WO2004019959A2

WO2004019959A2 - Dna repair enzymes

Info

Publication number: WO2004019959A2
Application number: PCT/GB2003/003776
Authority: WO
Inventors: Thomas Lindahl; Barbara Sedgwick; Sarah Trewick; Tod Duncan; Paul Bates
Original assignee: Imperial Cancer Research Technology
Priority date: 2002-08-30
Filing date: 2003-08-29
Publication date: 2004-03-11
Also published as: AU2003263296A1; AU2003263296A8; WO2004019959A3; WO2004019959A9

Abstract

The invention relates to methods of selection of agents and compositions for the treatment of proliferative diseases such as cancer. More specifically, the present invention relates to the methods of selecting compounds which modulate the activity of DNA repair enzymes HAB1 and HAB2 and their uses in the treatment of disorders, which would benefit from an increase or a decrease in HAB1 or HAB2 activity.

Description

DNA REPAIR ENZYMES

Background

The present invention relates to methods of treatment and to methods of identifying compounds which are potentially useful in therapy. In particular, the invention relates to methods of treating proliferative disease, such as cancer, and methods of identifying compounds which are potentially useful in treating proliferative disease such as cancer.

DNA alkylating agents occur endogenously and are present in the environment (Rebeck et al., 1991 ; Vaughan et al., 1991). More importantly, DNA alkylating agents are used as chemotherapeutic agents in the treatment of prolific diseases such as cancer. It is a well known problem that tumour cells very often possess or develop resistance against these DNA alkylating agents thereby severely debilitating this method of treating cancer patients. Conversely, DNA alkylating agents also damage non-tumours cells such as bone marrow cells leading to myelosuppression thereby limiting the application of these powerful drugs.

E. coli exposed to DNA alkylating agents respond by inducing the expression of four genes, ada, alkA, aidB and alkB. Ada protein is an O⁶-methylguanine- DNA methyltransferase and it also regulates this adaptive response. AlkA is a 3-methyladenine (3-meA)-DNA glycosylase and AidB is proposed to destroy certain alkylating agents (Lindahl et al., 1988; Seeberg et al., 1999).

AlkB processes cytotoxic DNA damage generated in single-stranded DNA by SN₂ methylating agents, such as methylmethanesulfonate (MMS), dimethylsulphate (DMS) and methyl iodide (Mel) (Dinglay et al., 2000). 1- Methyladenine (1-meA) and 3-methylcytosine (3-meC) are predominant forms of base damage only in single-stranded DNA because the modification sites are normally protected by base pairing (Bodell et al., 1979; Singer et al., 1983). These lesions are cytotoxic because they stall DNA replication and could prevent transcription, and are not removed by known DNA repair pathways. It was previously proposed that 1-meA and 3-meC in DNA are candidate substrates of the AlkB protein (Dinglay et al., 2000). Nevertheless, many attempts to develop assays for this enzyme were unsuccessful. Theoretical sequence profile and fold recognition searches suggested that AlkB may be either a hydrolase or a α-ketoglutarate (αKG) - Fe (II) - dependent dioxygenase (Aravind & Koonin, 2001 ; Aravind et al., 1999).

Expression of E. coli alkB confers alkylation resistance to human cells (Chen et al., 1994), and conversely, a human homologue (known as ABH or ABH1) (Accession number S64736) has been reported to convey methyl methanesulphonate (MMS) resistance to the E. coli alkB mutant (Wei et al., 1996). The data indicated that ABH only partially rescued the E. coli alkB mutant phenotype, whereas E. coli alkB gene can fully complement E. coli alkB mutant cells. This information is also disclosed in W096/12791. The DNA and polypeptide sequences disclosed in W096/12791 were subsequently corrected and amended (accession number Q13686) and published under accession number XM_007409 (herein, referred to as ABH2).

As it is discussed in more detail below we show here for the first time that AlkB protein repairs DNA alkylation damage by an unprecedented mechanism. AlkB has no detectable nuclease, DNA glycosylase or methyltransferase activity, yet Escherichia coli alkB mutants are defective in processing methylation (a form of alkylation) damage generated in single- stranded DNA. We have found that purified AlkB repairs the cytotoxic lesions 1 -methyladenine and 3- methylcytosine in single- and double-stranded DNA in a reaction dependent on oxygen, α-ketoglutarate and Fe(ll).The AlkB enzyme couples oxidative decarboxylation of α-ketoglutarate to the hydroxylation of these methylated bases in DNA, resulting in direct reversion to the unmodified base and the release of formaldehyde. Furthermore, by employing the aforementioned assay we surprisingly identified two genes to be human homologues of AlkB, in the following referred to as HAB1 (depicted in Figure 2) and HAB2 (depicted in Figure 3). We found that HAB1 and HAB2 are DNA repair enzymes. Moreover, we found that HAB1 and HAB2 complemented the mutant E. coli alkB mutants defective in processing methylation damage in single-stranded DNA, in the following referred to as E. coli complementation assay. In addition to the demethylation activity of AlkB, HAB1 and HAB2, we also discovered these enzymes possess a de-ethylation activity. Furthermore, we unexpectedly found that HAB1 and HAB2 are differentially expressed in different human tissues at the mRNA level. In particular HAB1 is expressed at high levels in spleen, prostate, bladder, lung and colon tissue, whereas HAB2 is expressed at high level in liver and bladder. In summary, surprisingly and unexpectedly we found that HAB1 and HAB2 possess a new biological activity, which plays a major role in the reaction of tumour and non-tumour cells to alkylating agents. In particular, HAB1 and HAB2 are thought to convey resistance against these alkylating agents by effecting a DNA repair mechanism of the cell.

It is an object of the present invention to make use of these unexpected findings in the field of therapy for proliferative diseases.

Accordingly the invention provides an assay for identifying AlkB and AlkB homologues by measuring the enzyme activity. The term "enzyme activity" includes the oxidative dealkylation, including demethylation and de-ethylation, of single and double stranded DNA. As extensively detailed below oxidative demethylation by AlkB and AlkB homologues comprises several distinct reactions including the oxidative decarboxylation of α-ketogluterate and the hydroxylation of the methylated DNA bases. Assays for measuring these enzymatic reactions are well known in the art.

For example the enzyme activity could be tested by a direct assay on the natural substrate of the enzyme, utilising the physical differences of the substrate and product (e.g. fluorescence/absorption spectra), or a chemically modified derivative of that substrate designed to give better absorption/fluorescent properties. As a further example the enzyme activity could be tested using an indirect assay of one or more of the reaction components, e.g. directly measuring α- ketoglutarate, aldehyde or succinate.

As a further example the enzyme activity could be tested using a radiolabelled assay. For example the methyl group can be tritiated allowing the direct measure of its removal through the enzyme activity.

Another object of the invention is a method of combating a proliferative disorder in a patient the method comprising administering to the patient an inhibitor of HAB1 (and/or HAB2) activity. In a preferred embodiement, the patient is in need of sensitising the tumour cells to alkylating agents.

