(Received for publication, July 2, 1996, and in revised form, November 20, 1996)
From the Center for Molecular Science, The University
of Texas Medical Branch, Galveston, Texas 77555 and the
§ Department of Chemistry, Vanderbilt University,
Nashville, Tennessee 37235
To determine the effect of various stereoisomers
of benzo[a]pyrene-7,8-dihydrodiol 9,10-epoxide (BPDE) on
translesion bypass by human immunodeficiency virus-1 reverse
transcriptase and its -helix H mutants, six 33-mer templates were
constructed bearing site- and stereospecific adducts. This in
vitro model system was chosen to understand the
structure-function relationships between the polymerase and damaged DNA
during replication. Comparison of the replication pattern between wild
type human immunodeficiency virus-1 reverse transcriptase and its
mutants, using primers which were 3
to the lesion, revealed
essentially similar patterns. While these primers terminated with all
three of the C10R and two of the C10S
BPDE-adducted templates 1 base 5
and 1 base 3
to the damaged site
respectively, (+)-anti-trans-(C10S)
BPDE-adducted DNA alone permitted the formation of full-length
products. Utilization of a primer with its 3
-hydroxyl 1 base beyond
the lesion resulted in full-length products with all the
C10S BPDE-adducted templates and the
(
)-syn-trans-(C10R)-BPDE-adducted template,
following replication with either the wild type or mutant enzymes.
However, the other two C10R BPDE-adducted templates failed
to allow any primer extension, even with the wild type enzyme. Although
T·P depletion studies further confirmed the differential primer
extension abilities using the C10R and C10S
adducted templates, their binding affinities were similar, yet distinct
from the unadducted template.
Benzo[a]pyrene is a ubiquitous by-product of
incomplete combustion and is one of the most potent carcinogens known
(1-4). In mammalian cells, it is metabolized by cytochrome P-450 and epoxide hydrolase to a variety of products, including the ultimate DNA-damaging agent BPDE1 (5-9). The
carcinogenic and mutagenic effects of BPDE depend on covalent
modification of DNA and subsequent interactions with various
polymerases. There are a large number of potential adducts which arise
by the cis and trans openings of the epoxide
group by the exocyclic N2, O6, or N-7 positions
of guanine or N6 position of adenine or N4
position of cytidine. Structure-activity studies in several polycyclic aromatic hydrocarbon systems have shown that closely related positional or stereoisomers have dramatically different tumorogenicities (10). In
this context, oligonucleotides containing site-specifically placed
adducts are powerful tools for exploring how individual chemical
lesions, formed in DNA, are converted into mutations, especially those
in critical target genes such as cellular proto-oncogenes or tumor
suppressors (11, 12). Activation of certain members of the
ras proto-oncogene family (13-15) is a result of mutations predominantly in the codons for amino acid residues 12 or 61 (16). Previous data derived from a model prokaryotic system concerning the
mutagenicity of six site- and stereospecific BPDE lesions placed on
adenine N6 at position 2 of N-ras codon 61 revealed a narrow range of mutation frequencies that were exclusively A
G transitions (17).
Furthermore, there are several reports to show that the chirality of
adducted molecules plays a major role in influencing biological
processes, such as exonucleolytic processing and DNA replication
(18-20). Structural data on double-stranded oligonucleotides adducted
with BPDE at N2 of guanine revealed that the pyrenyl moiety
of (+)-anti-trans-isomers with the S configuration at
C10 is directed toward the 5 end of the modified strand.
In contrast, the pyrenyl moiety of the (
)-anti-trans-isomer (C10R configuration) is
oriented toward the 3
end of the modified strand (18). However, NMR
studies on duplexed oligodeoxyribonucleotides containing these adducts on N6 of adenine showed that the C10R adducts
were partially intercalated from the 5
side, and the C10S
adducts, although existing in more than one conformation, appear to be
mainly intercalated from the 3
side (21, 22). In agreement with the
NMR studies, were the in vitro replication studies, using
seven polymerases that included both prokaryotic and eukaryotic
sources that suggested the C10R and C10S
adducts were oriented to the 5
and 3
direction on the template,
respectively (23).
