(Received for publication, October 21, 1996, and in revised form, June 25, 1997)
From the Laboratory of Molecular Genetics, NICHD, National Institutes of Health, Bethesda, Maryland 20892
Retroviral RNases H are similar in sequence and structure to Escherichia coli RNase HI and yet have differences in substrate specificities, metal ion requirements, and specific activities. Separation of reverse transcriptase (RT) into polymerase and RNase H domains yields an active RNase H from murine leukemia virus (MuLV) but an inactive human immunodeficiency virus (HIV) RNase H. The "handle region" present in E. coli RNase HI but absent in HIV RNase H contributes to the binding to its substrate and when inserted into HIV RNase H results in an active enzyme retaining some degree of specificity. Here, we show MuLV protein containing the C-terminal 175 amino acids with its own handle region or that of E. coli RNase HI has the same specific activity as the RNase H of RT, retains a preference for Mn2+ as the cation required for activity, and has association rate (KA) 10% that of E. coli RNase HI. However, with model substrates, specificities for removal of the tRNAPro primer and polypurine tract stability are lost, indicating specificity of RNase H of MuLV requires the remainder of the RT. Differences in KA, while significant, appear insufficient to account for the differences in specific activities of the bacterial and viral RNases H.
Reverse transcriptases (RT)1 are enzymes responsible for copying the retroviral genomic RNA into double-stranded DNA by a process involving both the polymerase and RNase H of the RT (1). For human immunodeficiency virus type 1 (HIV-1), RT comprises a heterodimeric protein consisting of p66 and p51 subunits (2-4). The p51 polypeptide is derived from p66 by proteolysis, removing the C-terminal (p15) RNase H domain. Both polymerase and RNase H catalytic sites are contributed by p66, with p51 acting primarily as a structural polypeptide (5, 6). Isolated p51 protein has very poor polymerase activity and, of course, no RNase H activity. Linker-scanning mutations in HIV-1 RT (7), as well as the structure determined by x-ray crystallography (8, 9), show an intermingling of polymerase and RNase H regions, with one activity relying on other portions of the protein for activity (10). The RT of murine leukemia virus (MuLV) differs from that of HIV-1 RT in that it is isolated as a single polypeptide chain of 76 kDa (11), which appears to form dimers when bound to nucleic acids (12, 13). Furthermore, MuLV RT can be divided into two separate polypeptides; the N-terminal two-thirds portion of the protein has very good polymerase activity and a protein from the C-terminal one-third is reported to be an active RNase H (14).
RNase H activity is essential for retroviral replication (15, 16) and is involved in several steps of replication including the following: removal of tRNA, which functions as a primer for minus-strand DNA synthesis (17); generation of polypurine tract (PPT), the primer for second strand DNA and its subsequent removal (18, 19); and degradation of the viral RNA genome (20-23). In vivo studies demonstrate that inactivation of RNase H results in production of noninfectious virions (24-26).
Amino acid sequence comparison shows that HIV and MuLV retroviral
RNases H share about 25% of their sequence with Escherichia coli RNase HI (27). Much greater similarity is found when
comparing the three-dimensional structures of E. coli RNase
HI (28, 29) and the RNase H of HIV-1 reverse transcriptase in the
complete RT (8, 9) or in the p15 (C-terminal domain) (30). A
significant exception to the structural similarity is the presence of
an additional handle region in E. coli RNase HI (helices
B,
C, and
D) not found in
the corresponding region of HIV-1 p15, which contains only helices
B and
D. Several studies demonstrate that
C is rich in basic amino acids and contributes
significantly to binding to RNA-DNA hybrids (31, 32). One construct of
E. coli RNase HI was made in which the handle region is
missing and in which enzymatic activity is lost due, in part, to a
decrease in binding to RNA-DNA hybrids by more than 3 orders of
magnitude (32). In another case, an E. coli RNase HI protein
with a different connection of the
B and
D helices retains some enzymatic activity but only in
the presence of Mn2+ ions (33). A protein containing only
the p15 domain has no RNase H activity (34-36) but acquires activity
by the following: (i) mixing with the HIV-1 polymerase domain (p51)
(36) or part of the connection region (37); (ii) adding a
His6-tag to the RNase H domain (37-39); and (iii)
inserting the
C region from the E. coli RNase
HI (39, 40). One somewhat surprising result obtained with the active
HIV-1 RNase H proteins is their ability to cleave model tRNA primer
substrates at the same site as intact RT, suggesting that at least some
of the specificity elements for cleavage are present in these
polymerase minus RNases H (37-40).
