In situ detection of transrenal gene mutations without DNA isolation and amplication using Array Piezoelectric Plate Sensor

(PEPS).

A Thesis Submitted to the Faculty of Drexel University by Ceyhun E. Kirimli in partial fulllment of the requirements for the degree of Doctor of Philosophy in Biomedical Engineering June 2013

c Copyright 2013 Ceyhun E. Kirimli. All Rights Reserved.

iv

Table of Contents List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Structure of DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.1 Hybridization of Single Stranded and Denaturation of Double Stranded DNA Fragments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Cell-free Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 1.3 1.4 6 7 1 1

Concentration of Cell Free Nucleic Acids . . . . . . . . . . . . . . 10

Transrenal DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 DNA Biosensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.4.1 1.4.2 Polymerase Chain Reaction . . . . . . . . . . . . . . . . . . . . . . 15 Other Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

1.5

Piezoelectric Microcantilever (PEMS) and Piezoelectric Plate Sensor . . . . 22 1.5.1 1.5.2 Piezoelectric Property Dependence . . . . . . . . . . . . . . . . . . 27 Insulation Quality Dependence . . . . . . . . . . . . . . . . . . . . 29

1.6

Objective and Specic Aims . . . . . . . . . . . . . . . . . . . . . . . . . 31

2 Detection of DNA hybridization in buffer and Urine . . . . . . . . . . . . . . . . 33 2.1 2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 2.2.1 PEPS fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

2.2.2 2.2.3 2.2.4 2.2.5 2.2.6 2.2.7 2.2.8 2.2.9

Electrical Insulation . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Target DNA, probe DNA, and reporter DNAs . . . . . . . . . . . . 41 Urine Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . 45 Flow Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Blocking of Non-specic Binding in Urine . . . . . . . . . . . . . . 46 Data acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Resonance peak frequency determination . . . . . . . . . . . . . . 49 Re-plotting Detection Experiments . . . . . . . . . . . . . . . . . . 54

2.2.10 Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 2.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 2.3.1 2.3.2 2.3.3 2.3.4 2.4 Analysis on Simulation Parameters . . . . . . . . . . . . . . . . . . 59 BSA blocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Dose response in 1 x PBS and urine and visual conrmation . . . . 63 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Summary of Resonance Frequency Peak Determination . . . . . . . . . . . 70

3 Flow-Enhanced Detection Specicity of Mutated DNA against Wild Type with Reporter Microspheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 3.1 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 3.1.1 3.1.2 3.1.3 3.1.4 3.1.5 Target DNAs, probe DNA, and reporter DNAs . . . . . . . . . . . . 75 Substrate Preparation and pDNA Immobilization . . . . . . . . . . 77 rDNA conjugation to FRMs . . . . . . . . . . . . . . . . . . . . . 78 tDNA Hybridization and Subsequent Capturing of FRMs . . . . . . 80 FRMs counting . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

vi

3.2

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 3.2.1 3.2.2 3.2.3 FRM conjugation and initial room-temperature testing without ow 84

Effect of Flow Rate and Temperature . . . . . . . . . . . . . . . . . 86 SMT /SW T at 10, 103 , 106 , and 107 WT/MT concentration ratios . . . 91

3.3 3.4

Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Summary of Temperature and Flow Enhanced Specicity using Fluorescent Reporter Microspheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

4 Detection of mutations of single-stranded DNA . . . . . . . . . . . . . . . . . . 97 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 4.1.1 4.2 Locked Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . 98

Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 4.2.6 4.2.7 PEPS Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Electrical Insulation . . . . . . . . . . . . . . . . . . . . . . . . . . 101 pDNA Immobilization . . . . . . . . . . . . . . . . . . . . . . . . 104 Immobilization of pDNA for 6 mutations on codon 12 . . . . . . . 105 Non-specic binding . . . . . . . . . . . . . . . . . . . . . . . . . 105 Target DNAs, probe DNA, reporter DNAs and FRMs . . . . . . . . 106 Urine samples and Experimental Setup . . . . . . . . . . . . . . . . 110

4.3

Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 4.3.1 4.3.2 4.3.3 HBV-DM single stranded DNA detections . . . . . . . . . . . . . . 113 kRas single stranded tDNA detections . . . . . . . . . . . . . . . . 119 Multiplexed Detection of kRas codon 12 Mutations . . . . . . . . . 125

4.4

Summary of Single Stranded DNA mutation detections . . . . . . . . . . . 128

vii

5 Detection of double stranded DNA . . . . . . . . . . . . . . . . . . . . . . . . . 131 5.1 5.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 5.2.1 PEPS fabrication, electrical insulation, pDNA immobilization and BSA blocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 5.2.2 5.2.3 5.3 Target DNAs, probe DNA, reporter DNAs and FRMs . . . . . . . . 133 Urine samples and Experimental Setup . . . . . . . . . . . . . . . . 134

Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 5.3.1 5.3.2 5.3.3 5.3.4 5.3.5 5.3.6 Comparison of Denaturation and tDNA capture methods . . . . . . 137 HBV-DM double stranded mutation detections . . . . . . . . . . . 138 HBV-DM double stranded mutation mixture detections . . . . . . . 140 kRas double stranded mutation detections . . . . . . . . . . . . . . 142 kRas double stranded mutations mixture detections . . . . . . . . . 144 SW480 Cell Line Extracted double stranded DNA detections . . . . 146

5.4

Summary of Double Stranded DNA mutation detections . . . . . . . . . . . 149

6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 6.1 6.2 6.3 6.4 6.5 Optimization of Temperature and Flow Rate . . . . . . . . . . . . . . . . . 151 Resonance Frequency Peak Determination Method . . . . . . . . . . . . . 151 Detection of Single Stranded DNA fragments . . . . . . . . . . . . . . . . 151 Multiplexed Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Detection of Double Stranded DNA . . . . . . . . . . . . . . . . . . . . . 152

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

viii

List of Tables

1.1 2.1 3.1 4.1 4.2 5.1

Comparison of other DNA biosensing platforms . . . . . . . . . . . . . . . The sequences and Tm of the pDNA, tDNA, urDNA and drDNA . . . . . . . The sequences and corresponding melting temperatures adjusted with salt concentration in 1xPBS . . . . . . . . . . . . . . . . . . . . . . . . . . . . The sequences and Tm of the pDNA, tDNA, urDNA and drDNA for detection of mutations of kRas and HBV-DM . . . . . . . . . . . . . . . . . . . The sequences of pDNAs and Tm s of the pDNA with MT and WT tDNA, for detection of mutations of 6 kRas mutations . . . . . . . . . . . . . . . . The sequences of tDNAs, used for detection of mutations of double stranded kRas codon 12 GGT to GTT point mutation and HBV-DM . . . . . . . . .

22 41 76 129 130 134

ix

List of Figures

1.1

1.2

1.3

1.4

1.5

1.6

1.7

Complementary base pairs making up the double helix are shown. Adenine hybridizes with Thymine and Guanine with Cytosine. GC pairs hybridize with 3 hydrogen bonds whereas AT pairs hybridizes with two and antiparallel backbones of complementary strands are also shown. (Image adopted from [1]). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DNA and its constituent substructures are shown on the top left. They covalently bond with each other in a sequence to form the single stranded DNA shown on top right. In the bottom left diagram the DNA is shown straightened out for ease of understanding however the real DNA structure which is shown on the right is twisted to form the double helix. (Image adopted from [1]). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A) A space-lling model of helical double stranded DNA is shown on the left. This 1.5 turn (10.4 base pairs per turn) model shows the major and minor groves in the structure. (B) A short section of double stranded DNA is also illustrated on the right to show the real bond patterns between nucleotides. 5 and 3 ends can also be shown together with the phosphate backbone.(Image adopted from [1]). . . . . . . . . . . . . . . . . . . . . Hypochromism investigated by spectroscopy. (A) Single-stranded DNA absorbs more light with respect to double stranded helical DNA. (B) The absorbance increases as the double stranded DNA denatures to form single stranded DNA which can be monitored in real time to determine the melting temperature Tm of any DNA sequence. (Image adopted from [2]). . . . . Not only mutations but also methylations, DNA integrity, microsatellite alterations and viral DNA has been shown to be associated with cf-DNA in blood. Both single stranded and double stranded DNA can be released in to the blood stream through different ways as discussed. Moreover individual tumor type can release more than one form of cf-DNA. (Image adopted from [3]). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transrenal DNA production in a patient with pulmonary tuberculosis is shown. M tuberculosis bacilli from infective foci in the lungs are destroyed by the immune response releasing cell-free nucleic acids in plasma. Filtered cf-DNA forms Tr-DNA which can then be detected. (Image adopted from [4]). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Illustration of Denaturation part of a cycle of PCR reaction. (Image adopted from www.btci.org). . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. 13 . 15

1.8 1.9 1.10

1.11 1.12 1.13

1.14 2.1

2.2

2.3

Illustration of annealing part of a cycle of PCR reaction. Forward and reverse primers anneal on complementary strands of the double stranded amplicon. (Image adopted from www.btci.org). . . . . . . . . . . . . . . . . Illustration of extension part of a cycle of PCR reaction. Forward and reverse primers anneal on complementary strands of the double stranded amplicon. (Image adopted from www.btci.org). . . . . . . . . . . . . . . . . Illustration of extension part of a cycle of PCR reaction. Forward and reverse primers anneal on complementary strands of the double stranded amplicon. Image adopted from www.wikimedia.org where it is published by its author (Ygonaar) for sharing under both GNU Free Documentation License and creative commons by-sa licenses with some rights reserved. . . A schematic of (a) a PEPS which consists of solely a piezoelectric layer thinly coated with two electrodes on the opposite sides and (b) a PEPS electrically insulated with MPS-W9 coating for in-liquid detection. . . . . (a) A schematic top view of the rst LEM, (b) WEM, and (c) PEM vibration where the shaded bars illustrated the initial position of the PEMS and the dash-dotted shapes illustrate the bent or extended positions. . . . . . . . Phase angle (blue) and impedance (red) versus frequency of a PEPS with a k31 = 0.34 around the WEM resonance peak at 3.17 MHz as indicated by the peak in the phase angle plot and the characteristic frequencies, fa and f p are the frequencies at which the minimum and maximum of the impedance occur, respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The in-air and in PBS resonance frequency spectra of a PEPS (a) with MPS5 (old) insulation and with MPS-W9 insulation. . . . . . . . . . . . . . . A schematic of (a) a piezoelectric plate sensor (PEPS), (b) the rst length extension mode (LEM), (c) width extension mode (WEM) vibration of a PEPS where the shaded bars illustrated the initial position of the PEPS and the dash-dotted shapes illustrate the extended positions, (d) a top-view optical micrograph. The gold color in (d) and the thin layers lining the top and the bottom of the PMN-PT are 110 nm Cr/Au electrodes. . . . . . . . (a) In-air (black) and in-PBS (red) phase angle-versus-frequency resonance spectra, and (b) relative resonance frequency shift, f/f, of the PMN-PT PEPS during the various steps of probe cDNA immobilization and target DNA detection. The insert in (b) shows a schematic of the molecules involved in the immobilization and detection. . . . . . . . . . . . . . . . . cDNA immobilization Steps on PEPS surface: (1) In the rst step, the maleimide of the maleimide-PEG-biotin linker reacted with the sulfhydryl of 3-(trimethoxysilyl)-propyl-methacrylate (MPS) on the MPS coating surface to form a thiol-ester bond that covalently linked the maleimide-PEGbiotin on the MPS surface; (2) In the second step, the biotin of the immobilized maleimide-PEG-biotin reacted with streptavidin to immobilize streptavidin on the PEPS surface; (3) In the third step, the biotin at the 5end of the cDNA reacted with the streptavidin bound on the biotin of the immobilized maleimide-PEG-biotin to nally immobilize cDNA on the PEPS surface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. 16 . 17

. 18 . 24 . 25

. 28 . 31

. 37

. 40

. 43

xi

2.4

A schematic of the immobilization steps and nucleotide sequences of target DNA and its hybridization on pDNA, upstream rDNA and downstream rDNA. 2.5 Three resonance frequency peaks are monitored in real time. Each resonance peaks raw impedance spectrum is plotted on top and their corresponding time versus peak position graph is plotted in the bottom. Peak positions are determined using the peak determination method discussed below . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Phase angle versus frequency of WEM resonance spectrum of a PEPS. black full circles represent the raw data. The blue data points represent a neighborhood of 10 (a) and maximum neighborhood available (b) for parabola tting (blue, in insets), raw peak value is indicated by an arrow on graphs, , polynomials tted on these neighborhoods (pink, in insets) and peak positions of the tted polynomials (red circles). . . . . . . . . . . . . 2.7 a. Maximum of raw and smoothed data is shown in blue and pink respectively. b. Polynomial tting is done on a neighborhood of 41 points on smoothed spectrum. c. Polynomial tting is done on maximum available neighborhood of on smoothed spectrum. d. Neighborhood size versus peak of polynomials tted on smoothed (black) and raw (red) data is plotted. . . 2.8 (a) Detection of 10 M of tDNA in 1x PBS is plotted using raw peak positions of phase angle versus frequency plots of WEM resonance peak. (b) Impedance spectra of resonance frequency peaks(red stars in (a)) at time 0 (lled symbols) and 30 (open symbols) min. are plotted for the xed (red circles) and moving (black squares) window impedance scanning methods . 2.9 Detection experiment involving simultaneous monitoring of 7 peaks of the same PEPS is done. Experiment consisted of the following steps; background signal checking in 1xPBS, maleimide activated biotin binding, streptavidin binding, biotinylated pDNA binding, tDNA hybridization and second back ground signal checking in 1 x PBS. Two peaks (LEM and WEM) is drawn in this re-plotting. A sample raw impedance spectra of each peak is drawn on the right of time versus frequency shift graphs for peaks . . . . 2.10 a. Simulated model of a detection experiment representing graph of time versus resonance frequency peak position. b. Gaussian and second order simulated peaks without noise. c. A third order polynomial, with R representing the extend of asymmetry of polynomial. d. Simulated model of third order polynomials with different R values. . . . . . . . . . . . . 2.11 Error analysis of the peak monitoring algorithm by simulations is shown. a. Effect of parameter, R, on relative and frequency shift errors are plotted (a and b). Effect of Noise added on simulated resonance peaks on relative and frequency shift errors are plotted (c and d). Effect of Window Size on relative and frequency shift errors are plotted (e and f). . . . . . . . . . . 2.12 Relative frequency shifts of PEPS B in urine treated after 1,2 and 3 percent BSA and no BSA followed by a washing step in 6ml/min ow rate. . . . . .

45

48

50

52

54

56

58

61 63

xii

2.13 Standard curve of HBV DM detection in urine (a) and in 1xPBS (b) with PEPS A and PEPS B respectively at room temperature with 2ml/min ow rate. (c) Relative frequency shifts at 30 min(after tDNA hybridization) and 60 min (after FRM hybridization) are plotted versus concentration of tDNA. (d) FRM micrographs after tDNA and FRM hybridizations in urine for control, and concentrations; 50zM, 100 zM, 1 aM, 10 aM and 100 aM of tDNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.14 FRM micrographs after tDNA and FRM hybridizations in urine for control, and concentrations; 50zM, 100 zM, 1 aM, 10 aM and 100 aM of tDNA. Width of PEPSB is 420 m and is shown by dashed parallel lines in micrographs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.15 (a) Noise parameter of simulated detection experiments with zero frequency shifts are plotted versus corresponding standard deviation of the raw peak positions for each simulated experiement. Standard deviation in raw peak position of WEM resonance peak of a PEPS in 1xPBS is shown with dashed line. (b) Experimental data in (a) and simulated experiment with noise parameter 27 is plotted. This value is used for further simulations . . . . . . 2.16 (a) Dose response experiment in 1x PBS is done recording peak resonance frequencies applying the peak determination algorithm and a xed window schem simultaneously. SNR of xed window resonance frequency monitoring is plotted versus concentration of tDNA with single parabola tting and raw peak positions and compared with that of peak resonance frequency determination method (b) Percentage of error in relative frequency shifts between 25th and 30th minutes are plotted using simulated resonance peaks with the same noise in experimental peaks. . . . . . . . . . . . . . . . . . . 3.1 A schematic of the nucleotide sequences of (a) MT and (b) WT and their hybridization on pDNA, upstream rDNA and downstream rDNA . . . . . . 3.2 A schematic of (a) pDNA immobilization on GCG surface using sulfoSMCC, and (b) schematic of rDNA conjugation on carboxylated FRMs with sulfo-NHS and EDC. . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 (a) A schematic of a FPM coated with upstream and downstream rDNA, (b) a schematic of ow cell where the GCG is placed in the center of laminar ow, and (c) a schematic of ow system. . . . . . . . . . . . . . . . . . . . 3.4 A schematic of a) tDNA sequence own over pDNA immobilized on GCG, b) Upstream and downstream rDNA-conjugated FRMs own over the GCG surface and FRMs hybridized on the anking regions of the tDNA sequences, and c) unbound or loosely bound GCG washed . . . . . . . . . . . 3.5 Number of FRMs per mm2 captured on the GCG (a) followed by 10-10 M MT detection versus rDNA concentration and, (b) followed by MT detection versus MT concentration with pDNA optimally immobilized in 330 nM pDNA. Both (a) and (b) obtained at room temperature and without ow 3.6 Number of FRMs per mm2 captured by a GCG followed by detection in the control (no MT or WT, black), that followed by detection at 10-10 M WT (red), and that followed by detection (blue) at 10-10 M MT, (a) at room temperature, (b) at 30o C, and (c) at 35o C. . . . . . . . . . . . . . . . . . .

65

66

68

70 75 78 80

82

85

88

xiii

3.7

3.8

3.9

4.1

4.2

4.3

4.4

4.5

(a) SW T /SC and (b) p value versus ow rate at room temperature (black), at 30o C (red), and at 35o C (blue) where SW T and SC are number of FRMs per mm2 captured by a GCG followed by detection at 10-10 WT and that captured by a GCG followed by detection in control (no WT or MT). . . . . SMT /SW T versus ow rate where SMT is number of FRMs per mm2 captured by a GCG followed by detection at 1010 MT and SW T is number of FRMs/mm2 captured by a GCG followed by detection at 1010 WT at room temperature (black), at 30o C (red), and at 35 o C (blue) . . . . . . . . . . . . (a) Number of FRMs per mm2 versus WT concentration and, (b) ratio of number of FRMs per mm2 of MT/WT and Mixture/WT versus WT where mixture is 1017 M MT mixed with different WT concentrations, and brown, blue and green horizontal bars represent number of FRMs per mm2 for control, 1017 M MT with ow and 1017 M MT without ow conditions. Both (a) and (b) are done at 30o C, in 4ml/min ow rate or in no ow conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1m indicates the melting temperature curve of a certain mismatching sequence (point mutation) with a pure DNA probe. 1p indicated the melting temperature curve of the same sequence with a perfectly matching pure DNA probe. 2m and 2p indicates the same melting curves with a LNADNA hybrid pDNA which contains 3 LNA bases centered around the mismatching base. Figure adopted from [5]. . . . . . . . . . . . . . . . . . . . (a) In-air (black) and in-PBS (red) phase angle-versus-frequency resonance spectra, and (b) relative resonance frequency shift, f/f, of the PMN-PT PEPS during the various steps of probe cDNA immobilization and target DNA detection. The insert in (b) shows a schematic of the molecules involved in the immobilization and detection. . . . . . . . . . . . . . . . . . A schematic of amine activated pDNA immobilization on GCG surface using sulfo-SMCC.(a) PEPS is immersed in sulfo-SMCC solution at pH 6.5. (b) 5 amine activated pDNA is immobilized on the sulfo-SMCC conjugated PEPS surface at pH 8.0. . . . . . . . . . . . . . . . . . . . . . . . . . Schematics of pDNA, Wild Type and Mutant tDNAs and upstream and downstream rDNAs designed for (a) HBV-DM and (b) kRas mutations. Mutation site(s) are shown with patterned boxes, dashed lines indicate the hybridization sites and mismatch(es) on Wild Type tDNAs is/are indicated.(c) Bright blue and yellow-green microspheres are conjugated with upstream and downstream microspheres to be used as Mt-FRM and Wt-FRM . . . . Schematic of the detection setup. pDNA (Pink) is immobilized on SMCC conjugated PEPS surface via its primary amine attached on the 5. WT tDNA (light green) and MT tDNA (dark blue) can hybridize on pDNA with mismathc(es) and no mismatches respectively. Bright blue FRMs (dark blue sphere) conjugated with MT rDNA (yellow) and Yellow Green FRMs (green sphere) hybridizes on MT tDNA and WT tDNA respectively. . . . .

90

91

93

100

103

105

108

110

xiv

Array of 6 PEPS is shown inside the ow cell (bottom), and a blow up of this array is shown on top. Each PEPS is shown in different colors indicating immobilization by different pDNAs, each specic to one type of mutation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7 (a)Detection of HBV-DM Wt tDNA at room temperature and 2ml/min ow rate. 100zM and control experiments are shown in the inset. Detections were conrmed by FRM hybridizations and PBS background. (b)Average relative frequency shifts of tDNA and FRM hybridizations of 10 aM and 100 zM of MT-tDNA and WT-tDNA of HBV-DM are plotted. Experiments are done at room temperature and 2ml/min ow rate . . . . . . . . . . . . . 4.8 Relative frequency shift versus time graph of detections of (a) 100zM, 100,1,5 and 10 aM of MT tDNA followed by MT-FRM hybridization (b)1aM, 25aM, 100 aM, 10 fM and 100 fM of WT tDNA followed by WT-FRM hybridization in urine at 30o C and 4ml/min ow rate (c) Relative frequency shifts in the last ve minutes of tDNA and FRM hybridization parts (signal, S) of experiments in (a) and (b) plotted versus concentration and (d) The ratios of SMT /SW T are plotted for tDNA and FRM hybridization parts seperately . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9 (a) Relative frequency shift versus time graph of HBV-DM mixture detection experiments with 100zM, 1, 10 and 100 aM of MT tDNA mixed with 250 fold more amount of WT tDNA (Concentrations of MT tDNA in the mixture experiments are shown on the x axis). (b) Signal (S) of tDNA and FRM hybridization parts of mixture experiments are compared with sum of that of MT-only and WT-only experiments. (c) Number of MT-FRMs and WT-FRMs on PEPS surface was counted and plotted versus concentration of MT tDNA of mixture experiments.(d) Micrographs of PEPS after HBV-DM mixture detection experiments are ahown. MT-FRMs are shown in blue and WT-FRMs are shown in green. . . . . . . . . . . . . . . . . . . 4.10 Relative frequency shifts of detections of 10 fM of MT and WT kRas tDNAs are plotted versus time at 55o , 60o and 63o . (b) Signal (S) from MT (black) and WT (red) tDNAs (y-axis on the left)and ratio SMT /SW T (blue, y-axis on the right) are plotted for the detections in (a) were plotted . . . . . 4.11 Relative frequency shifts of detections of (a)100 zM, 1, 10 and 100 aM of MT tDNA and (b) 100 zM, 1, 10, 100 aM, 10fM and 100 fM of WT-tDNA are plotted versus time. (c) Signal(S) (average of relative frequency shifts between 25th )-30th min.(tDNA hybridization) and 55th -60th min. (FRM hybridization) versus tDNA concentration of detections in (a) and (b). (d) SMT /SW T plotted versus tDNA concentration. . . . . . . . . . . . . . . . .

4.6

112

114

115

118

120

122

xv

4.12 (a) Relative frequency shift versus time graph of kRas mixture detection experiments with 100zM, 1, 10 and 100 aM of MT tDNA mixed with 1000 fold more amount of WT tDNA (Concentrations of MT tDNA in the mixture experiments are shown on the x axis).(Signal of tDNA and FRM hybridization parts of kRas and HBV-DM mixture detection experiments compared with with that of HBV-DM MT tDNA detection experiments carried out in room temperature (denoted by @RT) in chapter 2). (c) Number of MT-FRMs and WT-FRMs hybridized on PEPS surface after detections in (a). (d) Micrographs of PEPS surface after detections carried out in (a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 4.13 Impedance spectra of PEPS used in multiplexed detection. . . . . . . . . . 125 4.14 Relative frequency shifts of PEPS used in multiplexed detection of 6 mutations of kRas codon 12, known to be associated with colorectal carcinoma. 6 urine samples containing 6 different mutations are own one by one over the PEPS array. Each sample was own for 30 min. . . . . . . . . . . . . . 127 5.1 buraya caption yaz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 5.2 Schematic of the experimental setup used to detect dsDNA.(a) Urine samples are heated to boiling temperature for 15 min. Hot sample (shown in red piping) is own through a water bath that is inside an incubator which is equilibrated to detection temperature. (b) Schematic of the ow system is shown. Sample is own through 2 ow cells including capture DNA immobilized GCG and pDNA immobilized PEPS in order. Detected sample is then discarded in the waste container. Dashed line illustrates the incubator in (a), everything inside the dashed line is in the incubator. . . . . . . . . . 135 5.3 (a) Relative frequency shifts of different cooling regimes are plotted for detection of 10 aM of HBV-DM dsMT tDNA in room temperature with 1ml/min ow rate. Relative frequency shift of same concentration of shorter ssDNA is plotted for comparison (b) Percentage of relative frequency shifts with respect to ssDNA detection of same concentration is plotted for different cooling regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 5.4 Relative frequency shifts of (a) 100 zM, 1 aM, 10 aM and 100 aM of HBVDM dsMT tDNA and (b) 100 zM, 1 aM, 10 aM, 100 aM, 10 fM and 100 fM of HBV-DM dsWT tDNA is plotted versus time. Detections are followed by FRM hybridizations. (c) Signal (S) is plotted versus tDNA concentration for dsMT and dsWT tDNA and FRM hybridizations. (d) SMT /SW T ratios are plotted for tDNA and FRM hybridizations versus concentration of tDNA.139

xvi

5.5

5.6

5.7

5.8

(a) Relative frequency shift versus time graph of mixture of dsMT and dsWT tDNA followed by FRM hybridizations. Concentration of dsMT tDNA is indicated on gure. 250 fold more dsWT tDNA is mixed with dsMT tDNA in each experiment. (b) Plot of average relative frequency shifts in last ve minutes of tDNA and FRM hybridization parts of mixture detections in (a) versus dsMT tDNA concentration and comparison with sum of average of relative frequency shifts of dsMT tDNA-only and dsWT tDNA-only detections in gure 5.4. (c)Plot of number of FRMs hybridized on PEPS surface after each mixture detection versus dsMT tDNA concentration. (d) Micrographs of FRMs on PEPS surface after mixture detection experiments in (a) . . . . . . . . . . . . . . . . . . . . . . . . . . . Relative frequency shifts of (a) 100 zM, 1 aM, 10 aM and 100 aM of kRas dsMT tDNA and (b) 10 aM, 100 aM, 10 fM, 100 fM, and 1 pM of kRas dsWT tDNA is plotted versus time. Detections are followed by FRM hybridizations. (c) Signal (S) is plotted versus tDNA concentration for hybridizations of dsMT, dsWT tDNA and FRM. (d) SMT /SW T ratios are plotted for tDNA and FRM hybridizations versus concentration of tDNA. . . . (a) Relative frequency shift versus time graph of mixture of dsMT and dsWT tDNA followed by FRM hybridizations. Concentration of dsMT tDNA is indicated on gure. 1000 fold more dsWT tDNA is mixed with dsMT tDNA in each experiment. (b) Average relative frequency shifts in last ve minutes of tDNA and FRM hybridization parts of mixture detections in (a) is plotted versus dsMT tDNA concentration and compared with sum of average of relative frequency shifts of dsMT tDNA-only and dsWT tDNA-only detections made in gure 5.6. (c) Number of FRMs hybridized on PEPS surface after each mixture detection is plotted versus dsMT tDNA concentration. (d) Micrographs of FRMs on PEPS surface after mixture detection experiments in (a) is shown . . . . . . . . . . . . . (a) Relative frequency shift versus time graphs of ds-SW480 tDNA followed by FRM hybridizations. Concentration of ds-SW480 tDNA is indicated on gure. (b) Average relative frequency shifts in last ve minutes of tDNA and FRM hybridization parts of ds-SW480 tDNA detections in (a) is plotted versus tDNA concentration and compared with average of relative frequency shifts of synthetic dsMT tDNA detections made in gure 5.6a. (c) Number of FRMs hybridized on PEPS surface after each ds-SW480 tDNA detection is plotted versus tDNA concentration. (d) Micrographs of FRMs on PEPS surface after mixture detection experiments in (a) is shown

141

143

145

148

xvii

Abstract In situ detection of transrenal gene mutations without DNA isolation and amplication using Array Piezoelectric Plate Sensor (PEPS). Ceyhun E. Kirimli Wan Y. Shih, Ph. D. Wei-Heng Shih, Ph. D. Low molecular weight (LMW) DNA fragments from distant organs can pass through kidneys and exist in urine as transrenal DNA. Transrenal DNA provides a noninvasive means to access cell-free DNAs from distant organs such as liver, pancreas, colon that are otherwise difcult to obtain. Shih and Shihs laboratory at Drexel has developed a highly sensitive piezoelectric plate sensor (PEPS) based on 8- m thick highly piezoelectric lead magnesium niobate-lead titanate (PMN-PT) freestanding layer. For detection, receptor specic to a target analyte is immobilized on the PEPS. Binding of the target analyte to the receptor on the PEPS surface decreases its resonance frequency. In-situ, label-free detection of the target analyte is achieved by monitoring the PEPS resonance frequency shift. The detection resonance frequency shift of PEPS is enhanced more than 1000 times due to the PMN-PT layers capability to change its crystalline orientation upon analyte binding. As a result, PEPS has exhibited unprecedented attomolar (aM) sensitivity in labelfree detection of single-stranded DNA. The purpose of this study was to further develop PEPS as a tool for rapid (in < 40 min), label-free, multiplexed and highly sensitive detection of transrenal gene mutations of distant organs in urine. 1762T/1764A double mutation of hepatitis B virus (HBV), a risk factor associated with hepatocellular carcinoma (HCC) and Kras point mutation at codon 12, a biomarker present in 50% of colorectal carcinomas (CRC) will be the model transrenal gene mutations in this study. To achieve the goal, a signal processing algorithm was devel-

xviii

oped to reduce noise in the resonance spectrum to increase the resolution of the detection resonance frequency shift. A ow system was implemented with temperature-controlled compartments to effectively dehybridize the naturally double-stranded target DNA so that the target DNA can be detected without the need for DNA isolation, concentration, and amplication. The use of temperature, ow speed and locked nucleic acid (LNA) were exploited to enhance the detection specicity of the mutated DNA against the wild type and an array of PEPS is used to detect multiple target DNAs simultaneously. Results indicate that PEPS could directly detect Kras point mutations in urine with 0.1 aM (60 copies/ml) sensitivity and 1:1000 mutant to wild type specicity without labeling, DNA isolation and amplication.

1 1. Introduction

1.1 Structure of DNA A deoxyribonucleic acid (DNA) molecule is composed of two polynucleotide chains each of which contains four different types of nucleotides. Each of these chains is also called a strand of DNA. Each nucleotide contains base portions through which two strands are attached one another by hydrogen bonding. Four bases contained in nucleotide of a single DNA strand may be either Adenine (A), Cytosine (C), Guanine (G) or Thymine (T) (Figure 1.1).

Figure 1.1: Complementary base pairs making up the double helix are shown. Adenine hybridizes with Thymine and Guanine with Cytosine. GC pairs hybridize with 3 hydrogen bonds whereas AT pairs hybridizes with two and antiparallel backbones of complementary strands are also shown. (Image adopted from [1]). The nucleotides in a DNA strand are covalently attached to each other through the sugars and phosphates which creates the backbone of a DNA strand. Backbone of a DNA strand makes the structure negatively charged due to abundance of negatively charged and

3 perfectly aligned phosphate groups (Figure 1.2).

Figure 1.2: DNA and its constituent substructures are shown on the top left. They covalently bond with each other in a sequence to form the single stranded DNA shown on top right. In the bottom left diagram the DNA is shown straightened out for ease of understanding however the real DNA structure which is shown on the right is twisted to form the double helix. (Image adopted from [1]).

