Biochemical and Structural Properties of Entecavir-Resistant Hepatitis B Virus Polymerase with L180M/M204V Mutations

ABSTRACT

Entecavir (ETV) is a widely used anti-hepatitis B virus (HBV) drug. However, the emergence of resistant mutations in HBV reverse transcriptase (RT) results in treatment failure. To understand the mechanism underlying the development of ETV resistance by HBV RT, we analyzed the L180M, M204V, and L180M/M204V mutants using a combination of biochemical and structural techniques. ETV-triphosphate (ETV-TP) exhibited competitive inhibition with dGTP in both wild-type (wt) RT and M204V RT, as observed using Lineweaver-Burk plots. In contrast, RT L180M or L180M/M204V did not fit either competitive, uncompetitive, noncompetitive, or typical mixed inhibition, although ETV-TP was a competitive inhibitor of dGTP. Crystallography of HIV RT^{Y115F/F116Y/Q151M/F160M/M184V}, mimicking HBV RT L180M/M204V, showed that the F115 bulge (F88 in HBV RT) caused by the F160M mutation induced deviated binding of dCTP from its normal tight binding position. Modeling of ETV-TP on the deviated dCTP indicated that a steric clash could occur between ETV-TP methylene and the 3′-end nucleoside ribose. ETV-TP is likely to interact primarily with HBV RT M171 prior to final accommodation at the deoxynucleoside triphosphate (dNTP) binding site (Y. Yasutake, S. Hattori, H. Hayashi, K. Matsuda, et al., Sci Rep 8:1624, 2018, https://doi.org/10.1038/s41598-018-19602-9). Therefore, in HBV RT L180M/M204V, ETV-TP may be stuck at M171, a residue that is conserved in almost all HBV isolates, leading to the strange inhibition pattern observed in the kinetic analysis. Collectively, our results provide novel insights into the mechanism of ETV resistance of HBV RT caused by L180M and M204V mutations.

IMPORTANCE HBV infects 257 million people in the world, who suffer from elevated risks of liver cirrhosis and cancer. ETV is one of the most potent anti-HBV drugs, and ETV resistance mutations in HBV RT have been extensively studied. Nevertheless, the mechanisms underlying ETV resistance have remained elusive. We propose an attractive hypothesis to explain ETV resistance and effectiveness using a combination of kinetic and structural analyses. ETV is likely to have an additional interaction site, M171, beside the dNTP pocket of HBV RT; this finding indicates that nucleos(t)ide analogues (NAs) recognizing multiple interaction sites within RT may effectively inhibit the enzyme. Modification of ETV may render it more effective and enable the rational design of efficient NA inhibitors.

INTRODUCTION

Hepatitis B virus (HBV) chronically infects 257 million people (1), who are at increased risk of developing liver cirrhosis and hepatocellular carcinoma (2). Current anti-HBV treatments include nucleos(t)ide analogues (NAs), interferons (IFNs), or both (3, 4). NAs, including lamivudine (LMV) (cytosine analogue), entecavir (ETV) (guanosine analogue), adefovir (ADV) (adenosine analogue), and tenofovir (TFV) (adenosine analogue), suppress HBV replication via inhibition of viral polymerase. IFN-α and its pegylated form modulate the host immune response to HBV infection and/or directly inhibit viral replication in hepatocytes. NAs often lead to the selection of drug-resistant mutations that in many cases are cross-resistant to other NAs during long-term treatment. Moreover, IFN-α is associated with unfavorable side effects.

HBV polymerase, the target of NAs, is composed of four domains, i.e., terminal protein, spacer region, reverse transcriptase (RT), and RNase H (5–8). The RT domain consists of seven motifs (A through G). Motifs B and E are important for template binding, and motifs A and D are critical for nucleoside binding, as inferred from HIV RT (6, 9, 10). Generally, two types of mutations have been reported to be associated with NA treatment failure, namely, primary resistance mutations, which are directly responsible for drug resistance, and secondary (compensatory) mutations, which promote or enhance replication competence (8). ETV is one of the most effective anti-HBV drugs and has been used as a first-line reagent. Resistance to ETV is conferred by a combination of mutations in motifs B, C, or E of the viral polymerase, in addition to a background of substitutions at position M204 in motif C (8). Resistance to these NAs has been mapped to various mutations in RT (6, 8, 11, 12) (Fig. 1A). The primary mutation in the YMDD domain (RT domain amino acid positions 203 to 206) in motif C confers LMV resistance, and the secondary mutation at L180M (motif B) enhances viral replication as a compensatory mutation. V173L further enhances viral replication of the L180M/M204V variant (6, 13, 14). ETV resistance requires a secondary mutation in addition to LMV resistance mutations L180M and M204V; secondary mutations at I169, T184, S202, or M250 result in greater resistance to ETV than that of L180M/M204V (6). Although the amino acids responsible for ETV resistance have been identified, how HBV RT acquires this resistance is still unclear.

FIG 1.

Purification of HBV RT-RNase H. (A) Schematic of HBV polymerase. ETV and LMV resistance substitutions in motifs A to G of RT are indicated as the amino acid positions of the RT domain (8). (B) SDS-PAGE (Coomassie blue stain) and Western blotting of 10 pmol each of the partially purified RT-RNase H variants (each indicated above the lanes of the Coomassie blue-stained gel and the Western blot). Arrowheads indicate the position of RT-RNase H; the low-molecular-weight marker (Marker, Bio-Rad) is shown on the left. The position of the Western blot was adjusted by Kaleidoscope (KS) (Bio-Rad) to that of the low-molecular-weight marker.

Because NAs exhibit cross-resistance, it is crucial to understand the biochemical features of ETV-resistant HBV RT (13, 15–17). Therefore, to elucidate the mechanisms underlying ETV resistance, we focused on L180M and M204V mutations, which were originally identified as the core mutations of LMV resistance and also confer cross-resistance to other L-nucleotides and reduce sensitivity to ETV. We studied these mutants using a combination of kinetic analysis for ETV inhibition using traditional Lineweaver-Burk (L-B) plots, analysis of the soluble HBV RT fraction (18, 19), and experimental structural analysis of HBV-mimic HIV RT^{Q151M/Y115F/F116Y} (RT^3MB) with ETV-resistant F160M and M184V mutations (20), for a more detailed understanding than that obtained with in silico modeling analysis alone (5, 16, 17).

RESULTS

Purification of wild-type and mutant HBV RT-RNase H.

We expressed and purified recombinant HBV RT-RNase H proteins derived from patients with ETV-resistant HBV, L180M/M204V (14), L180M and M204V, and their RT knockout (KO) mutants (YMDD to YMAA of the wild-type [wt] strain and L180M or YVDD to YVAA of the L180M/M204V mutant and M204V) (Fig. 1B) (8, 21). The partially purified RT-RNase H fractions contained bacterial contaminants and intact-size RT-RNase H proteins with various lengths of degraded products, as detected by anti-His antibodies (Fig. 1B).

Effects of ETV-TP on resistant and wt HBV RT-RNase H.

First, the inhibitory effects of ddGTP and ETV-5′-triphosphate (ETV-TP) were roughly evaluated by adding 200, 100, 50, and 25 μM ddGTP or ETV-TP to 10 μM dGTP for 2 h of incubation (Fig. 2A and B). Both ddGTP and ETV-TP dose-dependently inhibited the primer extension by wt and mutant HBV RT-RNase H. In other words, with high doses of ddGTP and ETV-TP, the long RT elongation products disappeared and short products were observed. From the 4C RNA template, 5G products with an additional G were at times generated by the wt and mutant RT-RNase H in the presence of dGTP (see below). With 200 μM ddGTP, wt RT-RNase H generated products up to 4G; with the same concentration of ETV-TP, it exhibited only up to 3G. Because these fractions contained DNase activity (19) and less DNA degradation was observed when HBV RT-RNase H was incubated for 30 min, compared to that observed with 2 h of incubation, inhibition of ETV-TP was reevaluated by adding 200, 100, 50, 25, 12.5, 6.25, 3.13, 1.56, and 0.78 μM ETV-TP into 10 μM dGTP for 30 min of incubation (Fig. 2C). The amount of RT elongation products longer than 2 nucleotides (nt) was measured in each gel, and the 50% inhibitory concentration (IC₅₀) of ETV-TP was calculated and averaged from three experiments (Table 1).

FIG 2.

Evaluation of the inhibitory effect of ETV-TP on HBV RT. (A and B) Inhibitory effects of ddGTP (A) or ETV-TP (B) on HBV wt and mutant RT-RNase H, examined by adding 200, 100, 50, 25, and 0 (−) μM ddGTP (A) (from left to right) or ETV-TP (B) to 10 μM dGTP using the 4C RNA template for 2 h of incubation. (C) Measurement of the IC₅₀ of ETV-TP. The wt RT-RNase H and mutant RT-RNase H were incubated for 30 min with 100, 50, 25, 12.5, 6.25, 3.13, 1.56, 0.78, and 0 μM ETV-TP with 10 μM dGTP. Numbers at the top of the PAGE gel indicate the percentage of RT elongation products (2 to 5 nt) (%RT) versus those without ETV-TP in this example. (D) PAGE of the elongation activity of RT KO mutants (wt KO, M240V/L180M KO, M204V KO, and L180M KO). The RT KO mutants were incubated with 50, 25, 12.5, 6.25, and 3.13 μM dGTP (from left to right). The positions of FITC-conjugated primers are indicated by P, and those of the primer extension products are indicated by numbers. FITC-primer (10 fmol) is indicated below the gel of the wt enzyme (A).

TABLE 1.

IC₅₀ and estimated K_i of ETV-TP for HBV RT

Enzyme	IC50		Ki (μM)
	In vitro (μM)	Cell culture (nM)a
wt	13.8 ± 4.1	0.0865	3.9
L180M/M204V	139 ± 11	9.95	Not calculated
M204V	18.9 ± 0.9	Not done	27.9
L180M	29.0 ± 2.9	Not done	Not calculated

^aData from Yamada et al. (14).

The contamination of the host RT activity was evaluated with RT KO mutants because these fractions contained host nucleotide addition activity (Fig. 2D) (19). The KO mutants in which polymerase amino acids 551 and 552, DD, were substituted by AA were incubated with the series of concentrations of dGTP. Less than 10 fmol of products that were elongated by 1 or 2 nt were observed. This activity was similar to the terminal deoxynucleotide transferase (TdTase) activity detected in the RT fraction (19). This activity did not depend on the concentration of dGTP. Therefore, for the measurement of HBV RT elongation products, the bands longer than the second nucleotide (including the second nucleotide) were measured including this contaminating activity, and their values were averaged from several experiments in order to decrease the effect of the contaminating activity.

Incorporation of ETV-TP and LMV-TP in RNA and DNA templates.

Next, we examined the effect of ETV-TP on both HBV RT-RNase H (wt) and HIV RT, along with that of LMV-5′-triphosphate (LMV-TP), ddGTP, and ddCTP (Fig. 3). When RT-RNase H was incubated only with either 200 μM ddGTP or ETV-TP, the elongation stopped at 1 nt; with 200 μM dGTP, it proceeded up to the 4- and 5-nt products (Fig. 3A). In contrast, HIV RT generated long products even with 100 μM ETV-TP (alone), as well as with 100 μM dGTP; however, the products were shorter in the presence of ETV-TP than with dGTP at 30 min of incubation (Fig. 3B). HIV RT exhibited DNA termination at the first nucleotide when it was incubated with 500 μM ddGTP alone as a control. This result is consistent with the previous observation that ETV-TP has a limited and much lower inhibitory effect on HIV RT than on HBV RT (22). When the 18C DNA template was used, ETV-TP was not incorporated efficiently by HBV RT-RNase H, unlike with the 4C RNA template (Fig. 3A). Similar to ddCTP, LMV-TP stopped the elongation of wt HBV RT-RNase H at the first nucleotide on the 5G RNA templates (Fig. 3C). The wt HBV RT-RNase H produced longer products from 5G RNA templates, as observed previously (19).

FIG 3.

Incorporation of ETV-TP and LMV-TP by HBV RT-RNase H and HIV RT. (A) RT elongation assay of HBV wt RT-RNase H with 200 μM ETV-TP, ddGTP, and dGTP using the 4C RNA (left) and 18C DNA (right) templates. (B) RT elongation assay of HIV RT with 100 μM dGTP, ETV-TP, and ddGTP using the 19C RNA template. (C) Elongation assay of HBV wt RT-RNase H with 200 μM LMV-TP, ddCTP, and dCTP using the 5G RNA template. (D) Elongation assay of HBV wt, L180M/M204V (LM), L180M, and M204V RT-RNase H with ETV-TP (E lanes) or dGTP (G lanes) on RNA and DNA templates for 30 and 60 min. The positions of FITC-conjugated primers are indicated by P, and those of the primer-extension products are indicated by numbers. The long RT products from 18C DNA templates are indicated by arrows. The long RT products from G5 RNA templates are indicated by arrowheads.

The effect of DNA chain termination by ETV-TP was examined for both RNA and DNA templates with wt, L180M/M204V, L180M, and M204V HBV RT-RNase H. Thus, 100 μM dGTP or ETV-TP was incubated with wt or mutant RT-RNase H and RNA or DNA template for 30 or 60 min (Fig. 3D). The signals produced by RT elongation products incubated for 60 min were weaker than those produced with 30 min-incubation, which indicated degradation of DNA products or 5′-fluorescein isothiocyanate (FITC). For all tested RT-RNase H mutants, ETV-TP stopped DNA elongation at the first nucleotide on the RNA template, while the signal of the second nucleotide on the DNA template was observed. The signal on the 4C RNA template was weak only for L180M upon incubation with dGTP. The signals of L180M, M204V, and L180M/204V were weaker than that of wt enzyme at 30 min when the 18C DNA template was used. At 60 min, the signal of M204V became similar to that of the wt enzyme.

K_m and V_max of HBV wt and mutant RT-RNase H estimated from L-B plots.

We estimated the K_m and V_max of wt and mutant RT-RNase H using L-B plots obtained from the elongation assay with homopolymeric RNA templates after a 30-min incubation (Table 2; see Fig. 4 for the scanned data). We measured the DNA elongated further than 2 nt from the primer because ETV-TP terminated RT elongation at 1 nt from the primer. Furthermore, HBV RT-RNase H efficiently elongated 1 nt when the deoxynucleoside triphosphate (dNTP) concentration was very low (Fig. 4 and the top row of Fig. 5A). The single line of the L-B plot was drawn from all measurements averaged at each dNTP concentration in order to obtain the single crossing point on the x axis and y axis. The K_m of the wt enzyme for dGTP (0.68 μM) and dCTP (0.78 μM) was lower than those for dATP (6.93 μM) and TTP (2.38 μM). The V_max for dATP (2.27 pmol/min) was lower than those for the others (TTP, 6.00 pmol/min; dGTP, 7.58 pmol/min; dCTP, 8.62 pmol/min). L180M/M204V exhibited higher K_m values for three dNTPs except for dATP (dATP, 0.8-fold; TTP, 3-fold; dGTP, 4-fold; dCTP, 3-fold) and lower V_max values for all dNTPs (dATP, 0.8-fold; TTP, 0.5-fold; dGTP, 0.6-fold; dCTP, 0.3-fold), compared with those of the wt enzyme.

TABLE 2.

Apparent V_max and K_m values from in vitro elongation assays

Enzyme	Data obtained with:
	dATP		TTP		dGTP		dCTP
	Vmax (pmol/min)	Km (μM)	Vmax (pmol/min)	Km (μM)	Vmax (pmol/min)	Km (μM)	Vmax (pmol/min)	Km (μM)
wt	2.27	6.93	6.00	2.38	7.61	0.68	8.62	0.78
L180M/M204V	1.76	5.58	3.03	7.10	4.29	2.77	2.71	2.53
M204V	1.20	81.5	0.72	4.10	5.75	2.00	0.52	2.80
L180M	2.35	13.2	6.08	3.23	6.94	2.42	8.61	1.82

FIG 4.

Quantitation of HBV RT products with native substrates. Example kinetic analyses of HBV wt RT-RNase H are shown. Concentrations of dATP (A), TTP (B), dGTP (C), and dCTP (D) are 250, 125, 62.5, 31.3, 15.6, 7.81, 3.91, 1.95, 0.977, and 0.488 μM (from left to right). Standard curves of fluorescent signals were generated by loading the indicated amounts of FITC-primers on the same gels (C). The sequences of the 5′-FITC (F)-conjugated primers and the annealed RNA templates are indicated below the gels. The position of the primer is indicated as P. The numbers to the left of the gels indicate the length of the RT products.

FIG 5.

L-B plots of HBV RT-RNase H with ETV-TP. (A) Quantitation of HBV RT products at different concentrations of dGTP with different amounts of ETV-TP. HBV wt RT-RNase H was incubated with 30, 15, 7.5, 3.25, and 1.63 μM dGTP. L180M/M204V was incubated with 30, 15, 7.5, 3.25, 1.63, and 0.81 μM dGTP. M204V and L180M were incubated with 250, 125, 62.5, 31.3, 15.6, 7.81, 3.91, 1.95, and 0.98 μM dGTP. The amount of ETV-TP used is indicated below each gel. (B) L-B plots of HBV wt, L180M/M204V, L180M, and M204V RT-RNase H were drawn with and without (−) the indicated amounts of ETV-TP. These plots were drawn from dGTP concentration (S) (micromolar) and polymerization rate (V) (picomoles per hour) values. The 1/V_max values (y intercept) for the wt and M204V enzymes are 0.0022 and 0.0029, respectively. The standard deviation of 1/V for each measurement is indicated in the graph.

We then calculated the kinetic parameters for the mutants with the single mutations M204V (the key mutation for LMV resistance) and L180M. The M204V mutant showed higher K_m values (dATP, 12-fold; TTP, 1.7-fold; dGTP, 3-fold; dCTP, 3.6-fold) and lower V_max values (dATP, 0.5-fold; TTP, 0.1-fold; dGTP, 0.8-fold; dCTP, 0.06-fold) for all dNTPs, compared with those of the wt enzyme. The L180M mutant instead showed higher K_m values for all dNTPs (dATP, 1.9-fold; TTP, 1.4-fold; dGTP, 3.6-fold; dCTP, 2-fold) but similar V_max values (dATP, 1-fold; TTP, 1-fold; dGTP, 0.9-fold; dCTP, 1-fold), compared with those of the wt enzyme.

IC₅₀ and K_i of ETV-TP for HBV wt and mutant RT-RNase H.

The mechanism of ETV inhibition was visualized by the L-B plots drawn at various concentrations of dGTP with three different concentrations of ETV-TP (Fig. 5A). First, the IC₅₀ of ETV -TP was calculated at 10 μM dGTP for each RT-RNase H, and the K_i of ETV-TP for the wt and M204V enzymes was then calculated from the L-B plot (Table 1 and Fig. 2C and 5B). As the lines crossed the y axis at one point, the ETV-TP inhibition patterns for wt and M204V enzymes fit a competitive inhibition model; however, those for L180M and L180M/M204V did not fit either uncompetitive, noncompetitive, or typical mixed inhibition models, because their three lines crossed at different points. Therefore, it was not possible to obtain K_i values for L180M and L180M/M204V from their L-B plots.

Structural changes observed in HIV RT mimicking HBV RT L180M/M204V.

We previously designed a mutant HIV RT with the HBV-associated triple amino acid substitutions Q151M, Y115F, and F116Y (3MB) and found that these mutations rendered HIV highly susceptible to ETV (20). Therefore, to explore the dNTP binding site structure of ETV-resistant HBV RT with L180M and M204V mutations, we chose HIV RT^{3MB/F160M/M184V} as a surrogate for experimental structural studies. We successfully obtained crystals of HIV RT^{3MB/F160M/M184V}-DNA-dCTP and determined the three-dimensional structure to a resolution of 2.67 Å. In addition, we determined the structure of RT^3MB/M184V-DNA-dGTP at 2.43-Å resolution. Unfortunately, we could not obtain crystals of HIV RT^{3MB/F160M/M184V}-DNA-dGTP. The overall structures of RT^3MB/M184V-DNA-dGTP and RT^{3MB/F160M/M184V}-DNA-dCTP showed the closed conformation accommodating dGTP/dCTP at the dNTP binding site, very similar to findings reported previously for closed dNTP-bound HIV RT (23). The electron density for the bound dGTP/dCTP and Mg²⁺ is clear (Fig. 6). The binding position of dGTP in RT^3MB/M184V-DNA is nearly identical to that in wt HIV RT-DNA and RT^3MB-DNA (Fig. 7A). In contrast, there is a novel finding in the structure of RT^{3MB/F160M/M184V}-DNA-dCTP. The root mean square (RMS) superimposition of p66 subunit backbone atoms revealed that the bulky side chain of M160 in RT^{3MB/F160M/M184V} slightly moves the phenyl ring of F115 (F88 in HBV RT) toward the bound dNTP with ∼0.6 Å. This F115 bulge leads to deviated dCTP binding, compared to the tightly bound dGTP/dCTP observed in the structure of RT^3MB with a distance of ∼0.8 Å (22). In addition, Mg²⁺ and the side chains of R72 and D110 also moved in response to the deviated binding of dCTP. In particular, the movement of the bound Mg²⁺ with a distance of ∼1.6 Å results in the complete destruction of the octahedral oxygen coordination (Fig. 7B and C). The electron density peak for the Mg²⁺ was significantly low, and model refinement was completed with partial occupancies and high temperature factor values for Mg²⁺ in both chain A and chain C (Table 3). The movement of Mg²⁺ also affects the position of the D185 side chain in the YMDD loop, further compressing the conformation of the 3′-end nucleotide ribose. The modeling of ETV-TP on the deviated dCTP strongly suggests that the ETV-TP methylene could introduce steric hindrance with the C2 atom of the 3′-end nucleotide ribose (Fig. 7D).

FIG 6.

Simulated annealing F_o − F_c omit map for the bound dGTP/dCTP and Mg²⁺ in the structure of HIV RT^3MB/M184V and RT^{3MB/F160M/M184V}. The maps for dGTP-Mg²⁺ in RT^3MB/M184V and for dCTP-Mg²⁺ in RT^{3MB/F160M/M184V} are contoured at the 3.0σ and 2.5σ level, respectively.

FIG 7.

Deviated binding of dCTP to RT^{3MB/F160M/M184V}. (A to C) Active site superimposition of HIV RT^3MB/M184V-DNA-dGTP and RT^3MB-DNA-dGTP (22) (A), RT^{3MB/F160M/M184V}-DNA-dCTP and RT^3MB-DNA-dGTP (22) (B), and RT^{3MB/F160M/M184V}-DNA-dCTP and RT^3MB-DNA-dCTP (22) (C) with two perpendicular views. Carbon atoms for HIV RT^3MB/M184V-DNA-dGTP, RT^{3MB/F160M/M184V}-DNA-dCTP, RT^3MB-DNA-dGTP, and RT^3MB-DNA-dCTP are colored in cyan, yellow, light blue, and light gray, respectively. Interatomic interactions for Mg²⁺ coordination and the nucleotide base pairing are shown as gray dotted lines. Mg²⁺ in the previously reported RT^3MB structure is colored green, while relocated Mg²⁺ in the RT^{3MB/F160M/M184V} is colored orange. The interactions between relocated Mg²⁺ and D185/V111/β-phosphate oxygen are shown as orange dotted lines. The significant movement of the dCTP, side chain residues, and Mg²⁺ in RT^{3MB/F160M/M184V} are indicated by red arrows. (D) Close-up view of superimposition of the bound dGTP in RT^3MB/M184V and deviated dCTP in RT^{3MB/F160M/M184V}, with the 3′-end nucleotide, Mg²⁺, and D185 in the same color scheme as in panels A and B. The approximate location of the ETV methylene group in the deviated dCTP binding position is shown as a thick red line, indicating the serious steric clash with the C2 atom of the 3 ′-end nucleotide (red dotted line). In contrast, the ribose oxygen of the tightly bound dGTP in RT^3MB/M184V lies further apart from the C2. The interatomic distances between 3′-end C2 atoms and the deviated dCTP or tightly bound dGTP ribose oxygen atoms are 3.4 and 4.0 Å, respectively.

TABLE 3.

X-ray diffraction data and model refinement statistics

Parameter	Data for:
	RT3MB/M184V-DNA-dGTP	RT3MB/M184V/F160M-DNA-dCTP
Data collection
Beamline	PF BL-1A	PF BL-1A
Wavelength (Å)	1.1000	1.1000
Temperature (K)	100	100
Detector	Eiger X4M	Eiger X4M
Space group	H3	H3
Unit cell parameters (Å)	a = b = 283.5, c = 96.1	a = b = 285.7, c = 96.3
Resolution (Å)a	50–2.43 (2.47–2.43)	50–2.67 (2.83–2.67)
No. of unique reflections	107,363	83,121
Rmeasa,b	0.103 (0.819)	0.089 (0.982)
Mean I/σ (I)a	10.5 (2.1)	14.8 (2.0)
Completeness (%)a	100.0 (100.0)	100.0 (99.6)
Multiplicitya	5.4 (5.6)	5.4 (5.5)
Model refinement
Rworkc	0.184	0.178
Rfreed	0.228	0.226
No. of atoms	17,702	17,447
Avg B-factors (Å2)
All	66.0	73.0
DNA	63.3	71.9
dNTP
Chain A	46.4 (dGTP)	112.4 (dCTP)
Chain C	53.0 (dGTP)	97.7 (dCTP)
Mg2+
Chain A	47.7	42.1 (occupancy = 0.39)
Chain C	48.6	144.9 (occupancy = 0.84)
RMS deviation from ideal
Bond lengths (Å)	0.006	0.007
Bond angles (°)	0.818	0.932
Ramachandran plot
Favored (%)	96.59	96.64
Outliers (%)	0.16	0.16

^aThe values shown in parentheses are for the outermost resolution shell.

^bR_meas = Σ_hΣ_i |I_h,i − <I_h>|/Σ_hΣ_i I_h,i, where <I_h> is the mean intensity for a set of equivalent reflections.

^cR_work = Σ|F_obs – F_calc|/Σ F_obs for 95% of the reflection data used in the refinement. F_obs and F_calc are the observed and calculated structure factor amplitudes, respectively.

^dR_free is the equivalent of R_work except that it was calculated for a randomly chosen 5% of reflections, which were excluded from refinement.

IC₅₀ of ETV-TP and K_m and V_max with dGTP for HIV RT mimicking HBV RT.

We compared the IC₅₀ values of ETV-TP for HIV RT^3MB, RT^{3MB/F160M/M184V}, RT^3MB/F160M, and RT^3MB/M184V by the same method except for the use of RNA templates longer than those used in the HBV RT assay (Fig. 8A). Because HIV RT efficiently reverse transcribed DNA and because ETV-TP was incorporated efficiently by HIV RT (Fig. 3B), 595-nt-long RNA templates were used to measure IC₅₀ (Fig. 8B). HIV RT was incubated with 200, 150, 100, 50, 25, 12.5, 6.25, and 3.13 μM ETV-TP with 5 μM dNTPs. The longest four bands were measured, and their amounts were compared to those without ETV-TP. The ETV-TP IC₅₀ values were calculated from two experiments (Table 4), as follows: HIV RT^3MB, 3.39 ± 0.18 μM; RT^{3MB/F160M/M184V}, 94.1 ± 9.7 μM; RT^3MB/F160M, 5.73 ± 0.26 μM; RT^3MB/M184V, 34.8 ± 0.73 μM.

FIG 8.

Effects of ETV-TP and dGTP concentrations on HIV RT in vitro. (A) SDS-PAGE of the purified HIV RT. Each (5 pmol) of HIV RT^3MB, RT^{3MB/F160M/M184V}, RT^3MB/F160M, and RT^3MB/M184V was separated by SDS-PAGE and stained with Coomassie blue. The positions of P66 and P51 are indicated on the right, and those of the standard markers (Marker) are indicated on the left. (B) Example of ETV-TP effect on HIV RT. HIV RT^3MB, RT^3MB/M184V, and RT^3MB/F160M (each 40 nM) were incubated with 50, 25, 12.5, 6.25, 3.13, 1.56, 0.78, and 0.39 μM ETV-TP or without ETV-TP (−) with 5 μM dNTP. RT^{3MB/M184V/F160M} was incubated with 200, 100, 50, 25, 12.5, 6.25, 3.13, and 1.56 μM ETV-TP or without ETV-TP (−). The position of the primer (P) and the four largest primer elongation products measured are indicated. (C) Example of the effect of the concentration of dGTP on HIV RT. HIV RT (0.4 μM) was incubated for 10 min with 100, 50, 25, 12.5, 6.25, 3.13, 1.56, and 0.781 nM dGTP with the 19C RNA template. The RT primer extension products were separated on a 15% PAGE gel containing 6 M urea. The positions of primer (P) and the 19-nt primer elongation products are indicated.

TABLE 4.

IC₅₀ of ETV-TP and estimated K_m and V_max with dGTP for HIV RT

Enzyme	IC50		Vmax with dGTP (pmol/min)	Km with dGTP (nM)
	In vitro (μM)	Cell culture (nM)a
3MB	3.39 ± 0.18	96 ± 24	15.9	2.73
3MB/F160M/M184V	94.1 ± 9.7	1,136 ± 264	16.2	3.37
3MB/M184V	34.8 ± 0.73	451 ± 89	20.9	6.56
3MB/F160M	5.73 ± 0.26	Not done	20.6	2.73

^aData from Yasutake et al. (38).

For comparison with K_m and V_max values for HBV RT with dGTP, the HIV RT elongation products were measured from two experiments using 19C RNA templates for 10-min incubations with 100, 50, 25, 12.5, 6.25, 3.13, 1.56, and 0.781 nM dGTP. The single line of the L-B plot was drawn from the averaged values at each concentration of dGTP, and K_m and V_max values were calculated (Table 4). The K_m and V_max values for HIV RT^3MB were 2.73 nM and 15.9 pmol/min, respectively. The K_m and V_max values for RT^{3MB/F160M/M184V} were 3.37 nM and 16.2 pmol/min, respectively. The K_m and V_max values for RT^3MB/M184V were 6.56 nM and 20.9 pmol/min, respectively. The K_m and V_max values for RT^3MB/F160M were 2.73 nM and 20.6 pmol/min, respectively.

DISCUSSION

HBV mutations that confer resistance to NAs are a serious problem, making complete treatment more difficult. Substrate affinity of viral polymerases is one factor determining the sensitivity to NAs (24–26). ETV is a guanosine analogue and one of the most potent anti-HBV drugs. In order to understand how ETV-resistant mutations affect the affinity for natural substrates to discriminate NAs, we compared the kinetic constants (K_m and V_max), IC₅₀, and K_i values for HBV RT L180M, M204V, and L180M/M204V mutants, which showed 115-fold resistance to ETV in a previous report using a cell culture assay (Table 4) (14). Compared to the wt RT, HBV RT-RNase H L180M/M204V exhibited 10-fold resistance to ETV-TP in vitro, which was less than one-tenth of that in the cell culture (Table 2). ETV exhibited stronger inhibition in the cell culture than in in vitro assays because it inhibits protein priming, reverse transcription, and DNA elongation (7, 27). Kinetic analysis of ETV- and LMV-resistant HBV RT was previously performed using HBV core and endogenous DNA templates purified from HepG2 cells (15, 17). We estimated the K_m and V_max values for the mutant HBV RT using RNA templates (Table 2). Comparison of our results with those of three independent studies, which were calculated from different systems, revealed that the K_m values vary in different reports. Gaillard et al. (15) demonstrated that L180M/M204V, M204V, and L180M mutants exhibited higher K_m values for dCTP and dGTP than that did the wt enzyme. According to Walsh et al. (17), the K_m of L180M/M204V for dGTP, which was calculated from a L-B plot, was lower than that of the wt enzyme. Because our study included the background signals, the apparent K_m and V_max values measured under the same conditions could be compared. Our study shows that L180M, M204V, and L180M/M204V mutants had higher K_m values for dGTP and lower affinity for dGTP and dCTP than did the wt enzyme, consistent with the observation by Gaillard et al. (15) in which a L-B plot was not applied. Interestingly, L180M had V_max values similar to those of the wt enzyme, while those of M204V were lower, for all dNTPs. This indicates that, although both L180M and M204V mutations decreased the affinity for dNTPs, L180M did not substantially affect the DNA polymerization efficiency. This suggests that L180 does not play a role in DNA polymerization activity itself.

Two mechanisms of nucleoside RT inhibitor resistance were proposed from studies on HIV RT, i.e., (i) increased discrimination of the dNTP analogue at the RT active site, relative to the natural dNTPs, and (ii) repair of the analogue-terminated DNA chain. The LMV resistance mutation HIV RT M184V belongs to the former (28–30), and zidovudine (AZT) resistance mutations HIV RT M41L, D67L, K70R, L210W, T215Y/F, and K219(Q/E/N/R) belong to the latter (31, 32).

Our kinetic analysis showed that the decreased affinity of HBV RT M204V for dGTP coincided with that of a similar HIV RT mutant, M184V, which has a lower affinity for dGTP, ETV-TP, and LMV-TP than does the wt enzyme (Tables 1, 2, and 4) (33–35). Our data comparing K_m and V_max values with dGTP between HIV RT^3MB and RT^3MB/M184V indicated preferences similar to those for dGTP reported by Domaoal et al. (33). M204 in the YMDD motif is located at the RT polymerization center, as shown by structural analysis of HIV RT. Our kinetic study also showed that the M204V mutation did not alter the location of the nucleotide binding site but decreased dNTP and ETV-TP binding affinity and reduced the polymerization speed (Tables 1, 2, and 4 and Fig. 5B).

Molecular modeling proposed different interactions of ETV-TP with HBV RT, compared with that of dGTP, in the polymerization center (16, 17, 36). Structural analysis of HIV RT^3MB/M184V-DNA-dGTP confirmed that the binding position of dGTP is identical to that in RT^3MB (Fig. 7A) (20). The substitution of bulky M with smaller V could decrease the interactions with ribose and the cyclopentyl moiety of dNTP/ETV-TP, consequently decreasing their binding affinity. In contrast, the L-B plot of dGTP with ETV-TP did not fit the competitive, noncompetitive, uncompetitive, or mixed inhibition patterns for L180M or L180M/M204V (Fig. 5B). These kinetic data suggested the emergence of multiple ETV-TP binding positions caused by L180M mutation.

We have investigated the structure of HBV RT based on the idea that HIV and HBV RT assume similar structures, since the consensus sequences of motifs A to D indicate that their dNTP binding site is highly conserved (20). However, the inhibition kinetics of L180M (Fig. 5B) were difficult to interpret based on the previous modeling of HBV RT based on HIV RT (5, 16, 17, 36). Therefore, we analyzed the crystal structure of HIV RT^{3MB/F160M/M184V} mimicking HBV L180M/M204V in order to better understand the interaction between dNTPs and HBV RT (Fig. 7).

Modeling analysis of the HBV RT tertiary structure demonstrated that M204V decreased the stability of dNTPs (16, 17, 36), in agreement with the kinetic analysis indicating competitive inhibition (Fig. 5B and Tables 1 and 2). The X-ray structure of HIV RT^3MB/M184V-DNA-dGTP showed that the dGTP is bound in the dNTP binding site, similar to what was observed in the previously reported structures of HBV-mimicking HIV RTs (20, 37, 38). In contrast, analysis of the RT^{3MB/F160M/M184V}-DNA-dCTP structure clearly indicated the slight but significant deviation of the dCTP binding position within the dNTP binding site, which was due to the F115 bulge caused by the introduction of bulky M160. The deviated binding of dCTP also destroyed the stable Mg coordination with the triphosphate moiety of dCTP, thus leading to the decreased dNTP binding affinity observed in the kinetic data. More importantly, modeling of ETV-TP superimposed on the deviated dCTP indicates that a serious steric clash would be expected to occur between the cyclopentyl methylene of ETV-TP and the C2 atom of the 3′-end nucleotide ribose (Fig. 7D). Therefore, the ETV-TP could not fit in the bulged dNTP binding site shown in the RT^{3MB/F160M/M184V} structure, thereby leading to ETV resistance.

The ETV sensitivity of HIV carrying these HBV-mimic RTs was tested previously (38). We confirmed the sensitivity to ETV-TP in vitro by calculating IC₅₀ values (Table 4). The magnitude of ETV resistance of HIV RT^3MB mutants in vitro was greater than that in cell culture; HIV RT^{3MB/F160M/M184V} was 28-fold resistant in vitro and 11-fold resistant in cell culture, compared to HIV RT^3MB, HIV RT^3MB/M184V was 10-fold resistant in vitro and 4.5-fold resistant in cell culture, and HIV RT^3MB/F160M was 1.7-fold resistant in vitro. These differences from those of HBV RT-RNase H (Table 1) indicated that ETV inhibited three biochemical functions of HBV RT (7, 27), of which the priming step might play a key role. The K_m values with dGTP for HBV L160M/M204V and HIV RT^{3MB/F160M/M184V} also agreed with these models for our proposed resistance mechanism of HBV RT against ETV (Tables 1 and 4 and Fig. 7), although that for HBV RT L180M indicated that HBV RT L180 played a more important role in forming the nucleotide pocket with M204 than did the interaction of HIV RT F160 and M184.

Our previous structural study of HBV-mimic HIV RT together with antiviral assays revealed that the M151 in HIV RT is a key determinant for ETV susceptibility of HIV, i.e., ETV-TP could enter the dNTP binding site via transient interactions between the methylene of ETV-TP and the M151 side chain. M151 in HIV RT corresponds to M171 in HBV RT, which is almost completely conserved in all HBV isolates in the Hepatitis Virus Database (http://s2as02.genes.nig.ac.jp/db_noacct/index.php#hbv/ShowData.php?div=public%2Fhbv%2Faa%2Fpol&dtype=aln), and the same entry mechanism might therefore apply to HBV RT (Fig. 9). Considering this mechanism, since the L180M and M204V mutations would render the F88 bulge at the dNTP binding site unsuitable for ETV-TP binding, ETV-TP would be stuck at the M171 of HBV RT L180M/M204V. M171 is located at the entrance of the dNTP binding site and acts as a lid for the dNTP binding site; thus, the simultaneous binding of ETV-TP and dNTP at the M171 and dNTP binding sites would be unlikely. These distinct binding positions are in agreement with the strange inhibition pattern for HBV RT L180M/M204V.

FIG 9.

Proposed mechanism of ETV entry into the dNTP binding site that plausibly explains the strange inhibition of the HBV RT L180M/M204V mutant by ETV. (A) Based on the structural study of the HBV RT-mimic HIV RT (20), ETV-TP likely enters the dNTP binding site via a prerequisite transient interaction with M171 of HBV RT, which corresponds to Q151M of the mutant HIV RT. Such an entry mechanism could not be expected for dNTPs because the Q151M in HIV RT does not affect the polymerase activity. (B) In the present study, the structure of HIV RT^{3MB/F160M/M184V} implies that bulky M180 could push the F88 side chain and create the bulged dNTP binding pocket, precluding the tight binding of ETV-TP/dNTP, as shown by the dotted outline. In particular, a steric clash could be expected to occur between the methylene of ETV-TP and the 3′-end nucleotide ribose, represented as a red pentagon. The binding of ETV-TP at M171 and the deviated dNTP binding at the active site could explain the strange inhibition behavior shown by HBV RT L180M and L180M/M204V in kinetic analyses.

Regarding the L180M mutation, in the mimic X-ray model F160M caused the F115 bulge and consequently pushed dCTP slightly out (Fig. 7B and C). Such structural deformation at the dNTP binding site provides evidence of the possible conformational perturbation of the dNTP binding site caused by L180M, which was predicted by a previous in silico modeling study (5). Although kinetic and structural analyses agreed that M204V kept the position of the dGTP and ETV-TP binding site in the YMDD, these dNTPs did not stably remain there. L180M interfered with the entry of these dNTPs by flicking them out. Expanding to three other dNTPs, M204V slowed down the DNA polymerization rate because it destabilized the dNTP in the RT polymerization center. L180M did not affect the polymerization speed when the dNTP entered the YMDD motif because it did not affect the V_max (Table 2). Crystallography of the HBV RT-mimic HIV RT fit well with our kinetic data. Previously we proposed the additional association of ETV-TP with M171 of HBV RT (20). In HBV RT M204V, dNTP binding was weaker than that in wt enzyme because of loss of their interaction with the M184 side chain (Fig. 7).

With the additional mutation of L180M, ETV-TP fits worse into the dNTP binding pocket due to the possible F88 bulge caused by M180, in addition to the weak binding affinity caused by V204. The strange inhibition pattern of L180M mutants could be explained by the possible occurrence of ETV-TP stuck at M171. This unique feature of interaction between ETV-TP and HBV RT M171 is the reason why ETV-resistant HBV RTs have multiple mutations other than those of LMV resistance (8, 16). The profiles of resistance to ETV and LMV of HBV mutants L180M, M204V, and L180M/M204V vary slightly in different reports (6, 15, 39). HBV M204V showed high resistance to LMV and intermediate resistance to ETV. HBV L180M showed intermediate resistance to LMV and low resistance to ETV. The resistance to LMV and ETV increased in the double mutant (L180M/M204V). Moreover, the expected ETV-TP–M171 interaction may contribute to the lower resistance of HBV L180M to ETV than to LMV.

Many HBV drugs are still under development. LMV-TP is a chain terminator, while ETV-TP is a delayed chain terminator (40). ETV-TP worked as a chain terminator on RNA templates as efficiently as ddGTP (Fig. 3). ETV-TP also inhibited the HBV RT priming while ddGTP did not (41). However, its HBV RT inhibition is not complete, and ETV-resistant mutations arise frequently (6, 8, 27). In this paper, we clarified the mechanism of ETV resistance of L180M/M204V by a combination of kinetic analysis of soluble HBV RT and crystallography of HBV RT-mimic HIV RT. The wt HBV RT showed low affinity for dATP (K_m = 6.93 μM) and TTP (K_m = 2.38 μM) (Table 2) (19), which may in part explain the effective inhibition of RT by TFV and ADV (dATP analogues) and telbivudine (TTP analogue). Although ETV-TP methylene binding at M171 is presumably transient and highly dynamic, it is possible that stable binding of certain compounds to M171 could overcome the ETV resistance of HBV. Design of NAs that have additional binding sites in HBV RT beside the dNTP pocket may produce effective inhibitors like GG167 for influenza virus (42).

MATERIALS AND METHODS

Reagents.

dNTPs were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Dideoxynucleotides (ddNTPs) were purchased from Sigma-Aldrich (St. Louis, MO, USA). LMV-TP was purchased from Abcam (Cambridge, UK). ETV-TP was synthesized by GeneACT, Inc. (Kurume, Japan).

Plasmids.

A DNA fragment encoding an N-terminally His-tagged fragment of recombinant HBV polymerases (corresponding to amino acid residues 304 to 843 of genotype C HBV polymerases, clones A1 [GenBank accession number AB246344] and A1+MV) (14) was PCR amplified and cloned between the NdeI and EcoRI sites of pET28b (Merck, Darmstadt, Germany), resulting in pET28-wt and pET28-L180M/M204V. M204V and L180M were constructed from clone A1 with the QuikChange II site-directed mutagenesis kit (Stratagene, La Jolla, CA) and cloned into the NdeI and EcoRI sites of pET28b, generating pET28-M204V and pET28-L180M, respectively. The RT activity KO mutants in which polymerase amino acids 551A and 552A were mutated to D and D, respectively (wt and L180M, YMDD to YMAA; M204V and L180M/M204V, YVDD to YVAA), were constructed with the QuikChange II site-directed mutagenesis kit. The DNA sequences of the plasmid inserts were confirmed by DNA sequencing (FASMAC, Atsugi, Japan, or Eurofins Genomics, Tokyo, Japan).

Expression and purification of wt and mutant HBV RT-RNase H.

Recombinant RT-RNase H (HBV polymerase amino acids 304 to 843) proteins were expressed in Escherichia coli and purified under nondenaturing conditions, as described previously (18).

Expression and purification of HIV RT.

Escherichia coli XL1-Blue cells (Merck) were transformed with pQEHIV (93JP-NH1) RT p51 and p61 (kindly provided by T. Sato) (43). Expression of HIV RT p51 and p61 was induced with 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) for 3 h at 37°C when XL1-Blue cells reached an optical density at 600 nm (OD₆₀₀) of 0.6. Cells were then harvested, pooled, and lysed in 50 mM Tris-HCl (pH 8.0), 0.5 M NaCl, 0.1% Triton X-100, 1 mM EDTA, 1 mM dithiothreitol (DTT), with cOmplete protease inhibitor cocktail (Merck). The resulting p51/p61 heterodimer was purified on HisTrap FastFlow Crude and Superdex 200 Increase columns.

Western blotting.

After electrophoresis, the purified RT-RNase H fractions containing 10 pmol RT-RNase H were electroblotted onto polyvinylidene difluoride membranes. The membranes were incubated with 1 μg/ml of anti-His antibodies (Medical and Biological Laboratories, Nagoya, Japan), followed by alkaline phosphatase-linked anti-mouse IgG secondary antibodies (Promega, Madison, USA). The bands were detected with 5-bromo-4-chloro-3-indolyl-phosphate/nitroblue tetrazolium color development substrate (Promega).

Elongation assay for HBV RT-RNase H.

The recombinant proteins (0.4 μM) were mixed with 50 mM Tris-HCl (pH 8.0), 3 mM MgCl₂, 1 mM DTT, 0.02% Triton X-100, 0.25 mM MnCl₂, 0.4% RNase inhibitor (TaKaRa, Kusatsu, Japan), 0.02 μM primer (FITC)-TCACGGTGGTCTCCATGCG, 0.02 μM RNA template (21A, AAAAAAAAAAAAAAAAAAAAACGCAUGGAGACCACCGUGA; 18U, UUUUUUUUUUUUUUUUUUCGCAUGGAGACCACCGUGA; 4C, CCCCCGCAUGGAGACCACCGUGA; 5G, GGGGGCGCAUGGAGACCACCGUGA) or DNA template (18dC, CCCCCCCCCCCCCCCCCCCGCATGGAGACCACCGTGA), and different concentrations of each substrate (dATP, TTP, dCTP, dGTP, ETV-TP, LMV-TP, ddGTP, and ddCTP). The mixture was incubated at 37°C for 2 h. The RT products were separated on a PAGE gel containing 6 M urea, followed by image analysis on a Typhoon RGB imager (GE Healthcare, Chicago, IL, USA).

Elongation assay for HIV RT.

HIV RT (0.4 μM) was incubated in a mixture of 50 mM Tris-HCl (pH 8.0), 1 mM MgCl₂, 1 mM DTT, 50 mM KCl, 0.4% RNase inhibitor (TaKaRa), 0.02 μM primer 5′-FITC-TCACGGTGGTCTCCATGCG, 0.02 μM RNA template 19C (CCCCCCCCCCCCCCCCCCCCGCAUGGAGACCACCGUGA), and 100 μM (each) dGTP, ETV-TP, or ddGTP. The mixture was incubated at 37°C for 30 min. The RT products were analyzed as described above.

Kinetic assay for HBV RT.

HBV wt RT-RNase H and mutant RT-RNase H were incubated with various concentrations of nucleotides for 30 min. The reverse-transcribed products were analyzed with a Typhoon RGB imager, and then the values from at least two independent experiments were quantified with ImageQuant TL v8.2 (GE Healthcare). V_max and K_m values were estimated from L-B plots.

Expression and purification of HBV RT-mimic HIV RT mutants.

The plasmid constructs for expressing HIV RT^3MB, RT^3MB/M184V, and RT^{3MB/F160M/M184V} were described previously (38). To create the expression plasmid for RT^3MB/F160M, the F160M mutation was introduced by the inverse PCR method using pCDF-RT^3MB(p66) as a DNA template. The overexpression and purification of these HIV RT mutants were performed according to the method described previously (20, 38).

IC₅₀ and K_i of ETV-TP for HBV RT.

For measurement of the ETV-TP IC₅₀, the wt RT and the mutant RT were incubated for 30 min with 100, 50, 25, 12.5, 6.25, 3.13, 1.56, 0.78, and 0 μM ETV-TP with 10 μM dGTP using 0.02 μM 4C RNA template. The K_i for ETV-TP was estimated from L-B plots plotted for various concentrations of dGTP with and without ETV-TP.

IC₅₀ of ETV-TP and K_m and V_max with dGTP for HIV RT^3MB, RT^3MB/M184V, RT^3MB/F160M, and RT^{3MB/M184V/F160M}.

For calculation of ETV-TP IC₅₀ values for HIV RT mutants in vitro, 595-nt RNA templates were prepared by in vitro transcription from pET28bHBV97RNase H, which contains the RNase H domain (HBV polymerase amino acids 691 to 843) of HBV clone 97, linearized by XhoI digestion using the MEGAscript kit (Ambion, Thermo Fisher Scientific). HIV RT^3MB, RT^3MB/M184V, and RT^3MB/F160M (each 40 nM) were incubated with 50, 25, 12.5, 6.25, 3.13, 1.56, 0.78, and 0.39 μM ETV-TP or without ETV-TP in 50 mM Tris-HCl (pH 8.0), 2 mM MgCl₂, 1 mM DTT, 50 mM KCl, 0.4% RNase inhibitor (TaKaRa), 0.02 μM 595-nt RNA annealed with primer 5′-FITC-TCACGGTGGTCTCCATGCG, and 5 μM dNTP. RT^{3MB/M184V/F160M} was incubated with 200, 100, 50, 25, 12.5, 6.25, 3.13, and 1.56 μM ETV-TP or without ETV-TP. The RT primer extension products were separated on a 4% PAGE gel containing 6 M urea, followed by image analysis with a Typhoon RGB imager.

For calculation of dGTP K_m and V_max values, 0.4 μM HIV RT was incubated for 10 min with 100, 50, 25, 12.5, 6.25, 3.13, 1.56, and 0.781 nM dGTP with the 19C RNA template. The RT primer extension products were separated on a 15% PAGE gel containing 6 M urea, followed by image analysis with a Typhoon RGB imager. dGTP K_m and V_max values were estimated from a L-B plot.

X-ray structure determination of HIV-1 RT mutants.

The samples were further loaded on HiLoad 16/600 Superdex 200-pg size exclusion columns (GE Healthcare) with a buffer containing 10 mM Tris-HCl (pH 8.0) and 50 mM NaCl. Native PAGE analysis with Tris-borate buffer was performed to assess the constituents and purity of each eluted fraction, according to the previously described procedure (20). The highly purified RT-DNA complex was concentrated to approximately 20 mg/ml using the Amicon Ultra filtration device (Millipore). Crystals of HIV RT^3MB/M184V-DNA and RT^{3MB/M184V/F160M}-DNA were obtained with the hanging-drop vapor diffusion technique at 20°C, using a reservoir solution consisting of 20 mM bis-Tris (pH 6.0), 20 to 40 mM diammonium hydrogen citrate, 2.5 to 3.2% polyethylene glycol (PEG) 6000, 2.4% sucrose, and 4.8% glycerol. The crystals appeared 48 h after crystallization setup and were grown within 2 weeks to a maximum size of approximately 0.5 by 0.5 by 0.2 mm³. The crystals were briefly soaked in reservoir solution supplemented with 10% PEG 6000, 25.6% glycerol, and 2.5 mM dGTP/dCTP and were then flash-cooled using liquid nitrogen. X-ray diffraction experiments were performed using synchrotron radiation at beamline Photon Factory (PF) BL-1A (Tsukuba, Japan). The raw diffraction data were processed using the programs XDS (44) and AIMLESS (45). The model refinement was started using the program REFMAC5 (46) with the atomic coordinates of HIV-1 RT^3MB-DNA-dCTP (Protein Data Bank [PDB] code 6KDK) (38) as a starting model. Manual model building and corrections were conducted with the program Coot (47). Further model refinement was achieved using the program PHENIX (39) with TLS restraints (one TLS group per chain). Final refined models were evaluated using MolProbity (48). Molecular drawings were prepared using the program PyMol (Schrödinger LLC). Data collection and model refinement statistics are provided in Table 3. The atomic coordinates and structure factor amplitudes for HIV RT^3MB/M184V-DNA-dGTP and HIV RT^{3MB/M184V/F160M}-DNA-dCTP were deposited in the RCSB PDB under accession numbers 7DBM and 7DBN, respectively.