Extensive in vitro studies have been reported on the
replicative fate of DNAs damaged by bulky adducts (24-31). In
particular, BPDE-adducted templates exhibit replication blockage
opposite and/or one base 3 to the adduct, with translesion synthesis
rarely occurring (20, 26, 27, 31-34). In order to elucidate the molecular mechanisms by which polymerases could bypass a lesion site,
and to gain an understanding of the structures within the polymerase
that govern translesion synthesis, the study of enzymatic functions
such as fidelity and regulation of DNA replication is critical. HIV-1
reverse transcriptase (HIV-1 RT) is a hypermutable enzyme that
catalyzes the addition of approximately 20,000 nucleotides with
extremely low fidelity during replication, resulting in mutations at a
frequency of 1/1700 nucleotides inserted (35-37). Both the crystal
structure and the co-crystal structure with duplex DNA are known and
numerous site-specific mutants that were designed with
structure-function aspects in mind are readily available (38-41). This
facilitates the choice of HIV-1 RT and its mutants as a unique system
in which to monitor polymerase stalling as a diagnostic for adduct
directed perturbations.
Preliminary primer extension studies using HIV-1 RT revealed that the
C10R and C10S BPDE-adducted templates gave
different termination sites, but none formed full-length products,
except (+)-anti-trans-BPDE (20). Therefore, the present
study was undertaken in order to understand if termination occurred due
to pausing of the polymerase in the vicinity of the lesion or due to
dissociation of the enzyme from the T·P at this point and subsequent
inability to re-initiate synthesis. Furthermore, an attempt has been
made to characterize HIV-1 RT and two mutants of -helix H (G262A and W266A) by determining their polymerization behavior on site- and stereospecifically adducted BPDE templates. Because the amino acid
residues 262 and 266 are in contact with the duplex region in the
vicinity of the DNA minor groove, relatively close to the 3
terminus
of the primer, these two mutants were utilized in the present study.
Although these polycyclic aromatic hydrocarbon adducts are not likely
to be encountered by HIV-1 RT in vivo, they serve as an
ideal model for examining the effect of bulky adducts on DNA processing
by polymerases.
The p66/p51 heterodimeric form of HIV-1 RT was
expressed in Escherichia coli from a plasmid containing the
precise coding region from HXB2 and was a generous gift from Dr. S. H. Wilson (University of Texas Medical Branch, Galveston, TX) (42). The alanine-substituted mutant forms of HIV-1 RT, G262A and W266A, were
also gifts from Drs. S. H. Wilson and T. A. Kunkel (38). The T4
polynucleotide kinase and T4 DNA ligase were obtained from New England
Biolabs Inc. (Beverly, MA). The six stereochemically defined BPDE
adducts were constructed on adenine N6 at position two of
N-ras codon 61 within an 11-base oligodeoxynucleotide by the
post-oligomerization strategy (43). The six lesions examined herein
were (+)- or ()-anti-trans-, (+)- or
(
)-syn-trans-, and (+)- or (
)-anti-cis-BPDE
adducts (Fig. 1). Deoxynucleoside triphosphates (100 mM)
were obtained from Pharmacia Biotech Inc., and
[
-32P]ATP (3000 Ci/mmol) utilized for 5
end labeling
of oligodeoxynucleotides was purchased from DuPont NEN.
Synthesis of Primers and Construction of BPDE-adducted 33-mer Templates and Their Subsequent Purification
Oligonucleotides of
17, 27, and 29 bases that were used as primers for DNA replication
studies were synthesized on an automated Applied Biosystems 394 DNA/RNA
synthesizer (Perkins-Elmer/Applied Biosystems Division, Foster City,
CA), according to the manufacturer's protocol (Fig. 2). The six
BPDE-adducted 11-mers were analyzed for their purity on 15%
polyacrylamide sequencing gels prior to ligation with a 22-mer, using a
27-mer scaffold (23, 43). The primers and the adducted templates were
purified on 10% polyacrylamide gels containing 8 M urea.
