The repair patch of E. coli (A)BC excinuclease.

The size of repair patch made by E. coli DNA polymerase I (Poll) following the removal of a thymine-psoralen monoadduct by E. coli (A)BC excinuclease was determined by using an M13mp19 DNA with a single psoralen monoadduct at the polylinker region. Incubation of this substrate with (A)BC excinuclease, Poll and a combination of 3 dnTP plus 1 dNTP(alpha S) for each nucleotide, and DNA ligase resulted in a repair patch with phosphorothioate linkages. The preferential hydrolysis of phosphorothioate bonds by heating in iodoethanol revealed a patch size--with minimal nick translation--equal in length to the 12 nucleotide gap generated by this excision nuclease.     

Structure and function of the (A)BC excinuclease of Escherichia coli

(A)BC excinuclease is the enzymatic activity resulting from the mixture of E. coli UvrA, UvrB and UvrC proteins with damaged DNA. This is a functional definition as new evidence suggests that the three proteins never associate in a ternary complex. The UvrA subunit associates with the UvrB subunit in the form of an A2B1 complex which, guided by UvrA's affinity for damaged DNA binds to a lesion in DNA and delivers the UvrB subunit to the damaged site. The UvrB-damaged DNA complex is extremely stable (t1/2 congruent to 100 min). The UvrC subunit, which has no specific affinity for damaged DNA, recognizes the UvrB-DNA complex with high specificity and the protein complex consisting of UvrB and UvrC proteins makes two incisions, the 8th phosphodiester bond 5' and the 5th phosphodiester bond 3' to the damaged nucleotide. (A)BC excinuclease recognizes DNA damage ranging from AP sites and thymine glycols to pyrimidine dimers, and the adducts of psoralen, cisplatinum, mitomycin C, 4-nitroquinoline oxide and interstrand crosslinks.     

Binding and Incision Activities of UvrABC Excinuclease on Slipped DNA Intermediates that Generate Frameshift Mutations

Previousin vivostudies involving sequence 5′-CCCG₁G₂G₃-3′ (SmaI site) have demonstrated that adducts ofN-2-acetylaminofluorene (AAF) to any of the three guanine residues of theSmaI sequence induce, with different efficiencies, two classes of −1 frameshift events, namely −G and −C mutations, referred to as targeted and semitargeted mutations, respectively. It has been proposed that both events occur during replication as a consequence of slippage events involving slipped mutagenic intermediates (SMIs). In order to evaluate the potential role of the UvrABC excinuclease in frameshift mutagenesis, we have studied the interaction of this enzyme with DNA molecules mimicking SMIsin vitro.In all of our constructions, when present, the AAF adduct was located on the third guanine residue of theSmaI site (5′-CCCG₁G₂G₃-3′). This strand was referred to as the top strand, the complementary strand being the bottom strand. Double-stranded heteroduplexes mimicking the targeted and semitargeted SMIs contained a deletion of a C and a G within theSmaI sequence in the bottom strand and were designated ΔC/3 and ΔG/3 when modified with the AAF on the third guanine residue in the top strand or ΔC/O and ΔG/O when unmodified. The modified homoduplex was designatedSmaI/3.ΔC/O and ΔG/O were weakly recognized by UvrA₂B, but not incised. All three AAF-modified substrates were recognized with similar efficiency and much more efficiently than unmodified heteroduplexes. With AAF-monomodified substrates, dissociation of UvrA₂from the UvrA₂B- DNA complex occurred more readily in heteroduplexes than in the homoduplex.SmaI/3 and ΔC/3 were incised with equal efficiency, while ΔG/3 was less incised. The position of the AAF lesion dictated the position of the incised phosphodiester bonds, suggesting that the presence of a bulge can modulate the yield but not the incision pattern of AAF-modified substrates. The finding that UvrABC excinuclease acts on substrates that mimic SMIs suggests that the nucleotide excision repair pathway may help in fixing frameshift mutations before the following round of replication.     

