NL2031337B1 - Method for separating high-methylated DNA and low-methylated DNA - Google Patents

Abstract

The present invention provides a method for separating high-methylated DNA and low- methylated DNA, the method comprising a) providing a liquid sample comprising high- methylated DNA and low-methylated DNA; b) contacting the sample with a surface to which MBD2 proteins are immobilized to allow formation of MBD2-high-methylated DNA complexes on said surface, thereby separating high-methylated DNA and low-methylated DNA. The sample may be a body fluid, preferably a urine sample or a blood plasma sample.

Description

P35615NLOO/MJO

Method for separating high-methylated DNA and low-methylated DNA

Technical field

The present invention relates to a method for separating high-methylated DNA and low- methylated DNA, preferably for isolating methylated DNA from a sample comprising high- methylated DNA and low-methylated DNA.

Background of the invention

Cancer is a growing health concern worldwide. Despite the increase in cancer survival rates over the last decades, cancer remains one of the deadliest diseases. One of the key aspects of surviving cancer is to detect the disease as early as possible, thereby enlarging the window of opportunity of curing it (Crosby et al. Lancet Oncol. 2020, 21, 1397). A recent trend for early cancer diagnostics is the pre-symptomatic detection of cancer biomarkers in liquid biopsy samples and, therefore, improving the survival rate (Ignatiadis et al. Nat. Rev. Clin.

Oncol. 2021, 18, 297).

Hypermethylated DNA (hmDNA) is one of the typical biomarkers found in liquid biopsy samples of cancer patients (de Rubis et al. Trends Pharmacol. Sci. 2019, 40, 172). The presence of hMmDNA in blood and urine is correlated to multiple types of cancer, including gastric, lung and ovarian cancer (van Helder et al. Clin. Epigenetics 2020, 12, 165). From a genetic point of view, DNA methylation takes place predominantly at promotor regions with a relative high amount of cytosine bases that are followed directly by a guanine (CpGs) in the 5'-to-3 direction, the so-called CpG-rich regions (Zhang et al. Cancers (Basel). 2020, 12, 2123). The methylation of a CpG is an epigenetic alternation in which a methyl group is covalently bonded to the cytosine base at the fifth carbon. By CpG methylation, gene expressions are controlled in cells. As a consequence, methylation can play a role in tumor development when tumor suppressor genes are methylated, thereby repressing the transcription and thus silencing these genes (Herman et al. N. Engl. J. Med. 2003, 349, 2042). Hypermethylation refers to the situation that methylation of a promotor region occurs, which in a healthy situation does not occur.

Early disease detection as well as disease progress and the effectiveness of a therapy can be monitored by measuring the hmDNA concentration over time. Yet, the concentration of hmDNA, especially in early-stage cancers, can be as low as a few DNA copies per liquid biopsy sample (Zheng et al. Epigenetics 2014, 9, 483). The current approach to measure hmDNA employs DNA isolation and bisulfite conversion followed by quantitative PCR (of the region of interest; can be cancer dependent), making it time-consuming and labor intensive.

The usage of hmDNA as a biomarker for cancer detection is currently limited, because it is difficult to distinguish hmDNA from non-methylated DNA (Corcoran et al. N. Engl. J. Med. 2018, 379, 1754.). As a result, the method is not widely applicable in the clinic. The detection of specific hmDNA biomarkers by sequencing or biosensing approaches involves the ability to distinguish between hmDNA and healthy, non- or low-methylated DNA. Commonly a preselection step is applied, and three different approaches exist in order to differentiate between hmDNA and non-methylated “background” DNA (Olkhov-Mitsel et al. Cancer Med. 2012, 1, 237).

The first approach uses methyl-sensitive restriction enzymes (Hofner et al. Epigenetics

Methods, 2020, pp. 181-212). These enzymes are able to cleave DNA at a specific recognition sequence only when the CpG is (non-)methylated, depending on the enzyme used. For example, after digesting the DNA sample with a specific enzyme, all the non- methylated DNA is cleaved, while the methylated DNA remains intact. The second approach to differentiate between the two types of DNA is bisulfite conversion (Herman et al. Proc. Natl.

Acad. Sci. U. S. A. 1996, 93, 9821). Using this chemical modification method, all non- methylated cytosines are converted into the base uracil, while all methylated cytosines remain unconverted. As a result, the base sequence of the non-methylated DNA changes, allowing an easier differentiation using PCR.

Another method described to differentiate between hmDNA and non-methylated DNA uses affinity chromatography, for example in combination with a methyl binding domain (MBD) protein attached to the surface of beads employed in the method (Cross et al. Nat. Genet. 1994, 6, 236). The general approach of this method consists of three consecutive steps, namely: 1) immobilizing MBD protein (such as MBD2) onto a surface such as a column material or nano/micro-size particles; 2) binding of hmDNA to the MBD-modified surface; 3) the bound DNA is removed from the surface with an elution buffer to enable subsequent characterization.

A detailed performance study using this method has for example been described by Nair et al (Epigenetics 2011, 6, 34). There are several major limitations in the current MBD2-based enrichment systems; they are therefore deemed inappropriate for widespread use. A first limitation is that a significant amount of non-methylated DNA remains present in the enriched

DNA sample, meaning that there is much contamination of background DNA and the sensitivity and/or specificity is poor. The “enrichment factor” is commonly used to describe the efficiency of enrichment, and is defined as the enriched amount of methylated DNA divided by the enriched amount of non-methylated DNA. Wee at al. has proposed the incorporation of sperm DNA as blocking agent upon hmDNA enrichment. However, even with this method, enrichment factors of between 2.1 and 14.2 could only be achieved (Wee et al. Clin.

Epigenetics 2015, 7, 65).

Alternatively, Warton et al. aimed to improve the methylated DNA enrichment by decreasing the amount of MBD2-functionalized magnetic beads upon enrichment of DNA. However this method led to only a limited positive effect on the enrichment factor. A second limitation is that current systems are only able to differentiate well between hmDNA and non-methylated

DNA at high input concentrations but not at combined low and high input concentrations of hmDNA and non-methylated DNA, respectively. For example, enrichment factors of only 1 and 5 were achieved when using 11 ng non-methylated DNA and 0.03 or 0.1 ng of hmDNA, respectively (Yegnasubramanian et al. Nucleic Acids Res. 2006, 34, e19). Such low levels of of hmDNA can be found in the clinical setting, hence there is a need for far more sensitive methods MBD2-based enrichment systems.

At the moment, there is an unmet need for an enrichment method that overcomes one or more limitations of the MBD2-based enrichment systems, and which therefore is suitable for widespread (clinical) use. The present invention aims to provide such a method.

Summary of the invention

The inventors surprisingly found a method to greatly enhance the selectivity of hMDNA enrichment from a sample comprising hmDNA and low-methylated DNA. The method follows the finding by the inventors that a control of the MBD2 surface receptor density is a way to improve selectivity and affinity for hmDNA over low-methylated DNA. In particular, it is shown that high hmDNA selectivity is favored by a low multivalency enhancement factor and a high intrinsic selectivity of the receptors. The MBD2 surface receptor density can be effectively controlled by employing available methods in the art, for example by using a platform which involves a mixed thiol self-assembled monolayer (SAM) on a (gold) surface as shown by the present inventors. The platform preferably comprises a first spacer compound, preferably hydroxyl thiol, to provide an anti-fouling effect. The platform preferably comprises azide thiol as a second spacer and which can for instance be combined with a linker molecule to enable

MBD2 immobilization such as by click chemistry. In a preferred embodiment, the linker molecule preferably carries dibenzocyclooctyne (DBCO) functional groups such that allows it to react with azide thiol, while at the other end, the linker molecule preferably carries a nitrilotriacetic acid (NTA) functional group such that it can be complexed by Ni?* ions to form

NINTA moieties. By tagging the MBD2 protein with a histidine (His) tag, this allows immobilization of the MBD2 protein at the surface through interaction with the NiNTA complexes. After the binding step, the DNA mixture can be introduced on the surface. Then, the surface bound hmDNA can be successfully removed using a basic solution.

Using the method of the invention, it was found that a MBD2 surface receptor between 100- 500 ng/cm? leads to particularly high selectivity and affinity for hmDNA over DNA which is not hmDNA. In particular, it was found that that the stoichiometric ratio between hydroxyl and azide-functionalized thiols in the SAM can be fine-tuned to control the MBD2 surface receptor density. For example, having a percentage of spacers with an azide group of between 5-20% {of the total of spacers with a hydroxyl group and spacers with an azide group), leads to particularly high selectivity and affinity for hmDNA over DNA which is not hmDNA.

The method of the invention overcomes at least in part the limitations of MBD2-based enrichment systems using magnetic beads as according to the prior art. To illustrate, it was found that the amount of hypermethylated DNA content could be enriched by at least 20-fold, also when the amount of hmDNA concentration in a given sample is (very) low. For example, a high selectivity and affinity for hmDNA can be achieved when the DNA in a liquid sample in the nanogram/ml range or lower. The high enhance selectivity and affinity for hmDNA was confirmed for a DNA mixture comprising 1% hmDNA and 99% non-methylated DNA (i.e. the ratio comparable to the ratio found in liquid biopsy samples such as urine). Through the present disclosure, an MBD2-based method is achieved that is more appropriate for clinical use (as compared to MBD2-based enrichment systems using magnetic beads. For example, the inventors foresee applicability in a medical setting by implementing the current invention in an analytical process or a lab-on-a-chip device.

Detailed description of the invention

The present disclosure provides a method for separating high-methylated DNA and low- methylated DNA, the method comprising: a) providing a (liquid) sample comprising high-methylated DNA and low-methylated DNA, preferably in an amount of between 1-5000 ng high-methylated DNA and low-methylated

DNA per g or ml sample;

b) combining or contacting the sample with a surface to which MBD2 proteins (Methyl-CpG

Binding Domain Protein 2) are immobilized in a density of preferably between 100-500 ng/cm? to allow formation of MBD2-high-methylated DNA complexes on said surface, thereby separating high-methylated DNA and low-methylated DNA, wherein preferably high-methylated DNA concerns DNA sequence(s) comprising at least 2 methylated CpG sites, and wherein preferably low-methylated DNA concerns DNA sequence(s) comprising less methylated CpG sites in comparison to the high-methylated

DNA.

In the present disclosure, high-methylated DNA concerns DNA sequence(s) comprising at least 1, 2, 3, 4, 5, 6, 8, 10 methylated CpG sites, wherein the number of methylated CpG sites preferably refers total number of methylated CpG sites in the sequence. Low-methylated DNA may concern DNA sequences with fewer methylated CpG sites than high-methylated DNA.

The low-methylated DNA and high-methylated may be provided in a same (liquid) sample. In addition or alternatively, low-methylated DNA may concern DNA sequence(s) comprising at most 1, 2, 3 methylated CpG sites. The number of methylated CpG sites preferably refers total number of methylated CpG sites in the sequence. In particular, and in a preferred embodiment, CpG methylation may occur in CpG-rich regions which are regions in the DNA, such as a promotor region, with a relative high amount of cytosine bases that are followed directly by a guanine in the 5'-to-3’ direction. The methylation of a CpG is considered to be an epigenetic alternation in which a methyl group is covalently bonded to the cytosine base at the fifth carbon. Methylation encompasses methylation of cytosine and/or adenine. In cancer, methylation of cytosine is most frequently observed. The term “5-Methylcytosine (5mC)” means that a methyl group is present at the N5 position of cytosine. In an embodiment, the methylation of CpG sites encompasses 5mC methylation. In an embodiment, the methylation of CpG sites refers to 5mC methylation. The degree of methylation typically depends on the type of tissue (Ehrlich et al. Oncogene volume 21, pages 5400-5413. 2002). Low- methylation, normal methylation and high or methylation {e.g. hypermethylation) of DNA are relative terms and denote less or more methylation than in some standard DNA in the given context.

The skilled person is familiar with appropriate methods to measure DNA methylation. There are at least three primary methods to identify and quantify DNA methylation (Laird et al. Nat.

