The HIV-1 RNA genome is rapidly evolving and, although its sequence heterogeneity is widely recognized, its biological origins and consequences are poorly understood. Sequence variability at the population (interhost) and the individual (intrahost) level stems mainly from the inherent infidelity of the reverse transcriptase enzyme that lacks a proofreading activity. The high HIV-1 replicative turnover, the relative large size of the viral quasispecies population in an infected individual, and the propensity for recombination all add to the viral capacity to adapt to immunological, pharmaceutical, or other selection pressures. Despite the high evolution rate of HIV-1, there are indications that its evolution is not unlimited but confined to a restricted sequence space. We tried to assess the actual sequence space available for HIV-1 in a novel experimental setting.
The evolutionary capacity of HIV-1 was probed by putting a strong selective pressure on highly conserved sequences within the HIV-1 genome and comparing the natural with the selected variation. We used RNAi by means of short hairpin RNA (shRNA) inhibitors to target highly conserved 19-nucleotide sequences within the HIV-1 RNA genome . Multiple cell culture infections were followed to select for viral escape variants. Details of this massive virus evolution study can be found in another article . We compiled the RNAi-resistant mutations that were selected (approximately 500 sequences) with four shRNA inhibitors (Fig. 1a) and compared this with the variation observed in 625 natural HIV-1 strains.
We grouped all escape data with respect to the position of the acquired mutations within the 19-nucleotide targets, which yields a typical pattern (Fig. 1b). In theory, 19 × 3 = 57 point mutations are possible per target, which means 228 escape possibilities for the four targets. However, we scored only a restricted set of 36 escape routes. Some target positions are highly preferred for acquisition of escape mutations, whereas others are strongly avoided, especially at both ends of the target (positions 1, 2, 18, 19). The lack of escape mutations at the target ends is most likely dictated by the RNAi mechanism, which is known to allow mismatches between the siRNA effector and the target sequence at these sites [3–8]. However, we also observed considerable variation in escape pattern for the central target domain (positions 3–17) that cannot be explained by properties of the RNAi mechanism. As we selected the most conserved viral target sequences for the RNAi-therapeutic approach [1,5,7], we checked whether the induced sequence variation would mimic the limited sequence variation present in natural HIV-1 strains in the Los Alamos database. We observed a striking similarity between the RNAi-induced sequence variation (Fig. 1b) and natural sequence variation (Fig. 1c). These results imply that only a selective set of HIV-1 variants are allowed to evolve, both in nature and under RNAi pressure. This also indicates that the HIV-1 variants that are not selected in vitro and in vivo represent viruses with suboptimal replication capacity and fitness. This idea is further supported by the preponderance of silent codon changes among the RNAi-resistant viruses . These results suggest that HIV-1 evolution already reached the outer limits of the allowed sequence space, at least for the highly conserved domains that were probed in this study.
Several studies previously suggested that HIV-1 evolution is restricted to a relatively small domain of sequence space. Most importantly, Lukashov and Goudsmit  compared HIV-1 subtype B sequences early in the Amsterdam HIV-1 epidemic versus 10 years later. They described that the high evolution rate of HIV-1 leads to a steady increase in synonymous distance from the consensus sequence, but the nonsynonymous distance remained constant over time. It was concluded that HIV-1 sequences fluctuate within a sequence space with fixed distance from the subtype consensus. This study focused on the V3 domain of the Envelope protein, which may be particularly constrained in function as it is a major determinant for receptor/co-receptor interactions and the key determinant of the CCR5-CXCR4 coreceptor switch.
In natural infections, cytotoxic T lymphocytes (CTLs) put pressure on different HIV-1 protein epitopes. There is accumulating evidence that some of the Gag epitopes tolerate only limited sequence variation because of a fitness cost . These fitness problems also explain the reversion of modified CTL target sequences to the consensus sequence upon HIV-1 transmission to a new host [11–13]. These in-vivo results strongly argue against unremitting adaptation of HIV-1 to host immune pressures and strongly support the idea of restricted HIV-1 evolution. Similar fitness losses have been described for drug-resistant variants of the Protease enzyme [14,15], variation within the Tat protein  and the Integrase enzyme . Protease sequences can, in fact, be subject to dual pressure from antiretroviral therapy and CTLs . A complicating factor is that compensatory mutations are able to mask the deleterious effects of another mutation. This has been described for CTL-escape mutations in the Gag protein  and drug-resistance mutations in the viral Protease enzyme [7,20]. Such compensatory changes can even be selected in vitro and in vivo in the absence of drug pressure, demonstrating that repair of enzyme function and viral fitness is the driving force .
Our current findings also have important implications for the design of improved antiviral RNAi strategies, in particular with respect to attempts to avoid viral escape. We previously used one simple criterion to select target sites: optimal conservation of the targeted sequence among virus strains . We now propose additional criteria to stipulate target selection. First, selection of the most conserved targets should focus on the center of the 19-nucleotide target domain (positions 3–17), as RNAi is tolerant for sequence variation at the termini of the target. Second, this 3–17 domain should be selected with the least number of silent codon positions (mostly third codon positions). Third, it seems beneficial to find targets enriched for AUG and UGG codons encoding methionine and tryptophan, respectively, because these are the only two codons that lack options for silent escape. Fourth, our data suggest that one can actually use the available information on natural HIV-1 sequence variation in the design of improved shRNA inhibitors, as this should define the available sequence space for HIV-1 that can be used to obtain RNAi resistance. Specifically, one could define targets that restrict sequence variation to only one or two positions, in which case one could anticipate these viral escape routes and design a limited set of second generation shRNAs to block these escape routes . These considerations are of importance for the development of a durable RNAi-based gene therapy for HIV-AIDS.
