Recombination among genetically distinct subtypes and strains of HIV-1 is now recognized with increasing frequency in the global pandemic [1-11]. Recombination provides a mechanism to increase viral sequence diversity rapidly, unlike the slow accumulation of mutations that occurs through replication errors . For recombination to occur between distinct HIV-1 strains, a cell needs to be dually infected with different viruses . The progeny virions that result possess RNA genomes from each virus, permitting strand-switching to occur during the next round of reverse transcription . Therefore, recombination requires coinfection of viral strains in an individual. This dual infection may occur during the primary infection period, before the immune response is fully developed, or it may occur as a superinfection with a new viral strain after the initial strain has established a chronic infection. Both superinfection and recombination have the potential to complicate efforts to develop vaccines, and reinfection with a drug-resistant virus could jeopardize treatment .
Superinfection with a new strain may make recombination between the original and new strains possible. Although recombination following superinfection has been postulated, it has not been observed in individuals . Previous studies have identified fully formed recombinants in patients, but not recombination between the non-recombinant first and superinfecting strains occurring in the same individual [1–11]. Some individuals have been infected by recognized circulating recombinant forms, which were transmitted in primary infection [1–4,10,11]; in these cases, the recombination event presumably occurred in a previously infected patient [2,10]. A small number of unique recombinants have also been detected in the setting of primary infection [1–11]. With increasing recognition of recombinant strains globally, it was unclear whether recombination resulted from superinfection of chronically infected individuals. The distinct HIV-1 subtypes or clades seen in different regions of the world differ from one another by as much as 30% in the envelope gene (env) and up to 15% in the core proteins . Strains classified as different subtypes are genetically quite distinct, making it relatively easy to detect superinfection and recombination. To determine whether recombination has occurred at any site in a viral genome, it is best to analyze the complete genomic sequence of the virus. Detection of superinfection and recombination between strains of the same subtype can be challenging  because of the difficulty of distinguishing different viral species of the same subtype from variation within a strain as a result of evolution.
Instances of cross-clade and intraclade coinfection [7,8,16–20] and superinfection have been previously reported, particularly during and soon after primary infection and in the setting of treatment interruptions of chronic infections [15,21–24]. The determination of superinfection in two of these cases was aided by epidemiological and immunological investigation [22,23]. The possibility of dual transmission during the primary infection period, however, was not completely excluded by highly sensitive, virological analyses in these and other studies [11,14,21–23,25]. It has been unclear whether recombination occurs only in the primary infection period [1–10] or can result from superinfection during untreated, chronic infection.
The present study searches for evidence of superinfection and recombination in Africa, where multiple HIV-1 subtypes exist, by examining serial blood samples from women in Nairobi's Pumwani sex worker cohort. The dominant HIV-1 subtype in Kenya is clade A HIV-1, which accounts for ∼70% of HIV-1 infections, the remainder consisting of clades D (∼20%), C, and several minor species . The women in this cohort, like most infected individuals worldwide, have never received treatment and it was interesting to determine whether such seropositive, chronically infected individuals are susceptible to superinfection with a new strain and whether the superinfection could lead to recombination. Highly exposed persons, such as the women in the Pumwani cohort, may be particularly vulnerable to reinfection because their partners expose them to a range of different viral strains; this makes recombination more likely to be detected in this group. Full-length HIV-1 RNA sequences from seven long-term survivors have been analyzed for the appearance of new subtypes and recombinants.
Clinical study subjects
The subjects studied were women enrolled in the Pumwani sex worker cohort. This cohort was established in 1985 by the Universities of Nairobi and Manitoba to study HIV-1 and other sexually transmitted diseases. The cohort currently comprises over 1850 women who are commercial sex workers in a Nairobi slum area. They have never been treated with antiretroviral drugs. The research program was approved by the National Ethical and Scientific Review Committee of Kenyatta National Hospital, the University of Manitoba Use of Human Subjects in Research Review Committee, and the NY State Department of Health Institutional Review Board.
