Accelerated DNA recombination on a functionalized microfluidic chip

Abstract

Genetic recombination is a powerful tool to create functional DNA hybrids with high controllability and precision for applications such as target gene therapy and molecular diagnostics, but it is often limited by the slow and low-efficient reconstruction of novel DNA species, especially in bulky volume solutions. Here, we present a microfluidic platform to enable rapid and consecutive on-line DNA digestion and ligation. The microdevice, embedded with a temperature sensor, offers controllable temperature and confined space to enhance intermolecular collision, and thus improves the slicing and splicing of DNA strands of interest. The stepwise functionalization of the microchannel surface with APTES, biotin and streptavidin permits the immobilization of biotinylated double stranded DNA (dsDNA) labeled with fluorescent probes at the ends for real-time monitoring. Digestion enzyme (BamH1) and DNA T4 ligase, subsequently, flow through the channel to hydrolyze and ligate the target DNA segments separately, and both digestion and ligation occur to a large extent within 10 min, with an efficiency of up to 71% (maximal at 1 h) and 63% (maximal at 2 h) of the highest value, respectively. Compared to conventional molecular diagnostics methods, which take several hours for digestion and ligation and deplete large volumes of reagents, such a system requires minimal time and sample volumes that are desirable in low-cost molecular therapy and point-of-care tests.

---

Introduction

Genetic engineering-based targeted therapy can serve as a promising approach to address the questions arising from symptoms-based treatments by reducing undesirable side-effects.¹ In genetic engineering, subcloning is the most commonly used technique in which an interest gene (insert) is transported from a chosen vector to a destination vector. Typical subcloning is restriction cloning, the most straightforward and economical DNA recombination, in which the DNA fragments of interest are digested by restriction enzymes (REs).² The detailed DNA recombination process of restriction cloning includes RE digestion, gel electrophoresis, DNA extraction and insert-vector ligation.³ Apart from gene target therapy, emerging technologies such as DNA sequencing,^4,5 DNA nanostructuring,⁶ DNA computing^7,8 and synthetic biology^9,10 are also involved in DNA recombination. However, complicated and troublesome bench-top DNA digestion and ligation limits its advantages. The entire process usually takes from several hours to one day. Furthermore, the operation processes of gel cutting expose personnel to risks of both EtBr and UV hazards. Therefore, a highly efficient way with controllability for DNA recombination without the drawbacks mentioned above is highly expected.

Microfluidics has great appeal in nucleic acid research, with the advantages of high reaction efficiency, short analysis times, simple operation, easy integration, and low reagent consumption.^11–13 Most of the nucleic acid reactions involved in DNA recombination, such as hybridization,^14–16 digestion,^17,18 ligation^19,20 and PCR,^21–26 have been realized on microfluidic chips. Thus, it is possible to transfer the routine in-test-tube DNA recombination to a novel all-in-one microfluidic platform. Recently, several new developed microfluidic systems for DNA digestion and ligation have been reported. For instance, DNA digestion, gel electrophoresis, DNA extraction and insert-vector ligation have been integrated on a chip,³ and another system composed of a reaction unit (comprising DNA digestion and ligation) and detection unit (a homemade LIF detection) has been reported.²⁷

Although microfluidics has been widely used in nucleic acid assay, only a few studies were done on the further analysis of DNA reconfiguration, including its digestion with restriction endonucleases and ligation procedure needed for DNA recombination. Both REs and ligases play an important role in mapping the structure of chromosomes, sequencing long-chain DNA and creating new DNA molecules that are desirable in molecular assembly and diagnostics.^28,29 Nevertheless, such static processes are not compatible with the speed requirements of either rapid clinical test or point-of-care diagnostics.³⁰ In addition, these works offer a homogeneous view of sequential DNA digestion and ligation, which generally involves complex electrophoresis procedures and multifunctional micro-zones design that may require the integration of multiple detection modes.^3,27 Therefore, it is critical to build up a platform for continuous and quick break-up and splicing of target DNA sequences to realize accelerated DNA recombination.

