Oncogene (2001) 20, 7635 ± 7644
ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00
www.nature.com/onc
p53 enhances the ®delity of DNA synthesis by human immunode®ciency
virus type 1 reverse transcriptase
Mary Bakhanashvili*,1
1
Infectious Diseases Unit, Chaim Sheba Medical Center, Tel Hashomer 52621, Israel
The tumor suppressor protein p53 plays a critical role in
the maintenance of genetic integrity. p53 possesses 3'?5'
exonuclease activity, however, the signi®cance of this
function in DNA replication process remains elusive. It
was suggested that 3'?5' exonuclease activity of p53
may provide a proofreading function for DNA polymerases. In order to better understand the signi®cance of
this activity, the puri®ed wild-type recombinant p53 was
further evaluated for substrate speci®city and for
contribution to the accuracy of DNA synthesis. p53associated 3'?5' exonuclease displays 3' terminal
nucleotide excision from RNA/DNA template-primer
using ribosomal RNA as a template. The data
demonstrate that p53 is highly ecient in removing a
terminal mispair. Analysis of mispair excision opposite
the template adenine residue shows that p53 catalyzes
3' terminal mismatch excision with a speci®city of
A : G4A : A4A : C. Hence, the observed speci®city of
mismatch excision indicates that p53 exonucleolytic
proofreading preferentially repairs transversion mutations. The in¯uence of the p53 on the accuracy of DNA
synthesis was determined with exonuclease-de®cient
human immunode®ciency virus-1 (HIV-1) reverse transcriptase (RT), a key enzyme in the life cycle of the
virus, that contributes signi®cantly to the low accuracy
of proviral DNA synthesis. Using an in vitro biochemical
assay with recombinant puri®ed HIV-1 RT, p53 and
de®ned RNA/DNA or DNA/DNA template-primers, two
basic features related to ®delity of DNA synthesis were
studied: the misinsertion and mispair extension. The
misincorporation of non-complementary deoxynucleotides into nascent DNA and subsequent mispair extension by HIV-1 RT were substantially decreased in the
presence of p53 with both RNA/DNA and DNA/DNA
template-primers. In addition, the productive interaction
between polymerization (by HIV-1 RT) and exonuclease
(by p53) activities was observed; p53 preferentially
hydrolyzes mispaired 3'-termini, permitting subsequent
extension of the correctly paired 3'-terminus by HIV-1
RT. Taken together the data demonstrate that preferential excision of mismatched nucleotides by 3'?5'
exonuclease activity of wild-type p53 enhances the
®delity of DNA synthesis by HIV-1 RT in vitro, thus
*Correspondence: M Bakhanashvili; E-mail: [email protected]
Received 12 June 2001; revised 31 August 2001; accepted 4
September 2001
providing a biochemical mechanism to reduce mutations
caused by incorporation of mismatched nucleotides. The
fact that p53 is reactive with both RNA/DNA and
DNA/DNA template-primers raises an interesting possibility of the existence of functional cooperation between
p53 and HIV-1 RT in cytoplasm during the reverse
transcription process, which may be important for
maintaining HIV genomic integrity. Oncogene (2001)
20, 7635 ± 7644.
Keywords: p53; HIV-1 RT; 3'?5' exonuclease; DNA
replication ®delity
Introduction
The tumor suppressor protein p53 ensures the
integrity of the cell's genome by protecting it from
adverse e€ects of DNA damage (Albrechtsen et al.,
1999; Prives and Hall, 1999). This protein is an
important participant in the cellular responses to
genotoxic stress including ionizing radiation (Kastan
et al., 1991), UV radiation (Nelson and Kastan, 1994),
hypoxia (Graeber et al., 1994) or nucleotide deprivation (Linke et al., 1996). p53 levels are low in most
cells due to a short half-life (20 min) of the protein
(Soussi, 1995). In contrast, p53 is stabilized and its
level increased in response to DNA damaging stress.
The anti-proliferative e€ect of high levels of p53
induced in response to DNA damage correlates with
growth arrest at the G1 phase of the cell cycle,
allowing DNA repair prior to cell division or
elimination of cells with extensive DNA damage by
apoptosis (Cox et al., 1995). These observations have
led to the proposal that p53 is a `guardian of the
genome', monitoring the integrity of the genome
(Lane, 1992).
p53 was found in the nucleus, the cytoplasm or both
compartments of the cell. The sub-cellular localization
of p53 and the interaction with other cellular or viral
proteins plays a central role in the regulation of its
various biological activities (Moll et al., 1996; Komarova et al., 1997). The mechanism underlying its central
role in growth suppression is based on p53's function
as a potent transcriptional regulator of crucial growth
inhibitory genes (Prives and Hall, 1999; May and May,
1999). In addition, p53 exerts a full range of various
biochemical activities, that are directly related to its
p53 enhances the fidelity of HIV-1 RT
M Bakhanashvili
7636
Oncogene
function in preserving the integrity of the cell's genetic
information. After DNA damage: (1) p53 is able to
recognize and bind sites of DNA damage, such as
single-stranded (ss) DNA and double-stranded (ds)
DNA ends (Kern et al., 1991; Steinmeyer and Deppert,
1988), (2) p53 catalyzes DNA and RNA strand transfer
and promotes the annealing of complementary DNA
and RNA single-strands (Bakalkin et al., 1994;
Oberosler et al., 1993), (3) p53 binds insertion/deletion
mismatches and bulges (Lee et al., 1995) and (4) it can
bind DNA in a non-sequence-speci®c manner (El-Deiry
et al., 1992; Foord et al., 1993). Evidence suggesting a
role in DNA repair is supported by observations that
(1) p53 increases transcription-coupled nucleotide
excision repair (Hwang et al., 1999); (2) p53, like
classical mismatch repair factors, checks the ®delity of
homologous recombination processes by speci®c mismatch recognition (Dudenho€er et al., 1998); (3) p53
can markedly stimulate base excision repair (Zhou et
al., 2001); (4) p53 exhibits 3'?5' exonuclease activity
(Mummenbrauer et al., 1996) and may increase ®delity
of DNA replication mediated by cellular DNA
polymerase a (Huang, 1998). Our previous studies
revealed that wild-type (wt) p53 ful®ls the established
criteria for a proofreading exonuclease. It might act as
an external proofreader for errors produced by
exonuclease-de®cient retroviral murine leukemia virus
(MLV) reverse transcriptase (RT) (Bakhanashvili,
2001). Namely, recombinant wt p53-associated 3'?5'
exonuclease is able to preferentially remove mispaired
nucleotides from DNA, thus enhancing the accuracy of
MLV RT.