4 A DNA strand is also polarized in the sense that one end of a DNA chain ends with a 5 phosphate and the other end terminates with a 3 hydroxyl group as each nucleotide in the sequences of the strand is covalently attached to the next from these groups. This polarity is indicated by abbreviating the ends of a DNA strand as 5end and 3end. The three dimensional structure of a double stranded DNA is called the double helix (Figure 1.3). The helical structure of the double stranded DNA arises from the structural features of its polynucleotide chains held together by hydrogen bonding. The underlying mechanism can best be described by the hydrogen bonding taking place between the bases. The bases of the nucleotides are always on the inside of the helical structure whereas the sugar-phosphate back bone faces outwards. Moreover, in any hydrogen bonding event, a two-ring base (also called Purines) is paired with a single-ring base (also called Pyrimidines) as shown in Figure 1.3.

Figure 1.3: (A) A space-lling model of helical double stranded DNA is shown on the left. This 1.5 turn (10.4 base pairs per turn) model shows the major and minor groves in the structure. (B) A short section of double stranded DNA is also illustrated on the right to show the real bond patterns between nucleotides. 5 and 3 ends can also be shown together with the phosphate backbone.(Image adopted from [1]). In order to fold into a structure which is energetically favorable, the length of each hydrogen bond is minimized by winding nucleotides around each other to form a helical structure which makes a complete turn in every ten base pairs. In order for the hydrogen bonding to take place each strand should also antiparallel. Thus if one strand is in a direction of 5 to 3 ends the complementary should be in 3 to 5 direction. The alignment

6 of two perfectly complementary strands and formation of hydrogen bonding in between to generate a helical double stranded DNA (dsDNA) is also called a hybridization event, and the reverse process is often called de-hybridization or denaturation.

1.1.1 Hybridization of Single Stranded and Denaturation of Double Stranded DNA Fragments In cells, during replication or other processes such as transcription, two strands of the double helix is separated from each other, if not all, in a local region. This can only be accomplished by disrupting the hydrogen bonds between base pairs. This process can be done in the laboratory by heating a solution of DNA or by adding acid or alkali to ionize its bases. When heat treated, the dissociation of one strand from the other occurs precipitously at a certain temperature range as shown in Figure 1.4. The melting temperature (Tm ) of DNA is dened as the temperature at which half the helical structure is lost. The melting temperature of nucleic acids is usually monitored by measuring their absorption of light at 260 nm because stacked bases absorb less ultraviolet light than unstacked ones (hypochromism). When the temperature is lower than Tm , complementary sequences start forming the double helix by a process called renaturation or reannealing. Other than temperature, sequence of the strands (G-C content) and the sequence length is the major factors determining how fast this process would occur.

Figure 1.4: Hypochromism investigated by spectroscopy. (A) Single-stranded DNA absorbs more light with respect to double stranded helical DNA. (B) The absorbance increases as the double stranded DNA denatures to form single stranded DNA which can be monitored in real time to determine the melting temperature Tm of any DNA sequence. (Image adopted from [2]).

1.2 Cell-free Nucleic Acids The existence of cell-free nucleic acids (cf-NA) is rst discovered in 1948 by Metais and Mandel [6]. However, unlike what one would expect, the importance of this discovery is understood rather lately, as most of the reviews on the subjects are published in the last 3 years. The spotlight could only be turned on this phenomena by 1994, with the discovery of Ras gene fragments( [7, 8]) in the blood of cancer patients. Moreover not only DNA

8 but also mRNA and miRNA fragments are shown to be in high concentrations in blood of cancer patients. Different genetic and epigenetic signatures then shown to be associated with cfNAs. The formation of cfNA in blood can be stemmed from different pathways. Apoptosis (a normal, genetically regulated process leading to the death of cells and triggered by the presence or absence of certain stimuli, as DNA damage.) and necrosis (the death of one or more cells in the body, usually within a localized area, as from an interruption of the blood supply to that part) can cause the disruption of the cell membrane, hence release of genetic material into the blood stream. Both of these occurances are common in all all types of cancer. Secretion is another way cfNA forms. Phagocytosis of the necrotic and apoptotic cells by macrophages or other scavenger cells have been shown in in vitro cell cultures and secretion of digested DNA from macrophages is studied before [9].However it has been shown that there are not enough circulating cells to justify the amount of DNA found in the plasma [10] by lysis alone. It would have to be assumed that 10,000 tumor cells per mL are circulating in the blood to account for the DNA found in breast cancer patients which is far more than has ever been found in any patient [11]. It has been proven afterwards that cellular nuclear acid fragments can also be actively released or secreted [12], as shown in Figure 1.5.

Figure 1.5: Not only mutations but also methylations, DNA integrity, microsatellite alterations and viral DNA has been shown to be associated with cf-DNA in blood. Both single stranded and double stranded DNA can be released in to the blood stream through different ways as discussed. Moreover individual tumor type can release more than one form of cf-DNA. (Image adopted from [3]). . Although the size and quantity of the total DNA fragments in blood of cancer patients varies a lot. It has been estimated that, for a tumour that weighs 100 g, which corresponds to 31010 tumour cells, up to 3.3% of tumour DNA may enter the blood every day [13]. On average, the size of this DNA varies between small fragments of 70 to 200 base pairs and large fragments of approximately 21 kilobases.

10 The composition of cfNA not only consists of DNA derived from cancer cells but also from healthy cells as well. The ratio of DNA derived from cancer and normal cells depends on the cancer type and mostly on the stage of the cancer and size of the tumor as well [3]. Moreover nucleic acids also have a half-life. They are generally cleared from the blood stream in 15 minutes to several hours [14]. The structure of the DNA is also a very important factor affecting the half-life of certain fragments. For instance double stranded DNA has a signicantly higher half-life with respect to single stranded one, as shown by Bendich et. al. [15]. Moreover, circular DNA released into the blood stream by viruses may also have a longer half-life than the linear one as postulated in the same study.

1.2.1 Concentration of Cell Free Nucleic Acids It has been shown by various studies that amount of cf-DNA in cancer patients is higher than normal subjects, however, the amount of cf-DNA in blood or plasma from both groups varies considerably. Although cancer patients have higher cfDNA levels than healthy control donors, the concentrations of overall cfDNA vary considerably in plasma or serum samples in both groups [1618]. For example, the average cf-DNA in blood of cancer patients is 180 ng per ml, however this average is calculated from a range from 0 to over 1000 ng per ml. [1922] Average concentration of cf-DNA in blood of normal subjects is approximately 30 ng per ml, an average of a population with cf-DNA concentration ranging from 0-100 ng per ml. [14]. However these results are inconclusive due to the fact that DNA is isolated using different methods in different studies and sizes of the populations used in these studies are usually not suitable to perform any powerful statistical tests. [3]

11

cf-DNA not only consists of genomic DNA, but also of mitochondrial DNA, and the level at which these two types of DNA is affected by the existence of a cancer varies as shown by diverging results [3].

1.3 Transrenal DNA Transrenal DNA (Tr-DNA) was rst discovered in 2000 by Botezatu et. al. [23]. Labeled DNA has been injected into the peritoneal cavity of mice and the distribution of radioactivity was analyzed. Fragments of 150-160 bp. were detected in urine. This is also conrmed in humans as well. For example, DNA from Y chromosome was detected in urine of women pregnant with male fetuses. [2426]. This discovery lead the way for possibility of prenatal diagnosis from transrenal DNA.

Another study demonstrated that male derived DNA fragments were present in the urine of female subjects transfused with male blood [27]. Association with cancer on the other hand was shown when mutant kRas gene fragments were detected in urine of colorectal and pancreatic cancer patients who had tumors with the same exact kind of mutations [23]. This nding was later conrmed and further investigated and conrmed in precancerous tissue by Su et. al. [28].

As expected (judging by the studies on cf-DNA) Tr-DNA also includes DNA not only from cancerous tissue but also from infections caused by pathogens. For example EpsteinBarr virus can be excreted into urine in nasopharyngeal carcinoma patients and the fraction

12 of excretion is negatively associated with the size of the DNA molecules [29]. Moreover Tuberculosis can be diagnosed by detecting mycobacterial DNA in urine by nucleic acid amplication methods [4]. Figure 1.6 illustrates how the cf-DNA in blood is ltered through the kidney and form Tr-DNA which can then be tested for diagnosis.

13

Figure 1.6: Transrenal DNA production in a patient with pulmonary tuberculosis is shown. M tuberculosis bacilli from infective foci in the lungs are destroyed by the immune response releasing cell-free nucleic acids in plasma. Filtered cf-DNA forms Tr-DNA which can then be detected. (Image adopted from [4]).

14 DNA fragments derived from tumors of different cancer types were also found in urine. A study on kRas gene fragments associated with colorectal and pancreatic cancer has already been discussed above. A mutation, a G:C to T:A transversion at codon 249 (249T) of p53 gene found to be associated with Hepatocellular Carcinoma (HCC), is detected in urine of patients with HCC. [30]Another important region leading to the same cancer type is induced by Hepatitis B Virus (HBV). 1762T/1764A double mutation of HBV is found in the urine of patients with HBV induced HCC [31]. Not only mutations, but also other epigenetic factors associated with cancer is studied in Tr-DNA. For example, hypermethylated vimentin gene promoter fragments were detected in urine of patients with CRC [32]. To summarize, DNA fragments of dying cells can be released into the blood stream by several mechanisms such as apoptosis, necrosis or secretion by active transport, and some portion (especially low molecular weight (150-250 bp.)) of this cf-DNA can pass through kidney barrier to form Tr-DNA.

1.4 DNA Biosensing Although there are many different biosensor platforms used to detect DNA, most important and study extensively by others are Polymerase Chain Reaction (PCR) dependent, Quartz Crystal Microbalance (QCM) , Surface Plasmon Resonance (SPR) , nanostructure associated (such as carbon nanotubes), Piezoelectric microcantilever (PEMS), Atomic Force Microscopy (AFM), and Electrochemical biosensors. Because PCR is the Golden Standard of any detection scheme involving DNA markers, PCR will be discussed in detail and examples from other platforms will be given including their advantages and disadvantages.

15 1.4.1 Polymerase Chain Reaction PCR can be described in one sentence as in vitro DNA polymerization of a target DNA fragment which is usually called amplicon. As the name implies ampicon is the sequence which is amplied through making copies of it, in cycle of polymerizations. One cycle of PCR includes 3 main stages. Denaturation,Primer Annealing and Polymerization/Elongation/Primer Extension. These stages separated from each other by a sudden change in the temperature, which is required to increase the rate at which each reaction occurs.

In the denaturation part of the cycle, rst double stranded amplicon is denatured by heat treatment into its constituent complementary single stranded chains as shown in Figure 1.7

Figure 1.7: Illustration of Denaturation part of a cycle of PCR reaction. (Image adopted from www.btci.org). Then in the primer annealing part of the cycle, two primers required for the enzyme Taq. Polymerase (or subunits of this enzyme in different types of PCR) anneals on each strand. Primers are designed to be complementary to the denatured sister strands from their

16 3 ends, because Taq. polymerase has only a 5 to 3 polymerization capability as shown in Figure 1.8.

Figure 1.8: Illustration of annealing part of a cycle of PCR reaction. Forward and reverse primers anneal on complementary strands of the double stranded amplicon. (Image adopted from www.btci.org). In the extension or elongation part of the cycle, Taq. polymerase enzyme binds on the -primer annealed- double stranded portion of each strand, and elongates or extends the primers with the deoxyribonucleotidetriphosphates (dNTPs) dissolved in the solution as you can see in Figure 1.9.

17

Figure 1.9: Illustration of extension part of a cycle of PCR reaction. Forward and reverse primers anneal on complementary strands of the double stranded amplicon. (Image adopted from www.btci.org). Cycles of PCR follow each other until a predetermined number of cycles is reached for amplication. As number of cycles increases the number of amplicon copied increases exponentially. Thus a chain of cycles can potentially produce millions of copies of the region intended for amplication as illustrated in Figure 1.10.

18

Figure 1.10: Illustration of extension part of a cycle of PCR reaction. Forward and reverse primers anneal on complementary strands of the double stranded amplicon. Image adopted from www.wikimedia.org where it is published by its author (Ygonaar) for sharing under both GNU Free Documentation License and creative commons by-sa licenses with some rights reserved. The reason why PCR is considered the standard technique for DNA detection is its unprecedented sensitivity. PCR has been shown previously to be capable of amplifying single DNA molecules in a reaction chamber [3337]. All of the studies described above involving cf-DNA and Tr-DNA, involved PCR in one form or another. However PCR also suffers from some issues which is still being addressed by researchers.

19 One of the main disadvantage of PCR is its requirement for DNA isolation. Like most other reactions involving proteins and enzymes, contamination inhibits the reaction catalyzed by the enzyme, in this case Taq. polymerase. Moreover, in the case of isolation of Tr-DNA from urine for PCR, things are even more complicated. It has been shown by electrophoresis that, DNA isolated from urine reveals two major fractions: High-molecularweight DNA and Low-molecular-weight DNA. High molecular weight DNA is majorly DNA from intact cells including infectious disease agents as well. On the other hand low molecular weight DNA is usually 150-200 bp in length and constitutes the Tr-DNA [23]. This nding is further supported by a study on mutant kRas DNA fragments of approximately the same length, in urines of patients with colorectal tumors [28]. Most of the current isolation kits however are not designed to isolate such low molecular weight DNA, they are desgined to isolate nuclear DNA from intact cells with high molecular weights (>350 bp.), and there is a need for a simple, cost effective method for isolation of small DNA fragments from body uids [38]. Another disadvantage when working with PCR is inhibition by co-isolated factors. Not only nucleases but also other proteins may interfere with the reaction catalyzed by PCR. Moreover amplicon size also limits the use of PCR on urine. As it is described before, the amplicon by denition should be smaller than the DNA fragments to be analyzed. However in reality the fragments in urine, usually rarely randomly fragmented, meaning a target sequence usually resides in fragments cleaved differently. For example if there are 160 copies of 160 bp DNA fragments including a target sequence of interest, only 70 of them will be amplied with a 90 bp amplicons or only 40 of them will be amplied with a 120 bp amplicons. The rest of the copies will not have full amplicons inside [39]. This can be

20 generalized to deduce that as the amplicon size increases the sensitivity of detection from Tr-DNA by PCR drops signicantly. It has been claimed by the studies of Umansky et. al. that the sensitivity of the PCR reaction for detecting the presence of Tr-DNA in urine specimens falls drastically as the amplicon size exceeds 150 bp. and in all experiments, the sensitivity was reduced to zero with amplicons of more than 300 bp. [38].

1.4.2 Other Platforms For disease screening and personalized medicine, the ability to rapidly detect multiple genetic markers and examine the genetic prole of the disease at low cost is greatly needed. Current genetic detection technologies under development rely on uorescence [40], quartz crystal microbalance (QCM) [41, 42] electrochemical [43] binding to nano-metal particles [44], surface plasmon resonance (SPR) [45], silicon-based microcantilever sensor as well as piezoelectric microcantilever sensor [46, 47]. For DNA detection, QCM exhibited a concentration sensitivity of 0.1 fM [41, 48, 49]. Direct conductivity measurement of metal nanoparticles exhibited a concentration sensitivity of 500 fM [43]. The SPR exhibits concentration sensitivity of 1 fM [45,50]. The electrochemical methods also exhibit concentration sensitivity on the order of 1 aM [42, 51] Nanowires [52] and nanotubes [5356]; exhibit concentration sensitivity ranging from 100 fM to 0.2 fM. Microcantilevers coupled with nano-metal particles exhibited 0.01 nM concentration sensitivity [57]. Although many of these methods such as QCM, SPR, siliconbased microcantilever sensor as well as lead zirconate titanate (PZT) piezoelectric microcantilever sensor [46, 47] are label-free, the sensitivity is still many orders of magnitude away from the attomolar requirement. Similarly, the 1016 M sensitivity achieved by magnetic beads isolation coupled with electro-

21 chemical enhancement was not sufcient [58]. Although nano-scale mechanical imaging by atomic force microscopy (AFM) can differentiate unhybridized ingle-stranded DNAs (ssDNAs) from hybridized double-stranded DNAs (dsDNAs) at attomolar level it requires sophisticated instrument such as AFM [59]. Although several recent studies showed attomolar sensitivity detection, they were not label-free. For example, carbon nanotubes exhibited attomolar sensitivity using streptavidin horseradish peroxidase labeling for signal amplication [60]. Electrochemical biosensor involving magnetic beads exhibited attomolar sensitivity using target DNA biotinylation [61] or electrocatalytic amplication [62]. GaN nanowire based extended-gate eld-effect-transistor exhibited attomolar realtime ensitivity [63] requiring the dilution of the buffer more than 100 times, indicating that the sensitivity before dilution was only 0.1 fM. Table 1.1 at the end of the chapter compares advantages and disadvantages of these methods and indicates the limit of detection of each method.

22

Table 1.1: Comparison of other DNA biosensing platforms Method Sensitivity Disadvantages Advantages

Quartz Crystal Microbalance

0.1 fM [48, 49] Time Consuming

Low sensitivity

Low Cost [64]

Surface Plasmon Resonance (SPR)

1 fM [50]

Expensive Low Sensitivity [65]

Multiplexed [65] 9.5min-1.5 hrs. [66]

Carbon NanoTubes (CNTs)

35 fM [56]

Low Sensitivity

Low cost [67]

Piezoelectric Microcantilevers

10 pM [57]

Low Sensitivity

Low cost

Atomic Force Microscopy (AFM)

Attomolar [59]

Expensive Equipment

High Sensitivity

Electrochemical

Attomolar [51]

Time Consuming (4 Hrs.) [66]

High Sensitivity

1.5 Piezoelectric Microcantilever (PEMS) and Piezoelectric Plate Sensor This section and its subsections are written by the approval of the inventors (Wan Y. Shih, Wei-Heng Shih, Mehmet C. Soylu,and Wei Wu ) of a yet unpublished patent dis-

23 closure entitled Piezoelectric Plate Sensor (PEPS). PEPS is a sensor that consists solely of a thin piezoelectric layer such as lead magnesium niobate-lead titanate solid solution, (PbMg1/3 Nb2/3 O3 )0.63 (PbTiO3)0.37 (PMN-PT) or lead zirconate titanate (PZT) less than 33 m in thickness thinly coated with an electrode <150 nm in thickness on either side of the plate. A PEPS may be square, rectangular, circular in shape, or hollow in the center of the plate, and 100-2000 m in its lateral dimension, and the edge of a PEPS may be completely of partially xed on a substrate for handling. A schematic of the cross-section view of a PEPS is shown in Figure 1.11.

24

Figure 1.11: A schematic of (a) a PEPS which consists of solely a piezoelectric layer thinly coated with two electrodes on the opposite sides and (b) a PEPS electrically insulated with MPS-W9 coating for in-liquid detection. For detection in air, receptors specic to the target analyte will be coated on a PEPS surface and excited electrically at its length extension mode (LEM), width extension mode (WEM), or any other planar extension mode (PEM) resonance frequency, f,typically in the range of 0.4 to 50 MHz. A schematic of rst mode LEM, WEM, and PEM vibrations are shown in Figures 1.12a-c. Binding of the target analyte to the specic receptor on the PEPS surface shifted the LEM, WEM, or PEM resonance frequency. Detection of the target

25 analyte is achieved by monitoring a PEPS LEM, WEM, or PEM resonance frequency shift.

Figure 1.12: (a) A schematic top view of the rst LEM, (b) WEM, and (c) PEM vibration where the shaded bars illustrated the initial position of the PEMS and the dash-dotted shapes illustrate the bent or extended positions. For detection in a biological uid or water, a PEPS must be electrically insulated. The detection sensitivity as dened by the relative resonance frequency shift, f/f, is governed by (1) the piezoelectric property of the piezoelectric layer, (2) the thickness of the piezo-

26 electric layer, and (3) the quality of the electrical insulation. One important reason why PEPS is chosen to detect Tr-DNA in this thesis is its high sensitivity as a result of piezoelectric layers Youngs modulus change due to binding [68]. It has been shown in this study that exural resonance frequency shift was more than 300 times larger than could be accounted for by mass loading. Moreover PEPS is proven as a solid platform which can detect 1.6x1018 M target DNA sequences by monitoring the rst longitudinal extension mode (LEM) resonance frequency shift. One major advantage of PEPS over PCR is it does not require 2 primers which has to be specic. It may be very difcult in some cases to design probes that would specically hybridize to the region on interest with no cross hybridization. Moreover the probe DNA sequence could be shorter than PCR primers just because they only need to be specic to one position with respect to forward and reverse primers both of which has to be specic. Detection also depends solely on hybridization, and amplicon size would not affect the sensitivity in the sense that occurs in PCR.

Detection using PEPS can also multiplexed as shown before.Array PEPS are used for in situ, real-time, all-electrical detection of Bacillus anthracis (BA) spores in an aqueous suspension using the rst longitudinal extension mode of resonance. [69] All-electrical detection scheme allows for detection using multiple PEPS with relays. Moreover detections using PEPS is label-free. For example label-free detection of complex body uids such as serum is shown before. Highly sensitive detection of HER2 extracellular domain in theserumof breast cancer patients by piezoelectric microcantilevers is done. [70]

27 1.5.1 Piezoelectric Property Dependence The most unique characteristic of a PEPS is that the detection sensitivity as dened by the relative resonance frequency shift, f/f, can be controlled by the improvement in the piezoelectric property of the piezoelectric layer. In PEPS geometry, the electric eld is applied in the thickness direction to excite mechanical vibrations in the length and width directions. Under such conditions, the electromechanical coupling coefcient, k31 , or the piezoelectric coefcient, d31 , are the most relevant properties. The electromechanical coupling coefcient, k31 , measures the efciency of conversion of the electric energy associated with applying an electric eld in the thickness direction to mechanical energy associated with extensional vibration energy in the length or width direction. The piezoelectric coefcient, d31 , correlates the elastic strain in the lateral direction, 11 to the applied electric eld in the thickness direction, E3 , as 11 = d31 E3 . The k31 constant can be obtained by a resonance method using the LEM or WEM resonance peak. In Figure 1.13 , the phase angle, (blue) and the impedance, z, (red) resonance spectra are shown around the WEM resonance peak of a PEPS.

28
-20
Phase Angle Impedance

100

80
Phase Angle (Deg.)

31

=0.31

60 -60

40

-80

20

3.2

3.3

3.4

3.5

Frequency (MHz)

Figure 1.13: Phase angle (blue) and impedance (red) versus frequency of a PEPS with a k31 = 0.34 around the WEM resonance peak at 3.17 MHz as indicated by the peak in the phase angle plot and the characteristic frequencies, fa and f p are the frequencies at which the minimum and maximum of the impedance occur, respectively. As can be seen, the (electrical impedance) versus f plot exhibited the WEM peak at 3.17 MHz. The z versus f plot exhibits the characteristic frequencies, fa , which exhibits a minimum in z and f p , which exhibits a maximum in z. The electromechanical coupling constant k31 can be obtained by using f p and fa to determine k2 e f f as; [7175].

Impedance (Ohm)

-40

29

2 ke ff =

2 f2 fp a 2 fp

Once ke f f is determined, k31 can be related to ke f f numerically [71]. 1.5.2 Insulation Quality Dependence 3-mercaptopropyltrimethoxysilane (MPS) coating was investigated at various pH and water contents. It was found that PEPS electrically insulated with a new MPS insulation procedure MPS-W9 yielded much better insulation results than the earlier MPS-5 insulation. The old MPS-5 insulation entails soaking a PEPS in a 1% MPS solution in ethanol at pH = 4.5-5.5 for 12 hr three times. PEPS with MPS-5 insulation typically exhibited a maximum current density larger than 600 A/cm2 as measured by cyclic voltammetry (CV) [76]. Such high current densities created a large noise background rendering the PEPS not reliable enough for in-liquid detection. The large background noise level made it hard to track the peak frequency reliably, thus leaving the sensor not responsive to binding of target analyte to the sensor surface. The new MPS-W9 insulation procedure on the other hand called for soaking a PEPS in a 0.1% MPS solution in ethanol at pH 9 with 0.5% of water for 12 hr three times, which allows the condensation of the silanol groups of the MPS to occur more effectively. As a result, a denser MPS coating > 100 nm in thickness could be achieved that reduced the maximum current density of a PEPS in a CV test to less than 2 A/cm2 [76]. As a result, the noise level as judged by the baseline difference between the spectrum of a PEPS in a phosphate buffer saline (PBS) solution and that of the same PEPS in air was much reduced.

30 A schematic of a PEPS electrically insulated with MPS-W9 is shown in Figure 1.11b.

As an example, the in-air and in-PBS resonance spectra of two similar PEPSs -one electrically insulated with MPS-5 and the other with MPS-W9 are shown in Figures 1.14 a and b, respectively. As can be seen, both PEPS exhibit the rst LEM peak around 500-600 kHz and the rst WEM peak at around 3 MHz. Note that there is little damping effect on the LEM and WEM peak frequencies of a PEPS due to the much smaller vibration amplitude compared to those of bending modes. As a result, both PEPSs in Figs. a and b exhibited LEM and WEM peaks that showed minimal change in the peak height intensity and negligible change in peak frequencies when in PBS as compared to in air. However, the PEPS coated with MPS-5, the in-PBS baseline was about 13 degrees higher than the in-air baseline at the base of the 1st LEM resonance peak (Figure 1.14 a). In contrast, PEPS coated with MPS-W9, the in-PBS baseline was only 2 degrees higher than the in-air baseline at the base of the 1st LEM resonance peak (Figure 1.14 b). The higher in-PBS baseline of the PEPS coated with MPS-5 insulation layer indicates a much higher noise level, which would reduce the sensitivity of the PEPS.

31
(a)
PHASE ANGLE (DEG.)
st

PHASE ANGLE (DEG.)

40 20 0

AIR PBS

width extension

(b)
40

AIR PBS

st

width extension

mode (LEM)

mode (WEM)

20

With MPS-5 insulation

With MPS-W9 Insulation

-20 -40 -60 -80 0 1 2 3

1
st

-20

st

length extension

length extension

mode (LEM) -40

mode (LEM)

-60

(a)
4

-80 0 1 2 3

(b)
4

FREQUENCY (MHz)

FREQUENCY (MHz)

Figure 1.14: The in-air and in PBS resonance frequency spectra of a PEPS (a) with MPS-5 (old) insulation and with MPS-W9 insulation.

1.6 Objective and Specic Aims Sensitivity and specicity are the major factors determining the performance of a biosensor. In this thesis different aspects of the development of a PEPS is studied to increase the overall sensitivity and specicity. Clinical requirements for diagnosis of cancer from transrenal DNA are very stringent. The objective of this thesis is, to develop a method that can be applied to detection of transrenal DNA Mutations satisfying the following to be practical;

Have at least attomolar (1018 M) detection sensitivity.( [77]) Be specic enough to detect mutant DNA in a background with abundant Wild Type DNA (ratio of Wild Type to Mutant can exceed 240 [28]) Be multiplexed, as most clinical conditions require detections from multiple loci for di-

32 agnosis Be robust (sample-to-answer in less than 40 min.)

For this purpose the following specic aims are achieved;

Development of a signal processing algorithm to increase the sensitivity by reducing noise in the resonance spectrum to achieve 1019 M limit of detection (chapter 2) Initial optimization of ow speed and temperature for real time mutation detection using glass slides with uorescent microspheres (chapter 3) Development of a validation methodology of specic mutation detection using uorescent reporter microspheres (chapter 4) In situ mutation detection with optimal temperature and ow speed with at least 1:250 MT/WT specicity in detecting both single mutation and double mutation (chapter 4) Development of in situ double-stranded target DNA mutation. (chapter 5, target DNA can be detected without the need for DNA isolation, concentration, and amplication)

33 2. Detection of DNA hybridization in buffer and Urine

2.1 Introduction Cell free DNA is rst discovered by Mandel and Mtais in 1948 [6], a phenomena importance of which fathomed only after its association with cancer, as mutant Ras gene fragments were discovered in the blood of patients [7, 8]. Till that time circulating DNA has been studied extensively for its diagnostic and even prognostic association with different cancer types including but not limited to, bladder [7880], breast [8191], cervical [9294], colorectal [28, 32, 95108], hepatocellular carcinoma [30, 31, 109116], lung [117127], non-Hodgkins lymphoma [128132], melanoma [133141], ovarian [142146], pancreatic [100, 147149], and prostate cancer [150159]. These studies involve detection of cancer not only from mutations in chromosomal DNA but also from mitochondrial DNA [159, 160] and viral DNA, epigenetic factors, such as methylations [79, 84, 85, 92, 93, 102, 108, 111, 121, 122, 125, 141, 142, 147, 148, 150, 151, 153, 161] and microsatellite alterations [80, 126, 136138, 162164]. The passage of circulating DNA through the kidney barrier has been neglected for many years due to the selective behavior of the nephron. As a result of this incorrect assumption, DNA fragments observed in urine mostly attributed to have originated from organs and tissues of urogenital tract. However it was later found out that low molecular weight DNA fragments can actually pass through kidneys [23] as glomerular ltrate, moreover, kidneys play a major role in preparation of clinical specimens of cell-free DNA [38]. The standard method for detecting DNA fragments in body uids is Polymerase Chain Reaction (PCR). Although PCR has the ability to amplify one

34 molecule in a reaction chamber (reference), it has limitations such as amplicon size [39] and potential inhibition by co-isolated factors. Moreover PCR depends on DNA isolation techniques specialized for transrenal DNA as low molecular weight DNA is abundant in urine. Currently, many methods of DNA isolation do not preserve the short fragments, since they are normally intended for isolation of nuclear DNA from intact cells [38]. Other genetic detection technologies under development rely on uorescence [40], quartz crystal microbalance (QCM) [41, 42], electrochemical [43] binding to nano-metal particles [44], surface plasmon resonance (SPR) [45], silicon-based microcantilever sensor as well as piezoelectric microcantilever sensor. For DNA detection, nanoparticle amplied QCM exhibited a concentration sensitivity of 1 pM [165]. Nanoparticle enhanced SPR exhibits concentration sensitivity of 10-100 aM [166]. The electrochemical methods involving nanobers and nanotubes also exhibit concentration sensitivity on the order of 30 fM [167]. Nanowires [52, 168171] and nanotubes [53, 54] exhibit concentration sensitivity ranging from fM to 1 fM. Microcantilevers coupled with nano-metal particles exhibited 0.01 nM concentration sensitivity [57]. Although many of these methods such as QCM, SPR, silicon-based microcantilever sensor as well as lead zirconate titanate (PZT) piezoelectric microcantilver sensor (PEMS) [46, 47] are label-free, the sensitivity is still many orders of magnitude away from the attomolar requirement. Similarly, the 1016 M sensitivity achieved by magnetic beads isolation coupled with electrochemical enhancement was still not sufcient [58]. Nano-scale mechanical imaging by atomic force microscopy (AFM) can differentiate unhybridized single-stranded DNAs (ssDNAs) from hybridized double-stranded DNAs (dsDNAs) at attomolar sensitivity it requires sophisticated instrument such as AFM [59]. There are also other biosensor platforms involving very sensitive

35 detection of DNA. A GaN nanowire based extended-gate eld-effect-transistor is capable of detecting attomolar concentrations of target DNA in situ [63]. Streptavidin horseradish peroxidase functionalized carbon nanotubes are used for indirect amperometric detection of target DNA (tDNA) in attomolar concentrations [60]. In another study label free carbon nanotube impedance biosensors are used to detect 100 aM of target DNA which was again lower than required levels of clinical applications [55]. Another electrochemical biosensor based on an integrated chip is shown to reach attomolar sensitivity however that detection was also not in real time [62]. Recently a disposable electrochemical biosensor based on magnetic bead amplication and target DNA modication was able to reach attomolar sensitivity. Although very sensitive and disposable, this biosensor depended on amplication and labeling [61].

PMN-PT PEPS is a new type of piezoelectric sensor consisting of a PMN-PT freestanding lm 8 mm in thickness [172] thinly coated with gold electrodes on the two major surfaces and encapsulated with a thin electrical insulation as schematically shown in 2.1. Receptor specic to a biomarker is immobilized on the surface of the electrical insulation layer. Binding of the target biomarker to the receptor on the PEPS surface shifts the PEPS length-extension-mode (LEM) (Fig.2.1b) or width-extension mode (WEM) (Fig.2.1c) resonance peak frequency, f. Detection of a target protein or DNA marker is achieved by directly immersing a PEPS in the biological uid and monitoring the LEM or WEM resonance frequency shift, Df in real time. Detection of tDNA fragments using LEM of PEPS is studied before and a limit of detection of 1 aM is accomplished. [173] Since WEM resonance frequency was higher than the LEM (due to the geometry of PEPS as discussed in

36 chapter 1), limit of detection in theory should be possible as absolute frequency shifts in resonance peak due to stress generated by hybridization would be much larger. However the noise in the WEM resonance frequency peak was also increased preventing this. In order to do sensitive detections in higher frequencies noise levels in the background should be lowered.

