Genetic and Phenotypic Analyses of Human Immunodeficiency Virus Type 1 Escape from a Small-Molecule CCR5 Inhibitor

Abstract

We have described previously the generation of an escape variant of human immunodeficiency virus type 1 (HIV-1), under the selection pressure of AD101, a small molecule inhibitor that binds the CCR5 coreceptor (A. Trkola, S. E. Kuhmann, J. M. Strizki, E. Maxwell, T. Ketas, T. Morgan, P. Pugach, S. X. L. Wojcik, J. Tagat, A. Palani, S. Shapiro, J. W. Clader, S. McCombie, G. R. Reyes, B. M. Baroudy, and J. P. Moore, Proc. Natl. Acad. Sci. USA 99:395-400, 2002). The escape mutant, CC101.19, continued to use CCR5 for entry, but it was at least 20,000-fold more resistant to AD101 than the parental virus, CC1/85. We have now cloned the env genes from the the parental and escape mutant isolates and made chimeric infectious molecular clones that fully recapitulate the phenotypes of the corresponding isolates. Sequence analysis of the evolution of the escape mutants suggested that the most relevant changes were likely to be in the V3 loop of the gp120 glycoprotein. We therefore made a series of mutant viruses and found that full AD101 resistance was conferred by four amino acid changes in V3. Each change individually caused partial resistance when they were introduced into the V3 loop of a CC1/85 clone, but their impact was dependent on the gp120 context in which they were made. We assume that these amino acid changes alter how the HIV-1 Env complex interacts with CCR5. Perhaps unexpectedly, given the complete dependence of the escape mutant on CCR5 for entry, monomeric gp120 proteins expressed from clones of the fully resistant isolate failed to bind to CCR5 on the surface of L1.2-CCR5 cells under conditions where gp120 proteins from the parental virus and a partially AD101-resistant virus bound strongly. Hence, the full impact of the V3 substitutions may only be apparent at the level of the native Env complex.

Full text

JOURNAL OF VIROLOGY, Mar. 2004, p. 2790–2807 Vol. 78, No. 6
0022-538X/04/$08.00⫹0 DOI: 10.1128/JVI.78.6.2790–2807.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Genetic and Phenotypic Analyses of Human Immunodeficiency Virus
Type 1 Escape from a Small-Molecule CCR5 Inhibitor†
Shawn E. Kuhmann,
1
Pavel Pugach,
1
Kevin J. Kunstman,
2
Joann Taylor,
2
Robyn L. Stanfield,
3
Amy Snyder,
1
Julie M. Strizki,
4
Janice Riley,
4
Bahige M. Baroudy,
4
Ian A. Wilson,
3
Bette T. Korber,
5,6
Steven M. Wolinsky,
2
and John P. Moore
1
*
Department of Microbiology and Immunology, Weill Medical College of Cornell University, New York, New York 10021
1
;
Department of Medicine, The Feinberg Medical School, Northwestern University, Chicago, Illinois 60611
2
; Department of
Molecular Biology, The Scripps Research Institute, La Jolla, California 92037
3
; Schering Plough Research Institute,
Kenilworth, New Jersey 07033
4
; Theoretical Biology and Biophysics Group, Los Alamos National Laboratory,
Los Alamos, New Mexico 87545
5
; and Santa Fe Institute, Santa Fe, New Mexico 87501
6
Received 21 August 2003/Accepted 18 November 2003
We have described previously the generation of an escape variant of human immunodeficiency virus type 1
(HIV-1), under the selection pressure of AD101, a small molecule inhibitor that binds the CCR5 coreceptor (A.
Trkola, S. E. Kuhmann, J. M. Strizki, E. Maxwell, T. Ketas, T. Morgan, P. Pugach, S. X. L. Wojcik, J. Tagat,
A. Palani, S. Shapiro, J. W. Clader, S. McCombie, G. R. Reyes, B. M. Baroudy, and J. P. Moore, Proc. Natl.
Acad. Sci. USA 99:395–400, 2002). The escape mutant, CC101.19, continued to use CCR5 for entry, but it was
at least 20,000-fold more resistant to AD101 than the parental virus, CC1/85. We have now cloned the env genes
from the the parental and escape mutant isolates and made chimeric infectious molecular clones that fully
recapitulate the phenotypes of the corresponding isolates. Sequence analysis of the evolution of the escape
mutants suggested that the most relevant changes were likely to be in the V3 loop of the gp120 glycoprotein.
We therefore made a series of mutant viruses and found that full AD101 resistance was conferred by four amino
acid changes in V3. Each change individually caused partial resistance when they were introduced into the V3
loop of a CC1/85 clone, but their impact was dependent on the gp120 context in which they were made. We
assume that these amino acid changes alter how the HIV-1 Env complex interacts with CCR5. Perhaps
unexpectedly, given the complete dependence of the escape mutant on CCR5 for entry, monomeric gp120
proteins expressed from clones of the fully resistant isolate failed to bind to CCR5 on the surface of L1.2-CCR5
cells under conditions where gp120 proteins from the parental virus and a partially AD101-resistant virus
bound strongly. Hence, the full impact of the V3 substitutions may only be apparent at the level of the native
Env complex.
Several members of a new class of inhibitors based on block-
ing human immunodeficiency virus type 1 (HIV-1) entry into
target cells are now in, or approaching, human clinical trials (8,
52, 77, 80, 85, 90, 98). These various compounds antagonize
different stages in the multistep pathway by which HIV-1 fuses
with susceptible cells. For example, the CD4-immunoglobulin
G2 protein (CD4-IgG2; PRO 542) attaches to the viral enve-
lope glycoprotein gp120 to prevent it from interacting with the
CD4 receptor (3, 109). The T-20 and T-1249 peptides act later,
by inhibiting conformational changes in the viral gp41 glycop-
rotein that are necessary for membrane fusion to be initiated
(8, 40, 72, 115). All these inhibitors cause plasma viremia
reductions in HIV-1-infected people (8, 49, 53, 62, 66, 71, 80).
A step in the entry process intermediate between gp120-
CD4 attachment and gp41 conformational changes involves a
coreceptor for gp120 (21, 31, 37, 85, 98). Thus, after gp120 has
bound to CD4, it changes conformation to enable it to bind to
a coreceptor from the G-protein-coupled receptor superfamily
(21, 31, 37, 85, 98, 107, 116). The most physiologically relevant
coreceptors are the chemokine receptors CCR5 or CXCR4,
the former used by HIV-1 strains that usually dominate early in
infection and the latter used by viruses that sometimes emerge
several years later or that are detectable only transiently (21,
31, 85, 99, 122). The presence of viruses able to use CXCR4
(X4 strains) is associated with an accelerated disease course,
due in part to the loss of naive CD4
⫹
T cells that express
CXCR4 but not CCR5 (32, 44, 69). Viruses using CCR5 (R5
strains) target memory CD4
⫹
CCR5
⫹
T cells and are lethal in
their own right (32, 44, 57, 69).
Both CCR5 and CXCR4 are important targets for pharma-
cological intervention, and several inhibitors have been iden-
tified that are specific for each receptor (8, 51, 52, 77, 80, 85,
90, 98). Reductions in the amount of plasma X4 viruses were
observed during trials of the CXCR4 inhibitor AMD3100, but
the clinical development of this particular compound has been
discontinued because of pharmacological and toxicology con-
cerns (26). A small-molecule CCR5 inhibitor, SCH-C (SCH
351125), has caused significant viral load reductions in ongoing
phase I clinical trials (51, 71, 101).
It is inevitable that HIV-1 will escape from the selective
pressure exerted by any single replication inhibitor (65).
* Corresponding author. Mailing address: Department of Microbi-
ology and Immunology, Joan and Sanford I. Weill Medical College of
Cornell University, 1300 York Ave., W-805, New York, NY 10021.
Phone: (212) 746-4462. Fax: (212) 746-8340. E-mail: jpm2003@med
.cornell.edu.
† Contribution la-ur-03-5893 from the Los Alamos National Labo-
ratory.
2790

Hence, it is prudent to study escape pathways in vitro in order
to learn what might happen in clinical use. A particular con-
cern with CCR5 inhibitors is that they might drive the evolu-
tion of X4 variants in vivo (35, 51, 73, 77). We have previously
reported on the in vitro escape of an R5 HIV-1 isolate, CC1/
85, from AD101 (SCH 350581), a CCR5 inhibitor structurally
related to SCH-C (51, 101, 108). The escape variant continued
to use CCR5 rather than acquiring the ability to enter cells via
CXCR4 or any alternative coreceptor present in human pe-
ripheral blood mononuclear cells (PBMC) (108). In contrast,
X4 viruses later evolved naturally in the individual from whom
CC1/85 was isolated (23, 24).
We have now analyzed the sequence changes in the env gene
of HIV-1 CC1/85 that correlate temporally with the evolution
of AD101 escape mutants and have contrasted them with those
occurring during the evolution of X4 variants of the same virus
in vivo. The most relevant amino acid changes, i.e., those that
arose contemporaneously with increasing levels of resistance in
vitro, seemed likely to be four charges within the V3 loop of
gp120. We confirmed their relevance by reconstituting the
AD101-resistant phenotype in an NL4-3/env chimeric provirus
and then making a series of point substitutions. Individually,
each V3 change conferred partial resistance when introduced
into an AD101-sensitive clone of CC1/85; collectively, the four
changes were both necessary and sufficient for high-level resis-
tance. The use of a gp120-CCR5 binding assay suggests that
these amino acid substitutions alter the way in which HIV-1
interacts with CCR5, albeit in a manner that remains to be
identified.
MATERIALS AND METHODS
Reagents. AD101 and SCH-C were synthesized by Schering-Plough Research
Institute (Kenilworth, N.J.) (101, 108). RANTES was from Peprotech Inc.
(Rocky Hill, N.J.). The extracellular portion of CD4 encompassing domains D1
to D4 (sCD4) (27), the CD4-IgG2 (PRO 542) molecule (3), the murine anti-
CCR5 monoclonal antibody (MAb) PA14 (81), and recombinant JR-FL gp120
(107) were gifts from Bill Olson (Progenics Pharmaceuticals Inc., Tarrytown,
N.Y.). The human anti-gp120 MAb 17b (102) was provided by James Robinson
(Tulane University).
Cloning and nomenclature of env genes. The R5 HIV-1 isolate CC1/85 was
derived in January 1985 from individual “Case C”(23, 24). The escape of CC1/85
from AD101, the cloning of full-length env genes from infected PBMC, and a
preliminary analysis of their sequences have all been described (108). env clones
were named according to our previous nomenclature (108). Thus, CC1/85 is the
parental isolate and CC101.XX refers to the isolate made after XX passages
under selection pressure from AD101. Similarly, CCcon.XX refers to the control
virus that was cultured for XX passages in the same PBMC but without AD101.
The clones generated and used here extend this nomenclature, such that CC1/85
cl.XX refers to clone XX from the CC1/85 isolate. The nomenclature used to
designate other clones was also derived from that used elsewhere (24, 108).
Additional env genes were PCR amplified and cloned into the vector pAMP1
by using the CloneAmp system as specified by the manufacturer (Invitrogen,
Carlsbad, Calif.). The nested PCR products were generated from genomic DNA
from infected PBMC, as described previously (108). The outer primers were
5958F (5⬘-GGCTTAGGCATCTCCTATGGCAGGAGG AA-3⬘) and 9100R
(5⬘-TAGCCCTTCCAGTCCCCCCTTTTCTTTTA-3⬘), and the inner primers
were 6205F (5⬘-AGAAAGAGCAGAAGACAGTGGCAATGA-3⬘) and 8821R
(5⬘-TTTTGACCACTTGCCACCCAT-3⬘).
CC101.22 was isolated after 22 passages with AD101 in vitro and then pas-
saged 9 more times without AD101 (108). Isolates made during this secondary
culture were designated CC101.22RX, where X is the number of passages made
in the absence of AD101. The corresponding clones were named using the
system outlined above.
To facilitate comparisons with previous reports on structure-function relation-
ships in HIV-1 Env and to follow convention in the field, amino acid numbering
for all Env proteins was based on the numbering of the prototypic HXBc2 Env
(http://hiv-web.lanl.gov/content/hiv-db/LOCATE_SEQ/locate.html) (56).
Phylogenetic sequence analysis. Phlyogenetic analysis and bootstraps were
performed using the neighbor-joining method available in PHYLIP (Phylogeny
Inference Package, written by J. Felsenstein in 1993, version 3.6, distributed by
the author [http://evolution.genetics.washington.edu/phylip.html]), with the F84
model for base changes and a transition/transversion ratio of 2. In the final tree,
there were 243 sequences, with 2,416 bases left in the alignment after gap
stripping. The sequence from the initial sample most closely resembling those
that emerged in culture after AD101 selection (CC1/85 cl.7) was selected as an
outgroup for the tree; the use of this sequence from the initial sample gave an
orientation of the tree that generally radiated outward from the time of sampling
when drawn as a phenogram. Because of the reasonable outcome of the initial
neighbor-joining phylogenetic analysis and because the large number of se-
quences in this study make likelihood methods computationally very expensive,
the neighbor-joining method was deemed adequate for the purposes of this
study.
To exclude problematic sequences, representative sequences from each sam-
ple and a few sequences that gave peculiar outcomes in an initial phylogenetic
analysis (i.e., strikingly long branch lengths relative to the rest of the sample)
were carefully screened to explore the integrity of the original set of 249 se-
quences. A BLAST search of the HIV database (http://hiv-web.lanl.gov/content
/hiv-db/BASIC_BLAST/basic_blast.html) (4) revealed that one clone was nearly
identical to the common reference strain NL4-3; it was therefore excluded as a
probable contaminant. That NL4-3 clone was then used to probe the rest of the
sequence set, identifying three clones containing short stretches with regional
identity to NL4-3, including distinctive insertion and deletion patterns. These
three clones were also excluded because they were probably PCR recombinants
with the NL4-3 contaminant. Screening for hypermutation was performed using
the HYPERMUT program (http://hlv-web.lanl.gov/content/hiv-db/HYPER-
MUT/hypermut.html) (95), revealing one clone hypermutated in its C⬘-terminal
half, beyond position 2000; it, too, was excluded. Finally, one clone with an
extensive deletion near V3 was excluded. The other 243 sequences were deemed
to be valid after rigorous quality control testing.
Construction of chimeric NL4-3/env proviruses. Chimeric proviruses were
constructed from the pNL4-3 proviral plasmid (2) (AIDS Research and Refer-
ence Reagent Program, National Institute of Allergy and Infectious Diseases;
contributed by Malcolm Martin) by overlapping PCR. The gp160 coding se-
quences were amplified from the cloning vectors by using primers EnvF (5⬘-AG
CAGAAGACAGTGGCAATGAGAGTGAAG-3⬘) and EnvR (5⬘-TTTTGACC
ACTTGCCACCCATCTTATAGC-3⬘). A portion of the NL4-3 provirus from
nucleotides 5284 to 6232 was amplified with primers NL(5284)F (5⬘-GGTCAG
GGAGTCTCCATAGAATGGAGG-3⬘) and NL(6232)R (5⬘-CTTCACTCTCA
TTGCCACTGTCTTCTGCT-3⬘). This fragment encompasses the unique EcoRI
restriction site in pNL4-3. Another fragment from the NL4-3 provirus spanning
nucleotides 8779 to 9045 was amplified using primers NL(8779)F (5⬘-GCTATA
AGATGGGTGGCAAGTGGTCAAAA-3⬘) and NL(9045)R (5⬘-GATCTACA
GCTGCCTTGTAAGTCATTGGTC-3⬘). This fragment includes the unique
XhoI restriction site in pNL4-3. Overlapping PCR was used to join the gp160-
coding sequence from the desired clone to the fragment encompassing bases
8779 to 9045 that had been amplified from pNL4-3. The resulting fragment was
then similarly joined to the amplified fragment encompassing bases 5284 to 6232
from pNL4-3. The product was digested with EcoRI and XhoI and subcloned
into the corresponding sites in pBluescript KS(⫹) (Stratagene) for sequencing
and subsequent manipulation. The EcoRI-XhoI fragment was sequenced after
this and all subsequent manipulation steps to confirm the presence of the desired
env gene and the absence of other changes. The EcoRI-XhoI fragment for each
env gene was then subcloned back into pNL4-3. The end results were proviral
plasmids that differ from each other only in the env gene.
Construction of chimeric and mutant env genes. We selected one clone from
the parental CC1/85 isolate (CC1/85 cl.7) and one from the AD101-resistant
CC101.19 isolate (CC101.19 cl.7) for more detailed studies of genotype-pheno-
type associations. To generate constructs with chimeric env genes, pBluescript
KS(⫹) plasmids containing these env fragments were cut with StuI and BsaBI.
The StuI-BsaBI fragment from CC101.19 cl.7 was then ligated into the vector
fragment from CC1/85 cl.7 that had been cut with the same enzymes. This
chimeric construct is referred to as CC1/85 cl.7(8), the number 8 signifying that
it differs from CC1/85 cl.7 by 8 amino acid changes. The reverse chimeric env
gene, with the StuI-BsaBI fragment from CC1/85 cl.7 inserted into CC101.19 cl.7,
was also constructed and is referred to as CC101.19 cl.7(8).
Site-directed mutagenesis was performed with the QuickChange mutagenesis
kit as specified by the manufacturer (Stratagene), using the pBluescript KS(⫹)
VOL. 78, 2004 HIV-1 ESCAPE FROM A CCR5 ANTAGONIST 2791