The corresponding bands were excised, eluted in water,
ethanol-precipitated, and excess urea eliminated by six to eight washes
with cold absolute alcohol. Purified oligonucleotides and adducted
templates were dissolved in water and the concentrations calculated by
absorbance measured at 260 nm.
Primer Extension Assays and Calculations for Termination Efficiencies
In vitro replication assays using HIV-1
RT and its mutants, G262A and W266A, employed the following reaction
mixture, with the final concentrations being 0.1 mg/ml bovine serum
albumin, 1 mM dithiothreitol, 1.25 mM ATP, 0.3 mM each of the four deoxynucleotide triphosphates (dNTPs),
33 mM Tris-OAc (pH 7.8), 66 mM KOAc, and 10 mM MgOAc. To this reaction mixture were added the adducted or unadducted templates, which had been previously annealed to one of
three primers (17-, 27-, or 29-mer) that were 5-labeled with
[
-32P]ATP (3000 Ci/mmol). The T·P ratio was 5:1 (250 fmol:50 fmol), and 25 fmol of enzyme were added per reaction in a total
volume of 10 µl, 37 °C. At the end of 1 h, the reaction was
terminated by the addition of 5 µl of stop buffer (95% formamide, 20 mM EDTA, 0.05% bromphenol blue, and 0.05% xylene cyanol).
Primer extension products were separated on 15% polyacrylamide
sequencing gels (8.3 M urea) by electrophoresis at 2000 V
for 4 h and visualized by autoradiography using Hyperfilm MP
(Amersham Corp.). Primer annealing was determined by separation of the
annealed mixture on a 10% native polyacrylamide gel, and in each case,
more than 99% of the primer existed as a T·P complex prior to the
addition of the enzyme.
Quantitation of bands were performed by PhosphorImager analysis (Molecular Dynamics, Sunnyvale, CA). The percentage of partial and full-length products were determined in relation to the amount of unextended primers.
T·P Depletion Assay and Binding AssayThe conditions utilized in these assays were similar to those described for the primer extension studies, except that there was an initial extension of either unadducted or adducted templates using an unlabeled 29-mer primer. After an initial incubation of 30 min at 37 °C, unadducted templates annealed to labeled 29-mer primers were added in 5-fold excess, to serve as a hot chase and to monitor for the presence of free enzyme. All four dNTPs (0.3 mM each) were added to the reaction mixture prior to addition of 40 fmol of enzyme. A similar experiment was designed to measure the initial binding of the polymerase to the various substrates. In these experiments each of the three templates was annealed with unlabeled 29-mer primer (5:1 ratio, respectively) and incubated with HIV-1 RT (10:1, primer:enzyme) for 20 min in the absence of dNTPs. Then a 5-fold excess of template, and 17-mer 32P-labeled primer, were added and the reactions initiated by the addition of 0.33 mM dATP and 0.33 mM dGTP only. Aliquots were taken at 0, 0.5, 2, 10, and 20 min and analyzed as described previously.
The interactions between BPDE-adducted templates and polymerases under in vitro conditions most often result in replication termination in the vicinity of the DNA base lesion. The goal of this study was to determine whether correlations exist between three properties of polymerase·DNA interactions: (i) the extension of a primer on the damaged template and the orientation of the adduct; (ii) the elongation of a primer on an adducted template and the binding of the polymerase to the T·P complex; and (iii) the ability to form full length products and the specific identity of amino acid residues in the polymerase.
In order to study the structure-function interactions between
polymerases and damaged DNA during replication, HIV-1 RT was chosen
because of its processive nature, low fidelity, and lack of 3
5
exonucleolytic activity (36, 37). Furthermore, the resolution of its
crystal and co-crystal structures gives an insight into the various
subunits and subdomains and their probable functions (44, 45). HIV-1 RT
is a heterodimer with two subunits, p66 and p51, of which both the
polymerase and RNase H activities are limited to the p66 subunit.
