Effect of A and B Metal Ion Site Occupancy on Conformational Changes in an RB69 DNA Polymerase Ternary Complex

Abstract

Rapid chemical quench assays, as well as equilibrium and stopped-flow fluorescence experiments, were performed with an RB69 DNA polymerase (RB69 pol)–primer-template (P/T) complex containing 2-aminopurine (dAP) and a metal exchange-inert Rh(III) derivative of a deoxynucleoside triphosphate (Rh • dTTP). The objective was to determine the effect of catalytic metal ion (A site) occupancy on the affinity of an incoming Rh • dTTP for the RB69 pol–P/T binary complex and on the rate of the conformational change induced by Rh • dTTP binding. With Ca²⁺ in the A site, the affinity of the incoming Rh • dTTP for the RB69 pol–P/T binary complex and the conformational change rate can be determined in the absence of chemistry. When Mg²⁺ was added to a ternary complex containing Rh • dTTP opposite dAP, the templating base, nucleotidyl transfer occurred, but the rate of product formation was only one-tenth of that found with Mg • dTTP, as determined by rapid chemical quench assays. Rates of conformational change subsequent to formation of a ternary complex, in the absence of chemistry, were estimated from the rate of change in dAP fluorescence with an increase in the Rh • dTTP concentration. We have shown that there is an initial rapid quenching of dAP fluorescence followed by a second phase of dAP quenching, which has nearly the same rate as that of dTMP incorporation, as estimated from rapid chemical quench experiments. We have also demonstrated that the affinity of Rh • dTTP for occupancy of the B metal ion site is dependent on the presence of Ca²⁺. However, a saturating Rh • dTTP concentration in the absence of Ca²⁺ results in full quenching of dAP fluorescence, whereas a saturating Ca²⁺ concentration in the absence of Rh • dTTP gives only partial quenching of dAP fluorescence. The implications of these results for the mechanism of Fingers closing, metal ion binding, and base selectivity are discussed.

A key feature of replicative DNA polymerases is the high level of dNTP¹ discrimination exhibited during DNA synthesis. Reported error frequencies during replication are typically less than 10⁻⁶ due mainly to base selection during nucleotide incorporation (1–5). Several features of the enzymes and substrates contribute to base selectivity, including base-pair hydrogen bonding, base stacking, active site tightness, and base-pair geometry (6–10). These features manifest themselves kinetically as changes in the rate of prechemistry conformational changes and changes in the rate of chemistry (11–13). In addition, there is an editing function associated with many DNA polymerases that can excise incorrect nucleotide residues from the nascent primer strand, contributing further to fidelity (9, 14–18).

DNA polymerases catalyze a nucleotidyl transfer reaction that results in primer extension via a mechanism requiring two divalent metal ions (19). The metal ions occupy spatially distinct sites, A and B, although they share carboxylate ligands from two aspartyl side chains. The A site has been designated as the catalytic metal ion site because reactions occur only when this site is occupied by Mg²⁺ or Mn²⁺ ions. The presence of either one of these metal ions in the A site lowers the pK_a of the 3′-terminal OH group of the primer, generating an oxyanion that can attack the α-phosphorus atom of the incoming dNTP (20). The metal ion occupying the B site enters with the incoming dNTP and is ligated to the negatively charged oxygen atoms in its triphosphate tail. The function of the B metal ion is to stabilize the developing negative charge in the pentavalent transition state and to assist in the departure of the pyrophosphate as the nascent phosphodiester is formed (20).

We have chosen RB69 pol as a model B family DNA polymerase for our investigation into the mechanisms used by DNA polymerases to achieve high-fidelity DNA synthesis for the following reasons. (i) Crystal structures of RB69 pol–P/T binary and ternary complexes have been determined at high resolution (21–25). (ii) Extensive kinetic studies have been carried out on RB69 pol and T4 pol, its close relative, as well as on many single-site mutants of these enzymes that have helped to delineate the functional role of various conserved residues (26–29). (iii) RB69 pol has extensive sequence similarities in the highly conserved regions that characterize replicative DNA polymerases belonging to the B family, suggesting that the results with RB69 pol could be applicable to other eukaryotic B family members, including human α and δ DNA polymerases (30–32). A diagram showing the location of the A and B metal ion sites as observed in the crystal structure of the RB69 pol ternary complex is shown in Figure 1. Although two divalent metal ions are essential for the nucleotidyl transfer reaction, it is not known whether one or both of these metal ions are required for the conformational change, nor is it clear how the occupancy of either the A or B site, alone, affects either the affinity of incoming dNTP for the E • D binary complex or the rate of the conformational change that precedes chemistry.

Figure 1.

RB69 pol active site with Ca²⁺ occupying the A metal site and Rh³⁺ occupying the B metal site. This figure was adapted from Figure 2A of ref (19).

The key to understanding the individual contributions of the two metal binding sites to isomerization is to separately analyze each site. Bahktina and Tsai (11) recently reported that rhodium(III) can be used as an exchange-inert metal ion to form a complex with a dNTP without largely compromising the catalytic efficiency of human polymerase β, effectively allowing the investigators to analyze the B site binding (Rh • dNTP) in a manner independent of A site binding (catalytic Mg²⁺). The conclusion was that B site binding of Rh • dNTP was sufficient for isomerization to a closed ternary complex (i.e., closing of the Fingers domain to form a preactive complex), but that addition of Mg²⁺ to occupy the A site was necessary for catalysis.

Their analysis however was incomplete, given the necessity to analyze the effect on the isomerization steps upon binding of Rh • dNTP with the catalytic A site already occupied. The use of Mg²⁺ to saturate the A site would be problematic since subsequent addition of Rh • dNTP leads to phosphodiester bond formation, complicating the interpretation of the kinetic data. Preventing catalysis has been achieved in several ways: the most frequently used approach has been the removal of the 3′-OH group of the primer, but the resulting space within the active site may not provide an accurate representation of the prechemistry steps due to the absence of this functionality that coordinates with the catalytic metal ion and water molecules that keep the reactive centers optimized for catalysis. Because so many crystal structures of DNA polymerase ternary complexes were obtained with dideoxy-terminated primers, we wanted to compare the results of kinetic studies using ddP/Ts versus dP/Ts to see whether the presence or absence of the primer's 3′-OH group affected the rate of the prechemistry conformational change.

Ca²⁺ has been shown to stabilize a ternary complex (i.e., by increasing the binding affinity of the incoming dNTP) without supporting catalysis and may be better suited as a mimic with the 3′-OH group present. However, a recent report by Egli and Guengerich (33) claimed that DNA pol IV from Sulfolobus solfataricus could utilize Ca²⁺ for phosphodiester bond formation, albeit at a much slower rate than Mg²⁺, but it was unclear if they had removed all of the Mg²⁺ from their assays.

