Concentration and Size Separation of DNA Samples at Liquid–Liquid

LETTER pubs.acs.org/ac

Concentration and Size Separation of DNA Samples at Liquid
Liquid Interfaces Thomas Hahn and Steffen Hardt* Center of Smart Interfaces, TU Darmstadt, Petersenstrasse 32, 64287 Darmstadt, Germany ABSTRACT: This report introduces a new analytical concept utilizing the mass transfer resistance of a liquid
liquid interface to concentrate and separate DNA samples. DNA molecules can be electrophoretically accumulated at a liquid
liquid interface of an aqueous two-phase system (ATPS) of poly(ethylene glycol) (PEG) and dextran, two polymers that form two immiscible phases in aqueous electrolyte solutions. The detachment of DNA from the interface into the other phase can be triggered by increasing the applied electric field. We experimentally study the size dependence of the detachment process for a broad spectrum of DNA fragments. In a regime where the coiling of the chains does not play a significant role, the process shows a linear dependence on the diffusion coefficient, with shorter DNA chains detaching at lower electric field strengths than larger ones. The concept may enable novel separation protocols for preparative and analytical purposes.

O

wing to the far-reaching importance of DNA separation, this process has been intensely studied in microfluidic devices for analytical1 and preparative purposes.2,3 Often separation of biomolecules is achieved in gel-like materials that retard DNA chains or proteins according to their size.4 DNA can be also separated using entropic traps,5 tilted obstacles,6 or dielectrophoretic traps within a microchannel where DNA is hydrodynamically transported through several energy minima.7 Preparative protocols have been reported to isolate specific DNA fragments,2,6,8 also from complex samples.3 Especially in preparative protocols, a number of different steps have to be integrated: DNA preconcentration, separation, and recovery of the sample. This article focuses on a novel method tailored for integrating these different steps. A completely new approach to separate molecular samples is herein presented, utilizing the free-energy profile experienced by molecules at or close to a liquid
liquid interface. On the basis of a quite general theoretical picture, the energy minimum at a liquid
liquid interface depends on the size of molecules or nanoparticles,9 and if the free-energy profile can be modified by external parameters, it should be possible to trigger a size-specific adsorption or detachment. Commonly, researchers focus on organic phase/water interfaces and on ion/molecule/nanoparticle transfer across them by using one10 or multiple species.11 Only in some attempts ion/molecule transfer across liquid
liquid interfaces of aqueous biphasic systems has been considered,12,13 although their biocompatibility makes them a natural choice in separation sciences. These ATPSs are made from two polymer solutions containing, for example, poly(ethylene glycol) (PEG) and dextran that form two immiscible aqueous phases at sufficiently high concentrations, a denser bottom phase rich in dextran r 2011 American Chemical Society

and a top phase rich in PEG termed dextran phase and PEG phase, respectively. While little is known in general about the structure of the interface in ATPSs, they share some properties with what is referred to as immiscible electrolyte solutions (ITIESs), a term usually denoting systems comprising an aqueous and an organic phase. For instance, a Galvani potential of a few millivolts often builds up between a PEG and dextran phase,14 going along with a back-to-back double layer at the interface.15 The interfacial tension in ATPS16 is usually of the order of a few hundred microNewton meter
1, far lower than in organic/aqueous systems.17 Both of these phenomena, the back-to-back double layer and the interfacial tension, contribute to the free-energy profile experienced by a molecule or nanoparticle trapped at the interface. Recently we introduced a microfluidic cell to stabilize a liquid
liquid interface by capillary forces under quiescent conditions.13 Briefly, two electrodes are located in two adjacent buffer reservoirs, one containing a PEG, the other a dextran solution, and an electric field is applied to accumulate DNA at the liquid
liquid interface. The fluidic cell is fabricated in PDMS using soft lithography, bonded to a glass cover slide and silanized by introducing a solution of 0.5% (v/v) 3-(trimethoxysilyl)propyl methacrylate in 50 mM sodium acetate pH 3.9 for 12 h at 20 °C to fix a photopolymerized acrylamide gel within the comb structure (Figure 1).13 The ATPS of specific composition (7.72% (w/w) PEG (MW 6 000), 2.28% (w/w) PEG (MW 35 000) and 12.5% (w/w) dextran (MW 10 000) in 5 mM Tris/borate pH 8.0) is Received: May 14, 2011 Accepted: June 17, 2011 Published: June 17, 2011 5476

dx.doi.org/10.1021/ac201228v | Anal. Chem. 2011, 83, 5476–5479

Analytical Chemistry

Figure 1. Microfluidic cell containing a bilaminated arrangement of immiscible aqueous phases. An applied electric field transports DNA toward the interface.

