Variations in Reverse Transcriptase and RNase H Domain Mutations in Human Immunodeficiency Virus Type 1 Clinical Isolates Are Associated with Divergent Phenotypic Resistance to Zidovudine

Abstract

Mutations in the RNase H domain of human immunodeficiency virus type 1 RT have been reported to cause resistance to zidovudine (ZDV) in vitro. However, very limited data on the in vivo relevance of these mutations in patients exist to date. This study was designed to determine the relationship between mutations in the RNase H domain and viral susceptibility to nucleoside analogues. Viruses harboring complex thymidine analogue mutation (TAM) and nucleoside analogue mutation (NAM) profiles were evaluated for their phenotypic susceptibilities to ZDV, tenofovir (TNF), and the nonapproved nucleoside reverse transcriptase inhibitors (NRTIs) β-2′,3′-didehydro-2′,3′-dideoxy-5-fluorocytidine (Reverset), β-d-5-fluorodioxolane-cytosine, and apricitabine. As controls, viruses from NRTI-naïve patients were also studied. The pol RT region (codons 21 to 250) of the viruses were sequenced and evaluated for mutations in the RNase H domain (codons 441 to 560) and the connection domain (codons 289 to 400). The results showed that viruses from patients failing multiple NRTI-containing regimens had distinct TAM and NAM profiles that conferred various degrees of resistance to ZDV (0.9- to >300-fold). Sequencing of the RNase H domain identified five positions (positions 460,468, 483, 512, and 519) at which extensive amino acid polymorphisms common in both wild-type viruses and viruses from treated patients were identified. No mutations were observed at positions 539 and 549, which have previously been associated with ZDV resistance. Mutations in the RNase H domain did not appear to correlate with the levels of phenotypic resistance to ZDV. Although some mutations were also observed in the connection domain, the simultaneous presence of the L74V and M184V mutations was the most significant determinant of phenotypic resistance to ZDV in patients infected with viruses with TAMs.

Nucleoside reverse transcriptase inhibitor (NRTI) resistance mutations, also known as nucleoside analogue-associated mutations (NAMs), are associated with resistance to several NRTIs. The M41L, D67N, K70R, L210W, T215Y/F, and K219Q/E mutations are known as thymidine analogue mutations (TAMs). TAMs are a subset of NAMs that are selected by the thymidine analogues zidovudine (ZDV) and stavudine and that are associated with cross-resistance to most currently approved NRTIs (4, 12, 14-16). Furthermore, some NRTI resistance mutations have been preferentially associated with certain TAM profiles (20). TAMs commonly arise by two different pathways in patients receiving therapy with thymidine analogues. The first pathway (TAM-1) includes M41L, L210W, and T215Y; and the second pathway (TAM-2) includes D67N, K70R, T215F, and K219Q/E (11, 20). However, this distinction is not absolute because the pathways can overlap (19).

RNase H (amino acids 441 to 560), localized to the C terminus of the p66 subunit of reverse transcriptase (RT), performs the obligate step of degrading the RNA template after it is transcribed into DNA, permitting formation of the cDNA strand (5, 13). It has been reported in vitro that some substitutions in the RNase H domain (Q509L, H539N, and D549N) of human immunodeficiency virus (HIV) type 1 (HIV-1) might also confer a high level of resistance to ZDV and stavudine (3a, 23a, 24-26). It has also been shown that viruses from RNase H domains derived from treatment-experienced patients have increased resistance to ZDV (22). Furthermore, RNase H domain mutations at positions 469, 470, 554, and 558 have been shown to be more frequent in viruses from previously treated patients than treatment-naïve patients and have been associated with the presence of TAMs (27). Both increased and decreased RNase H activities following the use of nonnucleoside reverse transcriptase inhibitors (NNRTIs), which bind to RNase H domains, have also been described (10a). In a recent study, mutations in the connection domain (amino acids 289 to 423) were shown to significantly augment ZDV resistance in patients (23). However, very limited data that define the in vivo relevance of RNase H domain mutations on ZDV resistance exist to date.

In this study we determined the relationship between mutations in the RNase H domain and viral susceptibility to NRTIs in viruses with complex TAM and NAM profiles. This was done by evaluating the impacts of mutations in the connection and RNase H domains of RT on viral susceptibility to NRTIs by using both wild-type viruses and viruses that contained TAMs and/or NAMs. We also evaluated the effects of the L74V and/or M184V mutations in patients with viruses harboring complex TAM and NAM profiles on susceptibility to apricitabine (ATC).

(The research performed by Michel Ntemgwa was in partial fulfillment of the Ph.D., Faculty of Graduate Studies and Research, McGill University, Montreal, Quebec, Canada.)

MATERIALS AND METHODS

Patients and virus isolates.

Twenty-one heavily treated NRTI-experienced HIV-1-infected patients harboring viruses with complex TAM and NAM profiles were selected for this study. As controls, 21 NRTI-naïve patients, including 17 patients infected with wild-type virus and 4 patients infected with viruses with NNRTI resistance-conferring mutations, including K103N, V106A, and Y181C, were also included. Clinical isolates were obtained with ethics approval from patients at our clinics in Montreal, Quebec, Canada; and those from patients beginning with the designation ITA were from the Division of Infectious Diseases, L. Sacco Hospital, Milan, Italy, also with ethics approval.

Generation of recombinant clones.

The human MT-4 lymphoblastoid T-cell line was used in these experiments. RT from the infectious clone pNL4-3 (1) was used to generate variants of HIV-1 containing the L74V, M184V, and L74V-M184V mutations by site-directed mutagenesis, as described previously (7).

Drugs.

We obtained ZDV and tenofovir (TNF) as gifts from GlaxoSmithKline (Research Triangle Park, NC) and Gilead (Foster City, CA), respectively. ATC was kindly provided by Avexa Ltd. (Richmond, VIC, Australia). β-2′,3′-didehydro-2′,3′-dideoxy-5-fluorocytidine (Reverset) and β-d-5-fluorodioxolane-cytosine (d-FDOC) were kindly provided by R. F. Schinazi of Emory University (VA Medical Center, Decatur, GA).

d-FDOC is a nucleoside analogue that is currently in clinical development and that is an effective inhibitor of HIV-1. In vitro, molecular infectious clones containing M184V or M41L/D67N/K70R/T215Y/K219Q are as susceptible as wild-type virus (pNL4-3) to d-FDOC (30a).

Reverset is a cytidine analogue that has been shown in tissue culture to retain activity against HIV isolates with TAMs as well as the M184V mutation in RT, which is associated with resistance to lamivudine. Although its development has been discontinued due to side effects in patients, we retained it in our study because of its in vitro effectiveness against viruses with TAMs (30).

TNF and ZDV have disparate mutational pathways; none of the viruses from our patients with TAMs possessed the K65R mutation, which can sometimes be associated with resistance to TNF. It has been shown that patients harboring viruses with multiple TAMs show high-level resistance to TNF (36).

ATC is the (−) enantiomer of 2′-deoxy-3′-oxa-4′-thiocytidine (6, 31). ATC (also known as AVX754 and SPD754) is a new deoxycytidine analogue NRTI currently in clinical development for the treatment of NRTI-experienced HIV-1-infected patients. In vitro, ATC shows good activity against HIV-1 isolates harboring either the M184V or the L74V mutation and TAMs (3, 10).

Nucleic acid extraction, PCR amplification, and sequencing.

Viral RNA was extracted from plasma by using a QIAamp viral extraction kit (QIAGEN, Mississauga, Ontario, Canada). The RT gene (up to amino acid 250) was sequenced in both directions by using a Trugene automated DNA system (Siemens Inc., Toronto, Ontario, Canada), according to the manufacturer's instructions. For the amplification of regions of the RT gene that included the connection domain (amino acids 289 to 423), genotyping was performed by a published protocol (Virco, BVBA, Mechelen, Belgium), based on sequencing of a 1,497-bp fragment of the HIV-1 pol gene (positions 2253 to 3749) (2) encompassing up to 400 amino acids in RT. Despite the noninclusion of amino acids 401 to 423, the region studied included all mutations in the connection domain that have been implicated in ZDV resistance (23). We amplified and sequenced the RNase H domain of the viruses using standard procedures. RNA was reverse transcribed, and the RNase H domain (codons 441 to 560) was amplified by PCR with specific primer pairs, as described previously (27). Cycle sequencing of both strands was performed with a GeneAmp PCR 9700 instrument with a BigDye Terminator cycle sequencing kit (Applied Biosystems, Foster City, CA). The purified products were sequenced on an ABI Prism 3130 genetic analyzer (Applied Biosystems) with the same primers used for the nested PCR. The sequences were aligned with the strain HXB2 consensus sequence (GenBank accession number M15390) as a reference, and amino acid substitutions in the RNase H domain region were compared.

Phenotypic susceptibilities of viral isolates.

The drug susceptibilities of the viral isolates were measured by determining the extent to which select antiretrovirals inhibited in vitro HIV replication. Viruses or recombinant clones containing complex TAM and NAM profiles were phenotyped for their resistance to ZDV, RVT, d-FDOC, TNF, and ATC. Briefly, cord blood mononuclear cells (CBMCs) were infected over 2 h with wild-type or mutated viruses. The infected CBMCs were then plated on 96-well plates containing the drugs at various concentrations. After 3 days of incubation at 37°C under 5% CO₂, the cells were fed fresh medium containing appropriate drug dilutions. At day 7, RT enzyme assays were performed to determine the 50% effective concentrations (EC₅₀s) of these drugs by using Prism analytic software (GraphPad Inc.). The EC₅₀s obtained for mutated viruses were compared to the EC₅₀s determined for wild-type viruses in order to compute the fold resistance. The results are expressed as the mean EC₅₀ ± the standard deviation (18, 28).

Nucleotide sequence accession numbers.

The sequences used in this study have been deposited in the GenBank database as follows. The 48 RT sequences that included the connection domain can be found under accession numbers EU091155 to EU091202, respectively. The 42 sequences of the RNase H domain can be found under accession numbers EU091203 to EU091244, respectively.

RESULTS

Amino acid substitutions in RT and connection domain.

Table 1 shows the genotypes of the RT gene (codons 21 to 250) and the connection domain (codons 289 to 400) of all viruses studied. As shown, the viruses from 19 of 21 patients studied had complex TAM and NAM profiles involving mutations at position 41, 67, 70, 210, 215, or 219. The two patients (patients 5686 and ITA9) whose viruses did not have TAMs had other NAMs involving Q151M. The virus from one patient (patient 6169) had the Q151M mutation, in addition to TAMs. Of the four patients whose viruses exclusively had NNRTI resistance-conferring mutations, the viruses from three patients possessed K103N, while the viruses from two patients had Y181C and the virus from one patient had V106A. Patients who were not receiving any therapy had wild-type viruses (Table 1). There were 11 different mutational profiles among the viruses from the 19 patients with TAMs (data not shown). The most common profile (5 of 19 patients) involved M41L, D67K/N/G, L210W, T215Y, and K219R/Q. Many amino acid substitutions of unknown significance were observed in the connection domain. However, we did observe amino acid substitutions at five positions (positions 335, 348, 360, 365, and 376) that are associated with resistance to ZDV (Table 1).

TABLE 1.

Viral drug resistance mutational profiles

Virus group and patient no.	RT resistance-conferring mutation(s) (codons 21 to 250)	Connection domain amino acid substitutions (codons 289 to 400)a
Isolates with complex TAMb and NAM profiles
4205	D67N,b T69D, A98G, V118I, M184V, T215F, K219Q	V292I, Q334E, M357S, UA376TU, A400T
5403	M41L, D67N, K70R, L74V, L100I, K103N, V118I, M184V, T215Y, K219Q	I293V, V317A, I329V, Q334N, UV365IU, K390R, K395N
5555	M41L, D67N, M184V, T215Y, K219Q	E297K, K311R, S322T, I329V, G333E, T338S, G359S, A37IV
5652	M41L, D67N, T69D, K70R, L74V, K103N, V181C, M184V, H208Y, T215Y, K219Q	I293V, E297A, Y318F, I329V, A355G, R356K, G359S, A371V, K390R
5686	L74V, V75I, A98S, K103N, Y115F, Q151M, M184V, G190A	I293V, E297A, E370D, K390R
5705	M41L, A98G, K103N, M184V, T215Y	I293V, F346Y, M357T, UA360TU, Q394L, K395S, A400T
5865	M41L, E44A, T69D, A98G, V108I, V118I, Y181C, M184I, G190A, H208Y, L210W, T215Y, K219R	I293V, P294Q, E297R, I326IV, Q334E, P345Q, A355G, R356K, G359S, A371V, I375V, T377Q
5965	D67N, T69N, K70R, L74V, K101H, Y181C, M184V, G190A, T215Y, K219Q	P294T, D320DE, D324E, UN348IU, M357T, R358K, K366R, UA376NU, T386S, K390R
6049	M41L, E44D, D67N, T69D, A98S, K103N, V106M, V118I, L210W, T215Y, K219R	I293V, T296S, E297K, Q334L, UG335SU, R358K, K366R, T369V, A371V, I375V, UA376TU, T386AT, K390R, E399DE
6169	M41ML, D67N, K70R, L74V, K103S, F116Y, Q151M, M184V, G190S, L210FL, T215Y, K219Q	I293V, S322T, UN348IU, UA360VU, UA376TU, T386A
6517	M41L, E44A, D67N, L74I, K103N, V108IV, V118I, M184V, L210W, T215Y	I293V, E297K, Q334E, R356K, M357L, G359A, K366R, S379C, T386I, A400T
6824	M41L, E44A, D67N, V118I, K103N, V181C, M184V, L210W, T215Y, K219Q	V292I, E297K, I326V, G359A, K366R, S379C, T386I, A400T
6827	M41L, E44A, D67N, T69D, V75M, V106A, V108A, V118I, M184V, L210W, T215Y	I293V, E297R, L301I, Y318F, Q334L, R356K, UA360TU, E370D, T386I, K390R
6895	M41L, E44D, D67N, T69D, L74V, V118I, M184V, L210S, T215Y, K219R	I293V, E297K, G333E, Q334Y, M357S, I375V, T400A
6897	M41L, E44D, D67G, L74V, K101E, K103R, V118I, M184V, G190A, L210W, T215Y, K219Q	I293V, E297A, I326L, R358K, A371V, I375V, T386I, K390R
6971	M41L, D67N, T69D, K70R, A98S, K103N, M184V, T215F, K219Q	P294Q, Q334L, UN348IU, K385KR
ITA4	M41L, E44A, D67E, V75M, K103N, V118I, Y181C, L210W, T215Y	I293IV, M357T, G359S
ITA5	D67G, K70G, L74I, K101E, V106I, V181C, M184V, G190S, T215Y	P294A, E297R, M357T, UV365IVU, UA376TU, T386I
ITA7	M41L, D67KN, L74IV, K103N, V108I, V118I, Y181C, M184V, L210W, T215C, K219Q	P294KPQT, E297K, UG335SU, R356K, G359S, A371V, UA376STU, S379C, T386IT, K390R
ITA8	M41L, D67N, K101E, V118I, M184V, Y181C, G190A, L210W, T215Y	E291D, V292I, I293V, E297A, P321S, D324E, I329V, Q334L, UG335DU, F346Y, R356K, M357R, R358G, G359S, A371V, UA376SU, T386I, Q394R
ITA9	Q151M	I293V, P294Q, I326V, I329L, R356KR, K366R, T386V
Isolates with NNRTI resistance-conferring mutations only
4937	K103N	NDc
5002	K103N	ND
5004	K103N, Y181C	ND
NevRNIH	V106A,Y181C	ND
WTd	5269, 5310, 5322, 5323, 5326, 5328, 5329, 5331, 5332, 5346, 5512, 5785, 6316, 6682, 6718, 6853, 6879

^aOnly amino acids 289 to 400 of the connection domain were considered. The underlined amino acids in the connection domain are those at recently described positions (E312Q, G335C/D, N348I, A360I/V, V365I, and A376S) that augment ZDV resistance.

^bBoldface indicates TAMs at positions 41, 67, 70, 210, 215, and 219.

^cND, not determined.

^dWT, wild type.

Amino acid substitutions in the RNase H domain.

Viruses whose RT region (codons 21 to 250) was sequenced were evaluated for mutations in the RNase H domain. We sequenced the RNase H domain spanning amino acids 441 to 560. The 119 amino acids analyzed for each sequence were compared with the amino acids in a strain HXB2 consensus sequence. Amino acid substitutions were observed at 45 different positions (Fig. 1). Mutations in the RNase H domain were generally more frequent in viruses with complex TAM and NAM profiles than in the control group (185 and 145, respectively). Figure 1 shows that five regions contained extensive amino acid substitutions which were polymorphic in viruses from both treated and untreated patients. These included the amino acids at positions 460, 468, 483, 512, and 519. Mutations at positions 491 and 527 were more frequent in viruses with TAMs than in the control group (10/19 and 3/21, respectively, for position 491 [P = 0.02] and 8/19 and 3/21, respectively, for position 527 [P = 0.08]). Mutations at position 469 were more frequent in patients with viruses with TAMs than in patients with wild-type virus (10/19 and 2/17, respectively; P = 0.01) and were also present in two patients (patients 4937 and 5002) whose viruses had only NNRTI resistance-conferring mutations. We also observed that the Q524E mutation was exclusively found in viruses with complex TAM and NAM profiles.

FIG. 1.

Distribution of amino acid mutations and polymorphisms in the RNase H domain of wild-type viruses, viruses with NNRTI resistance-conferring mutations only, and viruses containing complex TAM and NAM profiles. Dots, identical amino acids; regions in gray, regions in which common polymorphisms were observed; +, <20-fold resistance to ZDV; +++, >50-fold resistance to ZDV.

No mutations were observed at positions 539 and 549, although these have been reported to augment ZDV resistance in vitro. We found two mutations at position 509 (Q509R and Q509K) exclusively in viruses with complex TAM and NAM profiles (Fig. 1). However, we did not find the Q509L mutation, which has been reported to be selected for in vitro. We further evaluated all RNase H domain sequence alignments obtained from the Los Alamos National Laboratory HIV database (n = 2,170), which included all available RNase H domain sequences from drug-naïve individuals (n = 458) (data not shown). We found only five sequences with mutations at position 539 and only one sequence that contained H539N. Of the other sequences, one contained H539R, while the others included H539P. The three sequences with H539P were from drug-naïve individuals. An evaluation of position 549 revealed amino acid substitutions at only seven positions, three of which were D549N. The other four substitutions were 549H (n = 1), 549A (n = 1), and 549G (n = 2). When we further evaluated drug-naïve patient virus RNase H domain sequences (n = 458) for substitutions at position 549, two sequences each had 549N and 549G, implying that only one virus from a treated patient in the entire database possessed the 549N substitution. In regard to the Q509L mutation, the frequency among viruses from treated patients (n = 1,712) was 0.2% (n = 3).

Phenotypic susceptibilities of virus isolates to NRTIs.

Table 2 shows the phenotypic susceptibilities of viruses containing multiple TAMs and NAMs to NRTIs. The mean EC₅₀s for wild-type viruses were 0.469 ± 0.023 μM for RVT, 0.128 ± 0.012 μM for d-FDOC, 0.033 ± 0.021 μM for ZDV, 0.25 ± 0.02 for TNF, and 1.41 ± 0.072 μM for ATC. The viruses from four patients (patients 5865, 6517, 6624, and 6895) were susceptible to RVT, and except for the viruses from patients 5555, ITA8, and 5686, the viruses from the others showed moderate levels of resistance (3.0- to 8.0-fold); the viruses from patients 5555, ITA8, and 5686 showed between 10.7- and 16-fold resistance. The viruses from four patients (patients ITA4, ITA8, ITA9, and 6517) showed high levels of resistance to d-FDOC (16.5- to 39.1-fold). The viruses from most patients showed high levels of resistance to TNF (>15-fold for the viruses from most patients). In regard to ATC, the viruses from most patients showed low levels of resistance; the viruses from four patients (patients 4205, 5555, 6827, and ITA9) had high levels of resistance (between 13.3- and >50-fold). We observed various levels of resistance to ZDV for the different virus isolates (5- to >333-fold for the viruses from most patients). In regard to the viruses from patients with complex TAM and NAM profiles, we distinguished two groups: group +++ comprised the viruses from patients with >50-fold resistance to ZDV, while group + were those with <20-fold resistance to ZDV. Similar numbers of mutations were observed in the RNase H domains of viruses with high and low levels of ZDV phenotypic resistance (94 and 90 mutations, respectively) (Fig. 1). Of the viruses from five patients that possessed the Q524E mutation (Fig. 1), the viruses from two patients (patients 6049 and 6827) had high levels of phenotypic resistance to ZDV, while the viruses from three patients (patients ITA7, ITA9, and 6169) had low levels resistance to this drug (Table 2). No significant differences in the growth rates of the viruses tested compared with those of the wild-type viruses were observed.

TABLE 2.

Phenotypic susceptibilities of virus isolates to NRTIs^a

Level of resistance to ZDV (group)	Patient no.	RVT		d-FDOC		TNF		ZDV		ATC
		EC50b (μM)	FRc	EC50 (μM)	FR	EC50 (μM)	FR	EC50 (μM)	FR	EC50 (μM)	FR
≥50-fold (+++)	6049	NDd		ND		ND		>10	>333	12.19 ± 0.144	8.6
	5865	0.693 ± 0.396	1.5	0.161 ± 0.278	1.3	4 ± 0.440	16	>10	>333	2.711 ± 0.409	1.9
	ITA4	3.390 ± 0.482	7.2	3.0 ± 0.494	23	9.1 ± 0.140	36.5	>10	>333	3.705 ± 0.125	2.6
	6827	3.725 ± 0.371	8.0	0.142 ± 0.244	1.1	6.375 ± 0.130	25.5	>10	>333	0.847 ± 0.389	0.6
	ITA8	6.0 ± 0.450	12.8	5.0 ± 0.220	39.1	7.5 ± 1.240	30	8 ± 1.201	267	18.8 ± 0.192	13.3
	6517	0.562 ± 0.237	1.2	5.0 ± 0.810	39.1	19.75 ± 3.33	79	8 ± 0.824	267	<1.0	-
	5403	2 ± 0.210	4.2	0.312 ± 0.255	2.4	4.75 ± 1.240	19	6 ± 0.832	200	1.396 ± 0.241	1.0
	5555	5 ± 1.309	10.7	0.869 ± 0.214	6.8	12.5 ± 0.070	50	>10	>333	26.03 ± 0.16	18.5
	5686e	7.5 ± 0.543	16	5 ± 0.01	39	0.675 ± 0.120	2.7	>10	>333	2.541 ± 0.736	1.8
	6824	0.027 ± 0.157	0.1	0.668 ± 0.696	5.2	25 ± 3.330	100	>10	>333	3.344 ± 0.274	2.37
	4205	ND	ND	ND		ND		1.65 ± 0.340	50	50 ± 0.367	35
<20-fold (+)	6971	ND	ND	ND	ND	ND		0.561 ± 0.616	17	7.779 ± 0.499	5.5
	5705	ND	ND	ND	ND	ND		0.33 ± 0.020	10	0.408 ± 0.126	0.3
	6895	0.279 ± 0.435	0.6	0.496 ± 0.365	3.9	1.875 ± 0.302	7.5	0.524 ± 0.314	17.4	6.692 ± 0.804	4.7
	6897	ND	ND	ND	ND	0.044 ± 0.489	0.2	0.16 ± 0.129	5	4.424 ± 0.326	3.1
	5965	ND	ND	ND	ND	ND	ND	0.035 ± 0.400	1.0	2.5 ± 0.192	1.8
	ITA7	ND	ND	ND	ND	ND	ND	0.026 ± 0.002	0.9	4.64 ± 0.062	3.3
	5652	1.379 ± 0.408	3.0	0.351 ± 0.236	2.7	7.25 ± 2.01	29	0.217 ± 0.281	7.2	8.2 ± 0.062	5.8
	616e	2.404 ± 1.030	5.1	3.060 ± 1.210	24	0.867 ± 0.42	1.7	0.040 ± 0.010	1.2	ND	ND
	ITA5	1.43 ± 0.194	3.1	1.184 ± 0.225	9.3	2.957 ± 0.193	5.9	0.058 ± 0.136	1.8	5.98 ± 0.048	4.2
	ITA9e	2.572 ± 0.262	5.5	2.109 ± 0.330	16.5	0.967 ± 0.223	1.9	0.033 ± 0.429	1.0	>50	>35.5

^aThe mean EC₅₀s for wild-type isolates were 0.469 ± 0.023 μM for Reverset (RVT), 0.128 ± 0.012 μM for d-FDOC, 0.25 ± 0.02 μM for TNF, 0.033 ± 0.021 μM for ZDV, and 1.41 ± 0.072 μM for ATC.

^bThe values represent the means ± standard deviations of at least two independent experiments, each performed in duplicate.

^cFR, fold resistance.

^dND, not determined.

^eIsolates with the Q151M mutation.

Amino acid substitutions in the connection domain of RT and ZDV resistance.

We evaluated all viruses with complex TAM and NAM profiles and various levels of resistance to ZDV in order to determine the possible role of mutations in the connection domain on ZDV resistance (Table 3). We also took into consideration the L74V and M184V mutations, which have been reported to enhance ZDV susceptibility. We included the viruses from 10 patients with high levels of phenotypic resistance (>50-fold) and those from 9 patients with lower levels (<20-fold) of resistance to ZDV and TAMs. Of the six positions that are known to augment ZDV resistance (positions E312Q, G335C, N348I, A360I/V, V365I, and A376S), we identified amino acid substitutions at five of them (positions 335, 348, 360, 365, and 376). We did not find any amino acid substitution at position 312. As shown in Table 3, substitutions at positions 335, 360, 365, and 376 were found in the viruses from 6 of 10 patients with high levels of resistance to ZDV, while the viruses from 6 of 9 patients with lower levels of resistance to ZDV had amino acid substitutions at positions 335, 348, 360, 365, and 376. N348I was not found in viruses with high levels of resistance to ZDV. The viruses from some patients (seven of nine) that possessed L74IV and M184V had only low levels of resistance to ZDV, despite the presence of TAMs. This is probably due to the effects of L74V and M184V on the enhancement of the excision reaction, which can augment sensitivity to ZDV (8, 9, 21). The viruses from only two patients harboring TAMs (patients 5403 and 6517) and with high levels of resistance to ZDV had L74I/V and M184V. However, the virus from patient 5403 possessed the V365I mutation in the connection domain, and the virus from patient 6517 possessed L74I instead of L74V. Mutations in the RNase H domain, which were more frequent among the viruses from treated patients (positions 469, 491 and 527), did not appear to be associated with either low- or high-level resistance to ZDV (Table 3).

TABLE 3.

ZDV resistance in the context of TAMs, connection domain, RNase H domain, and L74V and M184V mutations

Patient no.	TAM profile	Connection domain mutations (positions 289 to 400)a	RNase H domain mutation(s)b	L74V/M184V	Fold resistance to ZDV
4205	D67N, T215F, K219Q	A376T		184V	50
5403	M41L, D67N, K70R, T215Y, K219Q	V365I	469I, 491S	74V, 184V	200
5555	M41L, D67N, T215Y, K219Q	None	469I	M184V	>300
5865	M41L, L210W, T215Y, K219R	None	527N	184I	>333
6049	M41L, D67N, L210W, T215Y, K219R	G335S, A376T		None	>333
6517	M41L, D67NK, L210W, T215Y	None	469I, 491S, 527N	74I, 184V	267
6824	M41L, D67N, L210W, T215Y, K219Q	None	469I	184V	>333
6827	M41L, D67N, L210W, T215Y	A360T	491S	184V	>333
ITA4	M41L, D67DE, L210W, T215Y	None	469I, 491S, 527N	None	>333
ITA8	M41L, D67N, L210W, T215Y	G335D, A376S	491T	184V	267
ITA5	D67G, K70G, T215Y	V365I/V, A376T	491T	74I, 184V	1.8
ITA7	M41L, D67N, L210W, 215CHRY, K219Q	G335S, A376ST	469I	74I/V, 184V	0.9
5652	M41L, D67N, K70R, T215Y, K219Q	None	469T, 491S, 527N	74V, 184V	7.2
5705	M41L, T215Y	A360T	527N	184V	10
5965	D67N, K70R, T215Y, K219Q	N348I, A376N	527N	74V, 184V	1
6169	M41LM, D67N, L210FL, T215FV, K219Q	N348I, A360V, A376T	469F	74V, 184V	1.2
6895	M41L, D67N, L210S, T215Y, K219R	None	491S, 527N	74V, 184V	17.4
6897	M41L, D67G, L210S, T215Y, K219Q	None	491P, 527N	74V, 184V	1
6971	M41L, D67N, T215F, K219Q	N348I	469I, 491S	184V	17

^aOnly amino acids in the connection domain at positions that were recently described (E312Q, G335C/D, N348I, A360I/V, V365I, and A376S) to augment ZDV resistance were considered.

^bOnly mutations in the RNase H domain at positions that were more frequent in treated patients (positions 469, 491, and 527) are indicated.

Phenotypes of mutated viruses for ZDV and ATC resistance.

In order to study the impact of the L74I/V and M184V mutations on drug susceptibility, we reevaluated the mutational profiles of the viruses with or without the double presence of the L74I/V and 184V mutations and the susceptibilities of these viruses to ZDV and ATC. The latter agent is a novel NRTI in clinical development and possesses good activity against HIV-1 isolates harboring the M184V mutation and TAMs. The results showed that viruses with multiple NRTI resistance-conferring mutations are more sensitive to ZDV in the presence of TAMs if L74I/V and M184V are also present; however, this effect was not seen for ATC (data not shown).

We further tested recombinant clones of viruses harboring either L74V, M184V, or for their susceptibilities to both ZDV and ATC (Fig. 2). The results show that L74V and M184V each conferred a degree of increased sensitivity to ZDV and the double mutation conferred a higher degree of susceptibility, as shown above (Table 3). The L74V mutation did not confer any resistance to ATC, while the M184V mutation conferred slight resistance to this drug. The double mutant did not show further susceptibility, as was seen in the case of ZDV (Fig. 2).

FIG. 2.

Comparative phenotypes of recombinant clones of viruses harboring the L74V and/or M184V mutation for susceptibilities to ATC and ZDV in CBMCs. Bars with dots, fold resistance to ZDV; bars with slashes, fold resistance to ATC; wt, wild type. Recombinant clones containing the L74V or M184V mutation, or both, were tested for their susceptibilities to ATC and ZDV in cell culture assays. The mean EC₅₀s of the mutants were compared to those of wild-type strain pNL4-3, which were 5.87 ± 0.29 and 0.07 ± 0.03 for ATC and ZDV, respectively. The results are shown as means ± standard deviations.

DISCUSSION

Mutations at positions 469, 470, 554, and 558 were previously shown to be more prevalent in treated patients than in treatment-naïve patients; and mutations at position 558 have been associated with the presence of TAMs (27). Furthermore, the L469T and K558R mutations in the RNase H domain have been shown to be statistically more common in viruses from treated patients than in those from drug-naïve patients (20a). Although we found many amino acid substitutions at position 469, the L469T mutation was found in the virus from only one treated patient, while the majority of isolates possessed L469I. In our study, the L469I mutation was more frequent in patient isolates with complex TAM and NAM profiles than in isolates from the control group (8/21 and 3/21, respectively); however this difference was not statistically significant (P = 0.079). We did not find any statistically significant correlation between the presence of the K558R mutation and the treatment status of the patients. The K558R mutation was found in the viruses from only two treated patients with complex TAM and NAM profiles, the isolate from one patient infected with wild-type virus, and the isolate from one patient with only NNRTI resistance-conferring mutations. Mutations at positions 470 and 554 were not more prevalent in patients with complex TAM and NAM profiles than in the control group: for position 470, 10/21 and 10/21 patients with viruses with complex TAM and NAM profiles and the control group, respectively; and for position 554, 1/21 and 3/21 patients with viruses with complex TAM and NAM profiles and the control group, respectively.

It has been shown in vitro that mutations in the HIV-1 RNase H domain confer high-level resistance to NRTIs and that the H539N mutation increased resistance to ZDV ninefold more than several TAMs did. Furthermore, two mutations in the RNase H domain, H539N and D549N, have been shown to augment TAM-mediated ZDV resistance in vitro (23a). However, no substitutions at position 539 or 549 were seen in our study, which evaluated isolates with complex TAM and NAM profiles. These two mutations were also absent in other studies that looked at viruses from a larger group of treatment-naïve and NRTI-experienced patients (27, 29, 32, 34). It is probable that these two mutations are preferentially selected in vitro but are difficult to select in vivo, possibly for reasons of viral fitness.

We evaluated the RNase H domains of all 2,170 sequences available from the Los Alamos National Laboratory HIV database: only five sequences had amino acid substitutions at position 539 (H539N, n = 1; H539R, n = 1; and H539P, n = 3), while seven had substitutions at position 549 (D549H, n = 1; D549N, n = 3; D549A, n = 1; D549G, n = 2). Further analysis of sequences that excluded viruses from drug-naïve patients showed that the H539N and D549N mutations were present in the virus from only one treated patient (n = 1,712). Thus, H539N and D549N may augment ZDV resistance in vitro but are rare in treated patients.

The Q509L mutation is associated with other NRTI resistance-conferring mutations (67N, 70R, T215I/F, A371V) after in vitro passage in the presence of ZDV (3a); however, none of the patient isolates with complex TAM and NAM profiles in our study had the Q509L mutation. The frequency of the Q509L mutation in viruses from treated patients (n = 1,712) from the Los Alamos National Laboratory HIV database is only 0.2% (n = 3). Indeed, the viruses from two patients (patients 5686 and 6895) possessed either Q509R or Q509K mutations, but not in association with 67N, 70R, T215I/F, or A371. The last mutation, which is in the connection domain, was found in equal numbers in viruses from patients with high- and low-level resistance to ZDV (three patients in each group).

Our study also shows that mutations at positions 469, 491, and 527 were more frequent in isolates from patients infected with viruses with complex TAM and NAM profiles than in patients infected with wild-type virus. We also found that the Q524E mutation was present exclusively in patients infected with viruses with complex TAM and NAM profiles but not in patients infected with wild-type virus. This led us to predict that Q524E may be associated with TAMs and resistance to ZDV. However, evaluation of the 458 RNase H domain sequences from drug-naïve patients from the Los Alamos National Laboratory HIV database revealed that 76 (16.6%) sequences possessed this amino acid substitution, suggesting that it is likely a polymorphism. Although a variety of studies have reported an association between mutations in RT with those in RNase H in vitro (3a, 23a), none until now has evaluated patient viruses with complex TAM and NAM profiles. Databases, such as that of Stanford, contain phenotyping data but lack sequences of the RNase H domain. Our study suggests that RNase H mutations do not affect the levels of resistance to ZDV in patients infected with viruses with multiple TAMs and NAMs.

The high levels of ZDV resistance observed in some of the isolates from our patients are consistent with the results of other studies that have evaluated isolates harboring TAMs (4, 14, 15, 33, 35). Similar numbers of mutations were observed in the RNase H domains of viruses with high- and low-level resistance to ZDV. This suggests that the levels of phenotypic resistance to ZDV may be mediated in part by mechanisms that are not well understood and that do not necessarily involve mutations only in the RNase H domain.

Viruses from 5 of 11 patients with high-level resistance to ZDV had amino acid substitutions in the connection domain at positions known to augment ZDV resistance (positions 335, 360, 365, and 376), while viruses from 6 of 10 patients with low-level resistance to ZDV possessed substitutions at positions 348, 360, 365, and 376. In our cohort of patients in Montreal whose viruses were genotyped (n = 1,361), the frequency of the N348I mutation is 7.2% (n = 98). The isolates from 75.6% of these patients harbored the N348I mutation and also possessed M184V, indicating the increasing coassociation of these substitutions in viruses from treated individuals.

This is consistent with recent findings showing that N348IT is commonly selected in vivo, confers decreased susceptibility to ZDV, and is highly associated with the M184V mutation (37).

Viruses from seven of nine patients with low-level resistance to ZDV, despite the presence of TAMs and mutations in the connection domain, also possessed both the L74I/V and the M184V mutations, which are known to increase the level of sensitivity to ZDV (8, 9, 21). This suggests that the lower levels of ZDV resistance in these patients may be attributable to the hypersensitizing effects of L74I/V and M184V. The isolates from only two patients with TAMs (patients 5403 and 6517) and L74V and M184V mutations had high levels of resistance to ZDV. The virus from patient 5403 also possessed the V365I mutation in the connection domain, which is known to augment ZDV resistance. The virus from patient 6517 possessed L74I instead of L74V; L74I may have less of a sensitizing effect on ZDV than L74V. The high levels of resistance to ZDV in patient 5686, despite the presence of L74V and M184V, may have been due to the Q151M mutation (17).

These results were not reproduced when viruses harboring TAMs as well as both the L74I/V and the M184V mutations were evaluated for their susceptibilities to ATC, further suggesting that both of these mutations have the ability to hypersensitize isolates only to ZDV. Furthermore, we determined the levels of susceptibility to ATC of recombinant HIV-1 clones containing the L74V and/or M184V mutation in cell culture. The double presence of L74V and M184V did not result in increased susceptibility to ATC, probably because M184V only moderately compromises the responsiveness to ATC (3). Our results are also consistent with other data that showed that L74V plus M184V confer increased susceptibility to ZDV and resistance to most other NRTIs (19).

The introduction of these specific mutations in the context of an otherwise wild-type genotype (strain pNL4-3) permits the correlation of these observed phenotypes with the clinical isolates tested. Considering that the backbone used is considered “laboratory adapted” and relatively genetically “pure” compared with the backbones of the clinical isolates used, we cannot conclude with absolute certainty that the L74V and M184V mutations are exclusively responsible for the observed increase in ZDV sensitivity. We can, however, state that a strong association exists between ZDV resensitization and these clinically relevant mutations.

The patients who were evaluated for the collection of data for the Los Alamos National Laboratory HIV database had received antiretroviral therapy. Although the drugs used were not specified, it is likely that almost all of these individuals received NRTIs. The data available at this moment suggest that mutations in the RNase H domain have been shown to augment resistance to ZDV in vitro and are infrequent in isolates from patients exposed to NRTIs. We also analyzed 21 patients who failed treatment with NRTIs and whose isolates had multiple TAMs and NAMs. In all cases, mutations in the RNase H domain that have been reported to be associated with ZDV resistance in vitro were rare. While an analysis of the RNase H region in isolates from a larger group of patients might have yielded different results, our results do not support the inclusion of the RNase H domain as a component of routine drug resistance genotyping assays at this time. Although the connection domain may have some relevance to genotyping, care in regard to the interpretation of the results is advised, particularly if the L74V and M184V mutations are present.