Interestingly, increased p53 levels have been noted
following infection of cells with various viruses: SV-40
(Oren et al., 1981), adenovirus (Grand et al., 1994),
Epstein-Barr virus (EBV) (Szekely et al., 1995) and
retrovirus-human immunode®ciency virus (HIV), the
etiological agent of acquired immunode®ciency syndrome (AIDS) (Genini et al., 2001). HIV infection is
characterized by the high genetic variability found in
virus populations (Katz and Skalka, 1990). This
phenomenon is attributed to the inaccuracy of the
replication machinery that is unique to the retroviral
life cycle. RT has a key role in the early stages of HIV
infection; it catalyzes the conversion of the viral
genomic ssRNA into the proviral DNA in the
cytoplasm (Con, 1986). HIV-1 RT exhibits dual
template speci®city and is highly error-prone displaying
low accuracy of DNA synthesis during copying both
RNA and DNA templates into their complementary
DNA (Perrino et al., 1989; Bakhanashvili and Hizi,
1992a,b). Excision of incorrectly polymerized nucleotides during DNA synthesis by proofreading exonucleases is an important mechanism to decrease errors
during DNA synthesis (Kunkel, 1988). Several studies
demonstrate the functional interaction between DNA
polymerases and external proofreaders carried out by
the 3'?5' exonucleolytic activity of other DNA
polymerases or/and by a separate protein serving as a
proofreading exonuclease (Brutlag and Kornberg,
1972; Maki and Kornberg, 1987; Perrino and Loeb,
1990). HIV-1 RT lacks 3'?5' exonuclease activity and
is unable to edit errors during proviral DNA synthesis.
The fact that HIV proviral DNA is synthesized in
the cytoplasm and the availability of the exonuclease
(associated with p53) at the site of viral reverse
transcription (in cytoplasm), raise an interesting
possibility that p53 and HIV-1 RT may be functionally
linked. To this end, p53 was examined for its ability to
remove 3'-terminal nucleotides from previously untested RNA/DNA substrate and for its functional
interaction with exonuclease-de®cient HIV-1 RT. The
contribution of p53 to the ®delity has been reconstituted with puri®ed recombinant p53, HIV-1 RT and
3' mismatch-containing template-primers. The results
demonstrate the increase in the ®delity of DNA
synthesis by HIV-1 RT in the presence of p53, thus
indicating the functional cooperation between p53 and
HIV-1 RT. The fact that p53 is reactive with both
DNA/DNA and RNA/DNA suggests that it may
functionally interact with substrates participating in
the reverse transcription process.
Results
p53-associated 3'?5' exonuclease activity with
RNA/DNA template-primer
The previous studies on the exonuclease activity
associated with p53 were focused on the DNA/DNA
template-primer (Huang, 1998; Skalski et al., 2000;
Bakhanashvili, 2001). The current study is devoted to
the assessment of the excision capacity of p53 with
RNA/DNA template-primer. The p53-associated 3'?5'
exonuclease was tested on RNA/DNA substrate using
native ribosomal RNA (rRNA) as a template. A
mixture of 16S and 23S rRNA was primed with
16mer oligonucleotide that hybridizes to nucleotides at
positions 2112-2127 of the 16S rRNA thus producing
RNA/DNA substrate. In parallel, the exonucleolytic
removal of 3'-terminal nucleotide was examined on
DNA/DNA substrate. The 3'?5' exonuclease activity
was detected by a decrease in the length of
oligonucleotide primers to 15 nucleotides and smaller,
as determined by increased mobility during electrophoresis. The results obtained indicate that p53 is able
to release dNMPs from the 3' terminus of RNA/DNA
template-primer as well as from DNA/DNA templateprimer (Figure 1, lanes 2 and 3). It should be noted
that the exonuclease activity is not associated with
DNA polymerase, as oligonucleotide bands greater
than 16 nucleotides in length are not detected even
though dNTPs are included in reaction (Figure 1, lane
4). The fact that p53 degrades DNA primer from both
RNA/DNA and DNA/DNA substrates, demonstrates
no di€erence in nucleic acid nature (RNA vs DNA) of
the template.
To further evaluate the 3'-terminal base excision
capacity by p53-associated 3'?5' exonuclease with
RNA/DNA template-primer, a series of templates with
16mer primers were prepared that formed the 3'-
p53 enhances the fidelity of HIV-1 RT
M Bakhanashvili
Figure 1 The excision of 3'-terminal nucleotide by the wt p53associated 3'?5' exonuclease. The correctly paired RNA/DNA
and DNA/DNA template-primers were incubated with puri®ed wt
p53 protein (50 ng) for 10 min and reaction products were
analysed by polyacrylamide gel electrophoresis as described under
Materials and methods. Lane 1, RNA/DNA template-primer
incubated with no p53. Lane 2, RNA/DNA template-primer
incubated with p53. Lane 3, DNA/DNA template-primer
incubated with p53. Lane 4, RNA/DNA template-primer
incubated with p53 and dNTP (1 mM). The position of the
16mer primer is indicated by an arrow
terminal A : A, A : C and A : G mismatches or A : T
correct pair at position 2112 of 16S rRNA. Upon
addition of the puri®ed p53, degradation of the primer
was monitored in time course reactions. The ability of
p53-associated 3'?5' exonuclease to remove 3'-terminal
mispairs under non-polymerization conditions is illustrated in Figure 2a. It is apparent that p53 excises the
3' terminal nucleotide from all primers tested. However, removal of the 3' terminal mismatched (A : A,
A : C or A : G) nucleotide is more ecient than that of
a complementary terminal base pair (A : T). Densitometric analysis of the primer excision bands resolved
by PAGE con®rms fourfold preference for excision of
the A : G mispair over excision from the matched A : T
base pair (Figure 2b). The trend in the order of a 3'
terminal mispair excision opposite a template A residue
with RNA/DNA template-primer by p53 is
A : G4A : A4A : C at this site. Thus, the data suggest
that p53 exonucleolytic proofreading prefers transversion mutations over transition mutations. Notably, the
same type of preformed mismatches were chosen in the
same sequence context that was previously employed
with DNA/DNA substrate (Bakhanashvili, 2001). It
should be emphasized that the observations obtained
previously with DNA/DNA template-primers, that
excision following purine ± purine (A : A or A : G)
mispair is more ecient than following the purine ±
pryimidine (A : C) mispair, is maintained with RNA/
DNA template-primer.
