Increased Catalytic Activity and Altered Fidelity of Human DNA Polymerase ι in the Presence of Manganese*

All DNA polymerases require a divalent cation for catalytic activity. It is generally assumed that Mg2+ is the physiological cofactor for replicative DNA polymerases in vivo. However, recent studies suggest that certain repair polymerases, such as pol λ, may preferentially utilize Mn2+ in vitro. Here we report on the effects of Mn2+ and Mg2+ on the enzymatic properties of human DNA polymerase ι (pol ι). pol ι exhibited the greatest activity in the presence of low levels of Mn2+ (0.05–0.25 mm). Peak activity in the presence of Mg2+ was observed in the range of 0.1–0.5 mm and was significantly reduced at concentrations >2 mm. Steady-state kinetic analyses revealed that Mn2+ increases the catalytic activity of pol ι by ∼30–60,000-fold through a dramatic decrease in the Km value for nucleotide incorporation. Interestingly, whereas pol ι preferentially misinserts G opposite T by a factor of ∼1.4–2.5-fold over the correct base A in the presence of 0.25 and 5 mm Mg2+, respectively, the correct insertion of A is actually favored 2-fold over the misincorporation of G in the presence of 0.075 mm Mn2+. Low levels of Mn2+ also dramatically increased the ability of pol ι to traverse a variety of DNA lesions in vitro. Titration experiments revealed a strong preference of pol ι for Mn2+ even when Mg2+ is present in a >10-fold excess. Our observations therefore raise the intriguing possibility that the cation utilized by pol ι in vivo may actually be Mn2+ rather than Mg2+, as tacitly assumed.

It has been known for several decades that DNA polymerases (pol) 2 require a divalent cation as an activator for phosphotidyl transfer (1)(2)(3). Two metal ions are usually coordinated by three acidic amino acids within the active site of the polymerase, so as to form a metal bridge between the enzyme and the terminal phosphoryl group of the substrate (4,5), thereby facilitating the departure of the pyrophosphate moiety (6). Based upon its cellular abundance, it is generally believed that Mg 2ϩ is the activating cofactor in vivo. However, Mn 2ϩ , Co 2ϩ , Ni 2ϩ , and Zn 2ϩ have the capacity to substitute for Mg 2ϩ under certain conditions in vitro (2,3,7), but usually at the consequence of reduced fidelity and in some cases decreased processivity. The effect of substituting Mg 2ϩ with Mn 2ϩ on the activity of A-and B-family polymerases has been widely studied (8 -12). In addition to reducing fidelity, Mn 2ϩ also helps facilitate translesion synthesis by certain replicative polymerases, including Escherichia coli pol I (13) and herpes simplex virus-1 UL30 protein (14) but not others, such as T4 DNA polymerase, or bovine pol ␦ (12,14).
The effect of Mn 2ϩ on X-family polymerases has also been studied. In the case of pol ␤, Mn 2ϩ decreases the K m value of nucleotide incorporation by ϳ30-fold (15), such that there is little regard for the instructions provided by the templating base (16). Recent data also suggest that the phylogenetically related pol may have actually evolved to utilize Mn 2ϩ under physiological conditions, because the enzyme is active over a wide range of Mn 2ϩ concentrations and is inhibited by high levels of Mg 2ϩ (17,18).
To date, there have been few studies on the effects of Mn 2ϩ on the Y-family of DNA polymerases. In the single published report, Sulfolobus solfataricus Dpo4 behaved in a similar manner to the replicative polymerases, in that 5 mM Mn 2ϩ increased the efficiency of lesion bypass, but with a concomitant 3-4-fold decrease in the overall fidelity of the enzyme (19). The effect of Mn 2ϩ on other Y-family polymerases, including human DNA polymerases , , and , is largely unknown. We were therefore very interested in determining the effects of various metal ions on the activity of the human Y-family polymerases and in particular pol , because its reported biochemical properties are quite unusual. Indeed, pol exhibits a remarkable template-dependent fidelity in vitro. When replicating template T in the presence of Mg 2ϩ , pol preferentially misinserts G over the correct base A by a factor of 3-10-fold (20 -24). In contrast, when replicating template A, the enzyme is relatively accurate with misincorporations occurring in the range of 10 Ϫ3 -10 Ϫ4 (20 -22). Thus, the fidelity of pol can vary by up to 100,000-fold depending upon the template sequence being replicated. Here we report on the enzymatic properties of pol in vitro in the presence of various divalent cations, in particular Mg 2ϩ and Mn 2ϩ .

