DNA damage, repair and mutational spectrum

The integrity and stability of DNA is essential to life since it stores genetic information in every living cell. Chemicals from the environment will assault DNA to form various types of DNA damage, ranging from small covalent crosslinks between neighboring DNA bases as seen in cyclobutane pyrimidine dimers, to big bulky adducts derived from benzo[a]pyrene. This resultant damage will lead to replication block and mutation if remain unrepaired and will eventually cause cancer or other genetic diseases. The work presented in this dissertation has illustrated the important role of the AlkB family DNA repair enzymes in cancer and Wilson’s Disease. In addition, we discovered these enzymes can modify epigenetic markers that affect DNA regulation. We also studied sequence-dependent conformational heterogeneity of aminobiphenyl adduct on DNA replication. The AlkB family DNA repair enzyme is a family of α-ketoglutarate (αKG)and non-heme iron-dependent dioxygenases. Among all the homologs in this family, human ALKBH2 and ALKBH3, and E. coli AlkB have been proved to be the major enzymes that directly remove the alkyl adducts from alkylated DNA bases like 3-methylcytosine (3mC) and 1-methyladenine (1mA). These DNA adducts will cause strong replication block and mutagenicity in cell if AlkB enzymes are suppressed by toxicants. Cancerassociated mutations often lead to perturbed cellular energy metabolism and accumulation of potentially harmful oncometabolites. Chiral molecule 2hydroxyglutarate (2HG) and its two stereoisomers (Dand L-2HG) have been demonstrated to competitively inhibit several αKGand iron-dependent dioxygenases, including ALKBH2 and ALKBH3. In this work, we carried out detailed kinetic analyses of DNA repair reactions catalyzed by ALKBH2, ALKBH3 and the bacterial AlkB in the presence of Dand L-2HG in both double and single stranded DNA contexts. We not only determined kinetic parameters of inhibition, including kcat, KM, and Ki, but also correlated the relative concentrations of 2HG and αKG previously measured in tumor cells with the inhibitory effect of 2HG on the AlkB family enzymes. Both Dand L2HG significantly inhibited the human DNA repair enzymes ALKBH2 and ALKBH3 under pathologically relevant concentrations (73-88% for D-2HG and 31-58% for L2HG inhibition). This work provides a new perspective that the elevation of either Dor L-2HG in cancer cells may contribute to an increased mutation rate by inhibiting the DNA repair carried out by the AlkB family enzymes and thus exacerbate the genesis and progression of tumors. Another type of inhibitor of AlkB is toxic metals, such as, copper. Disturbed metabolism of copper ions can cause diseases, such as Wilson’s disease (WD). In this work, we investigated the inhibitory effect of Cu(II) ion on the AlkB family DNA repair enzymes, include human ALKBH2, ALKBH3, and E. coli AlkB proteins. None of the three proteins were significantly inhibited under normal cellular copper concentrations. But under WD related condition, we observed the activities of all three enzymes were strongly suppressed (inhibition from 95.2 to 100.0%). We also noted the repair efficiency under ds-DNA condition is less susceptible than ss-DNA to the inhibition. AlkB can repair many alkylated DNA bases including 3mC, 1mA, 3metheylthymine, 1-methylguanine, ethenoadenine and ethenocytosine. But in this work, we found a new DNA base substrate for AlkB, 5-Methylcytosine (5mC). 5mC in DNA CpG islands is an important epigenetic biomarker for mammalian gene regulation. It is oxidized to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5carboxylcytosine (5caC) by the ten-eleven translocation (TET) family enzymes, which are also α-KG/Fe(II)-dependent dioxygenases. In this work, we demonstrate that the epigenetic marker 5mC is biochemically modified to 5hmC, 5fC, and 5caC by ALKBH2, ALKBH3, and AlkB. Theoretical calculations indicate that these enzymes may bind 5mC in the syn-conformation, placing the methyl group comparable to 3methylcytosine, the prototypic substrate of AlkB. This is the first demonstration of the AlkB proteins to oxidize a methyl group attached to carbon, instead of nitrogen, on a DNA base. These observations suggest a broader role in epigenetics for these DNA repair proteins. Besides alkyl DNA adducts, there are bulky DNA adducts existing in human. Bulky organic carcinogens are activated in vivo and subsequently react with nucleobases of cellular DNA to produce adducts. Some of these DNA adducts exist in multiple conformations that are slowly interconverted to one another. Different conformations could contribute to different mutagenic and repair outcomes. Unfortunately, studies on the conformation-specific inhibition of replication, which is more relevant to cell survival, are scarce; this is presumably due to difficulties in studying the structural dynamics of DNA lesions at the replication fork. It is challenging to capture the exact nature of replication inhibition by traditional end-point assays, since they usually detect either the ensemble of consequences of all the conformers or the culmination of all cellular behaviors, such as mutagenicity or survival rate. One article reported an unusual sequence-dependent conformational heterogeneity involving FABP-modified (4′fluoro-4-aminobiphenyl) DNA under different sequence contexts. There were 67% Btype (B) conformation and 33% stacked (S) conformation in TG1*G2T sequence; whereas, 100% B conformation was observed in TG1G2*T sequence. In this study, we applied primer extension assay to compare the inhibition models between these two FABP-modified DNA sequences. We utilized a combination of surface plasmon resonance (SPR) and HPLC-based steady-state kinetics to reveal the differences in terms of binding affinity and inhibition with polymerase between these two conformers (67%B:33%S and 100%B). The conformational heterogeneities from these two sequences lead to different types of inhibition on replication.