Figure 2. Accumulation of DNA molecules at the liquid
liquid interface for 60 s, followed by their detachment triggered by a gradually increasing electric field. (a) A YOYO-1 stained 242 bp fragment is accumulated at the interface and detaches at around 150 s when the electric field reaches a critical value (scale bar = 50 μm). (b) The same experiment is performed individually for DNA fragments of different sizes. The different detachment fields become visible through the fluorescence recorded in a detection window close to the interface.

introduced in a two-lamella configuration as sketched in Figure 1. In such an ATPS, the two phases have a similar viscosity of about 2  10
2 Pa s
1. Each phase occupies half of the compartment, and the interface remains fixed even after the volumetric flow is stopped. First, various DNA fragments obtained from an electrophoretic agarose gel separation of a pBR328 mix ladder (298, 653, 1230, 2176 bp) and pUC19 x Msp I digest (190, 242, 331, 401 bp) are introduced together with the dextran phase in individual experiments with a final concentration of 6 pg μL
1. An electrostatic potential difference of 10 V is applied to accumulate DNA at the liquid
liquid interface, and after 60 s of accumulation it is gradually increased by 1 V each 15 s.

LETTER

Figure 3. The detachment field as a function of the translational diffusion coefficient of DNA molecules stained with YOYO-1. The linear curve represents a fit to the data points for DNA fragments between 190 and 2176 bp (R2 = 0.91). The 75 bp fragment (Dt = 1.3  10
8 cm2 s
1) is excluded from the fit because it is additionally labeled with a fluorophore at the 50 end, while the others are not. The inset shows the same quantity as a function of the DNA chain length. The data correspond to an initial DNA concentration of 6 pg μL
1 in the dextran phase.

The fluorescence within the PEG phase is recorded close to the liquid
liquid interface within a narrow window to detect the passage of YOYO-1 stained DNA at sufficiently large electric fields using a fluorescence microscope (Nikon, Ti). At a potential difference of 10 V, corresponding to an electric field strength of 96 V m
1, no molecules pass the interface. As the electric field in the compartment increases with a rate of 9.6 V m
1 each 15 s, the detachment of DNA strands and their migration into the PEG phase is triggered (Figure 2a). For each of the different DNA strands, the maximum fluorescence in the detection window is found at a specific field strength, termed “detachment field” in the following. The time of detachment of a DNA species is unambiguously related to the length of the nucleic acid chain. DNA fragments of less than 500 bp are efficiently separated (Figure 2b). For larger fragments, a comparatively weak separation is observed, and it occurs that 653 bp DNA detaches after the 1230 bp fragment. The detachment field for DNA larger than 653 bp increases only slightly (Figure 3, inset). Remarkably, the detachment field shows roughly a linear decrease with the calculated translational diffusion coefficient18 of the molecules (Figure 3). In this conclusion, we neglect the behavior of the 75 bp fragment which was additionally labeled with a fluorophore at the 50 end, increasing the hydrodynamic radius of the molecule. The calculated translational diffusion coefficient of 1.3  10
8 cm2 s
1 does not take this additional label into account and should be lower in reality. Hence, the observation is consistent with a thermally activated escape from a metastable state being responsible for the detachment of DNA from the interface.13 Shorter fragments with a larger translational diffusion coefficient escape more easily from the energy minimum at the interface. DNA molecules of less than 500 bp have a rodlike conformation, but larger fragments coil up. Owing to this fact, the translational diffusion coefficient changes more rapidly with size for smaller DNA fragments.19 This could explain the decrease in separation efficiency for coiled fragments of more than 653 bp. To simulate a realistic separation protocol and to corroborate the findings from the individual tests, experiments where a mixture of different DNA strands was introduced into the dextran phase are performed. Specifically, fragments have been synthesized 5477

dx.doi.org/10.1021/ac201228v |Anal. Chem. 2011, 83, 5476–5479

Analytical Chemistry

LETTER

Figure 4. Differently labeled DNA fragments, 75 bp-Cy3, 152 bp-fluorescein, and 363 bp-Cy5, are simultaneously introduced into the dextran phase, accumulated at the liquid
liquid interface and separated using a gradually increasing electric field. (a) The molecules at the interface are excited with different wavelengths (488, 543, and 633 nm), and the interface is monitored over time (scale bar =50 μm). (b) The interfacial fluorescence is measured as a function of time to compare the detachment process of the different molecular species. The background fluorescence level is about 0.3
0.4.