In this report, we used catalytically inert ternary complexes (with Ca²⁺ as one of the metal ions, 3′-OH, a dideoxy-terminated primer, and Rh • dTTP) to analyze, in all combinations (i.e., through occupancy of the A site first, B site second, and vice versa), the contributions of each metal binding site to the formation of the final isomerized RB69 pol–P/T–Rh • dTTP complex using rapid chemical quench and stopped-flow fluorescence techniques. Our results showed that isomerization to a ternary complex (“closing” of the Fingers domain) occurred without occupancy of the A site. In the following paper (38), we show that reverse isomerization (Fingers “re-opening”) in the presence of Rh • dTTP is required before the catalytic metal ion can enter the empty A site. The importance of these results in understanding the kinetic mechanism of catalysis and base selectivity by this DNA polymerase is discussed.

MATERIALS AND METHODS

Escherichia coli strain BL21(DE3) was obtained from Stratagene Corp. DH5α cells and dNTPs were purchased from Invitrogen. Sequenase was from United States Biochemical Company. [γ-³²P]ATP was from Perkin-Elmer Life Science Inc. Ni-NTA resin was obtained from Qiagen. Rhodium chloride hydrate was purchased from Sigma-Aldrich, Inc. AG1-X2-100 ion exchange resin was from ACROS Organics. Biogel P2 was ordered from Bio-Rad Laboratories. Other chemicals were analytical grade obtained from Sigma-Aldrich, Inc. RB69 DNA pol cDNA was a gift from J. Karam (Tulane University, New Orleans, LA).

Oligonucleotides were purchased from the W. M. Keck Foundation Biotechnology Resource Laboratory (Yale University). The sequences of the primer-templates (P/T) are shown in Figure 2.

Figure 2.

Sequences of the primer-templates used in this paper. C^d represents dideoxycytosine; P represents 2-aminopurine as the templating base.

Expression and Purification of Exo⁻ RB69 pol

Site-directed mutagenesis, as previously reported, was used to substitute alanine for aspartic acid at positions 222 and 327 to create the exonuclease-deficient RB69 DNA pol that was used in all of our experiments, and that we term RB69 pol throughout the text (34). We subcloned the cDNA for the exo⁻ RB69 pol into the pET21 vector (Invitrogen) so that the expressed protein would have six histidine residues appended to its C-terminus to facilitate purification which was carried out as previously described (29).

Preparation, Characterization, and Stability of the Rh•dTTP Complex

Rh• dTTP was prepared according to the procedure described for the synthesis of Rh • dCTP (11). To separate Rh • dTTP from the starting material and side products, the reaction mixture was loaded onto a 1.6 cm × 20 cm column of AG1-X2 in the chloride form. Prior to loading, the resin was washed with distilled, deionized (nanopure) water ([Mg²⁺] < 0.01 nM) that had been adjusted to a pH of 3 with HCl. After loading, elution was carried out by first passing nanopure water (pH 3) through the column until the eluate had an A₂₆₀ of <0.05 optical density unit at 260 nm. The column was then eluted with 100 mM HCl. Two well-separated components with A₂₆₀ absorption were obtained. The first component was a rhodium nucleotide complex that was inert as a substrate in primer extension reactions and was not characterized further. It was probably the dTTP equivalent of the dCTP derivative described by Dunlap and Tsai (35). The second component contained the desired Rh• dTTP. This material was concentrated by rotary evaporation, adjusted to pH 5 with NaOH, then loaded on a Biogel P2 gel filtration column, and eluted with a 2 mM sodium acetate solution (pH 5) to remove NaCl.

Rh• dTTP obtained using this procedure was characterized by ³¹P NMR using a NOVA-B400b NMR spectrometer. Spectra were recorded using a Rh • dTTP concentration of 20 mM in 2 mM sodium acetate buffer (pH 5) at 60 °C. The spectra showed the doublets expected for α- and γ-phosphates at −8.2 and −10.1 ppm with 20 Hz splitting for each peak (Figure 3A). The low-intensity peak at −7.8 ppm probably represents the β-phosphate. The peaks at 0.94 and 1.5 ppm are likely due to the complexation of dTTP with Rh³⁺. The spectra of Na • dTTP were recorded under the same conditions and concentrations as those of Rh• dTTP (Figure 3B). The spectrum has a triplet at −21.5 ppm representing the β-phosphate which is absent in the Rh•dTTP sample. There are also two doublets at −6 and −11 ppm with a 20 Hz split. These doublets represent the α- and γ-phosphates. The NMR results show that our Rh • dTTP sample was essentially free of Na • dTTP and had the expected chemical shift for the α- and γ-phosphates. The stability of Rh • dTTP was a concern even though it has been reported that Rh • dNTPs are stable up to pH 7 (11), which we have confirmed by ³¹P NMR; we did not know how Rh • dTTP would behave when it was incubated with RB69 pol–P/T complexes under conditions that were more stringent (5 h at pH 7 in MOPS buffer at 25 °C) than those used for our kinetic studies. Testing Rh • dTTP under these conditions showed that there was no exchange of rhodium with other cations. However, after incubation of Rh • dTTP with EDTA for 2 h at 25 °C, peaks corresponding to dTTP appeared in the ³¹P NMR spectra, indicating that EDTA competes with dTTP to sequester and chelate the rhodium cation. To test whether Rh • dTTP was stable under the conditions used for equilibrium and stopped-flow fluorescence experiments, we incubated a 20-fold molar excess of Rh • dTTP with an RB69 pol–dP/T binary complex and 2 mM Ca²⁺ for 16 h at 20 °C and then examined the ³¹P NMR spectra. There was no indication that the Rh • dTTP had been converted to Ca • dTTP.

Figure 3.

³¹P NMR spectra of Rh • dTTP (blue) and dTTP (black). (A) ³¹P NMR spectra of 20 mM Rh • dTTP in 2 mM acetate buffer (pH 5) at 60 °C. (B) ³¹P NMR spectra of 20 mM dTTP in 2 mM acetate buffer (pH 5) at 60 °C.

As a further test of purity, we subjected our Rh • dTTP preparation to thin layer chromatography. The Rh • dTTP sample was spotted on a thin layer, fluorescein-treated, polyethyleneimine plate which was then developed using a solution containing 0.5 M LiCl in 2 M acetic acid. This solution had been shown previously to separate mono-, di-, and trideoxynucleoside triphosphates (36). Only a single spot, which quenched the fluorescein fluorescence, appeared with a mobility that was somewhat greater than the dTTP control.