plasmids containing the EcoRI-XhoI fragments. EcoRI-XhoI fragments were
then subcloned into pNL4-3.
Cells, viruses, and viral replication assays. PBMC were pooled from the blood
of four healthy volunteers. CD8
⫹
T cells were removed using the RosetteSep
CD8
⫹
depletion cocktail as specified by the manufacturer (StemCell Technolo-
gies, Vancouver, Canada). The CD8
⫹
T-cell-depleted PBMC (referred to here-
after as primary CD4
⫹
T cells) were then activated and used for HIV-1 repli-
cation assays 3 days later (108). Primary CD4
⫹
T cells were used to limit the
interassay variability of PBMC replication assays, but similar results were ob-
tained without CD8
⫹
T-cell depletion (data not shown).
Clonal, replication-competent, chimeric NL4-3/env viruses were prepared by
transfecting the full-length proviral plasmid into 293T cells by using Lipo-
fectamine 2000, as specified by the manufacturer (Invitrogen). Supernatants
were collected at 48 h posttransfection, filtered, and stored at ⫺80°C. Replica-
tion assays were performed with PBMC as described for standard HIV-1 isolates
(108), except that 100 50% tissue culture infective doses (TCID
50
) of virus were
used per well of a 96-well plate.
Expression and purification of recombinant gp120. Recombinant gp120 pro-
teins from clones CC1/85 cl.6 and cl.7, CC101.19 cl.3, CC101.19 cl.7, and
CC101.6 cl.10 were transiently expressed from pPPI4-gp120 expression vectors.
These were constructed by subcloning the KpnI-BbvCI fragments from the de-
sired env gene into the pPPI4-JR-FL gp140 vector (13). Two consecutive in-
frame stop codons were then introduced by QuickChange mutagenesis (Strat-
agene) immediately following the lysine in the sequence REKR, the natural
cleavage site between gp120 and gp41. Other than in the gp120 coding se-
quences, the resulting pPPI4 vectors were identical to that used to express JR-FL
gp120 (107). The gp120 proteins were expressed and purified as previously
described (13, 107). Their concentrations were measured by the DC modified
Lowry protein assay (Bio-Rad, Hercules, Calif.). Their purity was ⬎90% as
assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis.
Binding of gp120 to CCR5. Capture enzyme-linked immunosorbent assays for
gp120 binding to sCD4 and MAb 17b were performed as described elsewhere
(76). To detect gp120 binding to cell surface CCR5, L1.2-CCR5 cells were
incubated with both purified gp120 (100 ␮g/ml) and biotinylated CD4-IgG2 (100
␮g/ml) for1hat37°C (34, 111). Bound CD4-IgG2 was then detected by staining
for 30 min at 25°C with phycoerythrin-conjugated streptavidin (BD BioSciences,
San Jose, Calif.). After being washed the cells were fixed in paraformaldehyde
and analyzed by flow cytometry.
Nucleotide sequence accession numbers. The 243 sequences that were deemed
valid have been submitted to GenBank (accession numbers AY357338 through
AY357580).
RESULTS
Changes in Env are responsible for CCR5 inhibitor resis-
tance. We reported previously that a primary R5 isolate, CC1/
85, became increasingly resistant to the CCR5 inhibitor AD101
on passaging in PBMC cultures in the presence of increasing
AD101 concentrations. After 19 passages, AD101 resistance
was substantial (⬎20,000-fold) (108). To confirm that resis-
tance was conferred by changes in the viral envelope glyco-
proteins, we cloned env genes from the resistant virus
(CC101.19) and used them to replace the env gene of the
infectious molecular clone, HIV-1 NL4-3. Similarly, we made
chimeric NL4-3 viruses by using env genes cloned from the
parental virus, CC1/85, to serve as controls. Seven clonal, chi-
meric infectious viruses were chosen for further study:
CC101.19 cl.3, CC101.19 cl.7, and CC101.19 cl.15 from the
passage 19 escape mutant isolate, CC101.19; CC1/85 cl.6,
CC1/85 cl.7, and CC1/85 cl.8 from the parental isolate, CC1/85;
and CCcon.20 cl.11 from the passage 20 control isolate,
CCcon.20.
The clonal viruses were confirmed to be replication compe-
tent in both PBMC and primary CD4
⫹
T cells. Their titer in
primary CD4
⫹
T cells on day 12 postinfection ranged from 10
4
to 10
5
TCID
50
/ml, when p24 production was typically 5 to 10
ng/ml. There was no major difference in the replication rates of
the CC1/85 or CC101.19 clones in primary CD4
⫹
T cells, and
the CC101.19 clones did not require AD101 for replication
(data not shown). Although we have not yet performed formal
fitness studies, the AD101 escape mutant isolates and clones
appear to be fully replication competent in primary CD4
⫹
T
cells.
The parental and escape mutant clones faithfully recapitu-
lated the phenotypes of the corresponding isolates. Thus, in a
representative experiment, all three clones derived from
CC101.19 were not inhibited by AD101 even at concentrations
as high as 20 ␮M whereas all three clones from the parental
isolate CC1/85 were sensitive to AD101, with an average 50%
inhibitory concentration (IC
50
)of⬃2 pM (Fig. 1a). Hence in
this experiment, the extent of resistance for the CC101.19-
derived clones, compared to clones from CC1/85, was ⬎10
7
-
fold. For all practical purposes, the CC101.19 clones are there-
fore completely resistant to the selecting compound. Clone
CCcon.20 cl.11 from the passage 20 control isolate also re-
tained the AD101-sensitive phenotype (data not shown).
The CC101.19 escape mutant isolate is completely cross-
resistant to SCH-C (108), a small-molecule CCR5 inhibitor
chemically related to AD101 (82, 83, 101, 103, 104, 108). We
therefore determined if the derived clones CC101.19 cl.3 and
cl.7 were also cross-resistant. Neither clone was inhibited by
SCH-C at the highest concentration tested (20 ␮M), whereas
both parental clones, CC1/85 cl.6 and cl.7, were sensitive to
SCH-C, with IC
50
s in the range 30 to 100 nM (Fig. 1b). The
escape mutant clones were therefore completely cross-resistant
to SCH-C, whereas the parental isolate clones retained the
sensitivity of the input CC1/85 virus.
To see if the resistance of the CC101.19 clones to the small-
molecule inhibitors reflected a more general resistance to
CCR5-targeted agents, we evaluated their sensitivity to the CC
chemokine RANTES and to a murine anti-CCR5 MAb, PA14
(81). We previously showed that the CC101.19 isolate was
⬃10-fold more resistant to RANTES than was the parental
CC1/85 isolate (108). However, the CC101.19 clones were at
least as sensitive as the CC1/85 clones to RANTES (Fig. 1c).
The CCR5 epitope for MAb PA14 has been mapped to extra-
cellular loop (ECL) 2 and the N terminus (NT) of CCR5 (81).
The CC101.19 isolate was ca. fivefold less sensitive than
CC1/85 to PA14 (108). The pattern of sensitivities of the cor-
responding CC1/85 and CC101.19 clones toward PA14 was
similar to what was observed with RANTES, in that all the
clones were inhibited by similar PA14 concentrations (Fig. 1d).
Hence, escape from the small-molecule inhibitor AD101 does
not involve the acquisition of strong resistance to other classes
of entry inhibitors that also target CCR5. This conclusion ap-
plies to both the CC101.19 isolate (108) and individual clones
derived from it (Fig. 1). Any phenotypic differences between
the CC101.19 isolate and its derivative clones are minor; pre-
sumably some CC101.19 clones will be modestly resistant to
RANTES and PA14, just as the isolate is (108). It is also
possible that minor differences in phenotype observed only
with the isolates, and not with the clones, map to determinants
outside the env gene, such as Gag regions that might influence
Env incorporation into the chimeric NL4-3-based viruses.
The sensitivities of the clonal escape mutant viruses to
RANTES and PA14 implies that, like the escape mutant iso-
late and the parental CC1/85 virus, they have an absolute
2792 KUHMANN ET AL. J. VIROL.

dependence on CCR5 for entry (108). We confirmed this by
showing that none of the NL4-3/env clones could replicate to a
detectable extent in PBMC from a CCR5-⌬32 homozygous
individual whereas all the clones replicated efficiently in PBMC
from CCR5 wild-type donors (data not shown).
Env sequence evolution associated with the generation of
AD101 resistance. Multiple passages in PBMC culture were
required to generate an escape mutant with full resistance to
AD101 (108). Thus, the CC101.19 isolate acquired over
20,000-fold resistance to AD101 after 19 weekly passages. Par-
tial resistance was detected sooner; after six passages, the
CC101.6 isolate was approximately fourfold less sensitive than
CC1/85 to AD101 (108). The full escape mutant had a stable
phenotype, since AD101 resistance survived nine further pas-
sages in PBMC culture in the absence of AD101 (108). These
observations, taken together, suggest that multiple, sequential
genetic changes had conspired to create full resistance to
AD101 (108). To learn how escape had occurred, we studied
Env sequence evolution by using samples of virus-infected cells
frozen from each stage of the long-term PBMC culture and its
extension in the absence of AD101.
Multiple sequence changes had accumulated in CC1/85
gp120 over time, but gp41 changes were rare and inconsistent
(see below). In a phylogenetic analysis, sequences from the 4-,
10-, and 20-week passage control viruses (CCcon.4, CCcon.10,
and CCcon.20, respectively) clustered with the parental
CC1/85 sequences and had a similar degree of diversity (Fig.
2). In contrast, the viruses isolated under the AD101 selection
pressure branched off to form a distinct lineage with very little
internal sequence diversity (Fig. 2). Sequences from escape
mutant viruses cultured for nine further passages without
AD101 were clustered within the same lineage, suggesting that
they had changed little during the extension to the culture (Fig.
2).
For comparison, we also analyzed gp160 sequences from
viruses that had diversified in vivo from the same genetic lin-
eage. The parental virus for the escape mutant study, CC1/85,
was isolated in January 1985 from an individual who, at that
time, harbored only R5 strains (23, 24). Three isolates ob-
tained 13, 18, and 23 months later had the R5X4 phenotype
and are designated CC2/86, CC7/86, and CC12/86 (23, 24). A
precipitous decline in CD4
⫹
T-cell counts occurred during this
23-month period (23, 24). Sequences from these R5X4 isolates
showed a gradual progression to new forms but on a lineage
FIG. 1. Infectious molecular clones recapitulate the phenotype of the AD101-resistant escape mutant CC101.19. (a) Chimeric molecular clones
containing env genes derived from the parental CC1/85 isolate (open symbols) or the passage 19 escape mutant (solid symbols) were cultured in
primary CD4
⫹
T cells in the presence of the AD101 concentrations indicated, and the extent of HIV-1 replication was determined. The individual
clones are designated CC1/85 cl.6 (open squares), CC1/85 cl.7 (open circles), CC1/85 cl.8 (open triangles), CC101.19 cl.3 (solid squares), CC101.19
cl.7 (solid circles), and CC101.19 cl.15 (solid triangles). (b to d) Four of the above clones (same symbols) were also tested for sensitivity to SCH-C
(b), RANTES (c), and PA14 (d). Panel a displays data from a single representative experiment. Panels b to d show data averaged from three
independent experiments, with the error bars indicating the standard error of the mean (SEM).
VOL. 78, 2004 HIV-1 ESCAPE FROM A CCR5 ANTAGONIST 2793

FIG. 2. Phylogenetic analysis of gp120 sequences from CC1/85 and related viruses. Sequence diversity is represented by horizontal distance.
CC1/85 sequences are represented by black bars. Sequences from the CCcon.4 (passage 4), CCcon.10 (passage 10), and CCcon.20 (passage 20)
isolates are shown by blue bars, with the darker bars denoting the later isolates. The sequences from viruses isolated under AD101 selection
pressure are represented in yellow, orange, and red. The sequences in green represent in vivo diversification that generates the CC2/86, CC7/86,
and CC12/86 isolates. The CC101.22R2, CC101.22R5, and CC101.22R9 sequences in different shades of purple are derived from escape mutant
viruses cultured in the absence of AD101 for a further nine passages.
2794 KUHMANN ET AL. J. VIROL.

