“9’
`
`Patent Office 333333?!”
`
`lllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll
`
`(11)
`
`EP 1 644 51 9 B1
`
`(12)
`
`EUROPEAN PATENT SPECIFICATION
`
`(45) Date of publication and mention
`of the grant of the patent:
`31.12.2008 Bulletin 2009/01
`-
`(21) Application number: 04737521]
`
`(51 ) Int CL:
`C120 1/6812005'”)
`
`(86) International application number:
`PCT/AU2004/000900
`
`(87) International publication number:
`(22) Date of filing: 05.07.2004
`WO 2005/003381 (13.01.2005 Gazette 2005/02)
`
`(54) METHOD FOR DETECTION OF ALKYLATED CYTOSINE IN DNA
`VERFAHREN ZUM NACHWEIS VON ALKYLIERTEM CYTOSlN IN DNA
`
`PROCEDE POUR LA DETECTION DE CYTOSINE ALKYLEE DANS L’ADN
`
`(84) Designated Contracting States:
`AT BE BG CH CY CZ DE DK EE ES FI FR GB GR
`HU IE IT LI LU MC NL PL PT RO SE SI SK TR
`
`(56) References cited:
`WO-A-20/05005660
`CA-A1- 2 462 928
`
`WO-A2-02/061 1 24
`
`(30) Priority: 04.07.2003 AU 2003903430
`04.08.2003 US 491995 P
`
`(43) Date of publication of application:
`12.04.2006 Bulletin 2006/15
`
`(73) Proprietor: JOHNSON & JOHNSON RESEARCH
`PTY LIMITED
`
`Eveleigh, NSW 1430 (AU)
`
`(72) Inventors:
`- TODD, Alison Velyian
`Glebe, NSW 2037 (AU)
`- FUERY, Caroline Jane
`Toongabbie, NSW 2146 (AU)
`‘ APPLEGATE, Tanya Lynn
`Greenwich, NSW 2065 (AU)
`
`(74) Representative: Brasnett, Adrian Hugh
`Mewburn EIIis LLP
`
`York House
`23 Kingsway
`London WCZB 6HP (GB)
`
`
`
`0 RAMIRO ALMUDENA R ET AL: "Transcription
`enhances AID-mediated cytidine deamination by
`exposing single-stranded DNA on the
`nontemplate strand." NATURE IMMUNOLOGY,
`vol. 4, no. 5, May 2003 (2003-05), pages 452-456,
`XP002302401 ISSN: 1529-2908
`
`- PHAM P. ET AL.: ’Processive AID-catalyzed
`cytosine deamination on single-stranded DNA
`simulates somatic hypermutation’ NATURE vol.
`424, no. 6944, July 2003, pages 103 - 107,
`XP002302398
`- BRANSTEITTER R. ET AL.: ’Activation-induced
`
`cytidine deaminase deaminates deoxycytidine
`on single-stranded DNA but requires the action
`of RNase’ PROCEEDINGS OF THE NATIONAL
`ACADEMY OF SCIENCES USA vol. 100, no. 7,
`April 2003, pages 4102 - 4107, XP002302399
`- PETERSEN-MAHRT S.K.: ’In vitro deamination of
`cytosine to uracil by apolipoprotein B editing
`complex catalytic subunit 1 (APOBEC1)’
`JOURNAL OF BIOLOGICAL CHEMISTRY vol. 278,
`no. 22, May 2003, pages 19583 - 19586,
`XP008063736
`. REIN T. ETAL.: ’Identifying 5-methylcytosine and
`related modifications in DNA genomes’ NUCLEIC
`ACID RESEARCH vol. 26, no. 10,1998, pages 2255
`- 2264, XP002143106
`- CLARK S.J. ET AL.: ’High sensitivity mapping of
`methylated cytosines’ NUCLEIC ACID
`RESEARCH vol. 22, no. 15, 1994, pages 2990 -
`2997, XP002210107
`
`Note: Within nine months of the publication of the mention of the grant of the European patent in the European Patent
`Bulletin, any person may give notice to the European Patent Office of opposition to that patent, in accordance with the
`Implementing Regulations. Notice of opposition shall not be deemed to have been filed until the opposition fee has been
`paid. (Art. 99(1) European Patent Convention).