Incubation of this Rh • dTTP preparation with the RB69 pol–dP/T complex (dA as the templating base) for 30 min at 25 °C showed no evidence of dTMP incorporation. However, upon addition of 10 mM MgCl₂, the Rh • dTTP was utilized and the primer was extended by one nucleotide residue, showing that the Rh • dTTP preparation was free of Mg²⁺ and that it could serve as a substrate when Mg²⁺ was present.

Preparation of Synthetic Oligonucleotide Substrates

The primer oligonucleotide used in the rapid chemical quench experiments was 5′-labeled with ³²P with T4 polynucleotide kinase as per the manufacturer's instructions (New England Biolabs). Termination of the reaction was accomplished by heating the mixture to 95 °C for 5 min. We carried out primer-template annealing by mixing equimolar ratios of 13mer primer and 20mer template, heating the mixture to 95 °C, and allowing it to cool slowly to room temperature. The 20mer oligonucleotide templates used in the stopped-flow experiments, containing 2-aminopurine (dAP) at the templating position, were annealed to unlabeled 13mer primers as described above. The sequences of the primers and templates used are listed in Figure 2.

Chemical Quench Experiments

All rapid chemical quench reactions were preformed using a KinTek RQF-3 quench-flow instrument (KinTek Corp., Austin, TX). Single-turnover experiments were conducted at 24 °C by mixing equal volumes of a solution of 50 mM MOPS (pH 7), 10 mM MgCl₂, 180 nM 5′-³²P-labeled dP/T (dA or dAP at the templating position) (Figure 2), and 200 nM RB69 pol with a solution containing varying dTTP or Rh • dTTP concentrations in 50 mM MOPS buffer (pH 7) and 10 mM MgCl₂. The reaction was quenched with 0.5 M EDTA at different times. Separation of the ³²P-labeled substrate and product was achieved by gel electrophoresis (20% polyacrylamide/50% urea). The gel was visualized on a Storm 840 phosphoimager, and then substrate and product band intensities were quantified using Image QuaNT (Molecular Dynamics, Sunnyvale, CA).

Equilibrium Fluorescence Titrations

The emission spectra of deoxy- or dideoxy-P/Ts containing dAP were recorded at 24 °C with a Photon Technology International Alphascan scanning spectrofluorometer. The emission spectra of RB69 pol–P/T complexes were obtained in the same way. The reaction mixture contained 200 nM P/T (13/20mer) with dAP as the templating base (Figure 2), 1 μM RB69 pol, and either 10 mM MgCl₂, 2 mM CaCl₂, or no divalent cations. It also contained 50 mM MOPS buffer (pH 7.0) and varying dTTP concentrations. Samples were excited at 313 nm so that inner filter effects from nucleoside bases were negligible (37). Fluorescence emission spectra were collected from 330 to 460 nm. The intensities were corrected for the intrinsic fluorescence of RB69 pol and Rh • dTTP. No photobleaching could be detected after 10 min. Peak emission intensities at 365 nm were plotted as a function of dTTP or Rh • dTTP concentration and fit to hyperbolic or quadratic equations (see Data Analysis) to obtain the overall dissociation constants (K_dg^app) for dNTP binding.

Stopped-Flow Fluorescence Experiments

Experiments were performed using an Applied Photophysics (Leatherhead, U.K.) SX18MV-R stopped-flow apparatus thermostated at 24 °C. The excitation wavelength for dAP was 313 nm. A 345 nm long pass glass filter was used in monitoring of the dAP fluorescence emission. The final reaction mixtures consisted of 50 mM MOPS (pH 7.0), 400 nM P/T, 1 μM RB69 pol, varying dTTP or Rh • dTTP concentrations, and either 10 mM MgCl₂, 2 mM CaCl₂, or no divalent cation.

Data Analysis

Data from rapid chemical quench or stopped-flow experiments were fitted to single-exponential or double-exponential equations (Origin). The dNTP concentration dependence of k_obs was fit to the hyperbolic equation k_obs = k_max[dNTP]/(K_d + [dNTP]) + k_r, where k_obs is the observed rate of the reaction at different dNTP concentrations and k_r is the extrapolated intercept at the Y axis when the dNTP concentration is zero. In experiments where chemistry is prevented, k_r approximates the reverse isomerization rate. Data from the equilibrium fluorescence experiments were fit to either a hyperbolic or a quadratic equation as shown below:

$$\Delta Q=\frac{Q_m[E_0+S_0+K_d-\sqrt{(E_0+S_0+K_d{)}^2-4E_0{S}_0}]}{2E_0}$$

where ΔQ is the fluorescence signal change, Q_m is the maximum value for the fluorescence signal change at saturating substrate concentrations, E₀ is the total enzyme concentration, and S₀ is the total dNTP concentration. K_d is the ground-state dissociation constant for dissociation of dNTP from the RB69 pol–dP/T binary complex. The solution of these equations provided an estimate of the apparent dissociation constants of dNTP using Origin.

RESULTS

One of the goals of this study was to determine how occupancy of the A and/or B metal ion sites affected the affinity of Rh • dTTP for the RB69 pol–P/T binary complex and whether filling the A metal ion site influences the rate of Fingers closing after addition of a deoxynucleoside triphosphate. In addition, we wanted to determine whether the binary complex consisting of a dP/T or a ddP/T with RB69 pol would have comparable affinities for Rh • dTTP, and if both types of P/Ts would produce the same rates of conformational change in the presence or absence of sufficient Ca²⁺ to fill the A site. To monitor conformational changes subsequent to metal ion or nucleotide binding, we used 2AP as the templating base (dAP). For primer-templates (P/Ts), we used 13/20mers (Figure 2) that bound to RB69 pol with K_dg^app values in the range of 20 nM, approximately the same as those found previously for the T4 pol–13/20mer binary complexes (26).

Incorporation Efficiency of Rh • dTTP versus Templating dA or dAP

Prior to initiating equilibrium and stopped-flow fluorescence studies, we wanted to know how effective Rh• dTTP would be as a substrate for the nucleotidyl transfer reaction. For this purpose, we carried out rapid chemical quench experiments with four combinations of substrates and P/Ts that included Mg• dTTP or Rh • dTTP, as incoming dNTPs and dA or dAP as the templating base embedded in 13/20mer P/Ts. The k_pol and K_dq^app values for this set of experiments are listed in Table 1. As shown in Figure 4, while Rh • dTTP can serve as a substrate, its k_pol value (7 s⁻¹), when paired opposite templating dAP, was much lower than the k_pol value of 80 s⁻¹ when Mg • dTTP was paired opposite dAP. Also, the K_dq^app for Rh • dTTP incorporation (150 μM) was almost 3 times higher than the K_dq^app for Mg• dTTP (45 μM) when it was paired opposite dAP. The observation that the k_pol for Rh•dTTP was considerably lower than that for Mg• dTTP prompted us to examine the kinetic parameters for insertion of Mg• dTTP opposite dA and dAP. As shown in Table 1, the k_pol value for Mg• dTTP incorporation, when paired opposite dAP, was only reduced by ~40% relative to the k_pol for Mg• dTTP paired opposite dA under these conditions. These data provide a basis for interpreting the fluorescence results when Rh • dTTP was paired opposite dAP.

Table 1.

Re-Steady-State Kinetic Parameters for Incorporation of Different Metal Ion-dTTP Complexes Paired Opposite dA or dAP as the Templating Base

Me•dNTP:templating base	kpol(s−1)	Kdqapp (μM)
Mg•dTTP:dA	131 ± 8	27 ± 4
Mg•dTTP:dAP	80 ± 7	45 ± 16
Rh•dTTP:dA	76 ± 5	124 ± 40
Rh•dTTP:dAP	7 ± 0.6	150 ± 40

Figure 4.

Time course for the incorporation of Rh • dTTP opposite dAP catalyzed by RB69 pol. (A) The rate of product formation was measured at various Rh • dTTP concentrations, from 10 μM (black) to 1000 μM (dark yellow) from bottom to top. (B) The observed rate was plotted against Rh • dTTP concentration. The curve fit best to a hyperbolic equation with a k_pol of 7.1 ± 0.6 s⁻¹ and a K_dq^app of 150 ± 40 μM.

Equilibrium Fluorescence Experiments Using Rh • dTTP Paired Opposite a Templating dAP

When Rh • dTTP was added to the binary E• dP/T complex in the presence of a divalent cation (2 mM CaCl₂) (Figure 5) and the extent of dAP fluorescence quenching determined as a function of Rh • dTTP concentration, a K_dg^app for Rh • dTTP of 9.1 μM was obtained (Figure 6A). This experiment was repeated in the absence of divalent cations, and the K_dg^app for Rh • dTTP under these conditions was estimated to be 108 μM (Figure 6B). The reduction in K_dg^app for Rh • dTTP was attributed to Ca²⁺ occupancy of the A metal site which, we presume, stabilizes the closed ternary complex. In this state, the templating base was stacked in the dP/T duplex in the presence of the correct dNTP, as observed in the crystal structure of the RB69 pol–ddP/T–dNTP complex (21). The stacking of the templating base occurred even though the nascent incoming base was not covalently linked to the 3′ end of the primer (21). In the absence of Rh • dTTP, addition of Ca²⁺ results in partial quenching of dAP fluorescence in the binary RB69 pol–dP/T complex. The extent of dAP quenching depended on Ca²⁺ concentration, yielding a K_dg^app for Ca²⁺ of 152 μM (Figure 6C). The amplitude of the fluorescence change was only 8400 AU compared to a decrease of 37000 AU when a saturating concentration of Rh • dTTP was added to the binary complex in the absence of divalent cations. Whether this K_dg^app for Ca²⁺ actually represents the Ca²⁺ concentration required for half-occupancy of the A metal site only remains to be determined.

Figure 5.

Equilibrium fluorescence titration of the RB69 DNA pol–dP/T complex with an increasing Rh • dTTP concentration. Fluorescence emission spectra of the RB69 pol–dP/T–Rh • dTTP complex are shown as solid lines in the presence of 2 mM CaCl₂. The concentration of dP/T was 200 nM and that of RB69 pol 1 μM. The Rh • dTTP concentrations (from top to bottom) were 0, 2.2, 5.5, 9.7, and 17.8 μM. The dotted line is the fluorescence spectrum of dP/T alone.

Figure 6.

Equilibrium fluorescence titrations of 1 μM RB69 pol complexed with 200 nM dP/T with an increasing Rh•dTTP or CaCl₂ concentration. The change in fluorescence signal vs Rh • dTTP or CaCl₂ concentration fit a hyperbolic equation. For the dP/T sequence, see Figure 2A2. (A) Titration of Rh • dTTP with the A site filled with Ca²⁺ (2 mM) gave a K_d2,g^app of 9.1 ± 0.9 μM. (B) With the A site empty, titration of Rh • dTTP gave a K_d1,g^app of 108 ± 6.6 μM. (C) Titration of Ca²⁺ with the B site empty gave a K_d3,g^app of 152 ± 19 μM. The fluorescence change (ΔF) with an increasing Rh•dTTP concentration indicates the extent of fluorescence quenching for the ternary RB69 pol–dP/T–Rh•dTTP complex compared with that of the binary RB69 pol–dP/T complex.

Although it has been well established that two divalent metal ions are essential for the nucleotidyl transfer reaction with DNA polymerases, it was not known whether occupancy of either the A or B metal ion site, or both, is absolutely necessary for the open to closed transition when a correct dNTP is introduced. To address this question, Rh • dTTP was selected as the incoming dNTP because of its metal exchange-inert properties and because it was complementary to dAP. This allowed us to determine the effect of B site occupancy, with the nucleotide binding pocket also filled, on the rate of the conformational change in the presence and absence of Ca²⁺ in the A site. It also allowed us to determine the affinity of Rh • dTTP for the binary complex when the A site was empty. To fill the A site, Ca²⁺ was chosen because it is presumably ligated by the same set of aspartate carboxyl groups as Mg²⁺ and because the crystal structure of the RB69 pol ternary complex was determined in the presence of Ca²⁺ (21). The amplitude of the quench with a saturating Rh • dTTP concentration was greater than the amplitude found when Ca²⁺ only was added to the binary complex. This implies that filling the B site and the nucleotide binding pocket with Rh • dTTP, the correct base, is sufficient to induce a complete transition to the closed ternary complex but also that Ca²⁺ alone may cause partial closure of the Fingers domain (Figure 7). A summary of the relationships between the affinities of Rh•dTTP for the binary RB69 pol–P/T complex in the presence and absence of Ca²⁺ and the affinity of Ca²⁺ for the RB69 pol–dP/T complex in the absence of Rh • dTTP is shown in Figure 7, and the K_d values are given in Table 2 for dP/T and ddP/T. The affinity of Ca²⁺ for the RB69 pol–dP/T–Rh • dTTP complex under conditions of saturating Rh • dTTP concentrations could not be determined because there was no further quenching of dAP when Ca²⁺ was added to the RB69 pol–dP/T–Rh• dTTP complex.

Figure 7.

Dependence of Ca²⁺ or Rh • dTTP binding to the RB69 pol–dP/T binary complex. The K_d for Ca²⁺ with the RB69 pol–dP/T complex was 152 μM. K_d1 and K_d2 values for Rh • dTTP binding to the RB69 pol–dP/T or RB69 pol–ddP/T complex are listed in Table 2.

Table 2.

Effect of A Metal Site Occupancy on the Binding Affinity of Rh•dTTP with a dP/T or ddP/T and Rates of Fingers Closing (k₊₃)

		dP/T		ddP/T
Mg2+	Ca2+	Kdgapp (μM)	k+3 (s−1)a	Kdgapp (μM)	k+3 (s−1)a
−	−	108 ± 7	>1000b	63 ± 5	>1000b
+	−	40000c	2000c	111 ± 12	362 ± 25
−	+	9.1 ± 0.9	812 ± 39	9 ± 0.6	511 ± 29

^ak₊₃ indicates the forward rate for the Fingers closing in the kinetic Scheme 1.

^bThis value represents the lower boundary rate.

^cThis value was estimated from Figure 9 and simulation with the kinetic Scheme 1 using KinTekSim.

Since the amplitude of dAP quenching was nearly the same, with or without Ca²⁺, when a saturating concentration of RhTTP was added to the RB69 pol–dP/T complex, it implies that the fully closed ternary complex can form even when the A site is empty. The fluorescence observed with a saturating Rh • dTTP concentration was also consistent with the fluorescence of the RB69 pol–dP/T binary complex, as well as with that of the RB69 pol–ddP/T–Rh • dTTP ternary complex where the dAP stacks with its 3′ neighbor at saturating Rh • dTTP concentrations even though it does not become covalently linked to the nucleotide residue at the 3′ terminus of the primer.

Quench Amplitudes of 2AP Fluorescence Due to Rh • dTTP Binding to an RB69 pol–dP/T Binary Complex: Comparison of a dP/T with a ddP/T

The maximum amplitude for dAP quenching, when a ddP/T was a component of the binary complex, occurred with 2 mM Ca²⁺ and saturating Rh• dTTP or Ca • dTTP concentrations. The amplitude of the quench produced by saturating Rh • dTTP concentrations without Ca²⁺ was 26% of that observed when both Ca²⁺ and saturating Rh • dTTP concentrations were present and was 69% of that value when the mixture contained 10 mM Mg²⁺ instead of Ca²⁺. The higher affinity of Rh • dTTP for the E• D complex when Ca²⁺ rather than Mg²⁺ occupied the A site may be related to size differences and differences in coordination modes of the two metal ions. The quench amplitudes from the stopped-flow fluorescence scans, when a dP/T or a ddP/T was used, with Ca²⁺ in the A site, and either Rh • dTTP or Ca • dTTP in the B site, were dependent on the dTTP concentration.

Comparison of Rh • dTTP Affinities for RB69 pol–dP/T or RB69 pol–ddP/T Binary Complexes

While Rh • dTTP enabled us to use a dP/T, so that the contribution of the primer's 3′-OH group to the binding affinity of a deoxynucleoside triphosphate could be determined in the presence and absence of divalent cations, we still had to evaluate whether Rh • dTTP was able to mimic Mg• dTTP in its ability to induce the transition from the open to the closed state. In particular, we wanted to know (i) the rate at which Rh • dTTP induces the transition to the closed ternary complex and (ii) the Rh • dTTP concentration that would produce half-maximal quenching of dAP fluorescence. To do this, we compared the rates of dAP quenching induced by Rh • dTTP using both a dP/T and a ddP/T in the presence and absence of Ca²⁺. The effect on the affinity of Rh • dTTP for the binary E•D complex in the presence and absence of the primer's 3′-OH group could also be evaluated by determining the change in dAP fluorescence as a function of Rh • dTTP concentration, allowing us to estimate binding constants for Rh • dTTP in these two situations. As shown in Table 2, the K_dg^app for Rh • dTTP binding to the RB69 pol–ddP/T complex with the A site empty was 63 μM (Figure 1A of the Supporting Information) and 108 μM with a dP/T. When Ca²⁺ occupied the A site, the K_dg^app for Rh • dTTP was 9 μM with either a ddP/T (Figure 1B of the Supporting Information) or a dP/T, indicating that the presence or absence of an OH group at the 3′ terminus of the primer did not affect the affinity of Rh • dTTP for the E•D complex.

Rates of the Conformational Change When Rh • dTTP Binds to an E•D Complex: Comparison of a dP/T with a ddP/T on the Rate of Fingers Closing

Having determined the K_dg^app values for the different substrates and metal ion combinations, we then decided to compare the rates of conformational change when the A site was empty or filled with Ca²⁺ and when the B site was occupied by Rh³⁺ complexed with dTTP. At the same time, we wanted to determine what effect the 3′-OH group at the primer terminus would have on the rate of the conformational change. For this purpose, we compared a ddP/T with a dP/T. The first experiments were carried out with Rh • dTTP, dP/T, and 2 mM Ca²⁺. As can be seen in Figure 8A,B, the k_max was ~812 s⁻¹, a value nearly equal to what we reported previously with Mg• dTTP and a ddP/T (37). In all these experiments, there was no change in start points and the quench rates followed a single-exponential equation. These results suggest that a single step was involved and that the presence or absence of the 3′-OH group at the primer terminus had only a minor effect on the rate of dAP quenching. The similarity between the results obtained with the RB69 pol–ddP/T–Mg• dTTP combination and the RB69 pol–dP/T–Mg• dTTP complex in the presence of Ca²⁺ suggests that the same process (Fingers closing) was being observed in both situations. Furthermore, the Y axis intercept was 6.7 s⁻¹ (Figure 8B), a value consistent with the relatively slow rate of Fingers opening (~ 5s⁻¹) that was derived from pulse chase chemical quench experiments with Mg²⁺ as described in the following paper (38). However, the apparent K_dr^app was 100 μM (Figure 8B), a value much higher than the K_dg^app for binding of Rh • dTTP to a dP/T in the presence of 2 mM Ca²⁺, determined from equilibrium fluorescence measurements (Figure 6A). The final equilibrium amplitudes of the dAP quenching, derived from the fluorescence end points at 0.1 s (Figure 8A), when plotted as a function of Rh • dTTP concentration, fit a hyperbolic equation giving a K_da^app of 19.7 μM. The observation that the changes in amplitudes increased as a function of Rh • dTTP concentration (Figure 8C) is consistent with the reversibility of the EDN ⇌ FDN isomerization and the fact that chemistry did not occur. As in the case of Mg• dTTP with ddP/T, there was no change in start points, so that we were able to estimate the rate of the conformational change, which followed a single-exponential equation.

Figure 8.

Stopped-flow fluorescence scans of the RB69 pol–dP/T complex with increasing Rh • dTTP concentrations. For the dP/T sequence, see Figure 2A2. (A) One syringe contained 800 nM DNA and 4 μM RB69 pol in 50 mM MOPS buffer (pH 7) at 24 °C. The other syringe contained varying Rh • dTTP concentrations (0, 5, 10, 20, 40, 60, 75, 100, 125, 150, 200, and 250 μM from top to bottom) in MOPS buffer. Both syringes contained 2 mM CaCl₂. The fluorescence signal vs time fit a single-exponential equation. The scans are shown in different colors for different Rh • dTTP concentrations, and the fitting curves are colored black. (B) Rates from the single-exponential vs Rh • dTTP concentration fit a hyperbolic equation with a k_max of 812 ± 38.9 s⁻¹, a K_dr^app of 100 ± 18.3 μM, and a Y axis intercept of 7 ± 22 s⁻¹. (C) The amplitude change for the fluorescence quench vs Rh• dTTP concentration also fit a hyperbolic equation with a K_da^app of 19.7 ± 3.4 μM.

The same experiment was then performed under the same conditions except that a ddP/T was used instead of a dP/T so we could see how the primer's 3′-OH group affected the rate of dAP quenching. The fluorescence change fit best to a double exponential (Figure 2A of the Supporting Information). The faster rate of the first phase, obtained from the double exponential, when plotted against Rh • dTTP concentration, fit a hyperbolic equation. The average rate for the second phase was 21.5 ± 6.4 s⁻¹ and was independent of Rh•dTTP concentration. This second phase was not observed in experiments with dP/T carried out under the same conditions, suggesting an off-pathway step or a local conformational change of dAP at the active site due to lack of the 3′-OH group at the primer terminus. As seen in Figure 8 and Figure 2B of the Supporting Information, the estimated k_max was faster with the dP/T (812 s⁻¹) than with the ddP/T (511 s⁻¹) for the first phase which we assumed was the rate of the Fingers domain closing. The K_dr^app was also larger with the dP/T than with the ddP/T. We note that the Y axis intercept for the experiments with ddP/T (8.4 s⁻¹) (Figure 2 of the Supporting Information) was nearly the same as that found with the dP/T (6.7 s⁻¹) (Figure 8), indicating that the rate (k₋₃) for the reversal of the isomerized FDN complex to the collision complex, EDN (Scheme 1), was not influenced by the nature of the 3′ terminus of the primer and appeared to be independent of the nature of the cation that occupies the B metal ion site.

Scheme 1.

Expanded Kinetic Scheme for Primer Extension

Effect of Ca • dTTP versus Rh • dTTP on the Rate of Conformational Change

From stopped-flow fluorescence experiments, which gave an estimate of the Rh • dTTP concentration required to provide 50% of the maximum quench rate, a K_dr^app of 100 μM was obtained with dP/T (Figure 8B). This value was 10 times greater than the K_dg^app for Rh • dTTP binding to a dP/T as estimated by equilibrium fluorescence titration (Figure 6A). We note that there was no change in start points (Figure 8A). This result resembled the situation that we observed previously when Mg• dTTP was added to the RB69 pol–ddP/T binary complex (37). Next we wanted to determine whether Ca • dTTP would behave in a fashion similar to that of Rh • dTTP so we repeated the stopped-flow fluorescence experiment using Ca • dTTP in place of Rh • dTTP. The fluorescence scans (Figure 3A of the Supporting Information) with increasing Ca • dTTP concentrations showed a time-dependent decrease in dAP fluorescence that followed a single exponential with no change in start points (Figure 2A of the Supporting Information). The plot of k_obs versus Ca • dTTP concentration fit best to a hyperbolic equation giving a k_max of 1200 s⁻¹ (Figure 3B of the Supporting Information). The K_dr^app for Ca • dTTP was estimated to be 6.8 μM (Figure 3B of the Supporting Information), a value 125 times that of the K_dg^app (53 nM) determined from equilibrium fluorescence experiments (Figure 4 of the Supporting Information). A plot of the final amplitudes from Figure 3A of the Supporting Information versus dTTP concentration fit a hyperbolic function with a K_da^app of 1.9 ± 0.3 μM, consistent with the reversibility of the isomerization (Figure 3C of the Supporting Information).

Rates of Conformational Change and Chemistry When Mg²⁺ and an RB69 pol–dP/T Binary Complex Are Incubated with Increasing Rh • dTTP Concentrations

In all of the experiments described above, we investigated rates of the conformational change under conditions that did not allow chemistry. We now wanted to determine what would happen to the conformational change rate if we permitted chemistry to occur by using an RB69 pol–dP/T complex and adding increasing Rh • dTTP concentrations in the presence of 10 mM Mg²⁺. The results, depicted in Figure 9, show that the start points decrease as the Rh • dTTP concentration is increased, indicative of a quench rate of >1000 s⁻¹, and that the second phase of the fluorescence quench has a rate approximately equal to the k_pol obtained from the rapid chemical quench. These data are consistent with the kinetic scheme (Scheme 1), and the agreement between the fluorescent scans and those generated by KinTekSim using the parameters in Table 3 is quite good. The calculations supporting this assertion are given in the Supporting Information. These results should be compared with those of the same experiment that was carried out with Ca²⁺ instead of Mg²⁺. With Ca²⁺, there was a rapid quench phase where the decay of fluorescence followed single exponentials with different final amplitudes indicating the reversibility of the EDN ⇌ FDN isomerization. With Mg²⁺, the final amplitudes converged at the same value because all the Rh • dTTP was converted to product and where dAP was fully quenched because it was stacked against its 3′ neighbor in the P/T duplex. By reducing the rate of chemistry using Rh • dTTP instead of Mg• dTTP, we might have expected to see another fluorescent intermediate after Fingers closing, but none was found.

Figure 9.

Stopped-flow fluorescence for the incorporation of Rh • dTTP opposite dAP with a dP/T. One syringe contained 800 nM dP/T (Figure 2A2) and 4 μM RB69 pol in 50 mM MOPS buffer (pH 7) at 24 °C. The other syringe contained varying Rh • dTTP concentrations [0 (black), 20 (red), 60 (green), 200 (blue), and 400 μM (cyan)] in the same MOPS buffer. Both syringes contained 10 mM MgCl₂. The fluorescence signal vs time fits the kinetic scheme shown in Scheme 1. The parameters for the kinetic scheme for Rh • dTTP incorporation, generated from the KinTek Simulator program, are listed in Table 3. The experimental scans and the fitting curves are shown in different colors for different Rh • dTTP concentrations.

Table 3.

Kinetic Parameters for Rh•dTMP Incorporation According to Scheme 1

step	scheme	k+	k −
1	$\mathrm{E}+\mathrm{D}\underset{k_{-1}}{\overset{k_{+1}}{\rightleftharpoons }}\mathrm{ED}$	10 s−1 μM−1	0.01 s−1
2	$\mathrm{ED}+\mathrm{N}\underset{k_{-2}}{\overset{k_{+2}}{\rightleftharpoons }}\mathrm{EDN}$	10 s−1 μM−1	4000000 s−1
3	$\mathrm{EDN}\underset{k_{-3}}{\overset{k_{+3}}{\rightleftharpoons }}\mathrm{FDN}$	2000 s−1	4 s−1
4	$\mathrm{FDN}+\mathrm{N}\stackrel{k_{+4}}{\to }\mathrm{EP}$	7 s−1	0 s−1
5	$\mathrm{EP}\underset{k_{-5}}{\overset{k_{+5}}{\rightleftharpoons }}\mathrm{E}+\mathrm{P}$	0.01 s−1	10 s−1 μM−1

DISCUSSION

Effect of Metal Ions in the A and B Sites on the Affinity of Rh • dTTP for the E•D Binary Complex and on the Rate of dTMP Incorporation

The catalytic role for two divalent cations in the polymerase active site of DNA polymerases is well-established (39, 40). While Mg²⁺ is the physiologically relevant cation (41), Mn²⁺ can also efficiently catalyze the nucleotidyl transfer reaction; however, in its presence, the ability of replicative polymerases to discriminate against incorrect dNTPs is weakened, and thus, Mn²⁺ is considered to be a mutagenic agent (42–44). Attempts to crystallize RB69 pol–P/T–dNTP ternary complexes having a nonextendable primer with Mg²⁺ have failed, but the same types of complexes containing Ca²⁺ have provided high-quality crystals that have enabled RB69 pol structures to be determined at high resolution (21). This prompted us to use Ca²⁺ for our fluorescence studies since we could potentially relate our findings to the information available from X-ray crystallography, thus providing a structural context for interpreting our kinetic results.

Another advantage of using Ca²⁺ together with Rh • dTTP is that Ca²⁺ does not support catalysis. This made it possible to use a deoxy-terminated rather than a dideoxy-terminated primer to investigate the effect of filling the A or B metal ion sites individually on the rates of conformational change, and on the binding affinity of Rh • dTTP. It also allowed us to compare deoxy- and dideoxy-P/Ts to determine what role the primer's 3′-OH group has on the affinity of the incoming Rh • dNTP, and on the rate of Fingers closing. Nevertheless, we are aware that Ca²⁺ is an imperfect mimic for Mg²⁺ in a number of respects. Its ionic radius is larger than Mg²⁺, making it less effective in polarizing the α P–O bond of the incoming dNTP for nucleophilic attack by the 3′-OH group of the primer terminus (33). This is likely to be part of the reason that Ca²⁺ is unable to catalyze the nucleotidyl transfer reaction. In addition, the coordination mode of Ca²⁺ is more variable than that of Mg²⁺, which is usually octahedral. Ca²⁺ has been observed to be coordinated between four and eight ligands. There is also a difference in hydration enthalpy, −1577 kJ for Ca²⁺ versus −1921 kJ for Mg²⁺ (45). This results in inner-sphere coordination of Ca²⁺ to DNA (46). Ca²⁺ does not lower the pK_a of a water molecule as much as Mg²⁺ [12.8 and 11.4, respectively (47, 48)], and this likely also contributes to the catalytically inert behavior of Ca²⁺ in phosphoryl transfer reactions. These properties may also account for the fact that Ca²⁺ in the A site increases the affinity of Rh • dTTP for the RB69 pol–P/T complex.

The dissection of the effect of A and B site occupancy on the affinity of the incoming dNTP for the E •D complex and on the rate of conformational change can be investigated only when a nonexchangeable metal ion is complexed with an incoming dNTP. Despite the fact that the phosphate oxygen–rhodium bond distances are similar to those of Mg •dNTPs (11), it is not possible to conclude, a priori, that Rh • dNTPs are good substrates compared to dNTPs complexed with Mg²⁺. The fact that full quenching could be achieved with the A site empty suggests that A site occupancy is not a prerequisite for formation of the closed ternary complex. It would seem that dAP can be fully stacked in the duplex, even when the incoming Rh • dTTP is not covalently attached to the primer. The low K_dg^app for Rh • dTTP, when Ca²⁺ is present, suggests that filling the A site stabilizes the closed, ternary structure. Note that Ca²⁺ alone, in the absence of any dNTP, results in dAP quenching, but with only 26% of the full amplitude that is observed when the Rh • dTTP concentration is at saturating levels. However, we do not have any independent evidence to show that the dAP quenching is actually caused by Ca²⁺ occupancy of the A site only.

Rapid chemical quench experiments clearly show that Rh • dTTP was incorporated at less than one-tenth of the rate of Mg• dTTP, when matched opposite a dAP as the templating base, supporting the notion that Rh • dTTP did not undergo metal ion exchange under the conditions used for the chemical quench (Table 1). It is interesting to compare the kinetic parameters obtained from rapid chemical quench experiments with Rh• dTTP and Mg• dTTP when matched opposite a templating dA or dAP. As shown in Table 1, the largest difference in k_pol (10-fold) was observed when Rh • dTTP was matched opposite dA or dAP, favoring dA. In contrast, when Mg• dTTP was matched against dA or dAP, these differences in k_pol values were only 1.5-fold, in favor of dA. This difference is in line with what has been reported previously (37, 49). On the other hand, when the k_pol for Rh • dTTP was compared with that of Mg• dTTP, positioned opposite dA, there was only a small difference (1.5-fold) in favor of Mg• dTTP, but the enzyme exhibits less base discrimination in the presence of dAP rather than dA when Rh • dTTP is used. This suggests that dAP is a good mimic of dA when opposite Mg• dTTP, but not when opposite Rh • dTTP. While Rh • dTTP does not behave exactly like Mg• dTTP, even when Mg²⁺ is in the A site, it still serves as a substrate and presumably as a ligand for inducing the critical conformational change when the A site is empty or occupied by a divalent cation that does not permit catalysis (11). It is likely that the differences in coordination between Rh³⁺ and Mg²⁺ for dTTP can account for changes in rates of dTMP incorporation but have little effect on the rates of the conformational change. Comparison of the K_dq^app values for Mg• dTTP and Rh • dTTP is instructive in that they show approximately the same 2:1 ratios of apparent binding affinities for dA relative to dAP. The single substitutions show 2.8- and 7.9-fold effects, compared with a 104-fold effect for the double substitution (2.8 × 7.9) = 22, which is 5-fold less than 104). These data and the molecular dynamics simulation suggest a 1 Å movement of dAP opposite Rh • dTTP under this condition compared to dA opposite Mg• dTTP, which is consistent with the observed cooperativity in binding and/or recognition of templating base and metal B. This confirms previous reports that when dAP is the templating base, the complementarity with dTTP is altered compared to that when the templating base is dA (49, 50). The decrease in k_pol for the Rh • dTTP–dAP pair is probably due to imperfect alignment of the attacking 3′-oxygen at the terminus of the primer with the R-phosphorus atom of Rh • dTTP. These results suggest that the nature of the metal ion in the B site affects the catalytic rate in addition to facilitating dNTP binding.