Herein, a functionalized microfluidic system was developed to digest and ligate target DNA fragments as a proof-of-concept for rapid genetic engineering. This system was integrated onto a silicon chip with heating components and enzymatic reactions take place within the functionalized chip thereafter. Such surface modification with stepwise functionalization is a key process for improving the binding affinity between the channel surface and the DNA of interest. Moreover, the microscale diffusion space of the channel promotes enhanced molecular interactions that are particularly significant for biomolecular assay. By conjugating different fluorescent dyes, the measurement of splicing and detachment segments of DNA is possible via real-time imaging. This chip platform allows an accelerated DNA recombination with the assistance of surface functionalization and temperature controllability that are critical to the enzyme activity. By comparing with previous research outcomes in this area, this work provides the following advantages: (1) a simple and electrophoresis-free strategy for rapid DNA recombination on a heterogeneous interface; (2) an in situ detection mode, like imaging, amenable to the real-time recording of the dynamic results of DNA digestion or ligation.

Experimental

Device design and fabrication

A semiconductor material, silicon wafer based microfluidic device was fabricated in the standard semiconductor clean room by several microfabrication steps such as wet etching, deposition of conductive materials, and photolithographic patterning. Once the silicon chip was manufactured and cleaned, PDMS pre-polymer and curing agents (10 : 1) (Sylgard 184, Dow Corning, USA) were mixed and degassed. The mixture was poured onto a glass slide and cured for 2 h at 80 °C. Then, the peeled PDMS was punched with inlet and outlet holes. The prepared PDMS sheet was permanently bonded to the silicon substrate with microchannel through plasma treatment (30 s, middle strength) (PDC-32G-2, Harrick, USA). A single serpentine fluidic channel was designed to deliver reagents with one inlet and one outlet (Fig. 1). It has a volume capacity of approximately 50 μl with a heater and sensor point embedded on the chip for combining the usage with a temperature controller for enzymatic digestion. This temperature controller was fabricated in-house by embedding its own heater and sensor electrode underneath the channel layer.³¹ Note that resistance calibration must be done prior to chip functionalization. The resistances of the chip were first measured at room temperature (23 °C) and then gauged at 55 °C. Both these values are important for the controller to recognize the degree of change of resistance with respect to temperature (Ω °C⁻¹), and the role of the controller is to heat the specimen in the chip. Moreover, a Scienion spotter machine was used to cool the samples that were needed for ligation after the DNA strand for DNA recombination was digested.

Fig. 1 On-chip rapid DNA recombination. (a) A photograph of a microfluidic chip with heater elements for DNA digestion and ligation. (b) A magnified schematic of microscale DNA immobilization and recombination in the channel. (c) Time requirement of microscale DNA recombination reduced from several hours to several minutes.

Surface functionalization

A method similar to the reported approach by Barrett et al. was used to modify the silicon chip surface.³² Ammonium hydroxide (NH₄OH) was first used to clean the photoresist residue from the silicon wafers and generate a hydrophilic surface with a large number of silanol groups by hydroxylating the channel surface. Such treatment was realized by the delivery of pure NH₄OH into the silicon chip at a rate of 0.5 ml h⁻¹ for 30 min using a syringe pump (PHD 2000, Harvard Apparatus) in a continuous flow mode. 3-Aminopropyl-triethoxysilane (APTES) (Sigma, St Louis, MO), a silane coupling agent, was used to attach a primary amine functional group to the surface. The chip was silanized by injecting 1% (v/v) solution of APTES mixed with ultrapure water and denatured ethanol into the silicon chip using a continuous flow syringe pump at the rate of 0.7 ml h⁻¹ for 2 h. The substrates were then washed with ethanol to remove the excess silane reagent and then dried using nitrogen gas. Sulfo–NHS–Biotin (Sigma, St Louis, MO), as a popular type of biotinylation reagent, can react efficiently with primary amines (–NH₂) to form stable amide bonds in pH 7–9 buffers. After silylation, 10 μg ml⁻¹ of Sulfo–NHS–Biotin was driven into the chip and incubated at room temperature for 1 h for surface biotinylation. The chip was then rinsed thoroughly with buffer (1× PBS) to remove any non-reacted biotinylated reagents. Subsequently, 1 μg ml⁻¹ of prepared streptavidin (Sigma, St Louis, MO) was delivered and flowed through the biotin-functionalized channel surface for avidin modification (1 h) via extraordinarily strong affinity between biotin and streptavidin.