The term "proliferative disorder" includes any proliferative disorder including both neoplastic and hyperplastic disorders. It is preferred that the disorder to be treated is a neoplastic disorder, in particular, cancer. The patient may have any cancer such as a brain cancer including glioma, glioblastoma multiforme, medulloblastoma, astrocytoma, ependymonas, oligodendrogliomas, lung cancer, liver cancer, cancer of the spleen, kidney, adrenal gland (including pheochromocytoma), thyroid gland, oesophagus, pituitary gland, lymph node, small intestine, pancreas, blood (lymphomas and leukaemias), colon, stomach, breast, endometrium, prostate including prostate adenocarcinoma, testicle, ovary, skin including melanoma, head and neck, oesophagus and bone marrow.

The terms "HAB1 activity" and "HAB2 activity" include the oxidative dealkylation, in particular demethylation and de-ethylation, of single and double stranded DNA. HAB1 activity and HAB2 activity are determined by measuring the enzyme activity as described above for the AlkB enzyme activity or by measuring the ability to complement Alk B function in the E. coli complementation assay. In particular, the present invention relates to the use of an HAB1 or HAB2 polypeptide which has the amino acid sequence as depicted in Figure 2 (HAB1) and Figure 3 (HAB2) as well as fragments, analogues and derivatives of such polypeptides. Furthermore, the present invention relates to the use of nucleic acids encoding the polypeptides depicted in Figure 2 and Figure 3, as well as fragments, analogues and derivatives of such polypeptides.

The terms "fragment", "derivative" and "analogue" mean polypeptides (and nucleic acid encoding polypeptides) which retain essentially the same biological function or activity as the HAB1 and HAB2 polypeptide. Methods of preparing fragments, derivative and analogues falling within this definition are well known in the art. For example, a polypeptide of the present invention may be a recombinant polypeptide, or a synthetic polypeptide. The polypeptide might contain deletions or truncations. A polypeptide of the present invention might be one in which one or more of the amino acid residues are substituted with a conserved or non-conserved amino acid residue. Furthermore, a polypeptide of the present invention might be fused to another polypeptide.

The term "alkylating agents" as used herein describes agents which may be selected from, but are not limited to, alkylating agents, such as Temozolomide, Procarbazine, nitrogen mustards such as mechlorethamine (HN2), cyclophosphamide, ifosfamide, melphalan (L-sarcolysin) and chlorambucil; ethylenimines and methylmelamines such as hexamethylmelamine, thiotepa; alkyl sulfonates and thiosulfonates such as busulfan, methyl methanesulfonate (MMS) and methyl methanethiosulfonate; nitrosoureas and nitrosoguanidines such as carmustine (BCNU), lomustine (CCNU), semustine (methyl-CCNU), streptozocin (streptozotocin) and N- methyl-N'-nitro-N-nitrosoguanidine (MNNG); and triazenes such as dacarbazine (DTIC; dimethyltriazenoimidazole-carboxamide).

Furthermore, another object of the present invention is the use of HAB1 (or HAB2) activity for treating or preventing DNA damage by alkylating agents. A preferred object of the present invention is a method to protect a cell against alkylating agents comprises the expression of HAB1 or HAB2 polypeptides or fragments, analogues or derivatives thereof in the cell so that the overall activity of HAB1 or HAB2 activity in said cell is elevated, thereby increasing the rate of DNA repair induced by alkylating agents.

A variety of methods are known in the art to express a polypeptide in a cell. Preferably the cell is transformed with a suitable DNA construct encoding the aforementioned polypeptides. Preferred examples of these DNA constructs are viral vectors such as adenovirus-based systems or retroviral based systems.

It will be appreciated from the foregoing that HAB1 and/or HAB2 are potential targets for drugs that may be useful in treating proliferative diseases, especially cancer. In particular, drugs that inhibit HAB1 and/or HAB2 are believed to be useful in sensitising a patient to alkylating agents.

Furthermore it will be appreciated that HAB1 is a particular useful target in proliferative diseases of the spleen, prostate, bladder, lung and colon tissue, whereas HAB2 is a particular useful target in proliferative diseases of liver and bladder tissues.

Thus, a further object of the invention provides a method for selecting a test compound as a modulator of HAB1 and/or HAB2 activity, the method comprising assaying HAB1 and/or HAB2 in the presence of the said test compound and selecting a compound which modulates HAB1 and/or HAB2 activity.

A still further object of the invention provides a method for selecting a test compound as an agent for combating proliferative disease, or a lead compound for making such an agent, the method comprising assaying HAB1 and/or HAB2 in the presence of the said test compound and selecting a compound which modulates HAB1 and/or HAB2 activity. The term "lead compound" means a compound, which whilst not itself suitable for use as a drug may provide a starting point for the design of other compounds that may have more desirable characteristics.

A yet still further object of the invention provides a method for selecting a test compound as an agent for modulating a patient's response to alkylating agents, or a lead compound for making such an agent, the method comprising assaying HAB1 and/or HAB2 in the presence of the said test compound and selecting a compound which modulates HAB1 and/or HAB2 activity.

Typically, the activity of HAB1 and/or HAB2 in the presence of the test compound, and the activity of HAB1 and/or HAB2 under substantially the same conditions but in the absence of the test compound are compared in order to decide what effect the test compound has on HAB1 and/or HAB2 activity. The source of HAB1 or HAB2 in the screening assay is typically recombinant HAB1 or HAB2 or HAB1 or HAB2 which has been purified from its natural source, for example by immunoprecipitation. However, it will be appreciated that the assays and screening methods of the invention may be carried out using any suitable source of HAB1 or HAB2, such as an extract of a cell. The HAB1 or HAB2-specific nature of the results can be readily determined using a selective HAB1 or HAB2 inhibitory antibody.

Further objects of the present invention are compounds, which modulate, preferably inhibit, HAB1 and/or HAB2 activity. A further object of the invention is the use of these compounds to modulate, preferably inhibit, HAB1 and/or HAB2 activity.

The aforementioned compounds may be any inhibitor of HAB1 (and/or HAB2) activity. Thus an inhibitor of HAB1 (and/or HAB2) activity may be an inhibitor of the protein or it may be an inhibitor of the production of the HAB1 (and/or HAB2) protein from its gene. It is preferred if the inhibitor is selective for HAB1 (and/or HAB2). By selective we mean that the inhibitor has a greater affinity for HAB1 (and/or HAB2) (or its gene or mRNA) than for other proteins or genes or mRNA within the cell. It will be appreciated that it is not necessary for the inhibitor to be completely specific for HAB1 and/or HAB2; rather, it is acceptable for the inhibitor to display a degree of selectivity for HAB1 and/or HAB2 even though it may also be active against other components of the cell.

Examples of inhibitors of HAB1 (and/or HAB2) include duplex RNA which mediates interference of HAB1 (or HAB2) RNA, antisense nucleic acid, ribozymes selective for HAB1 (or HAB2) mRNA, antibodies which bind to HAB1 (or HAB2), and small molecules which inhibit its or their function.

A further object of the present invention is the combined, separate or sequential administration of one or more alkylating agents and a compound which inhibits HAB1 and/or HAB2 activity to a patient suffering a proliferative disease, especially cancer.

Description of Figures

Figure 1

Schematic representation of the repair of 1-meA and 3-meC residues in DNA by oxidative demethylation catalysed by AlkB. Oxidation of 1-meA and 3-meC by AlkB requires O₂, αKG and Fe(ll) and generates C0₂ and succinate. Oxidised methyl groups are released as formaldehyde resulting in direct reversal of the lesions to the unmodified base residues. Note that the alkylation positions in 1-meA and 3-meC are equivalent, although these sites in the pyrimidine rings of A and C are numbered differently by standard nomenclature.

