Fig. 5 Proposed RNA-RNA contacts between the 5’ and 3’ ends of the RNA genome in HIV-1. The model shows the interactions between gag sequences and 3’ U3/PolyA sequences, and between tRNA3Lys (red), bound to the PBS region (black thick (more ...)
The sequences of motif 9 are complementary to nucleotides in the 3' part of the anticodon stem and part of the variable loop of tRNA3Lys
(Brule et al., 2000
). This region of the tRNA has been proposed to interact with genomic RNA sequences located upstream of the PBS during formation of the initiation complex with RT (Isel et al., 1995
). Marquet and coworkers proposed that interaction between U3 and the anticodon stem of tRNA3Lys
could be established after initiation of DNA synthesis, when the base pairings within U5 are unwound (Brule et al., 2000
). However, segment 1 resembles PBS, which interacts with 18nt at the 3’ end of tRNA3Lys
. Soon after minus strand transfer takes place and the PPT sequence is copied, the synthesis of plus strand DNA could start and proceed until tRNA3Lys
. It is unknown when tRNA3Lys
is displaced from PBS, but displacement is necessary to allow copying of the 3’ end of the tRNA into DNA in preparation for second strand transfer. The interaction with segment 1 of U3 could facilitate this process. Considering the time-dependent alternative structures, the putative interactions between U3 and tRNA3Lys
appear to be part of a series of dynamic events that hold the viral ends together as the steps of (-)ssDNA, transfer and subsequent synthesis into U3 proceed.
Overall, U3-tRNA3Lys interactions are suggested by our results to be part of a complex series of changing RNA-DNA structures that promote minus strand transfer. A central theme of structure formation and breakage is to transiently circularize the template genomic RNA to bring its ends into proximity. Disruption of the U3-tRNA3Lys interaction reduces the efficiency of transfer, but only moderately. This observation supports the conclusion that other proposed mechanisms that promote transfer are also important and effective. Significantly, it also suggests that the complex multi-step stimulation process has evolved because the most rapid and efficient minus strand transfer possible is strongly selected in evolution and necessary for viral fitness and long-term survival.