EXPERIMENTAL PROCEDURES
Enzymes-N-terminal glutathione S-transferase-tagged human pol was expressed and purified from baculovirus-infected insect Sf9 cells as described (20). C-terminal His-tagged human pol (25) was also expressed in baculovirus-infected Sf21 cells and purified by nickel-agarose, Q-Sepharose, and Bio-Gel HT hydroxyapatite chromatography. N-terminal His-tagged human pol (26) was purchased from Enzymax (Lexington, KY).
Primers and Templates-All primers, undamaged templates, and the abasic site-containing template, TXT30, were synthesized by Lofstrand Laboratories (Gaithersburg, MD) and gelpurified prior to use. The cyclobutane pyrimidine dimer containing template, TTA48, was synthesized and purified by Phoenix Biotechnologies, Inc. (Huntsville, AL). The pyrimidine-pyrimidone (6-4)-containing template, 6-4TT30, was a kind gift from Shigenori Iwai (Osaka University). The benzo-[a]pyrene diol epoxide-containing template, BPDE29, was a kind gift from Don Jerina (NIDDK, National Institutes of Health). For the experiments presented in Fig. 1, the template was T50, 5Ј-CGT CTA GAC GAA TTC ACG GCT CAA GCT TGC TTG CGC ATG CTC TGC AGG CG-3Ј, and the primer was a 22-mer that was complementary to the template nucleotides underlined. The template used for the experiments presented in Figs. 2 and 5 and Table 1, parts B and C, was T50M13, 5Ј-CGG TAA TGA TTC CTA CGA TGA AAA TAA AAA CGG CTT GCT TGT TCT CGA TG-3Ј, which corresponds to M13mp18 (nucleotides 3486 -3437) (27). The primer was a 20-mer complementary to the template nucleotides underlined. The experiments described in Fig. 3 and Table 1, Part A utilized 48-mer templates UTTA and TTA48, with the sequence 5Ј-TCG ATA CTG GTA CTA ATG ATT AAC GAA TTA AGC ACG TCC GTA CCA TCG-3Ј. The primer was a 16-mer that was complementary to the template nucleotides underlined. The adjacent Ts were covalently linked to form a cyclobutane pyrimidine dimer (CPD) in template TTA48. The experiments described in Fig. 4 utilized three lesion-containing templates. TXT30 was a 30-mer with the sequence 5Ј-CTC GTC AGC ATC TXC ATC ATA CAG TCA GTG-3Ј, where the X was a synthetic tetrahydrofuran moiety (abasic site). The primer was a 16-mer that was complementary to the template nucleotides underlined. The second template, BPDE29, was a 29-mer with the sequence 5Ј-GCT CGT CAG CAG ATT TAG AGT CTG CAG TG-3Ј, where the A has a benzo[a]pyrene adduct at its N 6 position (28). The primer was a 16-mer that was complementary to the template nucleotides underlined. The third template, 6-4TT30, was a 30-mer with the sequence 5Ј-CTC GTC AGC ATC TTC ATC ATA CAG TCA GTG-3Ј, where the two Ts indicate a 6-4 pyrimidine-pyrimidone dimer (29). The primer was a 16-mer that was complementary to the template nucleotides underlined. All primers were 5Ј-labeled with [␥-32 P]ATP (5000 Ci/mmol; 1 Ci ϭ 37 gBq) (GE Healthcare).
Annealing Primers to Templates-100 nM of 32 P-labeled primer was mixed with 150 nM of the corresponding template in 50 mM Tris⅐HCl (pH 8.0), 10 mM NaCl, 1.42 mM 2-mercaptoethanol, and 50 g/ml bovine serum albumin, heated at 95°C for 5 min, and then slowly cooled to room temperature over a period of several hours.
Primer Extension Reaction-Standard replication reactions (10 l) contained 40 mM Tris⅐HCl (pH 8.0), 2.5% glycerol, 0.1 mg/ml bovine serum albumin, 10 mM dithiothreitol, 10 nM of 5Ј-32 P-labeled primer-template DNA, 2 nM glutathione S-transferase-pol , 5 nM His-pol , or His-pol . Reaction times, divalent metal ions, in the form of -Cl 2 salt, and dNTPs concentrations are specified in the figures and legends. After incubation at 37°C, reactions were terminated by addition of 10 l of 95% formamide, 10 mM EDTA, and samples were heated to 100°C for 5 min and then immediately transferred on ice. The reaction products were separated by 15% polyacrylamide, 8 M urea gel electrophoresis and analyzed using a Fuji FLA-5100 PhosphorImager and ImageGauge software.
Steady-state Kinetics-Steady-state kinetic parameters V max and K m , for incorporation opposite template T or A, were measured in standing-start reactions as described previously (20,27,28). Reactions contained 0.4 nM of pol , and conditions were optimized to ensure that the reaction remains in a linear range.

Effects of Various Divalent Metal Ions on the Activity of Human Y-family Polymerases-
The ability of human DNA polymerases , , and to utilize various divalent cations as activators for polymerization was assayed in replication reactions using a radiolabeled 22-mer primer annealed to a 50-mer oli-gonucleotide template. Each metal ion was present as the chloride salt, at a concentration of either 0.5 or 5 mM. Quite surprisingly, of the eight divalent cations assayed, only Cu 2ϩ failed to support any catalysis. The remaining seven cations all promoted polymerization to some extent, with the degree of activation being acutely polymerase-specific ( Fig. 1). As expected, robust synthesis was observed with pol and pol in the presence of both 0.5 and 5 mM Mg 2ϩ . In contrast, however, the activity of pol was far greater in 0.5 mM Mg 2ϩ than in 5 mM Mg 2ϩ (Fig.  1). All three polymerases were highly active in the presence of Mn 2ϩ with greater activity observed in 0.5 mM Mn 2ϩ compared with 5 mM Mn 2ϩ . Co 2ϩ also served as an activator for all three polymerases, with pol being the most active of the three enzymes in 0.5 mM Co 2ϩ . Similarly, pol was more active than either pol or pol in the presence of Ca 2ϩ or Zn 2ϩ . In contrast, pol exhibited much greater activity in the presence of 5 mM Cd 2ϩ or Ni 2ϩ , compared with either pol or pol . We conclude from these studies that each human Y-family polymerase possesses its own unique ability to utilize different metals as cofactors for DNA synthesis. In the case of pol and pol , the preferred activator appears to be Mg 2ϩ . However, based upon the data presented in Fig. 1, pol either prefers low levels of Mg 2ϩ or Mn 2ϩ for optimal activity in vitro.
Titration of Mn 2ϩ or Mg 2ϩ in pol and pol Replication Reactions-We next determined the optimum concentration of Mg 2ϩ or Mn 2ϩ required to promote peak activity of pol or pol by performing primer extension assays in the presence of Mg 2ϩ or Mn 2ϩ ranging in concentrations from 0.05 to 8 mM. The template for these assays was a 50-mer oligonucleotide that has a run of five template As followed by a T and then four additional As immediately downstream of the primer. This sequence context was chosen, as we have previously shown that in the presence of 5 mM Mg 2ϩ pol readily incorporates T opposite the five template As, but then frequently misinserts G opposite template T (29). The efficiency of G (or A) insertion opposite T is much lower than that of the incorporation of T opposite A (20). Further elongation from the mispair is also reduced compared with that of the correct A:T base pair (24), and as a consequence, there is a strong pol -dependent pause 5-6 bases from the starting primer terminus (Fig. 2, top, 0.1-0.5 mM Mg 2ϩ ). The percentage of replication products longer than 6 bp therefore provides The arrows indicate the location of a template T that is located 6 bp from the end of the primer terminus. In the presence of Mg 2ϩ , pol frequently misincorporates G opposite template T. The mispair is poorly extended compared with the correct base pair, and as a consequence a strong pause site is observed 6 bp from the primer terminus. Thus, one measurement of the overall catalytic activity of pol is its ability to extend beyond the normal pause site at template T. When products 7 bp and longer are plotted as a function of Mn 2ϩ (OE) or Mg 2ϩ (‚), it is clear that pol exhibits greatest activity at 0.075 mM Mn 2ϩ and 0.25-0.5 mM Mg 2ϩ . Although pol does not pause at template T, we used the same method to determine the optimal conditions for pol activity. As can be seen (lower panels), pol exhibits greatest activity in 0.05-0.1 mM Mn 2ϩ and 0.1-2 mM Mg 2ϩ . a reasonably accurate method of measuring the catalytic activity of pol . By using this approach, it is readily obvious that pol is very active in the presence of low levels of Mn 2ϩ (Fig. 2, top, left-hand panel). Presumably, the ability of pol to overcome the kinetic block to replication when the enzyme encounters a template T reflects a change in the specificity of nucleotide (mis)insertion (see below for more detailed analysis), as well as through an increased ability to extend G:T mispairs. 3 The peak activity of pol was observed in the range of 0.05-0.2 mM Mn 2ϩ . Similarly, pol was also much more active in the presence of low levels of Mg 2ϩ , with peak activity occurring within the range of 0.2-0.5 mM Mg 2ϩ . Indeed, higher concentrations of Mg 2ϩ appear to inhibit pol , as under these assay conditions the limiting amount of pol in the reaction was barely sufficient to elongate the primer by 1-2 bp (Fig. 2, right-hand panel, 8 mM Mg 2ϩ ).
In parallel experiments with human pol , the enzyme was very active in the presence of 0.05-0.1 mM Mn 2ϩ , but it rapidly began to lose activity at higher concentrations and was severely inhibited in reactions containing Ͼ2 mM Mn 2ϩ . Peak activity of pol in the presence of Mg 2ϩ was observed in the range of 0.075-2 mM Mg 2ϩ , and similar to pol , the limiting amount of pol in the reaction appeared to have somewhat reduced activity in 5-8 mM Mg 2ϩ .
In an attempt to further optimize the in vitro reaction conditions for pol synthesis, we also performed replication assays in the presence of low Mn 2ϩ or Mg 2ϩ in buffers that ranged in pH from 7.5 to 9.0 and in the presence of increasing amounts of NaCl. In general, changing the pH of the assay buffer had little effect on the activity of the enzyme. In contrast, the activity of pol steadily decreased as the concentration of NaCl was increased from 0 to 150 mM. 3 Activity of pol and pol on a CPD-containing Template in the Presence of Mg 2ϩ and Mn 2ϩ -We have previously reported that in the presence of 5 mM Mg 2ϩ , pol is capable of inserting a base opposite the 3ЈT of a T-T CPD, as well as facilitating low levels of complete translesion synthesis (30,31). We therefore wanted to reinvestigate the ability of pol to facilitate translesion synthesis over a range of Mg 2ϩ and Mn 2ϩ concentrations (Fig. 3). In these reactions, the 3Ј end of the primer was juxtaposed to the CPD and are therefore considered "standing start" reactions. Similar to the observations reported above, maximal activity for pol was observed in the range of 0.075-0.25 mM Mn 2ϩ on the damaged template (Fig. 3A). Under these conditions, there was significant extension of primers past the CPD. Indeed, we estimate that approximately 60% of the primer was elongated past the CPD by pol in the presence of 0.2 mM Mn 2ϩ (Fig. 3A, right-hand panel). The results with Mg 2ϩ were similar to our earlier published observations (30,31), in that the 3ЈT of the CPD is a strong kinetic block to pol and elongation past the 3ЈT was greatest in the range of 0.25-2 mM Mg 2ϩ . Indeed, we estimate that in the presence of 0.5 mM Mg 2ϩ , ϳ8% of the primers are elongated past both Ts of the CPD, compared with ϳ2.5% in the presence of 5 mM Mg 2ϩ , which is consistent with our earlier observations (30,31).
The ability of pol to bypass a CPD in vitro is unrivalled. It does so with the same or even higher efficiency than it replicates undamaged DNA (32)(33)(34). In the presence of Mn 2ϩ , robust bypass activity was observed from 0.05 to 0.25 mM Mn 2ϩ , but bypass activity steadily declined as the Mn 2ϩ concentration increased (Fig. 3B). In contrast, pol exhibited strong bypass activity in the presence of 0.05-8 mM Mg 2ϩ .
We conclude from these studies that pol and pol have differing affinities for metal ion activation. Translesion synthesis of a CPD by pol is stimulated ϳ3-fold by low levels of Mg 2ϩ and up to 25-fold in the presence of 0.2 mM Mn 2ϩ compared 3 E. G. Frank and R. Woodgate, unpublished observations. with 5 mM Mg 2ϩ . In contrast, pol appears to exhibit efficient bypass a CPD over a wide range of Mg 2ϩ concentrations.
General Ability of Mn 2ϩ to Stimulate Translesion Synthesis by pol -We next examined whether low levels of Mn 2ϩ or Mg 2ϩ would stimulate pol -dependent synthesis of a synthetic abasic site, a BPDE adduct, and a 6-4PP lesion (Fig. 4). Previous in vitro studies in the presence of 5 mM Mg 2ϩ indicated that pol can efficiently incorporate a nucleotide opposite an abasic site, a benzo[a]pyrene lesion, and up to two bases opposite the 6-4PP; however, further extension is limited (21,30,35). The previous results are recapitulated here in the presence of 5-8 mM Mg 2ϩ (Fig. 4). Lower concentrations of Mg 2ϩ helped stimulate incorporation opposite each of the three lesions but had little effect on bypass of the lesion. In contrast, in the presence of 0.2 mM Mn 2ϩ , pol -dependent bypass of the abasic site increased to ϳ18%, and bypass of the BPDE adduct increased to ϳ10% (Fig. 4, A and B). However, whereas Mn 2ϩ appears to stimulate the ability of pol to incorporate opposite both bases of the 6-4PP, it did not appreciably alter the extent of 6-4PP lesion bypass.
pol Has Higher Affinity for Mn 2ϩ than Mg 2ϩ -As noted earlier, in the presence of Mg 2ϩ pol readily misinserts G opposite template T, causing the enzyme to pause. This pausing is much less evident in the presence of Mn 2ϩ , with the accumulation of a significant proportion of replication products longer than 6 bp (Fig. 2, top). These two distinguishable properties gave us an opportunity to assay which cation pol prefers to utilize when presented with both Mn 2ϩ and Mg 2ϩ .
In the experiments depicted in Fig.  5, pol was first preincubated in the presence of various concentrations of Mn 2ϩ (Fig. 5A), or Mg 2ϩ (Fig.  5B), for 5 min at room temperature. Reactions were started by the addition of various concentrations of either Mg 2ϩ or Mn 2ϩ and the radiolabeled primer/template, followed by incubation at 37°C for 20 min. As shown in Fig. 5, it is clear that when preincubated with Mn 2ϩ and then challenged with Mg 2ϩ , the pol -dependent replication products are reminiscent of those observed in the presence of Mn 2ϩ alone, with very little pausing at template T (Fig. 5A). This property occurs over a wide range of Mn 2ϩ /Mg 2ϩ concentrations, and is even evident when Mg 2ϩ is present in a 40-fold molar excess over Mn 2ϩ (see Fig. 5A, 0.05 mM Mn 2ϩ /2 mM Mg 2ϩ ). The preference of pol for Mn 2ϩ is more evident in the reactions in which pol was preincubated in the presence of Mg 2ϩ and then challenged with Mn 2ϩ (Fig. 5B). In the absence of added Mn 2ϩ , the Mg 2ϩ reactions exhibited the classic termination pattern, with a strong pause 5-6 bp downstream of the primer. However, in all reactions containing Mn 2ϩ this pausing is significantly reduced. Again, Mn 2ϩ stimulation occurred over a wide range of Mg 2ϩ /Mn 2ϩ concentrations and was obvious even when Mg 2ϩ was present in a 10 -40-fold molar excess over Mn 2ϩ . Based upon these observations, we conclude that pol preferentially utilizes Mn 2ϩ as the activator for polymerization, even in reactions containing a vast molar excess of Mg 2ϩ .
Kinetics of pol -dependent Incorporation in the Presence of Mn 2ϩ and/or Mg 2ϩ -Several groups have previously reported the kinetic parameters for pol -dependent nucleotide incorporation on an undamaged template. In all cases, these values were determined in the presence of 5 or 8 mM Mg 2ϩ (20 -22). Under these conditions, the two most striking properties of pol are its ability to preferentially misinsert G opposite T and to faithfully and efficiently incorporate A opposite T. However, our present studies indicate that pol is much more catalytically active at much lower concentra-tions of Mg 2ϩ and Mn 2ϩ . We have therefore determined the kinetic parameters for (mis)insertion opposite template T and A, in the presence of 0.075 mM Mn 2ϩ , 0.25 mM Mg 2ϩ , and 5 mM Mg 2ϩ either alone (Table 1, parts A and B), or in combination (Table 1, part C). These kinetic studies revealed that the increase in catalytic activity observed in our earlier experiments is largely derived from a dramatic decrease in the K m value for nucleotide incorporation. Indeed, whereas the V max was essentially unchanged under all assay conditions, the K m value varied by as much as 2400-fold at template T and 30,000-fold at template A. For example, the K m value for the correct incorporation of A opposite T decreased by ϳ245-fold in the presence of 0.25 mM Mg 2ϩ compared with 5 mM Mg 2ϩ , and there was a further 10-fold reduction in the presence of 0.075 mM Mn 2ϩ . The K m value for the misincorporation of G opposite T also dropped dramatically but to a slightly lower extent. As a result, there was an overall change in the frequency of misincorporation (f inc ) under the various assay conditions. Although misincorporation of G opposite T was favored by a factor of 2.5-fold in 5 mM Mg 2ϩ and by 1.4-fold in 0.25 mM Mg 2ϩ , in reactions containing 0.075 mM Mn 2ϩ , the correct incorporation of A was actually favored ϳ2.3-fold more than G. Thus, in contrast to its fidelity-reducing effects on all known polymerases, Mn 2ϩ actually increases the fidelity of pol at template T. It should be  stressed, however, that even though Mn 2ϩ altered the relative ratios of (mis)incorporation opposite template T so as to favor the incorporation of the correct base, pol nevertheless remains extremely error-prone under these conditions. At template A, 0.075 mM Mn 2ϩ caused the characteristic decrease in fidelity as there was a dramatic increase in the frequency of misincorporation of A, over the correct base T ( Table  1, part B). Similar to our observations above, these effects were largely driven by changes in the K m value for nucleotide incorporation. The largest effect was on the K m value for misinsertion of A, where it dropped from 70 to 90 M in the presence of Mg 2ϩ to 0.003 M for Mn 2ϩ (ϳ23,000 -30,000-fold difference).
Overall, the catalytic activity (V max /K m ) of pol increased in the presence of low levels of Mg 2ϩ or Mn 2ϩ . The smallest effect was observed for (mis)incorporation opposite A, where the catalytic activity only increased ϳ2-2.5-fold in the presence of 0.25 mM Mg 2ϩ compared with 5 mM Mg 2ϩ . However, much larger effects were observed at template T, where the catalytic activity increased ϳ45-80-fold in 0.25 mM Mg 2ϩ compared with 5 mM Mg 2ϩ . The effects of 0.075 mM Mn 2ϩ were much more dramatic, with an additional ϳ11-3300-fold increase in catalytic activity over that observed in the presence of 0.25 mM Mg 2ϩ .
In the simultaneous presence of 0.075 mM Mn 2ϩ and 0.25 mM Mg 2ϩ , the catalytic activity (V max /K m ) of pol for the correct incorporation of T opposite A was intermediate to that observed in the presence of each individual cation. For example, the catalytic activity of pol was ϳ4.5-fold lower than in the presence of 0.075 mM Mn 2ϩ alone but ϳ2.6-fold higher than in the presence of 0.25 mM Mg 2ϩ alone. With regard to fidelity, Mn 2ϩ appears to be dominant, in that pol exhibited ϳ200-fold lower fidelity in the presence of both Mn 2ϩ and Mg 2ϩ compared with that observed in the presence of 0.25 mM Mg 2ϩ alone.

DISCUSSION
In this study, we have investigated the effects of various divalent cations on the in vitro properties of human DNA polymerases , , and . The most abundant divalent cation is Mg 2ϩ , with an intracellular concentration of 0.21 to 0.24 mM (36,37). Given its abundance, Mg 2ϩ is therefore generally considered to be the physiologically relevant divalent metal cofactor for most DNA polymerases. In support of this notion, our analyses strongly suggest that the preferred cation for pol and pol is Mg 2ϩ , as both enzymes were active over a wide range of Mg 2ϩ concentrations. In contrast, pol exhibited greatest activity within a much more narrow concentration range of Mg 2ϩ (Fig.  2). Peak activity was observed at 0.25 mM Mg 2ϩ , and synthesis was reduced significantly in reactions containing Ͼ2 mM Mg 2ϩ . As far as we are aware, all of the published reports on the biochemical properties of pol in vitro were performed in the presence of 5 or 8 mM Mg 2ϩ and were therefore performed under suboptimal conditions (20 -24, 28, 30, 31, 35, 38 -42). Indeed, it appears that in the presence of low levels of Mg 2ϩ , pol may exhibit higher catalytic activity and fidelity than previously thought (Table 1).