oxidized to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5carboxylcytosine (5caC) by the ten-eleven translocation (TET) family enzymes, which are also α-KG/Fe(II)-dependent dioxygenases. In this work, we demonstrate that the epigenetic marker 5mC is biochemically modified to 5hmC, 5fC, and 5caC by ALKBH2, ALKBH3, and AlkB. Theoretical calculations indicate that these enzymes may bind 5mC in the syn-conformation, placing the methyl group comparable to 3methylcytosine, the prototypic substrate of AlkB. This is the first demonstration of the AlkB proteins to oxidize a methyl group attached to carbon, instead of nitrogen, on a DNA base. These observations suggest a broader role in epigenetics for these DNA repair proteins.
Besides alkyl DNA adducts, there are bulky DNA adducts existing in human. Bulky organic carcinogens are activated in vivo and subsequently react with nucleobases of cellular DNA to produce adducts. Some of these DNA adducts exist in multiple conformations that are slowly interconverted to one another. Different conformations could contribute to different mutagenic and repair outcomes. Unfortunately, studies on the conformation-specific inhibition of replication, which is more relevant to cell survival, are scarce; this is presumably due to difficulties in studying the structural dynamics of DNA lesions at the replication fork. It is challenging to capture the exact nature of replication inhibition by traditional end-point assays, since they usually detect either the ensemble of consequences of all the conformers or the culmination of all cellular behaviors, such as mutagenicity or survival rate. One article reported an unusual sequence-dependent conformational heterogeneity involving FABP-modified (4′fluoro-4-aminobiphenyl) DNA under different sequence contexts. There were 67% B-type (B) conformation and 33% stacked (S) conformation in TG1*G2T sequence; whereas, 100% B conformation was observed in TG1G2*T sequence. In this study, we applied primer extension assay to compare the inhibition models between these two FABP-modified DNA sequences. We utilized a combination of surface plasmon resonance (SPR) and HPLC-based steady-state kinetics to reveal the differences in terms of binding affinity and inhibition with polymerase between these two conformers (67%B:33%S and 100%B). The conformational heterogeneities from these two sequences lead to different types of inhibition on replication.        Adduct m1A is used here as an example to show the steps of enzymatic catalysis. b) The generation of D-and L-2HG and mechanisms of inhibition to the AlkB family DNA repair enzymes.
Oligonucleotides containing the methylated bases m1A and m3C were selected as substrates for the repair reactions because they are most efficiently repaired by these three enzymes. 24

RESULTS
To test the inhibitory effect of D-and L-2HG on the AlkB family enzymes, we first chemically synthesized oligonucleotides by site-specifically incorporating m1A and m3C, the major substrates of the AlkB family enzymes. 24,43 We also expressed and purified recombinant human ALKBH2 and ALKBH3 proteins, and the E. coli AlkB protein. 55 Then, we performed kinetic experiments to determine the kcat and KM of the three enzymes as they repair the two adducts in both ds-and ss-DNA. After that, we measured the Ki of D-and L-2HG on the repair reactions, and finally evaluated the inhibitory effect of the oncometabolites in the concentration range reported to occur in certain human cancers.
Oligonucleotide Synthesis and Protein Purification. Two 16mer oligonucleotides containing m1A and m3C were chemically synthesized with the sequence 5'-GAAGACCTXGGCGTCC-3' (X denotes the alkylated base). 43 After HPLC purification, the identity of the oligonucleotides was confirmed by comparing the theoretical m/z of the oligonucleotides with the observed m/z from high resolution LC-MS (Table S1). The genes for E. coli AlkB and its human homologs ALKBH2 and ALKBH3 were cloned into pET28a+ expression vector; the incorporation of the correct sequences was confirmed by sequencing the corresponding plasmids. The three proteins were then expressed in E. coli hosts, isolated and purified by affinity chromatography as described in Experimental Section. 55 Enzymatic Assay for Measuring Kinetic Constants. For each enzymatic reaction, the adduct-containing oligonucleotide was incubated with the necessary cofactors for the AlkB reaction: Fe(II), αKG, and ascorbic acid (see Experimental Section) in either ss-or ds-DNA. Below, m1A will be used as an example to explain the HPLC analyses.
For the ss-DNA reactions, the starting material 16mer m1A (1.5 min in Figure 2a) and product 16mer A (2.8 min in Figure 2b Therefore, we adopted a longer complementary oligonucleotide (23mer Tcp, 5.6 min in Figure 2c), which provided a similar repair efficiency as the 16mer complementary oligonucleotide. In the analysis of the 23mer reaction, the dsDNA of starting material (23mer Tcp plus16mer m1A, 7.5min in Figure 2c) and the dsDNA of product (23mer Tcp plus 16mer A, 7.7min in Figure 2c) still could not be fully separated under the HPLC condition. Consequently, we designed another 23mer oligonucleotide that was fully complementary to 23mer Tcp (23mer A, 5.5min in Figure 2d). After the dsDNA reaction with 23mer Tcp, 23mer A was added to the reaction mixture, and the mixture was heated to 80 °C for 10 min and then slowly cooled down to room temperature. The addition of 23mer A allowed the 23mer Tcp formed perfect dsDNA with 23mer A (9.1 min in Figure 2d), thus releasing 16mer m1A and 16mer A from their previous complementarity with 23mer Tcp. Under these conditions, the 16mer m1A and 16mer A in the dsDNA repair reaction were well separated and quantified by the HPLC analyses ( Figure 2d). A similar analytical strategy was successfully applied to m3C dsDNA repair reactions. For ALKBH2 repair of m1A in ds-DNA (Table 1, Table S2, and Figure S1), the kcat of αKG was 2.5 ± 0.1 min -1 and the KM was 7.3 ± 0.9 µM, which are comparable to the literature reported kinetics parameters of other αKG dependent enzymes. 7,25,56-61 The kcat/KM value of ds-repair reaction (0.34 min -1 ·µM -1 ) shows that the repair was more efficient than in ss-DNA (0.28 min -1 ·µM -1 ), which agrees with the literature on the reported strand preference of ALKBH2. 59 The kinetic data of ALKBH2 repair of m3C showed a similar trend (Table 1, Table S2, Table S4 and Figure S2). In contrast to ALKBH2's preference for ds-DNA substrates, the kinetic parameters of ALKBH3 repair of ds-DNA substrates could not be measured due to the low conversion ratio even with very high enzyme loading, such as 5.0 µM of ALKBH3 to 5.0 µM substrate.
These results confirm the previously reported preference of ALKBH3's repair of ss-DNA substrates. 48,55 The kinetic factors of the E. coli AlkB protein were also measured and the kcat and KM values agreed well with the literature reported kcat and KM of the reactions ( Table 1, Table S2 and S3, and Figure S3 and S4). 59  For the inhibition of the ALKBH2 repair reaction on m1A in ds-and ss-DNA, the Ki values for D-2HG are 280 ± 61 μM and 405 ± 61 μM, respectively (Table 2, Figure 3a and 3b, Table S5). These data indicate that D-2HG has a stronger binding affinity for the complex of ALKBH2 with ds-DNA than ss-DNA. For L-2HG reactions, the Ki values are similar but smaller (stronger inhibition) than with the corresponding D-2HG reactions (Table 2, Figure 3d and 3e, Table S5). The Ki values of N-OG show much stronger inhibition (with about 10 times more potency, Table 2 (Table 2). For the inhibition of AlkB-catalyzed reactions, there is no clear trend in the inhibitory potency between the D-and L-2HG; N-OG, however, is a stronger inhibitor than either of 2HG isomers. We also measured the IC50 of D-and L-2HG on the three enzymes (Table S7); in general, the IC50 values correlate well with the Ki values.  To make our experiments more relevant with regard to the anticipated cellular concentrations of metabolites/oncometabolites observed in human tumors, we also evaluated the extent inhibition of the ALKBH2 and ALKBH3 repair reactions by varying the ratios of D-or L-2HG to αKG. For D-2HG inhibition, we tested a ratio of concentrations for D-2HG:αKG = 373:1, which was observed in glioma patients with IDH mutations (detailed information see the Discussion section). 6 The concentration of αKG was fixed at 100 µM to make sure that the kinetic analyses reflected steady state catalysis ( Figure S5). We found that the repair efficiencies of ALKBH2 and ALKBH3 were 73-88% inhibited under such conditions, (Figure 3c, Table 3). For L-2HG inhibition, we tested a ratio of L-2HG:αKG= 28:1, which was reported in patients with kidney cancers (see Discussion section). 10 We found 48-58% of ALKBH2 and 31-40% of ALKBH3's activity was inhibited under this condition. These results suggest that the strong inhibition on DNA repair observed in the in vitro experiments may also occur in tumor cells of cancer patients.  7 Crystal structures of histone demethylases show that D-2HG binds to the same site as αKG in the catalytic center. 7 We tested the competition between 2HG and αKG in the DNA repair reactions. Using ALKBH2 repair of m1A as an example, the repair ratios without adding 2HG were controlled to be around 60% under different αKG concentrations (0.1, 0.5 and 1.0 mM, Figure 4 and Table S6). For the inhibition reactions, D-2HG was added at a fixed concentration (10 mM) in the reaction mixture, which contained ALKBH2 and necessary cofactors. Then, different concentrations of αKG were added and mixed. After that the reaction was initiated by adding the oligonucleotide substrates. When 0.1 mM αKG was present, the conversion decreased to 22%. When 0.5 mM and 1.0 mM αKG were added, the repair ratio increased to 35% and 38%, respectively ( Figure 4 and Table S6). This observed trend of reactivity recovery is consistent with the notion that D-2HG acts as a competitive inhibitor in the αKG-dependent DNA repair reactions. 7 Similar recovery patterns were observed for all other D-and L-2HG inhibition reactions on all three enzymes ( Figure 4 and Table S6). correspond to a concentration ratio between D-2HG and αKG of 373 to 1. 6,7 Under this ratio condition, the repair activities of ALKBH2 and ALKBH3 were 73-88% inhibited ( Table 3). The concentrations of L-2HG and αKG on average in kidney cancer cells are 1.15 µmol/g and 0.0484 µmol/g, respectively, which corresponds to a concentration ratio between L-2HG and αKG of 28 to 1. 10 Under this ratio condition, ALKBH2 and ALKBH3's repair activities are 31-58% inhibited (Table 3). Although the relative concentration of L-2HG (1.15 µmol/g) is more than 10 times lower than D-2HG (15.5 µmol/g), the ALKBH2 and ALKBH3 enzymes are still soundly inhibited by L-2HG partially due to the higher binding affinity of L-2HG (i.e., lower Ki) than D-2HG ( Table   2). The extent of inhibition in both cases was measured when the concentration of αKG was fixed at 100 µM, to ensure steady state catalysis. However, at lower concentrations of αKG, (i.e., 50 or 20 µM), the efficiency of adduct repair decreased even further. The cellular concentrations of αKG are typically around 40 to 50 µM (0.0415 and 0.0484 µmol/g or mM) in cancer patients, 6,10 which are near to the 50-100 µM range used in our experiments. Our data also show that, consistent with competitive inhibition of 2HG, the inhibition activity in the repair reaction reflects primarily the ratio between 2HG and αKG. ALKBH2 and ALKBH3 are enzymes that repair alkyl DNA damage; hence, inhibition of DNA repair leads to alkylation product accumulation, less cellular survival, and increased mutations, which affect the resistance/sensitivity balance to alkylating chemotherapeutics. The elevation of both D-and L-2HG in cancer cells may contribute to the increased mutation rate and exacerbate tumorigenesis and progression.
Strand Preference of the Three Repair Enzymes. According to the literature, ALKBH2 prefers to repair m1A and m3C in ds-DNA, whereas ALKBH3 and AlkB prefer to repair those adducts in ss-DNA. 24,30,55 We tested the repair activity in both ss-DNA and ds-DNA substrates in this study. The experimental results reported in this paper provide a strong kinetic basis for the previous observations. For ALKBH2, the kcat/KM values of ds-repair are higher than the repair in ss-DNA (Table 1). By contrast, the kcat/KM values of AlkB repair are higher for ss-DNA substrates than for ds-DNA substrates ( Table 1). For ALKBH3, we were only able to measure the kinetic parameters for ss-repair, as the ds-repair reactions were too inefficient to evaluate.
These results agree with and add quantitative detail to previous observations that ALKBH3 strongly prefers to repair adducts in ss-DNA.

