Continuous Separation of DNA Molecules by Size Using Insulator

Yokokawa , R.; Manta , Y.; Namura , M.; Takizawa , Y.; Le , N. C. H.; Sugiyama , S. Sens. Actuators, B 2010, 143 (2) 769– 775 DOI: 10.1016/j.snb.200...
0 downloads 0 Views 3MB Size
Subscriber access provided by Fudan University

Article

Continuous separation of DNA molecules by size using insulator-based dielectrophoresis Paul Vernon Jones, Gabriel L. Salmon, and Alexandra Ros Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b03369 • Publication Date (Web): 12 Dec 2016 Downloaded from http://pubs.acs.org on December 13, 2016

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

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

Page 1 of 29

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

Analytical Chemistry

Continuous separation of DNA molecules by size using insulatorbased dielectrophoresis Paul V. Jones, Gabriel L. Salmon, Alexandra Ros

School of Molecular Sciences, Arizona State University, Tempe AZ 85287 Center for Applied Structural Discovery, The Biodesign Institute, Tempe AZ 85281

Corresponding Author: Dr. Alexandra Ros Arizona State University School of Molecular Sciences P.O. Box 871604 Tempe, Arizona 85287-1604 [email protected] Phone: 480-965-5323, Fax: 480-965-7954

Abbreviations: dielectrophoresis (DEP), electrode-based dielectrophoresis (eDEP), insulator-based dielectrophoresis (iDEP), dielectrophoretic mobility (ߤୈ୉୔ ), region of interest (ROI), lambda phage DNA (λ-DNA),

Keywords: dielectrophoresis, sorting, fractionation, continuous-flow, DNA, microfluidics, bioanalysis

Total Words (excluding title page): ~ 6780

ACS Paragon Plus Environment

Analytical Chemistry

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

Page 2 of 29

Page 3 of 29

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

Analytical Chemistry

Abstract: Separation of nucleic acids has long served as a central goal of analytical research. Innovations in this field may soon enable the development of rapid, on-site sequencing devices that significantly improve both the availability and accuracy of detailed bioinformatics. However, achieving efficient continuous-flow operation and size-based fractionation of DNA still presents considerable challenges. Current methods have not yet satisfied the need for rapid fractionation of size-heterogeneous nucleic acid samples into specific and narrow size distributions. Dielectrophoretic (DEP) mechanisms integrated in microfluidic devices offer unique advantages for such applications, including short processing times, microscale reaction volumes, and the potential for low cost and portability. To facilitate such developments, we have adapted a microfluidic constriction sorter device to separate a wide range of nucleic acid analytes into distinct microchannel outlets. This work demonstrates selective and tunable deflection of DNA using alternating current (AC) insulator-based dielectrophoresis. We report conditions for size-based DEP sorting of 1.0, 10.2, 19.5, and 48.5 kbp dsDNA analytes, including both plasmid and genomic DNA. Applied potentials range from 200 to 2400 Vpp with frequencies ranging from 50 Hz to 20 kHz. These conditions result in sorting efficiencies up to 92% with a strong dependence on applied potentials and frequencies. In low-frequency AC fields, long DNA molecules form macro-ion clusters. This behavior is associated with an apparent shift from positive to negative DEP sorting behavior. Using a finite element model, we characterize the dynamics of sorting in the microdevice and link differential sorting to differences in dielectrophoretic mobility. We propose the use of a continuous-flow sorting strategy to facilitate future coupling to next generation sequencing approaches.

ACS Paragon Plus Environment

Analytical Chemistry

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

Introduction Size-dependent sorting of DNA is foundational for many analytical methods. As an example, early sequencing technologies pioneered by Frederick Sanger constructed gene sequences from the order of DNA fragments sorted using polyacrylamide gel electrophoresis.1 In the past decade, emerging next generation sequencing (NGS) methods have significantly increased sequencing parallelization and speed.2 However, both available and emerging NGS strategies can operate only within certain length ranges for target nucleic acids. Prior to analysis, isolated DNA fragments must be partitioned into size ranges compatible with a given technique. Continuous genomic maps are obtained through bioinformatic analysis of many fragment sequences in various size ranges. For all of these sequencing methods, length-based manipulation of DNA is an essential technical prerequisite.3 High-efficiency fractionation of DNA molecules by length is not only relevant for NGS. Other emerging techniques include the use of DNA for information storage4 and transmission5, as well as logical operations6 and computing.7 New challenges also accompany the recent, increasing interest in studying the size and conformation of especially short (e.g. microRNA8) or long (e.g. over 50 kbp DNA9) nucleic acid sequences. Realization of the wide-ranging potential for these technologies will be facilitated by creating robust and high throughput solutions for size-dependent manipulation of DNA. Existing protocols for NGS commonly fragment and fractionate DNA populations by size using methods such as Covaris ultrasonication followed by bead-based recovery or gel electrophoresis.10 These purification procedures can limit sample recovery and introduce greater complexity, time requirements, and cost. Simultaneous separation of both long- and short-chain molecules is not reliably or efficiently achieved using traditional gel or capillary electrophoresis.11 In general, the electrophoretic mobility of DNA in gels is nonlinear with size for moderate-length sequences and becomes independent of size for long sequences (above 10-20 kbp).12 These features limit the dynamic range and resolution of DNA separations achievable in a gel matrix. Pulsed field electrophoresis allows separation of

ACS Paragon Plus Environment

Page 4 of 29

Page 5 of 29

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

Analytical Chemistry

larger DNA molecules than achievable via conventional electrophoresis, but pulsed field separations demand significantly longer run times (on the order of days).13 A growing number of NGS and synthetic biology applications require size separation of small quantities (less than 100 pg) and volumes (less than 100 μL) of DNA sample. These stipulations have prompted development of a variety of microfluidic sorting and sizing schemes for DNA.14 Microfluidic approaches scale well for such needs, and offer the potential for direct integration with analytic and diagnostic platforms. Early microfluidic designs sorted single fluorescently-labeled DNA molecules using flow cytometry monitoring in nanofluidic channels.15-17 These methods involved a cumbersome level of apparatus and processing. A host of other microfluidic designs have been engineered to separate DNA based on size including post arrays,18 DNA "prisms,”19 entropic traps,20,21 nanofilters,22,23 and Brownian ratchets.24 To achieve improved separation performance compared to gel electrophoresis, these devices exert fine control through the physics of analyte-matrix interactions. In recent years, dielectrophoresis (DEP) has emerged as a powerful tool for microscale particle transport.25-27 The burgeoning interest in DEP is owed at least in part to its broad applicability for a wide variety of particles and biomolecules. It has been used to manipulate or capture targets ranging from relatively large cells28-34 to nanoscale bioanalytes such as viruses,35 proteins,36 and nucleic acids.37 In its earliest implementations, DEP was induced using closely-spaced, shaped electrodes (eDEP) to create large electric field inhomogeneities. More recently, insulator-based dielectrophoresis (iDEP) has emerged to offer considerable advantages over electrode-based designs.38,39 In these implementations of DEP, insulating features placed as obstacles or constrictions in a current-conducting medium create non-uniform fields upon application of potentials at a certain distance from the insulating features. Advantages of iDEP include simple fabrication using well-established and low-cost soft lithography, as well as compatibility with higher operating potentials at lower frequencies than those tolerated in electrode-based designs. Dielectrophoresis may occur either in conjunction with or independent of other electrokinetic phenomena such as electrophoresis and electroosmotic flow.40