Sequence effect on incision by (A)BC excinuclease of 4NQO adducts and UV photoproducts.

Nucleotide excision repair in Escherichia coli is initiated by (A)BC excinuclease, an enzyme which incises DNA on both sides of bulky adducts and removes the damaged nucleotide as a 12-13 base long oligomer. The incision pattern of the enzyme was examined using DNA modified by 4-nitroquinoline 1-oxide (4NQO) and UV light. Similar to the cleavage pattern of UV photoproducts and other bulky adducts, the enzyme incises the 8th phosphodiester bond 5' and 5th phosphodiester bond 3' to the 4NQO-modifed base, primarily guanine. The extent of DNA damage by these agents was determined using techniques which quantitatively cleave the DNA or stop at the site of the adduct. By comparison of the intensity of gel bands created by (A)BC excinuclease and the specific cleavage at the damaged site, the efficiency of (A)BC excinuclease incision at 13 different 4NQO-induced adducts and 13 different photoproducts was determined by densitometric scanning. In general, incisions made at 4NQO-induced adducts are proportional to the extent of damage, though the efficiency of cutting throughout the sequence tested varies from 25 to 75%. Incisions made at pyrimidine dimers are less efficient than at 4NQO-adducts, ranging from 13 to 65% incision relative to modification, though most are around 50%. The two (6-4) photoproducts within the region tested are incised more efficiently than any pyrimidine dimer.     

Reactivity of anti-human milk fat globule antibodies with synthetic peptides

The nucleotide sequence of partial cDNA clones coding for the core protein of a human polymorphic epithelial mucin has recently been obtained, this mucin consists of a highly conserved 60 bp tandem repeat and the amino acids commonly found are PDTRPAPGSTAPPAHGVTSA. We synthesized three peptides, 1) P1.24 containing the 20 amino acids and four amino acids (PDTR) of the adjoining repeat; 2) P1.15 consisting of the first fifteen (PDTRPAPGSTAPPAH) and P1.09 the second nine amino acids (GVTSAPDTR) of peptide P1.24. The reactivities of the synthetic peptides with mAb known to react with breast cancer (BC1, BC2, BC3, HMFG-1, 3E1.2, and RCC-1) were studied. The synthetic peptide, P1.24, corresponding to the antigenic sequence predicted from the tandem repeat reacted with antibodies BC1, BC2, and BC3 (known to react with human milk mucin and mucin expressed in breast cancer) and the antibody HMFG-1 which was used to select the cDNA clones. In addition, the epitopes recognized by BC1, BC2, and BC3 appear to be in the same region of the molecule represented by their reactions with the nine amino acids in peptide P1.09 (GVTSAPDTR). By contrast, other antibodies such as 3E1.2 which reacts only weakly with components of human milk, and RCC-1 that detects a low Mr component (95 kDa) in breast cancer, had no specific reaction with the synthetic peptides, indicating that their epitopes are distinct from those of BC1, BC2, BC3, and HMFG-1. Inasmuch as the antibodies HMFG-1, BC1, BC2, and BC3 react with the fully processed milk mucin, it is likely that some of the peptide is exposed, even in the fully glycosylated molecule. Identification of the different epitopes could lead to the development of "second generation" mAb with enhanced specificity for breast carcinoma using the appropriate synthetic peptides as immunogens.     