Figure 2

Human HAB1 (DEPC-1 Accession Number AB042029) cDNA and amino acid sequences.

Figure 3

Human HAB2 (DEPC-1 Like Accession Number XM_058581 ) cDNA and amino acid sequences. Figure 4

Release of ethanol soluble material from methylated poly(dA) by AlkB in the presence of Fe(ll) and αKG. a, [¹⁴C]Mel treated poly(dA) (1200 cpm) was incubated with AlkB in the complete reaction mixture, and the release of radioactive material was monitored, b, Requirements for AlkB activity (2 pmoles AlkB). c, Comparison of AlkB activity on [¹⁴C]-methylated poly(dA) (1200 cpm) (O) and after annealing to poly(dT) (λ ), assayed at 20°C for 30 min to maintain stable duplexes, d, Difference absorption spectrum of anaerobic AlkB (0.35 mM) in the presence of αKG (1 mM) and Fe(ll) (0.3 mM) minus the spectrum without Fe(ll).

Figure 5

Repair of 1-meA and 3-meC by AlkB. [¹⁴C]Mel treated a. poly(dA), or b. poly(dC) were incubated without AlkB (Δ) or with 2.5 pmoles AlkB (θ). The [¹⁴C]-labeIed methylated bases remaining in the substrates were analysed by HPLC and scintillation counting. Frame c. Direct reversion of 1-meA to adenine by AlkB. A thymine-rich oligodeoxynucleotide, containing five adenine residues was heavily methylated and then incubated with or without 920 pmoles of AlkB at 37°C for 30 min. The bases present in the oligonucleotide were analysed by HPLC and A₂₆o measurements. The early eluting peaks were oligo(dT) fragments.

Figure 6 Production of formaldehyde and consumption of O₂ by AlkB. Methylated oligo(dA)-oligo(dT) was incubated with AlkB. The products were derivatised with PFPH and analysed by GC/MS. GC traces are shown of derivatised products generated when the DNA was a. methylated or b. not methylated. Frame c. shows the mass spectrum of the product indicated by the arrow in panel a. Frame d. Consumption of O₂ during incubation of AlkB with either methylated or non-methylated oligo(dA)-oligo(dT) was determined by using an oxygen electrode. AlkB (9 μM) was added at the point indicated by the arrow. Figure 7

Amino acid sequence alignment of HAB1 and AlkB. Percent identity shown.

Figure 8 Amino acid sequence alignment of HAB2 and AlkB. Percent identity shown.

Figure 9

Amino acid sequence alignment of HAB1 and HAB2. Percent identity shown.

Figure 10

Amino acid sequence alignment of AlkB and ABH2. Percent identity shown.

Figure 11

Alternative amino acid sequence alignment of HAB1 , HAB2, AlkB, ABH2 and a Macacca AlkB homologue.

Figure 12

Complementation of an alkB mutant phenotype by overexpression of human

HAB1 or HAB2. Reactivation of MMS treated φ single stranded DNA phage was monitored in various E. coli strains. Left Panel, host strains were alkB⁺ or alkB22 derivatives of E. coli BL21.DΕ3. Plasmid pBAR54 encodes E.coli AlkB protein and pBAR65 encodes human ABH2. λ, α/fcδ γpET15b; r, alkB^' /pET15b; σ, alkB^'/ AlkB; D, alkB^'IABB.2. Right Panel, host strains were alkB⁺ or alkB22 derivatives of E. coli AB1157 gyrA. Plasmid pBAR32 encodes E.coli AlkB protein, pGST.HABI encodes GST.HAB1 and pGST.HAB2 encodes GST.HAB2. λ, α/AR pGEX6P-1 ; r, α/ΛS7pGEX6P-1 ; σ, α/&S7alkB; V, alkW /HAB1; D, α/AffVHAB2.

Figure 13 HAB2 removes 1 -methyladenine but not 3-methyladenine from DNA:

[14C]Mel treated poly(dA) was incubated without HAB2 (■) or with HAB2 (♦). The [14C]-labeled methylated bases remaining in the substrates were analysed by HPLC and scintillation counting. Figure 14

Repair of 3-meC by HAB1. [¹⁴C]Mel treated poly(dC) were incubated without HAB1 (♦) or with HAB1 (■). The [¹⁴C]-labeled methylated bases remaining in the substrates were analysed by HPLC and scintillation counting.

Figure 15

Northern blot analysis of HAB1 and HAB2 mRNA expression in different human tissues and cell lines.

Detailed Description of the Invention

One aspect of the present invention is an assay for identifying agents, which modulate, preferably inhibit, the activity of HAB1 and/or HAB2. For example, the assay tests the novel DNA repair activity for AlkB using DNA substrates containing 1-meA or 3-meC, αKG as a cosubstrate, and Fe(ll) as a cofactor.

A substrate containing [14C]methylated adenine residues was prepared by treating poly(dA) with [14C]Mel. In the presence of αKG and Fe(ll), purified His-tagged AlkB protein released ethanol soluble radioactive material from this methylated substrate (Figure 4a). Similar activity was also observed with FLAG-tagged AlkB protein (data not shown). The activity was dependent on αKG, inhibited by EDTA and stimulated by ascorbate (Figure 4b). Inhibition by EDTA was overcome by adding an excess of Fe(NH4)2(SO4)2.6H2O, thus demonstrating a requirement for Fe(ll). At 100-fold higher AlkB concentrations, the effect of ascorbate was reduced. Thus, ascorbate was not essential but stimulated AlkB activity, probably by regenerating Fe(ll) from Fe(lll) (data not shown). The conditions for this assay were optimised to 50 mM Hepes-KOH, pH 8, 75 μM Fe(NH4)2(SO4)2.6H2O, 1 mM αKG, 2 mM ascorbate, and 50 μg/ml bovine serum albumin (BSA). To determine whether AlkB could also act on double stranded DNA, the [14C]-methylated poly(dA) was annealed to poly(dT). AlkB was approximately 3-fold more active on the double- compared with the single-stranded substrate (Figure 4c). To examine which methylated adenines were processed by AlkB, the poly(dA) substrate was incubated with AlkB, acid hydrolysed and analysed by HPLC. In the [14C]Mel treated poly(dA), 1-meA was the major methylated base, 3- meA was also abundant, whereas 7-methyladenine (7-meA) was a minor product (Singer et al., 1983). Following incubation with AlkB, the levels of 1- meA were reduced whereas the amounts of 3-meA and 7-meA remained unchanged (Figure 5a). Thus, AlkB specifically catalysed removal of 1-meA from methylated poly(dA). From the data in Figure 4a, 0.1 pmole AlkB protein removed 1.7 pmoles 1-meA from methylated poIy(dA) in 15 min, indicating that AlkB acts enzymatically and is not consumed during the reaction. This distinguishes its mode of action from O6-methylguanine-DNA methyltransferase.

A second modification that is formed to a greater extent in single- compared with double-stranded DNA is 3-meC. This lesion was also considered as a candidate substrate of AlkB. AlkB protein released radioactive material from [14C]Mel treated poly(dC) (data not shown). HPLC analysis showed that 3- meC was the only detectable modified base in this substrate (Singer et al., 1983), and disappeared on incubation with AlkB protein (Figure 5b). We have therefore identified two substrates of AlkB, 1-meA and 3-meC, that are both generated in single stranded DNA on treatment with SN₂ methylating agents.