ACS Paragon Plus Environment

Analytical Chemistry

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

Page 6 of 29

Dielectrophoretic forces (ࡲୈ୉୔) are induced on polarizable particles by electric field inhomogeneities. The magnitude and direction of ࡲୈ୉୔ experienced by a suspended object within a given electric field are correlated with its electrophysical characteristics. For a polyelectrolyte like DNA, a permanent or transient electric field acts upon interfacial charge and the surrounding counterion cloud to produce an effective dipole moment. This action and its effects can be described in terms of the polarizability (ߙ). The arising DEP force scales with ߙ as well as the gradient of the local electric field vector (ࡱ).



ࡲୈ୉୔ = − ଶ ߙ∇ࡱଶ

(Eq. 1)

For a given analyte, ࡲୈ୉୔ may have positive or negative sign depending upon the specific particle polarization in a given medium and the applied frequency.41 In the case of positive DEP (pDEP), particles experience a force directed towards regions of higher field strength; in the case of negative DEP (nDEP), the force directed towards regions of lower field strength. Investigations of dielectrophoretic mobility (ߤୈ୉୔ ) for large dsDNA molecules have typically reported pDEP for frequencies ranging from 1 kHz to 15 MHz,42-44 with a few reports indicating crossover from pDEP to nDEP at high frequencies, including 75 kHz45 and 1 MHz46. Dielectrophoretic control of DNA has been achieved using both eDEP47,48 and iDEP39,49 paradigms. In 1991, Ajdari et al. suggested using transverse dielectrophoretic forces to supplement traditional electrophoretic separation of DNA.50 Fifteen years later Regtmeier et al. developed a functioning version of this concept, using a grid of insulating posts to create local electric field gradients and separate DNA in a size-dependent manner.51 While this approach achieved high separation efficiency, achieving higher throughput and continuous-flow operation would broaden the practical applications for DEP-based DNA fractionation. Recently, a limited number of continuous-flow implementations of iDEP-based DNA fractionation have emerged, using 90° corners in a channel52 or nanoslits.53,54 These examples represent cutting-edge uses of DEP for DNA sorting, but are still limited by low throughput due to nanoscale channel dimensions. Furthermore, these examples do not

ACS Paragon Plus Environment

Page 7 of 29

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

Analytical Chemistry

demonstrate physical fractionation by diverting separated analyte into different branches or channels of a microdevice. Previously, we developed a microfluidic sorting device to demonstrate size-based fractionation of protein crystals.55,56 The unique geometry of this microdevice created localized electric field non-uniformities near the corners of an axisymmetric constriction. Resulting DEP forces deflected crystals streaming through the constriction based on their size.57 Here, a similar constriction sorter design (Figure 1) has been adapted to demonstrate length-based sorting of DNA. This work builds upon crystal-sorting technology by using pressure-driven flow and AC potentials to achieve both pDEP and nDEP sorting. The purpose of this work is to demonstrate the general feasibility of size-dependent sorting of DNA in a constriction sorter microdevice, with a specific focus on next generation sequencing (NGS) applications. For such applications, a size cut-off is a primary technical prerequisite. We therefore show continuous-flow DEP sorting of four double-stranded DNA (dsDNA) analytes ranging from 1.0 to 48.5 kbp. We have also demonstrated feasibility of dual-analyte sorting experiments. These experiments revealed that the low-frequency electrohydrodynamic instability of DNA dispersions correlates with significant differences in sorting behavior. We therefore surmise that reversible field-induced agglomeration of DNA significantly affects net DEP mobility. Overall, these findings offer the potential to eliminate a significant bottleneck in current approaches to nucleic acid sample preparation and analysis, as well as facilitate future advances in microfluidic separations of DNA, RNA, and other nanoscale bioanalytes.

Experimental Section Microdevice fabrication Microfluidic constriction sorter devices were fabricated using standard softlithographic procedures, as previously reported.58 Briefly, channel structures were designed using AutoCAD software (Autodesk, USA). These designs were used to produce a chrome photomask (obtained from Photosciences, USA), which was then used to pattern structures

ACS Paragon Plus Environment

Analytical Chemistry

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

onto 4-inch silicon wafers with SU-8 negative photoresist (Microchem, USA). Resulting template wafers were used for elastomer molding. Sylgard 184 polydimethylsiloxane (PDMS) (Dow Corning, USA) was poured over a template wafer and cured at 60°C for 4 hours. PDMS casts were peeled from the template wafer, yielding a slab with the channel structures embossed in negative relief. Terminal reservoirs 2 mm in diameter were punched at the end of each channel for fluidic access. Separately, glass microscope slides were spin-coated with a thin layer of PDMS and cured using the same conditions. The PDMS casts and PDMS-coated slides were treated with oxygen plasma in a PDC-32G plasma cleaner oven (Harrick Plasma, USA) for 60 s at a power of 18 W, then brought into contact and allowed to irreversibly bond. This created a channel system composed entirely of PDMS surfaces. The assembled device featured an overall length of 5 mm with channel depth of 12 µm. The inlet channel was 100 µm wide, the constriction feature was 20 µm wide, and all outlet channels were 30 µm wide (see also Figure 1).

