Enzymatic bypass of an N6-deoxyadenosine DNA–ethylene dibromide–peptide cross-link by translesion DNA polymerases

Unrepaired DNA–protein cross-links, due to their bulky nature, can stall replication forks and result in genome instability. Large DNA–protein cross-links can be cleaved into DNA–peptide cross-links, but the extent to which these smaller fragments disrupt normal replication is not clear. Ethylene dibromide (1,2-dibromoethane) is a known carcinogen that can cross-link the repair protein O6-alkylguanine-DNA alkyltransferase (AGT) to the N6 position of deoxyadenosine (dA) in DNA, as well as four other positions in DNA. We investigated the effect of a 15-mer peptide from the active site of AGT, cross-linked to the N6 position of dA, on DNA replication by human translesion synthesis DNA polymerases (Pols) η, ⍳, and κ. The peptide–DNA cross-link was bypassed by the three polymerases at different rates. In steady-state kinetics, the specificity constant (kcat/Km) for incorporation of the correct nucleotide opposite to the adduct decreased by 220-fold with Pol κ, tenfold with pol η, and not at all with Pol ⍳. Pol η incorporated all four nucleotides across from the lesion, with the preference dT > dC > dA > dG, while Pol ⍳ and κ only incorporated the correct nucleotide. However, LC-MS/MS analysis of the primer-template extension product revealed error-free bypass of the cross-linked 15-mer peptide by Pol η. We conclude that a bulky 15-mer peptide cross-linked to the N6 position of dA can retard polymerization and cause miscoding but that overall fidelity is not compromised because only correct pairs are extended.

The preservation of genome integrity is vital for the proper development of an organism. Cells are subjected to multiple endogenous and exogenous agents capable of causing lesions and affect multiple DNA transaction processes (e.g., replication, repair, and transcription) (1). Covalent DNA-protein cross-links are bulky lesions and can be toxic if left unrepaired (2); they are formed from both exogenous and endogenous sources. Accumulation of these cross-links has been associated with aging, cancer, neurodegeneration, and Ruijs-Aalfs syndrome (2)(3)(4)(5)(6), and there has been considerable interest in both DNA-protein cross-links and proteases that can act on them (6). Several laboratories have shown that both reversible and irreversible DNA-protein cross-links can be formed under physiological conditions, that there are DNA-stimulated proteases that can act on these, and that DNA-protein cross-links can be bypassed and can miscode (2,(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19). The number of cross-links in a cell has been estimated to be high (20) but the exact number is not known. There is now evidence that DNA-protein adducts may be important in disease states, e.g., cross-linking was reported to be increased following ischemic reperfusion in cardiomyocytes (21). The list of cross-linked sites includes multiple DNA bases (G, C, T, A) and their modifications (e.g., N 7 -Me G (12), abasic sites (13), and 5-formyl dC (10,11)). The list of proteins in the cross-links includes histones (10,12), HMCES (13), and numerous other proteins (21)(22)(23)(24)(25). Some reversible lysine cross-links (Schiff bases) can also destabilize DNA and cause cleavage (12,26). However, the existence of a DNA-protein cross-link is not a priori evidence for miscoding (22,23).
Some DNA-protein cross-links are common, and cells have specific enzymes to act on them. One example is tyrosyl-DNA phosphodiesterases (TDP), which are enzymes capable of breaking down the covalent bond between DNA and DNA topoisomerases (27). The activity of these enzymes is limited by substrate accessibility, suggesting that the adduct needs first to be hydrolyzed to a peptide (28). Recently DNA-activated proteases (6) and the proteasome have been suggested to be involved in proteolysis by cleaving large DNA-protein cross-links to DNA-peptide cross-links (18,29,30). Some DNA-dependent proteases bind to ubiquitin, or small ubiquitin-like modifier (SUMO), indicative of a role for posttranslational modification in proteolysis (31)(32)(33). Ubiquitination is a key step for proteolysis by the nuclear DNA-dependent metalloprotease SPRTN (18). SPRTN uses an accessory process (i.e., ubiquitination) to repair cross-links because the ubiquitin controls its activity (34). SPRTN-mediated proteolysis also depends on the location of protein cross-links on the DNA strands (30). Thus, SPRTN has various limitations with respect to DNAprotein cross-link proteolysis. FAM111A is also known to act on DNA-protein cross-links and contains a trypsin-like domain (35).
Repair mechanisms can target different components of a DNA-protein cross-link (i.e., the DNA molecule, the cross-link bond, or protein) and may depend on the stage of the cell cycle (27,28,36). Most DNA-protein cross-links are too large for DNA polymerase bypass (37), and proteases are believed to cleave the proteins to smaller peptides that can be repaired by nucleotide excision repair (NER) (38) or homologous recombination (HR) or bypassed by translesion synthesis (TLS) polymerases (39). NER and HR have been shown to remove protein cross-links smaller than 10 kDa (40). Mammalian TLS polymerases include Y-family Pol η, ⍳, and κ and Rev1 (41), which have low processivity and fidelity (39). Only in a few cases has the mutagenic potential of a DNApeptide cross-link been analyzed (42), and essentially all of the work in this area has been done with either uncharacterized cross-links or synthetic models.
O 6 -Alkylguanine DNA-alkyltransferase (AGT, or MGMT), a 22 kDa protein, is a crucial repair system for protection from alkylating agents (43). It directly removes and transfers alkyl groups from the O6 position on guanine to the Cys-145 located in its active site. Ethylene dibromide (EDB, or 1,2dibromoethane) is a known carcinogen that had formerly been used as a gasoline additive, soil fumigant, and pesticide (44)(45)(46)(47). Overexpression of AGT paradoxically increased EDB toxicity and the incidence of base pair mutations in bacterial and mammalian cells (48,49). Cys-145 of AGT reacts with EDB via a nucleophilic substitution that results in a halfmustard intermediate, which cyclizes to form an unstable episulfonium ion that can form a covalent bond with DNA ( Fig. 1). The AGT-EDB episulfonium ion can react with nucleobases at different positions, including the N7, N2, N1, and O6 atoms of dG and the N6 atom of dA (50,51). The DNA-binding properties of AGT facilitate formation of these specific lesions. EDB is also known to cross-link DNA with the tripeptide glutathione (GSH) in a similar mechanism (52). Thus, this DNA-protein cross-link is directly relevant to a practical problem in toxicology, and other 1,2-dihaloalkanes and other bis-functional electrophiles are also relevant (53,54).
We developed a procedure for the synthesis and characterization of a DNA-peptide cross-link at the N6 position of dA with a 15-mer peptide from the active site of AGT in order to investigate its effect on DNA replication by TLS polymerases. Full-length extension and single nucleotide insertion assays, steady-state kinetics, and LC-ESI-MS/MS analysis were performed to determine the efficiency and fidelity at the N 6 -dA-adducted peptide.

Results
Synthesis and characterization of a 15-mer peptide cross-linked to an oligonucleotide A 15-mer peptide cross-linked to an oligonucleotide (N 6 -dA) was synthesized using the 15-mer peptide (acyl-PVPI-LIPCHRVVSSS-amide, AGT residues 138-152, with Cys-150 changed to Ser to permit exclusive modification at Cys-145) and a 6-chloropurine-containing oligonucleotide as outlined in Figure 2 (Scheme S1-S2, See Table S1 for oligonucleotide sequences). The peptide was treated with O-(mesitylsulfonyl) hydroxylamine (MSH), yielding amination of the cysteine to produce a dehydroalanine (dha) residue in its place (Scheme S1). The dha peptide was purified by HPLC (Fig. S1) and  Table S2). The 6-chloropurine-containing oligonucleotide was subjected to nucleophilic substitution with cystamine (Scheme S2), and the cystamine-containing oligonucleotide was purified by HPLC ( Fig. S4) and characterized by MALDI MS (Fig. S5).
Reduction of the cystamine-containing oligonucleotide with DTT yielded N 6 -(2-thioethyl)dA in the oligonucleotide, used for the next step without any purification (Fig. 2). Finally, the dha peptide was coupled with the thiol-containing oligonucleotide to obtain a 15-mer peptide cross-linked at the N6 position of dA via a two-carbon linker (Fig. 2). Alternatively, the 15-mer peptide cross-linked to DNA was also synthesized by a previously reported method (55).
The DNA-peptide cross-link was purified by gel electrophoresis (Fig. S6) and characterized by MALDI MS as well as ESI-MS (Figs. 3 and S7). This structure is analogous to the AGT-DNA cross-link induced by EDB (50), but an oxidized sulfur-bearing adduct was observed with an additional mass of 32 a. m. u. (Figs. 2, 3, and S7) which was characterized using LC-MS/MS analysis. The DNA moiety was hydrolyzed with HF to obtain a peptide adducted with adenine (Fig. S8). In mass spectral analysis (positive mode), two major m/z ions were observed, m/z 613.66 (+3) and 919.99 (+2) (Fig. S9), indicative of an oxidized sulfur atom, which was further confirmed from the fragmentation pattern ( Fig. S10 and Table S3).   Bypass of 15-mer peptide-N 6 -dA oligonucleotide cross-link by TLS hPols η, ι, and κ Full-length extension and single nucleotide insertion assays were carried out with TLS polymerases. Full-length primer extension reactions ("running start") were performed in the presence of all four dNTPs using a 12-mer primer (Fig. 4A). The N 6 -dA-peptide lesion in the template affected each polymerase differently. With the DNA-peptide cross-link, hPol η fully extended the primer with almost similar efficiency as the unmodified template (Fig. 4B, lanes 2-6 and 8-12). hPol ι mainly produced a single nucleotide incorporation product (Fig. 4C, lanes 20-24), similar to the unmodified template. The lesion also affected hPol κ activity, and a two-nucleotide incorporated product (stalled at lesion site) was produced along with a small amount of fully extended product (Fig. 4D, lanes 32-36).
The specificity constant (k cat /K m ) for the insertion of correct nucleotide (i.e., dTTP) opposite to the adduct by hPol η was 1.4 ± 0.2 μM −1 min −1 (Table 1), tenfold lower than for insertion across the unmodified template. However, the presence of the adduct did not greatly affect the specificity constant of hPol ɩ-specificity constants of dTTP insertion across the adduct and the control template were 0.28 ± 0.16 and 0.23 ± 0.08 μM −1 min −1 (Table 1), respectively. The specificity constant for incorporation by hPol κ was 220-fold lower than the control. hPol η had the highest specificity constant for insertion of dTTP across the lesion and inserted all four nucleotides. The specificity constants for dATP, dCTP, and dGTP insertion across from the adducted peptide were decreased 3.5-, 1.2-, and 2.8-fold, compared with the unmodified template.
hPol η-mediated primer extension analysis using LC-ESI-MS/MS Steady-state kinetic analysis showed insertion of all dNTPs opposite to the adducted peptide by hPol η as compared with hPol ⍳ and κ (Table 1), and mass spectral analysis was performed to determine if other miscoding events occurred with hPol η (56). For full-length extension reactions, a 2 0 -deoxyuridine (dU)-containing 12-mer primer was used. After fulllength extension, the reaction mixture was treated with hPol hPol hPol  Table 2.
hPol η yielded error-free products for the unmodified as well as modified template-primer complex with or without blunt end additions of A and G. For the unmodified templateprimer complex, the major m/z 1048.64 (−3) and 1053.91 (−3) ions were observed (Figs. S13 and S14; Tables S4 and S5). For the modified template-primer complex, the major m/z 944.27 (−3) ion observed indicates insertion of the correct base T (Fig. 7, S15 and Table S6). In addition to this, m/z 1048.64 (−3) and 1053.82 (−3) were also observed, indicating insertion of the correct base T with blunt end addition of A, as well as G (Figs. S16 and S17; Tables S7 and S8). Overall, mass spectral  11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 dATP  Table 1 Steady-state kinetic analysis of insertion opposite to N 6 -dA-cross-linked 15-mer peptide  analysis revealed only error-free bypass of the adducted peptide.
Discussion DNA-protein cross-links have been shown to be formed from both exogenous and endogenous sources. Proteinchemical-DNA cross-links are generally considered to be too large for DNA polymerase action (37), and it has been proposed that proteases cleave the protein to peptides. Walter's group reported the "repair" of an artificial DNA-protein crosslink by replication-coupled proteolysis in Xenopus egg extracts (57). DNA-dependent proteases have been identified in yeast (Wss1, Ddi1) (58), and mammalian SPRTN has been characterized (4, 16-18, 29, 59), with all three of these proteases acting on DNA-protein cross-links. These systems are operative in cultured human cells (8,9), yeast (58), amphibians (18,57), and plants (19). In addition to FAM111A, another mammalian protein that interacts with SPRTN (GCNA) has recently been identified, although protease activity has not been demonstrated (60,61). Degradation by the proteasome has been presented as an alternative to direct protease action (18,62). However, this process may require removal (from DNA) by nucleases (19,36) and replication proceeding following the latter steps of nucleotide repair, although nuclear proteasomes and ubiquitination are known and direct proteolysis of the protein in a DNA-protein cross-link cannot be ruled out (63,64). Human diseases have been associated with defective DNA-protein cross-link proteases (6), e.g., gracile bone dysplasia (GCLEB) and Kenny-Caffey syndrome type 2 are linked to mutations in FAM111A (65,66), pediatric germ cell tumors have been associated with mutations in GCNA (60), and Ruijs-Aalfs syndrome has been linked to mutations in SPRTN (67,68). DNA-peptide cross-links can be bypassed by TLS polymerases. TLS polymerases have been shown to bypass a GSHbased adduct (52,69,70) and (N-terminal-linked) peptides as long as 5-and 12-mers (71-73) but not a 23-mer peptide or green fluorescent protein (72). In vitro studies, including some in cells, have shown the mutagenic bypass of DNA-peptide cross-link by TLS polymerases (11,42,74). It is important to know how a DNA polymerase can bypass each cross-linked peptide entity and potentially cause base-pair mutations. As emphasized earlier, the cross-linking of a protein to DNA is not necessarily related to mutation (22,23), and exactly why some cross-links are associated with miscoding but others are not is not understood. The reason could be lack of degradation or the inherent miscoding potential of the peptides resulting from proteolysis.
EDB induces AGT-DNA cross-links in cellular environments ( Fig. 1) (48-51, 75). The possibility exists that proteolytic degradation of AGT-DNA cross-links occurs under cellular conditions by DNA-dependent proteases as well as the proteasome. Therefore, we were interested in understanding the biological consequences of peptide-DNA cross-link generated from the active site of AGT, the known major site of reaction (51). Following preliminary studies indicating the bypass of GSH (3-mer) and shorter AGT-derived peptides (3-, 7-, 11-mer), we checked and focused on a 15-mer peptide cross-linked to DNA for further studies. The chemical synthesis of the DNA-peptide cross-link was achieved by coupling (Michael addition) between the dha peptide and a thiol-containing oligonucleotide (Fig. 2). An oxidized DNApeptide cross-linked product was observed. We were unsuccessful in blocking its formation, so enzymatic and mass spectral studies were performed with this oxidized form of the DNA-peptide cross-link (Fig. 2). The same scenario was observed previously in the analysis of the DNA-AGT crosslink. Proteomic analysis of AGT-N 6 -DNA cross-link indicated the presence of a singly oxidized sulfur during in-gel (6.5% of oxidized product) as well as in-solution (0.8% of oxidized product) digestion by trypsin (data not shown) (43). We had not observed the presence of sulfoxides or sulfones in our analysis of AGT tryptic peptides cross-linked to DNA in previous work (50,51), but these assays were done more rapidly than the synthetic work described here and we had set up our searches based only on the thioether product.
The full-length extension assays showed that hPol η tolerated the 15-mer peptide cross-linked to DNA, as compared with hPol ⍳ and κ (Fig. 4). Single nucleotide insertion assays indicated that only hPol η had high levels of misincorporation of each dNTP opposite to unmodified template. hPol η mediated low levels of misincorporation of all three dNTPs across from the adducted site but hPol ⍳ and κ showed only high-fidelity bypass of the adduct (Fig. 5). These results are evidence that the outcomes are not related to the oxidation of the peptide-DNA cross-link, which we have localized to the sulfur atom and not some unusual modification of the purine or deoxyribose ring.
When we compared specificity constants for incorporation of each dNTP opposite to an unmodified template, hPol η inserted dATP 28-fold, dCTP 36-fold, and dGTP 67-fold less efficiently than dTTP (Figs. 6 and S11; Table 1). Across from  The bold and underlined T is the base inserted opposite the adduct X.
Bypass of DNA-protein cross-link the peptide-adducted site, a similar trend was seen but with lower specificity constants (Table 1): hPol η inserted dATP 10fold, dCTP 4.3-fold, and dGTP 19-fold less efficiently than dTTP. Overall, the misincorporation frequency was 37%. LC-MS/MS analysis of hPol η bypass of the 15-mer peptide cross-linked to DNA indicated that the major product resulting from insertion had only the correct base T opposite to the lesion site (with or without blunt end addition) ( Table 2). These findings are consistent with previous results showing efficient bypass of adducts in the major groove of DNA by hPol η (76)(77)(78). We conclude that other bases (A, C, G) can be inserted opposite to the adduct (Table 1) but that hPol η is not able to extend the primer to obtain full-length products. Whether or not these truncated products could be lethal to a cell is unknown at this point.
In conclusion, the steady-state kinetics showed misinsertion by hPol η while mass spectral analysis revealed only error-free product while bypassing the 15-mer peptide cross-linked to DNA. Previous studies with a number of N 6 -dA adducts (e.g., N 6 -oxopropenyl-dA, N 6 -carboxymethyl-dA, N 6 -(2-hydroxy-3butan-1,4-diyl)-dA) have not reported any mutagenic events (79)(80)(81). In the case of mutations arising from EDB, most are G to A transitions, and the numbers of A mutations we have observed in Escherichia coli (51, 82) yeast (83) or mouse liver (75) are too small, relative to background, to accurately assign a prominent base-pair substitution. However, hPol κ inserted C opposite to an N 6 -A dihydroxybutyl GSH adduct formed from diepoxybutane (70). The EDB adduct formed with the tripeptide GSH showed misincorporation of dCTP (6%) and a −1 frameshift (3%) in the LC-MS/MS analysis, as well as incorporation of mainly dGTP but also some dATP and dCTP (total 30%) in steady-state kinetic analysis (52). Thus, it appears that hPol η and (possibly other DNA polymerases) are capable of inserting bases opposite sites of bound peptides but the size of the peptide may be an issue in extension of mispairs (but not correct matches). Further investigation of this general hypothesis is in order.

Synthesis of MSH (87)
Ethyl hydroxamate (23 mmol) was stirred in a 2:1 (v/v) mixture of DMF and triethylamine (9 ml) and cooled to 0 C. Mesitylene sulfonyl chloride (23 mmol) was slowly added, and the reaction was stirred for an additional 15 min before adding CH 2 Cl 2 (100 ml). The reaction product was washed ten times with H 2 O, and the organic layer was collected and dried using brine and MgSO 4 before removing the solvent in vacuo. The reaction product was redissolved in dioxane (4 ml) and cooled to 0 C before adding 70% HClO 4 dropwise over 2 min (1.8 ml). The solution was stirred for an additional 2 min and transferred into chilled H 2 O (200 ml) prior to extraction with Bypass of DNA-protein cross-link diethyl ether (100 ml). The organic layer was neutralized, dried using K 2 CO 3 , and filtered. The filtrate was concentrated, and chilled hexane (150 ml) was added to crystalize the product overnight at −20 C. MSH crystals were collected, dried in a vacuum desiccator, and stored at −20 C.
Synthesis of 15-mer dehydroalanine peptide (dha peptide) K 2 CO 3 (9 mg) was dissolved in 100 μl of H 2 O and added to the 15-mer peptide (acyl-PVPILIPCHRVVSSS-amide). MSH (3.8 mg) was dissolved in 100 μl of anhydrous DMF and added dropwise into the peptide solution followed by incubation on ice for 20 min. After completion of the reaction, the reaction mixture was diluted in 800 μl of Mobile Phase A (95% H 2 O, 5% CH 3 CN, and 0.1% HCO 2 H, v/v/v) and purified by HPLC using a Phenomenex semipreparative octadecylsilane (C 18

Synthesis of N 6 -cystamine-dA oligonucleotide
The controlled pore glass (CPG, 1 μmol) of the 19-mer oligonucleotide (5 0 -TCTCXGTTTATGGACCACC-3 0 , where X is 6-Cl purine) was treated with cystamine⋅2HCl [10 μl of 500 mM stock in dimethylsulfoxide (DMSO)] in a mixture of DMSO (86 μl) and diisopropylethylamine (4.4 μl) to obtain the N 6 -cystamine-dA modified oligonucleotide. After incubation at 55 C for 22 h, the supernatant was removed. The oligonucleotide (still attached to the glass support) was washed with anhydrous DMSO (2 × 1 ml) and anhydrous CH 3 CN (3 × 1 ml) and then air-dried. The oligonucleotide was subjected to final deprotection using 0.4 M NaOH in CH 3 OH (1 ml) at room temperature for 17 h, with shaking. The oligonucleotide support was sonicated for 5 min, and the supernatant was collected in an Eppendorf tube. The oligonucleotide support was washed with H 2 O (2 × 400 μl), and the supernatant was collected in the same tube. The collected supernatant was neutralized to pH 7.0 using glacial CH 3 CO 2 H and concentrated in vacuo using a centrifugal evaporator. The dried pellet was resuspended in H 2 O (500 μl) for HPLC purification. Purification was carried out using a Phenomenex Clarity Oligo RP (C 18 (Fig. S4). The oligonucleotide fraction was collected and concentrated in vacuo with a centrifugal evaporator and resuspended in 10 mM Tris-HCl buffer (pH 8.0), 1 mM EDTA, and 300 mM NaCl. The modified oligonucleotide was desalted using C 18 Sep-Pak, and the mass was confirmed by MALDI-TOF (Fig. S5).

Synthesis of 15-mer peptide crosslinked at N 6 -dA position
To the N 6 -cystamine-dA oligonucleotide (3 nmol, in 44 μl nuclease free water), 10 mM potassium phosphate buffer (pH 8.0) and 2 mM DTT were added to obtain an N 6 -(2-thioethyl)-dA modified oligonucleotide (Scheme S2). The reaction was incubated at 95 C for 5 min and cooled to room temperature. The reaction mixture was desalted by passing through a Micro Bio-Spin chromatography column following the manufacturer's protocol. The dha peptide (1 mg) was dissolved in 10% CH 3 CN (v/v, 40 μl). The peptide solution was added to N 6 -(2thioethyl)-dA modified oligonucleotide (Fig. 2), and reaction mixture was incubated at 37 C for overnight. Details of the alternative synthetic route can be found in the reported procedure (55). The cross-linking reaction mixture was purified using gel electrophoresis (Fig. S6). The desired cross-link band was visualized using a UV lamp, cut out of the gel, and extracted using 10 mM Tris-HCl buffer (pH 8.0) containing 1 mM EDTA and 300 mM NaCl. The DNA-peptide cross-link was desalted using a C 18 Sep-Pak, and the mass was confirmed by MALDI as well as negative ion ESI-MS (Figs. 3 and S7). An additional mass of 32 a. m. u. was found.
The sample was further analyzed following hydrolysis of the DNA entity with HF to obtain a peptide adducted with adenine (Fig. S8). The cross-link (100 pmol) was dried and suspended in HF (48%, 50 μl), and incubated at 4 C for 14 h (43). The sample was dried under a stream of nitrogen, resuspended in anhydrous CH 3 OH (50 μl), and dried again under a stream of nitrogen. Finally, the digested cross-link was dissolved in 0.1% HCO 2 H (20 μl, v/v) and shaken for 10 min. Following centrifugation for 5 min at 21,000g at room temperature, the crude sample was directly (without any purification) analyzed in the positive mode using nano LC-ESI-MS (Figs. S9 and S10, Table S3). The fragmentation is consistent with the additional mass of 32 a. m. u. being on the linker between the peptide and the oligonucleotide.

Steady-state kinetics
A FAM-labeled 14-mer primer (5 0 -FAM-GGTGGTCCA-TAAAC-3 0 ) and a 19-mer template (5 0 -TCTCXGTTTATG-GACCACC-3 0 , where X is dA or N 6 -dA-peptide) were annealed (1:1 M ratio) at 95 C for 5 min and cooled to room temperature overnight. Reactions were carried out using 40 mM Tris-HCl buffer (pH 7.5) containing 5% glycerol (v/v), 10 mM DTT, 100 mM KCl, 5 mM MgCl 2 , and 0.1 mg ml −1 BSA. Varying concentrations of each dNTP were used to achieve ≤20% of product formation, with up to 11 different concentrations of each dNTP were used. Reactions were started by adding 1 μl of individual dNTP stock solutions. Aliquots of 3 μl were removed at different time points (depending on the polymerase, template, and dNTP) and quenched with 6.5 μl of 10 mM EDTA (pH 8.0) in 95% deionized formamide. Products were separated using PAGE (Figs. 6 and S11), and results were analyzed by ImageJ software by visualizing on a Typhoon scanner. Reactions were done in duplicates. Data points are shown as means (±SD) and estimated using fit to a hyperbolic equation in Prism (GraphPad, San Diego, CA) software to calculate the k cat and k cat /K m (k sp ), then deriving K m (Table 1) (88).

Data availability
All data are included in the article and supporting information.
Supporting information-This article contains supporting information.