ABSTRACT
Strand transfer catalyzed by HIV reverse transcriptase (RT) was examined. The
system consisted of a 142 nt RNA (donor) to which a 50 nt DNA primer was
hybridized. The primer bound such that its 3
'
terminal nucleotide hybridized to the 12th nt from the 5
'
end of the donor. The
3
'
terminal nucleotide on the primer was either a G, A or T residue. Since the
corresponding nucleotide of the donor was a C, the G formed a matched terminus
and the A or T a mismatched terminus. The efficiency with which DNA bound to
the donor transferred to a second RNA, termed acceptor, was monitored. The
acceptor was homologous to the donor for all but the last 9 nt at the 5'
end of the donor. Therefore, homologous strand transfer could occur at any point
prior to the DNA being extended into the nonhomologous region on the donor.
Strand transfer occurred approximately twice as efficiently with the mismatched
versus matched substrates. The mismatched nucleotide was fixed into transfer
products indicating that excision of the mismatch was not required for RT
extension or transfer. Results suggest that base misincorporations by RT may
promote recombination by enhancing strand transfer.
The human immunodeficiency virus (HIV) has been shown to contain a high degree of
genetic heterogeneity (
1
). The
fidelity
of HIV-reverse transcriptase (RT) is a major contributing factor in the
generation of diversity (
2
-
8
). This multifunctional enzyme converts the single-stranded RNA viral genome to double-stranded DNA through a series of steps (for a review see
9
). The RNA-dependent DNA polymerase activity synthesizes the negative strand DNA
while the RNase H activity hydrolizes the RNA template. The DNA-dependent DNA polymerase activity then synthesizes the positive strand DNA
to complete synthesis of the proviral DNA. Much of the genetic variability is
thought to arise from nucleotide misinsertions. HIV-RT has been found to incorporate more mistakes during replication than
reverse transcriptases from avian myeloblastosis virus (AMV) and Moloney murine
leukemia virus (MuLV) (
2
). One study performed
in vitro
estimated the frequency of base misincorporations for the polymerase at 1 in
6900 on an RNA template and 1 in 5900 on a DNA template (
8
). It should be noted that, although all studies suggest that HIV-RT has relatively low fidelity, error frequency estimates have varied
widely. One study suggests that fidelity is several-fold higher when RNA versus
DNA is used as template (
6
). Another study implies that RT fidelity
in vivo
is considerably greater than the
in vitro
estimates (
10
). As is the case with other retroviruses, HIV-RT contains no 3'-5' exonuclease, therefore, errors made by the enzyme
are not proofread (
3
).
Strand transfer occurs when DNA synthesized on one template is translocated to
another region on the same or a different template. Two such events, occurring
at the terminal regions of the retroviral genome, are an integral part of
retroviral replication. These are the transfer of the minus and plus strand
strong-stop DNAs (
11
,
12
). These DNA are initially synthesized at the 5' end of their respective viral templates, then transfer to homologous
regions at the 3' end of the template. Since the virus contains two copies of the genomic
RNA, the transfers could occur either intra- or inter-molecularly. It has also been shown that transfer of the growing DNA
strand can occur at internal regions of the viral RNA (
13
-
16
). Research suggests that pausing of the polymerase at a specific site may
promote transfer of the DNA strand to a homologous region of a different RNA
strand (
17
-
19
). A base misincorporation could provide the conditions necessary for pausing of
the polymerase and could set the stage for a strand transfer event. Results
have shown that even though RT can extend a mispaired 3' terminus more efficiently than RTs from other viruses, the frequency is
much lower when compared with the extension of the correct nucleotide (
20
-
22
). Pausing of the polymerase may allow the RNase H activity of RT to cleave the
RNA template, making the interaction between the primer and the RNA less
stable. This could lead to a subsequent dissociation of the DNA from the
original template and binding to a homologous region on a different RNA strand.
Alternatively, pausing may allow more time for strand invasion to occur (
18
). In this scenario the primer is displaced from the template RNA upon binding
to a second homologous RNA template which `invades' the duplex.
In this report we demonstrate that strand transfer to a homologous acceptor
template occurs more efficiently in the presence of a mismatched 3' termini
between the growing DNA strand and template RNA. The mismatched primers were
extended less efficiently than the primer containing the correct nucleotide. At
the same time, strand transfer from the
mismatched primers was enhanced. The mismatch was retained in the vast majority
of transfer products, thus there was no excision of the non-complimentary base. These results were found to be independent of enzyme
and acceptor template concentration.
Recombinant HIV-RT with native primary structure was graciously provided by Genetics
Institute (Cambridge, MA). The enzyme had a specific activity of ~40 U/[mu]g. One unit of RT is defined as the amount required to incorporate 1
nmol dTTP into nucleic acid product in 10 min at 37oC using poly(rA)-oligo(dT) as a template-primer. As we have previously reported, the enzyme
preparations contained very low levels of single strand nuclease activity. We
found that this activity could be inhibited by including 5 mM AMP in the assays
(
18
). The AMP, at this concentration, did not affect the polymerase or RNase H
activity of the RT (data not shown). Aliquots of HIV-RT were stored frozen at -70oC and a fresh aliquot was used for each experiment. T4 Ligase,
T4 polynucleotide kinase and Sequenase version 2.0 were obtained from United
States Biochemical Corp. Klenow fragment, restriction enzymes, T7 RNA
polymerase rNTPs and dNTPs were obtained from Boehringer Mannheim Biochemicals.
Superscript was obtained from Gibco BRL. The oligonucleotide DNA primers were
synthesized by Genosys Inc. All other chemicals were from Sigma. Radiolabeled
compounds were from New England Nuclear.
The strand transfer system used in these experiments is shown in Figure
1
A. The system consisted of a 142 nt RNA strand to which a 50 nt 5'-
32
P-labeled DNA primer was hybridized. This RNA is referred to as the `donor'
template, which is the template on which DNA synthesis initiates. In order to
observe strand transfer, a second RNA template termed `acceptor' was employed.
The acceptor was the template to which DNAs initially hybridized to the donor
would transfer. The acceptor was 189 nt in length and was homologous to the
donor for all but the last 9 nt at the 5' end of the donor. Full-length donor-directed DNA extension products were 61 nt in length while
DNAs which transferred to, and were subsequently fully extended on the acceptor
were 108 nt long. This difference in length allowed us to easily distinguish
transfer events from donor-directed extension using denaturing polyacrylamide gels (see Methods).
Primer DNAs extended to the end of the donor could not undergo homologous
strand transfer and subsequent extension on the acceptor since the last 9 nt of
these DNAs would be mismatched on the acceptor template (see above). Although
nonhomologous recombination could potentially occur, this type of recombination
is rare occurring at 1/100-1/1000 the frequency of homologous recombination (
27
). Consequently, we would expect that most of the observed transfer events
resulted from transfer of DNAs which transferred before being extended into the
region of the donor that was not homologous to the acceptor.
In these experiments strand transfer was expressed as a `percent efficiency'.
The transfer efficiency was defined as the amount of transfer products (T)
divided by the amount of full-length donor-directed (F) plus transfer products times 100 [(T/F + T) * 100]. The number reflects the proportion of DNA primers
extended to the end of the acceptor
versus
those extended to the end of the donor. This representation of the data, as
opposed to simply determining the gross level of transfer products, expresses
transfer relative to total DNA extension. Therefore, differences in the total
amount of primers extended with the match versus mismatch substrate are
compensated for.
Figure
3
A is an autoradiogram of a typical strand transfer experiment using the matched
or A mismatched substrate. In the standard assay 2 nM substrate and 20 nM
acceptor template were used. Under these conditions, transfer products were
detected ~4-8 min into the reactions and increased thereafter. At the final data
point (32 min) the gross level of transfer products with the matched substrate
was about half that with the mismatch. Figure
3
B shows a composite graph for six independent experiments of the efficiency of
transfer
versus time using the A mismatch or matched substrate. Between 8 and 32 min the
transfer efficiency was two to three times as great with the mismatched
substrate, reaching ~17% as opposed to 7% with the match. Between 32 and 64 min only a slight
increase in transfer efficiency was observed (data not shown).
Table 1
We have shown that the presence of a mismatched nucleotide at a 3' primer terminus enhances strand transfer from the mismatched substrate
to a homologous acceptor template. Given the high misincorporation rate of HIV-RT, and the lack of significant proofreading activity it is likely that
mismatched termini occur with relative frequency during RNA-directed DNA synthesis (see Introduction). This report suggests that such
events may generate genetic diversity not only through the possible fixation of
nonparental nucleotides, but also by stimulating strand transfer events.
In these experiments we examined the effect on strand transfer of two types of
mismatch. One of these mismatches (A:C) would result from a primer-slippage-type mechanism occurring within runs of the same nucleotide (
29
). The insertion within the primer strand of an additional nucleotide
complementary to the nucleotides within the run on the template strand is among
the most common errors occurring during HIV-RT-directed DNA synthesis (
6
,
29
). A mismatched C:A primer-template was extended ~1300 times less efficiently than a matched substrate by HIV-RT (
20
). Although in those experiments, unlike ours, the C was on the primer strand
and A on the template strand. AMV-RT extended an A:C primer-template mismatch ~300 times less efficiently (based on
V
max
/
K
m
) than the matched G:C pair (
21
). Since the kinetics of extension may vary depending on the surrounding
nucleotide sequence, the particular RT used, and the orientation of the
mismatch, it is not possible to draw a direct correlation between the AMV-RT work or previous HIV-RT work and our substrates. It was clear from our experiments (Fig.
4
) that extension of the mismatched template was less efficient than the match,
but a quantitative evaluation of efficiency was not done. We also evaluated the
effect of a T:C mismatch on strand transfer. Since the efficiency of extension
under the conditions employed in our experiments was approximately the same for
the A:C and T:C mismatches, it was not possible to determine if increased
transfer efficiency correlates with a decrease in extension efficiency. The
similarity between the extension efficiencies of the two mismatches may result
from the concentration of dNTPs used in our experiments which are well above
RTs
K
m
for nucleotides. Since the efficiency of mismatch extension as assessed by
V
max
/
K
m
indicates that the lower efficiency of mismatch extension is due mostly to
elevated
K
m
values (see Results), differences in extension kinetics would be less
pronounced at high dNTP concentrations. We were unable to evaluate transfer
with mismatched substrates at very low nucleotide concentrations due to the low
level of extended primer under these conditions.
We found that the A:C mismatch was fixed into transfer products by two different
mechanisms. Of the 18 sequenced transfer products in which the mismatch was
fixed, 15 resulted from extension directed from the next nucleotide (G)
downstream of the mismatched C on the template strand. This mechanism results
in `in frame' DNA products with a G to A substitution on the primer strand. The
other three fixed transfer products were DNAs in which extension was directed
from the mismatched C residue on the template. This mechanism produces
insertion `mutants' with +1 frameshifts. Frameshift mutants in open reading
frames are more likely to produce nonfunctional proteins than are nucleotide
substitutions. It is noteworthy that substituted DNA products were produced
with much greater frequency (15 as compared with 3) than frameshifts. It would
also be interesting to know whether the type and proportion of mutations
produced during strand transfer are different from those generated during
extension of the donor.
The increase in transfer efficiency with the mismatch versus matched templates
was modest (2-3-fold), but clearly significant (see Fig.
2
). With the two mismatched templates >10% of the extended primers had
transferred to the acceptor template by the 32 min time point. This proportion
seems unrealistically high since estimates for retroviral strand transfer
in vivo
are on the order of 1 event per 25 000 nt copied (
16
). Some possible reasons for the high
in vitro
rate include the small sizes of the acceptor and donor templates and the high
ratio of acceptor to donor in our reactions. The former explanation would lead
to more efficient transfer since it would be easier for regions of homology to
align appropriately on a very small template. The considerably greater length
of the normal retroviral genome probably makes alignment more difficult. The
latter explanation is clearly illustrated in Figure
4
. The efficiency of transfer decreased as the ratio of acceptor to donor
decreased. Although our
in vitro
system does not represent
in vivo
recombination in a quantitative sense, it is likely that there are mechanistic
similarities.
We did not attempt to elucidate the mechanism by which the mismatched terminus
enhanced strand transfer. However, one possibility, consistent with the reduced
extension of the mismatched versus matched substrates (Fig.
5
), is that the mismatch promotes pausing of the polymerase at the primer
terminus. It has been shown that pausing during RNA-directed DNA synthesis can promote strand transfer (
17
,
18
). Pausing allows greater time for the RT RNase H activity to cleave the RNA
beneath the terminal region of the DNA primer. Extensive cleavage may weaken
the interaction between the DNA and RNA allowing for invasion of the
heteroduplex by a homologous acceptor RNA, or dissociation and subsequent
binding of the DNA strand to a complementary acceptor (
18
). Further experiments will be necessary to elucidate the mechanism by which
mismatches can enhance strand transfer and to confirm this phenomena
in vivo
.
We wish to thank Drs Jasbir Seehra and John McCoy, representing Genetics
Institute, for the generous gift of HIV-RT. This work was supported by NIH grant GM-51140-01.
Standard strand transfer and primer extension reactions.
In the standard reaction primer-donor template (2 nM) and acceptor template (20 nM, unless otherwise
indicated) were preincubated for 3 min in a volume of 10.5 [mu]l at 37oC. Acceptor was omitted in primer extension assays designed to measure
donor-directed extension only. Reactions were initiated by the addition of 2 U (~35 nM final concentration) of HIV-RT in 2 [mu]l of 50 mM Tris-HCl (pH 8.0), 1 mM dithiothreitol and 80 mM KCl. The
final concentrations of reactions components were 50 mM Tris-HCl (pH 8.0), 5 mM AMP, 6 mM MgCl
2
, 1 mM dithiothreitol, 0.1 mM EDTA (pH 8.0), 100 [mu]M dNTPs and 80 mM KCl. Samples were incubated for 32 min at 37oC unless otherwise indicated, and reactions were terminated by addition
of 12.5 [mu]l of gel loading buffer (90% formamide, 10 mM EDTA pH 8.0, 0.1% xylene
cyanol, 0.1% bromophenol blue).
Hybridizations.
The 50 nt deoxyoligonucleotides for the matched and mismatched substrates (see
Fig.
1
) were labeled with
32
P at the 5' end using T4 polynucleotide kinase. The labeled primer was hybridized
such that the 3' end terminal nucleotide was positioned opposite the 12th nt from the 5' end of the RNA transcript. The hybrids were prepared by mixing
primer and transcript at a 4:1 ratio in 50 mM Tris-HCl, 1 mM dithiothreitol and 80 mM KCl. The mixture was heated 65oC for 10 min and then slowly cooled to room temperature. After
hybridization 6* native gel loading buffer [40% (w/v) sucrose, 0.25% (w/v) xylene cyanol
and bromophenol blue] was added and the mixture was electrophoresed on a non-denaturing 8% polyacrylamide gel as described below. The hybrid complex
was located by autoradiography, excised and eluted in a buffer containing 50 mM
Tris-HCl (pH 8.0), 80 mM KCl, 6 mM MgCl
2
and 1 mM dithiothreitol.
Gel electrophoresis.
Denaturing 8% polyacrylamide sequencing gels (19:1, acrylamide-bisacrylamide) containing 7 M urea or non-denaturing native gels (29:1, acrylamide-bisacrylamide) were prepared and subjected to
electrophoresis as described (
23
).
Transcription reactions.
Run-off transcription was performed as described (
24
). For the donor template pBSM13[Delta], prepared as described previously (
17
), was cleaved with
Mva
I and T7 RNA polymerase was used to prepare RNA transcripts 142 nt in length.
For the acceptor template, pBSM13+ was cleaved with
Mva
I and T7 RNA polymerase was used to prepare a transcript 189 nt in length. RNA
for both the donor and acceptor templates was gel purified on denaturing
polyacrylamide gels. The electrophoresed RNA was located on the gel by
ultraviolet shadowing, excized, and eluted in a buffer containing 150 mM NaCl,
50 mM Tris (pH 8.0), 1 mM EDTA and 0.1% SDS. The RNA was recovered from the
eluate as previously described (
25
).
Isolation of transfer products.
The strand transfer reactions were subjected to electrophoresis on an 8%
denaturing polyacrylamide gel. The strand transfer products were located by
autoradiography, excised from the gel, and recovered as described above. The
recovered DNA was amplified by PCR for 30 cycles as described (
26
). The reactions were carried out in buffer containing 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 2 mM MgCl
2
and 50 mM dNTPs. One primer, (5'-GGGCGAATTCGAGCTCGGTACCCGGGGATC-3') was complimentary to nt 78-108 on the transfer products while the other (5'-TACGCCAAGCTCGGAATTAA-3') was identical to nt
8-28 on the 5' end of the products. The PCR reactions were then subjected to
electrophoresis on a non-denaturing polyacrylamide gel. Products were located using ultra violet
light after staining the gels with ethidium bromide (
23
). The products were excised and eluted in 150 mM NaCl, 50 mM Tris-HCl (pH 8.0), 0.1% SDS and recovered by precipitation in ethanol. The PCR
products were then treated with the Klenow fragment in the presence of 100 [mu]M dNTPs. These blunt-ended products were ligated into pBCSK previously cleaved with
Eco
RV. The ligation mixture was transformed into
Escherichia coli
XL-1 Blue competent cells. Clones containing the insert were located by blue-white color selection and sequenced using Sequenase version 2.0
according to the manufacturer's instructions. The primer used for sequencing (5'-TAATACGACTCACTATAGGG-3') was complimentary to the T7 promoter on the plasmid.
Enzyme units
Transfer product (fmol)
Transfer efficiency (%)
match
mismatch
match
mismatch
0.5
0.15
0.39
1.0
6.8
1
0.15
0.52
1.1
6.6
2
0.19
0.91
1.2
8.9
4
0.25
0.74
1.5
6.5
REFERENCES
Return