Translesion synthesis of the major nitrogen mustard-induced DNA lesion by human DNA polymerase η

Nitrogen mustards are among the first modern anticancer chemotherapeutics that are still widely used as non-specific anticancer alkylating agents. While the mechanism of action of mustard drugs involves the generation of DNA interstrand cross-links, the predominant lesions produced by these drugs are nitrogen half-mustard-N7-dG (NHMG) adducts. The bulky major-groove lesion NHMG, if left unrepaired, can be bypassed by translesion synthesis DNA polymerases. However, studies of the translesion synthesis past NHMG have not been reported so far. Here we present the first synthesis of an oligonucleotide containing a site-specific NHMG. We also report kinetic and structural characterization of human DNA polymerase η (pol η ) bypassing NHMG. The templating NHMG slows dCTP incorporation ~130-fold, while it increases the misincorporation frequency ~10–30-fold, highlighting the promutagenic nature of NHMG. A crystal structure of pol η incorporating dCTP opposite NHMG shows a Watson-Crick NHMG:dCTP base pair with a large propeller twist angle. The nitrogen half-mustard moiety fits snugly into an open cleft created by the Arg61-Trp64 loop of pol η , suggesting a role of the Arg61-Trp64 loop in accommodating bulky major groove adducts during lesion bypass. Overall, our results presented here provide first insights into translesion synthesis of the major DNA adduct formed by nitrogen mustard drugs.


INTRODUCTION
Nitrogen mustards (NM) were among the first anticancer chemotherapeutics developed and have been widely used in the treatment of various blood cancers and solid tumors [1,2]. Nitrogen mustards such as cyclophosphamide and melphalan ( Figure 1A) are used in combination therapies for the treatment of non-Hodgkin's lymphoma and chronic lymphocytic leukemia. NM are latent electrophiles that generate an aziridinium ion, which can react with a wide variety of nucleophiles including DNA, proteins, and water. The cytotoxicity of NM is believed to be primarily caused by NM-induced DNA interstrand cross-links (ICLs) [3,4], which can impede DNA strand separation [5]. However, the NM-mediated ICLs represent only a small fraction of DNA lesions formed by the mustard drugs. The major lesions produced by NM including mechlorethamine (>90%), chlorambucil (~100%), and melphalan (~60%) [6,7] are nitrogen half-mustards-DNA monoadducts, which can be generated by hydrolysis of the initial NM-DNA monoadducts ( Figure 1B). The majority of NM-mediated DNA damage occurs at the N7 of guanine to produce nitrogen half-mustard-N7-dG (NHMG) adduct [6], while the N3 of adenine servers as a significant site of DNA alkylation for some NM [8]. The goal of the present work is to delineate the mutagenic potential of the major NM-induced DNA lesion NHMG.
Even after several decades of development of the first NM-based alkylating drugs, kinetic and structural studies of DNA polymerase bypassing NHMG have not been reported, thereby significantly limiting our understanding of the miscoding potential of NM-induced DNA damage. Systematic studies of NM-induced DNA lesions have been difficult due to the technical difficulty of preparing the oligonucleotide containing a site-specific incorporated N7-alkylG, which instead rapidly undergoes spontaneous depurination to generate abasic sites. We previously used a 2′-fluorine-mediated transition state destabilization method to incorporate N7-alkylG such as N7MeG and N7-benzyl-G in a site-specific manner [19][20][21][22]. Using the 2′-fluorination strategy, we discovered N7-methyl-G (N7MeG) can form Watson-Crick-like base pairing with thymine in the absence of protein contact ( Figure 1C and 1D) [20]. Guengerich and colleagues reported N7MeG promotes mutagenic replication by translesion synthesis (TLS) DNA polymerases [23].
Here we describe the first synthesis of a site-specific NHMG-containing oligonucleotide via the 2′-fluorination strategy. In addition, we present kinetic and structural assessment of human DNA polymerase η (polη) bypassing NHMG generated by an N-phenyl nitrogen mustard. Polη was chosen for our studies as it belongs to the Y-family DNA polymerases, which have more spacious active sites than replicative DNA polymerases and thus can accommodate a wide variety of DNA lesions [24,25]. Polη, which is encoded by the POLH gene in humans, plays a key role in accurate bypass of the UV-induced cyclobutane pyrimidine dimers in vivo [26][27][28][29]. Mutations in POLH gene cause the development of human syndrome xeroderma pigmentosum variant (XPV) [30]. Unlike Y-family polymerase κ that is generally blocked by major groove DNA lesions [31], polη is observed to bypass several types of major groove DNA lesions including O6-methylguanine [32], cisplatin-GpG intrastrand cross-links [33,34], and "unhooked" nitrogen mustards-GG interstrand crosslinks [35,36]. Therefore, our kinetic and structural characterization of polη-mediated bypass of NHMG provides new insights into the promutagenic nature and translesion synthesis of NM-induced DNA lesions

Preparation of oligonucleotide containing a site specific-incorporated NHMG.
Monoacetylation of commercially available N-phenyldiethanol amine 1 followed by treatment with methanesulfonyl chloride gave mesylated nitrogen half-mustard 3 (Scheme 1). The 2′-fluoro-2′-deoxyguanosine 5 was prepared from a ribose derivative 4 by following the published synthetic procedures [21]. The coupling of the 2′-fluorinated dG 5 with mesylated nitrogen half-mustard 3 took place smoothly in DMF at an ambient temperature to give N7-NHMG adduct 6 in 81% yield. The N7-NHMG 6 was subjected to sequential dimethoxytritylation and phosphatidylation to provide NHMG phosphoramidite 8, which was then incorporated into DNA via the solid phase DNA synthesis and the ultra-mild deprotection to furnish NHMG-modified oligonucleotide 9.

Protein expression and purification.
The catalytic core of human polη (amino acids 1-432) was cloned into pET28a plasmid with NcoI and BamHI restriction enzyme sites. The catalytic domain of polη was used in this study, because its TLS efficiency is comparable to that of the full length polη [29]. E. coli BL21 (DE3) cells transformed with this plasmid were grown at 37°C in LB medium until the OD 600 value reached 0.6. Polη expression was induced for eighteen hours at 20°C by adding 0.3 mM isopropyl-β-thiogalactoside. The induced cells were collected by centrifugation at 8,000g for 20 min at 4°C. Proteins were purified by Ni 2+ -NTA affinity (GE Healthcare), Heparin column and Superdex-75 gel filtration chromatography (GE Healthcare). Purified human polη was concentrated to 15 mg/ml in gel filtration buffer (25 mM Tris, pH 7.5, 300 mM KCl, 10% glycerol and 2 mM dithiothreitol), aliquoted and flash frozen in liquid nitrogen to store at −80°C.

Steady-state kinetics of single nucleotide incorporation opposite a templating NHMG by polη.
Steady-state kinetic parameters for nucleotide insertion opposite NHMG by polη were determined with slight modification of published protocols [37,38]. The primer for kinetic studies (5´-FAM/AGTGTGAG-3´) was purchased from Integrated DNA Technologies (Coralville, IA). An NHMG-containing template (5´-CAT(NHMG)CTCACACT-3´) was synthesized and purified by Midland Certified Reagent company (Midland, TX). To prepare DNA substrate containing NHMG lesion, the primer and template were annealed in hybridization buffer (10 mM Tris-HCl pH 7.5, 1 mM EDTA) at 90°C for 5 min. Polη activities were determined using a reaction mixture containing 40 mM Tris-HCl pH 8.0, 60 mM KCl, 10 mM dithiothreitol, 250 μg/ml bovine serum albumin, 2.5 % glycerol, 5 mM MgCl 2 , 50-100 nM recessed DNA, and varying concentrations of individual dNTP. To prevent the end-product inhibition and substrate depletion from interfering with accurate velocity measurement, the enzyme concentrations and reaction-time intervals were adjusted for every experiment (less than 20% insertion product formed). The reactions were initiated by the adding the enzyme and stopped with a gel-loading buffer (95% formamide with 20 mM EDTA, 45 mM Tris-borate, 0.1% bromophenol blue, 0.1% xylene cyanol). The quenched reaction samples were separated on 20% denaturing polyacrylamide gels. The gels were analyzed using Typhoon Imager (GE Healthcare) to quantify product formation. The k cat and K m were determined by fitting the reaction rate over dNTP concentrations to the Michaelis-Menten equation. Each experiment was repeated three times to measure the average of the kinetic results. The efficiency of nucleotide insertion was calculated as k cat / K m . The relative frequency of dNTP incorporation opposite NHMG was determined as f = Protein crystallization and structure determination.
The ternary polη complex with the templating NHMG was prepared and crystallized with slight modification of published protocols [34,39]. Purified polη (9 mg/ml) was mixed with a 1.2 molar excess of duplex DNA containing NHMG lesion (primer, 5´-AGTGTGAG-3´ and template, 5´-CAT(NHMG)CTCACACT-3´). Subsequently, a 10-fold molar excess of nonhydrolyzable dCMPNPP (Jena Bioscience) was added to the binary complex of the polη:NHMG duplex DNA. The ternary complex co-crystals with dCMPNPP paired with the templating NHMG were obtained using the hanging drop method in a buffer solution containing 100 mM MES pH 6.0-6.5 and 18% PEG2000 MME. Crystals were harvested and cryoprotected in mother liquor supplemented with 20% glycerol and were flash-frozen in liquid nitrogen. Diffraction data were collected at 100 K at the beamline 19-ID at the Advanced Photon Source, Argonne National Laboratory. The diffraction data were processed using HKL2000 and the scaled set was converted to structure factors by CCP4 [40]. Structures were solved by molecular replacement using polη-dG:dCTP* structure (PDB ID: 4O3N) as a search model. The model was built using COOT [41]and refined using PHENIX [42]. MolProbity in the PHENIX GUI was used to generate Ramachandran plots [43]. All the crystallographic figures were generated using PyMOL.

Single nucleotide insertion opposite templating N7-alkylguanine by human polη.
Polη has been shown to efficiently bypass N7-methylG in an error-prone manner [23,39]. To evaluate the impact of the steric bulkiness of the N7-alkyl group on the mutagenic potential of purine, we performed kinetic studies of polη incorporating nucleotide opposite templating dG, and NHMG. Our steady-state kinetic studies show that the size of the N7alkyl group significantly influences the catalytic efficiency (k cat /K m ) of correct insertion. Published kinetic results show small N7-methyl group moderately (~3-fold; 46 vs. 13.2) decreases the catalytic efficiency for dCTP insertion [39]. Furthermore, NHMG at the templating position decreases the dCTP-insertion efficiency by ~130-fold (k cat /K m : 0.36 vs. 45.6, Table 1 and Figure 2), highlighting the impact of steric bulkiness on the correct insertion opposite N7-alkylG. The K m value of the NHMG:dCTP insertion is ~43-fold greater than that of the dG:dCTP insertion (113.4 vs. 2.7 μM), while its k cat value is 3-fold less than that for the dG:dCTP insertion (40.9×10 −3 s −1 vs. 120.6×10 −3 s −1 ), indicating that the low catalytic efficiency (k cat /K m ) for the NHMG:dCTP insertion is mainly driven by increased K m .
The introduction of NHMG at the templating position significantly promotes misincorporation by polη. The misincorporation frequency of the dTTP insertion opposite NHMG insertion is 0.1, which is 10-fold greater than that of the dG:dTTP insertion ( Table  1). The K m values of the dG:dTTP and NHMG:dTTP insertions are very similar (159 vs. 146 μM), while k cat of the dG:dTTP insertion is ~15-fold greater than that for the NHMG:dTTP insertion (74×10 −3 s −1 vs. 5×10 −3 s −1 ), indicating the error-prone incorporation opposite NHMG is driven by increased k cat .

Structure of polη incorporating a non-hydrolyzable dCTP analog opposite templating NHMG.
To gain structural insights into polη catalyzing across the major adduct induced by nitrogen mustards, we determined a ternary complex structure of polη incorporating a nonhydrolyzable dCMPNPP (dCTP* hereafter) opposite a templating NHMG in the presence Mn 2+ (Figure 3). Extensive attempts for crystallizing the polη-NHMG:dCTP* complex with magnesium ions were not successful. The nonhydrolyzable dCTP* nucleotide was used because it is isosteric to dCTP and its coordination to the active-site metal ions is essentially identical to that of dCTP [34,44]. The ternary complex was crystallized in the P6 1 space group with the cell dimension of a = b = 98.9 Å, c = 82.1 Å, α = β = 90.0°, and γ = 120° and one polη was found in the asymmetric unit. The structure, which was refined to 2.69 Å with R work = 18.6% and R free = 23.8%, displayed the conserved secondary structures and the four characteristic domains (thumb, palm, finger, and little finger) of the Y-family DNA polymerases. The statistics for data collection and the refinement are summarized in Table. 2.