distinct from those of CC1/85 and the escape mutants selected
in vitro (Fig. 2).
Sequence changes in the V3 loop during the development of
AD101 resistance. Inspection of the predicted gp160 sequences
from throughout the course of AD101 selection revealed a
notable pattern of evolution within the gp120 V3 region (Fig.
3). One clone from the eight sequenced from CC1/85 con-
tained a single amino acid polymorphism; a proline residue
was present instead of the predominant histidine at position
308 (H308P) in the N-terminal side of the V3 loop. By passage
2 in the presence of AD101 (the CC101.2 clones), only a
proline was found at position 308; the originally dominant
histidine was completely absent (Fig. 3). With the exception of
a few minor polymorphisms, the V3 sequence of the passage 2
clones then remained unchanged until passage 8, when an
additional polymorphism (K305R) emerged nearby (Fig. 3).
This change, observed in 5 of the 19 sequenced CC101.8
clones, presumably arose by de novo mutation, since arginine
was not present at position 305 in any of the 41 clones from
earlier time points. By passage 10, 12 of 13 clones contained
arginine at position 305, and after pass as 12, arginine was the
only amino acid found at this position (Fig. 3). Two more
changes were observed in passage 14 clones, A316V and
G321E, again presumably arising by mutation. V3 sequences
carrying all four substitutions dominated from passage 16 on-
ward and were the only ones present by passage 18. Although
other scattered substitutions occurred in the V3 loop and
throughout gp160, none of them showed this pattern of stabi-
lization under selection (Fig. 3 and 4 and data not shown).
The H308P polymorphism was present in three of eight
sequences from the CC2/86 R5X4 isolate (Fig. 3). The A316V
substitution was absent from CC1/85 sequences but was evi-
dent in sequences from the CC2/86, CC7/86, and CC12/86
R5X4 isolates (Fig. 3). There, it appears in the context of a
very different V3 loop sequence and in a different lineage in
FIG. 3. V3 sequence alignments of Env clones. The first line shows
the consensus V3 sequence of the CC1/85 isolate based on eight
clones. The locations of specific amino acids and their corresponding
amino acid numbers, based on the HXBc2 sequence, are shown above
the consensus line. There are five blocks of sequence data below the
consensus line. Each block includes the following information (from
left to right): N, the number of clones from a given isolate having the
indicated V3 sequence; the source isolate, i.e., the isolate, based on the
nomenclature described in the text, from which the clones were gen-
erated; and the sequence. The sequence is shown in reference to the
consensus, with a dash representing identity to the consensus. Where
one or more clones differ from the consensus, the amino acid found in
that clone or clones is shown. Block 1 contains the eight clones from
the CC1/85 isolate, seven of which are identical to the consensus and
one of which contains the H308P polymorphism. This polymorphism
and others discussed in the main text are highlighted in red. Block 2
contains select isolates generated from the cultures treated with
AD101; the clones are shown first in order of the isolate from which
they were derived and then in order of the frequency at which each
clone is present in that isolate. Block 3 contains clones derived from
isolates cultured for 22 passages before AD101 was removed for an
additional 9 passages. Block 4 contains clones from three control
isolates passaged in the absence of AD101. Block 5 contains clones
from isolates derived from patient Case C at 13 (CC2/86), 18 (CC7/86),
and 23 (CC12/86) months after the CC1/85 isolate. The $ symbol in
one CC101.14 clone indicates the position of a premature stop codon.
VOL. 78, 2004 HIV-1 ESCAPE FROM A CCR5 ANTAGONIST 2795

FIG. 4. Alignments of the consensus gp160 sequences from critical time points in the AD101 escape process. The consensus amino acid
sequence from the parental isolate, CC1/85, is given on the top line. Capital letters indicate that there was identity among all eight of the aligned
clones. Lowercase letters indicate the amino acid sequence of most but not all of the clones. The question mark indicates either that there was
a tie between amino acids at a given position or that the majority of clones had a gap at this position. A gap found in all clones is indicated by
2796 KUHMANN ET AL. J. VIROL.

the phylogenetic tree with strong bootstrap support (Fig. 2 and
3). Hence, it most probably arose independently both in vivo
and in the in vitro escape lineage.
None of the above four V3 changes was ever seen in clones
of the control isolates from passages 4, 10, and 20, except for
the H308P polymorphism, which was present in 1 of 11 clones
from the CCcon.10 control isolate (Fig. 3). The latter obser-
vation is consistent with the hypothesis that H308P is a minor,
preexisting CC1/85 variant that was rapidly selected for in the
presence of AD101 but was not under selection pressure in the
control cultures. The V3 sequences from the passage 20 con-
trol clones also lacked the four V3 changes, although another
V3 polymorphism, F317L, was present in 9 of 12 CCcon.20
clones (Fig. 3). This particular polymorphism was absent from
clones from CC1/85 or the other passage control. isolates (Fig.
3). Its significance, if any, is unclear.
The four amino acid changes in V3 remained the dominant
form in the escape mutant viruses that had been cultured for 22
passages in the presence of AD101 and then for up to 9 addi-
tional passages without AD101 (designated CC101.22RX), al-
though some variants were evident as minor forms (Fig. 3).
Clearly, as with the AD101 escape phenotype itself (108), the
genetic changes associated with it are stable.
Sequence evolution elsewhere in gp160 during the evolution
of AD101 resistance. Other sequence changes occurred else-
where in Env as AD101 resistance developed (Fig. 4). These
changes arose with a different time course from the V3 substi-
tutions; they typically became fixed in the AD101-treated cul-
tures at or before passage 4, so their occurrence did not cor-
relate with the development of complete resistance to AD101.
It is, however, possible that some non-V3 changes may play a
role in the fourfold resistance to AD101 possessed by the
CC101.6 isolate (108). Most of the non-V3 changes are likely
to have been rapidly selected for from the input virus popula-
tion; many, but not all, of them were present as minor variants
in CC1/85 clones (Fig. 4 and data not shown).
Several distinct patterns of selection can be seen by inspec-
tion of the alignments of the consensus sequences from critical
time points (Fig. 4). Three such patterns are as follows. First,
some sequences were selected for in the AD101-treated cul-
tures but not in the control culture (CCcon.20) or the in vivo
isolates (CC12/86). These sequences are shown in red in Fig. 4.
Twelve of these changes are in gp120, and seven in gp41. Of
those in gp120, two were found in the V1 loop, three were
found in the V3 loop, three were found in the C3 region, and
four were found in the V4 loop. All the amino acid changes,
including H308P, that arose in the AD101-treated cultures but
not in the control cultures or in vivo, were selected for by
passage 4 or earlier, with the exception of the K305R and
G321E substitutions in the V3 loop (see above). Another of
these variants is part of a 4-residue insertion, VTNN, in V1
between amino acids 136 and 137. This insertion adds a new
N-linked glycosylation site at position 136. In contrast, an N-
to-K change at the first amino acid of V4 eliminates an N-
linked glycosylation site. The seven changes in gp41 are scat-
tered throughout the protein and are mostly conservative
substitutions (Fig. 4).
The second pattern observed, depicted in green in Fig. 4,
involves amino acid changes that were selected for in both the
AD101-treated and control cultures. Some of these are also
present in the in vivo sequences. Of the changes that meet this
criterion, one each can be found in C2, C3, and V4 while four
are present in V5 and four more are present in gp41 (Fig. 4).
The only substitution in this group that was not stabilized in
the AD101-treated cultures by passage 4 was D167N (Fig. 4).
This amino acid is polymorphic in the CC1/85 isolate and
remains so throughout the AD101 selection process. An Asn
residue gradually becomes more prevalent at this position but
does not become dominant until passage 20. Given the slow-
ness with which it emerged and given that it is also selected for
in the control cultures, the D167N change is not likely to be
involved in the development of AD101- resistance (Fig. 4 and
data not shown).
The final pattern seen in the consensus alignments, shown in
blue in Fig. 4, involves amino acids that were selected for both
in the AD101-treated culture and in vivo but not in the control
culture. All these changes also become fixed in the AD101-
treated cultures by passage 4, with the exception of the A316V
substitution in V3 (see above). Four of them occur in V1, three
as part of the 4-amino-acid insertion described above, one
occurs in C3, four occur in V4, and one occurs in gp41 (Fig. 4).
Overall, all the amino acid changes that occurred under
AD101 selection pressure, regardless of the pattern into which
they fall, remained the dominant sequences in the CC101.22R9
culture (Fig. 4). Hence they were stable after the AD101 se-
lection pressure was withdrawn. Unlike many of the gp120
substitutions, those within gp41 were conservative and did not
involve charge changes or alterations in glycosylation sites. The
V4 and V5 regions were highly polymorphic in the CC1/85
isolate but became fixed to essentially invariance by passage 4.
In contrast, while sequences from the control culture also lost
some variability in these regions, they did so to a much lesser
a dot. Aligned with the CC1/85 consensus are the consensus sequences from isolates from passages 4, 10, and 20 (CC101.4, CC101.10, and
CC101.20, respectively), derived under AD101 selection pressure. Also depicted are consensus sequences from the CC101.22R9 isolate, which was
cultured in the absence of AD101 for 9 passages after 22 passages in the presence of AD101; the passage control CCcon.20 isolate, which was
cultured for 20 passages in the same PBMC as the AD101-treated cultures, but in the absence of drug; and CC12/86, which was isolated from
individual Case C 23 months after the CC1/85 isolate was obtained. If the aligned sequence is identical to that of the CC1/85 consensus line, this
is indicated by a dash; otherwise, the symbols described above are used to depict differences from the consensus sequence. Amino acids highlighted
in red are those that were selected for in the AD101-treated cultures but not in the control cultures or in vivo. Amino acids highlighted in green
are those that were selected for in both the AD101-treated and control cultures. Amino acids highlighted in blue are those that were selected for
in the AD101-treated cultures and in vivo but not in the control cultures. The numbering system is based on that of HXBc2. The amino acid under
the ⬙0”of each number is the one intended to be indicated by the label. The variable and constant regions of gp120, and the gp120-gp41 cleavage
site, are also indicated. The outlined amino acids are those which have been implicated in the binding of, or neutralization by, MAb 17b (59, 94,
105, 118). Amino acids on a gray background are those implicated in forming part of a conserved coreceptor binding site that partially overlaps
the epitope for 17b (93, 94).
VOL. 78, 2004 HIV-1 ESCAPE FROM A CCR5 ANTAGONIST 2797

extent. The in vivo sequences also remained highly polymor-
phic in V4 and less so in V5 (Fig. 4 and data not shown).
Finally, and of particular significance, the selection pressure
of AD101 caused no detectable sequence variation in the
amino acids from the C4 region and the stem of the V1/V2
loops that together comprise the bridging sheet of the gp120
core, a structure which represents the most highly conserved
element of the CCR5 binding site (59, 93, 94) (Fig. 4). Mu-
tagenesis studies and the crystal structure of the complex
formed between the 17b Fab fragment, the conserved core of
gp120, and the first two domains of CD4 together indicate that
the CD4-induced epitope for MAb 17b is formed from gp120
elements that overlap the conserved CCR5 binding site (59, 94,
105, 118). The 17b epitope was also unchanged in the AD101
escape mutant clones (Fig. 4).
Changes in the V3 loop of CC1/85 confer AD101 resistance.
The availability of NL4-3/env chimeras that reproduced the
AD101-resistant and -sensitive phenotypes allowed us to in-
vestigate the genetic basis for AD101 resistance. We noted,
however, that individual CC1/85 env clones were extremely
sensitive to AD101 inhibition, significantly more so than the
corresponding isolates. The average IC
50
for a single clone,
CC1/85 cl.7, over four independent experiments, was ⬃4pM
(Table 1). In contrast, the average IC
50
for the CC1/85 isolate
was ⬃1 nM, 250-fold greater (108). The difference was not a
general property of the clonal virus stocks, because they were
no more sensitive than the CC1/85 isolate to inhibition by
SCH-C, RANTES, or PA14 (Fig. 1) (108). Instead, it presum-
ably reflects the method and other differences involved in
working with isolates and NL4-3/env chimeric clones (see be-
low). Hence, absolute values for the fold resistance of clones to
AD101 should not be directly compared with values for the
corresponding isolates. Fold resistance values for clones may,
however, be compared with values for other clones, and those
for isolates may be compared with values for other isolates.
The pattern of Env sequence evolution described above sug-
gested that genetic analyses should focus on the V3 loop.
Although amino acid changes in other regions of gp120 and
gp41 do become fixed early in the AD101 selection process
(Fig. 4), only V3 amino acids were under active selection at
times that correlated with the acquisition of complete AD101
resistance. Moreover, there is ample precedent for the V3 loop
influencing the HIV-1 phenotype, including coreceptor choice
(20–22, 31, 33, 39, 47, 48, 100, 114).
We therefore made chimeric env genes in which the coding
region for amino acids 271 to 386 of the parental CC1/85 virus
(clone CC1/85 cl.7) was replaced with the corresponding seg-
ment of the fully resistant AD101 escape mutant clone;
CC101.19 cl.7. This chimera is referred to as CC1/85 cl.7(8).
The exchanged fragment encodes part of the C2 region, the V3
and C3 regions, and the first amino acid of V4. The designation
“8”in the chimera identifier signifies that there are eight
amino acid changes between the clones in the region swapped.
Listing the CC1/85 cl.7 amino acid first, the changes were
I271V in C2; K305R, H308P, A316V, and G321E in V3;
N337Q and E351K in C3; and N386K, the first amino acid of
the V4 loop (Fig. 4; Table 2). The other three C3 polymor-
phisms (R343K, Q344H, and G379R) that were selected for in
the AD101-treated cultures were already present in CC1/85
cl.7. Hence, they are invariant among the various chimeras and
TABLE 1. IC
50
s for AD101 inhibition of NL4-3/env chimeric virus replication in primary CD4
⫹
T cells
env clone, chimera, or mutant AD101 IC
50
(nM)
a
Relative fold difference
b
CC1/85 cl.7 0.004 1
CC101.19 cl.7 ⬎20,000 ⬎5⫻10
6
,⬎
CC1/85 cl. 7(8) ⬎20,000 ⬎5⫻10
6
,⬎
CC101.19 cl.7(8) 0.6 150, ⬎
CC1/85 cl.7(H308P) 2 500, ⬎
CC1/85 cl. 7(K305R, H308P) 400 100,000, ⬎
CC1/85 cl.7(K305R, H308P, A316V, G321E) ⬎20,000 ⬎5⫻10
6
,⬎
CC1/85 cl.7(K305R) 0.0004 10, ⬍
CC1/85 cl.7(A316V, G321E) 0.0006 7, ⬍
CC1/85 cl.7(K305R, A316V, G321E) 0.000006 700, ⬍
CC101.19 cl.7(R305K, P308H, V316A, E321G) 0.02 5, ⬎
CC101.19 cl.7(P308H) 0.0002 20, ⬍
CC101.19 cl.7(R305K, P308H) 1 250, ⬎
a
Approximate IC
50
s for AD101 inhibition curves were calculated by fitting to a sigmoidal dose-response curve by nonlinear regression using the program Prism
(GraphPad Software) for the average data sets shown in Fig. 4 and 5.
b
Relative fold difference in IC
50
for AD101 compared to CC1/85 cl.7, which is used as a reference. ⬎and ⬍indicate that IC
50
is respectively greater or less than
that for CC1/85 cl.7.
TABLE 2. Comparison of amino acids exchanged in the CC1/85 cl.7(8) and CC101.19 cl.7(8) Env chimeras
Env Amino acid identity
a
at following position and location
271, C2 305, V3 308, V3 316, V3 321, V3 337, C3 343, C3 344, C3 351, C3 379, C3 386, V4
CC1/85 consensus I K H A G N R Q E G N
CC101.19 consensus V R P V E Q K H K R K
CC1/85 cl.7(8) IKHAGNKHERN
CC101.19 cl.7(8) VRPVEQKHKRK
a
The amino acids which differ between CC1/85 cl.7(8) and CC101.19 cl.7(8) are highlighted in bold type.
2798 KUHMANN ET AL. J. VIROL.

