(Received for publication, April 1, 1997)
From the Max Planck Institut für molekulare Physiologie, Abteilung Physikalische Biochemie, Rheinlanddamm 201, 44139 Dortmund, Germany and the ** Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio 44106
The highly conserved primer grip region in the
p66 subunit of HIV-1 reverse transcriptase (RT) is formed by the
12-
13 hairpin (residues 227-235). It has been proposed to play a
role in aligning the 3
-OH end of the primer in a position for
nucleophilic attack on an incoming dNTP. To analyze the importance of
the primer grip for RT function, mutant RTs were used that contain
single alanine substitutions of residues Trp229,
Met230, Gly231, and Tyr232 in the
p66 subunit of the heterodimeric p66/51 enzyme. Steady-state and
pre-steady-state kinetic analyses of the enzymes were performed. All
mutant enzymes revealed reduced polymerase activity. Mutation of Y232A
showed the smallest effect on polymerase function. Equilibrium fluorescence titrations demonstrated that the affinity of the mutants
for tRNA was only slightly affected. However, the affinity for
primer-template DNA was reduced 27-fold for mutant
p66W229A/51 and 23-fold for mutant p66G231A/51,
and the maximal pre-steady-state rate of nucleotide incorporation, kpol, was reduced 27-fold for
p66W229A/51 and 70-fold for p66G231A/51,
respectively. Mutant p66M230A/51 revealed no reduced
affinity for primer-template but showed a 71-fold reduced affinity for
dTTP. Additionally, the mutations Trp229 and
Gly231 affected the stability of the RT heterodimer.
Reverse transcriptase (RT)1 is pivotal to the replication of retroviruses. The RT of human immunodeficiency virus 1 (HIV-1) is a heterodimeric enzyme composed of a p66 and a p51 subunit that share a common amino terminus. The smaller subunit is a cleavage product of p66 and lacks the RNase H domain (1, 2). RT is a multifunctional enzyme that possesses a DNA-dependent and an RNA-dependent DNA polymerase activity and an RNase H activity that cleaves RNA in RNA/DNA hybrids. In addition to these activities it uses a host-derived tRNA3Lys molecule as a primer for the first round of polymerization. Furthermore, it has been shown that HIV RT is only active as a dimeric enzyme and that the monomeric subunits show no catalytic activities (3, 4).
For HIV-1 RT and for other DNA polymerases, a polymerization mechanism has been proposed where the binding of primer-template precedes the binding and incorporation of the first dNTP (5). A model for the mechanism of RT polymerization is shown in Scheme 1. In this model, the first step of the reaction is thought to be the formation of a collision complex of RT and DNA that is concentration-dependent. This is succeeded by a conformational change to form a tight complex, which in turn is a prerequisite for dNTP binding. A second conformational change is necessary after the initial binding of the dNTP to form a tight ternary complex. This was identified as the rate-limiting step prior to nucleotide incorporation (6).
![]() |
![]() |
![]() |
SCHEME 1. Model for the kinetic mechanism of polymerization by RT. The asterisks indicate different conformational states of the RT.
One of the highly conserved motifs in retroviral RTs spans
the region from residue 220 to 243 (7). Part of this region in p66
constitutes the "primer grip" and forms a hairpin structure that
ranges from residue 227 to 235 (12-
13 hairpin, adopting the
nomenclature of Ref. 8). The corresponding residues in the p51 subunit
are unraveled in the crystal structure (9). In the RT/DNA cocrystal the
primer grip is located in close vicinity to the polymerase active site
and was therefore proposed to be important for maintaining the 3
-OH
end of the primer in the appropriate orientation for nucleophilic
attack on an incoming dNTP (8, 9). Especially the amino acids in the
loop connecting
12 and
13 (Met230 and
Gly231) and the residues adjacent to the loop, namely
Trp229 at the end of
12 and Tyr232 at the
beginning of
13, are close to the nucleic acid and could play a role
in orienting the DNA primer. The analysis of regions that are necessary
for RT function may contribute to the design of more efficient
inhibitors targeting essential residues.
To investigate the importance of the primer grip region for RT function, mutant RTs were used that contain single alanine substitutions of residues Trp229, Met230, Gly231, and Tyr232 only in p66. The mutant p66 polypeptides were reconstituted into heterodimer in vitro with a wild type p51 subunit (10-12). Pre-steady-state and steady-state kinetic methods were used to analyze tRNA and primer-template binding, as well as nucleotide binding and incorporation of the mutant enzymes.
Selectively mutated p66/51 HIV-1 RT and WT RT were
prepared by in vitro reconstitution as described previously
(10, 11, 13). Briefly, biomass from Escherichia coli
cultures expressing the mutant p66 polypeptide was mixed with that of a
strain expressing His6-p51 and co-homogenized. The high
speed supernatant was loaded on a Ni2+-nitrilotriacetic
acid-Sepharose (Qiagen) column and eluted with a 0-0.5 M
imidazole gradient. RT-containing fractions were pooled, dialyzed, and
applied on a DEAE-Sepharose column. The flow-through of the column was
then applied onto a HiTrap SP column (Pharmacia Biotech Inc.) to
separate the dimeric fraction from excess His6-p51 and
eluted with a 0-0.5 M NaCl gradient. Samples were dialyzed against a buffer containing 50% glycerol and stored at 20 °C. The
purified RTs were free of nuclease contamination.
RT buffer contained 50 mM Tris-HCl, pH 8.0, 6 mM MgCl2, 0.05% (v/v) Triton X-100, 5 mM NaCl, and 5 mM dithiothreitol. Annealing buffer consisted of 20 mM Tris-HCl, pH 7.5, and 50 mM NaCl. Reaction buffer contained 50 mM Tris-HCl, pH 8.0, 10 mM MgCl2, 5 mM KCl, and 1 mM dithioerythritol. HPLC buffer consisted of 50 mM Tris-HCl, pH 8.0, and 150 mM KCl. Urea-loading buffer contained 7 M urea in Tris borate/EDTA buffer with 0.1% xylene cyanol and bromphenol blue.
Polymerase Activity DeterminationRT polymerase activity on poly(rA)/oligo(dT)12-18 substrates was measured by a standard assay (30 µl reaction volume) described previously (11) with 5-10 ng of enzyme in RT buffer. Under these conditions, 1 unit of RT activity catalyzes the incorporation of 1 nmol of dTTP into poly(rA)/oligo(dT)12-18 in 10 min.
HPLC Gel Filtration AnalysisHPLC gel filtration analysis was performed using a Superdex 200 HR10/30 column. 5-30 µg of enzyme were loaded onto the column. The column was eluted with HPLC buffer at 0.5 ml/min.
tRNAsIn vitro transcription and purification of tRNA3Lys was performed as described previously (14). E. coli tRNA2Trp was purchased from Sigma.
Fluorescence Equilibrium Titrations with tRNAFluorescence titrations were performed using an SLM Smart 8100 spectrofluorometer equipped with a PH-PC 9635 photomultiplier. Intrinsic fluorescence measurements using the tryptophan fluorescence of the enzymes were performed in reaction buffer in a total volume of 0.7 ml or 1 ml as described previously at an excitation wave length of 295 nm. The emission intensity was measured at 340 nm (slit widths of 1 and 16 nm for excitation and emission, respectively) (14). Values for the dissociation constant (Kd) were determined as described previously (15) using the data-fitting program Grafit (Erithacus Software). A quadratic equation analogous to the one given by Müller et al. (15) was used for the fitting procedure.
OligodeoxynucleotidesOligodeoxynucleotides were
synthesized on an Applied Biosystems 380 B DNA synthesizer and purified
by HPLC on a reversed phase column by standard procedures (15) or by
denaturing polyacrylamide gel electrophoresis (10% acrylamide, 7 M urea) and elution of the desired DNA band. The sequence
of the 18-mer primer oligonucleotide was 5-TCCCTGTTCGGGCGCCAC, and
that of the corresponding 36-mer template oligonucleotide was
5
-TGTGGAAAATCTCATGCAGTGGCGCCCGAACAGGGA. The complementary regions are
equivalent to the primer binding site sequence from HIV-1. The sequence
of the fluorescently labeled primer oligodeoxynucleotide was identical
to the one above with the fluorescent label at the 5
-end. It was
synthesized by coupling the phosphoramidite dye 6-FAM, a fluorescein
derivative, directly to the 5
-end of the oligodeoxynucleotide during
DNA synthesis according to the recommendations of the manufacturer
(Applied Biosystems). For the fluorescently labeled primer
oligonucleotide, 5
-TGTGGAAAATCTCTAGATGTGGCGCCCGAACAGGGA was used as a
template. Primer and template oligonucleotides were annealed by heating equimolar amounts in annealing buffer for 2 min at 90 °C, followed by cooling to room temperature over 1 h in a water bath.
Experiments with the 5-fluorescently labeled
FAM-primer-template (p/t) were performed and evaluated as described
above. The excitation was at 492 nm, and the emission intensity was
measured at 516 nm (slit widths set at 1 and 8 nm for excitation and
emission, respectively). For displacement titrations, a preformed
complex of 30 nM RT and 30 nM of fluorescently
labeled p/t DNA was titrated with unlabeled p/t in reaction buffer. The
increase in fluorescence was used as a signal for displacement of
labeled p/t. Data were analyzed by using the fitting program Scientist
to determine the Kd values for the unlabeled
p/t.
Primer oligodeoxynucleotides
were 5-end-labeled with T4 polynucleotide kinase (New England Biolabs)
using 7 µCi of [
-32P]ATP (3000 Ci/mmol) (Hartmann
Analytic, Braunschweig) per 100 pmol of DNA for 1 h at 37 °C.
The primer was separated from nucleotides by purification over a
NucTrap Probe Purification Column (Stratagene) according to the
procedure described by the manufacturer. Concentrations were determined
by thin layer chromatography using a phosphor imaging device (Bio-Rad)
for quantification.
Rapid quench experiments were carried out in a rapid quench apparatus built by KinTek Instruments (University Park, PA) in reaction buffer. 15 µl of a preformed complex of 100 nM RT or higher concentrations as indicated for the mutants in the figures and p/t DNA (200 nM) were rapidly mixed with 15 µl of dTTP (up to concentrations of 5 mM). Reactions were quenched at various time points with 0.6% trifluoroacetic acid. The products were analyzed on denaturing 10% polyacrylamide gels with 7 M urea. Quantification of the data was achieved by phosphor imaging. All concentrations reported are final concentrations after mixing in the rapid quench apparatus. Data were fitted to a burst equation (single exponential plus slope) using the program Grafit. The pre-steady-state rate is given by the exponential (6). The burst amplitude corresponds to the amount of RT bound to p/t at the beginning of the reaction (t = 0).
Affinities for dTTP were determined by the dependence of the
pre-steady-state burst rate on the dTTP concentration. The RT-p/t complex (200 nM p/t with RT as indicated in Fig. 5 for the
various enzymes) was rapidly mixed with increasing concentrations of
nucleotide and quenched after t1/2 of the maximal
pre-steady-state rate. The corresponding rates were calculated from the
elongated fraction of the primer by converting the single exponential
equation into k = ln((1
fraction of elongated
primer/burst amplitude))/t1/2. The observed rates
were then plotted against the nucleotide concentration, and the
Kd for dTTP was calculated by fitting the data to a
hyperbola using the program Grafit.
Primer Extension Assays
primer extension assays were
performed with 200 nM radioactively labeled p/t and 100 or
450 nM dimeric RT enzyme at 25 °C in reaction buffer in
a total reaction volume of 10 µl. Reactions were started by adding
dNTPs as indicated in Fig. 6 and were stopped after 5 min by the
addition of 10 µl of urea-loading buffer.
To obtain a first impression of whether the introduction of mutations into the primer grip region has an impact on RT function, we assayed the RNA-dependent DNA-polymerase activity of the reconstituted wild type and mutant enzymes on the homopolymeric substrate poly(rA)/oligo(dT)12-18. Table I shows that the RT activity of all mutants is reduced. The most severe reduction of RT polymerase activity can be detected with p66W229A/51. Previous primer extension experiments with a synthetic 36/71-mer p/t DNA had already indicated that this mutant is strongly impaired in polymerase activity (11). However, no detailed kinetic analyses of the polymerase function of the mutants were available.
|
The reduction of RT activity of the mutants may derive from different mechanisms. First, the mutation can have an effect on the dimer stability of p66/51. Second, the binding affinity to the nucleic acid substrate can be reduced; and third, the mutant enzyme can have a defect in nucleotide binding and incorporation (see Scheme 1). In the following sections we describe the analysis of these possibilities.
Effect on Dimer StabilitySince monomeric HIV-1 RT subunits do not show catalytic activity (3), it is important to determine the dimer content of the purified enzymes, especially since the enzymes used were reconstituted from subunits expressed separately.
N-terminal tagging of the p51 subunit with six additional histidines
allows purification by metal chelate chromatography. The wild type or
mutated p66 subunit, which in our case does not have a histidine
extension, can only be retained on the column if it interacts with p51
(10). Although metal chelate chromatography (16) of the in
vitro reconstituted HIV-1 RT mutants is a powerful method that
permits purification of the dimer fraction of an enzyme, it does not
give much information on the stability of the purified dimer. If the
dimer is not stable, a dimer/monomer equilibrium will be reestablished
after purification of the dimeric enzyme fraction. To investigate this
possibility, we determined the dimer content of the RT enzymes by HPLC
size exclusion gel filtration (Fig. 1) after the
proteins had been purified and dialyzed against storage buffer. The
results, shown in Table II, indicate that the mutants
p66W229A/51 and p66G231A/51 are not completely
dimerized. p66W229A/51 is only about 35% dimeric, and the
dimer content of p66G231A/51 is about 60%. Therefore, the
low polymerase activity of mutants p66W229A/51 and
p66G231A/51 is partly due to the low dimer content of the
purified stock solutions. This result indicates that the dimer, due to
its instability, tends to monomerize during the purification procedure.
To account for this result, for all the analyses shown below only the
dimer content of the enzymes was considered for calculations. The dimer interface of the mutants p66M230A/51 and
p66Y232A/51 appears to be intact, since we could not detect
appreciable amounts of monomer.
|
To examine whether the enzymes p66W229A/51 and p66G231A/51 are stable under the conditions used in our kinetic experiments (25 °C for up to 1.0 h) we isolated the dimer fraction of mutants p66W229A/51 and p66G231A/51 by separation on an HPLC gel filtration column. This is shown in Fig. 1, C and D, for p66W229A/51. Immediately after collection, an aliquot of the collected dimer fraction was loaded again on the HPLC gel filtration column to show the dimer content of the isolated peak fraction (Fig. 1C). Another aliquot was incubated for 1.0 h at 25 °C and then analyzed again by HPLC gel filtration (Fig. 1D). The mutant enzymes remained dimeric for up to 4.5 h of incubation at 25 °C (data not shown). This suggests that the proteins do not dissociate to a great extent under the conditions in which our experiments were performed. However, the wild type and mutant enzymes started to dissociate appreciably after 15 min at 37 °C, a temperature that in general is used for many enzyme assays with RT (data not shown).
tRNA Binding Is Not ImpairedtRNA3Lys is the natural
replication primer of HIV-RT. We have shown recently that the intrinsic
fluorescence of RT can be used to monitor p/t and tRNA binding (14,
17). Heterodimeric RT possesses 37 tryptophan residues. Binding of tRNA
or p/t to RT leads to a quenching of the tryptophan fluorescence of the
protein that can be used to determine the dissociation constant
(Kd). However, comparison of binding affinities for
WT RT to various tRNA species showed that RT does not bind
tRNA3Lys specifically; instead, for all
tRNA species tested, similar Kd values were
determined (14). Therefore, to determine whether the mutant RTs are
still able to bind tRNA we used E. coli
tRNA2Tyr for the intrinsic fluorescence
titrations (Fig. 2). The results presented in Table
III indicate that mutations in the primer grip do not
affect tRNA binding. This was confirmed by titration of wild type and
mutant RT p66W229A/51 with synthetic
tRNA3Lys. The Kd
values obtained for synthetic tRNA3Lys
were 11 nM for the wild type and 7 nM for
p66W229A/51. Our results demonstrate that the primer grip
region is not essential for tRNA binding.
|
The affinity of the
mutant RTs for p/t DNA was determined by equilibrium fluorescence
titrations using a 18/36-mer p/t DNA substrate labeled at the 5-end of
the primer with the fluorescein derivative FAM. The
Kd values obtained for the wild type enzyme with
this method are similar to those obtained by intrinsic fluorescence
measurements (Kd = 1.8 nM; Ref. 17).
Since some of the mutants have rather high Kd values
for p/t, for practical reasons we used the extrinsic fluorescence
determination for standard procedures because the signal change is
larger and more stable (Fig. 3). The binding affinities
for p66W229A/51 and p66G231A/51 were about
23-27-fold lower than for wild type, whereas the binding affinity for
mutant p66Y232A/51, which possessed the highest RT
polymerase activity (Table I), was only about 8-fold lower.
Interestingly, the mutation of p66M230A/51 does not appear
to have an effect on p/t binding. This result indicates that this
mutant, which retains only 11% of the WT RT polymerase activity must
be impaired in one of the steps following p/t binding. However, to
exclude the possibility that the fluorescence label has an influence on
the affinity of the mutant for p/t, we performed displacement
titrations. A preformed complex between RT and the fluorescent p/t was
titrated with the unlabeled 18/36-mer DNA. Displacement of the labeled
p/t is dependent on the Kd of both p/t DNAs.
Analysis of the displacement data for the wild type and
p66M230A/51 leads to Kd values for the
unlabeled p/t species of 5.2 and 5.3 nM, respectively,
which indicates that the label does not influence the binding
affinities of the mutant and confirms that the affinity of wild type
and p66M230A/51 for p/t is unchanged.
Kinetics of Nucleotide Incorporation Are Changed
To analyze
the kinetics of nucleotide binding and incorporation we used quench
flow methods (6). Elongation of the 18/36-mer DNA to the 19/36-mer
product was achieved by rapid mixing of the preformed enzyme-DNA
complex with various concentrations of the next correct deoxynucleotide
dTTP in a quench flow apparatus. Fig. 4 shows a typical
time dependence of dTTP incorporation at a concentration of 1 mM. All enzymes incorporated dTTP rapidly in a burst of
polymerization followed by a slower steady state reaction (slope of the
fitted curves), which is dependent on the off-rate of p/t binding. The
burst amplitude corresponds to the amount of RT bound to p/t at the
beginning of the reaction. Both the burst amplitude and the burst rate
are, among other things, dependent on the affinity of the enzyme for
p/t. Mutants having high Kd values for p/t DNA
(p66W229A/51, p66G231A/51, and
p66Y232A/51) showed smaller burst amplitudes and reduced
burst rates. For p66W229A/51 and p66G231A/51,
the enzyme concentration had to be increased to detect the burst (Fig.
4B). To reach the maximal pre-steady-state rate of nucleotide incorporation (maximal burst rate,
kpol), elevated dTTP concentrations of up to 5 mM had to be used for mutants p66W229A/51,
p66M230A/51, and p66G231A/51, indicating a
lower affinity for nucleotides (see below). With WT RT, the maximal
rate is already achieved at a dTTP concentration of 150 µM. The kpol values obtained for
dTTP incorporation for all enzymes are shown in Table
IV.
|
The method described here for the determination of kpol can also be applied to determine the apparent affinities of dTTP to the RT-DNA complex by measuring the dependence of the pre-steady-state burst rate on the concentration of dTTP (6). The rates increase with increasing dTTP concentrations used. The observed rates at different dTTP concentrations were plotted against the dTTP concentration, and the dissociation constant (Kd) for dTTP was calculated by fitting the data to a hyperbola (6) (Fig. 5). Again, mutant p66Y232A/51 was the least impaired. In addition to their low affinity for p/t DNA, the affinity of mutants p66W229A/51 and p66G231A/51 for dTTP was reduced 20- and 36-fold, respectively. Our data clearly show that the major effect of the mutation in p66M230A/51 is a severe increase (71-fold) in the Kd for dTTP. A summary of the results is given in Table IV.
Polymerization on a Heteropolymeric DNA SubstrateFig. 6 demonstrates that the results of the equilibrium titration and single turnover experiments can be confirmed in a polymerization assay. A primer extension assay was performed using the same radioactively labeled p/t (200 nM) as substrate as applied in the quench flow experiments and increasing amounts of dNTPs and RT. In this experiment, p66Y232A/51 was only slightly less active than the wild type enzyme even when low dNTP concentrations of 50 µM were used. p66M230A/51 needs higher dNTP concentrations to perform efficient polymerization to the end of the template. With mutants p66W229A/51 and p66G231A/51, as expected from the results shown above, both an increase in the enzyme and dNTP concentrations is necessary for proficient polymerization.
The aim of this work was to characterize the importance of the
primer grip of HIV-1 RT by kinetic analyses of four mutant RT enzymes.
Comparison of the primer grip region of HIV-1 RT with RTs from other
retroviruses possessing a p66/51 structure (feline immunodeficiency
virus, simian immunodeficiency virus, equine infectious anemia virus,
and HIV-2) shows that Pro225, Pro226,
Trp229, Gly231, and Leu234 are
invariant, indicating an important role of these residues for enzyme
function (7). Residues Trp229 and Leu234 of
HIV-1 p66 also participate in the binding of non-nucleoside inhibitors
(NNIs). Although in the cocrystal structure Trp229 and
Leu234 are in close contact with the inhibitor nevirapine
(3.8 Å), no mutant resistant to NNIs has been found with changes in
these amino acid residues (18-21). Targeting of residues in the NNI
binding site that are essential for RT function could be important for the design of novel inhibitors that do not lead to resistant virus easily.
Previous biochemical analyses with p66W229A/51 already
showed that this enzyme is strongly impaired in polymerase function
(11). The instability of the dimer we observed with
p66W229A/51 and p66G231A/51 is somewhat
surprising, since the amino acids constituting the primer grip are not
directly involved in dimer formation. However, it has been demonstrated
that a p66 subunit with a mutation of Leu234 to Ala
completely fails to yield heterodimeric RT (12). Crystallographic data
from unliganded HIV-1 RT show that the side chains of
Trp229, Met230, and Tyr232 are
oriented into the core region between the primer grip and the
polymerase active site (22). We therefore assume that the mutations in
Trp229, Gly231, or Leu234
indirectly influence dimerization, for example by changing or preventing interactions with amino acids that contact the dimer interface. Fig. 7 shows possible interactions of the
amino acids of the primer grip with surrounding residues as determined
by the RT crystal structure of Rodgers et al. (22). A
contact of Trp229 with Pro95 (distance ~3.79
Å) in the palm subdomain of p66 is possible by van der Waals
interactions. Pro95 is also involved in binding NNIs (19,
20, 23). Pro95 in turn may interact with Asn137
and Glu138 in the finger subdomain of p51 (distance ~3.81
Å). By changing the bulky Trp229 into Ala or by exchanging
Leu234 to Ala, a "deficiency of atoms" occurs in the
pocket harboring these amino acids. This could lead to a conformational
change of adjacent residues to fill the created space. As a
consequence, the loop (residues 135-140) between 7 and
8 (shown
in light purple in Fig. 7) of p51 containing
Asn137 and Glu138 (highlighted in
yellow in Fig. 7) and reaching into p66 is not able to
contact the corresponding residues in p66 properly. Furthermore, this
effect could lead to displacement of the
7-
8 loop of p51 away
from the pocket, thus destabilizing the dimer interface. At present we
are investigating the role of this loop region of p51 in dimer
formation. Interestingly, the only resistance mutation against NNIs
found in p51 is located in this region (residue Glu138
Lys), indicating a function of this loop in binding the NNI (24,
25).
For residue Gly231, the situation might be somewhat
different. It is located in the loop connecting 12 and
13 and
therefore might have an important architectural role for the stability
of the
12-
13 hairpin. Exchanging this residue against the bulkier alanine could affect the stability of the hairpin, which in turn by a
change in conformation of surrounding residues could have similar
consequences for dimer stability as already described above.
Additionally, our results demonstrate that the conformational changes
described influence the polymerization mechanism. Our experiments
indicate a function for Trp229 and Gly231 in
p/t binding and dNTP incorporation. Crystallographic data have
implicated the primer grip in orienting the primer terminus for
nucleophilic attack on an incoming dNTP (8). Our results with the
mutants p66W229A/51 and p66G231A/51 imply that
these enzymes are unable to align the primer properly, leading to a
reduced affinity for p/t and dNTP and to a reduced incorporation rate
for dNTP. However, the mutation in Met230 clearly shows
that the two effects, i.e. p/t binding and dNTP incorporation, can be separated. With mutant p66M230A/51,
dNTP binding and incorporation are impaired, but we do not find a
significant reduction of the affinity for p/t DNA (Table III). Modeling
experiments that dock a dTTP into the polymerase active site of the
HIV-1 RT/DNA/Fab cocrystal suggest that -strand
6 (residues
105-118), the
9-
10 hairpin (residues 178-191), and the region
11a-
11b (residues 214-222) of p66 might be involved in dNTP
binding (9, 26). This allows positioning of the
- and
-phosphate
of the dTTP to the active site residues Asp185 and
Asp186. Direct interaction of the primer grip hairpin with
a dNTP would be unfavorable, since the base of the dTTP in this case
could not stack onto the terminal base of the primer. Therefore, it is
reasonable to assume that the amino acids of the primer grip are not
directly involved in dNTP binding. Instead, the changes in the
orientation of the primer induced by the mutations probably lead to a
less favorable positioning of the
-phosphate of the dNTP for attack
by the 3
-OH of the primer terminus. However, minor changes of the
primer orientation do not necessarily lead to a reduced affinity of the
mutated enzyme for p/t but can have an affect on the stacking of the
incoming nucleotide onto the terminal base of the primer. This is
probably the case with mutant p66M230A/51. This effect
could also lead to a reduced fidelity of the mutant enzymes. In fact,
preliminary results with mutant p66W229A/51 indicate that
the fidelity of this enzyme has also changed (data not shown).
Comparison of the results obtained for tRNA binding with those of p/t
binding demonstrate that in the binary complex the primer grip region
is not important for the affinity of RT for tRNA. We have shown
previously by displacement titrations that a p/t DNA can be displaced
by tRNA3Lys and various other tRNAs
(14). This implies that tRNA and p/t bind to the same or at least
partially overlapping sites. Together with the data presented here,
these experiments suggest that the binding sites for tRNA and p/t are
overlapping but not identical. A function of the primer grip in tRNA
binding must be expected when the tRNA is hybridized to the primer
binding site on the viral RNA, since here its 3-OH end is used as a
primer for the start of minus strand synthesis.
Previous results indicate that introduction of the mutation G231A into the virus leads to infectious particles in cell culture assays, whereas after introduction of W229A infectious virus is not formed (11, 12). Our results show that in vitro both mutants are significantly impaired in polymerase function. The reason for this discrepancy is not clear, since most parameters we have measured for these two mutants (Tables I, II, III, IV) are similar. The only major difference (Table I) is in the steady-state activity assay, in which p66W229A/51 is considerably worse than p66G231A/51. This is initially surprising, since the single nucleotide incorporation kinetics of p66W229A/51 are better than those of p66G231A/51. We attribute this discrepancy to the poorer dimerization properties of p66W229A/51, which probably lead to extensive dissociation under the low concentration conditions of the steady-state assay. In contrast, the quenched flow results were obtained at much higher protein concentration and are corrected for dimer content. We therefore conclude tentatively that impairment of heterodimer formation may be responsible for the lack of production of infectious virus for this mutant. Another possibility that has not been analyzed yet are compensatory mutations that could have occurred during the cell culture experiment to overcome the impairments of p66G231A/51. In the absence of this, production of infectious virus after introduction of the G231A mutation remains a surprising result in view of the significantly impaired kinetic properties.
The observation that the mutation Y232A leads to a decrease in infectious particles in cell culture assays, although in in vitro assays the polymerization activity is not as drastically reduced as with mutant p66G231A/51, is probably due to the impaired RNase H activity of p66Y232A/51 demonstrated previously (12).
One general result of the work presented here is that single amino acid changes in HIV-1 RT can have far reaching and unexpected global effects on enzyme structure and function (e.g. dimerization), which can make interpretation of results difficult and misleading unless these factors are specifically taken into account. Such unexpected results can, however, be useful. As an example, the information obtained on the effects of mutations of the primer grip region on dimerization suggest that this region of the molecule might be included in the search for approaches to inhibition of dimerization.
We thank Dr. Jochen Reinstein for helpful discussions, Manfred Souquet for performing the HPLC gel filtration analyses, and Martina Wischnewski and Karin Vogel-Bachmayr for excellent technical assistance.