Host-Specific Driving Force in Human Immunodeficiency Virus Type 1 Evolution In Vivo

Transcription

1 JOURNAL OF VIROLOGY, Mar. 1997, p Vol. 71, No X/97/$ Copyright 1997, American Society for Microbiology Host-Specific Driving Force in Human Immunodeficiency Virus Type 1 Evolution In Vivo LINQI ZHANG, 1 * RICARDO S. DIAZ, 2 DAVID D. HO, 1 JAMES W. MOSLEY, 3 MICHAEL P. BUSCH, 2 AND ALLEN MAYER 1,2 Aaron Diamond AIDS Research Center, The Rockefeller University, New York, New York ; Irwin Memorial Blood Centers, San Francisco, California ; and Transfusion Safety Study, University of Southern California School of Medicine, Los Angeles, California Received 31 May 1996/Accepted 9 December 1996 To investigate the process of human immunodeficiency virus type 1 (HIV-1) evolution in vivo, a total of 179 HIV-1 V3 sequences derived from cell-free plasma were determined from serial samples in three epidemiologically linked individuals (one infected blood donor and two transfusion recipients) over a maximum period of 8 years. A systematic analysis of pairwise comparisons of intrapatient sequences, both within and between each sample time point, revealed a preponderance and accumulation of nonsynonymous rather than synonymous substitutions in the V3 loop and flanking regions as they diverged over time. This strongly argues for the dominant role that positive selection for amino acid change plays in governing the pattern and process of HIV-1 env V3 evolution in vivo and nullifies hypotheses of purely neutral or mutation-driven evolution or completely chance events. In addition, different rates of evolution of HIV-1 were observed in these three different individuals infected with the same viral strain, suggesting that the degree of positive pressure for HIV-1 amino acid change is host dependent. Finally, the observed similar rate of accumulation in divergence within and between infected individuals suggests that the process of genetic divergence in the HIV epidemic proceeds regardless of host-to-host transmission events, i.e., that transmission does not reset the evolutionary clock. A striking feature of human immunodeficiency virus type 1 (HIV-1) in vivo infection is the rapid generation and turnover of viral variants, resulting in a high degree of sequence diversity within and between infected individuals (5, 7, 10, 16, 18, 20, 21, 26, 33, 38, 40, 41). The rapid dynamics of HIV-1 replication in vivo is likely to be a consequence of intrinsic properties of the virus, such as a high replication rate and substantial numbers of viral progeny produced per replication cycle (8, 15, 39). The cause of the high degree of sequence diversity in vivo, however, is much less well understood. It is fairly clear that the error-prone nature of the viral reverse transcriptase provides the biochemical basis for the observed viral variation. What determines the ultimate fate of a mutated variant genome, however, is controversial. One viewpoint is that the accumulation of specific variants is largely determined by their relative fitness in a given selective environment. Immune surveillance (16, 20, 26, 34, 41) and viral cell tropism (4, 31, 37) are examples of plausible selective forces that may be shaping HIV diversity in vivo. Others discount the role of selection at the amino acid level and propose that the observed diversity could result from mutation-driven evolution (36), neutral evolution (14), or simply chance antigenic stimulation of lymphocytes carrying resident variant proviruses (38). Evidence that supports the existence of some degree of selective pressure on the V3 region of HIV-1 env comes from analyses of K a and K s values and the K a /K s ratio of sequence pairs, where K a is the observed frequency of nonsynonymous substitutions per replacement site and K s is the observed frequency of synonymous substitutions per silent site (22 24, 28, 32, 33). The higher the value of K a or K s, the more divergence there is between a pair * Corresponding author. Mailing address: Aaron Diamond AIDS Research Center, Rockefeller University, 455 First Ave., 7th Floor, New York, NY Phone: (212) Fax: (212) lzhang@adarc.org. of sequences, and the higher the K a /K s ratio, the stronger the selection pressure for amino acid change. To quantify and assess the temporal pattern of selective constraints that might underlie the process of HIV-1 evolution in vivo, Bonhoeffer and colleagues (2) used HIV-1 plasma V3 sequence data from samples