Identification of the different intermediates in the interaction of (A)BC excinuclease with its substrates by DNase I footprinting on two uniquely modified oligonucleotides

(A)BC excinuclease is the enzymatic activity resulting from the joint actions of UvrA, UvrB and UvrC proteins of Escherichia coli. The enzyme removes from DNA many types of adducts of dissimilar structures with different efficiencies. To understand the mechanism of substrate recognition and the basis of enzyme specificity, we investigated the interactions of the three subunits with two synthetic substrates, one containing a psoralen-thymine monoadduct and the other a thymine dimer. Using DNase I as a probe, we found that UvrA makes a 33 base-pair footprint around the psoralen-thymine adduct and that UvrA-UvrB make a 45 base-pair asymmetric footprint characterized by a hypersensitive site 11 nucleotides 5' to the adduct and protection mostly on the 3' side of the damage. Conditions that favor dissociation of UvrA from the UvrA-UvrB-DNA complex, such as addition of excess undamaged DNA to the reaction mixture, resulted in the formation of a 19 base-pair UvrB footprint. In contrast, a thymine dimer in a similar sequence context failed to elicit a UvrA, a UvrA-UvrB or UvrB footprint and gave rise to a relatively weak DNase I hypersensitive site typical of a UvrA-UvrB complex. Dissociation of UvrA from the UvrA-UvrB-DNA complex stimulated the rate of incision of both substrates upon addition of UvrC, leading us to conclude that UvrA is not a part of the incision complex and that it actually interferes with incision. The extent of incision of the two substrates upon addition of UvrC (70% for the psoralen adduct and 20% for the thymine dimer) was proportional to the extent of formation of the UvrA-UvrB-DNA (i.e. UvrB-DNA) complex, indicating that substrate discrimination occurs at the preincision step.     

Substrate overlap and functional competition between human nucleotide excision repair and Escherichia coli photolyase and (a)BC excision nuclease

Human cell free extract prepared by the method of Manley et al. (1980) carries out repair synthesis on UV-irradiated DNA. Removal of pyrimidine dimers by photoreactivation with DNA photolyase reduces repair synthesis by about 50%. With excess enzyme in the reaction mixture photolyase reduced the repair signal by the same amount even in the absence of photoreactivating light, presumably by binding to pyrimidine dimers and interfering with the binding of human damage recognition protein. Similarly, the UvrB subunit of Escherichia coli (A)BC excinuclease when loaded onto UV-irradiated or psoralen-adducted DNA inhibited repair synthesis by cell-free extract by 75-80%. The opposite was true also as HeLa cell free extract specifically inhibited the photorepair of a thymine dimer by DNA photolyase and its removal by (A)BC excinuclease. Cell-free extracts from xeroderma pigmentosum (XP) complementation groups A and C were equally effective in blocking the E. coli repair proteins, while extracts from complementation groups D and E were ineffective in blocking the E. coli enzyme. These results suggest that XP-D and XP-E cells are defective in the damage recognition subunit(s) of human excision nuclease.     

A new mechanism for repairing oxidative damage to DNA: (A)BC excinuclease removes AP sites and thymine glycols from DNA

Escherichia coli (A)BC excinuclease is the major enzyme responsible for removing bulky adducts, such as pyrimidine dimers and 6-4 photoproducts, from DNA. Mutants deficient in this enzyme are extremely sensitive to UV and UV-mimetic agents, but not to oxidizing agents, or ionizing radiation which damages DNA in part by generating active oxygen species. DNA glycosylases and AP1 endonucleases play major roles in repairing oxidative DNA damage, and thus it has been assumed that nucleotide excision repair has no role in cellular defense against damage by ionizing radiation and oxidative damage. In this study we show that the E. coli nucleotide excision repair enzyme (A)BC excinuclease removes from DNA the two major products of oxidative damage, thymine glycol and the baseless sugar (AP site). We conclude that nucleotide excision repair is an important cellular defense mechanism against oxidizing agents.     

Molecular Lego: design of molecular assemblies of P450 enzymes for nanobiotechnology.

