294 Electrophoresis 2001, 22, 294±299

Zsolt Ronai1,2 DNA analysis on electrophoretic microchips:

Csaba Barta1,2
Maria Sasvari-Szekely2 Effect of operational variables
Andrµs Guttman1
Applicability of modern microfabrication technology to electrophoresis microchips initi-
1
Novartis Agricultural ated a rapidly moving interdisciplinary field in analytical chemistry. Electric field medi-
Discovery Institute, ated separations in microfabricated devices (electrophoresis microchips) are signifi-
La Jolla, CA, USA cantly faster than conventional gel electrophoresis, usually completed in seconds to
2
Institute of Medical Chemistry, minutes. Electrophoretic separation of DNA molecules on microfabricated devices
Molecular Biology and
proved to have the potential to improve the throughput of analysis by orders of magni-
Pathobiochemistry,
Semmelweis University, tude. The flexibility of electrophoresis microchips allows the use of a plethora of sepa-
Budapest, Hungary ration matrices and conditions. In this paper, we report on electric field mediated sepa-
ration of fluorescent intercalator-labeled dsDNA fragments in polyvinylpyrrolidone
matrix-filled microchannel structures. The separations were detected in real time by a
confocal, single-point laser-induced fluorescence/photomultiplier setup. Effects of the
sieving matrix concentration (Ferguson plot), migration characteristics (reptation plot),
separation temperature (Arrhenius plot), as well as applied electric field strength and
intercalator concentration on the separation of DNA fragments are thoroughly dis-
cussed.

Keywords: Microchip / Electrophoresis / DNA fragments / Operational variables EL 4264

1 Introduction Successful incorporation of a PCR amplification chamber

with consequent DNA analysis on a glass microchip was
Microfabrication technology using planar glass substrates also recently reported [6].
enhances the performance of electric field mediated sep-
aration devices and leads to novel integrated microchip-
New advances in electric field-mediated separation of
based analytical tools [1, 2]. Well-established separation
dsDNA fragments evolved from recent exploration of
protocols can easily be transferred from capillary electro-
novel separation matrices. In spite of the early demon-
phoresis (CE) to electrophoretic microchips. Effenhauser
strations on the use of high-concentration linear polymer
et al. [3] were the first to describe the use of a microchan-
solutions in planar separation formats almost two deca-
nel system, filled with 10% noncross-linked polyacryl-
des ago [7, 8], noncross-linked polymers were just
amide gel, for the size separation of phosphorothioate
recently applied to novel separation formats, such as CE
oligonucleotides (10±25-mers). Applying electric field
[9]. Linear polymers, such as noncross-linked polyacryl-
strengths as high as 2300 V/cm and using a 38 mm sepa-
amide matrix introduced by Karger and co-workers [10],
ration distance, good resolution of all the components
derivatized celluloses thoroughly investigated by Barron
was obtained in less than 45 s. Separation of double-
et al. [11], polyethylene oxides [12] and polyvinylpyrroli-
stranded (ds) DNA molecules on CE chips was first dem-
done (PVP) recently reported by Yeung©s group [13]
onstrated by Woolley and Mathies [4]. They used derivat-
shown to be effective in capillary electrophoretic size sep-
ized cellulose solutions as sieving matrix for the analysis
aration of DNA molecules. The use of noncross-linked,
of FX174 HaeIII restriction digest mixture and attained
linear polymers for DNA fragment analysis proved advan-
good separation of all 11 fragments within 120 s in a
tageous in several respects. Noncross-linked polymers
35 mm channel length. Microchip-based electrophoresis
are supplied in a desiccated, dry form, providing long
analysis of DNA fragments was also reported in conjunc-
shelf life and good reproducibility [14]. In addition, lower
tion with on-the-chip restriction digestion using TOTO-1
viscosity noncross-linked polymer solutions are easily
fluorophore dye in the separation matrix for labeling [5].
replaced even within capillary formats [15], therefore, sim-
ilar to narrow bore capillary columns, microfabricated
Correspondence: Dr. Andràs Guttman, Novartis Agricultural channels supporting repetitive work can be readily filled
Discovery Institute, La Jolla, CA 92121, USA
with these matrices.
E-mail: andras.guttman@syngenta.com
Fax: +858-812-1097
Most of the important operational variables in gel electro-
Abbreviations: CGE, capillary gel electrophoresis; TBE, Tris- phoresis based analysis of DNA fragments, such as siev-
borate-EDTA buffer; UTLGE, ultrathin-layer gel electrophoresis ing matrix concentration, applied electric field strength,

 WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001 0173-0835/01/0202-294 $17.50+.50/0

Electrophoresis 2001, 22, 294±299 Microchip gel electrophoresis of DNA 295

separation temperature, etc., have been well character- 2.2 Detection/separation setup
ized in conventional slab-gel electrophoresis [16, 17],
capillary gel electrophoresis (CGE) [18, 19] and ultrathin- A common confocal, single-point laser-induced fluores-
layer gel electrophoresis (UTLGE) [20, 21]. Yet, the cence (LIF) detection system [28] was applied to the elec-
migration properties of dsDNA molecules are somewhat trophoresis microchip (Eclipse TE200, Nikon, Melville,
different in these various separation formats. This is NY, USA). The beam of the 532 nm frequency doubled
mainly caused by the significant differences both in NdYAG laser (15 mW, B&W TEK, Newark, DE, USA)
dimensions (regular vs. capillary) and in applied electric was projected through a dichroic beam-splitter into a high
field strength, the latter being orders of magnitude larger numerical aperture microscope objective (10 ´ 0.45). The
in CGE and UTLGE. The high separation field strength objective focused the laser beam in the center of the sep-
may alter the migration behavior of DNA fragments aration microchannel of the electrophoresis chip. The
through the sieving medium [22]. In conventional slab gel emitted fluorescent light from the fluorophore-labeled
electrophoresis using field strengths of several V/cm, sep- analyte molecules was collected and collimated by the
aration of DNA fragments smaller than 1000 base pairs same microscope objective and passed back through the
are well described by the Ogston model [23]. In capillary dichroic beam-splitter. The beam-splitter reflected the line
dimensions (CGE and UTLGE), due to the considerably of the laser light and passed through the longer wave-
higher applied electric fields, DNA fragments tend to length emission beam. An achromatic lens was used to
migrate by reptation, or in extreme cases by biased repta- focus the emission beam into a spatial pinhole filter. In
tion [24]. On the other hand, it is well-known, that high this way, only the fluorescence light emitted from the focal
field strengths increase Joule heat development within area of the electrophoresis microchip passed through.
the separation platform, that needs to be efficiently dissi- Thus, any scattered light possibly coming from the micro-
pated [25]. Using electrophoretic microchips, fields of chip surface and fluorescent light other than which origi-
thousands of V/cm can be applied because of their favor- nated from the separation channel were filtered out. The
able heat dissipation properties due to the increased light passing the pinhole was filtered again through a 585
aspect ratio (width to height) of the separation channels  25 nm band-pass filter (Omega Optical, Battleboro, NJ,
[26]. In this paper, we report the effects of operational var- USA) and directed into a photomultiplier tube (PR1; Prod-
iables such as sieving matrix concentration, separation ucts for Research, Danvers, MA, USA). The output signal
temperature, applied electric field strength, and intercala- of the photomultiplier tube was preamplified by a model
tor dye concentration on electrophoretic migration charac- SR570 low noise preamplifier (Stanford Research Sys-
teristics in microfabricated electrophoresis chips, using a tems, Sunnyvale, CA, USA), digitized using a PCI-6711
standard dsDNA ladder ranging up to a thousand base board (National Instruments, Austin, TX, USA) and ac-
pairs. quired by a PC for subsequent signal processing with
Caesar Workstation 7.0 software (CE Solutions, Mission
Viejo, CA, USA). Gold-plated beryllium electrodes were
2 Materials and methods

