HIV-1 is known to display a high degree of genetic diversity, in that each individual is host to a swarm of related viral genomes (quasispecies) that evolve over time . The role of host factors, such as the immune response, in the evolution of HIV can be highlighted in situations where different individuals are infected with the same strain of virus, and when viral divergence is tracked in these different individuals. This situation can be experimentally established using animal models, such as simian immunodeficiency virus infection of macaque monkeys , but in humans it can only be approximated in epidemiologically linked clusters of individuals exposed to the same infecting HIV-1 quasi-species.
Sequential quasispecies sampling and sequencing of cloned HIV-1 genomes from HIV-1-infected individuals belonging to such clusters has been performed in two studies [3,4] and showed variation to be host- dependent. In contrast, clinical and virological studies of one cohort of linked recipients and donors implicated viral strain as having a predominant influence on disease progression [5,6].
We present a study of a cluster of four HIV-1-infected individuals linked by blood transfusion of components derived from a single infected unit of blood donated in 1985 to two individuals, followed by sexual transmission to an additional individual. This study differs from previous studies in several respects. First, the blood donor (the source of the infection) was unambiguously known, was enrolled in the study, and was followed with multiple blood samples over time in parallel with the recipients. Furthermore, HIV sequences derived from the serum of the actual donated unit of blood have been analysed and are compared with the sequences diverging over time in the cluster. Thus, we know the origin of the cluster with respect to both time and majority HIV V3 sequences. In addition, we have analysed RNA sequences from plasma rather than viral DNA from peripheral blood mononuclear cells, thereby focusing on genomes that are being actively expressed and shed into blood. Finally, a unique feature of the cluster studied here is that it contained a sexual partner of one of the recipients, and hence we were able to study the course of viral evolution in a sexual couple over an extended period of time. Our study agrees with the previous studies [3,4] indicating the host-specific nature of viral variation, and shows that exchange of new viral variants did not appear to occur over time in the sexual couple after the initial viral transmission.
Materials and methods
The donation sera and plasma samples of the blood donor, both recipients (A and B), and the sexual contact of recipient B (SC) were obtained from the Transfusion Safety Study (TSS) Repositories [7–10]. The dates (month/year) of samples analysed by sequencing are as follows: donor, 1/1985, 9/1986, 2/1987, 1/1989, 1/1990, 3/1991, 3/1993; RA, 9/1986, 3/1987, 3/1989, 1/1990, 2/1991; RB, 11/1986, 3/1987, 2/1989, 1/1990, 8/1990; SC, 8/1988, 2/1989, 1/1990. The CD4 and CD8 counts were obtained from the TSS records.
Viral load determinations
Quantification of HIV RNA in the serum from the donation sample and sequential plasma samples was performed using the Amplicor HIV Monitor (Roche Diagnostics Systems, Somerville, New Jersey, USA) for the quantification of HIV-1 RNA as described previously [6,11]. Briefly, 200 µl plasma or serum was extracted with a lysis reagent containing guanidine thiocyanate and quantification standard (QS) RNA. Nucleic acids were precipitated with isopropanol and resuspended in a buffer containing carrier RNA. A 142 base-pair sequence in the gag gene of HIV-1 (defined by the biotinylated primers SK431 and SK462) was amplified by reverse transcriptase (RT) polymerase chain reaction (PCR) in a single reaction using the enzyme rTth DNA polymerase (Perkin Elmer-Cetus, Foster City, California, USA). The biotinylated HIV and QS PCR products were detected in separate wells of a microwell plate coated with HIV- and QS-specific oligonucleotide probes, respectively. In order to measure the HIV and QS-amplified products over a large dynamic range, fivefold serial dilutions were made in the HIV-1- and QS-specific wells. The bound, biotinylated products were quantified with an avidin-horseradish peroxidase (HRP) conjugate and a colorimetric reaction for HRP. The HIV-1 copy number was then calculated from the known input copy number of the QS RNA, the optical densities (450 nm) of the HIV and QS wells that fell within a defined range, and the dilution factors associated with the selected wells. All samples were tested for HIV RNA PCR under code in a single assay run, with the exception of the donation serum which was assayed last to ensure that this sample did not contaminate other samples. In contrast, for sequence analysis each sample was analysed individually from beginning to end to avoid PCR carry-over.