Limited synthesis by pol , pol , and pol was observed in the presence of many different divalent cations, but in most cases, the concentration required for replication occurred well outside the physiological concentration of the trace metal, and is therefore probably of little biological significance. The exception was the ability of pol to utilize low levels of Mn 2ϩ as an activator for catalysis (Fig. 2). The cellular concentration of Mn 2ϩ in mammalian cells is much lower than that of Mg 2ϩ and is thought to be in the range of 0.1 to 40 M (43-45). Our current studies indicate that in vitro, the optimum concentration of Mn 2ϩ required for maximal stimulation of pol on an undamaged template occurs around 0.075 mM (75 M). Although this concentration may lie outside the physiological range of Mn 2ϩ available to most cells, pol nevertheless exhibited a strong preference for Mn 2ϩ over Mg 2ϩ , even when Mg 2ϩ was in a 10 -20-fold excess. Our steady-state kinetic analyses revealed that Mn 2ϩ stimulated the catalytic activity of pol through a dramatic decrease in the K m value for nucleotide incorporation. Interestingly, when both Mn 2ϩ and Mg 2ϩ were present at optimal concentrations (0.075 and 0.25 mM, respectively), pol exhibited kinetic parameters that more closely resembled those observed in the presence of Mn 2ϩ alone ( Table  1, part C), which infers that Mn 2ϩ is the preferred activator for polymerization.
The structural basis for the enhanced catalytic activity of pol in the presence of Mn 2ϩ is presently unknown. However, it is clear that Mn 2ϩ has a more relaxed coordination than Mg 2ϩ , thereby facilitating reactions that are suboptimal or even inhibited in the presence of Mg 2ϩ (5). Indeed, the crystal structure of pol reveals that its active site is somewhat distorted and has large side chains protruding into the space normally occupied by the replicating base pair (40). Mn 2ϩ may therefore simply allow greater flexibility of the active site of pol , such that it can adopt a conformation that is more favorable for catalysis.
The effects of Mn 2ϩ on the fidelity of pol were unique. Although it caused a characteristic reduction in fidelity at template A, Mn 2ϩ actually increased fidelity at template T, by 3-5fold compared with the fidelity of pol in the presence of 0.25 or 5 mM Mg 2ϩ , respectively. Low levels of Mg 2ϩ and Mn 2ϩ also had dramatic effects on the ability of pol to traverse a T-T CPD. The efficiency of bypass increased from ϳ2.5% in the presence of 5 mM Mg 2ϩ to ϳ8% in the presence of 0.5 mM Mg 2ϩ and up to ϳ60% in the presence of 0.2 mM Mn 2ϩ . Thus, under certain conditions in vitro pol can bypass a T-T CPD relatively efficiently. Presumably conditions favorable for the pol -dependent bypass of CPDs also occur in vivo. Such an assumption is supported by the recent reports that implicate cellular roles for pol in UV-induced mutagenesis and carcinogenesis in mice and humans (46 -48). The stimulatory effect of Mn 2ϩ on pol -dependent lesion bypass was not limited to the T-T CPD, as bypass of an abasic site increased from ϳ1 to 2% in the presence of Mg 2ϩ to ϳ18% in the presence of Mn 2ϩ . Similarly, bypass of a BPDE lesion also increased from ϳ2.5 to ϳ10% in the presence of Mg 2ϩ and Mn 2ϩ , respectively, thereby raising the possibility that pol may also bypass these lesions in vivo.
In summary, pol is more active in the presence of physiological concentrations of Mg 2ϩ than at concentrations previously used in vitro to study the enzymatic properties of the polymerase. We therefore suggest that future in vitro studies on pol be conducted in buffers containing low Mg 2ϩ (0.25-2 mM). We also discovered that the enzyme is highly active in low concentrations of Mn 2ϩ . When asked to select between Mg 2ϩ and Mn 2ϩ , pol clearly prefers to utilize Mn 2ϩ , even when Mg 2ϩ is in large molar excess. Whether a similar situation occurs in vivo remains to be determined, but it nevertheless remains an intriguing possibility certainly worth considering.