Báo cáo Y học: Structural basis for the inhibitory efficacy of efavirenz (DMP-266), MSC194 and PNU142721 towards the HIV-1 RT K103N mutant doc

Structural basis for the inhibitory efficacy of efavirenz (DMP-266),
MSC194 and PNU142721 towards the HIV-1 RT K103N mutant
Jimmy Lindberg
1
, Snævar Sigurðsson
1
, Seved Lo¨ wgren
1
, Hans O. Andersson
1
, Christer Sahlberg
2
,
Rolf Nore
´
en
2
, Kerstin Fridborg
1
, Hong Zhang
2
and Torsten Unge
1
1
Department of Cell and Molecular Biology, Uppsala Biomedical Center, Uppsala University, Sweden;
2
Medivir AB, Huddinge,
Sweden
The K 103N substitution is a frequently observed HIV-1 RT
mutation in patients who do not respond to combination-
therapy. The drugs Efavirenz, MSC194 and PNU142721

belong to the r ecent generation of NNRTIs characterized by
an improved resistance profile to the most common single
point mutations within HIV-1 R T, including t he K103N
mutation. In the present study we present structural obser-
vations from Efavirenz i n complex with wild-type p rotein
and the K103N mutant and PNU142721 and MSC194 in
complex with the K103N mutant. The structures unani-
mously indicate that the K103N substitution induces only
minor positional adjustments of the three inhibitors and the
residues lining t he binding pocket. Thus, compared to t he
corresponding wild-type structures, these inhibitors bind to
the mutant in a conservative mode rather than through
major rearrangements. The structures implicate that the
reduced inhibitory efficacy should be attributed to the
changes in the chemical environment in the vicinity o f
the substituted N103 residue. T his i s s upported by c hanges in
hydrophobic and electrostatic interactions to the inhibitors
between wild-type and K103N mutant complexes. These
potent inhibitors accommodate to the K103N mutation by
forming new interactions to the N103 side c hain. Our results
are consistent with the proposal by Hsiou et al. [Hsiou, Y.,
Ding, J., Das, K., Clark, A.D. Jr, Boyer, P.L., Lewi, P.,
Janssen, P.A., Kleim, J.P., Rosner, M., Hughes, S.H. &
Arnold, E. (2001) J. Mo l. Biol. 30 9, 4 37–445] that inhibitors
with good activity against the K103N mutant would be
expected to have favorable interactions with the mutant
asparagines side chain, thereby compensating for resistance
caused by stabilization of the mutant enzyme due to a
hydrogen-bond network involving the N 103 and Y188 side
chains.

Keywords: drug-resistance; HIV; NNRTI; reverse tran-
scriptase.
The u se of highly active ant iretroviral t herapy (HAART)
involving multidrug combinations has s ignificantly reduced
the death rates of HIV-1 infected individuals receiving such
treatment [1]. Inhibitors of the HIV-1 reverse transcriptase
(RT) constitute a cornerstone in this therapy and are
commonly used in combination w ith inhibitors of the H IV-1
protease. The RT inhibitors belong to two classes, the
nucleoside inhibitors and the non-nucleoside inhibitors
(NNRTI). Whereas the NRTIs are nucleoside analogues
with chain-terminating properties and affinity to active site
residues, the NNRTIs include a wide range of series of
chemical compounds characterized by noncompetitive
binding to an allosteric site some 10 A
˚
away from the
active site. Structural comparison o f RT i n complex with
template/primer and NNRTIs together with native R T
complexes have shown that the NNRTIs inhibit the
polymerase activity through long-range and short-range
structural distortions in several of the RT subdomains. The
distortions involve repositioning o f residues in the no n-
nucleoside binding pocket (NNIBP) that impose steric
impediments on the thumb subdomain flexibility forcing it
to remain in the open conformation. In addition, the
RNase H activity as well as initiation of polymerization
may be affected by these NNRTI-induced distortions [2,3].
Despite the in itial efficacy in combating HIV infection,
NNRTIs select for multidrug resistant strains of HIV over

time [4,5]. The mutations occur exclusively among the
residues in the NNIBP. The ne w generation NNRTIs, e.g.
Efavirenz, s elect for a panel of resistance mutations K103N,
V106I, V108I, Y181C, Y188H Y188L, G190S, P225H, and
F227L, indicating that a majority of t he NNIBP residues
are potential sites for drug-resistant mutations [6].
The Ôfirst generationÕ NNRTIs, such as the c urrently
marketed drugs N evirapine and Delavirdine show orders of
magnitude decreases in binding as a r esult of single point
mutations [7,8]. The so-called Ôsecond generationÕ NNRTIs
such as Efavirenz (DMP-266) [9], carboxanilides [10], PETT
analogues [11] and the recent member S-1153 [12] de mon-
strate more favorable resistance p rofiles. Efforts are now
Correspondence to T. Unge, Department of Cell and Molecular
Biology, Uppsala Bi omedical C enter, Uppsala University, Box 596,
SE-751 24 Up psala, Sweden.
Fax: + 4618536971, Tel.: + 46184714985,
E-mail:
Abbreviations: HAART, h ighly a ctive antiretroviral therapy;
HIV-1, human immunodeficiency virus type 1; RT, reverse
transcriptase; rms, root mean square; NNRTI, non-nucleoside
RT inhibitor; NNIBP, non-nucleoside inhibitor NNIBP.
Note: The coordinates have b een deposited i n the Protein Data Bank
(PDB) with accession codes 1IKW, 1IKV, 1IKY and 1IKX for wild-
type RT-Efavirenz, K103N RT-Efavirenz, K103N RT-MSC194 and
K103N RT-PNU142721, respectively.
(Received 8 October 2001, revised 28 December 2001, accepted
24 January 2002)
Eur. J. Biochem. 269, 1670–1677 (2002) Ó FEBS 2002
being put on the design o f n ew inhibitors with improved