The three-dimensional structure of MuLV RNase H has not yet been
elucidated, but amino acid sequence alignment predicts that this RNase
H has a handle region, consisting of B,
C, and
D, similar to E. coli
RNase HI (28, 29). Compared with E. coli RNase HI, MuLV
reverse transcriptase-associated RNase H activity is very low, whether
associated with the polymerase domain or not. Furthermore, the RNase H
activity of MuLV RT prefers Mn2+, and MuLV RT cleaves
replication intermediates essential for retroviral replication at
specific sites that are different from those E. coli RNase
HI cleaves (21-22). Little is known about the biochemical properties
of the C-terminal RNase H of MuLV RT.
To further our understanding of retroviral RNases H, we have examined the C-terminal 175 amino acid portion of MuLV RT (C175-AKR-RNase H) for its activity, its binding to RNA-DNA hybrids, and its specificity of cleavage. For specificity studies, we have used a model tRNA primer site, the polypurine tract (PPT), a fragment of mRNA for ovalbumin annealed to a complementary DNA oligonucleotide, and a homopolymeric poly(rA)·poly(dT).
The E. coli strain MIC1066
(rnhA-339::cat recB270(TS)) (41) was used to
over-produce the C175-AKR-RNase H (AKR is the strain of mice from which
the virus was isolated), AEA (C175-AKR-RNase H with the E. coli RNase HI "handle region"), in addition to E. coli RNase HI. Bacterially expressed AKR-MuLV reverse
transcriptase and a polyclonal antibody against N-terminal AKR-MuLV
RNase H were kind gifts from Dr. Judith Levin at NICHD, National
Institutes of Health (42). The Mega RNA transcription kit was from
Ambion. Expression vectors pET15b(+) and pET21a(+), biotinylated
protease thrombin, streptavidin-agarose, and His-Bind® resin were
from Novagen. The Western LightTM chemiluminescent detection system was from Tropix, Inc. RNA36 oligonucleotide
(5)-biotin-GGGAUCAGUGGUUCCCAUAUCCCGGACGAGCCCCCA-(3
) was synthesized
by Oligos Etc. Inc. A partially complementary DNA36
(TGGGGGCTCTGCCGGGATATGGGAACCACTGATCCC) was from BioServe. The ultrapure
Sequagel sequencing system was from National Diagnostics, and 6 and 8%
gel mix were from Life Technologies, Inc. The sensor chip CM5, Tween
P20, and amine coupling kit were from Pharmacia Biotech Inc.
Streptavidin and biotin were from Pierce.
The DNA segment encoding Met1 to Val155 of E. coli RNase HI and Leu496 to Leu671 of the pol gene of AKR-MuLV (11) was modified by adding restriction sites (NdeI and EspI) and cloned either into the pET15b(+) vector, which expresses a His6-tag at the N terminus of the target proteins, or into the pET21a(+) vector for expression of proteins without His6-tag. The AEA plasmid carries the DNA of AKR-MuLV encoding amino acids Leu496 to Tyr581, the E. coli RNase HI sequence for Thr69 to His114, followed by AKR-MuLV Arg629 to Leu671.
Expression and PurificationFour-liter cultures of MIC1066
harboring the C175-AKR-RNase H, chimeric RNase H (AEA), or E. coli RNase HI plasmids were grown at 32 °C to an
A600 value about 0.8 and then induced with 1 mM isopropyl--D-thiogalactopyranoside for
3 h at 32 °C. Cells were harvested and sonicated in loading
buffer (5 mM imidazole, 0.1% Nonidet P-40, 0.5 M NaCl, and 20 mM Tris-HCl, pH 7.9). After
centrifugation at 10,000 rpm for 30 min, the supernatants were filtered
through a 0.45-micron filter and loaded on a 1-ml His-Bind® resin.
