Transmission of HIV-1 CTL Escape Variants Provides HLA-
Mismatched Recipients with a Survival Advantage
Denis R. Chopera1, Zenda Woodman1, Koleka Mlisana2, Mandla Mlotshwa3, Darren P. Martin1, Cathal
Seoighe1, Florette Treurnicht1, Debra Assis de Rosa3, Winston Hide4, Salim Abdool Karim2, Clive M.
Gray3, Carolyn Williamson1* and the CAPRISA 002 Study Team
1Institute of Infectious Diseases and Molecular Medicine, Division of Medical Virology, Faculty of Health Sciences, University of Cape Town, Cape Town, South Africa,
2Centre for the AIDS Programme of Research in South Africa, University of Kwa-Zulu Natal, Durban, South Africa, 3National Institute for Communicable Diseases,
Johannesburg, South Africa, 4South African National Bioinformatics Institute, University of the Western Cape, Cape Town, South Africa
One of the most important genetic factors known to affect the rate of disease progression in HIV-infected individuals is the
genotype at the Class I Human Leukocyte Antigen (HLA) locus, which determines the HIV peptides targeted by cytotoxic T-
lymphocytes (CTLs). Individuals with HLA-B*57 or B*5801 alleles, for example, target functionally important parts of the Gag
protein. Mutants that escape these CTL responses may have lower fitness than the wild-type and can be associated with
slower disease progression. Transmission of the escape variant to individuals without these HLA alleles is associated with
rapid reversion to wild-type. However, the question of whether infection with an escape mutant offers an advantage to
newly infected hosts has not been addressed. Here we investigate the relationship between the genotypes of transmitted
viruses and prognostic markers of disease progression and show that infection with HLA-B*57/B*5801 escape mutants is
associated with lower viral load and higher CD4+ counts.
Citation: Chopera DR, Woodman Z, Mlisana K, Mlotshwa M, Martin DP, et al. (2008) Transmission of HIV-1 CTL Escape Variants Provides HLA-Mismatched
Recipients with a Survival Advantage. PLoS Pathog 4(3): e1000033. doi:10.1371/journal.ppat.1000033
Editor: Richard A. Koup, National Institutes of Health-NIAID, United States of America
Received September 24, 2007; Accepted February 24, 2008; Published March 21, 2008
Copyright: ? 2008 Chopera et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was funded by National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health (NIH), US Department of Health and
Human Services Grant U19 A151794. CW is funded by the South African AIDS Vaccine Initiative; DPM is funded by the Wellcome Trust, Sydney Brenner and Harry
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
Avoidance of host anti-viral responses is a major factor
influencing the evolution of HIV genomes. Of particular impor-
tance for virus survival, and a major contributor to ongoing HIV-1
diversification worldwide, is continual escape from anti-HIV host
cytotoxic T lymphocyte (CTL) responses. CTLs can potentially
detect many small polypeptide epitope sequences encoded through-
out HIV genomes. Evasion of CTL responses involves mutations
within and around targeted epitopes that result in the peptide no
longer bindingto the Class I MHC grove, or non-recognition by the
CTL T cell receptor, or interference with peptide processing [1–5].
These so-called CTL escape mutations have been associated with
increased viral loads and more rapid disease progression [3,6–8].
However, mutations associated with CTL evasion can also incur
significant viral replicative fitness costs and some escape mutations
have therefore been associated with decreased viral loads. In the
macaque model, for example, in vitro replication of SIVmac239
variants carrying certain CTL escape mutations is impaired relative
to SIVmac239 without the mutations [9,10].Fitnesscosts associated
with CTL escape have also been demonstrated in HIV-1 infected
humans carrying either the B*57 or B*5801 HLA alleles. CTL
escape mutations that frequently arise in these individuals, such as
the T242N mutation in the Gag TW10 epitope and the A163X
(X=G, N, D or S) mutation in the KF11 epitope, have been found
to compromise viral replicative capacity [11,12].
Because of the fitness costs associated with CTL escape
mutations, specific HLA alleles backgrounds of HIV infected
individuals have an influence on rates of disease progression. For
example, HIV-infected individuals possessing the B*57, B*5801
and B*27 HLA-alleles tend to take significantly longer to progress
to AIDS than individuals without these alleles [13–16]. HIV-1
epitopes targeted by these HLA types occur within functionally
important protein domains and escape mutations in these domains
tend to decrease viral replicative fitness [9,11,12,17–19]. There-
fore decreased rates of disease progression in people carrying the
B*57, B*5801 and B*27 alleles is at least partially driven by HLA
associated virus attenuation.
From the virus’ perspective, the conflicting demands of
replicative fitness on one hand and immune evasion on the other,
are best illustrated by the high rates at which certain CTL escape
mutations revert to ancestral, presumably replicationally fitter,
states following transmission
[12,20,21]. Whenever CTL-escape mutations do not revert
following transmission to such hosts it is generally assumed either
that the fitness costs of the mutations are negligible [21,22], or that
the replicative fitness of escape mutants has been effectively
restored by compensatory mutations [11,12,17,21,23,24].
While much effort has been focused on demonstrating the
causal influence of host genetic features on reduced viral
replication and decreased rates of disease progression [7,25–28],
there are a few instances where viral genetic features alone have
PLoS Pathogens | www.plospathogens.org1 2008 | Volume 4 | Issue 3 | e1000033
been identified as the primary cause of slower disease progression.
For example, a study dealing with individuals infected with
contaminated blood from a common donor determined that long
term non-progression was due to transmitted HIV genomes
carrying deletions in nef and the terminal repeat region [29,30].
Here we describe the identification of two HIV-1 Gag
polymorphisms that are associated with low viral loads and high
CD4+ counts during both the acute and chronic phases of
infection. Although these polymorphisms have been previously
identified as attenuating CTL escape mutations in individuals
carrying either the HLA-B*57 or B*5801 alleles, we detect these
associations in a group of HLA-B*57/5801-negative individuals.