collected from an HIV-1-infected hemophiliac over a 7-year time period. These sequences were analyzed in terms of the changes in K a, K s, and the K a /K s ratio within each sampling point. The K s value within each sampling point was seen to increase over time, whereas the K a value remained roughly the same. Consequently, the K a /K s ratio within each sampling point decreased over time, which is interpreted as being due to a gradually weakening selective pressure for amino acid sequence change as the patient s CD4 counts dropped (2). A circumstance consisting of the transfusion in 1985 of components derived from the same HIV-1-contaminated unit of blood into two recipients, followed by the enrollment of the donor and recipients in the Transfusion Safety Study (11), has allowed us to compare the evolution of the same HIV-1 strain in three different hosts, namely, the two recipients (A and B) and the actual blood donor. The V3 regions of genomes present in sequential plasma samples obtained from these three epidemiologically linked individuals over a maximum period of 8 years were sequenced (10a). To assess the intensity and pattern of the selective constraints exerted by different individuals on the same viral strain, we have used this sequence data to calculate both intra- and intersample K a, K s, and the K a /K s values. Details of the donor and the two patients, the transfusion event, and the sequential samples obtained from them as part of the Transfusion Safety Study (11) up to 8 years after the transfusion are given elsewhere (10a). A total of 179 HIV-1 RNA genomes derived from cell-free plasma were PCR amplified for the V3 region after reverse transcription, as described in detail in reference 10a. V3 region sequences were 2555

2 2556 NOTES J. VIROL. Downloaded from FIG. 1. The amino acid alignments of donor (a), recipient A (b), and recipient B (c) sequences are shown aligned with the consensus sequence obtained from the donation sample in 1985 (D85-con). Dots represent amino acid identity to the consensus, dashes indicate gaps introduced to preserve the alignment, and pound signs represent sites of synonymous substitutions. Nonsynonymous substitutions are indicated by lowercase letters. The V3 loop is highlighted. This figure is a modified version of Fig. 2 in reference 10a. on October 5, 2018 by guest obtained by direct sequencing of nested PCR products derived from single cdna molecules. Ten genomes were obtained from the serum of the original donated unit of blood, 53 genomes were obtained from six sequential donor follow-up plasma samples, 57 genomes were obtained from five sequential recipient A plasma samples, and 59 genomes were obtained from six sequential recipient B plasma samples (Gen- Bank accession numbers U29433 to U29437, U29956 and U29957, U29959 to U30074, U30077 to U30145, U31573 to U31582, and U43035 to U43054). The 237-nucleotide-long sequences covered the entire V3 loop as well as flanking regions. The average frequencies of synonymous (K s ) and nonsynonymous (K a ) nucleotide substitutions for all pairwise sequence comparisons, both within and between samples, were calculated by the methods of Nei and Gojobori (29) and Li et al. (24). Because results obtained by these two methods were similar, only those calculated by the former method were used in the analysis. Intra- and intersample genetic distances were calculated with the program DNADIST, implemented in the PHYLIP package (version 3.5) (12). The Jukes-Cantor model of molecular evolution was used to correct for multiple substitutions in observed genetic distance, as well as in K a and K s

3 VOL. 71, 1997 NOTES 2557 Downloaded from (29). Regression curves were constructed by the program Curve Fitting, implemented in the DeltaGraph package (version 3.0) (9), and correlation coefficients (r values) were estimated by the principle of least squares (9). The amino acid alignments of the sequences from the donor and the two recipients are shown in Fig. 1 to highlight both synonymous and nonsynonymous changes. As discussed by Diaz et al. (10a), the general pattern seen is one of individualspecific divergence, but there are some common amino acid changes in the different individuals. No unusual mutations or foci of mutations with respect to published sequences are prominent. A phylogenetic study demonstrated that early sequences cluster around the origin of the tree whereas sequences from later time points are more distant (10a). The average number of nucleotide substitutions per nonsynonymous site (K a ) and per synonymous site (K s ) for all pairwise comparisons of sequences within each sampling point were calculated for the donor and the two transfusion recipients, the same method used in the study by Bonhoeffer et al. (2). The FIG. 1 Continued. means of intrasample K a, K s, and K a /K s values are shown in Fig. 2a to c, respectively, plotted versus the time at which the samples were collected. The intrasample K s values for the three individuals fluctuate substantially without a steady increase or decrease over time (Fig. 2b). This is different from the gradual increase observed in the individual studied by Bonhoeffer et al. (2). The intrasample K a values in our three individuals appear to increase with time, and they drop off late in recipient B. This contrasts with the relatively stable K a value in the individual in the study by Bonhoeffer et al. (2). The intrasample K a /K s ratios fluctuate with time like K s values, although there is a minor tendency toward an increase over time (K a /K s 1 during the later years of infection). The lack of concordance between the results from this study and those from the study by Bonhoeffer et al. (2) might be due to differences in host factors or, alternatively, the analytic approach used. In an attempt to explore the second possibility, we chose to go beyond the snapshot measurement provided by the intrasample K a and K s values to the calculation of the on October 5, 2018 by guest

4 2558 NOTES J. VIROL. Downloaded from FIG. 1 Continued. on October 5, 2018 by guest intersample K a and K s values obtained by pairwise comparison of sequences belonging to samples obtained at different times (termed rk a and rk s ). The means of intersample rk a and rk s values for the three individuals (one donor and two recipients) were plotted as a function of corresponding nucleotide divergence, calculated under the Jukes-Cantor model of molecular evolution (29). By using this approach, the relative contributions of nonsynonymous and synonymous changes to the overall nucleotide divergence can be readily estimated. The results are shown in Fig. 3, which demonstrates a strong linear correlation between the rk a values and genetic divergence for all three individuals. The r value for rk a and genetic divergence is highly significant, reaching 0.99 (P 0.001) in the donor and in recipient A and 0.96 (P 0.001) in recipient B. In contrast, the relatedness between rk s and genetic divergence is minimal. The r values for rk s and genetic divergence are all relatively small in both recipients A (0.15, P 0.10) and B (0.29, P 0.10), although a weak correlation can be seen in the donor (0.66, P 0.01). It is also interesting to note that in cases where genetic divergence is relatively small, rk s values are about the same as the corresponding rk a values. However, when the overall genetic divergence increases, rk a increases at a much faster rate than rk s. As a result, rk a values gradually surpass rk s values and become the dominant contributor to the overall genetic divergence (Fig. 3). The observed pattern is consistent with the notion that selective pressure in these three individuals actively promotes amino acid changes in the V3 loop and flanking regions. To investigate the temporal pattern of divergence over time in these three individuals, rk a,rk s, and genetic distance were

5 VOL. 71, 1997 NOTES 2559 FIG. 2. Average intrasample K a, K s, and K a /K s values from plasma-derived HIV-1 RNA sequences obtained from three epidemiologically linked individuals over the time period of the study. The open circles, squares, and crosses represent samples collected from the donor, recipient A, and recipient B, respectively. Each point represents the average value obtained from pairwise sequence comparisons within a particular sample. plotted versus time, using pairwise comparisons between the first and all subsequent samples of each individual. Figure 4 demonstrates the linear regression lines for rk a,rk s, and genetic distance versus time in the donor and the two recipients, A and B. Strong linear correlations for rk a and genetic distance were observed, showing an increase with time. The linear correlations shown are strongly supported by the high r values in all three of the individuals analyzed. With regard to the regression curve for rk a and time, the r value is 0.94 (P 0.001) in the donor, 0.98 (P 0.001) in recipient A, and 0.91 (P 0.001) in recipient B, whereas for genetic distance and time, the r value is as high as 0.94 (P 0.001) in the donor, 0.98 (P 0.001) in recipient A, and 0.91 (P 0.001) in recipient B. A linear correlation between rk s and time, on the other hand, is much less apparent, with r values of 0.83 in the donor, 0.42 in recipient A, and 0.67 in recipient B. The linear correlations between both rk a and genetic divergence and time suggest that the entire intrapatient viral population, as far as the V3 loop and flanking regions are concerned, progressively diverges away from the initial viral inoculum without significant back mutation to, or reappearance of, previous sequences. The linear relationships obtained by this type of analysis facilitate quantitative comparisons between different individuals. When the slopes of the rk a and genetic distance curves in Fig. 4 for the three different individuals are compared, differences are noted, indicating different evolutionary paces of HIV-1 V3 sequences within the different hosts. The slopes of the rk a regression lines for the donor and for recipients A and B are 0.007, 0.009, and 0.013, respectively, and the respective values for genetic distance are 0.007, 0.007, and The rates of amino acid and nucleotide change in recipient B are therefore approximately 1.7 times faster than those in the donor and recipient A. The different pace of viral evolution in recipient B, although this individual was initially infected with the same viral strain as the donor and recipient A, argues for the existence of different intensities of selective pressure in different individuals. In summary, we have analyzed the evolution of the hypervariable V3 region of a single HIV-1 strain over time in three epidemiologically related infected individuals. The existence, temporal pattern, and relative strength of positive pressure for HIV-1 amino acid change within the different individuals were assessed by calculating the relative contributions of nonsynonymous and synonymous nucleotide changes to V3 genotypic divergence over time. We observed marked fluctuations over time in the values of K s, K a, and K a /K s calculated from intrasample sequence comparisons for the three individuals belonging to this epidemiological cluster, which makes quantitation and comparison between the different individuals difficult. We show instead that plotting intersample rk a values versus time yields linear curves that are better suited for interpatient comparisons. Furthermore, plotting the intersample rk a and rk s values versus genetic divergence effectively demonstrates the predominant contribution of nonsynonymous rather than synonymous changes to the V3 nucleotide sequence divergence shown by this particular HIV-1 strain within all three of the individuals infected by it. If this strain is representative of HIV-1 in general, then selective pressure appears to be the dominant driving force for V3 sequence evolution leading to the vast diversification in FIG. 3. The average values of the frequencies of intersample synonymous (rk s ) (open squares) and nonsynonymous (rk a ) (solid circles) substitutions plotted versus total genetic distance in the donor, recipient A, (RA), and recipient B (RB). Each point represents the average value obtained from pairwise comparisons of sequences for two samples from a given individual. The ranges of the standard deviations (SD) for rk a and rk s were comparable (donor: SD rka, 0.7 to 3.8%, and SD rks, 1.3 to 2.9%; recipient A: SD rka, 0.7 to 2.7%, and SD rks, 1.1 to 2.4%; recipient B: SD rka, 0.6 to 3.3%, and SD rks, 1.5 to 3.3%).

6 2560 NOTES J. VIROL. FIG. 4. The average values of the frequencies of intersample synonymous (rk s ) (open squares) and nonsynonymous (rk a ) (solid circles) substitutions and genetic distances (open triangles) relative to the first available sample over the time period of the study. Each point represents the average value obtained from pairwise comparisons between sequences in each subsequent sample and sequences in the first sample. RA, recipient A; RB, recipient B. viral quasispecies in vivo (1, 5, 7, 10, 16, 18, 20, 21, 25, 26, 33, 34, 40, 41). If HIV diversity was instead largely a manifestation of purely neutral evolution (14), mutation-driven evolution (36), or completely random events (38), then one would not expect to see a more significant correlation between rk a and time than between rk s and time over the course of infection. The preponderance of nonsynonymous substitutions observed and the rapid turnover of the viral population observed by Ho et al. (15) and Wei et al. (39) make it clearly inappropriate to apply neutral and chance theory to the evolutionary process of HIV-1. The rapid accumulation of nonsynonymous substitutions in other retroviruses and RNA viruses, such as simian immunodeficiency virus (3) and foot-and-mouth disease virus (13), also argues for a similar nonneutral process of viral evolution in vivo. In addition, the observation here that the slopes of the rk a regression versus time curves are not all the same in these three individuals infected with the same viral strain suggests that the selective pressure for V3 region amino acid change is host dependent. A similar conclusion was reached from an analysis of HIV-1 proviral DNA from two time points per person in a cluster of five epidemiologically linked transfusion recipients (40). However, the correlation between the selective pressure for amino acid change and the rate of clinical progression is not clear-cut, since recipient A, who showed a more moderate clinical progression and V3 amino acid change than recipient B, showed about the same pressure for divergence as the donor, who did not progress clinically over the course of the study (10a). A study by Lukashov et al. (25) comparing two sequences, one obtained during seroconversion and one determined 5 years later, from each of 44 individuals showed a positive correlation between the length of the immunocompetent period (CD4 counts 200) and the degree of nonsynonymous substitution, suggesting that slower progressors have stronger selective pressures for amino acid change. However, in five HIV-1-infected infants, Strunnikova et al. (35) found an inverse correlation between average CD4 cell counts and the rate of sequence divergence. In a recent study of six patients, two moderate progressors and one slow progressor showed higher rates of nonsynonymous changes than one slow progressor and two rapid progressors (41). It must be emphasized that our population consists of only three patients and that they were not selected to represent extremes with regard to rate of clinical progression. The relationship between the rate of sequence divergence and CD4 cell decline warrants further study. Our results do not directly address the issue of the nature of the force for selection, be it related to immune escape or to the acquisition of a wider cellular host range for the virus within the infected individual. If the pressure for amino acid change is due to immune selection, then at least for the three individuals studied here, sharing the same viral strain, there was no clear-cut correlation between the degree of immune selection and the time to onset of immunodeficiency. Due to the significant linear correlation between genetic distance and time in these three individuals, we ventured further to estimate the rate of genetic divergence within each of the study subjects. The slopes of the genetic distance regression curves yield average rates of increase in genetic divergence per year of 0.7% in the donor, 0.7% in recipient A, and 1.2% in recipient B. The overall average among these three individuals is therefore 0.87% per year, which is in fact very similar to the rate of nucleotide divergence (1.10% per year) calculated from a large cross-sectional sequence data set (19, 20, 27, 28). The observed rate of variation is also surprisingly similar to that seen in macaques experimentally infected with molecularly cloned simian immunodeficiency virus (3, 17). The similar rates of accumulation of divergence within and between infected individuals suggests that the process of genetic divergence in the HIV epidemic proceeds regardless of host-to-host transmission events, i.e., that transmission does not reset the evolutionary clock. Thus, the homogeneous viral population observed during seroconversion (6, 30, 42, 43) may not be the result of selection back to a common universal transmission sequence but rather may be due to selection for a particular viral phenotype within a given incoming viral population. Indeed, distinct viral sequences that share a predominantly macrophage-tropic phenotype have been obtained from different individuals during seroconversion (6, 30, 43). The sequences of these transmitted viral genomes would be predicted to have evolved away from their ancestral sequence at a rate of about 1% per year, particularly during the early course of the HIV-1 epidemic in the human population. We thank Steve Wolinsky, Andrew Leigh Brown, and Sebastian Bonhoeffer for critical reading of the manuscript. Financial support for the collection and documentation of the subjects and their specimens was provided by contracts NO1-HB-47002, NO1-HB-47003, and NO1-HB from the NHLBI and is gratefully acknowledged. The sequencing was supported by NHLBI grant RO1-HL The visit by R.S.D. to Irwin Memorial Blood Centers was funded by the Fogarty International Program through the University of California, Berkeley. REFERENCES 1. Balfe, P., P. Simmonds, C. A. Ludlam, J. O. Bishop, and A. J. Leigh Brown Concurrent evolution of human immunodeficiency virus type 1 in patients infected from the same source: rate of sequence change and low

7 VOL. 71, 1997 NOTES 2561 frequency of inactivating mutations. J. Virol. 64: Bonhoeffer, S., E. Holmes, and M. Nowak Causes of HIV diversity. Nature 376: Burns, D. P. W., and R. C. Desrosiers Selection of genetic variants of simian immunodeficiency virus in persistently infected rhesus monkeys. J. Virol. 65: Cheng-Mayer, C., D. Seto, M. Tateno, and J. Levy Biological features of HIV-1 that correlate with virulence in the host. Science 240: Cichutek, K., H. Merget, S. Norley, R. Linde, W. Kreuz, M. Gahr, and R. Kurth Development of a quasispecies of human immunodeficiency virus type 1 in vivo. Proc. Natl. Acad. Sci. USA 89: Cichutek, K., S. Norley, R. Linde, W. Kreuz, M. Gahr, J. Lower, V. von Wangenheim, and R. Kurth Lack of HIV-1 V3 region sequence diversity in two haemophiliac patients infected with a mutative biologic clone of HIV-1. AIDS 5: Coffin, J Genetic diversity and evolution of retroviruses. Curr. Top. Microbiol. Immunol. 176: Coffin, J HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science 267: DeltaGraph Deltapoint, Inc., Monterey, Calif. 10. Delwart, E. L., H. W. Sheppard, B. D. Walker, J. Goudsmit, and J. I. Mullins Human immunodeficiency virus type 1 evolution in vivo tracked by DNA heteroduplex mobility assays. J. Virol. 68: a.Diaz, R. S., L. Zhang, M. P. Busch, J. W. Mosley, and A. Mayer. AIDS, in press. 11. Donegan, E., M. Stuart, J. C. Niland, H. S. Sacks, S. P. Azen, S. L. Dietrich, C. Faucett, M. A. Fletcher, S. H. Kleinman, E. A. Operskalski, H. A. Perkins, J. Pindyck, E. R. Schiff, D. P. Stites, P. A. Tomasulo, and J. W. Mosley Infection with human immunodeficiency virus type 1 (HIV-1) among recipients of antibody-positive blood donations. Ann. Intern. Med. 113: Felsenstein, J PHYLIP manual, version 3.5. Department of Genetics, University of Washington, Seattle. 13. Gebauer, F., J. C. de la Torre, I. Gomes, M. G. Mateu, H. Barahona, B. Tiraboschi, I. Bergmann, P. Augé de Mello, and E. Domingo Rapid selection of genetic and antigenic variants of foot-and-mouth disease virus during persistence in cattle. J. Virol. 62: Gojobori, T., E. N. Moriyama, and M. Kimura Molecular clock of viral evolution, and the neutral theory. Proc. Natl. Acad. Sci. USA 87: Ho, D. D., A. Neumann, A. Perelson, W. Chen, J. M. Leonard, and M. Markowitz Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 373: Holmes, E. C., L. Zhang, P. Simmonds, C. A. Ludlam, and A. J. Leigh Brown Convergent and divergent sequence evolution in the surface envelope glycoprotein of human immunodeficiency virus type 1 within a single infected patient. Proc. Natl. Acad. Sci. USA 89: Johnson, P. R., T. E. Hamm, S. Goldstein, S. Kitov, and V. M. Hirsch The genetic fate of molecularly cloned simian immunodeficiency virus in experimentally infected macaques. Virology 185: Korber, B. T. M., E. E. Allen, A. D. Farmer, and G. L. Myers Heterogeneity of HIV-1 and HIV-2. AIDS 9(Suppl. A):S5 S Kuiken, C. L., V. V. Lukashov, E. Baan, J. Dekker, J. A. M. Leunissen, and J. Goudsmit Evidence for limited within-person evolution of the V3 domain of the HIV-1 envelope in the Amsterdam population. AIDS 10: Kuiken, C. L., G. Zwart, E. Baan, R. A. Coutinho, J. A. R. van del Hoek, and J. Goudsmit Increasing antigenic and genetic diversity of the HIV-1 V3 domain in the course of the AIDS epidemic. Proc. Natl. Acad. Sci. USA 90: LaRosa, G. J., J. P. Davide, K. Weinhold, J. A. Waterbury, A. T. Prosy, J. A. Lewis, A. J. Langlois, G. R. Dreesman, R. N. Boswell, and B. P. Shadduck Conserved sequence and structural elements in the HIV-1 principal neutralizing determinant. Science 249: Leigh Brown, A. J., and P. Monaghan Evolution of the structural proteins of human immunodeficiency virus: selective constraints on nucleotide substitution. AIDS Res. Hum. Retroviruses 4: Li, W. H., and D. Graur Fundamentals of molecular evolution. Sinauer Associates, Inc., Sunderland, Mass. 24. Li, W. H., C. I. Wu, and C. C. Luo A new method for estimating synonymous and nonsynonymous rates of nucleotide substitution considering the relative likelihood of nucleotide and codon changes. Mol. Biol. Evol. 2: Lukashov, V. V., C. L. Kuiken, and J. Goudsmit Intrahost human immunodeficiency virus type 1 evolution is related to length of the immunocompetent period. J. Virol. 69: McNearney, T., A. Hornickova, R. Markhan, A. Birdwell, M. Arens, A. Saah, and L. Ratner Relationship of human immunodeficiency virus type 1 sequence heterogeneity to stage of disease. Proc. Natl. Acad. Sci. USA 89: Myers, G., K. MacInnes, and B. Korber The emergence of simian/ human immunodeficiency viruses. AIDS Res. Hum. Retroviruses 8: Myers, G., and G. N. Pavlakis Evolutionary potential of complex retroviruses, p In J. A. Levy (ed.), The retroviruses. Plenum Press, New York, N.Y. 29. Nei, M., and T. Gojobori Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol. Biol. Evol. 3: Roos, M. T. L., J. M. A. Lange, R. E. Y. Goede, R. A. Coutinho, P. T. A. Schellekens, F. Miedema, and M. Tersmette Viral phenotype and immune response in primary human immunodeficiency virus type 1 infection. J. Infect. Dis. 165: Schiutemaker, H., M. Koot, N. A. Kootstra, M. W. Dercksen, R. E. Y. de Goede, R. P. van Steenwijk, J. M. A. Lange, J. K. M. Eeftink Schattenkerk, F. Miedema, and M. Tersmette Biological phenotype of human immunodeficiency virus type 1 clones at different stages of infection: progression of disease is associated with a shift from monocytotropic to T-cell-tropic virus populations. J. Virol. 66: Shpaer, E. G., and J. I. Mullins Rates of amino acid changes in the envelope protein correlate with pathogencicity of primate lentiviruses. J. Mol. Evol. 37: Simmonds, P., P. Balfe, C. A. Ludlam, J. O. Bishop, and A. J. Leigh Brown Analysis of sequence diversity in hypervariable regions of the external glycoprotein of human immunodeficiency virus type 1. J. Virol. 64: Simmonds, P., L. Q. Zhang, F. McOmish, P. Balfe, C. A. Ludlam, and A. J. Leigh Brown Discontinuous sequence changes of human immunodeficiency virus (HIV) type 1 env sequences in plasma viral and lymphocyteassociated proviral populations in vivo: implications for models of HIV pathogenesis. J. Virol. 65: Strunnikova, N., S. C. Ray, R. A. Livingston, E. Rubalcaba, and R. P. Viscidi Convergent evolution within the V3 loop domain of human immunodeficiency virus type 1 in association with disease progression. J. Virol. 69: Temin, H. W Is HIV unique or merely different? J. Acquired Immune Defic. Syndr. 2: Tersmette, M., J. M. Lange, R. E. de Goede, F. de Wolf, J. K. M. Eeftink Schattenkerk, P. T. Schellekens, R. A. Coutinho, J. G. Huisman, J. Goudsmit, and F. Miedema Association between biological properties of human immunodeficiency virus variants and risk for AIDS and AIDS mortality. Lancet i: Wain-Hobson, S Human immunodeficiency virus type 1 quasispecies in vivo and ex vivo. Curr. Top. Microbiol. Immunol. 176: Wei, X., S. K. Ghosh, M. E. Taylor, V. A. Johnson, E. A. Emini, P. Deutsch, J. D. Lifson, S. Bonhoeffer, M. A. Nowak, B. Hahn, M. S. Saag, and G. M. Shaw Viral dynamics in human immunodeficiency virus type 1 infection. Nature 373: Wolfs, T. F., J. J. de Jong, H. Van den Berg, J. M. Tijnagel, W. J. Krone, and J. Goudsmit Evolution of sequences encoding the principal neutralization epitope of human immunodeficiency virus type 1 is host dependent, rapid, and continuous. Proc. Natl. Acad. Sci. USA 87: Wolinsky, S. M., B. T. Korber, A. U. Neumann, M. Daniels, K. J. Kunstman, A. J. Whetsell, M. R. Furtado, Y. Cao, D. D. Ho, J. T. Safrit, and R. A. Koup Adaptive evolution of human immunodeficiency virus type 1 during the natural course of infection. Science 272: Zhang, L. Q., P. MacKenzie, A. Cleland, E. C. Holmes, A. J. Leigh Brown, and P. Simmonds Selection for specific sequences in the external envelope protein of human immunodeficiency virus type 1 upon primary infection. J. Virol. 67: Zhu, T., H. Mo, N. Wang, D. S. Nam, Y. Cao, R. A. Koup, and D. D. Ho Genotypic and phenotypic characterization of HIV-1 in patients with primary infection. Science 261: