Methylsorb: A Simple Method for Quantifying DNA ... - ACS Publications

Sep 16, 2014 - ... College and Cooper Roads (Building 75), Brisbane QLD 4072, Australia .... Ripon Bhattacharjee , Shunsuke Tanaka , Sofia Moriam , Mo...
0 downloads 0 Views 801KB Size
Subscriber access provided by University of Newcastle, Australia

Article

Methylsorb: A Simple Method for Quantifying DNA Methylation Using DNA-Gold Affinity Interactions Abu Ali Ibn Sina, Laura G. Carrascosa, Ramkumar Palanisamy, Sakandar Rauf, Muhammad J. A. Shiddiky, and Matt Trau Anal. Chem., Just Accepted Manuscript • Publication Date (Web): 16 Sep 2014 Downloaded from http://pubs.acs.org on September 26, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Methylsorb: A Simple Method for Quantifying DNA Methylation Using DNA-Gold Affinity Interactions

Abu Ali Ibn Sina1,†, Laura G. Carrascosa1,†,*, Ramkumar Palanisamy1, Sakandar Rauf1, Muhammad J. A. Shiddiky1* and Matt Trau1, 2,* 1

Australian Institute for Bioengineering and Nanotechnology (AIBN), Corner College and Cooper

Roads (Bldg 75), The University of Queensland, Brisbane QLD 4072, Australia 2

School of Chemistry and Molecular Biosciences, The University of Queensland, Brisbane, QLD 4072,

Australia. Tel: +61-7-33464178; Fax: +61-7-33463973

1

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 25

ABSTRACT

The analysis of DNA methylation is becoming increasingly important both in the clinic and also as a research tool to unravel key epigenetic molecular mechanisms in biology. Current methodologies for the quantification of regional DNA methylation (i.e., the average methylation over a region of DNA in the genome) are largely affected by comprehensive DNA sequencing methodologies which tend to be expensive, tedious, and time-consuming for many applications. Herein, we report an alternative DNA methylation detection method referred to as “Methylsorb”, which is based on the inherent affinity of DNA bases to the gold surface (i.e., the trend of the affinity interactions is adenine > cytosine ≥ guanine > thymine).1 Since the degree of gold-DNA affinity interaction is highly sequence-dependent, it provides a new capability to detect DNA methylation by simply monitoring the relative adsorption of bisulfite treated DNA sequences onto a gold chip. Because the selective physical adsorption of DNA fragments to gold enable a direct read-out of regional DNA methylation, the current requirement for DNA sequencing is obviated. To demonstrate the utility of this method we present data on the regional methylation status of two CpG clusters located in the EN1 and MIR200B genes in MCF7 and MDA-MB-231 cells. The methylation status of these regions was obtained from the change in relative mass on gold surface with respect to relative adsorption of an unmethylated DNA source and this was detected using surface Plasmon resonance (SPR) in a label-free and real-time manner. We anticipate that the simplicity of this method, combined with the high level of accuracy for identifying the methylation status of cytosines in DNA, could find broad application in biology and diagnostics.

2

ACS Paragon Plus Environment

Page 3 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

INTRODUCTION DNA methylation is a prevalent epigenetic alteration in cancer,2 and hence considered one of the most promising biomarkers for cancer diagnosis, prognosis and therapy monitoring.3,4 Early detection of aberrant DNA methylation offers a great potential to improve survival rates and decrease the cost of treatment by ameliorating prescription of ineffective therapies. DNA methylation strategies typically involve bisulfite5 modification of DNA to convert methylation events into base changes followed by comprehensive sequencing to read each individual base pair in the genomic region of interest.6,7,8 Over the past few years, much attention has also been focused on detecting DNA methylation from bisulfite treated DNA using sequence-sensitive methods such as methylation specific PCR (MSP),9 methylation specific high resolution (MS-HR) melting10, MALDI-TOF mass spectrometry11, combined bisulfite restriction analysis (COBRA)12,13, hyperbranched rolling circle amplification (HRCA)14 or biosensing coupled with electrochemical15, 16, 17 or optical readouts18,19,20. While most of these methodologies have significantly improved the analysis performance, their practical application is restricted due to the complicated surface and conjugation chemistries, multistep analysis procedures or the use of labels. Therefore, we believe that the development of a simple, low cost and robust detection method for DNA methylation is essential to enable the uptake and full utilization of such biomarkers in the clinic. More recently, Lin et al.,21 has reported an advanced technique, where the strength of the colloidal interaction between DNA and gold nanoparticles has been used to detect DNA methylation in cancer cell lines by naked-eye. This method avoids complicated surface modification procedures involve in many conventional methods. Herein, we have envisioned an extremely simple method referred to as “Methylsorb”, which is an effective and potential alternative to traditional methods and offers a label-free and real-time quantitative detection of DNA methylation. The underlying principle of Methylsorb involves the use of the sequencedependent affinity interaction of DNA bases towards gold substrates. Since (i) bisulfite treatment converts methylation events into base changes, and (ii) DNA base-gold affinity interaction (i.e., DNA ad3

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 25

sorption) is highly sequence-dependent,1,22-28 we hypothesized that the direct detection/quantification of adsorbed (via DNA-gold affinity interaction) bisulfite treated-DNA sequences could be an extremely simple way to analyse DNA methylation events. The schematic of the Methylsorb approach is illustrated in Scheme 1 (see also Supplementary Movie 1). Initially, genomic DNA from cells was treated with sodium bisulfite to convert unmethylated cytosines into uracil, leaving methylated cytosines unchanged. Bisulfite converted samples were then amplified via a subsequent asymmetric PCR amplification step, which converts all uracils into thymines in the sense strand (i.e., the DNA strand running from 5’ to 3’ end) or into adenines in the antisense strand (i.e., the DNA strand running from 3’ to 5’ end). Because amplification is performed asymmetrically it ensures that only ss-DNA sequences corresponding to the antisense strand are amplified. This step therefore introduces an accurate and unique way to create specific base changes on DNA, which ultimately can alter DNA-gold affinity interaction, and thus give quantitative information on the original methylation status of the sample. This is because (i) it only generates ss-DNA, which is more prone than dsDNA to uncoil sufficiently to expose its bases and rapidly interact with the gold surface,29 and (ii) it correlates methylation profile with gold-DNA affinity behavior, as only the antisense strand, which becomes guanine-enriched for methylated or adenine-enriched for unmetylated samples, is amplified. Since adenine bases are reported to have stronger affinity towards gold than guanine,1,22-28 the adenine-enriched ssDNA sequences (i.e., unmethylated sequence) are expected to exhibit higher adsorption on gold than the methylated guanine-enriched sequences. The differential affinity interaction (i.e., differential adsorption) of adenine or guanine-enriched sequences, corresponding to the DNA base-changes associated with DNA methylation, can be determined in a single step via direct quantification of adsorbed DNA sequences in-real time. We used this approach to quantify the DNA methylation status of two CpG clusters located within the engrailed homeobox1 (EN1) and MIR200B gene promoter regions on DNA derived from MCF7 and MDA-MB-231 cells respectively. As a standard reference for unmethylated DNA sample, we used 4

ACS Paragon Plus Environment

Page 5 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

whole genome amplified (WGA) DNA. To quantify adsorbed DNA sequences on gold chip, we used a surface Plasmon resonance (SPR) biosensing approach that detects changes in refractive index over time (i.e., changes in the SPR spectral shift) at the sensing surface, which are directly proportional to the relative mass increase associated with target adsorption, thus enabling the real-time and label-free monitoring of targets. This method has previously been used to detect regional DNA methylation using proteins with affinity to CpG rich regions30,31 or molecular inversion probes.18 Detecting DNA methylation through this gold-driven adsorption process in combination with SPR offers several critical improvements to conventional methylation technologies that include: (i) minimal sample and operational requirements, (ii) no use of surface-bound receptors and thus avoids modification procedures for the recognition layer of the SPR surface and (iii) label-free and real-time monitoring that avoids any artifacts generally associated with the use of labels.

MATERIAL AND METHODS DNA Samples Preparation. Genomic DNA from MCF7 and MDA-MB-231 cells was extracted from 105 cells plates using the DNeasy blood and tissue kit (QIAGEN Pty. Ltd., Venlo, Netherlands) according to manufacturer's instructions. Briefly, the cells were suspended in lysis buffer to lyse and release the nucleic acids and proteins into the solution. To remove the protein and RNA in the solution a digestion step was performed using proteinase and RNase enzymes respectively. The digested proteins and RNA were removed by centrifuging the solution in a spin column. The purified DNA was eluted from the column in 100 μL of elution buffer and stored at -20 °C. Whole Genomic Amplified (WGA) DNA samples were prepared by amplifying 50 ng of the human genomic DNA (Roche, Germany) using a REPLI-g whole genome amplification kit (QIAGEN Pty. Ltd., Venlo, Netherlands) according to manu-

5

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 25

facturer's instructions. This amplification step removes all methylation marks from DNA sequences and provides a control genomic DNA sample with fully unmethylated status. Bisulfite Treatment. The purified DNA was bisulfite treated using a MethylEasyXceed kit (Human Genetic Signatures Pty. Ltd., New South Wales, Australia) according to the manufacturer's instructions. Briefly, 4 μg of DNA were incubated with 150 mM sodium hydroxide solution at 37 °C for 15 min followed by treatment with sodium bisulfite at 80 °C for 45 min. After treatment, the DNA solution containing sodium bisulfite was mixed with 800 μL of water, 2 μL of glycogen (20 mg/mL, Fermentas, USA) and 1 mL of iso-propanol. The mixture was incubated on ice for 30 min followed by centrifugation at 14000 g for 10 min. The supernatant was removed and washed with 70% ethanol to precipitate the DNA pellet. The washing step was repeated twice to completely remove residual sodium bisulfite salts from the precipitated DNA pellet. The pellet was then resuspended in elution buffer and desulphonicated at 95 °C for 20 min. DNA Copy Number Normalization. To normalize the DNA copy number from each DNA source, the relative amount of EN1 and MIR200B genes in bisulfite treated cell and WGA DNA samples were determined by real-time amplification of the housekeeping Col2A1 gene using a Rotor-Gene 6000 thermomixer (Corbett Research, Mortlake, Australia) and performing comparative analysis of Ct (i.e. the fractional PCR cycle number at which the reporter fluorescence is greater than the threshold). Each 20 μL of the PCR mixture contains 1 unit Taq DNA polymerase (AmpliTaq DNA Polymerase, Applied Biosystems, Australia), 2 mM MgCl2 1X PCR buffer (AmpliTaq 10X PCR buffer) 0.2 mM each dNTP, 1% triton X-100, 10 μM SYTO 9 dye and 250 nM of each Col2A1 forward and reverse primer (Table S1 of the Supporting Information). Equal volume of the target cell sample and WGA DNA were added as template to the PCR reaction. Thermal cycling was carried out in a Rotor-Gene 6000 thermocycler (Corbett Research, Mortlake, Australia) using the following conditions: denaturation at 94 ºC for 10 min followed by 40 cycles of 94 ºC for 30 s, 55 ºC for 30 s and 72 ºC for 30 s. 6

ACS Paragon Plus Environment

Page 7 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Asymmetric PCR: To obtain ss-stranded amplified amplicons containing the target sequence, bisulfite treat samples were amplified in a 60 μL reaction containing 1.5 unit Taq DNA polymerase (AmpliTaq DNA Polymerase, Applied Biosystems, Australia), 0.7X PCR buffer (AmpliTaq 10X PCR buffer) 0.2 mM each dNTP, 0.1% tween, 125 nM of forward primer and 375 nM reverse primer (Table S1 of the Supporting Information). Cycling was carried out in a Bio-Rad thermo cycler (MJ Mini Personal Thermal Cycler) using the following conditions: denaturation at 94 ºC for 10 min followed by 50 cycles of 94 ºC for 30 s, 58 ºC for 45 s and 72 ºC for 30 s. Methylation Detection with SPR Biosensor. All experiments were performed using a previously reported18 custom-made SPR platform working under Krestchman configuration. This platform uses a low-cost, CCD spectrometer (Ocean Optics, Spectrasuite Jaz Module) for analyzing spectral shifts. Initially, the SPR sensor chips (5nm Ti and 50 nm Au) were cleaned by rinsing with hot acetone, ethanol and deionised water and dried under flow of nitrogen gas. Subsequently, the chips were dipped in piranha solution (70% H2SO4–30% H2O2) for few seconds, rinsed with water and dried under flow of nitrogen gas. Following this step, the chip was placed in the SPR platform and sodium saline citrate (SSC) 5X buffer (0.75 M in NaCl, 0.075 M in sodium citrate, pH 7) was flowed in continuous at a flow rate of 0.6 mL/h. This buffer with high salt concentration was chosen to screen negative charges of the DNA strands (i.e., to minimize repulsive forces between DNA strands and between DNA strands and gold surfaces, a condition where DNA adsorption favours on flat gold surfaces32). Then, 250 μL of DNA samples (10 μL of asymmetric PCR output spiked in 250 μL of SSC5X buffer) were directly injected in the SPR flow system at the same flow rate. The SPR signals were monitored using custom-made Labview software. Bisulfite Sequencing. To investigate the methylation status of the interrogated region (MIR200B) in MDA-MB-231 cells, we employed bisulfite sequencing. Briefly, the resulting PCR amplicon was electrophoresed through an agarose gel and purified using a QIAquick Gel Extraction Kit (QIAGEN Pty. 7