The essential requirement for αKG and Fe(ll), inhibition by EDTA and stimulation by ascorbate strongly support the proposal that AlkB is an αKG- Fe(ll) dependent dioxygenase. Direct evidence that AlkB binds Fe(ll) and αKG was obtained by examining the absorption spectrum of the anaerobic protein in the presence or absence of Fe(ll) or αKG. The protein with bound metal and cofactor had an absorption peak at 500 nm (Figure 4d). This chromophore is attributed to a weak charge transfer from Fe(ll) to αKG, and is a spectroscopic signature of αKG-Fe(ll)-dependent dioxygenases (Hegg et al., 1999; Pavel et al., 1998; Ryle et al., 1999). The AlkB absorption is slightly shifted compared to the 530 nm transition observed in other family members, and was not further perturbed upon addition of methylated DNA.

We propose that AlkB repairs 1-meA and 3-meC in DNA by oxidative demethylation. In such a mechanism, the lesions would be reverted to adenine and cytosine, formaldehyde would be generated and O₂ consumed. To demonstrate that AlkB directly reverts 1-meA in DNA to adenine residues, a non-radioactive substrate was prepared in which 76% of the adenine residues were methylated to form 1-meA. This was achieved by repeated treatments with DMS of an oligonucleotide containing adenine residues interspersed between inefficiently methylated thymine residues (Singer et al., 1983). Only 4% of the adenines were recovered as 3-meA, and 2% were 7- meA (Figure 5c). The low amount of 3-meA might be caused by instability of the glycosyl bond and loss of this modification during the extensive DMS treatments. The heavily methylated substrate was incubated with AlkB in the optimised assay conditions, the DNA hydrolysed and individual bases quantified by HPLC and A260 measurements. In the presence of AlkB, a decrease in the amount of 1-meA correlated with a stoichiometric increase in the amount of adenine recovered (Figure 5c). We conclude that AlkB converts 1-meA directly to adenine in DNA.

To determine whether formaldehyde was a reaction product, AlkB was incubated with MMS-treated poly(dA) annealed to poly(dT), the products were derivatised with pentafluorophenylhydrazine (PFPH) and analysed by gas chromatography (GC). One derivatised product, as indicated by the arrow in Figure 6a, arose only when the DNA substrate was methylated and Fe(ll) and αKG were present (Figure 6a, b), and had a mass spectrum (MS) identical to that previously described for the HCHO-PFPH adduct (Heck et al., 1982) (Figure 6c). The release of HCHO was also monitored by a coupled spectrophotometric assay using formaldehyde dehydrogenase (Heck et al., 1982), and again was detected only in complete assay conditions with the heavily methylated DNA substrate (turnover number 1.5 s-1 , data not shown). A yield of 140 μM HCHO correlated stoichiometrically with the amount of 1- meA in the DNA substrate (7-10% of the adenines), the amount of O₂ consumed (Figure 6d) and the succinate generated (data not shown). This stoichiometric relationship further verifies the proposal that AlkB is an αKG- dependent dioxygenase. In the absence of methylated DNA, an observed slow consumption of oxygen (Figure 6d) and αKG was consistent with a partial uncoupling of αKG decomposition and hydroxylation of the methylated DNA bases. Such uncoupling is a well-known property of this family of enzymes (de Jong et al., 1982; Myllyla et al., 1984).

Oxidative demethylation is an unprecedented mechanism of DNA repair. The proposed reaction mechanism by which AlkB repairs 1-meA and 3-meC in DNA is shown in Figure 1. Due to the stability of the N-C bond in 1-meA and 3-meC, demethylation by hydrolysis would be energetically unfavorable; consequently oxidative demethylation by reactive iron-oxygen species is required. Direct reversal of this base damage to unsubstituted parent residues would be a highly accurate form of DNA repair and agrees with in vivo observations that repair by AlkB is non-mutagenic (Dinglay et al., 2000). E. coli alkB mutants are more sensitive to alkylating agents during active growth than in stationary phase probably because 1-meA and 3-meC are produced in single-stranded regions of DNA in replication forks and transcription bubbles (Dinglay et al., 2000). DNA unfolds only transiently during replication and transcription, so it is beneficial that AlkB repairs its substrates not only in single-strands but also, and even more efficiently, after DNA reannealing (Figure 4). Interestingly, Caulobacter crescentus alkB expression is cell cycle regulated with a pattern similar to activities required for DNA replication (Colombi, et al., 1997).

Bacterial and human 3-methyladenine-DNA glycosylases and O⁶- methylguanine-DNA methyltransferases also repair bulkier ethylated adducts, but usually with lesser efficiency (Tudek et al., 1998). AlkB uses a completely different mechanism to these enzymes to remove aberrant methyl groups from DNA. It was therefore of interest to determine whether this activity can also repair ethyl adducts. To investigate whether E. coli AlkB can repair ethyl lesions generated in single stranded DNA, the ability of an alkB mutant to reactivate an ethyl iodide (Etl) treated single stranded DNA phage was examined. Survival of the ethylated M13 phage was greatly decreased in the alkB mutant compared with the wild type indicating that AlkB protein repairs DNA ethylation damage (Data not shown). A single stranded substrate containing 1-ethyladenine (1-etA) was prepared by treating poly(dA) with [14C]-Etl. In standard assay conditions, E. coli AlkB protein released [14C]- ethanol soluble material from this substrate. Analysis of the ethylated bases remaining in the polynucleotide showed that AlkB had completely removed the 1-etA lesions. A small amount of 3-etA was also present in the substrate but was not removed by AlkB activity. AlkB therefore repaired 1-etA but not 3-etA, which parallels its observed activity on 1-meA but not 3-meA. By quantifying the release of [14C]-ethanol soluble material, AlkB was observed to repair 1- etA with a similar but slightly lower efficiency than 1-meA (data not shown).

AlkB oxidizes the methyl group of 1-meA in DNA to release formaldehyde. Depending on whether AlkB oxidises 1-etA in DNA initially at carbon-1 or carbon-2 of the ethyl group, the product released could be either formaldehyde or acetaldehyde. To determine the nature of the aldehyde released, AlkB was incubated with [14C]Etl treated poly(dA) and the ethanol soluble [14C]-product derivatised with DNPH and examined by HPLC. The radiolabeled product co-chromatographed with the DNPH derivative of acetaldehyde and not formaldehyde (data not shown). AlkB therefore oxidizes at carbon-1 of the ethyl adduct generating acetaldehyde.

We have extended the family of αKG-Fe(ll) dependent dioxygenases to include AlkB, as recently proposed on theoretical grounds (Aravind & Koonin, 2001). These enzymes catalyse a variety of reactions including hydroxylations, desaturations and oxidative ring closures, and account for oxidation of proline in collagen, steps in the biosynthesis of several antibiotics and cellular metabolites, as well as biodegradation of selected compounds (Prescott et al., 2000; Ryle & Hausinger, 2002). No other members of this family are presently known to act on DNA, but it is notable that a fungal enzyme, thymine-7-hydroxylase, oxidises the methyl group of free thymine (Thornburg et al., 1993).

