Conserved residue lysine165 is essential for the ability of O6-alkylguanine-DNA alkyltransferase to react with O6-benzylguanine

The role of lysine(165) in the activity of the DNA repair protein, O(6)-alkylguanine-DNA alkyltransferase (AGT), and the ability of AGT to react with the pseudosubstrate inhibitor, O(6)-benzylguanine (BG), was investigated by changing this lysine to all other 19 possibilities. All of these mutants (except for K165T, which could not be tested as it was too poorly active for assay in crude cell extracts) gave BG-resistant AGTs with increases in the amount of inhibitor needed to produce a 50% loss of activity in a 30 min incubation (ED(50)) from 100-fold (K165A) to 2400-fold (K165F). Lys(165) is a completely conserved residue in AGTs from many species, and all of the mutations at this site also reduced the ability to repair methylated DNA. The least deleterious change was that to arginine, which reduced the rate constant for DNA repair by approx. 2.5-fold. Mutant K165R resembled all of the other mutants in being highly resistant to BG, with an ED(50) value for inactivation by BG>200-fold greater than wild-type. Detailed studies of purified K165A AGT showed that the rate constant for repair and the binding to methylated DNA substrates were reduced by 10-20-fold. Despite this, the K165A mutant AGT was able to protect cells from alkylating agents and this protection was not abolished by BG. These results show that, firstly, lysine at position 165 is needed for optimal activity of AGT towards methylated DNA substrates and is essential for efficient reaction with BG; and second, even if the AGT activity towards methylated DNA substrates is impaired by mutations at codon 165, such mutants can protect tumour cells from therapeutic alkylating agents. These results raise the possibility that the conservation of Lys(165) is due to the need for AGT activity towards substrates containing more bulky adducts than O(6)-methylguanine. They also suggest that alterations at Lys(165) may occur during chemotherapy with BG and alkylating agents and could limit the effectiveness of this therapy.

Active and alkylated human AGT structures: a novel zinc site, inhibitor and extrahelical base binding

Human O(6)-alkylguanine-DNA alkyltransferase (AGT), which directly reverses endogenous alkylation at the O(6)-position of guanine, confers resistance to alkylation chemotherapies and is therefore an active anticancer drug target. Crystal structures of active human AGT and its biologically and therapeutically relevant methylated and benzylated product complexes reveal an unexpected zinc-stabilized helical bridge joining a two-domain alpha/beta structure. An asparagine hinge couples the active site motif to a helix-turn-helix (HTH) motif implicated in DNA binding. The reactive cysteine environment, its position within a groove adjacent to the alkyl-binding cavity and mutational analyses characterize DNA-damage recognition and inhibitor specificity, support a structure-based dealkylation mechanism and suggest a molecular basis for destabilization of the alkylated protein. These results support damaged nucleotide flipping facilitated by an arginine finger within the HTH motif to stabilize the extrahelical O(6)-alkylguanine without the protein conformational change originally proposed from the empty Ada structure. Cysteine alkylation sterically shifts the HTH recognition helix to evidently mechanistically couple release of repaired DNA to an opening of the protein fold to promote the biological turnover of the alkylated protein.

Characterization of human polymorphic DNA repair methyltransferase

The O6-methylguanine-DNA methyltransferase (MGMT) is a critical defence against alkylation-induced mutagenesis and carcinogenesis. More than a 20-fold interindividual difference in the MGMT activity is known to exist among human cultured fibroblasts. We previously reported three allelic variants of the human MGMT gene, namely V1, V2, and V3. Both V1 and V2 carry amino acid substitutions, Leu84Phe and Trp65Cys, respectively, while V3 has a silent mutation. In order to reveal the pharmacogenetic and ecogenetic significance of polymorphism in the human MGMT gene, we investigated the in-vivo characteristics of V1 and V2 methyltransferase enzyme. Escherichia coli strain KT233 (ogt-, ada-) and mer- HeLa MR cells carrying a V1 sequence exhibited almost the same level of sensitivity against N-methyl-N'-nitro-N-nitrosoguanidine (MNNG), as did those with a wild-type sequence. The level of methyltransferase protein in those cells was essentially the same as for the wild-type and V1 samples. On the other hand, E. coli and human cells expressing V2 cDNA showed a significantly reduced level of survival. In these cells, V2 protein was hardly detected, even though mRNA was produced normally. An in-vitro translation experiment revealed that the V2 sequence had the potential to produce methyltransferase protein, as did the wild-type and V1 sequences. There was also evidence for a small amount of V2 protein being produced but rapidly degraded, thus implying that the V2 molecule is unstable in vivo. Using purified recombinant proteins, we estimated the kinetic values of wild-type and variant form of enzymes, which would support these views. From these results, we concluded that the wild-type and V1 protein have similar enzymatic and physicochemical properties, while V2 protein is considered to be unstable and rare.