Rev. Genet. 11:191-203. PMID: 20125086): 1) sodium bisulfite conversion and sequencing; 2) differential enzymatic cleavage of DNA; and

3) affinity capture of methylated DNA, such as by methylated DNA immunoprecipitation (Me-

DIP) that uses methyl DNA specific antibody, or methyl capture using methyl-CpG binding domain (MBD) proteins.

The method according to the present disclosure is for separating high-methylated and low- methylated DNA, e.g. isolating high-methylated DNA from a sample comprising high- methylated and low-methylated DNA. The present method thus allows for selective enrichment of high-methylated DNA from a sample comprising high-methylated DNA and low- methylated DNA. The method according to the present disclosure may be used for determining a ratio between high-methylated DNA and low-methylated DNA in the sample, preferably by performing quantitative polymerase chain reaction (PCR), e.g. on the separated high-methylated DNA as well as on the low-methylated DNA.

In step a), the sample may be a liquid sample, such as an aqueous sample. Preferably, the sample is a body fluid sample, or comprises a body fluid sample, comprising high-methylated and low-methylated DNA, wherein the body fluid sample is preferably chosen from the group consisting of - a urine sample comprising high-methylated and low-methylated DNA; - a blood plasma sample comprising high-methylated and low-methylated DNA; - a saliva sample comprising high-methylated and low-methylated DNA; - a cerebrospinal fluid sample comprising high-methylated and low-methylated DNA; - a sputum sample comprising high-methylated and low-methylated DNA; - a stool (derived) sample comprising high-methylated and low-methylated DNA; or - a pleural fluid sample comprising high-methylated and low-methylated DNA.

The liquid samples as disclosed herein encompasses a body fluid that has been diluted (in an aqueous fluid), further concentrated, or processed in another way, before use in the method of the disclosure.

The sample may comprise high-methylated DNA and low-methylated DNA, in a total amount of between 1-50000, 1-10000, 1-5000, 5-2000 ng/ml, preferably 1-1000 ng/ml, more preferably 200-800 or 300-700 ng/ml. In addition or alternatively, the DNA in a sample may concern DNA sequences having an average length of between 10-500 bp, such as 10-400 bp or 10-300 bp, preferably between 20-200 bp. In addition or alternatively, the ratio between high-methylated DNA and low-methylated DNA in the sample may be 1:1 -1: 100000, 1:1 -

1:10000, or 1:1- 1:100.The term DNA sequence is interchangeable with the term nucleic acid sequence.

The sample may be obtained from a mammalian subject, preferably a human. The present disclosure preferably does not involve a method for treatment of the human or animal body by surgery or therapy and/or a diagnostic method practiced on the human or animal body.

In step b), the sample is combined and/or contacted with a surface to which MBD2 proteins are immobilized (or bound) in a density of 100-500 ng MBD2 proteins per cm?, preferably between 150-400, or between 150-250, 175-225, preferably 190-210, or between 200-350, preferably 250-300 or 265-285 ng/cm? to allow formation of MBD2-high-methylated DNA complexes on said surface, thereby separating high-methylated and low-methylated DNA.

A density of between 150-250, 175-225, preferably 190-210 ng MBD2 proteins per cm? may be particularly suitable for use in combination with a blood plasma sample. A density of between 200-350, preferably 250-300 or 265-285 ng MBD2 proteins per cm? may be particularly suitable for use in combination with a urine sample.

The MBD2 protein(s) according to the present disclosure preferably has an amino acid sequence having at least 70, 80, 90, 95, 99 or 100% sequence identity with SEQ ID NO: 1 and/or is encoded by a nucleotide sequence having at least 70, 80, 90, 95, 99 or 100% sequence identity with SEQ ID NO:2.

Although MBD2 proteins are preferred in the present disclosure, also Methyl-CpG binding protein 2 (MeCP2) or any of Methyl-CpG Binding Domain Protein 1, 2, 3 or 4 may be used alternatively. Human proteins MECP2, MBD1, MBD2, MBD3, and MBD4 comprise a family of nuclear proteins related by the presence in each of a methyl-CpG-binding domain (MBD).

Each of these proteins, with the exception of MBD3, is capable of binding specifically to high- methylated DNA.

The family of MBD proteins contains five members that have a particularly higher binding affinity for high-methylated compared to low-methylated DNA. These are methyl-CpG binding protein 2 (MeCP2), MBD1, MBD2, MBD3 and MBD4. Despite all are selective for a C*pG, the absolute binding affinities of the individual proteins within the MBD family for C*pG and CpG differ. The binding affinities of an MBD protein that is interacting with C*pG or CpG is important for the overall affinity with (non-)methylated DNA and for the selectivity of the enrichment of hmDNA.

In particular MeCP2, MBD1 and MBD2 possess low Ky values for C*pG, but all proteins show, rather comparable, affinities for CpG. In particular the ratio of the Ky values for CpG and

C*pG, which is a measure for the selectivity, can be a proper measure to identify the most optimal receptor for the enrichment. It can be expected that the factors complicating the comparison of Ky values between different proteins is largely eliminated by taking ratios, so by looking at the selectivities. High selectivities (= Ka,cpe/Ka, ps) are achieved by MeCP2,

MBD1 and MBD2, with values ranging from 50-295.

Good individual selectivities are obtained for MeCP2, MBD1 and MBD2, which are much better than those of MBD3 and MBD4. Most likely, any of the first three can be used in achieving enrichment. Yet, as will become clear from the multivalency and superselectivity as proposed herein, the intrinsic selectivity (/.e., the ratio of Ky values for individual CpG/C*pG sites) is not enough to predict the best selectivity in a multivalent binding event for (methylated) DNA that contains multiple C(*)pG sites. Qualitatively, the overall selectivity for binding hmDNA will be better when a good intrinsic selectivity is coupled to a relatively weak binding constant (high Ka), in particular for CpG, and high valency (number of interactions) leading to a “multiplication” of several individual selectivities. In that way, a large difference in overall K values for hmDNA and low-methylated can be achieved in the biologically relevant concentration range. MBD2 has been demonstrated to have one of the best intrinsic binding selectivities among the proteins present within the MBD protein family, with dissociation constant values ranging between 188 — 6500 nM for non-methylated CpG (CpG).

The method according to the present disclosure may comprise an additional step c) of elution of the high-methylated DNA from the surface to which MBD2 proteins are immobilized by applying a basic solution, thereby obtaining the high-methylated DNA. Instead of using a basic solution, elution may be achieved by using an aqueous solution with between 0.1-5 wt% salt relative to the weight of the solution, or using a aqueous solution of at least 30,40, 50, 60 degrees Celsius (preferably at most 80, 90, 99 degrees Celsius), or a aqueous solution comprising imidazole or with Ni-complexing ligand (e.g. EDTA).

In preferred embodiment according to the present disclosure, the surface to which MBD2 proteins are immobilized is a surface comprising

- spacers without reactive moiety (e.g. which do not couple to DBCQ), preferably spacers with a hydroxyl group, and spacers with reactive moiety (e.g. a click moiety), preferably spacers with an azide group, i.e. azide functionalized spacers, wherein preferably the percentage of spacers with reactive moiety (e.g. an azide group) is between 5-20%, preferably 5-15%, 8- 12% or 10-20%, 12-18%, relative to the total number of spacers; - linkers comprising at one end a complementary reactive group (e.g. a dibenzocyclooctyne (DBCO) moiety) and at another end a MBD2 protein, wherein the complementary reactive group (e.g. the DBCO moiety) of the linkers is coupled to the reactive group (e.g. azide group) of the spacers.

The skilled person is aware that also alternative methods may be applied to control the density of MBD2 proteins on the surface, for example by applying - silanes on silicon oxide; - polyelectrolytes (bv PLL) on various surfaces, such as metals, metal oxides and polymers.

All these methods allow to control density of reactive groups (for example azide) and thereby to control MBD2-density.

A density of between 150-250, 175-225, preferably 180-210 ng MBD2 proteins per cm? may correspond to a percentage of spacers with an azide group between 5-15%, preferably 8-12% relative to the total number of spacers. A density of between 200-350, preferably 250-300 or 265-285 ng MBD2 proteins per cm? may correspond to a percentage of spacers with an azide group between 10-20%, preferably 12-18% relative to the total number of spacers be particularly suitable for use in combination with a urine sample.

In a particularly preferred embodiment according to the present disclosure, the surface to which MBD2 proteins are immobilized is a surface comprising - spacers which do not couple to DBCO, preferably spacers with a hydroxyl group, and spacers with an azide group, i.e. azide functionalized spacers, wherein preferably the percentage of spacers with an azide group is between 5-20%, preferably 5-15%, 8-12% or 10-20%, 12-18%, relative to the total number of spacers; - linkers comprising at one end a dibenzocyclooctyne (DBCO) moiety and preferably at another end a nitrilotriacetic acid (NTA) moiety chelated or complexed to Nickel (II), wherein the DBCO moiety of the linkers is coupled to the azide group of the spacers; - MBD2 proteins preferably comprising a histidine tag, wherein the histidine tag of the proteins is bound to the nitrilotriacetic acid (NTA) moiety chelated to Nickel (II) of the linkers.

The present disclosure also provides for a surface to which MBD2 proteins are immobilized in a density of preferably between 100-500 ng/cm?, the surface comprising - spacers without reactive moiety (e.g. which do not couple to DBCO), preferably spacers with a hydroxyl group, and spacers with reactive moiety (e.g. a click moiety), preferably spacers with an azide group, i.e. azide functionalized spacers, wherein preferably the percentage of spacers with reactive moiety (e.g. an azide group) is between 5-20%, preferably 5-15%, 8- 12% or 10-20%, 12-18%, relative to the total number of spacers; - linkers comprising at one end a complementary reactive group (e.g. a dibenzocyclooctyne (DBCO) moiety) and at another end a MBD2 protein, wherein the complementary reactive group (e.g. the DBCO moiety) of the linkers is coupled to the reactive group (e.g. azide group) of the spacers.

The present disclosure also provides for a surface to which MBD2 proteins are immobilized in a density of preferably between 100-500 ng/cm?, the surface comprising - spacers which do not couple to DBCO, preferably spacers with a hydroxyl group, and spacers with an azide group, i.e. azide functionalized spacers, wherein preferably the percentage of spacers with an azide group is between 5-20%, preferably 5-15%, 8-12% or 10-20%, 12-18%, relative to the total number of spacers; - linkers comprising at one end a dibenzocyclooctyne (DBCO) moiety and at another end a

MBD2 protein, wherein the DBCO moiety of the linkers is coupled to the azide group of the spacers.

More preferably, the present disclosure provides for a surface to which MBD2 proteins are immobilized in a density of preferably between 100-500 ng/cm?, the surface comprising - spacers which do not couple to DBCO, preferably spacers with a hydroxyl group, and spacers with an azide group, i.e. azide functionalized spacers, wherein preferably the percentage of spacers with an azide group is between 5-20%, preferably 5-15%, 8-12% or 10-20%, 12-18%, relative to the total number of spacers; - linkers comprising at one end a dibenzocyclooctyne (DBCO) moiety and preferably at another end a nitrilotriacetic acid (NTA) moiety chelated to Nickel (II), wherein the DBCO moiety of the linkers is coupled to the azide group of the spacers; - MBD2 proteins preferably comprising a histidine tag, wherein the histidine tag of the proteins is bound to the nitrilotriacetic acid (NTA) moiety chelated to Nickel (11) of the linkers.

In a preferred embodiment, the MBD2 proteins are immobilized in a density of 100-500 ng (e.g. 100-400 ng) MBD2 proteins per cm?, preferably between 150-400, or between 150-250, 175-225, preferably 190-210, or between 200-350, preferably 250-300 or 265-285 ng/cm?.

In an embodiment, -the liquid sample as disclosed herein is a urine sample, or comprises a urine sample; and - the surface on which MBD2 proteins are immobilized as disclosed herein has a percentage of spacers with an azide group of 5-15%, preferably 7.5-12.5%, relative to the total number of spacers (e.g. relative to a mixture of thiol-functionalized and azide-functionalized spacers).

In an embodiment, -the liquid sample as disclosed herein is a blood sample, or comprises a blood sample; and - the surface on which MBD2 proteins are immobilized as disclosed herein has a percentage of spacers with an azide group of 10-20%, preferably 12.5-17.5%, relative to the total number of spacers (e.g. relative to a mixture of thiol-functionalized and azide-functionalized spacers).

The present disclosure also provides for a method for making a surface to which MBD2 proteins are immobilized in a density of between 100-500 ng/cm?, the method comprising a) cleaning a gold surface, preferably by exposing the gold surface to ultraviolet light and/or ozone, and optionally rinsing the gold surface (e.g. with ethanol); b) combining the cleaned gold surface with a thiol solution comprising spacers with a hydroxyl group and spacers with an azide group, wherein the percentage of spacers with an azide group is between 5-20, preferably 5-15 or 10-20% or 8-12% or 12-18%, relative to the total number of spacers, to thereby obtain a thiol based self-assembled monolayer (SAM) on the gold surface; c) combining the SAM with linkers comprising at one end a dibenzocyclooctyne (DBCO) moiety and at another end a nitrilotriacetic acid (NTA) moiety, to thereby obtain SAM comprising NTA functionalized linkers; d) combining the SAM comprising NTA functionalized linkers with NiClz to thereby obtain

SAM comprising nitrilotriacetic acid (NTA) moieties chelated or complexed to Nickel (II); e) combining the SAM comprising nitrilotriacetic acid (NTA) moieties chelated (or complexed) to Nickel (Il) with MBD2 proteins to thereby obtain a surface to which MBD2 proteins are immobilized in a density of between 100-500 ng/cm2.

Preferably the method is for making a surface to which MBD2 proteins are immobilized in a density of between 100-500, or between 150-400, or between 150-250, 175-225, preferably 190-210, or between 200-350, preferably 250-300 or 265-285 ng/cm?.

The spacers according to the present disclosure preferably form a thiol-based self-assembled monolayer (SAM) on the surface, which preferably is a gold surface. Accordingly, he spacers may be bound to the gold surface by a gold-sulfur bond. Hence, the surface according to the present disclosure is preferably a gold surface, preferably on a flat gold surface or a gold surface on gold beads having a size of between 1 nm and 1 Hm.

The spacers with a hydroxyl group are preferably hydroxyl ethylene glycol-alkanethiols and/or the spacers with an azide group are preferably azide-functionalized ethylene glycol- alkanethiols. An azide group can react with a DBCO group. In the present disclosure, the term “azide” is also used after the reaction has occurred.

The linkers according to the present disclosure are preferably DBCO-polyethyleneglycol-NTA chelated to Nickel (II).

Clauses

Herein, clauses are embodiments of the invention. Features of clauses (embodiments) herein can be combined.

Clause 1. Method for separating high-methylated DNA and low-methylated DNA, the method comprising: a) providing a liquid sample comprising high-methylated DNA and low-methylated DNA, in an amount of 1-5000 ng DNA per ml; b) contacting the sample with a surface to which (methyl binding domain 2) MBD2 proteins are immobilized in a density of between 100-500 ng/cm? to allow formation of MBD2-high- methylated DNA complexes on said surface, thereby separating high-methylated DNA and low-methylated DNA, wherein high-methylated DNA concerns DNA sequence(s) comprising at least 2 methylated

CpG sites, and wherein low-methylated DNA concerns DNA sequence(s) comprising less methylated CpG sites in comparison to the high-methylated DNA.

Clause 2. Method according to clause 1, wherein step a) provides a liquid sample which is a aqueous sample or a body fluid sample comprising high-methylated DNA and low-methylated

DNA, preferably a urine sample comprising high-methylated DNA and low-methylated DNA or a blood plasma sample comprising high-methylated DNA and low-methylated DNA.

Clause 3. Method according to any one of the previous clauses, wherein step a) provides a liquid sample comprising high-methylated DNA and low-methylated DNA, in an amount of 5- 2000, preferably 1-1000 ng/ml.

Clause 4. Method according to any one of the previous clauses, wherein step b) contacts the mixture with a surface to which MBD2 proteins are immobilized in a density of between 100- 500, 150-400, 150-250, 200-350, preferably 250-300 ng/cm? to allow formation of MBD2-high- methylated DNA complexes on said surface, thereby separating high-methylated DNA and low-methylated DNA.

Clause 5. Method according to any one of the previous clauses, wherein said MBD2 proteins have an amino acid sequence having at least 70, 80, 90% sequence identity with SEQ ID

NO:1.

Clause 6. Method according to any one of the previous clauses, wherein the method is for determining a ratio between high-methylated DNA and low-methylated DNA, preferably by performing quantitative polymerase chain reaction (PCR).

Clause 7. Method according to any one of the previous clauses, comprising step c) of elution of high-methylated DNA from the surface to which MBD2 proteins are immobilized preferably by applying a basic solution, thereby obtaining the high-methylated DNA.

Clause 8. Method according to any one of the previous clauses, wherein the surface to which

MBD2 proteins are immobilized is a surface comprising - spacers with a hydroxyl group and spacers with an azide group, wherein the percentage of spacers with an azide group is between 5-20%, preferably 5-15% or 10-20%, relative to the total number of spacers, - linkers comprising at one end a dibenzocyclooctyne (DBCO) moiety and at another end a nitrilotriacetic acid (NTA) moiety chelated to Nickel (Il), wherein the DBCO moiety of the linkers is coupled to the azide group of the spacers; - MBD2 proteins comprising a histidine tag, wherein the histidine tag of the proteins is bound to the nitrilotriacetic acid (NTA) moiety chelated to Nickel (11) of the linkers.

Clause 9. Method according to clause 8, wherein the spacers form a thiol-based self- assembled monolayer (SAM) and/or wherein the surface is a gold surface, wherein the spacers are bound to the gold surface by a gold-sulfur bond.

Clause 10. Method according to any one of clauses 8-9, wherein the spacers with a hydroxyl group are hydroxyl ethylene glycol-alkanethiols and/or wherein spacers with an azide group are azide-functionalized ethylene glycol-alkanethiols.

Clause 11. Method according to any one of clauses 8-10, wherein the linkers are DBCO- polyethyleneglycol-NTA chelated to Nickel (11).

Clause 12. Surface to which MBD2 proteins are immobilized in a density of between 100-500 ng/cm?, the surface comprising - spacers with a hydroxyl group and spacers with an azide group, wherein the percentage of spacers with an azide group is between 5-20, preferably 5-15 or 10-20%, relative to the total number of spacers; - linkers comprising at one end a dibenzocyclooctyne (DBCO) moiety and at another end a nitrilotriacetic acid (NTA) moiety chelated to Nickel (Il), wherein the DBCO moiety of the linkers is coupled to the azide group of the spacers; - MBD2 proteins comprising a histidine tag, wherein the histidine tag of the proteins is bound to the nitrilotriacetic acid (NTA) moiety chelated to Nickel (II) of the linkers.

Clause 13. Surface to which MBD2 proteins are immobilized according to clause 12, wherein the MBD2 proteins are immobilized in a density of between 100-500, 150-400, 150-250, 200- 350, preferably 250-300 ng/cm?.

Clause 14. Surface to which MBD2 proteins are immobilized according to any one of clauses 12-13, wherein the spacers form a thiol-based self-assembled monolayer (SAM) and/or wherein the surface is a gold surface, wherein the spacers are bound to the gold surface by a gold-sulfur bond.

Clause 15. Surface to which MBD2 proteins are immobilized according to any one of clauses 12-14, wherein the spacers with a hydroxyl group are hydroxyl ethylene glycol-alkanethiols and/or wherein spacers with an azide group are azide-functionalized ethylene glycol- alkanethiols.

Clause 16. Surface to which MBD2 proteins are immobilized according to any one of clauses 12-15, wherein the linkers are DBCO-polyethyleneglycol-NTA chelated to Nickel (II).

Clause 17. Method for making a surface to which MBD2 proteins are immobilized in a density of between 100-500 ng/cm?, the method comprising a) cleaning a gold surface, preferably by exposing the gold surface to ultraviolet light and/or ozone; b) combining the cleaned gold surface with a thiol solution comprising spacers with a hydroxyl group and spacers with an azide group, wherein the percentage of spacers with an azide group is between 5-20, preferably 5-15 or 10-20%, relative to the total number of spacers, to thereby obtain a thiol based self-assembled monolayer (SAM) on the gold surface; c) combining the SAM with linkers comprising at one end a dibenzocyclooctyne (DBCO) moiety and at another end a nitrilotriacetic acid (NTA) moiety, to thereby obtain SAM comprising NTA functionalized linkers; d) combining the SAM comprising NTA functionalized linkers with NiCl, to thereby obtain

SAM comprising nitrilotriacetic acid (NTA) moieties chelated to Nickel (ll); e) combining the SAM comprising nitrilotriacetic acid (NTA) moieties chelated to Nickel (II) with MBD2 proteins to thereby obtain a surface to which MBD2 proteins are immobilized in a density of between 100-500 ng/cm?.

Clause 18. Method according to clause 17, wherein the method is for making a surface to which MBD2 proteins are immobilized in a density of between 100-500, 150-400, 150-250, 200-350, preferably 250-300 ng/cm?.

Clause 19. Method according to any one of clauses 17-18, wherein in step b) the spacers form a thiol-based self-assembled monolayer (SAM) and/or wherein the spacers are bound to the gold surface by a gold-sulfur bond.

Clause 20. Method according to any one of clauses 17-19, wherein the spacers with a hydroxyl group are hydroxyl ethylene glycol-alkanethiols and/or wherein the spacers with an azide group are azide-functionalized ethylene glycol-alkanethiols.

Clause 21. Method according to any one of clauses 17-20, wherein the linkers are DBCO- polyethyleneglycol-NTA chelated to Nickel (11).

Sequences referred to:

SEQID NO:1:

: 10 20 30 40 50

MRAHPGGGRC CPEQEEGESA AGGSGAGGDS AIEQGGQGSA LAPSPVSGVE

: 60 70 80 90 100 :

REGARGGSEG RGRWKQAGRG GGVOGRGRGR GRGRGRGEGR GRGRGRPESSE

110 120 130 140 150 ì

GSCLGSGDGGG CGCGGSCGSG APRREPVPFP SGSAGPGPRG PRATESGKRM : 159 179 180 190 200 ®

DCPALPPGWK KEEVIRKSGL SAGKSDVYYF SPSGKKFRSE PQLARYLGNT : (VDLSSFDFRT GEMMPSKLOK NKQRLRNDFL NGNKGKPDLN TTLEIRQTAS 260 270 280 290 300

IFKOPVTEVT NHPSNKVKSD PORMNEQPRO LFWEKRLOGL SASDVTEQII

310 320 330 340 350

KTMELERGLG GVGPGSNDET LLSAVASALH TSSAPITGQV SAAVEKNPAV

360 370 380 390 400

WLNT SOPLCK AFIVTDEDIR KOQEERVOOVR KKLEEALMAD ILSRAADTEE

: 410

MDIEMDSGDE A

SEQ ID NO: 2: Gene sequence of His:sMBD2 based on the MBD2 amino acid sequence with