We suggest that oxidative demethylation requiring free-radical chemistry could be involved in other mechanisms of active removal of chemically stable adducts from DNA or histones, for example, the controversial 5- methylcytosine demethylase activity (Mayer et al., 2000; Smith, 2000) that is apparently important in epigenetic control.

Aravind and Koonin (2001 ) used sequence profile searches to identify several protein families that contain the 2-oxoglutarate (2OG) - Fe(ll) oxygenase fold. Protein families that form the 2OG-Fe(ll) oxygenase superfamily include the DNA repair protein AlkB, the extracellular matrix protein leprecan, RNA methylases, EGL-9 homologues implicated in smooth muscle biology, a series of uncharacterised proteobacterial proteins, e.g., those typified by YbiX, the prolyl hydroxylase family and the lysyl hydroxylase family. Despite being related to the 2OG-Fe(ll) oxygenase superfamily these proteins possess different functions as highlighted above.

A putative human homolog of the E.coli AlkB protein, ABH, was reported to partially complement the MMS sensitivity to an E. coli alkB mutant (Wei et al., 1996). We obtained 2 independent full length ABH cDNA clones from the IMAGE consortium (IMAGE IDs 1337358 and 2284109). Determination of their DNA sequences showed that both cDNAs contained ORFs encoding identical proteins, which we will refer to as ABH2, but their with amino- and carboxy- termini differed from those reported by Wei et al. (1996). These did not appear to be splice variants. The ABH2 amino acid sequence agreed with that predicted by conceptual splice of the genomic DNA sequence (accession AC008044) and a further cDNA (accession XM_007409). The ABH2 cDNA was amplified by PCR using Pfu polymerase, subcloned into the pET15b expression vector and overexpressed in an alkB22 derivative of E. coli BL21.DE3. Induction of the ABH2 protein in this strain by 50 μM IPTG at 37°C for 3hours resulted in 10 to 20 % solubility of the overexpressed protein. E. coli alkB mutants are defective in reactivation of MMS treated single stranded DNA phage (Dinglay et al., 2000). Despite previous reports overexpression of ABH2 protein failed to complement this phenotype (Figure 12), although overexpression of E. coli AlkB protein did complement the mutant phenotype. Again, despite claims (Wei et al., 1996) that ABH encoded an AlkB homolog, the purified ABH2 his-tagged protein did not release soluble radiolabelled material from a [¹⁴C]-methyl iodide treated poly(dA) substrate in our optimised AlkB assay conditions (data not shown) or after various modifications of this assay. We were therefore unable to support the previous suggestion that human ABH and therefore ABH2 is an AlkB homolog.

In order to identify a robust human AlkB homologue the inventors, through primary sequence analysis, have identified a number of additional cDNAs with albeit low sequence homologies (see below). These cDNA clones were then assessed in the complementation assay but more importantly in the in vitro DNA repair assay disclosed within this application in order to identify those with AlkB activity, e.g., removal of 1 -methyladenine from DNA.

The amino acid and cDNA sequence of HAB1 identified as a potential AlkB homologue through primary sequence analysis is shown in Figure 2. Utilising alignment and comparison with the E. coli AlkB amino acid sequence it is noted that HAB1 has 24.5% identity in relation to AlkB (Figure 7). This sequence comparison, and those that follow, were carried out using the GAP module of GCG Sequence Analysis Software Package Version 3.2. Gap uses the alignment method of Needleman and Wunsch (J. Mol. Biol. 48; 443-453 (1970)) that has been shown to be equivalent to Sellers (SIAM J. of Applied Math 26; 787-793 (1974). It should be noted that HAB1 has been previously described as DEPC-1 (Accession Number AB042029) encoding prostate cancer antigen-1.

Utilising the HAB-1 cDNA to search sequence databases a further cDNA, HAB2, previously defined as similar to prostate cancer antigen-1 (Accession number XM__058581 ), was identified. Alignment and comparison with the E. coli AlkB amino acid sequence demonstrate that HAB2 has 26.3% identity in relation to AlkB (Figure 8). Alignment and comparison of HAB1 and HAB2 is depicted in Figure 9. Both proteins contain the Fe(ll) binding motif, HXDX_NH, that is characteristic of α-ketoglutarate dependent dioxygenases and a conserved arginine that may be involved in α-ketoglutarate interactions. It should be noted that Aravind and Koonin (2001) identified a number of potential human AlkB homologues, including ABH2, although did not identify HAB1 or HAB2. In other words the prior art failed to provide functional human homologs based on conservative sequence alignment alone.

Full length HAB1 and HAB2 cDNAs were obtained from the IMAGE consortium (IDs 0624651 and 5179189, respectively). The ORFs were amplified by PCR and subcloned using Gateway technology (Invitrogen) to obtain plasmid constructs encoding GST-tagged fusion proteins. These plasmids were transformed into an E. coli alkB mutant and their ability to complement the defective reactivation of a MMS treated single stranded DNA phage was examined. Suprisingly, both plasmids fully complemented this alkB phenotype to regenerate reactvation equivalent to that in an alkB⁺ strain or in an alkB mutant overexpressing E. coli AlkB protein (Figure 12). This complementation provides the first evidence that both HAB1 and HAB2 are the first true human homologs of the E. co/7 AlkB protein. This is in clear contrast to ABH (and therefore, ABH2), previously identified as an AlkB homologue (see Wei et al. (1996) and Figure 12).

Recombinant HAB1 and HAB2 were overexpressed and purified to homogeneity utilising the GST fusion tags for further analysis in the in vitro AlkB assay described in Example 2. Unlike ABH2, HAB1 and HAB2 were shown to possess in vitro activity analogous to AlkB, e.g., repair of 1-meA and 3-meC DNA (see Figure 13 and Figure 14 and data not shown). Interestingly another cDNA identified in Macacca's (see Figure 11) did not possess this in vitro activity despite identification as a member of the 2OG-Fe(ll) oxygenase superfamily. The observation that ABH2 does not possess in vitro AlkB activity and that ABH1 and ABH2 poorly complement the E. coli alkB^' mutant is interesting in that amino acid sequence analysis indicates that ABH2 has a percent identity of 23.7% to AlkB (Figure 10) similar to that described for HAB1 and HAB2. Hence, primary sequence analysis at these levels of percent identity cannot be used to predict whether cDNAs possess AlkB activity, as assessed by phenotypic complementation or in an in vitro assay.

HAB1 and HAB2 proteins were also shown to remove 1-etA from [14C]- ethylated poIy(dA), however, both human proteins repaired 1-etA inefficiently, at less than 1 % of the rate of 1-meA (data not shown).

To investigate the tissue distribution of HAB1 and HAB2 the expression of the respective mRNAs were analysed as described in Example 3. The mRNAs of HAB1 and HAB2 were clearly detected in all of 10 different human tissues and several different cell lines derived from carcinomas by Northern hybridisation, but were present at variable levels (Figure 15). HAB1 mRNA was present at relatively high levels in spleen, prostate, bladder, lung and colon tissues, whereas HAB2 was highest in liver and bladder. In the cell lines derived from carcinomas, HAB2 mRNA was relatively high in HeLa while HAB1 mRNA was barely detectable in Hela cells or in two out of three cell lines derived from Burkitt's lymphomas (data not shown). Yellow fluorescent protein fusion constructs with the two human proteins showed that HAB2 localised diffusely throughout the nucleoplasm and accumulated in nucleoli, whereas HAB1 was present diffusely in both nuclei and cytoplasm.