Other αKG/Fe(II)-Dependent Enzymes may be Inhibited by Oncometabolites.
There are about 80 proteins in the αKG/Fe(II)-dependent enzyme family, including jmjc, prolyl hydroxylase, TET, and the AlkB family enzymes. [21][22][23] Studies have demonstrated that D-and L-2HG inhibit jmjc and TET family proteins. 7,62 In addition to 2HG, intermediates in the TCA cycle such as succinate and fumarate have also been found to exhibit higher-than-normal concentrations in different cancer cells (Figure 1b). 18 Given their structural similarities to αKG and 2HG, these metabolites could also perturb αKGdependent enzymatic activities in the cell, especially DNA repair processes that are related to the AlkB family enzymes. Systematic studies are needed to explore these possibilities and correlate these biochemical results with clinical observations. These studies are also pivotal for the design and development of therapeutic agents that target the abnormal metabolic pathways of cancer.

Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: XXX.
Tables of molecular weights and m/z values, initial rate for kinetic studies, and IC50 values of inhibition reactions; figures for steady-state kinetics, repair percentage of reactions, and inhibition curves of different repair enzymes (PDF)

Corresponding Author
Mailing address:

Notes
The authors declare no competing financial interest.

ACKNOWLEDGMENTS
The                                             To test the inhibitory effect of Cu(II) ion on the AlkB family enzymes, we first sitespecifically synthesized a 16mer oligonucleotide containing m3C, the major substrate of the AlkB family enzymes. 12, 13 We expressed and purified recombinant human ALKBH2 and ALKBH3 proteins, and the E. coli AlkB protein. 14 Then, we evaluated the repair efficiency of the enzymes ( Figure S1) and the inhibitory effect of Cu(II) ion under different conditions. It has been reported that the cellular concentrations of "free" copper and iron ions are very limited. Under normal cellular conditions, the concentration of free copper ions is 1.5 µM, which is 1.5 fold greater than free iron ions (1.0 µM). For the WD patients, the level of free copper ion is increased to 24.4 µM. 15 In this work, we tested different Fe(II) ion concentrations and selected 5.0 µM to ensure the maximal efficiency of repair by the AlkB family enzymes ( Figure S2); this concentration was similar to the Fe(II) ion concentrations used in other inhibition studies. 6 The inhibitory effects of Cu(II) ion were tested from 0.0 to 100.0 µM, among which 7.5 µM of Cu(II) was used to mimic normal cellular condition. Though the cellular levels of copper ion under WD condition vary considerably from different reports, all of the levels are more than 10 fold higher than the normal cellular concentration. 1,2,15 To make this work rigorous and relevant to pathological disease condition, we used 75.0 µM (10 fold) of copper ion to represent the WD condition.
For each enzymatic reaction, the m3C-containing oligonucleotides were incubated with the necessary cofactors for the AlkB reaction: Fe(II) ion, αKG, and ascorbic acid (see  Table   1). The IC50 values for AlkBH3 and AlkB were also measured ( Table 1). From the IC50 tests, we found the IC50 values are higher in ds-DNA condition than ss-DNA condition for all three enzymes, despite the preference of AlkBH3 and AlkB to repair m3C in ss-DNA and AlkBH2 preferring repair in ds-DNA. 12 These results shows the repair under ds condition is less susceptible than ss condition to the inhibition from Cu(II) ion. One possibility is that copper ion is harder to access to the active site of an enzyme and replace its iron ion due to the more crowded environment under ds-DNA binding condition.  We then tested the inhibition of Cu(II) ion under normal physiological condition with the concentration at 7.5 µM. We found that none of the three proteins were significantly inhibited under this concentration, with the highest inhibition at 7.0% for ALKBH3 repairing m3C in ds-DNA. (Table S2 in Table S2).
In this work, we used m3C, the best substrate of AlkB, to demonstrate the AlkB family enzymes are strongly inhibited by high concentration of Cu(II) ion under WD condition.
Other substrates, such as the etheno DNA lesions will be tested both in vitro and in vivo in the future, which will provide a new explanation to the previously observed elevation of etheno adducts in Wilson's disease. 1 Considering that there are about eighty Fe(II)/αKG-dependent enzymes in human body, 11 it is highly possible that some of those enzymes could be inhibited in a similar mechanism by the high concentrations of the copper ion under Wilson's disease condition.

Notes
The authors declare no competing financial interest.

ACKNOWLEDGMENTS
The       Figure S18b) to align the 5-methyl moiety for oxidation, which is similar to how the TET family enzymes bind 5mC. This paper is the first work to demonstrate the ability of the AlkB family enzymes to oxidize a methyl group that is attached to carbon, instead of nitrogen, on a DNA base.

Synthesis of oligonucleotides containing 5mC and other modifications.
All oligonucleotides used in this study were synthesized by solid-phase synthesis. (23)(24)(25)(26) The 5mC and other phosphoramidites was purchased from Glen Research. Synthetic oligonucleotides were purified by reverse-phase HPLC and identified by electrospray ionization mass spectrometry (ESI-MS).
Protein expression and purification. The expression and purification of ALKBH2, ALKBH3 and AlkB proteins were described by previous published papers. (23,24) ALKBH2 and ALKBH3 in storage buffer containing 50 mM N- The 23mer oligonucleotide is complementary to the 16mer sequence plus 7 nucleotides longer with the sequence of 5'-CTGGGACGCCGAGGTCTTCACTG-3'. The rest steps of dsDNA reaction were the same as ssDNA reaction described above. All reactions were carried out in triplicates.

