Anal. Chem. 1999, 71, 3598-3602
Sample Stacking during Membrane-Mediated Loading in Automated DNA Sequencing Andra´s Guttman
Genetic BioSystems, Inc., San Diego, California 92121
Microporous membrane-mediated loading is a novel and efficient sample injection technique for ultrathin slab gel electrophoresis-based automated DNA sequence analysis. The sequencing reaction mixture is spotted directly onto the tabs of the membrane loader, which is then inserted to close proximity of the straight edge of the separation gel. The use of a higher viscosity (>60 cSt), low ionic strength (conductivity 800 µS). Also, the pH of the well solution (7.0) is more than two pH units lower than that of the pK of the running buffer co-ions (borate, pK ) 9.3). Thus, when the electric field is applied and the running buffer co-ions (borate) enter the lower pH well solution zone, their degree of dissociation significantly declines. This causes severe electrophoretic mobility drop of the co-ions, despite the higher local field strength of this zone (Ew). On the other hand, the DNA molecules are still fully ionized at pH 7.0 (pKDNA ∼ 2); thus when they enter the well solution zone, the higher local potential drop results in very fast migration of the sample components toward the separation gel interface. This high local field strength tends to increase the sharpness of the sample zone (see eq 2, increase of E). Also, please note that the viscosity of the well solution was optimized (63 cSt) not to significantly effect the mobility of the DNA fragments, but ensuring higher stacking efficiency by reducing diffusion-induced band broadening (see eq 2, decrease D). At the interface of the separation gel, the DNA molecules stack up against the higher conductivity, κ3 ) 800 µS (lower electric field strength, E3), sieving matrix, forming the required sharp starting zones. During this stacking process, pH is the primary determinative factor mediating the actual mobilities of the borate co-ions in the different sections, forming the leading and terminating zones. When the borate ions leave the lower pH well solution 3600 Analytical Chemistry, Vol. 71, No. 16, August 15, 1999
regime and enter into the higher pH separation gel-buffer system (pH 8.3), their dissociation rate and concomitant electrophoretic mobility increases. As a result of that, these ions rapidly pass through the separating DNA fragments and proceed as a moving boundary through the separation gel matrix, leaving a continuous buffer system behind. Figure 2A exhibits a typical DNA sequencing separation using the conventional shark’s tooth loading method. As one can see, good base-calling accuracy was obtained up to almost 400 bases (99.7%). As the shark’s tooth loading column depicts in Table 1A, peak efficiencies (N) corresponding to arbitrary fragment sizes of 109, 211, and 305 (striped peaks in Figure 2A) vary from 93 000 to 578 000 theoretical plates per meter.15 The resolution values (Rs) between consecutive peaks in various arbitrary chain length domains of 112/113 (G-G), 205/206 (T-T), and 303/304 (T-T) (shaded doublets in Figure 2A) are shown in Table 1B, second column. The highest resolution is obtained between the 205- and 206-mer pair (Rs ) 0.59), probably because the separation gel composition provided maximum separation selectivity (R) for this regime (see eq 1). Figure 2B depicts the separation of the same sequencing reaction mixture using a membrane-mediated loading technique. Peaks of this electropherogram in general are apparently sharper compared to the peaks in the separation obtained with the regular loading method (Figure 2A). The membrane loading column in (15) Foley, J. P.; Dorsey, J. G. Anal. Chem. 1983, 55, 730-737.
Figure 2. Electropherograms of the BigDye-labeled cycle sequencing fragments prepared with AmpliTaq DNA polymerase FS and purified by ethanol precipitation. (A) Regular shark’s tooth sample injection method. (B) Membrane-mediated sample loading using low ionic strength, higher viscosity, pH 7.0 well solution. Conditions: running buffer 1× TBE; separation gel Long Ranger Singel Pack (36 cm separation length); voltage 1680 V (constant); separation temperature 52 °C; sample volume 1.0 µL.
Table 1A shows the numerical values of these higher separation efficiencies for the same size fragments (109, 211, and 305) ranging from 175 000 to 716 000 theoretical plates per meter (striped peaks in Figure 2B). The resolution values (Table 1B,
third column) almost doubled in all three fragment regimes, with Rs ) 1.09 being the maximum again for the 205- and 206-mer pair (shaded doublets in Figure 2B). Please also note, that basecalling accuracy was 100% in this instance. Analytical Chemistry, Vol. 71, No. 16, August 15, 1999
3601
Table 1. Comparison of Peak Efficiency (A) and Resolution (B) for Regular Shark’s Tooth and Membrane-Mediated Sample Loading Using Low Ionic Strength, Higher Viscosity, pH 7.0 Well Solution, in Automated DNA Sequencing fragment size 109 211 305
fragments 112/113 205/206 303/304
A. Theoretical Plate Numbers (N) shark’s tooth loading membrane loading 93 000 278 000 578 000
175 000 565 000 716 000
B. Resolution (Rs) shark’s tooth loading
membrane loading
0.41 0.59 0.27
0.97 1.09 0.68
CONCLUSIONS We have previously reported the ease of use of membranemediated sample loading in gel electrophoresis analysis of DNA fragments. In this paper, we have demonstrated that, in addition
3602 Analytical Chemistry, Vol. 71, No. 16, August 15, 1999
to that, the utilization of a low ionic strength (conductivity 60 cSt), pH 7.0 well solution can greatly enhance the separation efficiency of DNA sequencing fragments during automated DNA sequencing. We assume that, due to the higher peak efficiencies, better resolution can be obtained, resulting in easier base calling with higher confidence and accuracy. Also, in high-throughput sequencing mode, robots can readily accommodate automated sample spotting onto the loading membranes that can be bar-coded for later identification. ACKNOWLEDGMENT The author gratefully acknowledges the help of Dr. Stephan Johnston and Dr. Jing Tan for providing the DNA sequencing expertise. The stimulating discussion with Dr. Gyula Vigh is also greatly appreciated.
Received for review February 15, 1999. Accepted May 5, 1999. AC990173A