UniProtKB/Swiss-Prot ID: Q9UBB5.1. 5-3

ATGCACCATCACCATCATCACCATCATCACCACGATTGTCCTGCGTTGCCGCCC

GGATGGAAAAAAGAGGAAGTTATTCGTAAATCTGGTCITGAGTGCGGGCAAGT

CAGATGTATATTATTTCTCCCCTTCGGGTAAAAAGTTCCGTAGTAAACCTCAAC

TCGCGCGCTACCTTGGCAATACAGTGGATCTCAGTTCTTTCGATTTITCGCACTG

GAAAGATGATGCCATCAAAGCTGCAAAAAAATAAACAGCGCCTACGCAACGA

CCCACTTAACCAAAATAAAGGCAAACCAGATTTAAATACAACCTTACCCATTCG

TCAGACTGCGTCTATITTTAAACAACCGGTCACCAAAGTAACCAATCATCCGA

HisioMBD2 gene GTAACAAAGTTAAATCCGACCCGCAACGCATGAACGAGCAACCGCGGCAGTT

ATTTTGGGAGAAGCGCTTGCAGGGCCTGTCGGCGTCCGATGTCACCGAACAG

ATCATTAAGACCATGGAGTTGCCGAAAGGCCTGCAGGGCGTTGGTCCGGGTA

GCAACGACGAGACCCTGCTGTCAGCCGTGGCATCCGCGCTGCACACCAGCAG

CGCACCGATTACGGGTCAGGTGTCGGCTGCCGTGGAAAAAAACCCGGCCGTT

TGGCTGAACACGTCGCAGCCGCTGTGCAAAGCCTTTATCGTCACGGACGAAG

ATATCCGAAAACAGGAAGAACGTGTGCAGCAGGTGCGTAAAAAACTGGAAG

AAGCTCTGATGGCCGACATACTGAGCAGAGCAGCAGACACGGAAGAAATGG

ATATCGAAATGGATAGCGGTGATGAAGCTTAA

As used herein, the term “identity” refers to a measure of the identity of nucleotide sequences or amino acid sequences. In general, the sequences are aligned so that the highest order match is obtained. "Identity" per se has an art-recognized meaning and can be calculated using published techniques. See, e.g.: (COMPUTATIONAL MOLECULAR BIOLOGY, Lesk,

A. M., ed., Oxford University Press, New York, 1988; BIOCOMPUTING: INFORMATICS AND

GENOME PROJECTS, Smith, D. W., ed., Academic Press, New York, 1993; COMPUTER

ANALYSIS OF SEQUENCE DATA, PART |, Griffin, A. M., and Griffin, H. G., eds., Humana

Press, New Jersey, 1994; SEQUENCE ANALYSIS IN MOLECULAR BIOLOGY, von Heinje,

G., Academic Press, 1987; and SEQUENCE ANALYSIS PRIMER; Gribskov, M. and

Devereux, J., eds., M Stockton Press, New York, 1991). While there exist a number of methods to measure identity between two polynucleotide or polypeptide sequences, the term "identity" is well known to skilled artisans (Carillo, H., and Lipton, D., SIAM J. Applied Math (1988) 48:1073). Methods commonly employed to determine identity or similarity between two sequences include, but are not limited to, those disclosed in GUIDE TO HUGE

COMPUTERS, Martin J. Bishop, ed., Academic Press, San Diego, 1994, and Carillo, H., and

Lipton, D., SIAM J. Applied Math (1988) 48:1073. Methods to determine identity and similarity are codified in computer programs. For example NCBI Nucletide Blast with standard settings (blastn, https://blast.ncbi.nim.nih.gov/). Preferred computer program methods to determine identity and similarity between two sequences include, but are not limited to, GCS program package (Devereux, J., et al., Nucleic Acids Research (1984) 12(1):387), BLASTP, BLASTN,

FASTA (Atschul, S. F. et al., J. Molec. Biol. (1990) 215:403).

As an illustration, by a polypeptide (i.e. protein) or nucleotide sequence having at least, for example, 95% "identity" to a reference polypeptide or nucleotide sequence, it is intended that the polypeptide or nucleotide sequence is identical to the reference sequence, except that there may be up to five point mutations per each 100 amino acids or nucleotides of the reference polypeptide or nucleotide sequence. In other words, to obtain a polypeptide or nucleotide sequence being at least 95% identical to a reference polypeptide or nucleotide sequence, up to 5% of the amino acids or nucleotides in the reference sequence may be deleted and/or substituted with another residue or nucleotide, and/or a number of residues or nucleotides up to 5% of the total residues or nucleotides in the reference sequence may be inserted into the reference sequence. Preferably, the sequence identity refers to the sequence identity over the entire length of the sequence. It is further understood that, when referring to “sequences” herein, generally the actual physical molecules with a certain sequence of subunits (e.g. amino acids or nucleotides) are referred to.

In this document and in its claims, the verb "to comprise" and its conjugations is used in its non-limiting sense to mean that items following the word are included, but items not specifically mentioned are not excluded. In addition, reference to an element by the indefinite article "a" or "an" does not exclude the possibility that more than one of the elements is present, unless the context clearly requires that there be one and only one of the elements.

The indefinite article "a" or "an" thus usually means "at least one".

Brief description of the figures

Figure 1. Design of a superselective surface coating for hmDNA enrichment. A) Schematic illustrations displaying the crucial role of the MBD2 surface receptor density for the superselective binding of methylated DNA at the surface. Low, medium, and high MBD2 surface receptor densities are illustrated and the subsequent degree of interaction with methylated DNA (DNA with red hexagons to represent C*pG) and non-methylated DNA are shown. B) The DNA surface coverage on the MBD2-modified surface is displayed as a function of the MBD2 surface receptor density. The optimal MBD2 surface receptor density range to achieve optimal selectivity for enrichment of methylated DNA is indicated.

Figure 2. Surface chemistry employed to control the degree of His1oMBD2 immobilization on a gold surface modified with a SAM of functionalized ethylene glycol-alkanethiols. The mixed ethylene glycol-alkanethiol-based SAM on a gold surface with azide and hydroxyl functional groups is shown. Azide-functionalized thiols are modified with a linker molecule bearing

DBCO and NiNTA functional groups. The Hiss tag of MBD2 is used to achieve immobilization of the protein on the surface.

Figure 3. SDS-PAGE gel electrophoresis of the HisisMBD2 after isolation from E. coli

Rosetta DE3 cells using NiNTA column chromatography and SEC. The gel was stained with

Coomassie Blue prior to imaging. In the SDS-PAGE gel the Ladder (L), cell lysate (Lys), flow through (FT), wash of the NiNTA column {W1), elution of the MBD2 with imidazole (E1}, and the isolated His1oMBD2 after SEC (MBD2) is visible. The characteristic MBD2 band is visible at the expected molecular weight (= 32 kDa).

Figure 4. A) In situ monitoring of the His:5MBD2 immobilization process by QCM. Prior to

QCM measurement, a mixed SAM of hydroxyl and azide ethylene glycol-alkanethiols (95:5) was formed overnight. After obtaining a stable baseline, a solution with the DBCO-NTA linker molecule was flown over the SAM. The NTA groups were complexed by NiClz, followed by the immobilization of His:4MBD2. Washing steps with PBS, Milli-Q (H20) and immobilization buffer (IB) are indicated by the grey areas. B) Frequency shift upon MBD2 immobilization (determined after the 1.5 h washing step with IB) as a function of the percentage of azide- functionalized thiol in the underlying SAM. Each datapoint was obtained from at least 2 experiments. An exponential trendline was fitted to the data with R? > 0.99 to guide the eye.

Figure 5. A) MBD2 surface receptor density (ng/cm?) as a function of the percentage of azide-functionalized thiol in the underlying SAM. The surface receptor density was calculated from frequency shift data using the Sauerbry equation. B) MBD2 surface receptor density (molecules/cm?) as a function of the percentage of azide-functionalized thiol in the underlying

SAM. Each datapoint was obtained from at least 2 experiments. An exponential trendline was fitted to the data with R2 > 0.99 to guide the eye.

Figure 6. A) The model DNA targets, based on the Mal gene, ?>® are 40 bp or 90 bp in length and consist of 1 (40 bp), 2, 3, 4 or 5 C(*)pGs (90 bp). Location of cytosines in C(*)pG is displayed here by the red circles with nucleotide numbers in the 5’ — 3’ direction. The methylation-dependent restriction endonucleases Hhal and HPall recognition sites are indicated by the blue dashed lines. B) Electrophoresis gel images of Mal3C(*)pG treated with

Hhal and HPall, showing the Ladder (L), and Hhal and HPall-digested Mal3C*pG (1) and

Mal3CpG (2). The band at 15 bp in each lane is a marker.

Figure 7. A) Location of cytosines in C(*)pGs on the 90 bp long used model DNA target

Mal3C(*)pG in the 5 till 3’ direction. Location of the recognition sequence of methyl sensitive restriction enzymes Hhal and Hpall on Mal3C(*)pG are indicated with the vertical dashed line.

The primer binding region for qPCR amplification is indicated by dashed line forming an open sqaure. B) Gel electrophoresis image of amplified Mal3C*pG. To visualize the DNA the SYBR safe DNA stain was used. In the gel image the ladder (L), no template control (1) and amplified Mal3C*pG after reacting Mal3C*pG with Hhal and Hpall (2) is visible. The Mal 3C*pG band is visible at the expected DNA length of 90 bp.

Figure 8. Monitoring of the hypermethylated enrichment process from a DNA mixture measured with QCM. The DNA mixture consisted out a 1:100 ratio between Mal3C*pG versus Mal3CpG. The total DNA concentration of the DNA mixture was 500 nM. The QCM gold chip was modified with the SAM contained 8% azide-functionalized thiols prior to the

QCM measurement. In consecutive performed steps the DNA capture device is formed: binding of the linker molecule, activation of the NTA groups with NiCI2 and His10MBD2 immobilization. Washing steps with PBS, Milli-Q water (H20}, immobilization buffer (IB) and binding buffer (BB) are indicated by the grey areas.

Figure 9. Degree of DNA binding as function of the MBD2 surface receptor densities upon the enrichment of the DNA mixtures monitored by QCM. The DNA mixture consist out of 1:100 ratio between Mal3C*pG versus Mal3CpG with a total DNA concentration of 500 nM.

The total DNA binding time was 60 min.

Figure 10. Relationship between the Mal3C*pG versus Mal3CpG ratio in a DNA mixture and the qPCR Ct value. The total DNA concentration used in each DNA sample was 0.5 pg/uL.

Prior to gPCR analysis were the DNA samples subjected to methyl sensitive restriction enzyme digest with Hhal and Hpall.

Figure 11. A) Amount of DNA eluted from MBD2 modified surfaces as function of the percentage of azide-functionalized thiols in the SAM after the enrichment of the DNA mixture.

DNA mixture with a 1:100 ratio between Mal3C*pG versus Mal3CpG and a total DNA concentration of 500 nM was enriched for 60 min.

The amount of eluted DNA was determined with the fluorescent double stranded DNA binding dye.

B) Ct value of enriched DNA mixtures as function of the degree of azide functionalized thiols in the SAM.

The eluted DNA mixture was subjected to Hhal and Hpall restriction enzyme digest prior to qPCR analysis.

EXAMPLES

The following Examples illustrate the different embodiments of the invention.

Example 1

Example 1 shows the development of a superselective, multivalent, MBD2-coated platform to improve the selectivity of hmDNA enrichment. Adopting the findings shown herein into hmDNA enrichment of clinical samples has the potential to become more selective than current MBD2-based methods and, therefore, to improve cancer diagnostics.

Materials & Methods

Chemicals 30% Acrylamide/Bis solution 37 .5:1, gravity flow columns, Experion DNA chips, Experion™

DNA 1K Reagents and Supplies, 4x Laemmli Sample were obtained from BioRad.

Coomassie brilliant blue (Coomassie blue), NaCl, B-mercaptoethanol, dimethyl sulfoxide 299.7 (DMSO), EDTA, MgCl. = 98.0 % (MgCl), LB Broth, lysozyme from chicken egg white (290 % lysozyme), phenylmethanesulfonyl fluoride (PMSF), ribonuclease A from bovine pancreas (RNAse), deoxyribonuclease | from bovine pancreas (DNAse), kanamycin sulfate from Streptomyces kanamyceticus, nickel(lly chloride hexahydrate, Na, Na-bis(carboxymethyl)-

L-lysine hydrate = 97.0%, phosphate-buffered saline (PBS) tablets for 10 mM solution, H2SO4 95-97 %, 0.2 ym membrane filter and sodium dodecyl sulfate = 99% (SDS) were purchased from Sigma Aldrich. Tris(hydroxymethyl)aminomethane (TRIS), isopropyl-B-D- thiogalactopyranoside 2 99 % (IPTG), H2O2 33 % and ethanol absolute were bought from

VWR. Imidazole 99% and Triton X-100 were obtained from ACROS Organics. Ni-NTA- agarose beads were purchased from Protino®. PD-10 Sephadex™ desalting columns were obtained from GE Health. DBCO-PEG,-NHS ester was purchased from Click Chemistry

Tools. All the DNA sequences were bought from Eurofins Genomics. Nuclease-free water,

CpG methyltransferase (M.Sssl), 1X NEBuffer™ 2, S-adenosylmethionine, Cutsmart buffer,

Hhal, HPall and high fidelity polymerase Q5 were purchased from New England Biolabs.

HSC++(EG)s-OH and HSC4:(EG)s-N3 were bought from Prochimia. QCM gold-coated chips (QS-QSX301) were obtained from Quantum Design GmbH.

Cloning and transformation

The used bacterial strains and plasmids are displayed in Table 1 and all the required molecular reagents to enable cloning and transformation were obtained from New England

Biolabs. The His:41MBD2 gene was cloned in a pET-15b vector by Eurofins genomics. The

His1oMBD2 gene was amplified by PCR using the primers displayed in Table 2 and high fidelity polymerase Q5 and cloned into pRSFDuet vector (Novagen) using the SLIC method between Ndel-Kpnl (MCS2) restriction sites as described before (Jeong et al. Appl. Environ.

Microbiol. 2012, 78 (15), 5440-5443). The amplified DNA obtained from PCR was purified using Macherey Nagel PCR clean-up kit. Then E. coli NovaBlue competent cells were transformed with the SLIC product. Plasmids were purified using Qiagen miniprep kit. The plasmids were sequenced by Eurofins Genomics to confirm the sequence. Competent cells of

E. coli Rosetta strain were transformed with the plasmid. The plasmids used are shown in

Table 3.

Table 1. Bacterial strains used in this study.

E. coli strain Genotype/Characteristics Origin

Rosetta(DE3) F-ompT hsdSB(rB- mB-) gal dem (DE3) pRARE (CamR) Novagen

NovaBlue endA1 hsdR17 (rK12— mK12+) supE44 thi-1 recAl gyrA96 Novagen relAl lac F'[proA+B+ lacigZAM15::Tn10} (TetR)

Table 2. Primer sequence used to amplify the His:oMBD2 gene by PCR.

Primer 5=>3 Characteristics

HisioMBD2 fwd | AGAAGGAGATATACATATGCACCATCACCATCATCAC | For Ndel site in pDuet

CATCATCACCACGATTGTCCTGCGTTG MCS2

Hisi,MBD2 rev ACCAGACTCGAGGGTACCTTAAGCTTCATCACCGCT For Kpnl site in pDuet

MCS2

Table 3. Plasmid type used in this study.

Plasmids Genotype/Characteristics Origin pET-15b Novagen pRSFDuet Km}, 2 MCS, Pry, Ori RSF 1030 Novagen pRSFDuet pRSFDuet carrying MBD2 gene with His15-tag at N-terminus This study

His10MBD2 cloned in MCS 2

His:,MBD2 production and purification

The E. coli bacteria were grown up to an optical density (OD) of 0.5, at A = 600 nm at 37°C in

LB medium with 30 pg/ml of Kanamycin. The culture was cooled down until 17 °C, followed by the expression of the His:4MBD2 protein with 1 mM of IPTG for 15 h at 17°C while stirring at 210 rpm. The culture was centrifuged (Allegra™ 25R) at 5000 rpm for 15 min at 4 °C to sediment the bacteria. The bacteria were lysed with sonication in a lysis buffer of 50mM Tris-

HCI pH 7.2, 300 mM NaCl, 30mM imidazole, 0,1% B-mercaptoethanol, 1 mM EDTA, 20 mM

MgCl, 1 mM PMSF, 0,5 mg/mL lysozyme, 20 ug/mL DNAse, 20 pg/mL RNAse A. Sonication (Fisherbrand™ 120) was performed on ice twice for 30 s with a waiting step in between of 2 min. The sonicated sample was centrifuged at 3100 rpm for 15 min at 4 °C, and the supernatant was then centrifuged at 40000 rpm for 60 min at 4°C (WX Ultra 90, Thermo

Scientific). The supernatant was loaded on a NiNTA column and incubated for 30 min while shaking at 4°C . The column was washed with 25 mL washing buffer (50 mM Tris pH 7.2, 300 mM NaCl, 30 mM imidazole, 0.1% B-mercaptoethanol). His:4oMBD2 was removed from the

NIiNTA column with an elution buffer (50 mM Tris pH 7.2, 300 mM NaCl, 650 mM imidazole, 0.1% B-mercaptoethanol). Directly afterwards, the His:oMBD2 sample was purified with a

PD10 column and eluted in the IB buffer (50 mM TRIS pH 7.2, 300 mM NaCl, 0.1% B- mercaptoethanol). The His:1MBD2 sample was aliquoted and stored at -80°C until use.

SDS-Page gel electrophoresis

Protein fractions were mixed with 4x Laemmli Sample Buffer + 0.1% B-mercaptoethanol at a 1:4 ratio and analyzed on 15% polyacrylamide gels with gel electrophoresis (BioRad) in a running buffer of 25 mM Tris, 192 mM glycine and 0.1% SDS. The separated proteins were stained using Coomassie brilliant blue solution consisting out of 10% acetic acid, 40% ethanol, 50% Milli-Q water and 2 g/L Coomassie R250. The stained gel was then unstained with a solution of 10% acetic acid, 40% ethanol and 50% Milli-Q water followed by imaging.

DNA methylation

DNA methylation was performed by mixing 4 ug DNA with 8 units of M.Sss/ enzyme, S- adenosylmethionine at a concentration of 600 uM and 2.5 yL NEBuffer™ 2. The reaction volume was increased till 25 pL with nuclease free water. The reaction was performed at 37 °C for 15 hours in a T100 thermocycler (BioRad).

Methyl-sensitive restriction enzyme digest 10 ng of DNA was mixed with 1 pL of Cutsmart buffer, 5 units Hhal and 5 units HPall. The total reaction volume was increased till 10 uL using nuclease free water. Digestion was performed for 15 hours at 37 °C using a thermocycler. Afterwards were the enzymes deactivated by a heat treatment of 85 °C for 20 min.

DNA electrophoresis

DNA samples were analyzed using DNA Experion chips and Experion™ DNA 1K Reagents and Supplies according the instruction of the manufactures on an automated electrophoresis system (ExperionTM, BioRad).

Synthesis of linker molecule

DBCO-PEG.-NHS was dissolved in DMSO at 250 mM and directly aliquoted and stored at - 18 °C until use. Na,Na-Bis(carboxymethyl)-L-lysine hydrate was dissolved in PBS pH 7.4 at 1 mM before the start of the reaction. The dissolved DBCO-OEG4-NHS was added to the

Na, Na-bis(carboxymethyl)-L-lysine hydrate solution at a final concentration of 0.1 mM. The reaction components were stirred overnight at 180 rpm to ensure completion of reaction.

SAM (self-assembled monolayer) formation

Gold QCM chips were cleaned in a piranha solution for 10 s followed by immersion of the chips in Milli-Q water. Afterwards, the chips were rinsed extensively with ethanol, Milli-Q water and ethanol, followed by drying using N». The gold chips were then oxidized using UV- ozone (BioForce, Nanosciences) for 30 min. A thiol solution was prepared using HSC11(EG)s-

OH and HSC+:(EG)s-N3 at the desired ratio between the two components in ethanol with a total thiol concentration of 2 mM. The oxidized gold chips were completely immersed in the thiol solution overnight to form the SAM. After the SAM formation were the QCM chips rinsed extensively with ethanol, Milli-Q water and ethanol and dried in a stream of Na.

Quartz crystal microbalance (QCM)

Gold-coated QCM chips modified with the SAM were placed in the QCM Analyzer equipped with 4 individually addressable flow cells (Biolin Scientific). A flow rate of 30 yL/min was set during the experiments with a peristaltic pump (Biolin Scientific). All solution were filtered with a 0.2 um filter prior to use. Frequency and dissipation values used in this study are normalized for the used overtone. After each QCM experiment was the system cleaned with a 15 min washing step with 1% SDS solution continued by Milli-Q water.

DNA binding to MBD2 surfaces

After SAM formation on the gold QCM chip was the linker molecule at a concentration of 0.1 mM in PBS pH 7.4 flushed over the surface for 1.5 h followed by the addition of 25 mM NiCl, in Milli-Q water for 10 min, a washing step with Milli-Q for 5 min and IB until a constant QCM frequency change over time was established. The washing step with Milli-Q was used to prevent the reduction of NiClz by B-mercaptoethanol present in the IB. His:sMBD2 dissolved in IB at a concentration of 1 uM was added until the frequency change stabilized. After MBD2 immobilization, a washing step with IB for 1.5 h was applied followed by flushing with BB until stable frequency change over time. BB contains 50 mM TRIS, 350 mM NaCl and 0.1% Triton

X-100. Then 500 nM DNA dissolved in BB was added for 30 min followed by a washing step with BB. The MBD2 surface receptor density and degree of DNA binding was determined using the normalized frequency of the 5" overtone (fs).

Results

Figure 1 shows the superselective hmDNA enrichment platform that was developed by tuning the MBD2 surface receptor density.

As can be seen in Figure 1A, the avidity increases at higher degrees of DNA methylation, upon binding of a DNA sequence with a specific number of C(*)pGs at the MBD2-modified surface, thereby decreasing the threshold receptor density required for efficient DNA surface binding.

As can be seen in Figure 1B, the inventors show that there is an optimal density range at medium MBD2 surface receptor densities at which the MBD2 density is only sufficient for binding methylated DNA, while non-methylated DNA remains unbound. Moreover, they show that control over the MBD2 surface receptor density is crucial for superselective hmDNA enrichment. The MBD2 threshold density is defined as the minimal MBD2 density required for significant DNA binding. If the MBD2 density is too low, the possible number of interaction pairs between MBD2 and C(*)pG of DNA that can be formed is limited. In this situation the avidity is insufficient for the surface binding of both types of DNA. Increasing the MBD2 density results in an increase of the number of MBD2-C(*)pG interaction pairs and, as a result, the avidity increases. The avidity increases faster for methylated DNA compared to non-methylated DNA because of the higher intrinsic affinity of MBD2 for C*pG compared to

CpG. When the MBD2 density is increased further, both methylated and non-methylated DNA can bind strongly.

Figure 2 shows how the MBD2 surface receptor density is controlled by employing a mixed

SAM on a gold-coated surface of two ethylene glycol-alkanethiols; one with a hydroxyl and the other with an azide end groups. Through this method, the MBD2 density also controls the total affinity of any bound DNA, and this control is important to finetune the efficacy of the surface as a capture layer. The hydroxyl thiol is the major component and is used to create an anti-fouling surface (e.g. Estirado et al. JACS 2019, 141, 18030-18037), while the minor, azide thiol is used to enable MBD2 immobilization. A linker molecule, bearing on one end a dibenzocyclooctyne (DBCO) moiety and on the other end a nitrilotriacetic acid (NTA) functional group, was reacted at the azide groups of the SAM by employing the catalyst-free click chemistry reaction between the DBCO and azide functional groups (Kuzmin et al.

Bioconjug. Chem. 2010, 21 (11), 2076-2085). The NTA functional groups were subsequently complexed by Ni?* ions to form NiNTA moieties. Finally, the MBD2 protein bearing a histidine- 10 (Histo) tag at the N-terminus (His:i9MBD2)} was immobilized at the surface by interacting with the NINTA complexes. An individual NiNTA moiety has the possibility to interact with two histidines with a Ky of 14 uM (Lata et al. J. Am. Chem. Soc. 2005, 127 (29), 10205-10215).

Therefore, the number of NINTA moieties interacting with the Histo tag is maximally 5. The

MBD2 surface receptor density was varied by tuning the stoichiometric ratio between the hydroxyl and azide-functionalized thiols in the SAM, leading to controlled variation of the density of NINTA groups as well as that of the MBD2s binding to these groups.

Figure 3 shows the successful production of MBD2 as confirmed by SDS-PAGE gel electrophoresis. His:1,MBD2 was produced using E. coli Rosetta (DE3) competent cells. The genetic information for the protein was based on the work of Bird ef al. (Hendrich et al. Mol.

Cell. Biol. 1998, 18 (11), 6538-6547). The protein was isolated from lysed cells using NINTA affinity column chromatography. The HissMBD2 protein was eluted from the NiNTA column by employing an imidazole wash step. Thereafter, the His oMBD2 protein was purified by size exclusion chromatography (SEC) and eluted in the immobilization buffer (IB). The final purity of the His1oMBD2 (molecular weight = 32 kDa) was confirmed by the presence of a single strong band at the expected molecular weight upon SDS-PAGE gel electrophoresis, thereby confirming the purity of the isolated His1oMBD2 sample.

Figure 4 shows the in situ monitoring of the His:cMBD2 immobilization process by QCM.

SAMs on gold substrates for quartz crystal microbalance (QCM) analysis were made by overnight immersion in mixtures of azide (minor) and hydroxyl-functionalized (major) ethylene glycol-alkanethiols in varying ratios. After mounting the sample inside a QCM chamber, all subsequent steps to bind MBD2 were monitored in situ by QCM.

Figure 4A shows a typical example of MBD2 immobilization with 5% azide in the thiol mixture.

After obtaining a stable baseline in buffer, a solution with the linker molecule bearing the

DBCO and NTA functional groups was flown over the SAM substrate. After 1.5 h a stable baseline was obtained, indicating completion of the reaction. Next, the NTA functional groups were complexed with Ni?* by flushing a solution of NiCl2 over the chip, followed by a washing step with water. Thereafter, the His:4MBD2 protein was immobilized on the surface. Upon the immobilization of MBD2, a typical binding curve was observed, with an initially rapid frequency decrease which leveled off quickly due to surface saturation. The total frequency shift (Af) of irreversibly immobilized MBD2 is 13 Hz for 8% azide, which has been determined by comparing the baselines in IB prior to and after the addition of MBD2.

Figure 4B shows that the MBD2 surface receptor density is directly related to the percentage of azide in the underlying SAM. In comparison, a densely packed streptavidin layer typically shows a frequency shift of around 25 Hz (Hamming et al. ACS Appl. Mater. Interfaces 2021, 13 (48), 58114-58123). Even though streptavidin and MBD2 have different molecular weights and their degrees of hydration (which affects QCM frequencies) may differ as well, we regard the here observed MBD2 frequency values ranging from 35 to 55 Hz as indicative of dense packing because the frequency values stabilizes.

When using between 2% and 10% of azide in the SAM, the amount of immobilized MBD2 increases linearly as a function of the amount of azide in the SAM. The coverage levels off above 10% azide in the SAM, and reaches a maximally packed MBD2 surface density at approximately 40% azide in the SAM. The relationship between the stoichiometric ratio of functional groups in the SAM and the degree of protein immobilization is in agreement with previously reported protein immobilization studies on thiol-based SAM systems (e.g. Ostuni et al. Langmuir 2003, 19 (1), 1861-1872). Overall, it is concluded that the MBD2 surface receptor density can be tuned by the stoichiometric ratio of hydroxyl and azide-functionalized ethylene glycol-alkanethiols in the SAM.

Figure 5 shows the MBD2 surface receptor density as a function of the percentage of azide- functionalized thiol in the underlying SAM. In Figure 5A, the MBD2 surfac density is expressed as ng/cm?. In Figure 5A, the MBD2 surfac density is expressed as number of molecules/cm?. It is considered that the measured MBD2 frequency shifts are proportional to the MBD2 surface coverage. The surface receptor density was calculated from the frequency shift data (as shown in Figure 4B) using the Sauerbry equation, considering that 55% of the mass is bound to water. The QCM frequency shifts, Af, of MBD2 at different azide fractions in the SAM then become a mass comparison measure according to the Sauerbrey equation (Sauerbrey et al. Zeitschrift fur Phys. 1959, 155 (2), 206-222). This comparison seems valid as the trends for different overtones vary only marginally at both low and high MBD2 surface receptor densities, and the dissipation changes (AD) were lower than 2 x 108. This was established by the inventors by monitoring frequency and dissipation over time upon

His10MBD2 immobilization at overtones 3, 5, 7 and 9 using a SAM with 5% azide or 50% azide-functionalized thiols in the SAM. The Af of MBD2 can thus be used to express the

MBD2 surface receptor density. Potentially, the frequency shifts can be converted to absolute mass, but this requires making assumptions about the degree of hydration of the protein at the monolayer surface.

Figure 6A shows the binding response of DNA with varying degrees of CpG methylation on

MBD2-modified surfaces was studied using a model DNA target sequence based on the Mal gene (e.g. Overmeer et al. J. Pathol. 2009, 219 (July), 327-336). In cancer cells, each of the 5 CpGs of the natural Mal gene is methylated. The selected model DNA target is 40 or 90 bp long and contains up to 5 CpGs. To study the effect of different numbers of C(*)pG sites on binding to the MBD2 surfaces, the model DNA target was used with numbers of C(*)pGs ranging between 0 and 5 on the model DNA target. The used DNA sequences in this study are abbreviated as MalxC(*)pG, where x represents the number of C(*)pGs in the model DNA target. Varying the number of C(*)pG sites on the model DNA target was used, instead of using very different sequences, to minimize the effect of altering the binding kinetics upon the interaction of different DNA sequences with MBD2, as was reported by Fraga et al. (Nucleic

Acids Res. 2003, 31 (6), 1765-1774). Mal2C(*)pG was used in two different versions: the

C(*)pGs were located either close to or far away from each other with 22 or 66 bp separation between the C(*)pGs, respectively, and these were abbreviated as Mal2C(*)pG-far/close, respectively. For the Mal1C(*)pG, a length of only 40 bp was taken, but it is assumed this has little effect on the affinity since only monovalent binding is possible in this case. The location of the C(*)pGs in Mal4C(*)pG is different compared to the model DNA target in order to evaluate the role of the C(*)pG location in a DNA sequence.

Figure 6B shows electrophoresis gel images of Mal3C(*)pG treated with Hhal and HPall, showing the Ladder (L), and Hhal and HPall-digested Mal3C*pG (1) and Mal3CpG (2). The

CpGs of all MalxCpG constructs were methylated to C*pGs using the M.Sssl enzyme in the presence of S-adenosylmethionine. As a typical example, successful methylation of the

Mal3CpG sequence was confirmed by the blockage of digestion upon reacting the methylated sequence with the methylation-dependent restriction endonucleases Hhal and HPAII, as was characterized by the gel electrophoresis. Hhal and HPAII can digest their recognition sequences, 5-GCGC-3 and 5'-CCGG-3', respectively, in the absence of CpG methylation.

Both Hhal and HPAII have one recognition site on Mal3CpG, resulting in the reduction of the sequence length of 24 bp in case of absence in CpG methylation. Electrophoresis showed retention of the full length for Hhal and HPAII-treated Mal3C*pG, while full digestion into the 66 bp product was observed for Mal3CpG. These results therefore confirm the successful methylation of the former sequence.

Example 2

Example 2 shows the selectivity of hypermethylated DNA enrichment on the multivalent hypermethylated DNA binding platform from DNA mixtures containing both hypermethylated

DNA and non-methylated DNA.

Materials and methods

The MBD2 protein production and surface immobilization with MBD2 surface receptor density control was performed according to Example 1. A Teflon flow cell was used to determine the hypermethylated DNA enrichment selectivity. Throughout this study, a flow rate of 30 uL/min was used for all the step during hypermethylated DNA enrichment with a peristaltic pump (Biolin Scientific). After MBD2 immobilization and washing with immobilization buffer (IB), the system flushed with binding buffer (BB) for 30 min. BB contains 50 mM TRIS, 350 mM NaCl and 0.1% Triton X-100. Then the DNA mixtures were enriched using the MBD2 modified surfaces. The DNA mixture used contained Mal3C*pG versus Mal3CpG (Table 4) at a 1:100 ratio with a total DNA concentration of 500 nM dissolved in BB. Then a washing step of 90 min was applied with BB followed by elution of surface bounded DNA for 30 min using ammonia hydroxide pH 11.3 buffer. The eluted DNA fraction was collected followed by determination of the total DNA concentration using a fluorescent DNA binding dye (Next generation dsDNA quantification kits, AccuBlue®). The enriched DNA mixtures were then subjected to methyl sensitive restriction enzyme digest with 5 units Hhal and 5 units Hpall.

Upon the restriction enzyme digest 3 ng of DNA was used which was mixed with 1 pL of rCutsmart buffer in a total reaction volume of 10 pL. Digestion was performed for 15 hours at 37 °C using a thermocycler. Afterwards, the enzymes were deactivated by a heat treatment of 85 °C for 20 min. In the next step, the restriction enzyme treated enriched DNA mixtures were subjected to qPCR analysis using primers (300 nM, Table 5) and SsoAdvance Universal

SYBR® Green supermix. Denaturation of the DNA samples took place in the first step for 1 min at 98 °C, followed by a block of two step which was repeated for 39 times; 10 s at 98°C and 30 s at 64 °C followed by determining the fluorescence intensity of SYBR green. Then, 1 min at 72 °C and 5 s at 65 °C. Followed by determining the fluorescence intensity of SYBR green.

Table 4. DNA sequences of the used model DNA targets in the 5’ — 3'direction. Location of the C(*)pGs are highlighted. a! ATCC EC RA ARCA Sa AGT ACA RTC

Mal3C(*)pG COCERACCRGGOCTEGCTCAGTOUAGC CECRNGEGLCACHD

GCCTGOCCCTTCESGCTCCACTGAGCCAGGCCTGSTTESGGCCTGGGAT

CCTGCACCTGGCCTCTCTGCTTGGTGCTGCGCATCTGCCTG

Table 5. DNA sequences of the used primers in the 5’ — 3’direction.

FWD ee —

REV GCCTGCCCCTTCCG

Results

The hypermethylated DNA enrichment selectivity from enriched DNA mixtures using MBD2 based affinity chromatography both methyl sensitive restriction enzyme digest and quantitative PCR (qPCR) analysis is performed. First, hypermethylated DNA enrichment of the DNA mixture is applied on MBD2 modified surface. The DNA contains both methylated

DNA and non-methylated DNA to llustre a liquid biopsy sample. In this DNA binding step,

DNA binds to the MBD2 modified surface. The surface bounded DNA is removed from the surface in the elution step using an ammonia hydroxide pH 11.5 solution. The pH of the eluent is above the isoelectric point of MBD2 (Hendrich et al. Mol. Cell. Biol. 1998, 18 (11), 6538-6547), thus causing alternation in the three-dimensional structure of MBD2. As a consequence, the non-covalent interactions between MBD2 and C(*)pG will be disrupted causing elution of surface bounded DNA. The eluted DNA is the DNA mixture enriched towards methylated. To assess the ratio between hypermethylated versus non-methylated

DNA in the enriched DNA mixture, the DNA mixture is subjected to restriction enzyme digest with Hpall and Hhal. The non-methylated DNA becomes digested, while the methylated DNA is protected for digestion due to CpG methylation. In the last step, the restriction enzyme treated enriched DNA mixture is amplified by qPCR at a constant DNA concentration. The number of cycles needed to reached the qPCR threshold (C: value) is directly dependent on the ratio between methylated DNA versus non-methylated in the enriched DNA sample. In case the methylated DNA amount is higher in comparison to the non-methylated DNA, this causes a lower degree of restriction enzyme digest and, therefore, results into a lower Ci value. On the other hand, if the enriched DNA mixture contains more non-methylated DNA, the C: value increases as well. This method thus enables to determine quantitatively the selectivity of the methylated DNA enrichment process.

The methylated DNA enrichment was evaluated at a gold surface modified with the MBD2 protein using a ethylene glycol-alkanethiols based self-assembled monolayer (SAM). The

SAM consist out of hydroxyl and azide functionalized thiols. Prior to MBD2 immobilization, a linker molecule bearing a dibenzocyclooctyne (DBCO) and nitrilotriacetic acid (NTA) functional group is bounded to the azide functionalized thiols. The linker molecule is covalently bounded to the azide functional groups by the catalyst free catalyst-free click chemistry reaction between DBCO and azide moieties (Kuzmin et al. Bioconjug. Chem. 2010, 21 (11), 2076-2085). Afterwards, the NTA moieties are transformed into nickel NTA (NiNTA) by flushing nickel chloride (NiCl.) over the surface. The surface tethered NINTA moieties are then used to enable the immobilization of the histidine 10 tagged MBD2 protein (His1sMBD2) (Knecht et al. J. Mol. Recognit. 2009, 22 (4), 270-279.).

Figure 7A shows the 90 bp long sequence used, and which originates from the Mal gene found to be hypermethylated in cancer cells (e.g. Overmeer et al. . J. Pathol. 2009, 219 (July), 327-336). The DNA sequence length used is in agreement to main fragment size of DNA in urine liquid biopsies (Burnham et al. Nat. Commun. 2018, 9 (1), 1-10). The DNA sequence used in this study is the model DNA target and contains 3 CpGs, which either is used in its methylated or non-methylated form. The methylated and non-methylated form of the model

DNA target is abbreviated here as Mal3C*pG or Mal3CpG, respectively. The model DNA target possess a recognition site for the methyl sensitive restriction enzymes Hhal and HPall at the first and third CpG in the 5'- 3 direction, which both become digested in absence of

CpG methylation. In case of Mal3CpG restriction enzyme digest the primer cannot bind anymore, thus preventing PCR amplification.

Figure 7B shows that Mal3C*pG was successfully amplified with qPCR after reacting

Mal3C*pG with Hhal and Hpall, as confirmed by gel electrophoresis. Upon amplification of

Mal3C*pG the methylation status is lost, as non-methylated cytosines are used during the amplification. Only one band is visible in the gel for the Hhal and HPall treated Mal3C*pG proving the absence of any formed side products. The sequence length of the amplified DNA sequence is close 90 bp, which matches the expected sequence length based on the primer design. Thus, confirming successful and specific amplification of Mal3C*pG.

Figure 8 shows the monitoring of the hypermethylated enrichment process from a DNA mixture measured with quartz crystal microbalance (QCM). Before the start of the QCM measurement, a thiol based SAM was formed containing 9% azides on the gold surfaces.

Then, the linker molecule was reacted to the surface. As a result of the binding event, the frequency was reduced until completion of the reaction. In the next step, the NTA moieties were transformed into NiNTA by flushing with NiCl: solution. The MBD2 protein was successful immobilized on the surface due the interaction between the Hiss, tag of MBD2 with the NiINTA moieties on the surface. After the MBD2 immobilization, a washing step with immobilization buffer (IB} was applied to remove loosely bounded MBD2. The degree of

His1sMBD2 immobilization is determine by measuring the frequency shift (Af) after the MBD2 immobilization followed by the washing step with immobilization buffer (IB) and before MBD2 immobilization while flushing with IB. The Af observed equals to 23 Hz, which can be converted to the MBD2 surface receptor densities using the Sauerbrey equation. Additionally, the MBD2 surface receptor density is directly depending on the amount of azide functionalized thiols in the SAM. The MBD2 modified surface was subsequently used to monitor the enrichment process of a DNA mixture. The DNA mixture used in this study consist out of a 1:100 ratio between Mal3C*pG versus Mal3CpG, which is typical methylation levels also found in genomic DNA {e.g. Zeng et al. Epigenetics 2014, 9 (4), 483-491}, thus representing DNA originating from a liquid biopsy. Upon the enrichment a relatively small but constant decrease in frequency is observed over time. After the DNA was bounded towards the MBD2 surface, a washing step is applied to remove weakly attached DNA from the surface. The amount of DNA bound to the MBD2 surface is determined after the washing step with binding buffer (BB) and equals to a Af of 1.6 Hz.

Figure 9 shows the degree of DNA binding as function of the MBD2 surface receptor densities upon the enrichment of the DNA mixtures monitored by QCM. The degree of DNA binding on MBD2 surfaces was determined for MBD2 surface receptor densities between 12

Hz and 30 Hz. The degree of DNA binding was constant over the measured MBD2 density range indicating the absence of binding limitations. The average degree of DNA binding equals to 1.4 Hz over the measured MBD2 density range, which equals to an average DNA mass binding of 7.4 ng/cm? using the Sauerbrey equation and assuming that 70% of the adsorbed mass is due to the hydration of DNA.

Surface bounded DNA after the enrichment is removed from the surface using the eluent ammonia hydroxide pH 11.5. The DNA concentration of the eluted DNA sample is determined by a DNA binding dye that becomes fluorescent upon the binding to double stranded DNA.

This concentration determination step confirmed similar DNA input concentrations upon the determination of the ratio Mal3C*pG versus Mal3CpG of enriched DNA mixtures.

Figure 10 shows the the ratio between Mal3C*pG versus Mal3CpG in a DNA mixture quantitatively determined with qPCR after reacting the DNA mixtures with the methyl sensitive restriction enzymes Hhal and Hpall. DNA mixtures with 1:1, 1:10, 1:100 and 1:1000 ratio between Mal3C*pG versus Mal3CpG were characterized. Also samples containing only

Mal3C*pG and Mal3CpG were analyzed. After the restriction enzyme digest step, the DNA samples were analyzed with qPCR at constant total DNA input concentration. A clear dependency between the C: values and amount of Mal3C*pG in the DNA mixture was observed. The lowest C values was obtained for the samples containing only Mal3C*pG (1:0). On the other hand, the highest C: values was observed in case only Mal3CpG is present (0:1) due to that Mal3CpG is not protected for restriction enzyme digest by CpG methylation. Restriction enzyme treated Mal3CpG does not result into complete prevention of

Mal3CpG amplification. This is attributed to the low used DNA concentration upon the digest and limited number of methyl sensitive restriction enzymes recognition sites per sequence, both reducing the restriction enzyme digest efficiency. The difference in C values (AC) between Mal3C*pG and Mal3CpG equals to 7 cycles. This difference equals to = 100-fold concentration difference between Mal3C*pG and Mal3CpG attributed to the DNA digestion step, assuming that a 10-fold concentration increase equals to a AC of 3.6 cycles at 100% efficiency independent of the DNA mixture composition. Furthermore, the C value is increasing linearly when the amount of Mal3C*pG is decreasing in the DNA mixture. The increase in the C value is due to the presence of more Mal3CpG, thus resulting into more digestion of Mal3CpG by the restriction enzymes. With the developed method, the DNA mixture compositions up to a 1:100 ratio between Mal3C*pG versus Mal3CpG is quantifiable.

Therefore, enabling the determination of the enrichment selectivity using DNA mixtures with a 1:100 ratio of Mal3C*pG versus Mal3CpG as input.

Figure 11 shows that the methylation level of enriched 1:100 DNA mixtures is strongly dependent on the MBD2 surface receptor density used in the multivalent DNA binding platform. The MBD2 proteins were immobilized at varying surface receptor densities on SAM modified surfaces by varying the amount of azide-functionalized thiols between 4% and 25%.

Thus, the higher the azide percentage the higher the MBD2 surface receptor density. The enrichment was performed in a Teflon based flow cell and not in QCM set-up in order to increase the chip surface area (more DNA binding). After the DNA binding step, the surface was washed with BB followed by determination of the DNA concentration using the double stranded fluorescent DNA binding dye.

Figure 11A shows that the amounts of DNA eluted from the surface is roughly equal independent on the MBD2 density and is on average 30.9 ng. The surface area of the gold surface in the Teflon flow cell equals to 3.12 cm2, resulting in an average DNA binding of 9.8 ng/cm?, which is in excellent agreement to 7.4 ng/cm? amount of DNA bounded towards

MBD2 surface obtained with QCM. Furthermore, this indicates that upon elution a significant part of the surface bounded DNA is eluted.

Figure 11B shows that the methylated DNA enrichment selectivity is dependent on the used

MBD2 surface receptor density. This was shown by subjecting the eluted DNA samples to the restriction enzyme digest followed by qPCR analysis to determine the Mal3C*pG versus

Mal3CpG ratio. The highest enrichment selectivity is achieved upon usage of 15% azides in the SAM, resulting into a Mal3C*pG versus Mal3CpG ratio of 1:4, indicating that 20% of the enriched DNA mixture is Mal3C*pG. At lower azide percentage the enrichment selectivity decrease linearly, were the lowest selectivity is achieved at 4% of azides corresponding to a 1:10 ratio (9% Mal3C*pG). The decrease in selectivity is attributed to the fact that the avidity of Mal3C*pG towards the MBD2 decreases at lower MBD2 surface receptor density, resulting into more mass transport limitations. On the other hand, also an increase in enrichment selectivity is observed above 15% of azides in the SAM. This behavior can also be attributed to mass transport limitations as at those azide percentage the corresponding MBD2 surface receptor density results into steric hindrance of the MBD2 active sites by other MBD2 proteins. Reducing mass transport limitations at azide percentage lower than 15% can likely be reduced by increasing the contact area between the DNA mixture and the MBD2 surface.

A typical way to achieve this can be to increasing mixing in the flow cells or to make use of 3D structured surface. Overall, the control over the MBD2 density is of paramount importance to achieve high hypermethylated DNA enrichment selectivities.

Example 3

Preliminary observations of the inventors indicate that the optimal percentage of azide- functionalized thiol in the underlying SAM (hence the MBD2 surface receptor density depends onthe biofluid) is dependent on the type of biofluid.

In the current example, the inventors illustrate the difference between urine and blood, which are two (most) relevant bio-fluids for which the current invention can be applied. Urine sample typically contains DNA sequences with on average 20-200 bp length, whereas a blood sample typically contains DNA sequences with larger fragment sizes ( >200 bp length). The experimental design involves using a DNA mixture composition with a 1:100 ratio between

Mal3C*pG versus Mal3CpG (using the method in Example 2), a DNA length of 90 bp (for urine) or 250 bp (for blood), and varying the % Azide functionalized thiol.

Table 6 shows that the selectivity of hmDNA enrichment in optimal for 15% azide- functionalized thiol (i.e. 250-300 ng/cm? MBD2 surface receptor density). In comparison, the selectivity of hmDNA enrichment in optimal for 10% azide-functionalized thiol (i.e. 175-225 ng/cm? MBD2 surface receptor density). The inventors consider that the type of biofluid and/or

DNA fragment length a relevant parameter to consider when fine-tuning the selectivity of hmDNA enrichment in blood. Based on the trend as shown in Table 6, the inventors consider that a MBD2 surface receptor density below 100 ng/cm? (e.g. <5% azide), or above 500 ng/cm? (e.g. >30% azide) leads to inappropriate selectivity.

Table 6. Selectivity of hmDNA enrichment in blood for various % of azide-functionalized thiol in the underlying SAM. % Azide MBD2 surface | Selectivity of Selectivity of functionalized | receptor hmDNA hmDNA thiol density enrichment in urine | enrichment in (ng/cm?) blood

SEQUENCE LISTING

<110> Universiteit Twente <120> Method for separating high-methylated DNA and low-methylated DNA <130> P35615NL00 <160> 8 <170> PatentIn version 3.5 <210> 1 <211> 411 <212> PRT <213> Homo sapiens <400> 1

Met Arg Ala His Pro Gly Gly Gly Arg Cys Cys Pro Glu Gln Glu Glu 1 5 10 15

Gly Glu Ser Ala Ala Gly Gly Ser Gly Ala Gly Gly Asp Ser Ala Ile

Glu Gln Gly Gly Gln Gly Ser Ala Leu Ala Pro Ser Pro Val Ser Gly

Val Arg Arg Glu Gly Ala Arg Gly Gly Gly Arg Gly Arg Gly Arg Trp 60

Lys Gln Ala Gly Arg Gly Gly Gly Val Cys Gly Arg Gly Arg Gly Arg 65 70 75 80

Gly Arg Gly Arg Gly Arg Gly Arg Gly Arg Gly Arg Gly Arg Gly Arg 85 90 95

Pro Pro Ser Gly Gly Ser Gly Leu Gly Gly Asp Gly Gly Gly Cys Gly 100 105 110

Gly Gly Gly Ser Gly Gly Gly Gly Ala Pro Arg Arg Glu Pro Val Pro 115 120 125

Phe Pro Ser Gly Ser Ala Gly Pro Gly Pro Arg Gly Pro Arg Ala Thr 130 135 140

Glu Ser Gly Lys Arg Met Asp Cys Pro Ala Leu Pro Pro Gly Trp Lys 145 150 155 160

Lys Glu Glu Val Ile Arg Lys Ser Gly Leu Ser Ala Gly Lys Ser Asp 165 170 175

Val Tyr Tyr Phe Ser Pro Ser Gly Lys Lys Phe Arg Ser Lys Pro Gln 180 185 190

Leu Ala Arg Tyr Leu Gly Asn Thr Val Asp Leu Ser Ser Phe Asp Phe 195 200 205

Arg Thr Gly Lys Met Met Pro Ser Lys Leu Gln Lys Asn Lys Gln Arg 210 215 220

Leu Arg Asn Asp Pro Leu Asn Gln Asn Lys Gly Lys Pro Asp Leu Asn 225 230 235 240

Thr Thr Leu Pro Ile Arg Gln Thr Ala Ser Ile Phe Lys Gln Pro Val 245 250 255

Thr Lys Val Thr Asn His Pro Ser Asn Lys Val Lys Ser Asp Pro Gln 260 265 270

Arg Met Asn Glu Gln Pro Arg Gln Leu Phe Trp Glu Lys Arg Leu Gln 275 280 285

Gly Leu Ser Ala Ser Asp Val Thr Glu Gln Ile Ile Lys Thr Met Glu 290 295 300

Leu Pro Lys Gly Leu Gln Gly Val Gly Pro Gly Ser Asn Asp Glu Thr 305 310 315 320

Leu Leu Ser Ala Val Ala Ser Ala Leu His Thr Ser Ser Ala Pro Ile 325 330 335

Thr Gly Gln Val Ser Ala Ala Val Glu Lys Asn Pro Ala Val Trp Leu 340 345 350

Asn Thr Ser Gln Pro Leu Cys Lys Ala Phe Ile Val Thr Asp Glu Asp

355 360 365

Ile Arg Lys Gln Glu Glu Arg Val Gln Gln Val Arg Lys Lys Leu Glu 370 375 380

Glu Ala Leu Met Ala Asp Ile Leu Ser Arg Ala Ala Asp Thr Glu Glu 385 390 395 400

Met Asp Ile Glu Met Asp Ser Gly Asp Glu Ala 405 410 <210> 2 <211> 819 <212> DNA <213> Based on homo sapiens <400> 2 atgcaccatc accatcatca ccatcatcac cacgattgtc ctgcgttgcc gcccggatgg 60 aaaaaagagg aagttattcg taaatctggt ctgagtgcgg gcaagtcaga tgtatattat 120 ttetcccctt cgggtaaaaa gttccgtagt aaacctcaac tcgcgcgcta ccttggcaat 180 acagtggatc tcagttcttt cgattttcgc actggaaaga tgatgccatc aaagctgcaa 240 aaaaataaac agcgcctacg caacgaccca cttaaccaaa ataaaggcaa accagattta 300 aatacaacct tacccattcg tcagactgcg tctattttta aacaaccggt caccaaagta 360 accaatcatc cgagtaacaa agttaaatcc gacccgcaac gcatgaacga gcaacCgcgg 420 cagttatttt gggagaagcg cttgcagggc ctgtcggcgt ccgatgtcac cgaacagatc 480 attaagacca tggagttgcc gaaaggcctg cagggcgttg gtccgggtag caacgacgag 540 accctgctgt cagccgtggc atccgcgctg cacaccagca gcgcaccgat tacgggtcag 600 gtgtcggctg ccgtggaaaa aaacccggcc gtttggctga acacgtcgca gccgctgtgc 660 aaagccttta tcgtcacgga cgaagatatc cgaaaacagg aagaacgtgt gcagcaggtg 720 cgtaaaaaac tggaagaagc tctgatggcc gacatactga gcagagcagc agacacggaa 780 gaaatggata tcgaaatgga tagcggtgat gaagcttaa 819 <210> 3 <211> 64 <212> DNA <213> Artificial Sequence

<220>

<223> -

<400> 3 agaaggagat atacatatgc accatcacca tcatcaccat catcaccacg attgtcctgc 60 gttg 64 <210> 4

<211> 36

<212> DNA

<213> Artificial Sequence

<220>

<223> primer

<400> 4 accagactcg agggtacctt aagcttcatc accgct 36 <210> 5

<211> 90

<212> DNA

<213> Artificial Sequence

<220>

<223> primer

<400> 5 caggcagatg cgcagcacca agcagagagg ccaggtgcag gatcccaggc ccgaaccagg 60 cctggctcag tggagccgga aggggcaggc 90 <210> 6

<211> 90

<212> DNA

<213> Artificial Sequence

<220>

<223> primer

<400> 6 gcctgcccct tccggctcca ctgagccagg cctggttcgg gcctgggatc ctgcacctgg 60 cctctctgct tggtgctgcg catctgcctg 90 <210> 7

<211> 13

<212> DNA

<213> Artificial Sequence

<220>

<223> primer

<400> 7 caggcagatg cgc 13 <210> 8

<211> 14

<212> DNA

<213> Artificial Sequence

<220>

<223> primer

<400> 8 gcctgcccct tccg 14

Claims (21)

CONCLUSIONS A method for separating high-methylated DNA and low-methylated DNA, the method comprising: a) providing a liquid sample comprising high-methylated DNA and low-methylated DNA in an amount of 1-5000 ng DNA per ml; b) contacting the sample with a surface on which MBD2 (methyl-CpG binding domain protein 2) proteins are immobilized at a density of 100-500 ng/cm? to allow formation of MBD2-highly methylated DNA complexes on said surface, thereby separating high-methylated DNA from low-methylated DNA, where the high-methylated DNA is DNA sequence(s) containing at least 2 methylated CpG sites, and where low-methylated DNA refers to DNA sequence(s) that include less methylated CpG sites compared to the highly methylated DNA. A method according to claim 1, wherein step a) provides a liquid sample that is an aqueous sample or a body fluid sample that comprises high-methylated DNA and low-methylated DNA, preferably a urine sample containing high-methylated DNA and low-methylated DNA or a blood plasma sample that includes high-methylated DNA and low-methylated DNA. A method according to any one of the preceding claims, wherein step a) provides a liquid sample comprising high-methylated DNA and low-methylated DNA, in an amount of 5-2000 ng/ml, preferably 1-1000 ng/ml. Method according to any of the preceding claims, wherein step b) brings the mixture into contact with a surface on which MBD2 proteins are immobilized at a density between 100-500, 150-400, 150-250, 200-350, preferably 250-300 ng/cm? to allow formation of MBD2-highly methylated DNA complexes on said surface, thereby separating high-methylated DNA and low-methylated DNA. A method according to any one of the preceding claims, wherein the MBD2 proteins have an amino acid sequence with at least 70, 80, 90% sequence identity to SEQ ID NO:1. Method according to any of the preceding claims, wherein the method is for determining a ratio between high-methylated DNA and low-methylated DNA, preferably by carrying out quantitative PCR (polymerase chain reaction). Method according to any of the preceding claims, comprising step c) of eluting highly methylated DNA from the surface on which MBD2 proteins are immobilized, preferably by applying a basic solution, thereby obtaining the highly methylated DNA. Method according to any of the preceding claims, wherein the surface on which MBD2 proteins are immobilized is a surface comprising - spacers with a hydroxyl group and spacers with an azide group, wherein the percentage of spacers with an azide group is between 5-20%, preferably 5 -15% or 10-20%, compared to the total number of spacers; - linkers comprising a dibenzocyclooctyne (DBCO) moiety at one end and a nitrilotriacetic acid (NTA) moiety at the other end, chelated to nickel (I), with the DBCO moiety of the linkers linked to the azide group of the spacers; - MBD2 proteins that include a histidine tag, where the histidine tag of the proteins is bound to the nitrilotriacetic acid (NTA) moiety chelated to nickel (I1} of the linkers. 9. Method according to claim 8, wherein the spacers form a thiol-based self-assembled monolayer (SAM) and/or wherein the surface is a gold surface, wherein the spacers are attached to the gold surface by a gold-sulfur bond be bound to. A method according to any one of claims 8-9, wherein the spacers with a hydroxyl group are hydroxyethylene glycol alkanethiols and/or wherein spacers with an azide group are azide-functionalized ethylene glycol alkanethiols. A method according to any one of claims 8-10, wherein the linkers are DBCO-polyethylene glycol-NTA chelated to nickel (II). 12. Surface on which MBD2 proteins are immobilized at a density between 100-500 ng/cm 2 , the surface comprising - spacers with a hydroxyl group and spacers with an azide group, the percentage of spacers with an azide group between 5-20, preferably 5 -15 or 10- 20%, compared to the total number of spacers; - linkers comprising a DBCO moiety at one end and an NTA moiety at the other end, chelated to nickel (II), the DBCO moiety of the linkers being linked to the azide group of the spacers; - MBD2 proteins that include a histidine tag, where the histidine tag of the proteins is bound to the nitrilotriacetic acid (NTA) moiety chelated to nickel (II) of the linkers. A surface on which MBD2 proteins are immobilized according to claim 12, wherein the MBD2 proteins are immobilized in a density between 100-400, 150-400, 150-250, 200-350, preferably 250-300 ng/cm2. A surface on which MBD2 proteins are immobilized according to any one of claims 12-13, wherein the spacers form a thiol-based SAM and/or wherein the surface is a gold surface, wherein the spacers are bonded to the gold surface by a gold-sulfur bond . A surface on which MBD2 proteins are immobilized according to any one of claims 12 to 14, wherein the spacers with a hydroxyl group are hydroxyethylene glycol alkanethiols and/or wherein spacers with an azide group are azide-functionalized ethylene glycol alkanethiols. A surface on which MBD2 proteins are immobilized according to any one of claims 12-15, wherein the linkers are DBCO-polyethylene glycol-NTA chelated to nickel (II). 17. Method for making a surface on which MBD2 proteins are immobilized at a density between 100-500 ng/cm 2 , the method comprising a) cleaning a gold surface, preferably by exposing the gold surface to ultraviolet light and /or ozone; b) combining the cleaned gold surface with a thiol solution comprising spacers with a hydroxyl group and spacers with an azide group, wherein the percentage of spacers with an azide group is between 5-20, preferably 5-15 or 10-20%, relative to the total number of spacers, thereby obtaining a thiol-based SAM on the gold surface; c) combining the SAM with linkers comprising a DBCO portion at one end and an NTA portion at the other end, thereby obtaining SAM comprising NTA-functionalized linkers; d) combining the SAM comprising NTA-functionalized linkers with NiCl, thereby obtaining SAM comprising NTA units chelated to nickel (II); e) combining the SAM comprising NTA moieties and chelated to nickel (11) with MBD2 proteins to thereby obtain a surface on which MBD2 proteins are immobilized at a density between 100-500 ng/cm 2 . A method according to claim 17, wherein the method is to make a surface on which MBD2 proteins are immobilized at a density between 100-400, 150-400, 150-250, 200-350, preferably 250-300 ng/ cm?. 19. Method according to any of claims 17-18, wherein in step b) the spacers form a thiol-based SAM and/or wherein the spacers are bound to the gold surface by a gold-sulfur bond. A method according to any one of claims 17-19, wherein the spacers with a hydroxyl group are hydroxyethylene glycol alkanethiols and/or wherein the spacers with an azide group are azide-functionalized ethylene glycol alkanethiols. A method according to any one of claims 17-20, wherein the linkers are DBCO-polyethylene glycol-NTA chelated to nickel (II).