37

Figure 2.1: A schematic of (a) a piezoelectric plate sensor (PEPS), (b) the rst length extension mode (LEM), (c) width extension mode (WEM) vibration of a PEPS where the shaded bars illustrated the initial position of the PEPS and the dash-dotted shapes illustrate the extended positions, (d) a top-view optical micrograph. The gold color in (d) and the thin layers lining the top and the bottom of the PMN-PT are 110 nm Cr/Au electrodes. In this study, we examine the detection sensitivity of lead magnesium niobate-lead titanate (PbMg1/3 Nb2/3 O3 )0.65 (PbTiO3)0.35 (PMN-PT) piezoelectric plate sensor(PEPS) in real- time, label-free, in situ MT tDNA hybridization detection in full urine and 1x PBS

38 buffer solution without isolation and amplication. In order to accomplish this a peak determination method is developed that decreases the noise of a WEM resonance frequency peak that resulted in a detection limit of 100 zM in buffer and an even lower 50 zM in urine. HBV-DM tDNA is used as a model, since conditions on hybridization of this model is studied before in chapter 3 and is submitted for publication as of time this thesis is written, and more importantly this model is used published results on LEM resonance peak detections [173].

2.2 Experimental 2.2.1 PEPS fabrication Two PEPS (PEPS A and PEPS B) used in this study were 2.5 mm long and 0.4 mm wide. They were fabricated from (PbMg1/3 Nb2/3 O3 )0.65 (PbTiO3 )0.35 (PMN-PT) freestanding lms 8 m thickness (ref) that was coated with 110 nm thick Cr/Au electrodes by thermal evaporation (Thermionics VE 90) and cut into 2.5 mm by 0.4 mm strips by a wire saw (Princeton Scientic Precision, Princeton, NJ). Gold wires 10 m in diameter were glued to the top and bottom electrodes of each strip using conductive glue (XCE 3104XL, Emerson and Cuming Company, Billerica, MA). The rear end of the strip was xed on a glass substrate by a nonconductive glue (Loctite 1C Hysol Epoxy Adhesive) to form the PEPS geometry. It was then poled at 15KV/cm at 90o C for 60 min in an incubator (Digital Control Steel Door Incubator 10-180E, Quincy Lab). The dielectric constant of the PEPS was measured using an electrical impedance analyzer (Agilent 4294A) to be about 1800 with a loss factor of 2.5-3.7% at 1 kHz.

39 2.2.2 Electrical Insulation The PEPS were electrically insulated to stabilize the resonance peaks for in-liquid detection by a new 3-mercaptopropyltrimethoxysilane (MPS) (Sigma-Aldrich Co. LLC.) solution coating scheme involving enhanced MPS cross-linking at pH=9.0 and with the addition of water.(Ref. mehmets paper) First, the PEPS was cleaned in a Piranha solution (two parts of 98% sulfuric acid (Fisher) with one part of 30% hydrogen peroxide (Fisher)) for 1 min, followed by rinsing in water and ethanol. Before coating the PEPS with MPS at pH=9.0, we dipped the PEPS in 50 ml of a 0.01 mM MPS solution in ethanol (Fisher) with 0.5% of de-ionized (DI) water for 30 min to promote hydrolysis followed by rinsing with water and ethanol. It was then subject to 5 12-hr of MPS coating in 50 ml of a 0.1% MPS solution with 0.5% of DI water in ethanol at pH = 9.0 (adjusted by adding KOH (Fisher)). For each 12-hr of MPS coating, the PEPS was always rinsed with water and ethanol rst before being immersed in a fresh 0.1% MPS solution at pH=9.0 with 0.5% water. At the end of the 5th round of MPS coating, the PEPS was rinsed with DI water and ethanol before further coating with receptors for detection. A schematic of the cross-section of a PEPS is shown in Fig. 2.1. After insulation, the resonance spectra of the PEPS were measured using an impedance analyzer(AIM 4170 C, Array Solutions). The phase-angle-versus-frequency resonance spectra of the PEPS in air (black) and in phosphate buffer saline (PBS) solution (red) are shown in Fig. 2.2.

40

(a) 40 In Air 20
Phase Angle (deg.)

(b)
WEM

-20

In Air In PBS

In PBS

Phase Angle (deg.)

-40

-20

PEPS A
K LEM =0.32

PEPS B
-60 k
31

-40

31

=0.38

-60

-80

-80 0 1 2 Frequency (MHz) 3 4

2 Frequency (MHz)

(c) 0 -2 -4
f/f (10 )
-3

Biotin Streptavidin PBS Biotinylated Probe cDNA FRM tDNA

-6 -8 -10 -12

Reporter DNA Target DNA Biotinylated Probe cDNA Streptavidin Biotin PBS FRM

50

100 Time (min)

150

200

Figure 2.2: (a) In-air (black) and in-PBS (red) phase angle-versus-frequency resonance spectra, and (b) relative resonance frequency shift, f/f, of the PMN-PT PEPS during the various steps of probe cDNA immobilization and target DNA detection. The insert in (b) shows a schematic of the molecules involved in the immobilization and detection.

41 2.2.3 Target DNA, probe DNA, and reporter DNAs The target DNA (tDNA) used in this study was a 200-nucleotide (nt) long singlestranded DNA (Integrated DNA Technologies) containing the nucleotide sequence of the Hepatitis B virus genome (GeneBank Accession #X04615) centered around the 1762T/1764A double mutations. Part of the sequence of the tDNA containing the double mutations is shown in Table 2.1 where the two mutation sites were underlined. The probe DNA (pDNA) was a 16-nt long synthetic single-stranded DNA (Sigma) complementary to the 16-nt sequence of the tDNA centered around the double mutation sites as shown in Table 2.1. The pDNA had a biotin with a 12-polyethyleneglycol (PEG) spacer at the 5end. The melting temperature of the pDNA with the tDNA was 47 o C as estimated under the experimental conditions Refs and listed in Table 2.1. Table 2.1: The sequences and Tm of the pDNA, tDNA, urDNA and drDNA Type of DNA tDNA pDNA pDNA urDNA drDNA Sequence (5 to 3) 5-. . . GGTTAATGATCTTTGT . . . -3 5-. . . ACAAAGATCATTAACC . . . -3 Biotin-5-ACAAAGATCATTAACC-3 Amine-5-ACAGACCAATTTATGCCT ACAGCCTCCTAG-3 Amine-5-AATCTCCTCCCCCAACTC CTCCCAGTCTTT-3 Tm with tDNA(Co ) 47 76.3 77.4

To immobilize the biotin-activated pDNA on the PEPS surface, the MPS-coated PEPS was rst immersed in 200 l of 5 mg/ml of maleimide activated biotin (Maleimide-PEG11Biotin) (Pierce) in PBS for 30 minutes. The maleimide reacted with the thiol group on the MPS surface to immobilize the biotin on the PEPS surface. It was then followed by

42 immersing the PEPS in 200 l of 1 M of streptavidin in PBS to bind streptavidin to the biotin on the PEPS surface. Afterwards, the PEPS was immersed in 200 l of a 10 M solution of the probe DNA in PBS for an hour to allow the biotin at the 5 end of the pDNA to bind to the streptavidin on the PEPS surface. Steps of immobilization is illustrated in inset in Fig.2.2b. The details of the chemical reaction of the immobilization steps are contained in Fig.2.3 [173].

43

Figure 2.3:

cDNA immobilization Steps on PEPS surface:

(1) In the rst step,

the maleimide of the maleimide-PEG-biotin linker reacted with the sulfhydryl of 3(trimethoxysilyl)-propyl-methacrylate (MPS) on the MPS coating surface to form a thiolester bond that covalently linked the maleimide-PEG-biotin on the MPS surface; (2) In the second step, the biotin of the immobilized maleimide-PEG-biotin reacted with streptavidin to immobilize streptavidin on the PEPS surface; (3) In the third step, the biotin at the 5end of the cDNA reacted with the streptavidin bound on the biotin of the immobilized maleimide-PEG-biotin to nally immobilize cDNA on the PEPS surface.

44 There were two 30-nt long reporter DNAs (rDNAs) (Sigma). One was complementary to the sequence upstream of what was complementary to the pDNA and the other was complementary to the sequence downstream of what was complementary to the tDNA. The upstream rDNA was amine-activated with a 12-PEG spacer at the 5 end while the downstream rDNA was amine-activated with a 7-PEG spacer at the 3 end. The sequence of the upstream rDNA and that of the downstream rDNA are also shown in Table 2.1. In real DNA fragments, the mutation sites may be located anywhere in the fragments and in some cases the mutated sites may be too close to the edge for strong enough rDNA binding. Under such conditions, an rDNAs in the opposite stream would permit the binding of the rDNA to captured tDNA on the sensor surface. For this reason, we included both upstream and downstream rDNA in the study even though in the present synthetic tDNA the mutation sites were centrally located. The melting temperature for the binding the upstream rDNA (urDNA) to the tDNA was 76.3 o C and that of the downstream rDNA (drDNA) to the tDNA was 77.4 o C, which are also listed in Table 2.1. Fig.2.4 is a schematic illustrating the relations between the tDNA, the pDNA, the urDNA, and the drDNA.

45

Figure 2.4: A schematic of the immobilization steps and nucleotide sequences of target DNA and its hybridization on pDNA, upstream rDNA and downstream rDNA.

2.2.4 Urine Sample Preparation Urine samples were collected in a 50ml centrifuge in a rst morning sample collection fashion after emptying the bladder the previous evening. Samples are kept in 4o C refrigerator for detection. Blocking of non-specic binding is accomplished by dissolving 3% BSA in urine equilibrated to room temperature.

46 2.2.5 Flow Setup All the tDNA detections were carried out in a ow. A schematic of the ow system consisting of a polycarbonate detection chamber 18.5 mm long 3.5 mm wide and 5.5 mm deep (volume = 356 l), three reservoirs, and a peristaltic pump (Cole-Parmer 7712062) interconnected with 0.8-mm wide tubing as illustrated in Fig.3.3b and c. The PEPS was vertically placed in the center of the ow in the detection chamber with its major faces parallel to the ow. In each detection event, only one reservoir was connected to the detection chamber. The total volume of the liquid was 50 ml including the liquid in the reservoir, the detection chamber and the connecting tubing. In what follows, all detections were carried out with a ow rate of 1 ml/min corresponding to an average ow velocity of 1.4 mm/s at the PEPS surface. Furthermore, in this setup, the detection could transition from one detection experiment involving the sample in one reservoir to another detection experiment involving the sample of another reservoir by turning the valves. Typically, a 20-second period for valves turning without data recording was sufcient for a smooth transition from one detection experiment to another.

2.2.6 Blocking of Non-specic Binding in Urine In order to block binding of non-specic binding on PEPS in urine, PEPS surface is treated with Bovine Serum Albumin (BSA) following pDNA immobilization. This is done by dipping the PEPS in a 100 l of 3% BSA dissolved in 1 x PBS.

47 2.2.7 Data acquisition A detection experiment involves recording of impedance spectrum of a resonance frequency with time continuously until the recording of data is stopped by the user. Fig.2.5 illustrates the screen of the laptop in a real detection experiment. Recording of impedance spectrum of a resonance frequency is achieved by scanning a given frequency window (W) using the impedance analyzer (AIM 4170 C, Array Solutions). As can be seen in gure 2.5, there is always an inherent noise in an impedance spectrum due to the noise in the impedance analyzing device. Since the analyzer itself can be simulated with a RLC circuit, as any RLC circuit it generates a noise. The difculty is then to reduce this noise and nd out the real peak position of the resonance frequency peak of PEPS. In order to achieve this goal, rst, each impedance scan is time labeled and recorded in a text le. A peak nding algorithm determines the peak position of the resonance peak and adjusts the start and stop frequencies for the next frequency range to be scanned without changing the window size (frequency difference between a start and stop frequency) such that the next peak position remains closer in the window scanned. This Moving Window algorithm is developed to keep pace with the moving resonance peak during a detection experiment, due to binding on the PEPS surface. Simultaneously, a time versus peak position graph is plotted and recorded in another text le for each resonance peak monitored every time a scan is nished so that user can monitor the frequency shift in real time. The designed algorithm also enables use of a relay for multiplexed detection setups involving monitoring resonance peaks of multiple PEPS simultaneously. The order of resonance frequency scans is given by the user as an input before the detection software is run so that after each scan the software automatically adjusts the relay and for the next PEPS. The software includ-

48 ing the peak nding and peak monitoring algorithms are completely automated until the user stops it, requiring no input from the user during detection. To summarize the inputs for the software developed for multiplexed resonance frequency monitoring are the number of PEPS used in the detection, the sequence indicating the order for impedance scans PEPS used (if more than one PEPS is used), the initial start and stop frequencies for each resonance peak to be monitored, and the le name and position for recording the raw data.

Figure 2.5: Three resonance frequency peaks are monitored in real time. Each resonance peaks raw impedance spectrum is plotted on top and their corresponding time versus peak position graph is plotted in the bottom. Peak positions are determined using the peak determination method discussed below

49 2.2.8 Resonance peak frequency determination The control of the resonance spectrum measurement, the determination of the resonance peak frequency, and the output of the resonance frequency with time were programmed in MatLab and were completely automated. To start detection, an initial start frequency and an initial stop frequency are input to dene the initial frequency window in which the resonance spectrum is measured. To determine the resonance frequency from the measured spectrum, it is important to dene an origin around which data points will be chosen to t second order polynomials. A simple way to dene origin is to use the peak of the raw impedance spectrum as illustrated in Fig.2.6a. This can be followed by tting parabolas of different neighborhood sizes around the origin. A typical resonance spectrum includes 1000 data points and maximum number of data points around the origin is typically 480 points from either side of the origin. Parabola tting is carried out over a range of data points around origin with multiples of ten up to the maximum available. To illustrate, 20 points are chosen around the origin to t a parabola as shown in Fig.2.6a and 390 points is chosen in Fig.2.6b on the same spectrum. Parabolas tted are shown in pink.

50
(a)
-9.5
Maximum of raw data (origin of neighbourhood)

Raw Data Peak position Neighbourhood Fitted Data

(b)
-9.5

Raw Data Peak position Neighbourhood Fitted Data

Maximum of raw data (origin of neighbourhood)

Phase Angle

-10.0

-9.40 -9.45
Phase Angle

Phase Angle

-9.45
Phase Angle

-9.50 -9.55 -9.60 -9.65

-10.0

-9.48

-10.5

-9.51

-9.54 3.4194 3.4197 3.4200

-11.0 3.41

3.4165

3.4170

3.4175

Frequency (MHz)

Frequency (MHz)

3.42

3.43

-10.5 3.41 3.42 3.43

Frequency (MHz)
Window (W)

Frequency (MHz)

Figure 2.6: Phase angle versus frequency of WEM resonance spectrum of a PEPS. black full circles represent the raw data. The blue data points represent a neighborhood of 10 (a) and maximum neighborhood available (b) for parabola tting (blue, in insets), raw peak value is indicated by an arrow on graphs, , polynomials tted on these neighborhoods (pink, in insets) and peak positions of the tted polynomials (red circles). After outlier removal by an outlier removal method (American Society for Testing and Materials E178, k=1.5), the resonance frequency of the measured spectrum is nally obtained by averaging all the peak values of the tted parabolas (excluding the outliers) as illustrated in Fig.2.7d with black squares. A typical number of peak values used for averaging is close to 354. In order to maximize the predictive power of this method it is obvious that one needs a resonance peak spectrum which has a peak in the middle of the scanning frequency window which will maximize the available data points for parabola tting. It is important here to note that data points chosen for parabola tting is always dened by equal number of data points around a pre-determined origin. Thus, the distance

51 of the real peak frequency of a resonance peak from the midpoint of the frequency scan window is inversely proportional to the number of parabolas that are tted on available data points around the origin. The raw peak frequency may be further from the real peak position as can be seen in Fig.2.6a. Thus instead of choosing this point as the origin one may rst smooth the raw impedance spectrum by local regression using a weighted linear least square t to a second-degree polynomial model with a span of 0.1 (2.7a). Parabola tting can be done exactly as explained before on this smoothed spectrum instead of the raw spectrum as illustrated in Fig.2.7b and c (tting on 20 and 430 data points respectively). As can be seen in Fig.2.7d this will not only increase the number of parabolas that can be tted on the same spectrum with respect to an origin dened as the maximum raw spectrum but also decrease the variation in peaks of the parabolas tted, thus decreasing the noise in determination of peak frequency from a given impedance spectrum.

52

Figure 2.7: a. Maximum of raw and smoothed data is shown in blue and pink respectively. b. Polynomial tting is done on a neighborhood of 41 points on smoothed spectrum. c. Polynomial tting is done on maximum available neighborhood of on smoothed spectrum. d. Neighborhood size versus peak of polynomials tted on smoothed (black) and raw (red) data is plotted. Another problem during a real detection experiment is to keep the origin determined after smoothing the raw frequency in the middle of the frequency scan window. As analytes

53 bind or hybridize on the PEPS surface the resonance frequency shifts to higher or lower frequencies depending on the stress generated. In order to overcome this situation the following formula is used to determine the start and stop frequencies of the next frequency scan window;
i i i 2 f peak + fstart fstop

i+1 fstart

= =

2
i i i 2 f peak fstart + fstop

i+1 fstop

+1 i+1 where fi start and fstart are the start and stop frequencies of the next frequency scan win-

dow respectively as determined from start and stop frequencies of the current frequency
i scan window (fi start and fstart , respectively). This formula measures how far away the peak

position of a resonance peak from the middle of the frequency scan window and moves the scanning window in the next scan such that the current resonance frequency peak will be in the middle in the next scan. This algorithm makes sure as long as the shift of the resonance peak stops at the steady state of the hybridization/binding, the resonance frequency peak will be kept in the middle of the scanning window. This generated a lag in algorithm to estimate the resonance frequency peaks position. As a future work this lag can be minimized using improved algorithm taking invaluable information from historical data on the time versus frequency shifts of the current and past detection experiments into account. Fig.2.8a below shows the resonance frequencies raw peak positions versus time during hybridization on PEPS surface of 10 M tDNA dissolved in 1x PBS. This hybridization experiment is monitored using a xed scanning window regime and the moving scanning window regime discussed above. Impedance spectra recorded using xed window and

54 moving window are plotted before hybridization (t=0 min) and after hybridization (t =30 min) as illustrated in Fig.2.8b.
(a) 0
Frequency Shift (kHz)

(b)

9.0 8.5 8.0

-3

Phase Angle

-6

7.5 7.0 6.5 6.0

-9

Fixed Window t=0 Fixed Window t=30 Moving Window t=0 Moving Window t=30

-12

-15

5.5
0 5 10 15 20 25 30

3.39

3.40

3.41

3.42

3.43

Time (min)

Frequency (MHz)

Figure 2.8: (a) Detection of 10 M of tDNA in 1x PBS is plotted using raw peak positions of phase angle versus frequency plots of WEM resonance peak. (b) Impedance spectra of resonance frequency peaks(red stars in (a)) at time 0 (lled symbols) and 30 (open symbols) min. are plotted for the xed (red circles) and moving (black squares) window impedance scanning methods

2.2.9 Re-plotting Detection Experiments Detection experiments can be re-plotted using a MatLab routine developed enabling access to raw impedance spectra and time dependent behavior of peak positions of different experiments and different peaks simultaneously. Frequency shifts are plotted in kHz versus time as can be seen in Fig. 2.9. Moreover user can plot experiments done in an order (such as different steps of immobilization and/or tDNA or FRM hybridization steps) in one frequency shift versus time plot. Each experiment is plotted and the frequency shift

55 in that experiment is written on the curve together with the le name assigned to record that experiments raw impedance spectra data. As can be seen in Fig.2.9, background signal checking in 1xPBS, covalent binding of maleimide activated biotin, conjugation of streptavidin and biotinylated pDNA and hybridization of tDNA followed by a second background signal check in 1x PBS is plotted for the length and width extension modes of the same PEPS. User can also click on any of the data points on time versus frequency shift graphs of any peak and the raw impedance spectra at that point is automatically drawn on the right of the graph as illustrated in Fig.2.9.

56

Figure 2.9: Detection experiment involving simultaneous monitoring of 7 peaks of the same PEPS is done. Experiment consisted of the following steps; background signal checking in 1xPBS, maleimide activated biotin binding, streptavidin binding, biotinylated pDNA binding, tDNA hybridization and second back ground signal checking in 1 x PBS. Two peaks (LEM and WEM) is drawn in this re-plotting. A sample raw impedance spectra of each peak is drawn on the right of time versus frequency shift graphs for peaks

2.2.10 Simulations To determine the performance of the resonance peak frequency determination explained above, it is tested on a simulated detection experiment. The known resonance frequency shift versus time plot was is generated using a sigmoidal function (Fig.2.10a). Each data point on this plot represented the frequency of a resonance peak at a given time, and each

57 of these peaks is modeled using a second, or and third order polynomial and or a Gaussian function (Figures 2.10b,c,d). Gaussian function is adjusted so that peak position and iniction points intersect the second order polynomial (Fig. 2.10b). For each resonance frequency peak model, average increment in phase angle in used dened as 1 a unit of noise parameter. Normally distributed random noise is generated using multiples of this unit as the standard deviation. The noise is added on the modeled peaks and peak nding algorithm is used to determine the peak positions. Window size is another parameter, effect of which is studied by simulations. A frequency window is dened as the difference in frequency between start and stop frequencies of each impedance scan. Third order polynomials are used to generate asymmetrical peaks. The ratio of distances of start and stop frequencies to the peak position of the modeled third order function is used as the parameter, R, representing extend a measure of asymmetry of a generated model function. (Please observe that, if Pn (x) is dened as set of all polynomials of degree n, than P(x)|P(x) P3 (x); 1 < R < 2, Fig.2.10d)

58
(a)
Peak positions determined by data analysis
Phase Angle (deg.)

(b)
Gaussian

-30

Second order

Real peak positions

Frequency (MHz)

3.465

-45

3.462 f

-60

f
calculated

Real

3.459

-75

3.456 0 5 10 15 20 25 30

-90 3.84 3.92 4.00 4.08 4.16

Time (min)

Frequency (MHz)

(c)
60

(d) -20
-30
Phase Angle (deg.)

R=a/b 0 -60 -120 -180 -240 -300 1.5 2.0 2.5 3.0 3.5 a b (3,-40)

-40 -50 -60 R=1 -70 -80 -90 3.35 3.40 3.45 3.50 3.55 R=1.5 R=1.99

Frequency (MHz)

Figure 2.10: a. Simulated model of a detection experiment representing graph of time versus resonance frequency peak position. b. Gaussian and second order simulated peaks without noise. c. A third order polynomial, with R representing the extend of asymmetry of polynomial. d. Simulated model of third order polynomials with different R values.

59 2.3 Results and Discussion 2.3.1 Analysis on Simulation Parameters Width extension mode resonance frequency peaks are almost always asymmetrical due to the increase in baseline phase angle at high frequencies due to the inductive effect. Any circuit can be modeled as a serial RLC circuit, phase angle of which is can be calculated from; j L R
1 jC

where is phase angle, is angular frequency which is 2. .f, where f is frequency, L is inductance, R is resistance and C is capacitance. As the frequency increases while effect of capacitance gets smaller, effect of inductance on the phase angle increases. This, asymmetry is modeled using a third degree polynomial with a degree of asymmetry measured by the parameter R. The higher the R, value the more asymmetric the simulated resonance peak model is. Two types of error are measured in each simulation, relative error and error in frequency shift (f). For any simulation, frequency shift is calculated from the difference in frequencies of the peak positions with highest and lowest frequency peak positions (E f = f real - f calculated , where E f is error in frequency shift). Relative error is calculated by averaging the differences in peak positions in the simulated and calculated resonance frequency peak positions for every time values. As can be seen in Figures 2.11 and b, after 10 simulations although the relative error increased with increasing R, standard deviations did not change signicantly, whereas error in frequency shifts are not affected on average, however increased in standard deviation. This shows the peak monitoring algorithm

60 is capable of measuring the frequency shifts correctly although the calculated resonance frequency peaks are not close to the real values. This is due to the fact that the resonance frequency peaks retains their shape and asymmetry is not changing in an experiment due to binding, and even though there is an error due the asymmetry, because almost the same amount of error is retained in each peak position calculation, the net movement in the peak position is also retained. Figures 2.11c and d shows that, as the noise added on each modeled resonance frequency peak is increased both relative error and error in frequency shift increases. Window size on the other hand showed an increased effect in relative error in the asymmetrical model, with respect to the symmetrical models. However this effect is not seen in error in frequency shift as can be seen from Figures 2.11e and f.

61
(a) 0.015 0.06
f (kHz) Error in

(b) 0.08

0.012
Relative Error

0.04

0.009

0.02

0.006

0.00

0.003

-0.02 0.000 1.0 1.2 1.4 R

(c) 0.08 Third order (R=1.9) 0.06
Relative Error

1.6

1.8

2.0

-0.04 1.0 1.2 1.4 R

(d) 1.2 1.0
f (kHz)

1.6

1.8

2.0

Third Order (R=1.9) Second Order Gaussian

Second Order Gaussian Window Size = 30 in all cases

0.8 0.6 0.4 0.2 0.0 -0.2 0

Window Size = 30 kHz in all cases

0.04

0.02

0.00

Error in

20

40

60 Noise

80

100

20

40

60 Noise

80

100

(e) 0.030 Third order (R=1.9) Second Order

Relative Error

(f) 0.08 Third order (R=1.9) Second Order Gaussian

0.025

0.06
f (kHz) Error in

0.020

Gaussian Noise =10 in all cases

0.04

0.015

0.02

0.010

0.00

0.005

-0.02 0.000 0 5 10 15 20 25 30 35 40 45 -0.04 0 5 10 15 20 25 30 35 40 45

Window Size (kHz)

Window Size (kHz)

Figure 2.11: Error analysis of the peak monitoring algorithm by simulations is shown. a. Effect of parameter, R, on relative and frequency shift errors are plotted (a and b). Effect of Noise added on simulated resonance peaks on relative and frequency shift errors are plotted (c and d). Effect of Window Size on relative and frequency shift errors are plotted (e and f).

62 2.3.2 BSA blocking In order to determine the amount of BSA required to block the PEPS surface for detections done in urine, resonance peak frequency of PEPS treated with different amounts of BSA is monitored for non-specic binding in urine using a ow rate of 1ml/min. This is followed by washing with 1x PBS using a 6ml/min ow rate. It is important to determined the least amount of BSA required for completely blocking non-specic binding because as the concentration of BSA used to block the PEPS surface is increased the sensitivity of PEPS also decreases as pDNA immobilized on the surface is hindered by the bound BSA. In Fig. 2.12 relative resonance peak frequency shift in non-specic binding and washing steps of a PEPS treated with no BSA and 1, 2, and 3 % is plotted time. Relative resonance peak frequency shift refers to the ratio of the resonance peak frequency shift to the position of the resonance peak frequency. As can be seen 3% BSA totally blocked the PEPS surface, enabling detection in urine.

63

0.0

-0.1
f/f (10 )

-0.2

-3

-0.3

No BSA 1% BSA

-0.4

2% BSA 3% BSA

-0.5 0

10

20

30
Time (min)

40

50

60

Figure 2.12: Relative frequency shifts of PEPS B in urine treated after 1,2 and 3 percent BSA and no BSA followed by a washing step in 6ml/min ow rate.

2.3.3 Dose response in 1 x PBS and urine and visual conrmation Detection of tDNA concentrations from 100zM to 10 nM and from 50 zM to 10 nM are done in 1x PBS and in urine. tDNA hybridizations are followed by hybridizations of FRMs. Fig.2.13a and b show the relative resonance peak frequency shifts during hybridization of different concentrations of tDNA. Average relative frequency shifts in the last ve

64 minutes of each hybridization reaction (25th -30th min. and 55th -60th min for hybridizations of tDNA and FRMs respectively) are plotted versus tDNA concentration for in urine and 1 x PBS detections in Fig.2.13. Relative frequency shifts after FRM hybridizations are also plotted versus number of FRMs hybridized on PEPS surface as can be seen in Fig.2.13d. This indicated a direct proportion between relative frequency shifts in FRM hybridizations and number of FRMs htybridized on PEPS surface. As can be seen in Fig.2.13c, although in semi-log plot of concentration versus relative frequency shift graph there is a linear behavior between concentrations of 100 zM and 100 pM in detections in both urine and 1 x PBS, there is a steep decrease in the relative frequency shift in 50zM concentration with respect to 100 zM. This change in relative frequency shift can be best described by a threshold concentration mechanism, that when exceeded, generates detectable surface stress on PEPS surface. Moreover hybridization of FRMs on tDNA on PEPS surface generated almost an equal 1.2 times higher frequency shifts which occurred in tDNA hybridizations.

65
tDNA in Urine 0 -1 -2
) f/f (x10 )

(a)

FRM

(b)

tDNA in PBS

FRM

-2
3

f/f (10

-3 -4 -5 -6 -7 10 nM 100 pM 10 fM 10 aM 100 zM 0 15 1 fM 1 aM 50 zM 30 Time (min) 45 Control 60

-3

-4

-6

Control 100 zM 10 aM 1 aM 100 aM 1 pM 100 pM

-8 0

10 fM

10

20

30 Time (min)

40

50

60

(c) 4 tDNA in Urine FRM in Urine tDNA in PBS 3 FRM in PBS

(d) VI 80 V 60
# FRMs

f/f (10

-3

IV 40

20 I II

III

1E-18

1E-15

1E-12

1E-9

0.0

0.4

0.8 f/f (10 )

-3

1.2

1.6

tDNA Concentration (M)

Figure 2.13: Standard curve of HBV DM detection in urine (a) and in 1xPBS (b) with PEPS A and PEPS B respectively at room temperature with 2ml/min ow rate. (c) Relative frequency shifts at 30 min(after tDNA hybridization) and 60 min (after FRM hybridization) are plotted versus concentration of tDNA. (d) FRM micrographs after tDNA and FRM hybridizations in urine for control, and concentrations; 50zM, 100 zM, 1 aM, 10 aM and 100 aM of tDNA. Micrographs of PEPS surface after detection of some of the lowest concentrations in

66 urine are shown in Fig.2.14 . As can be seen from the images FRMs are not aggregated and are evenly distributed on PEPS surface at high concentrations of tDNA. Number of FRMs on PEPS surface at these concentrations are plotted in Fig.2.13d.

Figure 2.14: FRM micrographs after tDNA and FRM hybridizations in urine for control, and concentrations; 50zM, 100 zM, 1 aM, 10 aM and 100 aM of tDNA. Width of PEPSB is 420 m and is shown by dashed parallel lines in micrographs.

67 2.3.4 Simulation Results Random noise added on the simulated resonance peak frequencies are generated using normally distributed pseudorandom numbers as described before. By comparing the standard deviation of raw peak positions of simulated resonance peak frequencies generated by adding pseudorandom numbers with different standard deviations on a xed second order polynomial, with standard deviation of a WEM resonance peak frequency monitored in 1 x PBS where there is no hybridization on PEPS surface. In 1 x PBS any frequency shift in resonance peak is considered as background noise. Standard deviation in raw peak positions is chosen as the criteria to compare the simulations with real noise in the background to eliminate any effect from signal processing. Fig.2.15illustrates the noise versus standard deviation of raw peak positions of such simulated peaks and standard deviation in raw peak positions of WEM resonance peak in 1 x PBS is shown with a horizontal dashed line.