Within the cohort, there are several subgroups of HIV-1-infected individuals, including rapid progressors, long-term non-progressors, and long-term survivors. Long-term survivors are defined as surviving with HIV-1 infection for at least 10 years.
The women were evaluated clinically at approximately 6-month intervals. Lymphocyte subsets were determined using a FACscan flow cytometer (Becton Dickson, San Jose, California, USA). Plasma viral loads were determined using Amplicor assay (Roche Molecular Systems, Branchburg, New Jersey, USA).
HIV-1 virus isolation and amplification
HIV-1 virions were isolated from plasma. The plasma was clarified by centrifugation for 10 min at 400 × g at room temperature and then transferred to a sterile, siliconized 1.5 ml microcentrifuge tube; virions were collected by centrifugation at 18 500 × g for 90 min at 10°C. Virion-associated RNA was extracted from the virus using reagents from RNAgents Total RNA Isolation System Kit (Promega, Madison, Wisconsin, USA).
The HIV-1 RNA genome was amplified by long reverse transcription (RT) polymerase chain reaction (PCR) [26–29]. The RT reaction was carried out in a 20 μl mixture containing extracted HIV-1 viral RNA (5 μl), 1 U RNase inhibitor (PerkinELMER, Boston, Massachusetts, USA), 1.0 mmol/l of each deoxynucleoside trisphosphate (dNTP), 2.5 U (RNase H−) murine leukemia virus reverse transcriptase (PerkinELMER), 0.1 μg bovine serum albumin, 10 mmol/l Tris-HCL (pH 8.3), 8 mmol/l MgCl2, 50 mmol/l KCl, 10 mmol/l DTT, and 2.5 μmol/l oligo (dT) primer. The reaction mixture was incubated for 60 min at 42°C and then at 95°C for 5 min.
The reverse transcribed HIV-1 RNA genome was amplified by long RT PCR using rTth XL DNA polymerase (PerkinELMER). Specific primers were used with primer sequences listed in Table 1. When the HIV-1 RNA load is high, the entire genome can be amplified in two fragments. The 3′ part of the genome (∼ 5 kb) was amplified using F4650 and R9626; the 5′ part (4.5 kb) was amplified using AF2 and R5090 primers. For samples with moderate or low viral titer, genomes were amplified as four overlapping fragments by using nested PCR. The primers for the first PCR were F4650 and R9626; for the second PCR, primers F4956 and R9173 or F7695 and R9626 were employed (Table 1). The middle part of the viral genome (3 kb) was amplified using primers F2018 and R5220 for the first PCR and F2042 and R5090 for the second PCR. The 5′ end of the genome (1.5 kb) was amplified by using nested PCR, with primers AF 2 and AR 1603 employed for the first PCR and AF 6 and AR 1539 for the second PCR. Each cycle of PCR consisted of a 15 s denaturation step at 94°C and an annealing and elongation step of 3–4.5 min (1 min/kb) at 60°C. After 40 PCR cycles were completed, the samples were extended for an additional 10 min at 72°C to ensure that the PCR products were full length. The PCR products were checked for size using 0.8% SeaKem GTG agarose gel (FMC, Rockland, Maine, USA) and ethidium bromide staining.
Although effective in amplifying molecules of RNA, RT-PCR carries potential pitfalls of recombination between molecules and underestimates sequence diversity. To decrease the chance of underestimating diversity, multiple RT reactions were performed on limiting dilutions of RNA, followed by multiple PCR on complimentary DNA (cDNA) from each RT reaction. The DNA was sequenced from the dilution yielding PCR products in only half of the reactions. To minimize the potential for PCR-related recombination during long RT-PCR, the reaction conditions were controlled using the methods described in our previous studies [26–29].
