Presence of 18-Å Long Hydrogen Bond Track in the
Active Site of Escherichia coli DNA Polymerase I (Klenow
Fragment)
ITS REQUIREMENT IN THE STABILIZATION OF ENZYME-TEMPLATE-PRIMER
COMPLEX*
Kamalendra
Singh and
Mukund J.
Modak
From the Department of Biochemistry and Molecular Biology,
University of Medicine and Dentistry, New Jersey Medical
School, Newark, New Jersey 07103
Received for publication, November 11, 2002, and in revised form, December 23, 2002
 |
ABSTRACT |
The analysis of the active site region in the
crystal structures of template-primer-bound KlenTaq (Klenow fragment
equivalent of Thermus aquaticus polymerase I) shows the
presence of an ~18-Å long H-bonding track contributed by the Klenow
fragment equivalent of Asn845, Gln849,
Arg668, His881, and Gln677. Its
location is nearly diagonal to the helical axis of the template-primer. Four base pairs in the double stranded region proximal to 3' OH end of
the primer terminus appear to interact with individual amino acid
components of the track through either the bases or sugar moieties. To
understand the functional significance of this H-bonding network in the
catalytic function of Klenow fragment (KF), we generated N845A, N845Q,
Q849A, Q849N, R668A, H881A, H881V, Q677A, and Q677N mutant species by
site-directed mutagenesis. All of the mutant enzymes showed low
catalytic activity. The kinetic analysis of mutant enzymes indicated
that Km.dNTP was not significantly
altered, but KD.DNA was significantly increased. Thus the mutant enzymes of the H-bonding track residues had
decreased affinity for template-primer, although the extent of decrease
was variable. Most interestingly, even the reduced binding of TP by the
mutant enzymes occurs in the nonproductive mode. These results
demonstrate that an H-bonding track is necessary for the binding of
template-primer in the catalytically competent orientation in the pol I
family of enzymes. The examination of the interactive environment of
individual residues of this track further clarifies the mode of
cooperation in various functional domains of pol I.
| |
INTRODUCTION |
Sequence alignment of various nucleic acid polymerases has
revealed the presence of several conserved motifs (motifs A-E) in
diverse DNA polymerase families (1-3). Two of these motifs (motifs A
and C) are absolutely conserved in all DNA polymerase families, whereas
the conservation of other motifs is restricted to individual families.
Two conserved aspartates, one belonging, respectively, to motifs A and
C each, serve as ligands for divalent cations (4-6) and tags for the
active site location of DNA polymerases (1, 2, 7).
The Klenow fragment (KF)1 of
Escherichia coli DNA polymerase I has been extensively used
as a model system for understanding the structural basis for the
polymerase reaction mechanism. The crystal structures have shown the
68-kDa KF folds into two distinct domains, a ~200-amino acid long
3'-5' exonuclease domain at the N-terminal, and a ~400-residue
polymerase domain at the C-terminal (8-10). The polymerase domain,
which anatomically resembles a half-open right-hand, consists of a
cleft formed by three subdomains designated as fingers, palm, and thumb
(8, 10). These subdomains have been implicated in different functions
of polymerases (7). The active site constituted by motifs A and C
resides at the palm subdomain, the dNTP binding site at the fingers
subdomain, and the template-primer binding site at the thumb subdomain
(7).
The crystal structures of several DNA polymerases complexed with
template-primer and dNTP have shown the participation of individual
amino acid residues in the respective polymerase function (4, 6, 11,
12). In particular, the crystal structures of binary and ternary
complexes of the Klenow fragment of Thermus aquaticus DNA
polymerase I (KlenTaq) can be considered to represent the snapshots of
initial steps of the nucleotidyltransferase reaction pathway of DNA
polymerases (11). The analyses of these crystal structures together
with the crystal structure of apo-KlenTaq revealed that the binding of
template-primer to the enzyme induces a conformational rearrangement of
the side chains of several residues present at the active site. We
noted that resulting from this conformational change in the side chains
is the formation of a predominately hydrogen bonded network of ~18 Å long at the bottom of the cleft (Fig. 1). This network of hydrogen
bonding (called "H-bonding track" hereafter) is present in both
binary and in the "open" and "closed" conformations of the
ternary complexes. The residues that participate in the formation of
the H-bond track are highly conserved in the pol I family of DNA
polymerases. They are equivalent to the E. coli pol I
residues Asn845, Gln849, Arg668,
His881, and Gln677.
To understand the significance of the H-bonding track residues at the
active site of pol I and their functional contribution, we carried out
site-directed mutagenesis of Asn845, Gln849,
Arg668, His881, and Gln677. These
residues were mutated to both conserved and nonconserved (alanine)
residues. Properties of the mutant proteins were then investigated to
assess their individual participation in the catalytic function of KF.
All of the mutants showed decreased catalytic activity, albeit to
different degrees, suggesting the involvement of these residues in the
catalytic function of pol I. Most importantly, all of the mutant
enzymes showed significantly reduced (15-75-fold) template-primer
binding affinity compared with wild-type enzyme. Furthermore,
neither the conservative nor the nonconservative mutant enzymes showed
ability to form prepolymerase binary or ternary
complexes. Similarly, unlike the wild-type enzyme, the enzyme-TP
covalent complexes generated for individual mutant species of all five
residues failed to catalyze in situ addition of substrate dNTP. These results together with the structural analysis of the binary
and ternary complexes of pol I family DNA polymerases strongly implicate the function of H-bonding track residues in the binding of
template-primer at the proper position and orientation, which is
prerequisite for the nucleotidyltransferase function of polymerases. An
examination of the location of some of the highly conserved residues
present in the different motifs of pol I, in the context of the
H-bonding track, further reveals the coordinated participation of all
of these motifs in the catalytic function.
| |
EXPERIMENTAL PROCEDURES |
Materials--
All mutant enzymes were generated from the
plasmid (pCJ141) generously provided by Dr. Catherine Joyce of Yale
University. This construct contains the E. coli
polA gene encoding the Klenow fragment (13). The maintenance
and expression strains for plasmid pCJ141 (E. coli CJ406 and
E. coli CJ376, respectively) were also obtained from Dr. Joyce.
Enzymes--
PFU-turbo polymerase used for PCR-based
site-directed mutagenesis was from Stratagene. Restriction enzymes were
from Roche Molecular Biochemicals. Polynucleotide kinase from either
Invitrogen Corp. or PerkinElmer Life Sciences was used.
Reagents--
The PCR grade dNTPs were from Roche Molecular
Biochemicals. Radiolabeled nucleotides were obtained from PerkinElmer
Life Sciences. The DNA extraction kit was from Qiagen, whereas DNA
oligonucleotides were from MWG-Biotechnologies. All
32P-5'-end labeled oligomers were purified by denaturing
polyacrylamide-urea gel electrophoresis.
In Vitro Site-directed Mutagenesis--
We used the PCR-based
protocol described in Stratagene's QuikChange site-directed
mutagenesis kit to generate the desired mutations of KF. The plasmid
pCJ141 (13, 14) was used for the generation of KF protein and to
construct the desired H-bonding track mutant derivative. This plasmid
contains the E. coli KF gene carrying a mutation, D424A.
This mutation confers deficiency in 3'-5' exonuclease
activity. The wild-type KF used for this study and its
Asn845, Gln849, Arg668,
His881, and Gln677 mutants are therefore
exonuclease-deficient. All mutations were confirmed by DNA sequencing.