This paper reports on the application of the molecular Lego approach to P450 enzymes. Protein domains are used as catalytic (P450 BM3 haem domain and human P450 2E1) or electron transfer (flavodoxin and P450 BM3 reductase) modules. The objectives are to build assemblies with improved electrochemical properties, to construct soluble human P450 enzymes, and to generate libraries of new P450 catalytic modules based on P450 BM3. A rationally designed, gene-fused assembly (BMP-FLD) was obtained from the soluble haem domain of cytochrome P450 BM3 from Bacillus megaterium (BMP) and flavodoxin from Desulfovibrio vulgaris (FLD). The assembly was expressed successfully and characterised in its active form, displaying improved electrochemical properties. Solubilisation of the human, membrane-bound P450 2E1 (2E1) was achieved by fusing key elements of the 2E1 enzyme with selected parts of P450 BM3. An assembly containing the first 54 residues of P450 BM3, the whole sequence of P450 2E1 from residue 81 and the reductase domain of P450 BM3 was constructed. The 2E1-BM3 assembly was successfully expressed in the cytosol of Escherichia coli. The soluble form of 2E1-BM3 was reduced in carbon monoxide atmosphere and displayed the typical absorption peak at 450 nm, characteristic of a folded and active P450 enzyme. Finally, the alkali method previously developed in this laboratory was used to screen for P450 activity within a library of random mutants of P450 BM3. A number of variants active towards non-physiological substrates, such as pesticides and polyaromatic hydrocarbons were identified, providing new P450 catalytic modules. The combination of these three areas of research provide interesting tools for exploitation in nanobiotechnology.     

A digenic cause of cleft lip in A-strain mice and definition of candidate genes for the two loci

BACKGROUND: Nonsyndromic cleft lip with or without cleft palate, CL(P), is a common human birth defect with a complex unknown genetic cause. The mouse model is the "A/-" strains. Our previous studies mapped two loci: clf1 on Chr11 and clf2 on Chr13--with a strong genetic maternal effect on the level of risk. Here we test the hypothesis that CL(P) is digenic and identify candidate genes for clf1 and clf2. METHODS: We observed E14 CL(P) frequencies in backcross (BC1) embryos from a new cross of A/WySn to AXB-4/Pgn and from test crosses of three new "congenic RI" lines. Using new polymorphic markers from genes and our mapping panels of segregants and RI strains, we identified the candidate genes for clf1 and clf2. We sequenced the coding region of Ptch in A/WySn cDNA. RESULTS: Seventy new BC1 CL(P) segregants (4%) were obtained, as predicted. All three new congenic RI lines homozygous for both clf1 and clf2 had A/WySn-level CL(P) frequencies (10-30%) in test crosses. The clf1 region contains 10 known genes (Arf2, Cdc27, Crhr1, Gosr2, Itgb3, Mapt, Myl4, Nsf, Wnt3, and Wnt9b). The clf2 region contains 17 known genes with human orthologs. Both regions contain additional potential genes. No causal mutation in Ptch coding sequence was found. CONCLUSIONS: In A-strain mice, nonsyndromic CL(P) is digenic, suggesting that nonsyndromic human CL(P) may also be digenic. The orthologous human genes are on 17q (clf1) and 9q, 8q and 5p (clf2), and good candidate genes are WNT3 or WNT9B (17q), and PTCH (9q) or MTRR (5p).     

Echinococcus P29 Antigen: Molecular Characterization and Implication on Post-Surgery Follow-Up of CE Patients Infected with Different Species of the Echinococcus granulosus Complex