Examples

EXAMPLE 1

In vitro testing of Alk B activity

[14C]Mel treated substrates and assay of AlkB protein. PoIy(dA), 0.6 mg (average length 310 residues) (Amersham Biosciences) was treated with 1 mCi [14C]Mel (58mCi/mmol) (Amersham Biosciences) in 0.8 ml 10 mM sodium cacodylate pH 7 at 30°C for 6 hours. The methylated polymer was ethanol precipitated, dissolved in 10 mM Tris-HCI, pH 8, and had a specific activity of 1860 cpm/μg. To prepare a double stranded substrate, [14C]-methylated poly(dA) was annealed to poly(dT) at 20°C. Poly(dC), 1.2 mg (average length 370 residues) in 1.3 ml 50 mM Hepes-KOH, pH 8, was similarly treated with [14C]Mel to yield a specific activity of 700 cpm/μg. His- tagged AlkB was overexpressed and purified as previously described by nickel-agarose affinity chromatography (Dinglay et al., 2000), except that EDTA was added to the sonication buffer, and the extract was dialysed against the same buffer without EDTA. A plasmid encoding FLAG-tagged AlkB was constructed by digestion of pBAR54 (Dinglay et al., 2000) with Nde1 and Ncol to remove DNA encoding the His-tag. This DNA fragment was replaced by a double stranded oligonucleotide encoding the FLAG-tag. The FLAG-tagged protein was purified by immnunoaffinity chromatography (Sigma). Purified AlkB was incubated with the [14C]Mel treated substrates in 50 mM Hepes-KOH, pH8, 75 μM Fe(NH4)2(SO4)2.6H2O, 1 mM αKG, 2 mM ascorbate, and 50 μg/ml BSA for 15 min at 37°C. The reaction was stopped by adding EDTA to 11 mM, and the substrate ethanol precipitated in the presence of carrier calf thymus DNA. Ethanol soluble radioactive material was monitored by scintillation counting.

[14C]Etl treated substrates and assay of AlkB protein.

As described for [14C]Mel treated substrates, where 0.6mg poly(dA) was treated with 250μCi [14C]Etl (31.5mCi/mmol) (ICN Biochemicals) except that the treatment was at 37°C for 18 hours. The specific activity of the ethylated poly(dA) was 120 cpm/μg. To determine whether AlkB released formaldehyde or acetaldehyde from [14C]-ethylated poly(dA), the ethanol soluble [14C]- labelled material was derivatised with 2,4-dinitrophenylhydrazone (DNPH) (Houlgate et al., 1989). The DNPH derivatives were analysed by reverse phase HPLC on a Phenomenex Hypersil column using a linear 50-90% methanol gradient in water and quantitated by scintillation counting. DNPH derivatised formaldehyde and acetaldehyde markers were monitored at A254.

Hexahistidine-tagged or GST-tagged fusion protein expression and purification

Expression plasmids encoding GST- or hexahistidine- tagged HAB1 or HAB2 were transformed into E. coli Codon plus. Cultures were grown to Aβoo 0.6 when protein expression was induced by the addition of 0.1 mM IPTG. Expression continued for 3.5 hours at room temperature. Harvested cells expressing hexahistidine- tagged proteins were lysed in a buffer containing 25 mM Tris-CI (pH 8.0), 300 mM NaCI, 10 mM imidazole, 1 mM EDTA and 1 mM β-mercaptoethanol either by French press or addition of lysozyme to 1 mg/ml and sonication. Lysates were clarified by centrifugation at 17 OOOg for 30 minutes at 4 °C and dialysed against two changes of 100 volumes of lysis buffer lacking EDTA. The dialysed lysates were loaded onto Ni²⁺-NTA agarose resin in batch format for 45 minutes at 4 °C after which the resin was washed with 50 column volumes of lysis buffer containing 40 mM imidazole but no EDTA. Specifically bound proteins were eluted with buffer containing 25 mM Tris-CI (pH 8.0), 150 mM NaCI, and 250 mM imidazole. Lysis buffer for GST-tagged proteins contained 50 mM Tris-CI (pH 8.0), 300 mM NaCI, 1 mM EDTA and 1 mM DTT. Clarified lysates were loaded onto GSH-Sepharose resin as described for hexahistidine-tagged HAB1 and HAB2 and washed with 50 column volumes of lysis buffer lacking EDTA. Specifically bound proteins were eluted using buffer containing 50 mM Tris-CI (either pH 8.0 or 8.5), 150 mM NaCI, 1 mM DTT and 10 mM reduced glutathione.

[14C]Mel treated substrates and assay of HAB1 and HAB2.

This was carried out as described for AlkB, although it is well known to those skilled in the art that reaction conditions could be varied in order to optimise the in vitro activity of HAB1 and HAB2, e.g., variation in cofactor concentration such as Fe(ll) or variation in pH. HPLC analysis.

Methylated adenine residues were released from [14C]-methylated poly(dA) by hydrolysis in 0.1 M HCI at 95°C for 1 h, and 3-meC from methylated poly(dC) by treatment with 90% formic acid at 180°C for 20 min. Methylated bases were analysed by HPLC on a Whatman Partisil 10 cation exchange column in 0.1 M ammonium formate, pH 3.6. A gradient of MeOH from 20 to 40% was applied to separate the methylated adenine derivatives, and a gradient from 5 to 40% MeOH was used to analyse 3-meC.

High level methylation of an oligonucleotide.

A 41-mer oligonucleotide, TTTTTT(ATTTTTT)5, was treated with 50 mM DMS in 75 mM sodium cacodylate, pH 7.4, at 30°C, 2 times for 2 h and then 4 times for 1 hour. Between each treatment the DMS was removed by centrifugation through a G25 Sephadex mini-column equilibrated in the same buffer. The level of methylation of adenine residues was examined by acid hydrolysis, HPLC and A260 measurements. The A260 ratio of adenine to 1- meA was 1.04 which was determined by monitoring known amounts of these purines in the same conditions.

GC/MS.

Oligo(dA) (25-mer) was treated with 500 mM MMS at 30°C for 30 min and, after removal of the MMS, annealed to oligo(dT). AlkB (38 μM) was incubated with this substrate (50 μM) in 50 mM HEPES-KOH, pH 8, containing 1 mM αKG, 75 μM Fe(NH4)2(SO4)2.6H2O, and 2 mM ascorbate in a final volume of 200 μl. After 10 min at 30°C, 200 μl 15 mM PFPH in 1.2 M phosphoric acid was added, and incubation continued at 50°C for 2 h. The reaction mixtures were extracted with 100 μl of 85:15 hexane:dichloromethane, and 20 μl analysed by GC/MS as described (Heck et al., 1982).

O₂ consumption.

50 mM HEPES, pH 8, with 1 mM αKG, 75 μM Fe(NH4)2(SO4)2.6H2O, 100 μM ascorbate, and 50 μM oligonucleotide substrate, prepared as described for GC/MS, were mixed for 3 min in a YSI 5300 oxygen electrode at 30°C to allow equilibration with atmospheric oxygen. The electrode was precalibrated with air saturated water (236 μM O₂). AlkB was added through a gas tight syringe to a final concentration of 9 μM.

EXAMPLE 2

Bacterial expression and functional complementation of E. coli alkB mutants by HAB1 and HAB2

The DNA sequence encoding for HAB1 , HAB2, AlkB and ABH2 (Accession numbers AB042029, XM_058581 , P05050 and XM_007409 respectively), were amplified using PCR oligonucleotide primers corresponding to the 5' and 3' sequences of cDNAs incorporating restriction enzyme consensus sites. The amplified sequences were ligated into bacterial expression vectors pET15b or pGEX6P-1 as described by the procedure in Sambrook, J. et al., Molecular Cloning: A Laboratory Manual, Cold Spring Laboratory Press, (1989). PGEX6P-1 may be purchased from Amersham Biosciences (Amersham Biosciences Europe GmbH, Munzinger Strasse 9, D-79111 Freiburg, GERMANY) is a bacterial expression vector designed for high-level intracellular expression of fusions with glutathione S transferase (GST). Purification of a GST tagged protein is simple and well known in the art, using mild elution conditions that minimize the risk of damage to the functionality of the target protein. The GST tag is easily detected and can be removed if required. Expression constructs were confirmed by restriction analysis and sequence analysis.

Alternatively, oligonucleotides were also used to prime DNA synthesis contained either attB1 or attB2 sites to facilitate recombinational cloning when using Gateway technology (Invitrogen). The amplified ORFs were cloned into the entry vector p221 by recombination and were subsequently shuttled into pDEST GST or pDEST 17 for expression as GST- or hexahistidine- tagged fusion proteins in E. coli respectively, or into pcDNA DEST 53 for expression as GFP-fusions in human cells.

Expression constructs of HAB1 , HAB2, AlkB, ABH2 and vector alone controls were then examined for their ability to complement the marked defect in processing methylation damage in single stranded DNA. The results are depicted in Figure 12.

Phage reactivation

We have previously used M13 phage to examine reactivation of MMS treated single stranded DNA in alkB mutants (Dinglay et al., 2000). To avoid the requirement for F' strains, here we have used single stranded φK phage provided by Dr. Kodaira (Kodaira et al., 1996). MMS treatment of this phage and estimation of reactivation were as previously described (Dinglay et al., 2000). Cells were grown in the presence of 100 μM IPTG for 3 hours at room temperature to induce overexpression of the cloned genes before transfecting with φK phage.

EXAMPLE 3

Northern Hybridisation

Multiple tissue Northern blots (Ambion 3142/3143) (Ambion, Inc., 2130 Woodward, Austin, TX 78744-1832, USA.) were probed, washed and stripped according to the manufacturer's instructions. Each blot contained 2 μg of polyA÷ RNA from 10 different normal tissues or 10 different cell lines derived from various carcinomas. Briefly, blots were pre-hybridised in hybridization buffer (Ambion) for 30 min at 42°C. Full length ORF probes were labeled with [³²P]-dCTP by PCR and purified through G-50 Sephadex spin columns. The probe was added directly to the hybridisation buffer to a specific activity of 106 cpm/ml. Blots were hybridised for 16 h at 42 °C before being washed twice with 0.3M NaCI, 0.03M citrate, 0.1 % SDS and twice with 0.15M NaCI, 0.015M citrate, 0.1 % SDS at 42°C. The bound probe was visualized by fluorography. A β-actin probe was used as a control to verify integrity of the samples on the blots.

References

Aravind, L. & Koonin, E. V. The DNA-repair protein AlkB, EGL-9, and leprecan define new families of 2-oxoglutarate- and iron-dependent dioxygenases.

Genome Biology 2, RESEARCH0007 (2001 ).

Aravind, L., Walker, D. R. & Koonin, E. V. Conserved domains in DNA repair proteins and evolution of repair systems. Nucleic Acids Res. 27, 1223-1242

(1999). Bodell, W. J. & Singer, B. Influence of hydrogen bonding in DNA and polynucleotides on reaction of nitrogens and oxygens toward ethylnitrosourea.

Biochemisry 18, 2860-2863 (1979).

Chen, B. J., Carroll, P. & Samson, L. The Escherichia co//^'AlkB protein protects human cells against alkylation-induced toxicity. J. Bacteriol. 176, 6255-6261 (1994).

Colombi, D. & Gomes, S. L. An alkB homologue is differentially transcribed during the Caulobacter crescentυs cell cycle. J. Bacteriol. 179, 3139-3145

(1997).

Dinglay, S., Trewick, S. C, Lindahl, T. & Sedgwick, B. Defective processing of methylated single-stranded DNA by E. coli alkB mutants. Genes Dev 14,

2097-2105 (2000).

Heck, H. D. A., White, E. L. & Cassanova-Schmitz, M. Determination of formaldehyde in biological tissues by gas chromatography/mass spectrometry. Biomedical mass spectrometry 9, 347-353 (1982). Hegg, E. L. et al. Herbicide-degrading a-keto acid-dependent enzyme: metal coordination environment and mechanistic insights. Biochemistry 38, 16714-

16726 (1999).

Houlgate, P. R., Dhingra, K. S., Nash, S. J. & Evans, W. H. (1989) Analyst 114, 355-360. de Jong, L., Albracht, S. P. & Kemp, A. Prolyl 4-hydroxylase activity in relation to the oxidation state of enzyme-bound iron. The role of ascorbate in peptidyl proline hydroxylation. Biochim. Biophys. Ada 704, 326-332 (1982).

Kodaira, K. I., Oki, M. M. K., Kimoto, H. & Taketo, A. (1996) J. Biochem. 119, 1062-1069.

Lindahl, T., Sedgwick, B., Sekiguchi, M. & Nakabeppu, Y. Regulation and expression of the adaptive response to alkylating agents. Ann. Rev. Biochem.

57, 133-157 (1988).

Mayer, W., Niveleau, A., Walter, J., Fundele, R. & Haaf, T. Demethylation of the zygotic paternal genome. Nature 403, 501-502 (2000).

Myllyla, R., Majamaa, K., Gunzler, V., Hanauske-Abel, H. M. & Kivirikko, K. I.

Ascorbate is consumed stoichiometrically in the uncoupled reactions catalysed by prolyl 4-hydroxylase and lysyl hydroxylase. J Biol Chem 10,

5403-5405 (1984). Pavel, E. G. et al. Circular dichroism and magnetic circular dichroism spectroscopic studies of the non-heme ferrous active site in clavaminate synthase and its interaction with a-ketoglutarate co-substrate. J. Am. Chem.

Soc. 120, 743-753 (1998).

Prescott, A. G. & Lloyd, M. D. The iron(II) and 2-oxoacid-dependent dioxygenases and their role in metabolism. Nat. Prod. Rep. 17, 367-383

(2000).

Rebeck, G. W. & Samson, L. Increased spontaneous mutation and alkylation sensitivity of Escherichia coli strains lacking the ogt O⁶-methylguanine-DNA methyltransferase. J. Bacteriol. 173, 2068-2076 (1991). Ryle, M. J., Padmakumar, R. & Hausinger, R. P. Stopped-flow kinetic analysis of Escherichia coli taurine/a-ketoglutarate dioxygenase: interactions with a- ketoglutarate, taurine, and oxygen. Biochemistry 38, 15278-15286 (1999).

Ryle, M. J. & Hausinger, R. P. Non-heme iron oxygenases. Curr. Opin. Chem.

Biol. 6, 193-201 (2002). Seeberg, E. & Berdal, K. G. in Base Excision Repair of DNA Damage: Repair of Alkylation Damage to DNA (ed. Hickson, I. D.) 151-168 (Landes

Bioscience, Austin, 1999).

Singer, B. & Grunberger, D. in Molecular Biology of Mutagens and Carcinogens: Reactions of Directly Acting Agents with Nucleic Acids 45-96

(Plenum Press, New York, 1983).

Smith, S. S. Gilbert's conjecture: the search for DNA (cytosine-5) demethylases and the emergence of new functions for eukaryotic DNA

(cytosine-5) methyltransferases. J. Mol. Biol. 302, 1-7 (2000). Thomburg, L. D., Lai, M.-T., Wishnok, J. S. & Stubbe, J. A non-heme iron protein with heme tendencies: an investigation of the substrate specificity of thymine hydroxylase. Biochemistry 32, 14023-14033 (1993).

Tudek, B., VanZeeland, A. A., Kusmierek, J. T. & Laval, J. (1998) Mutat. Res.

407, 169-176. Vaughan, P., Sedgwick, B., Hall, J., Gannon, J. & Lindahl, T. Environmental mutagens that induce the adaptive response to alkylating agents in

Escherichia coli. Carcinogenesis 12, 263-268 (1991).

Wei, Y., Carter, K. C, Wang, R. & Shell, B. K. Molecular cloning and functional analysis of a human cDNA encoding an Escherichia coli AlkB homolog, a protein involved in DNA alkylation damage repair. Nucleic Acids

Res. 24, 931-937 (1996).

Claims

1. Method for selecting a test compound that modulates HAB1 activity and/or HAB2 activity comprising providing a composition comprising HAB1 or HAB2, providing a test agent and assaying HAB1 activity and/or HAB2 activity in the composition, wherein a change in the activity of HAB1 and/or HAB2 indicates that the test compound may modulate HAB1 and/or HAB2 activity.

2. Method according to claim 1 wherein HAB1 has the amino acid sequence shown in figure 2 or a fragment or derivative or analogue thereof and HAB2 has the amino acid sequence shown in figure 3 or a fragment or derivative or analogue thereof.

3. Method according to claims 1-2 wherein the test compound inhibits HAB1 and/or HAB2 activity.

4. Method according to claims 1-2 wherein the test compound increases HAB1 and/or HAB2 activity.

5. Method according to claims 1-4 wherein the test compound is an agent for combating a proliferative disease.

6. Agent that modulates HAB1 and/or HAB2 activity.

7. Agent according to claim selected by the method of claims 1-4.

8. Agent according to any of claims 6-7 that modulates HAB1 and/or HAB2 expression.

9. Agent according to any of claims 6-8 wherein the agent comprises an antisense RNA or a small interfering RNA that inhibits HAB1 or HAB2 expression.

10. Composition comprising an agent according to any of claims 6-9 and a cancer therapeutic agent.

11. Composition according to claim 10 wherein the cancer therapeutic agent is a chemical DNA alkylating agent.

12. Pharmaceutical composition comprising an agent according to any of claims 6-9 or a composition according to any of claims 10-11 and a pharmaceutically acceptable carrier.

13. Kit-of-parts for separate, sequential or simultaneous administration comprising an agent according to any of claims 6-9 and a DNA alkylating agent.

14. Method of combating a proliferative disorder in a patient the method comprising administering to the patient an inhibitor of HAB1 and/or HAB2 activity.

15. Method according to claim wherein the inhibitor identifiable by the method of any of claims 1-5.

16. Method according to any of claims 14-15 wherein the patient is also administered a DNA alkylating agent agent.

17. Method according to claim 16 wherein the DNA alkylating agent is selected from the group of Temozolomide, Procarbazine, nitrogen mustards such as mechlorethamine (HN2), cyclophosphamide, ifosfamide, melphalan (L-sarcolysin) and chiorambucil; ethylenimines and methylmelamines such as hexamethylmelamine, thiotepa; alkyl sulfonates and thiosulfonates such as busulfan, methyl methanesulfonate (MMS) and methyl methanethiosulfonate; nitrosoureas and nitrosoguanidines such as carmustine (BCNU), lomustine (CCNU), semustine (methyl-CCNU), streptozocin (streptozotocin) and N-methyl-N'-nitro-N-nitrosoguanidine (MNNG); and triazenes such as dacarbazine (DTIC; dimethyltriazenoimidazole- carboxamide).

18. Method according to any of claims 14-17 wherein the proliferative disorder is a neoplastic disorder.

19. Method according to any of claims 14-18 wherein the proliferative disorder is cancer.

20. Method according to any of claims 14-19 wherein the cancer is selected from the group of glioma, glioblastoma multiforme, medulloblastoma, astrocytoma, ependymonas, oligodendrogliomas, lung cancer, liver cancer, cancer of the spleen, kidney, adrenal gland (including pheochromocytoma), thyroid gland, oesophagus, pituitary gland, lymph node, small intestine, pancreas, blood (lymphomas and leukaemias), colon, stomach, breast, endometrium, prostate including prostate adenocarcinoma, testicle, ovary, skin including melanoma, head and neck, oesophagus and bone marrow cancer.

21. Use of an inhibitor of HAB1 and/or HAB2 activity in the manufacture of a medicament for combating a proliferative disorder in a patient.

22. Use according to claim 21 wherein the inhibitor identifiable by the method of any of claims 1 -5.

23. Use according to any of claims 21-22 wherein a DNA alkylating agent has been administered to the patient

24. Use according to claim 23 wherein the DNA alkylating agent is selected from the group of Temozolomide, Procarbazine, nitrogen mustards such as mechlorethamine (HN2), cyclophosphamide, ifosfamide, melphalan (L-sarcolysin) and chiorambucil; ethylenimines and methylmelamines such as hexamethylmelamine, thiotepa; alkyl sulfonates and thiosulfonates such as busulfan, methyl methanesulfonate (MMS) and methyl methanethiosulfonate; nitrosoureas and nitrosoguanidines such as carmustine (BCNU), lomustine (CCNU), semustine (methyl-CCNU), streptozocin (streptozotocin) and N-methyl-N'-nitro-N-nitrosoguanidine (MNNG); and triazenes such as dacarbazine (DTIC; dimethyltriazenoimidazole- carboxamide).

25. Use according to any of claims 21-24 wherein the proliferative disorder is a neoplastic disorder.

26. Use according to any of claims 21-25 wherein the proliferative disorder is cancer.

27. Use according to claim 26 wherein the cancer is selected from the group of glioma, glioblastoma multiforme, medulloblastoma, astrocytoma, ependymonas, oligodendrogliomas, lung cancer, liver cancer, cancer of the spleen, kidney, adrenal gland (including pheochromocytoma), thyroid gland, oesophagus, pituitary gland, lymph node, small intestine, pancreas, blood (lymphomas and leukaemias), colon, stomach, breast, endometrium, prostate including prostate adenocarcinoma, testicle, ovary, skin including melanoma, head and neck, oesophagus and bone marrow cancer.

28. Assay for identifying AlkB and Alk B homologues by measuring the enzyme activity.