68
(a) 5 (b)
0.8 4 Experimental 0.4

Experimental Simulation

( f/f, 10 )

-4

f/f (10 )

-3

0.0

-0.4 1 -0.8 0 8 16 Noise 32 64 0 20 40 60 80 100 120 140 Time (min)

Figure 2.15: (a) Noise parameter of simulated detection experiments with zero frequency shifts are plotted versus corresponding standard deviation of the raw peak positions for each simulated experiement. Standard deviation in raw peak position of WEM resonance peak of a PEPS in 1xPBS is shown with dashed line. (b) Experimental data in (a) and simulated experiment with noise parameter 27 is plotted. This value is used for further simulations In order to assess the effect of resonance peak frequency determination method developed a signal to noise ratio(SNR) is dened as the ratio of relative frequency shift in the last 5 minutes of any hybridization detection experiment to the standard deviation of resonance peak position in 1 x PBS. WEM of PEPS A is monitored for tDNA detections of different concentrations in 1 x PBS. Same WEM resonance peak is monitored using a xed window scanning method and the resonance frequency peak determination method discussed in this study simultaneously. In Fig.2.16a, three different methods are applied to determine the peak frequency, and SNR is plotted versus tDNA concentrations from 100 zM to 10 nM. In method one, raw peak positions are determined from xed window scan-

69 ning method. In the second method peak positions determined by xed window scanning and tting 100 data points around the raw peak position of each scan. In third and nal method peak determination methods discussed in this study is applied. As can be seen resonance frequency peak determination method results in an order of increase in SNR with respect to to method one and 4-5 times increase with respect to method two. In order to estimate the error done in relative frequency shifts simulations with the noise level determined in Fig.2.15a and b is used to generate simulations of detections of different tDNA concentrations. Percentage error of frequency shifts in the last ve minutes of each simulation is calculated using the following formula; | fsimulation factual | factual where fsimulation is the average of relative frequency shift in last ve minutes (25th min and 30th min) in each simulated tDNA detection

70
(a)
Peak Determination tDNA

(b) 12 10 8
f (%)

Peak Determination FRM Single Parabola tDNA Single Parabola FRM

100

Raw tDNA Raw FRM

50zM

SNR

f -

10

f |/ |

6 4 2 0 -2

1E-18 1E-16 1E-14 1E-12 1E-10 tDNA Concentration (M)

1E-8

0.3

0.6

0.9
-3 a

1.2

1.5

1.8

f /f(10 )

Figure 2.16: (a) Dose response experiment in 1x PBS is done recording peak resonance frequencies applying the peak determination algorithm and a xed window schem simultaneously. SNR of xed window resonance frequency monitoring is plotted versus concentration of tDNA with single parabola tting and raw peak positions and compared with that of peak resonance frequency determination method (b) Percentage of error in relative frequency shifts between 25th and 30th minutes are plotted using simulated resonance peaks with the same noise in experimental peaks.

2.4 Summary of Resonance Frequency Peak Determination We have investigated real-time, in situ DNA hybridization detection using piezoelectric plate sensors (PEPSs) consisting of a highly piezoelectric lead magnesium niobatelead titanate (PMN PT) layer 8 mm in thickness thinly coated with Cr/Au electrodes and electrically insulated with 3-mercaptopropyltrimethoxysilane (MPS)encapsulation. With probe complementary DNA (cDNA) immobilized on the PEPS surface and by monitoring the rst

71 width extension mode (WEM) resonance frequency shift of the PEPS we showed that we could detect hybridization of the target DNA (tDNA) to the probe DNA on the PEPS surface at 50zM in urine and 100 zM in 1 x PBS with a signal to noise ratio of 14 and without isolation and amplication, which was validated in situ by the detection of uorescently labeled microspheres coated with reporter DNAs complementary to the tDNA but different from the probe DNA following the detection of the tDNA. Simulations are done to measure the effectiveness of the developed method and a 5 fold increase in SNR is shown with respect to single parabola tting and less than 12% error in nal frequency shift in observed at the limit of detection. It has also been shown by simulations that, asymmetrical nature of WEM is not affecting the nal frequency shift after each experiment.

72 3. Flow-Enhanced Detection Specicity of Mutated DNA against Wild Type with Reporter Microspheres

Cancer is a genetic disease and gene mutation is an important form of genetic defects that play an important role in cancer pathways. Detecting gene mutation is essential for cancer diagnosis, cancer therapy decision, as well as therapy efcacy monitoring. Many genetic cancer markers are known to circulate in body uids such as serum and urine. Detecting circulating genetic markers in serum or urine is minimally invasive or non-invasive, which can be an integral part of the therapy monitoring when the primary tumor is removed or hard to get to. One challenge of detecting cancer genetic markers in serum or urine is that the wild type shed by the normal cells may be far more abundant than the defected gene shed by the cancer. For example, the wild type (WT) kRas gene is known to outnumber the mutant (MT) kRas gene by a factor of 240. [39] Gene mutation is among the most challenging to detect as one must be able to detect and differentiate the MT from the WT given that the WT outnumbers the MT and the genetic difference between the MT and WT is often one nucleotide. Currently WT can be discriminated using methods such as denaturing gradient gel electrophoresis, [174] temperature gradient gel electrophoresis, [175] single strand conformation polymorphisms, [176178] heteroduplex analysis, [179] chemical cleavage method, [180, 181] protein truncation tests, [182, 183] DNA chips [184, 185] and high resolution melting temperature analysis. [186] All of the above methods are based on the fact that the hybridization kinetics between the WT and the probe DNA (pDNA) is different from that between the MT and the pDNA. For a pDNA that is perfectly com-

73 plementary to MT, the melting temperature of the MT, the de-hybridization temperature between the pDNA and the MT, TMT , is higher than the melting temperature of the WT the de-hybridization temperature between the pDNA and the MT, TW T . This melting temperature difference is important in differentiating the MT from the WT. Typically MT is detected at a temperature between TMT and TW T . [186188] While raising the temperature within the TMT - TW T window may minimize the binding of the WT to improve specicity; it can also reduce the binding of the MT, thus reducing the sensitivity. Some single-nucleotide mutations may have a TMT - TW T window as small as less than 1 C, [189] making it difcult to differentiate MT from WT by the temperature means alone. Recent development of locked nucleic acid (LNA) pDNA that can help widen the window between TMT and TW T is one way to help better differentiate MT from WT. [5] Fluorescent polystyrene microspheres have been established as a means of detecting DNA hybridization and conjugation of DNA stabilizes the surface charge of microspheres. [190] In a systematic study involving shear ow, microspheres were coated with single stranded DNA to hybridize with pDNA on a surface under the inuence of various levels of shear stress. It was shown that shear stress played an important role similar to that of temperature in that DNA became increasingly de-hybridized with an increasing shear stress and that there existed a critical shear stress above which microspheres became detached from the surface and that DNAs with a single mismatch exhibited a somewhat lower number of attached microspheres per unit area. [191] In addition, ow had also been shown to minimize cross-binding of closely related species of bacillus anthrax (BA) such as B. thuringiensis (BT), B. cereus (BC), and B. subtilis (BS) to anti-BA antibody immobilized on a sensor situated at the center of a laminar ow. [69] Fluid ow in a narrow channel

74 has also been shown to help enhance detection sensitivity by reducing nonspecic binding. [192194] Therefore, it seems possible that uid ow may be an auxiliary tool to further improve specicity of mutation detection. The purpose of this study was to investigate how ow can affect the specicity of mutation detection in addition to temperature with the help of uorescent reporter microspheres (FRMs). The model mutant (MT) gene was the 1762T/1764A HBV double mutation which is present in 50-85% of hepatocellular carcinomas (HCC). [195199] The HBV-DM MT gene has two mutated sites that are close to each other as shown in the schematic in Fig. 3.1a and in Table 3.1. A gold-coated glass (GCG) with pDNA complementary to the MT covalently immobilized on its surface was vertically immersed in the center of a laminar ow of the MT or WT solution for the MT or WT to hybridize to the pDNA on GCG. Separately, FRMs are covalently coated with reporter DNA (rDNA) that is complementary to MT and WT but different from pDNA (see the schematic in Fig. 3.1 and Table 3.1). After MT or WT binding, the GCG was immersed in a ow of FRMs for the FRMs to hybridize to the captured MT or WT on the GCG. By varying the ow rate at various temperatures between TMT and TW T we will be able to determine if ow helps improve the specicity of mutation detection.

75

Figure 3.1: A schematic of the nucleotide sequences of (a) MT and (b) WT and their hybridization on pDNA, upstream rDNA and downstream rDNA

3.1 Experimental 3.1.1 Target DNAs, probe DNA, and reporter DNAs The MT used in this study was a 200-nucleotide (nt) long single-stranded DNA (Integrated DNA Technologies) containing the nucleotide sequence of the Hepatitis B virus genome (GeneBank Accession #X04615) centered around the 1762T/1764A double mutations. [31] Part of the sequence of the MT around the double mutations is shown in Table 3.1 where the two mutation sites were denoted by the underline. Partial sequence of the 200-nt long WT is also shown in Table 3.1. 16-nt long synthetic single-stranded pDNA and 30-nt reporter DNA (rDNA) were purchased from Sigma. The pDNA was complementary to MT targeting the 16-nt sequence focused around the double mutation sites of the MT. The sequence of the pDNA is also shown in Table 3.1 The pDNA was amine-activated and

76 had a 12-ploethyleneglycol (PEG) spacer at the 5 end. The melting temperature of the MT with pDNA was 47 C and that of WT with pDNA was 23 C as estimated using salt adjustment for phosphate buffered saline (1 x PBS) [200, 201]. These two melting temperatures are also listed in Table 3.1. Table 3.1: The sequences and corresponding melting temperatures adjusted with salt concentration in 1xPBS Type of DNA Sequence (5 to 3) Tm (Co) a MT 5-. . . GGTTAATGATCTTTGT . . . -3 47 WT 5-. . . GGTTAAAGGTCTTTGT . . . -3 23 pDNA Biotin-5-ACAAAGATCATTAACC-3 UpstrDNAb Amine-5-ACAGACCAATTTATGCCTACAGCCTCCTAG-3 76.3 DwstrDNAc Amine-5-AATCTCCTCCCCCAACTCCTCCCAGTCTTT-3 77.4 a Mutation sites indicated by underlines,b Upstream rDNA,c Downstream rDNA The sequences of the pDNA, MT, WT, upstream rDNA and downstream rDNA and the melting temperature, Tm, of the MT with pDNA, of the WT with pDNA, of the upstream rDNA with MT or WT, and of the downstream rDNA with MT or WT.

There were two different 30-nt long rDNAs. The rDNAs were complementary to the sequence of the MT or WT upstream or downstream of the 16-nt sequence that was complementary to the pDNA. The upstream rDNA was amine activated with a 12-PEG spacer at the 5 end and the downstream rDNA was also amine activated but with a 7-PEG spacer at the 3 end. The sequence of the upstream rDNA and that of the downstream rDNA are also shown in Table 3.1. The melting temperature of the upstream rDNA to the MT/WT was 76.3 o C and that of the downstream melting temperature of the upstream or downstream rDNA to the WT was the same as that of the upstream or downstream rDNA to the MT. The melting temperatures of the two rDNAs with the MT and those of the rDNA to the

77 WT are also listed in Table3.1. Fig 3.1a and 1b are schematics illustrating the relationship between MT, pDNA, and upstream and downstream rDNAs and that between WT, pDNA, and upstream and downstream rDNAs, respectively. The upstream and downstream rDNAs were designed to have much stronger binding to the target MT or WT than the pDNA to the target MT or WT. Therefore, when unbinding due to the ow-induced impingement force occurred it would occur at the binding sites between the pDNA and the MT or WT but not at that between the rDNA and the MT or WT. Because the binding of the pDNA to the WT was much weaker than that of pDNA to the MT, theoretically, the ow-induced impingement force could more easily overcome the weaker binding between the pDNA and the WT than that between the pDNA and the MT to allow us selectively detect MT but not WT.

3.1.2 Substrate Preparation and pDNA Immobilization Glass microscope coverslips (22 mm 22 mm) were deposited with 100 nm thick gold using thermal evaporation. The coverslips were then cut into small rectangular pieces of approximately 3mm3mm. In the following we will refer to these 3mm3mm Gold-Coated Glass coverslips as GCGs. The surface of these GCGs were cleaned by immersing them into 100 times diluted piranha solution (1:1 Sulfuric Acid: Hydrogen Peroxide by volume) for 2 minutes and then washed by deionized water and anhydrous ethanol. The GCGs were then immersed in 50 ml of 0.01 mM 3-mercaptopropyl trimethoxysilane (MPS) (Sigma) solution in ethanol to coat the MPS on the gold surface of the GCGs. The pDNA was immobilized on the MPS surface using Sulfosuccinimidyl-4-(N-maleimidomethyl) cyclohexane1-carboxylate (sulfo-SMCC) (Pierce) as the bi-function. The maleimide end of the SMCC

78 reacted with the thiol of the MPS on the GCG surface and the NHS ester end of the SMCC reacted with the amine at the 5 end of the pDNA, thereby, covalently immobilized the pDNA on the GCG surface (Fig. 3.2a). The 12-PEG spacer at the 5 end of the pDNA would allow the pDNA to be at a distance from the GCG surface for easier hybridization to the tDNA. After MPS coating, the MPS-coated GCGs were immersed in a 1M pDNA solution with 5 mM sulfo-SMCC for 1 hour to immobilize the pDNA. The GCG was then rinsed with DI water and phosphate buffer saline (PBS) solution and ready for MT or WT binding.

Figure 3.2: A schematic of (a) pDNA immobilization on GCG surface using sulfo-SMCC, and (b) schematic of rDNA conjugation on carboxylated FRMs with sulfo-NHS and EDC.

3.1.3 rDNA conjugation to FRMs Blue uorescent polystyrene microspheres (FRMs) (Bright Blue, excitation: 360 nm, emission: 407nm) (Polysciences) 6 m in diameter were conjugated with two rDNAs. First, 0.1 ml of 2.1108 particles/ml stock suspension of FRMs was diluted 10 times in PBS. After-

79 wards, the suspension went through the following washing steps three times: vortexing for 15 seconds, centrifuging at 3700 rpm (Centra, CL2, IEC, MA), discarding the supernatant, re-suspending the sediment in 10 ml PBS. For conjugation, FRMs suspensions at 2.1106 particles/ml were incubated with 3.3 M, 330 nM, 33 nM and 3.3 nM of mixed upstream and downstream rDNAs solutions at 1:1 ratio with 5 mg/ml 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) (Sigma, MA) and 5 mg/ml sulfonated N-Hydroxysuccinimide (sulfoNHS) (Pierce, IL) at pH=6 at room temperature for 1 hr (Fig 3.2b). The suspensions were then washed by centrifugation 3 times as described above. After the nal washing, 10 ml of four different stock conjugated FRMs suspensions at 2.1 106 particles/ml were obtained from four different rDNA concentrations. In each experiment, 1 ml of a stock conjugated FRMs suspension was further diluted by 10 times to a volume of 10 ml and a concentration 2.1 105 particles/ml. In the following all results were obtained at 2.1 105 particles/ml. Fig. 3.3 shows a schematic of a FRM covalently coated with both the upstream and downstream rDNAs.

80

Figure 3.3: (a) A schematic of a FPM coated with upstream and downstream rDNA, (b) a schematic of ow cell where the GCG is placed in the center of laminar ow, and (c) a schematic of ow system.

3.1.4 tDNA Hybridization and Subsequent Capturing of FRMs Hybridization of MT or WT to the probe DNA on a GCG and the subsequent hybridization of the FRMs by the bound MT or WT on the GCG were carried out in an open ow cell where a GCG was placed at the center of the detection chamber with the major faces of the GCG parallel to the direction of the ow as schematically shown in Fig. 3.3 at various temperatures controlled in an incubator (Digital Control Steel Door Incubator 10-180E, Quincy Lab). The custom-made ow cell was 18.5 mm long, 3.5 mm wide, and 5.5 mm deep (volume =356 l and cross-section area =19.25 mm2 ) driven with a peristaltic pump (model 77120-62, Cole-Parmers Master Flex, Vernon Hills, IL). A schematic of the ow system is shown in Fig. 3.3c. An open container of deionized water was included in the incubator to control the humidity in the incubator to minimize evaporation from the open

81 ow cell. To examine the effects of both the temperature and the ow rate, we carried out hybridization experiments at various temperatures and various ow rates. The temperatures examined were room temperature (RT), 30o C, and 35o C, which were between the melting temperature of the WT, 23 o C, and that of the MT, 47 o C. The ow rates examined were 0, 2, 4, and 6 ml/min. At the given temperature and ow rate, we rst circulated the MT or WT solution through the ow cell for 30 min for the MT or WT to bind to the pDNA on the GCG as illustrated by the schematic in Fig. 3.4a. Afterwards, we circulate the FRMs suspension through the ow cell for 30 min to allow the FRMs to bind to the MT or WT captured on the GCG as schematically illustrated by Fig. 3.4b. PBS was then own through the ow cell to wash off loosely bound FRMs at a ow rate of 2 ml/min for 30 in as schematically shown in Fig. 3.4c.

82

Figure 3.4: A schematic of a) tDNA sequence own over pDNA immobilized on GCG, b) Upstream and downstream rDNA-conjugated FRMs own over the GCG surface and FRMs hybridized on the anking regions of the tDNA sequences, and c) unbound or loosely bound GCG washed

3.1.5 FRMs counting After washing, the GCG was air dried and examined in a uorescent microscope (BX51, Olympus). The number of FRMs per unit area was determined with a custom made program turned into black and white using Otsus method. [202] As the brightness of each FRM was different in each image (due to different exposure times and FRMs uorescence

83 decay), the number of white pixels per FRM could be different in each image after black and white conversion. To determine an average white pixel per FRM, clusters of white pixels were labeled and cluster size distribution was determined. The number of the FRMs per unit area in each image was determined from the pixel cluster size distribution as follows. First, clusters of white pixels less than 20 pixels were considered to be noise due to the gray-scale to black-white conversion and were not included in FRM counting. Second, in addition to well-separated FRMs, some FRMs were close together, they might appear as one large cluster. These large clusters of white pixels were separated from clusters of white pixels corresponding to single FRMs when determining the average number of white pixels per FRM. This was accomplished by applying an outlier removal algorithm (ASTM E178, k=1.5, ASTM=American Society for Testing and Materials) with the large outlying clusters representing groups of FRMs that were too close to be seen as separate FRMs. After the large outlying groups of white pixels were neglected, the average FRM size in terms of number of white pixels was determined by dividing the total number of white pixels in the remaining clusters by the total number of the remaining clusters. Once the average FRM size was determined, the total number of white pixels in all clusters including the large outliers was divided to this average number of white pixels per FRM to estimate the total number FRMs per unit area.

84 3.2 Results 3.2.1 FRM conjugation and initial room-temperature testing without ow In order to determine the optimal condition for FRMs conjugation to the rDNAs, different concentrations (3.3, 33, 330, and 3,300 nM) were used to conjugate rDNAs to the FRMs as described above. To determine which concentration had the optimal conjugation conditions, we soaked GCGs with immobilized pDNAs in a 1-M solution of the 200-nt MT followed by rinsing with PBS. We then soaked the GCGs in suspensions of rDNAconjugated FRMs obtained with different rDNA concentrations for 30 min followed by rinsing with PBS. Fig. 3.5a shows the number of FRMs/mm2 on the GCG surface versus the rDNA concentration for rDNA conjugation to the FRMs. As can be seen from Fig. 3.5a, the FRMs conjugated in 330 nM rDNA solution exhibited the most number of FRMs/mm2 captured on the GCG surface with identical pDNA immobilization and tDNA hybridization conditions, indicating that the optimal rDNA concentration for the rDNA conjugation on the FRMs was 330 nM. In what follows, all FRMs were conjugated with rDNAs at this concentration.

85

Figure 3.5: Number of FRMs per mm2 captured on the GCG (a) followed by 10-10 M MT detection versus rDNA concentration and, (b) followed by MT detection versus MT concentration with pDNA optimally immobilized in 330 nM pDNA. Both (a) and (b) obtained at room temperature and without ow We then tested thus-obtained FRMs following MT hybridization to the GCG at various MT concentrations for 30 min followed by FRMs binding at 2.1105 particles/ml and for 30 min where binding of both MT and FRM were carried out at room temperature without ow. The resultant number of FRMs/mm2 versus MT concentration is plotted in Fig. 3.5b. As can be seen from Fig. 3.5b, at 10-16 M and 10-17 M of MT, the GCG showed about 450 and 300 FRMs/mm2, respectively while at 10-18 M and 10-19 M of MT tDNA, the number of FRMs/mm2 were comparable to that of the negative control at 0 M of MT. Sample uorescent images of FRMs obtained at 10-18 M, 10-17 M, and 10-16 M MT are shown in inserts I, II, and III, respectively, indicating that the concentration sensitivity of this method without ow was about 10-17 M of MT.

86 3.2.2 Effect of Flow Rate and Temperature Combinations of 3 different temperatures (room temperature, 30o C and 35o C) and 4 different ow rates (no ow, 2, 4 and 6 ml/min) were studied. The MT or WT was rst own for 30 min followed by the ow of FRMs for 30 min at the same temperature and ow rate, and nally washed with a ow of PBS at 2 ml/min at the same temperature as described above. The MT and the WT were kept at 10-10 M. The number of FRMs/mm2 hybridized to MT on the GCG surface, that to the WT on the GCG surface and that to the negative control versus ow rate at room temperature, 30o C, and 35o C are plotted in Fig. 3.6a-3.6c. It is interesting to note from Fig. 3.6a that at RT, the ow increased the binding of both the MT and the WT in the 2-4 ml/min range but decreased the binding of both MT and the WT as the ow rate increased. Such increase of binding at a low ow rate and decrease in binding upon further increase of ow rate was also observed in the Anthraces bacillus detection. [69] The enhancement of binding by a ow has also been observed in antibody binding on antigen [203]. The enhanced binding at a lower ow rate was presumably due to that ow helped to bring more tDNAs to GCG surface for binding. Without ow, the DNAs would have relied solely on diffusion to get to the GCG surface. While ow can help bring more tDNAs to the GCG surface, it could also generate an impingement force on the FRMs [69] which could overcome the binding force between the tDNAs and pDNAs to unbind them on GCG, thus reducing the overall number of FRMs bound on the GCG surface as the ow rate was further increased. At 30o C and 35o C, though, the binding of both the MT and the WT decreased monotonically with an increasing ow rate, presumably due to the weakening of the binding of the pDNA to both the MT and the WT. Although the binding of the pDNA to both MT and WT decreased with an increasing ow rate at 30o C

87 and 35o C, what was of interest is that the binding of the pDNA to WT was suppressed to the levels similar to the negative controls at higher ow rates, suggesting that ow made the detection more selective at these two temperatures. In the following, we will refer to the number of FRMs/mm2 captured on the GCG surface as the signal, S and SMT , SW T , and SC refer to the numbers of FRMs/mm2 captured on the GCG surface following owing a 10-10 M MT solution, a 10-10 M WT solution, and a control blank PBS, respectively.

88

Figure 3.6: Number of FRMs per mm2 captured by a GCG followed by detection in the control (no MT or WT, black), that followed by detection at 10-10 M WT (red), and that followed by detection (blue) at 10-10 M MT, (a) at room temperature, (b) at 30o C, and (c) at 35o C.

89 To examine if the ow helped increase the detection specicity, we plot the ratio, SW T /SC versus ow rate in Fig. 3.7a with black for room temperature, red for 30o C and blue for 35o C. As can be seen, with ow, SW T /SC was signicantly reduced to close to unity not only at 35o C but also at 30o C. We further compared the SW T and SC using the MannWhitney U test. [204] The resultant p value versus ow rate is plotted, again, with black for room temperature, red for 30o C and blue for 35o C. As can be seen, at low ow rates of 0 and 2 ml/min SW T and SC were statistically different (p<0.05) at room temperature and at 30o C while at high ow rates of 4 and 6 ml/min SW T and SC were only statistically different at room temperature but not at 30o C. These results indicated that with the diminished signicance of SW T against SC at a ow rate of 4-6 ml/min permitted specic MT detection at a lower temperature of 30o C as opposed to 35o C with a ow of 0-2 ml/min.

90

Figure 3.7: (a) SW T /SC and (b) p value versus ow rate at room temperature (black), at 30o C (red), and at 35o C (blue) where SW T and SC are number of FRMs per mm2 captured by a GCG followed by detection at 10-10 WT and that captured by a GCG followed by detection in control (no WT or MT). One measure of MT detection specicity with respect to WT is SMT /SW T . In Fig. 3.8, we plot SMT /SW T versus ow rate for all three temperatures: room temperature (black), 30o C (red) and 30o C (blue). As can be seen, by increasing the temperature alone, SMT /SW T increased only slightly from about 11 to about 12 at 35o C. With ow, SMT/SWT increased dramatically. More importantly, at 30o C ow weakened the binding of WT more than that of MT to allow SMT /SW T to reach around 24 at 4 ml/min, which was higher than the SMT /SW T at any ow rate at 35o C, indicating that ow could indeed enhance the detection specicity at a low temperature to allow more sensitive detection. The above results indicated that with the present ow setup the optimal detection conditions for the current HBV-DM MT occurred at 30o C and a ow rate of 4 ml/min.

91

Figure 3.8: SMT /SW T versus ow rate where SMT is number of FRMs per mm2 captured by a GCG followed by detection at 1010 MT and SW T is number of FRMs/mm2 captured by a GCG followed by detection at 1010 WT at room temperature (black), at 30o C (red), and at 35 o C (blue)

3.2.3 SMT /SW T at 10, 103 , 106 , and 107 WT/MT concentration ratios As can be seen from Fig. 3.5b, 1017 M was the lowest MT concentration where MT could be detected without ow by the current detection scheme. To see how specic the

92 MT detection was at this lowest detectable concentration, we used the optimal conditions, i.e., 30o C ow rate of 4 ml/min and carried out MT detection at 1017 M for 30 min followed by 30 min of FRMs hybridization. We then carried out WT detection at concentrations at 1016 , 1014 , 1011 , and 1010 M (i.e., 10-fold, 103 -fold, 106 -fold, and 107 fold that of the MT) followed with 30 min of FRMs hybridization. In Fig. 3.9a, we plot number of FRMs/mm2 captured by hybridized WT at different concentrations, with and without 4ml/min ow rate (empty and pattern lled blue bars respectively), and number of FRMs/mm2 captured by hybridized mixture of same concentrations of WT and 1017 M of MT with and without 4ml/min ow rate (empty and pattern lled red bars respectively) versus WT concentration at 30o C. Also plotted in Fig. 3.9a are the horizontal bars indicating the number of FRMs/mm2 captured by hybridized 1017 M of MT only, at 30o C with and without 4ml/min ow rate (violet and green respectively) and that non-specically bound of control (brown). In Fig. 3.9b, we plotted the SMT /SW T and SMix /SW T ratios with and without 4ml/min ow rate deduced from Fig. 9a. As can be seen, without ow, SW T remained different from SC . With a ow of 4 ml/min, SW T was not distinguishable from SC . Moreover, mixture experiments showed that, at 30o C, even the addition of 107 fold more amount of WT did not result in a signicant increase in number of hybridized FRMs, when 4ml/min ow is applied. On the other hand, in the case of no ow, number of FRMs hybridized increased signicantly. Moreover specicity increased signicantly in ow conditions with respect to no-ow conditions as can be seen from Fig. 3.9b, and mixing signicantly high concentrations of WT into a solution of 1017 M MT did not change this ratio. This indicates that the presence of WT of any concentration at 30o C and 4ml/min had no discernible contribution to detection signal even at a MT concentration as

93 low as 1017 M and that a ow of 4 ml/min made the detection as specic as 1017 M MT/1010 M WT, or 1 MT to 107 WT.

Figure 3.9: (a) Number of FRMs per mm2 versus WT concentration and, (b) ratio of number of FRMs per mm2 of MT/WT and Mixture/WT versus WT where mixture is 1017 M MT mixed with different WT concentrations, and brown, blue and green horizontal bars represent number of FRMs per mm2 for control, 1017 M MT with ow and 1017 M MT without ow conditions. Both (a) and (b) are done at 30o C, in 4ml/min ow rate or in no ow conditions.

3.3 Discussions With a cross section area of the detection chamber being 19.25 mm2 , the average ow velocity u was 1.7, 3.5, 5.2 mm/s for ow rates 2, 4, and 6 ml/min, respectively. The Reynolds number, Re = uw/ was just 6, 12, and 18 at ow rates 2, 4, and 6 ml/min, respectively where =1010 kg/m3 was the density of the uid, =1.05 cP the viscosity of the uid and w the width of the ow cell, well in the range of laminar ow. Furthermore

94 the entrance length, Ie , the length over which a fully developed velocity prole can be established once the ow entered the ow cell, can be calculated using;

I = 0.06d Re

(3.1)

where d was the width of the ow cell. According to Eq. (1) even at the highest ow rate which gave the largest Re, the entrance length was approximately 3.8 mm which was well below 7.7 mm, the distance from the inlet of the ow cell to GCG. Therefore, the ow in the detection chamber at the position of GCG, was laminar and the ow velocity prole in the width direction was parabolic as schematically shown in Fig. 3b: u was zero at the cell wall and a maximum at the center of the ow. Thus, unlike other systems where the capture surface is part of a wall of the ow channel, [69] at which point the uid velocity diminishes to zero, the present system situates the capture surface at the center of the ow where the ow velocity is at a maximum (as shown schematically in Fig. 3b). The concept of Goldman et al. [205] for a sphere parallel to a planar wall in a uniform ow could be applied to the FRMs bound on the GCG in the middle of a laminar ow where the ow velocity was uniform as discussed above. According to Goldman et al., in the middle of the ow cell where GCG is situated, one can obtain an analytical expression for the impingent force on a bound FRM as

F = (1.7)6 (1.5u)

(3.2)

95 where u was the average ow velocity and 1.5 u was the ow velocity at the centre of the ow, and a = 3 m the FRM radius. The deduced impingent force on the FRMs by the ow was about 262, 524, and 787 pN with u = 1.7, 3.5, 5.2 mm/s at 2, 4, and 6 ml/min, respectively. Although the exact binding forces between the pDNA and the WT and that between the pDNA and the MT were not known and it was unclear if there were more than one captured MT or WT bound to a FRM it sufces to say that the deduced force was consistent with 70-1500 pN found in the de-hybridization of a single double-stranded DNA [206, 207].

3.4 Summary of Temperature and Flow Enhanced Specicity using Fluorescent Reporter Microspheres We have investigated the effect of a laminar ow on enhancing the specicity of MT detection at a lower temperature by immersing the detection GCG surface at the center of the ow of the target MT or WT at various ow rates and temperatures. pDNA complementary to the target MT DNA was immobilized on the GCG surface. 30 min of a ow of the MT or WT DNA solution was followed by a ow of 6- m FRMs of 105 FRMs/ml for 30 min at the same ow rate and temperature. With HBV-DM as the model MT, we have shown that ow can increase the MT detection specicity by lowering the detection temperature to allow (1) a higher ratio of SMT /SW T and (2) a lower SW T not distinguishable from SC . For the present system, a ow rate of 4-6 ml/min reduced the specic MT detection temperature from 35o C without ow to 30o C with a ow rate of 4-6 ml/min. Furthermore, optimal specic MT detection was shown to occur at a lower temperature with ow than the temperature without ow. For example, the detection specicity as measured by SMT /SW T

96 was 24 at 30o C and 4 ml/min as opposed to 15 at 35o C without ow. These results clearly indicate that ow can be utilized to help increase mutation detection specicity at a lower temperature.

97 4. Detection of mutations of single-stranded DNA

4.1 Introduction Detection of mutations requires specicity that would differentiate one base differences in hybridization reactions. The melting temperature difference between MT and WT tDNAs can be less than 2 degrees Celcius as discussed before in chapter 3. As expected, as the number of mismatches increases the melting temperature difference between a MT and WT tDNA with a pDNA complementary to the MT tDNA would increase. Considering DNA denatures in an unzipping mechanism, the proximity of mismatches makes a big difference. Piezoelectric plate sensor (PEPS) demonstrates label-free detection of single stranded target DNA sequences in 1xPBS with a sensitivity level of 100 zM in width extension mode resonance peak without isolation and amplication as shown in chapter 2. In this study, we examine the detection sensitivity of lead magnesium niobate-lead titanate (PbMg1/3 Nb2/3 O3 )0.65 (PbTiO3)0.35 (PMN-PT) piezoelectric plate sensor(PEPS) in realtime, label-free, in situ MT tDNA mutation hybridization detection in full urine without isolation and amplication with high specicity that would allow detection in a background of excess amount of WT tDNA. A probe DNA (pDNA) sequence specic to HBV-DM used in chapters 3 and 2 and a pDNA specic to kRas codon 12 GGT GTT transversion are immobilized on the electrical insulation layer on PEPS surface. HBV-DM has been previously shown to be a risk factor for the development of Hepatocellular Carcinoma (HCC) [208]. A high percentage ( > 60 %) of HCC patients had HBV-DM in their sera [197,198]. On the other hand, It has been reported that kRas gene mutations occur in up to 50% of colorectal

98 adenocarcinomas; these are mainly point mutations at codons 12 and 13, and less frequent at codon 61 [209]. Among the seven most common kRas mutations, codon 12 GGT GTT transversion (Glycine Valine) was associated with signicantly higher colorectal cancer-specic mortality [210].Probe DNA sequence used to detect kRas single mutation is composed of DNA and Locked Nucleic Acid (LNA) hybrid nucleic acids. Non-specic binding is prevented by treating the PEPS surface with 3% Bovine Serum Albumin (BSA). Hybridization detection of synthetic target DNA sequences in urine are accomplished at 30o C and 63o C using laminar ow. Further conrmation of detection is accomplished by Fluorescent Reporter Microspheres (FRM) as described in chapters 2. Detection limit of 100 zM is achieved. Futhermore an array of 6 PEPS is used to detect not only the GGT to GTT transversion but also 5 other possible mutations in codon 12 of kRas gene that would lead to colorectal carcinoma. Multiplexed detection setup is only done at 1 fM concentration, to measure cross hybridization between different mutations of the same codon and different pDNA immobilized on each PEPS. No cross hybridization at this concentration (which is probably more than what would be dissolved in any patients urine), is observed, although no dose response experiments are done.

4.1.1 Locked Nucleic Acids Mismatch discrimination depends on differential hybridization between perfectly matched and mismatched strands hybridizing to form the double stranded duplex. A measure of efciency of this discrimination is the differences in melting temperatures (Tm) between these species. The Tm is usually small (0.53 o C) for a single mismatch. [211] Hybrid probes including both DNA and LNA nucleotides were reported to enhance both

99 duplex stability and mismatch discrimination. [211218] LNA monomers contain a modied ribose moiety. O-methyl group in LNA residues and bridges 2 and 4 carbons of the ribose ring. This covalent bridge locks the ribose in the N-type (3-endo) conformation.This conformation enhances base stacking and phosphate backbone pre-organization [219] and results in improved afnity for complementary DNA or RNA sequences. LNA-DNA hybrid pDNAs increases Tm of both perfectly matching and mismatching sequences, however increase in mismatching sequences is much less than that in perfectly matching sequences, widening the Tm difference of a mismatching and perfectly matching sequence as shown in gure 4.1.

100

Figure 4.1: 1m indicates the melting temperature curve of a certain mismatching sequence (point mutation) with a pure DNA probe. 1p indicated the melting temperature curve of the same sequence with a perfectly matching pure DNA probe. 2m and 2p indicates the same melting curves with a LNA-DNA hybrid pDNA which contains 3 LNA bases centered around the mismatching base. Figure adopted from [5]. The use of LNA-DNA hybrid pDNAs were necessary to differentiate MT and WT tDNA with high specicity. The design in gure 4.1 is used to in this study to maximize the effect of LNA nucleotides as discussed in [5]

101 4.2 Materials and Methods 4.2.1 PEPS Fabrication PEPS used in this study were fabricated from (PbMg1/3 Nb2/3 O3 )0.63 -(PbTiO3 )0.37 (PMNPT) freestanding lms 8 m thickness that was coated with 110 nm thick Cr/Au electrode by thermal deposition (Thermionics VE 90) cut into 2.5 mm by 0.7 mm strips by a wire saw (Princeton Scientic Precision, Princeton, NJ). Gold wires, 10 m in diameter, were glued to the top and bottom electrodes of each strip using conductive glue (XCE 3104XL, Emerson and Cuming Company, Billerica, MA). The rear end of the strip was xed on a glass substrate by a nonconductive glue (Loctite 1C Hysol Epoxy Adhesive) to form the PEPS geometry as illustrated in Fig. 2.1d and poled at 15KV/cm at 90o C for 60 min. in an incubator (Digital Control Steel Door Incubator 10-180E, Quincy Lab). The dielectric constant of the PEPS was measured using an electrical impedance analyzer (Agilent 4294A) to be about 1800 with a loss factor of 2.5-3.7% at 1 kHz.

4.2.2 Electrical Insulation The PEPS were electrically insulated to stabilize the resonance peaks for in-liquid detection by using a new 3-mercaptopropyltrimethoxysilane (MPS) (Sigma-Aldrich Co. LLC.) solutions coating scheme involving enhanced MPS cross-linking at pH=9.0 and with the addition of water.(135) First, the PEPS was cleaned in a Piranha solution (two parts of 98% sulfuric acid (Fisher) with one part of 30% hydrogen peroxide (Fisher)) for 1 min, followed by rinsing in water and ethanol. Before coating MPS at pH=9.0, we dipped the PEPS in 50 ml of a 0.01 mM MPS (98% v/v) solution in ethanol (Fisher) with 0.5% of de-

102 ionized water for 30 min to promote hydrolysis. It is then rinsed with water and ethanol and immersed in 50 ml of a 0.1% MPS solution in ethanol with 0.5% of DI water at pH=9.0 adjusted by adding KOH (Flakes/Technical, Fisher)for 12 hr. The PEPS was then rinsed with water and ethanol and then subjected to four more rounds of 12-hr soaking in a fresh 0.1% MPS solution at pH=9.0 with 0.5% water with each soaking followed by rinsing in water and ethanol. It has been shown such a pH=9.0 MPS coating scheme resulted in a dense and smooth MPS coating at a rate of about 1.6 nm/hr and resultant current density of 2 A/cm2, a more than 100 fold reduction than MPS coating obtained at pH=4.5 without water.(135) The in-air and in-phosphate buffer saline (PBS) solution phase angle versus frequency resonance spectra of a PEPS are shown in Figure 4.2a. As can be seen, the baseline and the length-extension-mode (LEM) resonance peak and the width-extension-mode resonance (WEM) peak of the in-liquid spectrum were close to that of the in-air spectrum, indicating the effectiveness of the new MPS coating at pH=9.0 with water.

103
(a)

-20 -30 -40 -50 -60 -70 -80 -90 0

In Air In PBS

Phase Angle (deg.)

2 Frequency (MHz)

(b)

PBS

SMCC

pDNA tDNA FRM PBS

-2

f/f (10 )

-4

-3

-6

-8

-10 0

20

40

60

80

100 120 140 160 180

Time (min)

Figure 4.2: (a) In-air (black) and in-PBS (red) phase angle-versus-frequency resonance spectra, and (b) relative resonance frequency shift, f/f, of the PMN-PT PEPS during the various steps of probe cDNA immobilization and target DNA detection. The insert in (b) shows a schematic of the molecules involved in the immobilization and detection.

104 4.2.3 pDNA Immobilization 4.2.3.1 Immobilization of pDNA for detections of kRas codon 12 GGT to GTT transversion and HBV-DM mutations To immobilize the amine-activated pDNA on the PEPS surface, the MPS-coated PEPS was rst immersed in a solution of 5mM of sulfo-SMCC dissolved in 200 l 1xPBS (pH adjusted to 6.5) for 1 hour. The sensor is then washed three times with deionized water and then it is immersed in a solution of 10 M of amine activated probe DNA dissolved in 200 l of 1 PBS(pH 8.0) as shown in Figure 4.3. The relative resonance frequency shift, f/f, of the rst WEM peak at various steps of the immobilization process is shown in Figure 4.2b. Also shown in the insert in Figure 4.2b is a schematic of the various steps involved in the immobilization process.

105

Figure 4.3: A schematic of amine activated pDNA immobilization on GCG surface using sulfo-SMCC.(a) PEPS is immersed in sulfo-SMCC solution at pH 6.5. (b) 5 amine activated pDNA is immobilized on the sulfo-SMCC conjugated PEPS surface at pH 8.0.

4.2.4 Immobilization of pDNA for 6 mutations on codon 12 The immobilization of biotinylated pDNAs designed for multiplexed detection are done using maleimide actived biotin as discussed in chapter 2.

4.2.5 Non-specic binding Non-specic binding was prevented treating the PEPS with different percentages of Bovine Serum Albumin (Sigma) after pDNA immobilization. BSA is dissolved in 1x PBS

106 and PEPS was immersed in this solution for 1 hour followed by washing 5 times with 1x PBS. WEM resonance peak of PEPS were monitored in urine in order to check for non-specic binding. This process was repeated systematically by increasing the BSA percentage used for blocking until no non-specic binding was observed as seen in Figure 2.12.

4.2.6 Target DNAs, probe DNA, reporter DNAs and FRMs The probe DNAs (tDNA) used to detect HBV-DM was a 16-nucleotide (nt) long singlestranded DNA (Sigma) containing the nucleotide sequence of the Hepatitis B virus genome (GeneBank Accession #X04615) centered around the 1762T/1764A double mutations.(136) The probe DNA used to detect kRas codon 12 mutation was a 17 nucleotide DNA-LNA hybrid fragment (Exiqon, Inc) derived from the kRas gene sequence (Gene ID: 3845) centred around the point mutation. Three LNA bases are centred around the mutation site and the rest of the sequence consists of DNA bases (Figure 4.4a-b). This design is chosen to have the highest melting temperature between wild type and mutant tDNA hybridization reactions.(137) Target DNA sequences (Sigma) designed for both diseases are 50 nucleotides long and in both cases MT tDNA sequences consists of the region complementary to the pDNA and upstream region of the gene and WT tDNA sequences consists of the region complementary to the pDNA except the mutation sites and the downstream region of the gene (Figure 4.4a-b). Sequences and melting estimated temperatures are shown on Table 4.1 . Both pDNA sequences had a biotin with a 12-polyethyleneglycol (PEG) spacer at the 5 end. The tDNA and pDNA sequences designed for multiplexed detections of kRas codon 12 mutations are shown on table 4.2. The tDNA sequences were 50 nucleotides long

107 (sequences not shown) as those in table 4.1, the only difference was in the mutation site where each tDNA were perfectly matching with the pDNA on that region.

108

Figure 4.4: Schematics of pDNA, Wild Type and Mutant tDNAs and upstream and downstream rDNAs designed for (a) HBV-DM and (b) kRas mutations. Mutation site(s) are shown with patterned boxes, dashed lines indicate the hybridization sites and mismatch(es) on Wild Type tDNAs is/are indicated.(c) Bright blue and yellow-green microspheres are conjugated with upstream and downstream microspheres to be used as Mt-FRM and WtFRM

109 There were two different 30-nt long rDNAs (Sigma) complementary to the sequence of the MT tDNA upstream of the pDNA binding site and complementary to that of WT tDNA downstream of the pDNA binding site for both diseases. The upstream rDNAs were amine activated with a 12-PEG spacer at the 5 end and the downstream rDNAs were also amine activated but with a 7-PEG spacer at the 3 end. The sequence of the upstream rDNA and that of the downstream rDNA are also shown in Table 4.1. Figure 4.5 is a schematic illustrating the relationship between tDNA, pDNA, and upstream and downstream rDNAs. FRMs are prepared as explained previously in chapter 3.

110

Figure 4.5: Schematic of the detection setup. pDNA (Pink) is immobilized on SMCC conjugated PEPS surface via its primary amine attached on the 5. WT tDNA (light green) and MT tDNA (dark blue) can hybridize on pDNA with mismathc(es) and no mismatches respectively. Bright blue FRMs (dark blue sphere) conjugated with MT rDNA (yellow) and Yellow Green FRMs (green sphere) hybridizes on MT tDNA and WT tDNA respectively.

4.2.7 Urine samples and Experimental Setup Urine mixtures used for detection are prepared done by mixing different concentrations of tDNAs (as a mixture of WT and MT or MT alone or WT alone) in 50 ml of urine. PEPS is immersed in a ow cell as seen in Figure 3.3b. Experimental setup involving solutions of tDNA, FRM, PBS washings and Peristaltic Pump are prepared as explained previously. The whole setup was installed inside an incubator (Digital Control Steel Door Incubator 10180E, Quincy Lab) for temperature control, including a 2 liter water bath for minimization

111 of evaporation during experimentation. Detection experiments involving HBV-DM were done using 4ml/min ow rate at 30o C as determined previously in chapter 3, and those involving kRas were done using 6ml/min ow rate at 63o C. This temperature is determined after experimentation with MT and WT at different temperatures. The array of 6 PEPS is shown in gure 4.6. Each PEPS is immobilized with one type of pDNA specic to one mutation shown in Table 4.2. Flow cell used for single PEPS detections is used. PEPS are connected to a relay which is controlled by the same MatLab routine discussed in chapter 2. Each time an impedance scan is performed by the analyzer, the relay is switched to the next PEPS (order of which is dened by the user) for the next impedance scan until the user stops the sequence.

112

Figure 4.6: Array of 6 PEPS is shown inside the ow cell (bottom), and a blow up of this array is shown on top. Each PEPS is shown in different colors indicating immobilization by different pDNAs, each specic to one type of mutation.

113 4.3 Results and Discussion 4.3.1 HBV-DM single stranded DNA detections As previously discussed in Chapter 3, HBV-DM mutations can be detected specically at 30o C with 4 ml/min ow rate 3.8. In Chapter 2 hybridization of HBV-DM MT tDNA is studied in room temperature and 2ml/min ow rate, 2.13a. However in this condition, hybridization of WT-tDNA is only studied in Chapter 3 on GCGs but not on PEPS.

4.3.1.1 Detection of HBV-DM MT and WT tDNA Figure 4.7a below shows the detection of 10 aM and 100 zM of HBV-DM WT-tDNA at room temperature with 2 ml/min ow rate. Figure 4.7b shows the comparison of average relative frequency shifts between 25th -30th min.(signal of hybridization of tDNA) and 55th 60th min.(signal of hybridization of FRM) of MT tDNA and WT tDNA detections and these concentrations at this condition. As can be seen relative frequency shifts are very close at 10 aM. It is important to note here that in real patient samples WT tDNA is almost 103 fold higher in concentration. Thus it is obvious that this condition is not suitable for detection of mutations.

114
(a)
0.0
0.002

Wt tDNA

FRM

PBS

(b)

2.1 Mt tDNA 1.8 1.5 1.2 0.9 0.6 0.3 0.0 Wt tDNA Mt FRM Wt FRM

-0.4
f/f (x10 )
3

0.000 -0.002 -0.004 -0.006 -0.008 0

f/f (x10 )

-0.8 -1.2 -1.6 -2.0 10 aM 100 zM Control Room Temp. 2ml/min 15 30

30

60

90

Time (min)

45 Time (min)

60

75

90

f/f (10 )

-3

10 aM

100 zM

tDNA Concentration (M)

Figure 4.7: (a)Detection of HBV-DM Wt tDNA at room temperature and 2ml/min ow rate. 100zM and control experiments are shown in the inset. Detections were conrmed by FRM hybridizations and PBS background. (b)Average relative frequency shifts of tDNA and FRM hybridizations of 10 aM and 100 zM of MT-tDNA and WT-tDNA of HBV-DM are plotted. Experiments are done at room temperature and 2ml/min ow rate Detections of MT-tDNA and WT-tDNA of HBV-DM are also carried out in 30o C and and 4 ml/min ow rate separately. 10, 5 and 1 aM and 100 zM of MT-tDNA and 100,25 and 1 aM of WT-tDNA are detected as shown in Figure 4.8a and b respectively. Average relative frequency shifts between 25th -30th min. (SMT and W T ) and 55th -60th min.(SFRM MT
FRM FRM and SFRM MT ) of these detections are plotted in Figure 4.8c. The SMT /SW T and SMT /SMT

ratio are plotted in Figure 4.8d.

115
(a)
0.0 Mt tDNA FRM

(b)
0.00 -0.03 -0.06 -0.09 -0.12 -0.15 -0.18 0

Wt tDNA

FRM

-0.2

30 C, 4ml/min

f/f (x10 )

-0.4
f/f (10 )
-3

-3

-0.6 Control -0.8 100 zM 1 aM 5 aM 10 aM 30 C, 4ml/min 10 20 30 Time (min) 40 50 60

o

Control 1 aM 25 aM 100 aM 10 fM 100 fM 10 20 30 40 50 60

-1.0

-1.2 0

(c)
0.7 0.6 0.5

Time (min)

(d)
Mt tDNA Mt FRM Wt tDNA Wt FRM
S S
MT

/S

WT FRM WT

FRM MT

/S

f/f (10 )

-3

0.4
WT

0.1 0.0 -0.1

10
1 -19 -18 -17

Control 10

-20

10

-18

10

-16

10

-14

0.2

MT

/S

0.3

10

10

10

10

tDNA Concentration (M)

tDNA Concentration (M)

Figure 4.8: Relative frequency shift versus time graph of detections of (a) 100zM, 100,1,5 and 10 aM of MT tDNA followed by MT-FRM hybridization (b)1aM, 25aM, 100 aM, 10 fM and 100 fM of WT tDNA followed by WT-FRM hybridization in urine at 30o C and 4ml/min ow rate (c) Relative frequency shifts in the last ve minutes of tDNA and FRM hybridization parts (signal, S) of experiments in (a) and (b) plotted versus concentration and (d) The ratios of SMT /SW T are plotted for tDNA and FRM hybridization parts seperately Figure 4.8b shows that at 25 aM of WT tDNA detection although there is detectable

116 hybridization of tDNA hybridization of FRMs was not detected. This is also conrmed by uorescent microscopy as there was no WT-FRMs on the surface. This indicates that detection of tDNA by PEPS may not always be conrmed by FRM hybridization, especially at low concentrations or hybridizations of low afnity, which may occur in unfavorable conditions for a hybridization reaction (such as, high temperature, low salt concentration or pH). Figure 4.8c shows that as the concentration increased the number of hybridized tDNA is increased for both MT and WT tDNAs. The increase in relative frequency shift in MT is apparent even at 100zM. On the other hand, such an increase can only be observed at a 1000 fold higher concentration for WT. Moreover signal (S) of WT tDNA at 100aM is close to that of MT at a 100 fold lower concentration of 1aM. In gure 4.8d, although the increasing behavior of SMT /SW T with MT tDNA concentration does not mean much since the SW T was already very close to control level in those low concentrations, the SMT /SMT was at least 23. However it can still be deduced from gure 4.8c that SMT at 10 aM (long before SMT reaches plateau), is more than 6 times SW T at 100 fM (plateau is reached for SW T ). 4.3.1.2 Detection of mixtures of HBV-DM MT and WT tDNA As discussed in chapter 3 patients usually have extensive amount of WT tDNA with respect to the amount of MT tDNA. Thus it is not enough to detect low concentrations of MT tDNA most of the time. MT tDNA has to be detected in urine samples with a background of abundant WT tDNA. In order to test if the conditions optimized for this demanding specicity requirements a 1:250 ratio of MT to WT ratio is kept constant for mixture detections of HBV-DM.

117 For this purpose, 100zM, 1, 10 and 100 aM of MT tDNA is mixed with WT tDNA concentrations such that 1:250 MT to WT ratio is kept constant. Figure 4.9a shows the relative frequency shifts versus time graphs of hybridizations of MT and tDNA hybridizations which is followed by hybridizations of MT-FRMs and WT-FRMs. In order to check the interference of excess amounts of WT tDNA on hybridization of MT tDNA the signal (S) from hybridizations of MT and WT tDNA detections of the same concentrations done seperately is added and compared with the mixture experiments as shown in gure 4.9b for both tDNA and FRM hybridization parts. As can be seen, as the concentration is increased from 100zM to 10aM the interference from WT tDNA becomes apparent as difference between the sum of signal from WT tDNA-only(red) and MT tDNA-only(black) experiments and mixture experiments (blue) increases. This was of course expected as WT tDNA interferes solely by its existence even though it does not hybridize with pDNA signicantly. The same behavior is also observed in the FRM hybridization parts (patterned bars of the corresponding colors from tDNA experiments).Number of FRMs hybridized on PEPS surface are plotted versus concentration for these detections as shown in 4.9c and the micrographs are also shown on the lower right in the same gure. As can be seen although WT FRM hybridized on the surface increases as concentrations of tDNAs are increased, the rate of this increase is less than that of MT-FRM as can be deduced from the nummber of MT-FRMs on the surface.

118
(a)
0.0 1.2 -0.2 1.0
f/f (10 )

tDNA

FRM

(b)

1.4

Mt tDNA WT tDNA Mix tDNA Mt tDNA+FRM WT tDNA+FRM Mix tDNA+FRM

-0.4
f/f (10 )
-3

-3

0.8 0.6 0.4 0.2

-0.6 -0.8 -1.0 -1.2 0 10 20 30 Time (min) 40 50 60 100 zM 1 aM 10 aM 100 aM

10

-19

10

-18

10

-17

tDNA Concentration (M)

(c)
60 Total FRM MT FRM 50 WT FRM

(d)

40
# FRMs

30

20 IV 10 I 0 10
-19

III II

10

-18

10

-17

10

-16

tDNA Concentration (M)

Figure 4.9: (a) Relative frequency shift versus time graph of HBV-DM mixture detection experiments with 100zM, 1, 10 and 100 aM of MT tDNA mixed with 250 fold more amount of WT tDNA (Concentrations of MT tDNA in the mixture experiments are shown on the x axis). (b) Signal (S) of tDNA and FRM hybridization parts of mixture experiments are compared with sum of that of MT-only and WT-only experiments. (c) Number of MT-FRMs and WT-FRMs on PEPS surface was counted and plotted versus concentration of MT tDNA of mixture experiments.(d) Micrographs of PEPS after HBV-DM mixture detection experiments are ahown. MT-FRMs are shown in blue and WT-FRMs are shown in green.

119 4.3.2 kRas single stranded tDNA detections 4.3.2.1 Determination of optimum temperature for detection of kRas mutations In order to determine the optimum temperature at which the ratio of hybridizations of MT and WT tDNAs is maximized, detections of 10 fM of MT and WT tDNA are carried out at 55o C, 60o C and 63o C with 4ml/min ow rate (Figure 4.10a). These temperatures are chosen considering the Tm of MT and WT tDNA are estimated at 70o C and 55o C respectively as shown in Table 4.1. As can be seen from gure 4.10b although the number of hybridized tDNAs on PEPS is inversely proportional with temperature for both MT (black) and WT (red) tDNAs (y axis on the left for relative frequency shifts), the ratio of signal from MT to WT (blue) was maximized at 63o C (y axis on the right for the ratio). Thus all the following detections of kRas mutations were carried out at 63o C at 4ml/min ow rate. 4.10

120
(a)
0.0

(b)

6 Mt Wt Mt/Wt

12

-0.5

10

f/f (x10 )

f/f (10 )

-1.0

4
-3

8
Wt Mt

-1.5
o o

-2.0

Mt 63 C Mt 60 C
o

Wt 63 C Wt 60 C Wt 55 C 15 20 25 30
o o

-2.5 0

Mt 55 C 5 10

63 C

60 C Temperature

55 C

Time (min)

Figure 4.10: Relative frequency shifts of detections of 10 fM of MT and WT kRas tDNAs are plotted versus time at 55o , 60o and 63o . (b) Signal (S) from MT (black) and WT (red) tDNAs (y-axis on the left)and ratio SMT /SW T (blue, y-axis on the right) are plotted for the detections in (a) were plotted

4.3.2.2 Detection of kRas MT and WT tDNA kRas mutation detections represented even an harder specicity problem with respect to the HBV-DM detections. Not only because kRas is a point mutation instead of a double mutation, but also the mismatch sites were also close (one base apart) in HBV-DM detections as discussed in chapter 1. This is overcomed with incorporation of LNA bases which required optimization of ow and temperature conditions once more. As shown previously in chapter 3, best specicity for the HBV-DM detections were achieved with a 4ml/min ow rate, thus mixture experiments of kRas mutation detections are carried out using this ow rate. Detections of 100 zM, 1, 10, and 100 aM of MT-tDNA (gure 4.11a) and 100 zM, 1 aM, 10 aM, 100 aM, 10 fM and 100 fM WT-tDNA (gure 4.11b) is done at this

121 condition. WT-tDNA did not show a signicant signal until a threshold concentration of 10 aM is exceeded. Moreover 10 fM and 100 fM showed a very similar signal indicating the hybridization of WT reaches a plateau much sooner than MT-tDNA as can be seen in gure 4.11c. As expected the signal from the FRM hybridization was around 1.2 fold of that of tDNA hybridization for every concentration. The signal ratio of MT/WT tDNA are plotted in gure 4.11d. Contrary to the HBV-DM mutation detections shown in gure 4.8d, signal ratio of MT/WT is observed inversely proportional with concentration, although the ratio is higher for any given concentration with respect to HBV-DM mutation detections. This inverse proportionality is probably meaningless as because as can be seen in gure 4.11b these three concentrations (1 aM, 10 aM and 100 aM) are very close to control level. However as can be extrapolated from the gure 4.11c, SMT at 100 aM tDNA is more than 10 times higher than SMT at 1pM tDNA. Moreover SMT did not reach plateau at 100 aM whereas SW T is already in the plateau in 1pM.

122
(a)
0.0 MT tDNA FRM

(b)
0.00

WT tDNA

FRM

-0.3

-0.05
-0.6
f/f (10 )

f/f (10 )

-3

-3

-0.9 100 zM -1.2 1 aM 10 aM 100 aM

-0.10

1 aM 10 aM

-0.15

100 aM 10 fM 100 fM

-1.5

-0.20

-1.8 0 10 20 30 Time (min) 40 50 60

1 pM 10 20 30 Time (min) 40 50 60

(c)
1.8 1.5 1.2
f/f (10 )
-3

(d)
MT tDNA MT FRM WT tDNA WT FRM
/S
WT

tDNA 10
4

FRM

0.9 0.6 0.3 0.0 10

-19

MT

10

10

10

-18

10

-17

10

-16

10

-15

10

-14

10

-13

10

-12

10

-18

10

-17

10

-16

tDNA Concentration (M)

tDNA Concentration (M)

Figure 4.11: Relative frequency shifts of detections of (a)100 zM, 1, 10 and 100 aM of MT tDNA and (b) 100 zM, 1, 10, 100 aM, 10fM and 100 fM of WT-tDNA are plotted versus time. (c) Signal(S) (average of relative frequency shifts between 25th )-30th min.(tDNA hybridization) and 55th -60th min. (FRM hybridization) versus tDNA concentration of detections in (a) and (b). (d) SMT /SW T plotted versus tDNA concentration.

123 4.3.2.3 Detection of mixtures of kRas MT and WT tDNA At the optimum condition discussed above mixture of MT and WT kRas tDNA detections were done at a xed ratio of MT:WT of 1:1000. 100zM, 1, 10 and 100 aM of MT tDNA is mixed with 1000 fold more amount of WT tDNA in urine samples. tDNA detections are followed by MT-FRM and WT-FRM hybridizations as shown in gure 4.12a. As expected, concentration and relative frequency shifts were directly proportional for both tDNA and FRM hybridization parts. Signal(S) of tDNA and FRM hybridization parts of kRas mixture experiments, HBV-DM mixture experiments and HBV-DM MT tDNA detection experiments (denoted by @RT refering to room temperature) done in chapter 2 are compared in Figure 4.12b. Highest signals were from HBV-DM MT tDNA detections which were carried out in room temperature. This was expected because room temperature and 2ml/min ow rate was a favorable condition for hybridization as there were no need of specicity against WT tDNA. However, kRas mixture experiments showed a higher signal for the tDNA hybridization with respect to that shown by HBV-DM mixture experiments. This was probably related to the high hybridization afnity of LNA containing pDNAs with tDNAs. However interestingly, FRM hybridization parts indicated the opposite. Although higher numbers of FRMs are hybridized on PEPS surface for the kRas mixture detections (as shown in Figure 4.12c), signal from the PEPS was higher in the HBV-DM mixture detections. This may be related to temperature dependence of stress generated on the PEPS surface by FRMs or to the extend that the stress is turned into resonance frequency shifts by PEPS at different temperatures. Micrographs of MT-FRMs and WT-FRMs are also shown in Figure 4.12d.

124
tDNA 0.0 -0.3 -0.6
f/f (10 ) f/f (10 )
-3 -3

mix.

FRM 3 tDNA
FRM HBVDM (@RT)

tDNA

-0.9 -1.2 -1.5 -1.8

HBVDM(@RT) FRM HBVDM FRM kRas kRas

tDNA 1

100 zM 1 aM 10 aM 100 aM

tDNA tDNA tDNA

0 30 Time (min) 40 50 60 10
-19

HBVDM

10

20

10

-18

10

-17

10

-16

10

-15

tDNA Concentration

Total FRM 75 WT FRM MT FRM 60

# FRMs

45

30

IV
15

I
0 10
-19

II
-18

III

10

10

-17

10

-16

tDNA Concentration (M)

Figure 4.12: (a) Relative frequency shift versus time graph of kRas mixture detection experiments with 100zM, 1, 10 and 100 aM of MT tDNA mixed with 1000 fold more amount of WT tDNA (Concentrations of MT tDNA in the mixture experiments are shown on the x axis).(Signal of tDNA and FRM hybridization parts of kRas and HBV-DM mixture detection experiments compared with with that of HBV-DM MT tDNA detection experiments carried out in room temperature (denoted by @RT) in chapter 2). (c) Number of MTFRMs and WT-FRMs hybridized on PEPS surface after detections in (a). (d) Micrographs of PEPS surface after detections carried out in (a)

125 4.3.3 Multiplexed Detection of kRas codon 12 Mutations The PEPS array composed of six sensors are illustrated in gure 4.6. Each PEPS is immobilized with one pDNA in table 4.2. In air impedance spectra of each PEPS used in the array is shown in gure 4.13 together with the k31 calculated from the impedance using the formula in chapter 1.
0 -20

Phase Angle (deg.)

-20

Phase Angle (deg.)

Phase Angle (deg.)

31

=0.39

-40

31

=0.38

31

=0.39

-40

PEPS 1
-40

PEPS 2

PEPS 3

-60

-60

-60

-80

-80

-80 0 1 2 3 4 5 0 1 2 0 1 2 3 4

Frequency (MHz)
0 -40

Frequency (MHz)
40

Frequency (MHz)

Phase Angle (deg.)

Phase Angle (deg.)

Phase Angle (deg.)

-20

31

=0.39

31

=0.39

20

31

=0.39

PEPS 4
-40

-60

PEPS 5

PEPS 6
-20

-40

-60

-80

-60

-80

-80 0 1 2 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0

Frequency (MHz)

Frequency (MHz)

Frequency (MHz)

Figure 4.13: Impedance spectra of PEPS used in multiplexed detection. In order to see if there is any cross hybridization of tDNAs with pDNAs on PEPS surfaces, 100 aM of all 6 tDNAs is dissolved in different urine samples. Each PEPS is immobilized with a different pDNA and 3% BSA blocking is done as explained before. Samples of urines containing different tDNAs are own one by one, through the ow cell including the array of PEPS. Detections are carried out in the same condition as before (63o C and 4 ml/min ow rate). Relative frequency shifts of all PEPS are recorded simul-

126 taneously and plotted as seen in gure 4.14. As can be seen no cross hybridization is observed. This was expected as not only one but 2 bases were mismatching for half (15) of the cross hybridizations and the rest (15) were one mismatch cross hybridizations, which would be no different than separation of MT and WT tDNA as shown in previous results. 4.14

127

f/f (10 )

-3

0.0 -0.5 -1.0 0 0.0 -0.5 -1.0 0

30

60

90

120

150

180

f/f (10 )

-3

30

60

90

120

150

180

f/f (10 )

-3

0.0 -0.5 -1.0 0

0.0 -0.5 -1.0 0 0.0 -0.5 -1.0 0

30

60

90

120

150

180

f/f (10 )

-3

30

60

90

120

150

180

f/f (10 )

-3

30

60

90

120

150

180

f/f (10 )

-3

0.0 -0.5 -1.0 0

30

60

90

120

150

180

Figure 4.14: Relative frequency shifts of PEPS used in multiplexed detection of 6 mutations of kRas codon 12, known to be associated with colorectal carcinoma. 6 urine samples containing 6 different mutations are own one by one over the PEPS array. Each sample was own for 30 min.

128 4.4 Summary of Single Stranded DNA mutation detections 100 zM concentrations of HBV-DM and kRas MT tDNA can be detected in urine with a background of 250 and 1000 fold more WT tDNA respectively. Possibility of multiplexed tDNA detection was shown for 6 possible mutations of a single codon. Considering mutations/SNPs and other chromosomal abnormalities associated to a disease would oocur in different regions rather than in a single codon, multiplexed detected can be easily accomplished using PEPS. LNA-DNA hybrid pDNAs showed an increase in specicity such that, even the ]DeltaTm difference between MT and WT was much lower in the single mutation with respect to double mutation (2o C and 20o C respectively) if pDNA were pure DNA, MT tDNA could be detected in urine with a background of higher concentration of WT (1000 fold with respect to 250 fold in single mutation).

129

Table 4.1: The sequences and Tm of the pDNA, tDNA, urDNA and drDNA for detection of mutations of kRas and HBV-DM Type of DNA drDNA (kRas) urDNA (kRas) pDNA (kRas) MT-tDNA (kRas) WT-tDNA (kRas) drDNA (HBV-DM) urDNA (HBV-DM) pDNA (HBV-DM) MT-tDNA (HBV-DM) WT-tDNA (HBV-DM) Sequence (5 to 3) Amine-C12-GCTGAAAATGACTGAATATA AACTTGTGGTAGT GCAAGAGTGCCTTGACGATA CAGCTAATTCAGA-C7 -Amine Amine-C12-TGGAGCT+G+T+TGGCGTAG TCTGAATTAGCTGTATCGTCAAGGCAC TCTTGCCTACGCCAACAGCTCCA CTACGCCACCAGCTCCAACTACCAC AAGTTTATATTCAGTCATTTTCAGC Amine-C12-ACAGACCAATTTATGC CTACAGCCTCCTAG AATCTCCTCCCCCAACTCCTC CCAGTCTTT-C7 -Amine Amine-C12-ACAAAGATCATTAACC TTTAAAGACTGGGAGGAGTTGG GGGAGGAGATTAGGTTAATGATCTTTGT GGTTAAAGGTCTTTGTACTAGG AGGCTGTAGGCATAAATTGGTCTGTTCA Tm 73

78

70

55

76

77

47

23

Mutation sites indicated by underlines in MT-tDNAs and WT-tDNAs, and in bold letters in pDNAs Fragments of tDNAs hybridizing with pDNA are shown in bold on pDNA sequences LNA bases are indicated by a + sign preceding the nucleic acid Tm s are melting temperatures with tDNA(Co )

130

Table 4.2: The sequences of pDNAs and Tm s of the pDNA with MT and WT tDNA, for detection of mutations of 6 kRas mutations PEPS name PEPS1 PEPS2 PEPS3 PEPS4 PEPS5 PEPS6 Codon 12 Mutation AGT CGT GAT GCT GTT TGT pDNA BioTEG-tggagc+t+A+gtgcgtag BioTEG-tggagc+t+C+gtgcgtag BioTEG-tggagct+g+A+tgcgtag BioTEG-tggagct+g+C+tgcgtag BioTEG-tggagct+g+T+tgcgtag BioTEG-tggagc+t+T+gtgcgtag
MT Tm WT Tm

Tm 15.3 20.9 15.3 20.9 15.7 15.7

68 71 70 72 70 69

52.7 50.1 54.7 51.1 54.3 53.3

Codon 12 complementary region on pDNAs are indicated in bold letters pDNAs are 5 biotinylated Mutation sites are indicated by capital letters LNA bases are indicated by a + sign preceding the nucleic acid All sequences are from 5 to 3 Biotin-TEG is a biotin attached to a 15-atom mixed polarity triethylene glycol spacer

131 5. Detection of double stranded DNA

5.1 Introduction Detection of any double stranded via hybridization may only be done by denaturation of the two complementary strands it composed of as described in chapter 1. Although hydrogen bonding can oocur between triplets (RNA triplets [220], or DNA triplets [221]) and even quadruple helices in cells as discovered recently [222] in different occasions, none of these kind of hybridizations or recongnitions would be sufciently strong to detect minute amounts of transrenal DNA in urine. Boiling a sample of urine would be sufcient in order to facilitate heat denaturation . However in order to prevent annealing or hybridization of heat denatured complementary strands in a sample, it has to be cooled down as fast as possible to the detection temperature. The denaturation or annealing of DNA can be monitored in real time using absorbance as it is discussed in chapter 1. In gure 5.1 [223], the effect of cooling rate is shown by measuring the absorption of light with 260nm wavelength. As can be seen, rapid cooling allows only the formation of local regions of dsDNA, formed by the base pairing or annealing of short regions of complementarity within or between DNA strands; the decrease in A260 is hence rather small. On the other hand, slow cooling allows time for the wholly complementary DNA strands to nd each other, and the sample can become fully double-stranded, with the same absorbance as the original native sample. 5.1

132

Figure 5.1: buraya caption yaz In this study the effect of cooling rate is investigated and the fastest cooling method is developed for the detection of kRas codon 12 GGT to GTT transversion and HBV 1762T/1764A double mutation. In order to prevent annealing of denatured complementary DNA fragments further, capture DNA fragments, having exactly the same sequence as the target DNA, are immobilized on GCGs. Effect of this competitive inhibition on annealing is investigated. Detection of kRas and HBVDM double stranded DNA fragments in urine is achieved with 100 zM limit of detection in a background of 1000 and 250 fold more concentration of WT double stranded DNA fragments.

133 5.2 Materials and Methods 5.2.1 PEPS fabrication, electrical insulation, pDNA immobilization and BSA blocking PEPS fabrication, electrical insulation and BSA blocking are done as discussed in chapters 4 and 2. pDNA immobilization is done using sulfo-SMCC as a linker between MPS covered PEPS surface and amine activated pDNA for detections of HBVDM double stranded and kRas codon 12 GGT to GTT transversion DNA mutations as discussed in chapters 3 and 4.

5.2.2 Target DNAs, probe DNA, reporter DNAs and FRMs pDNA sequences used in this study was the same with those used in chapter 4 for detection of HBV-DM and kRas single stranded mutations (4.1). 90 bp. tDNAs on the other hand were double stranded and supplied from IDT (Integrated DNA Technologies) sequence of which are shown in table 5.1. urDNAs and drDNAs were the same as in table 4.1 in chapter 4. Only Bright Blue FRMs were used in this study since MT and WT tDNA both have the upstream and downstream regions thus it was not possible to hybridize different color FRMs on MT and WT tDNA when mixture experiments were being conducted. Double stranded DNA is not only detected from synthetic sources, but also from SW480 cells cultured in cell medium. SW480 cell line is homozygous for kRas codon 12 GGT to GTT transversion. DNA is extracted from the cells using whole human genome DNA extraction kit and then sonicated to generate an average fragment size of 500 bps. The

134

Table 5.1: The sequences of tDNAs, used for detection of mutations of double stranded kRas codon 12 GGT to GTT point mutation and HBV-DM Type of DNA MT-tDNA (kRas) Sequence (5 to 3) GGTTCTGAATTAGCTGTATCGTCAAGGCA CTCTTGCCTACGCCAACAGCTCCAACTACC ACAAGTTTATATTCAGTCATTTTCAGCAGGC GGTTCTGAATTAGCTGTATCGTCAAGGCA CTCTTGCCTACGCCACCAGCTCCAACTACC ACAAGTTTATATTCAGTCATTTTCAGCAGGC TTGTTTAAAGACTGGGAGGAGTTGGGGGAGG AGATTAGGTTAATGATCTTTGTACTAGGA GGCTGTAGGCATAAATTGGTCTGTTCACCA TTGTTTAAAGACTGGGAGGAGTTGGGGGAGG AGATTAGGTTAAAGGTCTTTGTACTAGGA GGCTGTAGGCATAAATTGGTCTGTTCACCA

WT-tDNA (kRas)

MT-tDNA HBV-DM

WT-tDNA HBV-DM

Mutation sites indicated by underlines in MT-tDNAs and WT-tDNAs, Fragments of tDNAs hybridizing with pDNA are shown in bold on pDNA sequences extracted cells are then dissolved in urine samples of required concentrations and detections were same as in kRas synthetic double stranded tDNA at 63o C and 4ml/min ow rate.

5.2.3 Urine samples and Experimental Setup Urine samples were collected and prepared as it was done in chapters 2 and 4. Experimental setup involved a modied version of the setups used in those chapters. Two ow cells are used for detection as shown in gure 5.2.

135

Figure 5.2: Schematic of the experimental setup used to detect dsDNA.(a) Urine samples are heated to boiling temperature for 15 min. Hot sample (shown in red piping) is own through a water bath that is inside an incubator which is equilibrated to detection temperature. (b) Schematic of the ow system is shown. Sample is own through 2 ow cells including capture DNA immobilized GCG and pDNA immobilized PEPS in order. Detected sample is then discarded in the waste container. Dashed line illustrates the incubator in (a), everything inside the dashed line is in the incubator. 5 amine activated capture DNA, having the same sequence as tDNA is immobilized

136 on 3mm2 GCG as described in chapter 3. Capture DNA is used to hybridize to the already denatured double stranded DNA in urine. Because it has the same sequence as tDNA of interest, it competes with tDNA in hybridizing the complementary strand thus decreasing the probability of annealing. Moreover three different DNA cooling rates have been investigated. Three cooling regimes can be summarized as in the following.

1) Slow cooling rate (22 min): Boiled urine sample (50 ml)is passed through 1m of plastic tubing (Tygon R3603) with an inner diameter of 1/16 inches. The tubing is immersed in iced water bath (4-10 o C). Collected sample was waited to be equilibrated with room temperature and detection is done afterwards. This method took approximately 22 min.

2) Moderate cooling rate (4 min): Boiled urine sample is passed through 2 m of same tubing. Tubing was in contact with ambient air. Total volume of the tubing was approximately 4ml, which took 4 min. for the sample to pass through (ow rate was kept constant at 1ml/min for all the detections) and the collected sample was at room temperature. Detection is directly done without no need of waiting for equilibration.

3) Fast Cooling rate (1 min): Boiled urine sample is passed through 50 cm. of tubing which is immersed in room temperature water bath. It took 1 min. for the sample to pass through the piping with 1ml/min ow rate. Detection was made directly with no need for equilibration.

137 5.3 Results and Discussion 5.3.1 Comparison of Denaturation and tDNA capture methods Effect of cooling rate and existence of capture DNA is investigated detecting 10 aM of 90 bp. HBV-DM double stranded DNA is done at room temperature after the above described cooling regimes. Figure 5.3a below is the relative frequency shift versus time graph for the experiments consisting of 6 possible combinations of 3 different cooling rates and 2 cases of existence of capture DNA. Relative frequency shift of 50 nts long ssDNA of same target done in room temperature (2.13b) is also plotted for comparison. As expected although the length of ssDNA was less than the dsDNA used in these detections the relative frequency shift was higher than any of the dsDNA detections. This is due to inevitable inefciency of cooling. Most probably, some portion of denatured DNA is hybridized back with its complementary before sample hits the PEPS surface. The percentage of relative frequency shifts of different denaturation regimes are plotted in gure 5.3b. Black bars indicate the shifts in different cooling regimes without capture DNA. For each cooling regime difference in percentage relative frequency shifts of detections with capture and without capture DNA is plotted in red (In other words, for each cooling regime on the x axis, sum of black and red bars indicate the percentage relative frequency shift with capture DNA and black bars represents the detections with no capture DNA). As can be seen effect of capture DNA was least in the slowest cooling rate. This was expected since 22 min. was already enough time for renaturation of dsDNA. When the sample was run over the capture DNA it was already renatured. The effect of capture DNA in the fast cooling regime was less than the moderate cooling regime, most probably due to fast cooling already negated

138 the need of capturing any denatured complementary tDNA because the signal is already high when no capture is used in fast cooling regime.
(a)
0.0

(b)
80

Capture DNA (+) Capture DNA (-)

-0.2
f/f (10 )
ssDNA dsDNA
-3

-0.6

/(
0 5 10 15 Time (min) Slow, Capture (-) Moderate, Capture (-) Moderate, Capture (+) ssDNA Slow, Capture (+) Fast, Capture (-) Fast, Capture(+) 20 25 30

f/f)

-0.4

60

40

( f/f)

20

0 Slow Moderate Cooling Rate Fast

Figure 5.3: (a) Relative frequency shifts of different cooling regimes are plotted for detection of 10 aM of HBV-DM dsMT tDNA in room temperature with 1ml/min ow rate. Relative frequency shift of same concentration of shorter ssDNA is plotted for comparison (b) Percentage of relative frequency shifts with respect to ssDNA detection of same concentration is plotted for different cooling regimes

5.3.2 HBV-DM double stranded mutation detections Detection of 100 zM, 1 aM, 10 aM and 100 aM of HBV-DM dsMT tDNA (gure 5.4a) and 100 zM, 1 aM, 10 aM, 100 aM, 10 fM and 100 fM of HBV-DM dsWT tDNA (gure 5.4b) is done in urine, followed by FRM hybridizations. Average relative frequency shifts in the last ve minutes of tDNA (25th -30th min) and FRM hybridizations (55th -60th min) are plotted for dsMT and dsWT tDNAs in gure 5.4c. As can be seen from these graphs dsWT tDNA reached plateau at 10 fM, and the signal was almost equal to that of 100 zM of

139 dsMT tDNA. Ratio of signal of dsMT tDNA to dsWT tDNA is also plotted in gure 5.4d. As can be seen, in the range between 100 zM to 100 aM, ratio was always higher than or equal to 15.
(a)
0.0 MT tDNA FRM

(b)

WT tDNA 0.00

FRM

-0.03 -0.3 -0.06

f/f (10 )
-3

-0.6

f/f (10 )

-3

-0.09

100 zM 1 aM

-0.9

100 zM 1 aM

-0.12

10 aM 100 aM

-1.2

10 aM 100 aM

-0.15

10 fM 100 fM

-0.18 30 40 50 60 0

-1.5 0

10

20

10

20

30 Time (min)

40

50

60

Time (min)

(c)

1.5

(d)
75

1.2

MT tDNA + FRM MT tDNA

60
WT MT

f/f (10 )

-3

0.9

WT tDNA + FRM

0.6

/S

WT tDNA

45

30 tDNA 15 tDNA + FRM

0.3

0.0 10
-19

10

-18

10

-17

10

-16

10

-15

10

-14

10

-13

10

-19

10

-18

10

-17

10

-16

tDNA Concentration (M)

tDNA Concentration (M)

Figure 5.4: Relative frequency shifts of (a) 100 zM, 1 aM, 10 aM and 100 aM of HBV-DM dsMT tDNA and (b) 100 zM, 1 aM, 10 aM, 100 aM, 10 fM and 100 fM of HBV-DM dsWT tDNA is plotted versus time. Detections are followed by FRM hybridizations. (c) Signal (S) is plotted versus tDNA concentration for dsMT and dsWT tDNA and FRM hybridizations. (d) SMT /SW T ratios are plotted for tDNA and FRM hybridizations versus concentration of tDNA.

140 5.3.3 HBV-DM double stranded mutation mixture detections Detection of mixtures of HBV-DM dsMT and dsWT tDNA are done and conrmed by FRM hybridizations. However it is important to note that two color FRM scheme used in ssDNA detection in chapter 4 could not be used since both tDNAs have the upstream and downstream regions, thus FRMs were not specic to MT-tDNA or WT-tDNA, they could hybridize on both. Although a specic conrmation was not able to be performed a quantitative one was still possible. MT to WT tDNA ratio was kept at 250 as it was done in ssDNA detection of the same target. Figure 5.5a shows the relative frequency shift versus time graph for tDNA and FRM hybridization parts for mixture experiments involving dsMT tDNA of concentrations of 100 zM, 1 aM, 10 aM and 100 aM (250 fold more dsWT tDNA was mixed in each detection, concentration of dsMT was indicated on gure 5.5in a, b and c). The comparison of mixture experiments with dsMT tDNA-only and dsWT tDNAonly are also investigated. Average relative frequency shifts at the last ve minutes of tDNA and FRM hybridization experiments are plotted versus concentration for mixture experiments and separately performed experiments of dsMT and dsWT tDNA (data from gure 5.4a and b respectively) in gure 5.5b. The sum of signal of separately performed dsMT and dsWT tDNA experiments were always higher than that of mixture experiments as expected because existence of dsWT tDNA interfered with the hybridization of dsMT tDNA as shown in the ssDNA case in chapter 4. Moreover the inhibitive nature ofdsWT tDNA on dsMT tDNA was more aberrant in higher concentrations in terms of absolute relative frequency shift difference. Number of FRMs versus concentration graph was also plotted in gure 5.5c. As expected, a direct proportionality is observed between number of FRMs hybridized and concentration of tDNAs in the mixture. Micrographs of FRMs are

141 also shown on gure 5.5d. The width of each micrograph is approximately 400 nm.
(a)
0.0 tDNA FRM

(b)
1.8

MT tDNA WT tDNA

-0.2

1.5

Mix. tDNA MT tDNA+FRM

f/f (10 )

-3

-3

-0.4 100 zM 1 aM -0.8 10 aM 100 aM -1.0

f/f (10 )

1.2

WT tDNA+FRM Mix tDNA+FRM

0.9

-0.6

0.6

0.3

0.0 0 10 20 30 Time (min) 40 50 60 10

-19

10

-18

10

-17

10

-16

tDNA Concentration (M)

(c)
50

40

# FRMs

30

20

10

10

-19

10

-18

10

-17

10

-16

tDNA Concentration (M)

Figure 5.5: (a) Relative frequency shift versus time graph of mixture of dsMT and dsWT tDNA followed by FRM hybridizations. Concentration of dsMT tDNA is indicated on gure. 250 fold more dsWT tDNA is mixed with dsMT tDNA in each experiment. (b) Plot of average relative frequency shifts in last ve minutes of tDNA and FRM hybridization parts of mixture detections in (a) versus dsMT tDNA concentration and comparison with sum of average of relative frequency shifts of dsMT tDNA-only and dsWT tDNA-only detections in gure 5.4. (c)Plot of number of FRMs hybridized on PEPS surface after each mixture detection versus dsMT tDNA concentration. (d) Micrographs of FRMs on PEPS surface after mixture detection experiments in (a)

142 5.3.4 kRas double stranded mutation detections Detection of 100 zM, 1 aM, 10 aM and 100 aM of kRas dsMT tDNA (gure 5.6a) and 10 aM, 100 aM, 10 fM, 100 fM, and 1 pM of kRas dsWT tDNA (gure 5.6b) is done in urine, followed by FRM hybridizations. Average relative frequency shifts in the last ve minutes of tDNA (25th -30th min) and FRM hybridizations (55th -60th min) are plotted for dsMT and dsWT tDNAs in gure 5.6c. As can be seen from these graphs dsWT tDNA reached plateau at 100 fM, and the signal was almost equal to that of 100 zM of dsMT tDNA. Ratio of signal of dsMT tDNA to dsWT tDNA is also plotted in gure 5.6d. As can be seen, for concentration of 10 aM and 100 aM, ratio was higher than or equal to 70, which is much larger than that observed in detection of HBV-DM double stranded tDNA (15 in the latter). As can be seen from gures 5.4b and 5.6b SMT of HBV-DM double stranded DNA at 100 aM was 0.65 and that of kRas was 0.85. There was already more signal in the kRas case. Moreover, the detections of kRas are done at a higher temperature (63o )C, in which the noise in the control was less with respect to noise in the control experiments of HBV-DM experiments which are done at 30o C. This decrease in the background noise can be attributed to the behavior of PEPS in high temperatures. The higher SMT can also be attributed to again the same behavior of PEPS in high temperature. The reason for this behavior of PEPS is that, high temperatures are closer to the Curie temperature of PEPS where domain switching in the crystal structure occurs.

143
(a)
0.0 0.00 -0.3 -0.04
f/f (10 ) f/f (10 )

MT tDNA

FRM

(b)

WT tDNA

FRM

-0.6

-3

-3

-0.08

-0.9

10 aM 100 aM

100 zM 1 aM

-0.12

-1.2

10 fM 100 fM

10 aM 100 aM -0.16

-1.5

1 pM

10

20

30 Time (min)

40

50

60

10

20

30 Time (min)

40

50

60

(c)
1.5 MT tDNA + FRM MT tDNA 1.2 WT tDNA + FRM WT tDNA
f/f (10 )
-3

(d)

300 tDNA 250 tDNA + FRM

200
WT

0.9

0.6

MT

/S

150

100

0.3

50

0.0 10
-18

0 10
-16

10

-14

10

-12

10

-17

10

-16

tDNA Concentration (M)

tDNA Concentration (M)

Figure 5.6: Relative frequency shifts of (a) 100 zM, 1 aM, 10 aM and 100 aM of kRas dsMT tDNA and (b) 10 aM, 100 aM, 10 fM, 100 fM, and 1 pM of kRas dsWT tDNA is plotted versus time. Detections are followed by FRM hybridizations. (c) Signal (S) is plotted versus tDNA concentration for hybridizations of dsMT, dsWT tDNA and FRM. (d) SMT /SW T ratios are plotted for tDNA and FRM hybridizations versus concentration of tDNA.

144 5.3.5 kRas double stranded mutations mixture detections Detection of mixtures of kRas dsMT and dsWT tDNA are done and conrmed by FRM hybridizations. However again the two color FRM scheme used in ssDNA detection in chapter 4 could not be used since both tDNAs have the upstream and downstream regions, thus FRMs were not specic to MT-tDNA or WT-tDNA. MT to WT tDNA ratio was kept at 1000 in kRas case as it was done in ssDNA detection of the same target. Figure 5.7a shows the relative frequency shift versus time graph for tDNA and FRM hybridization parts for mixture experiments involving dsMT tDNA of concentrations of 100 zM, 1 aM, 10 aM and 100 aM (1000 fold more dsWT tDNA was mixed in each detection, concentration of dsMT was indicated on gure 5.7in a, b and c). The comparison of mixture experiments with dsMT tDNA-only and dsWT tDNA-only are also investigated. Average relative frequency shifts at the last ve minutes of tDNA and FRM hybridization experiments are plotted versus concentration for mixture experiments and separately performed experiments of dsMT and dsWT tDNA (data from gure 5.6a and b respectively) in gure 5.7b. The sum of signal of separately performed dsMT and dsWT tDNA experiments were always higher than that of mixture experiments as expected because existence of dsWT tDNA interfered with the hybridization of dsMT tDNA as shown in the ssDNA case in chapter 4. Moreover the inhibitive nature of dsWT tDNA on dsMT tDNA was more aberrant in higher concentrations in terms of absolute relative frequency shift difference but not as much as that occurred in HBV-DM double stranded tDNA detections. Number of FRMs versus concentration graph was also plotted in gure 5.7c. As expected, a direct proportionality is observed between number of FRMs hybridized and concentration of tDNAs in the mixture. Micrographs of FRMs are also shown on gure 5.7d. The width of each

145 micrograph is approximately 400 nm.

(a)
0.0 -0.2 -0.4
f/f (10 )
f/f (10 )
-3

tDNA

FRM

(b)
1.8

Mt tDNA WT tDNA

1.5

Mix tDNA Mt tDNA+FRM

1.2

WT tDNA+FRM Mix tDNA+FRM

-3

-0.6 100 zM -0.8 -1.0 -1.2 1 aM 10 aM 100 aM

0.9

0.6

0.3

0.0

10

20

30 Time (min)

40

50

60

10

-19

10

-18

10

-17

10

-16

(c)
60

tDNA Concentration (M)

(d)

50

40
# FRMs

30

20

10

10

-19

10

-18

10

-17

10

-16

tDNA Concentration (M)

Figure 5.7: (a) Relative frequency shift versus time graph of mixture of dsMT and dsWT tDNA followed by FRM hybridizations. Concentration of dsMT tDNA is indicated on gure. 1000 fold more dsWT tDNA is mixed with dsMT tDNA in each experiment. (b) Average relative frequency shifts in last ve minutes of tDNA and FRM hybridization parts of mixture detections in (a) is plotted versus dsMT tDNA concentration and compared with sum of average of relative frequency shifts of dsMT tDNA-only and dsWT tDNA-only detections made in gure 5.6. (c) Number of FRMs hybridized on PEPS surface after each mixture detection is plotted versus dsMT tDNA concentration. (d) Micrographs of FRMs on PEPS surface after mixture detection experiments in (a) is shown

146 5.3.6 SW480 Cell Line Extracted double stranded DNA detections Detection of SW480 cell line extracted DNA is done for 100 zM, 1 aM, 10 aM and 100 aM of tDNA. Calculation of concentration is done considering the cells are homozygous for the mutation (each cell has 2 copies of mutation). Relative frequency shift versus time graph of SW480 tDNA and FRM hybridizations are plotted in gure 5.8a. The relative frequency shifts in the last ve minutes of tDNA and FRM hybridization parts of ds-SW480 tDNA detections are compared with synthetic kRas dsMT tDNA (data from gure 5.6a). The major difference in these detections are in ds-SW480 detections fragment size is not exact, however on average it is 500 bp. Thus hybridization of a 500 nts tDNA would be different than a 90 nts tDNA in dsMT tDNA detections. On the other hand, FRMs have the potential to hybridize on any tDNA captured by the pDNA on PEPS surface in synthetic tDNA detection case, since all the tDNAs possessed both the upstream and downstream regions certainly. On ds-SW480 tDNA case however, all tDNAs may not necessarily have upstream and/or downstream regions, or they may simply have portions of one and/or both. Thus, number of FRMs that could hybridize were expected to be less than that occurred in synthetic dsMT tDNA detection case. As shown in gure 5.8b, signals in both tDNA and FRM hybridization parts of ds-SW480 detections are slightly less than (if not equal to) those of dsMT tDNA. The uncertainty in tDNA sequence in ds-SW480 should not only affect the hybridization of FRMs but also hybridization of tDNA with pDNA on PEPS surface since, not all the fragments necessarily have the pDNA complementary region, some of them might have some portion of it. The length of pDNA on the hand was much less than the fragments size (16 to 500), decreasing the probability of such occurrences (probability that any fragments not including the whole pDNA sequence was 16/500 or

147 approximately 3%). The number of FRMs hybridized on PEPS surface after ds-SW480 tDNA detection are also plotted in gure 5.8c. As can be seen the direct proportionality of number of FRMs on surface with tDNA concentration was conserved as in all previous detections. However in the ds-SW480 tDNA detection case, there was a signicant drop in number of FRMs hybridized on PEPS surface with respect to that occurred in the synthetic dsMT tDNA detection case (drop was observed in all concentrations at an approximately 60% rate). Micrographs of FRMs on PEPS surface after these detections are also shown in gure 5.8d.

148
(a)
0.0 tDNA FRM

(b)

1.8 SW480 tDNA 1.5 SW480 tDNA + FRM kRas MT tDNA 1.2 kRas MT tDNA + FRM

-0.3

f/f (10 )

-3

-0.6
f/f (10 )
-3

0.9

-0.9 100 zM -1.2 1 aM 10 aM -1.5 100 aM

0.6

0.3

0.0 30 40 50 60 10
-19

10

20

10

-18

10

-17

10

-16

(c) 40

Time (min)

tDNA Concentration (M)

(d)

30

# FRMs

20

10

10

-19

10

-18

10

-17

10

-16

tDNA Concentration (M)

Figure 5.8: (a) Relative frequency shift versus time graphs of ds-SW480 tDNA followed by FRM hybridizations. Concentration of ds-SW480 tDNA is indicated on gure. (b) Average relative frequency shifts in last ve minutes of tDNA and FRM hybridization parts of dsSW480 tDNA detections in (a) is plotted versus tDNA concentration and compared with average of relative frequency shifts of synthetic dsMT tDNA detections made in gure 5.6a. (c) Number of FRMs hybridized on PEPS surface after each ds-SW480 tDNA detection is plotted versus tDNA concentration. (d) Micrographs of FRMs on PEPS surface after mixture detection experiments in (a) is shown

149 An important point that should be mentioned is how would PEPS behave given a real patient sample with an unknown ratio of MT and WT. In order to best answer this, the worst case where there was abundant WT tDNA should be considered in healthy subjects. As described in the introduction part, normal subjects have approximately 30 ng of cfDNA per ml, an average of a population with cf-DNA concentration ranging from 0-100 ng per ml [14]. Thus in the worst case a normal subject would have 100 ng of cf-DNA in blood. Considering each cell have 3.2 x 109 bps and each bp has a molecular weight of 660 gr/mole on average, one cell would have approximately 3.5pg of DNA. Thus 100 ng of cf-DNA would mean approximately 29,000 cells per ml. Each cell would have 2 copies of the WT tDNA thus a normal subject would have at most approximately 100 aM of MT tDNA in blood. Of course not all of this DNA can pass through the kidney barrier but for simplicity we assume the concentration of DNA in urine and blood is same because DNA in urogenital tract can pass in to the urine.

As shown in gures 5.4 and 5.6, signal from hybridization of 100 aM of WT tDNA is approximately 4 and 5 times lower than that of 1 aM of MT tDNA in double mutation and point mutation cases respectively. Thus even in the worst case a positive detection of 1 aM of ds-tDNA is achievable by using PEPS.

5.4 Summary of Double Stranded DNA mutation detections Detection of HBV-DM and kRas double stranded MT tDNA was accomplished with a 100 zM limit of detection on a background of 250 and 1000 fold more amount of WT

150 tDNA. Using a heat denaturation and rapid cooling (taking approximately 15 sec. with a 4ml/min ow rate) regime, although the signal strength (relative frequency shift) dropped to about 80% with respect to single stranded tDNA detection, limit of detection remained unchanged. Detection of SW480 cell line extracted double stranded tDNA fragments with an average length of 500 bp is accomplished with the same limit of detection with 100 zM. Moreover detections are conrmed with FRM hybridizations. These resuls indicate that PEPS is capable of detecting less than 1 aM of point mutations specically, without need of amplication or isolation of DNA.

151 6. Conclusion

6.1 Optimization of Temperature and Flow Rate We have incorporated ow in gene mutation detection and used uorescent microspheres to optimize the ow rate together with temperature and achieved better mutation detection specicity than by changing the temperature alone.

6.2 Resonance Frequency Peak Determination Method Peak determination method has increased the SNR and allowed detection of <60 copies/ml tDNA with an SNR of 10. Less than 12% error has been estimated as shown by the simulations

6.3 Detection of Single Stranded DNA fragments Single-stranded tDNA of HBV-DM and kRas codon 12 GGT GTT point mutation have been detected in urine at the optimum temperature and ow rate. Detection of 100 zM single stranded tDNA has been accomplished in urine and unambiguous conrmation of MT and WT tDNA has been achieved using 2 color FRM hybridization MT tDNA was detected in urine and conrmed at the lowest concnetration of 100 zM in a background of 250 fold and 1000 fold more WT tDNA in double mutation and point mutation cases respectively.

152 LNA-DNA hybrid probes used to detect point mutations showed increased specicity (WT/MT = 1000 ) yet at a higher temperature (63o C) with respect to detections of double mutation (30o C).

6.4 Multiplexed Detection 1015 M of 6 different tDNAs (each perfectly matching with one pDNA immobilized on an array of 6 PEPS) have been run through the array PEPS one by one and hybridizations only detected on PEPS with pDNA perfectly matching with the tDNAs. At the optimum temperature of 63o C and 4 ml/min ow rate no cross hybridization has been observed.

6.5 Detection of Double Stranded DNA Sample at boiling temperature was cooled down to room temperature in 30 sec. for direct detection. Using capture DNA a 87% single stranded coverage was achieved. Without capture DNA the coverage dropped only 10% to yield 77% coverage 100 zM MT tDNA of HBV-DM and kRas codon 12 GGT GTT point mutation was detected in urine. 100 zM of SW480 cell line (homozygous for kRas codon 12 GGT GTT point mutation) extracted DNA was detected in urine and conrmed by FRM hybridization.

153 Bibliography [1] B. Alberts, J. H. Wilson, and T. Hunt, Molecular biology of the cell. New York: Garland Science, 5th ed., 2008. [2] J. M. Berg, J. L. Tymoczko, L. Stryer, and L. Stryer, Biochemistry. New York: W.H. Freeman, 6th ed., 2007. [3] H. Schwarzenbach, D. S. Hoon, and K. Pantel, Cell-free nucleic acids as biomarkers in cancer patients, Nat Rev Cancer, vol. 11, no. 6, pp. 42637, 2011. [4] C. Green, J. F. Huggett, E. Talbot, P. Mwaba, K. Reither, and A. I. Zumla, Rapid diagnosis of tuberculosis through the detection of mycobacterial dna in urine by nucleic acid amplication methods, The Lancet Infectious Diseases, vol. 9, no. 8, pp. 505511, 2009. [5] Y. You, B. G. Moreira, M. A. Behlke, and R. Owczarzy, Design of lna probes that improve mismatch discrimination, Nucleic Acids Res, vol. 34, no. 8, p. e60, 2006. [6] P. Mandel, Les acides nucleiques du plasma sanguin chez lhomme, CR Acad Sci Paris, vol. 142, pp. 241243, 1948. [7] G. D. Sorenson, D. M. Pribish, F. H. Valone, V. A. Memoli, D. J. Bzik, and S. L. Yao, Soluble normal and mutated dna sequences from single-copy genes in human blood, Cancer Epidemiology Biomarkers and Prevention, vol. 3, no. 1, pp. 6771, 1994. [8] V. Vasioukhin, P. Anker, P. Maurice, J. Lyautey, C. Lederrey, and M. Stroun, Point mutations of the n-ras gene in the blood plasma dna of patients with myelodysplastic syndrome or acute myelogenous leukaemia, Br J Haematol, vol. 86, no. 4, pp. 774 9, 1994. [9] J. J. Choi, r. Reich, C. F., and D. S. Pisetsky, The role of macrophages in the in vitro generation of extracellular dna from apoptotic and necrotic cells, Immunology, vol. 115, no. 1, pp. 5562, 2005. [10] M. Stroun, P. Maurice, V. Vasioukhin, J. Lyautey, C. Lederrey, F. Lefort, A. Rossier, X. Q. Chen, and P. Anker, The origin and mechanism of circulating dna, Ann N Y Acad Sci, vol. 906, pp. 1618, 2000.

154 [11] X. Chen, H. Bonnefoi, S. Diebold-Berger, J. Lyautey, C. Lederrey, E. Faltin-Traub, M. Stroun, and P. Anker, Detecting tumor-related alterations in plasma or serum dna of patients diagnosed with breast cancer, Clin Cancer Res, vol. 5, no. 9, pp. 2297 303, 1999. [12] P. B. Gahan and R. Swaminathan, Circulating nucleic acids in plasma and serum. recent developments, Ann N Y Acad Sci, vol. 1137, pp. 16, 2008. [13] F. Diehl, M. Li, D. Dressman, Y. He, D. Shen, S. Szabo, J. Diaz, L. A., S. N. Goodman, K. A. David, H. Juhl, K. W. Kinzler, and B. Vogelstein, Detection and quantication of mutations in the plasma of patients with colorectal tumors, Proc Natl Acad Sci U S A, vol. 102, no. 45, pp. 1636873, 2005. [14] M. Fleischhacker and B. Schmidt, Circulating nucleic acids (cnas) and cancera survey, Biochim Biophys Acta, vol. 1775, no. 1, pp. 181232, 2007. [15] B. Aaron, T. Wilczok, and E. Borenfreund, Circulating dna as a possible factor in oncogenesis, Science, vol. 148, no. 3668, pp. 374376, 1965. [16] P. Wimberger, C. Roth, K. Pantel, S. Kasimir-Bauer, R. Kimmig, and H. Schwarzenbach, Impact of platinum-based chemotherapy on circulating nucleic acid levels, protease activities in blood and disseminated tumor cells in bone marrow of ovarian cancer patients, Int J Cancer, vol. 128, no. 11, pp. 257280, 2011. [17] J. L. Boddy, S. Gal, P. R. Malone, A. L. Harris, and J. S. Wainscoat, Prospective study of quantitation of plasma dna levels in the diagnosis of malignant versus benign prostate disease, Clin Cancer Res, vol. 11, no. 4, pp. 13949, 2005. [18] A. A. Kamat, M. Baldwin, D. Urbauer, D. Dang, L. Y. Han, A. Godwin, B. Y. Karlan, J. L. Simpson, D. M. Gershenson, R. L. Coleman, F. Z. Bischoff, and A. K. Sood, Plasma cell-free dna in ovarian cancer: an independent prognostic biomarker, Cancer, vol. 116, no. 8, pp. 191825, 2010. [19] D. Allen, A. Butt, D. Cahill, M. Wheeler, R. Popert, and R. Swaminathan, Role of cell-free plasma dna as a diagnostic marker for prostate cancer, Annals of the New York Academy of Sciences, vol. 1022, no. 1, pp. 7680, 2004. [20] H. Schwarzenbach, J. Stoehlmacher, K. Pantel, and E. Goekkurt, Detection and monitoring of cell-free dna in blood of patients with colorectal cancer, Ann N Y Acad Sci, vol. 1137, pp. 1906, 2008.

155 [21] E. Sunami, A.-T. Vu, S. L. Nguyen, A. E. Giuliano, and D. S. B. Hoon, Quantication of line1 in circulating dna as a molecular biomarker of breast cancer, Annals of the New York Academy of Sciences, vol. 1137, no. 1, pp. 171174, 2008. [22] F. K. Chun, I. Muller, I. Lange, M. G. Friedrich, A. Erbersdobler, P. I. Karakiewicz, M. Graefen, K. Pantel, H. Huland, and H. Schwarzenbach, Circulating tumourassociated plasma dna represents an independent and informative predictor of prostate cancer, BJU Int, vol. 98, no. 3, pp. 5448, 2006. [23] I. Botezatu, O. Serdyuk, G. Potapova, V. Shelepov, R. Alechina, Y. Molyaka, V. Ananev, I. Bazin, A. Garin, M. Narimanov, V. Knysh, H. Melkonyan, S. Umansky, and A. Lichtenstein, Genetic analysis of dna excreted in urine: a new approach for detecting specic genomic dna sequences from cells dying in an organism, Clin Chem, vol. 46, no. 8 Pt 1, pp. 107884, 2000. [24] M. K. Al-Yatama, A. S. Mustafa, S. Ali, S. Abraham, Z. Khan, and N. Khaja, Detection of y chromosome-specic dna in the plasma and urine of pregnant women using nested polymerase chain reaction, Prenatal Diagnosis, vol. 21, no. 5, pp. 399402, 2001. [25] K. Koide, A. Sekizawa, M. Iwasaki, R. Matsuoka, S. Honma, A. Farina, H. Saito, and T. Okai, Fragmentation of cell-free fetal dna in plasma and urine of pregnant women, Prenat Diagn, vol. 25, no. 7, pp. 6047, 2005. [26] W. J. Shi, Y. H. Sun, L. J. Cui, Y. Zhang, and S. R. Li, [detection of cell-free fetal dna in the urine of pregnant women by nested polymerase chain reaction], Zhonghua Yi Xue Za Zhi, vol. 83, no. 6, pp. 4824, 2003. [27] J. Zhang, K. L. Tong, P. K. Li, A. Y. Chan, C. K. Yeung, C. C. Pang, T. Y. Wong, K. C. Lee, and Y. M. Lo, Presence of donor- and recipient-derived dna in cellfree urine samples of renal transplantation recipients: urinary dna chimerism, Clin Chem, vol. 45, no. 10, pp. 17416, 1999. [28] Y. H. Su, M. Wang, D. E. Brenner, A. Ng, H. Melkonyan, S. Umansky, S. Syngal, and T. M. Block, Human urine contains small, 150 to 250 nucleotide-sized, soluble dna derived from the circulation and may be useful in the detection of colorectal cancer, J Mol Diagn, vol. 6, no. 2, pp. 1017, 2004. [29] K. C. A. Chan, S. F. Leung, S. W. Yeung, A. T. C. Chan, and Y. M. D. Lo, Quantitative analysis of the transrenal excretion of circulating ebv dna in nasopharyn-

156 geal carcinoma patients, Clinical Cancer Research, vol. 14, no. 15, pp. 48094813, 2008. [30] S. Y. Lin, V. Dhillon, S. Jain, T. T. Chang, C. T. Hu, Y. J. Lin, S. H. Chen, K. C. Chang, W. Song, L. Yu, T. M. Block, and Y. H. Su, A locked nucleic acid clampmediated pcr assay for detection of a p53 codon 249 hotspot mutation in urine, J Mol Diagn, vol. 13, no. 5, pp. 47484, 2011. [31] X. D. Ren, S. Y. Lin, X. Wang, T. Zhou, T. M. Block, and Y.-H. Su, Rapid and sensitive detection of hepatitis b virus 1762t/1764a double mutation from hepatocellular carcinomas using lna-mediated pcr clamping and hybridization probes, J Virol Methods, vol. 158, no. 12, pp. 2429, 2009. [32] B. P. Song, S. Jain, S. Y. Lin, Q. Chen, T. M. Block, W. Song, D. E. Brenner, and Y. H. Su, Detection of hypermethylated vimentin in urine of patients with colorectal cancer, J Mol Diagn, vol. 14, no. 2, pp. 1129, 2012. [33] G. Ruano, K. K. Kidd, and J. C. Stephens, Haplotype of multiple polymorphisms resolved by enzymatic amplication of single dna molecules, Proc Natl Acad Sci U S A, vol. 87, no. 16, pp. 6296300, 1990. [34] A. J. Jeffreys, V. Wilson, R. Neumann, and J. Keyte, Amplication of human minisatellites by the polymerase chain reaction: towards dna ngerprinting of single cells, Nucleic Acids Research, vol. 16, no. 23, pp. 1095310971, 1988. [35] H. H. Li, U. B. Gyllensten, X. F. Cui, R. K. Saiki, H. A. Erlich, and N. Arnheim, Amplication and analysis of dna sequences in single human sperm and diploid cells, Nature, vol. 335, no. 6189, pp. 4147, 1988. [36] M. Nakano, J. Komatsu, H. Kurita, H. Yasuda, S. Katsura, and A. Mizuno, Adaptor polymerase chain reaction for single molecule amplication, Journal of Bioscience and Bioengineering, vol. 100, no. 2, pp. 216218, 2005. [37] A. J. Jeffreys, R. Neumann, and V. Wilson, Repeat unit sequence variation in minisatellites: a novel source of dna polymorphism for studying variation and mutation by single molecule analysis, Cell, vol. 60, no. 3, pp. 47385, 1990. [38] S. R. Umansky and L. D. Tomei, Transrenal dna testing: progress and perspectives, Expert Rev Mol Diagn, vol. 6, no. 2, pp. 15363, 2006.

157 [39] Y. H. Su, M. Wang, T. M. Block, O. Landt, I. Botezatu, O. Serdyuk, A. Lichtenstein, H. Melkonyan, L. D. Tomei, and S. Umansky, Transrenal dna as a diagnostic tool: important technical notes, Ann N Y Acad Sci, vol. 1022, pp. 819, 2004. [40] D. M. Hammond, A. Manetto, J. Gierlich, V. A. Azov, P. M. Gramlich, G. A. Burley, M. Maul, and T. Carell, Dna photography: an ultrasensitive dna-detection method based on photographic techniques, Angew Chem Int Ed Engl, vol. 46, no. 22, pp. 41847, 2007. [41] M. Passamano and M. Pighini, Qcm dna-sensor for gmos detection, Sensors and Actuators B: Chemical, vol. 118, no. 12, pp. 177181, 2006. [42] K. Feng, J. Li, J. H. Jiang, G. L. Shen, and R. Q. Yu, Qcm detection of dna targets with single-base mutation based on dna ligase reaction and biocatalyzed deposition amplication, Biosens Bioelectron, vol. 22, no. 8, pp. 16517, 2007. [43] R. Gasparac, B. J. Taft, M. A. Lapierre-Devlin, A. D. Lazareck, J. M. Xu, and S. O. Kelley, Ultrasensitive electrocatalytic dna detection at two- and three-dimensional nanoelectrodes, Journal of the American Chemical Society, vol. 126, no. 39, pp. 1227012271, 2004. [44] S. J. Park, T. A. Taton, and C. A. Mirkin, Array-based electrical detection of dna with nanoparticle probes, Science, vol. 295, no. 5559, pp. 15036, 2002. [45] L. He, M. D. Musick, S. R. Nicewarner, F. G. Salinas, S. J. Benkovic, M. J. Natan, and C. D. Keating, Colloidal au-enhanced surface plasmon resonance for ultrasensitive detection of dna hybridization, Journal of the American Chemical Society, vol. 122, no. 38, pp. 90719077, 2000. [46] K. Rijal and R. Mutharasan, Pemc-based method of measuring dna hybridization at femtomolar concentration directly in human serum and in the presence of copious noncomplementary strands, Anal Chem, vol. 79, no. 19, pp. 73927400, 2007. [47] S. Zheng, J. H. Choi, S. M. Lee, K. S. Hwang, S. K. Kim, and T. S. Kim, Analysis of dna hybridization regarding the conformation of molecular layer with piezoelectric microcantilevers, Lab on a Chip, vol. 11, no. 1, pp. 6369, 2011. [48] T. Liu, J. Tang, and L. Jiang, Sensitivity enhancement of dna sensors by nanogold surface modication, Biochemical and Biophysical Research Communications, vol. 295, no. 1, pp. 1416, 2002.

158 [49] T. Liu, J. Tang, and L. Jiang, The enhancement effect of gold nanoparticles as a surface modier on dna sensor sensitivity, Biochemical and Biophysical Research Communications, vol. 313, no. 1, pp. 37, 2004. [50] R. DAgata, R. Corradini, G. Grasso, R. Marchelli, and G. Spoto?, Ultrasensitive detection of dna by pna and nanoparticle-enhanced surface plasmon resonance imaging, ChemBioChem, vol. 9, no. 13, pp. 20672070, 2008. [51] K. Hu, D. Lan, X. Li, and S. Zhang, Electrochemical dna biosensor based on nanoporous gold electrode and multifunctional encoded dna-au bio bar codes, Anal Chem, vol. 80, no. 23, pp. 91249130, 2008. [52] G. Zheng, F. Patolsky, Y. Cui, W. U. Wang, and C. M. Lieber, Multiplexed electrical detection of cancer markers with nanowire sensor arrays, Nat Biotechnol, vol. 23, no. 10, pp. 1294301, 2005. [53] J. Wang, R. Polsky, A. Merkoci, and K. L. Turner, electroactive beads for ultrasensitive dna detection, Langmuir, vol. 19, no. 4, pp. 989991, 2003. [54] H. Chang, Y. Yuan, N. Shi, and Y. Guan, Electrochemical dna biosensor based on conducting polyaniline nanotube array, Anal Chem, vol. 79, no. 13, pp. 51115, 2007. [55] T. Kurkina, A. Vlandas, A. Ahmad, K. Kern, and K. Balasubramanian, Label-free detection of few copies of dna with carbon nanotube impedance biosensors, Angew Chem Int Ed Engl, vol. 50, no. 16, pp. 37104, 2011. [56] S. Niu, M. Zhao, R. Ren, and S. Zhang, Carbon nanotube-enhanced dna biosensor for dna hybridization detection using manganese(ii)-schiff base complex as hybridization indicator, J Inorg Biochem, vol. 103, no. 1, pp. 439, 2009. [57] M. Su, S. Li, and V. P. Dravid, Microcantilever resonance-based dna detection with nanoparticle probes, Applied Physics Letters, vol. 82, no. 20, pp. 35623564, 2003. [58] J. Wang, A. N. Kawde, and M. Musameh, Carbon-nanotube-modied glassy carbon electrodes for amplied label-free electrochemical detection of dna hybridization, Analyst, vol. 128, no. 7, pp. 912916, 2003. [59] S. Husale, H. H. J. Persson, and O. Sahin, Dna nanomechanics allows direct digital detection of complementary dna and microrna targets, Nature, vol. 462, no. 7276, pp. 1075U138, 2009.

159 [60] W. Gao, H. Dong, J. Lei, H. Ji, and H. Ju, Signal amplication of streptavidinhorseradish peroxidase functionalized carbon nanotubes for amperometric detection of attomolar dna, Chem Commun (Camb), vol. 47, no. 18, pp. 52202, 2011. [61] O. A. Loaiza, S. Campuzano, M. Pedrero, M. I. Pividori, P. Garcia, and J. M. Pingarron, Disposable magnetic dna sensors for the determination at the attomolar level of a specic enterobacteriaceae family gene, Anal Chem, vol. 80, no. 21, pp. 823945, 2008. [62] L. Soleymani, Z. Fang, S. O. Kelley, and E. H. Sargent, Integrated nanostructures for direct detection of dna at attomolar concentrations, Applied Physics Letters, vol. 95, no. 14, pp. 1437013, 2009. [63] C.-P. Chen, A. Ganguly, C.-Y. Lu, T.-Y. Chen, C.-C. Kuo, R.-S. Chen, W.-H. Tu, W. B. Fischer, K.-H. Chen, and L.-C. Chen, Ultrasensitive in situ label-free dna detection using a gan nanowire-based extended-gate eld-effect-transistor sensor, Anal Chem, vol. 83, no. 6, pp. 19381943, 2011. [64] S.-R. Hong, H.-D. Jeong, and S. Hong, Qcm dna biosensor for the diagnosis of a sh pathogenic virus vhsv, Talanta, vol. 82, no. 3, pp. 899903, 2010. [65] S. Paul, P. Vadgama, and A. Ray, Surface plasmon resonance imaging for biosensing, vol. 3, no. 3, pp. 7180, 2009. [66] L. Zanoli, R. DAgata, and G. Spoto, Functionalized gold nanoparticles for ultrasensitive dna detection, Analytical and Bioanalytical Chemistry, vol. 402, no. 5, pp. 17591771, 2012. [67] A. Star, E. Tu, J. Niemann, J.-C. P. Gabriel, C. S. Joiner, and C. Valcke, Label-free detection of dna hybridization using carbon nanotube network eld-effect transistors, Proc Natl Acad Sci U S A, vol. 103, no. 4, pp. 921926, 2006. [68] Q. Zhu, W. Y. Shih, and W. H. Shih, Mechanism of exural resonance frequency shift of a piezoelectric microcantilever sensor during humidity detection, Appl Phys Lett, vol. 92, no. 18, pp. 1835051835053, 2008. [69] J. P. McGovern, W. Y. Shih, R. Rest, M. Purohit, Y. Pandya, and W. H. Shih, Labelfree ow-enhanced specic detection of bacillus anthracis using a piezoelectric microcantilever sensor, Analyst, vol. 133, no. 5, pp. 64954, 2008.

160 [70] L. Loo, J. A. Capobianco, W. Wu, X. Gao, W. Y. Shih, W.-H. Shih, K. Pourrezaei, M. K. Robinson, and G. P. Adams, Highly sensitive detection of her2 extracellular domain in the serum of breast cancer patients by piezoelectric microcantilevers, Analytical Chemistry, vol. 83, no. 9, pp. 33923397, 2011. [71] H. Jaffe and D. Berlincourt, Piezoelectric Transducer Materials. IEEE, 1965. [72] Ieee standard on piezoelectricity, ANSI/IEEE Std 176-1987, pp. 01, 1988. [73] Ieee standard denitions of primary ferroelectric terms, ANSI/IEEE Std 180-1986, pp. 01, 1986. [74] Ire standards on piezoelectric crystals: Determination of the elastic, piezoelectric, and dielectric constants-the electromechanical coupling factor, 1958, Proceedings of the IRE, vol. 46, no. 4, pp. 764778, 1958. [75] Standards on piezoelectric crystals, 1949, Proceedings of the IRE, vol. 37, no. 12, pp. 13781395, 1949. [76] M. C. Soylu, W.-H. Shih, and W. Y. Shih, Insulation by solution 3mercaptopropyltrimethoxysilane (mps) coating: Effect of ph, water, and mps content, Industrial and Engineering Chemistry Research, vol. 52, no. 7, pp. 25902597, 2013. [77] F. Caruso, E. Rodda, D. N. Furlong, K. Niikura, and Y. Okahata, Quartz crystal microbalance study of dna immobilization and hybridization for nucleic acid sensor development, Anal Chem, vol. 69, no. 11, pp. 20439, 1997. [78] J. Ellinger, P. J. Bastian, N. Ellinger, P. Kahl, F. G. Perabo, R. Buttner, S. C. Muller, and A. Ruecker, Apoptotic dna fragments in serum of patients with muscle invasive bladder cancer: a prognostic entity, Cancer Lett, vol. 264, no. 2, pp. 27480, 2008. [79] M. T. Valenzuela, R. Galisteo, A. Zuluaga, M. Villalobos, M. I. Nunez, F. J. Oliver, and J. M. Ruiz de Almodovar, Assessing the use of p16(ink4a) promoter gene methylation in serum for detection of bladder cancer, Eur Urol, vol. 42, no. 6, pp. 6228; discussion 62830, 2002. [80] M. Utting, W. Werner, R. Dahse, J. Schubert, and K. Junker, Microsatellite analysis of free tumor dna in urine, serum, and plasma of patients: a minimally invasive method for the detection of bladder cancer, Clin Cancer Res, vol. 8, no. 1, pp. 35 40, 2002.

161 [81] H. Fiegl, S. Millinger, E. Mueller-Holzner, C. Marth, C. Ensinger, A. Berger, H. Klocker, G. Goebel, and M. Widschwendter, Circulating tumor-specic dna: a marker for monitoring efcacy of adjuvant therapy in cancer patients, Cancer Res, vol. 65, no. 4, pp. 11415, 2005. [82] E. Y. Rykova, P. P. Laktionov, T. E. Skvortsova, A. V. Starikov, N. P. Kuznetsova, and V. V. Vlassov, Extracellular dna in breast cancer: Cell-surface-bound, tumorderived extracellular dna in blood of patients with breast cancer and nonmalignant tumors, Ann N Y Acad Sci, vol. 1022, pp. 21720, 2004. [83] G. Sharma, S. Mirza, R. Parshad, A. Srivastava, S. Datta Gupta, P. Pandya, and R. Ralhan, Cpg hypomethylation of mdr1 gene in tumor and serum of invasive ductal breast carcinoma patients, Clin Biochem, vol. 43, no. 4-5, pp. 3739, 2010. [84] G. Sharma, S. Mirza, R. Parshad, A. Srivastava, S. D. Gupta, P. Pandya, and R. Ralhan, Clinical signicance of promoter hypermethylation of dna repair genes in tumor and serum dna in invasive ductal breast carcinoma patients, Life Sci, vol. 87, no. 3-4, pp. 8391, 2010. [85] T. E. Skvortsova, E. Y. Rykova, S. N. Tamkovich, O. E. Bryzgunova, A. V. Starikov, N. P. Kuznetsova, V. V. Vlassov, and P. P. Laktionov, Cell-free and cell-bound circulating dna in breast tumours: Dna quantication and analysis of tumour-related gene methylation, Br J Cancer, vol. 94, no. 10, pp. 14925, 2006. [86] H. Schwarzenbach, K. Pantel, B. Kemper, C. Beeger, F. Otterbach, R. Kimmig, and S. Kasimir-Bauer, Comparative evaluation of cell-free tumor dna in blood and disseminated tumor cells in bone marrow of patients with primary breast cancer, Breast Cancer Res, vol. 11, no. 5, p. R71, 2009. [87] J. M. Silva, J. Silva, A. Sanchez, J. M. Garcia, G. Dominguez, M. Provencio, L. Sanfrutos, E. Jareo, A. Colas, P. Espaa, and F. Bonilla, Tumor dna in plasma at diagnosis of breast cancer patients is a valuable predictor of disease-free survival, Clinical Cancer Research, vol. 8, no. 12, pp. 37613766, 2002. [88] B. Taback, A. E. Giuliano, N. M. Hansen, F. R. Singer, S. Shu, and D. S. Hoon, Detection of tumor-specic genetic alterations in bone marrow from early-stage breast cancer patients, Cancer Res, vol. 63, no. 8, pp. 18847, 2003. [89] N. Umetani, A. E. Giuliano, S. H. Hiramatsu, F. Amersi, T. Nakagawa, S. Martino, and D. S. Hoon, Prediction of breast tumor progression by integrity of free circulating dna in serum, J Clin Oncol, vol. 24, no. 26, pp. 42706, 2006.

162 [90] U. Deligezer, Y. Eralp, E. Akisik, E. Akisik, P. Saip, and E. Topuz, Effect of adjuvant chemotherapy on integrity of free serum dna in patients with breast cancer, Ann N Y Acad Sci, vol. 1137, pp. 175 9, 2008. [91] C. Kohler, R. Radpour, Z. Barekati, R. Asadollahi, J. Bitzer, E. Wight, N. Burki, C. Diesch, W. Holzgreve, and X. Zhong, Levels of plasma circulating cell free nuclear and mitochondrial dna as potential biomarkers for breast tumors, Molecular Cancer, vol. 8, no. 1, p. 105, 2009. [92] C. C. Ren, X. H. Miao, B. Yang, L. Zhao, R. Sun, and W. Q. Song, Methylation status of the fragile histidine triad and e-cadherin genes in plasma of cervical cancer patients, Int J Gynecol Cancer, vol. 16, no. 5, pp. 18627, 2006. [93] A. Widschwendter, H. M. Muller, H. Fiegl, L. Ivarsson, A. Wiedemair, E. MullerHolzner, G. Goebel, C. Marth, and M. Widschwendter, Dna methylation in serum and tumors of cervical cancer patients, Clin Cancer Res, vol. 10, no. 2, pp. 56571, 2004. [94] W. Pornthanakasem, K. Shotelersuk, W. Termrungruanglert, N. Voravud, S. Niruthisard, and A. Mutirangura, Human papillomavirus dna in plasma of patients with cervical cancer, BMC Cancer, vol. 1, p. 2, 2001. [95] T. Lecomte, A. Berger, F. Zinzindohou, S. Micard, B. Landi, H. Blons, P. Beaune, P.-H. Cugnenc, and P. Laurent-Puig, Detection of free-circulating tumor-associated dna in plasma of colorectal cancer patients and its association with prognosis, International Journal of Cancer, vol. 100, no. 5, pp. 542548, 2002. [96] B. Lefebure, F. Charbonnier, F. Di Fiore, J. J. Tuech, F. Le Pessot, F. Michot, P. Michel, and T. Frebourg, Prognostic value of circulating mutant dna in unresectable metastatic colorectal cancer, Ann Surg, vol. 251, no. 2, pp. 27580, 2010. [97] C. Trevisiol, F. Di Fabio, R. Nascimbeni, L. Peloso, C. Salbe, E. Ferruzzi, B. Salerni, and M. Gion, Prognostic value of circulating kras2 gene mutations in colorectal cancer with distant metastases, Int J Biol Markers, vol. 21, no. 4, pp. 2238, 2006. [98] J.-Y. Wang, J.-S. Hsieh, M.-Y. Chang, T.-J. Huang, F.-M. Chen, T.-L. Cheng, K. Alexandersen, Y.-S. Huang, W.-S. Tzou, and S.-R. Lin, Molecular detection of apc, k-ras, and p53 mutations in the serum of colorectal cancer patients as circulating biomarkers, World Journal of Surgery, vol. 28, no. 7, pp. 721726, 2004.

163 [99] B. M. Ryan, F. Lefort, R. McManus, J. Daly, P. W. Keeling, D. G. Weir, and D. Kelleher, A prospective study of circulating mutant kras2 in the serum of patients with colorectal neoplasia: strong prognostic indicator in postoperative follow up, Gut, vol. 52, no. 1, pp. 1018, 2003. [100] N. Umetani, J. Kim, S. Hiramatsu, H. A. Reber, O. J. Hines, A. J. Bilchik, and D. S. Hoon, Increased integrity of free circulating dna in sera of patients with colorectal or periampullary cancer: direct quantitative pcr for alu repeats, Clin Chem, vol. 52, no. 6, pp. 10629, 2006. [101] T. deVos, R. Tetzner, F. Model, G. Weiss, M. Schuster, J. Distler, K. V. Steiger, R. Grtzmann, C. Pilarsky, J. K. Habermann, P. R. Fleshner, B. M. Oubre, R. Day, A. Z. Sledziewski, and C. Lofton-Day, Circulating methylated sept9 dna in plasma is a biomarker for colorectal cancer, Clin Chem, vol. 55, no. 7, pp. 13371346, 2009. [102] R. Grutzmann, B. Molnar, C. Pilarsky, J. K. Habermann, P. M. Schlag, H. D. Saeger, S. Miehlke, T. Stolz, F. Model, U. J. Roblick, H. P. Bruch, R. Koch, V. Liebenberg, T. Devos, X. Song, R. H. Day, A. Z. Sledziewski, and C. Lofton-Day, Sensitive detection of colorectal cancer in peripheral blood by septin 9 dna methylation assay, PLoS One, vol. 3, no. 11, p. e3759, 2008. [103] Q. He, H. Y. Chen, E. Q. Bai, Y. X. Luo, R. J. Fu, Y. S. He, J. Jiang, and H. Q. Wang, Development of a multiplex methylight assay for the detection of multigene methylation in human colorectal cancer, Cancer Genet Cytogenet, vol. 202, no. 1, pp. 110, 2010. [104] M. Tanzer, B. Balluff, J. Distler, K. Hale, A. Leodolter, C. Rocken, B. Molnar, R. Schmid, C. Lofton-Day, T. Schuster, and M. P. Ebert, Performance of epigenetic markers sept9 and alx4 in plasma for detection of colorectal precancerous lesions, PLoS One, vol. 5, no. 2, p. e9061, 2010. [105] Y. H. Su, M. Wang, B. Aiamkitsumrit, D. E. Brenner, and T. M. Block, Detection of a k-ras mutation in urine of patients with colorectal cancer, Cancer Biomark, vol. 1, no. 2-3, pp. 17782, 2005. [106] Y. H. Su, M. Wang, D. E. Brenner, P. A. Norton, and T. M. Block, Detection of mutated k-ras dna in urine, plasma, and serum of patients with colorectal carcinoma or adenomatous polyps, Ann N Y Acad Sci, vol. 1137, pp. 197206, 2008.

164 [107] Y. H. Su, J. Song, Z. Wang, X. H. Wang, M. Wang, D. E. Brenner, and T. M. Block, Removal of high-molecular-weight dna by carboxylated magnetic beads enhances the detection of mutated k-ras dna in urine, Ann N Y Acad Sci, vol. 1137, pp. 8291, 2008. [108] M. Wang, T. M. Block, L. Steel, D. E. Brenner, and Y. H. Su, Preferential isolation of fragmented dna enhances the detection of circulating mutated k-ras dna, Clin Chem, vol. 50, no. 1, pp. 2113, 2004. [109] K. C. A. Chan, P. B. S. Lai, T. S. K. Mok, H. L. Y. Chan, C. Ding, S. W. Yeung, and Y. M. D. Lo, Quantitative analysis of circulating methylated dna as a biomarker for hepatocellular carcinoma, Clin Chem, vol. 54, no. 9, pp. 15281536, 2008. [110] J. Wang, Y. Qin, B. Li, Z. Sun, and B. Yang, Detection of aberrant promoter methylation of gstp1 in the tumor and serum of chinese human primary hepatocellular carcinoma patients, Clin Biochem, vol. 39, no. 4, pp. 3448, 2006. [111] I. H. Wong, Y. M. Lo, W. Yeo, W. Y. Lau, and P. J. Johnson, Frequent p15 promoter methylation in tumor and peripheral blood from hepatocellular carcinoma patients, Clin Cancer Res, vol. 6, no. 9, pp. 351621, 2000. [112] N. Ren, L. X. Qin, H. Tu, Y. K. Liu, B. H. Zhang, and Z. Y. Tang, The prognostic value of circulating plasma dna level and its allelic imbalance on chromosome 8p in patients with hepatocellular carcinoma, J Cancer Res Clin Oncol, vol. 132, no. 6, pp. 399407, 2006. [113] K. Szymanska, O. A. Lesi, G. D. Kirk, O. Sam, P. Taniere, J. Y. Scoazec, M. Mendy, M. D. Friesen, H. Whittle, R. Montesano, and P. Hainaut, Ser-249tp53 mutation in tumour and plasma dna of hepatocellular carcinoma patients from a high incidence area in the gambia, west africa, Int J Cancer, vol. 110, no. 3, pp. 3749, 2004. [114] G. D. Kirk, O. A. Lesi, M. Mendy, K. Szymanska, H. Whittle, J. J. Goedert, P. Hainaut, and R. Montesano, 249(ser) tp53 mutation in plasma dna, hepatitis b viral infection, and risk of hepatocellular carcinoma, Oncogene, vol. 24, no. 38, pp. 5858 67, 2005. [115] Y. W. Su, Y. W. Huang, S. H. Chen, and C. Y. Tzen, Quantitative analysis of plasma hbv dna for early evaluation of the response to transcatheter arterial embolization for hbv-related hepatocellular carcinoma, World J Gastroenterol, vol. 11, no. 39, pp. 61936, 2005.

165 [116] P. Tangkijvanich, N. Hourpai, P. Rattanatanyong, N. Wisedopas, V. Mahachai, and A. Mutirangura, Serum line-1 hypomethylation as a potential prognostic marker for hepatocellular carcinoma, Clinica Chimica Acta, vol. 379, no. 12, pp. 127133, 2007. [117] S. Wang, T. An, J. Wang, J. Zhao, Z. Wang, M. Zhuo, H. Bai, L. Yang, Y. Zhang, X. Wang, J. Duan, Y. Wang, Q. Guo, and M. Wu, Potential clinical signicance of a plasma-based kras mutation analysis in patients with advanced non-small cell lung cancer, Clin Cancer Res, vol. 16, no. 4, pp. 132430, 2010. [118] H. Kimura, Y. Fujiwara, T. Sone, H. Kunitoh, T. Tamura, K. Kasahara, and K. Nishio, Egfr mutation status in tumour-derived dna from pleural effusion uid is a practical basis for predicting the response to getinib, Br J Cancer, vol. 95, no. 10, pp. 13901395, 2006. [119] O. Gautschi, B. Huegli, A. Ziegler, M. Gugger, J. Heighway, D. Ratschiller, P. C. Mack, P. H. Gumerlock, H. J. Kung, R. A. Stahel, D. R. Gandara, and D. C. Betticher, Origin and prognostic value of circulating kras mutations in lung cancer patients, Cancer Lett, vol. 254, no. 2, pp. 26573, 2007. [120] G. Jian, Z. Songwen, Z. Ling, D. Qinfang, Z. Jie, T. Liang, and Z. Caicun, Prediction of epidermal growth factor receptor mutations in the plasma/pleural effusion to efcacy of getinib treatment in advanced non-small cell lung cancer, J Cancer Res Clin Oncol, vol. 136, no. 9, pp. 13417, 2010. [121] Q. An, Y. Liu, Y. Gao, J. Huang, X. Fong, L. Li, D. Zhang, and S. Cheng, Detection of p16 hypermethylation in circulating plasma dna of non-small cell lung cancer patients, Cancer Lett, vol. 188, no. 1-2, pp. 10914, 2002. [122] A. Bearzatto, D. Conte, M. Frattini, N. Zaffaroni, F. Andriani, D. Balestra, L. Tavecchio, M. G. Daidone, and G. Sozzi, p16(ink4a) hypermethylation detected by uorescent methylation-specic pcr in plasmas from non-small cell lung cancer, Clin Cancer Res, vol. 8, no. 12, pp. 37827, 2002. [123] Y. Liu, Q. An, L. Li, D. Zhang, J. Huang, X. Feng, S. Cheng, and Y. Gao, Hypermethylation of p16ink4a in chinese lung cancer patients: biological and clinical implications, Carcinogenesis, vol. 24, no. 12, pp. 1897901, 2003. [124] C. S. H. Ng, J. Zhang, S. Wan, T. W. Lee, A. A. Ari, T. Mok, D. Y. M. Lo, and A. P. C. Yim, Tumor p16m is a possible marker of advanced stage in nonsmall cell lung cancer, J Surg Oncol, vol. 79, no. 2, pp. 101106, 2002.

166 [125] J. L. Ramirez, R. Rosell, M. Taron, M. Sanchez-Ronco, V. Alberola, R. de Las Penas, J. M. Sanchez, T. Moran, C. Camps, B. Massuti, J. J. Sanchez, F. Salazar, and S. Catot, 14-3-3sigma methylation in pretreatment serum circulating dna of cisplatin-plus-gemcitabine-treated advanced non-small-cell lung cancer patients predicts survival: The spanish lung cancer group, J Clin Oncol, vol. 23, no. 36, pp. 910512, 2005. [126] N. Bruhn, T. Beinert, C. Oehm, B. Jandrig, I. Petersen, X. Q. Chen, K. Possinger, and M. Fleischhacker, Detection of microsatellite alterations in the dna isolated from tumor cells and from plasma dna of patients with lung cancer, Ann N Y Acad Sci, vol. 906, pp. 7282, 2000. [127] G. Sozzi, D. Conte, L. Mariani, S. Lo Vullo, L. Roz, C. Lombardo, M. A. Pierotti, and L. Tavecchio, Analysis of circulating tumor dna in plasma at diagnosis and during follow-up of lung cancer patients, Cancer Res, vol. 61, no. 12, pp. 4675 4678, 2001. [128] G. Hosny, N. Farahat, and P. Hainaut, Tp53 mutations in circulating free dna from egyptian patients with non-hodgkins lymphoma, Cancer Lett, vol. 275, no. 2, pp. 2349, 2009. [129] W. Y. Au, A. Pang, C. Choy, C. S. Chim, and Y. L. Kwong, Quantication of circulating epstein-barr virus (ebv) dna in the diagnosis and monitoring of natural killer cell and ebv-positive lymphomas in immunocompetent patients, Blood, vol. 104, no. 1, pp. 2439, 2004. [130] K. I. Lei, L. Y. Chan, W. Y. Chan, P. J. Johnson, and Y. M. Lo, Diagnostic and prognostic implications of circulating cell-free epstein-barr virus dna in natural killer/tcell lymphoma, Clin Cancer Res, vol. 8, no. 1, pp. 2934, 2002. [131] A. S. Machado, M. C. Da Silva Robaina, L. M. Magalhaes De Rezende, A. G. Apa, N. D. Amoedo, C. E. Bacchi, and C. E. Klumb, Circulating cell-free and epsteinbarr virus dna in pediatric b-non-hodgkin lymphomas, Leuk Lymphoma, vol. 51, no. 6, pp. 10207, 2010. [132] U. Deligezer, F. Yaman, N. Erten, and N. Dalay, Frequent copresence of methylated dna and fragmented nucleosomal dna in plasma of lymphoma patients, Clin Chim Acta, vol. 335, no. 1-2, pp. 8994, 2003. [133] R. E. Board, G. Ellison, M. C. Orr, K. R. Kemsley, G. McWalter, L. Y. Blockley, S. P. Dearden, C. Morris, M. Ranson, M. V. Cantarini, C. Dive, and A. Hughes,

167 Detection of braf mutations in the tumour and serum of patients enrolled in the azd6244 (arry-142886) advanced melanoma phase ii study, Br J Cancer, vol. 101, no. 10, pp. 172430, 2009. [134] P. Pinzani, F. Salvianti, R. Cascella, D. Massi, V. De Giorgi, M. Pazzagli, and C. Orlando, Allele specic taqman-based real-time pcr assay to quantify circulating brafv600e mutated dna in plasma of melanoma patients, Clin Chim Acta, vol. 411, no. 17-18, pp. 131924, 2010. [135] A. Fujimoto, S. J. ODay, B. Taback, D. Elashoff, and D. S. Hoon, Allelic imbalance on 12q22-23 in serum circulating dna of melanoma patients predicts disease outcome, Cancer Res, vol. 64, no. 12, pp. 40858, 2004. [136] Y. Fujiwara, D. D. Chi, H. Wang, P. Keleman, D. L. Morton, R. Turner, and D. S. Hoon, Plasma dna microsatellites as tumor-specic markers and indicators of tumor progression in melanoma patients, Cancer Res, vol. 59, no. 7, pp. 156771, 1999. [137] B. Taback, Y. Fujiwara, H. J. Wang, L. J. Foshag, D. L. Morton, and D. S. Hoon, Prognostic signicance of circulating microsatellite markers in the plasma of melanoma patients, Cancer Res, vol. 61, no. 15, pp. 57236, 2001. [138] B. Taback, S. J. ODay, P. D. Boasberg, S. Shu, P. Fournier, R. Elashoff, H. J. Wang, and D. S. Hoon, Circulating dna microsatellites: molecular determinants of response to biochemotherapy in patients with metastatic melanoma, J Natl Cancer Inst, vol. 96, no. 2, pp. 1526, 2004. [139] M. Shinozaki, S. J. ODay, M. Kitago, F. Amersi, C. Kuo, J. Kim, H. J. Wang, and D. S. Hoon, Utility of circulating b-raf dna mutation in serum for monitoring melanoma patients receiving biochemotherapy, Clin Cancer Res, vol. 13, no. 7, pp. 206874, 2007. [140] K. Koyanagi, T. Mori, S. J. ODay, S. R. Martinez, H. J. Wang, and D. S. Hoon, Association of circulating tumor cells with serum tumor-related methylated dna in peripheral blood of melanoma patients, Cancer Res, vol. 66, no. 12, pp. 61117, 2006. [141] T. Mori, S. J. ODay, N. Umetani, S. R. Martinez, M. Kitago, K. Koyanagi, C. Kuo, T. L. Takeshima, R. Milford, H. J. Wang, V. D. Vu, S. L. Nguyen, and D. S. Hoon, Predictive utility of circulating methylated dna in serum of melanoma patients receiving biochemotherapy, J Clin Oncol, vol. 23, no. 36, pp. 93518, 2005.

168 [142] A. Melnikov, D. Scholtens, A. Godwin, and V. Levenson, Differential methylation prole of ovarian cancer in tissues and plasma, J Mol Diagn, vol. 11, no. 1, pp. 60 5, 2009. [143] H. M. Muller, S. Millinger, H. Fiegl, G. Goebel, L. Ivarsson, A. Widschwendter, E. Muller-Holzner, C. Marth, and M. Widschwendter, Analysis of methylated genes in peritoneal uids of ovarian cancer patients: a new prognostic tool, Clin Chem, vol. 50, no. 11, pp. 21713, 2004. [144] E. M. Swisher, M. Wollan, S. M. Mahtani, J. B. Willner, R. Garcia, B. A. Goff, and M. C. King, Tumor-specic p53 sequences in blood and peritoneal uid of women with epithelial ovarian cancer, Am J Obstet Gynecol, vol. 193, no. 3 Pt 1, pp. 6627, 2005. [145] R. Zachariah, S. Schmid, N. Buerki, R. Radpour, W. Holzgreve, and X. Zhong, Levels of circulating cell-free nuclear and mitochondrial dna in benign and malignant ovarian tumors, Obstet Gynecol, vol. 112, no. 4, pp. 843 50, 2008. [146] R. Salani, B. Davidson, M. Fiegl, C. Marth, E. Muller-Holzner, G. Gastl, H. Y. Huang, J. C. Hsiao, H. S. Lin, T. L. Wang, B. L. Lin, and M. Shih Ie, Measurement of cyclin e genomic copy number and strand length in cell-free dna distinguish malignant versus benign effusions, Clin Cancer Res, vol. 13, no. 19, pp. 58059, 2007. [147] T. Liggett, A. Melnikov, Q. L. Yi, C. Replogle, R. Brand, K. Kaul, M. Talamonti, R. A. Abrams, and V. Levenson, Differential methylation of cell-free circulating dna among patients with pancreatic cancer versus chronic pancreatitis, Cancer, vol. 116, no. 7, pp. 167480, 2010. [148] A. A. Melnikov, D. Scholtens, M. S. Talamonti, D. J. Bentrem, and V. V. Levenson, Methylation prole of circulating plasma dna in patients with pancreatic cancer, J Surg Oncol, vol. 99, no. 2, pp. 11922, 2009. [149] A. Castells, P. Puig, J. Mora, J. Boadas, L. Boix, E. Urgell, M. Sole, G. Capella, F. Lluis, L. Fernandez-Cruz, S. Navarro, and A. Farre, K-ras mutations in dna extracted from the plasma of patients with pancreatic carcinoma: diagnostic utility and prognostic signicance, J Clin Oncol, vol. 17, no. 2, pp. 57884, 1999. [150] P. J. Bastian, G. S. Palapattu, X. Lin, S. Yegnasubramanian, L. A. Mangold, B. Trock, M. A. Eisenberger, A. W. Partin, and W. G. Nelson, Preoperative serum

169 dna gstp1 cpg island hypermethylation and the risk of early prostate-specic antigen recurrence following radical prostatectomy, Clin Cancer Res, vol. 11, no. 11, pp. 403743, 2005. [151] O. E. Bryzgunova, E. S. Morozkin, S. V. Yarmoschuk, V. V. Vlassov, and P. P. Laktionov, Methylation-specic sequencing of gstp1 gene promoter in circulating/extracellular dna from blood and urine of healthy donors and prostate cancer patients, Ann N Y Acad Sci, vol. 1137, pp. 2225, 2008. [152] C. Goessl, M. Muller, R. Heicappell, H. Krause, and K. Miller, Dna-based detection of prostate cancer in blood, urine, and ejaculates, Ann N Y Acad Sci, vol. 945, pp. 518, 2001. [153] C. Jeronimo, H. Usadel, R. Henrique, C. Silva, J. Oliveira, C. Lopes, and D. Sidransky, Quantitative gstp1 hypermethylation in bodily uids of patients with prostate cancer, Urology, vol. 60, no. 6, pp. 11315, 2002. [154] M. Roupret, V. Hupertan, J. W. Catto, D. R. Yates, I. Rehman, L. M. Proctor, J. Phillips, M. Meuth, O. Cussenot, and F. C. Hamdy, Promoter hypermethylation in circulating blood cells identies prostate cancer progression, Int J Cancer, vol. 122, no. 4, pp. 9526, 2008. [155] J. Ellinger, P. J. Bastian, K. I. Haan, L. C. Heukamp, R. Buettner, R. Fimmers, S. C. Mueller, and A. von Ruecker, Noncancerous ptgs2 dna fragments of apoptotic origin in sera of prostate cancer patients qualify as diagnostic and prognostic indicators, International Journal of Cancer, vol. 122, no. 1, pp. 138143, 2008. [156] J. Ellinger, S. C. Muller, N. Wernert, A. von Ruecker, and P. J. Bastian, Mitochondrial dna in serum of patients with prostate cancer: a predictor of biochemical recurrence after prostatectomy, BJU Int, vol. 102, no. 5, pp. 62832, 2008. [157] H. Schwarzenbach, C. Alix-Panabires, I. Mller, N. Letang, J.-P. Vendrell, X. Rebillard, and K. Pantel, Cell-free tumor dna in blood plasma as a marker for circulating tumor cells in prostate cancer, Clinical Cancer Research, vol. 15, no. 3, pp. 1032 1038, 2009. [158] E. Sunami, M. Shinozaki, C. S. Higano, R. Wollman, T. B. Dorff, S. J. Tucker, S. R. Martinez, R. Mizuno, F. R. Singer, and D. S. Hoon, Multimarker circulating dna assay for assessing blood of prostate cancer patients, Clin Chem, vol. 55, no. 3, pp. 55967, 2009.

170 [159] N. Mehra, M. Penning, J. Maas, N. van Daal, R. H. Giles, and E. E. Voest, Circulating mitochondrial nucleic acids have prognostic value for survival in patients with advanced prostate cancer, Clin Cancer Res, vol. 13, no. 2 Pt 1, pp. 4216, 2007. [160] J. Ellinger, P. Albers, S. C. Muller, A. von Ruecker, and P. J. Bastian, Circulating mitochondrial dna in the serum of patients with testicular germ cell cancer as a novel noninvasive diagnostic biomarker, BJU Int, vol. 104, no. 1, pp. 4852, 2009. [161] J. Ellinger, K. Haan, L. C. Heukamp, P. Kahl, R. Buttner, S. C. Muller, A. von Ruecker, and P. J. Bastian, Cpg island hypermethylation in cell-free serum dna identies patients with localized prostate cancer, Prostate, vol. 68, no. 1, pp. 429, 2008. [162] H. Nawroz, W. Koch, P. Anker, M. Stroun, and D. Sidransky, Microsatellite alterations in serum dna of head and neck cancer patients, Nat Med, vol. 2, no. 9, pp. 10357, 1996. [163] F. Coulet, H. Blons, A. Cabelguenne, T. Lecomte, O. Lacourreye, D. Brasnu, P. Beaune, J. Zucman, and P. Laurent-Puig, Detection of plasma tumor dna in head and neck squamous cell carcinoma by microsatellite typing and p53 mutation analysis, Cancer Res, vol. 60, no. 3, pp. 70711, 2000. [164] H. Nawroz-Danish, C. F. Eisenberger, G. H. Yoo, L. Wu, W. Koch, C. Black, J. F. Ensley, W. Z. Wei, and D. Sidransky, Microsatellite analysis of serum dna in patients with head and neck cancer, Int J Cancer, vol. 111, no. 1, pp. 96100, 2004. [165] X. Mao, L. Yang, X. L. Su, and Y. Li, A nanoparticle amplication based quartz crystal microbalance dna sensor for detection of escherichia coli o157:h7, Biosens Bioelectron, vol. 21, no. 7, pp. 117885, 2006. [166] L. K. Gifford, I. E. Sendroiu, R. M. Corn, and A. Luptak, Attomole detection of mesophilic dna polymerase products by nanoparticle-enhanced surface plasmon resonance imaging on glassied gold surfaces, J Am Chem Soc, vol. 132, no. 27, pp. 92657, 2010. [167] T. Yang, N. Zhou, Y. Zhang, W. Zhang, K. Jiao, and G. Li, Synergistically improved sensitivity for the detection of specic dna sequences using polyaniline nanobers and multi-walled carbon nanotubes composites, Biosens Bioelectron, vol. 24, no. 7, pp. 216570, 2009.

171 [168] G. J. Zhang, Z. H. Luo, M. J. Huang, G. K. Tay, and E. J. Lim, Morpholinofunctionalized silicon nanowire biosensor for sequence-specic label-free detection of dna, Biosens Bioelectron, vol. 25, no. 11, pp. 244753, 2010. [169] A. Andreu, J. W. Merkert, L. A. Lecaros, B. L. Broglin, J. T. Brazell, and M. ElKouedi, Detection of dna oligonucleotides on nanowire array electrodes using chronocoulometry, Sensors and Actuators B: Chemical, vol. 114, no. 2, pp. 1116 1120, 2006. [170] Z. Gao, A. Agarwal, A. D. Trigg, N. Singh, C. Fang, C. H. Tung, Y. Fan, K. D. Buddharaju, and J. Kong, Silicon nanowire arrays for label-free detection of dna, Anal Chem, vol. 79, no. 9, pp. 32917, 2007. [171] J.-i. Hahm and C. M. Lieber, Direct ultrasensitive electrical detection of dna and dna sequence variations using nanowire nanosensors, Nano Letters, vol. 4, no. 1, pp. 5154, 2003. [172] W. Y. Shih, H. Luo, H. Li, C. Martorano, and W.-H. Shih, Sheet geometry enhanced giant piezoelectric coefcients, Applied Physics Letters, vol. 89, no. 24, pp. 2429133, 2006. [173] W. Wu, C. E. Kirimli, W. H. Shih, and W. Y. Shih, Real-time, in situ dna hybridization detection with attomolar sensitivity without amplication using (pb(mg1/3nb2/3)o3)0.65-(pbtio3)0.35 piezoelectric plate sensors, Biosens Bioelectron, vol. 43, pp. 3919, 2013. [174] R. Fodde and M. Losekoot, Mutation detection by denaturing gradient gel electrophoresis (dgge), Hum Mutat, vol. 3, no. 2, pp. 8394, 1994. [175] R. B. Scholz, K. Milde-Langosch, R. Jung, H. Schlechte, H. Kabisch, C. Wagener, and T. Loning, Rapid screening for tp53 mutations by temperature gradient gel electrophoresis: a comparison with sscp analysis, Hum Mol Genet, vol. 2, no. 12, pp. 21558, 1993. [176] M. Orita, H. Iwahana, H. Kanazawa, K. Hayashi, and T. Sekiya, Detection of polymorphisms of human dna by gel electrophoresis as single-strand conformation polymorphisms, Proc Natl Acad Sci U S A, vol. 86, no. 8, pp. 276670, 1989. [177] P. V. Danenberg, T. Horikoshi, M. Volkenandt, K. Danenberg, H. J. Lenz, L. C. Shea, A. P. Dicker, A. Simoneau, P. A. Jones, and J. R. Bertino, Detection of point

172 mutations in human dna by analysis of rna conformation polymorphism(s), Nucleic Acids Res, vol. 20, no. 3, pp. 5739, 1992. [178] G. Sarkar, H. S. Yoon, and S. S. Sommer, Screening for mutations by rna singlestrand conformation polymorphism (rsscp): comparison with dna-sscp, Nucleic Acids Res, vol. 20, no. 4, pp. 8718, 1992. [179] D. Soto and S. Sukumar, Improved detection of mutations in the p53 gene in human tumors as single-stranded conformation polymorphs and double-stranded heteroduplex dna, PCR Methods Appl, vol. 2, no. 1, pp. 968, 1992. [180] R. G. Cotton, N. R. Rodrigues, and R. D. Campbell, Reactivity of cytosine and thymine in single-base-pair mismatches with hydroxylamine and osmium tetroxide and its application to the study of mutations, Proc Natl Acad Sci U S A, vol. 85, no. 12, pp. 4397401, 1988. [181] J. A. Saleeba and R. G. Cotton, 35s-labelled probes improve detection of mismatched base pairs by chemical cleavage, Nucleic Acids Res, vol. 19, no. 7, p. 1712, 1991. [182] P. A. Roest, R. G. Roberts, S. Sugino, G. J. van Ommen, and J. T. den Dunnen, Protein truncation test (ptt) for rapid detection of translation-terminating mutations, Hum Mol Genet, vol. 2, no. 10, pp. 171921, 1993. [183] R. van der Luijt, P. M. Khan, H. Vasen, C. van Leeuwen, C. Tops, P. Roest, J. den Dunnen, and R. Fodde, Rapid detection of translation-terminating mutations at the adenomatous polyposis coli (apc) gene by direct protein truncation test, Genomics, vol. 20, no. 1, pp. 14, 1994. [184] R. J. Lipshutz, D. Morris, M. Chee, E. Hubbell, M. J. Kozal, N. Shah, N. Shen, R. Yang, and S. P. Fodor, Using oligonucleotide probe arrays to access genetic diversity, Biotechniques, vol. 19, no. 3, pp. 4427, 1995. [185] E. M. Southern, Dna chips: analysing sequence by hybridization to oligonucleotides on a large scale, Trends Genet, vol. 12, no. 3, pp. 1105, 1996. [186] G. H. Reed and C. T. Wittwer, Sensitivity and specicity of single-nucleotide polymorphism scanning by high-resolution melting analysis, Clin Chem, vol. 50, no. 10, pp. 174854, 2004.

173 [187] K. M. Ririe, R. P. Rasmussen, and C. T. Wittwer, Product differentiation by analysis of dna melting curves during the polymerase chain reaction, Anal Biochem, vol. 245, no. 2, pp. 15460, 1997. [188] E. Lyon, Mutation detection using uorescent hybridization probes and melting curve analysis, Expert Rev Mol Diagn, vol. 1, no. 1, pp. 92101, 2001. [189] R. H. Lipsky, C. M. Mazzanti, J. G. Rudolph, K. Xu, G. Vyas, D. Bozak, M. Q. Radel, and D. Goldman, Dna melting analysis for detection of single nucleotide polymorphisms, Clin Chem, vol. 47, no. 4, pp. 63544, 2001. [190] P. H. Rogers, E. Michel, C. A. Bauer, S. Vanderet, D. Hansen, B. K. Roberts, A. Calvez, J. B. Crews, K. O. Lau, A. Wood, D. J. Pine, and P. V. Schwartz, Selective, controllable, and reversible aggregation of polystyrene latex microspheres via dna hybridization, Langmuir, vol. 21, no. 12, pp. 55625569, 2005. [191] Y. Zhang, V. T. Milam, D. J. Graves, and D. A. Hammer, Differential adhesion of microspheres mediated by dna hybridization i: experiment, Biophys J, vol. 90, no. 11, pp. 412836, 2006. [192] S. P. Mulvaney, C. L. Cole, M. D. Kniller, M. Malito, C. R. Tamanaha, J. C. Rife, M. W. Stanton, and L. J. Whitman, Rapid, femtomolar bioassays in complex matrices combining microuidics and magnetoelectronics, Biosens Bioelectron, vol. 23, no. 2, pp. 191200, 2007. [193] S. P. Mulvaney, C. N. Ibe, C. R. Tamanaha, and L. J. Whitman, Direct detection of genomic dna with uidic force discrimination assays, Anal Biochem, vol. 392, no. 2, pp. 13944, 2009. [194] S. P. Mulvaney, K. M. Myers, P. E. Sheehan, and L. J. Whitman, Attomolar protein detection in complex sample matrices with semi-homogeneous uidic force discrimination assays, Biosens Bioelectron, vol. 24, no. 5, pp. 110915, 2009. [195] C. C. Hsia, H. Yuwen, and E. Tabor, Hot-spot mutations in hepatitis b virus x gene in hepatocellular carcinoma, The Lancet, vol. 348, no. 9027, pp. 625626, 1996. [196] M. Baptista, A. Kramvis, and M. C. Kew, High prevalence of 1762(t) 1764(a) mutations in the basic core promoter of hepatitis b virus isolated from black africans with hepatocellular carcinoma compared with asymptomatic carriers, Hepatology, vol. 29, no. 3, pp. 94653, 1999.

174 [197] P. Arbuthnot and M. Kew, Hepatitis b virus and hepatocellular carcinoma, International Journal of Experimental Pathology, vol. 82, no. 2, pp. 77100, 2001. [198] S. Y. Kuang, P. E. Jackson, J. B. Wang, P. X. Lu, A. Munoz, G. S. Qian, T. W. Kensler, and J. D. Groopman, Specic mutations of hepatitis b virus in plasma predict liver cancer development, Proc Natl Acad Sci U S A, vol. 101, no. 10, pp. 3575 80, 2004. [199] S. Y. Kuang, S. Lekawanvijit, N. Maneekarn, S. Thongsawat, K. Brodovicz, K. Nelson, and J. D. Groopman, Hepatitis b 1762t/1764a mutations, hepatitis c infection, and codon 249 p53 mutations in hepatocellular carcinomas from thailand, Cancer Epidemiol Biomarkers Prev, vol. 14, no. 2, pp. 3804, 2005. [200] J. SantaLucia, J., A unied view of polymer, dumbbell, and oligonucleotide dna nearest-neighbor thermodynamics, Proc Natl Acad Sci U S A, vol. 95, no. 4, pp. 14605, 1998. [201] W. A. Kibbe, Oligocalc: an online oligonucleotide properties calculator, Nucleic Acids Res, vol. 35, no. Web Server issue, pp. W436, 2007. [202] N. Otsu, A threshold selection method from gray-level histograms, IEEE Transactions on Systems, Man and Cybernetics, vol. 9, no. 1, pp. 6266, 1979. [203] J. Liu, N. Agrawal, A. Calderon, P. Ayyaswamy, D. Eckmann, and R. Radhakrishnan, Multivalent binding of nanocarrier to endothelial cells under shear ow, Biophys J, vol. 101, no. 2, pp. 319326, 2011. [204] D. R. W. H. B. Mann, On a test of whether one of two random variables is stochastically larger than the other, Annals of Mathematical Statistics, vol. 18, no. 1, p. 11, 1947. [205] A. J. Goldman, R. G. Cox, and H. Brenner, Slow viscous motion of a sphere parallel to a plane wallii couette ow, Chemical Engineering Science, vol. 22, no. 4, pp. 653660, 1967. [206] T. Strunz, K. Oroszlan, R. Schafer, and H. J. Guntherodt, Dynamic force spectroscopy of single dna molecules, Proc Natl Acad Sci U S A, vol. 96, no. 20, pp. 1127782, 1999. [207] G. U. Lee, L. A. Chrisey, and R. J. Colton, Direct measurement of the forces between complementary strands of dna, Science, vol. 266, no. 5186, pp. 7713, 1994.

175 [208] A. Munoz, J. G. Chen, P. A. Egner, M. L. Marshall, J. L. Johnson, M. F. Schneider, J. H. Lu, Y. R. Zhu, J. B. Wang, T. Y. Chen, T. W. Kensler, and J. D. Groopman, Predictive power of hepatitis b 1762t/1764a mutations in plasma for hepatocellular carcinoma risk in qidong, china, Carcinogenesis, vol. 32, no. 6, pp. 8605, 2011. [209] V. Bazan, V. Agnese, S. Corsale, V. Calo, M. R. Valerio, M. A. Latteri, S. Vieni, N. Grassi, G. Cicero, G. Dardanoni, R. M. Tomasino, G. Colucci, N. Gebbia, and A. Russo, Specic tp53 and/or ki-ras mutations as independent predictors of clinical outcome in sporadic colorectal adenocarcinomas: results of a 5-year gruppo oncologico dellitalia meridionale (goim) prospective study, Ann Oncol, vol. 16 Suppl 4, pp. iv5055, 2005. [210] Y. Imamura, T. Morikawa, X. Liao, P. Lochhead, A. Kuchiba, M. Yamauchi, Z. R. Qian, R. Nishihara, J. A. Meyerhardt, K. M. Haigis, C. S. Fuchs, and S. Ogino, Specic mutations in kras codons 12 and 13, and patient prognosis in 1075 braf wild-type colorectal cancers, Clin Cancer Res, vol. 18, no. 17, pp. 475363, 2012. [211] L. A. Ugozzoli, D. Latorra, R. Puckett, K. Arar, and K. Hamby, Real-time genotyping with oligonucleotide probes containing locked nucleic acids, Anal Biochem, vol. 324, no. 1, pp. 14352, 2004. [212] M. P. Johnson, L. M. Haupt, and L. R. Grifths, Locked nucleic acid (lna) single nucleotide polymorphism (snp) genotype analysis and validation using real-time pcr, Nucleic Acids Res, vol. 32, no. 6, p. e55, 2004. [213] L. S. Chou, C. Meadows, C. T. Wittwer, and E. Lyon, Unlabeled oligonucleotide probes modied with locked nucleic acids for improved mismatch discrimination in genotyping by melting analysis, Biotechniques, vol. 39, no. 5, pp. 644, 646, 648 passim, 2005. [214] A. A. Koshkin, S. K. Singh, P. Nielsen, V. K. Rajwanshi, R. Kumar, M. Meldgaard, C. E. Olsen, and J. Wengel, Lna (locked nucleic acids): Synthesis of the adenine, cytosine, guanine, 5-methylcytosine, thymine and uracil bicyclonucleoside monomers, oligomerisation, and unprecedented nucleic acid recognition, Tetrahedron, vol. 54, no. 14, pp. 36073630, 1998. [215] N. Tolstrup, P. S. Nielsen, J. G. Kolberg, A. M. Frankel, H. Vissing, and S. Kauppinen, Oligodesign: Optimal design of lna (locked nucleic acid) oligonucleotide capture probes for gene expression proling, Nucleic Acids Res, vol. 31, no. 13, pp. 375862, 2003.

176 [216] E. Kierzek, A. Ciesielska, K. Pasternak, D. H. Mathews, D. H. Turner, and R. Kierzek, The inuence of locked nucleic acid residues on the thermodynamic properties of 2-o-methyl rna/rna heteroduplexes, Nucleic Acids Res, vol. 33, no. 16, pp. 508293, 2005. [217] H. Orum, M. H. Jakobsen, T. Koch, J. Vuust, and M. B. Borre, Detection of the factor v leiden mutation by direct allele-specic hybridization of pcr amplicons to photoimmobilized locked nucleic acids, Clin Chem, vol. 45, no. 11, pp. 1898905, 1999. [218] A. Simeonov and T. T. Nikiforov, Single nucleotide polymorphism genotyping using short, uorescently labeled locked nucleic acid (lna) probes and uorescence polarization detection, Nucleic Acids Res, vol. 30, no. 17, p. e91, 2002. [219] M. Petersen, K. Bondensgaard, J. Wengel, and J. P. Jacobsen, Locked nucleic acid (lna) recognition of rna: Nmr solution structures of lna:rna hybrids, J Am Chem Soc, vol. 124, no. 21, pp. 597482, 2002. [220] Y. Zhou, E. Kierzek, Z. P. Loo, M. Antonio, Y. H. Yau, Y. W. Chuah, S. GeifmanShochat, R. Kierzek, and G. Chen, Recognition of rna duplexes by chemically modied triplex-forming oligonucleotides, Nucleic Acids Res, 2013. [221] P. Rajagopal and J. Feigon, Triple-strand formation in the homopurine:homopyrimidine dna oligonucleotides d(g-a)4 and d(t-c)4, Nature, vol. 339, no. 6226, pp. 63740, 1989. [222] G. Bif, D. Tannahill, J. McCafferty, and S. Balasubramanian, Quantitative visualization of dna g-quadruplex structures in human cells, Nat Chem, vol. 5, no. 3, pp. 1826, 2013. [223] P. C. Turner, Molecular biology. The instant notes series, Oxford New York: BIOS ; Springer, 2nd ed., 2000.