DNA hybridization, digestion and ligation

Two oligonucleotides (Oligo 1 and Oligo 2) were synthesized and labeled with Cy5 and FAM fluorescent dyes, respectively (1st BASE, Singapore) (Table 1). Both these sequences were annealed using a thermal cycler (Bio-Rad) with a reaction volume of 20 μl—10 μl of Oligo 1 (with FAM, 100 μM) and 10 μl of Oligo 2 (with Cy5, 100 μM)—in a 0.2 ml PCR tube. The temperature profile starts from heating at 95 °C for 5 min, and the temperature gradually decreases to 4 °C at 0.1°C s⁻¹ and hybridized strands are maintained at 4 °C. The prepared DNA model molecules subsequently bind to the functionalized surface in the dark for 1 h. NEBuffer3 (New England Biolabs), BSA (Sigma, St Louis, MO) and BamHI (30 U μl⁻¹) (New England Biolabs) with nuclease free water (total: 100 μl) were first prepared in a 0.2 ml PCR tube with different concentrations (for BamHI) prior to the digestion of the immobilized DNA targets at 37 °C. After incubating for 1 h, the chip was cleaned with 1× PBS buffer and dried with nitrogen gas, followed by the fluorescence detection of digestion. Similarly, adaptors (20 μl) composed of Oligo 3 (10 μl, 50 nM) (with Cy5) and Oligo 4 (10 μl, 50 nM) were annealed in a 0.2 ml PCR tube (Table 1). Next, varied concentrations of T4 DNA ligase (350 U μl⁻¹) (Fementas) were prepared, containing identical amounts of DNA adaptors that were complementary to the digested strands moieties. The as-prepared mixture was subsequently transported through the corresponding channel to tether the strand moieties of interest by incubating from 5 min to an hour at 37 °C. After that, the chip was washed with buffer and dried with nitrogen gas for fluorescence characterization of ligation.

Table 1 DNA sequences w/o fluorescent labels

Oligonucleotides	Sequences (5′–3′)
Oligo 1	Cy5-AAA AAA GGA TCC CGT ACA (biotin) TCC GCC TTG GCC GT
Oligo 2	FAM-ACG GCC AAG GCG GAT GTA CGG GAT CCT TTT TT
Oligo 3	Cy5-AAA AAA AAA AAA AAG
Oligo 4	GAT CCT TTT TTT TTT TTT T

---

Fluorescence detection and image processing

Following the digestion and ligation of the target DNA fragments of interest that were labeled with fluorescent dyes, the remaining immobilized DNA sequences were rinsed with 1× PBS buffer for 5 min and dried with nitrogen gas. The microchannel surface-bound fluorescent DNA fragments were subsequently imaged using a fluorescence microscope (BX61, Olympus) with an in-built CCD camera under tunable filter cubes. DNA binding fluoresced positive for both Cy5 (red) and FAM (green), while DNA digestion fluoresced positive only for FAM (green). The recombined DNA segments, DNA ligation, also fluoresced positive for both Cy5 (red) and FAM (green). A 10× objective lens was used, and the exposure time for excitation was manipulated using the image acquisition software (Image-Pro AMS 6.0). For better understanding of the methodology, the images were captured at a 1000 ms interval starting from 1000 ms (red) to 6000 ms (red). Depending on the quality of the fluorescence intensity, a higher exposure time or gain might be needed. The captured image data were analyzed by the image acquisition software. A Matlab based algorithm was used to analyze the 8 bit true color image captured by the CCD camera. The program converts the entire image data into a histogram graph for the red and green channels. The intensity of the red channel is the main concern as it determines the efficiency of digestion and ligation methodology.

