Báo cáo khoa học: Structural insights into mechanisms of non-nucleoside drug resistance for HIV-1 reverse transcriptases mutated at codons 101 or 138

Structural insights into mechanisms of non-nucleoside drug resistance for HIV-1 reverse transcriptases mutated at codons 101 or 138 Jingshan Ren1, Charles E. Nichols1, Anna Stamp1, Phillip P. Chamberlain1, Robert Ferris2, Kurt L. Weaver2, Steven A. Short2 and David K. Stammers1

1 Division of Structural Biology, The Wellcome Trust Centre for Human Genetics, Henry Wellcome Building for Genomic Medicine,

University of Oxford, UK

2 Glaxo Smith Kline Inc., Research Triangle Park, NC, USA

Keywords drug resistance; HIV-1 reverse transcriptase mutants; Lys101Glu, Glu138Lys; non- nucleoside inhibitors; X-ray crystallography

Correspondence D.K. Stammers, Division of Structural Biology, The Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford OX3 7BN, UK Fax: +44 1865 287 547 Tel: +44 1865 287 565 E-mail: daves@strubi.ox.ac.uk

(Received 23 March 2006, revised 20 June 2006, accepted 22 June 2006)

doi:10.1111/j.1742-4658.2006.05392.x

Lys101Glu is a drug resistance mutation in reverse transcriptase clinically observed in HIV-1 from infected patients treated with the non-nucleoside inhibitor (NNRTI) drugs nevirapine and efavirenz. In contrast to many NNRTI resistance mutations, Lys101(p66 subunit) is positioned at the sur- face of the NNRTI pocket where it interacts across the reverse transcrip- tase (RT) subunit interface with Glu138(p51 subunit). However, nevirapine contacts Lys101 and Glu138 only indirectly, via water molecules, thus the structural basis of drug resistance induced by Lys101Glu is unclear. We have determined crystal structures of RT(Glu138Lys) and RT(Lys101Glu) in complexes with nevirapine to 2.5 A˚ , allowing the determination of water structure within the NNRTI-binding pocket, essential for an understanding of nevirapine binding. Both RT(Glu138Lys) and RT(Lys101Glu) have remarkably similar protein conformations to wild-type RT, except for sig- nificant movement of the mutated side-chains away from the NNRTI pocket induced by charge inversion. There are also small shifts in the posi- tion of nevirapine for both mutant structures which may influence ring stacking interactions with Tyr181. However, the reduction in hydrogen bonds in the drug-water-side-chain network resulting from the mutated side-chain movement appears to be the most significant contribution to nevirapine resistance for RT(Lys101Glu). The movement of Glu101 away from the NNRTI pocket can also explain the resistance of RT(Lys101Glu) to efavirenz but in this case is due to a loss of side-chain contacts with the drug. RT(Lys101Glu) is thus a distinctive NNRTI resistance mutant in that it can give rise to both direct and indirect mechanisms of drug resist- ance, which are inhibitor-dependent.

The emergence of resistant viruses resulting from drug treatment of HIV-infected patients poses one of the most significant problems in countering AIDS in West- ern countries [1]. Two virus specific proteins, reverse

transcriptase (RT) and protease, have been the main targets for the development of anti-HIV drugs used in multidrug combination therapy regimens [2]. HIV-1 RT contains two distinct inhibitor binding sites for

Abbreviations NRTI, nucleoside analogue inhibitors of RT; NNRTI, non-nucleoside reverse transcriptase inhibitor; NtRTI, nucleotide analogue; PETT, phenethylthiazolylthiourea; RT, reverse transcriptase; TSAO, (2¢,5¢-bis-O-(tert-butyldimethylsilyl)-b-d-ribofuranosyl]-3¢-spiro-5¢¢-(4¢¢-amino- 1¢¢,2¢¢-oxathiole-2¢¢,2¢¢-dioxide).

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3850

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

N

O

N

N 10

N 1

S

F

N

N

H

N H

Cl

O

N H

PETT-2

Nevirapine

Scheme 1. Structures of nevirapine and PETT-2 are shown.

charge, e.g. Lys101Glu or Glu138Lys ⁄ Arg. Pyridinone NNRTIs can select for the Lys101Glu mutation in HIV-1 RT, giving approximately eight-fold resistance to compounds in this series [14]. The Lys101Glu mutation in RT also gives eight-fold resistance to nevirapine, based on the mean of four published values [15–18]. Significantly, Lys101Glu in RT is observed in the clinic for drug regimens which include nevirapine [19,20] or efavirenz [21] as the NNRTI component. The Lys101Glu mutation in RT is also selected by passaging HIV in tissue culture in the presence of carboxanilide NNRTIs such as UC-38 [22], giving greater than 100-fold resistance [23]. In the crystal structure with wild-type HIV-1 RT, UC-38 has a van der Waals contact with the side-chain of Lys101 [24]. Consistent with the reports of the selec- tion of Lys101Glu in clinical use of efavirenz, the crystal structure of RT with this NNRTI shows a contact of the inhibitor with the CD atom of Lys101 [25]. In the case of nevirapine, however, there are only indirect contacts of Lys101 in the complex with wild-type RT, via a series of three water molecules that link the drug, the side-chain of Glu138 and the main-chain carbonyl of Lys101 [4]. The molecular mechanism of drug resistance for nevirapine induced by Lys101Glu mutant is thus unclear, although an indirect mechanism is present.

including

different classes of drugs within the larger subunit of the p66 ⁄ p51 heterodimer [3,4]. Nucleoside analogue inhibitors of RT (NRTIs), such as zidovudine, lamivu- dine, stavudine and the nucleotide analogue (NtRTI) tenofovir, bind, respectively, as activated tri- or di- phosphate forms at the RT polymerase active site. NRTIs and NtRTIs, as well as competing with cognate nucleotide substrates, are also incorporated into the primer strand causing termination of the growing DNA chain [5]. Non-nucleoside inhibitors (NNRTIs) form a second class of compounds that target HIV-1 RT [6]. NNRTIs bind at an allosteric site about 10 A˚ from the polymerase active site. The structural mech- anism of NNRTI inhibition involves distortion of the catalytically essential triad of aspartic acid residues [7,8]. The non-nucleoside site, although almost entirely contained within the p66 subunit, also has Glu138 from the p51 subunit located at the edge of the inhib- itor pocket [3,4]. First-generation NNRTI drugs such as nevirapine (Scheme 1) and delavirdine generally show significant reduction in potency in the presence of a wide range of single point mutations [9]. Second efavirenz, have generation compounds, greater resilience to many such mutations [10,11]. Whilst the majority of the mutation sites encoding resistance to NRTIs are distal to the dNTP site within RT [12,13], those which result in NNRTI drug resist- ance are generally proximal to the inhibitor binding pocket [3,4]. Many of

The RT(Glu138Lys) mutation in HIV is selected in tissue culture by inhibitors of the TSAO (2¢,5¢-bis- O-(tert-butyldimethylsilyl)-b-d-ribofuranosyl]-3¢-spiro- 5¢¢-(4¢¢-amino-1¢¢,2¢¢-oxathiole-2¢¢,2¢¢-dioxide) series [26]. TSAO compounds have been shown to be capable of disrupting the p66 ⁄ p51 heterodimer [27], indeed model- ling studies suggest that certain TSAO NNRTIs may bind at a novel site between the subunits of the RT heterodimer [28]. It is proposed that mutation of Glu138Lys causes loss of a key interaction of the gluta- mic acid carboxyl group to an amino group of TSAO inhibitors. Mutations in RT at codon 138 (to Lys ⁄ Arg) have been shown to give cross resistance to a number of NNRTIs including members of the PETT series [29] as well as to emivirine [17]. In contrast, RT(Glu138Lys) is reported as having minimal cross resistance to nevi- rapine [18,30]. PETT-2 (Scheme 1) forms a van der Waals contact with the carboxyl group of Glu138 in the complex with wild-type RT [31]. crystal

determinations

structure

date,

To

the most commonly observed NNRTI resistance mutations are located at hydrophobic resi- dues at the centre of the drug binding pocket (e.g. Leu100, Val106, Val108, Tyr181 and Tyr188). How- ever, charged residues positioned towards the surface of the drug pocket are also known sites of drug resistance mutations. Indeed, Lys103Asn is the muta- tion most frequently observed in clinical use of NNRTIs. Other NNRTI resistance mutations at sur- face residues in RT include Lys101 and Glu138 (the latter residue from the p51 subunit). Mutations at residues 101 and 138 that give resistance to NNRTIs usually involve a change in the sign of the side-chain

of NNRTI resistant HIV-1 RTs have mainly focussed on mutations at positions 103, 181 and 188 [25,32– 36]. More recently structures of RT with drug resist- ance mutations Leu100Ile, Val106Ala and Val108Ile have been reported, which indicate an indirect com- ponent in the mechanism of resistance via modulation

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3851

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

Results

Overall RT structure

the X-ray data

and RT(Lys101Glu),

for

gave

crystals

suitable

of interactions with the key aromatic side-chains of Tyr181 and Tyr188 [37]. For RT(Lys103Asn) it has been shown that the second generation NNRTI efavi- renz can bind to the enzyme in both a wild-type con- formation as well as with Tyr181 in a rearranged position [25,36]. The side-chain of Asn103 can form a hydrogen bond to the Tyr188 hydroxyl thereby stabil- izing the unliganded RT structure [7,35]. In work reported here, crystal structures of HIV-1 RTs either containing the mutation Lys101Glu or Glu138Lys in complexes with nevirapine and PETT-2 have been determined in order to investigate structural effects of such mutants on the drug binding pocket and how these might relate to resistance mechanisms. RT(Glu138Lys) and RT(Lys101Glu) mutants are logi- cal to study as a pair as they are both charged sur- face residues interacting across the p66 ⁄ p51 interface. The availability of a range of structures of NNRTI- resistant RTs drugs will provide greater understanding of the molecular basis of drug resistance mechanisms. Such data will contribute to the design of novel inhibi- tors, which are still needed for containing the effects of HIV infection.

collection and the Details of refinement statistics are shown in Table 1. Of the wide range of NNRTIs tested for co-crystallization three with RT(Glu138Lys) structure complexes determination: RT(Lys101Glu)-nevirapine (to 2.5-A˚ (2.5 A˚ ) and resolution), RT(Glu138Lys)-nevirapine RT(Glu138Lys)-PETT2 (3.0 A˚ ). The relatively high the mutant RT-nevirapine structures resolution of allowed details of the water structure associated with the protein relevant to NNRTI binding to be deter- mined. Water molecules are important for the interac- tions of nevirapine with RT, particularly via Lys101 and Glu138, as there are no direct contacts of the inhibitor with these side-chains. Fo-Fc omit maps show clear electron density for the bound NNRTI, some water molecules, as well as for the mutated side- chains (Fig. 1) in each case. Side-chain electron density relevant to NNRTI drug resistance is also shown in

Table 1. Statistics for crystallographic structure determinations.

K101E-nevirapine PF BL-6 A Fuji BAS III 1.000 138.9, 114.9, 65.6 30.0–2.5 149821 35057 94.4 9.8 0.100

E138K-nevirapine APS SBC-2 CCD SBC 0.82657 139.6, 115.0, 65.6 30.0–2.5 251696 37104 99.5 12.6 0.089

E138K-PETT-2 SRS PX14.2 CCD ADSC-Q4 0.979 138.8, 114.8, 65.9 30.0–3.0 57571 19095 88.0 11.8 0.083

2.59–2.50 3371 92.5 1.0

2.59–2.50 3627 99.5 1.1

3.11–3.00 1537 72.6 1.4

Data collection details Data set Data collection site Detector Wavelength (A˚ ) Unit cell dimensions (a,b,c in A˚ )a Resolution range (A˚ ) Observations Unique reflections Completeness (%) Average I ⁄ r(I) b Rmerge Outer resolution shell Resolution range (A˚ ) Unique reflections Completeness (%) Average I ⁄ r(I) Refinement statistics: Resolution range (A˚ ) Number of reflections(working ⁄ test) R-factorc (Rwork ⁄ Rfree) R-factorc (all data) Number of atoms (protein ⁄ inhibitor ⁄ water) Rms bond length deviation (A˚ ) Rms bond angle deviation ((cid:2)) Mean B-factor (A˚ 2)d Rms backbone B-factor deviation (A˚ 2)

30.0–2.5 33264 ⁄ 1740 0.198 ⁄ 0.270 0.194 7570 ⁄ 20 ⁄ 80 0.008 1.4 59 ⁄ 65 ⁄ 37 ⁄ 44 4.4

30.0–2.5 35199 ⁄ 1856 0.216 ⁄ 0.288 0.203 7614 ⁄ 20 ⁄ 79 0.008 1.4 62 ⁄ 68 ⁄ 39 ⁄ 48 4.2

20.0–3.0 18153 ⁄ 929 0.226 ⁄ 0.277 0.226 7554 ⁄ 23 ⁄ – 0.009 1.5 73 ⁄ 81 ⁄ 60 ⁄ – 3.4

aAll crystals belong to space group P212121 [40,41]; bRmerge ¼ S|I – < I > | ⁄ S < I > ; cR factor ¼ S|Fo- Fc| ⁄ SFo; dmean B factor for main-chain, side-chain, inhibitor and water molecules.

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3852

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

the NNRTI pocket

mic acid side-chain is well defined (Fig. 1A). The struc- ture of is overall remarkably similar for mutant and wild-type, with the exception of the mutated side-chain and adjacent interacting resi- due. There is a significant movement of Glu101 relat- ive to the position of the wild-type Lys101, such that the side-chain carboxyl group points away from the NNRTI site (distance of the carboxyl group to the equivalent wild-type lysine side-chain amino group of 3.4 A˚ ) (Fig. 2A). There is also a smaller displacement of the Glu138 side-chain with a shift of (cid:2)0.5 A˚ relative to wild-type. The result of these changes is the loss of the salt bridge that links the protonated amino group of Lys101 in the p66 subunit to the negatively charged carboxyl of the Glu138 side-chain from the p51 sub- unit in the wild-type RT nevirapine complex [4]. Accompanying the side-chain movement related to the Lys101Glu mutation there is also a shift in the posi- tion of nevirapine, with an average atom displacement of 0.3 A˚ and a maximum movement of 0.5 A˚ . The side-chains of Tyr181 and Tyr188 are however, main- tained in a wild-type conformation. In spite of the observed rearrangements of inhibitor and the side- chains of residues 101 and 138, the three water mole- cules that link nevirapine to Glu138 in the wild-type complex are retained in the mutant structure albeit with shifts in position in the same direction as the nevirapine, i.e. away from the binding pocket.

Glu138Lys

(B) Glu138Lys in p51 and nevirapine.

Fig. 1. Simulated annealing omit electron density maps contoured at 3.5r showing bound inhibitors, waters and mutated residues within one of the HIV-1 RT subunits as indicated. (A) Lys101Glu in (C) p66 and nevirapine. Glu138Lys in p51 and PETT-2.

Fig. 1, i.e. for Lys101Glu in the p66 subunit (Fig. 1A) and Glu138Lys in the p51 subunit (Fig. 1B,C).

Comparison of wild-type and mutant RT structures

Lys101Glu

In the crystal structure of RT(Lys101Glu) in complex with nevirapine, electron density for the mutant gluta-

The structures of HIV-1 RT(Glu138Lys) in complexes with nevirapine and PETT-2 complex have well defined side-chain density for Lys138 (p51) (Fig. 1B,C). As with the Lys101Glu mutant, the overall structure of the NNRTI pocket in RT(Glu138Lys) closely matches that of the equivalent nevirapine complex with wild- type RT, with the exception of the mutated side-chain itself, which together with Lys101 undergoes large positional changes. Lys138 moves away from the NNRTI pocket in the complex with nevirapine as indi- cated by the shift in position of the CD atom of 2.8 A˚ (Fig. 2B), whilst Lys101 becomes much less ordered than in the wild-type structure. The best fit of the side- chain to electron density indicates a rotation of close to 150(cid:2) about the Lys101 CA–CB bond, positioning it away from the NNRTI pocket (Fig. 2B) and resulting in the loss of interaction of residue 138 in p51 with Lys101 in the p66 subunit. There is also a shift in the position of nevirapine outwards from the binding pocket with an average atom displacement of 0.4 A˚ and a maximum displacement of 0.6 A˚ . The three that bridge between Glu138 and water molecules

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3853

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

Fig. 2. Stereo-diagrams comparing the NNRTI binding sites of wild-type and mutant RTs for the following complexes: (A) Lys101Glu and nevirapine, (B) Glu138Lys and nevirapine, and (C) Glu138Lys and PETT-2. The thinner and thicker bonds show the CA backbone and side- chains with wild-type RT coloured orange and the mutant RTs coloured blue, respectively. Inhibitors and water molecules are shown in ball- and-stick representation and coloured red for wild-type RT and cyan for mutant RTs. For clarity, the side-chains at the site of mutation are shown in magenta for wild-type and green for the mutants. Hydrogen bonds linking drug to protein and drug to water molecules are marked in dashed lines, coloured yellow for wild-type and blue for mutant RTs.

nevirapine in the wild-type complex remain in the Glu138Lys mutant but are displaced in position and the network of hydrogen bonds is altered. The remain-

ing side-chains within the mutant NNRTI binding site are essentially positioned the same as for the wild-type enzyme.

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3854

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

Comparison of RT wild-type and RT(Glu138Lys) complexes with PETT-2 shows that there appears to be a perturbation in the positions of a number of side- chains, with the largest changes at the mutation site. The wild-type Glu138 and mutant Lys138 residues overlap as far as the CB atoms, the remainder of the side chain of Lys138 is swung away from the NNRTI pocket by a rotation of (cid:2)170(cid:2) about the CA–CB bond giving a difference in position of amino and carboxyl groups of 6.3 A˚ (Fig. 2C). The difference in conforma- tion of Lys138 relative to Glu138 results in the loss of the van der Waals contact between the carboxyl and the pyridine nitrogen ring of PETT-2 in the wild-type complex. Additionally, there is a movement in the Lys101 side chain which involves a rotation about the CD–CE bond resulting in the amino group moving 3.4 A˚ relative to the wild-type structure, positioning it further away from the NNRTI pocket. The changes in side-chain conformations are also accompanied by a shift in the position of PETT-2 within the binding site with an average atom displacement of 0.4 A˚ and a maximum displacement of 0.7 A˚

(Fig. 2C).

Potency of PETT-2 against RT(Glu138Lys) in viral and enzyme assays

for wild-type

PETT-2

by

The results of EC50 determinations for PETT-2 in the HIV assay together with fold-resistance for the Glu138Lys mutant are shown in Table 2. IC50 values recombinant determined from inhibition assays of HIV-1 RT and RT(Glu138Lys) are also shown in Table 2. In both antiviral and enzyme assays, no significant difference in potency was observed for PETT2 between wild-type and the Glu138Lys mutant, hence no cross-resistance was detectable.

Discussion

The effect of many of the drug resistance mutations reported for NNRTIs can be rationalized at the molecular level in terms of direct alterations in the ste- reochemistry of inhibitor–RT interactions. Thus, in the

case of both RT(Tyr181Cys) and RT(Tyr188Cys), the loss of aromatic ring stacking with first generation NNRTIs, such as nevirapine, explains the observed drug resistance [32,34]. The mechanism whereby RT(Lys101Glu) gives resistance to nevirapine is more complex as no direct contacts between the drug and Lys101 exist, rather a network of three water mole- cules links the inhibitor to the side-chains of Lys101 and Glu138 in wild-type RT [4]. We were fortunate in being able to solve both RT(Glu138Lys) and RT(Lys101Glu) nevirapine complexes to the relatively high resolution for HIV-1 RT crystals of 2.5 A˚ , thereby allowing full structural refinement and deter- the associated water structure, of key mination of importance in understanding the interaction of the drug with these residues. There are some similarities in the structural changes seen in the nevirapine complexes for both the RT(Glu138Lys) and RT(Lys101Glu) mutants. Thus significant movements of side-chains away from the NNRTI pocket are observed in both structures due to the juxtaposition of like charges caused by the charge inversion resulting from these mutations. Although three water molecules, which link Glu138, Lys101 and nevirapine in wild-type RT are also present in both RT(Glu138Lys) and RT(Lys101- Glu) structures, they are shifted in position and hydro- gen bonding patterns are perturbed. A set of H-bonds linking waters to the main-chain carbonyl and amide groups of residue 101 to one of the nevirapine pyridine rings is retained in each case. However the side-chain movements away from the NNRTI pocket observed for RT(Lys101Glu) and RT(Glu138Lys) nevirapine complexes in both cases abolish the H-bond and salt bridge linking Lys101 to Glu138 in wild-type RT. Such movement leads to a reduction in the extent of the H-bond network so that there is no longer a water molecule H-bonded to residue 138 in the Glu138Lys mutant. Paradoxically the greater loss of H-bonding seen in RT(Glu138Lys) does not lead to a bigger reduction in binding potency for nevirapine, clearly there must be some compensatory indicating that changes. Detailed inspection of overlaps of the RT(Lys101Glu) and RT(Glu138Lys) structures with

Table 2. Potency of PETT-2 against HIV-1 RT(Glu138Lys).

EC50 virus assay (nM)

IC50 enzyme assay (nM)

WT

E138K

Fold resistance

WT

E138K

Fold resistance

PETT-2 Nevirapine

3.2 ± 2.0 81 ± 12

2.8 ± 0.5 97 ± 21

0.9 1.2

2.8 ± 0.3 1955 ± 376a

2.7 ± 0.4 1165 ± 375a

1.0 0.6

aData from reference [30].

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3855

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

the

crystal

in the position of nevirapine

that

is

that better meet both hydrophobic and hydrogen bonding requirements which might therefore have greater potency. As an example, it may be possible to introduce additional substituents into a nevirapine-like inhibitor in order to displace the bound waters, allow- ing direct interactions with the protein. Comparison of structures of wild-type RT and RT(Glu138Lys) with bound PETT-2 shows that there is a loss of a van der Waals contact between the carb- oxyl of Glu138 and the inhibitor. It would thus be expected that there is an accompanying reduction in binding affinity for PETT-2 for the Glu138Lys muta- tion. In fact the EC50 and IC50 data both from the mutant HIV in tissue culture and recombinant RT enzyme assays clearly indicate essentially equal potency for PETT-2 against WT and mutant forms. It is poss- ible that the observed PETT-2 contact with the Glu138 carboxyl may be induced as a result of the crystals being grown at pH 5.0, which approaches the range of pKa values for glutamic acid side-chains ((cid:2)4.4). Glu138 may therefore be protonated and as a conse- quence induce a contact which would not be found under conditions for assay of RT inhibition (pH 7.4). It should also be born in mind that since the resolution of the RT complexes with PETT-2 (both wild-type and for RT(Glu138Lys)) are limited to 3.0 A˚ , it is not possible to fully define all aspects of the structure, par- ticularly bound water molecules. It is thus not known whether, for example, changes in hydrogen bonding and water networks may occur as is the case in the work reported here on the RT(Lys101Glu) and RT(Glu138Lys) mutants in complexes with nevirapine. Such changes might compensate for the loss of the interactions between PETT-2 and van der Waals RT(Glu138Lys) compared with wild-type. This limita- tion in structural detail emphasizes the need to exercise caution in the interpretation of molecular details of inhibitor binding when using lower resolution X-ray structures of HIV-1 RT.

For

the Glu138Lys mutation,

resistance to the TSAO series of NNRTIs is thought to be due to the loss of interaction of the side-chain carboxyl group with a key amino group of the compounds [28]. Unfor- tunately it has not been possible to obtain crystal structures of TSAO bound to RT as inhibitors of this series cause the protein to precipitate prior to co- crystallization set-ups or induce disorder in pre-grown crystals following soaking attempts. Such properties may relate to the known disrupting effects of TSAO on the RT dimer structure.

The work described here shows how the charge inversion introduced by the Lys101Glu and Glu138Lys in a common structural mutations

in RT results

the equivalent nevirapine complex of wild-type RT indicates the NNRTI site protein conformation is remarkably similar in both mutants and wild-type RT, thus limiting possible options for compensatory chan- ges. Indeed, apart from movements of the mutated and immediately interacting side-chains and associated water molecules, the only other changes visible are shifts for both RT(Lys101Glu) and RT(Glu138Lys), with the biggest movement observed for the drug in the latter mutant. Previous structural work has shown that a shift of nevirapine relative to Tyr181 can compromise drug to side-chain aromatic ring stacking interactions, thereby contributing to weaker drug binding [37]. There are, however, distinctive features in the case of RT(Glu138Lys), in which the movement of nevirapine relative to the Tyr181 side-chain is by a translation in one direction rather than a rotation, thus potentially optimizing the favourable parallel offset ring stacking interactions with Tyr181. Normally, where changes of the interactions of nevirapine with the tyrosine side- chain leads to reduced binding, e.g. Val106Ala, there is a rotation of the nevirapine and significant movement of the side-chain [37]. Thus the nevirapine movement in RT(Glu138Lys) appears in this case to lead to an optimization of interactions rather than the reverse. A factor that is different in RT(Glu138Lys) compared with RT(Val106Ala) the former mutation slightly expands the NNRTI pocket in a way that is not possible for resistance mutations at the hydropho- bic core of the pocket which are more structurally con- strained. For RT(Lys101Glu), where the H-bonding network involving nevirapine is less disrupted (the link to Glu138 being maintained) there is a smaller shift in the nevirapine position and therefore suggesting there is less opportunity to optimize interactions, resulting in a greater overall loss of binding affinity. A corollary to this is that the interactions of nevirapine with wild- type RT would seem non-optimal. Indeed, there is evidence of this for the mutation Pro236Leu which, although it gives resistance to delavirdine [38], causes hypersensitivity to nevirapine. There thus appears to be an incompatibility between the competing require- ments of nevirapine hydrogen bonding via water mole- cules to side-chains at the outer edge of the pocket (Lys101, Glu138) and aromatic ring stacking to tyro- sines in the inner region. This would also explain why nevirapine, although a much bulkier molecule than ef- avirenz, and capable of making more interactions with RT, is actually of much lower potency ((cid:2)30-fold) than the latter drug. A further implication is that, as nevi- rapine does not appear to optimally fit the NNRTI pocket, there are possibilities for designing analogues

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3856

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

inhibitor

mutant RT inhibitor complexes used in this work were obtained by co-crystallization. X-ray data were either col- lected at SRS, Daresbury, UK and APS, Chicago, IL, USA, using the oscillation method or at the Photon Fac- tory Tsukuba, Japan, using a Weissenberg camera [42]. Data collected for each mutant–inhibitor complex were from crystals flash-cooled in liquid propane and maintained at 100 K in a nitrogen gas stream. Indexing and integration of data images were carried out with denzo, and data were merged with scalepack [43]. Details of the X-ray data sta- tistics are given in Table 1.

Structure solution and refinement

closeness of unit

The molecular orientation and position of HIV-1 RT het- erodimers in each unit cell were determined using rigid- body refinement with cns [44]. The initial model used as the starting point for the current structure determinations was chosen from our collection of RT–NNRTI complexes on the basis of cell parameters [4,24,25,34,45–48]. All structures were refined with cns [44] using positional, simulated annealing and individual B-fac- tor refinement with bulk solvent correction and anisotropic B-factor scaling. Model rebuilding was carried out using ‘o’ [49]. Table 1 gives the refinement statistics for the three structures.

the significant movement of mutated side- feature, chains away from the NNRTI pocket, due to the repulsion of like charges. The structural consequences of the mutant side-chain movement include partially disrupting the hydrogen bond network involving nevi- rapine, water and side-chain ⁄ main-chain interactions in wild-type RT. Compensatory changes must occur for the Glu138Lys mutation as this has little effect on nevirapine binding which we suggest to be optimiza- tion of ring stacking interactions with Tyr181. Lys101Glu provides an example of an indirect mechanism of inducing drug resistance in RT but via a different means to other reported examples viz Lys103Asn, which apparently stabilizes the unliganded RT conformation [7,35] whilst Val108Ile perturbs the interaction of Tyr188 and Tyr181 with nevirapine [37]. the Lys101Glu mutation in RT also Interestingly, appears to be capable of a direct mechanism of drug resistance. For the carboxanilide, UC-38, and for ef- avirenz there are van der Waals contacts between the side-chain of Lys101 and the inhibitors [24]. Movement the Lys101 side-chain away from the NNRTI of pocket on mutation to glutamic acid, as observed for the nevirapine complex, would thus cause loss of contacts and a consequent weakening in binding for UC-38 and efavirenz.

Structures of wild-type and mutant RTs were overlapped using the program SHP [50]. The wild-type RT complexes used for these comparisons were 1rtd (for nevirapine) [4] and 1dtt (for PETT-2) [31].

Data deposition

codes are as

The selection of drug-resistant virus remains a signi- ficant problem in controlling HIV infection and AIDS. The work reported here is part of ongoing studies aimed at understanding mechanisms of drug resistance in HIV RT by the use of X-ray crystal structure analy- sis. Combining such data with biochemical and virolo- gical studies will be of value in contributing to the development of further generations of NNRTIs with improved resilience to existing drug resistance muta- tions present in HIV. Such drugs are needed as an important priority for continued efforts to treat HIV.

Coordinates and structure factors for all of the HIV-1 RT mutant structures reported here have been deposited in the Protein Data Bank for immediate release on publication. The follows: RT(Lys101Glu)-nevirapine, 2HND; RT(Glu138Lys)-nevirapine, 2HNY; RT(Glu138Lys)- PETT-2, 2HNZ.

Experimental procedures

Potency of PETT-2 against HIV RT(Glu138Lys) in tissue culture and enzyme assays

Crystallization and data collection

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3857

The assay to determine EC50 values for PETT-2 used a modified HeLa MAGI system in which HIV-1 infection is detected due to the activation of the LTR-driven b-galac- tosidase reporter in HeLa CD4-LTR-b-gal cells as des- cribed previously [51]. Quantitation of the b-galactosidase was carried out by measurement of the activation of the chemoluminescent substrate, tropix, after 3 days infection. The value of the EC50 was the mean of three separate experiments. PETT-2 inhibition of recombinant wild-type HIV-1 RT and RT(Glu138Lys) were determined in an enzyme assay using rC-dG as template-primer and measur- Expression vectors for HIV-1 RT (HXB-2 isolate) contain- ing either Lys101Glu or Glu138Lys mutations were made using site-directed mutagenesis as previously reported [34]. Purification of mutant RTs from recombinant Escherichia coli was based on the ion-exchange procedures as described in an earlier report [39]. A wide range of NNRTIs were screened for crystallization with the mutant RTs, producing crystals of three complexes suitable for structure determin- ation. Crystals of complexes of RTs with NNRTIs were grown and soaked in 50% PEG 3400 prior to data collec- tion as described previously [40,41]. All the crystals of

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

ing the incorporation of radiolabelled dGTP [39]. Curve fit- ting of enzyme inhibition data and IC50 determination were carried out using origin (Microcal). highly potent nonnucleoside inhibitor of the human immunodeficiency virus type 1 reverse transcriptase. Antimicrob Agents Chemother 39, 2602–2605. 11 De Clercq E (2001) New developments in anti-HIV che- motherapy. Curr Medical Chem 8, 1543–1572.

Acknowledgements

12 Jacobo-Molina A, Ding JP, Nanni RG, Clark ADJ,

Lu X, Tantillo C, Williams RL, Kamer G, Ferris AL, Clark P et al. (1993) Crystal structure of human immu- nodeficiency virus type 1 reverse transcriptase com- plexed with double-stranded DNA at 3.0 A˚ resolution shows bent DNA. Proc Natl Acad Sci USA 90, 6320– 6324. 13 Huang H, Chopra R, Verdine GL & Harrison SC

We thank the staff of the following synchrotrons for their assistance in data collection: SRS, Daresbury, Japan and UK; The Photon Factory, Tsukuba, Advanced Photon Source, Chicago, IL, USA. Support- ing grants from the Medical Research Council, the Biotechnology and Biological Sciences Research Coun- cil and the EC (QLKT-2000–00291, QLKT2-CT-2002– 01311) to DKS are acknowledged.

(1998) Structure of a covalently trapped catalytic com- plex of HIV-1 reverse transcriptase: implications for drug resistance. Science 282, 1669–1675. 14 Byrnes VW, Sardana VV, Schleif WA, Condra JH,

References

1 Richman DD (2001) HIV chemotherapy. Nature 410, 995–1001. 2 Brettle RP, Wilson A, Povey S, Morris S, Morgan R,

Waterbury JA, Wolfgang JA, Long WJ, Schneider CL, Schlabach AJ, Wolanski BS et al. (1993) Comprehensive mutant enzyme and viral variant assessment of human immunodeficiency virus type 1 reverse transcriptase resistance to nonnucleoside inhibitors. Antimicrob Agents Chemother 37, 1576–1579.

Leen CL, Hutchinson S, Lewis S & Gore S (1998) Com- bination therapy for HIV: the effect on inpatient activ- ity, morbidity and mortality of a cohort of patients. Int J STD AIDS 9, 80–87. 3 Kohlstaedt LA, Wang J, Friedman JM, Rice PA &

Steitz TA (1992) Crystal structure at 3.5 A˚ resolution of HIV-1 reverse transcriptase complexed with an inhibi- tor. Science 256, 1783–1790. 15 Buckheit RW, Snow MJ, Fliakas-Boltz V, Kinjerski TL, Russell JD, Pallansch LA, Brouwer WG & Yang SS (1997) Highly potent oxathiin carboxanilide derivatives with efficacy against nonnucleoside reverse transcriptase inhibitor-resistant human immunodeficiency virus iso- lates. Antimicrob Agents Chemother 41, 831–837.

16 Petropoulos CJ, Parkin NT, Limoli KL, Lie YS, Wrin T, Huang W, Tian H, Smith D, Winslow GA, Capon DJ et al. (2000) A novel phenotypic drug susceptibility assay for human immunodeficiency virus type 1. Antimicrob Agents Chemother 44, 920–928.

4 Ren J, Esnouf R, Garman E, Somers D, Ross C, Kirby I, Keeling J, Darby G, Jones Y, Stuart D et al. (1995) High resolution structures of HIV-1 RT from four RT- inhibitor complexes. Nature Struct Biol 2, 293–302. 5 Goody RS, Muller B & Restle T (1991) Factors contri- buting to the inhibition of HIV reverse transcriptase by chain-terminating nucleotides in vivo. FEBS Lett 291, 1–5.

17 Pelemans H, Aertsen A, Van Laethem K, Vandamme AM, De Clercq E, Perez-Perez MJ, San-Felix A, Velazquez S, Camarasa MJ & Balzarini J (2001) Site- directed mutagenesis of human immunodeficiency virus type 1 reverse transcriptase at amino acid position 138. Virology 280, 97–106. 6 De Clercq E (1993) HIV-1-specific RT inhibitors: highly selective inhibitors of human immunodeficiency virus type 1 that are specifically targeted at the viral reverse transcriptase. Med Res Rev 13, 229–258.