RNA extraction and cDNA synthesis
RNA was extracted from frozen serum or plasma using the guanidine–thiocyanate–RNA matrix bead adsorption kit (RNaid Plus kit; Bio101, La Jolla, California, USA) . cDNA was synthesized using 5 U avian myeloblastosis virus RT (Invitrogen, La Jolla, California, USA), 1 µg random hexamer primer (Boehringer Mannheim, Indianapolis, Indiana, USA) in 20 µl of 100 mM Tris (pH 8.3), 40 mM KCl, 10 mM MgCl2, and 2 mM dithiothreitol containing 10 U RNase inhibitor (Boehringer Mannheim) at 42°C for 1 h. RNA–cDNA hybrids were denatured at 95°C for 3 min, and then chilled on ice.
cDNA preparations were serially diluted so that when used as targets for nested PCR, only fewer than one in five reactions yielded PCR products. This technique of endpoint PCR was used so that PCR product would for the most part be derived from targets consisting of only single molecules of cDNA, thus avoiding the problems inherent to bulk PCR, such as in vitro recombination and selective amplification [13,14]. cDNA molecules were amplified either for a region encompassing the V3 region of the env gene or the p17 region of the gag gene. PCR was performed in 10 mM Tris (pH 8.3), 50 mM KCl, 3 mM MgCl2, 200 µM each dNTP, 0.2 µM each primer, and 2.5 U Taq polymerase (Perkin Elmer) per 100 µl reaction mix. Five microlitres of first- round PCR product was used in the second round of nested PCR. V3-nested amplification was performed using primers of Wolfs et al.  and p17 gag-nested amplification was performed using primers of Zhang et al. . Negative controls were included in all PCR amplifications to detect possible contamination from product carry-over. Thermal cycling conditions were 95°C for 1 min, 55°C for 1 min, and 72°C for 2 min for 35 cycles, with a final extension at 72°C for 10 min.
env V3 and p17 gag products were directly sequenced as follows: single-stranded PCR product was generated from 5 µl nested PCR product in an additional 45-cycle asymmetric PCR reaction using the second round 3′ primer. DNA from this third PCR round was purified by ethanol precipitation and sequenced using the second round 5′ primer in the dideoxy-chain termination procedure .
Nucleotide sequence accession numbers
All sequences reported in this publication have been submitted to GenBank (accession numbers U29433–U29437, U29956 and U29957, U29959–U30074, U30077–U30145, U31573–U31582, and U43035–U43054).
Sequences were aligned by using LINEUP and PILE-UP programs in the University of Wisconsin Genetic Computer Group package (GCG Inc., Madison, Wisconsin, USA). Final alignments were adjusted by eye. HIV-1NL43 and HIV-1OY1 sequences were used for comparison in the analysis. No gaps were allowed in the alignment. Sequence distances were estimated and corrected for multiple hits by using the program DNADIST from the PHYLIP package , under an evolutionary model that discerns the different rates of transition and transversion (two-parameter model of Kimura ). Evolutionary relationships were analysed by various phylogenetic approaches using the programs NEIGHBOR and DNAPARS implemented in the PHYLIP package. A neighbor-joining tree was constructed on the basis of the distance matrix, whereas maximum parsimony tree was based on the actual nucleotide structures. Due to the virtual identity in the trees' structure and topology generated by the two programs, only the neighbor-joining tree is presented in these studies.