Miniaturization
used to provide electrical contact between the in-house
fabricated four-channel high-voltage power supply and
2.1 Chemicals the buffer/sample/waste reservoirs of the electrophoresis
Tris-base, boric acid and EDTA´Na2 were obtained from microchip (Fig. 1A). A computer program written in Lab-
Sigma Chemical (St. Louis, MO, USA), all in electro- View 4.1 (National Instruments) was used to automati-
phoresis grade. The 100 base pair (bp) DNA ladder (Life cally time and switch the appropriate voltages to the res-
Technologies, Gaithersburg, MD, USA) was diluted with ervoirs.
double deionized water (18 MW) to the working concen-
trations of 0.25±25 ng/mL and stored at ±20oC until use.
2.3 Electrophoresis microchip
PVP (Mr 1 300 000) (Aldrich, Milwaukee, WI, USA) was
dissolved in 1 ´ TBE buffer (89 mM Tris, 89 mM borate, The standard, low fluorescence borofloat glass, semicir-
2 mM EDTA´Na2, pH 8.3). Bodipy FL Hydrazide was used cular cross-section channel (~ 2.0 mm ´ 50 mm) electro-
as neutral marker for electroosmotic flow (EOF) measure- phoresis microchip was from AMC (Edmonton, Alberta,
ment (Molecular Probes, Eugene, OR, USA). The interca- Canada), (Fig. 1A). Injector design: 100 mm double T
lator dye Sytox Orange (ABSmax: 547 nm; EMmax: 570 nm (approximate volume: 100 pL). Total channel length was
[27]; from Molecular Probes) was added to the sample for 85 mm (80 mm to the injection cross); access holes were
noncovalent fluorophore labeling in 0.1±1.0 mM final con- 2 mm in diameter. Effective separation length was
centrations. All buffer and sample solutions were filtered 30 mm, unless specified otherwise. The separation plat-
through a 0.2 mm nylon membrane syringe filter (Fisher form was first flushed with 2% PVP (Mr 1 300 000) dis-
Scientific, Pittsburgh, PA, USA). solved in 1 ´ TBE buffer, then filled with the appropriate
296 Z. Ronai et al. Electrophoresis 2001, 22, 294±299

concentration polymer solution (0.5±2.5%) for the size

separation of the DNA fragments. This provided a simple
and reliable dynamic polymer coating, allowing easy
regeneration of the microchannel network as well as sta-
ble and reproducible separation means for multiple injec-
tion/separation cycles. Sample injection into the electro-
phoresis microchip was accomplished by first applying
480 V at the sample outlet reservoir (2) and grounding the
sample inlet reservoir (4) for 45 s. Reservoir positions are
depicted in Fig. 1A. A voltage of 240 V was applied at
both of the running buffer reservoirs (1, 3) during the
injection step. For the analysis the following voltages
were applied at the reservoirs: (1) = 0 V; (2) = 250 V;
(3) = 1650 V; (4) = 250 V (Fig. 1B). When the separation
voltage was changed for the field strength evaluation
study, the same voltage ratio of 6.6 was maintained be-
tween reservoirs (3)/(2) and (3)/(1), while reservoir (1)
was still kept at 0 V. The temperature of the electrophore-
sis microchip was regulated by a thermostated air bath
with a precision of  1oC. The actual separation tempera-
ture was measured at the top plate of the microchip by a
thermocouple. Figure 1. (A) Reservoir assignment of the electrophore-
sis microchip, (B) applied injection and separation vol-
tages and (C) a typical, rapid separation of the 100 bp siz-
3 Results and discussion ing ladder. Numbers on peaks correspond to the size of
DNA fragments in base pairs. Conditions: sieving matrix,
3.1 Effect of sieving matrix concentration and 2% PVP (Mr 1 300 000) solution in 1 ´ TBE buffer; effec-
DNA fragment size tive separation length, 30 mm; applied electric field
strength, 200 V/cm; room temperature; sample, 100 bp
In order to investigate the sieving capabilities of the sepa-
dsDNA ladder labeled with Sytox Orange (0.5 mM final
ration matrix, various concentration PVP solutions were
concentration); injection, 100 mm double-T (35 fg DNA/
filled into the narrow channels of the electrophoresis injection).
microchip (0.5±2.5% by the increments of 0.5%). The
100 bp dsDNA ladder ranging up to 1000 bp was used as
model test mixture in the experiments (see typical opti-
mized condition separation in Fig. 1C). Figure 2 exhibits
the nonlinear concave curves obtained when the natural
logarithm of the electrophoretic mobilities were plotted as
the function of sieving matrix concentration (Ferguson
plot [29]). Every data point represents the average of five
parallel measurements with less than 1% deviation. The
upward curvature of the plots in Fig. 2 is probably due to
possible conformational changes of the diluted sieving
polymer solution. First Chrambach and co-workers [17]
and later Karger©s group [10] reported similar concave
Ferguson plots in linear polyacrylamide gel based DNA
separations on slab and capillary gel electrophoresis, re-
spectively. It is interesting to note that it is apparently in
contrast to earlier observations of Holmes and Stellwagen
[30] in conventional agarose gel electrophoresis of DNA
fragments and those of Szoke et al. [21] in UTLGE using Figure 2. Ferguson plots of the separation of dsDNA
composite agarose-linear polymer matrices. Also note, fragments on the electrophoresis microchip. Numbers on
that during our studies the applied electric field strength plots correspond to the size of DNA fragments. Separa-
was orders of magnitude higher (200 V/cm) than in con- tion conditions were the same as in Fig. 1. The sieving
ventional slab-gel electrophoresis (1±4 V/cm) or in previ- matrix concentration ranged from 0.5 to 2.5%.
Electrophoresis 2001, 22, 294±299 Microchip gel electrophoresis of DNA 297

ously reported UTLGE (40±60 V/cm). As it was earlier individual plots were found to be decreasing with higher
suggested by Gao et al. [31] for narrow bore capillary col- gel concentrations. Figure 3B shows an exponential
umns, electroosmotic flow in the microchannels was decay of these slope values as the function of the sieving
minimized by the sieving PVP solution itself. The electro- matrix concentration. The actual values (ranging from
osmotic flow was measured according to Lengyel and ±0.07 to ±0.165) were significantly different from negative
Guttman [32] by using the neutral marker Bodipy FL unity (±1) that would be typical of reptation [24]. This sug-
hydrazide. gested that the separation of dsDNA fragments in the size
range of 100±1000 bp can be well characterized by the
Figure 3A exhibits the natural logarithms of the relative Ogston sieving model [23], even at higher polymer con-
electrophoretic mobility values of the dsDNA fragments at centrations of 2±2.5%. Based on the exponential decay of
various polymer concentrations as the function of the nat- the plot in Fig. 3B, at least a 25% PVP solution would be
ural logarithm of dsDNA fragment lengths (ln L). Every required to reach the reptation regime (slope = ±1) under
data point represents the average of five parallel meas- similar separation conditions (i.e., electric field strength,
urements with less than 1% deviation. The relative mobili- temperature, etc.).
ty values were calculated as m/m0, where the free solution
mobility of the solute was defined from the common Y-
axis interception of the extrapolated plots in Fig. 2 (m0 =
4.42 ´ 10±4 cm2/Vs). The average slope values of the

Figure 4. (A) Relationship between the logarithmic mo-

bility of the DNA fragments and the reciprocal absolute
temperature (Arrhenius plot) of the electrophoresis micro-
Figure 3. (A) Double logarithmic plot of the relative elec- chip separation. Numbers on plots correspond to the size
trophoretic mobility versus the chain length of the dsDNA of DNA fragments. Separation conditions were the same
molecules in the sample (reptation plot). Separation con- as in Fig. 1; temperatures: 20, 25, 30, 35, 40, 45, and
ditions were the same as in Fig. 3. (B) Slope values of the 50oC. (B) Activation energy (Ea) of the sieving polymer
plots in (A) as the function of sieving matrix concentration. solution as the function of DNA fragment length.
298 Z. Ronai et al. Electrophoresis 2001, 22, 294±299