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 25

Ltd., Venlo, Netherlands) according to manufacturer's instructions. The agarose gel containing the desired PCR amplicon was cut and dissolved in a dissolving buffer and centrifuged through a spin column to wash away impurities. The purified DNA was eluted in 40 μL of DNase free water and quantified using a ND-1000 Nanodrop spectrophotometer (Thermo Scientific, Wilmington, USA). The purified DNA was then mixed with the primer (20 ng DNA + 0.5 μL of 10 μM primer in 12 μL H2O) and sequenced using Sanger sequencing protocol by the Australian Genome Research Facility (AGRF, Brisbane). In silico Analysis of DNA Amplicons Sequences. BiQ analyser software33 was used to predict the sequence changes introduced by the bisulfite treatment and to generate the antisense strand sequences expected to be obtained during the asymmetric PCR processing.

RESULTS AND DISCUSSION To investigate the strategy of gold–DNA affinity interactions for detecting DNA methylation levels, we designed synthetic DNA oligos containing sequences (Figure 1A) similar to the bisulfite treated and asymmetric amplified sequences derived from the EN1 gene. The EN1 gene comprises a cluster of eight potentially methylated CpG sites within a span of 53 bases downstream to the transcription start site and identified as a potential biomarker in several cancer types.34-36 Samples containing synthetic methylated (200 nM) and/or unmethylated DNA oligos (200 nM) were driven through the SPR biosensor system using a syringe pump at the designated flow rate (see Supporting Information for SPR device setup and operation). We initially performed our experiment at pH 3 because DNA adsorption extent and kinetics are reported to be facilitated at pH 3.37,38 However, we have observed that adsorption at pH 3 limits the ability of the Methylsorb approach to differentiate methylated and unmethylated targets (see Figure S1 of the Supporting Information). At low pH (i.e. pH = 3) adenines and cytosines become protonated, leading to high DNA adsorption compared to that under neutral pH conditions. 37,38 Since both methylated and un8

ACS Paragon Plus Environment

Page 9 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

methylated sequences contain large numbers of cytosines and adenines bases, they could be strongly adsorbed on gold substrate within a short period of time at pH C ≥ G >T) of DNA onto gold surfaces. Notably, this outcome indicates that the Methylsorb approach can effectively distinguish methylated and unmethylated epigenotypes in synthetic samples even though they involved only 8 CpG sites (i.e., it has 8 A/G-base-changes “diluted” across 53 bases). Despite this high “dilution effect” the difference between the magnitude of SPR signals for methylated and unmethylated epigenotypes was found to be significantly large (i.e., Spectral Shift 1.5 nm). It is also important to note that for a given number of DNA bases, a DNA containing homo-polymer (e.g., a sequence with clustered poly-A motifs) and a DNA containing more distributed DNA bases (e.g., adenine bases spread across the sequence) might essentially not exhibit the same adsorption behaviour. We analysed the sequence composition of the methylated and unmethylated epigenotypes and investigated the presence of any high gold-affinity motif containing (i) homo-polymers (e.g., poly A) with a minimum of four bases and, (ii) polymers enriched with a mixture of adenine and cytosine bases. The identified homo-polymers (underlined) and polymers enriched with A, C nucleotide bases (highlighted in grey) are represented in Figure 1a. It is evident that the methylated and unmethylated DNA contains similar homo-polymers both in numbers (e.g., 3 poly A 9

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 25

and 1 poly-C motif) and in length. However, the polymers enriched with A,C bases were significantly larger in length for unmethylated sequence. This is because bisulfite treatment and PCR processing of DNA not only lead to A/G-enrichment, but also result in GC or AC enriched motifs in the methylated or unmethylated sequences (Scheme 1 and Figure 1A). Additionally, bisulfite conversion (followed by asymmetric PCR) of DNA also increases the adenine contents in the antisense strand for both epigenotypes. This enhances the possibility of forming larger AC motifs in the unmethylated sequences. We believe that this differential length of poly-AC motifs in methylated and unmethylated sequences is the key factor to offer a distinguishable SPR signal in our method. It is important to note that our adsorption (i.e., affinity) trend (A>C>G>T) is different than the trend (G>A>C>T) previously reported on planar surfaces22 where DNA was adsorbed in aqueous condition but their thermal desorption energy trend was measured under ultrahigh vacuum conditions. Because the G>A>C>T trend was observed under a very different experimental condition, we believe that this might not be directly related to the adsorption (i.e., affinity) trend in aqueous conditions on planar gold surfaces. To investigate the applicability of this approach for numerous other CpG sites, we used BiQ analyser software32 to predict the amplicon sequences that would result from bisulfite treatment and asymmetric PCR processing of the CpG rich regions from other cancer related genes. Table S2 (Supporting Information) represents the identified homo-polymers (underlined) and polymers enriched with A, C nucleotide bases (highlighted in grey) in the methylated and unmethylated epigenotypes of five other cancer related genes. We observed that unmethylated amplicons derived from cancer related genes contained larger polymers enriched with A, C bases than the methylated DNA (see Table S1 of the Supporting Information). This indicates that our approach can potentially be applicable for accurately analysing the methylation status of different cancer related genes. To assess the applicability of this methodology in genomic DNA samples, we chose to investigate (i) EN1 region in DNA derived from MCF7 cells, and (ii) a region containing 8 CpG sites in a span of 10

ACS Paragon Plus Environment

Page 11 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

65 bases within the promoter region of the MIR200B gene derived from the MDA-MB-231 cells. Both the genes were identified previously as a potential biomarker in several cancer types. 33-35, 38-40 We simultaneously analysed the same EN1 and MIR200B regions on whole genomic amplified (WGA) human DNA. This DNA is fully unmethylated (adenine enriched), and hence DNA amplicons derived from WGA are expected to give maximum achievable SPR signal for this region. Since methylated amplicons are guanine-enriched and they have lower adsorption affinity compared to the WGA DNA, there will be a clear correlation between the presence of methylation and lowering of the SPR signal from that of the WGA DNA. We compared the SPR response with that of the unmethylated WGA DNA sample. Figure 2a illustrates the SPR signals obtained for the direct adsorption of MCF7 and WGA DNA amplicons onto a gold surface. Similar to the responses obtained for EN1 synthetic DNA samples, we observed an enhanced adsorption behaviour in the case of unmethylated WGA DNA (A>G; dark blue line) that is evident from a larger spectral shift (Figure 2b, 5.85 ± 0.07 (WGA) versus 3.1 ± 0.2 nm (MCF7)) in comparison to methylated MCF7 DNA (G>A; red line) indicating the methylated status of this region in MCF7 cells. Responses were also very reproducible with a RSD value of less than 7% (n = 3) and were in agreement with our previous sequencing analysis data of this same region on MCF7 cells (see Supporting Information of reference 21). Subsequently, the Methylsorb analysis of MIR200B region resulted in almost identical SPR signals for MDA-MB-231 (Figure 2c; red) and WGA DNA (Figure 2c; blue). The run-to-run reproducibility of analysing methylation status of MIR200B was also high with a RSD value of less than 7% (n = 3). The negligible difference in SPR spectral shift (Figure 2d; 4.74 ± 0.09 (MDA-MB-231) versus 4.8 ± 0.1 nm (WGA)) values suggested that the MIR200B region in MDA-MB-231 could possibly be unmethylated. The in silico analysis of the MIR200B sequence using BiQ software32 supports this observation. As shown in the Table S1 (Supporting Information), DNA amplicons derived from a methylated source gave shorter poly-AC motifs (i.e., lower adsorption) while unmethylated sources generated amplicons with the longer poly-AC motifs (i.e., higher adsorp11

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 25

tion). Therefore, the similar level of adsorption for the MIR200B region in DNA derived from MDAMB-231 cells and WGA DNA can only be justified by the presence of longer poly-AC motifs in the MIR200B antisense sequence (i.e., unmethylation status of the promoter region of the MIR200B gene in MDA-MB-231 cells). Sequencing analysis data (Figure S2 of the Supporting Information) further confirmed the unmethylated status of MIR200B sequence in MDA-MB-231 DNA cell line. This data contrasts with previous observations, where a strong CpG methylation (i.e., hypermethylation) across the promoter of the MIR200B in MDA-MB-231 cell line has been identified.39,40,41 It has been reported that replicas of the same cell line in vitro could exhibit differential DNA methylation at certain regions,42 we believe that this could explain the discrepancy between our results and those from the literature. Methylation levels across a target region can vary among cancer types and stages of the disease43,44. Additionally, clinical samples often contain a mixture of tumour and normal cells where the presence of unmethylated sequences can potentially influence the sensitivity of the diagnostic assay45. This is particularly important at the early stages of cancer where tumour cells might be present in very low numbers. These factors contribute to generate heterogeneous DNA methylation. Because it has high relevance in cancer diagnosis, any new DNA methylation method should have ability to detect heterogeneous DNA methylation at different level. Therefore, we investigated the ability of our approach to detect heterogeneous DNA methylation (i.e., methylated targets in the presence of designated proportions of unmethylated DNA) by measuring the adsorption of DNA samples containing mixture of methylated and unmethylated sequences. Samples were prepared by quantifying the copy numbers of MCF7 and WGA DNA followed by mixing designated proportions of methylated and unmethylated targets (i.e., MCF7/WGA DNA at 0%:100%, 25%:75%, 50%:50%, 75%:25%, and 100%:0%). Figure 3a represents the SPR sensograms of the heterogeneously methylated DNA samples. It was evident that the SPR signal was a function of methylation percentage in the heterogeneous mixture. It was also noted that there was a linear correlation (R2 = 0.9988) in signal enhancement with a decrease in methylation percentage 12

ACS Paragon Plus Environment

Page 13 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

(Figure 3b). This is presumably due to the increase in the adenine nucleotide content with increase in the percentage of unmethylated DNA in the mixture resulting in an enhanced signal. Additionally, a significant difference in signal for samples containing 25% and 0% methylated targets (Figure 3b; 5.85 ± 0.07 versus 5.16 ± 0.11 nm) indicate that our approach was sensitive enough to detect at least 25% methylation changes. The level of sensitivity and excellent reproducibility demonstrated in this study indicates that our approach can potentially find its relevance in epigenetic investigations for heterogeneously methylated DNA samples. However, we believe that further optimization to the operating parameters (e.g., flow rate, buffer composition and multiplexed SPR platform) can improve the detection sensitivity of our method. It is also notable that our method is not suffered by PCR bias (see Supporting Information). Although there are many conventional methods able to detect heterogeneous methylation at much lower levels (e.g. DNA sequencing6,7,8 or MS-HR melting10), our method provides other advantages in compare to the conventional methodologies. Firstly, SPR is an extremely robust well-known technology for the quantitative analysis of molecular interactions46, and since our customized SPR system18 is easy-to-built and relatively inexpensive, it represents a cost-effective alternative to existing costly methods such as RT-PCR, sequencing, or mass-spectrometeric readouts. Secondly, as Methylsorb is a receptor-free approach, it avoids complicated surface functionalization steps (i.e., attaching capture probe on the Au surface for target hybridizations, etc). This also reduces variability associated with the surface modification process (i.e., it has high reproducibility) and makes the whole detection process relatively simple to accomplish. Thirdly, since detection is always performed as a relative measurement by comparing adsorption data from target sample with respect to fully unmethylated internal standard (i.e., WGA derived amplicons in this case), it further provides a way for controlling intra and inter-assay variability. Moreover, due to the real-time nature of the Methylsorb approach, it offers in-situ detection and absolute quantification in a single-step detection and does not require any complex data analysis. Finally, this assay could be addressed in a multiplex format using the SPR-imaging homologue. 13

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 25

CONCLUSION We have developed a simple, cost-effective and real-time approach with the capacity to sensitively detect DNA methylation levels in genomic DNA samples. Our experimental findings demonstrate the accurate interrogation of CpG regions on multiple genes using SPR biosensor and simultaneous validation using conventional bisulfite sequencing. This unique label-free and rapid (e.g., 40 minutes) approach based on the preferential nucleotide affinity towards gold can effectively detect DNA methylation levels thereby eliminating the need for sequencing approaches. We have demonstrated the feasibility of using this approach to sensitively (e.g., 25% methylated target) and specifically detect methylated targets in the presence of large excess of unmethylated DNA. This method has also the potential to detect global hypomethylation, one of the most important biomarkers of cancer, since the large cohort of CpG regions involved in this type analysis might promote a large difference in the adsorption profiles of normal and hypomethylated samples. We envisage that this simple and rapid approach can potentially find its relevance in diagnostic settings.

ASSOCIATED CONTENT Supporting Information Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org NOTES Authors declare no financial conflict of interest.

AUTHOR INFORMATION Corresponding Author 14

ACS Paragon Plus Environment

Page 15 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

* Corresponding authors: [email protected], [email protected] and [email protected] Author Contributions LGC, MJAS and MT designed the experiments and supervised the project. AAIS, LGC and RP conducted most of the experiments. SR designed and fabricated the SPR sensor chips. All authors discussed the data and wrote the paper. † These authors contributed equally. ACKNOWLEDGMENT This work was supported by the UQ fellowship (2012001456) and ARC DECRA and DP (DE120102503, DP140104006). We also acknowledge funding received by our laboratory from the National Breast Cancer Foundation of Australia (CG-12-07) to MT. This NBCF grant has significantly contributed to the environment that stimulates the research described here.

15

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 25

REFERENCES (1) Kimura-Suda, H., Petrovykh, D.Y., Tarlov, M.J. and Whitman, L.J. J. Am. Chem. Soc. 2003, 125, 9014-9015. (2) Suzuki, M. and Bird, A. Nat. Rev. Genet. 2008, 9, 465-476. (3) Heyn, H. and Esteller, M. Nat. Rev. Genet. 2012, 13, 679-92. (4) Heyn, H., Méndez-González, J. and Esteller, M. Expert. Rev. Mol. Diagn. 2013, 13, 473-479. (5) Clark, S.J., Statham, A., Stirzaker, C., Molloy, P.L. and Frommer, M. Nature Protocols. 2006, 1, 2364. (6) Lee, E., Luo, J., Wilson, J. and Shi, H. Cancer Lett. 2013, 340, 171-8. (7) Suzuki, M. and Greally, J. Semin. Hematol. 2013, 50, 70-7 (8) Tost, J. and I. G. Gut. Nat Protoc. 2007, 29: 2265-2275. (9) Herman, J.G., Graff, J.R., Myohanen, S., Nelkin, B.D. and Baylin, S.B. Proc. Natl. Acad. Sci. USA. 1996, 93, 9821-9826. (10) Wojdacz, T. K. and Dobrovic, A. Nucleic Acids Res. 2007, 35, e41. (11) Chatterton, Z., Burke, D., et al. Clin Chemistry. 2014, 60, 995-1003 (12) Xiong, Z. and Laird, P.W. Nucleic Acids Res. 1997, 25, 2532-2534. (13) Chen, H.-C., Chang, Y.-S., et al. J Chromatogr A. 2012, 1230, 123-129. (14) Cao, A. and Zhang, C.-Y. Anal. Chem. 2012, 84, 6199-6205. (15) Koo, K.M., Wee, E.J.H., Rauf, S., Shiddiky, M.J.A. and Trau, M. Biosens. Bioelectron. 2014, 15, 278-85. 16

ACS Paragon Plus Environment

Page 17 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

(16) Liu, S., Wu, P., et al. Chem. Commun. 2011, 47, 2844-2846. (17) Zhang, H., Li, M., et al. Chem. Commun. 2014,50, 2932-2934. (18) Carrascosa, L.G. et al. Chem. Commun. 2014, 50, 3585 - 3588. (19) Shin, Y., Perera, A. P., et al. Sens. Act. B: Chem. 2013, 177, 404-411. (20) Hu, J. and Zhang, C.-Y. Biosens. Bioelectron. 2012, 31, 451-457. (21) Lin, Y.-Z. & Chang, P.-L. ACS applied materials & interfaces. 2013, 5, 12045-12051. (22) Demers, L. et al. J. Am. Chem. Soc. 2002, 124, 11248-9. (23) Östblom, M., Liedberg, B., Demers, L.M. and Mirkin, C.A. J. Phys. Chem. B. 2005, 109, 151501516. (24) Opdahl, A., Petrovykh, D.Y., Kimura-Suda, H., Tarlov, M.J. and Whitman, L. J. Proc. Natl. Acad. Sci. USA. 2007, 104, 9-14. (25) Schreiner, S. et al. Anal. Chem. 2010, 82, 2803-10. (26) Nelson, E. and Rothberg, L. Langmuir. 2011, 27, 1770-7. (27) Zhang, X., Liu, B., Servos, M.R. and Liu, J. Langmuir. 2013, 29, 6091-6098. (28) Zhang, X., Servos, M. and Liu, J. Langmuir. 2012, 28, 3896-902. (29) Li, H. and Rothberg, L. Proc. Natl. Acad. Sci. USA. 2004, 101, 14036–14039 (30) Pan, S. et al. Biosens. Bioelectron. 2010, 26, 850-3. (31) Yu, Y. et al. Anal. Chem. 2010, 82, 5012-9. (32) Herne, T. M. and Tarlov, M. J., J. Am. Chem. Soc 1997, 119, 8916-8920 (33) Bock, C. et al. Bioinformatics 2005, 21, 4067-4068. (34) Karpinski, P., Ramsey, D., Grzebieniak, Z., Sasiadek, M.M. and Blin, N. Mol. Cancer Res. 2008, 6, 585. (35) Mayor, R. et al. Br. J. Cancer 2009, 100, 1534-1539. (36) Rauch, T. et al. Proc. Natl. Acad. Sci. USA. 2007, 104, 5527-32. 17

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 25

(37) Zhang, X., Liu, B., Dave, N., Servos, M. and Liu, J. Langmuir 2012, 28, 17053-60. (38) Jiang, H., Materon, E.M., Sotomayor, M.d.P. and Liu, J. J Colloid Interface Sci 2013, 411, 92-7. (39) Köhler, C. et al. Biochem. Biophys. Res. Commun. 2013, 438, 48-53. (40) Lim, Y. et al. J Cell Sci. 2013, 126, 2256-66. (41) Zhang, L. et al. Carcinogenesis. 2013, 34, 577-86. (42) Cocozza, S., Scala, G., Miele, G., Castaldo, I. and Monticelli, A. BMC Genomics 2013, 14, 692. (43) Varley, K. E., Mutch, D. G., Edmonston, T. B., Goodfellow, P. J., Mitra, R. D., Nucleic acids research 2009, gkp457. (44) Mikeska, T., Candiloro, I. L., Dobrovic, A., Epigenomics 2010, 2, 561-573 (45) Wang, S. et al. PLoS ONE. 2012, 7, e37928. (46) Homola, J., Chem. Rev. 2008, 108.

18

ACS Paragon Plus Environment

Page 19 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure Captions Scheme 1. Methylsorb Approach. Genomic DNA is derived from breast cancer cell lines. Bisulfite treatment and asymmetric PCR processing of these samples generate adenine-enriched (blue) or guanine-enriched (red) antisense ss-amplicons for unmethylated and methylated templates respectively. The ss-DNA amplicons are directly adsorbed on gold and their adsorption is monitored in real-time and label-free manner using surface Plasmon resonance readout. DNA adsorption generates an increase in the SPR signal with respect to the initial baseline, which is directly proportional to the amount of adsorbed DNA sequences on the SPR chip. (See also Supplementary Movie 1). Figure 1. Detection of DNA methylation in synthetic samples within the selected EN1 region. (a) Nucleotide composition of oligo sequences. High and low affinity sequence motifs are noted (i.e., DNA sequence motifs of minimum 4 identical bases are underlined; sequence motifs corresponding to poly A/C mixed bases of minimum 4 nucleotides in length are highlighted in grey colour; the A/G differences between the two strands are highlighted in red font). (b) SPR sensorgrams showing the spectral shift generated during the adsorption of methylated (red) and unmethylated (blue) synthetic EN1 region onto the gold chip. (c) Corresponding mean values of SPR spectral shift for methylated (M) and unmethylated (UM) regions. Each bar represents the average of three separate trials (n=3), and error bars represent the standard deviation of measurements within experiments (relative standard deviation (%RSD) was found to be