Expression and Purification of the Wild-type and Mutant
Proteins--
Isolation and purification of N845A, N845Q, Q849A,
Q849N, R668A, H881A, H881V, Q677A, and Q677N mutant enzymes were
carried out as described before (15). Plasmid DNA from mutant clones was used to transfect E. coli CJ376, an expression strain
used for this study (13, 16). Briefly, an overnight inoculum of the
expression strain was used to initiate a 500-ml cell culture at
30 °C, in an incubator shaker. At A595 = 0.3, the incubation temperature was raised to 42 °C to heat induce
overproduction of the enzyme. After 4-5 h of incubation, cells were
harvested, washed, and resuspended in cell lysis buffer (50 mM Tris-Cl, pH 8.0, 500 mM NaCl, 1 mM phenylmethylsulfonyl fluoride) containing 2 mg/ml
lysozyme. Following a 30-min incubation at 4 °C, the cell suspension
was sonicated and centrifuged (14,000 rpm for 30 min), and the
supernatant was passed through a DEAE column to remove DNA. The
flow-through was fractionated with ammonium sulfate, using 60 and 85%
saturations. The pellet obtained with 85% ammonium sulfate was
resuspended in 5 ml of buffer I (50 mM Tris-HCl, pH 7.0, 1 mM dithiothreitol (DTT), 1 mM EDTA), dialyzed
overnight against 1 liter of the same buffer, and applied to a Bio-Rex
70 column prewashed with Buffer I. A 50-500 mM linear
gradient of NaCl in Buffer I was used to elute the bound protein. Peak
fractions (representing a 68-kDa protein on SDS-polyacrylamide gel
electrophoresis) were pooled and concentrated with polyethylene glycol
(8000). The samples were further dialyzed in a buffer containing 50 mM Tris-HCl, pH 7.0, 1 mM DTT, 100 mM NaCl. Protein concentrations were determined by the
Bradford colorimetric assay (17) and the enzyme stocks (in 30%
glycerol) were stored at 20 °C.
Specific Activity Determination--
Enzymatic activity of the
various mutants was determined at 37 °C for 5 min on one
heteropolymeric (49/17-mer) and two homopolymeric (dA36/dT18 and
dC60/dG18) template-primers (Chart
1). The reaction was carried out in a
final volume of 100 µl containing 50 mM Tris-HCl, pH 7.8, 1 mM DTT, 100 µg/ml bovine serum albumin, 250 nM of the template-primer, 5 mM
MgCl2, and 25 µM [32P]dNTP (0.5 µCi/assay) corresponding to the homo- and heteropolymeric template-primers. Reactions with the heteropolymeric template-primer contained all four dNTPs at a concentration of 25 µM,
with two of them radiolabeled. Unless otherwise indicated, the final
enzyme concentration was 7.5 nM. The reaction was initiated
by the addition of MgCl2 and terminated by the addition of
5% ice-cold trichloroacetic acid containing 10 mM
PPi. The acid-precipitated material was collected on
Whatman GF/B filters and counted for radioactivity in a liquid
scintillation counter, as described elsewhere (18).
KD.DNA Determination--
To determine the
binding affinity of template-primer to the desired enzymes, gel
mobility shift assays were performed as described previously (19). The
binding of 33/20 + ddC-mer and 21/12 + ddC, the dideoxy-terminated
template-primers (Chart 1), to various concentrations of enzyme was
carried out in a reaction mixture containing 50 mM
Tris-HCl, pH 7.8, 5 mM MgCl2, 10% (v/v)
glycerol, and 0.1 mg/ml bovine serum albumin. The concentration of
32P-labeled template/primer was 50-100 pM.
Different protein concentrations were used to bracket the
KD.DNA value. Samples were
electrophoresed at 100 V for 1.5 h at 4 °C on a 6% nondenaturing
polyacrylamide gel, using 89 mM Tris borate, pH 8.2, buffer. Gels were dried, then scanned in a PhosphorImager and
quantitated by ImageQuant software (Amersham Biosciences). The
percent enzyme-TP binding was calculated by quantifying the amount of
uncomplexed TP in each lane. Percent binding values were then used for
the determination of KD.DNA by
interpolation, using nonlinear regression for one-site binding
(hyperbola) with GraphPad Prism software.
Determination of Steady-state Kinetics
Constants--
Determination of steady-state kinetic parameters was
carried out as described earlier (15,18) on homopolymeric
dC60/dG18 and dA36/dT18
template-primers, in the presence of varying amounts of complementary
dNTPs (dGTP and dTTP, respectively). Saturating concentrations of a
homopolymeric dA36/dT18 and
dC60/dG18 were used as the template-primers, to
mimic the first-order reaction conditions. The reaction
mixture contained 50 mM Tris-HCl, pH 7.8, 50 mM
KCl, 1 mM DTT, 0.01% bovine serum albumin, 1 µM respective template-primers, 1 µCi/reaction
[
-32P]dNTP, 1-1000 µM respective dNTPs,
5 nM WT KF, and activity normalized mutant enzymes. The
reaction was initiated by the addition of 5 mM
MgCl2. The acid-precipitable material was filtered on a
Whatman glass fiber filter and the radioactivity was quantitated by
scintillation counting. Steady-state parameters Km
and Vmax were determined from Eadie-Hoftsee
plots, using the Enzyme Kinetic program. Steady-state values of
kcat were calculated from the equation:
Vmax = [ E]total kcat.
Stable Ternary Complex Formation Assay--
Our method for
assessing stable ternary complex formation is based on the procedure
described by Scott and colleagues (20, 21), where nondenaturing gel
electrophoresis is used to identify the enzyme-TP-dNTP ternary complex.
Briefly, in a final volume of 10 µl, the desired enzyme quantity (1.2 nM of the wild-type and different amount of mutant enzymes)
is mixed with 21-mer 5'-end labeled dideoxynucleotide-terminated primer
annealed to a 33-mer template (final concentration of primer is ~100
pM) in a standard buffer containing 50 mM
Tris-HCl, pH 7.8, 10% (v/v) glycerol, 5 mM
MgCl2, and 0.1 mg/ml bovine serum albumin. The enzyme-DNA complexes were allowed to form on ice. To one of the duplicate samples,
cognate nucleotide was added at 200 µM concentration to
induce ternary complex formation. The stability of the enzyme-TP binary
complex and enzyme-TP-dNTP ternary complex was then assessed by the
degree of persistence of DNA in the enzyme-TP complex, upon addition of
500-fold excess of the same unlabeled template-primer. We find that
under these conditions, DNA from enzyme-TP binary complexes completely
dissociates, whereas ternary complexes exhibit significant retention of
radiolabeled template-primer by enzyme protein. All samples were
resolved on a native 6% polyacrylamide gel, and the positions of the
radioactive bands representing enzyme-DNA and free DNA were visualized
by exposure in a PhosphorImager.
Effect of Complementary or Noncomplementary Incoming Nucleotide
on Template-primer Binding Affinity--
The effect of the
complementary and noncomplementary nucleotide on the binding affinity
of wild-type and mutant enzymes was assessed in a manner similar to
that used for the stable ternary complex formation assay. The enzyme-TP
binary complexes were allowed to form in 50 mM Tris-HCl, pH
7.8, 10% (v/v) glycerol, 5 mM MgCl2, and 0.1 mg/ml bovine serum albumin for 10 min at 4 °C in duplicate. To one
of the duplicate samples, the correct incoming dNTP (200 µM dGTP) was added. The incorrect dNTPs (dATP, dCTP, and
dTTP, 100 µM each) were added to the second sample. The
reaction mixture was further incubated on ice for 10 min, prior to
loading on a 6% nondenaturing polyacrylamide gel, under the same
conditions as those used for KD.DNA determination.
Enzyme-Template-primer Cross-linking--
The UV-mediated
photochemical enzyme-TP cross-linking was performed essentially as
described before (18, 22, 23). The reaction mixtures containing 50 mM Tris-HCl, pH 7.8, 5 mM MgCl2, 1 mM DTT, 32P-5'-end labeled 33/20 + ddC or 21/12 + ddC template-primer and the desired amount of enzyme were incubated
on ice for 10 min. The samples were exposed to UV light
(wavelength = 254 nm) at a dose rate of 3700 mJ/cm2
for 3 min. The wild-type KF was 30 pmol, and the mutant proteins were
varied to obtain nearly equal extent of the enzyme-template-primer cross-linked species. Measurement of covalent attachment of labeled TP
or dNTP to enzyme protein was assessed by 8% SDS-polyacrylamide gel
electrophoresis, followed by the exposure of gel to PhosphorImager and
analysis of exposed gel by ImageQuant software (Amersham
Biosciences).