The crystal structure of polη in complex with a templating NHMG paired with an incoming dCTP* reveals, for the first time, the base pairing property of a nitrogen mustard-induced DNA lesion in the nascent base pair site of DNA polymerase. A 2F o -F c electron density map contoured at 1σ clearly shows the presence of NHMG and dCTP* in the replication site of polη ( Figure 4A). NHMG and dCTP* form Watson-Crick base pairing with three hydrogen bonds ( Figure 4B and 4C). The distances of interbase hydrogen bonds are 2.6 Å for between N2 of NHMG and O2 of dCTP*, 2.8 Å for between N1 of NHMG and N3 of dCTP*, and 3.3 Å for between O6 of NHMG and N4 of dCTP* ( Figure 4B). The distance between the C1´ carbons of NHMG and dCTP* is 10.4 Å and the λ angles are 53.3° (NHMG) and 59.2° (dCTP*), which are similar to the values observed in normal Watson-Crick base pairings [45]. The formation of three hydrogen bonds with Watson-Crick geometry indicates that the keto tautomer of NHMG is involved in the base pairing with dCTP. The propeller twist angle of the purine ring of NHMG and the pyrimidine ring of dCTP* is ~25° (Figure 4D), which is significantly greater than the values observed in normal Watson-Crick base pair. The nitrogen half-mustard moiety bends at the C2 position of NHMG and is positioned on the 5´ side of template strand ( Figure 4D). The hydroxyethyl moiety of NHMG is placed above the imidazole ring of NHMG and forms van der Waals interaction with the purine ring of NHMG ( Figure 4D). The phenyl ring of the nitrogen half-mustard moiety is almost perpendicular to the purine ring of NHMG. The aromatic moiety of NHMG points toward the Arg61-Trp64 loop in the finger domain.
The presence of NHMG at the templating position greatly changes DNA conformation near the lesion, particularly at the binding site of N+1 base of the template strand ( Figure 5). The nitrogen half-mustard moiety engages in extensive van der Waals interactions with the Arg61-Trp64 loop of polη ( Figure 5), which disrupts the possible stacking interaction between the bases at the N and N+1 positions. The nitrogen half-mustard moiety enters an open cleft generated by the Arg61-Trp64 loop and interacts with Ser62, Met63, and Trp64 ( Figure 5A and 5B). In the published polη structure with dG:dCTP (PDB ID:4O3N), the templating dG stacks with an adjacent 5´ base ( Figure 5C) [46]. In the NHMG structure, the templating NHMG does not stack with dT on the 5´ side. Instead, the nitrogen half-mustard moiety of NHMG occupies the space where the N+1 template base resides in the dG:dCTP structure, thereby precluding the stacking interaction between the templating base and the 5´ base. The N+1 template dT in the polη-NHMG:dCTP structure swings ~180° from the position in the polη-dG:dCTP structure and points toward the little finger domain of polη, residing where the N+2 template base is positioned in the dG:dCTP structure. This conformational difference suggests that, during the extension across NHMG, the N+1 template dT must undergo a large conformational reorganization to base pair with incoming nucleotide. The bulky NHMG moiety could also interfere the extension step by sterically interacting with the N+1 template dT. The large conformational change of the base 5´ to NHMG may contribute to the lack of extension in polη-mediated bypass of NHMG ( Figure  2A). The extension past NHMG could be conducted by other TLS polymerases such as polζ [47], which is specialized to extend distorted base pairs.
