Hydrogen-Bonding Interactions at the DNA Terminus Promote

Subscriber access provided by Washington University | Libraries

Article

Hydrogen bonding interactions at DNA terminus promote extension from methylguanine lesions by human extender DNA polymerase # Michael H. Räz, Shana J Sturla, and Hailey L. Gahlon Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.8b00861 • Publication Date (Web): 17 Sep 2018 Downloaded from http://pubs.acs.org on September 18, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Hydrogen bonding interactions at DNA terminus promote extension from methylguanine lesions by human extender DNA polymerase ζ Michael H. Räz, Shana J. Sturla, and Hailey L. Gahlon Department of Health Sciences and Technology, ETH Zürich, Schmelzbergstrasse 9, Zürich, 8092, Switzerland ABSTRACT: Chemically-induced DNA lesions can become DNA replication substrates that are bypassed by low-fidelity DNA polymerases. Following nucleotide misinsertion opposite a DNA lesion, the extension step can contribute to preserving such errors and lead to genomic instability and cancer. DNA polymerase ζ, a B-family polymerase, is proficient as an extender polymerase that catalyzes elongation, however, it is not understood what chemical factors impact its DNA replication. This study addresses the question of how DNA polymerase ζ achieves extension by examining the ability of recombinant human DNA polymerase ζ to extend from a series of methylated guanine lesions. The influence of Hbonding was examined by placing structurally altered nucleoside analogs and canonical bases opposite G, O6-MeG, N1MeG, and N2-MeG. We determined that terminal base pairs with the highest proclivity for H-bonding were most efficiently extended in both primer extension assays and steady-state kinetic analysis. In contrast, when no H-bonding was possible at the DNA terminus, the least efficient steady-state kinetics were observed. To evaluate H-bonding protein minor-groove interactions that may underlie this phenomenon, we performed computational modeling with E. coli DNA polymerase II, a homolog for DNA polymerase ζ. The modeling data together with the primer extension assays demonstrates the importance of having a carbonyl group on the primer strand that can interact with a lysine residue found to be conserved in many B-family polymerases, including human Pol ζ. These data provide a model whereby interbase Hbonding interactions at the DNA terminus promote lesion bypass and extension by human DNA polymerase ζ.

INTRODUCTION Exogenous and endogenous methylating agents can react with DNA to generate methyl-DNA adducts. For example, exposure to exogenous alkylating agents like Nmethyl-N-nitrosoguanidine and N-nitrosodimethylamine has been implicated in chemical carcinogenesis involving DNA damage by a reactive diazonium ion.1-2 In addition, the non-enzymatic transfer of the methyl group from Sadenosylmethionine to DNA is regarded as an endogenous methylation process introducing methylation adducts.3-4 If not repaired, DNA methylation damage can cause misinsertion of nucleotides or blockage of DNA replication by DNA polymerases (Pols) and lead to mutations and carcinogenesis. Translesion DNA synthesis (TLS) is a process that involves replication past sites of DNA damage and ensures the continuation of DNA synthesis but, carried out with low-fidelity (e.g. error frequency 10-2 – 10-4)5, it comes at a potential cost of initiating the development of cancer.6 In contrast, replicative B-family Pols like Pol ε and Pol δ, have high fidelity (e.g. error frequency 10-6–10-8)7-8 in the synthesis of unmodified DNA, however, they are often ineffective and error-prone in the replication 0f methylated DNA. Pols tend to misinsert T opposite the highly mutagenic O6-methylguanine (O6-MeG) adduct, resulting in G to A transition mutations.9-10 The O6-MeG:T base pair retains a pseudo-Watson-Crick geometry more closely than does O6-MeG:C,11 which forms the basis for the mis-

coding properties of O6-MeG. In addition, N2methylguanine (N2-MeG) adducts have also been shown to be miscoding and potentially responsible for G to A transition mutations.12 Replicative Pols are often hindered in bypassing DNA adducts, thus in order to avoid replication fork collapse and chromosomal instability, cells employ TLS to overcome DNA adducts. Although TLS is often performed by Y-family Pols with larger, more solvent exposed active sites,6 also Pol ζ, a Bfamily polymerase, prevents DNA damaged-induced replication fork stalling. Pol ζ has a similar fidelity (e.g. 10-5)13 as replicative Pols and the capacity to catalyze the insertion step for TLS in the bypass of thymine glycol lesions, cis-syn thymine-thymine dimers and cisplatin adducts,14-15 however, it is most notably invoked in extension. For example, yeast Pol ζ does not bypass the 3’-thymine of the UV-induced 6-4 photoproduct, but once a base is inserted by another polymerase, it bypasses the 5’-thymine.16 In the case of abasic sites, insertion is performed by DNA Pols δ, Rev1 and Pol η, whereas yeast Pol ζ performs the extension step.17 In a study concerning 8-oxoguanine and O6-MeG, yeast Pol ζ was inefficient at inserting a nucleotide opposite either lesion, but after Pol δ catalyzed the nucleotide insertion step it could perform extension.18 Pol ζ is implicated in TLS past methyl-DNA adducts. For example, DNA Pol ζ has been shown to contribute to drug resistance of mouse embryonic fibroblasts to temozolomide, a methylating agent.19 It was suggested that Rev3L may be important in the tolerance of N-alkylation

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and/or abasic sites, however, the absence of Rev3L did not sensitize cells to O6-MeG.19 In contrast, work with yeast DNA Pol ζ has shown an increased proclivity for extension from O6-MeG, suggesting the potential of Pol ζ to be involved in replication past this DNA lesion.18 Further, results of a recent study with cells deficient in human Pol ζ, suggested that it is required for the faithful processing of the O6-alkylG adduct O6-carboxymethylguanine (O6CMG).20 Thus, the role of human DNA Pol ζ in the bypass of alkylation-derived lesions and in particular methylated adducts, such as O6, N1 and N2-methylguanine remains unclear. Due to the high mutagenic potential of methylation adducts, understanding the ability for human Pol ζ to process these adducts is important to realize their role in the chemical basis of mutation.

Figure 1. Experimental strategy to evaluate DNA Pol ζ exten32 sion past methylated guanine DNA lesions. (A) A 5’- P labeled 24-mer primer (N = Cytosine, Uracil, Benzi, HBT, BIM and indole) is annealed opposite a 28-mer template (X = G, 6 1 2 O -MeG, N -MeG and N -MeG). (B) Methylation positions for 6 1 2 guanine are depicted at the O , N and N atoms. (C) Structures of the canonical and nucleoside probes situated at the 3’-end of the primer strand. Groups 1-3 are categorized based on the maximum number of H-bonds possible at the terminal base pair. Here, Cytosine and Uracil (Group 1) can form up to 3 hydrogen bonds, Benzi and HBT (Group 2) can form up to 2 hydrogen bonds and BIM and indole (Group 3) have the capacity for 1 or no hydrogen bond, respectively.