`
`Printed by Jouve, 75001 PARIS (FR)
`
`EP1644519B1
`
`
`
`

`

`
`
`EP1644519 B1
`
`Description
`
`FIELD OF THE INVENTION
`
`10
`
`15
`
`20
`
`25
`
`30
`
`'
`
`35
`
`40
`
`45
`
`50
`
`[0001] The present invention relates to methods for detecting alkylated cytosine in DNA.
`Methods of the invention employ enzymes that differentially modify alkylated cytosine and cytosine. The presence of
`alkylated cytosine in DNA is determined by evaluating the level of enzymatic modification of the DNA following incubation
`of the DNA with at least one such enzyme. The detection of alkylated cytosine in DNA is useful for diagnostic and other
`purposes.
`
`BACKGROUND OF THE INVENTION
`
`[0002] At least seven different covalent base modifications have been identified in prokaryotic, eukaryotic, bacteri-
`ophage and/or viral genomes (1 ). In higher order eukaryotes the most abundant covalently modified base is 5-methyl-
`cytosine located 5' to guanosine in CpG dinucleotides. it has been hypothesised that methylation patterns play a role
`' in gene transcription, X chromosome inactivation, genomic imprinting, cell differentiation and tumourigenesis (2).
`[0003] The abnormal phenotype of cancer cells is due to qualitative and/or quantitative change. Sequence-based
`qualitative changes (genetic mutations) are preserved in the genomic DNA and this has facilitated their detection and
`characterisation. The inheritance of information on the basis of gene expression is known as epi-genetics. Methylation
`of cytosine bases in nucleic acid can effect epigenetic inheritance by altering expression of genes and by transmission
`of DNA methylation patterns through cell division. Cancer cells have been frequently shown to harbour both genetic and
`epi-genetic mutations.
`[0004] Neoplastic cells simultaneously harbour multiple abnormalities relating to methylation patterns. They frequently
`have both widespread genomic hypomethylation as well as more regional areas of hypermethylation (1). Regional
`methylation of normally unmethylated CpG islands located within promoter regions of genes has been shown to be
`correlated with the down regulation of the corresponding gene. This hypermethylation can serve as an alternative
`mechanism to coding region mutations for the inactivation of tumour suppressor genes. Examples of genes which have
`CpG island hypermethylation in association with human tumours include p16 (lung, breast, colon, prostate, renal, liver,
`bladder, and head and neck tumours), E—cadherin (breast, prostate, colon, bladder, liver tumours), the von—Hippel Lindau
`(VHL) gene (renal cell tumours), BRCA1 (breast tumours), p15 (leukemias, Burkitt lymphomas), hMLH1 (colon), ER
`(breast, colon, lung tumours; leukemias), HlC1 (brain, breast, colon, renal tumours), MDG1 (breast tumours), GST—rr
`(prostate tumours), 05-MGMT (brain tumours), calcitonin (carcinoma, leukemia),and myo-D(bladder tumours) (1, 3).
`[0005] The converse situation has also been reported, whereby CpG hypomethylation is thought to contribute to
`neoplastic progression. For instance, the urokinase CpG island was found to be hypermethylated in early stage, non-
`metastatic breast tumour cells but was hypomethylated in highly metastatic breast tumor cells (4). Similarly, hypometh-
`ylation of a region within the metastasis-associated S100A4 gene has been hypothesized as the mechanism of gene
`activation in colon adenocarcinoma cell lines (5).
`[0006] At least eight different methods, along with several variatiOns, allow characterisation of 5-methylcytosine or
`related modified bases in DNA genomes (2). Each method has advantages and disadvantages in terms of specificity,
`resolution, sensitivity and potential artefacts.