Sequencing and sequence analysis of HIV-1 genome
After verification of fragment size by gel electrophoresis, the PCR products were purified using PCR Select Spin Columns (5′Prime to 3′Prime, Inc. Boulder, Colorado, USA). The purified DNA fragments were then quantified and analyzed directly by DNA sequencing. PCR fragments were fully sequenced on both strands by using fluorescent dye terminators and an Applied Biosystems DNA sequencer (Applied Biosystems, Foster City, California, USA). Long overlapping DNA sequences were obtained and those with the highest degree of similarity in the overlaps were assembled using Sequencher (Ann Arbor, Michigan, USA) and BioEdit programs. Sequence ambiguities were checked and resolved manually.
Full-length HIV-1 sequences derived from plasma viral RNA were aligned and subjected to computational analysis. A BLAST search was performed on all sequences and did not show any undue similarity to other strains . Phylogenetic trees were constructed using neighbor-joining trees based on Kimura's two-parametric distance estimates. HIV-1 subtypes were determined by using phylogenetic analysis, distance plotting, and bootscanning; the latter two with the help of SimPlot . Recombination was determined by distance plotting, followed by bootscanning and tree construction using the recombinant fragments.
Heteroduplex tracking assay
The frequencies of genetic HIV-1 variants isolated from ML170, the patient under study, at each time point were compared using a non-radioactive adaptation of the DNA heteroduplex tracking assay . To amplify a 288 bp fragment spanning the V1–V2 coding region of env, RT-PCR was performed using the following PCR primer pair: 5′-ATGGGATCAAA GCCTAAAGCCATGTG-3′ (6556F) and 5′-CTTTG GACAGGCCTGTGTAATGGGTGAGG-3′ (6844R). Each RT-PCR was repeated to verify representative sampling of the RNA templates. Viruses from two additional patients enrolled in the Nairobi sex worker cohort were included as local clade-specific controls. Their env sequences and subtypes were determined as described above. The virus env sequences are clade A and clade C specific and were given the names Nairobi A and Nairobi C, respectively.
A fluorescent probe was generated by PCR of cloned HIV-1 DNA derived from ML170 at each time point (1986, 1995, 1997). Cloned HIV-1 was used to ensure that each probe represented a single uniform viral variant; for each time point, nucleotide sequences were determined, compared with previously obtained sequence from patient ML170, and the clade was ascertained genotypically. For the PCR reaction, a 5′-fluorescein-labelled forward primer (6556F; Invitrogen, Carlsbad, California, USA) and an unlabelled reverse primer (6844R) were used. Heteroduplex formation was carried out using fluorescein-labelled probe mixed with a 10-fold excess of unlabelled target DNA. The resulting DNA heteroduplex molecules were electrophoresed through 10% non-denaturing polyacrylamide gels, which were scanned on a FluorImager 595 (Molecular Dynamics, Piscataway, New Jersey, USA).
Seven subjects in the Nairobi female sex worker cohort were chosen for study because they were long-term survivors, infected at least 10 years, and had detectable HIV-1 RNA plasma viral loads ranging from 2300 to 1 430 000 copies/ml. Because the median time from HIV-1 infection to development of AIDS in this cohort of untreated female sex workers is 4.4 years , survival of more than 10 years is remarkable. Complete HIV-1 genomic sequences were determined from plasma viral RNA from all seven subjects. Three patients demonstrated HIV-1 genomes in 1997 that were intersubtype recombinants. For two of these three, the same recombinants were detected in 1986. One of these subjects, however, ML170, had evidence of recombination and superinfection.