Sample preparation and labeling The analytes used in these experiments included genomic lambda phage dsDNA (λDNA, 48.5 kbp) (ThermoFisher Scientific, USA), as well as linearized plasmid dsDNA obtained from QIAGEN (Germany). The plasmids obtained from QIAGEN were proprietary noncatalogue plasmids with lengths of 19.5, 10.2, and 1.0 kbp (QIAGEN, Germany). Prior to use, DNA analytes were diluted to 2 or 5 ng/µL in 5 mM sodium phosphate buffer at pH 7.7, corresponding to final concentrations ranging from 63.5 pM to 7.7 nM. The final running buffer also contained 1.0 mM Pluronic F108 block copolymer (Sigma-Aldrich Co., USA) and 0.2% v/v β-mercaptoethanol (Sigma-Aldrich Co., USA). dsDNA analytes were labeled using either YOYO-1 or BOBO-3 intercalating dyes (ThermoFisher Scientific, USA) at a 1:10 molar ratio of dye molecules to DNA base pairs. For dual-analyte experiments, two DNA analytes were separately labeled with different dyes. Prior to mixing, labeled DNA samples were filtered with 30 kDa molecular weight cutoff filters (EMD Millipore Corporation, USA) to remove unbound dye and avoid cross-labeling.

ACS Paragon Plus Environment

Page 8 of 29

Page 9 of 29

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

Analytical Chemistry

Sorting experiments Prior to running experiments, all microdevice reservoirs were filled with a 5 µL aliquot of sample buffer. Channels were allowed to fill via capillary action. This produced a homogenous dispersion of analyte throughout the channel assembly. All reservoirs were then topped with 3-5 µL of mineral oil (Bio-Rad Laboratories, USA) to prevent evaporation and assist with hydrodynamic flow control. The pressure was first balanced between all reservoirs, then biased equally towards all five outlets by adding buffer to the inlet reservoir. During experiments, flow rates were maintained at approximately 1.3 µL/h, with corresponding analyte velocities of approximately 185 µm/s. Platinum wire electrodes (0.3 mm diameter) were inserted into the inlet and center outlet reservoirs, and attached to an AMT-3B20 high-voltage AC amplifier (Matsusada Precision Inc., USA). A sinusoidal AC signal was programmed using Labview software (National Instruments, USA) and produced with a USB 6343 DAQ device (National Instruments, USA). Applied potentials ranged from 100 to 2400 Vpp (peak-to-peak voltage), with frequencies ranging from 50 Hz to 20 kHz. The distance between electrodes (not the electrode geometry) will affect overall field strength within the microdevice. These electric potentials therefore correspond to average field strengths of approximately 70 to 1700 Vrms/cm. Microdevices were secured on the stage of an Olympus IX71 inverted fluorescence microscope (Olympus, USA) and viewed with ×4, ×20, or ×40 objectives. Single-analyte experiments used YOYO-1 labeled DNA; fluorescence was collected with a filter set containing a 470/40 nm excitation filter and a 525/50 nm emission filter. Dual-analyte experiments were conducted with a dual-bandpass filter set with 468/34 nm and 553/26 nm excitation regions and 553/26 nm and 630/91 nm emission regions (Semrock, USA; model GFP/DsRed-A-000). Images were captured using a monochromatic QuantEM:512SC CCD camera (Photometrics, USA) along with Micro-Manager 1.4.22 software (UCFS, USA). For these experiments exposure time was set to either 60 or 100 ms. For dual-analyte experiments,

ACS Paragon Plus Environment

Analytical Chemistry

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

color images were acquired with a SPOT Idea 1.3 Mp CMOS camera (Diagnostic Instruments, Inc., USA). Color images were acquired as single exposures of either 5 or 10 s. ImageJ 1.51a software (NIH, USA) was used for image analysis and fluorescence intensity determinations.

Data analysis In order to examine continuous-flow sorting behavior for single-analyte experiments, images were collected from an operating microdevice. Signal was averaged across 100 consecutive frames at a constant set of experimental parameters. The resulting images of time-averaged fluorescence intensity were used to qualitatively gauge the degree of DNA sorting. Fractionation or separation was considered to occur when a greater concentration of analyte was observed in any outlet channel relative to the others. Since the observed fluorescence intensity is proportional to DNA concentration, these images were used to quantitate mean fluorescence intensity within regions of interest (ROIs) corresponding to each of the channel system branches. Circular ROIs (of 5 pixel radius) were selected along a radial arc perpendicular to the outlet channels and located approximately 100 µm from the device constriction. In each case, ROIs corresponded to an area entirely within the outlet channel in question. For dual-analyte experiments, pre-processing steps were performed in ImageJ. Red and green color channel data were extracted from each color image, and a rolling ball operation59 (using a disc element radius of 250 pixels) was performed on each channel to correct for non-uniform background illumination.

Numerical modeling The models used for simulations were based upon prior work with a constriction sorter microdevice,55,56 and adapted to reflect the experiments described here. In brief, adaptations included the use of pressure-driven flow and AC electric potentials. The buffer conductivities and analyte properties were selected to match those used for DNA sorting experiments. In a manner similar to electrophoretic velocity, a DEP velocity (࢛ୈ୉୔ ) can be defined as:60

ACS Paragon Plus Environment

Page 10 of 29

Page 11 of 29

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

Analytical Chemistry



࢛ୈ୉୔ = ߤୈ୉୔ સࡱଶ = ଶ௙ સࡱଶ

(Eq. 2)

with ߤୈ୉୔ serving as the proportionality factor that accounts for particle characteristics affecting DEP transport, including particle polarization and the friction coefficient (݂).41,45,51,61 The details of the numerical modeling are described in the supporting information accompanying this manuscript.

Results and Discussion DNA sorting in an operating iDEP microdevice was investigated using time-averaged fluorescence microscopy images. These images yielded qualitative and quantitative assessments of analyte sorting performance. First, the behavior of each DNA analyte was systematically examined across different applied AC potentials in separate microdevices. Subsequently, pairs of DNA analytes were simultaneously sorted in a single microdevice, using different fluorescent intercalators for each analyte.

Four different dsDNA analytes were examined using a constriction sorter microdevice. DNA samples were introduced through an inlet reservoir, then carried downstream via pressure-driven flow toward a bilaterally symmetric constriction and outlet nexus (Figure 1). Five outlets diverged from this nexus, designated S1 for the outermost outlets, S2 for middle outlets, and C for the center outlet.