clones now under study (Fig. 4, Table 2). A reciprocal chimera
in which the corresponding region from CC1/85 cl.7 was in-
serted into CC101.19 cl.7 is referred to as CC101.19 cl.7(8).
The amino acid changes in the parental and chimeric env
clones are summarized in Table 2.
NL4-3-based viruses containing the above env genes were
used to generate clonal virus stocks, which were then tested for
AD101 sensitivity in replication assays using primary CD4
⫹
T
cells (Fig. 5). The CC101.19 cl.7(8) chimeric virus was highly
sensitive to AD101 (IC
50
,⬃0.6 nM), whereas CC1/85 cl.7(8)
resisted the highest AD101 concentration tested (20 ␮M) (Fig.
5; Table 1). In the context of these two env clones, complete
resistance to AD101 therefore maps to the interchanged gp120
fragment spanning amino acids 271 to 386. The CC101.19
cl.7(8) chimera was ⬃150-fold more resistant than CC1/85 cl.7
to AD101, with the average IC
50
s being ⬃600 and ⬃4 pM,
respectively (Fig. 5; Table 1). Hence, changes in gp120 outside
of residues 271 to 386 do make some contribution to AD101
resistance, but they cannot account for the complete resistance
of the CC101.19 clones (Fig. 5; Table 1).
Site-directed mutagenesis of specific residues in the V3 loop.
To define which specific amino acids within residues 271 to 386
can confer complete resistance to AD101, we used the AD101-
sensitive clone CC1/85 cl.7 as a basis for site-directed mutagen-
esis. This clone was chosen because, while it lacks many of the
Env substitutions that were apparently selected for during the
early passages with AD101 (i.e., prior to passage 4), including
the H308P polymorphism in the V3 loop, it is the most closely
related of the eight CC1/85 clones to the ones selected for
under pressure from AD101 (data not shown). We first
changed one or more of the four amino acids that differed
between the V3 regions of CC1/85 cl.7 and the AD101-resis-
tant clone, CC101.19 cl.7. One such mutant, CC1/85
cl.7(K305R, H308P, A316V, G321E), differs from CC1/85 by
possessing all four of the substitutions in the V3 loop high-
lighted by the sequence analysis (Fig. 3). The CC1/85 cl.7
(K305R), CC1/85 cl.7(H308P), CC1/85 cl.7(K305R, H308P),
CC1/85 cl.7(A316V, G321E), and CC1/85 cl.7(K305R, A316V,
G321E) env genes were also made; they differ from CC1/85 cl.7
at one, two, or three of the above V3 residues.
The mutated coding sequences were reconstituted into rep-
lication-competent viruses, and replication assays were per-
formed with primary CD4
⫹
T cells in the presence of increas-
ing AD101 concentrations (Fig. 6a and b). As noted above, the
IC
50
for AD101 inhibition of CC1/85 cl.7 was, on average, 4
pM. Consistent with our earlier studies of the viruses bearing
chimeric env genes (Fig. 5; Table 1) and the sequence evolu-
tion analysis (Fig. 3), the quadruple mutant CC1/85
cl.7(K305R, H308P, A316V, G321E) was completely resistant
to AD101, even at 20 ␮M (Fig. 6a; Table 1). However, the
single H308P change was alone sufficient to confer ⬃500-fold
resistance upon CC1/85 cl.7, with the IC
50
being ⬃2 nM (Fig.
6a; Table 1). A further ⬃200-fold resistance to AD101 was
created by adding the K305R change to the CC1/85
cl.7(H308P) virus to make the CC1/85 cl.7(K305R, H308P)
double mutant (Fig. 6a; Table 1). Clearly, the two further
changes (A316V and G321E) in the quadruple mutant gener-
ate yet further resistance (Fig. 6a; Table 1). Four amino acid
changes in the V3 loop were therefore sufficient to convert the
AD101-sensitive CC1/85 virus into one that was completely
resistant to AD101.
The sequence analysis revealed that the K305R change and
the later A316V and G321E changes arose in the context of a
V3 loop in which the H308P substitution was already present
(Fig. 3). We therefore determined whether the H308P substi-
tution was important for the action of the later changes. Thus,
the K305R, A316V, and G321E changes were introduced into
CC1/85 cl.7 in the absence of H308P (Fig. 6b; Table 1). The
mutant clones were all strongly inhibited by AD101 (Fig. 6b;
Table 1). Indeed, the CC1/85 cl.7(K305R) and CC1/85
cl.7(A316V, G321E) viruses were slightly (⬃10-fold) more sen-
sitive than was CC1/85 cl.7 and the triple mutant CC1/85
cl.7(K305R, A316V, G321E) was significantly more sensitive
(⬃700-fold). Hence, the phenotypic influence of the K305R,
A316V and G321E changes was strictly context dependent,
requiring the presence of proline at residue 308. The effect of
the three later-arising changes clearly differed from that of
H308P. They did not themselves impart any AD101 resistance
to the parental CC1/85 cl.7 clone, but they did markedly po-
tentiate the effect of H308P by conspiring with it to impart
even greater resistance to AD101.
Reverse mutagenesis confirms the role of the V3 loop in
conferring AD101 resistance. The above experiments clearly
showed that four amino acid changes in V3 were sufficient to
confer complete AD101 resistance on CC1/85 cl.7. We next
performed reverse mutagenesis to determine whether these
same changes were necessary for complete resistance. The four
critical V3 residues in the AD101-resistant clone, CC101.19
cl.7, were therefore replaced with the cognate amino acids
from the AD101-sensitive virus, CC1/85 cl.7 (Fig. 6c; Table 1).
The quadruple mutant CC101.19 cl.7(R305K, P308H, V316A,
E321G) was fully sensitive to AD101, with an IC
50
(⬃20 pM)
only ca. fivefold greater than that of CC1/85 cl.7. Two addi-
tional substitutions introduced into the CC101.19 cl.7 back-
ground, CC101.19 cl.7(P308H) and CC101.19 cl.7(R305K,
P308H), also caused an increase in AD101 sensitivity relative
FIG. 5. AD101 sensitivity of NL4-3/env chimeric viruses with inter-
changed C2, V3, and C3 regions. The AD101-sensitive clones CC1/85
cl.7 (squares) and AD101-resistant CC101.19 cl.7 (diamonds) and the
chimeric clones CC1/85 cl.7(8) (triangles) and CC101.19 cl.7(8) (cir-
cles) were all tested for AD101 sensitivity in an assay of HIV-1 repli-
cation in primary CD4
⫹
T cells. The results depicted are the average
of four or five independent experiments, with the error bars indicating
the SEM.
VOL. 78, 2004 HIV-1 ESCAPE FROM A CCR5 ANTAGONIST 2799

to the parental, AD101-resistant clone (Fig. 6c; Table 1).
Hence, the four V3 loop changes in CC1/85 that were high-
lighted by the sequence analysis (Fig. 3) were, in fact, critical
for generating AD101 resistance (Fig. 6c; Table 1).
Sequences outside the V3 loop influence the extent of AD101
resistance. Sequences outside the V3 loop can modify the way
in which sensitivity to AD101 is affected by changes within the
loop. For example, the CC101.19 cl.7(P308H) mutant was
more sensitive to AD101 than were the CC101.19 quadruple
mutant and the parental clone, CC1/85 cl.7 (Fig. 6c; Table 1).
Hence, in the context of CC101.19 cl.7, the single change
P308H was alone sufficient to create a highly AD101-sensitive
virus. Of note is the observation that the CC101.19 cl.7
(P308H) and CC1/85 cl.7(K305R, A316V, G321E) viruses
have identical V3 loops in different Env backgrounds. Both
clones were very sensitive to AD101, more so than the parental
clones containing the CC1/85 consensus amino acids at all four
V3 positions. The difference in sensitivity is most apparent
when comparing the IC
50
for CC1/85 cl.7 with that for CC1/85
cl.7(K305R, A316V, G321E) and when comparing the IC
50
for
CC101.19 cl.7(R305K, P308H, V316A, E321G) with that for
CC101.19 cl.7(P308H) (Fig. 6b and c; Table 1). Hence, the V3
sequence containing R305, H308, V316, and E321 seems able
to confer the phenotype of increased sensitivity to AD101,
irrespective of its context, whereas the H308P change yields
either a fully or a partially resistant phenotype depending on
the context into which it is introduced.
The other mutant made in the background of CC101.19 cl.7,
CC101.19 cl.7(R305K, P308H), underscores the influence of
regions outside the V3 loop in determining the AD101 resis-
tance phenotypes of clonal viruses bearing mutant env genes,
as follows. The CC1/85 cl.7 virus and CC101.19 cl.7 quadruple-
mutant virus have identical V3 loops that contain the CC1/85
consensus amino acids at all four V3 positions: K305, H308,
A316, and G321. As in the above example, these clones serve
as the reference viruses for the two mutants, CC1/85
cl.7(A316V, G321E) and CC101.19 cl.7(R305K, P308H). The
two mutants have identical V3 loops, each containing the V316
and E321 substitutions. However, these mutants have very
different phenotypes from the reference clones from which
they were derived. Thus, the CC101.19 cl.7(R305K, P308H)
mutant was ⬃50-fold less sensitive to AD101 than was the
CC101.19 cl.7 quadruple mutant (Fig. 6c; Table 1), whereas
CC1/85 cl.7(A316V, G321E) was slightly (⬃7-fold) more sen-
sitive than CC1/85 cl.7 to AD101 (Fig. 6c; Table 1).
Clearly, then, while the V3 loop appears to be the critical
determinant of complete AD101 resistance in these two env
clones, the V3 sequences do not act alone when influencing the
relative sensitivity of specific variants to AD101; the way in
which the V3 sequences interact with other regions of gp120 is
also important.
The above results demonstrate that, taken together, the V3
loop changes at positions 305, 308, 316, and 321 are both
necessary and sufficient for complete AD101 resistance in the
backgrounds of both CC1/85 cl.7 and CC101.19 cl.7. Differ-
ences in how specific amino acid substitutions act in different
mutant viruses, alone and in combination, probably reflect
differences in the genetic backgrounds of CC101.19 cl.7 and
CC1/85 cl.7 and the consequent influence of residues in re-
gions outside V3 that vary between the two Env proteins.
FIG. 6. Effect of specific amino acid substitutions in the V3 loop on
sensitivity to AD101. (a) The parental CC1/85 cl.7 virus (squares) and
mutants CC1/85 cl.7(H308P) (diamonds), CC1/85 cl.7(K305R, H308P)
(triangles), and CC1/85 cl.7(K305R, H308P, A316V, G321E) (circles)
were all tested for AD101 sensitivity in an assay of HIV-1 replication
in primary CD4
⫹
T cells. (b) Three additional mutants derived from
CC1/85 cl.7 (squares) were also tested for AD101 sensitivity: CC1/85
cl.7(K305R) (diamonds), CC1/85 cl.7(A316V, G321E) (triangles), and
CC1/85 cl.7(K305R, A316V, G321E) (circles). (c) Reverse mutants
made in the context of the AD101-resistant parental clone CC101.19
cl.7 (squares) were evaluated: CC101.19 cl.7(R305K, P308H, V316A,
E321G) (diamonds), CC101.19 cl.7(P308H) (triangles), and CC101.19
cl.7(R305K, P308H) (circles). The results depicted in all panels are the
average of four or five independent experiments, with the error bars
indicating the SEM.
2800 KUHMANN ET AL. J. VIROL.

Receptor binding properties of gp120 monomers derived
from AD101-sensitive and -resistant viruses. To gain initial
insights into which alterations in the properties of Env were
created by the genetic changes associated with AD101 resis-
tance, we expressed and purified monomeric gp120 proteins
from AD101-sensitive and -resistant clones. In a standard
gp120 capture ELISA (76), the 50% effective concentrations
(EC
50
s) for half-maximal CD4-IgG2 binding to gp120 were as
follows: JR-FL, 60 ng/ml; CC1/85 cl.6, 69 ng/ml; CC1/85 cl.7,
40 ng/ml; CC101.19 cl.3, 56 ng/ml; CC101.19 cl.7, 94 ng/ml;
CC101.6 cl.10, 75 ng/ml. The EC
50
s for 17b binding ranged
from 33 to 190 ng/ml in the absence of sCD4 and from 6.1 to
9.3 ng/ml in the presence of 1 ␮g/ml sCD4, the induction of 17b
binding by sCD4 being from 4- to 20-fold (data not shown).
Thus in the presence of sCD4, all the CC1/85 and CC101.19
gp120s bind 17b in a similar manner. Hence all the expressed
gp120 proteins were folded properly, in that they could bind
with high affinity to CD4 and a MAb, 17b, which is often used
as surrogate for CCR5.
We therefore tested the ability of the gp120 proteins to bind
CCR5 directly, by using flow cytometry to measure the extent
of binding of gp120–CD4-IgG2 complexes to CCR5-expressing
L1.2 cells (34, 111). No significant binding of any gp120-CD4-
IgG2 complex to CCR5-negative parental L1.2 cells could be
detected (data not shown). In the absence of CD4-IgG2 or
gp120, no gp120 binding to L1.2-CCR5 cells could be detected,
as expected (Fig. 7 and data not shown). However, when added
as CD4-IgG2 complexes, gp120 proteins from two different,
AD101-sensitive clones of CC1/85 bound efficiently to L1.2-
CCR5 cells, and their binding was completely inhibited by
AD101 (Fig. 7). In contrast, CD4-IgG2 complexes of gp120
proteins from two different clones of the fully AD101-resistant
virus, CC101.19, failed to bind to L1.2-CCR5 cells, even at the
very high input concentration of 100 ␮g/ml and whether or not
AD101 was present. We therefore tested gp120 from the par-
tially AD101-resistant virus CC101.6. This gp120 bound effi-
ciently to L1.2-CCR5 cells, and its binding was inhibited by
AD101 (Fig. 7).
Both the CC101.6 clone and the two CC101.19 clones con-
tain the amino acid changes selected for during the first four
passages in the presence of AD101, including the H308P
change in V3, whereas only the two CC101.19 clones have all
four V3 substitutions (K305R, H308P, A316V, and G321E)
(see above) (Fig. 3 and 4). Indeed, CC101.6 cl.10 gp120 differs
from CC101.19 cl.3 gp120 at only four positions: G78D,
K305R, A316V, and G321E (the CC101.6 cl.10 amino acid is
listed first, and the CC101.19 cl.3 residue is listed last). The
presence of G at position 78 of CC101.6 cl.10 probably repre-
sents a sporadic polymorphism, since this substitution was not
found in any other clone. Moreover, like the two CC101.19
gp120s, the two CC1/85 gp120s both contain D78. Hence, the
glycine residue present at position 78 in CC101.6 cl.10 is highly
unlikely to be relevant to why the CC101.19-derived gp120s fail
to bind cell surface CCR5. In addition to the G78D, K305R,
A316V, and G321E substitutions, the CC101.6 cl.10 gp120
protein differs from the CC101.19 cl.7 protein by only one
additional substitution, D167N, a polymorphic residue dis-
cussed above. Again, an asparagine residue at position 167 is
unlikely to have any impact on CCR5 binding; CC1/85 cl.7
possesses N at this position and binds to CCR5, whereas
CC101.19 cl.7 also has an N but fails to bind. Finally, neither
amino acid 78 nor 167 is known to be in a region associated
with CCR5 binding (59, 93, 94).
The most likely explanation of why gp120 from CC101.19
FIG. 7. Binding of monomeric gp120s from AD101-sensitive and -resistant viruses to L1.2-CCR5 cells in the presence and absence of AD101.
The monomeric gp120 proteins indicated (100 ␮g/ml) with or without biotinylated CD4-lgG2 (100 ␮g/ml) were added to L1.2-CCR5 cells with and
without AD101 (1 ␮M). The extent of binding of the gp120-CD4-IgG2 complex was measured as the mean fluorescence intensity (MFI) and is
shown in arbitrary units (A.U.). Uncorrected mean fluorescence intensity values from a single, representative experiment are presented. The
intrinsic fluorescence in the absence of gp120 is indicated in the “no gp120”column.
VOL. 78, 2004 HIV-1 ESCAPE FROM A CCR5 ANTAGONIST 2801

