EP2464748A1

EP2464748A1 - Detection and analysis of methylation in nucleic acid sequences

Info

Publication number: EP2464748A1
Application number: EP10807798A
Authority: EP
Inventors: Horace R. DREW; Peter Laurence Molloy; Glenn S. BROWN
Original assignee: Commonwealth Scientific and Industrial Research Organization CSIRO
Current assignee: Commonwealth Scientific and Industrial Research Organization CSIRO
Priority date: 2009-08-12
Filing date: 2010-08-12
Publication date: 2012-06-20
Also published as: WO2011017760A1; AU2010282225A1; EP2464748A4

Abstract

Provided herein are methods for the identification of methylated cytosine bases in a nucleotide sequence, the method comprising the steps of: (a) generating fragments of a DNA molecule comprising the nucleotide sequence such that the fragments comprise at least one single-stranded terminus comprising one or more bases and wherein at least one base within the single-stranded terminus is an unmethylated cytosine or methylcytosine; (b) incubating the fragments with bisulfite under non-denaturing conditions such that only unmethylated cytosines within the at least one single- stranded terminus are converted to uracil; and (c) detecting fragments comprising methylcytosine or uracil bases in the at least one single-stranded terminus. Also provided are methods for the diagnosis of, or determination of susceptibility to, diseases associated with aberrant cytosine methylation.

Description

Detection and analysis of methylation in nucleic acid sequences

Field of the Invention

The present invention relates generally to methods for the detection and analysis of methylation at cytosine residues within nucleotide sequences. The methods may be used for the analysis of methylation extent and/or pattern within DNA segments or genes of any size, or on a genome-wide basis. Accordingly, the methods described herein find application, inter alia, in the analysis of gene expression regulation and in the diagnosis of, or determination of susceptibility or predisposition to, diseases associated with aberrant DNA methylation levels or patterns.

Background of the Invention

The reference in this specification to any prior publication (or information derived from it), or to any matter which is known, is not, and should not be taken as an acknowledgment or admission or any form of suggestion that that prior publication (or information derived from it) or known matter forms part of the common general knowledge in the field of endeavour to which this specification relates. The methylation of position 5 on the cytosine ring, giving 5-methylcytosine (meC), is one of the most common post-synthetic modifications of DNA and is found in organisms from bacteria to higher eukaryotes including plants and mammals. The context of cytosine methylation varies from the dcm system of bacteria (methylation at CCWGG sites), methylation of cytosines in CpG sites as found widely in animals and plants, methylation of CpNpG sites in plant genomes, and asymmetric cytosine methylation that has been observed in particular in plant and fungal systems.

Cytosine methylation plays an important role in gene regulation. For example methylation within gene promoters typically leads to transcriptional silencing. In turn, cytosine methylation is involved in the regulation of a variety of cellular processes as well as in development. Increasingly, aberrant cytosine methylation patterns have been implicated in human disease, in particular cancer.

Accordingly, there is much interest in the analysis of the extent of cytosine methylation and the locations of methylated cytosine residues. However the approaches currently available for studying cytosine methylation suffer from a number of drawbacks that limit their applicability to genome wide analysis with sufficient resolution.

Most commonly, methylation analysis is carried out by treating the DNA with sodium bisulphite. Under standard conditions of treatment of single-stranded DNA with 2M Sodium bisulfite at pH 5 and temperatures of 55° - 80⁰C unmethylated cytosines are converted to uracil (U) while methylated cytosines remain unreactive. The resulting sequence difference in products from methylated compared with unmethylated DNA allows a range of approaches for sequence specific analysis of DNA methylation (see Clark et a/. 2006, Nature Protocols I: 2353-2364). For genome-wide application a genomic DNA sample is typically first fragmented randomly by sonication and/or mechanical shearing, and PCR-amplifiable synthetic linkers are added by ligation to both ends of the many small fragments so obtained (typically of size 200 to 1000 bp). Those small, ligated, double-stranded DNA molecules are next denatured fully into single strands using heat and/or high pH, and are treated with sodium bisulfite thereby providing a means of discriminating cytosine from methylcytosine at any location within a complex genome. Such fragmented, ligated and chemically-converted DNA is then amplified by PCR, using sequences from its two synthetic linkers as PCR primers, and may be analysed for its detailed informational nature using a microarray or high-throughput sequencing.

However this typical bisulfite treatment involves complete or near complete conversion of unmethylated cytosines to uracil. The low, essentially three-base complexity (A, G or T = U) of such completely-converted DNA makes this a difficult approach to use with microarray-based detection methods, which rely typically on four-base hybridization (A, G, T or C) to obtain specificity. Similarly, when such three-base samples (A, G or T = U) are analysed by high-throughput sequencing, the mapping of those sequences back to unique sites in complex genomes becomes a major challenge.

Disclosed herein is a novel approach to the identification of cytosine methylation at single nucleotide resolution that is viable both for genome-wide analysis and the analysis of individual genes.

Summary of the Invention

According to a first aspect of the invention there is provided a method for the identification of methylated cytosine bases in a nucleotide sequence, the method comprising the steps of:

(a) generating fragments of a DNA molecule comprising the nucleotide sequence such that the fragments comprise at least one single-stranded terminus comprising one or more bases, and wherein at least one base within the single-stranded terminus is an unmethylated cytosine or methylcytosine;

(b) incubating the fragments with bisulphite under non-denaturing conditions such that only unmethylated cytosines within the at least one single-stranded terminus are converted to uracil; and

(c) detecting fragments comprising methylcytosine or uracil bases in the at least one single-stranded terminus. The methylcytosine to be detected may be within a CG (CpG) dinucleotide or in any other nucleotide context, for example, CCWGG or CNG (CpNpG).

Fragments comprising at least one single-stranded terminus may be generated using one or more restriction endonucleases that cleave the DNA such that at least a one base single-stranded overhang is produced, the single-stranded overhang comprising at least one cytosine. Typically, the restriction endonuclease(s) recognises and cleaves methylated DNA and unmethylated DNA with substantially the same efficiency. The restriction endonuclease(s) may be selected from Λfspl, Taq\, ApeK\, Avή\,

BbvC\, Bcl\, BgIW, BsaWI, SseYI, Bsm\, BsoBI, SssSI, Btg\, Dpnll and H/ndlll. Typically the recognition sequence cleaved by the restriction endonuclease(s) includes the CG dinucleotide or one base thereof.

Fragments comprising at least one single-stranded terminus may be generated using two restriction endonucleases that cleave the DNA such that a fragment comprising at least a one base single- stranded overhang is produced. In one embodiment the restriction endonucleases are Mspl and Taq\, recognising the sequences C/CGG and T/CGA respectively.

Alternatively, fragments comprising at least one single-stranded terminus may be generated by digestion of linear fragments with an enzyme possessing 3' to 5' exonuclease activity in the absence of at least dGTP. For example, the enzyme may be a DNA polymerase such as T4 DNA polymerase.

The bisulfite may be sodium bisulfite, sodium metabisulfite or other suitable bisulfite salt. In one embodiment the non-denaturing conditions may comprise incubation at about 37⁰C for about 24 hours or less, typically for about 4 hours.

The detection step (c) may comprise:

(i) incubating bisuphite-treated fragments with DNA polymerase and dNTPs wherein either dATP or dGTP is labelled with a detectable marker; and

(ii) selectively detecting a desired subpopulation of fragments by detecting the detectable marker, wherein labelled dATP facilitates the selective detection of fragments comprising uracil in the single- stranded terminus, and labelled dGTP facilitates the selective detection of fragments comprising methylcytosine within the single-stranded terminus.

The detectable marker may be biotin, detectable using avidin or streptavidin. It will be appreciated that the detection step (c) may comprise, or be followed by, the selective isolation of a desired - A - subpopulation of fragments comprising either methylcytosine or uracil within the single-stranded terminus. Such isolation may be facilitated by, for example, the capture of fragments incorporating biotin labelled dATP or dGTP using avidin or streptavidin in solution or attached to a solid support. In an alternate embodiment, the detection step (c) may comprise the steps of:

(i) ligating to the single-stranded terminus of the bisulfite-treated fragments a detection facilitation linker complementary to bases of the single-stranded terminus and comprising at least one adenine or guanine at a position complementary to the at least one cytosine in the fragments of step (a); and (ii) analysing the ligated fragments by any suitable means such as PCR amplification, hybridisation and/or sequencing,

wherein a detection facilitation linker comprising at least one adenine at a position complementary to the at least one cytosine in the fragment of step (a) facilitates detection of uracil in the single-stranded terminus of said fragment and a detection facilitation linker comprising at least one guanine at a position complementary to the at least one cytosine in the fragment of step (a) facilitates detection of methylcytosine in the single-stranded terminus of said fragment.

Typically in accordance with the above embodiment the bisulfite-treated fragments are separated into two pools, wherein a detection facilitation linker comprising at least one adenine at a position complementary to the at least one cytosine in the fragment of step (a) is ligated to the bisulfite-treated fragments of one pool, and a detection facilitation linker comprising at least one guanine at a position complementary to the at least one cytosine in the fragment of step (a) is ligated to the bisulfite-treated fragments of the other pool.

Further in accordance with the above embodiment, in each pool a competitor linker may be incubated with the bisulfite-treated fragments and detection facilitation linker, wherein the competitor linker is specific for the non-selected fragment end.

The detection facilitation linker may comprise sequences from which amplification or sequencing reactions can be initiated using primers capable of hybridising to said sequences. In embodiments in which a competitor linker is employed, the competitor linker will typically differ sufficiently in sequence from the detection facilitation linker such that the primers used to initiate an amplification or sequencing reaction cannot hybridise to the competitor linker.

In an embodiment, where the fragments in step (a) of the present method are generated using the restriction endonucleases Msp\ and Taq\, the detection facilitation linkers comprise 5'-CG and 5'-CA to facilitate detection of 5'-methylCG and 5'-UG, respectively, within the single-stranded terminus of the treated fragment.

In accordance with the present method, the DNA molecule comprising the nucleotide sequence to be analysed is a genomic DNA sample. The genomic DNA sample may be pre-treated prior to step (a) by, for example, sonication or restriction endonuclease digestion to reduce fragment size.

According to a second aspect of the invention there is provided a method for the separation of methylated and unmethylated fractions within a DNA sample, the method comprising the steps of: (a) generating fragments of a DNA molecule such that the fragments comprise at least one single- stranded terminus comprising one or more bases and wherein at least one base within the single- stranded terminus is an unmethylated cytosine or methylcytosine;

(b) incubating the fragments with bisulfite under non-denaturing conditions such that only unmethylated cytosines within the at least one single-stranded terminus are converted to uracil;

(c) separating fragments into a methylated fraction comprising methylcytosine bases in the at least one single-stranded terminus, and an unmethylated fraction comprising uracil bases in the at least one single-stranded terminus; and

(d) optionally, isolating fragments within the fractions. The DNA sample may comprise a genome or a fragment or portion thereof.

The separation step (c) may comprise the steps of:

(i) incubating bisulfite-treated fragments with DNA polymerase and dNTPs, wherein either dATP or dGTP is labelled with a detectable marker; and

(ii) separating the methylated and unmethylated fractions by virtue of the detectable marker, wherein labelled dATP facilitates the selective isolation of fragments comprising uracil in the single- stranded terminus, and labelled dGTP facilitates the selective detection of fragments comprising methylcytosine within the single-stranded terminus. According to a third aspect of the invention there is provided a method for the diagnosis of, or determination of susceptibility to, a disease associated with aberrant cytosine methylation in a subject, the method comprising the steps of:

(a) providing a genomic sample from the individual comprising a nucleotide sequence to be analysed; (b) generating fragments of the genomic DNA such that the fragments comprise at least one single- stranded terminus comprising one or more bases and wherein at least one base within the single- stranded terminus is an unmethylated cytosine or methylcytosine;

(c) incubating the fragments with bisulfite under non-denaturing conditions such that only unmethylated cytosines within the at least one single-stranded terminus are converted to uracil; and

(d) detecting fragments comprising methylcytosine or uracil bases in the at least one single-stranded terminus,

wherein the extent and/or pattern of cytosine methylation detected is indicative of the disease, or is predicative of a susceptibility or predisposition thereto. Brief Description of the Drawings

The present invention is described herein, by way of non-limiting example only, with reference to the accompanying drawings.

Figure 1: Amplification of bisulfite-treated pUC19 fragments produced by Msp\ cleavage and ligated with CG or CA detection facilitation linkers. Numbers indicate temperature of incubation of fragments with bisulfite. U, unmethylated DNA; M, methylated DNA.

Figure 2: Amplification of bisulfite-treated human genomic DNA fragments produced by Msp\ cleavage and ligated with phosphorylated (p) or non-phosphorylated (OH) CG or CA detection facilitation linkers. U₁ unmethylated DNA; M, methylated DNA; m, molecular weight markers.

Figure 3: Amplification of bisulfite-treated fragments of human single copy genes HPRT and Xist produced by Msp\ cleavage and ligated with CG or CA detection facilitation linkers. Samples were obtained from male (m) and female (f) derived colorectal cancer cell lines. S, fully methylated DNA; H, DNA methylated with Hpall methylase; U, unmethylated DNA; and UH, unmethylated DNA cut with HpaW.

Figure 4: Microarray hybridisation of amplified bisulfite-treated HCT116 DNA cleaved with Msp\ and Taqrl and ligated with CG or CA detection facilitation linkers. Boxed regions indicate probes that selectively hybridise with the methylated DNA fraction (black), selectively hybridise with the unmethylated DNA fraction (grey), or hybridise with both fractions (white).

Figure 5: Amplification of bisulfite-treated DNA fragments produced by Msp\ and Taq\ cleavage and ligated with CG or CA detection facilitation linkers. DNA samples were (from right to left for each panel), fully methylated Sss I control (sonicated), fully unmethylated phi-29 control (sonicated), fully methylated Sss I control (Cvi Kl cleaved), fully unmethylated phi-29 control (Cvi Kl cleaved), human cell line HT29 (sonicated), human blood DNA (sonicated), human cell line SW480 (Cvi Kl cleaved), and human blood DNA (Cvi Kl cleaved).

Figure 6: Microarray hybridisation of amplified bisulfite-treated HT29 DNA fragmented and en-labelled with biotin. Boxed regions indicate probes that selectively hybridise with the methylated DNA fraction (black), selectively hybridise with the unmethylated DNA fraction (grey), or hybridise with both fractions (white).

Figure 7: Panels show plots of hybridisation signals across parts of the Nimblegen promoter tiling regions of four genes, SDPR, SDC2, CA1 and ALDH 1 A respectively. The signal for the methylated fractions is shown in the upper half of each panel and for the unmethylated fractions inverted in the lower half. Taq\ and Msp\ restriction enzyme sites are shown by the dashed vertical lines.

The nucleotide sequences of linkers and primers described in the Examples are provided in the formal Sequence Listing appearing at the end of the specification.

Detailed Description

The articles "a" and "an" are used herein to refer to one or to more than one {i.e., to at least one) of the grammatical object of the article. By way of example, "an element" means one element or more than one element.

Throughout this specification and the claims which follow, unless the context requires otherwise, the word "comprise", and variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integers or steps.

As used in the context of the present invention, the term "amplification" refers to making one or more copies of a targeted nucleotide sequence via a thermocyclic enzymatic reaction, and includes, but is not limited to the amplification of nucleic acid molecules by the polymerase chain reaction (PCR). As used herein the term "Polymerase Chain Reaction" (PCR) refers to a technique for producing a large number of copies of a specific segment of a nucleic acid sequence via oligonucleotide primer directed enzymatic amplification carried out in a thermocycling process. As used herein the term "linker" refers to an oligonucleotide designed to be ligated to one or both ends of a DNA fragment. The term "oligonucleotide" refers to a single-stranded sequence of deoxyribonucleotide or ribonucleotide bases, known analogues of natural nucleotides, or mixtures thereof. Oligonucleotides are typically short (for example less than 50 nucleotides in length) sequences which may be prepared by any suitable method, including, for example, direct chemical synthesis or cloning and restriction of appropriate sequences. Typically in the context of the present invention an oligonucleotide is designed to recognise and bind to a specific complementary nucleotide sequence located within another nucleic acid molecule. An oligonucleotide may also include bases in addition to those required to hybridise to the complementary sequence. The oligonucleotide may or may not be able to act a primer for polymerase-mediated extension. Not all bases in an oligonucleotide need be complementary to the sequence to which the oligonucleotide is designed to bind; the oligonucleotide need only contain sufficient complementary bases to enable the oligonucleotide to recognise and bind to that sequence. The oligonucleotide sequence may be an unmodified nucleotide sequence or may be chemically modified or conjugated by a variety of means known to those skilled in the art.

