Department of Human Retrovirology, Academic
Medical Center, University of Amsterdam, 1105 AZ
Amsterdam,1 Primagen, 1105 BA
Amsterdam,2 and Theoretical Biology,
Utrecht University, 3584 CH Utrecht,3 The
Netherlands
Sequence analysis of human immunodeficiency virus type 1 (HIV-1)
from 74 persons with acute infections identified eight strains with
mutations in the reverse transcriptase (RT) gene at positions 41, 67, 68, 70, 215, and 219 associated with resistance to the nucleoside
analogue zidovudine (AZT). Follow-up of the fate of these resistant
HIV-1 strains in four newly infected individuals revealed that they
were readily replaced by sensitive strains. The RT of the resistant
viruses changed at amino acid 215 from tyrosine (Y) to aspartic acid
(D) or serine (S), with asparagine (N) as a transient intermediate,
indicating the establishment of new wild types. When we introduced
these mutations and the original threonine (T)-containing wild type
into infectious molecular clones and assessed their competitive
advantage in vitro, the order of fitness was in accord with the in vivo
observations: 215Y < 215D = 215S = 215T. As detected by
real-time nucleic acid sequence-based amplification with two molecular
beacons, the addition of AZT or stavudine (d4T) to the viral cultures
favored the 215Y mutant in a dose-dependent manner. Our results
illustrate that infection with nucleoside analogue-resistant HIV leads
in newly infected individuals to mutants that are sensitive to
nucleoside analogues, but only a single mutation removed from
drug-resistant HIV. Such mutants were shown to be transmissible,
stable, and prone to rapid selection for resistance to AZT or d4T as
soon as antiretroviral therapy was administered. Monitoring of patients for the presence of new HIV-1 wild types with D, S, or N residues at
position 215 may be warranted in order to estimate the threat to
long-term efficacy of regimens including nucleoside analogues.
 |
INTRODUCTION |
Since 1987, when antiviral drug
therapy of human immunodeficiency virus type 1 (HIV-1) infection
started with the use of zidovudine (AZT), drug-resistant viral mutants
have rapidly emerged (3, 12). Due to the error-prone
reverse transcriptase (RT), viral mutants are generated during every
replication cycle (16). The mutational pathway is in
essence the result of a stochastic process in which a single mutation
appears more frequently than double or multiple mutations
(13). A newly formed viral mutant that has a higher
replicative capacity than its parent will overgrow the parent. This
process of selection results in the appearance of mutants with the
highest fitness or replicative capacity in their environment
(7). When an antiretroviral drug is present, mutants with
the highest level of resistance to the drug will have an advantage. As
soon as antiviral therapy is stopped, the original virus, which is
usually well adapted to an environment without the drug, will win the
competition with the resistant viruses (1). However, a
different situation occurs when a drug-resistant virus is transmitted
to a previously uninfected and untreated person. Separated from the
original drug-sensitive virus and lacking this dominant competitor in
the new drug-free environment, the drug-resistant virus will serve as a
new starting point for the evolution of HIV-1. The process does not
necessarily lead back to a virus with a genotype identical to that of
the original, drug-sensitive virus, but the phenotype (i.e.,
replicative capacity) will adapt to be optimally fit in the new
drug-free environment (7).
Of 74 new HIV-1 infections that occurred in the period 1992 to 1999, 8 were caused by viruses with mutations at positions associated with AZT
resistance. After their transmission to new hosts, those with the
resistance-conferring T215Y mutation in the RT gene showed rapid
evolution to better-replicating, more-fit viruses that acquired novel
residues at amino acid 215 of RT. The novel evolutionary pathway led to
a viral phenotype adapted to a drug-free environment and having a
replicative capacity similar to that of the original wild type. There
was evidence that the new viruses were transmissible and rapidly
converted by a single mutation to a drug-resistant phenotype when
suboptimal antiviral therapy was introduced.
| |
MATERIALS AND METHODS |
Study population.
The Amsterdam cohort studies (ACS) have
monitored a cohort of homosexual men and a cohort of intravenous drug
users, with quarterly screening of participants for the presence of the
HIV-1 antigen and antibodies. New HIV-1 infections are identified by a
seroconversion to the HIV-1 antigen and/or antibody with confirmation by Western blotting and HIV-1 viral load assay. In addition, patients admitted to the clinic of the Academic Medical Center (AMC) in Amsterdam with symptoms of suspected acute HIV-1 infection have been
examined. We studied a total of 74 new infections identified by ACS and
AMC between 1992 and 1999. Of these 74 new HIV-1 infections, 8 had RT
mutations that are associated with AZT-resistance (see below for the
sequencing method). Mutations specifically associated with RT
inhibitors other than AZT (zalcitabine [ddC], didanosine, stavudine
[d4T], lamivudine (3TC), nevirapine and other nonnucleoside RT
inhibitors) were not found. Of the 28 new HIV-1 infections identified
during or after 1996, the year of the introduction of protease
inhibitors, none involved viruses that had mutations that confer
resistance to protease inhibitors. Sequences of the 5' end of the RT
gene with mutations associated with AZT resistance were deposited in
GenBank (see below for sequence numbers).
Viral load and CD4 counts.
Viral load was determined by
using the NucliSens assay of Organon Teknika BV (Boxtel, The
Netherlands). CD4+ cells in blood were counted by standard
flow cytometry using a FACScan flow cytometer (Becton Dickinson, San
Jose, Calif.) and commercially available monoclonal antibodies (Becton Dickinson).
Screening for resistance-conferring mutations.
Viral RNA was
isolated from 200 µl of serum drawn from each of the 74 subjects at
the first point of seroconversion to the HIV-1 antigen or antibody by
using the method described by Boom et al. (2). Part of the
HIV-1 pol region including the part of the RT gene encoding
the amino terminus was amplified by RT-PCR, essentially according to
the protocol described by Nijhuis et al. (18). To
facilitate accurate two-stranded sequence analysis of the amino
terminus-encoding part of the RT gene, we adapted the nested PCR in the
Nijhuis protocol by amplifying two overlapping fragments instead of
one. The 5' fragment was amplified using primer
SP6- RT19new, ATTTAGGTGACACTATAGCACCTGTCAACATAATTGGAAG (SP6
sequence is in italics; HXB-2 nucleotides [nt] 2491 to 2512), and
T7-B-SEQ, TAATACGACTCACTATAGGGAATATTGCTGGTGATCCTTTCCA
(T7 sequence is in italics; HXB-2 nt 3030 to 2006). The 3'
fragment was amplified with primers SP6-C-SEQ,
ATTTAGGTGACACTATAGGTATACTGCATTTACCATACC (SP6
sequence is in italics; HXB-2 nt 2927 to 2947), and T7-ET10, TAATACGACTCACTATAGGGCTGCCAGTTCTAGCTCTGCTTC
(T7 sequence is in italics; HXB-2 nt 3462 to 3441). Both strands
of the nested PCR fragments were directly sequenced using the SP6 and
T7 primer sequences. Sequencing was performed with Taq dye
primers (Applied Biosystems, Foster City, Calif.) and the
ThermoSequenase fluorescence-labeled primer cycle sequencing kit
(Amersham International, Little Chalfont, England). The sequence
products were analyzed on an automatic sequencer (Applied Biosystems
DNA sequencer model 373A stretch or 377). The sequence of the RT gene
was screened at positions associated with drug resistance as described
by Hirsch et al. (10). For new infections caused by
viruses with mutations at positions associated with AZT resistance,
serum samples drawn after the seroconversion point were analyzed for
changes compared to the seroconversion sample. Mixtures of different
viral RT sequences were quantified by assessing the ratio between the
relevant peak areas of the signals of the corresponding nucleotides on
the electropherogram using the sequences of both strands. In cases
identified during or after 1996, the year of introduction of protease
inhibitors in The Netherlands, we sequenced the protease gene as well
as the RT gene. The protease gene was amplified according to the procedure described by de Jong et al. (5). Nested PCR was
performed by using primers SP6-A-SEQ,
ATTTAGGTGACACTATAGGAGCCAACAGCCCCACCAG (SP6
sequence is in italics; HXB-2 nt 2149 to 2167) and T7-A-SEQ, TAATACGACTCACTATAGGGTAAAGAAGACAGTTACCG (T7
sequence is in italics; HXB-2 nt 2639 to 2621). Sequence analysis was
performed as above for the RT gene.
Infectious molecular clones.
Biological clones of the virus
of patient 4 were prepared by a limiting-dilution series of human
peripheral blood mononuclear cells (PBMC) derived from a sample taken 1 month after seroconversion (14). Viral RNA was isolated
from the culture supernatants of individual biological clones. The
complete RT gene of the virus was reverse transcribed and amplified
(5) but by using primers ET44
(GACATTTATCACAGCTGGCTAC; antisense; nt 4338 to 4359 of
HIVHXB2CG; GenBank accession no. K03455) and 5'-PROT-OUT
(GAGCCAACAGCCCCACCAG; nt 2149 to 2167; sense) in the RT
reaction and first PCR and by using primers ET42
(CTGTTGACTCAGATTGGTTGCACTTTAAATTTTCC; nt 2517 to 2551;
sense) and ET17 (AGGTGGCAGGTTAAAATCAACTAGCCATGCCATTGCTCTCC; nt 4319 to 4285; antisense) in the nested PCR. The PCR product was cloned in plasmid pCR-2.1 using a TA vector system (InVitrogen BV,
Groningen, The Netherlands). Sequence analysis of the 5' end of the RT
gene was performed, and a clone with the T215Y mutation (AZT resistant)
was chosen. The insert was subcloned in the EcoRI site of
vector pUC19 to remove the flanking BstXI sites, creating p4-Y. PCR-mediated mutagenesis of amino acid 215 of RT was performed essentially as described by de Jong et al. (5) except that primer ET10 (CTGCCAGTTCTAGCTCTGCTTC; nt 3462 to 3441) and
the sense mutagenesis primer for the first PCR were used, whereas primer ET07 (GGAAGTTCAATTAGGAATACC; nt 2812 to 2832) and the
antisense mutagenesis primer were used for the second PCR. The two PCR
products were combined, and a third PCR was performed using primers
ET10 and ET07. The resulting PCR product was digested with
BstXI and cloned into a BstXI-digested p4-Y.
Mutagenesis primers ET47
(GAAGTGGGGATTCGACACACCAGACAAAAAAC; nt 3170 to
3209; sense; codon 215 is underlined) and ET48
(GTTTTTTGTCTGGTGTGTCGAATCCCCACTTC; nt 3209 to
3170; complementary) were used to introduce the 215 aspartic acid (GAC
codon) mutation; ET49
(GAAGTGGGGATTCTCCACACCAGACAAAAAAC; sense) and
ET50 (GTTTTTTGTCTGGTGTGGAGAATCCCCACTTC;
complementary) were used to introduce the 215 serine (TCC
codon) mutation; ET51 (GAAGTGGGGATTCACCACACCAGACAAAAAAC; sense) and
ET52 (GTTTTTTGTCTGGTGTGGTGAATCCCCACTTC; complementary) were used to introduce the 215 threonine (ACC
codon) wild-type mutation. Sequencing was performed to verify that
the exchanged BstXI fragment contained the introduced
mutations. By this procedure, we created plasmids p4-D, p4-S, and p4-T
encoding the indicated (one-letter code) amino acids at the 215 position.
At week 154, viral RNA was isolated from a sample drawn from patient 4, and RT-PCR was performed as described above using primers ET42 and
ET10; the product was cloned in the TA cloning vector pCRII (InVitrogen
BV). Sequence analysis showed that individual clones were
representative of the quasispecies that contained an S68G change
relative to the p4-Y clone. In addition, these clones had naturally
occurring variations at amino acids 39 and 135 (T39A, I135T). The 3'
end of one of these clones (p4-68G) starting from the SspI
site (HXB-2 nt 3025) was replaced by the 3' end of p4-Y or p4-D
creating p4-Y-68G and p4-D-68G, respectively. The infectious molecular
clone in which RT genes derived from patient 4 were recombined was
pHIV-Lai, lacking its RT gene (pLAI-
RT). The RT gene deletion was
obtained by PCR-mediated mutagenesis. Briefly, the deletion was
constructed in a subclone of HIV-Lai containing the
ApaI-NcoI fragment by using primers ET40
(TGACCGCGGAAAATTTAAAGTGCAAC; nt 2550 to 2532;
antisense; SacII site underlined) and ET23
(AGTCCGCGGAGAGCAATGGCTAGT; nt 4286 to 4300;
sense; SacII site underlined) and primers 5' and 3' of
ApaI and NcoI sites, respectively. This created
molecular clone pLAI-RT with a SacII site replacing the
RT gene.