`[0007] The total nucleic acid base composition of a genome can be determined by hydrolysing DNA to its constituent
`nucleotides, either chemically or enzymatically, and then fractionating and analysing the composition by standard meth-
`ods (chromatography, electrophoresis and high pressure liquid chromatography). This approach quantifies the amount
`of modified bases present in the genome, but does not give any information on which part of the genome was originally
`modified. Dinucleotide composition and frequency can be determined by nearest-neighbour analysis, but again this
`method produces only limited sequence information. Neither of these methods are genome specific, and contamination
`of samples by DNA from viruses and other endoparasites can lead to misleading results.
`[0008] More specific methods exist which can provide data on exactly where in the sequence of the genome modified
`bases exist. Genomic DNA can be analysed by restriction enzymes that are sensitive to methylation. With this method,
`however, the number of sites that can be examined is limited by the number of sequences recognized by methylation
`sensitive restriction enzymes. Sequencing would provide sequence-specific information, but methylation patterns are
`not preserved during PCR or when eukaryotic DNA is amplified in bacteria through molecular cloning.
`[0009]
`It is necessary to differentially modify the bases, in a methylation specific manner, to produce a modified
`sequence where the methylation-specific changes are retained during sequencing protocols. There are currently three
`protocols that rely on analysis of differential base modification. All of these protocols involve modification of DNA,~induced
`by chemical treatment of the DNA followed by analysis of the DNA sequence.
`[0010] Hydrazine (N2H4), permanganate (Mn04'), and bisulfite (HSOs') all differentially modify cytosine bases in
`genomic DNA depending on the methylation status of the cytosine base.
`
`

`

`
`
`EP 1 644519 B1
`
`[0011] Hydrazine has a lower reactivity with 5-methylcytosine than with cytosine or thymine. After incubation of DNA
`with hydrazine the DNA is run on a sequencing gel. Comparison of the hydrazine-treated DNA with DNA treated with
`other base-specific chemical cleavage compounds allows the sequence of the DNA to be determined. In hydrazine-
`treated DNA samples 5—methylcytosine—containing sequence positions produce an absence or reduced intensity of bands
`compared to the cytosine and cytosine + thymidine specific ladders of sequencing reactions from genomic DNA. Thus
`the hydrazine protocol produces a negative result that correlates with the presence of 5-methylcytosine. Unambiguous
`identification of 5-methylcytosine requires the generation of a positive signal. A further disadvantage of hydrazine mod-
`ification for the identification of 5—methylcytosine is that pg of template DNA is required.
`[0012] Potassium permanganate, at weakly acidic pH and room temperature, reacts preferentially with thymine and
`5—methylcytosine, and only weakly with cytosine and guanine. After incubation of DNA with permanganate the DNA is
`run on a sequencing gel. Comparison of the pennanganate-treated DNA with DNA treated with other base-specific
`chemical cleavage compounds allows the sequence of the DNA to be determined. Permanganate oxidation of DNA can
`therefore be used to discriminate between cytosine and 5-methylcytosine (6). Although the permanganate protocol
`produces a positive result, and thus has an advantage over the hydrazine protocol, permanganate does react weakly
`with cytosine and hence discrimination of cytosine versus 5-methylcytosine depends on a difference in the intensities
`of their bands on the sequencing gel. A further disadvantage of permanganate modification is that pg of template DNA
`is again required.
`[0013] Bisulfite treatment of genomic DNA deaminates unmethylated cytosine bases in the nucleic acid template to
`uracil, whereas 5-methylcytosine is resistant to deamination. Bisulfite has little activity on cytosine bases in double
`stranded DNA and so genomic double stranded DNA is preferably denatured to single stranded DNA. The standard
`bisulfite modification protocol uses incubation in alkali (NaOH) to denature double stranded DNA to single stranded DNA
`(7). Bisulfite deaminates cytosine slowly and incubation times have to achieve a compromise between complete deam-
`ination of all cytosine and fragmentation of DNA after long incubations. Protocols for bisulfite modification use a range
`of incubation times, generally from 4 to 20 hours incubation in bisulfite.