As used herein the term "bisulfite treatment" means incubation of a DNA fragment(s) with bisulfite under non-denaturing conditions such that such that the double-stranded portion of the fragment is not denatured or is not sufficiently denatured to enable reaction between bases within the double-stranded portion and the bisulfite. Thus the term "bisulfite treatment" is to be interpreted in the context of the present disclosure as "non-denaturing bisulfite treatment" or "bisulfite treatment under non-denaturing conditions".

Accordingly, with respect to bisulfite treatment the term "non-denaturing conditions" as used herein refers to conditions under which a fragment is incubated with bisulfite such that the double-stranded portion of the fragment is not denatured or is not sufficiently denatured to enable reaction between bases within the double-stranded portion and the bisulfite. Thereby, only bases within a single- stranded terminus of the fragment will be available to react with the bisulfite. Those skilled in the art will appreciate that the incubation conditions constituting "non-denaturing conditions" will typically be a combination of temperature and time (and optionally pH of the bisulfite solution) as described herein suitable to achieve the desired result. Those skilled in the art will also appreciate that the exact "non- denaturing conditions" may vary and may depend on factors such as the length and/or base constitution of the fragment and the bisulfite salt used. As used herein the terms "nucleic acid", "nucleotide sequence", "oligonucleotide" and "primer" designate any nucleic acid-based molecule, including DNA, cDNA, RNA, mRNA, cRNA , PNA or any combination thereof. Thus, a nucleic acid molecule, an oligonucleotide or a primer may comprise naturally occurring nucleotides, non-naturally nucleotides or a combination thereof.

As used herein the term "primer" refers to an oligonucleotide which binds to a specific region of a single stranded template nucleic acid molecule and initiates nucleic acid synthesis via an enzymatic reaction, extending from the 3' end of the primer and complementary to the sequence of the template molecule. By convention, primers are typically referred to as 'forward' and 'reverse' primers, one of which is complementary to a nucleic acid strand and the other of which is complementary to the complement of that strand.

As used herein the term "sample" refers to any biological sample that comprises nucleic acid molecules, typically comprising DNA and/or RNA. Samples may be tissues, cells or extracts thereof, or may be purified samples of nucleic acid molecules comprising, for example, an entire genome or a portion thereof. Use of the term "sample" does not necessarily imply the presence of target sequence within nucleic acid molecules present in the sample.

Described herein is novel methodology for the detection and analysis of the methylation status of cytosine residues in a nucleic acid sample. The methodology described also readily allows the physical separation of DNA fragments into those comprising methylated cytosines and those comprising unmethylated cytosines. Broadly, the methodology described herein involves generating small fragments of DNA so as to leave one or more terminal bases within a single-stranded region at one or both ends (termini). This is typically achieved using restriction enzymes that cleave the DNA to leave "sticky" ends, although alternative methods for generating fragments with single-stranded termini are contemplated. The DNA fragments are reacted with bisulfite under mild conditions in which only the terminal bases within the single-stranded region from any DNA fragment react with the bisulfite.

The inventors have termed this novel methodology "bisulfite tagging", as the ends of many small DNA fragments are effectively "tagged" as either "methyl" or "non-methyl" following a limited reaction with bisulfite under mild, non-denaturing conditions, while leaving the vast majority of DNA bases within the fragments unreacted and intact in their usual form A, G, T or C.

By virtue of the vast majority of cytosine bases within any DNA fragment remaining unreacted with bisulfite, a desired four-base character (A, G, T or C) for the fragment is maintained. This is particularly advantageous for analysis methods such as microarray hybridisation and high-throughout sequencing. Only one (or a few) cytosine bases within any single-stranded terminal location will be converted fully from cytosine (C) to uracil (U), while methylcytosine bases will not react. The sequence differences in these short single-stranded ends may then be used to selectively isolate the DNA populations containing methylcytosine or unmethylated cytosine (converted to uracil) in these terminal bases, thus allowing both the methylated and unmethylated fractions of, for example, a genome, to be separately and directly assayed.

Accordingly, one aspect of the invention provides a method for the identification of methylated cytosine bases in a nucleotide sequence, the method comprising the steps of:

(a) generating fragments of a DNA molecule comprising the nucleotide sequence such that the fragments comprise at least one single-stranded terminus comprising one or more bases and wherein at least one base within the single-stranded terminus is an unmethylated cytosine or methylcytosine;

(b) incubating the fragments with bisulfite under non-denaturing conditions such that only unmethylated cytosines within the at least one single-stranded terminus are converted to uracil; and (c) detecting fragments comprising methylcytosine or uracil bases in the at least one single-stranded terminus.

According to a further aspect there is provided a method for the separation of methylated and unmethylated fractions within a DNA sample, the method comprising the steps of:

(a) generating fragments of a DNA molecule such that the fragments comprise at least one single- stranded terminus comprising one or more bases and wherein at least one base within the single- stranded terminus is an unmethylated cytosine or methylcytosine;

(d) optionally, isolating fragments within the fractions. The methods described herein may be used for the detection and analysis of methylation that occurs in any sequence context and thus are amenable for use in any organism. Whilst of particular use in higher eukaryotes in which cytosines within the dinucleotide CpG (CG) are typically methylated, the invention is not so limited and may be employed, for example, in microorganisms where methylcytosine typically occurs within the sequence CCWGG (dcm methylation), and in plants for the detection of methylcytosine within the CpNpG (CNG) triplet, or in other cases where methylation is asymmetric.

In accordance with methods described herein, DNA fragments having short single-stranded regions at one or both termini may be generated by a variety of means. In one embodiment, one or more restriction endonucleases that cleave DNA so as to produce a single-stranded overhang or "sticky ends" of one or more bases, are used to generate the fragments. Suitable restriction endonucleases leave an unpaired cytosine base in the 5' and/or 3' single-stranded overhang or "sticky end". For detection of methylation at CpG sites that are commonly methylated in many higher eukaryotes, one of the "sticky ends" produced must contain the important dinucleotide CpG. Further, suitable restriction endonucleases typically cleave methylated and unmethylated DNA with similar or substantially similar efficiency. The use of restriction endonucleases to generate fragments having single-stranded regions at one or both termini offers some advantages for subsequent analysis following bisulfite treatment. For example, the fact that the restriction endonuclease cleavage sites are known allows for efficient microarray design by limiting the sequences of probes to correspond to those regions of the fragments adjacent the cleavage sites at the fragment ends to be analysed.

Examples of restriction endonucleases that may be used to interrogate cytosine methylation in different sequence contexts in accordance with the methods described herein are provided in Table 1.

Table 1. Exemplary restriction endonucleases

In particular embodiments, two restriction endonucleases are employed. For example Msp\ (recognising and cleaving the sequence C/CGG) and Taq\ (recognising and cleaving the sequence T/CGA) are particularly useful for the detection and analysis of methylation in higher eukaryotes such as mammals as each enzyme produces two-base single-stranded or sticky ends of sequence 5'-CpG or 5'-meCpG. Similarly, both Msp\ and Taq\ cleave methylated and unmethylated sites with equal efficiency.

