This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ceccherini-Silberstein, F.
Right arrow Articles by Perno, C. F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ceccherini-Silberstein, F.
Right arrow Articles by Perno, C. F.

 Previous Article  |  Next Article 

Journal of Virology, October 2007, p. 11507-11519, Vol. 81, No. 20
0022-538X/07/$08.00+0     doi:10.1128/JVI.00303-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Characterization and Structural Analysis of Novel Mutations in Human Immunodeficiency Virus Type 1 Reverse Transcriptase Involved in the Regulation of Resistance to Nonnucleoside Inhibitors{triangledown}

Francesca Ceccherini-Silberstein,1,{dagger}* Valentina Svicher,1,{dagger} Tobias Sing,2 Anna Artese,3,4 Maria Mercedes Santoro,1 Federica Forbici,5 Ada Bertoli,1 Stefano Alcaro,3,4 Guido Palamara,6 Antonella d'Arminio Monforte ,7,{ddagger} Jan Balzarini,8 Andrea Antinori ,5,§ Thomas Lengauer,2 and Carlo Federico Perno1,5,§

Department of Experimental Medicine, University of Rome Tor Vergata, Rome,1 Department of Pharmacobiological Sciences, University of Catanzaro Magna Graecia, Roccelletta di Borgia (CZ),3 National Institute of Nuclear Physics, Frascati, Rome,4 National Institute for Infectious Diseases L. Spallanzani, Rome,5 San Gallicano Hospital, Rome,6 Institute of Infectious and Tropical Diseases, University of Milan, Milan, Italy,7 Max-Planck-Institute for Informatics, Saarbrücken, Germany,2 Rega Institute for Medical Research, Katholieke Universiteit Leuven, Leuven, Belgium8

Received 12 February 2007/ Accepted 2 August 2007


arrow
ABSTRACT
 
Resistance to antivirals is a complex and dynamic phenomenon that involves more mutations than are currently known. Here, we characterize 10 additional mutations (L74V, K101Q, I135M/T, V179I, H221Y, K223E/Q, and L228H/R) in human immunodeficiency virus type 1 (HIV-1) reverse transcriptase which are involved in the regulation of resistance to nonnucleoside reverse transcriptase inhibitors (NNRTIs). These mutations are strongly associated with NNRTI failure and strongly correlate with the classical NNRTI resistance mutations in a data set of 1,904 HIV-1 B-subtype pol sequences from 758 drug-naïve patients, 592 nucleoside reverse transcriptase inhibitor (NRTI)-treated but NNRTI-naïve patients, and 554 patients treated with both NRTIs and NNRTIs. In particular, L74V and H221Y, positively correlated with Y181C, were associated with an increase in Y181C-mediated resistance to nevirapine, while I135M/T mutations, positively correlated with K103N, were associated with an increase in K103N-mediated resistance to efavirenz. In addition, the presence of the I135T polymorphism in NNRTI-naïve patients significantly correlated with the appearance of K103N in cases of NNRTI failure, suggesting that I135T may represent a crucial determinant of NNRTI resistance evolution. Molecular dynamics simulations show that I135T can contribute to the stabilization of the K103N-induced closure of the NNRTI binding pocket by reducing the distance and increasing the number of hydrogen bonds between 103N and 188Y. H221Y also showed negative correlations with type 2 thymidine analogue mutations (TAM2s); its copresence with the TAM2s was associated with a higher level of zidovudine susceptibility. Our study reinforces the complexity of NNRTI resistance and the significant interplay between NRTI- and NNRTI-selected mutations. Mutations beyond those currently known to confer resistance should be considered for a better prediction of clinical response to reverse transcriptase inhibitors and for the development of more efficient new-generation NNRTIs.


arrow
INTRODUCTION
 
Important progress in the development and clinical use of drugs for treating human immunodeficiency virus type 1 (HIV-1) infection has been made. To date, 21 antiretroviral drugs are approved. Most of them target two viral enzymes, reverse transcriptase (RT) and protease, while one drug, enfuvirtide, targets envelope glycoprotein gp41, which is involved in viral entry. The combined use of these drugs has substantially improved the clinical management of HIV-1 infection in terms of delaying disease progression, prolonging survival, and improving quality of life. Nevertheless, when antiretroviral therapy fails to be fully suppressive, new viral variants emerge, thus allowing HIV-1 to become resistant to one or more drugs by accumulating mutations, either alone or in multiple and complex patterns (13, 24, 18, 35, 37, 39). Understanding the genetic basis of resistance and cross-resistance is essential for optimizing the use of existing drugs and for designing new antiviral agents.

The HIV-1 RT enzyme is responsible for the conversion of the single-stranded RNA genome into a double-stranded DNA that is integrated into the host genomic DNA (19, 44). Owing to its pivotal role in the HIV-1 life cycle, RT still represents a very important target for antiviral therapy. More than half of the currently approved drugs for the treatment of HIV-1 infection are RT inhibitors: seven nucleoside RT inhibitors (NRTIs) (zidovudine [ZDV], stavudine [d4T], lamivudine [3TC], didanosine [ddI], abacavir [ABC], zalcitabine [ddC], and emtricitabine), one nucleoside phosphonate (tenofovir [TDF]), and three nonnucleoside RT inhibitors (NNRTIs) (nevirapine [NVP], efavirenz [EFV], and delavirdine) (2, 13, 15, 31).

NNRTIs are small molecules with a strong affinity for a hydrophobic pocket located close to the catalytic domain of RT. The binding of the inhibitors induces a structural alteration in RT, thus blocking its polymerase activity. Unfortunately, despite their high potencies of NNRTIs, coupled with their low toxicities, the use of NNRTIs is hindered by the rapid emergence of resistant viral strains (13, 24, 37, 39).

To date, 29 mutations at 16 sites of HIV-1 RT have been associated with resistance against the three currently approved NNRTIs (24; Stanford HIV Drug Resistance Database [http://hivdb.stanford.edu]). These mutations affect NNRTI binding directly by altering the size, shape, and polarity of different parts of the NNRTI binding pocket or indirectly by affecting the access of NNRTIs to the pocket itself (17, 22).

There is increasing evidence that additional mutations besides those currently known are involved in the development of NNRTI resistance, therefore leading to NNRTI therapy failure. For instance, recent studies have identified novel mutations positively associated with NNRTI treatment (9, 10, 18, 34, 36); however, the exact role of these mutations in contributing to NNRTI resistance remains unclear. Moreover, other studies have shown also the involvement in NNRTI resistance of mutations conferring resistance to NRTI, thus suggesting an interplay between NRTI- and NNRTI-characteristic mutations (12, 27, 40, 41, 45, 46). Finally, it has been suggested that the development of resistance to NNRTIs might be more complex than the classical one-step model of significant resistance via a single mutation so far considered (1).

Based on these assumptions, using statistical and computational methods together with structural analysis and molecular dynamics simulations (MDSs), we have focused our work on defining the roles in NNRTI resistance of nine previously uncharacterized mutations in HIV-1 RT and the as yet unclear role of the L74V NRTI resistance mutation. The involvement of novel mutations in NNRTI resistance may be particularly relevant in view of the upcoming approval of newer NNRTIs, reported to be active against strains carrying several mutations conferring resistance to currently available NNRTIs but whose efficacy is decreased by the presence of ≥3 NNRTI resistance mutations.

(Parts of these data have been presented at the 4th European HIV Drug Resistance Workshop: From Basic Science to Clinical Implications, Monte Carlo, Monaco, 29 to 31 March 2006; XV International HIV Drug Resistance Workshop: Basic Principles and Clinical Implications, Sitges, Spain, 13 to 17 June 2006; and the 7th Annual Symposium on Antiviral Drug Resistance, Chantilly, VA, 12 to 15 November 2006.)


arrow
MATERIALS AND METHODS
 
Patients. The study included 1,904 patients enrolled either in the Italian Cohort of Antiretroviral Naïve Patients (I.Co.N.A.) or in different clinical centers in central Italy; 758 patients were naïve to treatment with antiretroviral drugs, 592 were failing NRTIs but not NNRTIs, and 554 were failing NRTIs and NNRTIs at the time of genotypic resistance testing. Drug-naïve patients underwent genotypic tests between January 1997 and August 2005, while NRTI-treated patients were tested between January 1998 and August 2005.

Data for all patients were stored in a specifically designed anonymous database that included genotypic, demographic, immunologic, virologic, and therapeutic parameters.

HIV sequencing. HIV genotype analysis was performed on plasma samples by means of a commercially available kit (ViroSeq HIV-1 genotyping system; Abbott) (9, 32). Briefly, RNA was extracted, retrotranscribed by murine leukemia virus RT, and amplified with AmpliTaq Gold polymerase enzyme by using two different sequence-specific primers for 40 cycles. The full lengths of polymerase-amplified products (containing the entire protease and the first 335 amino acids of the RT open reading frame product) were sequenced in sense and antisense orientations by using seven different overlapping sequence-specific primers for an automated sequencer (ABI 3100). Sequences having a mixture of wild-type and mutant residues at a single position were considered to have the mutant(s) at that position. The isolates were subtyped by comparison to reference sequences of known subtypes (Stanford HIV Drug Resistance Database [http://hivdb.stanford.edu]). Patients carrying non-B subtypes were excluded from the analysis. Thus, all patients carried a B-subtype strain of HIV. The majority of nucleotide sequences of drug-naïve and RTI-treated patients have already been submitted to GenBank (9, 10, 32, 43).

Statistical analysis. (i) Mutation prevalence. To assess the association of RT mutations with NNRTI treatment, we calculated their respective frequencies in isolates from 758 drug-naïve patients, 592 patients failing their last antiretroviral regimen containing NRTIs but not NNRTIs, and 554 patients failing their last antiretroviral regimen containing NRTIs and NNRTIs. We then performed chi-square tests of independence (based on a 2-by-2 contingency table) to identify statistically significant differences in frequency between the following groups of patients: (i) drug-naïve patients versus NRTI- but not NNRTI-treated patients, (ii) drug-naïve patients versus patients treated with NRTIs plus NNRTIs, (iii) NRTI- but not NNRTI-treated patients versus patients treated with NRTIs plus NNRTIs. We also performed a generalized linear model that accounts for three levels of NRTI treatment (1 or 2, 3 or 4, and >4 NRTIs).

Based on this preliminary analysis (see Results), we focused our attention on the NRTI resistance mutation L74V and nine mutations (K101Q, I135M/T, H221Y, K223E/Q, L228H/R, and V179I), to which we refer as novel mutations for NNRTI resistance in the course of this article, as they have not yet been reported to be associated with NNRTI resistance by the Stanford HIV Drug Resistance Database (http://hivdb.stanford.edu) and by the International AIDS Society (24).

To assess the association of each novel mutation with treatment of a specific NNRTI, we compared its rate of occurrence in the subpopulation that underwent treatment with a specific NNRTI with the rate of mutation in the subpopulation that did not undergo treatment with that NNRTI; we then performed chi-square tests of independence to verify statistically significant differences (P < 0.05).

In our analysis, the Cochran rule, a conventional criterion for the chi-square test to be valid, was fully satisfied. In fact, in each contingency table performed with our data set, 80% of the expected frequencies exceeded 5, and all the expected frequencies exceeded 1. In addition, in those few cases in which the expected frequency in a single cell of the contingency table was less than 5, the significance was also confirmed by using a Monte Carlo procedure (21).

We used the Benjamini-Hochberg method (4) to identify results that were statistically significant in the presence of multiple-hypothesis testing. A false-discovery rate of 0.05 was used to determine statistical significance. This analysis was performed on the subset of 1,146 patients failing an antiretroviral regimen containing NRTIs with or without NNRTIs.

To assess the association of novel RT mutations with viremia and CD4 cell count, we compared viremias and CD4 cell counts of patients harboring isolates with a specific novel mutation with those of patients harboring isolates with a wild-type amino acid at that position. Viremia and CD4 cell count values analyzed were concomitant with the genotype resistance testing (or ±30 days). To verify statistically significant differences, the median test was performed and was corrected for multiple tests by using the Benjamini-Hochberg method at a false-discovery rate of 0.05.

To assess the association between the presence of novel mutations at the baseline and the presence of NNRTI resistance mutations at NNRTI failure, we selected 36 out of 554 NRTI- and NNRTI-treated patients that underwent genotypic resistance testing both at the onset and at the time of failure of an NRTI- and NNRTI-containing regimen. The former genotype was taken at most 90 days before the beginning of the NRTI- and NNRTI-containing regimen. Chi-square tests of independence were used to assess statistically significant associations.

(ii) Mutation covariation. For the set of isolates from 554 patients treated with NRTIs plus NNRTIs, we exhaustively analyzed patterns of pairwise interactions among mutations associated with NNRTI treatment, including the 10 novel mutations. Specifically, for each pair of mutations and corresponding wild-type residues, Fisher's exact test was performed to assess whether co-occurrence of the mutated residues differed significantly from what would be expected under an assumption of independence. Again, the Benjamini-Hochberg method was used to correct for multiple tests, in this case at a false-discovery rate of 0.01. Samples having a mixture of two or more mutations at a given pair of positions were ignored in the calculation of the covariation, due to the impossibility of identifying whether these mutations were indeed located in the same viral genome.

(iii) Cluster analysis. In order to identify and summarize higher-order interactions of mutations, we transformed the pairwise phi correlation coefficients into dissimilarity values. Based on these pairwise dissimilarity values, a dendrogram was computed by hierarchical clustering. Finally, the stability of the resulting dendrogram was assessed from 100 bootstrap replicates. The details of this explorative data analysis procedure have been described elsewhere (42, 43).

Briefly, the transformation from phi coefficients to dissimilarity values was done by mapping phi at 1 (maximal positive association) to dissimilarity 0 and phi at –1 (maximal negative association) to dissimilarity 1, with linear interpolation in between. The dissimilarity of mutations at the same position was left undefined, as such pairs never co-occur in a single sequence (except from mixtures) and would lead to artifacts in the resulting dendrogram. The resulting partial dissimilarity matrix was then used as the input for average-linkage hierarchical agglomerative clustering, and undefined dissimilarity values were ignored in computing average dissimilarities between clusters. In order to assess the stability of the resulting dendrogram, confidence values for all subtrees in the dendrogram were computed by 100 replicates of the clustering procedure on sequence sets bootstrapped from the original 554 sequences (42). For instance, a bootstrap value of 1 for an edge in the dendrogram means that the set of mutations in the induced subtree occurs as a subtree in all dendrograms from the different bootstrap replicates. Thus, higher bootstrap values indicate that the association of mutations into a group is not due to sampling bias.

Association with phenotypic NNRTI susceptibility. We analyzed genotype-phenotype correlations from the HIV Stanford Drug Resistance Database (http://hivdb.stanford.edu) to assess the association of mutations with NNRTI susceptibility. In particular, for each NNRTI, we compared the median change in resistance (n-fold) with or without specific novel mutations. Resistance (n-fold) was measured by Virco's Antivirogram assay. In addition to this univariate analysis, we measured the impact of novel NNRTI resistance mutations in multivariate computational models for predicting phenotypic resistance from genotype. In contrast to the univariate setting, multivariate analyses allow the assessment of the impact of mutations relative to other mutations. Specifically, we analyzed the support vector regression (SVR) models used in the geno2pheno Web-based prediction service (3; www.geno2pheno.org) by exploiting the bilinearity of the kernel used in geno2pheno, according to the method described in reference 42. These models are based on approximately 850 matched genotype-phenotype pairs derived from another recombinant assay (50) and are part of the German Arevir database.

Structural analysis. The X-ray crystallographic coordinates of HIV-1 RT deposited in the Protein Data Bank (PDB [http://www.rcsb.org/PDB/]) with code 1jkh were used for the structural analysis. To calculate the interatomic distances between residues that are significantly associated in statistical analysis, a script to measure the distance between each couple of atoms within two different amino acids has been developed. We then defined the interresidue distance as the shortest interatomic distance between any atoms in the two residues. Residues within a distance of 8 Å in the RT structure were considered near enough to interact (51). Both subunit p51 and subunit p66 were considered, and the shortest distances are reported in Table 2. Finally, the feasibility of direct side chain-side chain interaction between two mutated amino acids (both intra- and intersubunit) was assessed by the LigPlot program (49).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Significantly correlated pairs of mutations

MDS. Our computational work started with the analysis of the available crystallographic unbound models of RT. The 1DL0 PDB model (22) was judged as the most complete (971 residues; resolution = 2.70 Å) and was selected to generate the starting geometries for the MDS, adding on each subunit missing amino acids. MDSs were carried out by using the GROMACS software (6). All models were refined by including explicit solvation and Cl– counterions to neutralize the overall charge. We analyzed the wild-type RT model and the two RT mutants with mutations K103N and K103N/I135T, submitting to MDS the starting geometries obtained in the previous step by single-residue substitution. For each structure, we used the GROMOS96 53a6 force field with the explicit water implementation (30) and an octahedral box, setting the box edge at a distance of 3.0 nm from the target. All studied mutants were energy minimized for 50,000 steps, using the steepest descent conjugated gradient algorithm, with an energy convergence criterion fixed at 1,000 kJ/mol · nm. The Linear Constraints Solver algorithm was adopted for all atoms to prevent bond distortions as suggested for time steps larger than 1 fs (20). After energy minimization, we performed three MDSs by using Berendsen's temperature and pressure coupling methods (5) and particle mesh Ewald electrostatic treatment under coulombtype (14). In the first simulation, for 100 ps, we kept RT frozen and only water molecules were free to move; in the second one, water molecules and added residues were allowed to move for 500 ps, while the entire enzyme was still fixed; and in the last one, the whole enzyme was free to move for 1 ns. We analyzed several geometrical parameters by adopting different GROMACS utilities, like g_hbonds, which computes and analyzes hydrogen bonds during the whole simulation. We used the default distance cutoff of 2.5 angstroms and the default angle cutoff of 60 degrees. All calculations were assessed by a Linux cluster of five Intel Xeon dual processors at 3.2 GHz with 2 Gb of RAM. In order to consider the protein-protein contribution term at the p66-p51 interface, we used the GROMACS utility g_energy, and we selected the Coul-SR protocol to analyze electrostatic interactions in the short range (28).

Nucleotide sequence accession numbers. A total of 68 nucleotide sequences from NNRTI-treated patients have been submitted to GenBank under accession numbers EU019761 through EU019835.


arrow
RESULTS
 
Patient characteristics. The study included 1,904 HIV-1 B-subtype pol sequences from 758 drug-naïve patients, 592 NRTI-treated/NNRTI-naïve patients, and 554 patients treated with NRTIs plus NNRTIs. The 592 NRTI- but not NNRTI-treated patients were exposed to an average of four NRTIs, and 515 (87.0%) were receiving highly active antiretroviral therapy. Specifically, 84.5% of patients had been treated with 3TC, 79.7% with ZDV, 59.1% with d4T, 40.7% with ddI, 12.5% with ddC, 10.1% with ABC, and 5.2% with TDF. At the time of genotypic analysis, 72.6% of patients were receiving treatment with 3TC (for a median of 617 days), 46.5% with ZDV (median, 541 days), 46.1% with d4T (median, 663 days), 22.5% with ddI (median, 519 days), 4.6% with TDF (median, 262 days), 7.8% with ABC (median, 447 days), and 4.6% with ddC (median, 366 days). The 554 patients treated with NRTIs plus NNRTIs were all receiving highly active antiretroviral therapy and were exposed to averages of four NRTIs and one NNRTI. Regarding the NRTIs, 94.0% of these patients had been treated with 3TC, 80.9% with ZDV, 76.4% with d4T, 64.9% with ddI, 19.3% with ABC, 17.2% with ddC, and 13.7% with TDF. Regarding the NNRTIs, 57.8% of patients had been treated with EFV and 59.6% with NVP, with 114 out of 554 patients treated with both EFV and NVP. Regarding the RTIs received at the time of genotypic analysis, 63.2% of patients were receiving treatment with 3TC (for a median of 421 days), 50.0% with EFV (median, 428 days), 50.0% with NVP (median, 403 days), 45.5% with d4T (median, 479 days), 36.1% with ddI (median, 377 days), 29.6% with ZDV (median, 481 days), 15.6% with TDF (median, 222 days), 9.6% with ABC (median, 381 days), and 0.9% with ddC (median, 356 days).

Novel RT mutations and their association with NNRTI treatment. By evaluating the first 335 amino acids in HIV-1 RT sequences derived from 758 drug-naïve and 1,146 RTI-treated patients (592 failing their last antiretroviral regimen containing NRTIs but not NNRTIs and 554 failing their last antiretroviral regimen containing NRTIs and NNRTIs), we found that the NRTI resistance mutation L74V and nine novel uncharacterized RT mutations were associated with NNRTI treatment, on the basis of their different frequencies of occurrence in NNRTI-naïve subjects and patients failing NNRTI therapy (Fig. 1).


Figure 1
View larger version (13K):
[in this window]
[in a new window]

  		FIG. 1. Frequencies of novel RT mutations in isolates from drug-naïve patients and in isolates from patients failing their last antiretroviral regimen containing NRTIs but not NNRTIs and from patients failing their last antiretroviral regimen containing NRTIs plus NNRTIs. Statistically significant differences were assessed by chi-square tests of independence. P values were significant at a false-discovery rate of 0.05 following correction for multiple tests. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

In particular, eight mutations at six positions (L74V, K101Q, I135M, H221Y, K223E/Q, and L228H/R) were completely absent or occurred with a frequency of <1.2% in isolates from drug-naïve patients, and they had significantly increased frequencies (P values from 0.05 to <0.00001) in isolates from patients failing NNRTI treatment (Fig. 1). This increase in frequency was statistically significant not only with respect to drug-naïve patients but also in comparison with the frequencies of mutations in NRTI-treated but NNRTI-naïve patients (with the exception of K223E/Q). In particular, the NRTI resistance mutation L74V occurred at a frequency of 2.5% in patients failing only NRTIs (NNRTI naïve) and increased by 5.6-fold to 13.9% in patients failing treatment with NRTIs plus NNRTIs (P < 10–11). To further confirm the association of L74V with NNRTI treatment, we repeated the analysis, excluding all patients that experienced treatment with ddI and/or ABC, the two already known NRTIs selecting the L74V mutation. In particular, we selected 219 NRTI-treated patients naïve to NNRTIs, ddI, and ABC and also 169 NNRTI-treated patients naïve to ddI and ABC. In isolates from the first group of patients, the L74V mutation was completely absent. In contrast, in the isolates from the group of NNRTI-treated patients naïve to ddI and ABC, the frequency of L74V showed a statistically significant increase up to 4.1% (P = 0.003). The L74V mutation remained significantly associated with NNRTI treatment also in a multivariate model that controlled for the number of NRTIs received.

None of the other NRTI-characteristic resistance mutations was associated with NNRTI treatment in both univariate and multivariate analyses.

Differently, mutations I135T and V179I were already present in isolates from drug-naïve patients (at a frequency of >3.5%), and their frequencies significantly increased (P values from 0.05 to <0.0001) in isolates from NNRTI-treated patients (Fig. 1). In particular, I135T occurred at a frequency of around 32% in both drug-naïve and NNRTI-naïve patients and reached 42.8% in patients failing NNRTI treatment.

To confirm the association of mutations with NNRTI treatment also in patients receiving an initial antiretroviral regimen and to exclude biases introduced by sequential use of specific regimens, we selected 463 patients receiving an initial antiretroviral regimen containing NRTI but not NNRTI and 161 patients receiving an initial antiretroviral regimen containing NRTIs and NNRTIs. For both groups of patients, we calculated the frequencies of the novel RT mutations. Based on this analysis, the association with NNRTI treatment was fully confirmed for the novel RT mutations K101Q, I135M, H221Y, L228H, L228R, and V179I (P values from 0.02 to <0.001).

Most mutations were also significantly associated (P < 0.05) with the use of specific NNRTIs (Table 1) . Specifically, K101Q, I135M, K223Q, and L228R were associated with the use of EFV but not NVP, while H221Y, L228H, and V179I were associated with the specific use of NVP and L74V and I135T were significantly associated with the use of both NVP and EFV, thus confirming the existence of somewhat different resistance pathways occurring for NVP and EFV. In addition, the frequency of H221Y in patients failing NNRTI-containing regimens without ZDV was higher than that in patients failing NNRTI- and ZDV-containing regimens (P = 0.004) (data not shown), thus showing a negative association of H221Y with the use of ZDV.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Association of RT mutations with the use of specific NNRTIs

Associations among RT mutations. A further step of our study was to assess the association of these mutations with other RT mutations observed in the 554 NNRTI-treated patients, focusing our attention on the mutations involved in NNRTI resistance.

To identify significant patterns of pairwise correlations between novel mutations and specific RT mutations observed in isolates from NNRTI-treated patients, we calculated the binomial correlation coefficient (phi) and its statistical significance for each pair of mutations (Table 2).

Novel RT mutations involved in positive and negative interactions with NNRTI resistance mutations. Many mutations for which data are shown in Fig. 1 were positively correlated as pairs with other NNRTI resistance mutations (Table 2). In particular, the most frequently selected NNRTI resistance mutation, K103N (the prevalence was 54.7% in our cohort of NNRTI treatment-failing patients), was correlated with I135M (phi = 0.14) and I135T (phi = 0.12). Interestingly, 91.7% of isolates from patients with I135M carried the K103N mutation. In contrast, K103N was negatively associated with V179I (phi = –0.15) and K223E (phi = –0.10). The Y181C mutation (prevalence, 24.9%) was correlated with H221Y (phi = 0.18), L228H (phi = 0.16), L228R (phi = 0.12), and K223E (phi = 0.14), while Y188C (prevalence, 0.7%) was preferentially associated with I135M (phi = 0.25). In addition, A98G (prevalence, 4.7%) was correlated with K223E (phi = 0.17), I135M (phi = 0.15), and L228H (phi = 0.14), while F227L (prevalence, 3.2%) was correlated with K223E (phi = 0.23). The novel mutation K101Q was not correlated with any NNRTI resistance mutations and was positively correlated only with K102R (phi = 0.18). Also, the NRTI resistance mutation L74V was positively correlated with specific NNRTI resistance mutations (P < 0.05). In particular, L74V was positively correlated with Y181C (phi = 0.13), L100I (phi = 0.20), and the novel mutations H221Y (phi = 0.18), K223E (phi = 0.14), and L228H (phi = 0.11) (Table 2). The association of L74V with Y181C was fully confirmed (57% covariation frequency in patients with L74V, phi = 0.20, P = 8.4e–3) in the subset of 169 NNRTI-treated patients naïve to ddI and ABC (data not shown).

The novel RT mutation H221Y is involved in negative interactions with NRTI resistance mutations. Beyond the positive correlations with the classical NNRTI resistance mutations, H221Y, which was negatively associated with the use of ZDV, showed a strong negative correlation with some NRTI resistance mutations (P < 0.05). In particular, H221Y was negatively associated with the type 2 thymidine analogue mutations (TAM2s) D67N (phi = –0.21), K70R (phi = –0.18), K219Q (phi = –0.14), and T215F (phi = –0.12) (Table 2). Weak negative correlations were observed also between H221Y and TAM1s, even though these correlations were not statistically significant (data not shown). None of the other novel mutations showed negative correlations with the classical NRTI resistance mutations.

Clusters of correlated mutations. Pairwise analysis suggested that most novel mutations are associated with specific pathways of known NNRTI resistance mutations; therefore, we performed an average linkage hierarchical agglomerative cluster analysis (42) to investigate this hypothesis in more detail.

The topology of the dendrogram (Fig. 2) suggests the existence of five distinct clusters involving the novel and the already known NNRTI resistance mutations. In particular, a strong cluster was formed by K103N and L100I (bootstrap value = 0.89), which was linked with the novel I135T mutations (bootstrap value = 0.70) and the already known P225H and K238T mutations (bootstrap value = 0.65) (Fig. 2). As well, V179I clustered with K103S and K101P (bootstrap value = 0.47), while H221Y clustered with Y181C and with the NRTI resistance mutation L74V (bootstrap value = 0.41). A strong cluster was observed for G190A and K101E (bootstrap value = 0.94), and a strong cluster was also observed for F227L and V106A (bootstrap value = 0.71), to which the novel K223E was linked as well (bootstrap value = 0.44) (Fig. 2).


Figure 2
View larger version (20K):
[in this window]
[in a new window]

  		FIG. 2. Dendrogram obtained from average linkage hierarchical agglomerative clustering, showing clusters of known NNRTI resistance mutations and novel mutations. The length of branches reflects distances between mutations in the original distance matrix. Bootstrap values, indicating the significance of clusters, are reported in the boxes.

Association of novel RT mutations with viral load and CD4+ cell count. Another goal of our study was to investigate the association of novel RT mutations with the virologic and immunologic values at the times of therapeutic failure and genotype testing.

Among all mutations analyzed, two novel mutations, K223Q and L228H, were significantly associated with a worse virologic outcome in cases of NNRTI failure. In fact, the presence of K223Q or L228H in isolates from patients treated with NRTIs and NNRTIs was significantly associated with a 6.4-fold or 3.2-fold increase, respectively, in viremia compared to viremia in the absence of such a mutation (48,179 copies/ml versus 7,515 copies/ml [P = 0.01] or 23,401 copies/ml versus 7,269 copies/ml [P = 0.01], respectively) (data not shown). In a similar fashion, the copresence of L74V with K103N/L100I was significantly associated with a ≥2.8-fold increase in viremia at therapeutic failure compared to patients without L74V (20,611 copies/ml versus 7,400 copies/ml, P = 0.02) (data not shown). In this analysis, no statistically significant association with changes in CD4+ cell count were observed.

Association of novel RT mutations with reduced NNRTI susceptibility. A further step in our study was to assess the direct contribution of novel mutations to NNRTI resistance.

By analyzing sequences with known NNRTI susceptibility (measured in vitro by Virco's Antivirogram assay) from the Stanford HIV Drug Resistance Database, we found that the presence of some mutations at therapeutic failure was significantly associated with greater EFV or NVP resistance (n-fold) (Table 3). In particular, the presence of K101Q or I135T with K103N was associated with greater EFV resistance and also with an increase in NVP resistance (albeit less extensive). Similarly, the presence of I135M with K103N was associated with a 9.5-fold increase in EFV resistance, while no effect was observed for NVP. These results confirmed the association of I135M/T with K103N identified in the covariation analysis (Table 2 and Fig. 2).


View this table:
[in this window]
[in a new window]

  		TABLE 3. Novel RT mutations associated with a decreased NNRTI susceptibility

In contrast, the copresence of H221Y and Y181C resulted in a 12.4-fold increase in NVP resistance and a less extensive increase in EFV resistance (Table 3). Interestingly, the presence of NRTI resistance mutation L74V along with Y181C was also associated with a 19.8-fold increase in NVP resistance and a 2.7-fold in EFV resistance (Table 3). Again, these results fully confirmed the association observed for H221Y and L74V with Y181C (Table 2 and Fig. 2).

To investigate the contribution of novel mutations to NNRTI resistance, we also analyzed an independent data set of 850 matched genotype-phenotype pairs for each NNRTI by feature ranking based on SVR. In contrast to the univariate analysis, this multivariate procedure allows the quantification of the impact of mutations relative to other mutations. SVR showed that all known NNRTI resistance mutations (24; Stanford HIV Drug Resistance Database) appeared in the 50 mutations with the greatest weight in NVP and/or EFV resistance among 6,365 mutations analyzed (Fig. 3), thus supporting that this model may accurately capture the current knowledge regarding resistance to NNRTIs. In addition, the ranking of mutations by SVR weight provided evidence of the different contributions of a specific mutation to EFV and NVP resistance. For instance, mutation Y181C, present in 40% of patients failing NVP, has the third-greatest weight in the SVR model for NVP resistance. In contrast, its prevalence remarkably decreased to 6% and also its weight moved down to the rank number 9 in the context of EFV failure. When we investigated the role of novel mutations in the model, we found that many of them were significantly involved in determining NVP or EFV resistance. In particular, H221Y, K101Q, I135T, V179I, and L74V were among the top 25 mutations significantly associated with NVP resistance (Fig. 3). Among them, H221Y was included in the top 10 mutations that contribute to resistance to this drug (Fig. 3). Similarly H221Y, L74V, K101Q, I135M/T, and V179I were among the top 30 mutations associated with EFV resistance, with H221Y, L74V, and K101Q ranking even above some classical EFV mutations, such as V108I and G190E (24; Stanford HIV Drug Resistance Database [http://hivdb.stanford.edu]). Interestingly, the I135L mutation was included among the top 10 and top 25 mutations associated with NVP and EFV resistance, respectively, even if this mutation did not show a specific association with NNRTI treatment and NNRTI resistance mutations.


Figure 3
View larger version (36K):
[in this window]
[in a new window]

  		FIG. 3. Mutation ranking based on SVR showing the major mutations significantly associated with resistance to NVP and EFV among 6,365 mutations analyzed. Asterisks indicate Z scores of the effects of mutations on EFV and NVP resistance. Z scores were calculated on the basis of approximately 850 matched genotype-phenotype pairs that derived from a recombinant assay described by Walter et al. (50) and are part of the German Arevir database. The prevalences of each mutation in isolates from the subset of 223 patients that were failing an NVP-containing regimen at the time of genotyping but were naïve to EFV and in isolates from the subset of 217 patients that were failing an EFV-containing regimen at the time of genotyping but were naïve to NVP were assessed.

Since our analysis showed a negative association of H221Y with the use of ZDV and with some TAMs, we investigated the impact of this mutation on ZDV susceptibility. The presence of H221Y with mutations in the TAM2 cluster (D67N, K70R, K219Q/E and T215F) was associated with a remarkable increase in ZDV susceptibility (P = 0.003) (Table 4). This increase was even more marked than that observed for mutation Y181C already known to be associated with ZDV resensitization (38) (Table 4). H221Y also conferred an increase in ZDV susceptibility in the presence of the TAM1 cluster (M41L, L210W, and T215Y), albeit less extensive than that observed for Y181C (Table 4).


View this table:
[in this window]
[in a new window]

  		TABLE 4. Association of Y181C and H221Y with an increased ZDV susceptibility

Association between the presence of novel RT mutations at the baseline and the presence of NNRTI resistance mutations at NNRTI failure. Since I135T and V179I were already present in drug-naïve patients with a frequency higher than 3% (for I135T, >30%), we investigated whether their presence at the baseline was correlated with the presence of some classical NNRTI resistance mutations at NNRTI therapy failure. Thus, within the overall cohort of 554 patients treated with NRTIs and NNRTIs, we selected 36 patients that underwent genotypic resistance testing at the time of therapy onset and at the time of therapy failure. Among the mutations analyzed, the presence of I135T at the baseline was significantly correlated with the presence of K103N at NNRTI failure. In fact, 73.3% of patients with HIV-1 isolates containing K103N at the time of NNRTI failure harbored the I135T mutation at the baseline of NRTI and NNRTI therapy, compared to the overall prevalence of 54.7% (P = 0.028).

Differently, the presence of V179I at the baseline was neither positively nor negatively associated with the presence of any NNRTI resistance mutations at therapy failure.

Structural analysis of novel RT mutations. All novel mutations are localized in the NNRTI binding pocket, precisely, in the p66 subunit, with the exception of I135T, which is located in the p51 subunit (Fig. 4).


Figure 4
View larger version (106K):
[in this window]
[in a new window]

  		FIG. 4. Localization of novel mutations in the structure of HIV-1 RT. The NNRTI binding pocket with bound EFV is shown. The domains are color coded (p66, yellow; p51, magenta). NNRTI resistance mutations are shown in cyan, novel mutations are shown in red, and the NRTI resistance mutation L74V is shown in green. EFV is shown as blue.

Among the 26 significantly correlated pairs of residues, four pairs of residues were within 10 Å of each other and were involved in chemical interactions. In particular, residue 101 may establish hydrogen bonds with EFV and may directly interact with correlated residue 102 by van der Waals interactions. Similarly, correlated residues 227/228 and 227/223 engage with each other through van der Waals interactions, while residue 179 may interact with residue 190 by a hydrogen bond.

In studying the other pairs of correlated residues that were separated by >10 Å, we focused our attention on the correlated pair I135T and K103N, since we observed that I135T was associated with an increase in the level of K103N-mediated EFV resistance and, in addition, its presence at the baseline (before NNRTI treatment) correlated with the presence of K103N at NNRTI failure.

By performing MDS, we confirmed previous findings concerning the formation of a hydrogen bond between the phenol oxygen of tyrosine (Y) 188 and the side chain carboxamide of asparagine (N) 103 (Fig. 5). It has been suggested that this hydrogen bond induces the closed-pocket form of the NNRTI binding pocket, thus hampering the entrance of the drug (22). Therefore, we investigated the impact of the 135T mutation on the formation of this crucial hydrogen bond, by monitoring over 1 ns, every 5 ps (for 201 total observations), the distance and the frequency of the occurrence of the hydrogen bond between 188Y and 103N (Fig. 5). Our MDS showed that the phenol oxygen of 188Y and the side chain carboxamide of 103N are closer in the presence of 135T than they are in the presence of the wild-type I residue at position 135 (distance of 1.94 Å versus 2.90 Å) (Fig. 5B). In addition, the number of occurrences of a hydrogen bond over 1 ns between 188Y and 103N in the presence of 135T was significantly higher than that in the presence of the 135I wild type (152 occurrences versus 18 occurrences; P < 0.001). This suggests that I135T may favor the stabilization of the closed form of the NNRTI binding pocket induced by K103N.


Figure 5
View larger version (76K):
[in this window]
[in a new window]

  		FIG. 5. Insight into the NNRTI binding pocket showing the distance in angstroms (d) between NH2 of 103N and the OH of 188Y in the presence of 135I (A) or 135T (B). Side chains of residues 103N, 188Y, and 135I/T are highlighted in wire-frame. The p66 subunit is shown as yellow, while the p51 subunit is shown as magenta.

In addition, analyzing within the NNRTI binding pocket the energy related to the protein-protein electrostatic interaction at the heterodimer p66-p51 interface, we found that the mutation K103N itself determined a reduction of this energetic term with respect to the wild-type sequence (average Coul-SR energy, –38,004.6 versus –37,457.6 kJ/mol) and this average energy decreased much more with the presence of double mutation I135T/K103N (average Coul-SR energy, –38,679.1 kJ/mol).


arrow
DISCUSSION
 
Our study suggests that other HIV-1 RT mutations, beyond those currently known as NNRTI resistance mutations, may contribute to the development of NNRTI resistance in vivo. In fact, 10 novel mutations, including the L74V NRTI resistance mutation, have been identified and characterized.

In particular, the mutations L74V, K101Q, I135M, H221Y, K223E/Q, and L228H/R were rare or completely absent in isolates from drug-naïve patients. In contrast, their frequencies in isolates from patients failing an antiretroviral regimen containing NNRTIs showed significant increases, thus suggesting their specific association with NNRTI treatment and failure. Moreover, the fact that these mutations occurred predominantly in viral proteins carrying at least one known NNRTI resistance mutation (data not shown) suggests that they emerge after a prolonged NNRTI exposure when the virus has already accumulated "primary" NNRTI resistance mutations. As a confirmation of this hypothesis, a previous study demonstrated that the emergence of the K101Q mutation might require as a prerequisite the presence of K103N (1). Thus, a prolonged NNRTI exposure may trigger the accumulation of additional mutations, leading to even higher levels of drug resistance.

All these mutations established positive interactions with the currently known NNRTI resistance mutations. In particular, H221Y was strongly associated with the use of NVP and showed positive interactions with Y181C, which is known to confer high levels of resistance to NVP by reducing the contribution of the aromatic interaction crucial for the binding of this NNRTI (33). In the SVR model, H221Y was included in the top 10 and 15 determinants for NVP and EFV resistance, respectively, ranking even above some classical NNRTI resistance mutations, such as K101E, V108I, and G190E (24; Stanford HIV Drug Resistance Database [http://hivdb.stanford.edu]).

Interestingly, the H221Y mutation was also negatively associated with the use of ZDV and with TAMs (particularly TAM2s) and was then associated, when rarely present with TAMs, with an increased susceptibility to ZDV. In particular, when H221Y was copresent with TAM2s, the increase in ZDV susceptibility was even greater than that observed when TAM2s were copresent with Y181C, which is already known to be associated with ZDV resensitization by reducing nucleoside excision (38).

Our study showed that also L74V, an important mutation conferring resistance to ddI and ABC and conferring hypersusceptibility to ZDV, was closely associated with NNRTI failure and NNRTI resistance mutations. These results were observed to be independent of patient experience with ddI and ABC and are consistent with others' data suggesting that L74V/I mutations can be selected in patients exposed to NNRTIs and also with no prior use of ABC or ddI (8a). In addition, another previous study showed that the prolonged exposure to quinoxaline NNRTIs may determine the selection of L74I/V and V75I (25), again highlighting the significant interplay occurring between mutations involved in NRTI and NNRTI resistance (7, 12, 38, 40, 41, 46). In particular, it has been postulated that L74V may play a compensatory role to mitigate the deleterious effects of NNRTI resistance mutations such as G190A/E/S on the positioning of the primer template and, consequently, on the catalytic activity of the mutant RT (7). A more recent study confirms this compensatory role of L74V in the restoration of replicative capacity impaired by the classical NNRTI resistance mutations (26). Consistent with this finding, our study showed that the copresence of L74V together with K103N/L100I was significantly associated with higher levels of viremia at therapeutic failure. In addition, L74V is included in the top 15 and 25 determinants of EFV and NVP resistance, respectively, confirming that L74V, in addition to acting as a compensatory NNRTI resistance mutation for viral fitness, may also contribute directly to NNRTI resistance.

Thus, taken together, these data suggest that two novel mutations (L74V and H221Y) that cluster frequently with Y181C, alone or together, share with Y181C the ability to increase NNRTI resistance (particularly to NVP) and at the same time the ability to increase ZDV susceptibility. Additional in vitro and in vivo studies will be necessary to better distinguish and highlight the individual roles of these mutations and their mechanisms of action.

Unlike the previously discussed group of mutations, I135T and V179I are mutations already present in isolates from drug-naïve patients, and these mutations showed a positive association with NNRTI failure and NNRTI resistance mutations. V179I also has been recently associated with a significant decrease in TMC125 susceptibility when present with some classical NNRTI resistance mutations (47), thus confirming its role in NNRTI resistance.

Interestingly, the I135T mutation, observed in 32.6% of drug-naïve patients, seems to play an unexpected role in the development of NNRTI resistance. For instance, it already has been associated with drug resistance and/or positive selection pressure in a previous study (11), based on ratios of nonsynonymous to synonymous mutations of >1, similar to those observed for the already known RTI resistance mutations. In another study, it has been shown that I135T alone (and together with another novel mutation, L283I) may reduce susceptibility to NVP and delavirdine in vitro (8), and the authors suggested that these two mutations may play a role in the reduced susceptibility to NNRTI observed for individuals with primary HIV infections without the presence of other NNRTI resistance mutations.

For our cohort of patients, I135T was significantly associated with the failure of EFV and with K103N. It clustered with K103N together with L100I, P225H, and K238T mutations, which are known to occur almost exclusively with K103N and to contribute mainly to EFV failure (1). In particular, P225H (as well as V108I) has been demonstrated to increase the ability of K103N to stabilize the closed form of the NNRTI binding pocket (23). Similarly, our MDSs showed that I135T can contribute to stabilize the K103N-induced closure of the NNRTI binding pocket by reducing the distance and increasing the number of hydrogen bonds between 103N and 188Y residues.

In addition, the presence of I135T at the baseline (in the absence of NNRTI therapy) was significantly associated with the appearance of K103N at NNRTI failure, confirming a previous finding showing the association of mutations at position 135 with the development of NNRTI resistance mutations at therapeutic failure (46a).

Therefore, there is increasing evidence that the common polymorphism I135T may play a role itself to reduce NNRTI susceptibility and may represent a crucial determinant of NNRTI resistance evolution also in first-line treatment, influencing the occurrence and accumulation of the classical drug resistance mutations.

Interestingly, all novel mutations, with the exception of two (L228H/R), confer changes in known HLA-restricted cytotoxic-T-lymphocyte epitopes in HIV-1 RT (HIV Molecular Immunology Database [http://www.hiv.lanl.gov]), thus suggesting that they may also reflect the adaptation of HIV-1 to the selective pressure imposed by the immune system. In particular, position 135 has been shown to be the anchor position of the HLA-restricted B*5101 epitope encompassing positions 128 to 135 in RT (29). In vitro experiments demonstrated the ability of I135T to abrogate the epitope-HLA binding, thus contributing to the loss of cytotoxic-T-lymphocyte responses in vivo (16). This may explain why the I135T mutation occurs with high frequency in both drug-naïve and NNRTI-treated populations. Most likely, the I135T mutation contributes to the escape of HIV from the HLA-restricted immune response as well as from the inhibitory activity of the NNRTIs, thus representing an interesting example of synergistic interactions between immune and drug pressures. In contrast, the NNRTI-associated position 101 was found to be negatively associated with the HLA-A2 genotype (29), thus suggesting that in the presence of HLA-A2-restricted immune response, position 101 may be under negative selective pressure that favors the preservation of the wild-type amino acid (in fact, the frequency of K101Q in the absence of NNRTI pressure remained ≤1%). Although further studies are necessary to better understand synergistic and/or antagonistic interactions between immune pressure and drug pressure, these findings underline the fact that the evolution of drug resistance may be constrained by host genetic factors.

In summary, our data suggest that novel mutations may actively participate in the regulation of NNRTI resistance, thus supporting the concept that the development of NNRTI resistance may be more complex than the classical one-step resistance given by a single mutation. This may represent a concept of remarkable importance and may also be relevant in view of the upcoming approval of new expanded-spectrum NNRTIs, reported to be active against virus strains carrying a few mutations conferring resistance to narrow-spectrum NNRTIs but whose efficacy is markedly decreased by the presence of ≥3 NNRTI-related mutations. Further in vitro and clinical studies are therefore necessary to confirm the efficacy and power of these new expanded-spectrum NNRTIs and to understand better the mechanisms underlying the development of drug resistance.

These results also suggest the importance of investigating the prevalence and the role of these novel mutations in other non-B subtypes. In fact, it is known that pathways of viral evolution toward drug resistance may proceed through distinct steps and at different rates among different HIV-1 subtypes (48). Moreover, we cannot exclude the possibility that the observed patterns of correlated mutations may be the result of pharmacological pressure imposed by the drug regimens that were used by our cohort, while other treatment regimens may lead to the development of pathways which are partly different from those that we observed. Further analyses on enlarged databases (with genotypic and clinical and/or phenotypic data), complemented by experimental validation, will provide insights regarding these important open points and will lead to an improved understanding of the impact of these mutations on viral replication and drug resistance.

In conclusion, our study reinforces the diversity and complexity of resistance to RT inhibitors and the significant interplay between NRTI- and NNRTI-selected mutations. Concordant results from large clinical and genotype-phenotype data sets provide strong evidence that other mutations beyond those currently known to confer resistance are involved in the evolutionary adaptation of HIV-1 to NNRTI-containing regimens. We suggest that these additional mutations should be considered for the improvement of algorithms that predict clinical responses to antiretroviral drugs and for assessing the efficacies of next-generation drugs.


arrow
ACKNOWLEDGMENTS
 
This work was financially supported by grants from the Italian National Institute of Health, the Ministry of University and Scientific Research, Current and Finalized Research of the Italian Ministry of Health, and the European Community (QLK2-CT-2000-00291 and the Descartes Prize HPAW-90001). The research at the Max-Planck-Institute for Informatics was performed in the context of the EU BioSapiens Network of Excellence (European Union grant no. LSHG-CT-2003-503265). The computational work was supported by the LNF-INFN AMICO project coordinated by Vitaliano Chiarella (Laboratori Nazionali di Frascati, Istituto Nazionale di Fisica Nucleare, Frascati, Rome, Italy). The I.Co.N.A. network is supported by unrestricted educational grants from Glaxo-Smith-Kline, Italy.

We thank Caterina Gori, Roberta D'Arrigo, Fabio Continenza, Daniele Pizzi, Andrea Biddittu, Sara Bono, Lavinia Fabeni, Francesca Stazi, and Sandro Bonfigli for sequencing and data management, all other participants and members of INMI-Collaborative Group for Clinical Use of HIV Genotype Resistance Test, and the I.Co.N.A. Study Group.

Members of the I.Co.N.A. Study Group in Italy are as follows: from Ancona, M. Montroni, G. Scalise, A. Zoli, and M. S. Del Prete; from Aviano (PN), U. Tirelli and G. Di Gennaro; from Bari, G. Pastore, N. Ladisa, and G. Minafra; from Bergamo, F. Suter and C. Arici; from Bologna, F. Chiodo, V. Colangeli, C. Fiorini, and O. Coronado; from Brescia, G. Carosi, G. P. Cadeo, F. Castelli, C. Minardi, and D. Vangi; from Busto Arsizio, G. Rizzardini and G. Migliorino; from Cagliari, P. E. Manconi and P. Piano; from Catanzaro, T. Ferraro and A. Scerbo; from Chieti, E. Pizzigallo and M. D. Alessandro; from Como, D. Santoro and L. Pusterla; from Cremona, G. Carnevale and D. Galloni; from Cuggiono, P. Viganò and M. Mena; from Ferrara, F. Ghinelli and L. Sighinolfi; from Firenze, F. Leoncini, F. Mazzotta, M. Pozzi, and S. Lo Caputo; from Foggia, G. Angarano, B. Grisorio, and S. Ferrara; from Galatina (LE), P. Grima and P. Tundo; from Genova, G. Pagano, N. Piersantelli, A. Alessandrini, and R. Piscopo; from Grosseto, M. Toti and S. Chigiotti; from Latina, F. Soscia and L. Tacconi; from Lecco, A. Orani and P. Perini; from Lucca, A. Scasso and A. Vincenti; from Macerata, F. Chiodera and P. Castelli; from Mantova, A. Scalzini and G. Fibbia; from Milano, M. Moroni, A. Lazzarin, A. Cargnel, G. M. Vigevani, L. Caggese, A. d'Arminio Monforte, D. Repetto, R. Novati, A. Galli, S. Merli, C. Pastecchia, and M. C. Moioli; from Modena, R. Esposito and C. Mussini; from Napoli, N. Abrescia, A. Chirianni, C. Izzo, M. Piazza, M. De Marco, V. Montesarchio, E. Manzillo, and M. Graf; from Palermo, A. Colomba, V. Abbadessa, T. Prestileo, and S. Mancuso; from Parma, C. Ferrari and P. Pizzaferri; from Pavia, G. Filice, L. Minoli, R. Bruno, and S. Novati; from Perugia, F. Balzelli and K. Loso; from Pesaro, E. Petrelli and A. Cioppi; from Piacenza, F. Alberici and A. Ruggieri; from Pisa, F. Menichetti and C. Martinelli; from Potenza, C. De Stefano and A. La Gala; from Ravenna, G. Ballardini and E. Briganti; from Reggio Emilia, G. Magnani and M. A. Ursitti; from Rimini, M. Arlotti and P. Ortolani; from Roma, L. Ortona, F. Dianzani, G. Ippolito, A. Antinori, G. Antonucci, S. D. Elia, P. Narciso, N. Petrosillo, V. Vullo, A. De Luca, L. Del Forno, M. Zaccarelli, R. Acinapura, P. De Longis, M. Ciardi, G. D. Offizi, M. P. Trotta, P. Noto, M. Lichtner, M. R. Capobianchi, E. Girardi, P. Pezzotti, and G. Rezza; from Sassari, M. S. Mura and M. Mannazzu; from Torino, P. Caramello, A. Sinicco, M. L. Soranzo, L. Gennero, M. Sciandra, and M. Bonasso; from Varese, P. A. Grossi and C. Basilico; from Verbania, A. Poggio and G. Bottari; from Venezia, E. Raise and S. Pasquinucci; from Vicenza, F. De Lalla and G. Tositti; and from Taranto, F. Resta, and A. Chimienti. The member of the I.Co.N.A. Study Group from London, United Kingdom, is A. Cozzi-Lepri.

Members of the Collaborative Group for Clinical Use of HIV Genotype Resistance Test at the National Institute for Infectious Diseases Lazzaro Spallanzani in Rome are as follows: Andrea Antinori (co-chair), Gianfranco Anzidei, Francesco Baldini, Rita Bellagamba, Maria Concetta Bellocchi, Ada Bertoli, Sandro Bonfigli, Evangelo Boumis, Francesca Ceccherini-Silberstein, Bruno Christian Ciancio, Fabio Continenza, Roberta D'Arrigo, Patrizio De Longis, Gianpiero D'Offizi, Federica Forbici, Sara Giannella, Enrico Girardi, Caterina Gori, Giuseppina Liuzzi, Patrizia Lorenzini, Patrizia Marconi, Pasquale Narciso, Emanuele Nicastri, Pasquale Noto, Carlo Federico Perno (co-chair), Pietro Sette, Fabio Soldani, Maria Paola Trotta, Valerio Tozzi, Ilaria Uccella, Ubaldo Visco-Comandini, Mauro Zaccarelli, and Daniela Zinzi.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Department of Experimental Medicine, University of Rome Tor Vergata, Via Montpellier 1, 00133 Rome, Italy. Phone: 39-06-72596553. Fax: 39-06-72596039. E-mail: ceccherini{at}med.uniroma2.it Back

{triangledown} Published ahead of print on 8 August 2007. Back

{dagger} F.C.-S. and V.S. contributed equally to this work. Back

{ddagger} For the I.Co.N.A. Study Group (see Acknowledgments for a list of the members of this study group). Back

§ For the INMI-Collaborative Group for Clinical Use of HIV Genotype Resistance Test (see Acknowledgments for a list of the members of this study group). Back


arrow
REFERENCES
 
    1
  1. Bacheler, L. T., E. D. Anton, P. Kudish, D. Baker, J. Bunville, K. Krakowski, L. Bolling, M. Aujay, X. V. Wang, D. Ellis, M. F. Becker, A. L. Lasut, H. J. George, D. R. Spalding, G. Hollis, and K. Abremski. 2000. Human immunodeficiency virus type 1 mutations selected in patients failing efavirenz combination therapy. Antimicrob. Agents Chemother. 44:2475-2484.[Abstract/Free Full Text]
    2. 2
  2. Balzarini, J. 2004. Current status of the non-nucleoside reverse transcriptase inhibitors of human immunodeficiency virus type 1. Curr. Top. Med. Chem. 4:921-944.[CrossRef][Medline]
    3. 3
  3. Beerenwinkel, N., M. Daumer, M. Oette, K. Korn, D. Hoffmann, R. Kaiser, T. Lengauer, J. Selbig, and H. Walter. 2003. Geno2pheno: estimating phenotypic drug resistance from HIV-1 genotypes. Nucleic Acids Res. 31:3850-3855.[Abstract/Free Full Text]
    4. 4
  4. Benjamini, Y., and Y. Hochberg. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B 57:289-300.
    5. 5
  5. Berendsen, H., J. Postma, W. Van Gunsteren, A. Di Nola, and J. Haak. 1984. Molecular dynamics with coupling to an external bath. J. Chem. Phys. 81:3584-3590.
    6. 6
  6. Berendsen, H., D. Van der Spoel, and R. Van Drunen. 1995. GROMACS: a message-passing parallel molecular dynamics implementation. Comp. Phys. Comm. 91:43-56.[CrossRef]
    7. 7
  7. Boyer, P. L., H. Q. Gao, and S. H. Hughes. 1998. A mutation at position 190 of human immunodeficiency virus type 1 reverse transcriptase interacts with mutations at positions 74 and 75 via the template primer. Antimicrob. Agents Chemother. 42:447-452.[Abstract/Free Full Text]
    8. 8
  8. Brown, A. J., H. M. Precious, J. M. Whitcomb, J. K. Wong, M. Quigg, W. Huang, E. S. Daar, R. T. D'Aquila, P. H. Keiser, E. Connick, N. S. Hellmann, C. J. Petropoulos, D. D. Richman, and S. Little. 2000. Reduced susceptibility of human immunodeficiency virus type 1 (HIV-1) from patients with primary HIV infection to nonnucleoside reverse transcriptase inhibitors is associated with variation at novel amino acid sites. J. Virol. 74:10269-10273.[Abstract/Free Full Text]
    9. 8
  9. Camacho, R., K. Theys, A. Abecasis, K. Deforche, A. P. Carvalho, J. Cabanas, P. Gomes, and A. M. Vandamme. 2006. XV Int. HIV Drug Resist. Workshop Basic Princ. Clin. Implications, abstr. 148.
    10. 9
  10. Ceccherini-Silberstein, F., F. Erba, F. Gago, A. Bertoli, F. Forbici, M. C. Bellocchi, C. Gori, R. D'Arrigo, L. Marcon, C. Balotta, A. Antinori, A. D'Arminio Monforte, and C. F. Perno. 2004. Identification of the minimal conserved structure of HIV-1 protease in the presence or absence of drug pressure. AIDS 18:11-19.[CrossRef]
    11. 10
  11. Ceccherini-Silberstein, F., F. Gago, M. Santoro, C. Gori, V. Svicher, F. Rodríguez-Barrios, R. D'Arrigo, M. Ciccozzi, A. Bertoli, A. D'Arminio Monforte, J. Balzarini, A. Antinori, and C. F. Perno. 2005. High sequence conservation of human immunodeficiency virus type 1 reverse transcriptase under drug pressure despite the continuous appearance of mutations. J. Virol. 79:10718-10729.[Abstract/Free Full Text]
    12. 11
  12. Chen, L., A. Perlina, and C. J. Lee. 2004. Positive selection detection in 40,000 human immunodeficiency virus (HIV) type 1 sequences automatically identifies drug resistance and positive fitness mutations in HIV protease and reverse transcriptase. J. Virol. 78:3722-3732.[Abstract/Free Full Text]
    13. 12
  13. Clark, S. A., N. S. Shulman, R. J. Bosch, and J. W. Mellors. 2006. Reverse transcriptase mutations 118I, 208Y, and 215Y cause HIV-1 hypersusceptibility to non-nucleoside reverse transcriptase inhibitors. AIDS 20:981-984.[Medline]
    14. 13
  14. Clavel, F., and A. J. Hance. 2004. HIV drug resistance. N. Engl. J. Med. 350:1023-1035.[Free Full Text]
    15. 14
  15. Darden, T., D. York, and L. Pedersen. 1993. Particle mesh Ewald. J. Chem. Phys. 98:10089-10092.[CrossRef]
    16. 15
  16. De Clercq, E. 2004. Antivirals and antiviral strategies. Nat. Rev. Microbiol. 2:704-720.[CrossRef][Medline]
    17. 16
  17. De Groot, A. S., B. M. Jesdale, E. Szu, J. R. Schafer, R. M. Chicz, and G. Deocampo. 1997. An interactive Web site providing major histocompatibility ligand predictions: application to HIV research. AIDS Res. Hum. Retrovir. 13:529-531.[Medline]
    18. 17
  18. Geitmann, M., T. Unge, and U. H. Danielson. 2006. Interaction kinetic characterization of HIV-1 reverse transcriptase non-nucleoside inhibitor resistance. J. Med. Chem. 49:2375-2387.[CrossRef][Medline]
    19. 18
  19. Gonzales, M. J., T. D. Wu, J. Taylor, I. Belitskaya, R. Kantor, D. Israelski, S. Chou, A. R. Zolopa, W. J. Fessel, and R. W. Shafer. 2003. Extended spectrum of HIV-1 reverse transcriptase mutations in patients receiving multiple nucleoside analog inhibitors. AIDS 17:791-799.[CrossRef][Medline]
    20. 19
  20. Götte, M., X. Li, and M. A. Wainberg. 1999. HIV-1 reverse transcription: a brief overview focused on structure-function relationships among molecules involved in initiation of the reaction. Arch. Biochem. Biophys. 365:199-210.[CrossRef][Medline]
    21. 20
  21. Hess, B., H. Bekker, H. Berendsen, and J. Fraaije. 1997. A linear constraint solver for molecular simulations. J. Comp. Chem. 18:1463-1472.[CrossRef]
    22. 21
  22. Hope, A. C. 1968. A simplified Monte Carlo significance test procedure. J. R. Stat. Soc. Ser. B 30:582-598.
    23. 22
  23. Hsiou, Y., J. Ding, K. Das, A. D. Clark, Jr., S. H. Hughes, and E. Arnold. 1996. Structure of unliganded HIV-1 reverse transcriptase at 2.7 A resolution: implications of conformational changes for polymerization and inhibition mechanisms. Structure 4:853-860.[Medline]
    24. 23
  24. Hsiou, Y., J. Ding, K. Das, A. D. Clark, P. L. Boyer, P. Lewi, P. A. Janssen, J. P. Kleim, M. Rosner, S. H. Hughes, and E. Arnold. 2001. The Lys103Asn mutation of HIV-1 RT: a novel mechanism of drug resistance. J. Mol. Biol. 309:437-445.[CrossRef][Medline]
    25. 24
  25. Johnson, V. A., F. Brun-Vezinet, B. Clotet, D. R. Kuritzkes, D. Pillay, J. M. Schapiro, and A. D. D. Richman. 2006. Update of the drug resistance mutations in HIV-1: fall 2006. Top. HIV Med. 14:125-130.[Medline]
    26. 25
  26. Kleim, J. P., M. Rosner, I. Winkler, A. Paessens, R. Kirsh, Y. Hsiou, E. Arnold, and G. Reiss. 1996. Selective pressure of a quinoxaline nonnucleoside inhibitor of human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) on HIV-1 replication results in the emergence of nucleoside RT-inhibitor-specific (RT Leu-74->Val or Ile and Val-75->Leu or Ile) HIV-1 mutants. Proc. Natl. Acad. Sci. USA 93:34-38.[Abstract/Free Full Text]
    27. 26
  27. Koval, C. E., C. Dykes, J. Wang, and L. M. Demeter. 2006. Relative replication fitness of efavirenz-resistant mutants of HIV-1: correlation with frequency during clinical therapy and evidence of compensation for the reduced fitness of K103N + L100I by the nucleoside resistance mutation L74V. Virology 353:184-192.[CrossRef][Medline]
    28. 27
  28. Larder, B. A. 1994. Interactions between drug resistance mutations in human immunodeficiency virus type 1 reverse transcriptase. J. Gen. Virol. 75:951-957.[Abstract/Free Full Text]
    29. 28
  29. Lebowitz, J., S. Kar, E. Braswell, S. McPherson, and D. L. Richard. 1994. Human immunodeficiency virus-1 reverse transcriptase heterodimer stability. Protein Sci. 3:1374-1382.[Medline]
    30. 29
  30. Moore, C. B., M. John, I. R. James, F. T. Christiansen, C. S. Witt, and S. A. Mallal. 2002. Evidence of HIV-1 adaptation to HLA-restricted immune responses at a population level. Science 296:1439-1443.[Abstract/Free Full Text]
    31. 30
  31. Oostenbrink, C., A. Villa, A. Marck, and W. Van Gunsteren. 2004. A biomolecular force field based on the free enthalpy of hydration and solvation: the GROMOS force-field parameter sets 53A5 and 53A6. J. Comp. Chem. 25:1656.[CrossRef]
    32. 31
  32. Parniak, M. A., and N. Sluis-Cremer. 2000. Inhibitors of HIV-1 reverse transcriptase. Adv. Pharmacol. 49:67-109.[Medline]
    33. 32
  33. Perno, C. F., A. Cozzi-Lepri, C. Balotta, F. Forbici, M. Violin, A. Bertoli, G. Facchi, P. Pezzati, G. Cadeo, G. Tositti, S. Pasquinucci, S. Pauluzzi, A. Scalzini, B. Salassa, A. Vincenti, A. N. Phillips, F. Dianzani, A. Apice, G. Angarano, L. Monno, G. Ippolito, M. Moroni, and A. D'Arminio Manforte. 2001. Secondary mutations in the protease region of human immunodeficiency virus and virological failure in drug-naïve patients treated with protease inhibitor based therapy. J. Infect. Dis. 184:983-991.[CrossRef][Medline]
    34. 33
  34. Ren, J., C. Nichols, L. Bird, P. Chamberlein, K. Weaver, S. Short, D. I. Stuart, and D. K. Stammers. 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. 312:795-802.[CrossRef][Medline]
    35. 34
  35. Rhee, S. Y., W. J. Fessel, A. R. Zolopa, L. Hurley, T. Liu, J. Taylor, D. P. Nugyen, S. Slome, D. Klein, M. Horberg, J. Flamm, S. Follansbee, J. M. Schapiro, and R. W. Shafer. 2005. HIV-1 protease and reverse-transcriptase mutations: correlations with antiretroviral therapy in subtype B isolates and implications for drug-resistance surveillance. J. Infect. Dis. 192:456-465.[CrossRef][Medline]
    36. 35
  36. Rhee, S. Y., J. Taylor, G. Wadhera, A. Ben-Hur, D. L. Brutlag, and R. W. Shafer. 2006. Genotypic predictors of human immunodeficiency virus type 1 drug resistance. Proc. Natl. Acad. Sci. USA 103:17355-17360.[Abstract/Free Full Text]
    37. 36
  37. Saracino, A., L. Monno, L. Scudeller, D. C. Cibelli, A. Tartaglia, G. Punzi, C. Torti, S. Lo Caputo, S. Mazzotta, G. Scotto, G. Carosi, and G. Angarano. 2006. Impact of unreported HIV-1 reverse transcriptase mutations on phenotypic resistance to nucleoside and non-nucleoside inhibitors. J. Med. Virol. 78:9-17.[CrossRef][Medline]
    38. 37
  38. Schinazi, R. F., B. A. Larder, and J. Mellors. 2000. Mutations in retroviral genes associated to drug resistance: update 2000-2001. Int. Antivir. News 8:65-91.
    39. 38
  39. Selmi, B., J. Deval, K. Alvarez, J. Boretto, S. Sarfati, C. Guerreiro, and B. Canard. 2003. The Y181C substitution in 3'-azido-3'-deoxythymidine-resistant human immunodeficiency virus, type 1, reverse transcriptase suppresses the ATP-mediated repair of the 3'-azido-3'-deoxythymidine 5'-monophosphate-terminated primer. J. Biol. Chem. 278:40464-40472.[Abstract/Free Full Text]
    40. 39
  40. Shafer, R. W. 2002. Genotypic testing for human immunodeficiency virus type 1 drug resistance. Clin. Microbiol. Rev. 15:247-277.[Abstract/Free Full Text]
    41. 40
  41. Shulman, N. S., R. J. Bosch, J. W. Mellors, M. A. Albrecht, and D. A. Katzenstein. 2004. Genetic correlates of efavirenz hypersusceptibility. AIDS 18:1781-1785.[CrossRef][Medline]
    42. 41
  42. Shulman, N. S., J. Delgado, R. J. Bosch, M. A. Winters, E. Johnston, R. W. Shafer, D. A. Katzenstein, and T. C. Merigan. 2005. Nonnucleoside reverse transcriptase inhibitor phenotypic hypersusceptibility can be demonstrated in different assays. J. Acquir. Immune Defic. Syndr. 39:78-81.[CrossRef][Medline]
    43. 42
  43. Sing, T., V. Svicher, N. Beerenwinkel, F. Ceccherini-Silberstein, M. Daumer, R. Kaiser, H. Walter, K. Korn, D. Hoffmann, and M. Oette. 2005. Characterization of novel HIV drug resistance mutations using clustering, multidimensional scaling and SVM-based feature ranking, p. 285-296. In A. Jorge, L. Torgo, P. Brazdil, R. Camacho, and J. Gama (ed.), Knowledge discovery in databases: PKDD. Springer, New York, NY.
    44. 43
  44. Svicher, V., T. Sing, M. M. Santoro, F. Forbici, F. Rodríguez-Barrios, A. Bertoli, N. Beerenwinkel, M. C. Bellocchi, F. Gago, A. D'Arminio Manforte, and C. F. Perno. 2006. Involvement of novel human immunodeficiency virus type 1 reverse transcriptase mutations in the regulation of resistance to nucleoside inhibitors. J. Virol. 80:7186-7198.[Abstract/Free Full Text]
    45. 44
  45. Telesnitsky, A., and S. P. Goff. 1997. Reverse transcriptase and the generation of retroviral DNA, p. 121-160. In J. M. Coffin, S. H. Hughes, and H. E. Varmus (ed.), Retroviruses. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
    46. 45
  46. Tozzi, V., M. Zaccarelli, P. Narciso, M. P. Trotta, F. Ceccherini-Silberstein, P. De Longis, G. D'Offizi, F. Forbici, R. D'Arrigo, E. Boumis, and C. F. Perno. 2004. Mutations in HIV-1 reverse transcriptase potentially associated with hypersusceptibility to nonnucleoside reverse-transcriptase inhibitors: effect on response to efavirenz-based therapy in an urban observational cohort. J. Infect. Dis. 189:1688-1695.[CrossRef][Medline]
    47. 46
  47. Van Laethem, K., M. Witvrouw, C. Pannecouque, B. Van Remoortel, J. C. Schmit, R. Esnouf, J. P. Kleim, J. Balzarini, J. Desmyter, E. De Clercq, and A. M. Vandamme. 2001. Mutations in the non-nucleoside binding-pocket interfere with the multi-nucleoside resistance phenotype. AIDS 15:553-561.[CrossRef][Medline]
    48. 46
  48. Vavro, C., A. Florence, D. Irlbeck, B. Wine, and M. St. Clair. 2004. Abstr. 11th Conf. Retrovir. Opport. Infect., abstr. 658.
    49. 47
  49. Vingerhoets, J., H. Azijn, E. Fransen, I. De Baere, L. Smeulders, D. Jochmans, K. Andries, R. Pauwels, and M. P. De Bethune. 2005. TMC125 displays a high genetic barrier to the development of resistance: evidence from in vitro selection experiments. J. Virol. 79:12773-12782.[Abstract/Free Full Text]
    50. 48
  50. Wainberg, M. A. 2004. HIV-1 subtype distribution and the problem of drug resistance. AIDS 18:63-68.[CrossRef]
    51. 49
  51. Wallace, A. C., R. A. Laskowski, and J. Thornton. 1995. LIGPLOT: a program to generate schematic diagrams of protein-ligand interactions. Protein Eng. 8:127-134.[Abstract/Free Full Text]
    52. 50
  52. Walter, H., B. Schmidt, K. Korn, A. M. Vandamme, T. Harrer, and K. Uberla. 1999. Rapid, phenotypic HIV-1 drug sensitivity assay for protease and reverse transcriptase inhibitors. J. Clin. Virol. 13:71-80.[CrossRef][Medline]
    53. 51
  53. Wu, T. D., C. A. Schiffer, M. J. Gonzales, J. Taylor, R. Kantor, S. Chou, D. Israelski, A. R. Zolopa, W. J. Fessel, and R. W. Shafer. 2003. Mutation patterns and structural correlates in human immunodeficiency virus type 1 protease following different protease inhibitor treatments. J. Virol. 77:4836-4847.[Abstract/Free Full Text]

Journal of Virology, October 2007, p. 11507-11519, Vol. 81, No. 20
0022-538X/07/$08.00+0     doi:10.1128/JVI.00303-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Garriga, C., Perez-Elias, M. J., Delgado, R., Ruiz, L., Perez-Alvarez, L., Pumarola, T., Lopez-Lirola, A., Gonzalez-Garcia, J., Menendez-Arias, L., on behalf of the Spanish Group for the Study of An, (2009). HIV-1 reverse transcriptase thumb subdomain polymorphisms associated with virological failure to nucleoside drug combinations. J Antimicrob Chemother 64: 251-258 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ceccherini-Silberstein, F.
Right arrow Articles by Perno, C. F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ceccherini-Silberstein, F.
Right arrow Articles by Perno, C. F.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»