3.2 Effect of separation temperature high field densities might lead to field-dependent mobili-
ties of DNA fragments in slab-gel electrophoresis [16, 24,
The consequence of separation temperature on the elec-
37]. Comparable behavior was found more recently in
trophoretic mobility of dsDNA fragments was investigated
CGE separation of dsDNA fragments [10] exhibiting con-
at 20, 25, 30, 35, 40, 45, and 50oC, respectively. Arrhe-
cave Ferguson plots similar to those shown in Fig. 2.
nius plots, defined as natural logarithms of the mobilities
These effects were considered to arise from the orienta-
of the various size DNA fragments ranging from 200 to
tion and stretching of the coiled configuration of the
1000 bp, versus the reciprocal absolute temperature are
dsDNA molecules by the high applied field [38]. Figure 6
shown in Fig. 4A. Every data point represents the aver-
depicts an increasing relationship between the apparent
age of five parallel measurements with less than 0.2%
electrophoretic mobility of the DNA fragments and the
deviation. Because the viscosity of the sieving polymer
applied electric field strength in microchip gel electro-
solution changes with temperature, the observed electro-
phoresis. Every data point represents the average of five
phoretic mobilities were adjusted accordingly (1.1% / oK)
[33]. As Fig. 4A suggests, electrophoretic mobilities of the
DNA fragments increased with elevating temperature.
The negative slope values of the plots in Fig. 4A were
multiplied by the universal gas constant to obtain the re-
spective activation energy values (Ea). Then the resultant
activation energy values were plotted as function of the
fragment length of the analyte molecules in order to esti-
mate the amount of activation energy necessary to estab-
lish the appropriate sieving structure for the various size
solute molecules. As Fig. 4B depicts, Ea values for elec-
trophoresis microchip separation of dsDNA fragments
using PVP solution exhibited a slightly elevating tendency
with increasing solute size (23±23.25 kJ/mol), suggesting
that electric field mediated migration of the larger dsDNA
fragments require a somewhat higher activation energy to
warp through the dynamically structured polymer net-
work. Similar increasing plots were observed earlier in Figure 5. Effect of the staining dye concentration in the
CGE of ssDNA sequencing fragments in cross-linked sample on the electrophoretic mobility of the dsDNA frag-
ments. The staining dye concentration in the sample
polyacrylamide gels (19±24 kJ/mol) [34] and in UTLGE of
ranged from 0.1 to 1 mM. All other separation conditions
similar size dsDNA fragments using composite hydroxy-
were the same as in Fig. 1.
ethyl cellulose-agarose (22±24 kJ/mol) and polyethylene
oxide-agarose gels (18±20 kJ/mol) [35].

3.3 Effects of staining dye concentration and

the applied electric field strength
Figure 5 delineates the electrophoretic mobility values as
the function of staining dye concentration ranging from
0.1 to 1 mM. Every data point represents the average of
five parallel measurements with less than 0.25% devia-
tion. The apparent slight mobility decrease with increas-
ing dye concentration is probably due to the positive
charge of the intercalator dye that decreases the overall
charge-to-mass ratio of the DNA-dye complexes. This fig-
ure shows that the use of 0.5 mM complexing dye in the
sample is favorable providing efficient labeling of the DNA
fragments. This dye was reported earlier to possess no
significant self-quenching characteristics [36]. Figure 6. Relationship between the electrophoretic mo-
bility and the applied electric field strength in microchip
Electrophoretic mobility of biopolymers during gel electro- electrophoresis of dsDNA fragments. Separation condi-
phoresis should in general be independent of the strength tions were the same as in Fig. 1, except the applied vol-
of the applied electric field. However, as reported earlier, tages were 600, 1237, 1650, 2475, 2973, and 3717 V.
Electrophoresis 2001, 22, 294±299 Microchip gel electrophoresis of DNA 299