fails to bind CCR5 on L1.2 cells is that one or more of the
K305R, A316V, and G321E substitutions in V3 has a decisive
influence on a critical aspect of Env conformation. We pre-
pared a model of the V3 loop crown regions of CC1/85 and
CC101.19 gp120s, derived from the structures of complexes
between V3 loop peptides and Fab fragments from the HIV-1
neutralizing antibodies 50.1 (92) and 59.1 (41, 42). The model
indicates that the replacement of H308 by proline does not
increase the probability of a tight turn but that the normal
␤-strand observed in this region for all V3 peptide-Fab com-
plexes would probably be disrupted (data not shown). Hence,
it is clearly possible that in the context of a native Env struc-
ture, the H308P change could have a marked impact on the
geometry of the V3 loop. Additional studies are required to
investigate how this change could influence the resistance to
CCR5 inhibitors.
DISCUSSION
Genetic correlates of escape from a small-molecule CCR5
inhibitor. Clearly, HIV-1 can escape from the pressure exerted
in vitro by a CCR5-specific, small-molecule inhibitor of virus
entry (108). This is not surprising, given the propensity of
HIV-1 to mutate and the ability of its Env complex to accu-
mulate sequence changes without functional impairment (86,
119). What was less predictable was the finding that an R5
virus evolved to continue to use CCR5 rather than switching its
coreceptor preference to CXCR4, in an experimental system
where CXCR4 was clearly available (35, 108). We are now
trying to understand why HIV-1 did not follow what was intu-
itively the line of least resistance but instead, adopted an ap-
parently more complex escape route. An explanation of this
point might enable us to better understand what might happen
if CCR5 inhibitors are used clinically for sustained periods as
front-line antiviral drugs.
We previously studied two isolates (CC101.6 and CC101.19)
that were derived from the parental R5 isolate CC1/85 after
passaging for 6 and 19 passages, respectively, in the presence of
AD101 (108). We showed that evolution of AD101 resistance
was at least a two-step process; a partial (⬃4-fold) loss of
sensitivity to the selecting agent was acquired after 6 passages,
and essentially complete (⬎20,000-fold) resistance was ac-
quired by 19 passages (108). There was no overt loss of repli-
cation competence in PBMC associated with the development
of full AD101 resistance. The genetic data we have now ob-
tained are consistent with, and help explain, the phenotypic
changes.
By passage 4 of the AD101 selection process, the Env se-
quence had become highly homogeneous. Multiple, previously
minor, amino acid sequences became dominant among the
sequenced clones, to the extent that they were essentially fixed.
We propose that the partial AD101 resistance of the CC101.6
isolate is explained by selection of relatively resistant clones
from among the preexisting CC1/85 population. A likely mech-
anism is the increased ability of viruses bearing the variant Env
proteins to better exploit lower levels of free CCR5; such
viruses would be relatively well suited to replicate during the
early stages, when AD101 concentrations were only moder-
ately elevated (108). Because the clones derived from CC101.6
contain all the amino acid changes selected for early (by pas-
sage 4) but none of those that arose and were selected for later
(after passage 8), some or all of the initially selected sequences
must play a role in partial resistance to AD101. Among them
is the H308P substitution in the V3 loop, and we have con-
firmed by mutagenesis that this change does confer partial
resistance. Hence, we believe that the H308P substitution was
the most critical difference selected for from the CC1/85 pop-
ulation during the first four passages with increasing AD101
concentrations. Other changes occurring or selected for over
the same time may represent adaptations to long-term PBMC
culture or may be mere hitchhikers on the galaxy of Env pro-
teins that happened to carry the critical H308P change.
Sequence analysis of the clones obtained from isolates after
passage 4 indicates that the complete AD101 escape pheno-
type of the passage 19 isolate is associated with three addi-
tional, sequential V3 substitutions: K305R, A316V, and
G321E. These changes arise and are selected for at a time that
correlates with the acquisition of full resistance (108). The four
V3 loop changes associated with AD101 resistance remained
stable, with only a few exceptions, when the resistant isolate
CC101.22 was cultured for an additional nine passages without
AD101. This is consistent with the stability of the resistance
phenotype over the same time (108). We are now studying
whether a more prolonged culture without AD101 allows the
V3 changes to revert and whether this is associated with reac-
quisition of AD101 sensitivity.
In contrast to the four changes in the V3 loop, there was no
detectable selection pressure on amino acids in the C4 region
or the stem of the V1/V2 loop that make up the bridging sheet
of the gp120 core, the most highly conserved element of the
CCR5 binding site (59, 93, 94). Nor was there any change in
the footprint of MAb 17b to the CD4-induced epitope that
overlaps the CCR5 binding site (59, 94, 105, 118).
In vivo, X4 viruses evolved in the individual from whom
CC1/85 was obtained, such that an R5X4 virus, CC2/86, was
isolated 13 months later (23, 24). Additional studies are now in
progress, using clonal viruses of defined phenotype derived
from CC2/86, to try to define which sequence changes in a
CC1/85 R5 clone are sufficient to confer CXCR4 usage. It will
be instructive to compare such changes with those that create
AD101 resistance while preserving the R5 phenotype.
Sequence changes in the V3 loop confer resistance to AD101.
To better define the genetics of AD101 resistance, we inserted
env clones from CC1/85 and CC101.19 into the pNL4-3 back-
bone, in the absence of other changes. The clonal viruses
qualitatively reconstitute the AD101-sensitive and -resistant
phenotypes of the CC1/85 and CC101.19 isolates, respectively.
There are, however, quantitative differences in the degree of
resistance shown by the clonal viruses and the corresponding
isolates. Much of the difference is probably methodology
based. NL4-3-based Env chimeric viruses may differ from nat-
ural viruses in the extent of Env incorporatation, which could
affect fusion efficiency and its inhibition (55, 87). Viruses pro-
duced by transfection of 293T cells and by natural infection of
PBMC may also vary in how they attach to target cells because
of differences in their complement of host-derived proteins
(106). There are also more subtle issues; for example, it is not
obvious why AD101 inhibits the CC1/85 cl.7 clone 250-fold
more potently than it inhibits the CC1/85 isolate when no
differential sensitivity was observed using the related inhibitor,
SCH-C.
2802 KUHMANN ET AL. J. VIROL.

Mutagenesis studies using the chimeric Env clones refined
our understanding of the resistance mechanism. Thus, the
H308P change was sufficient to cause a 500-fold increase in the
AD101 resistance of a CC1/85 clone, and the addition of both
the H308P and R305K substitutions caused a further 200-fold
resistance increase. The introduction of all four amino acid
changes highlighted by the sequence analysis (K305R, H308P,
A316V, and G321E) created a virus completely resistant to
AD101 concentrations as high as 20 ␮M. Conversely, when the
reverse changes were made at these four V3 positions, the
AD101-resistant virus CC101.19 cl.7 was converted into one
that was highly sensitive to AD101. Taking the mutagenesis
and the evolutionary genetics data together, the ⬎20,000-fold
resistance of the CC101.19 isolate to AD101 is due primarily to
the four substitutions within the V3 loop.
It is no surprise that the V3 loop plays a dominant role in
altering the coreceptor interactions of an HIV-1 isolate. There
is ample precedent for a major influence of the V3 region on
viral tropism, which we now understand to be largely a conse-
quence of coreceptor choice (20–22, 31, 33, 39, 47, 48, 100,
114). Moreover, the V3 loop is a component of the gp120
binding sites for both CCR5 and CXCR4 (12, 22, 25, 33, 61, 93,
94). Its ability to vary in sequence while preserving its functions
renders it an obvious place for amino acid substitutions to
accumulate in response to AD101 selection pressure. Other
elements of the CCR5 binding site on gp120 are more highly
conserved, and so sequence changes there are inherently less
likely (33, 93, 94, 119).
There have been previous studies of escape from coreceptor
inhibitors. The development of partial (four- to sixfold) resis-
tance of the R5 isolate JR-FL to macrophage inflammatory
protein 1␣was also associated with a V3 sequence change,
acting in concert with another substitution in the V2 loop (68).
The generation of ⬃100-fold resistance to the anti-CCR5 MAb
2D7 when using the JR-CSF isolate was not associated with V3
sequence changes, although the genetic correlates of resistance
were not fully identified (1). A significant proportion of the
amino acid changes that accumulated in gp120 from the X4
virus NL4-3 during prolonged passage in vitro with the CXCR4
antagonists AMD3100, SDF-1, and T134 were also located in
V3 (30, 50, 97). In each case, the resistant virus continued to
use CXCR4, but in an inhibitor-insensitive manner (30, 50, 97).
It was not determined whether the V3 loop changes were
responsible for the generation of the resistant viruses, but it is
a reasonable assumption that they were, at least, relevant (30,
50, 97).
A clone of HIV-1 JR-CSF can adapt to growth on cells that
express CCR5 extracellular domain mutants with changes in
the defined binding sites for HIV-1 (84). The virus adapted to
each CCR5 substitution by selecting a specific pattern of
changes in the V3 loop. Engineered mutants of the parental
virus containing the V3 substitutions replicated efficiently, in a
strictly CCR5-dependent manner, in cells expressing either
wild-type or mutated CCR5, whereas the parental virus could
use only wild-type CCR5. Hence, as we have found in the
present study, the interaction of gp120 with CCR5 can be
highly plastic, with changes in the V3 loop being sufficient to
significantly influence it (84).
Functional aspects of AD101 resistance. Although the
CC101.19 clones are fully resistant to AD101 and to other
small-molecule CCR5 inhibitors such as SCH-C, they retained
their sensitivity to CCR5-specific MAbs and to the CC chemo-
kine RANTES. Hence, just like the CC101.19 isolate from
which they were cloned, they still use CCR5; indeed, their
inability to replicate in PBMC from CCR5-⌬32 homozygous
individuals shows they absolutely require CCR5 for entry. Re-
sistance to CCR5-targeted inhibitors is therefore not global;
instead, resistance to the different classes of inhibitor is likely
to depend on the nature of their binding sites on CCR5. The
small molecules AD101 and SCH-C bind within a pocket
formed between CCR5 transmembrane helices 1, 2, 3, and 7,
without involvement of the NT or ECLs (111). Their binding
sites are broadly similar, albeit not identical, so that cross-
resistance among small molecule CCR5 inhibitors is to be
anticipated (111). The CCR5 MAbs and RANTES, in contrast,
have binding sites that involve, wholly or in part, the NT and
ECLs (14, 15, 67, 81, 96, 117). It is perhaps not surprising, then,
that the AD101-resistant virus is still sensitive to inhibitors that
bind elsewhere on CCR5.
It was, however, unexpected that monomeric gp120 proteins
expressed from fully AD101-resistant clones (CC101.19 cl.3
and CC101.19 cl.7) failed to bind to CCR5 on L1.2-CCR5 cells.
Considering that the corresponding infectious viruses abso-
lutely require CCR5 to enter PBMC, the inability of their
gp120 proteins to bind CCR5 appears paradoxical. The gp120
proteins are properly folded; they have appropriate affinities
for CD4 and MAb 17b, so that their lack of binding to CCR5
is unlikely to have a trivial explanation. Moreover, gp120 from
a partially AD101-resistant virus, CC101.6 cl.10, bound CCR5
perfectly well. The three, later-arising de novo substitutions in
V3, K305R, A316V and G321E, are therefore responsible for
the inability of CC101.19 gp120 to bind CCR5.
We can imagine two plausible explanations for why gp120
from the fully AD101-resistant virus does not bind to L1.2-
CCR5 cells. The use of monomeric gp120 proteins may not
recapitulate what happens when trimeric Env on the virion
surface interacts with CCR5, a concept for which there is
ample precedent from studies of the CD4 and MAb interac-
tions of gp120 and virions (58, 74, 75, 86, 119). If so, the three
changes in V3 render the gp120 monomers incapable of CCR5
binding while still permitting the trimeric Env complex to do
so, albeit differently from the binding of Env complexes from
AD101-sensitive viruses. Alternatively, the Env complex on the
fully AD101-resistant virus may have evolved to recognize a
specific CCR5 isoform or conformation present only on human
PBMC (or on primary cells in general) and absent from L1.2-
CCR5 cells. The latter are CCR5 transfectants of a murine
pre-B-cell lymphoma line (116), so that post-translational
modifications of CCR5 may be subtly different from what hap-
pens in primary human cells. Again, there is precedent for
chemokine receptors, including CCR5, existing in multiple iso-
forms and conformations (6, 10, 11, 18, 36, 63, 64, 67, 70). If
this is the explanation, then full AD101 resistance involves
HIV-1 acquiring the ability to use a distinctive configuration of
CCR5 that, somehow, is not affected by AD101 and other
small-molecule inhibitors while still being vulnerable to CCR5-
specific MAbs and RANTES. Clearly, additional studies are
required to reveal what, exactly, is the explanation of AD101
resistance and how the critical sequence changes in the V3
loop affect the CCR5 binding of resistant viruses.
VOL. 78, 2004 HIV-1 ESCAPE FROM A CCR5 ANTAGONIST 2803

Effects of Env context on the V3 loop changes. Several ob-
servations suggested that introducing sequence changes into
the V3 loop of chimeric env clones can have context-dependent
effects. Hence, the same genetic substitutions do not always
have the same effect on phenotype; their precise impact in-
stead depends on residues elsewhere in gp120. The determin-
ing factor is presumably how the different V3 loops interact
with other gp120 domains, before or after CD4 binding. Qua-
ternary interactions within the native Env complex may also be
relevant. We do not yet know how any V3 loop is oriented and
functions in its native context; there is no gp120 structure that
includes the V3 loop. Moreover, we have no knowledge of how
sequence changes in the V3 loop might directly or indirectly
affect its orientation in a native Env trimer, for which we have
only models, not a structure (60). It is suspected, however, that
the C-terminal flank of the V3 loop abuts elements of the C4
region proximal to the CD4 binding site (17, 78, 79, 120, 121).
There are probably also associations between V3 and other
variable regions (16, 19, 123). Hence, interdomain interactions,
either physical or functional, that are important in CD4 and
coreceptor binding might differ in viral variants with V3 se-
quence changes (7, 16, 17, 19, 46, 78, 79, 120, 121, 123).
The H308P change is sufficiently radical to perhaps have a
significant impact on the geometry of the V3 loop, the way in
which it interacts with other regions of gp120, and hence the
way in which CCR5 recognition is affected, particularly in the
context of a native Env trimer. However, the effect of the
H308P substitution must be reinforced by the three later-aris-
ing V3 substitutions. Thus, CC101.6 contains the H308P sub-
stitution and is only partially AD101 resistant, and gp120 ex-
pressed from an H308P-containing CC101.6 clone binds CCR5
on L1.2-CCR5 cells.
Implications of in vitro resistance pathways for the clinical
use of CCR5 inhibitors. Resistance to small-molecule CCR5
inhibitors will no doubt also occur in vivo if these compounds
are used for sustained periods in the clinic. However, our
studies suggest that escape will probably not be a simple, one-
step process. Multiple amino acid changes must accumulate in
gp120 to generate first partial and then full resistance to the
selecting compound. In the case of complete resistance to
AD101, as many as four amino acid substitutions in V3 are
needed. The particular combination of the four residues that
were present at positions 305, 308, 316, and 321 in the fully
AD101-resistant virus, CC101.19, is very rare in naturally oc-
curring V3 sequences. Hence, among 6,265 sequences from
North American subtype B viruses in the Los Alamos Se-
quence Database, none had V3 loops that contained all four,
or even three, of the residues associated with AD101 resis-
tance. Two of the residues were naturally present in 188 of the
sequences (3%), while 30% of the sequences contained one of
the four.
Partial resistance may be easier to generate, however, as
exemplified by the selection during the first four passages of
preexisting variants with natural, albeit modest resistance to
AD101. It is unclear how many amino acid changes are re-
quired for the ca. fourfold resistance of CC101.6; the H308P
change in V3 may or may not be sufficient, and it is not
uncommon among naturally occurring V3 sequences (H was
found in 57% of B clade sequences, an P was found in 16%).
A more dogmatic assertion is precluded by difficulties in mak-
ing direct quantitative comparisons between the IC
50
s for in-
hibition of the NL4-3/env clones and the corresponding iso-
lates. Resistance of this magnitude, a fewfold, could be
significant in vivo. However, given the diversity of Env (and
gp120 in particular), inherent variability in the response to
entry inhibitors is to be expected. For example, in a standard
PBMC inhibition assay using 13 subtype B isolates, IC
50
s for
SCH-C spanned a 22-fold range from 0.4 to 8.9 nM, with a
mean of 2.3 ⫾0.8 nM (101). It will be important to establish
the normal sensitivity range for patient isolates both in vitro
and in vivo before resistance can be properly defined and
identified (9, 43, 45, 65). It will also be instructive to compare
the rate and genetic complexity of escape from CCR5 inhibi-
tors in vitro and in vivo with the development of resistance to
reverse transcriptase and protease inhibitors under compara-
ble conditions (9, 43, 45, 65).
Deriving patient isolates for in vitro testing is a laborious
and costly process, and so the use of more practical methods
for gauging the potency of, and resistance to, entry inhibitors
should be considered. Clonal virus systems, based on infectious
chimeras or Env-pseudotyped viruses, are becoming an in-
creasingly common way to study entry inhibitors, including
neutralizing antibodies (9, 29, 38, 43, 45, 65, 91, 110, 112, 113).
We noticed, however, that quantitative, but not qualitative,
estimates of AD101 resistance differed when they were derived
by using clonal, infectious chimeric viruses and uncloned iso-
lates, even with the same target cells. Thus, the infectious
chimera CC1/85 cl.7(H308P) has 500-fold higher resistance to
AD101 than does CC1/85 (Fig. 6a; Table 1). However, when
present in the PBMC-derived isolate CC101.6, the H308P sub-
stitution confers at most a fourfold increase in resistance to
AD101 over that of CC1/85 (108). As outlined above, several
factors could contribute to such differences, and some probably
also apply to other assay systems that use clonal or Env-
pseudotyped viruses. Hence, although these types of assay can
be very valuable for monitoring the relative sensitivity or re-
sistance of different viruses and entry inhibitors, care must be
taken not to overinterpret the results by comparing numerical
values with those obtained using the more biologically relevant
uncloned patient isolates.
Genetic methods are an alternative to phenotypic assays and
are very valuable when, for example, a single, dominant amino
acid change confers reverse transcriptase inhibitor resistance
(9, 43, 45, 65). Our study suggests, however, that such a sce-
nario is unlikely to apply to CCR5 inhibitors. The V3 region is
a likely site for resistance mutations to accumulate, both in
vitro and in vivo, but the precise pattern of sequence changes
we have observed may be unique to the particular virus, CCR5
inhibitor, and PBMC donor combination that we used. It may
never be seen again, even in vitro. Assays that can identify V3
loop genetic variants irrespective of the precise sequence
changes that occur may therefore be useful (5, 28, 54, 88, 89).
However, the influence of Env context on V3 loop function
may be a significant complication.
In clinical use, CCR5 inhibitors will be combined with other
entry inhibitors or with more conventional suppressors of viral
replication (8, 52, 77, 80, 85, 90, 98). The use of a drug com-
bination may mean that HIV-1 will not accumulate sufficient
mutations to escape from the CCR5 inhibitor before resistance
to the other components of the combination develops. Of
2804 KUHMANN ET AL. J. VIROL.

course, when escape does occur, it will be critical to determine
whether HIV-1 has switched to use CXCR4 or has instead
taken a pathway similar to the one we describe here.
ACKNOWLEDGMENTS
We are grateful to Ruth Connor for providing HIV-1 isolates CC1/
85, CC2/86, CC7/86, and CC12/86; Bill Olson for JR-FL gp120 protein,
sCD4, MAb PA14, and CD4-IgG2; and James Robinson for MAb 17b.
We thank Mika Vesanen and Maciej Paluch for their assistance in the
purification of recombinant gp120 proteins. We thank Fred Lee, Tom
Morgan, and Tom Ketas for technical assistance.
This work was funded by NIH grants AI41420 (J.P.M.), GM46192
(I.A.W. and R.L.S.), HD37356 (S.M.W.), and immunology training
grant T32 AI07621 (S.E.K. and P.P.); by IAVI (J.P.M. and I.A.W.);
and by Schering Plough Research Institute (J.P.M.). B.T.K. is an Eliz-
abeth Glaser Scientist of the Pediatric AIDS Foundation. J.P.M. is a
Stavros S. Niarchos Scholar. The Department of Microbiology and
Immunology at the Weill Medical College gratefully acknowledges the
support of the William Randolph Hearst Foundation.
REFERENCES
1. Aarons, E. J., S. Beddows, T. Willingham, L. Wu, and R. A. Koup. 2001.
Adaptation to blockade of human immunodeficiency virus type 1 entry
imposed by the anti-CCR5 monoclonal antibody 2D7. Virology 287:382–
390.
2. Adachi, A., H. E. Gendelman, S. Koenig, T. Folks, R. Willey, A. Rabson, and
M. A. Martin. 1986. Production of acquired immunodeficiency syndrome-
associated retrovirus in human and nonhuman cells transfected with an
infectious molecular clone. J. Virol. 59:284–291.
3. Allaway, G. P., K. L. Davis-Bruno, G. A. Beaudry, E. B. Garcia, E. L. Wong,
A. M. Ryder, K. W. Hasel, M. C. Gauduin, R. A. Koup, J. S. McDougal, and
P. J. Maddon. 1995. Expression and characterization of CD4-IgG2, a novel
heterotetramer that neutralizes primary HIV type 1 isolates. AIDS Res.
Hum. Retrovir. 11:533–539.
4. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990.
Basic local alignment search tool. J. Mol. Biol. 215:403–410.
5. Alves, K., M. Canzian, and E. L. Delwart. 2002. HIV type 1 envelope
quasispecies in the thymus and lymph Nodes of AIDS patients. AIDS Res.
Hum. Retrovir. 18:161–165.
6. Babcock, G. J., M. Farzan, and J. Sodroski. 2003. Ligand-independent
dimerization of CXCR4, a principal HIV-1 coreceptor. J. Biol. Chem.
278:3378–3385.
7. Bagnarelli, P., L. Fiorelli, M. Vecchi, A. Monachetti, S. Menzo, and M.
Clementi. 2003. Analysis of the functional relationship between V3 loop
and gp120 context with regard to human immunodeficiency virus corecep-
tor usage using naturally selected sequences and different viral backbones.
Virology 307:328–340.
8. Baldwin, C. E., R. W. Sanders, and B. Berkhout. 2003. Inhibiting HIV-1
entry with fusion inhibitors. Curr. Med. Chem. 10:1773–1782.
9. Balzarini, J. 1999. Suppression of resistance to drugs targeted to human
immunodeficiency virus reverse transcriptase by combination therapy. Bio-
chem. Pharmacol. 58:1–27.
10. Bannert, N., S. Craig, M. Farzan, D. Sogah, N. V. Santo, H. Choe, and J.
Sodroski. 2001. Sialylated O-glycans and sulfated tyrosines in the NH
2
-
terminal domain of CC chemokine receptor 5 contribute to high affinity
binding of chemokines. J. Exp. Med. 194:1661–1673.
11. Baribaud, F., T. G. Edwards, M. Sharron, A. Brelot, N. Heveker, K. Price,
F. Mortari, M. Alizon, M. Tsang, and R. W. Doms. 2001. Antigenically
distinct conformations of CXCR4. J. Virol. 75:8957–8967.
12. Basmaciogullari, S., G. J. Babcock, D. Van Ryk, W. Wojtowicz, and J.
Sodroski. 2002. Identification of conserved and variable structures in the
human immunodeficiency virus gp120 glycoprotein of importance for
CXCR4 binding. J. Virol. 76:10791–10800.
13. Binley, J. M., R. W. Sanders, B. Clas, N. Schuelke, A. Master, Y: Guo, F.
Kajumo, D. J. Anselma, P. J. Maddon, W. C. Olson, and J. P. Moore. 2000.
A recombinant human immunodeficiency virus type 1 envelope glycopro-
tein complex stabilized by an intermolecular disulfide bond between the
gp120 and gp41 subunits is an antigenic mimic of the trimeric virion-
associated structure. J. Virol. 74:627–643.
14. Blanpain, C., B. J. Doranz, A. Bondue, C. Govaerts, A. De Leener, G.
Vassart, R. W. Doms, A. Proudfoot, and M. Parmentier. 2003. The core
domain of chemokines binds CCR5 extracellular domains while their amino
terminus interacts with the transmembrane helix bundle. J. Biol. Chem.
278:5179–5187.
15. Blanpain, C., B. J. Doranz, J. Vakili, J. Rucker, C. Govaerts, S. S. Baik, O.
Lorthioir, I. Migeotte, F. Libert, F. Baleux, G. Vassart, R. W. Doms, and M.
Parmentier. 1999. Multiple charged and aromatic residues in CCR5 amino-
terminal domain are involved in high affinity binding of both chemokines
and HIV-1 Env protein. J. Biol. Chem. 274:34719–34727.
16. Carrillo, A., and L. Ratner. 1996. Cooperative effects of the human immu-
nodeficiency virus type 1 envelope variable loops V1 and V3 in mediating
infectivity for T cells. J. Virol. 70:1310–1316.
17. Carrillo, A., and L. Ratner. 1996. Human immunodeficiency virus type 1
tropism for T-lymphoid cell lines: role of the V3 loop and C4 envelope
determinants. J. Virol. 70:1301–1309.
18. Chelli, M., and M. Alizon. 2002. Rescue of HIV-1 receptor function
through cooperation between different forms of the CCR5 chemokine re-
ceptor. J. Biol. Chem. 277:39388–39396.
19. Cho, M. W., M. K. Lee, C. H. Chen, T. Matthews, and M. A. Martin. 2000.
Identification of gp120 regions targeted by a highly potent neutralizing
antiserum elicited in a chimpanzee inoculated with a primary human im-
munodeficiency virus type 1 isolate. J. Virol. 74:9749–9754.
20. Choe, H., M. Farzan, Y. Sun, N. Sullivan, B. Rollins, P. D. Ponath, L. Wu,
C. R. Mackay, G. LaRosa, W. Newman, N. Gerard, C. Gerard, and J.
Sodroski. 1996. The beta-chemokine receptors CCR3 and CCR5 facilitate
infection by primary HIV-1 isolates. Cell 85:1135–1148.
21. Clapham, P. R., and A. McKnight. 2002. Cell surface receptors, virus entry
and tropism of primate lentiviruses. J. Gen. Virol. 83:1809–1829.
22. Cocchi, F., A. L. DeVico, A. Garzino-Demo, A. Cara, R. C. Gallo, and P.
Lusso. 1996. The V3 domain of the HIV-1 gp120 envelope glycoprotein is
critical for chemokine-mediated blockade of infection. Nat. Med. 2:1244–
1247.
23. Connor, R. I., H. Mohri, Y. Cao, and D. D. Ho. 1993. Increased viral burden
and cytopathicity correlate temporally with CD4
⫹
T-lymphocyte decline
and clinical progression in human immunodeficiency virus type 1-infected
individuals. J. Virol. 67:1772–1777.
24. Connor, R. I., K. E. Sheridan, D. Ceradini, S. Choe, and N. R. Landau.
1997. Change in coreceptor use correlates with disease progression in
HIV-1-infected individuals. J. Exp. Med. 185:621–628.
25. Cormier, E. G., and T. Dragic. 2002. The crown and stem of the V3 loop
play distinct roles in human immunodeficiency virus type 1 envelope gly-
coprotein interactions with the CCR5 coreceptor. J. Virol. 76:8953–8957.
26. De Clercq, E. 2003. The bicyclam AMD3100 story. Nat. Rev. Drug Discov.
2:581–587.
27. Deen, K. C., J. S. McDougal, R. Inacker, G. Folena-Wasserman, J. Arthos,
J. Rosenberg, P. J. Maddon, R. Axel, and R. W. Sweet. 1988. A soluble form
of CD4 (T4) protein inhibits AIDS virus infection. Nature 331:82–84.
28. Delwart, E., M. Magierowska, M. Royz, B. Foley, L. Peddada, R. Smith, C.
Heldebrant, A. Conrad, and M. Busch. 2002. Homogeneous quasispecies in
16 out of 17 individuals during very early HIV-1 primary infection. AIDS
16:189–195.
29. Derdeyn, C. A., J. M. Decker, J. N. Sfakianos, X. Wu, W. A. O’Brien, L.
Ratner, J. C. Kappes, G. M. Shaw, and E. Hunter. 2000. Sensitivity of
human immunodeficiency virus type 1 to the fusion inhibitor T-20 is mod-
ulated by coreceptor specificity defined by the V3 loop of gp120. J. Virol.
74:8358–8367.
30. de Vreese, K., V. Kofler-Mongold, C. Leutgeb, V. Weber, K. Vermeire, S.
Schacht, J. Anne, E. de Clercq, R. Datema, and G. Werner. 1996. The
molecular target of bicyclams, potent inhibitors of human immunodefi-
ciency virus replication. J. Virol. 70:689–696.
31. Doms, R. W. 2000. Beyond receptor expression: the influence of receptor
conformation, density, and affinity in HIV-1 infection. Virology 276:229–
237.
32. Douek, D. C., L. J. Picker, and R. A. Koup. 2003. T cell dynamics in HIV-1
infection. Annu. Rev. Immunol. 21:265–304.
33. Dragic, T. 2001. An overview of the determinants of CCR5 and CXCR4
co-receptor function. J. Gen. Virol. 82:1807–1814.
34. Dragic, T., A. Trkola, D. A. Thompson, E. G. Cormier, F. A. Kajumo, E.
Maxwell, S. W. Lin, W. Ying, S. O. Smith, T. P. Sakmar, and J. P. Moore.
2000. A binding pocket for a small molecule inhibitor of HIV-1 entry within
the transmembrane helices of CCR5. Proc. Natl. Acad. Sci. USA 97:5639–
5644.
35. Farber, J. M., and E. A. Berger. 2002. HIV’s response to a CCR5 inhibitor:
I’d rather tighten than switch! Proc. Natl. Acad. Sci. USA 99:1749–1751.
36. Farzan, M., G. J. Babcock, N. Vasilieva, P. L. Wright, E. Kiprilov, T.
Mirzabekov, and H. Choe. 2002. The role of post-translational modifica-
tions of the CXCR4 amino terminus in stromal-derived factor 1 alpha
association and HIV-1 entry. J. Biol. Chem. 277:29484–29489.
37. Feng, Y., C. C. Broder, P. E. Kennedy, and E. A. Berger. 1996. HIV-1 entry
cofactor: functional cDNA cloning of a seven-transmembrane, G protein-
coupled receptor. Science 272:872–877.
38. Fikkert, V., P. Cherepanov, K. Van Laethem, A. Hantson, B. Van Remoor-
tel, C. Pannecouque, E. De Clercq, Z. Debyser, A. M. Vandamme, and M.
Witvrouw. 2002. env chimeric virus technology for evaluating human im-
munodeficiency virus susceptibility to entry inhibitors. Antimicrob. Agents
Chemother. 46:3954–3962.
39. Fouchier, R. A., M. Groenink, N. A. Kootstra, M. Tersmette, H. G. Huis-
man, F. Miedema, and H. Schuitemaker. 1992. Phenotype-associated se-
VOL. 78, 2004 HIV-1 ESCAPE FROM A CCR5 ANTAGONIST 2805