Cluster members, clinical progression
As part of TSS, aliquots of serum from all blood donations from September 1984 to February 1985 in five US cities were frozen for storage. When a serologic test for HIV became available in 1985, the stored samples were tested , and donors and recipients of HIV-positive units were contacted and enrolled for follow-up [7,8,10]. The particular cluster studied here involved a blood donation on 22 January 1985, with transfusions of packed red cells to recipient A on 29 January 1985, and transfusion of platelets to recipient B on 24 January 1985. Recipient A was aged 57 years when transfused as part of treatment for a bleeding ulcer. Recipient B was aged 20 years when transfused for postpartum idiopathic thrombocytopenic purpura. Her husband (SC) was found to be HIV-seropositive in November 1986. The donor's risk factors were bisexuality and sexual intercourse with an injecting drug user.
The donor remained asymptomatic during the course of the study, and had stable and normal CD4 and CD8 counts, and a low viral load (Fig. 1). Recipient A showed a modest drop in CD4 count and an increase in viral load to the highest level seen in this cluster (Fig. 1). Zidovudine was started in January 1988, and when lost to follow-up in February 1991, recipient A had weight loss, diarrhoea, and a CD4 count below 200 × 106/l in July 1990, but no AIDS-defining opportunistic infection. Recipient B also had a progressive decrease in CD4 and CD8 counts, and an increase to moderate levels in viral load (Fig. 1). Cryptococcal meningitis was diagnosed in June 1990 despite zidovudine treatment since 1989, and death from AIDS occurred on 14 February 1991. Recipient B's sexual contact (SC) entered the study in August 1988 with a CD4 count of 520 × 106/l, and follow-up until 30 January 1990 showed no evidence of clinical or immunological progression and no antiretroviral medication.
HIV quasispecies transmitted by transfusion
The V3 sequences of 10 randomly-sampled genomes that were present in plasma virions from the donated blood unit which infected the two recipients in 1985 were examined (D85 sequences; Fig. 2a). The sequences were very homogenous and were close to the North American/European V3-loop consensus sequence . This is consistent with the donor being at an early asymptomatic stage of infection [15,19]. These homogenous V3 sequences represent the majority of the cell-free HIV genomes present in the donor's blood at the time of donation, and provide a window into the quasispecies distribution from which HIV sequences evolved over time in the donor, the two recipients, and recipient B's sexual partner.
Amino-acid substitutions in HIV quasispecies over time
V3 regions of randomly-selected plasma viral genomes present in donor, the two transfusion recipients (A and B), and the sexual contact (SC) of recipient B, were sequenced from samples obtained at various timepoints from 1986 through 1993 to compare the patterns of V3 sequence variation that would develop in each of the four members of this cluster. Fig. 2 shows the progression of predicted amino-acid sequences over time in the four individuals using the donation consensus sequence as reference.
An overall view of these data shows a pattern of individual-specific divergence over time. Most sequences, especially later ones, show changes from the donation consensus sequence that are shared within that individual and differ from the other individuals in the cluster. With respect to the immunodominant tip of the V3 loop, the GPGR motif in the donation sequences remains unchanged in recipient B, and remains predominant over time in individual SC. However, GPGG almost completely replaces GPGR in RA sequences from 1990 and 1991. In the donor, a unique subset of sequences (designated D2 sequences), showing first a change from GPGR to GSGR followed by a further change to GSGK, appears and comes to represent the majority sequence several years after donation, only to disappear in the last sample analysed. The changes from Pro to Ser and from Arg to Lys both represent single base-pair changes.
Despite the general pattern of individual-specific divergence with respect to the V3 region of these genomes viewed as a whole, the appearance of new amino acids that are shared between the individuals at several sites suggests parallel or convergent evolution at these sites. For instance, with respect to the immunodominant tip of the V3 loop, GPGK-containing sequences crop up as minority sequences in all individuals except recipient B. GPGG, which becomes predominant in recipient A, is also found in one 1987 RB sequence. Fig. 2 shows other sites both within and flanking the V3 loop that have amino-acid changes that appear in common in more than one individual. These could be a result of true convergent evolution resulting from multiple independent mutations. Convergent evolution results from limitations on tolerated substitutions introduced by high mutation rates. Alternatively, these shared amino acids could be due to re-emergence of amino acids that were already present in minority sequences in the donor which were transmitted and reappeared in the various members of the cluster. Analysis of codon usage suggests that at least in some instances true convergent evolution took place, in that some shared amino acids showed different codon usage.