Chemical determinants for correct dNTP selection vary amongst polymerase families.21-24 Replicative polymerases, such as T7 or E. coli Pol I, employ a nucleotide selection that depends on active site tightness.25 This selection requires a high degree of geometric complementarity for the incoming dNTP within the active site of the enzyme, rather then depending on H-bonding interactions.25-26 For TLS polymerases, however, H-bonding seems to be more influencial than geometric fit, which could be due to their larger active sites. It was reported that TLS polymerases η and κ replicate with lower efficiency with nucleotides that lack H-bonding, suggesting that H-bonding promotes

Page 2 of 13

efficient DNA synthesis.23-24, 27 Recent computational data further supports the importance of H-bonding interactions in the bypass of O6-alkylG adducts28 or non-natural nucleobases29 by the Y-family DNA polymerase Dpo4. However, it remains unclear if efficient DNA extension during TLS by the B-family Pol ζ is dependent on geometric selection or H-bonding. Difficulties in the purification and characterization of mammalian Pol ζ, mainly due to the large size and low expression levels, contributed to previously unavailable human DNA Pol ζ. Pol ζ has a catalytic- (Rev3) and an accessory subunit (Rev7). Human Pol ζ consists of four subunits Rev3, Rev7, PolD2, and PolD3. The catalytic subunit Rev3 contains over 3,000 amino acid residues and is twice as large as the yeast homolog. In 2014, Yang and co-workers reported the first purification of human DNA Pol ζ.30 By deleting internal regions of the REV3L gene, ~ 800 amino acids, they successfully expressed and purified human DNA Pol ζ.30 To date, the active four subunit complex has only been used in very few in vitro studies and was shown to be efficient in the bypass of a cisplatincross-link.30 In addition Pol ζ was able to extend from an O6-CMG adduct but unable to insert a nucleotide opposite the adduct.31 Thus, in the absence of structural data, there is a gap in understanding the adduct substrate scope of human Pol ζ and to what extent Pol-DNA-adduct interactions are affecting the activity of the Pol. In this study, we characterized the capacity of human DNA polymerase ζ to extend from nucleoside analogs as well as the canonical bases C and U paired opposite G and a series of methylated DNA adducts including O6-MeG, N1-MeG, and N2-MeG (Figure 1). With these structurally altered nucleoside analogs, we were able to systematically evaluate how H-bonding, base pair size and shape, and stacking interactions influence Pol ζ extension past methylated DNA damage. MATERIALS AND METHODS Materials and chemicals. [γ-32P]ATP was purchased from PerkinElmer Life Sciences (Waltham, MA). dNTPs were purchased from Invitrogen (Switzerland) and Bioconcept (Switzerland). T4 polynucleotide kinase was purchased from Promega (Madison, WI). Human DNA Pol ε was purchased from Enzymax (Lexington, KY). TrisHCl (pH 8.0 at 25°C), NaCl, MgCl2, dithiothreitol (DTT) and glycerol were purchased from Invitrogen (Switzerland). Bovine serum albumin (BSA) was purchased from New England Biolabs (Ipswich, MA). Sep-Pak® C18 classic cartridges were purchased from Waters. All other reagents were purchased from Sigma Aldrich (Switzerland) and used without further purification. The human DNA Pol ζ is the four-subunit-complex (Rev3, Rev7, PolD2 and PolD3) containing an internal deletion described previously (TR4-2 construct).30 The multi-subunit enzyme complex was expressed and purified by Trenzyme GmbH, Konstanz, Germany. Synthetic details and characterization of nucleoside intermediates are described in supplementary information.


Page 3 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Phosphoramidite synthesis. BIM, Benzi and the indole-phosphoramidite were synthesized following literature protocols.32-34 Synthetic methods and characterization for the preparation of the 2-hydroxybenzthiazol(HBT), the N1-methylguanine- (N1-MeG) and N2methlyguaninephosphoramidites (N2-MeG) are described in the supplementary information. Preparation of oligonucleotides. The sequences for the oligonucleotide templates were as follows: (28mer), 5’-ACT CXT CTC CCT ATA GTG AGT CGT ATT A-3’ where X = G, O6-MeG, N1-MeG and N2-MeG. For the Gand O6-MeG-containing templates, they were purchased from Eurogentec (Belgium). N1-MeG- and N2-MeGcontaining templates were synthesized on a Mermade 4 DNA synthesizer from Bio Automation Corporation (Plano, TX) using standard conditions and base-labile phosphoramidites (iPr-Pac-dG-CE, Ac-dC-CE, and dA-CE). The synthesis was performed in trityl-on mode and the oligonucleotides were post-synthetically purified on a Sep-Pak® C18 cartridge (www.waters.com) to remove truncated sequences. Elution of the desired oligonucleotide consisted of cleaving the DMT group and collection of the desired DNA. All templates were purified by reverse phase high-performance liquid chromatography (RP-HPLC) with a linear gradient from 9-13 % (v/v) acetonitrile in 50 mM triethylammonium acetate over 25 min. The sequences for the oligonucleotide primer strands were as follows: (24mer), 5’-TAA TAC GAC TCA CTA TAG GGA GAN-3’ where N indicates C, U, BIM, Benzi, HBT and indole. C- and U-containing primers were purchased from Eurogentec (Belgium). BIM and Benzi terminated primers were prepared as previously described. HBT- and indole-containing primers were synthesized on a Mermade 4 DNA synthesizer from Bio Automation Corporation (Plano, TX). The DNA synthesis was performed with Universal support III PS from Glen Research under identical conditions as described for template DNA. Primers were purified by polyacrylamide gel electrophoresis (PAGE). The presence of the desired DNA products was confirmed by mass spectrometry performed on an Agilent MSD ion trap mass spectrometer with electrospray ionization, operated in negative ion mode. The concentration was determined by UV spectroscopy at 260 nm on a NanoDrop 1000. Primer extension assays. T4 polynucleotide kinase and [γ-32P] ATP were used to label the primers strands at the 5' end. Labeled primers (N = C, U, BIM, Benzi, HBT, indole or 5-nitroindole) were annealed to complementary templates(X = G, O6-MeG, N1-MeG or N2-MeG). The annealing reactions were carried out by heating at 95 °C for five minutes and allowing to cool to room temperature for 3 hours. The primer extension reactions were carried out in a reaction buffer containing 20 mM Tris-HCl (pH 8.0, 25 °C), 100 mM NaCl, 5 mM MgCl2, 0.5 mM TCEP (pH 7.0, 25 °C), 100 µg/ml BSA and 3% glycerol. Further, 40 nM 5’[32P]primer-template DNA and 20 nM human DNA polymerase ζ and ε were used. The reaction was started by addition of 100 µM dNTP mix (25 µM of each canonical dNTP) to the sample containing the buffer, 5'-[32P]primer-

template DNA and human polymerase ζ. The reaction mixture was incubated for 25 minutes at 37 °C and the reaction was terminated by adding 10 µL of quenching solution (80 % formamide with 10 mM NaOH, 0.1 M EDTA and a trace of bromophenol blue and xylene cyanol). The resulting solution was loaded on a 15 % polyacrylamide 7 M urea gel, run for four hours at 1500 V and visualized with a phosphorimager (BioRad, Hercules, CA). Steady-state kinetics. To measure single-nucleotide incorporation steady-state kinetics, a 5’-[32P]labeled primer strand was annealed to the complementary template for the primer extension assays. For the enzymatic reactions, 5 nM human polymerase ζ was allowed to react with 50 nM 5’-[32P]primer-template DNA at 37 °C in a reaction buffer containing 20 mM Tris-HCl (pH 8.0, 25 °C), 100 mM NaCl, 5 mM MgCl2, 0.5 mM TCEP (pH 7.0, 25 °C), 100 µg/ml BSA and 3% glycerol and increasing dGTP concentrations. All the reactions were performed in replicates of three. Reactions were quenched at various time points in stop dye (described above). The aliquots were loaded on to a 15 % polyacrylamide 7 M urea gel and separated by electrophoresis (1500V, 4h). The gel was visualized with a phosphorimager (BioRad, Hercules CA) and the substrate and product/primer bands were quantified by densitometry using Quantity One software (BioRad). The nucleotide incorporation rate was plotted as a function of dGTP concentration and fit to Michaelis-Menten equation using GraphPad Prism 6 software. KM and kcat parameters were derived from the hyperbolic plots. In silico modeling experiments. Model structures were computed with Molecular Operating Environment (MOE) software suite (Chemical Computing Group Inc., Montreal, QC, Canada). A tertiary crystal structure of E.coli polymerase II (PDB: 3MAQ) with incoming ddGTP, a 17 mer template (5’-TACGTACGCTAGGCACA-3’) and a 13 mer primer (5’-GTGCCTAGCGTACdd-3’ where Cdd represents a dideoxynucleotide) was used for all modeling experiments. The crystal structures was modified with the MOE builder tool, to match the sequence context in our in vitro assays. The template was changed to 5’TACCXACGCTAGGCACA-3’ with X= G, O6-MeG, N1-MeG and N2-MeG and the primer to 5’-GTGCCTAGCGTNG-3’, with N= C, U, Benzi, HBT, BIM, or indole. The polymerase structure was prepared with the MOE QuickPrep function with default settings, including corrections of structural errors, addition of hydrogens, 3D optimization of Hbonding networks and deletion of water molecules further than 4.5 Å from the protein, and restrained minimization within 8 Å from the modified base pairs. For energy minimization, the Amber99:EHT force field was used. Graphical visualization of the modeling results was performed in PyMol software (Schrödinger, New York).