ACS Paragon Plus Environment

Analytical Chemistry

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

Figure 1. Illustration of the constriction sorter microdevice design and its basic operation. A) Schematic: Sample containing DNA analyte was introduced through the circular reservoir at left. From there, analyte flowed through the inlet reservoir towards a constriction and five outlet branches, designated S1 or S2 for side outlets, or C for the center outlet. Platinum wire electrodes were inserted into the inlet and C outlet reservoirs and used to apply electric potential. B) An expanded view of the constriction region and the outlet branch nexus. Shaded blue areas show numerically-simulated regions of high-magnitude ∇ࡱଶ that generate DEP forces. Based on simulations using root mean square values for electric potential, centerline magnitudes of ∇ࡱଶ at the constriction ranged from 1.8 × 1014 to 7.0 × 1016 V2/m3. DEP forces selectively divert analyte from initial flow paths. Analyte may

be deflected and concentrated in the side outlets or the C outlet. This behavior depends on both analyte characteristics and electric field operating parameters. C) Bright-field image of an assembled constriction sorter microdevice, prior to introduction of analyte.

ACS Paragon Plus Environment

Page 12 of 29

Page 13 of 29

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

Analytical Chemistry

The constriction generates electric field gradients and concomitant dielectrophoretic forces. Under continuous-flow conditions, and with the DEP force not exceeding magnitudes sufficient for trapping, susceptible targets will be deflected as they pass the constriction. Both the direction and degree of DEP response are characteristic to a given analyte (Eq. 2). Analyte molecules will move either towards the channel centerline (in the case of nDEP) or the channel periphery (in the case of pDEP). Analyte experiencing pDEP at the constriction would thus be selectively diverted into S1 or S2 outlets, and depleted in the C outlet. Conversely, analyte experiencing negative DEP would be selectively diverted into the C outlet, and depleted in the S1 and S2 outlets. Each of these scenarios are illustrated in Figure 1B, with green arrows indicating pDEP deflection of analyte into side outlets and red arrows indicating nDEP deflection of analyte into the center outlet.

DNA sorting experiments Sorting phenomena for individual DNA analytes were investigated by varying the amplitude of the applied potential from 100 to 2400 Vpp and varying the frequency from 50 Hz to 20 kHz. Within these ranges, sorting was reproducibly observed for all four DNA analytes with magnitude and direction depending strongly upon the specific DNA analyte and the electric field characteristics. Figure 2 shows key qualitative results for λ-DNA across a range of potentials and frequencies. The top three images (Figure 3A – 3C) represent increasing potential at a constant frequency. At low potentials, little or no sorting was observed. DNA remained evenly distributed between all outlets. However, when the potential increased, concentrated bands were observed streaming through S1 outlets. This behavior became increasingly pronounced at higher potentials. The bottom three images (Figure 3D – 3F) represent increasing frequency at a constant potential. In the low frequency range, λ-DNA was deflected though the C outlet. As frequency increased, this behavior switched to deflection through the S1 outlets. When frequency increased past 5 kHz, the overall degree of sorting decreased. This switch in deflection from C to S1 was not observed for all DNA analytes.

ACS Paragon Plus Environment

Analytical Chemistry

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

Instead, this behavior was only associated with the two larger DNAs, 48.5 and 19.5 kbp in length.

Figure 2. Key results illustrating electric field dependence of DEP sorting behavior for λ-DNA. Images show average FI near the constriction and outlet nexus of the microdevice. A-C) The top three images show analyte distribution within the device for increasing potentials at a constant frequency of 1000 Hz. At an applied potential of 200 Vpp (A), little or no deflection is observed. The degree of deflection into S1 outlets increases with higher potentials. At 2000 Vpp (C), a high degree of deflection and enrichment is observed in S1 outlets, as well as depletion in the C outlet. D-F) The bottom three images show analyte distribution for increasing frequencies at a constant applied potential of 1000 Vpp. At a frequency of 50 Hz (D), λ-DNA is deflected into the C outlet, and depleted in both S outlets. At 250 Hz (E), DNA is still deflected into C outlets. At 1000 Hz (F), DNA is deflected into S1 outlets. At 5000 Hz and higher frequencies (not shown), little or no deflection occurs.

Overall trends in sorting behavior were markedly different for each of the four DNA analytes. The largest differences were observed in the low frequency range of applied potential, where longer DNA molecules exhibited C sorting while shorter DNA molecules

ACS Paragon Plus Environment

Page 14 of 29

Page 15 of 29

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

Analytical Chemistry

exhibited S1 sorting. A comparison of sorting behavior for all four DNA analytes is shown in Figure 3. To facilitate comparison, all images were collected with an AC potential of 2000 Vpp and 100 Hz. Under these conditions, 1.0 and 10.2 kbp DNA exhibited S1 sorting (Figure 3A and 3B), while 19.5 and 48.5 kbp DNA exhibited C sorting (Figure 3C and 3D). The low-frequency C-outlet deflection of larger DNA observed in individual experiments provided an opportunity to simultaneously separate two analytes. Pairs of DNA analytes were simultaneously introduced into a microdevice. Each analyte was labeled with a different fluorescent intercalator: either YOYO-1 (green emission) or BOBO-3 (red emission). An AC potential was chosen to elicit C sorting for larger DNA and S1 sorting for smaller DNA. The results of dual-analyte experiments corresponded with single-analyte experiments. Figure 3E shows an example of simultaneous deflection of 10.2 and 48.5 kbp DNA into different outlets of the microdevice. Reversing the labels had no effect on outcome, which confirmed that the resulting separation was independent of labeling strategy (data not shown).

Figure 3. Comparison of DEP sorting behavior of four different DNA analytes: three linearized plasmids (1.0, 10.2, and 19.5 kbp), and λ-DNA (48.5 kbp). These images show time-averaged FI near the constriction region. All images reflect a single set of electrical parameters: 2000 Vpp and 100 Hz. A) 1.0 kbp DNA displayed a weak S1 sorting effect. B) For 10.2 kbp DNA, pronounced S1 sorting was observed. C) 19.5 kbp displayed C sorting behavior. D) 48.5 kbp DNA demonstrated pronounced C sorting. E) Experimental results showing simultaneous sorting of two DNA analytes labeled with different fluorophores. 10.2 kbp plasmid DNA was deflected and concentrated in S1 outlets while 48.5 kbp genomic DNA was deflected and concentrated in the C outlet.