`[0014] Grunau et al (8) studied optimum conditions for bisulfite-mediated deamination of cytosine and found that 4
`hours incubation at 55°C gave 99% deamination of cytosine, but under these conditions 84 to 96% of the DNA was
`degraded, reducing yields for subsequent steps. Further, 5-methylcytosine is deaminated by heat at a greater rate than
`cytosine. For example, the rate of deamination of 5-methylcytosine at 60°C is 1.5 times higher than that of cytosine (9).
`lncubations in bisulfite at lower temperatures reduce fragmentation of DNA but the incubation times have to be extended
`to 14 to 20 hours to achieve full deamination of cytosines. Bisulfite modification requires approximately 10 ng of DNA
`for subsequent analysis using PCR-based methods.
`[001 5] The modified DNA sense and anti-sense strands produced by bisulfite modification are no longer complementary
`and therefore subsequent amplification by PCR must be performed with primers that are designed to be strand specific
`that is, the primers are complementary to either the modified sense strand or the modified anti-sense strand. When the
`region of interest is amplified by PCR, uracil (previously cytosine) is converted to thymine and 5—methylcytosine is
`converted to cytosine (7). The PCR products (amplicons) can be subsequently analysed by standard DNA sequencing
`(7) or other PCR—based techniques that produce sequence information such as methylation-specific PCR (10) or REMS—
`PCR (36), and analysis with restriction enzymes (3) or methylation-specific probes (11).
`[001 6] Although the bisulfite method has advantages in terms of ease of use and sensitivity over other existing protocols,
`potential artefacts can arise from the experimental protocol (2) namely not all cytosines are converted to uracil, a small
`percentage of 5—methylcytosine is converted to thymidine (12) (DNA polymerases do not distinguish between uracil and
`thymine) and there can be a loss of DNA from fragmentation caused by the long incubations and non-physiological
`buffers required (8). The full protocol is long and laborious involving 2 to 3 days of manipulation and at least 4 to 20
`hours of incubation in bisulfite before results are obtained. The rate-limiting step in all epigenetic studies is sample
`preparation using the bisulfite modification protocol.
`[0017] DNA extracted from many types of specimens including normal and tumour tissue, paraffin embedded tissues,
`as well as plasma and serum has been shown to contain aberrantly methylated sequences using the combination of
`bisulfite treatment and analysis by PCR-based methods (4,13,14).
`[0018] A variety of enzymes with the ability to deaminate cytosine bases have been described. Cytidine Deaminase
`(EC 3.5.4.5.) converts cytidine to uridine and is widely distributed in prokaryotes and eukaryotes. Cytosine Deaminase
`(EC 3.5.4.1 .) converts cytosine to uracil. Deoxycytidine Deaminase (EC 3.5.4.14.) converts deoxycytidine to deoxyuridine
`and Deoxycytidilate Deaminase converts deoxycytidine-S-phosphate to deoxyuridine-S-phosphate. These enzymes
`show different degrees of substrate specificity depending on the source of the enzyme. The ability of Cytidine Deaminase
`and Cytosine Deaminase to discriminate between 5—methylcytidine and 5-methylcytosine and their unmethylated ana-
`logues as substrates (respectively) is species specific. Cytidine Deaminase from humans can deaminate, with varying
`efficiency, numerous cytidine derivatives including cytosine, deoxycytidine, and 5-methylcytidine (15,16). Cytosine
`Deaminase from Pseudomonas can utilise 5—methylcytosine (17) while the enzyme produced by enterics can only use
`cytosine as a substrate. Cytosine Deaminase from the fungus Aspergillus fumigatus and the yeast enzyme can utilise
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`50
`
`55
`
`

`

`EP1644519 B1
`
`5-methylcytosine as a substrate (18,19).
`[0019] Apolipoprotein B mRNA Editing Enzyme (ApoBRe) is the central component of an RNA editsome whose phys—
`iological role is specifically to deaminate the cytosine base at position #6666 ofthe apoB mRNA to uracil in gastrointestinal
`tissues creating a premature stop codon (20, 21). The catalytic component with cytidine deaminase activity is called
`Apolipoprotein B mRNA Editing Enzyme Catalytic Polypeptide 1 (APOBEC1). Although mRNA is the physiological
`substrate of this enzyme there is some evidence that it has activity on DNA in vivo. Misexpression of Apolipoprotein B
`mRNA Editing Enzyme in transgenic mice predisposes to cancer (22) and expression of human Apolipoprotein B mRNA
`Editing Enzyme in E. coli results in a mutator phenotype where there is a several 1000-fold enhanced mutation frequency
`seen at various loci in UNG-deficient strains.
`
`[0020] UNG is an enzyme involved in the repair of U:G mismatches caused by spontaneous cytosine deamination
`and deficiency in this enzyme prevents cells from repairing deaminated cytosines in their genome (23). Sequencing of
`DNA showed that mutations were triggered by conversion of cytosine to uracil in DNA. There appears to be some context
`specificity in the small stretches of DNA studied in this model (23) with a requirement for a 5’ flanking pyrimidine. This
`is despite that fact that the cytosine base (#6666) exclusively targeted for deamination by this enzyme in the physiological
`RNA substrate has a 5’flanking purine (adenosine). Deamination of cytosines with 5'flanking pyrimidines by Apolipoprotein
`B mRNA Editing Enzyme may require factors not supplied in the E. coli model.
`[0021] Recent work by Petersen-Mahrt & Neuberger (24) investigated the deamination activity of Apolipoprotein B
`mRNA Editing Enzyme in vitro on DNA substrates. They found no activity on double stranded DNA but cytosine bases
`in chemically synthesized single stranded DNA substrates were readily deaminated with 57% deamination of three
`cytosine bases in 120 minutes of incubation with a crude extract of enzyme. The activity of the enzyme appeared to be
`slightly higher when treated with RNase. The authors calculated that one molecule of Apolipoprotein B mRNA Editing
`Enzyme in their crude extract couid deaminate a single cytosine base in a chemically synthesised single stranded DNA
`substrate in 10 minutes. They attributed this slow rate of deamination to the fact that their assay was likely to be sub-
`optimal. This was attributed to the lack of other factors required for activity that were not expressed in the E. coli host,
`that the human enzyme might not properly fold in the E. coli host, and the fact that any post—translation modifications
`required for activity would not be supplied by the E. coli host.
`[0022] Activation-Induced Cytidine Deaminase (known as AID or AlCDA) is a Bocell specific protein. Expression of
`Activation-Induced Cytidine Deaminase is a pre-requisite to class—switch recombination, a process mediating isotype
`switching of immunoglobulin, and somatic hypermutation, which involves the introduction of many point mutations into
`the immunoglobulin variable region genes. The mode of action of Activation-lnduced Cytidine Deaminase is unknown.
`Current theories focus on the fact that Activation-Induced Cytidine Deaminase has sequence motif homology with Apol-
`ipoprotein B mRNA-Editing Enzyme and Cytidine Deaminase.
`[0023] An early theory on the mode of action of Activation—induced Cytidine Deaminase suggested that the hypothe-
`sised RNA-editing function of the enzyme might be involved in editing mRNAs that encode proteins essential for class-
`switch recombination and somatic hypermutation. The theory with most experimental support suggests that Activation-
`lnduced Cytidine Deaminase functions as a DNA-specific cytidine deaminase. This model suggests that Activation-
`lnduced Cytidine Deaminase deaminates cytosine bases in somatic hypermutation hotspot sequences to produce G:U
`mismatches and that these are differentially resolved to effect somatic hypermutation or class switch recombination (25).
`Evidence for the latter theory includes the suggestion that somatic hypermutation is initiated by a common type of DNA
`lesion, and that there is a first phase of hypermutation that is specifically targeted to dC/ dG pairs. This would require
`Activation-Induced Cytidine Deaminase to have cytidine deaminase activity on DNA. All published work on Activation-
`lnduced Cytidine Deaminase has focused on determining the in vivo substrate to elucidate the role of the enzyme in
`somatic hypermutation and isotype switching of immunoglobulin.
`[0024] Research by various laboratories has showed that human Activation-Induced Cytidine Deaminase can deam—
`inate cytosine on single stranded DNA in vitro (26-29) but not on single stranded RNA (26, 27). Activity of Activation-
`lnduced Cytidine Deaminase on double-stranded DNA in vitra is limited to DNA coupled to transcription factors. It has
`been hypothesised that transcription allows deamination of double stranded DNA by generating secondary structures
`that provide single-stranded DNA substrates such as stable R loops and stem loops (28). These secondary structures
`can be mimicked in vitro by producing bubbles, or loops, of centrally located noncomplementary regions of DNA, which
`will be single stranded, between complementary regions ofdouble stranded DNA. Activation—Induced Cytidine Deaminase
`deaminates cytosines in such bubbles. The efficiency of deamination depends on the length of the single stranded
`bubble. Bransteitter et al. (27) measured the percent of a chemically synthesised double stranded DNA substrate deam»
`inated in 5 minutes of incubation and showed that substrates with 1 nucleotide bubbles were notdeaminated, 3 nucleotide
`bubbles showed 5% deamination, 4 nucleotide bubbles showed 8 % deamination, 5 nucleotide bubbles showed 35 %
`deamination and 9 nucleotide bubbles showed 56 % deamination.
`
`It has been hypothesised that Activation—Induced Cytidine Deaminase activity would be restricted to the phys-
`[0025]
`iological target (the immunoglobulin loci) because rampant DNA deaminase activity would be harmful to the cell. There
`is some suggestion that the deaminase activity of Activation-Induced Cytidine Deaminase is sequence specific (30),
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`

`

`EP1 644 519 B1
`
`and it is hypothesised that Activation-Induced Cytidine Deaminase would show greatest activity on the somatic hyper-
`mutation hot-spot sequence RGYW (a sequence commonly mutated in the variable region of the immunoglobulin gene).
`Bransteitter et al. (27) showed that in vitra Activation-Induced Cytidine Deaminase had approximately three-fold higher
`activity on two hot-spot sequences compared with non—hot-spot sequences. Conversely, Dickerson et al. (26) found that
`the deaminase activity of Activation—Induced Cytidine Deaminase was sequence specific, but that cold-spot sequences
`(sequences of the variable region of the immunoglobulin gene that have never been found to be mutated in vivo) were
`deaminated equallywell as hot-spot sequences, and that some hot-spot sequences were deaminated at only background
`levels.
`
`[0026] Work by Pham et al. (31) tested the ability of Activation-induced Cytidine Deaminase to deaminate cytosine
`bases in vitro using a large single stranded DNA template. in these experiments, the single stranded DNA template was
`a phage circular DNA substrate containing a 230-nucleotide target of the IacZa reporter sequence as part of a 365-
`nucleotide single—stranded gapped region. incubations were carried outwith 500 ng of the double-stranded phage DNA
`substrate with a 40-fold excess of enzyme in a 10 mM TRIS buffer (pH 8.0) with 1 mM EDTA and 1 mM dithiothreitol at
`37°C for 20 minutes. The spectra of mutations were assessed by transfecting mutated phage (which gave white or light
`blue plaques) into UNG-deficient E. coli with subsequent sequencing of clones. Under the test conditions used the
`deamination activity of Activation-Induced Cytidine Deaminase was found to vary with sequence context, and the authors
`hypothesised that their results suggested the enzyme was a mobile molecule that processively deaminated cytosine
`molecules in the single stranded DNA.
`[0027]
`Pham etal. (31 ) also described a protocol for measuring the deaminiation activity of Activation-Induced Cytidine
`Deaminase in a trancripticnally active version of their Phage substrate. lncubations were carried out with 30 nM of the
`double-stranded phage DNA substrate in a 50 mM HEPES buffer (pH 7.5) with 1 mM EDTA and 10 mM MgCl2 at 37°C
`for 30 minutes. The incubations included T7 RNA polymerase and rNTPs to produce transcriptionally active DNA which
`is a more accessible substrate for the Activation—Induced Cytidine Deaminase (27). These incubations showed that
`deamination mediated by Activation-induced Cytidine Deaminase on the non-transcribed strand required RNA polymer-
`ase (active transcription) and that deamination on the transcribed strand, "protected" as an RNA-DNA hybrid, occurs at
`an approximately 15-fold lower rate. These incubations also demonstrated favoured deamination occurred in hotspot
`motifs.
`
`[0028] Models that involve ectopic expression of Activation-Induced Cytidine Deaminase in vivo show untargeted
`cytosine deamination, that is deamination of genes other than the variable region of the immunoglobulin gene. For
`example, human Activation—Induced Cytidine Deaminase expressed in E. coli, which obviously lacks the human immu-
`nogiobulin target gene, produces context specific deaminations in genes screened for mutations (30). The reason for
`this context specific deamination was not examined.
`[0029] Bransteitter et al. (27) recently incubated human Activation—Induced Cytidine Deaminase with a variety of
`chemically synthesized nucleic acid substrates in vitro. This work showed that, in a very simple model, Activation-induced
`Cytidine Deaminase was capable ofdeaminating cytosine bases with 1 O-fold higherspecific activity than 5-methylcytosine
`bases. The model involved incubating Activation-induced Cytidine Deaminasewith chemically synthesized single strand-
`ed DNA molecules with either 27 or 33 nucleotides, including either 1 or 2 cytosine bases, with no complimentary DNA
`strand present. These artificial substrates were present in high concentration, 100 nM, in a two-fold excess of Activation-
`Induced Cytidine Deaminase. The ability of Activation-induced Cytidine Deaminase to differentially convert cytosine
`bases to uracil, with no or little activity on 5-methylcytosine, in a complex mixture of genomic DNA extracted from an
`individual where there are a multiplicity of mega-base fragments with a multiplicity of different sequence contexts of
`cytosine bases with both sense and complementary antisense strands present was neither tested nor considered.
`[0030] The deaminase activity of Activation-induced Cytidine Deaminase is inhibited by 1 ,10-phenanthrcline, a strong
`chelator, but not by EDTA, a weaker chelator. This suggests that Activation-Induced Cytidine Deaminase requires a
`tightly bound metal ion, possibly zinc, for deaminase activity (27, 29). Activation-Induced Cytidine Deaminase retains
`deaminase activity over salt levels of 50 to 150 mM, can tolerate moderate levels of EDTA (5 to 10 mM), works at a wide
`range of pH (from 7.6 to 9.0 were tested) and works with varying efficiencies from room temperature to 37°C (26). These
`conditions are conducive to retaining the integrity of genomic DNA without fragmentation. Activation—induced Cytidine
`Deaminase is still active after being heated at 65°C for 30 minutes (26).
`[0031] Mutant forms of enzymes can exist in nature (e.g. allelic variants and forms arising from in viva mutations) or
`can be artificially generated. Methods for generating mutant proteins are known in the art (39). Mutations can be artificially
`generated either following a rational approach, such as where specific amino acid substitutions, deletions or additions
`are generated. or they can be randomly generated, and the mutant form of the protein tested for the desired activity.
`[0032] Enzymes which modify DNA require only a few hours incubation. Purified restriction enzymes, for example,
`require only 1 hour incubation in optimal conditions to fully cleave double stranded DNA. Bransteitter eta]. (27)measured
`95 % conversion of cytosine to uracil by Activation-Induced Cytidine Deaminase in a chemically synthesized single-
`stranded DNA substrate in 16 minutes, and 56 % conversion of cytosine to uracil in a synthetic substrate with a 9
`nucleotide single stranded bubble after 5 minutes. This is thus a fast reaction. However, work by other groups, with
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`4o
`
`50
`
`55
`
`

`

`
`
`
`
`
`
`EP1644 519 B1
`
`different reaction conditions, have shown that only 10 % of a chemically synthesized single stranded DNA substrate
`containing one cytosine was converted to uracil after 30 minutes of incubation with Activation-Induced Cytidine Deam-
`inase (26).