quence variation in the third variable domain of the human immunodefi-
ciency virus type 1 gp120 molecule. J. Virol. 66:3183–3187.
40. Gallo, S. A., A. Puri, and R. Blumenthal. 2001. HIV-1 gp41 six-helix bundle
formation occurs rapidly after the engagement of gp120 by CXCR4 in the
HIV-1 Env-mediated fusion process. Biochemistry 40:12231–12236.
41. Ghiara, J. B., D. C. Ferguson, A. C. Satterthwait, H. J. Dyson, and I. A.
Wilson. 1997. Structure-based design of a constrained peptide mimic of the
HIV-1 V3 loop neutralization site. J. Mol. Biol. 266:31–39.
42. Ghiara, J. B., E. A. Stura, R. L. Stanfield, A. T. Profy, and I. A. Wilson.
1994. Crystal structure of the principal neutralization site of HIV-1. Science
264:82–85.
43. Hammer, S. M. 2002. HIV drug resistance: implications for management.
Top. HIV Med. 10:10–15.
44. Hazenberg, M. D., D. Hamann, H. Schuitemaker, and F. Miedema. 2000. T
cell depletion in HIV-1 infection: how CD4
⫹
T cells go out of stock. Nat.
Immunol. 1:285–289.
45. Hirsch, M. S., F. Brun-Vezinet, R. T. D’Aquila, S. M. Hammer, V. A.
Johnson, D. R. Kuritzkes, C. Loveday, J. W. Mellors, B. Clotet, B. Conway,
L. M. Demeter, S. Vella, D. M. Jacobsen, and D. D. Richman. 2000. Anti-
retroviral drug resistance testing in adult HIV-1 infection: recommenda-
tions of an International AIDS Society—USA Panel. JAMA 283:2417–
2426.
46. Hoffman, N. G., F. Seillier-Moiseiwitsch, J. Ahn, J. M. Walker, and R.
Swanstrom. 2002. Variability in the human immunodeficiency virus type 1
gp120 Env protein linked to phenotype-associated changes in the V3 loop.
J. Virol. 76:3852–3864.
47. Hoffman, T. L., and R. W. Doms. 1999. HIV-1 envelope determinants for
cell tropism and chemokine receptor use. Mol. Membr. Biol. 16:57–65.
48. Hwang, S. S., T. J. Boyle, H. K. Lyerly, and B. R. Cullen. 1991. Identifica-
tion of the envelope V3 loop as the primary determinant of cell tropism in
HIV-1. Science 253:71–74.
49. Jacobson, J. M., I. Lowy, C. V. Fletcher, T. J. O’Neill, D. N. Tran, T. J.
Ketas, A. Trkola, M. E. Klotman, P. J. Maddon, W. C. Olson, and R. J.
Israel. 2000. Single-dose safety, pharmacology, and antiviral activity of the
human immunodeficiency virus (HIV) type 1 entry inhibitor PRO 542 in
HIV-infected adults. J. Infect. Dis. 182:326–329.
50. Kanbara, K., S. Sato, J. Tanuma, H. Tamamura, K. Gotoh, M. Yoshimori,
T. Kanamoto, M. Kitano, N. Fujii, and H. Nakashima. 2001. Biological and
genetic characterization of a human immunodeficiency virus strain resistant
to CXCR4 antagonist T134. AIDS Res. Hum. Retrovir. 17:615–622.
51. Kazmierski, W., N. Bifulco, H. Yang, L. Boone, F. DeAnda, C. Watson, and
T. Kenakin. 2003. Recent progress in discovery of small-molecule CCR5
chemokine receptor ligands as HIV-1 inhibitors. Bioorg. Med. Chem. 11:
2663–2676.
52. Kilby, J. M., and J. J. Eron. 2003. Novel therapies based on mechanisms of
HIV-1 cell entry. N. Engl. J. Med. 348:2228–2238.
53. Kilby, J. M., S. Hopkins, T. M. Venetta, B. DiMassimo, G. A. Cloud, J. Y.
Lee, L. Alldredge, E. Hunter, D. Lambert, D. Bolognesi, T. Matthews, M. R.
Johnson, M. A. Nowak, G. M. Shaw, and M. S. Saag. 1998. Potent sup-
pression of HIV-1 replication in humans by T-20, a peptide inhibitor of
gp41-mediated virus entry. Nat. Med. 4:1302–1307.
54. Kitrinos, K. M., N. G. Hoffman, J. A. Nelson, and R. Swanstrom. 2003.
Turnover of Env variable region 1 and 2 genotypes in subjects with late-
stage human immunodeficiency virus type 1 infection. J. Virol. 77:6811–
6822.
55. Klasse, P. J., and J. P. Moore. 1996. Quantitative model of antibody- and
soluble CD4-mediated neutralization of primary isolates and T-cell line-
adapted strains of human immunodeficiency virus type 1. J. Virol. 70:3668–
3677.
56. Korber, B., B. T. Foley, C. L. Kuiken, S. K. Pillai, and J. Sodroski. 1998.
Numbering positions in HIV relative to HXB2CG, p. 102–111. In B.
Korber, C. L. Kuiken, B. T. Foley, B. H. Hahn, F. McCutchan, J. W.
Mellors, and J. Sodroski (ed.), Human retroviruses and AIDS. Theoretical
Biology and Biophysics Group, Los Alamos National Laboratory, Los
Alamos, NM.
57. Kwa, D., J. Vingerhoed, B. Boeser, and H. Schuitemaker. 2003. Increased in
vitro cytopathicity of CC-chemokine receptor 5-restricted human immuno-
deficiency virus type 1 primary isolates correlates with a progressive clinical
course of infection. J. Infect. Dis. 187:1397–1403.
58. Kwong, P. D., M. L. Doyle, D. J. Casper, C. Cicala, S. A. Leavitt, S. Majeed,
T. D. Steenbeke, M. Venturi, I. Chaiken, M. Fung, H. Katinger, P. W.
Parren, J. Robinson, D. Van Ryk, L. Wang, D. R. Burton, E. Freire, R.
Wyatt, J. Sodroski, W. A. Hendrickson, and J. Arthos. 2002. HIV-1 evades
antibody-mediated neutralization through conformational masking of re-
ceptor-binding sites. Nature 420:678–682.
59. Kwong, P. D., R. Wyatt, J. Robinson, R. W. Sweet, J. Sodroski, and W. A.
Hendrickson. 1998. Structure of an HIV gp120 envelope glycoprotein in
complex with the CD4 receptor and a neutralizing human antibody. Nature
393:648–659.
60. Kwong, P. D., R. Wyatt, Q. J. Sattentau, J. Sodroski, and W. A. Hendrick-
son. 2000. Oligomeric modeling and electrostatic analysis of the gp120
envelope glycoprotein of human immunodeficiency virus. J. Virol. 74:1961–
1972.
61. Labrosse, B., C. Treboute, A. Brelot, and M. Alizon. 2001. Cooperation of
the V1/V2 and V3 domains of human immunodeficiency virus type 1 gp120
for interaction with the CXCR4 receptor. J. Virol. 75:5457–5464.
62. Lalezari, J. P., K. Henry, M. O’Hearn, J. S. Montaner, P. J. Piliero, B.
Trottier, S. Walmsley, C. Cohen, D. R. Kuritzkes, J. J. Eron, Jr., J. Chung,
R. DeMasi, L. Donatacci, C. Drobnes, J. Delehanty, and M. Salgo. 2003.
Enfuvirtide, an HIV-1 fusion inhibitor, for drug-resistant HIV infection in
North and South America. N. Engl. J. Med. 348:2175–2185.
63. Lapham, C. K., T. Romantseva, E. Petricoin, L. R. King, J. Manischewitz,
M. B. Zaitseva, and H. Golding. 2002. CXCR4 heterogeneity in primary
cells: possible role of ubiquitination. J. Leukoc. Biol. 72:1206–1214.
64. Lapham, C. K., M. B. Zaitseva, S. Lee, T. Romanstseva, and H. Golding.
1999. Fusion of monocytes and macrophages with HIV-1 correlates with
biochemical properties of CXCR4 and CCR5. Nat. Med. 5:303–308.
65. Larder, B. 2001. Mechanisms of HIV-1 drug resistance. AIDS 15(Suppl.
5):S27–S34.
66. Lazzarin, A., B. Clotet, D. Cooper, J. Reynes, K. Arasteh, M. Nelson, C.
Katlama, H. J. Stellbrink, J. F. Delfraissy, J. Lange, L. Huson, R. DeMasi,
C. Wat, J. Delehanty, C. Drobnes, and M. Salgo. 2003. Efficacy of enfu-
virtide in patients infected with drug-resistant HIV-1 in Europe and Aus-
tralia. N. Engl. J. Med. 348:2186–2195.
67. Lee, B., M. Sharron, C. Blanpain, B. J. Doranz, J. Vakili, P. Setoh, E. Berg,
G. Liu, H. R. Guy, S. R. Durell, M. Parmentier, C. N. Chang, K. Price, M.
Tsang, and R. W. Doms. 1999. Epitope mapping of CCR5 reveals multiple
conformational states and distinct but overlapping structures involved in
chemokine and coreceptor function. J. Biol. Chem. 274:9617–9626.
68. Maeda, Y., M. Foda, S. Matsushita, and S. Harada. 2000. Involvement of
both the V2 and V3 regions of the CCR5-tropic human immunodeficiency
virus type 1 envelope in reduced sensitivity to macrophage inflammatory
protein 1␣. J. Virol. 74:1787–1793.
69. McCune, J. M. 2001. The dynamics of CD4
⫹
T-cell depletion in HIV
disease. Nature 410:974–979.
70. McKnight, A., D. Wilkinson, G. Simmons, S. Talbot, L. Picard, M. Ahuja,
M. Marsh, J. A. Hoxie, and P. R. Clapham. 1997. Inhibition of human
immunodeficiency virus fusion by a monoclonal antibody to a coreceptor
(CXCR4) is both cell type and virus strain dependent. J. Virol. 71:1692–
1696.
71. Meanwell, N. A., and J. F. Kadow. 2003. Inhibitors of the entry of HIV into
host cells. Curr. Opin. Drug Discov. Dev. 6:451–461.
72. Melikyan, G. B., R. M. Markosyan, H. Hemmati, M. K. Delmedico, D. M.
Lambert, and F. S. Cohen. 2000. Evidence that the transition of HIV-1 gp41
into a six-helix bundle, not the bundle configuration, induces membrane
fusion. J. Cell. Biol. 151:413–423.
73. Michael, N. L., and J. P. Moore. 1999. HIV-1 entry inhibitors: evading the
issue. Nat. Med. 5:740–742.
74. Moore, J. P., Y. Cao, L. Qing, Q. J. Sattentau, J. Pyati, R. Koduri, J.
Robinson, C. F. Barbas III, D. R. Burton, and D. D. Ho. 1995. Primary
isolates of human immunodeficiency virus type 1 are relatively resistant to
neutralization by monoclonal antibodies to gp120, and their neutralization
is not predicted by studies with monomeric gp120. J. Virol. 69:101–109.
75. Moore, J. P., J. A. McKeating, Y. X. Huang, A. Ashkenazi, and D. D. Ho.
1992. Virions of primary human immunodeficiency virus type 1 isolates
resistant to soluble CD4 (sCD4) neutralization differ in sCD4 binding and
glycoprotein gp120 retention from sCD4-sensitive isolates. J. Virol. 66:235–
243.
76. Moore, J. P., and J. Sodroski. 1996. Antibody cross-competition analysis of
the human immunodeficiency virus type 1 gp120 exterior envelope glyco-
protein. J. Virol. 70:1863–1872.
77. Moore, J. P., and M. Stevenson. 2000. New targets for inhibitors of HIV-1
replication. Nat. Rev. Mol. Cell. Biol. 1:40–49.
78. Moore, J. P., M. Thali, B. A. Jameson, F. Vignaux, G. K. Lewis, S. W. Poon,
M. Charles, M. S. Fung, B. Sun, P. J. Durda, L. Akerblom, B. Wahren, D. D.
Ho, Q. J. Sattentau, and J. Sodroski. 1993. Immunochemical analysis of the
gp120 surface glycoprotein of human immunodeficiency virus type 1: prob-
ing the structure of the C4 and V4 domains and the interaction of the C4
domain with the V3 loop. J. Virol. 67:4785–4796.
79. Morrison, H. G., F. Kirchhoff, and R. C. Desrosiers. 1993. Evidence for the
cooperation of gp120 amino acids 322 and 448 in SIVmac entry. Virology
195:167–174.
80. O’Hara, B. M., and W. C. Olson. 2002. HIV entry inhibitors in clinical
development. Curr. Opin. Pharmacol. 2:523–528.
81. Olson, W. C., G. E. Rabut, K. A. Nagashima, D. N. Tran, D. J. Anselma,
S. P. Monard, J. P. Segal, D. A. Thompson, F. Kajumo, Y. Guo, J. P. Moore,
P. J. Maddon, and T. Dragic. 1999. Differential inhibition of human im-
munodeficiency virus type 1 fusion, gp120 binding, and CC-chemokine
activity by monoclonal antibodies to CCR5. J. Virol. 73:4145–4155.
82. Palani, A., S. Shapiro, J. W. Clader, W. J. Greenlee, K. Cox, J. Strizki, M.
Endres, and B. M. Baroudy. 2001. Discovery of 4-[(Z)-(4-bromophenyl)-
(ethoxyimino)methyl]-1⬘-[(2,4-dimethyl-3-pyridinyl)carbonyl]-4⬘-methyl-
1,4⬘-bipiperidine N-oxide (SCH 351125): an orally bioavailable human
2806 KUHMANN ET AL. J. VIROL.