The column was washed with 10 ml of loading buffer and 10 ml of washing buffer (60 mM imidazole, 0.5 M NaCl, and 20 mM Tris-HCl, pH 7.9). The His6-tagged
C175-AKR-RNase H, AEA, and E. coli RNase HI were eluted with
0.5 M imidazole, 0.5 M NaCl, and 20 mM Tris-HCl, pH 7.9. After dialysis against buffer A (50 mM Tris-HCl, pH 7.9, 1 mM dithiothreitol, 50 mM NaCl, 10% glycerol, 0.5 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride), samples were applied to
consecutively connected high performance liquid chromatography columns
of TSK gels DEAE-5PW (7.5 × 75 mm) and SP-5PW (7.5 × 75 mm). The His6-tagged C175-AKR-RNase H was present in the
flow-through fraction, and the His6-tagged E. coli RNase HI bound to the SP column, eluting at about 150 mM NaCl from a 50 mM to 1 M NaCl
gradient, and the chimeric AEA RNase H bound to the SP column, eluting
in a manner similar to the E. coli protein. The proteins of
interest were pooled and concentrated by centrifugation at 5,000 rpm in
a filtron device (cutoff, 10 kDa). Enzymes were stored at
70 °C in
a solution containing 45% glycerol, 25 mM Tris-HCl, pH
7.9, 0.5 mM dithiothreitol, 25 mM NaCl, 0.25 mM EDTA, and 0.5 mM phenylmethylsulfonyl
fluoride.
Removal of His6tag was performed by treating 100 µg of His6-tagged proteins with 1 unit of biotinylated thrombin protease at 25 °C for several hours. His-Bind® resin was added to the reaction mixture and mixed for 30 min at room temperature. After spinning, the supernatant containing the protein without His6-tag was further mixed with 100 µl of streptavidin-agarose slurry for another 30 min at room temperature. After centrifugation, the supernatant was collected and dialyzed overnight at 4 °C against buffer A.
RNase H Activity AssayRNase H activity with
-32P-labeled poly(rA)·poly(dT) as substrate was
determined essentially as described (43) by measuring trichloroacetic
acid-soluble radioactivity. In situ renaturation gel assays
were performed as described (44).
Four different substrates were used to determine differences in
products generated by the various enzymes. Details of the conditions of
the assays are given in the figure legends for each substrate.
Polyacrylamide gel analysis of the products was performed according to
Zhan et al. (45). The substrate for minus-strand primer
removal assay was prepared according to Smith and Roth (38) and Stahl
et al. (39). The uniformly
[-32P]CTP-labeled polypurine substrate was synthesized
according to Guo et al. (13) and uniformly
[
-32P]CTP-labeled ovalbumin mRNA160
was prepared as described by Oyama et al. (46) and Post
et al. (42). The final products were visualized after
electrophoresis on a 6 or 8% sequencing gel by autoradiography with
Kodak X-Omat film or were exposed to a phosphor screen and quantified
using the ImageQuant program (Molecular Dynamics).
The 5-biotinylated RNA36
oligonucleotide was annealed to its complementary DNA36
oligonucleotide at a ratio 1:20 in the diethyl pyrocarbonate-treated
HBS buffer (20 mM HEPES, pH 7.5, 150 mM NaCl,
0.5 mM EDTA, 0.05% Tween P20, a non-ionic detergent). A sensor chip with streptavidin-modified surface was first immobilized with 20 µl of RNA-DNA hybrid and then blocked with biotin (for details see Ref. 32). 45 µl of four different concentrations of RNase
H, which were dialyzed against to HBS buffer before analysis, were
injected onto the RNA-DNA surface at 20 µl/min. The surface was
regenerated with one injection of 10 µl of 2 M NaCl. The
sensorgrams were analyzed by subtracting sensorgrams obtained from the
control surface, in which a sensor chip is loaded with streptavidin
followed by biotin but not with RNA-DNA. Kinetic constants were
calculated using BIAcoreTM evaluation software (see also Ref. 32 for
details).
RNases H of E. coli,
C175-AKR-RNase H, and chimeric AEA were purified from E. coli MIC1066 (rnhA-339::cat recB270(TS)) to ensure that only the overexpressed proteins contribute to the activity
observed. Versions of both proteins with or without an N-terminal
His6-tag were studied. Purification of the
His6-tagged proteins was facilitated by affinity
chromatography using nickel nitrilotriacetic acid-agarose columns.