Furthermore, p66 forms a nucleic acid binding cleft constituting the
fingers, palm, and thumb subdomains (45). The amino acid residues from
257 to 266 are highly conserved and play an important role in catalytic cycling (41). These constitute part of the
-helix H of the thumb
subdomain and play an important role in the movement of the T·P
strand during polymerization. Both the
-helices (H and I) track
along the DNA 4 to 5 nucleotides from the 3
primer terminus, with the
-helix H in close contact with the primer strand (45). The selection
of alanine substitution mutants, G262A and W266A, was pertinent because
of their role in reduction of processivity and fidelity. Our
expectation was that the results obtained using these mutants would
shed light on how specific amino acid residues influence the
replication fate of BPDE-adducted templates (Figs. 1 and
2). Although enzyme-substrate interactions occur in the minor groove of
the DNA, perturbations can be translated as alterations in the major
groove, because the two regions are coupled.
In
order to study the factors that influence the replication of adducted
DNAs, three primers were chosen that were 17, 27, and 29 bases in
length with a common 5 end (Fig. 2). These
deoxyribonucleotides position their 3
-OH 11 and 1 base(s) upstream and
1 base beyond the adduct and are indicated by a,
b, and c, respectively (Fig. 3).
The T·P complexes were formed with unadducted DNA (panel
1), and each of the six site- and stereospecifically modified
BPDE-adducted templates (panels 2-7), and primer extension
studies were performed using HIV-1 RT. The C10S and
C10R adducted DNA are grouped as B and
C, respectively.
As shown in the control portion of this experiment (Fig. 3A), when any of the three primers were annealed to unadducted templates, the wild type HIV-1 RT was able to extend these primers to full-length 33-mers and also catalyze a 1 base, blunt-end addition with an efficiency greater than 50%. This non-template-directed extension can possibly be attributed to slippage of the newly synthesized primer strand and the polymerase. However, it seems less likely, because of the absence of homopolymeric or repeated sequences at the end of the single-stranded template (46).
In contrast to the results obtained by using the unadducted DNAs, two
of the C10S adducted templates ((+)-syn-trans-
and ()-anti-cis-) adducts caused the polymerase to
terminate 1 base 3
to the damaged site, using a 17-mer primer
(panel 3, lane a, and panel 4, lane a,
respectively), resulting in the accumulation of products that were 27 bases in length. Little to no extension was visible with the 27-mer
primer (panel 3, lane b, and panel 4, lane b,
respectively). However, by using either of these primers, full-length
products were obtained when an (+)-anti-trans-BPDE-adducted
template (C10S) was utilized for primer extension
(panel 2, lanes a and b). Also in contrast, all adducts
bearing the R configuration at the C10 position (group C)
of the BPDE moiety ((
)-anti-trans,
(
)-syn-trans, and (+)-anti-cis) (panel 5, lanes a and b, and panel 6, lanes a and
b, and panel 7, lanes a and b,
respectively) did not promote termination 1 base prior to or opposite
the lesion, but caused HIV-1 RT to form truncated products 1 base
beyond the lesion, with generally negligible accumulation of
full-length products. These reactions were performed under multiple hit
conditions. However, observations under single hit conditions showed
that both the C10R and C10S adducted templates
formed truncated products that terminated one base prior to the lesion
(data not shown).
The primer extension results obtained using the adducted templates with
a 29-mer primer, which initiates synthesis 1 base beyond the lesion are
presented in lane c of panels 2-7. All the C10S BPDE-adducted templates ((+)-anti-trans,
(+)-syn-trans, and ()-anti-cis-) (panel
2, lane c, panel 3, lane c, and panel 4, lane c,
respectively) allowed for the formation of full-length products when
the lesion was located in the T·P stem. These fully extended products
were formed at levels similar to the unadducted templates. In contrast,
two of the three C10R adducted templates examined allowed
negligible formation of full-length products. The
(
)-syn-trans-BPDE-adducted template (panel 6, lane
c) was unique in that a prominent termination site 4 bases beyond
the site of lesion was observed in addition to a significant amount of
fully extended products. In our subsequent experiments, similar observations were made when 73-mer unadducted and adducted DNAs were
used as templates in which there is a 44-nucleotide single-stranded overhang (data not shown). These results indicate that the orientation of C10S and C10R BPDE adducts in the template
strand has a major influence on protein-DNA interactions, independent
of the length of the single-stranded template overhangs (data not
shown). End effects are considered to play a role in nucleotide
discrimination, especially in regard to the incorporation of
dideoxynucleotides. However, there are reports to suggest that the
catalytic efficiency of nucleotide incorporation is independent of the
length of the template overhang and that the ability to place a
nucleotide opposite the 5
end of the template is the same as it is to
place a nucleotide at any other position (35).