Other methods of generating fragments with short single-stranded termini can be used. For example, nucleases or DNA polymerases (such as T4 DNA polymerase) possessing 3' to 5' exonuclease activity may be employed in the absence of dGTP and one or more of dATP, dCTP and dTTP. Those skilled in the art will recognise that any means of producing fragments with short single-stranded termini comprising a cytosine (or methylcytosine) base may be used in accordance with the present methods. Fragments so produced are incubated with bisulfite under mild non-denaturing conditions such that only those cytosine bases located in the single-stranded termini are accessible to react with the bisulfite. As defined herein, the exact incubation conditions to ensure "non-denaturing" conditions will vary with application and will typically involve a combination of any suitable time and temperature such that the double-stranded region of a fragment is not denatured or substantially not denatured. Typically, incubation temperature will be less than about 55⁰C. A suitable temperature may be between about O⁰C and 55⁰C, between about 4⁰C and 5O⁰C₁ between about 1O⁰C and 45⁰C, between about 15⁰C and 4O⁰C or between about 2O⁰C and about 37⁰C. A suitable incubation time may be about 24 hours or less, for example about 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 10 hours, 15 hours, 20 hours or 24 hours. In a particular embodiment as exemplified herein, fragments may be incubated with 2 M sodium bisulfite at 37°C for 4 hours, while omitting any prior denaturation of the nearby double-stranded DNA by heat and/or high pH. In the case of prior cleavage of DNA with Msp\ and Taq\ in the embodiment described above, upon reaction with the bisulfite the two new 5' termini produced are 5'-UpG (where the cytosine was unmethylated) or 5'-meCpG. The detection and/or selective separation of fragments into fractions in which the bases within the single-stranded termini of the fragments comprise methylcytosine or uracil (representing unmethylated cytosine prior to bisulfite treatment) can be achieved via a variety of methods. In one embodiment bisulfite-treated fragments may be end-filled in separate reactions with DNA polymerase and biotin labelled dGTP (for incorporation opposite methylcytosine) or biotin labelled dATP (for incorporation opposite uracil, corresponding to unmethylated cytosine in the unreacted DNA). Avidin or streptavidin can then be used to bind to biotin and thereby selectively isolate either the methylated or unmethylated fraction of the fragment population. Those skilled in the art will appreciate that the use of biotin-avidin/streptavidin represents only one means of achieving such selective separation or isolation and that other combinations of detectable marker and binding partner or detection system may also be employed. For example, dNTPs may be fluorescently labelled.

In a particular embodiment, detection is facilitated by the sequence-specific ligation of two different synthetic linkers, one specific for methylated DNA (methylcytosine) and the other specific for unmethylated DNA (uracil). Individual populations of specifically-ligated molecules may then be analysed by, for example, selective PCR amplification using unique primers sequences directed to each linker. Again, considering the embodiment in which Msp\ and Taq\ are used to generate the fragments subsequently treated with bisulfite, the 5'-UpG and 5'-meCpG termini may be readily distinguished by ligation of detection facilitation linker oligonucleotides to the termini of bisulfite-treated fragments. A detection facilitation linker comprising 5'-CpA will be specific for the 5'-UpG termini, whereas a detection facilitation linker comprising 5'-CpG will be specific for the 5'-meCpG termini. It will be understood that 5'-CpA and 5'-CpG comprising linkers are exemplary only. Detection facilitation linkers can be designed specifically depending on the nature of the single-stranded termini to be distinguished.

Accordingly, detection facilitation linkers may be employed by (i) ligating to the single-stranded terminus of the bisulfite-treated fragments a detection facilitation linker complementary to bases of the single-stranded terminus and comprising at least one adenine or guanine at a position complementary the at least one cytosine in the fragments of step (a); and

(ii) analysing the ligated fragments by any suitable means such as PCR amplification, hybridisation and/or sequencing,

Typically the bisulfite-treated fragments are separated into two pools, wherein a detection facilitation linker comprising at least one adenine at a position complementary to the at least one cytosine in the fragment of step (a) is ligated to the bisulfite-treated fragments of one pool, and a detection facilitation linker comprising at least one guanine at a position complementary to the at least one cytosine in the fragment of step (a) is ligated to the bisulfite-treated fragments of the other pool. Depending on the application it may be desirable to increase the specificity of ligation of detection facilitation linkers to their corresponding single-stranded fragment termini. This may be achieved by the addition of competitor linkers with the bisulfite-treated fragments and detection facilitation linkers, wherein a competitor linker is designed to be specific for the fragment end that is not to be analysed (non-selected end). Further, for the purposes of subsequent analysis, only detection facilitation linkers, and not competitor linkers, possess the appropriate sequence to allow subsequent PCR amplification.

The methods described herein are suitable for whole genome analysis of methylation. For whole genome analysis bisulfite-treated fragments may be denatured and extended randomly using a suitable DNA polymerase so as to generate a smaller distribution of fragment sizes (for example 50- 300 bp) amenable to whole genome analysis. Further, DNA polymerase-extended fragments may be optionally treated with a "proofreading" restriction enzyme, for example PmH for 5'-CG ligation products, or BamH\ for 5'-CA ligation products, to remove a low level of A-C or G-U mismatches that may have been incorporated during ligation, such as where 5'-CA may have ligated incorrectly to 5'- meCG ends, or 5'-CG may have ligated incorrectly to 5'-UG ends. Thus, if the sequence of a detection facilitation linker is 5'-CA-GATCC, then any ligation product using that linker will be cleaved by BamHI if the copied sequence is incorrectly CG-GATCC instead of CA-GATCC.

Analysis of bisulfite-treated fragments isolated and/or separated in accordance with methods described herein may be achieved by one or more of a variety of methods well known to those skilled in the art, for example PCR amplification, microarray hybridisation and DNA sequencing, including using high-throughput sequencing techniques. Such techniques and methods are well within the skills and capabilities of those skilled in the art and require no more than mere routine experimentation (see for example Sambrook, J. and Russell, D.W. Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, 3rd Edition, 2001; the disclosure of which is incorporated herein by reference in their entirety).

Those skilled in the art will appreciate that the specific sequences of suitable oligonucleotides, such as detection facilitation and competitor linker oligonucleotides and primers, are typically optimised depending on the DNA sequences to be analysed and the particular application in which the method will be used. The design of suitable oligonucleotide and primer sequences is well within the capabilities of those skilled in the art and require no more than mere routine experimentation. Methods described herein enable methylation analysis to be conducted across an entire genome with high resolution. For example, where fragments to be bisulfite treated are generated using restriction endonucleases, DNA methylation may be determined individually at the CpG sequence of any individual cutting site for the restriction endonucleases used, rather than being averaged across a broad region of DNA related to the mean fragment size. Further the methods described herein are advantageously relatively insensitive to variations caused by sequence specificity of capture or density of methylation sites. In comparison to prior art methods using methylation sensitive restriction enzymes (either those that preferentially cleave unmethylated DNA or preferentially cleave methylated DNA₁ such as McrBC), the ability to measure both the fraction of methylcytosine and cytosine bases at each individual site provides higher resolution, whilst the ability to choose the restriction endonuclease(s) used to cleave the fragments enables regions of sequence to be analysed that are otherwise not able to be targeted due to enzyme recognition site constraints.

Methods described herein find application in any circumstance where it is desirable to investigate patterns and/or frequency of methylation in any given DNA sequence. Thus, the methods find application, inter alia, in the research environment, in clinical settings and in diagnostics. Referring to clinical and diagnostic applications, specific patterns and/or aberrations in frequency of DNA methylation are increasingly observed to be related to, or characteristic of, a range of diseases in humans for example a variety of cancers and malignancies, atherosclerosis and developmental disorders. Those skilled in the art will recognise that the scope of the present invention is not to be limited by reference to any specific clinical application or in relation to any specific disease or disorder. Any disease or condition found to be associated with specific patterns and/or aberrations in frequency of DNA methylation are amenable to analysis using the methods disclosed herein. Specific methylation patterns may also be characteristic of specific organisms, and thus the methods described herein also find application in the detection and identification of, for example, organisms, such as pathogenic organisms. In one embodiment there is provided a method for the diagnosis of, or determination of susceptibility to, a disease associated with aberrant cytosine methylation in a subject, the method comprising the steps of:

(a) providing a genomic sample from the individual comprising a nucleotide sequence to be analysed;

(b) generating fragments of the genomic DNA such that the fragments comprise at least one single- stranded terminus comprising one or more bases and wherein at least one base within the single- stranded terminus is an unmethylated cytosine or methylcytosine;

wherein the extent and/or pattern of cytosine methylation detected is indicative of the disease, or is predicative of a susceptibility or predisposition thereto.

All essential components required for detecting and analysing methylation in accordance with methods of the present invention may be assembled together in a kit. The kits may optionally include appropriate components for DNA manipulation such as restriction endonuclease cleavage, exonuclease digestion, end-filling, ligation, bisulfite treatment, sequencing, PCR amplification and/or hybridisation, including for example suitable reagents, primers, enzymes, cofactors, buffers, positive and negative controls and the like. Typically, the kits comprise instructions for performing the methods of the present invention.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

The present invention will now be described with reference to the following specific examples, which should not be construed as in any way limiting the scope of the invention. Examples

General methods

Random Fragmentation of DNA

For physical fragmentation, 1 to 10 μg of genomic DNA was treated with a Bioruptor UCD-200 sonicator (Diagenode, Belgium) in 300 μL of 10 mM Tris, 0.1 mM EDTA, pH 7.5 at a power setting of "high" for 60 minutes on ice, with alternating cycles of 30 seconds "on" or "off¹. A small part of each sample was then applied to a 2% agarose gel, to confirm that a mean size range of 200 to 300 bp had been reached. Each fully sonicated sample of genomic DNA, in a volume of 300 μl, was next concentrated by ethanol precipitation (using blue glycogen as carrier) to a volume of 10 μl in 1 mM Tris, 0.1 mM EDTA, pH 8. All of its irregularly broken molecular ends were then repaired to blunt, in a volume of just 15 μl using a standard kit (DNA End-Repair Kit, Epicentre #ER81050). The end-repair reaction involving Klenow polymerase, T4 DNA polymerase and T4 polynucleotide kinase was stopped by heating at 80 C for 20 minutes.

Enzymatic fragmentation, was achieved by treating 1 μg of genomic DNA with the restriction enzyme Cvi Kl (RG/CY) under "relaxed" conditions in restriction enzyme buffer containing 20% DMSO. Partial digestion for two hours at 37°C in a volume of 15 μl was sufficient to reduce the mean size of a typical genomic DNA to 200 or 300 bp, as confirmed by application to a 2% agarose gel. The enzymatic reaction was stopped by heating at 80°C for 20 minutes.

Linker Ligation to sheared DNA

Each blunt-ended sample (whether produced by sonication or by Cvi Kl) was treated with Taq DNA polymerase plus all four dNTPs at 200 μM each, in a total volume of 20 μL for 20 minutes at 70⁰C. Simple cooling of the sample to room temperature stops most Taq activity, before proceeding to the next step.

Each fragmented genomic sample, with all of its DNA ends having been converted from blunt to 3'-A, 1-2 μg of DNA was then ligated to a 20-fold molar excess of synthetic P2 linker (see Table 2) having complementary 3'-T ends, in a volume of 50 μL using a standard kit (Quick Ligation Kit from New England Biolabs, #M2200) in accordance with manufacturer's instructions. The PEG-assisted ligation was achieved in 15 minutes at 20⁰C. lmmediately after ligation, each sample was diluted from 50 to 200 μl_ with water, and purified on a standard silica column (Promega, Wizard) to remove excess DNA linkers or proteins, and subsequently eluted with 25 μL of 1 mM Tris, 0.1 mM EDTA, pH 8. Restriction enzyme digestion

DNA was cut to completion under standard conditions using the restriction enzymes Msp\ (C/CGG) and/or Taq\ (T/CGA) in a reaction volume of 20 to 30 μL. Reactions were stopped by addition of 2 μL of 50OmM EDTA. Bisulfite treatment of DNA

The cut DNA (1 μg) was reacted with sodium bisulfite, using the MethylEasy kit from Human Genetic Signatures (North Ryde, Australia), but omitting any prior heating or alkali denaturation step. Briefly: (a) 25-30 uL of restriction digested DNA was combined in a 2 mL tube with 250 μl of sodium bisulfite, then incubated at 37°C for 4 hours while omitting mineral oil.

(b) 20 μg of blue-dye-labelled glycogen was added and DNA precipitated with 750 μL of

MethylEasy solution 4 and 1 ml of isopropanol. Following centrifugation and washing of the pellet with 70% ethanol DNA was dissolved in 12 μL of MethylEasy Solution 3. Alternatively excess bisulfite was removed using a small silica column, and sample eluted from the column in 16 μL of MethylEasy Solution 3.

(c) Desulfonation was achieved by heating in MethylEasy solution 3 for 60 minutes at 72° C, while spinning down several times to reduce evaporation.

Linker ligation to bisulfite-treated DNA

Following bisulfite treatment under non-denaturing conditions, the 5'-methyl-CG or 5'-UG ends so produced were detected accurately at a single site level, by their specific ligation to complementary DNA linkers ending in two protruding bases as 5'-CG or 5'-CA respectively ("detection facilitation linkers"), for the two genomic ends 5'-methyl-CG or 5'-UG respectively. The reacted DNA was split into two pools (aliquots) for separate ligations. In each ligation a competitor linker was included that matched the non-desired end of the fragment. Sequences of linkers are shown in Table 2 (below). Double stranded linkers were prepared from the respective single strands by adding 1 μL of 100 μM stock solution of each (or 2 μL for competitor linkers) to 100 μL of 1 mM Tris, 0.1 mM EDTA, pH 8. The mixes were heated briefly to 65°C, then cooled to room temperature. Table 2. Sequences of detection facilitation and competitor linkers

Examples 1 to 4 (below) used the CG-1/CG-2 or CA-1/CA-2 detection facilitation linkers.

For the methylated DNA fraction the ligation reaction included both detection facilitation and competitor linkers and comprised:

6 μl_ bisulfite reacted DNA (-500 ng) with 20 μl_ of the following master mix:

13 μL 10XT4 Ligase buffer (New England Biolabs)

20 μL of CG1 -detection facilitation linker

20 μL of CG2-detection facilitation linker

40 μL of CA-competitor facilitation linker

4 μL T4 DNA ligase

The reaction was incubated at 20⁰C for 2 hours, followed by 4°C overnight. For the unmethylated DNA fraction successive ligations were performed as follows.

To 6 μl_ bisulfite reacted DNA (-500 ng) was added 10 μL of the following mix:

13 μL 10X T4 Ligase buffer (New England Biolabs)

40 μL of CG- -competitor linker

40 μ L water

4 μL T4 DNA ligase.

The reaction was incubated at 20⁰C for 2 hours, followed by 4°C overnight. To the first ligation mix, then add 10 μL of the following mix:

13 μL 10X T4 Ligase buffer (New England Biolabs)

20 μL of CA1 -detection facilitation linker

20 μL of CA2-detection facilitation linker

40 μ L water

4 μL T4 DNA ligase.

The reaction was incubated at 2O⁰C for 2 hours, followed by 4°C overnight.

For Examples 5 and 6 below, P1-CG or P1-CA linkers were used.

For the methylated DNA, 30 μl of the following master mix was added to 8 μL of bisulfite-con verted genomic DNA.

50 μl of 10X ligase buffer

100 μl of P1-CG detection facilitation linker (800 ng)

50 μl of CA-competitor linker (1000 ng)

300 μl of water

10 μl of T4 DNA ligase (New England Biolabs).

The combined volume of 38 μl was incubated at 20⁰C for two hours, then 4°C overnight.

For the unmethylated DNA, 12 μL of the following master mix was added to 8 μL of bisulfite-con verted genomic DNA

25 μl of 10X ligase buffer

50 μl of CG-competitor linker (1000 ng)

125 μ I of water

8 μl of T4 DNA ligase (New England Biolabs).

The combined volume of 20 μL was incubated at 2O⁰C for two hours, then 4°C overnight. 15 μL of the following master mix containing the detection facilitation linker was then added and the combined volume of 35 μL incubated at 20⁰C for two hours, then 4°C overnight. 25 μL of 1OX ligase buffer

100 μ L of P1-CA detection facilitation linker (800 ng)

125 μ L of water

8 μL of T4 DNA ligase (New England Biolabs).

As an optional step the completed CG ligations (but not CA) were cut with restriction enzyme BsftJI (CGCG) in a total volume of 30 μL at 60°C for one hour (spin down several times to reduce evaporation). Such CG ligations should be stored at -70⁰C if left for several months, because an excess of SsflJI will slowly cut CG-mCG ligation junctions even at -20⁰C. This optional digestion step serves to (i) remove any unconverted C bases from the CG ligations, by means of a CGCG sequence that has been placed on one end of both CG 1 and CG2 detection facilitation linkers (if the genomic DNA which ligates is methylated, then BsftJI will not cut); and (ii) remove a large excess of linker dimers between CG1 and CG2 detection facilitation linkers, which would otherwise interfere with PCR amplification.

Random priming on template DNA

Bisulfite-treated and ligated samples were heat-denatured briefly at 95-100⁰C into single strands, and chilled on ice. The denatured genomic DNA (of size 80 to 2000 bp) was then extended randomly using a T7-N6 primer and Klenow DNA polymerase, to generate a significantly smaller distribution of DNA sizes (typically 50 to 300 bp), suitable for whole-genome analysis.

To each cold, denatured sample, was added a master mix of equal volume (10 μl) containing Klenow DNA polymerase, all four dNTPs at 200 μM, and a T7-N6 primer (at PCR concentrations) that can use its 3'-N6 to generate new single strands, anywhere along a heat-denatured DNA molecule. The primer extension reaction was carried out for 10 minutes at 4⁰C, then for another 10 minutes at 20 C, before being stopped by direct transfer to 75⁰C, and further incubation at 75⁰C for 20 minutes.

PCR amplification

PCR amplification was carried out using Promega Hotstart Taq Polymerase in the buffer provided by the manufacturer. The appropriate primers (see Table 3) were used at a concentration of 300 nM. Table 3. Sequences of the PCR primers

Table 4. Sequences of the PCR primers for amplification of human HPRT and Xist genes

Denaturation was for 3 min at 95°C followed by a variable number of cycles (as indicated in the examples), typically 20 to 30 cycles of 95°C for 15 seconds, 55°C for 30 seconds, 72°C for 1 minute, and finally 72⁰C for 2 minutes.

Microarray Analysis

For preparation of the PCR product for hybridisation with an Affymetrix microarray trial amplifications were performed using different ratios of dUTP:dTTP (2:1, 3:1 and 4:1) followed by digestion with USER enzyme mix to determine the best ratio to give fragmented DNA in the size range 50-100 bp. Trial amplifications were also performed to determine the optimal number of amplification cycles.

Preparative amplification of both the methylated and unmethylated DNA fractions was done in 8 X 50 μl_ reactions. Preparative PCR products were purified according to Affymetrix protocols, digested with DNasel or UDG/APE to the desired 50-100 base length and labeled with biotin dNTP and terminal transferase according to Affymetrix protocols.

Example 1■ Non-denaturing conditions for bisulfite treatment

Conditions for bisulfite treatment of 5' single-strand overhangs produced by digestion with Msp\ were evaluated using pUC19 plasmid DNA₁ either unmethylated or after methylation in vitro with Hpall DNA methyltransferase, which methylated the internal cytosines of the Msp\ CCGG recognition site. Digested DNAs were incubated with sodium bisulfite for 4 hours at a range of temperatures from 4°C to 55°C. Following the bisulfite reaction samples were ligated with either CG 1 detection linkers (or CG1 plus CG2) or CA1 detection linkers in the absence of competitor linkers. Ligation products were subsequently amplified for 20 rounds of PCR.

The results in Figure 1 show that bisulfite treatment at 37°C allows efficient ligation of the CG linker the methylated DNA digest, but not the unmethylated digest. This indicates that the large majority of cytosine bases in the overhangs were converted to uracil. Similarly the CA linker is able to efficiently ligate with the unmethylated DNA from the 37°C bisulfite treatment but not to any of the DNAs derived from methylated DNA.

Example 2 - Methylation analysis in human genomic DNA

Amplification of fully methylated human genomic DNA (prepared by in vitro methylation with M.Sss I methylase) was compared with that of essentially unmethylated human genomic DNA, prepared by whole genome amplification using phi29 DNA polymerase. DNAs were cut with Msp\ and bisulfite treated for 4 hours at 37°C as described above. Following bisulfite treatment CG1 or CA1 linkers in which the 5' end of the linker overhang was either phosphorylated or non-phosphorylated were separately ligated with the treated DNA (in the absence of competitor linkers). To generate populations of molecules in which one end was derived from Msp\ digestion and the other end was random, and to generate a library of shorter DNA sequences flanking the restriction enzyme sites, DNAs were denatured and used as templates for random priming using a T7-N6 primer as described in methods. The T7-N6 primer was extended using Klenow DNA polymerase on ice and the DNA samples amplified using a T7 primer in combination with either the CG or CA primer Such randomly- primed samples where then amplified by PCR for 30 cycles using a T7 promoter sequence as one PCR primer, and a linker sequence (CG or CA) as the other primer. As a control, the T7 primer alone gave no PCR products except at much higher cycle numbers (40).

Figure 2 shows PCR products as obtained from the T7-N6 random priming reaction, for either fully methylated or fully unmethylated genomic DNA controls. On the left, it can be seen that a fully methylated control "M" is well amplified by a T7 primer plus its own CG linker primer, whereas a fully unmethylated control "U" is not well amplified. On the right, it can be seen that a fully unmethylated control "U" is well amplified by a T7 primer plus its own CA linker primer, whereas a fully methylated control "M" is not well amplified. The overall size of PCR-amplified random products is 100 to 400 bp, with a mean size of 200 to 300 bp. For amplification of methylated DNA using the CG primer, efficient amplification was seen for both phosphorylated and non-phosphorylated linkers. However, a higher background of amplification from unmethylated DNA was evident when the non-phosphorylated linker was used. All subsequent ligations therefore used phosphorylated linkers.

For amplification of unmethylated DNA using the CA primers, efficient amplification was seen for both phosphorylated and non-phosphorylated linkers. Some background amplification of methylated DNA was evident for both linkers, in the absence of competitor linkers that were subsequently used. The background amplification results from mis-ligation of the CA linker to meCG overhangs.

This example demonstrates that bisulfite treatment and selective ligation at single restriction sites can allow selective amplification of methylated or unmethylated DNA sequences from essentially pure populations of methylated or unmethylated DNA.

Example 3 - Methylation analysis in DNA from colorectal cancer cell lines Genomic DNAs from different colorectal cancer cell lines were digested with Msp\, and treated at 37°C for 4 hr as described above with sodium bisulfite. CG or CA linkers were ligated to separate aliquots of treated DNA and amplified libraries prepared using CG or CA primers. Single copy genes of known DNA methylation status were analysed to determine if they had been enriched in the methylated or unmethylated fraction. The HPRT gene is located on the X-chromosome and is active and unmethylated in male cells that carry only a single X chromosome. In female cells with two X chromosomes one copy is active and unmethylated, while the other is inactive and methylated. For amplification of the HPRT and Xist genes, the primers used are shown in Table 4 (see above).

Figure 3 shows the amplification of the expected band (arrow) from the methylated DNA fraction of the female cell line (G; HT29), but a lack of amplification from male cell lines (B, D, E; HCT116, CaCo2 and SW480 respectively). In contrast to the HPRT gene, the Xist gene is methylated on the active X chromosome and unmethylated on the inactive X chromosome present in female cells. As expected, amplification from the methylated fraction is seen for both male and female cells, but only from the unmethylated fraction from female cells.

Example 4 - Microarray analysis of DNA methylation from a colorectal cancer cell line

DNA from the colorectal cancer cell line HCT116 was digested with Msp\ and Taq\, treated with sodium bisulfite and CG1 and CG2 and CA1 and CA2 linkers ligated to separate fractions of the treated DNA as described above. Linkered DNAs were amplified and prepared for microarray hybridisation as described above. In this example amplification of fragments relies on both ends of the restriction fragment having the same methylation state.

Figure 4 shows the results of hybridisation of the amplified methylated and unmethylated DNA fractions with an Affymetrix promoter tiling microarray. Many spots show clear differences in the level of hybridisation of the separated fractions. In the segment of the array shown, a number of specific regions are boxed to show examples of probes showing strong selective hybridisation with either the methylated or the unmethylated fraction of the genome.