ACS Paragon Plus Environment

Analytical Chemistry

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

Page 16 of 29

Bimodal sorting efficiency Depending on experimental parameters, analyte may either be enriched or depleted in the S1 outlets. The possibility for bimodal sorting distribution created the need for a normalizing expression which compares the relative concentration of a single analyte between both S1 and C outlets. The bimodal sorting efficiency (Σ∗ ) is defined as: ூ

ିூ

Σ∗ = ூ౏భ ାூి ౏భ

(Eq. 3)

ి

where ‫ܫ‬ୗଵ represents the fluorescence intensity of analyte streaming though S1 outlets and IC represents the fluorescence intensity of analyte in the C outlet. This expression serves as de

facto normalization for each sorting experiment. Its output serves as a bimodal sorting efficiency where the sign denotes into which outlet material is deflected. Output values can range from -1 to 1: values between 0 and 1 indicate higher analyte concentration in the S1 outlet and values between 0 and –1 indicate higher concentration in the C outlet. Values of 1 or -1 would reflect complete deflection of analyte into a given outlet, while a value of 0 would indicate equal distribution between all outlets. Mean fluorescence intensities were measured within ROIs representative of each outlet channel. These values were used with Eq. 3 to determine and compare experimental sorting efficiency values for the four DNA analytes across a range of applied potentials and frequencies (Figure 4). Sweeping the potential from 200 to 2400 Vpp at a constant frequency of 500 Hz demonstrated generally increasing trends for sorting efficiency at higher potentials (Figure 4A). This was true for all four lengths of DNA, except for λ-DNA which exhibited decreased sorting efficiency at 2400 Vpp. Also noteworthy was the higher sorting efficiency for larger DNA molecules compared to the smaller plasmids.

ACS Paragon Plus Environment

Page 17 of 29

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

Analytical Chemistry

Figure 4. Plots showing experimentally determined sorting efficiency for four different DNA analytes: 1.0, 10.2, 19.5, and 48.5 kbp DNA. Positive and negative values indicate side-outlet or center-outlet sorting, respectively, with normalized absolute magnitudes of 1.0 representing 100% sorting of the given analyte. A) Sorting efficiency values at fixed AC frequency of 1000 Hz with AC amplitude varying from 200 to 2400 Vpp B) Sorting efficiency values at a fixed AC amplitude of 1000 Vpp with AC frequency varying from 50 Hz to 20 kHz. C) Sorting efficiency values for 48.5 kbp DNA obtained by sweeping AC frequency from 50 Hz to 5.0 kHz at a higher amplitude of 2000 Vpp. This data range exhibited the highest-efficiency sorting observed, with values exceeding -0.92 ± 0.03 for 50 Hz.

ACS Paragon Plus Environment

Analytical Chemistry

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

Page 18 of 29

Observed trends match expectations based on DEP theory. The magnitude of DEP force scales with the square of the electric field strength (Eq. 1). Thus the higher sorting efficiencies observed at higher AC amplitudes correspond with an expected increase in DEP force. In general, larger DEP forces are also expected to occur for larger analytes. This results from analyte polarizability scaling with radius (for particles) or length (for DNA molecules). Traditional treatments of DEP assume that particles act as lossless spheres and thus fail to address the dynamic geometric complexity of DNA molecules. In dilute buffers, dsDNA molecules longer than 6.0 kbp form coils that can be considered to behave as porous spheres.62 When exposed to AC fields, these molecules align with the field and adopt the form of either prolate spheroids in weak fields or linear rods in strong fields.63 These effects and phenomena serve as the subject of ongoing research64 but our present study underlines previously reported AC DEP-based observations of increased DNA polarizability with increasing length.51 In the case of DNA sorting behavior, this is observed as higher sorting efficiency for larger DNA molecules. Figure 4B shows sorting efficiencies determined by sweeping AC frequency from 50 Hz to 20 kHz at a constant amplitude of 1000 Vpp. In all cases, maximal sorting efficiencies were observed at the lower end of the frequency range. Above 1000 Hz, sorting efficiency magnitude progressively decreased for all four analytes. A decreased sorting efficiency indicates that less analyte is displaced from initial flow paths. Reduced analyte displacement likely stems from a decrease in DEP force, which in this case would be associated with lower ߤୈ୉୔ for these electric field parameters. This explanation is bolstered by independent observations of decreasing DEP response for dsDNA when AC frequencies increase above 10 kHz.42,45,65 Various reports have observed frequency-dependent DEP response for DNA, generally in the range of 10 kHz to 5 MHz.43,44,66 However, attempts to quantify molecular ߤୈ୉୔ may be complicated by the effects of analyte choice and experimental setup. Many of these experiments use DNA analytes of different chain-length, or use different methods for observing and quantifying DEP response and thus cannot be directly compared with the present case.

ACS Paragon Plus Environment

Page 19 of 29

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

Analytical Chemistry

Frequency-dependent data demonstrated an intriguing switch between C- and S1sorting for λ-DNA. At low frequencies, λ-DNA was selectively deflected into the C outlet, while at higher frequencies λ-DNA was instead deflected through the S1 outlets. The frequency at which this switch occurred depended upon the magnitude of applied potential. Specifically, larger AC amplitudes increased the switching frequency. To confirm this effect, another frequency sweep was performed for λ-DNA at a higher AC amplitude of 2000 Vpp (Figure 4C). For an applied potential of 1000 Vpp, a switch from C- to S-sorting occurred between 100 Hz and 200 Hz; however, when applying a 2000 Vpp potential, the switching frequency increased to between 500 Hz and 1000 Hz. Results obtained from dual-analyte sorting experiments qualitatively agreed with the single-analyte experiments (Figure 3E). Average sorting efficiency values calculated from these data were approximately 0.32 ± 0.09 for 10.2 kbp DNA and -0.43 ± 0.18 for 48.5 kbp DNA. The sign of the sorting efficiency values and the modality of these results agreed with those reported for single-analyte experiments. The decreased magnitude of sorting efficiency values relative to single analyte experiments may be attributed in part to spectral overlap between fluorescence excitation and emission filters, overlap between the green and red filters in the color CCD’s Bayer mask, and pixel crosstalk.67 Variations between fabricated devices and differences in flow rate may also contribute to disparity. The switch between center-deflection and side-deflection for large DNA molecules was not anticipated based on an a priori understanding of the dielectrophoretic response of single DNA molecules, and merits further investigation. The change may result from a lowfrequency-induced crossover from positive to negative values of ߤୈ୉୔ , along with the associated switch from pDEP to nDEP. We suggest two arguments supporting a switch from pDEP to nDEP and the associated sorting behavior in the constriction device. First, we used numerical modeling to verify if the sorter could (in principle) operate using either pDEP or nDEP for DNA (see supporting information for model details). Figure 5A shows the simulated concentration distribution for λ-DNA affected by pDEP (ߤୈ୉୔ > 0). In this case, analyte is deflected into S1 outlets. This is illustrated by the concentrated bands (shown in red) that form along the edges of the outlet channels. This result is consistent with

ACS Paragon Plus Environment

Analytical Chemistry

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

streaming pDEP deflection of analyte towards higher electric field strength along the channel periphery. Figure 5B shows concentration distribution for DNA affected by nDEP (ߤୈ୉୔ < 0). In this scenario, S1 and S2 outlets are depleted of analyte relative to the C outlet. This result is similarly consistent with streaming nDEP deflection of analyte towards lower electric field strength along the channel centerline. With no DEP mobility (ߤୈ୉୔ < 0), or no potential applied, analyte remains uniformly distributed throughout the microdevice (data not shown). This result is consistent with an absence of DEP effects, reflecting pressuredriven bulk transport alone.