resistance profiles to the most frequently drug-induced
mutations generated within RT.
The K 103N mutation is the most frequent mutation
observed within R T resulting from therapeutic interventions
involving NNRTIs [6,13–15]. As indicated above K 103 is
one of the NNIBP residues. The position of t he K103
residue is close t o the entrance of the pocket and contributes
through its aliphatic carbons to the hydrophobic character
of that part of the pocket that interacts with wing2 of t he
inhibitor compounds [16]. The combination of a broad
cross-resistance to the K103N mutant and the fact that
previous crystal s tructures o f a number of inhibitors did not
indicate a direct contact to this residue, h ave encouraged
additional explanations to the resistance phenomenon.
In support of an alternative resistance mechanism is the
observation of a hydrogen-bond network as a direct result
of the mutation. The network could stabilize the closed
conformation of the NNIBP [17]. This result is consistent
with kinetic data, which indicate the presence of a s teric
barrier in the K103N mutant affecting NNRTI entrance to
the NNIBP [8]. Despite these negative effects on drug
binding, no reduction in viral replication capacity has been
observed [18].
In this study we present the structural indications for the
role of K103 and N103 in drug binding and t he structural
implications for the inhibitory efficacy of the inhibitors
Efavirenz, PNU 142721, and MSC194 against the K103N
mutant. The results are deduced from comparisons of
crystal s tructures of inhibitor complexes of wild-type a nd
the K103N mutant.

MATERIALS AND METHODS
Protein expression, purification, and crystallization
The RT gene (HIV-1, BH10 isolat e, nucleotides 1908–3587)
was isolated by PCR, and ligated into the pET 11a
expression vector at the NdeI/BamHIsitesaspreviously
described [19]. Through this construct, th e protein sequence
of 560 amino acids was provided with an N-terminal
methionine. However, the methionine was processed by
bacterial proteases and never detected in the electron
density. In order to extinguish the RNase H activity, r esidue
E478 (GAG) was mutated to Q (CAG) by site-directed
mutagenesis. RT was expressed in the Escherichia coli,strain
BL21 (DE3) and purified as described previously [19] with
the following modifications. Instead of allowing HIV-1
protease or bacterial proteases to process the RT p66/p66
homodimer to the p66/p51 h eterodimer, processing was
performed by chymotrypsine D digestion of the total
bacterial lysate for 60 min immediately prior to purification
(1 mg to 30 mL lysate). In the chromatographic steps
the ion exchange and affinity matrices POROSÒ HQ,
POROSÒ SandPOROSÒ HE (PerSeptive B iosystems)
were used. Purified protein was used in evaluating the
antiviral activity of the three inhibitors in th e HIV-1 RT
enzyme assay described previously [20]. Prior to crystalliza-
tion, RT was concentrated by precipitation with 2
M
(NH
4
)
2

SO
4
and red issolved in distilled water. C rystalliza-
tion was performed by vapor diffusion as f ollows. Drops
consisting of 5 lLpremixedRT(20mgÆmL
)1
) and twofold
molar access inhibitor (30 m
M
in dimethylsulfoxide) toge-
ther with 5 lL of crystallization buffer [1.4
M
(NH
4
)
2
SO
4
,
50 m
M
Hepes pH 7.2, 5 m
M
MgCl
2
, 300 m
M
KCl] were
equilibrated against the same buffer at room temperature.
Typically, crystals appeared within two weeks and grew to a

size of 0.3 · 0.2 · 0.2 mm w ithin two mon ths. Crystals
belong to the orthorhombic space group, C222
1
.
Data collection and processing, structure solution
and refinement
X-ray data were collected at 4 °C using the Max-Labora-
tory beam line 711 and ESRF beam line B M14. Indexing
and integration of data were performed using
DENZO
,the
data were merged together with
SCALEPACK
[21], and further
processing was performed with the
CCP
4 program suite [22].
Essential details of data collection and processing are given
in Table 1. T he structures were refined b y e mploying the
software program
CNS
[23]. The protein model coordinates
from 1hni were used for rotation and translation functions.
The r ms deviations for the C a between the initial model and
the final structures were in the range of 1.9–2.1 A
˚
.
Inhibitory parameters were generated with
XPLO
2

D
[24].
The refinement proceeded with energy minimization, simu-
lated annealing, and individual B-factor refinemen t and
were monitored by the statistical values R
work
/R
free
[25].
Model building was carried out using the software program
MAPMAN
[26],
LSQMAN
[27] and
O
[28]. Difference F ourier
maps were calculated with ligand and residue 103 omitted
employing the omit-map option in
CNS
. All figures were
produced u sing
SWISS
-
PDB VIEWER
v. 3.51 [29] and 3D-ren-
dered with
POV
-
RAY
v. 3.1 [30].

RESULTS
We have determined the X-ray structures of wild-type and
the K103N mutant in complex with Efavirenz at 3.0 A
˚
resolution. Two additional K103N mutant complexes
were structurally determined together with a PETT ana-
logue (MSC194) and a pyrimidine thioether analogue,
PNU142721 at 3.0 and 2.8 A
˚
resolution, respectively. The
mean temperature factors was t ypically 55 A
˚
2
for all at oms,
30–60 A
˚
2
for the inhibitor atoms and lining residues. The
temperature factors for K103 a nd N103 atoms were i n
the range of 50–60 A
˚
2
. All complexes crystallized with the
symmetry of space group C222
1
.
Overall hydrophobic NNRTI interactions to wild-type
and K103N mutant RT
The interactions of Efavirenz, MSC194 and PNU142721 to
wild-type RT and the K103N mutant correspond to

previously described RT/NNRTI complexes [11,31]. The
NNRTI interactions are predominantly of hydrophobic
nature to pivotal r esidues from p66 and p51 lining the
NNIBP. The aromatic residues Y181, Y188 and W229
surround wing1 whereas wing2 is sandwiched between L100
and V106, while also making edge-on contacts with V179
and Y318A. A prominent nonhydrophobic contact is the
hydrogen bond formed between the NNRTIs and the
backbone carbonyl of K101. Omit electron density maps of
the R T/NNRTI complexes clearly show the orientation and
conformation of the inhibitors in the NNIBP, including the
mutated side-chain at position 103 (Fig. 2A,B).
Ó FEBS 2002 The HIV-1 RT K103N mutant and inhibitor efficacy (Eur. J. Biochem. 269) 1671
Antiviral activity
The three NNRTIs in this study, Efavirenz, MSC194, and
PNU142721, were t ested in HIV-RT enzyme assays wit h
wild-type RT and the K103N mutant [20]. The IC
50
values
from these assays are presented in Table 2. The activity
measurements rank the inhibitors MSC194, PNU142721
and Efavirenz according to potency. The inhibitory efficacy
to wild-type RT is within subnanomolar range fo r all three
inhibitors. PNU142721 and MSC194 show a 3–10-fold
reduction in efficacy for the K103N mutant. The effect of
the mutation is more p ronounced for E favirenz where t here
is a 200- fold reduction in efficacy compared t o wild-type
RT. This corresponds to a 5–10-fold larger v alue than
obtained by others [9]. T his may be due to the use of
homopolymeric rC-dG template in the assay. The assay

shows that PNU142721 is the most potent of the three
towards the K103N mutant with an IC
50
value of 9 n
M
.In
contrast the first generation NNRTI, Nevirapine is 20-fold
lesspotentcomparedtowild-typewithanIC
50
value of
3800 n
M
.
Wild-type and K103N mutant RT/Efavirenz complexes
are structurally conserved
Structural analysis of Efavirenz i n complex with the K103N
mutant revealed that the overall position of the inhibitor as
well as the r esidues lining the NNIBP corresponds to the
wild-type–Efavirenz complex (Fig. 3). C omparison between
the two complexes shows an rms-deviation of 0.4 A
˚
(all
atoms) and 0.20 A
˚
(Ca atoms) fo r r esidues w ithin 4.0 A
˚
of
the i nhibitor. The only significant difference is the K103N
substitution. In wild-type RT K103 is protruding into a
negatively charged patch composed of the backbone

carbonyls of K102 and G191 and the s ide chain of D192
Table 1. C rystallographic structure determination statistics.
WT-Efavirenz K103N-Efavirenz K103N-MSC194 K103N-PNU142721
Data collection details
Data collection site MAX-lab beam line 711 MAX-lab beam line 711 ESRF beam line BM14 MAX-lab beam line 711
Image plate MAR-research MAR-research MAR-research MAR-research
Space group C222
1
C222
1
C222
1
C222
1
Wavelength (A
˚
) 1.0232 1.0232 0.931 1.0159
Unit cell dimensions (A
˚
) 119.54, 157.31, 157.17 119.63, 157.17, 156.19 120.34, 156.54, 156.47 119.90, 156.40, 156.90
Resolution range (A
˚
) 25–3.0 25–3.0 50–3.0 25–2.8
Observations 254 669 48 782 261 528 289,120
Unique reflections 30 024 29 948 32 188 36,534
Completeness (%) 96.4 97.6 85.2 90.3
Reflections with F/rF > 3 19 795 16 687 24 601 27,906
R
merge
a

0.113 0.152 0.114 0.079
Outer resolution shell
Resolution range (A
˚
) 3.11–3.00 3.11–3.00 3.16–3.09 2.90–2.80
Unique reflections 2873 2856 1170 3,306
Completeness (%) 97.9 97.8 73.6 92.3
Reflections with F/rF > 3 765 383 124 1,488
Refinement statistics
Resolution range (A
˚
) 25.0–3.0 25.0–3.0 25–3.0 25–2.8
Reflections (working/test) 29,828/1505 25,992/1309 27,346/1385 32,921/1,661
R-factor
b
(R
work
/R
free
) 0.218/0.272 0.229/0.292 0.208/0.266 0.210/0.273
Rms bond length deviation
c
(A
˚
) 0.008 0.008 0.008 0.007
Rms bond angle deviation (°) 1.4 1.4 1.4 1.3
Rms dihedral angle deviation
c
(°) 22.6 23.1 22.7 22.6
Rms improper angle deviation

c
(°) 0.96 0.95 0.93 0.93
Mean B-factor (A
˚
2
)
d
75.3 53.0 60.6 54.9
a
R
merge
¼ S|I–<I>|/S<I>.
b
R-factor ¼ S|F
o
–F
c
|/SF
o
.
c
Ideal parameters are those defined by Engh and Huber.
d
Mean B-factor for
main chain, side chain, inhibitor and water atoms, respectively.
Fig. 1. Struc tures of NNRTIs. Chemical structure o f t he NN RTIs ( A)
Efavirenz, (B) PNU142721, and (C) MSC194. Atom numbering was
included for clarification of Table 3.
1672 J. Lindberg et al. (Eur. J. Biochem. 269) Ó FEBS 2002
forming a weak hydrogen bond interaction to D192

(3.16 A
˚
). However, in the K103N mutant complexes the
orientation of the N103 amide is undefined. Modeling of the
amide i n the Efavirenz complex resulted in an optimal
distance to residue D192 of 3.5 A
˚
.
The binding mode of Efavirenz to wild-type RT only
allows a few contacts between the K103 residue and the
inhibitor ( Table 3). The c ontact distances from the back-
bone N a nd the C b and Cc atoms to th e inhibitor a re all
within optimal van der Waals d istance for close packing
interactions. T he interactions of Efavirenz to the substituted
N103 residue is conserved compared to t he wild-type except
for the contact with Cc. The K103 Cc methylene group is
replaced by a bulky amide of N 103 that consequently
abolish the interaction.
Introduction of the asparagine at position 103 in the
NNIBP induces a m inute orientational shift of Efavirenz.
The effect on the i nhibitor is observed a s a minor rotation
around the branching carbon of Efavire nz. Consequently,
the trifluoromethyl g roup is repositioned 0 .2 A
˚
away and
the O10 of the benzoaxine-2-one ring 0.3 A
˚
towards the
N103 amide. These subtle changes are accompanied by
repositioning of the side chain of V179 0.4 A

˚
towards and of
D192 0.7 A
˚
away from the inhibitor with respect to the
wild-type-Efavirenz complex.
Fig. 2. Orientation and Conformation of Efavirenz and Residue 103 in the w ild-type RT and K103N mutant NNIBPs. (A) Simulated annealing o mit
electron density map covering E favirenz and residue K103 (green) i n the wild-type NNIB P. In ( B) the same v iew is shown for E favirenz an d residue
N103 (maroon) in the mutated N NIBP. Re sidue s ide-chains characteristic of the wild-type and K103N mutant NNIBPs are colored a ccordingly.
The map was calculated w ith ligand and residu e 103 om itted employing the o mit-map op tion in CN S and contou red at 1.5 r.
Table 2. Inhibition of HIV-1 RT.
HIV-1RT (rCdG), IC
50
(n
M
)
a
Wild-type K103N
MSC194 5.5 52
PNU142721 2.5 7.0
Efavirenz 2.5 520
Nevirapine 170 3800
a
The HIV-1 RT assay which used (poly)rCÆ(oligo)dG as the tem-
plate/primer is described in [23].
Fig. 3. Superimposition o f Efavirenz bound to wild-type a nd K 103N mutant RT NNIBPs. Stereoview of the superimposition of Efavirenz bound t o
the NNIBP of wild-type RT and the K103N mu tant. Residue side chain s c haracteristic of the NNIBP are included from e ach inhibitor com plex a nd
colored green for wild-type and m aroon for the K103N mu tant. T he superimposition was c arried out using all atoms from the residues within 4.0 A
˚
from the inhibitors (V189, K101, K103N, V179, Y181, Y188, F227, W229, L234, H235, Y318 and E138).

Ó FEBS 2002 The HIV-1 RT K103N mutant and inhibitor efficacy (Eur. J. Biochem. 269) 1673
Similar overall NNRTI binding mode to K103N mutant RT
In Fig. 4 the binding modes of the three inhibitors are
superimposed in the mutant NNIBP. The cyclopropyl-
group of MSC194 and the methyl group of PNU142721
partly overlap the trifluoromethyl group of Efavirenz.
Furthermore, wing2 composed of the heterocyclic ring
structure of MSC194 and the substituted pyrimidine
functionality of PNU142721 occupy the same part of the
NNIBP as the benzoaxine-2-one ring of Efavirenz. In a
similar manner, the position of wing1 composed of the
substituted phenyl ring of MSC194 and the fused ring-
structure of PNU142721 overlap the less bulky cyclopropyl
group of Efavirenz.
The overall similarity in binding mode of the inhibitors
to the K103N mutant means t hat several pivotal inter-
actions are shared to key residues i n the NNIBP, des-
pite the chemical differences among the inhibitors. This
similarity is apparent when the three structures are
superimposed (Fig. 4 ). Only subtle changes in the over-
all positioning and orientation of residues lining the
NNIBP can be observed among the three complexes.
A structural comparison of the K103N mutant-MSC194
and mutant-PNU142721 complexes w ith Efavirenz shows
an rms-deviation of 0.9/0.4 A
˚
(all/Ca atoms) and 0.7/0.3 A
˚
(all/Ca atoms), respectively, for residues within 4.0 A
˚

of the
inhibitors.
Regardless of the overall s imilarities in the NNIBP of the
K103N mutant structures t he chemically different inhibitors
affect the position and orientation of particular residues
differently. This is apparent in the K103N mutant-MSC194
complex where a flip of E138 in p51 to a downward rotamer
is observed in contrast to Efavirenz and PNU142721.
Furthermore, both MSC194 and PNU142721 bound to the
mutant show minu te displacements of Y181 and Y 188 away
from the inhibitors compared to Efavirenz. These rear-
rangements allow for a rotation o f the side chain of F227
with respect to t he position in the K103N mutant-Efavirenz
complex and the phenyl ring is observed with the partially
positively charged side pointing towards MSC194 and
PNU142721. In addition, the difference in c hemical struc-
ture of wing2 in PNU142721 with respect to MSC194 and
Efavirenz induces a change in t he rotamer of D192, thereby
disrupting the negative patch.
DISCUSSION
Anti-HIV compounds belonging to the new generation of
NNRTIs have increased inhibitory efficacy with respect to
wild-type and a number of d rug resistant RT mutants.
Table 3. I nter-atomic distances for the inhibitors Efavirenz, MSC194, PNU142721 to residue 103 in the wild-type and K103N mutant NNIBPs.
Distance units are in A
˚
. Interactions to Od and Nd of the N103 amide have been left out due to the undefined amide orientation. The atom
numbering is clarifi ed in Fig. 1.
Residue 103 WT-Efavirenz K103N-Efavirenz K103N-MSC194 K103N-PNU142721
Cb C6 4.2 C6 4.0 C6 3.8

C5 3.9
N8 3.9
N2 3.8
N C6 4.1 C6 3.9 C6 3.7 N18 3.4
Cc N8 3.7 N8 3.7 N2 4.0
F1 3.8 C9 3.9
S1 4.0
Fig. 4. Supe rimposition of three K103N mutant NNIBPs. Stere oview of the sup erimposition of Efavirenz (m aroon), MSC194 (light blue ) and
PNU142721 (yellow) bound to the K103N mutant NNIBP. Residue side chains characteristic of the NNIBP are included from each inhibitor
complex a nd colored accordingly. The superimposition was carried ou t using all atom s from the residu es w ithin 4.0 A
˚
from the inhibitors ( V189,
K101, K103N, V179, Y181, Y188, F227, W229, L234, H235, Y318 and E1138).
1674 J. Lindberg et al. (Eur. J. Biochem. 269) Ó FEBS 2002
Accordingly, the NNRTIs Efavirenz, PNU142721, and
MSC194 have IC
50
-values in the nanomolar range t o wild-
type RT as well as to the f requently occurring K103N
mutant. Not unexpectedly the efficacy towards this mutant
differs significantly among the three inhibitors. The
IC
50
-values range from 7.0 n
M
(PNU142721) to 520 n
M
(Efavirenz) (Table 2). These inhibitory constants are orders-
of-magnitude lower than the corresponding values for the
first generation NNRTI, Nevirapine. A detailed structural

analysis of the binding mode of the inhibitors in complex
with wild-type RT and the K103N mutant should give
insight i nto the structural basis f or the inhibitory efficacy
and the resistance phenomenon. Previously, a few studies of
inhibitors in complex with RT mutants have been reported
[17,32]. The analysis of the inhibitors HBY 097 in complex
with the Y188L mutant and 8-Cl TIBO in complex with the
Y181C mutant, revealed that the retained efficacy of these
inhibitors was due to only minor alterations in the binding
mode compared to wild-type RT [32,33]. A significantly
different result was obtained by Ren et al. for the inhibitor
Efavirenz in complex with the K103N mutant determined to
2.9 A
˚
resolution [34]. In this case the binding mode was
different from the previous inhibitor complex structures.
Indicating the flexibility of t he RT structure. Accompanying
this alteration in binding mode was a repositioning of Y181
into a position close to what has been found in the
RT/DNA complex [35]. In addition, the structure of the p66
subunit displayed a more open conformation compared to
our K103N mutant structure. Ren et al. [34] c oncluded t hat
the e fficacy of Efavirenz against t he K103N mutant was du e
to this novel binding mode.
The data presented herein indicate that the K103N
mutation induces minute repositions of the inhibitors
PNU142721, MSC194 and efavirenz, w ith minor readjust-
ments in t he positions of residues lining t he bind ing site.
Thus, compared to the corresponding wild-type structures,
these inhibitors bind to the mutant in a conservative mode

rather than through major rearrangements of the inhibitor
and binding site.
The consequences of a conserved binding mode
on inhibitory efficacy
In order to allow for bigger readjustments of the inh ibitor and
lining amino-acid residues, the compound needs t o b e s mal-
ler than the accessible volume. This requirement is fulfilled
for Efavirenz, whereas MSC194 and PNU142721 effectively
occupy the binding volume with extensive interactions.
Accordingly, MSC194 has a similar b inding mode to the
K103N mutant a s the chemically related inhibitors MSC204
and M SC215 h ave to wild-type R T [11]. Superimposition o f
the structure complexes of MSC194 and PNU142721
show that these inhibitors bind to the K103N mutant in
essentially the same way. Only minor adjustments are seen
in the positioning of the inhibitors an d lining r esidues.
Interestingly, our structural studies of efavirenz in
complex with wild-type RT and the K103N mutant
revealed the same conservative binding mode for Efavirenz
as observed for PNU142721 and MSC194. In addition, the
Efavirenz complexes superimpose well with t he structures
of PNU142721 and MSC194. Thus, the reasons for
resistance, and the individual differences in efficacy
exhibited by t hese inhibitors, should be found among the
minor local structural and chemical differences in the
vicinity of the mutation. The substitutions of a charged
and linear lysine for a uncharged and branched asparagine
at position 103 result in a drastic change in the chemical
environment in the proximity of the mutation. This has
mainly two consequences for the binding of NNRTIs:

changed hydrophobic and electrostatic properties of the
NNIBP.
Changes in hydrophobic interactions induced
by the K103N mutation
The aliphatic carbons of K103 make hydrophobic close
packing contacts with wing2 of t he inhibitor compounds.
The extent of these interactions is more abundant for
MSC194 than for PNU142721 and Efavirenz (Table 3).
The effects of the K103N mutation on the hydrophobic
interactions reveal individual differences among the t hree
compounds. Though t he electron density for the a mide part
of the residue is not very well defined for any of the inhibitor
complexes, the inhibitors can still be clearly ranked with
respect to the extent of the van der Waals interactions:
MSC194, PNU142721 and Efavirenz. The more extensive
interactions of MSC194 and PNU142721 to the asparagine
residue are in agreement with the higher e fficacy of these
compounds compared to Efavirenz, shown by the antiviral
data (Table 2).
Changes in electrostatic interactions induced
by the K103N mutation
The e lectron density fo r Cc of the K 103N mut ant st ructures
is well defined but the quality of t he map does n ot allow
assessment of the orientation of the amide p lane. There are,
however, marginal differences in the quality of the electron
density, with the most featured density for MSC194. In the
case of Efavirenz, it is difficult to model the amide dipole in
such a way that electrostatic repulsion will not occur
with neighboring amino-acid residues. In the p51 subunit
the N 103 am ide i s orientated w ith N d2 positioned in the

negatively charged patch composed of D192, w hile the
backbone carbonyls of G191 and K102 impose a stabilizing
effect on that regio n of RT. A similar orientation in the p 66
subunit positions O d1 in close proximity t o the highly
electronegative trifluoromethyl moiety of Efavirenz and the
sulfur atoms of PNU142721 and MSC194, with repulsion
as a c onsequence. However, in the cases of MSC194 and
PNU142721 the position of the sulfur atom is such that the
repulsive forces are less apparent. In t he MSC194 mutant
complex the rotational f reedom of the amide is reduced by
the s tacking of O d1 in between the plane of the thiourea
moiety and the Ca of G190.
Hence, the undefined orientation of the N103 amide Od1
and Nd2 atoms may r eflect the repulsive forces exerted on
the inhibitors.
Other factors of importance for resistance induced
by the K103N mutation
In conclusion, our results indicate that the K103N mutation
leads to changes in hydrophobic and electrostatic interac-
tions. Moreover, the significance of these changes on
binding, for the individual compounds, is i n agreement
Ó FEBS 2002 The HIV-1 RT K103N mutant and inhibitor efficacy (Eur. J. Biochem. 269) 1675
with the r anking of the compounds with respect to their
inhibitory efficacy. However, these factors may not solely
account for t he total reduction in inhibitory efficacy caused
by the K103N mutation. An additional factor was presented
by Hsiou et al. were they showed, in a study of unliganded
RT, that the K103N mutation led to the formation of a
network of h ydrogen bonds that was not present in the wild-
type enzyme [17]. In particular the hydrogen bond between

N103 and Y 188 was suggested to s tabilize the closed form of
the NNIBP. Hsiou et al . suggested that this stabilization of
the closed conformation of the RT structure could interfere
with NNRTI binding by imposing an energy barrier for
NNRTI entrance, consistent with kinetic data. Our results
are complementary to those of Hsiou et al. and support
their p roposal that individual differences in efficacy between
related NNRTIs can arise from d ifferential interactions
between the inhibitors and the N103 side chain [8,36].
The compounds Efavirenz, MSC194 and PNU142721
have one property in common, namely that they contain a
hydrogen-bond donor in wing2. This property is of general
importance for the efficacy of the inhibitor. Substitution of
the hydrogen-donating amide group for a methylene group
completely abolishes the activity of a MSC194-related
compound (unpublished results). Whether this property is
of importance f or competition with the hydrogen-bond
network i n the K103N mutant remains to b e shown. An
interesting observation is, however, that the first generation
inhibitor Nevirapine lacks this property.
We have presented new insights in drug resistance that
could explain the reduced susceptibility of the K103N
mutant to NNRTIs. The mutation leads to changes in the
chemical environment of the NNIBP which affect the
interactions to NNRTIs. The implication of t hese changes
for NNRTI-binding is described as changes among two
properties influencing the inhibitory efficacy: hydrophobic
and electrostatic factors. The potent i nhibitor compounds
accommodate the K103N mutation by the formation of
new interactions to the N103 side chain and minor

rearrangements of the inhibitor position in the binding site.
These results should be useful for design of improved
NNRTIs to the K103N mutant.
ACKNOWLEDGEMENTS
This work was supported by the Swedish Medical Research Council
(MFR, K79-16X-09505-07A), the Swedish National Board for Indus-
trial a nd Technical Develop ment (NUTEK). We thank the staffs of
station 711 of the MAX synchrotron, L und, Sweden, and the beam l ine
BM14, ESRF, 6 rue Jules Horowitz, BP 220, F-38043 G renoble Cedex,
France, fo r their assistance. Terese Bergfors is addressed t hanks for
proofreading the manuscript.
REFERENCES
1. Palella, F.J., Delaney, K.M., Moorman, A.C., Loveless, M.O.,
Fuhrer, J., Satten, G.A., Aschman, D.J. & Holmberg, S.D. (1998)
Declining morbidity and mortality among patien ts with advanced
human immunodeficiency virus infection. HIV Outpatient Study
Investigators. N.Engl.J.Med.338, 853–860.
2. Esnouf,R.,Ren,J.,Ross,C.,Jones,Y.,Stammers,D.&Stuart,
D. (1995) Mechanism of inhibition of HIV-1 reverse transcriptase
by non-nucleoside inhibitors. Nat. Struct. Biol. 2, 303–308.
3. Hsiou, Y., Ding, J., Das, K., Clark, A.D. Jr, Hughes, S.H. &
Arnold, E. (1996) Structure of unliganded HIV-1 reverse tran-
scripta se at 2.7 A
˚
resolution: implications of conformational
changes for polymerization and i nhibition mechanisms. Structure
4, 853–860.
4. Schmit, J.C., Cogniaux, J ., Hermans, P., Van Vaeck, C., Sprecher,
S., V an Remoortel, B., Witvrouw, M., Balzarini, J., D esmyter, J.,
De Clercq, E. & Vandamme, A.M. (1996) Multiple drug resis-

tance to nucleoside analogues and nonnucleoside reverse trans-
criptase inhibitors in an efficiently replicating human
immunodeficienc y v irus type 1 patient strain. J. Infect. Dis. 174,
962–968.
5. Iversen, A.K., Shafer, R.W., W ehrly, K., Winters, M.A., Mullins,
J.I., Chesebro, B. & Merigan, T.C. ( 1996) Mul tidrug-resistant
human immunodefi ciency virus type 1 strains resulting from
combination antiretroviral therapy. J. Virol. 70, 1086–1090.
6. Bacheler,L.,Jeffrey,S.,Hanna,G.,D’Aquila,R.,Wallace,L.,
Logue, K ., C ordova, B., Hertogs, K., Larder, B., B uckery, R.,
Baker, D., Gallagher, K., Scarnati, H., Tritch, R. & Rizzo, C.
(2001) Genotypic correlates of phenotypic resistance to efavirenz
in virus isolates from patients failing nonnucleoside reverse tran-
scriptase inhibitor therapy. J. Virol. 75, 4999–5008.
7. Demeter, L.M., Shafer, R.W., M eehan, P.M., Holden-Wiltse, J.,
Fischl, M.A., Freimuth, W.W., Para, M.F. & Reichman, R.C.
(2000) Delavirdine susceptibilities and associated reverse tran-
scriptase mutations in human immunodeficiency virus type 1
isolates from patients in a phase I/II trial of delavirdine
monotherapy ( ACTG 260). Antim icrob. Agents C hemother. 44,
794–797.
8. Maga, G., Amacker, M., Ruel, N., Hubscher, U. & Spadari, S.
(1997) Resistance to nevirapine of HIV-1 r everse transcr iptase
mutants: loss of stabilizing interactions and thermodynamic or
steric barriers are induced by different single ami no a cid s ub-
stitutions. J. Mol. Biol. 274, 738–747.
9. Young, S.D., Britcher, S.F., Tran, L.O., Payne, L.S., Lumma,
W.C., Lyle, T.A., Huff, J.R., Anderson, P.S., Olsen, D.B., Carroll,
S.S. & Emini, E.A. (1995) L-743, 726 (DMP-266): a novel, highly
potent nonnucleoside inhibitor of the human immunodeficiency

virus type 1 reverse transcriptase. Antimicrob. Agents Chemother.
39, 2602–2605.
10. Balzarini, J., Brouwer, W.G., Felauer, E.E., D e C lercq, E.
& Karlsson, A. (1995) Activity of various thiocarboxanilide
derivatives against wild-type and several mutant human
immunodeficienc y v irus type 1 strains. Antiviral Res. 27, 2 19–236.
11. Hogberg, M., Sahlberg, C., Engelhardt, P., N oreen, R., Kan-
gasmetsa, J., Johansson, N.G., Oberg, B., Vrang, L., Zhang, H.,
Sahlberg, B.L., Unge, T., Lovgren , S., Fridborg, K. & Backbro,
K. (1999) Urea-PETT compounds as a n ew class of HIV-1 rev erse
transcriptase inhibitors: Synt hesis and further s tructure-act ivity
relationship studies o f PETT analogues. J. Med. Chem. 42,
4150–4160.
12. Ren, J., Nichols, C., Bird, L.E., Fujiwara, T., Sugimoto, H.,
Stuart, D.I. & Stammers, D.K. (2000) B inding o f t he sec ond
generation non-nucleoside inhibitor S-1153 to HIV-1 reverse
transcriptase involves e xtensive main c hain hydrogen b onding.
J. Biol. Chem. 275, 14316–14320.
13. Miller,V.,Sturmer,M.,Staszewski,S.,Groschel,B.,Hertogs,K.,
de Bethune, M.P., P auwels, R ., Harrigan, P.R., Bloor, S., Kemp,
S.D. & Larder, B.A. (1998) The M184V mutation in HIV-1
reverse transcriptase (RT) conferring lamivudine resistance does
not result in broad cross-resistance to nuc leosid e analogue RT
inhibito rs. AIDS 12 , 705–712.
14. Torti, C., Pozniak, A., Nelson, M., Hertogs, K. & G azzard, B.G.
(2001) Distribution of K103N and/or Y181C HIV-1 m utations by
exposure to zidovudine and non- nucleoside reverse transcriptase
inhibito rs. J. Antimicrob. Chemother. 48, 113–116.
15. Conway, B ., Wainberg, M.A., Hall, D., Harris, M., R eiss, P.,
Cooper,D.,Vella,S.,Curry,R.,Robinson,P.,Lange,J.M.&

Montaner, J.S. ( 2001) Development of drug resistance in patients
1676 J. Lindberg et al. (Eur. J. Biochem. 269) Ó FEBS 2002
receiving c ombination s of zidovudine, didanosine and nevirapine.
AIDS 15, 1269–1274.
16. Ding, J., Das, K., Moereels, H., Koymans, L., Andries, K.,
Janssen, P.A., Hughes, S.H. & Arnold, E. (1995) Structure of
HIV-1 RT/TIBO R. 86183 complex reveals similarity in the
binding of diverse no nnucleoside inh ibitors. Nat. Struct. Biol. 2,
407–415.
17. Hsiou,Y.,Ding,J.,Das,K.,Clark,A.D.Jr,Boyer,P.L.,Lewi,P.,
Janssen, P.A., K leim, J .P., Rosner, M., Hughes, S.H. & Arnold, E.
(2001) The lys103asn mutation o f HIV-1 rt: a novel mechanism of
drug resistance. J. Mol. Biol. 309, 437–445.
18. Dykes, C., Fox, K., Lloyd, A., Chiulli, M., Morse, E. & Demeter,
L.M. (2001) Impact of clinical reverse transcriptase sequences on
the replication capacity of HIV-1 drug-resistant mutants. Virology
285, 193–203.
19. Bhikhabhai, R., Carlsson, T., Unge, T., Lo
¨
vgren,S.&Strandberg,
B. (1993) Increased yield of homogeneous H IV-1 reverse t ran-
scriptase (p66/p51) using a slow purification approach . J. Chro-
matogr. 639, 67–74.
20. Zhang, H., Vrang, L., Unge, T. & Oberg, B. (1993) Character-
ization of HIV-1 reverse t ranscriptase with Y181C and L100I
mutations. Antiviral Chem. Chemother. 4, 301–308.
1
21. Otwinowski, Z. & M inor, W. (1997) X-ray diffraction data col-
lected in oscillation mode. Methods Enzymol. 276, 307–326.
22. nr4, C.C. & P. (1994 ) T he CCp4 s uite: p rograms for protein

crystallography. Acta Crystallogr. D50, 760–763.
23. Brunger, A.T., Adams, P.D., Clore, G.M., Delano, W.L., Gros,
P., Grosse, K.R., Jiang, J.S., Kuszewski, J., Simonson, T. &
Warren, G.L. (19 98) Crystallography & NMR system: a new
software suit for macromolecular structure determination. Acta
Crystallogr. D54, 905–921.
24. Kleywegt, G.J. & Jones, T.A. (1998) Databases in protein crys-
tallography. Acta Crystallogr. D54, 1119–1131.
25. Kleywegt, G.J. & Brunger, A.T. (1996) Checking your imagina-
tion: applications of the free R. value. Structure 4, 897–904.
26. Kleywegt, G.J. & Jones, T.A. (1996) xdlMAPMAN and xdlDA-
TAMAN – programs for r eformatting, a nalysis and manipulation
of biomacromolecular electron-density maps and reflection data
sets. Acta Crystallogr. D52, 826–828.
27. Kleywegt, G.J. (1996) Use of non-crystallographic symmetry in
protein structure refinement. Acta Crystallogr. D52, 842–857.
28. Jones, T.A., Zou, J Y., Cowan, S.W. & K jeldgaard, M. (1991)
Improved methods f or building protein models in electron density
maps and the location of errors in these models. Acta Crystallogr.
A47, 110–119.
29. Guex, N. & Peitsch, M.C. (1999) SWISS-MODEL and the Swiss-
PdbViewer: An environment for comparat ive protein modeling.
Electrophoresis 18, 2714–2723.
30. Persistence of Vision (1999) The Ray Tracer POV-Ray 3.1,
available from
31. Kohlstaedt, L.A., W ang, J ., Fried man, J.M., R ice, P.A. & Steitz,
T.A. (1992) Crystal structure at 3.5 A
˚
resolution of HIV-1 reverse
transcriptase complexed with an inhibitor. Science 256, 1783–

1790.
32. Hsiou,Y.,Das,K.,Ding,J.,Clark,A.D.Jr,Kleim,J.P.,Rosner,
M., Winkler, I., Riess, G., Hughes, S.H. & Arnold, E. (1998)
Structures of Tyr188Leu m utant and wild-type HIV-1 r everse
transcriptase complexed with th e non-nucleoside inhibitor HBY
097: inhibitor fle xibility is a u seful de sign feat ure for redu cing drug
resistance. J. Mol. Biol. 284, 313–323.
33. Das, K., Ding, J., Hsiou, Y., Clark, A.D., Moereels, H., Koy-
mans,L.,Andries,K.,Pauwels,R.,Janssen,P.A.,Boyer,P.L.,
Clark, P., Smith, R .H., Kroeger S mith, M.B., Michejda, C.J.,
Hughes, S.H. & Arnold, E. (1996) Crystal structures of 8-Cl and
9-Cl TIBO complexed with wild-type HIV-1 RT and 8-Cl TIBO
complexed with the Tyr181Cys HIV-1 RT drug-resistant mutant.
J. Mol. Biol. 264, 1085–1100.
34. Ren, J., Milton, J., Weaver, K.L., Sh ort, S.A., Stuart, D .I. &
Stammers, D.K. (2000) Structural basis for the resilience of efa-
virenz (DMP-266) to drug resistance mutations in HIV-1 reverse
transcriptase. Struct. Fold. Des. 8, 1089–1094.
35. Ding, J., Hughes, S.H. & Arnold, E. (1997) Protein–nucleic
acid interactions and DNA conformatio n in a complex of
human immunodeficiency virus type 1 reverse transcriptase with
a double-stranded DNA template-primer. Biopolymers 44, 125–
138.
36. Miller, V., de Bethune, M.P., Kober, A., Sturmer, M., Hertogs,
K., Pauwels, R., Stoffels, P. & Staszewski, S. (1998) Patterns of
resistance and cross-resistance to human immunodeficiency virus
type 1 reverse transcriptase inhibitors in patients treated with the
nonnucleoside reverse tran scriptase inhibitor lo viride. Antimicrob.
Agents Chemother. 42, 3123–3129.
Ó FEBS 2002 The HIV-1 RT K103N mutant and inhibitor efficacy (Eur. J. Biochem. 269) 1677