Crystal structure of the human O(6)-alkylguanine-DNA alkyltransferase

The mutagenic and carcinogenic effects of simple alkylating agents are mainly due to O(6)-alkylation of guanine in DNA. This lesion results in transition mutations. In both prokaryotic and eukaryotic cells, repair is effected by direct reversal of the damage by a suicide protein, O(6)-alkylguanine-DNA alkyltransferase. The alkyltransferase removes the alkyl group to one of its own cysteine residues. However, this mechanism for preserving genomic integrity limits the effectiveness of certain alkylating anticancer agents. A high level of the alkyltransferase in many tumour cells renders them resistant to such drugs. Here we report the X-ray structure of the human alkyltransferase solved using the technique of multiple wavelength anomalous dispersion. This structure explains the markedly different specificities towards various O(6)-alkyl lesions and inhibitors when compared with the Escherichia coli protein (for which the structure has already been determined). It is also used to interpret the behaviour of certain mutant alkyltransferases to enhance biochemical understanding of the protein. Further examination of the various models proposed for DNA binding is also permitted. This structure may be useful for the design and refinement of drugs as chemoenhancers of alkylating agent chemotherapy.

Protein phosphorylation is a regulatory mechanism for O6-alkylguanine-DNA alkyltransferase in human brain tumor cells

The biochemical regulation of human O6-alkylguanine-DNA alkyltransferase (AGT), which determines the susceptibility of normal tissues to methylating carcinogens and resistance of tumor cells to many alkylating agents, is poorly understood. We investigated the regulation of AGT by protein phosphorylation in a human medulloblastoma cell line. Incubation of cell extracts with [gamma-32P]ATP resulted in Mg(2+)-dependent phosphorylation of the endogenous AGT. Immunoprecipitation after exposure of the cells to 32P-labeled inorganic phosphate showed that AGT exists as a phosphoprotein under physiological conditions. Western analysis and chemical stability studies showed the AGT protein to be phosphorylated at tyrosine, threonine, and serine residues. Purified protein kinase A (PKA), casein kinase II (CK II), and protein kinase C (PKC) phosphorylated the recombinant AGT protein with a stoichiometry of 0.15, 0.28, and 0.44 (mol phosphate incorporated/mol protein), respectively. Residual phosphorylation of the endogenous AGT by the PKs present in cell homogenates and phosphorylation of the recombinant AGT by purified serine/threonine kinases, PKA, PKC, and CK II reduced AGT activity by 30-65%. Conversely, dephosphorylation of cell extracts by alkaline phosphatases stimulated AGT activity. We also identified consensus phosphorylation motifs for many cellular kinases, including PKA and CK II in the AGT protein. These data provide the first and conclusive evidence of AGT phosphorylation and suggest that reversible phosphorylation may control the activity of this therapeutically important DNA repair protein in human normal and cancer cells.

Hyperthermostable protein structure maintained by intra and inter-helix ion-pairs in archaeal O6-methylguanine-DNA methyltransferase