The functional interaction between p53 and HIV-1 RT
The ecient misinsertion and extension of mismatched
3'-termini of the nascent DNA are major determinants
for the low ®delity of HIV-1 RT (Perrino et al., 1989;
Bakhanashvili and Hizi, 1992a,b). The 3'?5' exonuclease-de®ciency of HIV-1 RT permits an analysis of
the in¯uence of p53 on the accuracy of DNA synthesis
by this enzyme. To discern the relative contribution of
p53-associated 3'?5' exonuclease activity to each of
these mechanisms, a system that allows simultaneous
detection of both degradation (exonucleolysis) and
extension (polymerization), was utilized. Model in vitro
systems that mimic the steps of proviral DNA
replication, namely both RNA and DNA-directed
DNA synthesis, permits an assessment of the ability
of p53 to function as proofreader in these processes.
Several approaches were undertaken to evaluate the
e€ect of p53 on the accuracy of DNA replication.
To elucidate the in¯uence of p53 on the insertional
accuracy of HIV-1 RT, the misinsertion by the enzyme
was investigated with both RNA/DNA and DNA/
DNA template-primers in the presence of wt p53. The
insertional ®delity of HIV-1 RT was examined using
running-start template-primer at a ®xed concentration
of 0.5 mM dATP, allowing the extension of the 16mer
primer. The results of the primer extension assays show
that HIV-1 RT displays misinsertion activity with both
substrates, incorporating wrong dATP opposite the
template G at site 2009 of template RNA (Figure 3,
lane 1) or at site 10 of template DNA (Figure 4, lane
1), following the incorporation of two running-start
A's. Interestingly, addition of increasing amounts of
p53 alter this pro®le. Namely, there is no detectable
misincorporation opposite the template G with HIV-1
RT in the presence of p53 with both RNA/DNA
(Figure 3, lanes 2 ± 5) and DNA/DNA templateprimers (Figure 4, lanes 2 ± 5). When higher concentrations of wt p53 were added, the decrease in the amount
of 17mer and 18mer products was observed and
products lower than 16mer were gradually formed.
Thus, the experiments described in Figures 3 and 4
show that the p53 protein a€ected the polymerase
selectivity for base substitution errors during both
RNA-dependent and DNA-dependent DNA polymerizations; it substantially reduced the number of
mismatched nucleotides incorporated into DNA. The
template-primer sequence requiring incorporation of
two A's prior to reaching the target G was chosen to
maximize proofreading, i.e. proofreading is most
e€ective in removing misinserted nucleotides adjacent
to relatively unstable DNA regions. This latter point
serves to emphasize the ability of p53 to carry out
e€ective error correction.
It was suggested that an important component of
retroviral genetic variability may be attributable to the
ecient mismatch extension during the copying of both
RNA and DNA templates. The in¯uence of the p53 on
the mispair extension capacity of HIV-1 RT was
evaluated by analysing the extension of 3' mismatchcontaining RNA/DNA and DNA/DNA templateprimers during DNA polymerization in the presence
of the next complementary dATP (as the only dNTP
present-complementary nucleotide to the template base
immediately downstream from the mispair). It was
recently shown that HIV-1 RT binds 3'-terminal
mismatched template-primer to about the same extent
(Bakhanashvili and Hizi, 1996). Therefore, the ability
of the enzyme to extend such template-primers was
assessed with no incubation prior to the addition of
7637
Oncogene
p53 enhances the fidelity of HIV-1 RT
M Bakhanashvili
7638
5' - 32
—
Figure 2 Excision of 3'-terminal mispairs on natural rRNA by wt p53 protein. (a) Time course of 3'-terminal nucleotide excision
by p53 was examined by incubation of the RNA/DNA substrates with puri®ed wt p53 protein for the indicated times. Exonuclease
reactions for terminal mismatched A : A (lanes 1 ± 5), A : C (lanes 6 ± 10) and A : G (lanes 11 ± 15) or matched A : T (lanes 16 ± 20)
nucleotide excision by p53 protein were performed as described in Materials and methods. The reaction was started by addition of
p53 (50 ng). After 1-, 2-, 5-, 15-, and 25-min incubation at 378C, 5 ml aliquots were withdrawn and analysed on 16% polyacrylamide
gel. The position of the 16mer primer is indicated by an arrow. (b) The percentage of substrate digested at the various times was
determined by measuring the radioactivity associated with the excision products by densitometric analysis and was expressed as the
percentage of the total input radioactivity
dATP. Under the assay conditions, the polymerase
exhibited a characteristic mispair extension pro®le with
RNA/DNA (Figure 5) and DNA/DNA templateprimers, both standing starts (Figure 6). Indeed, the
extension from the preformed mispairs (A : A, A : C
and A : G) with HIV-1 RT alone is shown by
elongation of the 16mer primers to 17mer or greater
with both substrates. As presented in Figures 5 and 6,
the addition of high concentration of p53 had a
dramatic e€ect on mispair extension catalyzed by HIVOncogene
1 RT. There was a signi®cant decrease in the yield of
the 17mer and increase in smaller (15mer and 14mer)
excision products. Hence, the results indicate that in
the presence of p53, HIV-1 RT exhibits a reduced
ability of extension 3' mispaired termini with RNA/
DNA and DNA/DNA template-primers employed.