Effect of Rh • dTTP on the Rate of the Conformational Change

It is interesting to compare the rates of conformational change induced when Mg• dTTP or Rh • dTTP is the incoming dNTP. We have previously shown that there were two phases for the rate of conformational change when Mg• dTTP was matched opposite dAP: a very rapid phase with a rate of >1000 s⁻¹ which we assigned to Fingers closing and a slower phase (rate of 290 s⁻¹) that was approximately the same as the rate of dTMP incorporation as determined by rapid chemical quench (37). With Rh • dTTP opposite dAP, there was also a rapid fluorescence quench (>1000 s⁻¹) followed by a slower quench with approximately the same rate as dTMP incorporation. This slower phase of fluorescence quenching may be reporting a change in dAP stacking that occurs just prior to or simultaneously with the chemical step. Modeling the RB69 pol complex as it proceeds to the transition state shows that there must be an ~1 Å shift of the templating base with respect to its 3′ neighbor that could create additional stacking of dAP, causing an increase of ~20% in the amplitude of the quench if the 3′ neighboring base is a purine (J. Wang, personal communication). This was indeed the case with our template sequence, and the amplitude of the second phase was ~25% of the total quench amplitude. If there is a concerted movement of the substrates in their approach to the transition state, and in progressing through it to product, the rate of the additional stacking interaction cannot be separated from the rate of the chemical step. Note that this second phase is not observed when chemistry does not occur and that the first rate, which reflects Fingers closing, is reduced from >1000 to ~500 s⁻¹ at a saturating Rh • dTTP concentration. In contrast to the situation with dP/Ts, where the quench amplitudes of stopped-flow fluorescence scans converge at a single value, the amplitudes of the scans with ddP/Ts are dependent on the Rh • dTTP concentration because the quenched state is reversible, as depicted in Scheme 1.

Among the reports that have estimated the rate of Fingers closing, the most recent have come from FRET experiments with Klentaq (51, 52) and with T7 DNA polymerase (T7 pol) (53) which serve to illustrate the scope and limitation of this approach. With Klentaq, the rate of Fingers closing was very rapid, exceeding the rate of dNMP incorporation, as monitored by chemical quench experiments, suggesting that chemistry, or a kinetically silent step before chemistry, limited the nucleotidyl transfer rate (51, 52). With T7 pol, labeling with a fluorescent dye at a single Cys residue in the Fingers domain provided a convenient probe for monitoring conformational changes. Upon addition of a correct dNTP, the fluorescence was partially quenched, whereas addition of an incorrect dNTP caused fluorescence enhancement. The rates of these fluorescent changes were measured and ascribed to conformation changes involving opening and closing of the Fingers domain (53). Other papers that have dealt with the rate of Fingers closing include studies with Klenow fragment (13), T4 pol (54), RB69 pol (37), and pol β (55).

Effect of Individual Metal Ions on the Rates of Conformational Change in pol β

Attempts have been made previously to dissect the potential roles of the two metal ions, known to be essential for catalysis, in facilitating the conformational change that occurs after dNTP binding. Experiments with this aim were carried out using polymerase β (pol β) and Cr(III) nucleotide complexes because of their exchange-inert properties (56). These authors suggested that binding of the Cr •dNTP to the pol β–-P/T complex facilitated Mg²⁺ binding to the catalytic metal ion site (site B in the nomanclature Zhong et al. but termed the A metal ion site in our paper). Our results with RB69 pol differed from those reported for pol β in that we observed closing of the Fingers domain with Rh • dTTP even when the A site was empty. Instead of Rh • dTTP facilitating Mg²⁺ binding in the A site, Mg²⁺ binding did not occur until the Fingers reopened, a step that occurred more slowly than Fingers closing (38). We also found that the rate of Fingers reopening limited the rate of dTMP incorporation under conditions where Mg²⁺ was added to a ternary RB69 pol–dP/T–Rh • dTTP complex that had been assembled in the absence of divalent cations (38).

Dynamic simulations were also performed with pol β by Yang et al. (56), leading to the conclusion that the dNTP–metal ion complex alone cannot induce pol β closing before chemistry (56). Although a comparable dynamic simulation study has not been performed with RB69 pol, our experimental results with RB69 pol differ from the pol β simulations, suggesting that either the two polymerases behave very differently with respect to Fingers closing. In a subsequent report (57), the authors revised their scheme indicating that closing of the Fingers could be induced by Cr • dNTP binding alone with the A site empty but that the slow step reflected Mg²⁺ binding when the polymerase was still in the closed state (57). This revised model is closer to what we have proposed for RB69 pol but still differs, in that we suggest that the Fingers have to reopen, at least partially, to allow for Mg²⁺ binding and that it is the rate of Fingers reopening that is slow, rather than the rate of Mg²⁺ binding to the A site in the closed ternary complex (38).

In a more recent report, Rh(III) rather than Cr(III) • dNTP complexes were employed with pol β because of reasons stated here and in the paper by Bakhtina et al. (11). In this version, the authors stated that the slow phase following Fingers closing occurs after Mg²⁺ binding and that the slow step reflected the rate of the chemical step. While this revised scheme seems to be the one that best fits with all of the pol β data, it still does not square completely with our RB69 pol results. Further experimentation along the lines suggested by Bakhtina et al. (11) will be required to determine whether pol β represents a special case or whether the conclusions drawn from the extensive studies on pol β can be applied generally to all DNA polymerases.

SUMMARY

By employing Rh • dTTP as the incoming nucleotide matched opposite dAP, we have been able to separate the effect of A and B metal ion site occupancy on the rates of Fingers closing and on the affinity of Rh • dTTP for the E • D binary complex. We show the dependence of RhdTTP binding as a function of Ca²⁺ concentration. We show that the presence or absence of the 3′-OH group on the primer does not appreciably affect the rate of Fingers closing with the A site empty or when it was filled with Ca²⁺ when Rh • dTTP is the substrate. In the following paper (38), we show that the rate of Fingers reopening is slow in the presence of correct dNTPs but very rapid when an incorrect dNTP is part of the ternary complex. The implications for mechanisms that control polymerase fidelity are discussed in that paper.