The crystal structure of O6-methylguanine-DNA methyltransferase (EC 2.1.1.63) of hyperthermophilic archaeon Pyrococcuskodakaraensis strain KOD1 (Pk -MGMT) was determined by single isomorphous replacement method with anomalous scattering (SIRAS) at 1.8 A resolution. The archaeal protein is extremely thermostable and repairs alkylated DNA by suicidal alkyl transfer from guanine O6 to its own cysteine residue. Archaea constitute the third primary kingdom of living organisms, sharing characteristics with procaryotic and eucaryotic cells. They live in various extreme environments and are thought to include the most ancient organisms on the earth. Structural studies on hyperthermophilic archaeal proteins reveal the structural features essential for stability under the extreme environments in which these organisms live, and will provide the structural basis required for stabilizing various mesophilic proteins for industrial applications. Here, we report the crystal structure of Pk-MGMT and structural comparison of Pk-MGMT and methyltransferase homologue from Escherichia coli (AdaC, C-terminal fragment of Ada protein). Analyses of solvent-accessible surface area (SASA) reveals a large discrepancy between Pk-MGMT and AdaC with respect to the property of the ASA. In the Pk-MGMT structure, the intra-helix ion-pairs contribute to reinforce stability of alpha-helices. The inter-helix ion-pairs exist in the interior of Pk-MGMT and stabilize internal packing of tertiary structure. Furthermore, structural features of helix cappings, intra and inter-helix ion-pairs are found around the active-site structure in Pk-MGMT.

Reaction and binding of oligodeoxynucleotides containing analogues of O6-methylguanine with wild-type and mutant human O6-alkylguanine-DNA alkyltransferase

O6-Alkylguanine-DNA alkyltransferase (AGT) repairs DNA by transferring the methyl group from the 6-position of guanine to a cysteine residue on the protein. We previously found that the Escherichia coli Ada protein makes critical interactions with O6-methylguanine (O6mG) at the N1- and O6-positions. Human AGT has a different specificity than the bacterial protein. We reacted hAGT with double-stranded pentadecadeoxynucleotides containing analogues of O6mG. The second-order rate constants were in the following order (x10(-)5 M-1 s-1): O6mG (1.4), O6-methylhypoxanthine (1.6) > Se6-methyl-6-selenoguanine (0.1) > S6-methyl-6-thioguanine (S6mG) (0.02) >> S6-methyl-6-thiohypoxanthine (S6mH), O6-methyl-1-deazaguanine (O6m1DG), O6-methyl-3-deazaguanine (O6m3DG), and O6-methyl-7-deazaguanine (O6m7DG) (all <0.0001). Electrophoretic mobility shift assays were carried out to determine the binding affinity to hAGT. Oligodeoxynucleotides containing O6mG, S6mG and O6m3DG bound to AGT in the presence of competitor DNA with Kd values from 5 to 20 microM, while those containing G, S6mH, O6m1DG, and O6m7DG did not (Kd > 200 microM). These results indicate that the 1-, N2-, and 7- positions of O6mG are critical in binding to hAGT, while the 3- and O6-positions are involved in methyl transfer. These results suggest that the active site of ada AGT is more flexible than hAGT and may be the reason ada AGT reacts with O4mT faster than hAGT.

Alteration of the conserved residue tyrosine-158 to histidine renders human O6-alkylguanine-DNA alkyltransferase insensitive to the inhibitor O6-benzylguanine

The DNA repair protein O6-alkylguanine-DNA alkyltransferase (AGT) protects cells from alkylation damage. O6-Benzylguanine (BG) is a potent inactivator of human AGT (ED50 of 0.1 microM) that is currently undergoing clinical trials to enhance chemotherapy by alkylating agents. In a screen of AGT mutants randomly mutated at position glycine-160, we found that the double mutant Y158H/G160A protected Escherichia coli from killing by N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) even in the presence of BG and that the AGT activity of this mutant was strongly resistant to BG (ED50 of 180 microM). Because the single mutant G160A was not resistant to BG, this suggested that the presence of the charged histidine residue at position 158 was responsible. This hypothesis was confirmed by the construction of the single mutation Y158H. The Y158H-mutant AGT was slightly less active than wild-type AGT for the repair of methylated DNA in vitro, but it protected E. coli from killing by MNNG even in the presence of BG and had an ED50 for the inactivation by BG of 620 microM. In contrast, mutant Y158F had an ED5o of 0.2 microM. Previous studies (M. Xu-Welliver et al., Cancer Res., 58: 1936-1945, 1998) have shown that mutant P140K is highly resistant to BG (ED50 of >1200 microM). Models of human AGT suggest that the side chain of the lysine inserted into this mutant is close to tyrosine-158 and that the positively charged lysine side-chain may interfere with BG binding. The double mutants P140K/Y158H and P140K/Y158F resembled P140K and Y158H in being highly resistant to BG, but the use of a sensitive assay for reaction of BG with AGT indicated that their abilities to react were in the order P140K/ Y158H < P140K < P140K/Y158F. These results confirm that the presence of a positively charged residue close to the active site of human AGT renders it highly resistant to BG without substantially affecting activity toward methylated DNA substrates. Such mutants may limit the value of BG therapy if they arise in malignant cells during chemotherapy, but the mutant sequences may be useful for gene therapy approaches in which BG-resistant human AGTs are used to prevent hematopoietic toxicity. At least 28 AGT sequences (from 25 species) have now been described. In 25 of these, the position equivalent to 158 in the human AGT is also a tyrosine, and in the other 3, it is a phenylalanine. The importance of an aromatic ring side chain at this position is emphasized by previous studies (S. Edara et al., Carcinogenesis, 16: 1637-1642, 1995), which show that the replacement by alanine renders human AGT inactive. Our results show that histidine can also substitute for tyrosine at this position.