We thank Vladimir Lukashov for critical reading of the manuscript. RNAi research in the Berkhout laboratory is sponsored by ZonMw (Vici grant and Translational Gene Therapy grant).
1. ter Brake O, Konstantinova P, Ceylan M, Berkhout B. Silencing of HIV-1 with RNA interference: a multiple shRNA approach. Mol Ther 2006; 14:883–892.
2. von Eije KJ, ter Brake O, Berkhout B. Human immunodeficiency virus type 1 escape is restricted when conserved genome sequences are targeted by RNA interference. J Virol 2008; 82:2895–2903.
3. Schwarz DS, Ding H, Kennington L, Moore JT, Schelter J, Burchard J, et al
. Designing siRNA that distinguish between genes that differ by a single nucleotide. PLoS Genet 2006; 2:e140.
4. Haley B, Zamore PD. Kinetic analysis of the RNAi enzyme complex. Nat Struct Mol Biol 2004; 11:599–606.
5. Du Q, Thonberg H, Wang J, Wahlestedt C, Liang Z. A systematic analysis of the silencing effects of an active siRNA at all single-nucleotide mismatched target sites. Nucleic Acids Res 2005; 33:1671–1677.
6. Sabariegos R, Gimenez-Barcons M, Tapia N, Clotet B, Martinez MA. Sequence homology required by human immunodeficiency virus type 1 to escape from short interfering RNAs. J Virol 2006; 80:571–577.
7. Berkhout B. HIV-1 evolution under pressure of protease inhibitors, climbing the stairs of viral fitness. J Biomed Sci 1999; 6:298–305.
8. Pusch O, Boden D, Silbermann R, Lee F, Tucker L, Ramratnam B. Nucleotide sequence homology requirements of HIV-1-specific short hairpin RNA. Nucleic Acids Res 2003; 31:6444–6449.
9. Lukashov VV, Goudsmit J. Evolution of the human immunodeficiency virus type 1 subtype-specific V3 domain is confined to a sequence space with a fixed distance to the subtype consensus. J Virol 1997; 71:6332–6338.
10. Peyerl FW, Barouch DH, Yeh WW, Bazick HS, Kunstman J, Kunstman KJ, et al
. Simian-human immunodeficiency virus escape from cytotoxic T-lymphocyte recognition at a structurally constrained epitope. J Virol 2003; 77:12572–12578.
11. Li F, Horton H, Gilbert PB, McElrath JM, Corey L, Self SG. HIV-1 CTL-based vaccine immunogen selection: antigen diversity and cellular response features. Curr HIV Res 2007; 5:97–107.
12. Fernandez A, Tawfik DS, Berkhout B, Sanders R, Kloczkowski A, Sen T, et al
. Protein promiscuity: drug resistance and native functions – HIV-1 case. J Biomol Struct Dyn 2005; 22:615–624.
13. Wolinsky SM, Korber BT, Neumann AU, Daniels M, Kunstman KJ, Whetsell AJ, et al
. Adaptive evolution of human immunodeficiency virus-type 1 during the natural course of infection. Science 1996; 272:537–542.
14. Fernandez G, Clotet B, Martinez MA. Fitness landscape of human immunodeficiency virus type 1 protease quasispecies. J Virol 2007; 81:2485–2496.
15. Rouzine IM, Coffin JM. Search for the mechanism of genetic variation in the pro gene of human immunodeficiency virus. J Virol 1999; 73:8167–8178.
16. Meyerhans A, Cheynier R, Albert J, Seth M, Kwok S, Sninsky J, et al
. Temporal fluctuations in HIV quasispecies in vivo are not reflected by sequential HIV isolations. Cell 1989; 58:901–910.
17. Reinke R, Steffen NR, Robinson WE Jr. Natural selection results in conservation of HIV-1 integrase activity despite sequence variability. AIDS 2001; 15:823–830.
18. Karlsson AC, Deeks SG, Barbour JD, Heiken BD, Younger SR, Hoh R, et al
. Dual pressure from antiretroviral therapy and cell-mediated immune response on the human immunodeficiency virus type 1 protease gene. J Virol 2003; 77:6743–6752.
19. Crawford H, Prado JG, Leslie A, Hue S, Honeyborne I, Reddy S, et al
. Compensatory mutation partially restores fitness and delays reversion of escape mutation within the immunodominant HLA-B*5703-restricted Gag epitope in chronic human immunodeficiency virus type 1 infection. J Virol 2007; 81:8346–8351.
20. Maisnier-Patin S, Andersson DI. Adaptation to the deleterious effects of antimicrobial drug resistance mutations by compensatory evolution. Res Microbiol 2004; 155:360–369.
21. Borman A, Paulous S, Clavel F. Resistance of HIV-1 to protease inhibitors: selection of resistance mutations in the presence and in the absence of the drug. J Gen Virol 1996; 77:419–426.
22. ter Brake O, Berkhout B. A novel approach for inhibition of HIV-1 by RNA interference: counteracting viral escape with a second generation of siRNAs. J RNAi Gene Silencing 2005; 1:56–65.