parallel measurements with less than 0.2% deviation. In [6] Khandurina, J., McKnight, T. E., Jacobson, S. C., Waters,
agreement with earlier observations, a changing electric L. C., Foote, R. S., Ramsey, J. M., Anal. Chem. 2000, 72,
2995±3000.
field may result in altered electrophoretic mobility of DNA [7] Bode, H. J., in: Radola, B. J. (Ed.), Electrophoresis ©79,
fragments [10, 22]. The temperature of the electrophore- Walter de Gruyter & Co, New York, NY 1980, pp. 39±52.
sis microchip was kept at 25oC during the field strength [8] Tietz, D., Chrambach, A., Electrophoresis 1992, 13,
evaluation experiments. It was also reported that the per- 286±294.
[9] Guttman, A., in: Landers, J. P. (Ed.), Handbook of Capillary
sistent length of dsDNA fragments increases with interca- Electrophoresis, CRC Press, Boca Raton, FL 1994.
lation, due to the increase of the intrinsic viscosity of the [10] Heiger, D. N., Cohen, A. S., Karger, B. L., J. Chromatogr. A
DNA-intercalator complex [39]. This enlarged length and 1990, 516, 33±48.
stiffness of the fragments may also be responsible for the [11] Barron, A. E., Sunada, W. M., Blanch, H. W., Electrophore-
sis 1996, 17, 744±757.
apparent changes in electrophoretic migration properties.
[12] Chang, H. T., Yeung, E. S., J. Chromatogr. B 1995, 669,
113±123.
4 Concluding remarks [13] Gao, Q., Yeung, E. S., Anal. Chem. 1998, 70, 1382±1388.
[14] Schwatz, H., Guttman, A., Separation of DNA by Capillary
Application of microfabricated devices to electric field Electrophoresis, Beckman Instruments, Fullerton, CA 1995
mediated separation of DNA fragments proved to have [15] Guttman, A., US Patent 5,370,777, 1994.
[16] Stellwagen, N. C., Biochemistry 1983, 22, 6186±6193.
the potential to further modern separation science. In this
[17] Tietz, D., Gottlieb, M. H., Fawcett, J. S., Chrambach, A.,
paper, we have studied the effects of gel concentration, Electrophoresis 1986, 7, 217±222.
separation temperature, applied electric field strength, [18] Guttman, A., Cooke, N., J. Chromatogr. 1991, 559,
and staining dye concentration during electrophoresis 285±294.
[19] Fang, Y., Zhang, J. Z., Hou, J. Y., Lu, H., Dovichi, N. J.,
microchip analysis of dsDNA fragments. During the evalu-
Electrophoresis 1996, 17, 1436±1442.
ation of the sieving capability of PVP solution, we ob- [20] Ewing, A. G., Gavin, P. F., Hietpas, P. B., Bullard, K. M.,
served concave Ferguson plots similar to those observed Nature Med. 1997, 3, 97±99.
earlier in conventional polyacrylamide gel electrophoresis [21] Szoke, M., Sasvari-Szekely, M., Guttman, A., J. Chroma-
togr. A 1999, 830, 465±471.
of DNA fragments. The activation energy (Ea) values de-
[22] Guttman, A., Cooke, N., Anal. Chem. 1991, 63, 2038±2042.
rived from the slopes of the Arrhenius plots of ln m vs. 1/T [23] Ogston, A. G., Trans. Faraday. Soc. 1958, 54, 1754±1757.
exhibited elevating characteristics with increasing frag- [24] Lumpkin, O. J., Dejardin, P., Zimm, B. H., Biopolymers
ment size, suggesting the higher activation energy re- 1985, 24, 1573±1593.
quirement for the migration of the larger fragments [25] Karger, B. L., Chu, Y. H., Foret, F., Annu. Rev. Biophys.
Biomol. Struct. 1995, 24, 579±610.
through the sieving medium. The use of 0.5 mM intercala-
[26] Cifuentes, A., Poppe, H., Electrophoresis 1995, 16,
tor dye in the sample proved to be favorable for efficient 2051±2059.
labeling with no considerable self-quenching. The high [27] Haugland, R. P., Handbook of Fluorescent Probes and
applied electric field strength apparently affected the elec- Research Chemicals, Molecular Probes Inc., Eugene, OR
1999
trophoretic mobility of the solute molecules.
[28] Mathies, R. A., Huang, X. C., Nature 1992, 359, 167±169.
[29] Ferguson, K. A., Metab. Clin. Exp. 1964, 13, 985±1002.
The authors greatly appreciate the help and stimulating [30] Holmes, D. L., Stellwagen, N. C., Electrophoresis 1990, 11,
discussions of Drs. Julia Khandurina and Aran Paulus. 5±15.
[31] Gao, Q., Pang, H. M., Yeung, E. S., Electrophoresis 1999,
Received August 29, 2000 20, 1518±1526.
[32] Lengyel, T., Guttman, A., J. Chromatogr. A 1999, 853,
511±518.
5 References [33] Demorest, D., Dubrov, R., J. Chromatogr. 1991, 559,
43±51.
[1] Manz, A., Graber, N., Widmer, H. M., