DNA Capillary Electrophoresis in Entangled Dynamic Polymers of

Anal. Chem. 2001, 73, 1776-1783

DNA Capillary Electrophoresis in Entangled Dynamic Polymers of Surfactant Molecules Wei Wei and Edward S. Yeung*

Ames LaboratorysUSDOE and Department of Chemistry, Iowa State University, Ames, Iowa 50011

Aqueous solutions of monomeric nonionic surfactants, n-alkyl polyoxyethylene ethers (C16E6, C16E8, C14E6), can be used as sieving matrixes for the separation of DNA fragments by capillary electrophoresis. Unlike ordinary polymer solutions, these surfactant solutions behave as dynamic polymers. By combining the “reversible gel” theory of DNA electrophoresis and the static and dynamic properties of wormlike surfactant micelles, a model is developed for describing the migration behavior of DNA molecules in these solutions. According to the model, the separation limit can be extended at low surfactant concentrations. Surfactant solutions as a separation medium provide many advantages over ordinary polymers, such as ease of preparation, solution homogeneity, stable structure, low viscosity, and self-coating property for reducing electroosmotic flow. More importantly, the properties of wormlike micelles (micelle size, entanglement concentration) can be adjusted by simply changing the monomer concentration, denaturant, and temperature to allow the separation of different size ranges of DNA fragments. Fast separation is achieved for DNA fragments ranging from 10 bp to 5 kb by using bare fused-silica columns. DNA sequencing fragments of BigDye G-labeled M13 up to 600 bases were separated within 60 min. Capillary electrophoresis (CE) for DNA analysis has been studied extensively, both experimentally and theoretically, since 1988.1-22 The most popular and effective method for DNA (1) Cohen, A. S.; Najarian, D.; Smith, J. A.; Karger, B. L. J. Chromatogr. 1988, 458, 323-333. (2) Cohen, A. S.; Najarian, D. R.; Karger, B. L. J. Chromatogr. 1990, 516, 4960. (3) Ruiz-Martinez, M. C.; Berka, J.; Belenkii, A.; Foret, F.; Miller, A. W.; Karger, B. L. Anal. Chem. 1993, 65, 2851-2858. (4) Zhang, J.; Fang, Y.; Hou, J. Y.; Ren, H. J.; Jiang, R.; Roos, R.; Dovichi, N. Anal. Chem. 1995, 67, 4589-4593. (5) Madabhushi, R. S. Electrophoresis 1998, 19, 224-230. (6) Lindberg, P.; Righetti, P. G.; Gelfi, C.; Roeraade, J. Electrophoresis 1997, 18, 2909-2914. (7) Bashkin, J.; Marsh, M.; Barker, D.; Johnston, R. Appl. Theor. Electrophor. 1996, 6, 23-28. (8) Fung, E. N.; Yeung, E. S. Anal. Chem. 1995, 67, 1913-1919. (9) Chang, H.-T.; Yeung, E. S. J. Chromatogr., B 1995, 669, 113-123. (10) Wei, W.; Yeung, E. S. J. Chromatogr., A 2000, 745, 221-230. (11) Menchen, S. M.; Johnson, B.; Winnik, M. A.; Xu, B. Electrophoresis 1996, 6, 23-28. (12) Gao, Q.; Yeung, E. S. Anal. Chem. 1998, 70, 1382-1388. (13) Wu, C.; Liu, T.; Chu, B. Macromolecules 1997, 30, 4574-4583. (14) Slater, G. W.; Kist, T. B. L.; Ren, H.; Drouin, G. Electrophoresis 1998, 19, 1525-1541.

1776 Analytical Chemistry, Vol. 73, No. 8, April 15, 2001

separations by CE is capillary gel electrophoresis (CGE) due to the similar mobility in free solution for DNA fragments with sizes above tens of base pairs (bp). Thus, most studies have focused on developing different sieving matrixes. Early on, modeling after traditional slab gel electrophoresis, cross-linked polyacrylamide was first introduced to CE for DNA separation by Cohen2 and others.23 Since then, many hydrophilic polymers have been used successfully for DNA analysis, such as linear polyacrylamide,3,4 polydimethylamide,5 N-acrylaminoethoxyethanol,6 cellulose derivatives,7 poly(ethylene glycol),11 poly(ethylene oxide),8-10 poly(vinylpyrrolidone),12 and PEO-PPO-PEO triblock copolymers (Pluronics).13 These polymer matrixes have advantages over crosslinked polymers due to their replaceable property. However, because the polymers are not small molecules, they are not always stable due to degradation with time, environment, and mechanical shearing during preparation.24,25 The separation performance deteriorates with degradation due to the loss of long-chain molecules. Nonionic polymeric surfactants have been used successfully as a separation medium for DNA analysis.13 These block copolymers have a hydrophobic core of PO blocks and a strongly hydrated shell of EO blocks that can form micelles in solution. The aggregation number is determined by the length of the PO block. With increasing temperature, desolvation of the EO groups continues and the effective volume fraction decreases. These globular micelles overlap and entangle each other at high concentrations. Finally, lyotropic liquid crystals are formed13,26,27 that are useful as a sieving medium. Many monomeric nonionic surfactants have the same properties as the above polymeric surfactants, especially the family of (15) Ogston, A. G. Trans. Faraday Soc. 1958, 54, 1754-1757. (16) De Gennes, P. J. J. Chem. Phys. 1971, 55, 572-579. (17) Duke, T.; Viovy, J. L. Phys. Rev. E 1994, 49, 2408-2416. (18) Grossman, P. D.; Soane, D. S. Biopolymers 1991, 31, 1221-1228. (19) Duke, T.; Viovy, J. L.; Sememov, A. N. Biopolymers 1994, 34, 239-247. (20) Viovy, J. L.; Duke, T. Electrophoresis 1993, 14, 322-329. (21) Slater, G. W.; Mayer, P.; Hubert, S. J.; Drouin, G. Appl. Theor. Electron. 1994, 4, 71-79. (22) Slater, G. W.; Mayer, P.; Drouin, G. Electrophoresis 1993, 14, 961-966. (23) Kasper, T. J.; Melera, M.; Gozel, P.; Brownlee, R. G. J. Chromatogr. 1988, 458, 303-312. (24) Minoura, Y. J.; Kasuya, T.; Kawamure, S.; Nakano, A. K. J. Polym. Sci., Part A-2 1967, 5, 125-142. (25) Molyneux, P. Water-Soluble Synthetic Polymers: Properties and Behavior; CRC Press: Boca Raton, FL, 1983. (26) Becher, P., Ed. Nonionic Surfactants: Physical Chemistry; Marcel Dekker: New York, 1967. (27) Schick, M. J., Ed. Nonionic Surfactants: Physical Chemistry; Marcel Dekker: New York, 1986. 10.1021/ac0012997 CCC: $20.00

© 2001 American Chemical Society Published on Web 03/13/2001

n-alkyl polyoxyethylene ethers.26-31 Surfactants of the n-alkyl polyoxyethylene ether family consist of a hydrophobic core of alkyl chains surrounded by hydrophilic polyoxyethylene chains. Due to the relatively long alkyl chains compared to the EO chains, micelles undergo one-dimensional growth by balancing the intermolecular forces. The micellar structures change from sphere to rodlike with such one-dimensional growth.26 Finally, a giant wormlike micelle forms with huge aggregation numbers. For example, the aggregation number of C16E6 surfactant at room temperature is 10 500 and the apparent micellar molecular weight is ∼5 100 000.26,29-31 These wormlike or rodlike micelles become sufficiently long and flexible and act as a dynamic polymer solution32-40 although they are not chemically linked as in traditional linear polymers. As with other long-chain polymers, it should be possible to use these dynamic polymers of surfactants as a separation medium for DNA analysis. Surfactants have been used to separate DNA before.1,25 However, those are spherical micelles that rely on partitioning based on micellar electrokinetic chromatography25 or adsorption of the monomers along the DNA chain.1 In this work, we will demonstrate separation of both dsDNA fragments and Sanger sequencing ladders in dynamic polymer solutions. The theoretical consideration for DNA electrophoresis in these solutions will also be presented. THEORETICAL MODEL To understand how surfactant polymer solutions effect DNA separation, we recall the separation mechanism of DNA electrophoresis in traditional polymers. Many theoretical models have been proposed to describe DNA separation in sieving matrixes, including Ogston sieving model,15 reptation,16 biased reptation model (BRM),14 constrained release (CR) of entangled polymers,19-22 and “reversible gel” model.17 Here, we invoke the reversible gel model developed by Duke and Viovy.17 In this model, they assumed that the topological constraints on DNA motion in entangled long-chain polymer solutions are perpetually changing as intermolecular bonds break and form or as the polymers diffuse. The mobility of DNA molecules in a temporary gel µ is,17

µ/µ0 ∼ wτ ∼ (c/c*)-15/4ξ/b for wτ < 1

(1)

where µ0 is the mobility of DNA in free solution, ξ and b are the screening length of the polymer and the Kuhn length of DNA, and c and c* are the solution concentration and entanglement limit, (28) Tanford, C.; Nozaki, Y.; Rohde, M. F. J. Phys. Chem. 1977, 81, 1555-1560. (29) Balmbra, R. R.; Clunie, J. S.; Corkill, J. M.; Goodman, J. F. Trans. Faraday Soc. 1964, 60, 979-985. (30) Cummins, P. G.; Staples, E. Langmuir 1989, 5, 1195-1199. (31) Lin, Z.; Scriven, L. E.; Davis, H. T. Langmuir 1992, 8, 2200-2205. (32) Cates, M. E.; Candau, S. J. J. Phys.: Condens. Matter 1990, 2, 6869-6892. (33) Turner, M. S.; Marques, C.; Cates, M. E. Langmuir 1993, 9, 695-701. (34) Cates, M. E. J. Phys.: Condens. Matter 1996, 8, 9167-9176. (35) Kato, T.; Terao, T.; Tsukada, M.; Seimiya, T. J. Phys. Chem. 1993, 97, 39103917. (36) Carale, T.; Blankschtein, D. J. Phys. Chem. 1992, 96, 459-467. (37) Yamakawa, H. Modern Theory of Polymer Solutions; Harper and Row: New York, 1971. (38) Cantor, C. R.; Schimmel, P. R. Biophysical Chemistry; W. H. Freeman and Co.: New York, 1980. (39) Cates, M. E. Macromolecules 1987, 20, 2289-2296. (40) Groswasser, A. B.; Wachtel, E.; Talmom, Y. Langmuir 2000, 16, 41314140.

respectively. The separation limit (useful upper limit for size separation) is

N/ ∼ (wτ)-1

(2)

where w is the rupture frequency of the cross linking in the temporary gel, such that

w ) 1/τrep

(3)

where τrep is the reptation time of the polymer. This model predicates that size fractionation can be extended to higher molecular weight compared to a permanent gel with equivalent pore size. Static and Dynamic Properties of Entangled Surfactant Micelles. In a micellar system, the reversibility of the selfassembly process ensures that the molecular weight distribution of the wormlike or rodlike polymeric species is in thermal equilibrium, in contrast to traditional long-chain polymer solutions.32-34 The static properties of nonionic surfactants, such as osmotic compressibility, show behaviors similar to those observed in semidilute solutions of long-chain polymers. For wormlike or rodlike micelles, the aggregation number (n) is concentration dependent and can be written as35

n ∝ cR

(c > cmc)

(4)

where cmc is the critical micellar concentration. The value of R depends on the model for association equilibrium and is typically equal to 0.5. The contour length of n-monomers, Ln, can be described as,36

Ln )

nvc πlc

2

+

2lc + 2lh 3

(5)

where vc and lc are the volume and minor radius of the micelle core of the surfactant hydrophobic moiety, respectively, and lh is the effective length of the surfactant hydrophilic moiety. The radius of gyration of n-monomers is given as,36

〈Rg2〉 )

[

( )]

Lnξm 2ξm3 2ξm4 Ln 1 - exp - ξm2 + 2 3 Ln ξm L n

(6)

For ξm . Ln

〈Rg2〉 ≈ Ln2/12

(6a)

〈Rg2〉 ≈ Lnξm/3

(6b)

and for ξm , Ln

where ξm is the persistence length of the micelle (i.e., the finite diameter of the micelle) which results from steric interactions between the hydrophilic moiety at the micelle core/water interface. When the surfactant concentration is above the crossover Analytical Chemistry, Vol. 73, No. 8, April 15, 2001

1777

Figure 1. Electropherograms of DNA fragments at different surfactant concentrations. (a) 1-kb ladder, 0.1% C16E6; (b) 5-kb ladder, 0.1% C16E6; (c) 1-kb ladder, 0.25% C16E6; (d) 5-kb ladder, 0.25% C16E6; and (e) 1-kb ladder, 0.5% C16E6. Conditions: Leff ) 60 cm, E ) 250 V/cm, 1 µM TO in 100 mM HEPES-TEA buffer (pH 7.0), room temperature, and other conditions described in the text.

point from dilute to semidilute solution regimes, the monomers entangle each other and form a transient network of overlapping micelles.32,35,36 According to scaling law, the corrected length of the micelles (ξb), which is similar to the gel pore size,35 is 1778

Analytical Chemistry, Vol. 73, No. 8, April 15, 2001

ξb ) (R/g/x3)(c/c*)v/(1-3v)

(7)

where c* is crossover concentration, R/g is the radius of gyration at the crossover concentration, and v is a constant in the range of 0.5-0.588.

From eqs 4-7, the static properties (effective molecular weight, radius of gyration, etc.) of the transient network of overlapping micelles are determined by surfactant concentration and physical structure (hydrophilic and hydrophobic moieties), which are therefore different from those governing traditional long-chain polymers. The dynamic properties of self-assembled micelles (viscoelasticity, self-diffusion, etc.) are also distinguishable from entangled long-chain polymers.32,34,35 The flexible wormlike micelles behave like dynamic polymers, whose chains are subject to reversible breakage and formation. The dynamic properties of such polymers in the entangled state can be described by a modified reptation model, in which the scission and recombination reactions of the chains are introduced. The lifetime of a chain with mean length (Lh ) before breaking into two pieces (τb) is defined as,32, 39

τb ) 1/kL h

(8)

where the mean length (Lh ) can be expressed as,32,

40

L h = φ1/2 exp(Ec/2kBT)

(9)

Figure 2. Reptation plots of 6-FAM-labeled 100-bp DNA ladder at various surfactant concentrations. Conditions: Leff ) 30 cm, 100 mM HEPES-TEA buffer (pH 7.0) with 6 M urea, and E ) 250 V/cm.

Separation Limit in Dynamic Polymers. By combining eqs 2-3 and 8-12, we obtain the mobility of DNA molecules in dynamic polymers, 1/2

and Ec, the end-cap energy (in units of kBT), is the difference in the free energy of adding surfactant molecules to the wormlike core versus adding molecules to the two spherical end caps of the micelle. Ec is independent of concentration and is linearly dependent on temperature.40 φ is the total volume fraction of the surfactant defined as

φ)

∑L c(L ) ∝ ∑L n

n

n exp(-Ln/L)

(11)

where τrep is relaxation time of chains disentangled from a tubelike environment, i.e., the reptation time.

h 2/Dc τrep = L

(12)

and Dc is the collective diffusion constant, which is related to the hydrodynamic correlation length (ξH),

Dc ) kBT/6πηsξH ∝ c

x

(13)

ξH scales like ξb, in semidilute regime, and decreases with the increase of surfactant concentration. x is a constant. In the low concentration regime, x ) -5/3,32 and in the high concentrated regime, x ) 2/3.35,41 Thus, Dc first decreases with increasing concentration and then increases. (41) Kato, T.; Terao, T.; Seimiya, T. Langmuir 1994, 10, 4468-4474.

(14)

and the separation limit, 1/2

N/ ∼ L3/2/Dc

(15)

(10)

where c(Ln) is the number density of chains of length Ln. When the lifetime is long (τb . τrep), the dynamic polymer is like traditional unbroken polymers. However, for short lifetimes (τb , τrep), the stress relaxation time scale (τ) is given by

τ ) (τrepτb)1/2

Dc µ ∼ 3/2 µ0 L h

The present results are different from traditional long-chain polymer solutions. To achieve long reads, a large Lh a small Dc are preferred. Lh , from eq 9, is determined by the temperature and the total volume fraction (surfactant concentration). Higher temperatures and concentrations give rise to large Lh .29,32,40 However, Dc shows a different concentration dependence. In the low concentration regime, Dc decreases with increasing concentration (eq 13). Thus, long reads should be obtained by increasing the surfactant concentration. At the high concentration regime, Lh increases with concentration as φ1/2 while Dc also increases with concentration as φ2/3. Thus, N* ∼ φ5/12 according to eq 15 and increasing the concentration is worse for long reads. In addition, we note that Mw ∼ Lh . At a given set of conditions, we prefer to use surfactants with a large aggregation number (large Mw) to achieve a large Lh . MATERIALS AND METHODS Instrumentation. The CE instrument with laser-induced fluorescence (LIF) detection, built in-house, has been described in previous work.10 Briefly, ∼10 mW of 488-nm laser light from an argon ion laser (model Innova 90, Coherent, Palo Alto, CA) was used for excitation. The laser beam was focused on the detection region of the capillary by a 1-cm focal length lens at an angle of 90° to the laser beam. A 530-nm long-pass filter was used to eliminate the scattered light before imaging onto the photomultiplier tube (PMT). The fluorescence signal from the PMT was transferred directly through a 10-kΩ resistor to a 24-bit A/D Analytical Chemistry, Vol. 73, No. 8, April 15, 2001

1779

Figure 3. Electropherograms of φX174/HaeIII digest fragments: (a) 1% C16E6; (b) 3% C16E6. All other conditions are the same as in Figure 1.

converter (Lawson Labs, Kalispell, MT) and stored in a 486/33 computer at 4 Hz. Fused-silica capillaries with 50-µm i.d. and 363-µm o.d. were purchased from Polymicro Technologies (Phoenix, AZ). The separation capillary was enclosed in a 0.5-cm-i.d. copper heating jacket. The jacket was connected to a water bath circulator (Fisher Scientific). A model HH23 microprocessor thermometer was directly connected to the outside of the copper tube so that the separation temperature can be directly read with a precision of 0.1 °C. Chemicals and Materials. All chemicals were obtained from Sigma (St. Louis, MO). The buffer for ds-DNA fragments analysis is composed of 100 mM combined HEPES and triethylamine (TEA) at pH 7.0. For DNA sequencing, the buffer consists of 75 mM 3-[[tris(hydroxymethyl)methyl]propanesulfonic acid (TAPS), 75 mM histidine, 50 mM tris(hydroxymethyl)aminomethane (Tris), and 2 mM EDTA with 7 M urea. 10-bp, 25-bp, 1-kb, and 5-kb DNA ladders were purchased from Life Technologies (Frederick, MD). 6-FAM-labeled 100-bp size standard was from Transgenomics (Omaha, NE). The intercalated dyes for DNA labeling (1:5 dye/DNA), Thiazole orange (TO) and SYBR Gold nucleic acid stain, were from Molecular Probes (Eugene, OR). M13(-21) DNA samples were prepared at the Nucleic Acid Facility (Iowa State University, Ames, IA) by using cycle sequencing, BigDye-primer, Ampli Taq FS polymerase and standard Applied Biosystems reagents. The DNA samples were denatured 1780 Analytical Chemistry, Vol. 73, No. 8, April 15, 2001

by heating in a denaturing solution [1:1 (v/v) formamide/saturated urea] at 95 °C for 3 min and then put onto ice for 3 min. RESULTS AND DISCUSSION In capillary electrophoresis, DNA usually migrates against electroosmotic flow (EOF). Coated capillaries or buffer additives are often used for reducing EOF.2,3,42-44 Nonionic surfactants, such as Brij 35 (C12E23), have been employed for decreasing EOF in CE.42-44 Ellipsometry studies also confirmed that nonionic surfactants are easily absorbed onto the polar or hydrophilic silica/ water interface to form a constant-thickness layer.45 In fact, because the monomers are small compared to other DNA sieving matrixes, the coating is expected to be more uniform and more easy to put on and to wash off. Thus, a bare capillary is used in this work. n-Alkyl polyoxyethylene ethers (C16E6, C16E8) have low cmc and large aggregation numbers.26,27,29-31 For instance, the cmc of C16E6 is 1 µM26 and the apparent micelle length for 1% C16E6 at 28 °C is 340 nm, which is comparable to the length of 1-kb ds-DNA. When the temperature increases to 38 °C, the micelle length reaches 700 nm.30,31 The size of the micelles can be adjusted by changing both the surfactant concentration and the temperature.29,40 Moreover, the micelle size is determined by the structure of the surfactants. At the same concentration and temperature, the aggregation number of C16E6 is larger than that of C16E8.

Figure 4. Electropherograms of 10-bp DNA ladder at different temperatures: (a) 25, (b) 71. and (c) 74 °C. Conditions: Leff ) 50 cm, 7% C16E8 with 1:10 000 diluted SYBR Gold nucleic acid stain, 100 mM HEPES-TEA buffer (pH 7.0) with 3 M urea, and E ) 200 V/cm.

Therefore, the desirable dynamic polymer size can be easily controlled. Effect of Surfactant Concentration on Separation. The surfactant concentration is a very critical factor for separation performance because it affects both the static and the dynamic properties of entangled wormlike aggregates. Eventually, it determines the separation limit. As discussed above, in the lowconcentration regime, increasing the concentration is favorable for separating large DNAs. Figure 1 shows the concentration dependence on the separation limit. In Figure 1a and b, the resolution for the 1- and 5-kb DNA ladders is poor at 0.1% of C16E6. When the concentration increases to 0.25%, both 1- and 5-kb ladders can be resolved well (see Figure 1c and d). According to eq 15, when the concentration increases in the low-concentration regime, Lh increases and Dc decreases such that the separation limit is extended. Therefore, at these conditions, 0.25% separates the low from the high-concentration regimes. However, when the surfactant concentration is increased further (up to 0.5%), for the 1-kb ladder, we did not observe improved resolution, as shown in Figure 1e. On the other hand, the resolution for large DNA is decreased, which results from the increase of Dc with concentra-

tion. In Figure 1c-e, the larger fragments show asymmetric peaks characteristic of mismatch between their mobilities and those of the buffer ions usually observed only in zone electrophoresis.46 We plotted the reptation curves at several different surfactant concentrations in Figure 2. The mobilities were independent of dye/DNA ratio up to 1:5 and field strength up to 250 V/cm. The transition from the Ogston regime to the reptation regime (loss of linearity) was less evident at 250 V/cm for the low surfactant concentrations. Once again, at low concentrations, the separation limit is extended. As an example, Figure 3 shows the results for the separation of φX174/HaeIII digest fragments at two different monomer concentrations. The spurious peaks in Figure 3a were due to random noise spikes in the system. The loss of resolution for large DNA and improved separation for small DNA with an increase in the surfactant concentration is obvious. Effect of Temperature on Separation. n-Alkyl polyoxyethylene ethers usually have low cloud points (Tc). When the temperature reaches Tc, micelles could not be formed.26 However, Tc can be adjusted by adding urea.47 C16E8 is chosen here due to the relatively high cloud point so that a wide temperature range can be employed. Figure 4 shows electropherograms for the

(42) Towns, J. K.; Regnier, F. E. Anal. Chem. 1991, 63, 1126-1132. (43) Salmanowicz, B. P. Chromatographia 1995, 41, 99-106. (44) Durkin, D.; Foley, J. P. Electrophoresis 2000, 21, 1997-2009. (45) Tiberg, F.; Jonsson, B.; Lindman, B. Langmuir 1994, 10, 3714-3722.

(46) Mikkers, F. E. P.; Everaerts, F. M.; Verheggen, T. P. E. M. J. Chromatogr. 1979, 169, 11-20. (47) Briganti, G.; Puvvada, S.; Blankschtein, D. J. Phys. Chem. 1991, 95, 89898995.

Analytical Chemistry, Vol. 73, No. 8, April 15, 2001

1781

Figure 5. Raw electropherograms of BigDye primer G-labeled M13(-21) sequencing sample. Conditions: Leff ) 65 cm, E ) 150 V/cm, 10% C16E6 in 75 mM TAPS-75 mM histidine-50 mM Tris-2 mM EDTA (pH 8.2) with 7 M urea at 34 °C. The samples were injected at a constant electric field of 150 V/cm for 30 s.

separation of 10-bp DNA ladders at different temperatures. At room temperature, the resolution is poor. When the temperature rises to 54 °C, improved resolution is observed (data not shown). The best separation was achieved at 71 °C (Figure 4b). When the temperature increases further to 74 °C, the resolution is lost again, as shown in Figure 4c. This phenomenon can be explained by the various micellar structures at different temperatures. At low temperatures, the aggregation number (hence the micellar molecular weight) for C16E8 is too small to form an effective network even when the concentration is much higher than the cmc. With increasing temperature, the micelles become larger and larger until they entangle one another. Sieving separation is thus possible. When the temperature reaches its cloud point, phase separation occurs. As a result, the surfactant solution loses its ability to separate DNA fragments because of the absence of dynamic long chains of micelles in solution. DNA Sequencing in a Surfactant Solution. From the discussion above, it should be possible to perform DNA sequencing in these surfactant solutions. According to our previous work,10 Tris-TAPS-His-EDTA buffer with 7 M urea gives improved performance for DNA sequencing in a PEO matrix. Figure 5, as an example, is the electropherogram of BigDye primer G-labeled only M13(-21) Sanger fragments. Figure 5 shows that singlebase resolution of 0.5 (the minimum required for DNA sequencing) up to 600 bp is obtained within 60 min at 34 °C. Further improvements in the maximum number of bases read may be possible through additional optimization of the separation condi1782

Analytical Chemistry, Vol. 73, No. 8, April 15, 2001

tions. This is because strain on the polymer network is rapidly dissipated by equilibration with the monomer units, a feature that is analogous to the favorable separation of large DNA fragments at higher temperatures.48,49 Work in this area and toward mutation detection50 is underway in our laboratory. CONCLUSIONS We introduce the use of small surfactant molecules as the sieving medium for DNA capillary electrophoresis. The equilibrium solution of self-assembled surfactant micelles results in dynamic long polymer chains that are different from traditional high molecular weight polymers. The separation limit follows the scaling laws for aggregation and entanglement. By adjusting the surfactant concentration, separation temperature, and buffer additives, both ds-DNA fragments and ss-DNA sequencing ladders have been resolved successfully. Since these are dynamic polymers, they do not degrade like traditional linear polymers. Also, they are expected to be easily washed out of capillary columns because they disentangle when diluted. The surfactant solution here was replaced after each run. No deterioration in performance was observed over many runs in a one-month period. The monomers are self-coating on the capillary walls so that bare fusedsilica columns can be used. The temperature dependence of the (48) Salas-Solano, O.; Carrilho, E.; Kotler, L.; Miller, A. W.; Goetzinger, W.; Sosic, Z.; Karger, B. L. Anal. Chem. 1998, 70, 3996-4003. (49) Zhou, H.; Miller, A. W.; Sosic, Z.; Buchholz, B.; Barron, A. E.; Kotler, L.; Karger, B. L. Anal. Chem. 2000, 72, 1045-1052. (50) Gao, Q.; Yeung, E. S. Anal. Chem. 2000, 72, 2499-2506.

aggregation number may provide unique opportunities for imposing a size-selective gradient during electrophoresis.

Biological and Environmental Research, and by the National Institutes of Health.

ACKNOWLEDGMENT The Ames Laboratory is operated for the U.S. Department of Energy by Iowa State University under Contract W-7405-Eng-82. This work was supported by the Director of Science, Office of

Received for review December 15, 2000.

November

3,

2000.

Accepted

AC0012997

Analytical Chemistry, Vol. 73, No. 8, April 15, 2001

1783