Recombinant viruses with a recipient-derived RT gene.
To
obtain recombinant viruses, C33A cells (human cervix carcinoma cell
line) were seeded in 24-well plates and grown to 80% confluence in 1 ml of RPMI 1640 medium (Life Technologies, Breda, The Netherlands)
supplemented with 10% fetal calf serum and antibiotics. Cotransfection
of 300 ng of pLAI-RT and 200 ng of the purified RT gene-containing
EcoRI fragment of p4-Y, -D, -S, -T, -Y-68G, or -D-68G was
performed using the TFX-50 reagent protocol (Promega Benelux BV,
Leiden, The Netherlands). One day after transfection, 200,000 MT2 cells
per well were added, and C33A and MT2 cells were cocultured. After 2 days, the MT2 cells including medium were transferred to a
25-cm2 flask, with the addition of fresh medium and MT2
cells. Cells were monitored for the formation of syncytia, and the
supernatant containing the recombinant virus was harvested after the
spread of syncytia throughout the culture. The virus was frozen into aliquots at 
70°C. The virus titer (50% tissue culture infective dose [TCID50]) was determined by limiting-dilution
titration on MT2 cells. Resistance to antiviral drugs was assayed by
VIRCO (Edegem, Belgium) (9).
Competition experiments were performed by mixing a total amount of
1,000 TCID50 of the recombinant viruses in different
ratios. These virus mixtures were used to infect 2 ml of
phytohemagglutinin-stimulated PBMC (2 × 106
cells/ml), which were maintained in a six-well plate in RPMI 1640 medium supplemented with 10% fetal calf serum and interleukin-2 (6). Twice a week, cells were passaged, and 10 µl of the
culture was transferred to 2 ml of a new culture of 2 × 106 PBMC/ml. Similarly, 500,000 MT2 cells were infected
with 1,000 TCID50 and maintained in 1 ml of RPMI medium
supplemented with 10% fetal calf serum in a 24-well plate. When
syncytia in the MT2 cells had formed, 5 µl of the culture was
transferred to 1 ml of a new culture composed of 1 ml of medium
containing 500,000 MT2 cells. Competition experiments were performed in
the absence or presence of antiviral drugs (i.e., AZT, d4T, or ddC). At
each passage, samples from the supernatant from which to isolate viral RNA for sequence analysis were drawn. Viral RNA was isolated according to the method of Boom et al. (2). The RNA was directly
sequenced as described above. The ratios between the different
recombinant viruses were determined from the direct sequence by
assessing the ratios between the relevant peak areas of the signals on
the electropherogram corresponding to the nucleotides encoding amino acid 215 of RT. Alternatively, the ratios between the different viruses
were determined by using real-time nucleic acid sequence-based amplification (NASBA) with molecular beacons, as described in the next section.
Quantification of mutant mixtures using real-time NASBA and
molecular beacons.
RNA was isolated from 100 µl of culture
supernatant according to the method of Boom et al. (2).
The nucleic acids were eluted in 50 µl of water, and 5 µl of the
eluate was used as input for amplification by NASBA. NASBA was
performed by using the basic kit of Organon Teknika supplemented with
primers 215-P2 GACTTAGAAATAGGGCAGCA (HXB-2 nt 2705 to 2724)
and 215-P1
AATTCTAATACGACTCACTATAGGGGTTCATAACCCATCCAAAGGAAT GGA
(T7 sequence in italics; HXB-2 nt 2831 to 2806) and the molecular beacons (20) RT 215GAC,
ccgactcTCGACACACCAGACAAAAAACgagtcgg (FAM, dabcyl), and RT 215TAC,
ccgactcTCTACACACCAGACAAAAAACgagtcgg (6-ROX, dabcyl) (stem sequence is in lowercase; HXB-2 nt 2772 to
2792; position 215 [underlined, boldface] is either a GAC [D] or
TAC [Y] codon). The two primers and two beacons were each added to a final concentration of 200 nM. Fluorescence was measured in real
time on a Fluoroscan Ascent (Labsystems Oy, Helsinki, Finland). The
assays were calibrated by mixing different quantities of in
vitro-synthesized RNA containing the GAC or TAC position 215 codon.
To synthesize the RNA, PCR fragments were generated from p4-D-68G and
p4-Y-68G by using primers AATTCTAATACGACTCACTATAGGG (T7 site
located in the plasmid) and ET10 (HXB-2 nt 3462 to 3441). These PCR
fragments were used to produce RNA containing the GAC or TAC position
215 codon by using T7 RNA polymerase (Amersham Pharmacia Biotech
Benelux, Roosendaal, The Netherlands) in an in vitro transcription
reaction. Under our reaction conditions with 105 molecules
of in vitro-synthesized RNA per reaction, the beacon specific for the
GAC codon did not react with the RNA containing the TAC codon
and vice versa. In vitro-synthesized RNA containing the GAC and TAC
codons was mixed at different percentages (0, 2, 4, 10, 33, 50, 67, 90, 96, 98, and 100%). Repeated (n = 5) experiments
showed that quantification of the mixtures using real-time NASBA could
be achieved until 4% of either the TAC or GAC variant (see Fig. 4A and
B). Below 4% of the variant, the variant could be observed but not
quantified. Quantification of mixtures by real-time NASBA and direct
sequencing were compared for the competitions in the presence of AZT
(see Fig. 4C). The methods gave comparable percentages (±10%) of a
variant mixture, provided that direct sequencing was performed using
the ET dye primer on the ABI automatic sequencers and not the big-dye
techniques on the ABI machines. However, when a variant was present in
small amounts (less than 10%), quantification by direct sequencing was
no longer reliable due to background signals. Real-time NASBA was
broader in dynamic range; it could reliably quantify a variant at a
level as low as 4% and could detect a variant at a level as low as
1%.
Fitness calculation.
Selection coefficients were calculated
by a novel method involving the estimation of the total viral
replication during the observation period (16a). The
method can be used both for steady-state viral populations, to which
most previously used methods apply (8), and for decreasing
or expanding viral populations, such as those observed, for example, in
cell cultures. Briefly, the model is derived as follows. Because
mutations in the RT are expected to mainly influence the replication
rate of a mutant virus and to hardly influence its half-life, we write
dW/dM = rW 
W and dM/dt = r(1 + s)M M for the dynamics of
wild-type virus W and mutant virus M. Here
r is the (wild-type) replication rate (which may change over
time [t]), 1/ is the generation time, and s
is the classical coefficient of selection. Because this model combines the dynamics of productively infected cells and free virions, the
generation time should be about 2 days (11). The same
model has been used before for estimating s from in vivo
data with approximately steady-state viral loads (8). One
conventionally defines the frequency of the mutant genotype by the
equation P = M/(W + M), and hence 1 p is the frequency of the wild-type virus. If r is a constant, one writes the solutions W(t) = W(0)e(r 
)t and
M(t) = M(0)e[r(1 + s) ]t so that s is computed from the logarithms
of the genotype ratios (H = M/W = p/[1 p]) at time 0 and time t, i.e., s = ln[H(t)/H(0)]/rt.
Thus, for estimating the relative fitness one generally needs to know
the ratio of the genotype frequencies and the total replication,
rt. r may vary over time, however. This requires the integration of the replication rate over the experiment.
Fortunately, s can still be estimated by the simple formula
(16a) s = ln[H(t)/H(0)]/{ln[W(t)/W(0)] + t}, which
is available at
http://www-binf.bio.uu.nl/~rdb/fitness.html. If the total viral loads and the percentages of wild-type virus at
times 0 and t are known, the website equation can be used to compute the change in the wild-type virus load, W(t)/W(0),
and the change in the ratio of the genotype frequencies,
H(t)/H(0), to give s. H is undefined
when the percentage of mutant virus is estimated as 0 or 100%.
Whenever possible, we therefore ignored such data points. However, in a
few instances (patients 2 and 4) no alternative data points were
available. In such instances, we took 1 or 99% for the genotype
frequency, which was a conceivable assumption (see Table 1 for the
example of patient 1). For the fitness calculations for the viral
culture we took the cumulative p24 value as the viral load, i.e., a p24
value of 1,000 pg/ml after a 1:100 dilution of a viral culture was
100,000 pg/ml, allowing us to keep a dilution factor of 1 in the
formula on the website.
We calculated s over the interval between samples 1 and 6 (Table 1) (using 99 and 1% for the
percentages of virus with Y at position 215 [215Y virus]
in samples 1 and 6, respectively); the fitness difference between 215Y
and 215N was 2.8%. Corresponding fitness differences were 2.3, 3.3, 2.6, 2.5, and 3.0% for intervals between samples 2 and 5, 2 and 4, 2 and 3, 3 and 4 (all of which used detected percentages of 215Y virus),
and 2 and 6 (which used 1% 215Y virus for sample 6), respectively. The
one-sample t test of the SSPS, version 8.0, statistical
package was used to calculate the mean and the 95% confidence
interval. We concluded that 215N had a relative fitness of 103% (95%
confidence interval: 102 to 104%), with 215Y fitness = 100%.
We calculated s over the interval between samples 1 and 4 for the virus culture of the competition between 215Y and 215T viruses (Table 2); the fitness difference between
215Y and 215T was 5.1%. Corresponding fitness differences were 5.4, 6.0, 4.7, and 4.9% for intervals between samples 2 and 4, 3 and 4, 1 and 3, and 2 and 3, respectively. We concluded that 215T had a relative
fitness of 105% (95% confidence interval: 104 to 106%), with 215Y
fitness = 100%.
View this table:
[in this window]
[in a new window]
|
TABLE 2.
Example of the fitness calculation with the virus culture
of the competition between 215Y and 215T viruses
|
|
Nucleotide sequence accession numbers.
Accession numbers for
viral RT gene sequences obtained in this study were as follows: patient
1 (H0671), AF265569; patient 2 (I6056), AF265572; patient 3 (H0095),
AF265570; patient 4 (I7052), AF265571; patient 5 (H0137), AF265568;
patient 6 (I3234), AF265573; patient 7 (M12690), AF265574; patient 8 (2202575), AF265567.
| |
RESULTS |
In vivo evolution of viruses with a mutated RT gene in the absence
of an antiviral drug.
Among 74 new HIV-1 infections occurring
between 1992 and 1999, 8 were found to be due to viruses that have
mutations in the RT gene that are associated with AZT resistance. At
seroconversion, the corresponding mutations in RT were M41L and T215Y
(patients 1, 2, and 4); T215Y (patient 3); D67N, K70R, and T215F
(patient 5); D67N, K70R, and K219Q (patient 6); M41L and the T215D
mutation, which involved an unusual amino acid not seen associated with AZT resistance (patient 7); and K70R (patient 8). The K70R mutation in
the virus infecting patient 8 could be the result of a natural polymorphism, although in the absence of AZT the K70R variant rarely
becomes the dominant variant in the viral quasispecies (17). The eight infections were monitored until antiviral
therapy (Fig. 1). Infections 1 to 5, which all were found at seroconversion to have an AZT
resistance-conferring mutation at amino acid 215 of RT (either 215Y or
215F) showed evolution in their RT genes. Since infections 1 to 4 could
be monitored for at least 1 year before antiviral therapy was given,
the evolution of the RT gene, and in particular the evolution at amino
acid 215 of RT, was analyzed in these cases (Fig.
2A, 1 to 4).
View larger version (30K):
[in this window]
[in a new window]
|
FIG. 1.
In vivo evolution of transmitted viruses with drug
resistance-conferring mutations in the absence of antiretroviral
therapy. Indicated are the positions and amino acids (minor species in
italics) in which the viruses differ from the wild type at positions
known to be involved in drug resistance (10). Our subjects
were monitored until antiviral therapy.
|
|
View larger version (21K):
[in this window]
[in a new window]
|
FIG. 2.
In vivo evolution at amino acid 215 of RT. (A) Relative
abundances of viruses with the indicated amino acids at position 215 of
RT after transmission of AZT-resistant viruses. The evolution of
transmitted viruses is shown in patients 1 to 4. All start with 215Y
and evolve to 215D (patients 1, 2, and 4) and/or 215S (patients 3, and
4), with 215N observed as an intermediate (patients 1 and 2). (B) In
vivo mutational pathways as observed (patients 1, 2, and 4) or
hypothesized (patient 3) at amino acid 215 of RT after transmission of
AZT-resistant viruses.
|
|
In patient 1 at seroconversion, the RT of the infecting viruses had the
M41L and T215Y mutations conferring AZT resistance. While the codon
at position 41 of the RT gene did not change, amino acid codon 215 changed from TAC, encoding tyrosine, to AAC, encoding asparagine. The
215N virus had completely replaced 215Y within 100 weeks (2 years)
after seroconversion. Over the next 50 weeks, the AAC codon at
position 215 was gradually replaced by a GAC codon encoding
aspartic acid, which remained the dominant species thereafter.
In patient 2 at seroconversion, the M41L and T215Y mutations were
likewise present and the mutants evolved as in patient 1, but on a far
smaller time scale. During the 36 weeks that patient 2 was monitored
215Y evolved to a mixture of 215N and 215D, in which 215D eventually
became the dominant species, a process that took 150 weeks in patient 1.
In patient 3, RT had only the T215Y mutation at seroconversion, with
codon 215 evolving from TAC (215Y) to AGC, encoding serine (215S),
although at week 11 GAC (215D) was transiently observed. Unlike the
other mutations we studied, which involve a one-nucleotide change
within the codon 215, the change from TAC to AGC involved a
two-nucleotide change, and no obvious intermediate was detected. However, in this infection, the 215 AAC codon may have been
transiently formed as an intermediate to the 215 AGC codon encoding
serine. A notable mutation was observed at codon 68 of the RT gene,
which changed from wild-type AGC (serine) to a mixture of AGA and AAC (asparagine).
In patient 4, RT had the M41L and T215Y mutations at seroconversion. At
week 3 after seroconversion, the 215 TAC (Y) codon was already
being replaced by the 215 GAC (D) codon, with full replacement
observed at week 20. At week 37, we detected a mixture of the 215 GAC
codon and the 215 TCC codon (S), in which the relative abundances of the two fluctuated until week 150, when the last sample
was drawn before therapy began. The changes to an aspartic acid and to
a serine at position 215 of RT required only a single mutation, from
the TAC codon to GAC and TCC codons, respectively. Analogous to
what was observed in patient 3, the RT gene of patient 4 showed a
change at position 68. In this case, however, the wild-type 68 AGC
codon (serine) changed to a GGC codon encoding glycine (68G).
At week 120, wild-type 68S had been completely replaced by 68G, which
persisted thereafter. An overview of the mutational pathways in the
different patients is shown in Fig. 2B.
The rate of RT evolution in these cases varied considerably. In patient
1, replacement of 215Y by 215N and subsequently by 215D took place over
a time period of more than 150 weeks. Concurrent with the replacement
of 215Y by 215N, the viral load increased from below 1,000 copies/ml to
10,000 copies/ml, and it increased further to 30,000 copies/ml when
215D became the dominant species. During this period, the CD4 T-cell
counts remained stable and high at approximately 1,000/µl. In patient
2, replacement of 215Y by 215D occurred within less than 40 weeks, with
a viral load between 30,000 and 100,000 copies/ml and CD4 T-cell counts
hovering between 400 and 600/µl. In patient 3, replacement of 215Y by
215S occurred after 40 weeks; the viral load declined from 300,000 copies/ml at seroconversion to 10,000 to 30,000 copies/ml from week 30 to 55 and CD4 T cells remained at 200 to 400/µl. In patient 4, replacement of 215Y by 215D occurred within 25 weeks, with a viral load
that declined from above 100,000 copies/ml at seroconversion to about
25,000 copies/ml at week 100 and with CD4 T-cell counts varying from
approximately 300 to 600/µl.
In all three patients who had been infected with a virus in which the
RT gene contained both the 41L and 215Y mutations, the dominant viral
species that emerged contained an RT with the 41L and 215D mutations.
Patient 7 was infected with a virus containing an RT with these 41L and
215D mutations, indicating that a virus with these mutations is transmissible.
In vivo fitnesses of the newly emerged viruses with mutations in
the RT gene.
The increased replicative capacity, or fitness, of
the newly emerged viruses with changes in their RT genes was expressed relative to the replicative capacity of the 215Y variant at
seroconversion, which was arbitrarily set at 100%. The fitnesses of
the variants newly formed in an environment without an antiviral drug
were calculated. In patient 1 the available serum samples allowed the calculation of the fitness of the 215N variant, which was 103% (95%
confidence interval: 102 to 104%) of that of 215Y in an environment without an antiviral drug. By the same method, the fitness of the 215S
variant (patient 3) was calculated to be 106% (103 to 109%) of that
of 215Y and the fitness of the 215D variant was calculated to be 107 (patient 2) or 110% (patient 4) (96 to 124%) of that of the 215Y
variant. By way of comparison, the S68G change observed in the RT of
patient 4 resulted in a relative fitness of the 68G viruses of 101%
compared to that of the 68S virus, which was set at 100%. To give an
impression of the competition between the new variants and the
AZT-resistant 215Y, we plotted the fractions (relative to 215Y) of the
different variants in the viral population (on a log scale) against the
elapsed time after seroconversion (Fig.
3A).
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 3.
In vivo and in vitro relative fitnesses of viruses
marked by their amino acids at position 215 of RT. (A) Fitness plot of
viruses differing at amino acid 215 (N, S, and D), which was derived
from the observed replacement of 215Y by viral variant 215N, 215S, or
215D in the four patients (two of them had a 215D). The data points
correspond to the observed viral mixtures in the patients (Fig. 2A).
Shown are the rates (weeks after seroconversion) at which and to what
extent (on a log scale) 215Y virus is replaced by a different variant.
(B) Fitness plot of viruses differing only at amino acid 215 of RT,
which competed in PBMC in the absence of an antiretroviral drug. Shown
are the rates (number of days of culture) at which and to what extent
(on a log scale) variants 215D ( ), 215S ( ), and 215T ( )
replaced 215Y. When 215D was competed with 215T or 215S, 215D was found
to be as fit as 215T or 215S.
|
|
In vitro fitnesses of viruses differing at amino acid 215 of
RT.