Results of analysis of the purified proteins by SDS-polyacrylamide gel
electrophoresis are shown in Fig. 1 as
follows: a shows a Coomassie-stained gel, and b
shows a renaturation gel assay for RNase H activity. E. coli
RNase H labeled with the His6-tag purified essentially as it would without the His6-tag. Expression of the
C175-AKR-RNase H enzyme is relatively poor, and therefore, we purified
the His6-tagged version of this protein and, in some
studies, removed the His6-tag by proteolysis with thrombin.
C175-AKR-RNase H is found in the flow-through of both DEAE and SP
columns, in contrast to the E. coli protein, which flows
through the DEAE column but binds to the SP column, eluting at around
150 mM NaCl. Some preparations of the C175-AKR-RNase H
contain small amounts of a 12-kDa protein corresponding to the
N-terminal portion of C175-AKR-RNase H, as judged by its binding to
nickel nitrilotriacetic acid-agarose and its interaction with antisera
directed against the N-terminal portion of the RNase H region. This
fragment probably results from cleavage at a site previously shown to
be susceptible to proteolysis in E. coli when AKR-MuLV
reverse transcriptase is expressed and purified (47). We could not
detect any contribution of this AKR-MuLV fragment to RNase H activity
or to binding to RNA-DNA hybrids. The chimeric AEA RNase H purifies in
a similar manner to C175-AKR-RNase H and appears to be homogeneous
(data not shown). For simplicity in the presentation in the figures, the C175-AKR-RNase H is designated as A, E. coli
RNase HI as E, and the chimeric RNase H as
AEA.
Catalytic Activity of the C175-AKR-RNase H
Previous studies of MuLV RNase H have been more qualitative than quantitative but demonstrate clearly that this domain has RNase H activity (14). One rather novel feature of MuLV reverse transcriptase is a strong preference for Mn2+ (2.5 mM) rather than Mg2+ (5 mM) as the divalent cation when assaying for either polymerization or RNase H activity. To examine activity and metal ion preference, we assayed the C175-AKR-RNase H using poly(rA)·poly(dT) as substrate. The purified soluble C175-AKR-RNase H degrades RNA, preferring Mn2+ over Mg2+ as the divalent cation (Table I and Fig. 2). Products of C175-AKR-RNase H digestion of poly(rA)·poly(dT) were analyzed by gel electrophoresis and quantified using Molecular Dynamics PhosphorImager and ImageQuant (Table I). The specific activity of the C175-AKR-RNase H is very similar to MuLV reverse transcriptase-associated RNase H activity with a preference for Mn2+ over Mg2+ (about 10-20-fold) but differs significantly from that of E. coli RNase HI, which prefers Mg2+ over Mn2+ more than 150-fold.
|
The size distribution of products obtained with poly(rA)· poly(dT) as substrate differs markedly depending on the enzyme (Fig. 2). Since the specific activity of each enzyme is different, various amounts of RNase H were added, and the samples were removed at various times and products analyzed on polyacrylamide gels. The distribution of products in either Mg2+ or Mn2+ is very similar when AKR-MuLV RT provides the RNase H activity with the most abundant products at 10 min being 17- and 18-mer. At later times, when little or no initial substrate remains, smaller products accumulate. E. coli RNase HI products are significantly smaller than those seen with AKR-MuLV RT using Mg2+ as cation and are even smaller when Mn2+ is present in the reaction. C175-AKR-RNase H (A in Fig. 2) generates a very different set of products from RT in the presence of Mg2+ or Mn2+ with a long ladder of oligomers of increasing length at 30 min (lane 3) and products ranging from monomers to 14-mers at 150 min (lane 4). In contrast to E. coli RNase HI, C175-AKR-RNase H gives larger products in the presence of Mn2+, almost as large as those seen with AKR-MuLV RT (lanes 4).
Products of Cleavage Using a Synthetic mRNA-d20-mer HybridTo examine cleavage of a well characterized mRNA
substrate for E. coli and MuLV RNases H, we used a portion
of the ovalbumin mRNA (RNA160) annealed to a 20-mer DNA
primer, which starts 2 nt away from the 3-end of the RNA (Fig.