We propose that these ‘‘attenuating’’ polymorphisms probably
arose during virus passage through HLA-B*57/5801-positive
individuals and provide evidence that they have enabled better
control of viral replication for up to at least a year following
transmission to their current hosts. This is the first demonstration
that transmission of HIV variants carrying HLA-associated escape
mutations may also afford improved control of virus replication in
HLA-mismatched recipients. Our results suggest a dependency
between the rate of disease progression in the newly infected host
and the genotype of the individual from whom the virus was
HLA-B*27, -B*57 and -B*5801 alleles are associated with long
term non-progression [13–16], and participants with these HLAs
were therefore excluded from this investigation. Of an initial
twenty-four study participants enrolled, three were HLA B*5801
positive and none were HLA-B*27, -B*57 positive. The remaining
twenty-one HLA-B*27, -B*57 and -B*5801 negative individuals,
estimated to be between 22 and 62 days post HIV infection
(median=42 days) at their time of enrollment, were recruited and
followed-up for at least 12 months (Table S1). The median log
viral load and CD4+ count went from 4.71 copies.ml21(range
2.95 to 6.28) and 509 cells.ul21(range 255 to 1358), respectively,
at three months post-infection to 4.59 copies.ml21(range 2.60 to
6.09) and 367 cells.ul21(range 202 to 1030), at twelve months. For
each participant, complete gag genes were amplified and
sequenced from the earliest available HIV positive sample, and
samples taken at 3 and 6 months postinfection.
Gag amino acids 146 and 242 are associated with control
of virus replication
We tested for statistical association between polymorphic amino
acid positions and both viral load and CD4+ cell counts. This
identified amino acid polymorphisms at two sites in Gag, at HXB2
positions 146 (n=9) and 242 (n=6), that were associated with
higher than average CD4+ counts and lower than average viremia
(Table 1). Nine of the 21 study participants were infected with
viruses carrying the A146X (X=P, or S) polymorphism and the
viruses in six of these nine individuals also carried the T242N
mutation. Gag amino acid 146 is adjacent to the HLA-B*57/5801
restricted ISW9 epitope and the A146X polymorphism has
previously been identified as an epitope processing mutation
associated with CTL escape  (Table 1). Similarly, position 242
occurs in the immunodominant TW10 epitope and the T242N
polymorphism has also previously been associated with CTL
escape in HLA-B*57/5801 positive individuals . We, therefore
suspected that viruses carrying either one or both of these two
polymorphisms may have been CTL escape variants that had been
transmitted from HLA-B*57/5801 positive individuals. Whereas
in vitro studies have shown that the A146X mutation does not incur
a replicative fitness cost, the T242N is known to decrease viral
fitness [11,22]. There was only marginal statistical significance
(p=0.0733) for an association between the presence of both
mutations (T242N/A146X) (n=6) and lower viral loads, however
when three additional infections involving viruses carrying the
A146X mutation only were included in the analysis (n=9), the
association was strengthened (p=0.0275), suggesting that the
T242N mutation is not solely responsible for the association
T242N and A146X mutations are consistent with
transmission from B*57/B*5801 positive individuals
If viruses carrying the T242N and A146X mutations were
transmitted from either B*57 or B*5801 positive donors, we
hypothesised that selection should be evident at sites within
immunodominant B*57 and B*5801 specific epitopes. We
analyzed the three B*57 and B*5801 immunodominant epitopes,
TW10, (TSTLQEQIAW; HXB2 positions 241–249), ISW9
(ISPRTLNAW; HXB2 positions 147–155) and KF11 (KAFSPE-
VIPMF; HXB2 positions 162–172) for evidence of selection.
Comparing sequences from the 21 individuals to the subtype C
consensus sequence we calculated the proportions of non-
synonymous (i.e. amino acid-changing) nucleotide differences that
fell within or close to these epitopes (one flanking amino acid on
Table 1. The identification of sites associated with high CD4+
counts and low viral loads at 12 months post infection.
Mutationn Epitope +
- p-value +
A146X9ISW9544 348 0.01723.494.87 0.0275
T242N6 TW10 538390 0.0175 3.264.690.0733
a(+) Presence of mutation; (2) absence of mutation
Following infection with HIV, it is well established that a
person’s genetic makeup is a major determinant of how
quickly they will progress to AIDS. Particularly important is
the class I Human leukocyte antigen (HLA) gene that is
responsible for alerting the immune system to HIV’s
presence. One of the reasons our immune systems are
unable to beat HIV is that the virus can mutate to forms
that our HLA genes no longer recognise. However, some
people have versions of the HLA gene (for example HLA-
B*57 and HLA-B*5801) that are known to force HIV to
tolerate mutations that damage its ability to reproduce.
Slower HIV reproduction is thought to be one reason that
HLA-B*57 and HLA-B*5801 positive people progress to
AIDS more slowly than most other HIV infected persons.
We report here on a study of HLA-B*57 and HLA-B*5801
negative women in which better control of disease tended
to be associated with their being infected with viruses
carrying mutations that have been previously shown to
reduce replication. These mutations characterise viruses
found infecting HLA-B*57 and HLA-B*5801 positive peo-
ple. This indicates for the first time that HLA-B*57 or HLA-
B*5801 negative people that are infected by such
reproductively compromised viruses may also experience
better survival prospects.
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org2 2008 | Volume 4 | Issue 3 | e1000033
either side of the epitope was included to allow for possible epitope
processing escape mutations) using the SNAP program (www.hiv.
lanl.gov). This analysis indicated that non-synonymous differences
from the consensus subtype C sequences were more often
associated with B*57 and B*5801 immunodominant epitopes for
viruses with the A146X and/or T242N mutations than was the
case for viruses without these mutations (p=0.0010; Figure 1).
These results suggests that these sequences had experienced
greater selective pressure from the immune response around these
immunodominant epitopes and supports our hypothesis that the
women from which they were isolated were infected with CTL
escape mutants that had arisen in HLA B*57/B*5801 positive
This analysis also revealed additional evidence of selection by
B*57/B*5801 restriction. Apart from the potential immune
evasion mutations at Gag positions 242 and 146, three sequences
(CAP088, CAP228 and CAP255) also carried the well-studied
A163X mutation in the HLA B*57/B*5801 restricted KF11
epitope (Figure 2) [12,31,32]. Viruses carrying the T242N and
A146X escape mutants from two of the nine individuals (CAP045
and CAP061), also carried the H219Q compensatory mutation
that has been shown to partially restore replicative fitness losses
incurred by the T242N mutation [11,21,33].
The T242N and A146X mutations revert over time
The T242N escape mutation is rare in chronically infected
HLA-B*57/B*5801-negative individuals and has been known to
revert rapidly in these individuals upon transmission from HLA-
B*57/B*5801-positive donors . Following limiting cDNA
dilution and amplification, Gag sequences from the nine study
participants infected with T242N/A146X mutants were analyzed
over time to detect reversion mutations (Figure 2). Reversion of
the T242N mutation was observed between six and 24 months
post infection in five of the six (83%) individuals initially infected
with viruses carrying this mutation (Figure 2). Reversion of the
A146X mutation was only observed in two of the nine (22%)
individuals infected with viruses carrying this mutation. These
reversions were observed at 16 and 24 months post-infection.
Sequences from the two individuals infected with viruses carrying
the A163X mutation in the KF11 epitope did not show any
reversion of this mutation.
To investigate the proportion of T242N and wild-type variants
in the six participants infected with T242N mutants, bulk PCR
was performed and amplicons from three time points were cloned
and sequenced. Although there was complete replacement of the
escape mutation (N242) with the consensus amino acid (T242) in
three participants (CAP061, CAP085 and CAP200), in another
participant (CAP228) no reversion was observed (Figure 3). In the
remaining two participants (CAP088 and CAP225), a mixed viral
population consisting of both escape mutants and wild-type
variants were detected at the final time point assayed, indicating
that complete replacement of the escape mutant with the wild-type
variant had not occurred. In CAP225 at 14.2 months, the escape
mutation was detected in 8/14 clones (57%), the reversion
intermediate, S242 was identified in 4/14 clones (29%), and
T242 occurred in 2/14 clones (14%). In CAP088 the T242 wild-
type was the dominant population member at both 12.6 and
18.9 months post infection with only, 3/11 and 4/12 of the
sampled sequences at these respective timepoints displaying the
N242 polymorphism. Reversion of the A146X mutation was only
observed in one (CAP061) of the six individuals infected with
viruses carrying the T242N mutation. However, although the
wild-type A146 polymorphism was observed in 3/10 sampled
sequences at 11.9 months post infection, it was not detectable
amongst ten sequences sampled at 23.9 months post infection
(Figure 3). The transience of this reversion indicates that in this
participant at least, it may not have provided any substantial
Viral load and CD4+ count dynamics were plotted over time to
investigate whether reversion of T242N was associated with either
increased viral loads or decreased CD4+ counts (n=6 , Figure 4a
and b). Overall, there was no significant change in geometric mean
log viral loads or CD4+ counts between either 6–12 and 12–
18 months post-infection, or 12–18 and 18–24 months post
infection (p.0.2; Wilcoxon matched pairs test) (Table 2).
However, one of the six study participants (CAP085) had an
increase in log viral load of 1.05 and a corresponding decrease in
CD4+ count of 209 cells.ml21between 12–18 and 18–24 months.
The T242 reversion polymorphism was observed in this individual
at 6.8 months post infection suggesting that the loss of viral control
was not concomitant with reversion.
These results confirm earlier reports  that the T242N
mutations revert earlier than other HLA-B*57/B*5801 associated
escape mutations and that A146X and other escape mutations
persist as ‘‘footprints’’ of prior viral exposure to HLA-B*57/
B*5801 alleles. In addition, reversion of T242N mutations to the
wild-type consensus sequence does not have an obvious immediate
impact on viral load.
Phylogenetic clustering of A146X/T242N mutants
The 38% frequency of study participants infected with
A146X/T242N mutants is higher than the 16.5% combined
population frequency of HLA-B*57/B*5801 alleles . How-
ever, this higher frequency is not inconsistent with all of the
individuals infected with A146X/T242N mutants having re-
ceived these viruses from HLA-B*57/B*5801 positive individuals.
Given a sample of 21 individuals from this population we would
expect between two and eleven individuals, to be either
heterozygous or homozygous for at least one of these two alleles.
We nevertheless sought to test whether the observed A146X and
T242N mutations had all arisen independently. We analysed our
21 sequences together with 102 other sequences sampled from
the same population and within five years of those sampled in our
Figure 1. The accumulation of non-synonymous mutations
within the three B57/B5801-specific immunodominant epi-
topes. Comparison of the odds ratio of non-synonymous mutations
within TW10, ISW9 and KF11 between those B57/B5801-negative
participants infected with variants carrying the T242N and A146X
mutations and those not carrying the mutations.
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org3 2008 | Volume 4 | Issue 3 | e1000033
study . We constructed a maximum likelihood tree from these
sequences after discarding both nine potentially recombinant
sequences and codons corresponding to known HLA-B*57/
B*5801 associated escape mutations (Figure S1). Sequences
carrying HLA-B*57/B*5801 associated escape mutations clus-
tered significantly within this tree (p=0.0157) indicating that
there is a degree of epidemiological linkage amongst viruses
carrying these mutations. This result also reiterates the notion
that HLA-B*57/B*5801 associated mutations may persist for
extended periods within circulating viruses.
There is no enrichment of HLA types amongst study
Several other HLA alleles besides B*57 and B*5801 have also
been associated with improved viral control . It was possible that
the reduced viremia and increased CD4+ counts, apparently
associated with viruses carrying the A146X/T242N mutations,
may have in fact been due to individuals infected with these viruses
carrying HLA-alleles that are effective in controlling HIV.
Compared to the remainder of the study population, we found no
detectable enrichment of any alleles in the nine individuals infected
Figure 2. Sequence alignment of ISW9, KF11 and TW10 epitopes. Sequence changes of the 9 participants carrying the T242N and A146X
mutations in B57/B5801-specific immunodominant epitopes over time showing reversion to consensus sequence at position 242. One flanking amino
acid residue was included on either side of the epitope. The position of the H219Q compensatory mutation is indicated relative to the epitopes. MO.
PI refers to months post-infection.
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org4 2008 | Volume 4 | Issue 3 | e1000033
Figure 3. Proportion of HLA-B*57/5801-associated escape mutations at different time-points. Gag bulk PCR products were cloned and
sequenced at three available timepoints as indicated on the x-axis. A median of 12 sequences (range 7–20) were analyzed per participant per time-
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org5 2008 | Volume 4 | Issue 3 | e1000033
with viruses carrying A146X/T242N mutations (data not shown),
indicating that lower viremia and increased CD4+ counts were not
obviously associated with over-representation of anyone HLA allele.
There is no difference in magnitude or breadth of
responses to Gag between individuals infected with
The number of Gag peptides recognised by CD8+ T-cells in
ELIspot assays have been shown to be associated with viral control
in subtype C HIV-1 infection . To determine if the lower
viremia observed in the 9 individuals infected with viruses carrying
the putative transmitted CTL escape mutants was associated with
the strength or breadth of Gag responses, IFN-gamma ELIspot
responses were assessed at 9–15 weeks post-infection (Table 3).