The 3'?5' exonuclease activity is expected to confer
proofreading capability, thus increasing the ®delity of
DNA synthesis by many DNA polymerases (Kunkel,
1988). In order to explore the contribution of p53-
p53 enhances the fidelity of HIV-1 RT
M Bakhanashvili
7639
5 -32
Figure 3 The e€ect of wt p53 on misincorporation by HIV-1 RT
with RNA/DNA template-primer. The RNA/DNA templateprimer was incubated with HIV-1 RT and 0.5 mM dATP with no
p53 (lane 1) or in the presence of increasing amounts of p53
(lanes 2 ± 5). Lane 2, incubation with 25 ng p53. Lane 3,
incubation with 50 ng p53. Lane 4, incubation with 100 ng p53.
Lane 5, incubation with 150 ng p53. The position of the 16mer
primer is indicated by an arrow
5' - 32
Figure 4 The e€ect of wt p53 on misincorporation by HIV-1 RT
with DNA/DNA template-primer. The DNA/DNA templateprimer was incubated with HIV-1 RT and 0.5 mM dATP with no
p53 (lane 1) or in the presence of increasing amounts of p53
(lanes 2 ± 5). Lane 2, incubation with 25 ng p53. Lane 3,
incubation with 50 ng p53. Lane 4, incubation with 100 ng p53.
Lane 5, incubation with 150 ng p53. The position of the 16mer
primer is indicated by an arrow
5' - 32
Figure 5 The e€ect of wt p53 on mispair extension by HIV-1
RT and RNA/DNA template-primer. The 3'-terminal mismatch
containing RNA/DNA template-primers were extended by HIV-1
RT in the presence of 0.5 mM dATP with no p53 or with 150 ng
p53 (as described in Materials and methods). Lane A, templateprimer with 3'-terminal A : A mispair. Lane C, template-primer
with 3'-terminal A : C mispair. Lane G, template-primer with 3'terminal A : G mispair. The position of the 16mer primer is
indicated by an arrow
5' - 32
Figure 6 The e€ect of wt p53 on mispair extension by HIV-1
RT with DNA/DNA template-primer. The 3'-terminal mismatch
containing DNA/DNA template-primers were extended by HIV-1
RT in the presence of 0.5 mM dATP with no p53 or with 150 ng
p53 (as described in Materials and methods). Lane A, templateprimer with 3'-terminal A : A mispair. Lane C, template-primer
with 3'-terminal A : C mispair. Lane G, template-primer with 3'terminal A : G mispair. The position of the 16mer primer is
indicated by an arrow
Oncogene
p53 enhances the fidelity of HIV-1 RT
M Bakhanashvili
7640
associated 3'?5' exonuclease activity to the accuracy of
DNA synthesis, the interaction between p53 and HIV-1
RT was examined by analysing the products of a DNA
polymerization reaction. Mismatch-containing DNA is
an excellent model substrate for studying the contribution of p53 to the ®delity of DNA synthesis. A
biochemical proofreading assay utilized, exploits the
relatively delayed chain elongation from a 3' terminal
mispair by DNA polymerase, providing an opportunity
for a separate 3'?5' exonuclease to remove the
terminal mispair before the next correct base insertion
event. HIV-1 RT is able to extend 3'-terminal
mismatch-containing template-primer in the presence
of high concentration of next correct nucleotide.
Consequently, the functional interaction between p53
and HIV-1 RT was assessed in the presence of low
dNTP concentration, to avoid the mispair extension by
the enzyme at high dNTP concentration without the
contribution of external exonuclease. Obviously, no
elongation products are detected by HIV-1 RT with
mismatch-containing template-primers (A : A, A : C or
A : G) in reactions containing all four dNTPs at
equimolar concentrations (1 mM) (Figure 7, lanes 1 ±
3). However, in the presence of wt p53, HIV-1 RT was
able to extend the primer giving rise to the full-length
copy of the template (Figure 7, lanes 4 ± 6). Since the
p53-associated exonuclease activity itself is not a€ected
by the presence of dNTP, elongation presumably
occurs by initial hydrolysis of the terminally mispaired
nucleotide by the p53 and subsequent extension from
5' - 32
3
Figure 7 Interaction between HIV-1 RT and wt p53. The
reaction mixture contained 3'-terminal mismatch containing
template-primers and HIV-1 RT. Lanes 1 ± 3, incubation with
HIV-1 RT in the presence of 1 mM (each) of the four dNTPs.
Lanes 4 ± 6, incubation with HIV-1 RT in the presence of all four
dNTPs and 100 ng of wt p53. Lanes 7 ± 9, incubation with HIV-1
RT in the presence of 1 mM of dATP, dCTP and dGTP and
100 ng of p53. Lanes 1, 4, 7, template-primer with 3'-terminal
A : A mispair. Lanes 2, 5 and 8, template-primer with 3'-terminal
A : C mispair. Lanes 3, 6, 9, template-primer with 3'-terminal
A : G mispair. The position of the 16mer primer is indicated by an
arrow
Oncogene
the new, correctly paired 3'-terminus by HIV-1 RT. To
further con®rm that the mismatch was removed by
p53-associated 3'?5' exonuclease before extension of
the primer, the same reaction was performed with three
dNTPs, under omission of the dTTP (that correctly
replaces the mismatched nucleotide). Under these
conditions (do not allow correction of the mismatch
for which dTTP is needed), no detectable extension of
the 3' terminal mismatches occurred (Figure 7, lanes
7 ± 9). Instead, substantial hydrolysis of the 3'-terminal
mispair was detected, thus indicating that the products
observed in lanes 4 ± 6, are formed after correction of
the mismatch. The results demonstrate that hydrolysis
of the mispair is required to provide a 3'-terminus that
is eciently extended by HIV-1 RT and that
polymerization will not occur until excision of the
mispaired nucleotide has taken place. These proofreading results imply a functional association of
polymerization (by HIV-1 RT) and excision (by p53)
activities, which act in a coordinated manner during
DNA synthesis at the template-primer site to achieve
high ®delity.