`
`SUMMARY OF THE INVENTION
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`In one aspect of the present invention there is provided a method for detecting the presence or level of alkylated
`[0033]
`cytosine in a sample of genomic or mitochondrial double stranded DNA from an individual, the method comprising:
`
`(a) converting at least one region of the double stranded DNA to single stranded DNA;
`(b) reacting a target region of the single stranded DNA from step (a) with at least one enzyme having cytidine
`deaminase activity, the enzyme differentially modifying alkylated cytosine and cytosine; and
`(c) determining the level of enzymatic modification of the target region by the enzyme.
`
`[0034] Generally, the reaction conditions under which the enzyme is used will be such that the enzyme reacts sub—
`stantially only with either alkylated cytosine or cytosine but not both.
`[0035]
`Preferably, the enzyme will be capable of reacting substantially with only one of alkylated cytosine or cytosine.
`[0036]
`Preferably, the conversion of the region of the double stranded DNA to single stranded DNA will comprise at
`least partially separating the two strands. Separation of the strands may for instance be achieved by heat denaturation
`of the DNA or the use of strand displacement probes. Other techniques that may be employed include chemical or
`enzymatic denaturation of the double stranded DNA. The method may also comprise inhibiting annealing of the two
`strands of the double stranded DNA together once they have been separated to facilitate access to the target region of
`the single stranded DNA by the enzyme.
`[0037] One or more probes capable of hybridising with a respective strand of the double stranded DNA may be utilised
`to inhibit annealing of the separated strands. When a plurality of probes are used, the probe(s) may hybridise with only
`one of the strands, or one or more of the probes may hybridise with one strand and the remaining probe or probes with
`the other strand.
`
`[0038] Accordingly, a method of the invention may further comprise hybridising at least one probe with a strand of the
`double stranded DNA following separation of the two strands to inhibit annealing of the strands together and thereby
`facilitate access to the target region of the single stranded DNA by the enzyme.
`[0039] The or each probe will normally be an oligonucleotide and may be selected from the group consisting of sense
`probes, looping probes for forming a loop in the single stranded DNA for access of the enzyme to the target region,
`antisense probes, and combinations thereof. More generally, a probe may hybridise with a single contiguous region of
`a strand of the double stranded DNA, or separate discrete upstream and downstream regions of the strand which flank
`the target region of the strand being evaluated for the presence or level of alkylated cytosine.
`[0040]
`In the former instance, at least two such probes may be utilised, wherein one of the probes hybridises with a
`region of the strand downstream of the target region, and a further of the probes hybridises with the strand upstream of
`the target region such that hybridisation of the other strand of the double stranded DNA to the target region is inhibited.
`[0041]
`In the latter instance, the probe may have a sequence such that when hybridised with the strand the spaced
`apart upstream and downstream regions of the strand are drawn toward each other forming a loop or bubble which
`incorporates the target region. The probe may for instance have opposite end regions which hybridise with the strand
`and a middle region of non-complementary sequence that does not hybridise with the target region of the strand such
`that a loop or bubble incorporating the target region is formed and hybridisation of the other strand of the double stranded
`DNA with the target region is thereby inhibited. To facilitate the formation of the loop or bubble, the middle region of the
`probe may incorporate inverted repeats that hybridise together following hybridisation Of the probe with the strand.
`[0042] To detect the presence or level of alkylated cytosine in the target region of the single stranded DNA reacted '
`with the enzyme, the target region will typically be amplified and the resulting amplicon(s) analysed for sequence mod-
`ifications arising from the enzymatic modification of the target region by the enzyme. Hence, a method of the invention
`may further comprise:
`
`amplifying the target region of the single stranded DNA reacted with the enzyme utilising a process involving ther-
`mocycling and primers to obtain an amplified product; and
`
`analysing the amplified p