There was no significant difference between the group of
individuals infected with the T242N/A146X mutants and that
infected with the wild-type virus with respect to (1) the number of
responders, (2) the magnitude of responses to Gag (as measured by
Figure 4. A) Viral load and B) CD4+ counts over time for the 6 participants infected with virus carrying the T242N mutation at
enrolment. The amino acid residue at position 242 is indicated for the time-points cloned and sequenced. Red squares indicate N242 and blue
squares indicate T242. Where there was more than one variant, the dominant residue is shown in upper case whereas lower case letters indicate the
Table 2. Changes in viral load and CD4+ counts for the 6 participants infected with virus carrying the T242N mutation at
D CD4+ count
6–12 and 12–18 months 12–18 and 18–24 months6–12 and 12–18 months 12–18 and 18–24 months
20.08 0.0911 77
Geometric means were calculated for viral loads and CD4+ counts between 6212 months, 12218 months and 18224 months post infection and the differences in the
geometric means over the three periods were tabulated.
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org6 2008 | Volume 4 | Issue 3 | e1000033
total sfu/106PBMC for each Gag pool) and (3) the breadth of
responses (as measured by the number of Gag pools recognized).
Four out of nine individuals infected with T242N/A146X mutants
had detectable responses to Gag with three recognizing single
peptides (CAP085, 225, 228) (Table 3) and the fourth individual
recognizing two peptides (CAP255). Of the 11 individuals infected
with the wild-type variant, five responded to Gag, with two
individuals recognising single peptides (CAP008, 084) and three
targeting two peptides (CAP210, 256, 257) (Table 3).
No responses were detected to Gag fragments carrying the
TW10 and ISW9 epitopes. Only one individual (CAP228) showed
a response to a peptide overlapping with the KF11 epitope. The
optimal epitope within this peptide restricted by the host HLA
(HLA-A*2601) is EVIPMFSAL and the observed KF11 A163S
escape mutation falls outside this epitope.
Individuals infected with T242N/A146X mutants have
Together these data support our hypothesis that the T242N and
A146X mutations detected in viruses sampled from HLA-B*57/
5801 negative individuals are genetic ‘footprints’ of prior passage
through HLA-B*57/B*5801 positive donors. Initial identification
of these sites was based on a naive scan of all Gag amino acid
polymorphisms to identify those associated with high CD4+ counts
and low viremia. The proposed mechanism whereby HLA-B*57/
5801 individuals achieve good control of HIV replication is
unclear, although targeting of the Gag TW10, ISW9 and KF11
epitopes is thought to contribute to this control [16,21,34].
Improved control of viral replication in such individuals is due, at
least in part, to the fitness costs incurred by the T242N escape
mutation in the TW10 epitope. We were therefore interested in
determining the specific association of these escape mutations with
viremia and CD4+ counts following their transmission to HLA-
B*57/B*5801 negative recipients.
Clinical data was available from all individuals at 62 days post
infection and we compared viral load and CD4+ dynamics up to
15 months post infection in the nine individuals infected with
T242N/A146X mutants to those of the rest of the cohort
(Figure 5a and b; Figure S2a and b). At all time points the mean
log viral load and CD4+ count was lower in the individuals
infected with the T242N/A146X mutants. We found that, relative
to the rest of the study participants, individuals infected with the
T242N/A146X mutants had significantly lower viral loads and
higher CD4+ counts at three months post-infection (median log
VL 4.53 vs. 5.09, p=0.0077 and median CD4+ count 652.0 vs.
460.0, p=0.0129), (Figure 6a and b). At 12 months post infection,
these individuals also had significantly lower viral loads and higher
CD4+ counts (median log VL 4.26 vs. 4.92, p=0.0275 and
median CD4+ count 499.0 vs. 322.5, p=0.0172), (Figure 6c and
d). This suggests that, in HLA-B*57/B*5801 negative individuals,
HIV-1 variants carrying the Gag T242N and A146X mutations
tend to be less pathogenic than those which do not carry the
mutations. The H219Q mutation is a compensatory mutation
reported to partially restore viral fitness . However, we
observed that the two individuals infected with H219Q mutant
viruses tended to have lower viral loads within the T242N/
A146X+group (Figure 6a and c).
Examination of Gag sequences from acutely infected HLA-
B*57/5801 negative women has revealed two polymorphisms,
A146X and T242N, that are associated with lower viral loads and
higher CD4+ counts in these woman up to a year post infection.
As both polymorphisms have been previously identified as HLA-
B*57/5801 immune evasion mutations we propose that they are
probably genetic footprints of prior virus passage through HLA-
B*57/5801 positive individuals. While HLA imprinting of
circulating HIV sequences is an established phenomenon [35–
38], our demonstration that such imprinting might enable better
control of virus replication following transmission to HLA
mismatched recipients is entirely novel.
While we do not provide definitive evidence that any of the
studied viruses has been directly transmitted from HLA-B*57/
Table 3. Interferon-gamma ELIspot responses for the 9/21 study participants who responded to Gag.