Example 5 - Random fragmentation of genomic DNA prior to methylation analysis

In order to allow interrogation of the methylation state of individual sites libraries of bisulfite-tagged DNA were prepared in which one end corresponded to a Taq\ or Msp\ site and the other was random. Initially DNA was first randomly fragmented by sonication or by digestion with CviKI at RG/CY sequences and more randomly under "star" conditions. The DNA was the end-repaired and ligated with the P2 linker. The P2-linkered DNA was then digested with Msp\ and Taq\ and bisulfite treated for 4 hours at 37°C. The treated DNA was separated into two fractions and ligated with either the P1-CG or P1-CA linker in the presence of the respective competitor linkers as described above. Using this protocol DNA fragments were obtained that contain one random end, generated by shearing, and one end corresponding to an Msp\ or Taq\ restriction site. The methylated DNA fractions were amplified using primers P1-CG and P2, and the unmethylated fractions amplified using primers P1-CAI and P2.

Figure 5 illustrates the large-scale PCR amplification of eight different DNA samples each of which had been ligated with four different "detection linkers" beforehand, as shown in four different parts of the gel. The eight DNA samples were: (1) fully methylated Sss I control (sonication), (2) fully unmethylated phi-29 control (sonication), (3) fully methylated Sss I control (Cvi Kl), (4) fully unmethylated phi-29 control (Cvi Kl), (5) human cell line HT29 (sonication), (6) human blood DNA (sonication), (7) human cell line SW480 (Cvi Kl) or (8) human blood DNA (Cvi Kl).

The four "detection linkers" were: (A) upper left, P1-CG from Table 2, (B) upper right a mutant form of P1-CG with a 3'-terminal A-C base pair, (C) lower left, P1-CA from Table 2, or (D) lower right, a mutant form of P1-CA with a 3'-terminal G-T base pair. Two mutant linkers were tested in order to vary how many cycle numbers would be required for suitable amplification.

All samples from human cell lines HT29, SW480 or human blood DNA were amplified efficiently at cycle numbers of 23 for linker A, 26 for linker B, 28 for linker C, or 29 for linker D. Fully methylated controls (M) were amplified efficiently for both P1-CG linkers A or B₁ while fully unmethylated controls (U) were not amplified efficiently. Similarly, fully methylated controls (M) were not amplified efficiently for P1-CA linkers C or D₁ while fully unmethylated controls (U) amplified well.

These results demonstrate that the methods described herein are highly selective and efficient for producing both "methyl" and "no methyl" fractions of the human genome, in suitably large quantities for study on microarrays or by high-throughput sequencing. The linkers used in this example are based on those used in the Applied Biosystems SOLiD sequencing system, enabling the amplified libraries to be studied either by microarray hybridisation or by high throughput sequencing.

Example 6 - Microarray analysis of DNA methylation from HT29 cell line Amplified methylated and unmethylated DNA fractions from the HT29 cell line were fragmented, end- labeled with biotin and hybridised with Affymetrix promoter tiling arrays. In the section of the microarray shown in Figure 6 there are many differences in hybridisation evident between the two fractions. A number of corresponding probe regions are boxed to show examples of probes hybridising with clear selectivity with either the methylated or unmethylated DNA fraction. This example demonstrates that bisulfite modification and selective ligation at a single site is sufficient to provide for clear differential amplification based on the methylation state of the cytosine in the 5' overhang generated by restriction digestion.

In this example and Example 4 commercially available promoter tiling arrays were used. However those skilled in the art will appreciate that embodiments of the invention may advantageously be used with custom designed microarrays in which the probe sequences are designed to correspond to those regions of the fragments adjacent the cleavage sites at the fragment ends to be analysed.

Example 7 - Methylation analysis using two-color hybridisation of methylated and

unmethylated DNA fractions

The ligation of linkers to the ends of sonicated DNA can be a relatively inefficient step. In a further optimised protocol random ends were generated by reverse priming using a P2-N6 primer. Methylated and unmethylated DNA fractions were prepared as described below and separately labelled with different fluors prior to co-hybridisation to Nimblegen promoter tiling arrays. Steps in the protocol were as follows:

1. DNAs from blood and the colorectal cancer cell lines HCT116, SW480 and LIM1215 were sheared by sonication as described in preceding examples, precipitated and resuspended in water. The volume was adjusted to 20 μl_ in Antarctic Phosphatase reaction buffer (New England Biolabs) and DNA de-phosporylated using 5 units of Antarctic phosphatase at 37°C for 60 min. The enzyme was then inactivated by heating to 65°C for 10 min.

2. Sequential restriction enzyme digestion: 3 μl_ of 1OX New England Biolabs Buffer 4 and 2 μl_, 20 units, of Msp\ enzyme were added and incubation continued at 37°C for 2 hr. 20 units, 2 μl_, of

Taq\ and BSA to 100 μg/mL were added to give a volume of 28 μl_ and incubation continued at 65°C for a further 2 hr. Incubation was stopped by addition of 2 μL of 500 mM EDTA.

3. Bisulfite treatment of DNA was performed as described in preceding examples, for 4.5 hr at 37°C, and after precipitation, resuspended in 20 μL of MethylEasy Buffer 3 and incubated for 90 min at 72°C.

4. Ligation of CG and CA linkers. Separate ligations were set up with P1-CG or P1-CA linkers as described below. Both P1-CG and P1-CA linkers (Table 2) were modified by incorporation of a biotin at the 3' end of the upper strand to allow subsequent capture. Mixes of combined linkers and competitors were prepared as in Table 5 below. These amounts, and ratios, are for IOOOng of DNA of average length 200 bp.

Table 5

Ligation reactions with 1 μg of bisulfite treated DNA were set up in 50 μL with 0.6 μL (12 units) of T4 DNA ligase and incubated at room temperature for 5 hr. After ligation, DNA was purified through a Promega Wizard column and eluted with 2 x 50 μL of 10 mM Tris.HCI, 0.1 mM EDTA, pH8. This step removes the unligated linkers.

5. Bead capture and release of single-strand: 25 μL of Dynal M280 streptavidin beads were first washed twice with 500 μl of 2X salt wash buffer (2M NaCI, 1OmM Tris pH7.5, 1mM EDTA, 0.2% Tween 20) and resuspended in 100ul of 2x wash buffer. The 100μL of eluted DNA was added and mixed on a shaker at room temperature with the beads for 15min at 500rpm. Beads were magnetically captured for 1 min and the supernatant removed. This was followed with two washes in 500 μL of 1X wash buffer, once with 500 μL of 1 x SSC and then with 500 μL of 0.5 x SSC at 65⁰C for 3 min with shaking. The beads were then resuspended in 50 μL of 10 mM Tris.HCI, 0.1 mM EDTA, pH8 and transferred to a new tube. After incubation at 98°C for 5 min and chilling for 1 min on ice, the beads were magnetically captured for one minute on ice and the supernatant (50 μL) transferred to a fresh tube. After drying, the DNA was resuspended in 10 μL of water.

6. Random priming

The P2 primer sequence was incorporated by random priming using a P2-N6 primer (5'- CTGCCCCGGGTTCCTCATTCTCTNNNNNN-3' (SEQ ID NO:32)). The 10 μL of eluted single-stranded DNA was mixed with 500 ng (5 μL) of P2-N6 primer, heated at 98°C for 2 min and chilled on ice. To this mix was added, 2.5ul 1Ox New England Biolabs buffer 2, 0.625ul 1OmM dNTPs, 0.625ul Klenow fragment of DNA polymerase (5000u/ul) and the volume adjusted to 25 μL. After incubation on ice for 10 min, then at room temperature for 60 min, the reaction was terminated by heating at 75⁰C for 20 min. 7. Amplification

After trial reactions using 1 μL of random primed DNA, large to estimate amplification cycles, 300 μL amplification reactions for each of the methylated and unmethylated fractions were set up as follows:

10 μL random primed DNA sample

150 μL Promega GoTaq 2x Hot start mastermix

18 μL P1-CA or P1-CG primer (Table 2, 5uM stock)

18 μL P2-long SOLiD primer (δ'-CTGCCCCGGGTTCCTCATTCT-S' (SEQ ID NO:33); 5uM stock)

7.5 μL Sybr Green (1 :3000 dilution of 1000Ox concentration)

96.5 μL water Samples were run on a Roche LightCycler 480, SYBR green programme: 95°C 2min x1 cycle, and 95°C 15sec, 60⁰C 1min x 40 cycles. Fluorescence was monitored and reactions stopped at the top of the exponential phase. DNA was purified using Promega Wizard columns, eluted in 30 μL water and quantified by spectrophotometry. 8. Labelling and hybridisation

20 μg samples of each fraction were fluorescently labelled using the Nimblegen Dual-Color DNA labelling kit but with the Cy3 and Cy5 random nonamers being substituted with Cy3 and Cy5 random hexamers (Geneworks, Adelaide, Australia). Methylated DNA fractions were labelled with Cy3 and unmethylated with Cy5 dyes. Arrays were hybridised with Nimblegen 720K promoter tiling arrays following Nimblegen protocols and scanned on an Axon scanner.