7 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. Nature Struct Biol 2, 303–308.

18 Chan JH, Freeman GA, Tidwell JH, Romines KR, Schaller LT, Cowan JR, Gonzales SS, Lowell GS, Andrews CW 3rd, Reynolds DJ, et al. (2004) Novel benzophenones as non-nucleoside reverse transcriptase inhibitors of HIV-1. J Med Chem 47, 1175–1182. 19 Schweighardt B, Ortiz GM, Grant RM, Wellons M,

8 Spence RA, Kati WM, Anderson KS & Johnson KA (1995) Mechanism of inhibition of HIV-1 reverse transcriptase by nonnucleoside inhibitors. Science 267, 988–993. 9 Schinazi RF, Larder BA & Mellors JW (2000) Muta- Miralles GD, Kostrikis LG, Bartlett JA & Nixon DF (2002) Emergence of drug-resistant HIV-1 variants in patients undergoing structured treatment interruptions. AIDS 16, 2342–2344.

tions in retroviral genes associated with drug resistance. Int Antiviral News 8, 65–71.

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3858

20 Hanna GJ, Johnson VA, Kuritzkes DR, Richman DD, Brown AJ, Savara AV, Hazelwood JD & D’Aquila RT (2000) Patterns of resistance mutations selected by treat- ment of human immunodeficiency virus type 1 infection 10 Young SD, Britcher SF, Tran LO, Payne LS, Lumma WC, Lyle TA, Huff JR, Anderson PS, Olsen DB, Carroll SS et al. (1995) L-743, 726 (DMP-266): a novel,

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

derivatives trovirdine and MSC-127. Antiviral Res 28, 331–342. with zidovudine, didanosine, and nevirapine. J Infect Dis 181, 904–911.

30 Balzarini J, Kleim JP, Riess G, Camarasa MJ, De Clercq E & Karlsson A (1994) Sensitivity of (138 Glu?Lys) mutated human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) to HIV-1-specific RT inhibitors. Biochem Biophys Res Commun 201, 1305–1312. 21 Bacheler LT, Anton ED, Kudish P, Baker D, Bunville J, Krakowski K, Bolling L, Aujay M, Wang XV, Ellis D et al. (2000) Human immunodeficiency virus type 1 mutations selected in patients failing efavirenz combina- tion therapy. Antimicrob Agents Chemother 44, 2475– 2484. 31 Ren J, Diprose J, Warren J, Esnouf RM, Bird LE, 22 Balzarini J, Brouwer WG, Felauer EE, De Clercq E &

Karlsson A (1995) Activity of various thiocarboxanilide derivatives against wild-type and several mutant human immunodeficiency virus type 1 strains. Antiviral Res 27, 219–236. Ikemizu S, Slater M, Milton J, Balzarini J, Stuart DI et al. (2000) Phenylethylthiazolylthiourea (PETT) non- nucleoside inhibitors of HIV-1 and HIV-2 reverse tran- scriptases: structural and biochemical analyses. J Biol Chem 275, 5633–5639. 32 Das K, Ding J, Hsiou Y, Clark AD, Moereels H,

Koymans L, Andries K, Pauwels R, Janssen PA, Boyer PL et al. (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. 23 Buckheit RW, Kinjerski TL, Fliakas-Boltz V, Russell JD, Stup TL, Pallansch LA, Brouwer WG, Dao DC, Harrison WA, Schultz RJ et al. (1995) Structure-activity and cross-resistance evaluations of a series of human immunodeficiency virus type-1-specific compounds rela- ted to oxathiin carboxanilide. Antimicrob Agents Chemo- ther 39, 2718–2727.

24 Ren J, Esnouf RM, Hopkins AL, Warren J, Balzarini J, Stuart DI & Stammers DK (1998) Crystal structures of HIV-1 reverse transcriptase in complex with carboxani- lide derivatives. Biochemistry 37, 14394–14403. 25 Ren J, Milton J, Weaver KL, Short SA, Stuart DI & 33 Hsiou Y, Das K, Ding J, Clark AD, Kleim J-P, Rosner M, Winkler I, Riess G, Hughes SH & Arnold E (1998) HIV-1 reverse transcriptase complexed with the non- nucleoside inhibitor HBY 097: inhibitor flexibility is a useful design feature for reducing drug resistance. J Mol Biol 284, 313–323.

Stammers DK (2000) Structural basis for the resilience of efavirenz (DMP-266) to drug resistance mutations in HIV-1 reverse transcriptase. Structure Fold Des 8, 1089– 1094. 26 Balzarini J, Velazquez S, San-Felix A, Karlsson A,

34 Ren J, Nichols C, Bird L, Chamberlain P, Weaver K, Short S, Stuart DI & Stammers DK (2001) Structural mechanisms of drug resistance for mutations at codons 181 and 188 in HIV-1 reverse transcriptase and the improved resilience of second generation non-nucleoside inhibitors. J Mol Biol 312, 795–805.

35 Hsiou Y, Ding J, Das K, Clark AD Jr, Boyer PL, Lewi P, Janssen PA, Kleim JP, Rosner M, Hughes SH et al. (2001) The Lys103Asn mutation of HIV-1 RT: a novel mechanism of drug resistance. J Mol Biol 309, 437–445.

Perez-Perez MJ, Camarasa MJ & De Clercq E (1993) Human immunodeficiency virus type 1-specific [2¢,5¢-bis- O-(tert- butyldimethylsilyl)-beta-d-ribofuranosyl]-3¢- spiro-5¢-(4¢-amino-1¢,2¢-oxathiole-2¢,2¢-dioxide)-purine analogues show a resistance spectrum that is different from that of the human immunodeficiency virus type 1- specific non-nucleoside analogues. Mol Pharmacol 43, 109–114.

36 Lindberg J, Sigurosson S, Lowgren S, Andersson HO, Sahlberg C, Noreen R, Fridborg K, Zhang H & Unge T (2002) Structural basis for the inhibitory efficacy of efavirenz (DMP-266), MSC194 and PNU142721 towards the HIV-1 RT K103N mutant. Eur J Biochem 269, 1670–1677. 27 Sluis-Cremer N, Arion D & Parniak MA (2002) Desta- bilization of the HIV-1 reverse transcriptase dimer upon interaction with N-acyl hydrazone inhibitors. Mol Pharmacol 62, 398–405.

37 Ren J, Nichols CE, Chamberlain PP, Weaver KL, Short SA & Stammers DK (2004) Crystal structures of HIV-1 reverse transcriptase mutated at codons 100, 106 and 108 and mechanisms of resistance to non-nucleoside inhibitors. J Mol Biol 336, 569–578.

28 Rodriguez-Barrios F, Perez C, Lobaton E, Velazquez S, Chamorro C, San-Felix A, Perez-Perez MJ, Camarasa MJ, Pelemans H, Balzarini J et al. (2001) Identification of a putative binding site for [2¢,5¢-bis-O-(tert-butyldi- methylsilyl)-beta-d-ribofuranosyl]-3¢-spiro-5¢¢-(4¢¢-amino- 1¢¢,2¢¢-oxathiole-2¢¢,2¢¢-dioxide) thymine (TSAO) derivatives at the p51–p55 interface of HIV-1 reverse transcriptase. J Med Chem 44, 1853–1865. 29 Zhang H, Vrang L, Backbro K, Lind P, Sahlberg C,

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3859

Unge T & Oberg B (1995) Inhibition of human immu- nodeficiency virus type 1 wild-type and mutant reverse transcriptases by the phenyl ethyl thiazolyl thiourea 38 Dueweke TJ, Pushkarskaya T, Poppe SM, Swaney SM, Zhao JQ, Chen ISY, Stevenson M & Tarpley WG (1993) A mutation in reverse transcriptase of bis(heter- oaryl)piperazine-resistant human immunodeficiency virus type 1 that confers increased sensitivity to other nonnucleoside inhibitors. Proc Natl Acad Sci USA 90, 4713–4717.

J. Ren et al.

HIV-1 RT drug resistance mutations at 101 and 138

46 Ren J, Esnouf RM, Hopkins AL, Stuart DI &

39 Wilson JE, Wright LL, Martin JL, Haire SE, Ray PH, Painter GR & Furman PA (1996) Recombinant human immunodeficiency virus type 1 reverse transcriptase is heterogeneous. J Acquir Immune Defic Syndr Hum Retrovirol 11, 20–30. Stammers DK (1999) Crystallographic analysis of the binding modes of thiozolo-isoindolinone non-nucleo- side inhibitors to HIV-1 reverse transcriptase and comparison with modelling studies. J Med Chem 42, 3845–3851.

40 Stammers DK, Somers DON, Ross CK, Kirby I, Ray PH, Wilson JE, Norman M, Ren JS, Esnouf RM, Garman EF et al. (1994) Crystals of HIV-1 reverse transcriptase diffracting to 2.2 A˚ resolution. J Mol Biol 242, 586–588. 41 Esnouf RM, Ren J, Garman EF, Somers DON, Ross

47 Hopkins AL, Ren J, Esnouf RM, Willcox BE, Jones EY, Ross C, Miyasaka T, Walker RT, Tanaka H, Stammers DK et al. (1996) Complexes of HIV-1 reverse transcriptase with inhibitors of the HEPT series reveal conformational changes relevant to the design of potent non-nucleoside inhibitors. J Med Chem 39, 1589–1600.

CK, Jones EY, Stammers DK & Stuart DI (1998) Con- tinuous and discontinuous changes in the unit cell of HIV-1 reverse transcriptase crystals on dehydration. Acta Crystallogr D54, 938–954. 42 Sakabe N (1991) X-ray diffraction data collection sys-

48 Hopkins AL, Ren J, Tanaka H, Baba M, Okamato M, Stuart DI & Stammers DK (1999) Design of MKC-442 (Emivirine) analogues with improved activity against drug resistant HIV mutants. J Med Chem 42, 4500– 4505.

tem for modern protein crystallography with a Weissen- berg camera and an imaging plate using synchrotron radiation. Nucl Instr Methods Phys Res 303, 448–463. 43 Otwinowski Z & Minor W (1996) Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol 276, 307–326. 49 Jones TA, Zou JY, Cowan SW & Kjeldgaard M (1991) Improved methods for building protein models in elec- tron density maps and the location of errors in these models. Acta Crystallogr A47, 110–119. 50 Stuart DI, Levine M, Muirhead H & Stammers DK

(1979) The crystal structure of cat pyruvate kinase at a resolution of 2.6 A˚ . J Mol Biol 134, 109–142. 51 Chan JH, Hong JS, Hunter RN 3rd, Orr GF, Cowan

44 Brunger AT, Adams PD, Clore GM, Delano WL, Gros P, Grosse KRW, Jiang JS, Kuszewski J, Nilges M, Pannu NS et al. (1998) Crystallography and NMR sys- tem: a new software suite for macromolecular structure determination. Acta Crystallogr D54, 905–921. 45 Ren J, Esnouf R, Hopkins A, Ross C, Jones Y,

FEBS Journal 273 (2006) 3850–3860 ª 2006 The Authors Journal compilation ª 2006 FEBS

3860

JR, Sherman DB, Sparks SM, Reitter BE, Andrews CW 3rd, Hazen RJ et al. (2001) 2-Amino-6-aryl- sulfonylbenzonitriles as non-nucleoside reverse tran- scriptase inhibitors of HIV-1. J Med Chem 44, 1866– 1882. Stammers D & Stuart D (1995) The structure of HIV-1 reverse transcriptase complexed with 9-chloro-TIBO: lessons for inhibitor design. Structure 3, 915–926.