Oligonucleotide digestion.
The procedures of oligonucleotide digestion to deoxyribonucleoside were adopted from published procedure. (27)  were assigned epsilon protonation, while H72, H131, and H187 were assigned delta protonation. For ALKBH2, H55, H59, H106, H144, and H228 were assigned epsilon protonation, H199 and H220 were assigned delta protonation, and H167 was modelled as cationic. All crosslinks induced to facilitate crystallization between DNA and protein were removed for both complexes. The G169C, C67S, C165S, and C192S mutations to ALKBH2 were reverted to generate the wild-type enzyme, and the overhanging DNA ends (residues 259, 260, and 284) were truncated. To generate the AlkB-5mC or ALKBH2-5mC complexes, the anti and syn conformations of the 5mC nucleotide were overlaid onto the 3mC substrate in the ALKBH2-or AlkB-DNA complex using chemical intuition to minimize steric clashes between the bound substrate and active site amino acids.
Each complex was modelled using the AMBER parm14SB force field. Parameters Representative structures (clustering methodology below) were chosen from these preproduction simulations to initiate the final 100 ns production simulations. Coordinates were saved from each simulation every 5 ps and analyzed over the same interval.
The CPPTRAJ (37) (Figure 2a and Table S1). In the same -3 charge envelope, we also observed the ions of 5mC + Na + (1632.608) and 5mC + K + (1637.928). The observed m/z values of these species are consistent with the corresponding calculated m/z values (Table S1).
Previously, our lab has purified the three AlkB family enzymes mentioned above and  Table S1). Other oligonucleotide species containing metal ions, such as Na + and K + , were also observed. The complete assignments of the major species generated from MS analyses are summarized in Figure S1 and Table S1. The product oligonucleotides were digested into single nucleosides, analyzed by LC-MS, and compared with standard nucleosides to confirm the oxidative products generated from enzymatic reactions are indeed 5hmC, 5fC, and 5caC (see the Product oligonucleotides analyses section in SI, Figure S2 to S10). Also, to make sure the oxidations were carried out by AlkB and its homologs, we generated the catalytically inactive protein variants of AlkB: H131A, D133A, and H187A ( Figure S11). The three substituted amino acids in AlkB are the key residues that coordinate the Fe(II) ion. (13,38) The sequences of wild type and variant proteins were confirmed by trypsin digestion with MS analyses ( Figure S11 and Table S2). None of the AlkB variants showed any detectable oxidative product when reacted with 5mC ( Figure S12); these observations suggest that the oxidative modifications were carried out by AlkB and not by a contaminating enzyme.
For the oxidation of 5mC, the formation of products 5hmC, 5fC, and 5caC had different To probe the molecular basis by which the AlkB enzymes are able to oxidize 5mC, we performed molecular dynamics (MD) simulations to examine how 5mC is accommodated in the active sites of ALKBH2 and AlkB (Figure 4). We found that the 5-methyl of anti-5mC in our model is far from the Fe(IV)-oxo moiety (~5.3 Å for ALKBH2 in Figure 4a and ~7.9 Å for AlkB in Figure 4d; also in Figures S13-S14, and Tables S3-S6). In contrast, the distances between the Fe(IV)-oxo moiety and the 5methyl group in the models with syn-5mC bound to ALKBH2 or AlkB are much shorter for AlkB) through water (Tables S3-S4 and Figures S13-S14), as well as the Y122 hydroxy group in ALKBH2 (Table S3). These interactions likely facilitate oxidative catalysis by positioning the C5 methyl group near the Fe(IV)-oxo moiety (~3.6 -3.8 Å; Tables S5-S6).
Interestingly, a crystal structure of TET2 co-crystallized with 5mC-containing DNA reveals syn-5mC in the active site, (40) and the reported χ torsion angle is consistent with that predicted for syn-5mC in ALKBH2/AlkB ( Figure S15 and Tables S5-S6). Based on our combined experimental and theoretical data, we propose that the AlkB family enzymes are able to oxidize 5mC bound only in the syn-conformation. To provide insight into the variable activities of ALKBH2 and AlkB for subsequent nucleobase oxidation (Figure 3), MD simulations were performed on syn-5hmC and 5fC bound in the active sites. For 5hmC, the C5 substituent is further from the Fe(IV)-oxo moiety for ALKBH2 (~4.5 Å, Figure S16 and Tables S5) compared to AlkB (3.4 Å, Figure S17 and Tables S6), which is consistent with the relative low abundance of the 5fC product for ALKBH2 (Figures 3, S19 to S21 and Tables S3 to S6). For 5fC, increased flexibility of the bound nucleobase may permit enhanced catalysis and generate more 5caC for ALKBH2 comparing to AlkB (Figures 3a, 3b, and S22). In                             anti-3mC, anti-5mC, syn-5mC, syn-5hmC, and syn-5fC in the ALKBH2 complex. Table S4. Summary of important hydrogen bonds formed during MD simulations of anti-3mC, anti-5mC, syn-5mC, syn-5hmC, and syn-5fC in the AlkB complex. Table S5. Summary of important distances and dihedral angles adopted during MD simulations of anti-3mC, anti-5mC, syn-5mC, syn-5hmC, and syn-5fC in the ALKBH2 complex. Table S6. Summary of important distances and dihedral angles adopted during MD simulations of anti-3mC, anti-5mC, syn-5mC, syn-5hmC, and syn-5fC in the AlkB complex. Table S7. Time to half maximum conversion (T1/2, min) for the oxidation of different alkyl DNA modifications by the three AlkB family enzymes.

Product oligonucleotides analyses
It was important to assure that the observed new oligonucleotide species indeed contain the proposed oxidative products. The product oligonucleotides were digested with benzonase into single nucleosides, which were analyzed by LC-MS and compared with standard nucleosides for retention time and molecular weight. In the digestion of the product generated from the AlkB reaction, a nucleoside was eluted out at 7.  Figures S2 and S5). Similarly, nucleosides 5fC and 5caC were also discovered in the mixture of the digested reaction product and their identities were confirmed by comparing to standard nucleosides ( Figures S2, S7 to S10).
These observations support that the oxidative products generated from enzymatic reactions are indeed 5hmC, 5fC, and 5caC.

Product distribution for the oxidation of 5mC
For the oxidation of 5mC, the formation of products 5hmC, 5fC, and 5caC had different distribution patterns for the three enzymes; and the three enzymes had different preferences of oxidation in ss-or ds-DNA reactions (Figure 3). The relative amount of each product in the final reaction mixture was quantified according to the abundance of its corresponding ion at -3 charge state in the MS analysis. The product oligonucleotides containing 5hmC was the dominant species for the reactions of ALKBH2 and 3 in both ds-and ss-DNA (Figures 3a and 3b). For example, 5hmC represents 78.7% of the total amount of the three oxidative products in the reaction of ALKBH2 with 5mC in ds-DNA; 5fC and 5caC are quantified as 10.9% and 10.4% correspondingly (Figure 3a).
For the oxidation of 5mC by AlkB, 5hmC is the major species (51.2%) in ds-DNA reaction; however, 5fC is the most abundant species (65.3%) in ss-DNA reaction (Figures 3a and 3b). We were also able to quantify the yield of the oxidative products by comparing them to the total oligonucleotide species. The ratios of the oxidative products show ALKBH2 prefers to oxidize 5mC in ds-DNA (8.7%) over in ss-DNA

Modifications on different alkyl substrates by the AlkB enzymes
We compared the conversion efficiency of 5mC by the three AlkB family enzymes to other known methyl substrates of the AlkB proteins. Those substrates include 3mC, 1mA, 3mT, and 1mG; and all of them have been demonstrated to be repaired by AlkB both in vitro and in vivo.(1, 3) The results show the conversion of 5mC to the corresponding oxidative products are comparable to the demethylation of 3mT and 1mG; and the reactions of these three methyl modifications are slower than 3mC and 1mA ( Figure S23 and Table S7). For example, T1/2 of the reactions for 5mC (16.9 min) is slightly shorter than 3mT (17.4 min) but longer than 1mG (1.7 min) in the AlkB reactions. For the ALKBH2 reactions, T1/2 for 5mC are comparable to 3mT and 1mG (all between 16.4 to 18.7 min). For the ALKBH3 protein, 5mC is oxidized faster than 3mT and 1mG with about 5 minutes shorter in T1/2 than the other two ( Figure S23 and Table S7). Those biochemical results may indicate similar modification could happen in cell: the modification of 5mC is comparable to the weaker AlkB substrates 3mT and 1mG but less efficiently than the stronger substrates 3mC and 1mA.(3)