PID HLA Reactive peptideSFU x 106(9–15weeks)
A3002 A3002 B0801 B4501 Cw0701 Cw1601TGTEELRSLYNTVATLY
CAP225A0101 A3001 B4202 B8101 Cw0701 Cw1801
CAP228 A2301 A2601 B4403 B5101 Cw0303 Cw0701
CAP255 A0301 A8001 B0801 B1807 Cw0202 Cw0702 IAPGQMREPRGSDIA*
A0301 A8001 B0801 B1807 Cw0202 Cw0702
A2301 A2301 B0801 B1510 Cw0701 Cw1601GKKHYMLKHLVWASREL
A2902 A7401 B1503 B4407 Cw0210 Cw0701TGTEELRSLYNTVATLY
CAP210 A6802 A6802 B1510 B1510 Cw0304 Cw0304
A6802 A6802 B1510 B1510 Cw0304 Cw0304
CAP256 A2902 A6601 B1503 B5802 Cw0401 Cw0602
A2902 A6601 B1503 B5802 Cw0401 Cw0602
A2301 A2902 B4202 B4403 Cw1701 Cw1701GKKHYMLKHLVWASREL
A2301 A2902 B4202 B4403 Cw1701 Cw1701MREPRGSDIAGTTSTL*
In bold are the 4 individuals with the transmitted viruses carrying T242N/A146X mutations. Underlined are the known epitopes and their corresponding restricting HLA
*No predicted epitope within the peptide matches the participant’s HLA
aDerived from deconvoluting the pool matrix ELIspot response and subsequently confirming with single peptides
bDerived from deconvoluting the pool matrix ELIspot response only
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org7 2008 | Volume 4 | Issue 3 | e1000033
B*5801 positive individuals, we have detected a strong signal of
selection in the Gag HLA-B*57/B*5801 restricted epitope
sequences of viruses carrying the A146X and T242N polymor-
phisms. This is suggestive of at least some of the viruses having
been passaged through HLA-B*57/B*5801 positive individuals at
some time in the past. Our speculation that the A146X/T242N
mutants have been transmitted from either HLA-B*57 or HLA-
B*5801 positive individuals is also consistent with previous reports
dealing with the persistence of HLA-B*57/B*5801 associated
immune evasion mutations. Two Gag mutations, H219X (X=Q,
P or R) and the A146X polymorphism dealt with here, have been
previously identified as relatively stable HLA-B*57/B*5801
genetic imprints on Gag . These mutations appear to be
epistatically associated with the T242N mutation but unlike the
T242N mutation which reverts following transmission to HLA-
B*57/5801-negative hosts, the H219X and A146X mutations are
often maintained even in the absence of the selective forces exerted
by these alleles [12,21]. Whereas the H219X mutation partially
alleviates the fitness deficit incurred by the T242N mutation ,
the A146X mutation has not been associated with any significant
fitness loss . Importantly, we detected the H219X mutation in
two participants, one associated with a T242N mutation and the
other was associated with the A146X mutation. In the latter case
the presence of the H219X mutation suggests that this virus may
have descended from a T242N mutant.
The possibility that most, if not all, of the six women infected
with viruses carrying the T242N mutation were directly infected
by HLA-B*57/5801-positive individuals is additionally supported
by our observation that the mutation reverted in five of the
individuals during the study period. It is, however, more uncertain
whether viruses carrying the A146X mutation but not the T242N
mutation were transmitted directly from HLA-B*57/5801-positive
individuals. It cannot be discounted that the A146X mutation in
these viruses might be a persistent imprint of a more distant
passage through an HLA-B*57/5801 positive individual. While we
have detected reversions of this mutation, it has persisted in the
viruses infecting seven of the nine individuals studied here for
more than two years and is clearly more stable than the T242N
mutation. If the mutation had a reversion half-life of four or more
years (as is suggested by our data) it would not be surprising if
some of the A146X mutants we have studied had been serially
transmitted two or more times since they first arose. Our detection
of significant phylogenetic clustering of gag sequences carrying
HLA-B*57/5801 associated escape mutations supports the notion
that many of these mutations may persist for epidemiologically
significant time periods in HLA-B*57/5801 negative individuals.
Reversion of the T242N escape mutation did not result in a
concomitant increase in viral load. There is a relationship between
viral load during primary infection and viral load set-point 
and it is possible that these escape mutations provided a long term
benefit by reducing viremia during acute infection. It is also
possible that, amongst the viruses studied, reversion mutations in
gag were not sufficient on their own to fully restore fitness and that
there may have been other HLA-B*57/5801 associated escape
and compensatory mutations elsewhere in their genomes that
impacted on their fitness. Although there was only a marginal
correlation between viral load and infection with variants
harbouring both T242N and A146X mutations (p=0.0733), this
relationship was much stronger when individuals infected with
only the A146X mutation were included in the analysis
(p=0.0275). This supports the existence of a network of B*57/
5801 associated mutations that could contribute to viral control.
Brockman et. al.  recently reported several novel compensatory
mutations, associated with the T242N escape mutation, which
correlated with higher viral loads. It might have been expected
that the two individuals (CAP045 and CAP061) whose viruses had
compensatory mutations would have higher viral loads as
compared to those infected with the T242N/A146X-carrying
viruses [11,18]. Viral loads in these individuals were, in fact, lower.
While our data suggests that, within the first year of infection,
B*57/5801-negative individuals infected with viruses carrying
these escape mutations have lower viral loads, the long-term
impact on disease progression is unknown.
The T-cell responses determined in this study could not explain
the differences in viral loads observed for the individuals infected
with the escape mutants compared to those infected with the wild-
type variants. The number of individuals that responded to Gag
did not differ between the two groups and there were no significant
differences in the magnitude or breadth of Gag IFN-gamma
ELIspot responses between the two groups. However, it is likely
that we are underestimating the T-cell responses due to the use of
consensus based subtype C reagents compared to autologous
peptides. In addition, experimental limitations could also be a
contributing factor as the IFN-gamma ELIspot assay does not
detect all T-cell responses.
Figure 5. Mean and standard error of A) log viral loads and B)
CD4+ counts over a 15 month period for the 21 study
participants. The T242N/A146X+ participants are shown in red and
the T242N/a146X- participants are shown in blue.
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org8 2008 | Volume 4 | Issue 3 | e1000033
We have therefore demonstrated that, during the acute phase of
infection at least, individuals who are infected with viruses carrying
markers indicative of previous selection in HLA-B*57/5801
positive individuals experience both significantly lower viral loads
and higher CD4+ counts than individuals infected with viruses
without these markers. These lower initial viral loads and higher
CD4+ counts at the onset of infection may slow disease
progression in these individuals. The possibility that an interacting
network of attenuating mutations may be responsible for the lower
viral loads experienced by people infected with viruses passaged
through HLA-B*57/5801 positive individuals should be investi-
gated further as the existence of such networks could profoundly
influence our understanding of HIV pathogenesis. Current
opinion is that first generation CTL based vaccines are likely to
be only partially effective. Our study suggests that such vaccines
should contain epitopes where escape is associated with a fitness
cost to the virus as this might drive the attenuation of viruses in
individuals who become infected despite vaccination.
Participants in this study are part of the CAPRISA 002 cohort
investigating the role of viral and immunological factors in acute
and early HIV-1 infections. The cohort includes high risk HIV
negative women monitored monthly for recent HIV-1 infection
using two HIV-1 rapid antibody tests and PCR (Roche Amplicor
v1.5). HIV-1 infection was confirmed using an enzyme immuno-
assay (EIA) test. Women were enrolled in the present study within
3 months of infection from both the HIV negative cohort, and
other seroincidence cohorts in Durban, South Africa. The timing
of infection was estimated to be either at the midpoint between the
last HIV-1 negative test and the first antibody positive test, or as
14 days where individuals were PCR positive-antibody negative.