Subject ML170 enrolled in the Pumwani female sex worker cohort in Nairobi, Kenya in 1985, when she was 22 years old and had already been a prostitute for 4 years. At the time of enrollment, February 1985, ML170 was found to be HIV-1 seropositive. She was followed at the Pumwani sex worker clinic at least every 6 months, where she reported having an average of two to three clients per day and using condoms less than 5% of the time. Most of her clinic visits were occasioned by symptoms of sexually transmitted infections, including pelvic inflammatory disease, genital herpes, and genital warts. In April, 1992, she had an acute febrile illness with chills, the only documented febrile illness in her medical records from 1985 to 1997. Her CD4 lymphocyte count plummeted from 794 × 106 cells/l in March 1991 to 136 × 106 cells/l in April 1992, the time of her acute febrile illness, and it remained < 100 × 106 cells/l from October 1992 until 1998. The HIV-1 RNA load in plasma rose from 2300 copies/ml in 1986 to 87 000 in 1997 (Fig. 1). After maintaining a stable weight of 65 kg annually from 1988 to 1993, she began to lose weight in 1994 and continued to lose yearly until her death. Her genotypes for CCR5 and CCR2 were wild type. ML170 never received antiretroviral therapy and she died in 1998 from tuberculosis and cryptococcal infection complicating her HIV-1 infection.
Complete HIV-1 RNA sequences were obtained from subject ML170's plasma in November 1986, when she had been infected for at least 21 months, and in 1995 and 1997 (GenBank accession numbers AF539404–6). Full-length viral genomes were sequenced directly, then aligned and subjected to computational analysis. A BLAST search revealed no evidence of contamination. Recombination was determined by computational analyses of the serial sequences. The entire genome obtained in 1986 was determined to be subtype A, shown in the bootscan in Fig. 2a. In 1995 and 1997, however, the sequences were subtype A/C recombinants (Fig. 2b). Comparison of full-length sequences from 1986, 1995, and 1997 with sequences from the Los Alamos HIV Database using SimPlots demonstrated that the A regions of the 1995 and 1997 A/C recombinants were derived from the subtype A sequence present in 1986 in the same individual (Fig. 3).
To determine whether subtype C sequences were present as minor species in 1986, heteroduplex tracking assays were performed. These assays are capable of detecting rare recombinant species making up as little as 0.2% of the viral population , a degree of sensitivity empirically confirmed in our own laboratory (data not shown). DNA fragments from ML170's env genes obtained in 1986, 1995, and 1997 were annealed with subtype-specific env probes. The subtype C sequences were not detectable in the 1986 sample (Fig. 4).
These data confirm the interpretation that this chronically infected woman was superinfected with a new subtype C virus at some time after 1986, with subsequent recombination between her subtype A and C strains.
With frequent recombination between distinct viral strains now recognized in the HIV-1 epidemic [2,11] and reports of superinfection in the literature [21–24], it has been hypothesized that recombination is likely to occur following superinfection . Recombination, however, was not found in the cases of superinfection recently reported [21–24]. It has been unclear whether the viral recombination now recognized globally arises only during primary infection or can result from superinfection of an untreated chronically infected individual.
We observed HIV-1 recombination following superinfection of a highly exposed woman from Nairobi, Kenya. The woman was chronically infected at the time of superinfection and had never received antiretroviral therapy. We examined blood plasma using complete viral genomic sequencing, computational and phylogenetic analyses including SimPlots, and heteroduplex tracking assays, a highly sensitive technique to detect minor species.
We used these techniques to the limit of detection to document superinfection with resulting recombination between the original 1986 HIV-1 subtype A and the superinfecting subtype C strains. The subtype A/C recombinant replaced the 1986 subtype A virus of this patient. Other studies in the literature have also documented replacement of one strain by another in coinfected patients [22,23]. Because of an insufficient number of useful specimens, we were unable to detect the superinfecting subtype C virus prior to recombination. Similarly, we were unable to study ML170's immunological responses or viral loads in greater detail because of a lack of samples. Although our data are highly suggestive of superinfection, ultimate proof of superinfection, as opposed to dual infection, has not been obtained in this or any of the previous reports.
This report adds to the growing literature documenting superinfection [21–24]. It is unclear what effect, if any, superinfection may have on disease progression. In the case described here, superinfection may have been associated with the abrupt decline in the CD4 T cell count, which did not rise again. Considering ML170's frequent HIV-1 exposure, the rapid decline in her CD4 T cell count between 1991 and 1992, and her steady weight loss beginning in 1994, it is possible that superinfection occurred during the period of her febrile illness in 1992. Superinfection may be associated with immunodeficiency and disease progression, particularly if it is a consequence of an escape from immune control . It is unclear whether superinfection may result from the loss of immunity, lead to immune deficiency, or both. In addition, superinfection could result in the acquisition of drug-resistant HIV-1 .
The frequency of superinfection and recombination has not been studied in a systematic way and needs to be investigated in a variety of populations. Cases of superinfection and recombination may escape detection when the circulating strains in a community are all of the same clade and the analyses focus on fragments of the viral genome rather than the complete HIV-1 sequence. Highly exposed individuals may be at higher risk of superinfection and resulting recombination than those with limited exposures. It will be of interest to determine the frequency of superinfection in highly exposed populations to address whether those individuals who do not become reinfected are resistant to superinfection. Such studies may help to evaluate the correlates of protection.
The case described here demonstrates that superinfection during chronic infection may play a significant role in the formation of recombinants between HIV-1 strains. This finding helps to explain the rising prevalence of recombinant HIV-1 in the world [2,11,12]. Furthermore, it illustrates that chronic infection with one strain may not provide protection against challenge from another. Recombination resulting from superinfection with diverse strains may pose problems for eliciting the broad immune responses necessary for an effective vaccine.
We thank Kimdar Sherefa Kemal for helpful discussions, Chih-Hsiung Chen for technical assistance, and Suzanne Beck for manuscript preparation. We thank Tim Moran, Matt Shudt, and the Wadsworth Center Molecular Genetics Core for oligonucleotide synthesis and DNA sequence analysis.
Sponsorship: This project was funded by the MRC and NIH grant RO1-AI42555.
1. Gao F, Robertson DL, Carruthers CD, Morrison SG, Jian B, Chen Y, et al
. A comprehensive panel of near full length clones and reference sequences for non-subtype B isolates of human immunodeficiency virus type 1. J Virol
2. Kuiken C, Foley B, Hahn B, Korber, B, McCutchan F, Marx P, et al
. HIV Sequence Compendium 2000.
Los Alamos, NM: Los Alamos National Laboratory; 2000.
3. Dowling WE, Kim B, Mason CJ, Wasunna KM, Alam U, Elson L, et al
. Forty-one near full-length HIV-1 sequences from Kenya reveal an epidemic of subtype A and A-containing recombinants. AIDS
4. Sabino EC, Shpaer EG, Morgado MG, Korber BT, Diaz RS, Bongertz V, et al
. Identification of human immunodeficiency virus type 1 envelope genes recombinant between subtypes B and F in two epidemiologically linked individuals from Brazil. J Virol
5. Lole KS, Bollinger RC, Paranjape RS, Gadkari D, Kulkarni SS, Novak NG, et al
. Full-length human immunodeficiency virus type 1 genomes from subtype C-infected seroconverters in India, with evidence of intersubtype recombination. J Virol
6. Robertson DL, Sharp PM, McCutchan FE, Hahn BH. Recombination in HIV-1. Nature
7. Salminen MO, Carr JK, Robertson DL, Hegerich P, Gotte D, Koch C, et al
. Evolution and probable transmission of intersubtype recombinant human immunodeficiency virus type 1 in a Zambian couple. J Virol
8. Diaz RS, Sabino EC, Mayer A, Mosley JW, Busch MP. Dual human immunodeficiency virus type 1 infection and recombination in a dually exposed transfusion recipient. The Transfusion Safety Study Group. J Virol
9. Kampinga GA, Simonon A, van de Perre P, Karita E, Msellati P, Goudsmit J. Primary infections with HIV-1 of women and their offspring in Rwanda: findings of heterogeneity at seroconversion, coinfection, and recombinants of HIV-1 subtypes A and C. Virology
10. Thomson MM, Perez-Alvarez L, Najera R. Molecular epidemiology of HIV-1 genetic forms and its significance for vaccine development and therapy. Lancet Infect Dis
11. Najera R, Delgado E, Perez-Alvarez L, Thomas MM. Genetic recombination and its role in the development of the HIV-1 pandemic. AIDS
12. Malim MH, Emerman M. HIV-1 sequence variation: drift, shift, and attenuation. Cell
13. Jung A, Maier R, Vartanian JP, Bocharov G, Jung V, Fischer U, et al
. Recombination: multiply infected spleen cells in HIV patients. Nature
14. Levy JA. Is HIV superinfection worrisome? Lancet
15. Goulder PJ, Walker BD. HIV-1 superinfection-a word of caution. N Engl J Med
16. Zhu T, Wang N, Carr A, Wolinsky S, Ho DD. Evidence for coinfection by multiple strains of human immunodeficiency virus type 1 subtype B in an acute seroconvertor. J Virol
17. Artenstein AW, VanCott TC, Mascola JR, Carr JK, Hegerich PA, Gaywee J, et al
. Dual infection with human immunodeficiency virus type 1 of distinct envelope subtypes in humans. J Infect Dis
18. Sala M, Pelletier E, Wain-Hobson S. HIV-1 gp120 sequences from a doubly infected drug user. AIDS Res Hum Retroviruses
19. Janini LM, Tanuri A, Schechter M, Peralta JM, Vicente AC, Dela Torre N, et al
. Horizontal and vertical transmission of human immunodeficiency virus type 1 dual infections caused by viruses of subtypes B and C. J Infect Dis
20. Iversen AK, Learn GH, Fugger L, Gerstoft J, Mullins JI, Skinhoj P. Presence of multiple HIV subtypes and a high frequency of subtype chimeric viruses in heterosexually infected women. J Acquir Immune Defic Syndr
21. Ramos A, Hu DJ, Nguyen L, Phan KO, Vanichseni S, Promadej N, et al
. Intersubtype human immunodeficiency virus type 1 superinfection following seroconversion to primary infection in two injection drug users. J Virol
22. Altfeld M, Allen TM, Yu XG, Johnston MN, Agrawal D, Korber BT, et al
. HIV-1 superinfection despite broad CD8+ T-cell responses containing replication of the primary virus. Nature
23. Jost S, Bernard MC, Kaiser L, Yerly S, Hirschel B, Samri A, et al
. A patient with HIV-1 superinfection. N Engl J Med
24. Koelsch KK, Smith DM, Little SJ, Ignacio CC, Macaranas TA, Leigh Brown AJ, et al
. Clade B HIV-1 superinfection with wild-type virus after primary infection with drug-resistant clade B virus. AIDS
25. McMichael AJ, Rowland-Jones SL. HIV: bad news for stop-start therapy? Nature
26. Fang G, Weiser B, Visosky A, Moran T, Burger H. PCR-mediated recombination: a general method applied to construct chimeric infectious molecular clones of plasma-derived HIV-1 RNA. Nat Med
27. Fang G, Zhu G, Burger H, Keithly JS, Weiser B. Minimizing DNA recombination during long RT-PCR. J Virol Meth
28. Fang G, Weiser B, Visosky AA, Townsend L, Burger H. Molecular cloning of full-length HIV-1 genomes directly from plasma viral RNA. J Acquir Immune Defic Syndr Hum Retrovirol
29. Fang G, Burger H, Chappey C, Rowland-Jones SL, Visosky A, Chen CH, et al
. Analysis of transition from long-term nonprogressive to progressive infection identifies sequences that may attenuate HIV type 1. AIDS Res Hum Retroviruses
30. Delwart EL, Gordon CJ. Tracking changes in HIV-1 envelope quasispecies using DNA heteroduplex analysis. Methods
31. Anzala AO, Nagelkerke N, Bwayo JJ, Holton D, Moses S, Ngugi E, et al
. Rapid progression to disease in African sex workers with human immunodeficiency virus type 1 infection. J Infect Dis