Discussion
The present study was undertaken to characterize the
p53-associated exonuclease with an emphasis on its
substrate speci®city and for its interaction with HIV-1
RT. Recent studies revealed that p53 removes 3'terminal nucleotides from various nucleic acid substrates: ssDNA, blunt and cohesive double-strand
DNA breaks, dsDNA containing single nucleotide
mismatch (Huang, 1998; Skalski et al., 2000; Bakhanashvili, 2001). The current study identi®ed novel
property of p53 which distinguishes it from the large
majority of the known nucleases. Apparently, the
spectrum of nucleic acid substrates utilized by the
p53 includes RNA/DNA substrate; p53-associated
3'?5' exonuclease excises nucleotides from RNA/
DNA template-primers (Figure 1). Consequently, p53
protein could potentially have an additional editing
activity during RNA-dependent DNA polymerization,
as suggested by its preference for terminal mispairs
over the correctly paired residue (Figure 2). The
experiments reported herein demonstrate that p53
excises eciently 3'-terminal mispairs from all three
RNA/DNA template-primers tested. However, it is
evident from Figure 2 that p53 excises di€erent
mispairs with di€erent eciencies. Interestingly, p53
displays the same pattern of mispair excision speci®city
observed previously with DNA/DNA template-primer
(Bakhanashvili, 2001); namely, the general trend of
mispair excision is A : G4A : A4A : C. Thus, with
both RNA/DNA and DNA/DNA template-primers
(i.e. regardless the nature of template RNA vs DNA),
purine ± purine transversion mispairs are excised more
eciently by p53 than a purine ± pyrimidine transition
mispair. Generally, the extent of proofreading depends
on the nature of the DNA terminus and the local
template sequence context (Sinha, 1987). The mispair
p53 enhances the fidelity of HIV-1 RT
M Bakhanashvili
excision pattern obtained with identical RNA and
DNA sequences indicates that the p53-associated 3'?5'
exonuclease activity for di€erent mismatches is dependent upon the nature of the mispair. A comprehensive
study of various template substrates (with di€erent
purine/pyrimidine content of sequences) are needed to
further elucidate the in¯uence of base composition of
sequences on the substrate speci®city of the p53associated 3'?5' exonuclease.
The data show that p53 is highly ecient in
removing a terminal mismatch in the direct exonuclease
assay, thus indicating that p53 is very active when ®rst
binding to a 3'-terminus. However, the proofreading
capacity of p53 was observed during ongoing DNA
synthesis in vitro with DNA/DNA template-primer,
e.g. with cellular DNA polymerase a (Huang, 1998)
and viral MLV RT (Bakhanashvili, 2001). The results
of the current study demonstrate that the 3'?5'
exonuclease activity of p53 may provide a proofreading
function for HIV-1 RT as well; namely, wt p53 may
a€ect the ®delity of DNA synthesis carried out by
HIV-1 RT during both RNA?DNA and DNA?
DNA replication steps. HIV-1 RT extends mismatched
primer termini less eciently than matched termini.
This property could provide an opportunity for
external exonucleases to proofread HIV-1 RT-induced
replication errors. While HIV-1 RT alone exhibits low
accuracy of DNA replication, synthesis by the enzyme
in the presence of p53 is less error-prone. Support for
this assertion is derived from the observations that (1)
p53 enhances the ®delity of HIV-1 RT for the base
substitution ®delity; namely, both the formation
(Figures 3 and 4) and utilization of mispaired
template-primers by exonuclease-pro®cient HIV-1 RT
(Figures 5 and 6) decrease in the presence of p53; and
(2) p53 and HIV-1 RT could act coordinately in order
to establish an accurate replication reaction; p53associated 3'?5' exonuclease proofreads during DNA
synthesis by excising non-complementary nucleotides
prior to primer extension, that is eciently utilized by
HIV-1 RT (Figure 7). Collectively these ®ndings
suggest, that p53 may have a direct role in the reverse
transcription process and may control the accuracy of
DNA replication with retroviral HIV-1 RT by acting
as an external proofreader.
The DNA polymerase itself apparently plays a major
role in determining the frequency of formation of
various mispairs. Based upon a consideration of double
helix and individual mispair geometries, the frequencies
of mispairs are expected to be formed in the following
proportions: purine : pyrimidine4purine : purine4pyrimidine : pyrimidine. The mispair extension frequencies
of the HIV-1 RT were studied previously for the DNAdependent and RNA-dependent DNA polymerase
activities using the same sequences (Perrino, et al.,
1989; Bakhanashvili and Hizi, 1992a,b). This enzyme,
similar to other DNA polymerases (Mendelman et al.,
1990; Bakhanashvili and Hizi, 1993), exhibits a
variation in the eciency of mispair extension from
both DNA/DNA and RNA/DNA template-primers,
with a speci®city of A : C4A : A4A : G. There is a
possibility that the error speci®city is a€ected by the
proteins involved in DNA polymerization. Hence, in
an e€ort to better understand the contribution of
proofreading to the ®delity of DNA replication, two
di€erent mutational spectrums (mispair extension
spectrum by HIV-1 RT and mispair excision spectrum
by p53 with both RNA/DNA and DNA/DNA
template-primers) were compared. The mispair excision
spectrum, obtained in the previous study with DNA/
DNA substrate (Bakhanashvili, 2001) and in the
current study with RNA/DNA substrate, namely, the
preferential excision of purine ± purine mispairs over
purine ± pyrimidine, reveals an interesting results with
respect to the contribution of proofreading to ®delity.