Analyses of the data presented in Fig. 3 revealed
that the most significant contrast between a matched set of
C10S and C10R BPDE adducts was for the (+)- and
()-anti-trans-lesions, such that all primers could be
readily utilized for the (+)-anti-trans, while replication
was blocked 1 base beyond the lesion for the (
)-anti-trans-containing template (panels 2 and
5, respectively). These data raised further questions
regarding the mechanism by which the (
)-anti-trans-lesion
blocked replication, while the (+)-anti-trans-lesion allowed
facile bypass. One possibility for this apparent distinction could be
the inability of the polymerase to bind to the
(
)-anti-trans-BPDE-adducted template, once the primer had
been extended 1 base beyond the lesion site. In order to determine
whether there was differential binding and subsequent dissociation from
a primer that is 1 base beyond the lesion, the following experiment was
designed and carried out (Fig. 4). Three T·P complexes
were formed wherein the template (33-mer) was either unadducted or
contained a (+)- or (
)-anti-trans-BPDE lesion and the
primer utilized (29-mer) was unlabeled. The T·P ratio was 5:1 to
ensure complete primer annealing. Limiting amounts of polymerase were
added in the absence of dNTPs, such that the primer:polymerase ratio
was 10:1. This mixture was allowed to equilibrate initially for 20 min
and then a 5-fold excess of unadducted T (33-mer)·P (17-mer)
complexes were added to each of the three samples, in which the primer
was 32P-labeled. This incubation was carried out in the
presence of dATP and dGTP alone, which allowed replication to occur
exclusively on the labeled 17-mer primer strand, since the immediate
upstream bases on its template were C, T, T, while the immediate bases that would need to be incorporated on the unlabeled 29-mer primer were
A, G, G. Thus, the rate of extension of the labeled primer would be a
measure of the relative binding and dissociation from the original
template primer. These results are shown in Fig. 4, B and
C. In Fig. 4B, lanes 1-3 are the unadducted and
(+)anti-trans- and
(
)-anti-trans-BPDE-containing complexes, respectively, and A-D represent increasing chase times of 0.5, 2, 10, and 20 min, respectively. Quantitative analyses of these data reveal that the
enzyme dissociates more rapidly (approximately 2-fold) from the
unadducted than from either of the adducted DNAs. Furthermore, these
data strongly suggest that the binding of HIV-1 RT to the 29-mer primer
on the C10R BPDE adduct containing DNA is comparable with
that observed on the same primer complexed with the C10S BPDE-adducted template.
To further explore this stereospecific inhibition, a T·P depletion
assay was designed wherein the unadducted, C10S
(()-anti-trans) and C10R
((+)-anti-trans) adducted templates were individually annealed to an unlabeled 29-mer primer in the ratio of 5:1 (Fig. 5A). These T·P complexes were preincubated
for 30 min with HIV-1 RT and all four dNTPs. The initial pre-chase
would allow for synthesis to occur, given that the nucleotide
incorporation and subsequent extension were not hindered. Thus,
following the formation of products, the enzyme would then be free in
solution for further primer extension. The prediction was that the
primers complexed with either the unadducted and C10S
templates would be extended well, allowing the polymerase to be free to
initiate the chase portion of the experiment, while the T·P complex
with the C10R BPDE-adducted template would still have HIV-1
RT bound to it. This prediction was in concurrence with what was
determined experimentally (Fig. 5, B and C).