Results and discussion

DNA recombination with a functionalized microfluidic chip

The ability to individually modify small areas of a surface (i.e., microfluidics based patterning) has revolutionized multiplex immunoassays.^33,34 It is critical to obtain homogeneously functionalized surfaces on the submicron scale for sensitive biochemical analysis.³⁵ Thus, a reliable derivatization strategy for the homogeneous immobilization of DNA probes on small scales is increasingly required. Moreover, the excellent recognition capability between biotin and avidin has been used extensively for DNA immobilization.³⁶ Once biotin or biotinylated conjugate is linked to a desired surface domain, it would be simple to immobilize any biomolecules modified with an avidin label and vice versa. Actually, avidin is a tetrameric protein that has four identical binding sites available for biotin and the binding is almost irreversible with the binding strength comparable to a covalent bond.³⁷ Prior to streptavidin modification, here, the chip surface was silanized by APTES after NH₄OH activation that can generate adequate silanol groups to facilitate subsequent silanization (Fig. 2). Such silanization and amination at the surface serves as a ‘sticky’ substrate that can be further functionalized with a versatile biotin–avidin conjugate to realize the immobilization of any target of interest, only if modified with biotin or avidin.³² Therefore, the biotinylated DNA hybrids with fluorescent dyes (FAM and Cy5) are amenable to complex the as-prepared functionalized interface with streptavidin. The bright red fluorescence proves that the FAM labeled oligonucleotides have successfully tethered onto the functionalized sites (Fig. 3). A chip without a functional surface (Fig. 3, left) results in poor binding affinity, giving rise to a nearly invisible red fluorescence.

Fig. 2 Schematic of stepwise functionalization of silicon substrate of the chip by following DNA digestion and ligation. Silicon substrate surface was first activated by NH₄OH; the generated silanol groups facilitate the silanization and animation of the interface. The resultant NH₂-end can be further biotinylated with biotin conjugated agents, Sulfo–NHS–Biotin, to form a biotinylable site that is amenable to complex streptavidin. The powerful biotin–avidin conjugate, subsequently, binds with biotinylated DNA fragments for the next DNA digestion and ligation via BamHI and DNA T4 ligase, respectively.

Fig. 3 Comparison between functionalized and unfunctionalized substrate surface of the channel by immobilizing biotinylated fluorescent DNA sequences. (a) Untreated substrate surface for the immobilization of biotinylated fluorescent DNA probe. (b) Streptavidin-coated substrate surface for immobilization of biotinylated fluorescent DNA probe. Scale bars, 50 μm.

On-chip DNA digestion

To make a proof-of-concept DNA digestion and ligation on chip, the restriction endonuclease, BamHI, is used and it selectively recognizes the sequence G′GATCC.³⁸ Since the same recognition sequence occurs in both strands of the DNA duplex, the endonucleases can bind to and cleave both strands of the DNA molecule. Thus, the nucleotides at both the ends of these recognition sequences are often complementary to each other. After cleavage, each DNA fragment has a few nucleotides long single-stranded end, namely sticky end. These single-stranded ends can pair with each other again.

With the microfluidic advantages of the minimal sample reagent requirements and a fast enzymatic reaction, we optimized the concentration of BamHI and the enzymatic reaction time to perform highly efficient DNA digestion. By changing the volume of BamHI (30 U μl⁻¹) with nuclease free water (up to 100 μl), the enzyme concentration is varied. The digestion solution was then transferred to and incubated on the chip for 1 h at 37 °C. By comparing the fluorescence intensities of Cy5 between undigested DNA and digested DNA, the efficiency can be calculated. It was concluded that the maximum efficiency was achieved at 60 U of BamHI (Fig. 4a).

Fig. 4 Optimization of (a) the concentration of restriction endonuclease BamHI (inset), and (b) digestion time for hydrolyzation of surface-bound ds-DNA in the microchannel.

By fixing the optimized concentration at 60 U of BamHI, another important parameter, digestion time, was also investigated. By incubating BamHI and the immobilized DNA sequences with time length varied from 5 min to 3 h at 37 °C, it was observed that the highest efficiency appeared at 1 h and then reduced gradually (Fig. 4b). The low digestion efficiency after 1 h incubation may stem from the loss of fluorescence intensity due to photobleaching, and the low efficiency for time less than 30 min can be attributed to the incomplete enzymatic reaction. Actually, the digestion efficiencies for 10 min and 30 min incubation can reach up to 71% and 88%, respectively, of that for 1 h incubation. This means that highly efficient on-chip 10 min digestion can meet the requirements of rapid DNA digestion to some extent.