Simulations of AlkBH2 and AlkB bound to 3mC, 5mC, 5hmC, and 5fC
To validate that the simulated structures of the ALKBH2 or AlkB complexes are consistent with the observed catalytic activity, we examined the active site conformations adopted upon binding of both enzymes to anti-3mC-, syn-5mC-, syn-5hmC-, or syn-5fC-containing DNA. Similar DNA-protein interactions form when 3mC and 5mC are bound. Specifically, the 3mC N 4 amino group forms direct hydrogen bonds with the D174 or E175 sidechains when bound to ALKBH2 ( Figure S16b and Table S3), and direct or water-mediated hydrogen bonds with D135 and E136 when bound to AlkB (Table S4 and Figure S17). As a result, the N3 methyl group of 3mC occupies an equivalent active site position as the C5 methyl group of 5mC for both enzymes (Figures S16b and S17b), with a distance between the Fe(IV)-oxo and the 3mC methyl moiety of ~ 3.3 Å for both enzymes (Tables S5-S6), and the syn-5mC moiety of ~3.6 Å for AlkB and 3.8 Å for ALKBH2. Thus, the simulation data is consistent with the proposal that the AlkB family of enzymes is able to oxidize 5mC in the syn orientation.
Similar to ALKBH2 or AlkB bound to 5mC, syn-5hmC is stabilized by direct or watermediated hydrogen bonds between the N 4 amino group of the nucleobase and the carboxylate moieties of D174 and E175 (D135 and E136 in AlkB; Tables S3-S4 and Figure S20). For 5hmC bound by ALKBH2 or AlkB, hydrogen bonds are formed between the C5 substituent hydroxy group and the Fe(IV)-oxo moiety (occupancy = 34% for ALKBH2 and 52% for AlkB; Tables S3-S4 and Figure S21). In ALKBH2, an additional hydrogen bond is formed between the C5 substituent hydroxy group and D173 (59% , Table S3), which results in one orientation of the C5 substituent throughout the simulation ( Figure S21a). In contrast, the C5 substituent hydroxy group does not interact with D133 in AlkB, which results in two conformations of 5hmC within the active site ( Figure S21b-d). More importantly, the C5 substituent is at an optimum distance from the Fe(IV)-oxo moiety in AlkB (~3.4 Å), which matches the prototypic substrate (3mC; ~3.3 Å), while the equivalent distance is longer in the ALKBH2-5hmC complex (~4.5 Å; Tables S5-S6 and Figure 3), which would impede catalysis. This helps explain the observed higher abundance of 5fC for AlkB compared to ALKBH2 catalyzed oxidation on 5hmC, although several other factors could also be significant such as DNA binding and unique base flipping mechanism for each enzyme.
As discussed for 5mC and 5hmC, the syn-conformation of 5fC is stabilized by hydrogen bonds between the N 4 amino group of the nucleobase and carboxylate sidechains of active site residues (Tables S3-S4 and Figure S20). When bound to either ALKBH2 or AlkB, 5fC is planar due to an intramolecular hydrogen bond between the N 4 amino group and the carbonyl of the C5 substituent. A hydrogen bond also forms between the carbonyl of the C5 substituent and an active site arginine, which is notably more persistent for AlkB (R210, occupancy = 100.0%) than ALKBH2 (R254, occupancy = 29.2%, Tables S3-S4 and Figure S20). As a result, the distance between the hydrogen atom of the C5 substituent and the Fe(IV)-oxo group is longer for AlkB (~3.8 Å) compared to ALKBH2 (3.3 Å; Tables S5-S6 and Figure S22). Although the difference in distance for 5fC bound by ALKBH2 and AlkB is not as significant as observed for 5hmC, QM/MM studies on TET2-catalyzed oxidation of 5mC, 5hmC, and 5fC reveal that the initial hydrogen atom abstraction step is rate limiting, with the barrier increasing as 5mC < 5hmC < 5fC.(4) This suggests that the position of the substituent relative to the Fe(IV)-oxo group is even more crucial in the case of 5fC. Thus, our predicted structures correlate with the lower abundance of 5caC relative to 5fC for AlkB compared to ALKBH2 (Figure 3). Nevertheless, as discussed for 5hmC, nucleotide recognition and the base flipping mechanism could also play different roles in oxidative conversion of 5fC to 5caC.  Table S1.            Table S1.                 Figure S19) or C5 (5mC, 5hmC, or 5fC; Figure S19). b χ of bound pyrimidine defined as ∠(O4′C1′N1C2) (Figures S18 and S19). Table S6. Summary of important distances and dihedral angles adopted during MD simulations of anti-3mC, anti-5mC, syn-5mC, syn-5hmC, and syn-5fC in the AlkB complex.  Figure S19) or C5 (5mC, 5hmC, or 5fC; Figure S19). b χ of bound pyrimidine defined as ∠(O4′C1′N1C2) (Figures S18 and S19).

INTRODUCTION
The human genome is under constant assault from the environment. DNA-damageinduced mutations are known to trigger chemical carcinogenesis [1][2][3][4][5]. Therefore, understanding the biological responses of cells to DNA mutations is important.

4-Aminobiphenyl (ABP) is a major etiological agent of human bladder carcinogen
and a potent urinary-bladder carcinogen in experimental animals. As such, commercial production of ABP is banned, however, exposure to ABP can still take place from cigarette smoke. ABP is activated by cellular N-acetyltransferase to produce dG-C8substituted adduct as a major DNA lesion (dG-C8-ABP, Figure 1b) [12,[16][17][18][19][20]. The majority of human bladder cancer has a mutation in the p53 gene. Compared with other cancers, the ABP-induced mutations are more evenly distributed along the p53 gene and the mutation hotspots occur at both CpG, such as codons 175, 248, and 273, and non-CpG sites, such as codons 280 and 285, the latter two being unique mutational hotspots for bladder and other urinary tract cancers [21]. The major induced mutation was is G to T transversion mutation. Translesion synthesis (TLS) over dG-C8-ABP in two different sequences (CCG*GAGGC and CCGGAG*GCC, G*=dG-C8-ABP), which represent codon 248 and 249 sequences of the human p53 tumor suppressor gene, respectively, has confirmed that codon 248 is both hot spot of ABP adduct formation and G to T mutation [22]. These results suggest that the efficiency of TLS over dG-C8-ABP is influenced by the surrounding DNA sequences. The structurally similar liver

Desorption/ Ionization-Time of Flight Mass Spectrometry)
The FABP-modified biotin-31-mer oligonucleotides were characterized by exonuclease enzyme digestion followed by MALDI-TOF MS analysis in accordance with the published procedures [27,36]. In the present case, 5′-3′ exonuclease digestion on DNA was difficult to carry out due to the binding hindrance of 5'-biotin motif to the enzyme. Figure S1b shows the MALDI-TOF MS spectra of the 3′-5′ snake venom The m/z = 6,788 (theoretical 6,787) fragment, which persisted from 6 min to 10 min was assigned to the G1*-FABP-modified 21-mer. These results confirm the first eluting peak (peak 1) from the HPLC profile ( Figure S1a) is biotin-31-mer TG1*G2T. Figure   S1c presents the MALDI-TOF MS spectra of the peak 2 on HPLC with 3′-5′ The 84-and 85-mer biotinylated oligonucleotides were purified and identified by 15% denaturing polyacrylamide gel (Hercules, CA) [27]. Figure S2

HPLC-based Steady-state Kinetics
We conducted steady-state experiments to investigate the impact of conformational heterogeneity on nucleotide insertion kinetics [37]. The E. coli exonuclease-deficient Kf-exo − was used for single-nucleotide incorporation. Although the modified base could pair with dCTP to complete the primer extension reaction induced by Kf-exo − , the reaction efficiency was much poorer than the regular DNA template. The change of efficiency is represented by the enzyme kinetic parameters, Km and kcat, and the results are summarized in Table 1. The bulky C8 adduct on guanine does not directly block the Watson-Crick base pairing, but it could either physically interfere with the dNTP binding pocket in Kf-exo − when the FABP-G holds an "S" conformation; or distort Kfexo − structure in the ternary complexes and influence the geometry at the active site of forming phosphodiester bonds when FABP-G holds the "B" conformation ( Figure 1a).
In both scenarios, the bulky adduct acts as an inhibitor, but in two different ways ( Figure   3). In order to apply the inhibition kinetic model, the whole primer extension assay was performed by maintaining the concentration of inhibitor (FABP-containing DNA duplex), and varying the concentration of substrate, dNTP (dCTP or dATP). dATP was  Figure   3b) and n-3 for G1* (Figure 3a). Figure S3a presents the results at 10 min. In the control, ~70% of the 8-mer was extended to 9-mer. With adduct on G1, ~40% of the 8-mer was converted to 9-mer, indicating a moderate pre-lesion effect at G1*. However, G2* blocked ~80% of the 8-mer converting to 9-mer, which indicates a much stronger prelesion effect than that of G1*. These results are not surprising because the 8-mer primer is closer to the G2* adduct than the G1* adduct. In the 11-mer ( Figure S3b), unmodified control exhibited 90% of dATP insertion. However, only ~20% of the 11-mer was elongated to 12-mer opposite the G1* lesion. This finding indicates a strong post-lesion effect. Moreover, only ~55% of the 11-mer was converted to 12-mer in the n+1 for G2*.
These results suggest significant retardation of insertion close to the lesion site.

DISCUSSION
We previously conducted [23] systematic spectroscopic, thermodynamic, and chip binding ( 19 F-NMR, CD, DSC, and SPR) studies for the extension of 16-mer TG1*G2T and TG1G2*T (G*=FABP) sequences. These protein free model systems mimic a translesion synthesis of the bulky FABP lesion in two very distinctive sequencedependent conformational heterogeneities at G1 (67%B:33%S) and G2 (100%B). The results indicate that the sequence-dependent conformational complexities appear to persist at various elongation positions, including the ss/ds junction. The B-conformer is a major thermodynamic stabilizer in duplex settings, whereas the S-conformer is a destabilizer. However, the opposite is the case for adduct at the ss/ds junction. In particular, the S-conformation promotes lesion stacking with nascent base pairs at the replication fork. SPR results reveal that the S-conformation increases the binding affinity with the complementary strands in the order of G1* > G2*. In the present work, we examined the effects of these unusual sequence effects on Klenow polymerase binding (binary and ternary) and nucleotide insertion kinetics.

Improvement on Model Hairpin Oligonucleotide Construction
We previously reported a construction of biotinylated hairpin-based templateprimer strand for DNA-polymerase SPR binding studies [27,[39][40]. The general strategy was to ligate a biotinylated arylamine-modified 31-mer sequence with 52-mer hairpin DNA to form an 83-mer hairpin. This was then followed by addition of a ddNTP to the 3′-end to prevent potential primer elongation. However, the ddNTP addition step was low yielding and the purification of ligated products was difficult. In the present study, we succeeded in direct ligation of a biotinylated modified 31-mer with ddNTPcontaining 53-and 54-mer hairpin DNA. This process improved the yield significantly (from ~10% to ~ 30%), and the products were readily separated by HPLC. Also, previously, individual dNTPs (100 μM) were mixed with varying concentrations of Kfexoin sample buffers and injected over the surface without any dNTP in the running buffer. This would create a complication resulting in two variables, concentrations of dNTPs and Kf-exo - [39]. In the present work, we circumvented the complication by adding individual dNTP in running buffer and introducing Kf-exodirectly onto the chip surface. This system has only one variable, thus increasing the system stability and accuracy for the translesion synthesis.

Lesion and Sequence Effects on SPR Binding Affinities and Kinetics
Tight binding of Kf-exowith the unmodified dG control was observed when the correct nucleotide dCTP was presented in both 85-mer and 84-mer. This result is in a good agreement with expectation and shows high nucleotide selectivity (KD dCTP ≪ KD dTTP < KD dATP ~ KD dGTP). Meanwhile, a remarkably tighter binding of Kf-exowas found for the FABP-modified sample relative to the control: i.e., the KD of the 85-mer G1* and 84-mer G2* interactions are 8.3 and 5.6 fold higher than the corresponding unmodified controls, respectively. Adduct-induced tighter binding affinity with Kfexohas been reported [27,41] and may be due to the interactions between the bulky FABP and the nearby hydrophobic amino-acid residues in the active site of the Kf-exo -. In the ternary system, nucleotide selectivity is low. In particular, the KD of the dCTP at the opposite G1* and G2* are only ~5-and ~2-fold tighter than the incorrect dATP, dGTP, and dTTP, respectively. This is a strong lesion effect. We observed that the usual 1:1 model SPR simulation did not provide a clean fit of dCTP for G1* in the ternary system. It is possible that the present DNA adduct-Kf complex exist in multiple stages due to the FABP-induced S/B conformational heterogeneity at G1* replication fork.
However, in most cases conformational changes in protein are much faster than SPR time scale and thus additional studies would be necessary to confirm these possibilities.
Our HPLC-based steady-state kinetics data indicate competitive inhibition for the S/B-conformer replication fork at G1*. In other words, the S/B conformational heterogeneity at G1* is inhibitory to replication probably due to some unfavorable clash of the bulky FABP lesion with the incoming dCTP in a competitive manner. The unfavorable interactions may be caused by the competition of S-conformer with dCTP opposite of dG. However, the exclusively B-conformeric G2* accommodates well for the incoming dCTP, resulting in non-competitive inhibition. The equal affinity (Km) between G2* and its unmodified dG reveals that the B-conformation in the major grove may not interfere with the insertion of the correct dCTP. This finding might also explain the greater KD of dCTP at G1* over G2* of the SPR results and indicates a weaker binding affinity for G1* over G2*. Alternatively, the B-conformer may not interfere with the Watson-Crick base pair, but it may alter the native conformations or hinder the formation of phosphodiester bond. The S-conformer portion of G1* cannot provide a proper Watson-Crick base pair for replication and it may need to convert back to the Bconformation to be replicated properly. Thus, the S-conformer may function as a competitive inhibitor for replication. We observed a competitive-inhibition model for G1* but a non-competitive-inhibition model for G2*. These types of intermediate interactions may be a necessary step in the DNA polymerase proofreading process as well.
heating up to 70 °C and slowly cooling down to room temperature. Five different concentrations of dCTP or dATP (0, 12.5, 25, 50, and 100 µM) were used to initiate nucleotide insertion. All reactions were quenched by adding 10mM EDTA followed by immediately denaturing at 80 °C. The initial velocity of each reaction within the steadystate range was then obtained by performing every reaction within a short time period

SPR Measurements
FABP-modified Hairpin Template/Primer Constructs. 5′-Biotinylated 31-mer containing dG-C8-FABP in the "-TG1G2T-" sequence context was used in SPR analysis following the reported procedures [27,[39][40]. The two FABP-modified biotinylated 31-mer G1* and G2* adducts were separated by RP-HPLC and characterized by MALDI-TOF MS [36]. The 84-mer and 85-mer hairpin-template-primer were prepared by following the reported protocols [27,39]. Briefly, two different lengths of hairpin DNA sequences (53-and 54-mer) were phosphorylated at their 5′-ends, but their 3′-ends were modified with ddA and ddC, respectively, to prevent further primer elongation ( Figure 2a). The biotinylated 31-mer modified G1* adduct and 54-mer hairpin were desalted by G-25 spin columns and annealed together by heating to 95 °C for 5 min and cooling to room temperature. The mixture solution was ligated in a buffer containing 4,000 units of T4 ligase enzyme for 16 h at room temperature. The resulting 85-mer oligonucleotide was purified in 15% denaturing polyacrylamide gel and extracted using the electroelution method followed by desalting by again using G-25 spin columns. The corresponding biotinylated 85-mer was finally purified by RP-HPLC. G2* adduct biotinylated 84-mer and unmodified biotinylated 84-and 85-mers were also prepared similarly. All the 84-and 85-mer oligonucleotides were identified by the 15% denaturing polyacrylamide gel ( Figure S2).

Immobilization of Streptavidin (SA) on CM5 S Chip and DNA Coating. SPR
experiments were carried out using Biacore T200 (GE Healthcare, Piscataway, NJ).
The SA via the amine coupling kit was immobilized on flow cells on the carboxymethylated dextran-coated CM5 S chip by following the reported procedures [27,39]. After SA immobilization at around 2,000 RU on the flow cells, the chip surface was washed with 50 mM NaOH for 60 s five times to reach below 20 RU.
Then, the running buffer was injected three times, and the system was equilibrated with running buffer for 2 h. The 84-and 85-mer unmodified and modified G1* and G2* biotinylated DNA sequences (2 nM) were injected over the flow cells (2 to 4) for 90-120 s to achieve 5-6 RU. The surface was stabilized with running buffer for 3 h before conducting SPR binding affinity experiments.

Real-time Kinetic Analysis by SPR.
The SPR system was first primed at least three times with running buffer and zeroconcentration injections to condition the plasmon surface. The DNA was coated on the SA surface of flow cell 2 to 4 (cell 1 as blank reference). Surface testing, regeneration buffer scouting, and mass transport limitation test were conducted prior to kinetic experiments following previous reports [27,39]. DNA coating around 4-5 RU did not show any impact of mass transport. The steady-state affinity analysis of Kf-exo − binding to unmodified and modified DNA was analyzed in the absence (binary) and presence of dNTPs (ternary) by varying Kf-exo − concentrations. For the binary system, Kf-exo − was injected without dNTPs over the DNA surface in varying concentrations (0-25 nM) and repeated twice as described previously [27].

CONCLUSIONS
In this paper, we conducted SPR and HPLC-based steady-state kinetics studies to probe adduct-induced conformation-dependent replication block during translesion synthesis (TLS). The FABP-modified DNA adduct adopts a mixture of B and S conformations in the TG1*G2T (67%B:33%S), but shows exclusively B conformation in the TG1G2*T sequence context [23]. According to the present Lineweaver-Burk enzyme inhibition model, the S/B-conformeric mix G1* adduct exhibits a competitive-inhibition, whereas the B-conformeric G2* adduct behaves as a non-competitive inhibitor in the nucleotide insertion step. These results indicate that the S-conformer may not be able to accommodate the incoming dCTP and exhibits a competitive behavior with incoming dNTPs, thus blocking replication. By contrast, the exclusive B-conformer G2* does not interfere with the Watson-Crick base pairing, resulting in a proper dCTP insertion. As such, the B-conformer shows a non-competitive-inhibition behavior. The SPR binding results implicate an adduct-induced tight binding with Kf-exoin a binary system. In the ternary system, nucleotide selectivity decreases when G1 and G2 are modified by FABP.
From these experiments, we observed the effect of conformational heterogeneity induced by bulky lesion on replication block.
It has been reported that compared with other cancers, the ABP-induced mutations are more evenly distributed along the p53 gene and the mutation hotspots occur through the genome with the major mutation being G to T transversion [21][22]. TLS over dG-C8-ABP in two different sequences (CCG*GAGGC and CCGGAG*GCC, G*=dG-C8-ABP), which represent codon 248 and 249 sequences of the human p53 tumor suppressor gene, respectively, has confirmed that codon 248 is a hot spot for adduct formation and G to T mutation. These results suggest that the efficiency of TLS over dG-C8-ABP is affected by the surrounding DNA sequences of the ABP lesion, consequently the B/S conformational heterogeneity as described here. Elucidation of conformation-specific bypass, mutational and repair processes over the ABP adducts in cell should clarify the molecular mechanisms underlying ABP-induced mutagenesis and carcinogenesis.
In the present paper, we demonstrated the combination of SPR binding and HPLC steady-state kinetics as a power tool in investigating FABP-induced conformational heterogeneity in TLS. This approach can be applied to studying other bulky DNA adducts.