Samples were collected at enrolment, weekly for three weeks,
fortnightly until 3 months, monthly until a year and quarterly
thereafter. CD4+ T cells counts were assessed using a FACSCa-
Figure 6. Viral Load and CD4+ counts of study participants grouped according to the presence or absence of the T242N and/or
A146X mutations at enrolment. The 21 B57/B5801 negative individuals were grouped into those infected with viral strains comprising both the
TW10 escape mutation and the ISW9 processing escape mutation (n=9) and those that did not (n=12). The viral load and CD4+ counts at 3 and
12 months post-infection were compared between these two groups. HLA-B*5801 positive individuals were excluded from the analysis. m denotes
viral loads and CD4+ counts for the two individuals infected with viruses carrying the H219Q compensatory mutation.
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org9 2008 | Volume 4 | Issue 3 | e1000033
libur flow cytometer and viral loads were measured using a
COBAS AMPLICORTMHIV-1 Monitor Test v1.5 (Roche
Diagnostics). Plasma collected in EDTA was stored at 270uC
until use. Written informed consent was obtained from all
participants. This study received ethical approval from the
University of KwaZulu-Natal, University of the Witwatersrand
and University of Cape Town. All study participants in the cohort
who had reached 12 months postinfection were included in this
study, excluding 3 HLA-B*5801-positive individuals.
RT-PCR and viral sequencing
RNA isolated from plasma samples using the Magna-Pure
Compact Nucleic Extractor (Roche) was reverse transcribed using
the Invitrogen Thermoscript Reverse transcription kit (Invitrogen)
and the primer, Gag D reverse (59-AAT TCC TCC TAT CAT
TTT TGG-39; HXB pos 2382-2402). Limiting dilution nested
PCR was carried out by serial end-point dilution of the cDNA
. The first round PCR primers were Gag D forward (59-TCT
CTA GCA GTG GCG CCC G-39; HXB pos 626–644) and Gag
D reverse (59-AAT TCC TCC TAT CAT TTT TGG-39; HXB
pos 2382–2402). The second round PCR primers were Gag A
forward (59-CTC TCG ACG CAG GAC TCG GCT T-39; HXB
pos 683–704) and Gag C reverse (59-TCT TCT AAT ACT GTA
TCA TCT GC-39; HXB pos 2334–2356). PCR products were
either directly sequenced or cloned using the -T Easy vector
system (Promega). Sequencing was carried out using an ABI
PRISM dye terminator cycle-sequencing kit (Applied Biosysytems)
and the primers Gag A forward, Gag A reverse (59-ACA TGG
GTA TCA CTT CTG GGC T-39; HXB pos 1282–1303), Gag B
forward (59-CCA TAT CAC CTA GAA CTT TGA AT-39; HXB
pos 1226–1246), Gag B reverse (59-CTC CCT GAC ATG CTG
TCA TCA T-39; HXB pos 1825–1846), Gag C forward (59-CCT
TGT TGG TCC AAA ATG CGA-39; HXB pos 1748–1768) and
Gag C reverse for direct sequencing. For cloned sequences, only
the Gag B forward and Gag B reverse primers were used
generating p24 gag sequences. Sequences were assembled using
the CAPRISA Assembly Pipeline tool (http://tools.caprisa.org/)
and aligned using ClustalW (with default settings ).
High resolution (four digit) HLA typing was performed on all
participants. DNA was extracted from either PBMCs or
granulocytes using the Pel-Freez DNA Isolation kit (Pel-Freez).
HLA-A, -B and -C typing was performed by sequencing of exons
2, 3 and 4 using Atria AlleleSeqr kits (Abbott) and Assign-SBT 3.5
(Conexio Genomics). Any ambiguities resulting from either
polymorphisms outside the sequenced exons or identical hetero-
zygote combinations, were resolved using sequence-specific
IFN-c Elispot assay
PBMC were prepared and HIV-1 specific T cell responses were
quantified by gamma interferon (IFN-c) Elispot assay .
Synthetic overlapping peptides(15- to 18-mer peptides overlapping
by 10 amino acids) spanning the entire HIV-1 clade C Gag protein
corresponding to the HIV-1 consensus C were used in the assay. T
cell responses were derived using either a deconvoluted pool
matrix approach or confirmed using individual peptides. The
epitopes within peptides showing responses were predicted from
the published epitopes on the Los Alamos HIV database (www.
Phylogenetic and recombination analyses
Phylogenetic trees were constructed using the maximum
likelihood method implemented in PHYML  (100 full
maximum likelihood bootstrap replicates, HKY model+gamma
with four substitution rates and transition:transversion ratio
determined from the data). Seven different recombination analysis
methods implemented in RDP3  were used, with default
settings, to test for the presence of recombination amongst the 21
acute infection sequences and an additional 102 publicly available
gag sequences sampled from a matched cohort. Evidence of
phylogenetic clustering of viruses carrying particular Gag poly-
mphisms was assessed using a permutation test (with 10000
iterations) implemented in RDP3 that is similar to that described
in Poss et al. .
Wilcoxon rank-sum tests were used to identify amino acid sites
(encoded by the earliest gag sequences determined post infection)
that were associated with low viral loads and high CD4+ counts at
12 months postinfection. These tests compared the median viral
loads and CD4+ counts between groups of viruses with the
consensus or an alternative amino acid at each site independently
(without correction for multiple testing). Fisher’s exact test was
used to test each HLA allele for enrichment among individuals
with either the A146X or T242N mutations and to test for
associations between Gag ELIspot responses in the two groups
(with and without T242N/A146X mutations). Changes in viral
loads were tested using the Wilcoxon matched pairs test. Statistical
tests were implemented in the R statistical computing environment
 and GraphPad Prism 4.0 (GraphPad Software, Inc.).
Nucleotide sequence accession numbers
Sequence data are available from GenBank under accession
sequences sampled in Durban South Africa. Whereas blue symbols
represent sequences carrying nucleotide sequence polymorphisms
characteristic of immune evasion mutations that occur in HLA-
B*57/B*5801 positive individuals, red symbols indicate sequences
without these polymorphisms. Blue bars to the right of the figure
indicate clades in which sequences carrying the polymorphisms
predominate. Sequences denoted with triangles are those deter-
mined in this study. Whereas branches labeled with filled circles
have .50% bootstrap support, those labeled with open circles
have between 25 and 50% bootstrap support.
Found at: doi:10.1371/journal.ppat.1000033.s001 (0.13 MB TIF)
A) Viral load kinetics and B) CD4+ count kinetics
over a 15 month period. Red lines represent T242N/A146X+
individuals and blue lines represent T242N/A146X2 individuals.
Thick red and blue lines represent the medians for the T242N/
A146X+ and T242N/A146X2, respectively.
Found at: doi:10.1371/journal.ppat.1000033.s002 (1.72 MB TIF)
Maximum likelihood tree of HIV-1 subtype C gag
the study. Shown in bold are the 9 participants with the T242N/
A146X escape mutations at enrolment.
Found at: doi:10.1371/journal.ppat.1000033.s003 (0.09 MB
HLA alleles and viral loads for the 21 individuals in
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org10 2008 | Volume 4 | Issue 3 | e1000033
We thank the participants, clinical and laboratory staff at CAPRISA for the
specimens. The HIV sequences quality analysis and chromatogram
assemble were done by the CAPRISA Assemble Pipeline, which was
developed by Winston Hide, Adam Dawe, Allan Kamau, Ruby van
Rooyen, Alan Powell, Anelda Boardman and Heikki Lehvaslaiho at the
South African National Bioinformatics Institute, University of the Western
Cape, South Africa.
Conceived and designed the experiments: DC ZW KM SA CG CW.
Performed the experiments: DC MM FT DA. Analyzed the data: DC ZW
MM DM CS FT DA CG CW. Contributed reagents/materials/analysis
tools: KM DM CS WH CG. Wrote the paper: DC ZW DM CS SA CG
1. Phillips RE, Rowland-Jones S, Nixon DF, Gotch FM, Edwards JP, et al. (1991)
Human immunodeficiency virus genetic variation that can escape cytotoxic T
cell recognition. Nature 354(6353): 453–459.
2. Goulder P, Price D, Nowak M, Rowland-Jones S, Phillips R, et al. (1997) Co-
evolution of human immunodeficiency virus and cytotoxic T-lymphocyte
responses. Immunol Rev 159: 17–29.
3. Borrow P, Lewicki H, Wei X, Horwitz MS, Peffer N, et al. (1997) Antiviral
pressure exerted by HIV-1-specific cytotoxic T lymphocytes (CTLs) during
primary infection demonstrated by rapid selection of CTL escape virus. Nat
Med 3(2): 205–211.
4. Price DA, Goulder PJ, Klenerman P, Sewell AK, Easterbrook PJ, et al. (1997)
Positive selection of HIV-1 cytotoxic T lymphocyte escape variants during
primary infection. Proc Natl Acad Sci U S A 94(5): 1890–1895.
5. Smith SM (2004) HIV CTL escape: At what cost? Retrovirology 1: 8.
6. Koenig S, Conley AJ, Brewah YA, Jones GM, Leath S, et al. (1995) Transfer of
HIV-1-specific cytotoxic T lymphocytes to an AIDS patient leads to selection for
mutant HIV variants and subsequent disease progression. Nat Med 1(4):
7. Goulder PJ, Phillips RE, Colbert RA, McAdam S, Ogg G, et al. (1997) Late
escape from an immunodominant cytotoxic T-lymphocyte response associated
with progression to AIDS. Nat Med 3(2): 212–217.
8. Barouch DH, Letvin NL (2002) Viral evolution and challenges in the
development of HIV vaccines. Vaccine 20 Suppl 4: A66–8.
9. Friedrich TC, Dodds EJ, Yant LJ, Vojnov L, Rudersdorf R, et al. (2004)
Reversion of CTL escape-variant immunodeficiency viruses in vivo. Nat Med
10. Peyerl FW, Bazick HS, Newberg MH, Barouch DH, Sodroski J, et al. (2004)
Fitness costs limit viral escape from cytotoxic T lymphocytes at a structurally
constrained epitope. J Virol 78(24): 13901–13910.
11. Martinez-Picado J, Prado JG, Fry EE, Pfafferott K, Leslie A, et al. (2006) Fitness
cost of escape mutations in p24 gag in association with control of human
immunodeficiency virus type 1. J Virol 80(7): 3617–3623.
12. Crawford H, Prado JG, Leslie A, Hue S, Honeyborne I, et al. (2007)
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 81(15):
13. Goulder PJ, Bunce M, Krausa P, McIntyre K, Crowley S, et al. (1996) Novel,
cross-restricted, conserved, and immunodominant cytotoxic T lymphocyte
epitopes in slow progressors in HIV type 1 infection. AIDS Res Hum
Retroviruses 12(18): 1691–1698.
14. Kaslow RA, Carrington M, Apple R, Park L, Munoz A, et al. (1996) Influence of
combinations of human major histocompatibility complex genes on the course of
HIV-1 infection. Nat Med 2(4): 405–411.
15. Klein MR, van der Burg SH, Hovenkamp E, Holwerda AM, Drijfhout JW, et al.
(1998) Characterization of HLA-B57-restricted human immunodeficiency virus
type 1 gag- and RT-specific cytotoxic T lymphocyte responses. J Gen Virol 79
( Pt 9)(Pt 9): 2191–2201.
16. Kiepiela P, Ngumbela K, Thobakgale C, Ramduth D, Honeyborne I, et al.
(2007) CD8+ T-cell responses to different HIV proteins have discordant
associations with viral load. Nat Med 13(1): 46–53.
17. Schneidewind A, Brockman MA, Yang R, Adam RI, Li B, et al. (2007) Escape
from the dominant HLA-B27-restricted cytotoxic T-lymphocyte response in gag
is associated with a dramatic reduction in human immunodeficiency virus type 1
replication. J Virol 81(22): 12382–12393.
18. Brockman MA, Schneidewind A, Lahaie M, Schmidt A, Miura T,
et al. (2007) Escape and compensation from early HLA-B57-mediated
cytotoxic T-lymphocyte pressure on human immunodeficiency virus
type 1 gag alter capsid interactions with cyclophilin A. J Virol 81(22):
19. Peyerl FW, Barouch DH, Letvin NL (2004) Structural constraints on viral escape
from HIV- and SIV-specific cytotoxic T-lymphocytes. Viral Immunol 17(2):
20. Li B, Gladden AD, Altfeld M, Kaldor JM, Cooper DA, et al. (2007) Rapid
reversion of sequence polymorphisms dominates early human immunodeficiency
virus type 1 evolution. J Virol 81(1): 193–201.
21. Leslie AJ, Pfafferott KJ, Chetty P, Draenert R, Addo MM, et al. (2004) HIV
evolution: CTL escape mutation and reversion after transmission. Nat Med
22. Draenert R, Le Gall S, Pfafferott KJ, Leslie AJ, Chetty P, et al. (2004) Immune
selection for altered antigen processing leads to cytotoxic T lymphocyte escape in
chronic HIV-1 infection. J Exp Med 199(7): 905–915.
23. Kelleher AD, Long C, Holmes EC, Allen RL, Wilson J, et al. (2001)
Clustered mutations in HIV-1 gag are consistently required for escape from
HLA-B27-restricted cytotoxic T lymphocyte responses. J Exp Med 193(3):
24. Peyerl FW, Barouch DH, Yeh WW, Bazick HS, Kunstman J, et al. (2003)
Simian-human immunodeficiency virus escape from cytotoxic T-lymphocyte
recognition at a structurally constrained epitope. J Virol 77(23): 12572–12578.
25. Fellay J, Shianna KV, Ge D, Colombo S, Ledergerber B, et al. (2007) A whole-
genome association study of major determinants for host control of HIV-1.
Science 317(5840): 944–947.
26. Fang G, Burger H, Chappey C, Rowland-Jones S, Visosky A, et al. (2001)
Analysis of transition from long-term nonprogressive to progressive infection
identifies sequences that may attenuate HIV type 1. AIDS Res Hum
Retroviruses 17(15): 1395–1404.
27. Kirchhoff F, Greenough TC, Brettler DB, Sullivan JL, Desrosiers RC (1995)
Brief report: Absence of intact nef sequences in a long-term survivor with
nonprogressive HIV-1 infection. N Engl J Med 332(4): 228–232.
28. Alexander L, Weiskopf E, Greenough TC, Gaddis NC, Auerbach MR, et al.
(2000) Unusual polymorphisms in human immunodeficiency virus type 1
associated with nonprogressive infection. J Virol 74(9): 4361–4376.
29. Deacon NJ, Tsykin A, Solomon A, Smith K, Ludford-Menting M, et al. (1995)
Genomic structure of an attenuated quasi species of HIV-1 from a blood
transfusion donor and recipients. Science 270(5238): 988–991.
30. Churchill MJ, Rhodes DI, Learmont JC, Sullivan JS, Wesselingh SL, et al.
(2006) Longitudinal analysis of human immunodeficiency virus type 1 nef/long
terminal repeat sequences in a cohort of long-term survivors infected from a
single source. J Virol 80(2): 1047–1052.
31. Yu XG, Lichterfeld M, Chetty S, Williams KL, Mui SK, et al. (2007)
Mutually exclusive T-cell receptor induction and differential susceptibility to
human immunodeficiency virus type 1 mutational escape associated with a two-
amino-acid difference between HLA class I subtypes. J Virol 81(4): 1619–1631.
32. Bailey JR, Williams TM, Siliciano RF, Blankson JN (2006) Maintenance of viral
suppression in HIV-1-infected HLA-B*57+ elite suppressors despite CTL escape
mutations. J Exp Med 203(5): 1357–1369.
33. Leslie A, Kavanagh D, Honeyborne I, Pfafferott K, Edwards C, et al. (2005)
Transmission and accumulation of CTL escape variants drive negative
associations between HIV polymorphisms and HLA. J Exp Med 201(6):
34. Kiepiela P, Leslie AJ, Honeyborne I, Ramduth D, Thobakgale C, et al. (2004)
Dominant influence of HLA-B in mediating the potential co-evolution of HIV
and HLA. Nature 432(7018): 769–775.
35. Moore CB, John M, James IR, Christiansen FT, Witt CS, et al. (2002) Evidence
of HIV-1 adaptation to HLA-restricted immune responses at a population level.
Science 296(5572): 1439–1443.
36. Bhattacharya T, Daniels M, Heckerman D, Foley B, Frahm N, et al. (2007)
Founder effects in the assessment of HIV polymorphisms and HLA allele
associations. Science 315(5818): 1583–1586.
37. Klenerman P, McMichael A (2007) AIDS/HIV. Finding footprints among the
trees. Science 315(5818): 1505–1507.
38. McMichael A, Klenerman P (2002) HIV/AIDS. HLA leaves its footprints on
HIV. Science 296(5572): 1410–1411.
39. Kelley CF, Barbour JD, Hecht FM (2007) The relation between symptoms, viral
load, and viral load set point in primary HIV infection. J Acquir Immune Defic
Syndr 45(4): 445–448.
40. Rodrigo AG, Goracke PC, Rowhanian K, Mullins JI (1997) Quantitation of
target molecules from polymerase chain reaction-based limiting dilution assays.
AIDS Res Hum Retroviruses 13(9): 737–742.
41. Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: Improving the
sensitivity of progressive multiple sequence alignment through sequence
weighting, position-specific gap penalties and weight matrix choice. Nucleic
Acids Res 22(22): 4673–4680.
42. Masemola AM, Mashishi TN, Khoury G, Bredell H, Paximadis M, et al. (2004)
Novel and promiscuous CTL epitopes in conserved regions of gag targeted by
individuals with early subtype C HIV type 1 infection from southern africa.
J Immunol 173(7): 4607–4617.
43. Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate
large phylogenies by maximum likelihood. Syst Biol 52(5): 696–704.
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org 11 2008 | Volume 4 | Issue 3 | e1000033
44. Martin DP, Williamson C, Posada D (2005) RDP2: Recombination detection Download full-text
and analysis from sequence alignments. Bioinformatics 21(2): 260–262.
45. Poss M, Rodrigo AG, Gosink JJ, Learn GH, de Vange Panteleeff D, et al. (1998)
Evolution of envelope sequences from the genital tract and peripheral blood of
women infected with clade A human immunodeficiency virus type 1. J Virol
46. R Development Core Team (2005) R: A language and environment for
Transmission of Attenuated HIV-1 Variants
PLoS Pathogens | www.plospathogens.org12 2008 | Volume 4 | Issue 3 | e1000033