On-chip DNA ligation

DNA ligases have become an indispensable tool in genetic engineering for generating recombinant DNA sequences.³⁹ After digestion, another complementary fragment (adaptor) can be conjugated with the remnant digested moieties under the assistance of DNA ligase, which can re-form the phosphodiester bonds of DNA strands. Here, T4 DNA ligase was employed on an integrated microfluidic chip to test the DNA ligation efficiency. After the removal of the remnant exonuclease solution, a variety of concentrations of T4 DNA ligase mixed with adequate adaptors that need to be ligated and labeled with Cy5 were pumped into the channel destination and kept for incubation onto the chip for 1 h at 16 °C. We then optimized the concentration of T4 DNA ligase, and the highest efficiency was achieved at 600 U of T4 DNA ligase (Fig. 5a).

Fig. 5 Optimization of (a) the concentration of DNA T4 ligase (inset), and (b) ligation time for splicing of surface-bound ds-DNA in the microchannel.

Similarly, by fixing the optimized concentration of T4 DNA ligase, we characterized another significant parameter, ligation time, to realize rapid on-chip DNA ligation. As shown in Fig. 5b, the incubation time of both 1 h and 2 h resulted in better ligation efficiency. From incubation time of 5 min to 2 h, ligation efficiency increased gradually and began to decrease with ligation time greater than 2 h. After 2 h, the reduced ligation efficiency of T4 DNA ligase may reflect the loss of fluorescence intensity resulting from long time incubation and light influence, and that for time less than 30 min could be due to the incompletion of the enzymatic reaction. Moreover, we found that the ligation efficiency for an incubation time of 10 min and 30 min can reach 63% and 81%, respectively, of the highest ligation efficiency at 2 h. This result means that the vast majority of the enzyme reaction occurs within 30 min and the on-chip incubation of 10 min may meet the requirements of rapid genetic engineering to some extent.

Proof-of-concept for DNA digestion and ligation on chip

In order to develop an integrated microfluidic platform for highly efficient DNA recombination, we made a proof-of-concept for DNA digestion and ligation on chip with screened optimal parameters, such as 60 U of BamHI for 10 min incubation and 600 U of T4 DNA ligase also for 10 min incubation. First, two strands of synthesized oligonucleotides (Oligo 1 and Oligo 2), one labeled with Cy5 and the other with FAM at the opposite ends, respectively, were annealed and transferred to the functionalized chip for immobilization. After 1 h incubation, the biotinylated dsDNA stably tethered on the streptavidin-modified substrate (Fig. 6a). Then, 60 U of BamHI were pumped into the channel to digest the surface-bound fluorescent DNA strands. The undigested DNA sequence was intact, thereby resulting in the detection of both FAM and Cy5 successfully. Red fluorescence from Cy5 was almost invisible in Fig. 6b because Cy5 labeled DNA fragments were digested largely by BamHI, and the strong green fluorescence from FAM was still apparent (Fig. 6b). This signifies that BamHI selectively catalyzes the target DNA. Moreover, the solution of 600 U of T4 DNA ligase containing sufficient Cy5 labeled DNA adaptors with a specific end that is complementary to the surface-bound residual DNA sequences was transferred into the channel for 10 min in order to splice the two DNA fragments together. As demonstrated in Fig. 6c, red fluorescence of high intensity from Cy5 was recovered, indicating that a highly efficient ligation reaction occurs between the two segmented oligonucleotides.

Fig. 6 Sequential dsDNA binding, digestion and ligation. (a) Biotinylated dsDNA molecules labeled with fluorescence dyes (FAM and Cy5) bind with the streptavidin functionalized surface, and the simultaneous imaging of green (FAM) and red fluorescence (Cy5) is realized to confirm the successful immobilization of target DNA sequences. (b) After digestion by BamHI, the fragment labeled with Cy5 is hydrolyzed and removed, resulting in invisible red intensity, but the other moieties (FAM) of the digested segment are still anchored on the original domain, which exhibit bright green fluorescence. (c) In the presence of adaptors (Cy5) and T4 ligase, the remnant fragment can bind with the complementary sequence of adaptors to create a recombinant that also permits the simultaneous imaging of green (FAM) and red fluorescence (Cy5) to demonstrate the successful splice of the two DNA fragments. Scale bars, 50 μm.

Conclusions

Overall, rapid and sequential DNA digestion and ligation has been demonstrated on a microfluidic chip equipped with a temperature controller to enable highly efficient enzymatic activity. Based on the stepwise functionalized chip platform, four significant parameters, including the concentrations of digestion enzyme (BamHI) and T4 DNA ligase and the incubation time of both enzymes, have been optimized. The optimal conditions, such as 60 U of BamHI, 600 U of T4 DNA ligase, and 10 min incubation time, have been employed to verify the feasibility of the accelerated on-chip DNA recombination. Despite the fact that a microscale channel can improve the molecular recognition rates via shortening the diffusion distance, we believe a special microstructure, herringbone structure, would considerably enhance DNA recombination.⁴⁰ Moreover, this platform can lead to a broader prospect of genetic engineering and personal medicine by implementing techniques involving electric field acceleration^41–43 and nanofluidics for chromosome manipulation that are compatible with the microfluidic platform.⁴⁴ We believe such a rapid recombination strategy can serve as a potential tool to address some issues in molecular diagnostics in the future.

Acknowledgements

The authors acknowledge the support of National Natural Science Foundation of China (no. 21275040 and 21305034).

References

1. C. E. Thomas, A. Ehrhardt and M. A. Kay, Nat. Rev. Genet., 2003, 4, 346–358.
2. J. Sambrook, E. F. Fritsch and T. Maniatis, Molecular Cloning: A Laboratory Manual,Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1989.
3. A.-B. Wang, C.-W. Cheng, I. C. Lin, F.-Y. Lu, H.-J. Tsai, C.-C. Lin, C.-H. Yang, P.-T. Pan, C.-C. Kuan, Y.-C. Chen, Y.-W. Lin, C.-N. Chang, Y.-H. Wu, T. Kurniawan, C.-W. Lin, A. M. Wo and L.-C. Chen, Electrophoresis, 2011, 32, 423–430.
4. J. Shendure and E. L. Aiden, Nat. Biotechnol., 2012, 30, 1084–1094.
5. M. Meyer, U. Stenzel and M. Hofreiter, Nat. Protoc., 2008, 3, 267–278.
6. C. Lin and H. Yan, Nat. Nanotechnol., 2009, 4, 211–212.
7. Y. Benenson, T. Paz-Elizur, R. Adar, E. Keinan, Z. Livneh and E. Shapiro, Nature, 2001, 414, 430–434.
8. W. H. Grover and R. A. Mathies, Lab Chip, 2005, 5, 1033–1040.
9. T. Ellis, T. Adie and G. S. Baldwin, Integr. Biol., 2011, 3, 109–118.
10. G. J. Cost, Nat. Protoc., 2007, 2, 2198–2202.
11. D. Mark, S. Haeberle, G. Roth, F. von Stetten and R. Zengerle, Chem. Soc. Rev., 2010, 39, 1153–1182.
12. S. Choi, M. Goryll, L. Sin, P. Wong and J. Chae, Microfluid. Nanofluid., 2011, 10, 231–247.
13. C.-M. Chang, W.-H. Chang, C.-H. Wang, J.-H. Wang, J. D. Mai and G.-B. Lee, Lab Chip, 2013, 13, 1225–1242.
14. M. Noerholm, H. Bruus, M. H. Jakobsen, P. Telleman and N. B. Ramsing, Lab Chip, 2004, 4, 28–37.
15. K.-Y. Lien and G.-B. Lee, Analyst, 2010, 135, 1499–1518.
16. R. Peytavi, F. R. Raymond, D. Gagné, F. J. Picard, G. Jia, J. Zoval, M. Madou, K. Boissinot, M. Boissinot, L. Bissonnette, M. Ouellette and M. G. Bergeron, Clin. Chem., 2005, 51, 1836–1844.
17. C.-H. Lin, Y.-N. Wang and L.-M. Fu, Biomicrofluidics, 2012, 6, 012818.
18. C.-H. Wang, H.-C. Lai, T.-M. Liou, K.-F. Hsu, C.-Y. Chou and G.-B. Lee, Microfluid. Nanofluid., 2013, 15, 575–585.
19. Y.-J. Ko, J.-H. Maeng, Y. Ahn and S. Y. Hwang, Sens. Actuators, B, 2011, 157, 735–741.
20. C.-C. Lee, T. M. Snyder and S. R. Quake, Nucleic Acids Res., 2010, 38, 2514–2521.
21. Y.-H. Chang, G.-B. Lee, F.-C. Huang, Y.-Y. Chen and J.-L. Lin, Biomed. Microdevices, 2006, 8, 215–225.
22. C. Zhang and D. Xing, Chem. Rev., 2010, 110, 4910–4947.
23. F. Shen, B. Sun, J. E. Kreutz, E. K. Davydova, W. Du, P. L. Reddy, L. J. Joseph and R. F. Ismagilov, J. Am. Chem. Soc., 2011, 133, 17705–17712.
24. C. Zhang, H. Wang and D. Xing, Biomed. Microdevices, 2011, 13, 885–897.
25. J. Wu, W. Cao, W. Wen, D. C. Chang and P. Sheng, Biomicrofluidics, 2009, 3, 012005.
26. N. Han, J. H. Shin and K.-H. Han, RSC Adv., 2014, 4, 9160–9165.
27. H. Xie, B. Li, J. Qin, Z. Huang, Y. Zhu and B. Lin, Electrophoresis, 2009, 30, 3514–3518.
28. G. Cost and N. Cozzarelli, Biotechniques, 2007, 42, 86–89.
29. L. Gojová, E. Jansová, M. Külm, S. Pouchlá and L. Kozák, Clin. Genet., 2008, 73, 441–452.
30. L. Wang and P. C. H. Li, Anal. Chim. Acta, 2011, 687, 12–27.
31. L. T.-H. Kao, L. Shankar, T. G. Kang, G. Zhang, G. K. I. Tay, S. R. M. Rafei and C. W. H. Lee, Biosens. Bioelectron., 2011, 26, 2006–2011.
32. B. J. Nehilla, K. C. Popat, T. Q. Vu, S. Chowdhury, R. F. Standaert, D. R. Pepperberg and T. A. Desai, Biotechnol. Bioeng., 2004, 87, 669–674.
33. W. Pan, W. Chen and X. Jiang, Anal. Chem., 2010, 82, 3974–3976.
34. M. Hu, J. Yan, Y. He, H. Lu, L. Weng, S. Song, C. Fan and L. Wang, ACS Nano, 2010, 4, 488–494.
35. I. Wong and C.-M. Ho, Microfluid. Nanofluid., 2009, 7, 291–306.
36. C. Larsson, M. Rodahl and F. Höök, Anal. Chem., 2003, 75, 5080–5087.
37. E. P. Diamandis and T. K. Christopoulos, Clin. Chem., 1991, 37, 625–636.
38. M. McClelland and M. Nelson, Gene, 1988, 74, 169–176.
39. A. J. Doherty and S. W. Suh, Nucleic Acids Res., 2000, 28, 4051–4058.
40. J. Liu, B. A. Williams, R. M. Gwirtz, B. J. Wold and S. Quake, Angew. Chem., Int. Ed., 2006, 45, 3618–3623.
41. C. F. Edman, D. E. Raymond, D. J. Wu, E. Tu, R. G. Sosnowski, W. F. Butler, M. Nerenberg and M. J. Heller, Nucleic Acids Res., 1997, 25, 4907–4914.
42. J. Cheng, E. L. Sheldon, L. Wu, A. Uribe, L. O. Gerrue, J. Carrino, M. J. Heller and J. P. O'Connell, Nat. Biotechnol., 1998, 16, 541–546.
43. M. L. Y. Sin, T. Liu, J. D. Pyne, V. Gau, J. C. Liao and P. K. Wong, Anal. Chem., 2012, 84, 2702–2707.
44. E. T. Lam, A. Hastie, C. Lin, D. Ehrlich, S. K. Das, M. D. Austin, P. Deshpande, H. Cao, N. Nagarajan, M. Xiao and P.-Y. Kwok, Nat. Biotechnol., 2012, 30, 771–776.

---