RESULTS To evaluate the ability of Pol ζ to elongate from a terminal G, O6-MeG, N1-MeG, and N2-MeG lesions paired opposite nucleotide probes with systematic variations in size, shape and H-bonding, three groups of nucleoside probes were employed. Group 1 (GR1, Figure 1) consisted


Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of the canonical bases Cytosine and Uracil. Group 2 (GR2, Figure 1) included the synthetic base surrogate Benzi, which has one H-bond donor and one acceptor, and HBT, which has a similar core structure as Benzi, but replaces an N-H donor with S, thus it has two H-bond acceptors. Group 3 (GR3, Figure 1) contained BIM, which has a single H-bond acceptor, and indole, which lacks the possibility to H-bond. GR2 probes are larger than the GR1 probes,

Page 4 of 13

but share similarities in their ability to form up to two hydrogen bonds. The GR2 probes contain a carbonyl group in an identical position as the O2 atom in pyrimidines as do GR1 probes. GR3 probes share similarities in their size and shape as GR2 probes, however, their capacity to form H-bonds is limited (i.e. BIM) or entirely excluded (i.e. indole), and they lack the minor groove carbonyl of the GR1 and GR2 probes.

6

1

2

Figure 2. Extension products observed from the synthesis of human DNA Pol ζ to elongate from G, O -MeG, N -MeG and N MeG-containing templates paired with canonical and artificial probes (see Figure 1 for structures). Below each gel, the percentage of primer extension is shown for each corresponding reaction. In total, 24 base pair variations were tested that contain different H-bond donor and acceptor functionalities. Below the gel images, renderings of the proposed structural pairing between 6 1 the base pairs at the primer-template terminus are shown. The templating base is shown on the outside left (i.e. G, O -MeG, N 2 MeG and N -MeG) with the corresponding base pair on the primer strand (i.e. Cytosine, Uracil, Benzi, HBT, BIM and indole) labeled above the gel images. Reactions were incubated at 37°C for 25 min in the presence of 10 nM human polymerase ζ, 100 μM dNTPs and 20 nM DNA.

Full-length extension by DNA Pol ζ from methylguanine adducts promoted with increasing number of H-bonds at the DNA terminus. The ability for DNA Pol ζ to extend from G, O6-MeG, N1-MeG and N2-MeG was evaluated by primer extension analysis (Figure 1A). The nucleoside analogs Benzi, HBT, BIM and indole and the canonical bases Cytosine and Uracil, were placed at the 3’ end of the primer strand. The canonical G, O6-MeG, N1MeG and N2-MeG were incorporated in a 28-mer template DNA and hybridized with a complementary primer (Figure 1A). Phosphoramidites of BIM, Benzi and the indole probe were prepared following literature protocols.32-34

HBT phosphoramidite was prepared as described in the materials and methods. Hybridization of the six different primer strands with the four complementary templates resulted in 24 permutations of DNA substrates (Figure 2). Throughout the manuscript, the shorthand notation N:X (e.g. BIM:G) will be used where the first base N originates from the primer strand and the second base X from the template strand (Figure 1A). DNA Pol ζ extended from each of the 24 DNA duplexes, however, to varying extents. The primers terminated with C, U, Benzi and HBT paired with G and modified G displayed a similar primer extension pattern and all four


Page 5 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

primers were extended to the full-length product (Figure 2). In contrast, DNA Pol ζ did not perform full-length extension from BIM and indole, rather a prominent n+1 primer extension band was observed. Full-length extension by DNA Pol ζ was observed for canonical C and U primers opposite G and modified G substrates. In case of C, full-length product bands were observed when paired with canonical G and methyl G substrates (Figure 2, lanes 1-4), however, the least extension was for C:N1-MeG (15%). For C:N2-MeG, the overall primer extension was most efficient (55%). Uracil has a similar size and shape as Cytosine, but changes in H-Bond donor and acceptor moieties (i.e. N3 hydrogen becomes a donor and the O4 oxygen an acceptor). Uracil-containing primers were evaluated for DNA Pol ζ extension ability. For the U:O6-MeG DNA, the most notable full-length extension was observed. In addition, prominent n+1 extension products were observed for U:G, U:N1-MeG and U:N2-MeG and less full-length product extension was detected (Figure 1, lanes 5, 7, and 8, respectively). DNA Pol ζ extended the Benzi and HBT DNA similarly. For the Benzi:O6-MeG DNA, the full-length product was the most prominent for the entire Benzi series (41%), and was comparable to C:O6-MeG (44%) and U:O6-MeG (42%) DNA. For G, N1-MeG and N2-MeG paired with Benzi, fulllength extension was inefficient (12%, 6%, and 7%, respectively, Figure 2, lanes 9, 11, and 12). For the HBT:O6-MeG DNA, an identical extension pattern was observed for U and Benzi paired with O6-MeG, but the overall extension was reduced 2-fold. HBT:N2-MeG and HBT:G were both extended by DNA Pol ζ to yield full-length products, however with reduced extension products in comparison to HBT:O6-MeG. A prominent n+1 product was observed for HBT paired with G, N1-MeG and N2-MeG, however HBT: N1-MeG was the strongest block for further catalysis by DNA Pol ζ (Figure 2, lane 15). Absence of H-bonding reduces the extension efficiency by Pol ζ but can be restored by increased stacking capacity with the incoming nucleotide. When BIM or indole were paired with methylated guanine adducts, no full-length extension of the primer by DNA Pol ζ was observed, rather, an n+1 band. The percentage of the n+1 product was higher for BIM-containing DNA as compared to indole. For example, BIM:N2-MeG had the most n+1 product (40%, Figure 2, lane 20), whereas for indole:N2-MeG, 5% of product was observed (Figure 2, lane 24). These data suggest that an absence of Hbonding interactions at the primer-template terminus drastically reduce primer extension efficiency of Pol ζ (Figure 2, lanes 21-24). Aside from such interbase hydrogen bonding interactions, base stacking interactions of nucleobases may enhance extension, in this case between the incoming dNTP and the base at the 3’-end of the primer.35-36 In order to test if increased base stacking between the incoming dNTP and the primer terminus can restore the efficiency of extension by Pol ζ in the absence of Hbonding, i.e. in the indole:X base pairs, we performed

primer extension assays with 5-nitroindole at the 3’-end of the primer strand. 5-nitorindole has an enhanced base stacking ability in comparison to indole, and studies of T4 with 5DNA polymerase (gp43 exo-) nitroindoletriphosphate suggested that π-π stacking interactions can stabilize the ternary Pol-DNA-dNTP complex by interacting with aromatic amino acid residues in the Pol active site but also by stabilizing the hydrophobic DNA duplex during DNA synthesis.35-36 DNA Pol ζ indeed extended from 5-nitroindole more than from indole in all duplexes when placed opposite G, O6-MeG, N1-MeG and N2-MeG (Figure S12). For example, more extension by DNA Pol ζ was observed from 5-nitroindole:G (52%) than for indole:G (4%) (Figure S12). For both 5-nitroindole and indole duplexes, DNA Pol ζ only inserted a single nucleotide and no further primer extension was observed. These data demonstrate that increased stacking interactions with the incoming dNTP and the base at the 3’-end of the primer also promote DNA Pol ζ catalysis, consistent with similar observations made for T4 DNA polymerase. 35-36 The fidelity of Pol ζ is unaffected irrespective of Hbonding. Since H-bonding and stacking interactions at the primer terminus impacted the efficiency of extension catalyzed by Pol ζ, we were interested to evaluate the capacity for dNTP discrimination in these cases by DNA Pol ζ. Based on single nucleotide insertion assays, we characterized the nucleotide insertion by DNA Pol ζ for each of the canonical dNTPs with all DNA substrates. We observed in all cases that DNA Pol ζ only inserted the correct dNTP (i.e. dGTP) after N:X terminal base pairs (Figure S9 and Figure S13). In addition, we compared this fidelity to human DNA polymerase ε. We determined that this accurate and proofreading proficient enzyme also replicates from the C:G DNA in an error-free manner (Figure S14). Thus, DNA Pol ζ can catalyze replication in an error-free manner from methylguanine lesions irrespective of the H-bonding or stacking interactions present at the DNA terminus, but the differences in dGTP insertion efficiency for human DNA Pol ζ remained an open question and led us to evaluate the kinetics of the process in further detail. Catalytic efficiency of dGTP insertion by DNA Pol ζ from methylguanine lesions enhanced with increased H-bonding at the DNA terminus. To quantitatively asses the catalytic efficiency for DNA Pol ζ to insert dGTP from methylguanine DNA adducts, steady-state kinetic analysis was performed. Kinetic parameters KM and kcat were determined under enzyme-limiting conditions with a 10-fold excess of DNA and varying dGTP concentrations. Amongst all DNA duplexes tested, Cytosine paired with G, O6-MeG and N2-MeG had the most favorable catalytic efficiencies of 1.3 ± 0.3, 1.1 ± 0.2 and 1.5 ± 0.3 μM-1 min-1, respectively (Table 1, entries 1, 2 and 4). Further, for these DNA substrates, identical KM values were obtained, but the higher catalytic efficiency for C:N2MeG was attributed to a more favorable kcat of 1.1 ± 0.04 min-1. In contrast, the kcat for C:N1-MeG was 0.25 ± 0.01 min-1 and the KM was 2.5 ± 0.5 µM, which resulted in a 13fold decreased kcat/KM compared to C:G. For the Uracil


Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

duplexes, the most favorable catalytic efficiencies were for U:O6-MeG (0.057 ± 0.01 μM-1min-1) and U:N2-MeG (0.064 ± 0.01 μM-1min-1), however these rates were approximately 20-fold slower compared to C:G (1.3 ± 0.3 μM-1min-1). For U:G and U:N1-MeG, a KM of 7.6 ± 2.0 μM and 6.3 ± 0.9 μM was determined, respectively, which are the least favora-

Page 6 of 13

ble KM values obtained for the Uracil-containing DNA tested. The trend in catalytic efficiencies for Benzi and HBT was similar. Moreover, Benzi and HBT paired with O6MeG had the most favorable catalytic efficiencies of 0.14 ± 0.03 μM-1min-1 and 0.063 ± 0.01 μM-1min-1, respectively.

Table 1. Steady-state kinetic analysis for the correct insertion of dGTP on templates comprised of G, O6-MeG, N1MeG and N2-MeG paired opposite canonical Cytosine and Uracil and modified Benzi, HBT, BIM and indolecontaining primers. -1

-1

-1

Primer

Entry

Template

kcat [min ]

KM [μM]

kcat/KM [μM min ]

fold change

Cytosine

1 2 3 4

G 6 O -MeG 1 N -MeG 2 N -MeG

0.66 ± 0.02 0.62 ± 0.02 0.25 ± 0.01 1.1 ± 0.04

0.50 ± 0.1 0.60 ± 0.1 2.5 ± 0.5 0.70 ± 0.1

1.3 ± 0.3 1.1 ± 0.2 0.10 ± 0.02 1.5 ± 0.3

1.0 -1.2 -13 1.1

Uracil

5 6 7 8


0.24 ± 0.01 0.21 ± 0.01 0.23 ± 0.01 0.22 ± 0.01

7.6 ± 2.0 3.6 ± 0.6 6.3 ± 0.9 3.5 ± 0.5

0.032 ± 0.01 0.057 ± 0.01 0.038 ± 0.01 0.064 ± 0.01

Benzi

9 10 11 12


0.22 ± 0.01 0.82 ± 0.03 0.20 ± 0.01 0.16 ± 0.01

6.3 ± 1.5 6.0 ± 1.2 10 ± 2.0 7.2 ± 1.4

0.036 ± 0.01 0.14 ± 0.03 0.019 ± 0.01 0.022 ± 0.01

-41 -23 -34 -20 -36 -9 -68 -59

HBT

13 14 15 16


0.15 ± 0.01 0.20 ± 0.01 0.17 ± 0.01 0.19 ± 0.01

5.0 ± 1.0 3.1 ± 0.6 9.6 ± 2.0 5.2 ± 1.0

0.030 ± 0.01 0.063 ± 0.01 0.017 ± 0.01 0.036 ± 0.01

-43 -21 -76 -36

BIM

17 18 19 20


0.22 ± 0.01 0.16 ± 0.01 0.26 ± 0.01 0.29 ± 0.01

12 ± 2.0 12 ± 2.0 5.8 ± 0.9 4.5 ± 0.7

0.018 ± 0.01 0.013 ± 0.01 0.045 ± 0.01 0.065 ± 0.01

-72 -100 -29 -20

Indole

21 22 23 24

G O -MeG 1 N -MeG 2 N -MeG

0.19 ± 0.01 0.18 ± 0.01 0.15 ± 0.01 0.15 ± 0.01

40 ± 7.0 15 ± 3.0 20 ± 4.0 14 ± 3.0

0.005 ± 0.001 0.012 ± 0.002 0.007 ± 0.002 0.011 ± 0.002

-260 -108 -186 -125

6

Compared with C:G, Benzi:O6-MeG had a 9-fold reduction in catalytic efficiency and HBT:O6-MeG had a 21-fold reduction. Further, for Benzi:O6-MeG a kcat of 0.82 ± 0.03 min-1 (Table 1, entry 10) was the second highest turnover observed for all DNA duplexes tested. In addition, HBT:O6-MeG had a similar efficiency as Uracil paired with O6-MeG or N2-MeG. The least favorable catalytic efficiency for the Benzi and HBT DNA was for Benzi and HBT paired opposite N1-MeG. Here, KM values of 10 ± 2.0 μM and 9.6 ± 2.0 μM were determined (Table 1, entries 11 and 15, respectively). DNA Pol ζ had the lowest catalytic efficiencies for BIM and indole-containing primers. For all 24 DNA duplexes tested, the least efficient were for indole:G and indole:N1MeG with kcat/KM’s of 0.005 ± 0.001 μM-1 min-1 and 0.007 ± 0.002 μM-1 min-1 (Table 1, entries 21 and 23, respectively). The KM for indole:G was 40 µM, which was the least favorable in this study. The optimal catalytic efficiencies for the BIM-containing primers were for BIM:N1-MeG and BIM:N2-MeG with kcat/KM’s of 0.045 ± 0.01 μM-1min-1 and

0.065 ± 0.01 μM-1min-1 (Table 1, entries 19 and 20, respectively). These efficiencies were similar to Uracil paired with N1-MeG (0.038 ± 0.01 μM-1min-1) and N2-MeG (0.064 ± 0.01 μM-1min-1) (Table 1, entries 7 and 8, respectively). Further, BIM:N1-MeG and BIM:N2-MeG showed the most favorable turnover numbers (kcat 0.26 ± 0.01 min-1 and 0.29 ± 0.01 min-1, respectively) for all the modified primers tested, except for Benzi:O6-MeG with 0.82 ± 0.03 min-1 (Table 1, entry 10). Computed structures of artificial probes paired opposite G with the Pol ζ homolog, E. coli polymerase II. Intrigued by the observation that DNA Pol ζ is unable to insert more than a single dGTP for BIM and indole DNA (i.e. a single n+1 band, Figure 2, lanes 17-24), we performed computational modeling to gain insight into molecular interactions. We examined differences in terminal DNA interbase pairing and DNA-polymeraseminor groove H-bonding interactions that might influence the ability for proficient DNA synthesis by Pol ζ. Due to the lack of available structural data for Pol ζ, we em-


Page 7 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

ployed structural modeling studies with E. coli Pol II, a homolog for Pol ζ.37 Sequence alignment of the polymerase catalytic subunit REV3 against E. coli polymerase II, indicated similar sequence overlap for residues reported to be involved in minor groove interactions, i.e. DNA Pol II K615 and D545 that align with the corresponding Rev3

residues K2833 and D2781 (Figure 3B).37 Additionally, functional studies have shown similarities in lesion bypass by means of looping out the template strand for human Pol ζ and E. coli Pol II.30, 37 Due to complementary sequence alignment and functional data, we employed the ternary crystal PDB 3MAQ of E.coli polymerase II.37

Figure 3. (A) Minor groove H-bonding at position n-1. N = C, Benzi, BIM or indole. X = G. (B) Sequence alignment of B-family polymerases. The aspartic acid and lysine residue, involved in minor groove H-bonding are conserved in E. coli Pol II, RB69, human Pol δ, and human Pol ζ (Rev3 catalytic subunit). (C) Structural model of the canonical C:G base pair. Minor groove inter2 actions between the DNA and E. coli Pol II show coordination of the ε-amino group of K615. A H-bond between the O -carbonyl of C, D545 and a water molecule (W) align the ε-amino group in a tetrahedral fashion. (D) Benzi paired with G can form identical minor groove contacts with K615 like for the canonical C:G base pair . (E) Structural model of BIM:G. The ε-amino group is coordinated by a water molecule (W) and D545. (F) indole paired with G has a similar coordination like the BIM:G pair.

Molecular mechanics simulations were performed with the Molecular Operating Environment (MOE) software. The DNA structure was slightly modified to match the sequence context used in our in vitro studies (Figure 3A). The modified structures were prepared with the MOE QuickPrep functionality and energy minimized with the Amber 99 force field. To investigate H-bonding of the minor groove carbonyl group at the O2 position, we compared structural models of Benzi:G, BIM:G and indole:G with the C:G base pair in the n-1 position (Figure 3A). The canonical C:G base pair forms a direct H-bond between the O2-carbonyl of C and the ε-amino group of K615 (Figure 3C). The ε-amino group further interacts by H-bonding coordination of a water molecule with a hydrogen from the N2 of G (Figure 3C). The model for Benzi:G in the active site of E. coli Pol II revealed that Benzi has identical H-bonding interactions with G as does C (Figure 3D and 3C, respectively), originating from the Benzi carbonyl group in the 5membered ring that coordinates with the ε-amino group of K615. In contrast, the ε-amino group of K615 does not maintain a H-bonding interaction when BIM or indole are paired opposite G. This arises from the fact that BIM and indole lack a carbonyl group (Figure 3E and 3F, respectively). For Benzi:G, BIM:G, indole:G and C:G, H-bonding interactions are present in all cases with D545 via a water molecule coordinated by the N2 position of G. In addition to BIM and indole, there are instances where a prominent n+1 band was observed (e.g. C:N2-MeG and U:N2-MeG, Figure 2). Regarding N2-MeG paired opposite C and U, the methyl group at the N2 position could

be oriented in a distal or proximal position. To investigate this orientation, we performed computation modeling with the C:N2-MeG base-pair in the Pol II active site. In the instance when the N2 methyl group points away from the Watson-Crick H-bonding face (distal position, Figure S10) it prevents H-bond with a water (W) molecule that is involved in the positioning of the ε-amino group of K615. The loss of the H-bond reduces constraints on the interaction of the water molecule with the Lysine ε-amino group.

DISCUSSION While replicative polymerases select nucleotides largely on the basis of geometric fit, and Y-family polymerases more on the basis of H-bonding, it is unknown how these molecular features influence replication fidelity for human DNA polymerase ζ. DNA Pol ζ is a B-family polymerase, but functions in lesion bypass like Y-family TLS polymerases. In this study, we used synthetic nucleoside probes with altered H-boding and base stacking capabilities to probe extension properties and dissect molecularlevel interactions for human DNA Pol ζ replication from methylated guanine adducts. Our steady-state kinetic data suggests that interbase H-bonding promotes efficient extension from methylguanine DNA adducts by DNA polymerase ζ. Further, our computational modeling suggests minor groove H-bonding interactions between the O2-carbonyl group on the primer strand and a lysine residue on the polymerase to be important for elongation of the primer strand.


Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

A model for interbase H-bonding promoting DNA synthesis by DNA polymerase ζ. DNA Pol ζ inserted dGTP from all 24 permutations of terminal primertemplate base pairs created in this study (Table 1) with catalytic efficiencies that corresponded with increasing H-

Page 8 of 13

bonds at the DNA terminus. This trend is displayed in Figure 4, where base pairs are plotted based on their capacity to form H-bonds (x-axis) as a function of their kcat/KM value (y-axis).

Figure 4. Catalytic efficiency as a function of terminal base pair for human DNA polymerase ζ in extending from guanine and 6 1 2 methylated guanine adducts O -MeG, N -MeG and N -MeG. From left to right, terminal base pairs are categorized on their ability to hydrogen bond from 0-, 1-, 2- and 3-H-bonds. The circles with a color-filled number correspond to the numbering system in Figure 2. Here, each base pair is given a unique number and a color that is based on the maximum number of hydrogen bonds, i.e. pink (Group 1, 3-H-bonds), green (Group 2, 2-H-bonds) and blue (Group 3, 0- and 1-H-bond). The structural renderings for all 24 permutations of base pairing is shown in Figure 2. Our model shows that H-bonding is an important molecular feature that promotes DNA polymerase ζ to catalyze the most efficient DNA synthesis. Here, the most efficient catalysis was for 2 the 3-H-bonding base pairs C:G and C:N -MeG DNA (1 and 4, respectively). In contrast, the least efficient enzymatic catalysis was observed for the indole analogs (21, 23, 24 and 22). For the indole-containing primers, no H-bonding is possible.

Overall, the C:G and C:N2-MeG base pairs that can form 3 H-bonds were extended most efficiently, followed by C:O6-MeG, Benzi:O6-MeG and C:N1-MeG, which have the potential to form 2 H-bonds. In contrast, the least efficient enzymatic catalysis was observed for the indole probe (Figure 4, 21, 22, 23 and 24) with no H-bonding possibility. There are instances where the H-bonding trend is less clear as some terminal base pairs with 1- and 2-H-bonds overlap in their extension efficiency. For example, for BIM:O6-MeG, BIM:G, HBT:N1-MeG, Benzi:N1-MeG and Benzi:N2-MeG with 1 H-bond, inefficient Pol ζ catalysis was observed and all kcat/KM values were below 0.03 μM1 min-1 (Figure 4). Whereas for Benzi:G with 1 H-bond, a similar catalytic efficiency was observed like that for U:G, HBT:G and HBT:N2-MeG with 2 H-bonds. This discrepan-

cy could be due to Benzi forming a wobble base pair with G that would allow for 2 H-bonds. The catalytic efficiency for HBT paired with O6-MeG (1 H-bond) was similar or more efficient than for the 2-Hbond base pairs HBT:G, U:G, HBT:N2-MeG, U:O6-MeG and U:N2-MeG, suggesting other potentially relevant molecular interactions. Although HBT:O6-MeG can only form 1 H-bond, the O6-methoxy group may interact with the sulfur moiety, thereby stabilzing the base pair and increasing extension efficiency. For example, previous computational and structural studies demonstrated a stabilizing interaction between a synthetic base pair containing a sulfur atom that orients itself opposite a methoxy group of the pairing base.38-39 Major vs. minor-groove DNA alkylation can affect the ability of DNA Pols to bypass and extend from damaged


Page 9 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

basepairs.40-41 In this study, the placement of the methyl lesion at varying positions on guanine had different effects on DNA Pol ζ bypass. For example, DNA Pol ζ had an increased propensity for full-length elongation from the O6-MeG major groove lesion, whereas for the N2-MeG minor groove lesion a significant n+1 insertion occurred, and further elongation was inefficient. Regardless whether the methyl moiety was in the major or minor grove, the overall fidelity of DNA Pol ζ remained high (Figure S9), but its efficiency was significantly altered (Table 1). These data suggest a variation in tolerance for human DNA Pol ζ as a function of the position of the lesion within the major and minor groove. In addition, a recent study of 3-deaza-adenosine analogues containing methyl and bulky aromatic adducts in the minor groove were not bypassed by human Pol ζ.42 However, in a cellbased study yeast Pol ζ appears to be involved in bypassing a 3-methyl-adenosine lesion.43 For Pol δ, also a B-family Pol, it was shown to bypass both major and minor grove methylation-derived lesions, suggesting a higher tolerance to bypass minor groove lesions as compared to Pol ζ.31, 41, 44 Work to characterize this lesion preference in vivo would help elucidate its role in resistance towards alkylation-based chemotherapy. A model for DNA-protein minor groove H-bonding interactions that influence extension ability by DNA polymerase ζ. It is known that the carbonyl O2 atom of pyrimidines and N3 atom of purines can form H-bonding contacts with amino acid side chains on DNA polymerases, either directly or via water molecules.45 These interactions provide the chemical basis for minor groove interactions that are important for efficient DNA synthesis.46-52 Further, these minor groove interactions, between DNA and DNA polymerase, can occur at the insertion and postinsertion site.53 Herein, we observed that DNA Pol ζ could extend from all 24 DNA substrates (Figure 2), however only a single n+1 extension product was observed for BIM and indole DNA. We predicted that structural features shared between the BIM and indole probes may explain this observation. Structurally, BIM and indole lack the O2-carbonyl moiety that is found in the C, U, Benzi and HBT bases. While subsequent elongation by DNA Pol ζ was prevented in the instance that BIM:X and indole:X base pairs were at the n-1 position (Figure 3A), we hypothesized that the inability of elongation by Pol ζ originates from the absence of the O2-carbonyl in these probes. To investigate our hypothesis, we used molecular modeling with our modified DNA base pairs, to visualize potential DNA-polymerase H-bonding interactions during DNA synthesis by Pol ζ. In the computed structures, C:G and Benzi:G can form a direct DNA-polymerase H-bond contact emanating from the O2-carbonyl group (Figure 3C and 3D, respectively), while no direct- or water-mediated H-bonding interactions were observed between BIM:G and indole:G in the active site of the polymerase (Figure 3E and 3F, respectively). In fact, the missing O2-carbonyl in BIM and indole prevents a direct or water-mediated interaction with the ε-amino group of K615 of Pol II that

may explain the Pol ζ stalling during DNA synthesis from these base pairs. A possible explanation of what impairs the elongation ability by a DNA polymerase has been reported by Konigsberg and coworkers. In this study, a 900-fold decrease in catalytic activity of the B-Family polymerase RB69 was observed upon loss of minor groove interactions between polymerase and DNA at the penultimate base pair (i.e. the analogous n-1 position in our study).51 They demonstrated that minor groove interactions between the primer strand and K706 and D621 of RB69 are involved in the coordination of metal ions that are necessary to activate the 3’-OH on the primer terminus and are thus important for catalysis.51 Indeed, the lysine and aspartic acid residues in RB69 K706 and D621 are highly conserved in B-family polymerases. In our sequence alignment, we observe corresponding residues for yeast DNA Pol δ (K814 and D762)54, E. coli DNA Pol II (K615 and D545)37, and in φ29 DNA Pol (K498 and D456)55. Additionally, sequence alignment with human Pol ζ correspond to residues K2833 and D2781 (Figure 3B). Given the similar sequence motifs for B-family DNA polymerases, we predict that key aspartic acid and lysine residues (Figure 3B in red) play a role in the ability for DNA synthesis by DNA polymerase ζ. The distinct n+1 product band and unsuccessful full-length extension for BIM and indole by DNA Pol ζ (Figure 2) could result from the loss of these minor groove contacts, which are maintained for the C, U, Benzi and HBT probes. In the case of BIM and indole, it may be that the tetrahedral geometry of the ε-amino group of K615 is distorted, which could perturb the coordination of a metal ion and alter the correct alignment of the α-phosphorus atom of the incoming dNTP with the 3’-OH of terminal base in the primer strand. Using the E. Coli Pol II structure as a close homolog for DNA Pol ζ suggests that residues K2833 and D2781 might be directly involved in minor groove H-bonding interactions of Pol ζ with DNA during replication. This hypothesis is further supported by a recent study with mutant mouse strains where one or both aspartic acid residues in the YGDTDS motif of the Rev3L gene, which is conserved in B-family polymerases, were mutated to alanine.56 Mutation of the first Asp reduced the elongation capacity of Pol ζ, but viability was not affected. Upon mutating both Asp residues yielded an inactive enzyme and was embryonically lethal.56

CONCLUSION This work provides a basis for understanding interactions that influence how the human extender DNA Pol ζ can replicate from methylguanine DNA adducts. Using synthetic nucleobase surrogates as chemical probes, we found that increased interbase H-bonding interactions at the terminal base pair promotes DNA synthesis by human DNA Pol ζ. While our model reveals in detail how increased H-bonding promotes catalysis, other chemical features such as nucleobase size, shape and base stacking 27, 32, 35-36 with the incoming nucleotide also influence catalysis. For example, DNA Pol ζ could extend from an


Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

indole probe, which has no capacity for H-bonding, albeit with lower proficiency. Our analysis suggests that lysine and aspartate residues conserved amongst B-family Pols and Pol ζ interact with the penultimate base on the 3’ end of the primer strand and allows Pol ζ to promote the extension of DNA synthesis from methylated guanine lesions. Overall, these results elucidate chemical interactions that promote human DNA Pol ζ in performing promutagenic DNA synthesis, while also demonstrating the utility of synthetic nucleoside analogs as chemical tools for probing catalytic mechanisms of DNA polymerases.

ASSOCIATED CONTENT Supporting Information. Synthesis details, characterization of nucleoside intermediates, HPLC traces, mass spectrometry analysis of methylated guanine and synthetic probe-containing DNA and human polymerase ζ analysis are described in the supporting information. This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding Author To whom correspondence should be addressed: Hailey L. Gahlon, [email protected] Correspondence may also be addressed to Shana J. Sturla, Email: [email protected]

Notes The authors declare no competing financial interest.

ACKNOWLEDGMENT We thank Dr. Wei Yang, National Institutes of Health, USA, for kindly sharing plasmids and Dr. Young-Sam Lee for helpful input regarding enzyme expression and purification. This work was supported by the Swiss National Science Foundation (156280).

REFERENCES 1. Edara, S.; Kanugula, S.; Pegg, A. E., Expression of the inactive C145A mutant human O6-alkylguanineDNA alkyltransferase in E.coli increases cell killing and mutations by N-methyl-N′-nitro-N-nitrosoguanidine. Carcinogenesis 1999, 20 (1), 203-108. 2. Cheng, G.; Wang, M.; Upadhyaya, P.; Villalta, P. W.; Hecht, S. S., Formation of Formaldehyde Adducts in the Reactions of DNA and Deoxyribonucleosides with α-Acetates of 4(Methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), 4-(Methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL), and N-Nitrosodimethylamine (NDMA). Chem. Res. Toxicol. 2008, 21 (3), 746-751. 3. Barrows, L. R.; Magee, P. N., Nonenzymatic methylation of DNA by S-adenosylmethionine in vitro. Carcinogenesis 1982, 3 (3), 349-51. 4. Rydberg, B.; Lindahl, T., Nonenzymatic methylation of DNA by the intracellular methyl group donor S-adenosyl-L-methionine is a potentially mutagenic reaction. EMBO J. 1982, 1, 211-216.

Page 10 of 13

5. Rattray, A. J.; Strathern, J. N., Error-prone DNA polymerases: when making a mistake is the only way to get ahead. Annu. Rev. Genet. 2003, 37 (1), 31-66. 6. Vaisman, A.; Woodgate, R., Translesion DNA polymerases in eukaryotes: what makes them tick? Critical Reviews in Biochemistry and Molecular Biology 2017, 52 (3), 274-303. 7. Kunkel, T. A., DNA replication fidelity. J. Biol. Chem. 2004, 279 (17), 16895-8. 8. Lawrence, C. W., Cellular functions of DNA polymerase zeta and Rev1 protein. Adv. Protein Chem. 2004, 69, 167-203. 9. Mitra, G.; Pauly, G. T.; Kumar, R.; Pei, G. K.; Hughes, S. H.; Moschel, R. C.; Barbacid, M., Molecular analysis of O6-substituted guanine-induced mutagenesis of ras oncogenes. Proc. Natl. Acad. Sci. 1989, 86 (22), 8650-4. 10. Pauly, G. T.; Moschel, R. C., Mutagenesis by O6-methyl-, O6-ethyl, and O6-benzylguanine and O4methylguanine in human cells: effects of O6alkylguanine-DNA transferease and mismatch repair. Chem. Res. Toxicol. 2001, 14, 594-900. 11. Eoff, R. L.; Irimia, A.; Egli, M.; Guengerich, F. P., Sulfolobus solfataricus DNA polymerase Dpo4 is partially inhibited by "wobble" pairing between O6methylguanine and cytosine, but accurate bypass is preferred. J. Biol. Chem. 2007, 282 (2), 1456-1467. 12. Yasui, M.; Matsui, S.; Ihara, M.; Laxmi, Y. R.; Shibutani, S.; Matsuda, T., Translesional synthesis on a DNA template containing N2-methyl-2'deoxyguanosine catalyzed by the Klenow fragment of Escherichia coli DNA polymerase I. Nucleic Acids Res 2001, 29 (9), 1994-2001. 13. Zhong, X.; Garg, P.; Stith, C. M.; McElhinny, S. A. N.; Kissling, G. E.; Burgers, P. M. J.; Kunkel, T. A., The fidelity of DNA synthesis by yeast DNA polymerase zeta alone and with accessory proteins. Nucleic Acids Research 2006, 34 (17), 4731-4742. 14. Johnson, R. E.; Yu, S. L.; Prakash, S.; Prakash, L., Yeast DNA polymerase zeta is essential for errorfree replication past thymine glycol. Genes Dev. 2003, 77-87. 15. Nelson, J. R.; Lawrence, C. W.; Hinkle, D. C., Thymine-Thymine dimer bypass by yeast DNA polymerase ζ. Science 1996, 272 (5268), 1646-1649. 16. Johnson, R. E.; Haracska, L.; Prakash, S.; Prakash, L., Role of DNA polymerase eta in the bypass of a (6-4) TT photoproduct. Mol. Cell. Biol. 2001, 21 (10), 3558-3563. 17. Haracska, L.; Unk, I.; Johnson, R. E.; Johansson, E.; Burgers, P. M.; Prakash, S.; Prakash, L., Roles of yeast DNA polymerases δ and ζ and of Rev1 in the bypass of abasic sites Genes Dev. 2001, 15, 945-954. 18. Haracska, L.; Prakash, S.; Prakash, L., Yeast DNA polymerase ζ is an efficient extender of primer ends opposite from 7,8-dihydro-8-oxoguanine and O6-


Page 11 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

methylguanine. Mol. and Cell. Biol. 2003, 23 (4), 14531459. 19. Roos, W. P.; Tsaalbi-Shtylik, A.; Tsaryk, R.; Guvercin, F.; de Wind, N.; Kaina, B., The translesion polymerase Rev3L in the tolerance of alkylating anticancer drugs. Mol Pharmacol 2009, 76 (4), 927-34. 20. Wu, J.; Wang, P.; Li, L.; Williams, N. L.; Ji, D.; Zahurancik, W. J.; You, C.; Wang, J.; Suo, Z.; Wang, Y., Replication studies of carboxymethylated DNA lesions in human cells. Nucleic Acids Res 2017, 45 (12), 72767284. 21. Morales, J. C.; Kool, E. T., Efficient replication between non-hydrogen-bonded nucleoside shape analogs. Nat Struct Biol 1998, 5 (11), 950-954. 22. Morales, J. C.; Kool, E. T., Varied Molecular Interactions at the Active Sites of Several DNA Polymerases: Nonpolar Nucleoside Isosteres as Probes. J. Am. Chem. Soc. 2000, 122 (6), 1001-1007. 23. Washington, M. T.; Helquist, S. A.; Kool, E. T.; Prakash, L.; Prakash, S., Requirement of Watson-Crick Hydrogen Bonding for DNA Synthesis by Yeast DNA Polymerase Mol. Cell. Biol. 2003, 23 (14), 5107-5112. 24. Wolfle, W. T.; Washington, M. T.; Kool, E. T.; Spratt, T. E.; Helquist, S. A.; Prakash, L.; Prakash, S., Evidence for a Watson-Crick hydrogen bonding requirement in DNA synthesis by human DNA polymerase kappa. Mol. Cell. Biol. 2005, 25 (16), 713743. 25. Kool, E. T., Active site tightness and substrate fit in DNA replication. Annu. Rev. Biochem. 2002, 71, 191-219. 26. Kool, E. T., Hydrogen bonding, base stacking, and steric effects in dna replication. Annu. Rev. Biophys. Biomol. Struct. 2001, 30, 1-22. 27. Gahlon, H. L.; Boby, M. L.; Sturla, S. J., O6Alkylguanine postlesion DNA synthesis is correct with the right complement of hydrogen bonding. ACS Chem. Biol. 2014, 9, 2807-2814. 28. Wilson, K. A.; Wetmore, S. D., Molecular Insights into the Translesion Synthesis of BenzylGuanine from Molecular Dynamics Simulations: Structural Evidence of Mutagenic and Nonmutagenic Replication. Biochemistry 2017, 56 (13), 1841-1853. 29. Albrecht, L.; Wilson, K. A.; Wetmore, S. D., Computational Evaluation of Nucleotide Insertion Opposite Expanded and Widened DNA by the Translesion Synthesis Polymerase Dpo4. Molecules 2016, 21 (7). 30. Lee, Y. S.; Gregory, M. T.; Yang, W., Human Pol ζ purified with accessory subunits is active in translesion DNA synthesis and complements Pol eta in cisplatin bypass. Proc. Natl. Acad. Sci. 2014, 111 (8), 2954-2959. 31. Raz, M. H.; Dexter, H. R.; Millington, C. L.; van Loon, B.; Williams, D. M.; Sturla, S. J., Bypass of Mutagenic O(6)-Carboxymethylguanine DNA Adducts

by Human Y- and B-Family Polymerases. Chem Res Toxicol 2016, 29 (9), 1493-503. 32. Gahlon, H. L.; Schweizer, W. B.; Sturla, S. J., Tolerance of base pair size and shape in postlesion DNA synthesis. J. Am. Chem. Soc. 2013, 135 (17), 63846387. 33. Gahlon, H. L.; Sturla, S. J., Hydrogen bonding or stacking interactions in differentiating duplex stability in oligonucleotides containing synthetic nucleoside probes for alkylated DNA. Chemistry 2013, 19 (33), 11062-7. 34. Lai, J. S.; Kool, E. T., Selective pairing of polyfluorinated DNA bases. J. Am. Chem. Soc. 2004, 126 (10), 3040-1. 35. Reineks, E. Z.; Berdis, A. J., Evaluating the Contribution of Base Stacking during Translesion DNA Replication. Biochemistry 2004, 43 (2), 393-404. 36. Zhang, X.; Lee, I.; Berdis, A. J., Evaluating the contributions of desolvation and base-stacking during translesion DNA synthesis. Org Biomol Chem 2004, 2 (12), 1703-11. 37. Wang, F.; Yang, W., Structural insight into translesion synthesis by DNA Pol II. Cell 2009, 139 (7), 1279-89. 38. Betz, K.; Malyshev, D. A.; Lavergne, T.; Welte, W.; Diederichs, K.; Dwyer, T. J.; Ordoukhanian, P.; Romesberg, F. E.; Marx, A., KlenTaq polymerase replicates unnatural base pairs by inducing a WatsonCrick geometry. Nat. Chem. Biol. 2012, 8 (7), 612-4. 39. Negi, I.; Kathuria, P.; Sharma, P.; Wetmore, S. D., How do hydrophobic nucleobases differ from natural DNA nucleobases? Comparison of structural features and duplex properties from QM calculations and MD simulations. Phys. Chem. Chem. Phys. 2017, 19 (25), 16365-16374. 40. Washington, M. T.; Minko, I. G.; Johnson, R. E.; Haracska, L.; Harris, T. M.; Lloyd, R. S.; Prakash, S.; Prakash, L., Efficient and error-free replication past a minor-groove N2-guanine adduct by the sequential action of yeast Rev1 and DNA polymerase zeta. Mol Cell Biol 2004, 24 (16), 6900-6. 41. Choi, J. Y.; Guengerich, F. P., Adduct size limits efficient and error-free bypass across bulky N2guanine DNA lesions by human DNA polymerase eta. J Mol Biol 2005, 352 (1), 72-90. 42. Malvezzi, S.; Angelov, T.; Sturla, S. J., Minor Groove 3-Deaza-Adenosine Analogues: Synthesis and Bypass in Translesion DNA Synthesis. Chemistry 2017, 23 (5), 1101-1109. 43. Monti, P.; Ciribilli, Y.; Russo, D.; Bisio, A.; Perfumo, C.; Andreotti, V.; Menichini, P.; Inga, A.; Huang, X.; Gold, B.; Fronza, G., Rev1 and Polζ influence toxicity and mutagenicity of Me-lex, a sequence selective N3-adenine methylating agent. DNA Repair 2008, 7 (3), 431-438. 44. Choi, J. Y.; Chowdhury, G.; Zang, H.; Angel, K. C.; Vu, C. C.; Peterson, L. A.; Guengerich, F. P.,


Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Translesion synthesis across O6-alkylguanine DNA adducts by recombinant human DNA polymerases. J Biol Chem 2006, 281 (50), 38244-56. 45. Seeman, N. C.; Rosenberg, J. M.; Rich, A., Sequence-specific recognition of double helical nucleic acids by proteins. Proc. Natl. Acad. Sci. 1976, 73 (3), 804-8. 46. Doublie, S.; Tabor, S.; Long, A. M.; Richardson, C. C.; Ellenberger, T., Crystal structure of a bacteriophage T7 DNA replication complex at 2.2 A resolution. Nature 1998, 391 (6664), 251-8. 47. Morales, J. C.; Kool, E. T., Functional hydrogen-bonding map of the minor groove binding tracks of six DNA polymerases. Biochemistry 2000, 39 (42), 12979-88. 48. Franklin, M. C.; Wang, J.; Steitz, T. A., Structure of the replicating complex of a pol alpha family DNA polymerase. Cell 2001, 105 (5), 657-67. 49. Johnson, S. J.; Beese, L. S., Structures of mismatch replication errors observed in a DNA polymerase. Cell 2004, 116 (6), 803-16. 50. Kool, E. T.; Sintim, H. O., The difluorotoluene debate--a decade later. Chem. Commun. 2006, (35), 3665-75. 51. Xia, S.; Christian, T. D.; Wang, J.; Konigsberg, W. H., Probing minor groove hydrogen bonding interactions between RB69 DNA polymerase and DNA. Biochemistry 2012, 51 (21), 4343-53. 52. Ludmann, S.; Marx, A., Getting it Right: How DNA Polymerases Select the Right Nucleotide. Chimia 2016, 70 (3), 203-6. 53. Kiefer, J. R.; Mao, C.; Braman, J. C.; Beese, L. S., Visualizing DNA replication in a catalytically active Bacillus DNA polymerase crystal. Nature 1998, 391 (6664), 304-7. 54. Swan, M. K.; Johnson, R. E.; Prakash, L.; Prakash, S.; Aggarwal, A. K., Structural basis of highfidelity DNA synthesis by yeast DNA polymerase delta. Nat. Struct. Mol. Biol. 2009, 16 (9), 979-86. 55. Berman, A. J.; Kamtekar, S.; Goodman, J. L.; Lazaro, J. M.; de Vega, M.; Blanco, L.; Salas, M.; Steitz, T. A., Structures of phi29 DNA polymerase complexed with substrate: the mechanism of translocation in Bfamily polymerases. EMBO J 2007, 26 (14), 3494-505. 56. Fritzen, R.; Delbos, F.; De Smet, A.; Palancade, B.; Canman, C. E.; Aoufouchi, S.; Weill, J. C.; Reynaud, C. A.; Storck, S., A single aspartate mutation in the conserved catalytic site of Rev3L generates a hypomorphic phenotype in vivo and in vitro. DNA Repair 2016, 46, 37-46.


Page 12 of 13

Page 13 of 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry TOC Figure


Hydrogen-Bonding Interactions at the DNA Terminus Promote

Recommend Documents