3). Experiments performed with this and
subsequent substrates used Mg2+ as the divalent cation,
since it is thought to be the relevant ion in vivo. From
previous studies (42, 46), it is known that cleavage of
RNA160 is due to binding of the polymerase active site at
the 3
-OH of the d20-mer, placing the RNase H active site 15 base pairs
from the 3
-OH giving an RNA product of 153 nucleotides (5- and 10-min
time points for RT in Fig. 3). Incubation for longer times yielded
additional products resulting from 3
-OH-independent cleavages (40-min
RT in Fig. 3). E. coli RNase HI degrades this substrate very
differently from RT-RNase H with some products arising from hydrolysis
as close as three base pairs from the 3
-OH of the d20-mer.
C175-AKR-RNase H gives products very similar to E. coli
RNase HI; particularly notable is the absence of the 153-nt product
seen with RT.
tRNA-primer Substrate Cleavage
The primer of minus-strand DNA
of RT of MuLV is from a tRNAPro annealed to a site some 145 nt from the 5-end of the genomic RNA of MuLV. Synthesis of the first
strand of DNA is followed by initiation of plus-strand DNA synthesis at
the polypurine tract (PPT). Replication of the
tRNAPro-primer yields an RNA-DNA hybrid that is cleaved by
RT-RNase H removing the tRNAPro. To test for C175-AKR-RNase
H specificity, we used a model substrate tRNAPro, similar
to those used by Smith and Roth (38) and Stahl et al. (39),
to analyze tRNALys3 removal by the RNase H of HIV-1. This
substrate is a synthetic d69-mer with a primer binding site sequence
(PBS) at its 3
-terminus to which is annealed an 18-mer RNA containing
the 3
-terminal tRNAPro sequence complementary to the PBS.
The Klenow fragment of E. coli DNA polymerase I is used to
synthesize and label (using [
-32P]dGTP) the model
substrate (Fig. 4). This defined
substrate has an RNA-DNA hybrid with an RNA-DNA covalent junction
similar to the natural in vivo substrate. RNase H of MuLV
reverse transcriptase cleaves similar substrates to produce two
products, one with a single ribonucleotide A attached to the DNA and a
second which has the ribonucleotide A removed, resulting from
hydrolysis at the RNA-DNA junction (48). We find AKR-MuLV RT yields two
products, one whose size is consistent with cleavage at the RNA-DNA
junction and the other leaving the ribonucleotide A attached to the DNA (marked by the arrow in Fig. 4). E. coli RNase HI
rarely cleaves at an RNA-DNA junction (38-40, 42) but generates the
products shown in Fig. 4 resulting from hydrolysis at a point 3 to 4 nt from the RNA-DNA junction. C175-AKR-RNase H products are very similar
to those derived with the E. coli enzyme, and only after extensive degradation (A at 20 min in Fig. 4) is there even
a small amount of product containing a single ribonucleotide A.
PPT Substrate
For initiation of plus strand synthesis, RNase
H of RT cleaves genomic RNA, avoiding degradation within the PPT
sequence to yield specific short RNA primers (18-20). E. coli RNase HI has no sequence specificity and cleaves within the
PPT. We investigated cleavages of substrates containing PPT RNA by
annealing any of four d40-mers to a synthetic 32P-labeled
RNA containing the PPT sequence. Each d40-mer positions the PPT
different distances from the 3- and 5
-ends of the DNA oligomer.
Previous studies (13, 49, 50) demonstrate that processing of these PPT
RNA substrates depends on the RNA-DNA hybridized region. As seen in
Fig. 5 for substrate I, RT cleaves mainly
18 nt from the 3
-OH and generates two primary products (211 and 256 nts). In contrast, E. coli RNase HI cleaves throughout the
d40-mer region producing fragments much smaller than those seen with RT
(about 196 and 236 nts). The pattern of products for C175-AKR-RNase H
is very similar to that seen with E. coli RNase HI which is
true for all of the PPT substrates (Fig. 5).
In the case of the RNase H domain of AKR-MuLV reverse transcriptase,
the cleavage pattern is basically similar to that of E. coli
RNase HI. With substrate II, as analyzed by Guo et al. (13),
the cleavage is somewhat 3-OH-dependent, i.e.
the reverse transcriptase was positioned about 18 nt from the
3
-terminus of the DNA strand and, thus, in the PPT region, resulting
in degradation in this region. As expected, the cleavages catalyzed by
the C175-AKR-RNase H again are similar to those seen with E. coli RNase HI, which does not display a 3
-OH dependence and whose
5
products are much smaller (about 4 nt from 3
-terminus of the DNA
strand). With substrate III and IV, reverse transcriptase produces
larger 5
products, in contrast to E. coli RNase HI and
C175-AKR-RNase H, which accumulate smaller 5
products.
The N terminus of the RNase H protein used in the preceding experiments has six consecutive histidine residues and a thrombin cleavage site between the His6-tag and the RNase H. The possibility exists that the His6-tag may affect activity and cleavage specificity (38). When we treated His6-tagged proteins with thrombin or when we used purified proteins expressed without the additional N-terminal amino acids, no differences were seen in specific activity or cleavage specificity for any of the substrates tested (data not shown). Hence, the N-terminal His6-tag and thrombin cleavage site do not alter the cleavage patterns. From product analysis with either nonspecific substrates such as poly(rA)·poly(dT) and ovalbumin RNA160 or specific substrates, such as the PBS sequence and PPT sequence, the altered cleavage specificity appears to be an intrinsic property of the isolated C175-AKR-RNase H.
Previously, we have shown that the HIV-1 RNase H domain protein can be
converted from an inactive protein to an active enzyme by inserting the
C region of E. coli RNase HI into the
homologous region on the HIV-1 protein (39). To determine if
replacement of the basic protrusion (
B,
C, and
D) of the C175-AKR-RNase H alters
any of its enzymatic properties, we constructed a gene that produces a
chimeric RNase H with the amino portion derived from MuLV, the basic
protrusion corresponding to that of E. coli RNase HI, and
the C-terminal part coming from the MuLV protein (we designate this
RNase H as AEA). No significant differences were found between the
C175-AKR-RNase H and the AEA enzyme for (i) divalent cation preference,
(ii) specific activity, or (iii) any of the products generated using
all of the substrates described earlier.
As
reported elsewhere (32), we have shown that E. coli RNase HI
has fast association and dissociation rates when interacting with
RNA-DNA hybrids. It was also possible to demonstrate that an E. coli RNase HI lacking a significant portion of the basic protrusion has no enzymatic activity and a dramatically lower on-rate
compared with the wild-type protein. Furthermore, a point mutation
changing a highly conserved Asp residue to Ala results in a protein
that has no RNase H activity but retains the ability to interact with
RNA-DNA at rates similar, if not identical, to the wild-type enzyme. To
determine the binding properties of the C175-AKR-RNase H and the
AEA-RNase H, we used the BIAcore instrument to determine kinetic
parameters. Results of experiments in which we immobilized a
36-base pair RNA-DNA hybrid via a
streptavidin-biotin linkage (see "Experimental Procedures") to the
sensor chip and passed solutions of purified RNases H over the sensor
surface are shown in Fig. 6 . Table II
summarizes the kinetic parameters obtained from these experiments
which show that both versions of the MuLV RNases H bind with an
association rate (KA) of about 10% and a
dissociation rate (KD) approximately half that of
E. coli RNase HI. This is in contrast to a 103
difference in KA between that of native E. coli RNase HI and that of E. coli RNase HI whose basic
protrusion region had been deleted (32). Thus, the 10-fold difference
in binding of MuLV RNase H and E. coli RNase HI can account
for some of the difference in specific activities of the two enzymes,
but most of the difference must be due to difference in catalysis
rates.
|
Here we present evidence that a protein comprising the RNase H
domain of MuLV RT has RNase H activity comparable to the RNase H of
AKR-MuLV RT, and it maintains a preference for Mn2+ rather
than Mg2+. Thus, most or all of the determinants necessary
for RNase H activity reside in the C-terminal 175 amino acids of RT.
This retention of activity is in marked contrast to similar RNase H domain proteins derived from HIV-1 RT which has no RNase H activity (34-36). By adding either a His6-tag to the HIV-1 RNase H
protein (38) or by inserting the C region of E. coli RNase HI (39, 40), which is normally absent from HIV-1 RNase
H, enzymatic activity can be restored. Interestingly, these HIV-1
RNases H retain specificity for cleavage of a tRNALys3
model substrate (38-40).
In contrast, we show here that the C175-AKR-RNase H cleaves several substrates in a very different manner from AKR-MuLV RT, suggesting that specificity of hydrolysis by the retroviral RNase H is dictated by the remainder (non-RNase H domain) of the protein (polymerase and connection regions). In addition, substitution of the basic protrusion region of MuLV RNase H by the homologous region from E. coli RNase HI does not alter any of the properties of the RNase H that we tested. Except when using poly(rA)·poly(dT) as substrate (Fig. 2), C175-AKR-RNase H yields products similar to E. coli RNase HI from the tested substrates. Even with the poly(rA)·poly(dT) substrate, products formed in the presence of Mg2+ resemble those found for the E. coli enzyme, although the size of the final products tends to be slightly larger for the C175-AKR-RNase H than for E. coli RNase HI. When the divalent cation in the reaction is Mn2+, oligonucleotide A of 11-16 nt are the most abundant products, more closely resembling the fragments produced from the same substrate by RT.
E. coli RNase HI, MuLV RNase H, and HIV-1 RNase H all
probably have similar three-dimensional structures, but HIV-1 lacks the
C loop of E. coli RNase HI. Mutations in the
C-loop, including a deletion (31, 32), provide evidence
that this region is involved in substrate binding. Lack of RNase H
activity of the HIV-1 RNase H protein results from its poor ability to
bind to the substrate, as suggested by recovery of activity when a
His6-tag is added at the N or C terminus of the protein or
when the
C-loop of E. coli RNase HI is
inserted into the HIV-1 RNase H protein. Activity of these HIV-1 RNases
H is highest when Mn2+ is present in the reaction mixture,
with less than 0.1% activity being detected with Mg2+
(38-40). Studies using surface plasmon resonance (Fig. 6 and
Table II) indicate that the binding of the C175-AKR-RNase H is similar to AEA but only 10% that of E. coli RNase HI. Although this
difference in binding is significant, it seems unlikely to
account for 104 differences between the specific activities
of the E. coli and MuLV enzymes (42). However, the loss of
specificity described here is remarkable only when compared with the
HIV-1 RNase H results.
The physical arrangement of the two catalytic sites suggests that
polymerase and RNase H are coupled during reverse transcription (8, 9,
46). A number of experiments underscore the strong influence of
polymerase on RNase H activity, sometimes directing the RNase H not to
cleave (e.g. at the PPT) and sometimes dictating the site of
cleavage. Palaniappan et al. found (49) that nevirapine, an
inhibitor of the DNA polymerase activity of HIV-1 reverse
transcriptase, alters the cleavage specificity of the RNase H. These
authors propose that interruption of binding of the polymerase domain to nucleic acids permits reverse transcriptase to move away from the
3-terminus of the DNA strand, resulting in random cleavages. The
nucleocapsid protein NCp7 facilitates delivery of RNA-DNA into the
active site of the RNase H domain of HIV-1 reverse transcriptase, thereby altering the rate and, to a slight extent, the pattern of
cleavage (50). The results with NCp7 are also explained in terms of
restricting the RT to its RNase H activity. Thus, AKR-MuLV RNase H has
different specificity when free of the constraints imposed when it is
attached to the polymerase domain. This is another example of the
important role of the polymerase region for RNase H activity documented
in the literature.
Judging from the specificity of cleavage of RNase H of RT, and from the strong influence of the polymerase domain on RNase H activity, it is not surprising that the site of cleavage is dictated by the non-RNase H region. What remains unexplained is how HIV-1 RNase H domain proteins, whose RNase H activity has been restored by adding a His-tag or insertion of the E. coli RNase HI handle region, retain the ability to correctly cleave a model tRNALys3 substrate. E. coli RNase HI has no known cleavage specificity, has a very high specific activity, and has a strong preference for Mg2+. During evolution, the RNase H domain of HIV-1 may have lost the handle region that greatly affects the binding properties of the E. coli enzyme to its substrate, whereas the MuLV RNase H retains this basic loop and, as a result, binds with an association rate that is 10% that of the E. coli protein. The amino acids and structures responsible for the differences in these biochemical properties remain to be defined.
We thank Dr. Judith Levin and Dr. Jianhui Guo for AKR-MuLV reverse transcriptase as well as for valuable discussions.