In vivo, if HIV-1 isolates found in two different samples
drawn from a patient show genetic differences, they differ usually at
multiple positions, of which only a few may significantly contribute to
a change in replicative capacity. To examine the significance of RT
amino acid 215 for replicative capacity, viruses that differed only at
this amino acid were constructed. In the RT isolated from patient 4, containing the M41L and T215Y mutations, only the 215 position was
changed by in vitro mutagenesis to either 215D, 215S, or to 215T, the
wild-type amino acid present before the introduction of AZT. The
wild-type 215T was found also in the viruses of the donor of patient 4 before he received AZT and resistant viruses developed. The various RT
genes were recombined into an HIV-1 Lai background, creating viruses
that differed only at amino acid 215 of RT. Competition between the
resulting viruses was observed in PBMC in the absence of an antiviral
drug. In that environment, the 215Y virus was the least fit virus. The
relative fitnesses of all the viruses were calculated. The in vitro
relative fitnesses of 215D, 215S, and 215T compared to that of 215Y
(100%) were 104 (95% confidence interval: 102 to 106%), 107 (104 to
109%), and 105% (104 to 106%), respectively (Fig. 3B). Competition
among the 215D and the 215S or 215T viruses showed them to be equally fit, i.e., having a fitness difference of less than 1%. The order of
in vitro fitness of the viruses differing only at amino acid 215 was
215Y < 215D = 215S = 215T, corresponding to the in
vivo-observed order.
Characterization of the viruses after antiviral therapy by using
real-time NASBA.
Patient 1 was not treated because his CD4 counts
remained stable and high at approximately 1,000 CD4 cells/µl.
Patients 2 and 3 were treated with a combination of protease and RT
inhibitors (indinavir plus ritonavir plus d4T plus 3TC and indinavir
plus d4T plus 3TC), and their viral loads declined to below detection level (<400 copies/ml). Patient 4 was treated at 160 weeks after seroconversion with RT inhibitors d4T and ddC. No change in the viral
load was observed between week 154 (pretherapy) and week 170 (after
therapy), indicative of the development of resistance to the antiviral
drug. In this case, the only observed mutation in RT was at amino acid
215, which was changed from a mixture of 215D and 215S to 215Y by a
single mutation (GAC and/or TCC to TAC). Phylogenetic analysis of the
RT genes of the viruses found before and after therapy showed that the
posttherapy 215Y virus was derived from viruses present before therapy
but not present at transmission. The newer viruses shared 215Y with the transmission virus but had, in addition, mutations at positions 68 (S68G), 39 (T39A), and 135 (I135T). To examine whether the change of
environment caused by the therapy with d4T and ddC was involved in the
selection of a viral population with the 215D- or 215S-to-215Y
mutation, viruses that differed only at amino acid 215 of RT were
studied. The viruses tested had either 215D or 215Y with 68G (present
directly before and after therapy) or 68S (present directly after
transmission). By using competition experiments with MT2 cells, the
replicative capacities of the viruses were compared at increasing
concentrations of AZT, d4T, or ddC. To facilitate the analysis of the
competition experiments, we developed a real-time NASBA with two
molecular beacons differing at one nucleotide. The dynamic range of
quantification and the throughput of the real time NASBA are higher
than those in sequence analysis, which generally cannot quantify
mixtures with less than 10% of a variant. The molecular beacons, among
which the GAC beacon contained a FAM label and the TAC beacon contained
a ROX label, were able to discriminate viruses differing at the 215 codon of the RT gene (Fig. 4A). The
real-time NASBA was able to quantify mixtures containing as little as
4% of either the GAC- or TAC-containing virus (Fig. 4B) and to detect
the GAC- or TAC-containing virus in a mixture in which they composed no
more than 1%. Using the real-time NASBA, the analysis of the
competition experiments showed that the background of the viruses with
the 68S or 68G mutation (either those directly after transmission or
those present at therapy) did not influence the phenotype. However, the
215D-to-215Y change was crucial to obtain a replicative advantage in
the environment with both AZT and d4T (Fig. 4C). The replicative
advantage of a virus with a 215Y codon in the presence of AZT or
d4T varied in a dose-dependent manner. The replicative advantage of
215Y virus could be expressed as an average fitness difference of 5 and
11% in the presence of 45 and 170 nM AZT, respectively (0.5 and 2 times, respectively, the 50% inhibitory concentration
[IC50] of a sensitive virus), and 2 and 28% in the
presence of 3 and 12 µM d4T, respectively (0.5 and 2 times,
respectively, the IC50 of a sensitive virus), indicating
that the replicative advantage of 215Y increased with higher drug
concentrations. At 1.5 and 4.5 µM (0.5 and 1.5 times, respectively,
the IC50 of a sensitive virus), ddC did not appear to
influence the relative replicative capacities of the viruses. In
classical phenotypic resistance assays, the 215Y virus is >100-fold
more resistant to AZT and 2-fold more resistant to d4T than the wild
type and sensitive to ddC, whereas the 215D and 215S viruses are
sensitive to AZT, d4T, and ddC.
View larger version (26K):
[in this window]
[in a new window]
|
FIG. 4.
In vitro competitions of 215Y and 215D viruses in the
absence and presence of antiviral drugs as analyzed by using real-time
NASBA. (A) Representative experiment in which relative fluorescence
units (RFU) are shown as a function of time when RNA mixtures (total of
105 molecules) of 215D and 215Y (ranging from a 25:1 to a
1:25 ratio) were assayed by real-time NASBA using GAC (FAM-labeled) and
TAC (ROX-labeled) molecular beacons indicative of 215D and 215Y,
respectively. (B) Calibration curve derived from panel A. The log
function of the slope of the increase of fluorescence in time is
proportional to the relative abundances of 215D and 215Y in a sample,
which can be inferred from the known mixtures. (C) Competition between
215Y and 215D viruses in the presence of increasing concentrations of
d4T (left) or AZT (right). Viruses containing RT with 215Y and 215D
competed in MT2 cells in the presence of 0, 3, or 12 µM d4T or 0, 45, or 170 nM AZT. The ratio of 215Y/215D virus in the supernatant at
passage was determined by using real-time NASBA.
|
|
| |
DISCUSSION |
Our study indicates that between 1992 and 1999, 10 to 15% of new
HIV-1 infections in Amsterdam, The Netherlands, were caused by HIV-1
strains with mutations in the RT gene at positions that are associated
with AZT resistance. We found no mutations associated with resistance
to other antiretroviral drugs, including the protease inhibitors
introduced in 1996. When the HIV-1 strains with drug resistance-conferring mutations are transmitted, a new environment without antiviral drugs in a new host is encountered. After
transmission of usually a small amount of virus, the wild-type strains
from which the drug-resistant strains were derived are lost from the (memory of the) quasispecies and thereby are excluded from competition in the new environment (19). The separation of individual
viruses from the remainder of competing viruses by bottleneck
transmission is reminiscent of the founding of an island population and
may lead to new directions in evolution (15). In the new,
drug-free host, viruses that have an increased replicative capacity or
fitness compared to that of their drug-resistant parent will be
selected. Although the replicative properties of the viruses we studied will most likely be determined by more than one gene, we focused on the
resistance-conferring RT gene. We found that the
drug-resistance-conferring RT amino acid 215Y in viruses changed after
transmission, but not to the wild-type threonine. Instead, novel amino
acids at position 215 were observed (see also reference
21). Single mutations in the TAC codon (215Y)
gave rise to AAC (asparagine), GAC (aspartic acid), or TCC (serine). In
addition, in one of the new infections an AGC codon (serine, 215S)
was formed by a two-step mutation. Selection occurred for the fittest
variant, the 215N virus being less fit than the 215D or 215S virus.
Compared to the transmitted AZT-resistant 215Y virus the 215N viruses
were 3% more fit, the 215S viruses were 6% more fit, and the 215D
viruses were 7 to 10% more fit. Our calculations assumed a constant
viral generation time; however, variation among the respective
infections could affect the relative fitness values. In this respect,
patient 1, with an initially low viral load and slow replacement of the
different variants, is illustrative. Nevertheless, the order of viral
fitness is 215-tyrosine < 215-asparagine < 215-aspartic
acid = 215-serine. In vitro experiments in a drug-free environment
showed that wild-type viruses with 215-threonine were as fit as the
215D or 215S viruses and fitter than the 215Y viruses. Threonine (ACC
codon; 215T) can be formed by a two-step mutational process from
the TAC codon (215Y). Such a two-step mutant, generated by
stochastic mutational processes, will usually be formed later than
one-step mutants such as 215D (GAC) and 215S (TCC). The principle of
competitive exclusion could explain why we observed no 215T viruses in
the new infections started with AZT-resistant viruses, since an equally fit variant will not replace an existing virus that already dominates the population (4). In this group, the phenotype of a 215D or 215S virus will to a large extent be the new wild type. This is most
clearly shown in the virus of patient 7, in which the 41L and 215D
mutations appeared directly at seroconversion, strongly pointing at the
transmissibility of the 215D virus.
When antiviral therapy is applied, the environment changes once yet
again, and under conditions of suboptimal therapy in which sufficient
residual viral replication can occur, a new selection for the most fit
variant takes place. For patient 4, a single mutation (GAC or TCC to
TAC) within the codon of amino acid 215 of RT appeared sufficient
to confer a decisive replicative advantage in the presence of d4T and
ddC. By using a newly developed real-time NASBA with a high throughput
and dynamic quantification range, in vitro competitions with increasing
concentrations of AZT or d4T showed that the change from 215D to 215Y
by a single nucleotide mutation conferred the replicative advantage.
Apparently, the AZT-resistant genotype of the virus of patient 4 is
also advantageous in the presence of d4T. This finding shows that
although the transmitted AZT-resistant viruses that have adapted to the
drug-free environment closely resemble the 215T wild type in their
phenotypes, their genotype retains many characteristics of the
transmitted AZT-resistant 215Y virus.
Our observations confirm the high plasticity of the genome of HIV-1 and
in particular of the RT gene. Transmission from a treated host to an
untreated host immediately results in adaptation of the virus to its
new, drug-free environment. Such adaptation becomes manifest as a
fixation of a 215D or 215S mutation instead of reversion to the wild
type of the pretherapy era: 215T. These new wild types share AZT
sensitivity with the old wild type, 215T. However, the new wild types
are only one step away from 215Y and 215Y is therefore expected to
arise rapidly upon administration of antiretroviral therapy when
therapy is suboptimal. As we have observed, the newly established wild
types can be transmitted. The monitoring of new wild types with D, S,
or N at position 215 may thus be warranted in order to estimate the
threat to long-term efficacy of a therapy regimen that includes
nucleoside analogues.
We thank M. Bakker and S. Jurriaans for providing serum samples;
K. Lindenberg and R. Coutinho of the Amsterdam Municipal Health Service
for patient data; V. Benes and H. Voss of EMBL for initial help with
sequencing; B. Hemmelder and J. Maas for technical assistance; L. Phillips for editorial assistance; B. Berkhout, M. Cornelissen, and V. Lukashov for discussion and ideas; and the participants of the ACS for
their cooperation over many years.
| 1.
|
Albert, J.,
J. Wahlberg,
J. Lundeberg,
S. Cox,
E. Sandstrom,
B. Wahren, and M. Uhlen.
1992.
Persistence of azidothymidine-resistant human immunodeficiency virus type 1 RNA genotypes in posttreatment sera.
J. Virol.
66:5627-5630[Abstract/Free Full Text].
|
| 2.
|
Boom, R.,
C. J. Sol,
M. M. Salimans,
C. L. Jansen,
P. M. Wertheim-van Dillen, and J. van der Noordaa.
1990.
Rapid and simple method for purification of nucleic acids.
J. Clin. Microbiol.
28:495-503[Abstract/Free Full Text].
|
| 3.
|
Boucher, C. A.,
E. O'Sullivan,
J. W. Mulder,
C. Ramautarsing,
P. Kellam,
G. Darby,
J. M. Lange,
J. Goudsmit, and B. A. Larder.
1992.
Ordered appearance of zidovudine resistance mutations during treatment of 18 human immunodeficiency virus-positive subjects.
J. Infect. Dis.
165:105-110[Medline].
|
| 4.
|
Clarke, D. K.,
E. A. Duarte,
S. F. Elena,
A. Moya,
E. Domingo, and J. Holland.
1994.
The Red Queen reigns in the kingdom of RNA viruses.
Proc. Natl. Acad. Sci. USA
91:4821-4824[Abstract/Free Full Text].
|
| 5.
|
de Jong, J. J.,
J. Goudsmit,
V. V. Lukashov,
M. E. Hillebrand,
E. Baan,
R. Huismans,
S. A. Danner,
J. H. ten Veen,
F. de Wolf, and S. Jurriaans.
1999.
Insertion of two amino acids combined with changes in reverse transcriptase containing tyrosine-215 of HIV-1 resistant to multiple nucleoside analogs.
AIDS
13:75-80[CrossRef][Medline].
|
| 6.
|
de Ronde, A.,
B. Klaver,
W. Keulen,
L. Smit, and J. Goudsmit.
1992.
Natural HIV-1 NEF accelerates virus replication in primary human lymphocytes.
Virology
188:391-395[CrossRef][Medline].
|
| 7.
|
Domingo, E., and J. J. Holland.
1997.
RNA virus mutations and fitness for survival.
Annu. Rev. Microbiol.
51:151-178[CrossRef][Medline].
|
| 8.
|
Goudsmit, J.,
A. de Ronde,
E. de Rooij, and R. de Boer.
1997.
Broad spectrum of in vivo fitness of human immunodeficiency virus type 1 subpopulations differing at reverse transcriptase codons 41 and 215.
J. Virol.
71:4479-4484[Abstract].
|
| 9.
|
Hertogs, K.,
M. P. de Bethune,
V. Miller,
T. Ivens,
P. Schel,
A. van Cauwenberge,
E. C. van den Eynde,
V. van Gerwen,
H. Azijn,
M. van Houtte,
F. Peeters,
S. Staszewski,
M. Conant,
S. Bloor,
S. Kemp,
B. Larder, and R. Pauwels.
1998.
A rapid method for simultaneous detection of phenotypic resistance to inhibitors of protease and reverse transcriptase in recombinant human immunodeficiency virus type 1 isolates from patients treated with antiretroviral drugs.
Antimicrob. Agents Chemother.
42:269-276[Abstract/Free Full Text].
|
| 10.
|
Hirsch, M. S.,
B. Conway,
R. T. D'Aquila,
V. A. Johnson,
F. Brun-Vezinet,
B. Clotet,
L. M. Demeter,
S. M. Hammer,
D. M. Jacobsen,
D. R. Kuritzkes,
C. Loveday,
J. W. Mellors,
S. Vella, and D. D. Richman.
1998.
Antiretroviral drug resistance testing in adults with HIV infection: implications for clinical management. International AIDS Society USA Panel.
JAMA
279:1984-1991[Abstract/Free Full Text].
|
| 11.
|
Ho, D. D.,
A. U. Neumann,
A. S. Perelson,
W. Chen,
J. M. Leonard, and M. Markowitz.
1995.
Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection.
Nature
373:123-126[CrossRef][Medline].
|
| 12.
|
Kellam, P.,
C. A. Boucher, and B. A. Larder.
1992.
Fifth mutation in human immunodeficiency virus type 1 reverse transcriptase contributes to the development of high-level resistance to zidovudine.
Proc. Natl. Acad. Sci. USA
89:1934-1938[Abstract/Free Full Text].
|
| 13.
|
Keulen, W.,
C. Boucher, and B. Berkhout.
1996.
Nucleotide substitution patterns can predict the requirements for drug-resistance of HIV-1 proteins.
Antivir. Res.
31:45-57[CrossRef][Medline].
|
| 14.
|
Koot, M.,
A. B. van't Wout,
N. A. Kootstra,
R. E. de Goede,
M. Tersmette, and H. Schuitemaker.
1996.
Relation between changes in cellular load, evolution of viral phenotype, and the clonal composition of virus populations in the course of human immunodeficiency virus type 1 infection.
J. Infect. Dis.
173:349-354[Medline].
|
| 15.
|
Losos, J. B.,
T. R. Jackman,
A. Larson,
K. Queiroz, and L. Rodriguez-Schettino.
1998.
Contingency and determinism in replicated adaptive radiations of island lizards.
Science
279:2115-2118[Abstract/Free Full Text].
|
| 16.
|
Mansky, L. M., and H. M. Temin.
1995.
Lower in vivo mutation rate of human immunodeficiency virus type 1 than that predicted from the fidelity of purified reverse transcriptase.
J. Virol.
69:5087-5094[Abstract].
|
| 16a.
|
Maree, A. F. M.,
W. Keulen,
C. A. B. Boucher, and R. J. De Boer.
2000.
Estimating relative fitness in viral competition experiments.
J. Virol.
74:11067-11072[Abstract/Free Full Text].
|
| 17.
|
Nájera, I.,
Á. Holguín,
M. E. Quiñones-Mateu,
M. Á. Muñoz-Fernández,
R. Nájera,
C. López-Galíndez, and E. Domingo.
1995.
pol gene quasispecies of human immunodeficiency virus: mutations associated with drug resistance in virus from patients undergoing no drug therapy.
J. Virol.
69:23-31[Abstract].
|
| 18.
|
Nijhuis, M.,
C. A. Boucher, and R. Schuurman.
1995.
Sensitive procedure for the amplification of HIV-1 RNA using a combined reverse-transcription and amplification reaction.
BioTechniques
19:178-180[Medline].
|
| 19.
|
Ruiz-Jarabo, C. M.,
A. Arias,
E. Baranowski,
C. Escarmis, and E. Domingo.
2000.
Memory in viral quasispecies.
J. Virol.
74:3543-3547[Abstract/Free Full Text].
|
| 20.
|
Tyagi, S.,
D. P. Bratu, and F. R. Kramer.
1998.
Multicolor molecular beacons for allele discrimination.
Nat. Biotechnol.
16:49-53[CrossRef][Medline].
|
| 21.
|
Yerly, S.,
A. Rakik,
S. K. De Loes,
B. Hirschel,
D. Descamps,
F. Brun-Vezinet, and L. Perrin.
1998.
Switch to unusual amino acids at codon 215 of the human immunodeficiency virus type 1 reverse transcriptase gene in seroconvertors infected with zidovudine-resistant variants.
J. Virol.
72:3520-3523[Abstract/Free Full Text].
|
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
Present address: Primagen, 1105 BA Amsterdam, The Netherlands.