The individuals show the emergence of sequences with changes in the net positive charge of the V3 loop. A high net positive charge in the V3 loop is thought to be associated with the syncytium-inducing (SI) pheno- type . Recipient B, who died from AIDS, is seen to have a V3-loop sequence with a net positive charge of 5 as early as 1986, and with a positive charge of 6 by 1987. The only other member of the group to develop V3-loop sequences with these degrees of net positive charge is the donor, and only in the 1991 and 1993 samples. Sequence similarities are not evident when these positively-charged V3 loops are compared between the donor and recipient B. High positive net charges were not found in RA sequences, although recipient A progressed in his disease. It has been shown that positively charged amino acids at specific positions of the V3 loop (Arg and/or Lys at position 306 or 320) are associated with the SI phenotype [21,22]. Both the donor and recipient B show a minority of sequences with a Glu to Lys or Arg change at position 320. It has been shown previously that disease progression can occur, sometimes rapidly, in patients from whom SI virus is not detected .
Phylogenetic tree of sequences obtained from the cluster over time
A phylogenetic tree was constructed using the nucleotide sequences obtained from the donor's transfused blood unit, and from the four individuals at various timepoints spanning 1986 through 1993. The tree is shown in Fig. 3. It can be seen that the donation sequences (D85) and the early samples, especially some D86 and RA86 sequences, are located near the centre of the tree, reflecting the similarity of these early sequences. Sequences obtained from later specimens from the donor as well as the other members of the cluster are seen to be located further from the centre of the tree. This is consistent with derivation and evolution of the later sequences in all four individuals from sequences detected in the donor's donation plasma (sample D85). Sequences from each of the individuals in the cluster are to a large measure segregated on separate branches of the tree, as opposed to sequences from the various individuals being distributed and shared over many branches. This shows that shared early sequences evolved and diverged differently over time in each individual. It is worthy of note that the sequences of recipient B's sexual partner (SC) seem to originate in the tree from donation sequences rather than from RB sequences; even the earliest SC sequences analysed (dating from 1988) appear closer to the donation sequences of 1985 than to RB sequences dating from November 1986.
gag-p17 sequences, with a small portion of p24
Two subsets of gag sequences were detected in the cluster (Fig. 4). It is likely that both were present and transmitted by the donor in 1985, although only one of the subsets is detected in all four gag sequences from the donation sample D85.
In this study, we have analysed the evolution of the hypervariable V3 region of a shared swarm, or quasi- species, of HIV-1 that had infected four individuals, and find a general pattern of divergence within and between the infected individuals. The great majority of variation in the HIV quasispecies that occurred over time did so in an individual-specific manner, with resulting divergence from shared early genomes. This agrees with previous studies on epidemiologic clusters [3,4], and reiterates the pliability of this region of the viral genome. It also emphasizes the role of physical isolation and replication in distinct environments (i.e., the different hosts) in the evolution of HIV-1 over time.
The cluster under study here differs from previous studies in that it involved sexual as well as parenteral transmission. Transmission as a result of transfusion of HIV-containing blood components challenges the recipient with a large inoculum of virus, and hence we can be confident that both recipients were most likely to be exposed to identical representatives of the viral quasispecies present in the donor at the time of donation. The viral quasispecies distribution in the serum of the donated blood also appears to be homogenous, further strengthening the argument that the two recipients were inoculated with very similar HIV genomes. Our results clearly demonstrate that in the recipients the subsequent evolution of the quasispecies was individual-specific, presumably reflecting distinct immune pressures in the different hosts.
With respect to the sexual transmission, the first sample analysed from recipient B's sexual contact (SC), dating from August 1988, appears in the phylogenetic analysis to be more closely related to donations sequences from 1985 than to recipient B's sequences dating from November 1986. The most likely explanation for this finding is that sexual transmission occurred soon after transfusion in January of 1985, when recipient B's HIV quasispecies were still undiverged from the donation sequences. Alternatively, the sexual transmission occurred at a somewhat later time when the recipient B sequences had already started to diverge away from the donation sequences, but perhaps undiverged donation genomes still present in recipient B were preferentially transmitted to recipient B's sexual contact [15,19,24]. The exact time when recipient B's sexual contact was infected is not known, although infection occurred sometime within the 2-year period after the January 1985 transfusion, since the sexual contact tested HIV-seropositive in November of 1986.
Perhaps surprisingly, our data show the absence of continual exchange of HIV genomes between the sexual couple over time after the initial sexual transmission, in that recipient B and her sexual partner occupy separate, rather than share, branches on the phylogenetic tree. The observation could have several possible explanations. First, perhaps the couple used condoms, or only rarely had sexual intercourse, after the infecting transfusion. Condom use for prevention of HIV transmission is unlikely: for the first 2 years after the initial infection by blood transfusion the couple did not know they were HIV-infected, because the diagnosis of HIV infection in both members of the couple was made only in November 1986, and after the diagnosis was made, they knew they were both infected. Another possible explanation is that these individuals' immune systems, primed by the pre-existing infections, were able to prevent new infections. Alternatively, ongoing exchange of virus might have occurred but the newly-infecting genomes were not able to achieve high levels of viremia, and hence were not detected in our limited sampling of circulating genomes. Again, pre-existing immunity might have damped the new infections, or alternatively, the newly infecting viruses could not compete with the more numerous pre-existing viruses for a perhaps limited number of susceptible target cells. Finally, the sequences detected in the two members of the sexual couple might represent continuously transmitted sequences that are selectively amplified in only one member of the couple, and are present below our detection limit in the other member of the couple.
A unique subset of sequences (D2 in the phylogenetic tree), containing the GSGR/GSGK motif at the V3- loop tip, appears in the donor in 1987, persists and becomes a major constituent of the cell-free quasi- species distribution from 1989 through 1991, and then is not detected at all in the eight plasma HIV genomes sequenced from the 1993 samples. The presence of two distinct sets of V3 sequences evolving in parallel within one individual has been seen before . It is unlikely that this observation is due to sample contamination. The four samples in question were collected and aliquoted separately at 1–2-year intervals over the space of 4 years. RNA extraction, reverse transcription, and PCR amplification were likewise not performed simultaneously or sequentially. Cross-contamination by PCR product of these four timepoints is also unlikely in that sequences with GSGR or GSGK have not been detected in any other samples ever analysed in our laboratory. The appearance and turnover of these two lineages suggest dynamic selection pressure favoring one lineage over the other at a given timepoint in the donor, and are consistent with the process of immune selection occurring in this infected individual.
Reagents were generously provided at no charge by Roche Molecular Systems, Inc. (Somerville, New Jersey, USA), through the courtesy of S. Herman. The authors thank M. Royz for her superb technical assistance, and B. Johnson for assisting with manuscript preparation.
1. Goodenow M, Huet Th, Saurin W, Kwok S, Sninsky J, Wain-Hobson S: HIV-1 isolates are rapidly evolving quasispecies: evidence for viral mixtures and preferred nucleotide substitutions
. J Acquir Immune Defic Syndr
2. Chackerian B, Morton WR, Overbaugh J: Persistence of simian immunodeficiency virus Mne variants upon transmission
. J Virol
3. Wolfs TFW, de Jong JJ, van den Berg H, Tijnagel JMGH, Krone WJA, Goudsmit J: 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
4. Balfe B, Simmonds P, Ludlam CA, Bishop JO, Leigh Brown AJ: Concurrent evolution of human immunodeficiency virus type 1 in patients infected from the same source: rate of sequence change and low frequency of inactivating mutations
. J Virol
5. Deacon NJ, Tsykin A, Solomon A, et al.
: Genomic structure of an attenuated quasi species of HIV-1 from a blood transfusion donor and recipients