Synthesis of DNA oligonucleotides containing site-specifically incorporated O6-[4-oxo-4-(3-pyridyl)butyl]guanine and their reaction with O6-alkylguanine-DNA alkyltransferase

DNA pyridyloxobutylation occurs during the metabolic activation of the tobacco-specific nitrosamines, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) and N'-nitrosonornicotine (NNN). This pathway contributes significantly to the carcinogenic and mutagenic activity of these nitrosamines. In general, the chemical structure of pyridyloxobutyl adducts are not well understood. Recently, an AGT reactive pyridyloxobutyl adduct was identified as O6-[4-oxo-4-(3-pyridyl)butyl]guanine (O6-pobG). To better understand the importance of this adduct to the biological activity of pyridyloxobutylating agents, we developed a method for site-specifically incorporating O6-pobG into DNA oligonucleotides. They were synthesized using the phosphoramidite of the precursor 2'-deoxy-O6-{3-[2-(3-pyridyl)-1,3-dithian-2-yl]propyl}guanosine. The dithiane group was oxidatively removed with N-chlorosuccinimide in a final postoligomerization reaction to generate the desired product. Human AGT with a polyhistidine tag was able to repair the O6-pobG-containing DNA oligonucleotide, generating unmodified oligonucleotide. These results are consistent with an alkyl group transfer mechanism for the repair of O6-pobG by AGT.

Crystallization and preliminary X-ray crystallographic analysis of archaeal O6-methylguanine-DNA methyltransferase

Crystals of archaeal O6-methylguanine-DNA methyltransferase (MGMT) from hyperthermophilic archaeon Pyrococcus kodakaraensis strain KOD1 have been grown at room temperature using polyethylene glycol as a precipitant. The diffraction pattern of the crystal extends to 2.0 A resolution at room temperature upon exposure to Cu Kalpha radiation. The crystal belongs to the space group P212121 with unit-cell dimensions of a = 52.8, b = 86.6 and c = 39.9 A. The presence of one molecule per asymmetric unit gives a crystal volume per protein mass (Vm) of 2.3 A3 Da-1 and a solvent content of 48% by volume. A full set of X-ray diffraction data was collected to 2.0 A Bragg spacings from the native crystal.

Investigation of the role of tyrosine-114 in the activity of human O6-alkylguanine-DNA alkyltranferase