The protein P29 is a potential serological marker for post-treatment monitoring of cystic echinococcosis (CE) especially in young patients. We now have demonstrated that P29 is encoded in the Echinococcus genus by a single gene consisting of 7 exons spanning 1.2 kb of DNA. Variability of the p29 gene at inter- and intra-species level was assessed with 50 cDNA and 280 genomic DNA clones isolated from different E. granulosus s.l. isolates (E. granulosus sensu stricto (G1), E. equinus (G4), E. ortleppi (G5), E. canadensis (G6), E. canadensis (G7) and E. canadensis (G10)) as well as four E. multilocularis isolates. Scarce interspecies polymorphism at the p29 locus was observed and affected predominantly E. granulosus s.s. (G1), where we identified two alleles (A1 and A2) coding for identical P29 proteins and yielding in three genotypes (A1/A1, A2/A2 and A1/A2). Genotypic frequencies expected under Hardy-Weinberg equilibrium revealed a high rate of heterozygosity (47%) that strongly supports the hypothesis that E. granulosus s.s. (G1) is predominantly outbreeding. Comparative sequence analyses of the complete p29 gene showed that phylogenetic relationships within the genus Echinococcus were in agreement with those of previous nuclear gene studies. At the protein level, the deduced P29 amino acid (AA) sequences exhibited a high level of conservation, ranging from 97.9% AA sequence identity among the whole E. granulosus s.l. group to 99.58% identity among E. multilocularis isolates. We showed that P29 proteins of these two species differ by three AA substitutions without implication for antigenicity. In Western-blot analyses, serum antibodies from a human CE patient infected with E. canadensis (G6) strongly reacted with recombinant P29 from E. granulosus s.s. (G1) (recEg(G1)P29). In the same line, human anti-Eg(G1)P29 antibodies bound to recEcnd(G6)P29. Thus, minor AA sequence variations appear not to impair the prognostic serological use of P29.     

Isolation of human monoclonal antibodies that neutralize human rotavirus

A human antibody library constructed by utilizing a phage display system was used for the isolation of human antibodies with neutralizing activity specific for human rotavirus. In the library, the Fab form of an antibody fused to truncated cp3 is expressed on the phage surface. Purified virions of strain KU (G1 serotype and P[8] genotype) were used as antigen. Twelve different clones were isolated. Based on their amino acid sequences, they were classified into three groups. Three representative clones-1-2H, 2-3E, and 2-11G-were characterized. Enzyme-linked immunosorbent assay with virus-like particles (VLP-VP2/6 and VLP-VP2/6/7) and recombinant VP4 protein produced from baculovirus recombinants indicated that 1-2H and 2-3E bind to VP4 and that 2-11G binds to VP7. The neutralization epitope recognized by each of the three human antibodies might be human specific, since all of the antigenic mutants resistant to mouse monoclonal neutralizing antibodies previously prepared were neutralized by the human antibodies obtained here. After conversion from the Fab form of an antibody into immunoglobulin G1, the neutralizing activities of these three clones toward various human rotavirus strains were examined. The 1-2H antibody exhibited neutralizing activity toward human rotaviruses with either the P[4] or P[8] genotype. Similarly, the 2-3E antibody showed cross-reactivity against HRVs with the P[6], as well as the P[8] genotype. In contrast, the 2-11G antibody neutralized only human rotaviruses with the G1 serotype. The concentration of antibodies required for 50% neutralization ranged from 0.8 to 20 micro g/ml.     

Directed evolution of a glycosynthase from Agrobacterium sp. increases its catalytic activity dramatically and expands its substrate repertoire

The Agrobacterium sp. beta-glucosidase (Abg) is a retaining beta-glycosidase and its nucleophile mutants, termed Abg glycosynthases, catalyze the formation of glycosidic bonds using alpha-glycosyl fluorides as donor sugars and various aryl glycosides as acceptor sugars. Two rounds of random mutagenesis were performed on the best glycosynthase to date (AbgE358G), and transformants were screened using an on-plate endocellulase coupled assay. Two highly active mutants were obtained, 1D12 (A19T, E358G) and 2F6 (A19T, E358G, Q248R, M407V) in the first and second rounds, respectively. Relative catalytic efficiencies (kcat/Km) of 1:7:27 were determined for AbgE358G, 1D12, and 2F6, respectively, using alpha-D-galactopyranosyl fluoride and 4-nitrophenyl beta-D-glucopyranoside as substrates. The 2F6 mutant is not only more efficient but also has an expanded repertoire of acceptable substrates. Analysis of a homology model structure of 2F6 indicated that the A19T and M407V mutations do not interact directly with substrates but exert their effects by changing the conformation of the active site. Much of the improvement associated with the A19T mutation seems to be caused by favorable interactions with the equatorial C2-hydroxyl group of the substrate. The alteration of torsional angles of Glu-411, Trp-412, and Trp-404, which are components of the aglycone (+1) subsite, is an expected consequence of the A19T and M407V mutations based on the homology model structure of 2F6.     

Catalytic properties of the human cytochrome P450 2E1 produced by cDNA expression in mammalian cells.

A full-length cDNA encoding human cytochrome P450 2E1 was expressed in mammalian cell lines using the vaccinia virus expression system. Immunoblot analysis showed that the expressed protein reacted with a polyclonal antibody against rat 2E1 and comigrated with P450 2E1 from human liver microsomes. P450 2E1 expressed in Hep G2 cells, a human cell line which contains both cytochrome b5 and NADPH:P450 oxidoreductase, was able to metabolize several known P450 2E1 substrates: N-nitrosodimethylamine (NDMA), N-nitrosomethylbenzylamine (NMBzA), p-nitrophenol, phenol, and acetaminophen. Apparent Km and Vmax values for NDMA demethylation were 22 microM and 173 pmol/min/mg microsomal protein, respectively. P450 2E1 expressed in TK-143 cells, which do not contain b5, displayed Km and Vmax values of 31 microM and 34 pmol/min/mg microsomal protein, respectively. Incorporation of purified rat liver b5 into TK-143 microsomes increased the Vmax 2.2-fold and decreased the Km to 22 microM. Addition of b5 to Hep G2 microsomes resulted in a 1.6-fold increase in Vmax, but showed no effect on the Km. P450 2E1 expressed in Hep G2 cells was shown to metabolize NMBzA with a Km of 47 microM and Vmax of 213 pmol/min/mg microsomal protein. Addition of b5 lowered the Km to 27 microM, but had no effect on Vmax. These results demonstrate conclusively that P450 2E1 is responsible for the low Km forms of NDMA demethylase and NMBzA debenzylase observed in liver microsomes and that these activities are affected by cytochrome b5.     

Nucleotide excision repair: from E. coli to man

Nucleotide excision repair is both a 'wide spectrum' DNA repair pathway and the sole system for repairing bulky damages such as UV lesions or benzo[a]pyrene adducts. The mechanisms of nucleotide excision repair are known in considerable detail in Escherichia coli. Similarly, in the past 5 years important advances have been made towards understanding the biochemical mechanisms of excision repair in humans. The overall strategy of the repair is the same in the two species: damage recognition through a multistep mechanism involving a molecular matchmaker and an ATP-dependent unwinding of the damaged duplex; dual incisions at both sides of the lesion by two different nucleases, the 3' incision being followed by the 5'; removal of the damaged oligomer; resynthesis of the repair patch, whose length matches the gap size. Despite these similarities, the two systems are biochemically different and do not even share structural homology. E. coli excinuclease employs three proteins in contrast to 16/17 polypeptides in man; the excised fragment is longer in man: the procaryotic excinuclease is not able by itself to remove the excised oligomer whereas the human enzyme does. Thus, the excinuclease mode of action is well conserved throughout evolution, but not the biochemical tools: this represents a case of evolutionary convergence.     

Biosynthesis of the cancer-associated sialyl-Lex determinant in human amniotic fluid

Biosynthesis of the sialyl-Lex determinant (NeuAc alpha 2-3Gal beta 1-4(Fuc alpha 1-3)-GlcNAc beta 1-3-R) in human amniotic fluid has been shown to proceed via the same sequence of glycosylation steps established previously for lung carcinoma PC 9 cells (Holmes, E. H., Ostrander, G.K. & Hakomori, S. (1986) J. Biol. Chem. 261, 3737-3743): sialylation of type-2-chain-precursor substrates (paragloboside) by an amniotic alpha 2-3-sialyltransferase precedes fucosylation of sialylated intermediates (sialosyl paragloboside) by an organ-characteristic alpha 1-3-L-fucosyltransferase.     

RecA-dependent incision of psoralen-crosslinked DNA by (A)BC excinuclease.

Previous work to elucidate the mechanism of crosslink repair by (A)BC excinuclease has shown that a psoralen-crosslinked duplex is selectively incised in the furan-side strand, while a three-stranded structure is incised in the pyrone-side strand of the crosslink. These observations support a sequential incision and recombination model for the complete error-free repair of a psoralen crosslink. The work presented here extends these findings by demonstrating that in the presence of RecA protein and a homologous DNA oligonucleotide, (A)BC excinuclease is induced to incise the pyrone-side strand of a crosslinked double-stranded plasmid molecule. This finding adds further support to the current model for error-free crosslink repair.     

Cleavage efficiencies of model substrates for ribonuclease P from Escherichia coli and Thermus thermophilus.

We compared cleavage efficiencies of mono-molecular and bipartite model RNAs as substrates for RNase P RNAs (M1 RNAs) and holoenzymes from E. coli and Thermus thermophilus, an extreme thermophilic eubacterium. Acceptor stem and T arm of pre-tRNA substrates are essential recognition elements for both enzymes. Impairing coaxial stacking of acceptor and T stems and omitting the T loop led to reduced cleavage efficiencies. Small model substrates were less efficiently cleaved by M1 RNA and RNase P from T. thermophilus than by the corresponding E. coli activities. Competition kinetics and gel retardation studies showed that truncated tRNA substrates are less tightly bound by RNase P and M1 RNA from both bacteria. Our data further indicate that (pre-)tRNA interacts stronger with E. coli than T. thermophilus M1 RNA. Thus, low cleavage efficiencies of truncated model substrates by T. thermophilus RNase P or M1 RNA could be explained by a critical loss of important contact points between enzyme and substrate. In addition, acceptor stem--T arm substrates, composed of two synthetic RNA fragments, have been designed to mimic internal cleavage of any target RNA molecule available for base pairing.     

Interactions of the human, rat, Saccharomyces cerevisiae and Escherichia coli 3-methyladenine-DNA glycosylases with DNA containing dIMP residues

In DNA, the deamination of dAMP generates 2′-deoxyinosine 5′-monophosphate (dIMP). Hypoxanthine (HX) residues are mutagenic since they give rise to A·T→G·C transition. They are excised, although with different efficiencies, by an activity of the 3-methyladenine (3-meAde)-DNA glycosylases from Escherichia coli (AlkA protein), human cells (ANPG protein), rat cells (APDG protein) and yeast (MAG protein). Comparison of the kinetic constants for the excision of HX residues by the four enzymes shows that the E.coli and yeast enzymes are quite inefficient, whereas for the ANPG and the APDG proteins they repair the HX residues with an efficiency comparable to that of alkylated bases, which are believed to be the primary substrates of these DNA glycosylases. Since the use of various substrates to monitor the activity of HX-DNA glycosylases has generated conflicting results, the efficacy of the four 3-meAde-DNA glycosylases of different origin was compared using three different substrates. Moreover, using oligonucleotides containing a single dIMP residue, we investigated a putative sequence specificity of the enzymes involving the bases next to the HX residue. We found up to 2–5-fold difference in the rates of HX excision between the various sequences of the oligonucleotides studied. When the dIMP residue was placed opposite to each of the four bases, a preferential recognition of dI:T over dI:dG, dI:dC and dI:dA mismatches was observed for both human (ANPG) and E.coli (AlkA) proteins. At variance, the yeast MAG protein removed more efficiently HX from a dI:dG over dI:dC, dI:T and dI:dA mismatches.