Furthermore, the rate of primer utilization was indistinguishable when
either the unadducted or C10S BPDE-adducted DNAs served as
templates, whereas the availability of HIV-1 RT for chase synthesis
after preincubation with the labeled primer complexed with
C10R BPDE-adducted template was 3-fold less.
Role of the
In order to
determine whether the minor groove binding -helix H in HIV-1 RT
plays a role in adduct-directed terminations, wild type enzyme was
compared with two
-helix H mutants, G262A and W266A (Fig.
6). These residues have previously been shown to affect
termination probabilities on unadducted templates (38). Panel
1 of Fig. 6 shows data for unadducted templates, while
panels 2-7 show data for each of the six BPDE-containing
DNAs. The lanes a-c refer to wild type, G262A, and W266A
enzymes, respectively.
As shown in Fig. 6 and Table I, all of the
C10S-configured BPDE-adducted templates were readily copied
to full length, albeit at rates much reduced relative to unadducted
templates, by the alanine substitution mutants of the -helix H thumb
subdomain of HIV-1 RT when primed with the +1 (29-mer) primer. However, the overall degree of extension of the primer on the three adducted templates and the unadducted template differed moderately with these
enzymes. As tabulated in Table I, fully extended products with the
C10S adducted templates and unadducted template,
polymerized with the wild type HIV-1 RT, showed a range from 66 to
79%. With the G262A mutant, the values ranged from 25 to 83%, whereas
with the W266A, the values covered a span from 5 to 79%. The
(
)-anti-cis-BPDE-adducted template allowed the least
extension, with each of the three enzymes utilized for polymerization.
In all these reactions, the amount of degraded primers accounted for
less than 1% of the total primer used per lane.
|
Two of the C10R BPDE-adducted templates
(()-anti-trans and (+)-anti-cis), when copied
individually by the two mutant enzymes (G262A and W266A), showed no
polymerase extension at all. However, the
(
)-syn-trans-BPDE-adducted template (also with
C10R configuration), which yielded 24% full-length
products when polymerized by HIV-1 RT, allowed some extension with
G262A and W266A mutant enzymes. The W266A mutant enzyme, lacking
tryptophan, showed strong termination 3 bases 5
to the adducted
site (Fig. 6).
Polycyclic aromatic hydrocarbon carcinogens such as BPDE are known to bind covalently to cellular nucleic acid, leading to mutations and alterations of gene expression that could result ultimately in tumorigenesis and carcinogenesis (47-52). Insight into the precise nature of the interactions between the different stereospecific BPDE-adducted templates and various polymerases has involved numerous in vitro studies, where generally, termination occurs at or one base prior to the site of the adduct (26, 27, 32-34). However, translesion synthesis with these bulky adducted templates has to date been observed with relatively few polymerases (20, 23). HIV-1 RT is one polymerase that was able to bypass the site of lesion.
Our earlier in vitro studies with this polymerase involved
six site- and stereospecific BPDE-adducted templates, each being annealed to a 17-mer primer located 11 bases downstream of the adducted
base. Although some translesion synthesis by HIV-1 RT was observed past
the 5-oriented C10R BPDE adducts, using
11 primers, no
full-length products were formed. Furthermore, the 3
-oriented
C10S BPDE-adducted templates, in general, inhibited polymerization beyond the adduct, and this was unaltered by increased dNTP concentrations (20, 53). In the current study, although
1
primers were designed in an attempt to force polymerization beyond the
site of lesion, there was no change in the pattern of products
produced. In contrast, the utilization of +1 primers led to full-length
products with the C10S BPDE-adducted templates, suggesting
the influence of orientation on primer extension, given that the adduct
extends nearly two bases in the direction of the tilt.