using polymerase chain reactions with labeled primers, either Cy3 for 75 bp, fluorescein for 152 bp, and Cy5 for 363 bp DNA. The fragments are simultaneously introduced into the dextran phase and accumulated at the liquid
liquid interface as done in the previous experiments. The fluorescence microscope enables the concurrent imaging of the liquid
liquid interface (Figure 4a) while the sample is excited with a 488 nm, a 543 nm, or a 633 nm laser with a time shift of 1 s controlled by a shutter box. The fluorescence detected at the interface confirms the separation of DNA as expected from the individual experiments (Figure 4b). While the electric field is gradually increased, the 75 bp fragment detaches first, followed by the 152 bp and 363 bp DNA. The electric fields to trigger a detachment are larger in these experiments than in the individual experiments presented above, as the total DNA concentration in the dextran phase is about 1 ng μL
1 to yield a sufficient visualization. Recently, we have shown that the energy minimum DNA fragments experience at the liquid
liquid interface is also dependent on the concentration of these molecules itself, an effect that has to be taken into account when studying the E-field induced detachment.20 In the presented experiments, the 75 bp and 152 bp fragments are sufficiently separated to distinguish them for analytical purposes (Figure 4b). It should be emphasized that a fine adjustment of the parameters, e.g., a slower increase of the electric field strength, can still improve the separation. For long accumulation times, the time-dependent change of the energy barrier at the liquid
liquid must be also taken into account. Accumulation times longer than 300 s cause a spontaneous detachment of DNA.20 Moreover, we remark that in the course of the experiments, molecules migrate from the dextran phase toward the interface, giving a normalized background fluorescence between 0.3 and 0.4. However, all molecules are detached from the interface when the fluorescence levels off (Figure 4b). If desired, a multilaminated arrangement20 with a thinner dextran lamella containing the sample can be chosen to reduce or eliminate the background fluorescence. In the literature, DNA separation has been reported based on various approaches but rarely directly combined with an accumulation process. Concentration and separation of molecules across liquid
liquid interfaces as reported here elegantly combines preconcentration and separation and, furthermore, easily allows recovering the sample from the PEG phase for preparative

purposes. The technique may advance previous procedures to preconcentrate and isolate free fetal DNA for noninvasive prenatal diagnosis3,21 or for other purposes. Drawing a bigger picture, this new technique could open up a toolbox for the concentration and separation of various nanoscopic objects, including molecules as well as nanoparticles, via the detachment from liquid
liquid interfaces.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Fax: +49 6151 16-72021.

’ ACKNOWLEDGMENT We thank Anika Kohlstedt and Darina Riemer for photolithography and Andreas Kurz and Jos
e Pinzon Caballero for technical assistance. Funding by the German Research Foundation (DFG) (SPP 1164, Grant Number HA 2696/4-3) is also gratefully acknowledged. ’ REFERENCES (1) Liu, Y.; Ganser, D.; Schneider, A.; Liu, R.; Grodzinski, P.; Kroutchinina, N. Anal. Chem. 2001, 73, 4196–4201. Smith, E. M.; Xu, H.; Ewing, A. G. Electrophoresis 2001, 22, 363–370. Zhou, H.; Miller, A. W.; Sosic, Z.; Buchholz, B.; Barron, A. E.; Kotler, L.; Karger, B. L. Anal. Chem. 2000, 72, 1045–1052. Tegenfeldt, J. O.; Prinz, C.; Cao, H.; Huang, R. L.; Austin, R. H.; Chou, S. Y.; Cox, E. C.; Sturm, J. C. Anal. Bioanal. Chem. 2004, 378, 1678–1692. (2) Fu, J.; Schoch, R. B.; Stevens, A. L.; Tannenbaum, S. R.; Han, J. Nat. Nanotechnol. 2007, 2, 121–128. (3) Hahn, T.; Drese, K. S.; O’Sullivan, C. K. Clin. Chem. 2009, 55, 2144–2152. (4) Liu, C. L.; Xu, X.; Wang, Q.; Chen, J. J. Chromatogr., A 2007, 1142, 222–230. Stellwagen, N. C.; Stellwagen, E. J. Chromatogr., A 2009, 1216, 1917–1929. (5) Han, J.; Craighead, H. G. Anal. Chem. 2002, 74, 394–401. (6) Huang, L. R.; Cox, E. C.; Austin, R. H.; Sturm, J. C. Anal. Chem. 2003, 75, 6963–6967. (7) Regtmeier, J.; Duong, T. T.; Eichhorn, R.; Anselmetti, D.; Ros, A. Anal. Chem. 2007, 79, 3925–3932. Tuukkanen, S.; Kuzyk, A.; Toppari, J. J.; Hakkinen, H.; Hytonen, V. P.; Niskanen, E.; Rinkio, M.; Torma, P. Nanotechnology 2007, 18, 295204. 5478

dx.doi.org/10.1021/ac201228v |Anal. Chem. 2011, 83, 5476–5479

Analytical Chemistry

LETTER

(8) Lao, A. I.; Hsing, I. M. Lab Chip 2005, 5, 687–90. (9) Flatt
e, M. E.; Kornyshev, A. A.; Urbakh, M. J. Phys.: Condens. Matter 2008, 20, 073102. Bresme, F.; Oettel, M. J. Phys.: Condens. Matter 2007, 19, 1–33. (10) Samec, Z. Chem. Rev. 1988, 88, 617–632. Schweighofer, K. J.; Benjamin, I. J. Phys. Chem. 1995, 99, 9974–9985. Su, B.; Abid, J.-P.; Ferm
in, D. J.; Girault, H. H.; Hoffmannov
a, H.; Krtil, P.; Samec, Z. J. Am. Chem. Soc. 2004, 126, 915–919. Herzog, G.; Roger, A.; Sheehan, D.; Arrigan, D. W. M. Anal. Chem. 2010, 82, 258–264. Kivlehan, F.; Lanyon, Y. H.; Arrigan, D. W. M. Langmuir 2008, 24, 9876–9882. (11) Kakiuchi, T.; Nogushi, J.; Kotani, M.; Senda, M. J. Electroanal. Chem. 1990, 296, 517–535. Osakai, T.; Komatsu, H.; Goto, M. J. Phys.: Condens. Matter 2007, 19, 375103. (12) M€unchow, G.; Hardt, S.; Kutter, J. P.; Drese, K. S. Lab Chip 2007, 7, 98–102. M€unchow, G.; Sch€onfeld, F.; Hardt, S.; Graf, K. Langmuir 2008, 24, 8547–8553. Levine, M. L.; Bier, M. Electrophoresis 1990, 11, 605–611. Strutwolf, J.; Manning, M.; Arrigan, D. W. M. Anal. Chem. 2009, 81, 8373–8379. (13) Hahn, T.; M€unchow, G.; Hardt, S. J. Phys.: Condens. Matter 2011, 23, 184107. (14) Brooks, D. E.; Sharp, K. A.; Bamberger, S.; Tamblyn, C. H.; Seaman, G. V. F.; Walter, H. J. Colloid Interface Sci. 1984, 102, 1–13. Reitherman, R.; Flanagan, S. D.; Barondes, S. H. Biochim. Biophys. Acta 1973, 297, 193–202. Theos, C. W.; Clark, W. M. Appl. Biochem. Biotechnol. 1995, 54, 143–157. (15) Verwey, E. J. W.; Niessen, K. F. Phil. Mag. 1939, 28, 435–446. (16) Helfrich, M. R.; El-Kouedi, M.; Etherton, M. R.; Keating, C. D. Langmuir 2005, 21, 8478–8486. (17) Bier, M.; Zwanikken, J.; van Roij, R. Phys. Rev. Lett. 2008, 101, 046104. dos Santos, A. P.; Diehl, A.; Levin, Y. Langmuir 2010, 26, 10778–10783. (18) Tirado, M. M.; Martin
ez, C. L.; De la Torre, J. G. J. Chem. Phys. 1984, 81, 2047–2052. (19) Stellwagen, E.; Lu, Y.; Stellwagen, N. C. Biochemistry 2003, 42, 11745–11750. (20) Hahn, T.; Hardt, S.; , Size-dependent detachment of DNA molecules from liquid-liquid interfaces. Soft Matter 2011, in press, DOI: 10.1039/c1sm05309d. (21) Hahn, T.; O’Sullivan, C. K.; Drese, K. S. Anal. Chem. 2009, 81, 2904–2911.

5479

dx.doi.org/10.1021/ac201228v |Anal. Chem. 2011, 83, 5476–5479