As can be seen, the variances in the extension and
excision spectrum generated are di€erent for these two
reactions. Apparently, the most easily formed and
extended purine ± pyrimidine mispair (e.g. A : C) by
HIV-1 RT is less eciently excised by p53 protein, and
the less eciently formed and extended purine ± purine
mispairs (e.g. A : A or A : G) by the enzyme, are
eciently excised by p53 protein. Several factors may
be responsible for the preferential excision of the
transversion mispairs. Discrimination may be determined by hydrogen bonding between substrate and
template bases. Indeed, according to the geometric
selection principle the most stable misinsertions will
involve base pairs that are closest to Watson-Crick
geometry (e.g. A : C) (Sloane et al., 1988). Other base
pairs that deviate markedly from Watson-Crick
geometry are discriminated against more eciently.
The base substitution errors probably re¯ect the
properties of the proteins involved in polymerization.
It is of interest that, with the most accurate eukaryotic
DNA polymerase g, which contains a 3'?5' exonuclease, A : C mispair predominate over A : G and A : A
mispairs at the fX174 amber 3 locus (Kunkel, 1985).
In contrast, with DNA polymerase a, that lacks 3'?5'
exonuclease activity. A : A mispair predominate over
A : C and A : G mispairs (Kunkel and Alexander,
1986). These results, together with our ®ndings, raise
the interesting possibility that interaction of the
®delity-enhancing component (e.g. p53) of the DNA
polymerase might a€ect the base substitution speci®city
and that this in¯uence may be speci®c for each
proofreading exonuclease. In all, the magnitude and
location of mutations depend on a complex interplay
between the parameters of the catalytic `triad' involved
in DNA polymerization: DNA polymerases, proofreading exonucleases and the nature of the mispair.
During the ongoing DNA synthesis, the ability of
the polymerase to extend a speci®c mispaired primer is
a major factor that contributes to the overall eciency
of proofreading. Two mechanisms are possible for the
functional interaction of DNA polymerases and editing
exonucleases. The proofreading potential may depend
in part on whether the enzyme corrects the error
during a processive polymerization ± excision cycle
(without dissociating from the DNA)-intramolecular
pathway, or corrects it after dissociating-intermolecular
pathway. In both cases the eciency of editing
7641
Oncogene
p53 enhances the fidelity of HIV-1 RT
M Bakhanashvili
7642
Oncogene
misinserted nucleotides by a 3'?5' exonuclease would
be directly dependent on the DNA polymerase capacity
to extend from a misincorporated nucleotide. Polymerase dissociation at a mispair may be an important
consideration for proofreading in those systems for
which no polymerase-associated exonuclease has been
detected, thus allowing error correction by a separate
3'?5' exonuclease. It is notable, that the polymerase
reaction is generally faster than the exonuclease
reaction. Thus, the slower rate of extension of the
mispaired primer most likely re¯ects the time required
for the exonuclease to excise the non-complementary
nucleotide. It was suggested that DNA polymerase
could gain enormous bene®t from proofreading even
from a relatively weak exonuclease, if the polymerase
has diculty extending from a particular mispair
(Kunkel, 1988). Furthermore, proofreading exonuclease need not exhibit a stronger preference for
excision of a mismatched rather than matched base
in a direct exonuclease assay. All that is required is
discrimination against extension from a mispair within
the polymerase active site. Our results do not provide
evidence for the physical association between the
exonuclease de®cient HIV-1 RT and p53. The current
studies demonstrate the possible contribution of the
intermolecular pathway of the p53-associated 3'?5'
exonuclease proofreading. The experiments shown in
Figure 7, in which HIV-1 RT was tested for the
extension of a preformed 3'-terminally mispaired
substrates in the presence of p53 (conditions that
mimic a situation of intermolecular editing), points to a
mechanism of mismatch removing prior to subsequent
elongation. The most logical interpretation of the
interaction between p53 and HIV RT is that nucleotide
misinsertion results in RT dissociating from the
template-primer, leaving the 3'-terminal mispair accessible to the p53. Upon excision of the mispair the
exonuclease dissociates allowing the RT to reassociate
with the 3'-terminus and resume DNA synthesis.
Following misinsertion of a nucleotide, a DNA
polymerase may either dissociate from the templateprimer or it can extend the mispair, resulting in the
®xation of either transition or transversion mutations.
The errors produced during DNA synthesis could
result from three ®delity determining processes: nucleotide misinsertion into the nascent DNA, exonucleolytic
proofreading and extension of mismatched 3'-termini
of DNA (Echols and Goodman, 1991). The low ®delity
of HIV-1 RT presumably re¯ects a combination of the
absence of intrinsic proofreading, high frequency of
error formation and ecient mispair extension. Some
template-primers with terminal mispairs remain unextended by the polymerase. Interestingly, unextended
free template-primers (already dissociated from the
enzyme following the misinsertion) may be further
recognized by other HIV-1 RT molecules and undergo
a rebinding process with a subsequent 3'-mismatch
extension (Bakhanashvili and Hizi, 1996). The fact that
p53 excises terminal nucleotides independent of DNA
polymerase raises the possibility that the dissociated
unextended 3'-mismatch containing template-primer
may be recognized and utilized by p53 to remove
terminal mispairs thus enabling a new polymerization
cycle starting from the annealed state.
The experiments presented here demonstrate the
following characteristic features of the p53-associated
3'?5' exonuclease activity: (1) it hydrolyzes 3'-terminal
nucleotide from both RNA/DNA and DNA/DNA
template-primers; (2) its preferred substrate is 3'mismatch containing RNA/DNA and DNA/DNA
template-primers; (3) it exerts a preferential excision
of transversion mutations (e.g. purine ± purine), that
are less eciently extended by the DNA polymerase
(e.g. HIV-1 RT); (4) it may contribute to the accuracy
of DNA synthesis by excision 3'-terminal nucleotides
during the ongoing DNA synthesis i.e. coupled with
DNA polymerization and following direct binding to
template-primer i.e. independently from DNA polymerase. Consequently, p53 carrying these properties
may be relevant to the ®delity of HIV-1 RT.
These data raise questions about the biological
signi®cance of p53-associated exonuclease activity in
vivo. It was suggested, that this biochemical function of
p53 may be subject to stimulation by exogenous stimuli
(Skalski et al., 2000). HIV infections seem to increase
the intracellular level of p53 (Genini et al., 2001).