This differential extension of C10R and C10S BPDE-adducted templates initially suggested a differential affinity of polymerase binding. However, based on binding/dissociation assays (Fig. 4C), our data showed clearly that both of these sets of adducted templates have similar binding affinities and dissociation rates that are at least 3-fold slower than the unadducted template. As a corollary, this disparity in primer extension is likely to be attributed to some hindrance for nucleotide incorporation and/or for subsequent elongation. There are reports to show that, if there are no significant differences in polymerase binding affinities, then one can infer that reduced extension efficiencies are caused by an inherent difficulty that the enzyme has in extending a mismatch or a damaged site (54). The hydrophobic properties of the bulky adducts could be one of the contributing factors that stabilize binding.
The 3 orientation of the C10S BPDE adducts on adenine
rationalizes the formation of truncated products prior to the lesion, on utilizing either a 17- or 27-mer primer, given that the
directionality of the polymerase movement is from 5
3
. However,
the replication pattern on the (+)-anti-trans-BPDE-adducted
template with HIV-1 RT allowed for facile bypass and accumulation of
full-length products. Given that there is a wide range over which the
adduct is positioned, whether it is a 5
or 3
orientation relative to
the template, it is possible that this lesion is oriented with a
minimal distortion and inhibitory angle, thus offering very little
interference or hindrance to the polymerase movement along its
substrate. The same reasoning can be applied to the 5
-oriented
(
)-syn-trans-BPDE-adducted template, which unlike the
other C10R adducted templates examined, allows for the
formation of full-length products on utilizing a 29-mer. A correlation
between the spatial conformation and orientation of the whole
polycyclic aromatic hydrocarbon moiety may be important. There are
preferred conformations of the hydroxyl groups on C7 and C8 in each
isomer ((anti-trans-, syn-trans-, and
anti-cis-BPDE)) which may be axial or equitorial or some
intermediate states such as pseudoaxial or pseudoequatorial (22). These
differences in the conformation of the tetrahydrobenzo ring may
possibly be described as contributing to the inhibitory angle. Thus,
our results suggest that not only is the stereochemistry at the
attachment point of the adduct (C10) significant but also
the stereochemistry within the aralkyl ring. For example, in
nucleosides with anti-trans-BPDE adducts, this ring adopts a
half-chair conformation with the hydroxyl groups at 7 and 8 in
pseudoequatorial positions, whereas the anti-cis-compounds prefer a half-chair conformation with these groups in pseudoaxial orientations. However, in oligonucleotides or duplexed DNA, the preferred conformations may be different depending upon interference from neighboring bases. NMR studies on a duplexed oligonucleotide containing a dA
N6-(10R)-syn-trans-BPDE adduct showed
that the aralkyl ring is in a boat-like conformation with the hydroxyl
groups at C7 and C8 pseudoequatorial, whereas in the corresponding
nucleoside, the aralkyl ring prefers a half-chair with pseudoaxial
oxygen substituents at C7 and C8 (21). Each stereoisomer will perturb
the DNA structure differently depending upon the orientation of the
hydroxyl group, e.g. the angle between the adduct and the
helix axis and the extent of intercalation may vary significantly with
both adduct and sequence. Our studies show clearly that polymerase
activity, at least in the case of HIV-1 RT, is highly sensitive to such
perturbation.
An additional factor in considering termination probabilities of HIV-1
RT is the perturbing influence of the adduct on the configuration of
the DNA when placed in the 4-base window of the T·P duplex that is
accessible to chemical cleavage (55). Prominent termination sites were
visualized 4 bases 5 to the lesion, thus emphasizing the distant
affects (Fig. 6). Although the intercalation of BPDE could effect only
the immediate base 5
to the adducted site, it is possible that the
bulky lesion prevents the duplex from unwinding 3-5 bases upstream and
consequently blocks replication. It is possible that the perturbation
of the T·P is far greater if a lesion is in the minor groove
(AdeN6-BPDE lesions are semi-intercalated from the major
groove), even if the size of the bulky adduct is much smaller than
BPDE, given that the damaged site is in closer proximity to the thumb
subdomain. Observations along these lines have been made in our
laboratory with guanine N2 styrene oxide lesions (56).
Studies with alanine-scanning mutants of -helix H of HIV-1 RT have
shown that one face of the core of this helix interacts with newly
synthesized DNA 3-6 base pairs from the catalytic site, where the DNA
is bent at about 45° (Fig. 7). Mutations within this
region decrease polymerase fidelity and processivity relative to the
wild type enzyme (41). Our study with the G262A and W266A mutants of
-helix H has shed light on their role in association with the
adducted T·P stem at a distance (Fig. 7). As shown in Fig. 6,
although all the C10S BPDE-adducted templates allowed the
formation of full-length products, the degree of extension varied
between enzymes. With special reference to the
(
)-anti-cis-BPDE-adducted template, prominent stop sites
were seen at positions 2 and 4 bases 5
to the adduct on replicating
with G262A, whereas with W266A, the truncated products mainly
accumulated at a position 2 bases 5
to the lesion, with minimal
formation of full-length products. Similarly, the unique
C10R (
)-syn-trans-BPDE-adducted template,
which allowed formation of some fully extended molecules, did not show
significant amounts of full-length products, when mutant enzymes were
employed. W266A in particular exhibited strong termination sites 3 bases 5
to the site of lesion. Although N6-adenine lesions
are in the major groove, and the
-helix H interacts through the
minor groove, it was still not surprising that the G262A and W266A
mutants affected the polymerization pattern. The interactions with the
minor groove are significantly hydrophobic, and therefore, the removal
of tryptophan in W266A is suggestive of leaving a gap in the vicinity
of the enzyme-T·P contact. This in turn could lead to the
destabilization of the T·P complex, subsequently resulting in the
fraying of DNA and dissociation of the enzyme (57). Our findings are
also supported by the fact that G262A and W266A mutants are far less
processive than HIV-1 RT (38). Furthermore, our data also suggest that
aromatic amino acid groups prevent destabilization of DNA associated
with the polymerase. As shown in Fig. 6 and Table I, the extension of the (
)-anti-cis-BPDE-adducted T·P complex is only half
as efficient with the W266A mutant as compared with the G262A mutant.
This indicates that lack of tryptophan inhibits polymerization.
Furthermore, although alanine is also a hydrophobic residue, it has a
considerably smaller side chain replacing the bulkier side chain of
tryptophan. This possibly suggests that the formation of a gap in the
T·P complex by substituting alanine for tryptophan creates a less stable enzyme-substrate complex. These results highlight the importance of maintaining not only a hydrophobic environment in the minor groove
of DNA, wherein the HIV-1 RT and substrate interact, but one with the
right geometry.
Thus overall, for adduct bypass to occur, the polymerase must overcome
several barriers presented by the specific lesion. First, the adduct in
the single-stranded template must allow the polymerase to synthesize up
to the lesion and then assume a catalytic geometry to incorporate a
nucleotide opposite the damaged base. Second, the polymerase must be
able to translocate and incorporate a nucleoside 1 base beyond the
adduct (38). Third, the damaged DNA must be able to translocate through
the thumb and palm subdomains without causing major pausing and
termination. All these factors could be altered in the presence of
modified residues in the -helix H.
In summary, our data show that differential in vitro
replication of C10S and C10R BPDE-dA-adducted
templates is not only controlled by the stereochemistry of the adduct
at the C10 position, but also by the C7 and C8 positions in
the aralkyl ring. Furthermore, our studies on RT and two of its
mutants, with BPDE-damaged templates, provide evidence for
protein-nucleic acid interactions in the minor groove that are
manifested catalytically at a distance. Finally, primer extension on
C10R versus C10S adducted templates are independent of their binding affinities to the polymerase, even
though the catalytic site is engaged with the 3-hydroxyl terminus of
the primer.
We are grateful to Drs. S. H. Wilson, W. A. Beard, and T. A. Kunkel for their generous gifts of the human recombinant HIV-1 RTs; Dr. G. J. Latham and A. McNees for reading of the manuscript; Dr. T. A. Darden for providing the coordinates for the model of the DNA/Helix H and Helix I complex; and Dr. W. A. Beard for helpful comments throughout the course of this study and for the creation of Fig. 7. We are indebted to P. V. R. Chary for the typing of this manuscript.