The nitrogen half-mustard moiety is oriented toward the 5´ end of the template and almost perpendicular to the purine ring of NHMG. The N-phenyl moiety of NHMG does not engage in stacking interaction with the adjacent bases. The terminal hydroxyl group of the nitrogen mustard monoadduct forms two hydrogen bonds with the backbone carbonyl groups of Gly46 and Ile47 ( Figure 5B). To accommodate the arylalkyl moiety of NHMG, the backbone carbonyl group and the sidechain hydroxy group of Ser62 moves by ~1.5 Å and ~3.0 Å, respectively, relative to the positions of Ser62 in the polη-dG:dCTP structure ( Figure 5C).
In the catalytic site of the polη-NHMG:dCTP* structure, despite the formation of Watson-Crick NHMG:dCTP base pairing, only the nucleotide-binding metal ion (the B-site metal ion) is observed, which may explain the low catalytic efficiency of the NHMG:dCTP insertion by the enzyme. This is in contrast with published polη-dG:dCTP* [46] and polη-N7MeG:dCTP* [39] structures showing both metal ions in the catalytic site. Superposition of the polη-NHMG:dCTP* structure with the published polη-dG:dCTP* structure reveals that the conformation of the incoming dCTP* is essentially identical ( Figure 5C). The conformation of the purine ring of NHMG, however, differs significantly from that of dG, resulting in a greater twist angle value (25°) for NHMG:dCTP* base pair. This canonical but non-optimal base pairing would contribute to the reduced insertion efficiency across NHMG lesion. The favorable interaction of the nitrogen half-mustard moiety with the Arg61-Trp64 loop would lead to a high propeller-twist angle, which could decrease K m for dCTP insertion opposite NHMG and deter binding of metal ion at the metal-A site.
The overlay of the polη-NHMG:dCTP* and published polη-N7MeG:dCTP* (PDB ID: 6UI2) [39] structures reveals that the bulky nitrogen half-mustard group may induce a large conformational change of the templating bases in the N+1 and N+2 positions ( Figure 5D). The polη structure with templating N7MeG and incoming dCTP* shows the conformations of DNA and protein are virtually identical to those of the polη-dG:dCTP structure. The N+1 dT engages in π-π stacking with N7MeG, which forms Watson-Crick hydrogen bonds with dCTP with a moderate propeller twist angle (~10°). These indicate the steric bulkiness of the templating N7-alkyl group plays a role in the large conformational reorganization of the template strand.

DISCUSSION
Aromatic nitrogen mustards such as chlorambucil and melphalan are known to produce nitrogen half-mustard-G adducts as major lesions [6,7]. Our NHMG adduct is structurally very similar to the nitrogen half-mustard-G lesion formed by melphalan. To gain structural insight into how TLS polymerases bypass aromatic nitrogen half-mustard-G adducts, we built a modeled structure where polη incorporates dCTP opposite melphalan-induced NHMG adduct ( Figure 6). The modeled structure predicts melphalan-induced NHMG may be accommodated in the replicating base site of polη by generating favorable polar and nonpolar interactions. The melphalan moiety would enter the open cleft generated by the Arg61-Trp64 loop and the amino acid moiety of melphalan would make hydrogen bond interactions with the phosphate backbone of the templating DNA and the amino acid residues in Arg61-Trp64 loop. We therefore propose our polη-NHMG:dCTP* structure represents a close approximation of the structure of polη bypassing the aromatic nitrogen mustard-induced NHMG lesions.
The polη-NHMG:dCTP* complex represents the first crystal structure of DNA polymerase bypassing nitrogen mustard-induced DNA lesions. In contrast, there are published structures of TLS polymerase incorporating nucleotide opposite bulky N7G adducts formed by phenanthriplatin and aflatoxin B1 [48,49]. Phenanthriplatin is a monofunctional platinumbased anticancer agent with an aromatic phenanthridine moiety. Unlike the NHMG adduct, phenanthriplatin-N7G adduct is efficiently bypassed by polη [48]. Comparison of the polη-NHMG:dCTP* structure with published polη-phenanthriplatin-N7G:dCTP* (PDB ID:4Q8E) structure reveals that the Arg61-Trp64 loop may play a role in accepting bulky major-groove lesions in the enzyme active site ( Figure 7A). The overall conformation of the phenanthriplatin-N7G complex is similar to that of the NHMG complex. The phenanthridine moiety of the phenanthriplatin-N7G engages in van der Waals interaction with Ser62 and Trp64, suggesting that the Arg61-Trp64 loop may stabilize bulky N7G lesions at the replicating base pair site. The flexibility of the Arg61-Trp64 loop ( Figure 5C) would play an important role in accepting a wide variety of bulky major-groove lesions. In the published polη-phenanthriplatin-N7G:dCTP* structure, unlike in the polη-NHMG:dCTP* structure, both A-and B-site metal ions are present. This is consistent with the efficient incorporation of dCTP opposite the phenanthriplatin-N7G by polη. The crystal structure of Dpo4 incorporating dCTP* opposite aflatoxin B1 (AFB1)-N7G adduct [49] is significantly different from the polη-NHMG:dCTP* structure ( Figure 7B). Dpo4 lacks an Arg61-Trp64like open cleft near the replicating base pair site and has a spacious, solvent accessible active site. The aflatoxin B1 moiety engages in π-π stacking interaction with the modified guanine and incoming dCTP*.
Our polη-NHMG:dCTP* complex was crystallized in the presence of Mn 2+ but not with Mg 2+ , highlighting the effect of metal cofactor on the crystallization of polymerase-DNA complexes. Mn 2+ cofactor has more lenient coordination requirement than Mg 2+ and is more tolerant to the changes (e.g., distortion, misalignment, mutation) in the active site of DNA polymerases [50,51]. For example, in the case of TLS DNA polymerase Dpo4, replacing Mg 2+ with Mn 2+ dramatically increases the catalytic efficiency of dGTP insertion opposite thymine [52]. In addition, the use of Mn 2+ cofactor facilitates the bypass of cyclobutane pyrimidine dimer (CPD) lesion by polη and polι [53]. The fact that polη-NHMG:dCTP* complex was crystallized only with Mn 2+ suggests that the presence of templating NHMG causes structural distortion in the active site of polη, which could be tolerated by Mn 2+ but not Mg 2+ .
The bulky nitrogen half-mustard moiety may primarily contribute to the inefficient incorporation of dCTP opposite NHMG by polη. The introduction of the small N7methylguanine lesion at the templating position only slightly reduces the catalytic efficiency of polη and Solfolobus solfataricus Dpo4 [23], indicating TLS DNA polymerases can readily traverse a small N7-alkylguanine lesion. The presence of NHMG and N7methylguanine at the templating position increases the frequency of dTTP misincorporation 7 to 10-fold (Table 1), highlighting the promutagenic properties of N7-alkylG lesions. Replication across N7-alkylG by TLS polymerases would, therefore, promote G-to-A mutations, which have been shown to be induced by bulky N7-alkylG lesions in vivo [54].
The high promutagenicity of NHMG lesion suggests the NHMG-induced mutagenesis could be a contributing factor for the development of secondary cancers often found in patients receiving NM-based chemotherapy [6,9,10].
N7-alkylG adducts, which exhibit half-lives of several hours to days in duplex DNA [55], can undergo spontaneous depurination to generate abasic sites, which promotes G to T transversions [56]. The primary N7-alkylG lesions with a small alkyl group (e.g. methyl, ethyl) have been considered non-mutagenic [57,58]. However, recent studies show N7-methylG facilitates dTTP misincorporation opposite the lesion [23,39]. In addition, the bulky AFB1-N7G adduct has been shown to predominantly induce G to T mutations in primate cells [59], which appears to be caused by intercalation of aflatoxin moiety on the 5´ side of the damaged guanine rather than AFB1-N7G-induced apurinic sites. Indeed, the crystal structure of Dpo4 bound to AFB1-N7G adduct demonstrates the intercalation of the AFB1 moiety above the 5′-face of the modified guanine [49]. Interestingly, nitrogen halfmustard-N7G adducts formed by the acridine half-mustard mutagen ICR-191 strongly induce +1 and +G frame-shift mutations as well as G to A transitions [54,60], indicating that the N7-alkylG adducts of AFB1 and ICR-191 have different mutagenesis mechanisms. Nevertheless, the observed potent mutagenicity of some bulky N7-alkylG lesions (e.g. ICR-191-N7G) in vivo suggests that the presence of bulky N7-alkylG in DNA may have negative effects on genome integrity. For example, guanine N7-alkylation has been shown to increase the acidity of N1-H of guanine [61,62], which in turn could destabilize N7-alkylG:C base pairing and promote N7-alkylG:T and N7-alkylG:A mispairs during replication. Bulky N7-alkylG lesions such as AFB1-N7G can be removed by base and/or nucleotide excision DNA repair pathways [17,63]. If the N7-alkylG lesions are not detected prior to replication, TLS polymerases such as polη could involve in bypassing the lesions. Our studies show the bulky NHMG at the templating position slows catalysis by TLS polymerases, but it increases the misincorporation frequency of dTTP and dATP opposite the lesion, which could contribute to G to A transitions and G to T transversions, respectively.

CONCLUSIONS
The first kinetic and structural studies of NHMG provide insights into how TLS polymerases bypass aromatic NM-induced DNA lesions. We incorporated a 2´-F-mediated transition destabilization strategy to enable a site-specific incorporation of NHMG in DNA for further study. The presence of NHMG at the templating position greatly reduces the dCTP insertion efficiency of polη, suggesting TLS DNA polymerases would slowly catalyze across bulky N7-alkylG mono-adducts formed by aromatic nitrogen mustards (e.g., melphalan, chlorambucil). The templating NHMG increases the misincorporation frequency of dATP ~30-fold, implicating NHMG may promote G to T transversions. A crystal structure of polη-NHMG:dCTP* ternary complex shows a Watson-Crick NHMG:dCTP* base pair with a high degree of propeller twist, which appears to be caused by the insertion of the nitrogen half-mustard moiety of NHMG into an open cleft created by the Arg61-Trp64 loop. This underscores a possible role of the Arg61-Trp64 loop in accommodating bulky major groove N7-alkylG lesions including those produced by aromatic nitrogen mustards. Overall, the results reported here further our understanding of the miscoding potential of the major NMinduced DNA lesions as well as the impact of bulky N7-alkylG on the catalysis by TLS polymerases.
Photon Source with the support of GM/CA. GM/CA@APS has been funded in whole or in part with Federal funds from the National Cancer Institute (ACB-12002) and the National Institute of General Medical Sciences (AGM-12006   (A) A denaturing urea PAGE gel for nucleotide incorporation opposite NHMG by polη. Polη (5 nM), NHMG-containing 12-mer template (50 nM), 8-mer primer (50 nM), and incoming nucleotides (9 to 2,500 μM) were used, and the reaction lasted for 10 min. The formation of an N+1 band was negligible even for dATP, which is a correct insertion for the extension. X denotes NHMG. (B-D) Denaturing gel images for the insertion kinetics of dCTP, dTTP, and dATP. NHMG and dCTP* are shown in magenta and black, respectively. The nitrogen half-mustard moiety of the NHMG adduct is shown in magenta spheres. The nucleotide-binding metal ion complexed with dCTP* is shown in a green sphere. The template except for NHMG and the primer are shown in yellow and orange, respectively. The conserved four domains are indicated. The loop containing Arg61, Ser62, Met63, and Trp64 (R61-W64 loop) from the finger domain is colored in blue.  (A) Van der Waals interactions of the nitrogen-half mustard moiety with Ser62, Met63, and Trp64. Distances are indicated as double-headed arrows. (B) Van der Waals interactions of the nitrogen-half mustard moiety with Ser62, Met63, and Trp64 from another perspective. (C) Comparison of the conformations of the templating base and the N+1 base in the polη-NHMG:dCTP* structure and the polη-dG:dCTP* structure (PDB code: 4O3N). The conformational shift of the dT at the N+1 position is shown in a curved arrow. The shift of Ser62 is indicated as arrows. (D) Comparison of the conformations of the templating base and the N+1 base in the polη-NHMG:dCTP* structure and the polη-N7MeG:dCTP* structure (PDB code: 6UI2). The conformational shift of the dT at the N+1 base is shown in a curved arrow. The shift of Ser62 is indicated as blue arrows. The melphalan-induced nitrogen half mustard-G lesion is shown in magenta and an incoming dCTP* is shown in green. The amino acid moiety of melphalan-NHMG adduct would form additional hydrogen bonds with the phosphate backbone of template strand and the Arg61-Trp64 loop. d R free = Σ|F(obs)-F(calc)|/ΣF(obs), calculated using 5% of the data.