CCR5 antagonist for the treatment of HIV infection. J. Med. Chem. 44:
3339–3342.
83. Palani, A., S. Shapiro, H. Josien, T. Bara, J. W. Clader, W. J. Greenlee, K.
Cox, J. M. Strizki, and B. M. Baroudy. 2002. Synthesis, SAR, and biological
evaluation of oximino-piperidino-piperidine amides. 1. Orally bioavailable
CCR5 receptor antagonists with potent anti-HIV activity. J. Med. Chem.
45:3143–3160.
84. Platt, E. J., S. E. Kuhmann, P. P. Rose, and D. Kabat. 2001. Adaptive
mutations in the V3 loop of gp120 enhance fusogenicity of human immu-
nodeficiency virus type 1 and enable use of a CCR5 coreceptor that lacks
the amino-terminal sulfated region. J. Virol. 75:12266–12278.
85. Po¨hlmann, S., and R. W. Doms. 2002. Evaluation of current approaches to
inhibit HIV entry. Curr. Drug Targ. Infect. Dis. 2:9–16.
86. Poignard, P., E. O. Saphire, P. W. Parren, and D. R. Burton. 2001. gp120:
Biologic aspects of structural features. Annu. Rev. Immunol. 19:253–274.
87. Reeves, J. D., S. A. Gallo, N. Ahmad, J. L. Miamidian, P. E. Harvey, M.
Sharron, S. Pohlmann, J. N. Sfakianos, C. A. Derdeyn, R. Blumenthal, E.
Hunter, and R. W. Doms. 2002. Sensitivity of HIV-1 to entry inhibitors
correlates with envelope/coreceptor affinity, receptor density, and fusion
kinetics. Proc. Natl. Acad. Sci. USA 99:16249–16254.
88. Resch, W., N. Hoffman, and R. Swanstrom. 2001. Improved success of
phenotype prediction of the human immunodeficiency virus type 1 from
envelope variable loop 3 sequence using neural networks. Virology 288:51–
62.
89. Resch, W., N. Parkin, E. L. Stuelke, T. Watkins, and R. Swanstrom. 2001.
A multiple-site-specific heteroduplex tracking assay as a tool for the study
of viral population dynamics. Proc. Natl. Acad. Sci. USA 98:176–181.
90. Richman, D. D. 2001. HIV chemotherapy. Nature 410:995–1001.
91. Richman, D. D., T. Wrin, S. J. Little, and C. J. Petropoulos. 2003. Rapid
evolution of the neutralizing antibody response to HIV type 1 infection.
Proc. Natl. Acad. Sci. USA 100:4144–4149.
92. Rini, J. M., R. L. Stanfield, E. A. Stura, P. A. Salinas, A. T. Profy, and I. A.
Wilson. 1993. Crystal structure of a human immunodeficiency virus type 1
neutralizing antibody, 50.1, in complex with its V3 loop peptide antigen.
Proc. Natl. Acad. Sci. USA 90:6325–6329.
93. Rizzuto, C., and J. Sodroski. 2000. Fine definition of a conserved CCR5-
binding region on the human immunodeficiency virus type 1 glycoprotein
120. AIDS Res. Hum. Retroviruses 16:741–749.
94. Rizzuto, C. D., R. Wyatt, N. Hernandez-Ramos, Y. Sun, P. D. Kwong, W. A.
Hendrickson, and J. Sodroski. 1998. A conserved HIV gp120 glycoprotein
structure involved in chemokine receptor binding. Science 280:1949–1953.
95. Rose, P. P., and B. T. Korber. 2000. Detecting hypermutations in viral
sequences with an emphasis on G 3A hypermutation. Bioinformatics
16:400–401.
96. Samson, M., G. LaRosa, F. Libert, P. Paindavoine, M. Detheux, G. Vassart,
and M. Parmentier. 1997. The second extracellular loop of CCR5 is the
major determinant of ligand specificity. J. Biol. Chem. 272:24934–24941.
97. Schols, D., J. A. Este, C. Cabrera, and E. De Clercq. 1998. T-cell-line-tropic
human immunodeficiency virus type 1 that is made resistant to stromal
cell-derived factor 1␣contains mutations in the envelope gp120 but does
not show a switch in coreceptor use. J. Virol. 72:4032–4037.
98. Schwarz, M. K., and T. N. Wells. 2002. New therapeutics that modulate
chemokine networks. Nat. Rev. Drug Discov. 1:347–358.
99. Shankarappa, R., J. B. Margolick, S. J. Gange, A. G. Rodrigo, D. Upchurch,
H. Farzadegan, P. Gupta, C. R. Rinaldo, G. H. Learn, X. He, X. L. Huang,
and J. I. Mullins. 1999. Consistent viral evolutionary changes associated
with the progression of human immunodeficiency virus type 1 infection.
J. Virol. 73:10489–10502.
100. Shioda, T., J. A. Levy, and C. Cheng-Mayer. 1991. Macrophage and T
cell-line tropisms of HIV-1 are determined by specific regions of the enve-
lope gp120 gene. Nature 349:167–169.
101. Strizki, J. M., S. Xu, N. E. Wagner, L. Wojcik, J. Liu, Y. Hou, M. Endres,
A. Palani, S. Shapiro, J. W. Clader, W. J. Greenlee, J. R. Tagat, S. Mc-
Combie, K. Cox, A. B. Fawzi, C. C. Chou, C. Pugliese-Sivo, L. Davies, M. E.
Moreno, D. D. Ho, A. Trkola, C. A. Stoddart, J. P. Moore, G. R. Reyes, and
B. M. Baroudy. 2001. SCH-C (SCH 351125), an orally bioavailable, small
molecule antagonist of the chemokine receptor CCR5, is a potent inhibitor
of HIV-1 infection in vitro and in vivo. Proc. Natl. Acad. Sci. USA 98:
12718–12723.
102. Sullivan, N., Y. Sun, Q. Sattentau, M. Thali, D. Wu, G. Denisova, J.
Gershoni, J. Robinson, J. Moore, and J. Sodroski. 1998. CD4-Induced
conformational changes in the human immunodeficiency virus type 1 gp120
glycoprotein: consequences for virus entry and neutralization. J. Virol.
72:4694–4703.
103. Tagat, J. R., S. W. McCombie, R. W. Steensma, S. Lin, D. V. Nazareno, B.
Baroudy, N. Vantuno, S. Xu, and J. Liu. 2001. Piperazine-based CCR5
antagonists as HIV-1 inhibitors. I: 2(S)-methyl piperazine as a key phar-
macophore element. Bioorg. Med. Chem. Lett. 11:2143–2146.
104. Tagat, J. R., R. W. Steensma, S. W. McCombie, D. V. Nazareno, S. I. Lin,
B. R. Neustadt, K. Cox, S. Xu, L. Wojcik, M. G. Murray, N. Vantuno, B. M.
Baroudy, and J. M. Strizki. 2001. Piperazine-based CCR5 antagonists as
HIV-1 inhibitors. II. Discovery of 1-[(2,4-dimethyl-3-pyridinyl)carbonyl]-4-
methyl-4-[3(S)-methyl-4-[1(S)-[4-(trifluoromethyl)phenyl]ethyl]-1-piperazi-
nyl]-piperidine N
1
-oxide (Sch-350634), an orally bioavailable, potent CCR5
antagonist. J. Med. Chem. 44:3343–3346.
105. Thali, M., J. P. Moore, C. Furman, M. Charles, D. D. Ho, J. Robinson, and
J. Sodroski. 1993. Characterization of conserved human immunodeficiency
virus type 1 gp120 neutralization epitopes exposed upon gp120-CD4 bind-
ing. J. Virol. 67:3978–3988.
106. Tremblay, M. J., J. F. Fortin, and R. Cantin. 1998. The acquisition of
host-encoded proteins by nascent HIV-1. Immunol. Today 19:346–351.
107. Trkola, A., T. Dragic, J. Arthos, J. M. Binley, W. C. Olson, G. P. Allaway,
C. Cheng-Mayer, J. Robinson, P. J. Maddon, and J. P. Moore. 1996.
CD4-dependent, antibody-sensitive interactions between HIV-1 and its co-
receptor CCR-5. Nature 384:184–187.
108. Trkola, A., S. E. Kuhmann, J. M. Strizki, E. Maxwell, T. Ketas, T. Morgan,
P. Pugach, S. Xu, L. Wojcik, J. Tagat, A. Palani, S. Shapiro, J. W. Clader,
S. McCombie, G. R. Reyes, B. M. Baroudy, and J. P. Moore. 2002. HIV-1
escape from a small molecule, CCR5-specific entry inhibitor does not in-
volve CXCR4 use. Proc. Natl. Acad. Sci. USA 99:395–400.
109. Trkola, A., A. B. Pomales, H. Yuan, B. Korber, P. J. Maddon, G. P. Allaway,
H. Katinger, C. F. Barbas, 3rd, D. R. Burton, D. D. Ho, and J. P. Moore.
1995. Cross-clade neutralization of primary isolates of human immunode-
ficiency virus type 1 by human monoclonal antibodies and tetrameric CD4-
immunoglobulin G. J. Virol. 69:6609–6617.
110. Trouplin, V., F. Salvatori, F. Cappello, V. Obry, A. Brelot, N. Heveker, M.
Alizon, G. Scarlatti, F. Clavel, and F. Mammano. 2001. Determination of
coreceptor usage of human immunodeficiency virus type 1 from patient
plasma samples by using a recombinant phenotypic assay. J. Virol. 75:251–
259.
111. Tsamis, F., S. Gavrilov, F. Kajumo, C. Seibert, S. Kuhmann, T. Ketas, A.
Trkola, A. Palani, J. W. Clader, J. R. Tagat, S. McCombie, B. Baroudy, J. P.
Moore, T. P. Sakmar, and T. Dragic. 2003. Analysis of the mechanism by
which the small-molecule CCR5 antagonists SCH-351125 and SCH-350581
inhibit human immunodeficiency virus type 1 entry. J. Virol. 77:5201–5208.
112. Wei, X., J. M. Decker, H. Liu, Z. Zhang, R. B. Arani, J. M. Kilby, M. S.
Saag, X. Wu, G. M. Shaw, and J. C. Kappes. 2002. Emergence of resistant
human immunodeficiency virus type 1 in patients receiving fusion inhibitor
(T-20) monotherapy. Antimicrob. Agents Chemother. 46:1896–1905.
113. Wei, X., J. M. Decker, S. Wang, H. Hui, J. C. Kappes, X. Wu, J. F.
Salazar-Gonzalez, M. G. Salazar, J. M. Kilby, M. S. Saag, N. L. Komarova,
M. A. Nowak, B. H. Hahn, P. D. Kwong, and G. M. Shaw. 2003. Antibody
neutralization and escape by HIV-1. Nature 422:307–312.
114. Westervelt, P., H. E. Gendelman, and L. Ratner. 1991. Identification of a
determinant within the human immunodeficiency virus 1 surface envelope
glycoprotein critical for productive infection of primary monocytes. Proc.
Natl. Acad. Sci. USA 88:3097–3101.
115. Wild, C., T. Oas, C. McDanal, D. Bolognesi, and T. Matthews. 1992. A
synthetic peptide inhibitor of human immunodeficiency virus replication:
correlation between solution structure and viral inhibition. Proc. Natl.
Acad. Sci. USA 89:10537–10541.
116. Wu, L., N. P. Gerard, R. Wyatt, H. Choe, C. Parolin, N. Ruffing, A. Borsetti,
A. A. Cardoso, E. Desjardin, W. Newman, C. Gerard, and J. Sodroski. 1996.
CD4-induced interaction of primary HIV-1 gp120 glycoproteins with the
chemokine receptor CCR-5. Nature 384:179–183.
117. Wu, L., G. LaRosa, N. Kassam, C. J. Gordon, H. Heath, N. Ruffing, H.
Chen, J. Humblias, M. Samson, M. Parmentier, J. P. Moore, and C. R.
Mackay. 1997. Interaction of chemokine receptor CCR5 with its ligands:
multiple domains for HIV-1 gp120 binding and a single domain for che-
mokine binding. J. Exp. Med. 186:1373–1381.
118. Wyatt, R., P. D. Kwong, E. Desjardins, R. W. Sweet, J. Robinson, W. A.
Hendrickson, and J. G. Sodroski. 1998. The antigenic structure of the HIV
gp120 envelope glycoprotein. Nature 393:705–711.
119. Wyatt, R., and J. Sodroski. 1998. The HIV-1 envelope glycoproteins: fuso-
gens, antigens, and immunogens. Science 280:1884–1888.
120. Wyatt, R., M. Thali, S. Tilley, A. Pinter, M. Posner, D. Ho, J. Robinson, and
J. Sodroski. 1992. Relationship of the human immunodeficiency virus type
1 gp120 third variable loop to a component of the CD4 binding site in the
fourth conserved region. J. Virol. 66:6997–7004.
121. Ye, Y., Z. H. Si, J. P. Moore, and J. Sodroski. 2000. Association of structural
changes in the V2 and V3 loops of the gp120 envelope glycoprotein with
acquisition of neutralization resistance in a simian-human immunodefi-
ciency virus passaged in vivo. J. Virol. 74:11955–11962.
122. Zhang, Y. J., and J. P. Moore. 1999. Will multiple coreceptors need to be
targeted by inhibitors of human immunodeficiency virus type 1 entry?
J. Virol. 73:3443–3448.
123. Zwick, M. B., R. Kelleher, R. Jensen, A. F. Labrijn, M. Wang, G. V.
Quinnan, Jr., P. W. H. I. Parren, and D. R. Burton. 2003. A novel human
antibody against human immunodeficiency virus type 1 gp120 is V1, V2,
and V3 loop dependent and helps delimit the epitope of the broadly
neutralizing antibody immunoglobulin G1 b12. J. Virol. 77:6965–6978.
VOL. 78, 2004 HIV-1 ESCAPE FROM A CCR5 ANTAGONIST 2807