Therefore, the observed functional interaction between
the recombinant p53 and HIV-1 RT is valuable. The
presence of p53 by associated 3'?5' exonuclease
activity in cytoplasm may represent a possible pathway
that can excise mispaired bases (manuscript in
preparation). There is a possibility that upon HIV
infection, that replicates in cytoplasm, p53 may serve
as an exonuclease and compensate for a lack of
proofreading by HIV-1 RT during the process of
proviral DNA synthesis.
Materials and methods
Materials
Wild-type p53 was expressed employing a recombinant
baculovirus and puri®ed to apparent homogeneity (Ragimov
et al., 1993), quick frozen in small portions and stored at
7708C. The HIV-1 RT is the recombinant enzyme expressed
in E.coli from BH-10 proviral clone (Hizi et al., 1988) and
puri®ed according to Clark et al., 1990. The speci®c activity
of the enzyme was 4000 ± 5000 units/mg. One unit is de®ned
as the amount of enzyme that catalyzes the incorporation of
1 pmol of dTMP into DNA in the poly (rA)n.oligo(dT)12 ± 18directed reaction in 30 min at 378C. T4 polynucleotide kinase
was purchased from MBI (Fermentas).
Template-primers
Two template-primer substrates were used for the analysis
3'?5' exonuclease activity of p53 and characterization of
interaction of p53 and HIV-1 RT. First, ribosomal RNA
(rRNA) (a mixture of 16S and 23S E. coli rRNA) as a native
substrate, was primed with 16-mer oligonucleotide primer,
that hybridizes to the nucleotides at positions 2112-2127 of
the 16S rRNA. Four versions of primers were synthesized
separately. They are identical except for the 3'-terminal
p53 enhances the fidelity of HIV-1 RT
M Bakhanashvili
nucleotide (N), which is either A, C, G or T. The sequence of
these primers is 5'-ATTTCACATCTGACTN-3'. Second, is
an (N)30 synthetic oligonucleotide DNA template identical to
nucleotides 2110-2129 of E. coli 16S rRNA (5'AGGCGGTTTGTTAAGTCAGATGTGAAATCC-3'). This
DNA is primed with an (N)16 oligonucleotide, (utilized with
an rRNA template) that hybridizes to the nucleotides at
positions 13 ± 28 of the template DNA. Consequently, by
hybridizing the primer oligonucleotides to the rRNA
template at position 2112, or DNA template at position 13,
the A : A, A : C, A : G mispairs or the A : T correctly paired
terminus were produced. The primers were end labeled at the
5'-end with T4 polynucleotide kinase and [g-32P]ATP.
Unincorporated radioactivity was removed by using G-25
microspin columns (Pharmacia Biotech.), according to the
instructions of the producer. The end-labeled primers were
annealed to the template RNA or DNA as described
(Bakhanashvili and Hizi, 1992a,b).
Exonuclease assays
The 3'?5' exonucleolytic activity was measured as the removal
of 3'-terminal nucleotides (correct or incorrect) from the 5'[g-32P] end-labeled oligonucleotide, determined by the increase
in the mobility during the gel electrophoresis. The incubation
mixture (12.5 ml) contained 50 mM Tris HCl (pH 7.5), 5 mM
MgCl2, 1 mM DTT, 0.1 mg/ml BSA and 5'-end labeled various
substrates. The reaction was started by addition of p53 protein.
The variables, including reaction time and amounts of p53
protein, are given in the legends to the ®gures. Assays were
carried out at 378C. At appropriate time points the reactions
were stopped by addition of an equal volume of formamide gel
loading bu€er, followed by incubation at 908C for 3 min. The
reaction products were resolved by electrophoresis through a
16% polyacrylamide sequencing gel (PAGE). Exonuclease
activity, measured as the reduction in size of labeled
oligonucleotides, was detected by autoradiography. The
excision products were quanti®ed by densitometric scanning
of gel autoradiographs and the percentage of the total amounts
of primer excised were calculated.
7643
Exonuclease/polymerase coupled assays
To evaluate the interaction between p53 and HIV-1 RT, the
same template-primers used for the exonuclease activity were
utilized as substrates. When present, HIV-1 RT was at
1.5 U/ml. Proofreading reactions were performed with all four
dNTPs at 1 mM. For site-speci®c nucleotide misinsertion, a
correctly paired template-primer substrate was used to
analyse the dATP misincorporation opposite the G residues
at position 10 of the DNA template or at position 2009 of the
RNA template. For the mispair extension, elongation of the
5' end-labeled template-primers was examined in the presence
of dATP as the only dNTP present. The dNTPs used were of
the highest purity available (Pharmacia Biotech.) with no
detectable traces of contamination by other dNTPs. The
reaction products were analysed by electrophoresis through
16% PAGE as described above. Degradation or extension of
the 5'-end labeled primers were detected by autoradiography.
Acknowledgments
The author is indebted to Prof M Oren (from Weizmann
Institute, Rehovot, Israel) for the supply of puri®ed P53
protein and Prof A Hizi (from Tel-Aviv University, TelAviv, Israel) for supply of puri®ed HIV-1 RT. The author
thanks Prof E Rubinstein, Prof Y Sidi and Dr G Lilling for
critically reading the manuscript. This research was
supported by a grant from Israel Cancer Research Fund
(ICRF) and by a grant from Israel Cancer Association
(20000062 ± B).
References
Albrechtsen N, Dornreiter I, Grosse F, Kim E, Weismuller L
and Deppert W. (1999). Oncogene, 18, 7708 ± 7717.
Bakalkin G, Yakovleva T, Selivanova G, Magnusson KP,
Szekely L, Kiseleva E, Klein G, Terenius L and Wiman
KG. (1994). Proc. Natl. Acad. Sci. USA, 91, 413 ± 417.
Bakhanashvili M. (2001). Eur. J. Biochem., 268, 2047 ± 2054.
Bakhanashvili M and Hizi A. (1992a). FEBS Lett., 306,
151 ± 156.
Bakhanashvili M and Hizi A. (1992b). Biochemistry, 31,
9393 ± 9398.
Bakhanashvili M and Hizi A. (1993). Biochemistry, 32,
7559 ± 7567.
Bakhanashvili M and Hizi A. (1996). Arch. Bioch. Bioph.,
334, 89 ± 96.
Brutlag D and Kornberg A. (1972). J. Biol. Chem., 247, 241 ±
248.
Clark PK, Ferris AL, Miller DA, Hizi A, Kim KW,
Deringer-Boyer SM, Mellini ML, Clark AD, Arnold
GF, Lebherg III WB, Arnold E, Muschik GM and Hughes
SH. (1990). AIDS Res. Hum. Retroviruses, 6, 753 ± 764.
Con JM. (1986). Cell, 46, 1 ± 4.
Cox LS, Hupp T, Midgley CA and Lane DP. (1995). EMBO
J., 14, 2099 ± 2105.
Dudenho€er C, Rohaly G, Will K, Deppert W and
Wiesmuller L. (1998). Mol. Cell. Biol., 18, 5332 ± 5342.
Echols H and Goodman MF. (1991). Ann. Rev. Biochem., 60,
477 ± 511.
El-Deiry W, Kern SE, Pietenpol JA, Kinzler KW and
Vogelstein B. (1992). Nature Genet., 1, 45 ± 49.
Foord O, Navot N and Rotter V. (1993). Molec. Cell. Biol.,
13, 1378 ± 1384.
Genini D, Sheeter D, Rought S, Zaunders JJ, Susin SA,
Kroemer G, Richman DD, Carson DA, Corbeil J and
Leoni LM. (2001). FASEB J., 15, 5 ± 6.
Graeber T, Peterson J, Tsai M, Monica K, Fornace AJ and
Giaccia A. (1994). Mol. Cell. Biol., 14, 6264 ± 6277.
Grand RJA, Grant MI and Gallimore PH. (1994). Virology,
203, 229 ± 240.
Hizi A, McGill C and Hughes SH. (1988). Proc. Natl. Acad.
Sci. USA, 85, 1218 ± 1222.
Huang P. (1998). Oncogene, 17, 261 ± 270.
Hwang BJ, Ford JM, Hanawalt PC and Chu G. (1999). Proc.
Natl. Acad. Sci. USA, 96, 424 ± 428.
Kastan MB, Onyekwere O, Sidransky D, Vogelstein B and
Craig RW. (1991). Cancer Res., 51, 6304 ± 6311.
Katz R and Skalka AM. (1990). Ann. Rev. Genet., 24, 409 ±
445.
Kern SE, Kinzler KW, Baker SJ, Nigro JM, Rotter V, Levine
AJ, Friedman P, Prives C and Vogelstein B. (1991).
Oncogene, 6, 131 ± 136.
Komarova EA, Zelnick CR, Chin D, Zemerski M, Gleiberman AS, Bacus SS and Gudkov AV. (1997). Cancer Res.,
57, 5217 ± 5220.
Kunkel TA. (1985). J. Biol. Chem., 260, 12866 ± 12874.
Oncogene
p53 enhances the fidelity of HIV-1 RT
M Bakhanashvili
7644
Oncogene
Kunkel TA. (1988). Cell, 53, 837 ± 840.
Kunkel TA and Alexander PS. (1986). J. Biol. Chem., 261,
160 ± 166.
Lane DP. (1992). Nature, 358, 15 ± 16.
Linke SP, Clarkin KC, Di Leonardo A, Tsou A and Wahl
GM. (1996). Genes Dev., 10, 934 ± 947.
Lee S, Elenbaas B, Levine A and Grith J. (1995). Cell, 81,
1013 ± 1020.
Maki H and Kornberg A. (1987). Proc. Natl. Acad. Sci. USA,
84, 4389 ± 4392.
May P and May E. (1999). Oncogene, 18, 7621 ± 7636.
Mendelman LV, Petruska JS and Goodman MF. (1990).
J. Biol. Chem., 265, 2338 ± 2346.
Moll UM, Ostermeyer AG, Haladay R, Wink®eld B,
Frazier M and Zambetti G. (1996). Mol. Cell. Biol., 16,
1126 ± 1137.
Mummenbrauer T, Janus F, Muller B, Wiesmuller L,
Deppert W and Grosse F. (1996). Cell, 85, 1089 ± 1099.
Nelson W and Kastan MB. (1994). Mol. Cell. Biol., 14,
1815 ± 1823.
Oberosler P, Hloch P, Rammsperger U and Stahl H. (1993).
EMBO J., 12, 2389 ± 2396.
Oren M, Maltzman W and Levine AJ. (1981). Mol. Cell.
Biol., 1, 101 ± 110.
Perrino FW, Preston BD, Sandell LL and Loeb LA. (1989).
Proc. Natl. Acad. Sci. USA, 86, 8343 ± 8347.
Perrino FW and Loeb LA. (1990). Biochemistry, 29, 5226 ±
5231.
Prives C and Hall PA. (1999). J. Pathol., 187, 112 ± 126.
Ragimov N, Krankopf A, Navot N, Rotter V, Oren M and
Aloni I. (1993). Oncogene, 8, 1183 ± 1193.
Sinha NK. (1987). Proc. Natl. Acad. Sci. USA, 84, 915 ± 919.
Skalski V, Lin Z, Choi BY and Brown KR. (2000). Oncogene,
19, 3321 ± 3329.
Sloane DL, Goodman MF and Echols H. (1988). Nucleic
Acid Res., 16, 66465 ± 66475.
Soussi T. (1995). In: Molecular genetics of cancer. Cowell JK
(ed.) Bios Scienti®c: Oxford, pp 135 ± 178.
Steinmeyer K and Deppert W. (1988). Oncogene, 3, 501 ±
507.
Szekely L, Pokrovskaja K, Jiang WQ, Selivanova G,
Lowbeer M, Ringertz N, Wiman KG and Klein G.
(1995). Oncogene, 10, 1869 ± 1974.
Zhou J, Ahn J, Wilson SH and Prives C. (2001). EMBO J.,
20, 914 ± 923.