6. Lin HJ, Myers LE, Yen-Lieberman B, et al.
: Multicenter evaluation of quantification methods for plasma human immunodeficiency virus type 1 RNA
. J Infect Dis
7. Donegan E, Stuart M, Niland JC, et al.
: Infection with human immunodeficiency virus type 1 (HIV-1) among recipients of antibody-positive blood donations
. Ann Intern Med
8. Busch MP, Operskalski EA, Mosley JW, et al.
: Epidemiologic background and long-term course of disease in human immunodeficiency virus type 1-infected blood donors identified before routine laboratory screening
9. Kleinman SH, Niland JC, Azen SP, et al.
: Prevalence of antibodies to human immunodeficiency virus type 1 among blood donors prior to screening: the Transfusion Safety Study/NHLBI Donor Repository
10. Operskalski EA, Stram DO, Lee H, et al.
: Human immunodeficiency virus type 1 infection: relationship of risk group and age to rate of progression to AIDS
. J Infect Dis
11. Mulder J, McKinney N, Christopherson C, Sninsky J, Greenfield L, Kwok S: A rapid and simple PCR assay for quantitation of HIV-1 RNA in plasma: application to acute retroviral infection
. J Clin Microbiol
12. Sabino E, Pan L-Z, Cheng-Mayer C, Mayer A: Comparison ofin vivoplasma and peripheral blood mononuclear cell HIV-1 quasispecies to short-term tissue culture isolates: an analysis oftatand C2-V3 envregions
13. Simmonds P, Balfe P, Peutherer JF, Ludlam CA, Bishop JO, Leigh Brown AJ: Human immunodeficiency virus-infected individuals contain provirus in small numbers of peripheral mononuclear cells and at low copy numbers
. J Virol
14. Simmonds P, Balfe P, Ludlam CA, Bishop JO, Leigh Brown AJ: Analysis of sequence diversity in hypervariable regions of the external glycoprotein of human immunodeficiency virus type 1
. J Virol
15. Zhang LQ, MacKenzie P, Cleland A, Holmes EC, Leigh Brown AJ, Simmonds P: Selection for specific sequences in the external envelope protein of human immunodeficiency virus type 1 upon primary infection
. J Virol
16. Felsenstein J: Phylogenies from molecular sequences: inference and reliabilities
. Annu Rev Genet
17. Kimura M: A simple model for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences
. J Mol Evol
18. LaRosa GJ, Davide JP, Weinhold K, et al.
: Conserved sequence and structural elements in the HIV-1 principal neutralizing determinant
19. Zhu T, Mo H, Wang N, et al.
: Genotypic and phenotypic characterization of HIV-1 in patients with primary infection
20. Fouchier RAM, Groenink M, Kootstra NA, et al.
: Phenotype- associated sequence variation in the third variable domain of the human immunodeficiency virus type 1 gp120 molecule
. J Virol
21. de Jong JJ, Goudsmit J, Keulen W, et al.
: Human immunodeficiency virus type 1 clones chimeric for the envelope V3 domain differ in syncytium formation and replication capacity
. J Virol
22. de Jong JJ, de Ronde A, Keulen W, Tersmette M, Goudsmit J: Minimal requirement for the human immunodeficiency virus type 1 V3 domain to support the syncytium-inducing pheno- type: analysis by single amino acid substitution
. J Virol
23. Tersmette M, Gruters RA, de Wolf F, et al.
: Evidence for a role of virulent human immunodeficiency virus (HIV) variants in the pathogenesis of acquired immunodeficiency syndrome: studies on sequential HIV isolates
. J Virol
24. Cornelissen M, Mulder-Kampinga GM, Veenstra J, et al.
: Syncytium-inducing (SI) phenotype suppression at seroconversion after intramuscular inoculation of a non-syncytium-inducing/SI phenotypically mixed human immunodeficiency virus population
. J Virol
25. Holmes EC, Zhang LQ, Simmonds P, Ludlam CA, Leigh Brown AJ: 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
26. Myers G, Korber B, Berzofsky JA, Pavlakis GN (Eds): Human Retroviruses and AIDS
. Los Alamos: Los Alamos National Laboratory; 1995.