In Situ Addition of Nucleotide on Immobilized TP with
Enzyme--
Nucleotidyltransferase activity of the wild-type and
mutant enzymes, containing covalently cross-linked template-primer, was carried out as described previously (22). The wild-type KF and various
mutant enzymes were cross-linked to 32/19-mer template-primer in a
reaction mixture, in a final volume of 50 µl, containing 35 pmol of
the unlabeled TP, 50 mM Tris-HCl, pH 7.8, 1 mM
DTT, and 5 mM MgCl2. The reaction contained 30 pmol of the wild-type KF, whereas the amount of individual mutant
enzymes was varied to yield near equal quantities of enzyme-TP
cross-linked species. The nucleotidyltransferase reactions were then
initiated by the addition of 1 M NaCl together with 5 µCi
of complementary [-32P]dTTP at a final concentration
of 0.5 µM. The reaction mixture was incubated for 30 min
at room temperature and terminated by the addition of protein
solubilizing solution. An aliquot of the reaction mixture was subjected
to SDS-polyacrylamide gel electrophoresis, followed by phosphorimaging
and analysis by ImageQuant software (Amersham Biosciences).
| |
RESULTS |
The structural analysis of the template-primer and
template-primer-dNTP bound crystal structures of KlenTaq and
Bacillus stearothermophilus (11, 24) showed the presence of
a network of interactions (mainly H-bonds) near the active site of
these polymerases. The constituent residues in these two enzymes
(KlenTaq and Bst) are equivalent to Asn845,
Gln849, Arg668, His881, and
Gln677 of E. coli pol I (Fig.
1). In the template-primer-bound binary complex of KlenTaq (Protein Data Bank file 4ktq), the H-bonds exist between Asn845-Gln849,
Gln849-Arg668, and
His881-Gln677. The interaction between
Arg668 and His881 (which would complete the
H-bonding track) is van der Waals in nature. As mentioned by Franklin
et al. (12) for T7 DNA polymerase I, the equivalent residue
to Arg668 in KlenTaq (Arg573) is buttressed by
the His881 equivalent amino acid (His784). The
fact that these crystal structures have been solved at 2.3 Å or lower
resolution, it is possible that slightly altered modeling of the side
chain conformation of His784 (equivalent to
His881 of KF) would result in an H-bond between
Arg573 and His784. Therefore, for simplicity,
we have the interactions among these five residues designated as the
H-bonding track.
View larger version (61K):
[in this window]
[in a new window]
|
Fig. 1.
Presence of an 18-Å long
H-bonding track formed by five amino acid residues of
KlenTaq. This figure shows the positions of five amino acid
residues equivalent to Asn845, Gln849,
Arg668, His881, and Gln677 of
E. coli DNA polymerase I, which participate in the formation
of H-bonding track (shown as broken cylinder in gray
color) in the Klenow fragment of Taq polymerase. This
figure has been generated from the crystal structure of the enzyme-TP
binary complex of KlenTaq (11) (Protein Data Bank file 4ktq). In
panel A, the phosphate backbones of the template
(blue) and primer (gold) are shown as
ribbons. The constituent amino acid residues of the
H-bonding track are colored in cyan and are shown as
sticks. The two metals (in magenta balls marked
A and B) at the active site and two essential
aspartic acids (shown as green sticks) as seen in the
ternary complex of KlenTaq have also been included as reference points.
Because Asp610 of KlenTaq (equivalent to Asp705
of KF) assumes two different conformations in the binary and ternary
complexes, its orientation in this figure is as seen in the "closed
ternary complex" crystal structure (Protein Data Bank file 3ktq). The
dotted lines among H-bonding track residues represent the
hydrogen bond. An extra pink dotted line between
Arg668 and His881 equivalent residues of
KlenTaq represents the van der Waals interactions between these two
side chains. The potential donor atoms of H-bonding track residues that
can form the hydrogen bond with the minor groove moieties of
template-primer (purine N3 and pyrimidine C = O groups) are shown
as blue balls. The amino acid residues and positions in
KlenTaq are labeled with KF numbers based on their structural and
positional equivalence. In panel B, a stereo-diagram
depicting another view of the H-bonding track residues is shown.
This view has been generated by rotating the figure in panel
A by about 90° in the clockwise direction. In this figure, the
template and primer are colored blue and gold,
respectively. The bonds of the template and primer are rendered in
sticks, whereas the bases and sugar moieties have been
filled by plates. The amino acid residues, rendered in
sticks, are also shown with their van der Waals surface in
different colors. These figures were generated by MolMol (44).
|
|
Site-directed Mutagenesis and Purification of the Mutant
Enzymes--
A total of 9 mutants, viz. N845A, N845Q,
Q849A, Q849N, R668A, H881A, H881V, Q677A, and Q677N of Klenow fragment
were generated by site-directed mutagenesis as described under
"Experimental Procedures" (18, 22). All side chains were replaced
by alanine, a nonhomologous side chain, or by homologous side chain,
except in the H881V mutant. The H881V mutation results in a KF sequence (VDE), which mimics the sequence present at topologically conserved positions in reverse transcriptases (VDD/MDD). The mutant enzymes were
greater than 95% pure as judged by SDS-polyacrylamide gel electrophoresis. The levels of expression, solubility, and yield as
well as the chromatographic characteristics of all mutant proteins were
identical to that of the wild-type enzyme, suggesting no significant
change in the folding pattern of the mutant enzymes.
For studies involving characterization of the DNA polymerization
reaction, template-primer binding, cross-linking of dNTP, and
template-primer and in situ addition, the WT and H-bonding track mutant proteins also contained the D424A mutation at the 3'-5'
exonuclease active site, rendering them exonuclease-deficient. Thus,
the mutant protein D424A is referred to as WT and R668A (also
containing D424A mutant) as R668A mutant for the polymerase activity studies.
Specific Activities of the Mutant Enzymes with Different
Template-primers--
The polymerase activity of mutant enzymes was
determined with two homopolymeric (dA36/dT18
and dC60/dG18) and one heteropolymeric (49/17-mer) template-primer in the presence of Mg2+ as
metal cofactor. The activities of different mutant enzymes expressed as
a percentage of WT are summarized in Table
I. All mutants showed a significantly
decreased activity with dA36/dT18 ranging
between 0.2 and 32% of the level in the wild-type enzyme. Generally,
the two His881 mutants (H881A and H881V) appeared to retain
relatively higher activity with all three template-primers, namely
dA36/dT18, dC60/dG18 (homopolymers), and 49/17-mer (heteropolymer). Among the three template-primers, the least activity of individual mutant enzymes was
seen with dA36/dT18. The most severe effect on
the catalytic activity was noted with the N845Q, Q849N, R668A, and
Q677N mutant species of KF. Surprisingly, the conserved substitutions
of residues Asn845, Gln849, and
Gln677 consistently exhibited greater loss of activity with
all of the template-primers, compared with nonconserved substituents at
these positions. These activity data suggested that the side chains of
Asn845, Gln849, Arg668,
His881, and Gln677 have some important role in
the polymerase function of KF.