Tyrosine-114 is one of 13 totally conserved amino acids in all known sequences of O6-alkylguanine-DNA alkyltransferase (AGT). The importance of this amino acid in repair of alkylated DNA by AGT was studied by changing it to phenylalanine (F), alanine (A), threonine (T), or glutamic acid (E) in human AGT. The activities of the mutant proteins were then compared to those of the wild type with regard to abilities to do the following: (a) protect Escherichia coli from the methylating agent N-methyl-N'-nitro-N-nitrosoguanidine (MNNG); (b) repair methylated DNA in vitro; (c) bind to oligodeoxynucleotides containing O6-methylguanine; and (d) react with the low molecular weight pseudosubstrate, O6-benzylguanine. When expressed at high levels in E. coli strain GWR109, lacking endogenous AGT, the wild type and the Y114F mutant were highly effective in reducing mutations and cell killing by MNNG. The Y114A mutant had a much smaller protective effect, and mutants Y114T and Y114E were inactive. Purified preparations of all four AGT mutants showed an approximately similar degree (74-120-fold) of reduction in the rate of reaction with O6-benzylguanine. In contrast, the degree of reduction in activity toward methylated DNA substrates in vitro varied according to the mutation with the more conservative Y114F producing only a 30-fold reduction and the most drastic change of Y114E abolishing activity completely. Alteration Y114A produced a 1000-fold reduction whereas Y114T reduced activity by 10000-fold. All of the mutations affected the binding of AGT to single- or double-stranded oligodeoxynucleotides containing O6-methylguanine. The extent of increase in the Kd varied according to the amino acid with 2-5-fold (F), 7-11-fold (A), 167-200-fold (T), and 600-1000-fold (E) increases. These results are consistent with tyrosine-114 playing a role both in the binding of AGT to its DNA substrate and in facilitating the transfer of the alkyl group. It is probable that AGT resembles other DNA repair proteins in bringing about a "flipping out" of the target base from the DNA helix. Tyrosine-114 is therefore an excellent candidate for a key role in the interaction with the flipped O6-methylguanine. The results also show that when large amounts of AGT are produced in the cell, substantial decreases in the efficiency with which AGT can repair methylated DNA do not prevent the ability to protect E. coli from toxic alkylating agents. Mutant Y114F, whose activity was reduced by 30-fold, was equal to wild-type AGT in bringing about this protection.

Modified phased translation functions and their application to molecular-fragment location

Direct methods at high resolution have depended on the resolution of atomic like features in the map. At data resolutions more typical for protein structures (2-3 A) individual atoms may not be resolved, so larger features must be identified. At one extreme the whole molecule may be located using the diffraction magnitudes alone by the molecular-replacement method. At the other extreme it is possible to locate individual residues in a well phased map. In this paper an intermediate problem is addressed: the location of multi-residue fragments on the basis of weak phase information. An agreement function based on the mean-squared difference between model and map over a masked region is shown to be more effective than a simple overlap integral, and may be efficiently calculated by Fourier methods. The techniques are compared using poorly phased electron-density maps at approximately 3 A for the proteins RNAse and O6-methylguanine-DNA-methyltransferase.

Reaction of O6-benzylguanine-resistant mutants of human O6-alkylguanine-DNA alkyltransferase with O6-benzylguanine in oligodeoxyribonucleotides

Inactivation of the human DNA repair protein, O6-alkylguanine-DNA alkyltransferase (AGT), by O6-benzylguanine renders tumor cells susceptible to killing by alkylating agents. AGT mutants resistant to O6-benzylguanine can be made by converting Pro140 to an alanine (P140A) or Gly156 to an alanine (G156A). These mutations had a much smaller effect on the reaction with O6-benzylguanine when it was incorporated into a short single-stranded oligodeoxyribonucleotide. Such oligodeoxyribonucleotides could form the basis for the design of improved AGT inhibitors. AGT and mutants P140A and G156A preferentially reacted with O6-benzylguanine when incubated with a mixture of two 16-mer oligodeoxyribonucleotides, one containing O6-benzylguanine and the other, O6-methylguanine. When the 6 amino acids located in positions 159-164 in AGT were replaced by the equivalent sequence from the Escherichia coli Ada-C protein (mutant AGT/6ada) the preference for benzyl repair was eliminated. Further mutation incorporating the P140A change into AGT/6ada giving mutant P140A/6ada led to a protein that resembled Ada-C in preference for the repair of methyl groups, but P140A/6ada did not differ from P140A in reaction with the free base O6-benzylguanine. Changes in the AGT active site pocket can therefore affect the preference for repair of O6-benzyl or -methyl groups when present in an oligodeoxyribonucleotide without altering the reaction with free O6-benzylguanine.

Reversible folding of Ada protein (O6-methylguanine-DNA methyltransferase) of Escherichia coli

The multifunctional 39 kDa Escherichia coli Ada protein (O6-methylguanine-DNA methyltransferase) (EC 2.1.1.63), product of the ada gene, is a monomeric globular polypeptide with two distinct alkylacceptor activities located in two domains. The two domains are of nearly equal size and are connected by a hinge region. The Ada protein accepts stoichiometrically the alkyl group from O6-alkylguanine in DNA at the Cys-321 residue and from alkyl phosphotriester at the Cys-69 residue. This protein functions in DNA repair by direct dealkylation of mutagenic O6-alkylguanine. The protein methylated at Cys-69 becomes a transcriptional activator of the genes in the ada regulon, including its own. Each of the two domains functions independently as an alkyl acceptor. The purified homogeneous protein is unstable at 37 degrees C and spontaneously loses about 30% of its secondary structure in less than 30 min concomitant with a complete loss of activity. However, sedimentation equilibrium studies indicated that the inactive protein remains in the monomeric form without aggregation. Furthermore, electrospray mass spectroscopic analysis indicated the absence of oxidation of the inactive protein. This temperature-dependent inactivation of the Ada protein is inhibited by DNA. In the presence of increasing concentrations of urea or guanidine, the protein gradually loses more than 80% of its structure. The two alkyl acceptor activities appear to be differentially sensitive to unfolding and the phosphotriester methyltransferase activity is resistant to 7 M urea. The partial or complete unfolding induced by urea or guanidine is completely reversed within seconds by removal of the denaturant. The heat-coagulated protein can also be restored to full activity by cycling it through treatment with 8 M urea or 6 M guanidine. These results suggest that the nascent or unfolded Ada polypeptide folds to a metastable form which is active and that the thermodynamically stable structure is partially unfolded and inactive.

Probing of conformational changes in human O6-alkylguanine-DNA alkyl transferase protein in its alkylated and DNA-bound states by limited proteolysis

Human O6-alkylguanine-DNA alkyl transferase (hAGT) is a DNA repair protein that protects cells from alkylation damage by transferring an alkyl group from the O6-position of guanine to a cysteine residue in the active site (-PCHR-) of the protein. The structure of the hAGT protein (23 kDa) has been probed by limited proteolysis with trypsin and Glu-C endoproteases and analysis of the polypeptide fragments by SDS/PAGE. The native hAGT protein had limited accessibility to digestion with trypsin and Glu-C in spite of a number of potential cleavage sites. Initial cleavage by trypsin occurred at residue Lys-193 to give a 21 kDa polypeptide fragment, and this polypeptide underwent further cleavage at residues Arg-128 and Lys-165. These trypsin-cleavage sites became more accessible to digestion in the presence of double-stranded DNA (dsDNA), indicating that hAGT undergoes a change in its conformation on binding to DNA. However, the trypsin cutting site at the Arg-128 position was less available for digestion in the presence of single-stranded DNA (ssDNA), suggesting that the hAGT protein has a different conformation when bound to ssDNA compared with dsDNA. When protease digestion was carried out on wild-type protein, preincubated with the low-molecular-mass pseudosubstrate O6-benzylguanine, increased susceptibility to proteases was observed. A mutant C145A hAGT protein, which cannot repair O6-alkylguanine because the Cys-145 acceptor site in the active site of the protein is changed to Ala, showed identical trypsin cleavage to the wild type, but its digestion was not affected by O6-benzylguanine. These results suggest that alkylation of hAGT leads to an altered conformation. The acquisition of increased susceptibility to proteases upon DNA binding and alkylation demonstrates that hAGT undergoes considerable conformational changes in its structure upon binding to DNA and after repair of alkylation damage.

The potential role of glycine-160 of human O6-alkylguanine-DNA alkyltransferase in reaction with O6-benzylguanine as determined by site-directed mutagenesis and molecular modelling comparisons

O6-Alkylguanine DNA-alkyltransferase (ATase) repairs toxic, mutagenic and carcinogenic O6-alkylguanine (O6-alkG) lesions in DNA by a highly conserved reaction involving the stoichiometric transfer of the alkyl group to the active centre cysteine residue of the ATase protein. In the Escherichia coli Ada ATase, which is effectively refactory to inhibition by O6-benzylguanine (O6-BzG), the residue corresponding to glycine-160 (G160) for the mammalian proteins of this class is replaced by a tryptophan (W). Therefore, to investigate the potential role of the G160 of the human ATase (hAT) protein in determining sensitivity to O6-BzG, site-directed mutagenesis was used to produce a mutant protein (hATG160W) substituted at position 160 with a W residue. The hATG160W mutant was found to be stably expressed and was 3- and 5-fold more sensitive than hAT to inactivation by O6-BzG, in the absence and presence of additional calf-thymus DNA respectively. A similar, DNA dependent increased sensitivity of the hATG160W mutant relative to wild-type was also found for O6-methylguanine mediated inactivation. The potential role of the W160 residue in stabilising the binding of the O6-alkG to the protein is discussed in terms of a homology model of the structure of hAT. The region occupied by G/W-160 forms the site of a putative hinge that could be important in the conformational change that is likely to occur on DNA binding. Three sequence motifs have been identified in this region which may influence O6-BzG access to the active site; YSGG or YSGGG in mammals (YAGG in E. coli Ogt, YAGS in Dat from Bacillus subtilis), YRWG in E. coli Ada and Salmonella typhimurium (but YKWS in Saccharomyces cerevisiae) or YRGGF in AdaB from B. Subtilis. Finally,conformational and stereoelectronic analysis of the putative transition states for the alkyl transfer from a series of inactivators of hAT, including O6-BzG was undertaken to rationalise the unexpected weak inhibition shown by the alpha-pi-unsaturated electrophiles.

What structural features determine repair enzyme specificity and mechanism in chemically modified DNA?

A crucial question in repair is how do enzymes recognize substrates. In surveying the relevant literature, it becomes evident that there are no rules which can be clearly applied. At this time it appears that uracil glycosylase is the only repair enzyme for which all the known substrates can be rationalized on the basis of chemical structure. When surveying the multiplicity of substrates for m3A-DNA glycosylase, it is difficult, on the basis of present knowledge, to explain why 1,N6-etheno-A (epsilon A) is as good a substrate, if not better, than m3A for which the enzyme is named. There is no apparent unifying chemical structure which is required for recognition. It should also be noted that many studies of the mechanism of m3A-DNA glycosylase only utilized-N-3- and N-7-alkylpurines. On this basis, an electron-deficient purine, and later pyrimidine, was considered to be the recognition signal. Since epsilon A and Hx do not fall in this class, this is one illustration of why exploring new substrates becomes important in elucidating enzyme mechanisms. Ubiquitous enzymes, such as 5'-AP endonucleases, are present in both prokaryotes and eukaryotes. The primary function is the same, i.e., repair of an AP site which occurs through natural processes or from the action of DNA glycosylases. There are, however, completely unrelated substrates such as the exocyclic adducts pBQ-dC and pBQ-dG. pBQ-dC is repaired by both the human HAP1 and E. coli Exo III and Endo IV, while pBQ-dG is only repaired by the E. coli enzymes. Yet, when repair of these adducts occurs, it is by the same unusual pathway which differs from the usual base excision repair mechanism. This finding may ultimately not be as unusual as it now seems. The understanding of substrate recognition by repair enzymes, which can have different repair pathways, is complex. For example, three exocyclic derivatives which each have either the same modification (1,N4-epsilon dA and 3,N4-epsilon dC) or the same base with different modifying groups (3,N4-epsilon dC and 3,N4-pBQ-dC) are repaired by three separate enzymes and two mechanism (Figure 9). Investigators have also reported that two separate enzymes and pathways can be found for simple adducts such as m6G and O4T. It is not clear why, for these adducts, both MGMT and excision repair can be utilized. This could be visualized as a "backup" system and may be more common than now known. We cannot think like an enzyme or vice versa. In the absence of enough necessary information, we can only be descriptive. What information is necessary for further understanding? (1) More detailed structural studies of adducts in defined oligonucleotides would be useful. (2) New substrates should be explored. For example, is the mechanism for PBQ-dC (and pBQ-dG) repair unique? This involves guesswork and intuition. (3) For the adducts mentioned in this Perspective and others, understanding enzyme/substrate recognition will be facilitated by cocrystallography and site-directed mutagenesis. (4) Genetic approaches, such as knockouts or targeted mutations in repair genes, should be expanded in order to focus on the physiological role of a specific enzyme. Above all: structure, structure, structure! Enzymologists, organic chemists, physical chemiste, X-ray crystallographers, and others must combine forces if the fundamental problems addressed here are to be understood.