Figure 5. Numerical simulations showing concentration distribution of λ-DNA molecules in the constriction sorter. Model parameters are detailed in the supporting information. A) When potential is applied to the C outlet and λ-DNA is assigned a positive ߤୈ୉୔ a higher concentration of analyte is calculated in the S1 outlets, indicating pDEP deflection. B) With the same electric field parameters, if

ACS Paragon Plus Environment

Page 20 of 29

Page 21 of 29

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

Analytical Chemistry

λ-DNA is assigned a negative ߤୈ୉୔ a higher concentration of analyte is calculated in the C outlet, indicating nDEP deflection. In each case, the surface plot color indicates analyte concentration. For pDEP simulations (A), sorting is indicated by the formation of enriched streams in the side outlets. For nDEP simulations (B), sorting is indicated by the formation of depleted streams in the side outlets.

These simulations indicate that ߤୈ୉୔ values varying in sign (yielding pDEP vs nDEP) may result in sorting analyte into different outlet branches of the microdevice. DEP mobility for DNA molecules has been shown to vary with the magnitude and frequency of an applied AC potential, but the physical phenomena underlying these shifts are complex and not well understood. Interestingly, qualitative empirical findings showed that all instances of C-sorting for λ-DNA were associated with a visible change in the dispersion characteristics of suspended DNA. Upon application of low-frequency and high-amplitude potentials, homogeneously well-dispersed DNA molecules formed large, segregated globular clusters. This effect is not readily apparent in the time-averaged images of sorting shown in Figures 2 and 3, but becomes quite evident with higher frame rate capture. Agglomeration of DNA sample was investigated in a straight microchannel using experimental parameters that mirrored those used for sorting experiments (Figure 6). After applying AC potential to a homogeneous dispersion of λ-DNA, distinct clusters formed over a period of 10 – 20 seconds. Figure 6B shows DNA clusters under the influence of an AC field. Clusters take on an elongated allantoid appearance due to electrokinetic center-ofmass oscillation during 60 – 100 ms exposure times.64 Upon removing the AC potential, discrete globular clusters were observed before gradually dispersing (Figure 6C).

ACS Paragon Plus Environment

Analytical Chemistry

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

Figure 6. Images showing AC-induced reversible agglomeration of λ-DNA in a straight microchannel, at 2000 Vpp and 50 Hz. A) Prior to applying AC potential, suspended λ-DNA molecules are freely dispersed throughout the microchannel. B) After approximately 15 seconds of applied potential, agglomeration of λ-DNA produces segregated clusters. The image appears blurred due to the appearance of allantoid or dumbbell-like shapes. These shapes result from center-of-mass oscillation during the 60 ms exposure time. C) Immediately after removing the potential, agglomerates are still clearly visible.

Theoretical and experimental work has investigated low-frequency agglomeration of colloids and polyelectrolytes into macro-ion clusters.68,69 Such agglomerates have been shown to occur for λ-DNA dispersions with AC frequencies, amplitudes, and buffers similar to those used in the experiments presented here.70 Reports have demonstrated clusters consisting of up to hundreds of loosely-interacting molecules.70 We estimate that clusters in our study contain 10 or more molecules based on experimental observations. Cluster formation manifests as a result of electrohydrodynamic instability of macro-ion dispersions.71 This instability has been attributed to attraction forces between neighboring chains, which arises from long-wavelength fluctuations in charge distribution and molecular conformation.71,72 While low-frequency nDEP of single DNA molecules has not been widely reported, it is plausible that transiently interacting clusters of DNA molecules may be governed by different electrophysical characteristics than the individual constituent molecules. DNA

ACS Paragon Plus Environment

Page 22 of 29

Page 23 of 29

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

Analytical Chemistry

clusters may thus behave as a new quasiparticle with a unique ߤୈ୉୔ , possibly owing to a larger effective radius as well as concurrent changes for net quasiparticle conductivity and permittivity, ultimately affecting polarizability. A physical basis for this phenomenon is not presently understood; elucidating such a mechanism is beyond the scope of this work. We aim to further illuminate the electrohydrodynamic properties of DNA suspensions under the experimental conditions suitable for DNA sorting in the future.

Conclusions This study demonstrates selective sorting of DNA in a constriction sorter microdevice. Four different dsDNA analytes, ranging in size from 1.0 to 48.5 kbp, were selectively enriched and depleted in different outlets of the microdevice. These results elucidate conditions that enable size-based iDEP fractionation of small to large DNA fragments. This work expands existing knowledge regarding the use of DEP to selectively manipulate DNA and will facilitate future improvements to microdevices designed for separation and fractionation of DNA and other nanoscale bioanalytes. Experimental values for sorting efficiency ranged between approximately 0.59 ± 0.04 (reflecting side-outlet sorting) and -0.92 ± 0.03 (reflecting center-outlet sorting). These values indicate that over 90% sorting efficiency is achievable. Under certain conditions, AC electrohydrodynamic agglomeration of long DNA molecules facilitated sorting and was associated with an apparent shift from pDEP to nDEP behavior. While formation of such clusters has been reported previously by other researchers, the possible effects of cluster formation on DEP mobility have not been investigated and will serve as the subject of future studies. The majority of presently existing NGS methods are limited to read lengths below a few thousand base-pairs, but some approaches function with DNA lengths up to 40 kbp. High-efficiency sorting of relatively long-chain DNA molecules into different outlets than short-chain DNA molecules is therefore particularly relevant for NGS analytics. Furthermore, the dependency of sorting behavior on AC frequency could enable tunable separation of

ACS Paragon Plus Environment

Analytical Chemistry

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

DNA molecules into fractions more suited for NGS analysis of either short (< 10 kbp) or long (> 10 kbp) read lengths. In principle, the sorting performance of the dielectrophoresisbased microdevice presented here is tunable based on applied potential and frequency. These parameters can therefore be used to determine effective thresholds for diversion of select DNA populations into particular output channels. This approach represents a viable mechanism for tailoring DNA fractionation to a variety of downstream analysis applications.

