Anal. Chem. 1995,67, 3219-3222
Dynamics of DNA during Pulsed Field Electrophoresis in Entangled and Dilute Polymer Solutions Xuelong Shi, Richard W. Hammond, and Michael D. Morris* Department of Chemistry, The University of Michigan, Ann Arbor, Michigan 48 109-1055
Using fluorescence video microscopy, DNA electrophoretic behavior under field inversion conditions has been investigated in hydroxyethyl cellulose (HEC) solutions both above and below its entanglement limit. DNA confonnationalfluctuation periods are found to be strongly influenced by the frequency of the applied electric field. DNA maximum extension is found to be dependent on both the frequency and the strength of the applied field. It is proposed that both above and below the HEC entanglement limit, field inversion serves to keep the average DNA conformation in a size-dependent regime intermediate between full extension and random coil. In this time-averaged geometry, efficient long-chain DNA electrophoretic separation is enabled. The introduction of ultradilute polymer solutions for capillary electrophoretic separations of double-stranded nucleic acids by Barron and co-~orkers'-~ has allowed rapid separations of fragments as long as 23 x lo3 base pairs (23 kbp). At any given column length and field strength, separations are 5-10 times faster than possible in conventional entangled polymer solutions. The short times are a direct consequence of the low viscosities of the unentangled polymer solutions. We have shown that field inversion can be used to extend the range of ultradilute polymer solution electrophoresis to nucleic acids as long as 1.6 x lo6 base pairs (1.6 M ~ Pwith ) ~ running times of just over 13 min. Despite these spectacular experimental results, the separation mechanism remains largely c o n j e c t ~ r a l . ~ ~ ~ Recently, we have used video microscopy of fluorescencelabeled long-chain nucleic acids to probe DNA electrophoretic motions in entangled and unentangled (ultradilute) hydroxyethyl cellulose solutions during constant field (dc) nucleic acid electrophore~is.~That work demonstrated that the nucleic acid motions were qualitatively similar in all polymer concentration regimes. Generally, molecules migrated in V- or U-shaped topologies, consistent with a dominant interaction at a single point or region. In concentrated polymer solutions, the interaction region moved only slowly, because it consists of cellulose strands which are part of an extended network. However, below the cellulose entanglement threshold (also called entanglement limit, (1) Barron, A. E.; Soane, D. S.; Blanch, H. W. J. Chromatogr. 1993,652,3-16. (2) Barron, A. E.; Blanch, H. W.; Soane, D. S. Electrophoresis 1994,15, 597615. (3) Barron, A. E.; Sunada, W. M.; Blanch, H. W. Electrophoresis 1995,16,6474. (4) Kim, Y.; Moms, M. D. Anal. Chem. 1995,67, 784-786. (5) Shi, X.;Hammond. R. W.; Moms, M. D. Anal. Chem. 1995,67, 11321138.
0003-2700/95/0367-3219$9.00/0 0 1995 American Chemical Society
C*), the interaction region moved at almost the same velocity as the DNA center of mass, because the cellulose molecules or clusters were not retarded by entanglement with an extended network. At any time, only some of the DNA molecules were in extended topologies in dilute polymer solutions. Most of the remainder were present as random coils which appeared as compact masses under the light microscope. In this article, we describe the extension of our fluorescence microscopy studies to pulsed field electrophoresis conditions in entangled and unentangled hydroxyethyl cellulose solutions. Fluorescence microscopy has previously been used to study nucleic acid motions in field inversion agarose gel electrophoresis6 and in orthogonal field agarose gel electroph~resis.~ We know of no analogous studies in linear polymer solutions, either dilute or concentrated.
EXPERIMENTAL SECTION
The microscopy system has previously been d e ~ c r i b e d . ~ Briefly, it consisted of an Olympus BH-2 epifluorescence microscope equipped with a Reichert 50x/1.0 NA water immersion objective and a Photometrics Star-I scientific CCD camera containing a 576 x 384 pixel chip. The excitation source was 532 nm light from a frequency-doubled CW diode-pumped Nd:YAG laser. The laser power into the microscope was about 12 mW, and the exposure time for each image frame was 0.4 s. Electrophoretic experiments were performed in a locally constructed cell consisting of two polyacrylamide-coated cover slips resting on a microscope slide, to which two electrodes were cemented with a 2.5 cm gap.: The DNA samples were yeast chromosomal DNA (New England Biolabs, PFG Marker, 225 kbp-1.9 Mbp). The staining dye for visualization was ethidium homodimer I (Molecular Probes). The sieving polymer was hydroxyethyl cellulose (HEC) with M,,= 438, 800) and entanglement threshold 0.09% (w/w).l Samples were prepared in 0 . 5 TBE ~ (Ris-borate-EDTA) buffer. To increase viscosity and thereby slow DNA motions, solutions were made 55-60% in s ~ c r o s e . ~ , ~ As in our previous work,: quantitative measurements were made of DNA center of mass velocity (vJ and radii (RIand R,), as defined by Oana et aL9 Physically, RI and R, are the two principal axes of a DNA molecule modeled as an elliptical object. In contrast to the dc field case, a sign has been assigned to RI in (6) Rampino, N. J.; Chrambach, A.Biopolymers 1991,31, 1297-1307. (7) Gumeri S.; Rizzarelli, E.; Beach, D.; Bustamante, C. Biochemistry 1990, 29, 3396-3401. (8) Perkins, T.T.; Smith, D. E.; Chu, S. Science 1994,264,819-822. (9) Oana, H.; Masubuchi, Y.; Matsumoto, M.; Doi, M.; Matsuzawa, Y.; Yoshikawa, K. Macromolecules 1994,27,6061-6067.
Analytical Chemistry, Vol. 67, No. 78,September 75,7995 3219
P
Figure 1. Time sequence of yeast chromosomal DNA motions under pulsed field electrophoresisin 0.16% HEC, 55% sucrose. and 0 . 5 TBE ~ solution. Forward field was 25 Vlcm. and backward field was -20 Vlcm. The frames were taken at 4.0 s intervals. The field was switched every 20 s (fiveframes).Contrast is inverted (fluorescenceregions black) for clarity. The migration direction is indicated by the arrows, and the scale bar represents 10 pm.
analysis of pulsed field data in order to define the direction of DNA orientation. The center of mass and intensity moments of the DNA were calculated using the IPLab image processing program (Signal Analytics Corp.). RESULTS AND DtSCUSSlON
Figure 1shows an example of the electrophoretic behavior of an isolated DNA molecule in a periodically reversed electric field in 0.16%HEC solution (55%sucrose, viscosity 161 cp). In order to maximize the viewing time of a given DNA, the field was chosen to impart only slight net el&ophoretic motion to the nucleic acid while keeping the times in the forward and reverse directions identical. The HEC concentration was chosen to be above the entanglement threshold. When the electric field was applied to the sample, the DNA molecule started to move toward the positive electrode, as during dc electrophoresisin the absence of electroosmosis. As expected, the polyacrylamide coatings on the coverslip surfaces served to largely or completely suppress electroosmosis. The sequence of frames A-E in Figure 1 shows collision of an initially compact DNA molecule with the HEC network and extension of DNA into a characteristic U-shaped configuration. In repeated experiments, the exact conformation of the extended DNA molecule was found to depend on its conformation at the time of collision. Occasionally, the DNA molecule was slightly elongated along the direction normal to the electric field at the collision. When this happened, the DNA molecule became extended into a W-like shape before assuming the characteristic U- or V-shape. Predominantly, the DNA molecule formed a Uor V-shape directly after the collision with the polymer network. Both arms continuously grew during migration until the applied electric field was reversed. The effect of field inversion (direction reversal) is shown in frames F-J of Figure 1. When the direction of the applied field was reversed, the DNAmolecule started to move toward the newly defined anode. In response to the new field direction, both arms of the elongated DNA molecule began to retreat back along their 3220 Analytical Chemishy, Vol. 67, No. 18, September 15, 1995
stretching paths. After the DNA molecule collapsed back to a globular conformation, it would usually move some distance in this cotuiguration before e n d i n g into the U- or V-shape, with arms pointing toward the anode. This process was repeated when the field direction was again reversed, as shown in frames K-0 of Figure 1. Although there was alternation in the orientation of the extended DNA, the exact shapes were not repeated on successive cycles. Consistent with our earlier observations under dc electro phoresis conditions? the apex of a U-sbape moved in 0.16%HEC solutions under pulsed electric field. The movement direction of the apex was always toward the anode and reversed immediately with elechic field d m t i o n . This observation further confirmed that the movement of the apex was due to partial rupture of the polymer network caused by the force the DNA molecule exerted on the network. We investigated the cause of collapse of an extended DNA into a globular form upon electric field reversal. In principle, the cause could be either the reversed elechic field force or the elastic restoring (entropic) force. Experimentally, collapse due to an elastic restoring force alone can be examined by simply turning off the field instead of reversing the field. Figure 2 shows the DNA entropic collapse (relaxation) in 0.16%HEC solution under field-free conditions. Entropic collapse required more than 50 s. By contrast, collapse to globular form required only about 9 s under field inversion conditions. While an elastic restoring force may contribute to the collapse of DNA, the process is primarily fielddriven. Rapid collapse requireda large velocity dBerence between the apex and the ends of the arms and demonstrated the large difference in the resistance forces experienced by the apex and the arm ends. The arm ends were expected to experience very little resistance during collapse because they retreated hack along their stretching paths. The apex of the molecule experienced a much larger resistance force betause of its direct interaction with the polymer network.
0
20
40
60
BO
Time (sec) Figure 4. Time evolution of 9 of a yeast chromosomal DNA. Electric field was cycled at 1/24 Hz. Other conditions as in Figure 3.
c
A
0
10
20
30
40
zE o5
50
2! .P
Time (sec) Figure 2. Relaxation of a yeast chromosomal DNA molecule in TBE solution in the absence of 0.16% HEC, 55% sucrose, and 0 . 5 ~ electric field. The points on the plot correspond to the sequence of DNA images. 3 1 .o
2 $
E 2!
.P
cr,
1.5
0.5
4
0.0
f
0.0
k - 5
9 g W
-0.5
-10 I
I
0
20
40
Time (sec)
8 0
8 0
Figure 5. Time evolution of 9 (0)and Rs (*) of a yeast chromosomal DNA. The field was cycled between 8 and -6.4 V/cm. Other conditions as in Figure 3.
X
1
8
-1
0.5
E
l.ow
- 2
a
20
40
Time (sec)
A
3 0-I
60
8 0
4
K
204
The ultimate extension of DNA was a strong function of both field strength and frequency. In Figure 4, the behavior of RI at higher frequency, 1/24 Hz, is shown. On the average, the maximum extension is less than that at 1/40 Hz (Figure 3B). In Figure 5, we show the behavior of Ri and Rsat 1/40 Hz, but at lower field strength than in Figure 3B. Again, the maximum extension is less than that at higher field. Together these data demonstrate the strong effects of both field strength and field frequency on the average extension of DNA. Field inversion frequency is an important variable not present in dc electrophoresis. In the dc case, the periodicity of DNA conformational fluctuations has been shown to depend on the field strength and the separation medium for a given size DNAg It has long been known that in the limit of complete extension, frequently called strongly biased reptation, DNA mobility becomes size-independent.loJ The optimum geometry for size-dependent mobility is intermediate between a random coil and an extended chain. In the loosely entangled polymer solutions, field inversion serves to keep the average DNA conformation somewhere near the optimum. This optimization occurs if the field inversion takes place in less than the dc field entanglement/disentanglement time. In 0.01% HEC polymer solution (60% sucrose, viscosity -42 cp), which is about 1 order of magnitude lower than the 0.09% entanglement threshold, U-or V-shaped DNA molecules were still observed, but less than 10%of the nucleic acids were extended at any given time (Figure 6). The maximum extension was much smaller than that observed in 0.16% HEC under the same field conditions. Similarly, reversal of the orientation of the U- or V-shape was observed when the field direction was changed (frames F-J of Figure 6). These effects were caused by interac-
-20b\l-
-30
4 0
0
2 0
80
80
Time (sec)
Figure 3. Time evolution of v,, 9, and Rs of a yeast chromosomal DNA under field inversion in 0.16% HEC, 55% sucrose, and 0 . 5 ~ TBE solution. The square waves in the figures are the electric field values. The field was cycled at 1/40 Hz between 25 and -20 V/cm. (A) Response of v, (0)to the pulsed field. (B) Response of RI (0) and Rs (*) to the pulsed field.
Figure 3 shows the behavior of v,, RI, and R, during 1/40 Hz field reversal cycles in 0.16%HEC. Within the time resolution of our measurements, the center of mass velocity of a DNA molecule responded instantaneously to the change in field direction, but the DNA conformationalresponse to the electric field change was retarded. At a field strength of 20-25 V/cm, a DNA molecule required about 9 s to reorient. Most of the reorientation is along the field axis, as shown by the large change in RI. Less conformational change occurred perpendicular to the field axis, as demonstrated by much smaller fluctuations in R,.
(10) Lumpkin, 0. J.; Dejardin, P.: Zimm, B. H.Biopolymers 1985, 24, 15731593. (11) Slater, G. W.; Noolandi, J. Biopolymers 1986,25, 431-454.
Analytical Chemistry, Vol. 67, No. 18, September 15, 1995
3221
t-
B.
"I
I
. G 3 ,'.'.
c
I 0
1 0
4'0
so
Time (see)
Flgum 6. Time sequence of yeast chromosomal DNA motions under pulsed field electrophoresisin 0.01% HEC, 60% sucrose, and 0 . 5 ~TEE solution. Other mndiiions as in Figure 1.
tions between DNA and HEC molecules. No U-or V-shape with appreciable size could be observed in 0% HEC solutions. At HEC concentrations well below the entanglement limit the largest fraction of DNA molecules were in almost globular conformations. These globular conformationshad small fluctuations and were diflicult to distinguish from the DNA conformations in HEC-free solution. In the absence of HEC, DNA molecules assumed largely globular conformationsduring migration and had small conformational fluctuations too. Such small random conformational fluctuationswere observed even in the absence of an applied electric field. They may originate from segmental random motions of DNA. Qualitatively, the behavior of an observed Ushaped DNA molecule in 0.01%HEC solution was similar to that in 0.16%HEC solution, as suggested in Figure 7. In both cases, measurements were made on molecules of apparent radius of gyration of -2.0 pm. As in the case of loosely engtangled HEC (Figure 3). the center of mass velocity response to the elechic field was instantaneous within the time resolution of our instruments. Uncorrected for solvent (sucrose solution) viscosity,the response of RIto the elechic field, a lag of -15 s, was more retarded than that in entangled HEC solution. However, if this number is corrected for the ratio of sucrose solution viscosities (22 CPat 55%, 40 CPat @ the I lag% time is) reduced , to about 8.3 s. That number is similar to the RI lag time in 0.16%HEC, 55%sucrose, about 9 s. The smaller size of the Ushaped DNA suggests that the difference in resistance forces experienced by the apex and the ends of a U-shaped DNA molecule is smaller in 0.01% HEC solution than in 0.16%HEC solution. This behavior is consistent with the picture of ultradilute HEC as k l y floating obstades and 0.16% HEC as a loosely entangled but extended network As in 0.16%HEC polymer solution. the periods of the conformational fluctuations of the U- or V-shaped DNA were also strongly iduenced by the field reversal intervals of the applied electric field. The low magnitudes of R, in 0.01%HEC solution prevented
3222 AnaWcal Chemistry. Vol. 67, No. 18, September 15, 1995
0
1 0
4.0
Time (see)
so
8 0
ngum 7. Time evolution of &% and F4 of a yeast "nosomal DNA under field inversion in 0.01% HEC, 60% sucrose, end 0 . 5 ~TEE
solution. Other mnditions as in Figure 3.
us from making detailed examinations of field strength and field frequency effectson DNA conformational evolution during field reversal cycles. CONCLUSIONS Our observations underscore the basic qualitative similarity of DNA electrophoreticmigration under field inversion conditions in entangled and unentangled polymer solutions. The differrnces are quantitative rather than qualitative. We propose that the pulsed field confers sizedependent mobility only on those DNA molecules which are actualiy entangled with HEC. Subentangle ment separations of long-chain DNA can be successful because there is a sizedependent average mobility which depends upon the fraction of the time that DNA is entangled with HEC and on its average conformation during entanglement Good separation conditions can be maintained if the field reversal interval is less than the lifetine of the transient DNMHEC adduct The optimum separation conditions for field inversion electre phoresis in u1h;ldilute solution are not yet known. Further protocol exploration is certainly needed. But this work must be complemented by more detailed imaging studies and by theore& cal studies which de!ine the optimum geomehy for sizedependent mobility and the conditions for maximizing its presence.
We thank the National Institutesof Health for financial support through Grant GM-37006 to M.D.M. Received for review April 20, 1995. Accepted June 30, 1995." AC950389M AbsW plbli'shed in Aduonrr ACS Abs,h&, August 1,1995.