View this table:
[in this window]
[in a new window]
|
Table I
Relative specific activity of mutant enzymes of H-bonding track in KF
As described under "Experimental Procedures," DNA polymerization
assays were carried out under standard conditions with the indicated
template/primers. The values are expressed relative to the wild-type
and were performed using Mg2+ as the divalent cation. One
hundred percent activity of KF with dA36/dT18 and
dC60/dG18 was 1.10 × 104 and 1.16 × 104 units/mg protein, respectively. For the 49/18-mer
template/primer, 100% activity of KF equals 1.21 × 104
units/mg of protein. One unit is defined as the activity necessary to
incorporate 1 nmol of dNMP into acid-insoluble material in 30 min at
37 °C.
|
|
Determination of Steady-state Kinetics Constants--
The
steady-state kinetic parameters, kcat for the
polymerase reaction and Km for dNTP utilization,
were determined for the WT and each mutant protein, with two
homopolymeric DNA substrates (dA36/dT18 and
dC60/dG18). The rates of incorporation of
-32P-labeled dNTP into products at increasing
concentrations of dNTP were measured, and kcat
for the polymerase reaction and Km for dNTP were
determined by Eadie-Hofstee plots from
the rate data. The results are summarized in Table II. The
wild-type enzyme exhibited a Km.dNTP of
~5 µM with both homopolymeric template-primers. The
mutant N845A showed almost no change in Km.dNTP with either template-primer,
whereas the catalytic efficiency was decreased by ~6-fold with
dA36/dT18 template-primer. The catalytic
efficiency with dC60/dG18 template-primer was
not changed significantly for the N845A mutant. However, the conserved mutant, N845Q, showed a drastic change of ~85-fold in catalytic efficiency. The mutants of Gln849 exhibited a moderate
change in Km.dNTP only with
dC60/dG18 template-primer (~5-10-fold).
However, the steady-state catalytic rate (kcat)
was significantly reduced with both template-primers (20-fold with
dA36/dT18 and 60-fold with
dC60/dG18). The change in both
Km.dNTP and kcat
was reflected in significantly decreased catalytic efficiency
(14-75-fold) of Gln849 mutants. For both,
Asn845 and Gln849, the greater effect was noted
with conserved substitutions suggesting that not only the chemical
nature of the side chain but also the size of the side chain at a
specific position is critical for KF to function optimally. The two
His881 mutants did not show significant change in
Km.dNTP; however, the
kcat of both enzymes (H881A and H881V) was
significantly (~65-fold) reduced with
dA36/dT18 template-primer. This effect was not
as pronounced with dC60/dG18 template-primer,
as only a moderate (~6-fold) change was noted in the catalytic rate
of both His881 mutants.
View this table:
[in this window]
[in a new window]
|
Table II
Kinetic constants for DNA polymerase mutant enzymes
The steady-state kinetic parameters of
Km.dNTP and kcat for
wild-type Klenow fragment and mutants of H-bonding track enzymes were
measured under conditions where template-primer was saturating and dNTP
variable. Kinetic analysis was performed with the indicated
template-primers and corresponding dNTP substrate as described under
"Experimental Procedures." KD.DNA was
measured by gel retardation assays (38).
|
|
The mutants of Gln677 (Q677A and Q677N) also showed a
defect in their catalytic efficiency. The efficiency of the Q677A
mutant was reduced by ~10- and 120-fold on
dA36/dT18 and dC60/dG18
template-primers, respectively. As noted for Asn845 and
Gln849, the efficiency of the homologous mutant of
Gln677 (Q677N) was significantly more reduced than that of
its nonhomologous counterpart (Q677A). The catalytic efficiency was
reduced by nearly 3000-fold on dA36/dT18 and by
~300-fold with dC60/dG18 template-primer. Similarly, a reduction of about 300-fold in the catalytic efficiency of
the R668A mutant enzyme was noted with
dA36/dT18 template-primer. Similar (type of)
template-specific inactivation was also noted with the mutants of
O-helix residues (18). These steady-state kinetic data
revealed that most of the mutant enzymes have a compromised DNA
polymerization efficiency, suggesting some defect in the mutant of
enzymes at one or more steps of the catalytic mechanism.
Ability of Mutant Enzymes to Form E·TP Binary
Complexes--
Because the mutants of Asn845,
Gln849, Arg668, His881, and
Gln677 amino acid residues of the KF displayed reduced
polymerase activity, we examined if this decrease was because of a
defect in template-primer binding ability of mutant proteins. To
determine the binding affinity of various mutant enzymes with the
template-primer, we carried out gel mobility shift assay of the
wild-type and mutant enzymes with 33/20 + ddC and 21/12 + ddC, two
template-primers that differ in duplex length (Chart 1). The primer
moiety of 33/20 + ddC template-primer was 5'-end
32P-radiolabeled, whereas in the 21/12 + ddC
template-primer, the template strand was radiolabeled. The individual
enzyme-TP complexes were resolved on 6% nondenaturing polyacrylamide
gels. The representative patterns of the migration of enzyme-TP
complexes for each of the two template-primers (33/20 + ddC and 21/12 + ddC) are shown in Fig. 2. Panels
A and B in this figure show the formation of the enzyme-TP complex between wild-type KF and 33/20 + ddC and 21/12 + ddC
template-primers, respectively, with increasing concentrations of the
wild-type KF ranging between 0.3 and 20 nM for the 33/20 + ddC and 0.04 nM to 1.24 nM for the 21/12 + ddC.
From the comparison of the distribution of radiolabeled species between
complexed and uncomplexed positions in panels A and
B, it is clear that at 1.24 nM concentration of
the wild-type KF, nearly 85% of the 21/12 + ddC template-primer has
been stably bound. In contrast, the same amount of enzyme (4th lane
from left in panel A of Fig. 2) could bind
~55% of the 33/20 + ddC template-primer. These data suggest that the
affinity of wild-type KF for a template-primer with a short duplex
region (21/12 + ddC with only 13 base pairs) is greater than that for a
template-primer containing a longer duplex region (33/20 + ddC with 21 base pairs). The KD.DNA values for 33/20 + ddC and 21/12 + ddC template-primers were 0.60 and 0.14 nM, respectively, calculated by plotting the concentration of enzyme against that of complexed template-primer and fitting the
data to the hyperbolic curve shown in Fig. 2, panels C and D. Another interesting observation pertaining to the binding
of KF with two template-primers is the gel mobility pattern of the enzyme-TP complex, particularly at high enzyme concentrations. It
appears that with increasing concentrations of KF, a regular 1:1
enzyme-TP complex forms first, which is then shifted to a slower
migrating species (supershift position) (Fig. 2, panel A).
This kind of supershift is not seen with the use of 21/12 + ddC
template-primer (Fig. 2, panel B). Note that the last
two lanes in each panel show the pattern obtained with
75 and 150 nM enzyme concentration. Thus it appears that
the template-primer containing 21 base pairs may support sequential
binding of two enzyme molecules, albeit with different affinity.
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 2.
The template-primer binding affinity of the
wild-type KF with 33/20 + ddC and 21/12 + ddC. This figure shows
the template-primer binding affinity of wild-type Klenow fragment with
two template-primers containing different lengths in the double
stranded region. Panel A depicts the binding affinity of KF
to 33/20 + ddC (containing 21-base pair duplex), whereas panel
B shows the binding profile of 21/12 + ddC (containing 13-base
pair duplex). The lanes marked "input" represent the total amount
of radiolabeled template-primer present in the individual reaction. The
enzyme concentrations used differ for the two template-primers as the
one containing shorter duplex (21/12 + ddC) has greater affinity than
the one with longer double stranded DNA (33/20 + ddC). Thus, the enzyme
concentration used in panel A ranges between 0.3 and 20 nM, whereas that used in panel B ranges between
0.04 and 1.24 nM. The template-primer concentration used
was ~100 pM. In both panels A and
B, the last two lanes depict the pattern of
migration of the enzyme-TP complex with saturating concentrations of
the enzyme (75 and 150 nM KF, respectively). Note that with
33/20 + ddC template-primer, the migration of enzyme-TP complex changes
from the regular position to the supershifted position as the
concentration of enzyme is increased. With saturating concentration,
only the supershifted complex is seen (panel A). In
contrast, no supershifted species is formed when 21/12 + ddC is used at
any enzyme concentration (panel B). In panels C
and D, the amount of enzyme-TP complex has been plotted as a
function of increasing concentrations of the wild-type KF. The percent
enzyme-TP complexes were calculated by subtracting the free
template-primer in the individual reaction from the total input value.
The data were fit to a hyperbolic function by nonlinear regression for
the determination of KD.DNA.
|
|
The template-primer binding affinity of mutant enzymes of
Asn845, Gln849, Arg668,
His881, and Gln677 was determined in a similar
fashion as described above for the wild-type KF. A representative
profile of DNA binding by R668A and Q849N with two template-primers is
shown in Fig. 3. The estimated KD.DNA values for various mutant enzymes
with 33/20 + ddC template-primer are listed in Table II. It is clear
from this table that the template-primer binding affinity of all mutant proteins was significantly decreased. The
KD.DNA difference varied in the range of
15-75-fold with this template-primer. The difference in
KD.DNA was further increased when
measured with 21/12 + ddC template-primer. For example, the
KD.DNA for the R668A mutant with this
template-primer was ~100-fold greater than that for wild-type enzyme.
Curiously, the migration pattern of enzyme-TP complexes formed with the
21-base pair long template-primer by all the mutant enzymes appeared
only as the supershifted species (Figs. 3 and
4). Furthermore, no clear band of
enzyme-TP complex with the 21/12 + ddC template primer was observed.
Two representative patterns of these effects are shown in Fig. 3. Fig.
3A shows the binding affinity of R668A (left) and
Q849N (right) mutants with 33/20 + ddC template-primer. The
concentrations of R668A and Q849N ranged between 4.8 and 153.6 nM. It is clear that the enzyme-TP complex is detectable at
high enzyme concentrations and at the migration position of the
supershifted species. The two lanes (lanes 8 and
9 in the left panel) show the mobilities of the
enzyme-TP complex formed between wild-type KF and 33/20 + ddC. In
lane 8, the concentration of wild-type KF (4.8 nM) has been chosen such that it shows both (shift and
supershift) positions while in lane 9, the concentration is
higher (75 nM), which results in the formation of only
supershifted species. Similarly, lanes 8-10 in the
right panel show the mobility of wild-type enzyme-TP
complexes formed with 3.6, 7.2, and 25 nM enzyme and 33/20 + ddC. A transition from shift and supershift position is clearly
visible here with this template-primer. The binding of R668A and Q849N
mutant KF to 21/12 + ddC template-primer is shown in panel
B of Fig. 3. It is clear from the results that both R668A and
Q849N mutant enzymes fail to form a distinct species resembling the
enzyme-TP complex, although the amount of free TP is continually
decreasing as the concentration of the protein is increased. A shade of
gray above the free template-primer for R668A mutant enzyme (possibly a
dissociated complex) is nonetheless seen at higher concentrations of
the enzyme. This shade of gray, which is not so obvious for the Q849N
mutant enzyme, suggests that there are further differences among
different mutants with regards to the binding of the shorter duplex
containing template-primer. Panel C of Fig. 4 shows curves representing the binding affinity of R668A for 33/20 + ddC and 21/12 + ddC template-primers and Q849N for 33/20 + ddC. The
KD.DNA values of the R668A mutant for
two template-primers (33/20 + ddC and 21/12 + ddC), estimated from
these plots, are 14 and 33 nM, respectively. The
KD.DNA of the Q849N mutant enzyme with 33/20 + ddC template-primer is 31 nM. The affinity of Q849N
for 21/12 + ddC template-primer appears very low and could not be calculated from these data. These results strongly suggest that there
are two binding modes of KF to template-primer and that the mutants of
H-bonding track residues bind only in the supershifted mode.
View larger version (63K):
[in this window]
[in a new window]
|
Fig. 3.
The template-primer binding affinity and its
mode of binding with H-bonding track mutant enzymes. The affinity
of H-bonding track mutant enzymes (represented by R668A and Q849N) for
template-primer was determined in a manner similar to that used for
wild-type enzyme (Fig. 2). Whereas the template-primer concentration
was maintained at ~100 pM, much higher quantities of
mutant enzymes were required, because of their decreased and variable
affinity for template-primer. The panels A and B
show the binding of 33/20 + ddC and 21/12 + ddC template-primers,
respectively, to the indicated enzymes. The concentrations of both
R668A and Q849N mutant enzymes varied between 4.6 and 153.6 nM. The quantitation of enzyme-TP complex formation with
the mutant enzymes was carried out based on the quantity of the
uncomplexed template-primer, because no clear-cut band migrating at the
enzyme-TP position can be detected. For example, in panel A,
the binding of R668A to 33/20 + ddC shows a progressive decrease in the
free template-primer with increasing concentration of R668A. However,
in panel B, with the same mutant enzyme and 21/12 + ddC
template-primer, a progressive decrease in free template-primer is
accompanied by a "smudge" above the free template-primer position.
We have considered the smudge as a part of enzyme-TP complex, although
the strength and stability of this binary complex is certainly
compromised. In all cases, the quantitation of the area corresponding
to free TP position has been used to draw the curves shown
in panel C. Because of the extremely low affinity of Q849N
for 21/12 + ddC template-primer, the binding curve could not be
obtained. The KD.DNA was estimated by
fitting the data to a hyperbolic function by nonlinear regression. To
demonstrate that the mutants bind the long duplex containing
template-primer only in the supershifted mode, two wild-type enzyme
concentrations (lanes 8 and 9) were included
along with R668A. Three concentrations of wild-type (lanes
8-10) enzyme were included to show the same phenomenon with
Q849N. The concentrations of wild-type corresponding to lanes
8 and 9 of the left two figures was 4.8 and
75 nM, respectively. The concentrations of wild-type enzyme
for lanes 8-10 was 3.2, 7.2, and 25 nM,
respectively. The comparison of the migration of enzyme-TP complexes
with mutant enzymes and wild-type shows that the mutant enzymes bind
33/20 + ddC only in supershifted form. This observation was noted for
all mutants of the H-bonding track.
|
|
View larger version (107K):
[in this window]
[in a new window]
|
Fig. 4.
The effect of correct dNTP and three
incorrect dNTPs on the binding affinity of wild-type and mutant
enzymes. This figure shows the effects of incoming complementary
dNTP (dGTP for both template-primers) and three incorrect dNTPs (dATP,
dCTP, and dTTP) on the affinity of wild-type (panels A and
B), R668A (panels C and D), and Q849N
(panels E and F) for two template-primers.
Panels A, C, and E represent the
migration of the complexes of 33/20 + ddC with wild-type, R668A, and
Q849N mutant enzymes, respectively. Similarly, panels B,
D, and F represent the binding characteristics of
21/12 + ddC with wild-type, R668A, and Q849N mutant enzymes,
respectively. The effect of dNTPs on template-primer binding by the
wild-type and mutant enzymes was assessed using preformed enzyme-TP
binary complexes. To one group of the enzyme-TP complexes, 200 µM dGTP was added, whereas the duplicate received three
incorrect dNTPs (100 µM each). The concentration of the
wild-type (0.3-10 nM in panel A and 0.04-1.24
nM in panel B) and mutant enzymes (4.8-153.6
nM) was the same for lanes marked TP only,
TP + dGTP, and TP + 3dNTP-dGTP. The comparison of lanes marked TP
only with TP + dGTP and/or TP + 3dNTP  dGTP in panels A and B shows that the
presence of correct incoming substrate enhances the affinity of
template-primer to wild-type enzyme for both template-primers. In
contrast, the presence of 3 noncomplementary dNTPs inhibits the
template-primer binding to the wild-type enzyme. It is interesting to
note that such effects are not seen for the mutant enzymes, suggesting
a difference in the template-primer binding characteristics of mutant
enzymes, compared with wild-type enzyme. In panels C-F, the
lanes marked WT represent the wild-type enzyme, which was
included as control.
|
|
Effect of Correct and Incorrect dNTPs on E-dNTP Binary
Complex--
The other means that we employed to investigate the
differences in the binding modes of WT and mutants of H-bonding track residues was the determination of the effect of correct (complementary) and incorrect (noncomplementary) nucleotides on the binding affinity of
the wild-type and mutant enzymes for template-primer. It has been shown
before that the presence of correct incoming nucleotide enhances the
ability of wild-type KF to bind the template-primer (19, 25, 26). In
contrast, the presence of incorrect incoming nucleotide significantly
decreases the binding affinity of wild-type KF for template-primer (25,
26). We utilized this property to find out if the mutant enzymes
interacted differently with template-primer in the presence of correct
and incorrect dNTPs. The effect of the correct dNTP (dGTP for both
33/20 + ddC and 21/12 + ddC template-primers) and incorrect (3 other
dNTPs, viz. dATP, dTTP, and dCTP) on the binding affinity of
wild-type KF for two templates was assessed under the conditions
similar to those used for KD.DNA
determination. Briefly, the enzyme-TP complexes were allowed to form
for 10 min in a reaction mixture followed by the addition of correct
dNTP (dGTP, 200 µM) or a mixture of 3 incorrect ones
(dATP, dCTP, and dTTP, 100 µM each). The results of this
experiment with the wild-type, R668A, and Q849N mutant enzymes are
shown in Fig. 5. In panels A
and B, the effect of dNTPs on the binding characteristic of
wild-type KF with 33/20 + ddC (panel A) and 21/12 + ddC
(panel B) are shown. The comparison of the gel patterns
corresponding to "TP only" with "TP + dGTP" and "TP + 3dNTPs dGTP" clearly indicates that in the presence of
correct incoming nucleotide (dGTP) the binding affinity of wild-type
enzyme is increased. The KD.DNA of KF
for both 21/12 + ddC and 33/20 + ddC template-primer is increased by
~3-4-fold. In contrast, there is no significant change in the
binding affinity of either R668A or Q849N for both template-primers in
the presence of substrate or nonsubstrate dNTPs (panels
C-F). In the presence of correct incoming dNTP, the
KD.DNA of the R668A mutant was increased
only from 14 to 21 nM, whereas for Q849N, it was decreased
from 31 to 25 nM. In the presence of three incorrect dNTPs,
there was a small increase in KD.DNA of
R668A (from 14 to 24 nM), whereas Q849N showed little
change in KD.DNA (31 to 27 nM) under these conditions. Similar results were obtained with all the mutants of the H-bonding track residues (data not shown).
It is therefore clear that the presence or absence of dNTP minimally
affects the DNA binding properties of H-bonding mutants of the Klenow
fragment suggesting that the template-primer binding seen at the
supershift position may not be related to the polymerase mode of
binding but may represent an altered mode of binding with significant
decrease in the binding affinity.
View larger version (59K):
[in this window]
[in a new window]
|
Fig. 5.
Ability of wild-type KF and H-bonding track
mutants to form a stable ternary complex. The ability of the
wild-type and three mutant enzymes, H881A, N845A, and Q677A, to form
stable ternary complexes in the presence of cognate incoming dNTP was
assessed following the protocol described under "Experimental
Procedures." The sensitivity of radiolabeled template-primer (33/20 + ddC) in the desired enzyme-TP complexes to the addition of 500-fold
excess unlabeled DNA, (which completely dissociates the radiolabeled
DNA) provided an indication of ternary complex formation. The
radiolabeled template-primer in the complex formed by the wild-type
enzyme in the presence of cognate nucleotide (lane labeled
Enzyme + TP + dNTP) is not competed
out by addition of excess unlabeled DNA (shown in lane
Enzyme + TP + dNTP + Trap).
In contrast, except for the H881A enzyme, which forms ~5% stable
ternary complex compared with the wild-type KF, all other mutant
enzymes failed to form stable ternary complexes. Note that the addition
of excess nonradiolabeled template-primer readily competes out the
bound template-primer in the binary as well as ternary complexes for
all mutant enzymes (lanes marked enzyme + TP + Trap). The concentration of wild-type, H881A, N845A, and
Q677A enzymes was 1.2, 18.3, 150, and 75 nM,
respectively.
|
|
Assessment of the Stable Ternary Complex Formation by the Wild-type
and Mutant Enzymes--
To assess the ability of mutant enzymes to
form a stable ternary complex (i.e. enzyme-TP-dNTP complex),
we used the stable ternary complex formation assay initially developed
by Tong et al. (20). This assay has been successfully used
to demonstrate the ability or inability of mutant enzymes to form a
stable ternary complex, as judged by the effect of the trap on the
bound template-primer (15, 19). The results of this investigation for
the wild-type and three mutant enzymes are shown in Fig. 5. The lane
marked input represents the mobility of the template-primer
alone. The next four consecutive lanes marked "enzyme + TP,"
"enzyme + TP + trap," "enzyme + TP + dNTP," and
"enzyme + TP + dNTP + trap" show: (i) the binding of
template-primer alone, (ii) the effect of the addition of ~500-fold
nonradioactive template-primer (trap), (iii) the putative ternary
complex in the presence of complementary incoming dNTP, and (iv) the
susceptibility of the ternary complex to trap, respectively. It is
clear from the figure that enzyme-TP complex between the wild-type (or
mutant species) is readily competed out by the addition of the trap
template-primer (see lanes marked Enzyme + TP + Trap and
compare with Enzyme + TP lanes). However, the
complex between wild-type KF and template-primer + dNTP could not be
competed out by the same (trap) template-primer (see lane marked
Enzyme + TP + dNTP + Trap
for WT enzyme). For the mutant enzymes, the enzyme + TP complex in the
presence of dNTP was competed out by the nonradioactive template-primer
trap suggesting that the addition of dNTP to the enzyme-TP complex
failed to form a stable ternary complex. A small amount (~5% of
wild-type) of stable ternary complex was noted for H881A. This is
consistent with some catalytic activity seen with His881
mutants of KF.
UV-mediated Cross-linking of Enzyme-TP and the Addition of
Complementary dNTP in Situ by E-TP Complex--
The data presented
above suggest that the mutant enzymes of Asn845,
Gln849, Arg668, His881, and
Gln677 are generally defective in template-primer binding
and that the little binding that occurs is probably in the
nonpolymerase mode. To further assess if the small extent of binding of
template-primer to various mutant enzymes is in the polymerase mode, we
used UV-mediated photocross-linking of enzyme to template-primer and
examined the ability of individual enzyme-TP complexes for their
ability to incorporate the first complementary nucleotide. It has been
shown previously that the Klenow fragment of E. coli DNA
polymerase I, covalently cross-linked to template-primer, is capable of
adding one nucleotide onto the photocross-linked template-primer (18, 22). When covalently cross-linked to the template-primer, translocation ability of the enzyme along the template-primer is compromised. Under these conditions, the nucleotidyl transfer reaction is restricted to the incorporation of the first incoming nucleotide, provided that
the 3'-OH of the primer is correctly positioned in the complex for an
in-line attack on the -phosphate of incoming dNTP. In general, a
larger quantity of mutant protein was required to obtain approximately
equal extent of cross-linking as that of wild-type KF. To determine the
efficiency of enzyme-template-primer cross-linking by mutants compared
with wild-type KF, the radioactive template-primers were used to
produce enzyme-TP cross-link adducts. The quantitation of these adducts
with varying concentration of individual mutant enzymes (data not
shown) provided the guidance for the generation of approximately equal
quantity of enzyme-TP complexes for all mutant enzymes. For the dNTP
addition reaction, all mutant enzymes as well as the wild-type KF were
covalently cross-linked to nonradioactive 32/19-mer template-primer and
addition of [-32P]dCTP (the first incoming substrate
dNTP for this template-primer) was assessed. The in situ
catalytic addition of dCTP was performed in 1 M NaCl to
ensure that the addition reaction would be restricted only to the
cross-linked enzyme-template-primer species. The results are shown in
Fig. 6. A comparison of the intensity of
radiolabeled band (representing the addition of radiolabeled
nucleotide) shows that none of the mutant enzymes were able to
incorporate dNTP onto the cross-linked template-primer to any
significant extent (Fig. 6). Compared with the wild-type enzyme, less
than 2% incorporation was seen with N845A, R668A, and H881A mutant
enzymes. No visible incorporation was noted for Q849A and Q677A
enzymes. These results strongly suggest that the 3'-OH of the
template-primer cross-linked or bound to mutant proteins was not
properly placed at the active site for catalysis to occur.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 6.
In situ incorporation of dNTP onto
enzyme-TP covalent complexes of H-bonding track mutant enzymes.
The wild-type (30 pmol) and different amounts of Asn845,
Gln849, Arg668, His881, and
Gln677 mutant proteins were cross-linked with unlabeled
32/19-mer as described under "Experimental Procedures." The
irradiated mixture containing the E-TP covalent complex was further
incubated in a reaction mixture containing 1 M NaCl and 5 µCi of [-32P]dCTP (as incoming first nucleotide is
dCTP). The extent of product synthesis was determined by
SDS-polyacrylamide gel electrophoresis, followed by exposing the gel to
a PhosphorImager (Amersham Biosciences) and scanning. The Coomassie
Blue-stained wild-type and various mutant enzymes are designated as
Enzyme quantity, whereas the extent of incorporation of
[-32P]dCTP by the individual enzyme-TP species is
labeled as nucleotide addition.
|
|
| |
DISCUSSION |
The reaction mechanism of DNA polymerases has been well studied
(27-29). In addition, the crystal structures of several
template-primer and/or dNTP-bound DNA polymerases have provided
significant insight into the understanding of the DNA replication
mechanisms offered by the nucleic acid polymerases (4-7, 11, 24, 30,
31). The first step in the DNA replication kinetic pathway is the
binding of enzyme with TP. The comparison of the crystal structures of KlenTaq and B. stearothermophilus (Bst) DNA polymerases in
apo (Protein Data Bank files 1ktq and 1xwl) and DNA bound forms (Protein Data Bank files 4ktq and 2bdp), suggests that the thumb domain
of these polymerases undergoes a conformational change to accommodate
the double stranded region of the template-primer. As a result of this
conformational change, the thumb undergoes two changes: (i) a 12°
opening followed by 7° movement toward the cleft, and (ii) the
transition of J-helix to a coiled structure. In addition to these large
conformational changes, the binding of template-primer also induces new
intramolecular network(s) involving the side chains of several amino
acid residues in its vicinity. Most prominent among these interactions
is a cluster of five invariant amino acid residues interacting via
H-bonding in almost a straight line fashion. The involved residues are
the KF equivalents of Asn845, Gln849,
Arg668, His881, and Gln677 (Fig.
1). The distance between the C atoms of residues at either end of
the straight line (Gln677 at one end and Asn845
at the other) is ~18 Å and hence we have described it as an ~18-Å long H-bonding track. The binding of the substrate dNTP to enzyme-TP complex causes a major conformational change in the fingers domain of
pol I (5,11). The process of enzyme-TP-dNTP complex formation is
popularly referred to as "fingers closing" or "prepolymerase" ternary complex formation (32). Interestingly, the straight line
H-bonding network of interactions remains unaltered in the ternary
complex. To understand the function of this interaction network, we
employed site-directed mutagenesis to generate mutant derivatives of
all five residues and biochemically characterized individual mutant
enzymes for its polymerase activity, steady-state kinetic properties,
template-primer binding affinity, and the ability to incorporate dNTP
substrate onto the primer terminus in situ.
Architecture of the H-binding Track--
In the crystal structure
of KF and analogous DNA polymerases, the Arg668 equivalent
residue is at the center of the H-bonding track and is located on 
8.
This residue is a part of the highly conserved TGR motif found in the
polymerase I family of enzymes (33). In the template-primer bound
crystal structures of KlenTaq, the Arg668-equivalent
residue (Arg573) is flanked by Gln754
(Gln849 in KF), Gln849 and His786
(His881 in KF). The two distal residues are the equivalents
of Asn845 and Gln677 (Fig. 1). In the binary
complex of KlenTaq, the amino acid residue equivalent to
Arg668 of KF forms a hydrogen bond with Gln849
(3.44 Å), which in turn forms a hydrogen bond with Asn845
(3.69 Å). On the other side, Arg668 has van der Waals
interactions with His881. The distance between two of the
nearest atoms of Arg668 (N
) and His881
(C
2) is 3.38 Å. There also exists a possibility of a H-bond between
atom N
2 of Arg668 and N1 of His881.
However, the distance between the two H-bond forming atoms (4.16 Å) is
slightly unfavorable. Furthermore, His881 forms a hydrogen
bond with Gln677 (3.33 Å). There are also two
intermolecular hydrogen bonds mediated by Arg668 in the
enzyme-TP binary complex (11). One is with the primer base (2.39 Å)
and the other is with template nucleotide (3.8 Å) base. Similarly,
Gln849 and Asn845 form hydrogen bonds with
template strand. Gln849 hydrogen bonds with the base moiety
of the template nucleotide (3.15 Å) paired with the base at the primer
terminus and Asn845 forms a hydrogen bond with the sugar
oxygen of the same template base. Amino acid residue Gln677
is positioned between the 3rd and 4th base pairs from the primer terminus and has potential to form a hydrogen bond with the base moiety
of 3rd primer base in the double strand region of the template-primer. In the binary complex crystal structures of KlenTaq (11) and Bst
polymerase (24), Gln677 equivalent residues do not form a
hydrogen bond through their side chains. However, the possibility of a
water-mediated hydrogen bond with the base moieties of the 3rd and 4th
base pairs in the double stranded region from the primer end cannot be
ruled out as the side chain of Gln677 equivalent residues
is positioned within the minor groove of the template-primer. A
possibility of hydrogen bond between Gln677 and base moiety
of the primer strand can also be extrapolated from the ternary complex
crystal structure of T7 DNA polymerase. Gln439 in the
J-helix region of T7 DNA polymerase, which can be considered equivalent
to Gln677 of KF, forms a hydrogen bond with base moiety of
the 3rd nucleotide from the primer end (5).
From the detailed analysis of the properties of mutant species of
various residues, it is clear that the substitution of the H-bonding
side chain in any of the five residues disrupts the track; however, the
degree of disturbance is probably different, because the catalytic
activity of the individual mutant protein shows variation (Table I). We
shall first discuss the participation of individual members and its
mutant derivatives deduced from the structural and the experimental results.
Properties and Environment of Asn845 and
Gln849--
In KF, the sequence
845NAPMQG850 contains two highly conserved
uncharged polar residues common to all members of the polymerase I
family of enzymes. These are Asn845 and Gln849.
Previous studies have suggested that both Asn845 and
Gln849 play some role in catalysis of the polymerase
reaction (13, 14). Our kinetic characterization shows that the
Km.dNTP for N845A and N845Q did not
change significantly. However, a significant change was noted in
kcat values for the two Asn845
mutant enzyme species. It is rather curious that the N845Q mutant enzyme was significantly more affected compared with N845A. The loss of
catalytic activity of the Gln849 mutant derivatives (Q849A
and Q849N) was more severe suggesting that the side chain of
Gln849 participates directly or indirectly in the catalytic
process. Previously, Polesky et al. (13) have suggested that
Gln849 interacts with the DNA primer terminus and has some
role in dNTP turnover, because Q849A and Q849E exhibited deficiency in
the turnover of -thio-dNTPs. The latter observation also indicates function of Gln849 at the chemical step of the reaction.
However, the crystal structures of KlenTaq (Klenow fragment of T. aquaticus DNA polymerase I) in three different forms (binary, open
ternary, and closed ternary complexes) (11) as well as the crystal
structure of T7 DNA polymerase (5) show that the Gln849
equivalent glutamine interacts with template nucleotides and it does
not have any interaction with dNTP. Therefore, the suggested participation of Gln849 in the chemical step (13) is most
likely a consequence of a loss in the template-primer binding affinity
and/or altered position of 3' OH relative to the active site.
Properties and Environment of Arg668--
Extensive
biochemical characterization of R668A has been reported previously by
Polesky et al. (13). The results showed a moderate change in
Km.dNTP (~2-fold). Significant changes in KD.DNA (~20-fold) and
kcat (~400-fold) were noted for the R668A
mutant enzyme. Our estimations of
Km.dNTP, KD.DNA, and kcat
for the same mutant enzyme show similar changes. In fact, the
KD.DNA was increased by ~25- and ~160-fold with 33/20 + ddC and 21/12 + ddC, respectively. The kcat was reduced by ~350-fold with
dA36/dT18 template-primer. These data suggest
that the defect in the catalytic ability of the R668A mutant was
predominately because of the defect in the binding of the
template-primer. The interaction of Arg668 seen in the
crystal structure complexes with template-primer supports the decreased
affinity of the R668A mutant enzyme. The side chain position of
Arg668 in ternary complex crystal structures is favorable
for a hydrogen bond with the sugar moiety of dNTP. However, the major
participants in the binding of dNTP to enzyme-TP complex appear to be
the members of the O-helix and metal ligated
Asp705 and Asp882. Therefore, the weak hydrogen
bond between Arg668 and the sugar oxygen of dNTP may not be
contributing significantly to binding and stabilization of dNTP at the
active site.
The role of an amino acid residue at the chemical step of nucleotide
incorporation by DNA polymerases has been judged by the difference in
utilization of normal versus -thio-substituted dNTPs
(32). This is called an elemental effect. R668A mutant protein has been
shown to utilize dNTPS about 10-fold less efficiently than the
wild-type enzyme, suggesting the involvement of Arg668 at
the chemical step. Arg668 has been seen to form the ion
pairs with Glu710, which has been proposed to ligate a
metal ion coordinating with the 3'-OH of dNTP (34). Thus the negative
effect of R668A on the chemical step may be because of the loss of its
reactivity with Glu710, and the elemental effect seen for
the R668A mutant may be the reflection of some indirect effect mediated
through Glu710.
Properties and Environment of His881--
The
biochemical characterization of His881 shows that the H881A
mutant is ~30% as active as WT. There is no significant change in
Km.dNTP, whereas a 6-66-fold decrease
in kcat with different template-primers was
noted, suggesting the requirement of His881 for optimal
catalysis by KF. The location of His881 is on a -turn
between 12 and 13 in the palm subdomain. It is highly conserved
among pol I type DNA polymerases. In the crystal structure of T7 and
TaqDNA polymerases, the equivalent residues interact with
the sugar moiety of the primer terminus nucleotide (5, 11). Therefore,
this residue has been proposed to position the primer terminus for
catalytic activity (5). Comparison of the apo and binary complex of
TaqDNA polymerase crystal structures shows that the turn
between 12 and 13 undergoes a "springboard-like" conformational change upon DNA binding, which does not change further
upon ternary complex formation. The C distance between two
His881 conformations (in apo-KlenTaq and template-bound
KlenTaq) is ~3 Å. Superposition of apo onto the binary complex shows
that if such a sharp springboard-like conformational change of the turn
between 12 and 13 were not to occur, His881 would
interfere with the sugar moiety of the primer terminus nucleotide.
Analyses of these crystal structures suggest that two additional
residues probably contribute to induce a springboard-like conformational change. These residues are Gln677 and
Gln879. In the binary and/or ternary complex,
His881 side chain forms a hydrogen bond with
Gln677, whereas the main chain carbonyl forms a hydrogen
bond with Gln879. In addition, Gln879 forms a
hydrogen bond with Gln677. It appears that the triangular
H-bond arrangement at this site may be responsible for the
conformational change in the -turn, which can then accommodate the
template-primer terminus at a proper position.
Properties and Environment of
Gln677--
Gln677 belongs to the J-helix in
pol I-type DNA polymerases. This helix is formed by five amino acid
residues flanked by proline on either ends. The sequence of this region
is highly conserved. The J-helix undergoes a conformational change
(from helix to coil) upon TP binding (5, 11). However, the helical
structure is maintained when the binding of the template-primer occurs
in the 3'-5' exonuclease mode (9). We have previously shown that helix to coil transition (or vice versa) of the J-helix plays a crucial role
in the switching of the template-primer from the polymerase to the
exonuclease site, as judged from the properties of P680G (35). Upon the
binding of template-primer to the enzyme in the polymerase mode,
repositioning of the side chain of Gln677 has also been
noted. This in turn allows Gln677 to make contacts with
His881 and Gln879. The mutant Q677A enzyme is
an inactive enzyme (35, 36). Our activity and steady-state kinetic data
also show severe loss in the polymerase activity of both nonhomologous
Q677A and homologous Q677N mutant enzymes (Table II). In fact, the
polymerase activity of the P680G mutant of KF was also reduced to the
same extent as that seen for Gln677 mutant enzymes (35).
Therefore, it is possible that the reduced activity of P680G was a
reflection of the activity of the enzyme because of the altered
position of Gln677. A change in the position of
Gln677 would result in the destabilization of the H-bonding
track arrangement, which in turn would not permit the binding of enzyme
to template-primer. Whereas the polymerase activity of the P680G mutant
enzyme was severely reduced, the mutant enzyme was extremely efficient
in the 3'-5' proofreading activity (35). We have noted a similar increase in the efficiency of 3'-5' exonuclease activity with the Q677A
mutant species.2
Thus, the comparison between the activity and steady-state kinetic
constants of the wild-type and mutant enzymes (Table II) suggests that
all of the H-bonding track residues participate in the catalytic
function of Klenow fragment of E. coli DNA polymerase I. The
major participation of H-bonding track residues is in the process of
the template-primer binding, as judged by the significant decrease in
KD.DNA (Table II). These observations
are consistent with the crystal structure data of the binary and
ternary complexes of KlenTaq, which show that these residues interact with the template-primer (11). The
KD.DNA for some of the mutant enzymes
studied here was also determined by others (13). They noted that the
N845A mutant showed no change in template-primer binding affinity
compared with the wild-type KF. However, the other mutant of
Asn845 (N845D) had a 15-fold decrease in template-primer
binding affinity. The KD.DNA for H881A
was altered by only 2.5-fold, whereas this value for Q849A and R668A
mutants was reported to be ~10- and 17-fold greater, respectively
(13). The change in KD.DNA for N845A,
Q849A, R668A, H881A, and Q677A with 33/20 + ddC, in our study, is 16-, 43-, 22-, 25-, and 70-fold, respectively. Thus, the
KD.DNA values determined here are
significantly different from those reported by Polesky et
al. (13). The differences in KD.DNA
values may be attributed to two different methods and the
template-primers used in two studies. In fact, we have noted that
KD.DNA values vary significantly, based
upon the method of determination as well as the concentration of
template-primer (19).
The structural data on enzyme-DNA binary complexes of KlenTaq and
B. stearothermophilus Klenow fragment and the ternary
complexes of KlenTaq and T7 DNA polymerase I show that 8-10 base pairs
of the double stranded DNA are occluded by these enzyme proteins. The
template-primers in these crystal structure complexes have between 30 and 45 interactions. A majority of these interactions are in
TOP
ABSTRACT
INTRODUCTION