Acknowledgements Financial support from the National Science Foundation (NSF PFI:AIR-TT grant #1445006) is gratefully acknowledged. We also thank QIAGEN for project support and provision of key materials. We would also like to thank Anikki Geissler and Kamran Bodushev for assistance with microdevice preparation.

Supporting Information Details regarding numerical models of DNA sorting. This material is available free of charge online at http://pubs.acs.org.

References (1)

Sanger, F.; Coulson, A. R. J. Mol. Biol. 1975, 94 (3), 441–448.

(2)

Shendure, J.; Ji, H. Nat. Biotechnol. 2008, 26 (10), 1135–1145.

(3)

Metzker, M. L. Nature Rev. Genet. 2010, 11 (1), 31–46.

(4)

Farzadfard, F.; Lu, T. K. Science 2014, 346 (6211), 1265272.

(5)

Portney, N. G.; Wu, Y.; Quezada, L. K.; Lonardi, S.; Ozkan, M. Langmuir 2008, 24 (5), 1613–1616.

(6)

Jonoska, N.; Karl, S. A.; Saito, M. BioSystems 1999, 52 (1-3), 143–153.

(7)

Boruah, K.; Dutta, J. C. Int. Conf. on Electronic Design, Computer Networks &

ACS Paragon Plus Environment

Page 24 of 29

Page 25 of 29

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

Analytical Chemistry

Automated Verification 2015, 52–57. (8)

Mitchell, P. S.; Parkin, R. K.; Kroh, E. M.; Fritz, B. R.; Wyman, S. K.; PogosovaAgadjanyan, E. L.; Peterson, A.; Noteboom, J.; O'Briant, K. C.; Allen, A.; Lin, D. W.; Urban, N.; Drescher, C. W.; Knudsen, B. S.; Stirewalt, D. L.; Gentleman, R.; Vessella, R. L.; Nelson, P. S.; Martin, D. B.; Tewari, M. Proc. Natl. Acad. Sci. U.S.A. 2008, 105 (30), 10513–10518.

(9)

Rothemund, P. W. K. Nature 2006, 440 (7082), 297–302.

(10)

Quail, M. A.; Kozarewa, I.; Smith, F.; Scally, A.; Stephens, P. J.; Durbin, R.; Swerdlow, H.; Turner, D. J. Nat. Methods 2008, 5 (12), 1005–1010.

(11)

Lin, Y.-W.; Huang, M.-F.; Chang, H.-T. Electrophoresis 2005, 26 (2), 320–330.

(12)

Duke, T.; Semenov, A.; Viovy, J. Phys. Rev. Lett. 1992, 69 (22), 3260–3263.

(13)

Herschleb, J.; Ananiev, G.; Schwartz, D. C. Nat. Protoc. 2007, 2 (3), 677–684.

(14)

Dorfman, K. D.; King, S. B.; Olson, D. W.; Thomas, J. D. P.; Tree, D. R. Chem. Rev. 2013, 113 (4), 2584–2667.

(15)

Chou, H.-P.; Spence, C.; Scherer, A.; Quake, S. R. Gourley, P. L., Ed.; International Society for Optics and Photonics, 1998; Vol. 3258, pp 181–187.

(16)

Chou, H.-P.; Spence, C.; Scherer, A.; Quake, S. Proc. Natl. Acad. Sci. U.S.A. 1999, 96 (1), 11–13.

(17)

Goodwin, P. M.; Johnson, M. E.; Martin, J. C.; Ambrose, W. P.; Marrone, B. L.; Jett, J. H.; Keller, R. A. Nucleic Acids Res. 1993, 21 (4), 803–806.

(18)

Thomas, J. D. P.; Dorfman, K. D. Biomicrofluidics 2014, 8 (3), 034115.

(19)

Huang, L. R.; Tegenfeldt, J. O.; Kraeft, J. J.; Sturm, J. C.; Austin, R. H.; Cox, E. C. Nat.

Biotechnol. 2002, 20 (10), 1048–1051. (20)

Han, J.; Craighead, H. G. Science 2000, 288 (5468), 1026–1029.

(21)

Fu, J.; Schoch, R. B.; Stevens, A. L.; Tannenbaum, S. R.; Han, J. Nat. Nanotechnol. 2007,

2 (2), 121–128. (22)

Fu, J.; Mao, P.; Han, J. Appl. Phys. Lett. 2005, 87 (26), 263902.

(23)

Thomas, J. D. P.; Joswiak, M. N.; Olson, D. W.; Park, S.-G.; Dorfman, K. D. Lab Chip 2013, 13 (18), 3741–3746.

(24)

Duke, T.; Austin, R. H. Phys. Rev. Lett. 1998, 80 (7), 1552–1555.

(25)

Pethig, R. Biomicrofluidics 2010, 4 (2), 022811.

(26)

Tay, F. E. H.; Yu, L.; Iliescu, C. Defence Science Journal 2009, 59 (6), 595–604.

(27)

Çetin, B.; Li, D. Electrophoresis 2011, 32 (18), 2410–2427.

(28)

Fiedler, S.; Shirley, S. G.; Schnelle, T.; Fuhr, G. Anal. Chem. 1998, 70 (9), 1909–1915.

ACS Paragon Plus Environment

Analytical Chemistry

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

(29)

Gagnon, Z. R. Electrophoresis 2011, 32 (18), 2466–2487.

(30)

Voldman, J. Annu Rev Biomed Eng 2006, 8 (1), 425–454.

(31)

Bhattacharya, S.; Chao, T. C.; Ros, A. Electrophoresis 2011, 32 (18), 2550–2558.

(32)

Jones, P. V.; Staton, S. J. R.; Hayes, M. A. Anal. Bioanal. Chem. 2011, 401 (7), 2103– 2111.

(33)

Jones, P. V.; DeMichele, A. F.; Kemp, L.; Hayes, M. A. Anal. Bioanal. Chem. 2014, 406 (1), 183–192.

(34)

Jones, P. V.; Huey, S.; Davis, P.; McLemore, R.; McLaren, A.; Hayes, M. A. Analyst 2015,

140 (15), 5152–5161. (35)

Ding, J.; Lawrence, R. M.; Jones, P. V.; Hogue, B. G.; Hayes, M. A. Analyst 2016, 141 (6), 1997–2008.

(36)

Camacho-Alanis, F.; Gan, L.; Ros, A. Sensor. Actuat. B-Chem. 2012, 173, 668–675.

(37)

Gan, L.; Chao, T. C.; Camacho-Alanis, F.; Ros, A. Anal. Chem. 2013, 85 (23), 11427– 11434.

(38)

Cummings, E. B.; Singh, A. K. Anal. Chem. 2003, 75, 4724–4731.

(39)

Chou, C.; Tegenfeldt, J.; Bakajin, O.; Chan, S.; Cox, E.; Darnton, N.; Duke, T.; Austin, R.

Biophys. J. 2002, 83 (4), 2170–2179. (40)

Pohl, H. A. J. Appl. Phys. 1951, 22 (7), 869–871.

(41)

Hölzel, R. IET Nanobiotechnol 2009, 3 (2), 28–45.

(42)

Nishikawa, A.; Tsukahara, S.; Fujiwara, T. Bunseki Kagaku 2005, 54 (6), 459–465.

(43)

Henning, A.; Henkel, J.; Bier, F. F.; Hölzel, R. PMC Biophys. 2008, 1 (1), 4.

(44)

Bakewell, D. J.; Morgan, H. IEEE Trans. Nanobiosci. 2006, 5 (1), 1–8.

(45)

Martinez-Duarte, R.; Camacho-Alanis, F.; Renaud, P.; Ros, A. Electrophoresis 2013, 34 (7), 1113–1122.

(46)

Yokokawa, R.; Manta, Y.; Namura, M.; Takizawa, Y.; Le, N. C. H.; Sugiyama, S. Sensor.

Actuat. B-Chem. 2010, 143 (2), 769–775. (47)

Washizu, M.; Kurosawa, O. IEEE Trans. Ind. Appl. 1990, 26 (6), 1165–1172.

(48)

Lao, A. I. K.; Hsing, I.-M. Lab Chip 2005, 5 (6), 687–690.

(49)

Washizu, M.; Nanba, T.; Masuda, S. IEEE Trans. Ind. Appl. 1990, 26 (2), 352–358.

(50)

Ajdari, A.; Prost, J. Proc. Natl. Acad. Sci. U.S.A. 1991, 88 (10), 4468–4471.

(51)

Regtmeier, J.; Duong, T. T.; Eichhorn, R.; Anselmetti, D.; Ros, A. Anal. Chem. 2007, 79 (10), 3925–3932.

(52)

Parikesit, G. O. F.; Markesteijn, A. P.; Piciu, O. M.; Bossche, A.; Westerweel, J.; Young, I. T.; Garini, Y. Biomicrofluidics 2008, 2 (2), 24103.

ACS Paragon Plus Environment

Page 26 of 29

Page 27 of 29

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

Analytical Chemistry

(53)

Viefhues, M.; Wegener, S.; Rischmüller, A.; Schleef, M.; Anselmetti, D. Lab Chip 2013,

13 (15), 3111–3118. (54)

Viefhues, M.; Regtmeier, J.; Anselmetti, D. Analyst 2013, 138, 186–196.

(55)

Abdallah, B. G.; Chao, T. C.; Kupitz, C.; Fromme, P.; Ros, A. ACS Nano 2013, 7 (10), 9129–9137.

(56)

Abdallah, B. G.; Roy-Chowdhury, S.; Coe, J.; Fromme, P.; Ros, A. Anal. Chem. 2015, 87 (8), 4159–4167.

(57)

Abdallah, B. G.; Zatsepin, N. A.; Roy-Chowdhury, S.; Coe, J.; Conrad, C. E.; Dörner, K.; Sierra, R. G.; Stevenson, H. P.; Camacho-Alanis, F.; Grant, T. D.; Nelson, G.; James, D.; Calero, G.; Wachter, R. M.; Spence, J. C. H.; Weierstall, U.; Fromme, P.; Ros, A. Struct.

Dyn. 2015, 2 (4), 041719. (58)

Nakano, A.; Camacho-Alanis, F.; Chao, T. C.; Ros, A. Biomicrofluidics 2012, 6 (3), 34108.

(59)

Sternberg, S. R. Computer 1983, 16 (1), 22–34.

(60)

Regtmeier, J.; Eichhorn, R.; Viefhues, M.; Bogunovic, L.; Anselmetti, D. Electrophoresis 2011, 32 (17), 2253–2273.

(61)

Regtmeier, J.; Eichhorn, R.; Bogunovic, L.; Ros, A.; Anselmetti, D. Anal. Chem. 2010,

82 (17), 7141–7149. (62)

Zhao, H. Phys. Rev. E 2011, 84 (2), 021910(1)–021910(6).

(63)

Cohen, A. E. Phys. Rev. Lett. 2003, 91 (23), 235506.

(64)

Zhou, C.; Riehn, R. Phys. Rev. E 2015, 92 (1), 012714.

(65)

Asbury, C. L.; Diercks, A. H.; van den Engh, G. Electrophoresis 2002, 23 (16), 2658– 2666.

(66)

Tuukkanen, S.; Toppari, J. J.; Kuzyk, A.; Hirviniemi, L.; Hytoenen, V. P.; Ihalainen, T.; Toermae, P. Nano Lett. 2006, 6 (7), 1339–1343.

(67)

Hirakawa, K. Image Processing 2008, 677–680.

(68)

Magnúsdóttir, S.; Isambert, H.; Heller, C.; Viovy, J. L. Biopolymers 1999, 49 (5), 385– 401.

(69)

Isambert, H.; Ajdari, A.; Viovy, J. L.; Prost, J. Phys. Rev. E 1997, 56 (5), 5688–5704.

(70)

Mitnik, L.; Heller, C.; Prost, J.; Viovy, J. L. Science 1995, 267 (5195), 219–222.

(71)

Castañeda-Priego, R.; Grünberg, von, H. H. J. Phys. Condens. Matter 2004, 16, S3987–S3998.

(72)

Bruinsma, R. F.; Riehn, R. ChemPhysChem 2009, 10 (16), 2871–2875.

ACS Paragon Plus Environment

Analytical Chemistry

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

Page 28 of 29

Page 29 of 29

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

Analytical Chemistry

For TOC Only

ACS Paragon Plus Environment