Figure 7 shows examples of hybridisation signals of some genome regions adjacent to specific restriction enzyme sites. The lower two panels show the effect of distance from the restriction enzyme cut sites in that signal intensity drops rapidly with distance. The different relative levels of methylation in cell lines and blood are evident in the examples shown. For three of the genes SDC2, ALDH 1 and the level of methylation in blood DNA (methyl fraction G, unmethylated H) is low, while significant methylation is seen for CA1. Among the cell lines different patterns of methylation are seen for different genes - CA1 is unmethylated in SW480 (C - methyl; D - unmethylated), but unmethylated in HCT116 and LIM1215; ALDH1A is methylated only in LIM1215 but largely unmethylated in HCT116 and SWS480; SDPR and SDC2 show significant methylation in all three cell lines, but are unmethylated in blood.

Example 8 - Separation of methylated and unmethylated DNA Following reaction with bisulfite the sequence difference of the ends, 5'-meCG and 5'-UG from Taq\ or Msp\ digestion, may also be used to provide physical separation of methylated and unmethylated fractions. The methylated fraction can be tagged with biotin opposite the meC by incubation with a DNA polymerase, dCTP and biotin-dGTP. Likewise the unmethylated fraction can be tagged in a separate incubation with a DNA polymerase, dCTP and biotin-dATP. Binding of biotin to avidin or streptavidin, for example coupled to magnetic beads can then be used to isolate the separate methylated and unmethylated fractions.

Claims

1. A method for the identification of methylated cytosine bases in a nucleotide sequence, the method comprising the steps of:

(b) incubating the fragments with bisulfite under non-denaturing conditions such that only unmethylated cytosines within the at least one single-stranded terminus are converted to uracil; and

(c) detecting fragments comprising methylcytosine or uracil bases in the at least one single-stranded terminus.

2. The method of claim 1 wherein the methylcytosine to be detected is within a CG (CpG) dinucleotide.

3. The method of claim 1 wherein the methylcytosine to be detected is within the sequence CCWGG or CNG (CpNpG).

4. The method of any one of claims 1 to 3 wherein the fragments comprising at least one single- stranded terminus are generated using one or more restriction endonucleases that cleave the DNA such that at least a one base single-stranded overhang is produced, the single-stranded overhang comprising at least one cytosine.

5. The method of claim 4 wherein the restriction endonuclease(s) is selected from Msp\, Taq\, /ApeKI, AwW₁ BbvC\, BcW, BgIW, BsaWI, BseYI, Bsm\, BsoBI, SssSI, Btg\, DpnW and H/ndlll.

6. The method of claim 4 or 5 wherein the recognition sequence cleaved by the restriction endonuclease(s) includes the CG dinucleotide or one base thereof.

7. The method of any one of claims 4 to 6 wherein the fragments comprising at least one single- stranded terminus are generated using two restriction endonucleases that cleave the DNA such that a fragment comprising at least a one base single-stranded overhang is produced.

8. The method of claim 7 wherein the restriction endonucleases are Msp\ and Taq\, recognising the sequences C/CGG and T/CGA respectively.

9. The method of any one of claims 1 to 3 wherein the fragments comprising at least one single- stranded terminus are generated by digestion of linear fragments with an enzyme possessing 3' to 5' exonuclease activity in the absence of at least dGTP.

10. The method of claim 9 wherein the enzyme is a DNA polymerase.

11. The method of any one of claims 1 to 10 wherein the bisulfite is selected from sodium bisulfite, sodium metabisulfite or other suitable bisulfite salt.

12. The method of any one of claims 1 to 11 wherein the non-denaturing conditions comprise incubation at about 37⁰C for about 24 hours or less.

13. The method of any one of claims 1 to 12 wherein the detection step (c) comprises the steps of: (i) incubating bisulfite-treated fragments with DNA polymerase and dNTPs wherein either dATP or dGTP is labelled with a detectable marker; and

14. The method of claim 13 wherein the detection step (c) comprises, or is followed by, the selective isolation of a desired subpopulation of fragments comprising either methylcytosine or uracil within the single-stranded terminus.

15. The method of claim 13 or 14 wherein the detectable marker is biotin, detectable using avidin or streptavidin.

16. The method of any one of claims 1 to 12 wherein the detection step (c) comprises the steps of: (i) ligating to the single-stranded terminus of the bisulfite-treated fragments a detection facilitation linker complementary to bases of the single-stranded terminus and comprising at least one adenine or guanine at a position complementary to the at least one cytosine in the fragments of step (a); and (ii) analysing the ligated fragments by any suitable means such as PCR amplification, hybridisation and/or sequencing,

17. The method of claim 16 wherein the bisulfite-treated fragments are separated into two pools, wherein a detection facilitation linker comprising at least one adenine at a position complementary to the at least one cytosine in the fragment of step (a) is ligated to the bisulfite-treated fragments of one pool, and a detection facilitation linker comprising at least one guanine at a position complementary to the at least one cytosine in the fragment of step (a) is ligated to the bisulfite-treated fragments of the other pool.

18. The method of claim 17 wherein in each pool a competitor linker is incubated with the bisulfite- treated fragments and detection facilitation linker, wherein the competitor linker is specific for the non- selected fragment end.

19. The method of any one of claims 16 to 18 wherein the detection facilitation linker comprises sequences from which amplification or sequencing reactions can be initiated using primers capable of hybridising to said sequences.

20. The method of claim 18 wherein the competitor linker differs sufficiently in sequence from the detection facilitation linker such that primers used to hybridise to the detection facilitation linker and initiate an amplification or sequencing reaction therefrom cannot hybridise to the competitor linker.

21. The method of any one of claims 16 to 20 wherein the fragments comprising at least one single-stranded terminus in step (a) are generated using the restriction endonucleases Msp\ and Taq\, and the detection facilitation linkers comprise 5'-CG and 5'-CA to facilitate detection of 5'-methylCG and 5'-UG, respectively, within the single-stranded terminus of the treated fragment.

22. The method of any one of claims 1 to 21 wherein the DNA molecule comprising the nucleotide sequence to be analysed is a genomic DNA sample.

23. The method of claim 22 wherein the genomic DNA sample is pre-treated prior to step (a) by sonication or restriction endonuclease digestion to reduce fragment size.

24. A method for the separation of methylated and unmethylated fractions within a DNA sample, the method comprising the steps of:

(d) optionally, isolating fragments within the fractions.

25. The method of claim 24 wherein the DNA sample comprises a genome or a fragment or portion thereof.

26. The method of claim 24 or 25 wherein the separation step (c) comprises the steps of:

(i) incubating bisulfite-treated fragments with DNA polymerase and dNTPs wherein either dATP or dGTP is labelled with a detectable marker; and

(ii) separating the methylated and unmethylated fractions by virtue of the detectable marker, wherein labelled dATP facilitates the selective isolation of fragments comprising uracil in the single- stranded terminus, and labelled dGTP facilitates the selective detection of fragments comprising methylcytosine within the single-stranded terminus.

27. A method for the diagnosis of, or determination of susceptibility to, a disease associated with aberrant cytosine methylation in a subject, the method comprising the steps of: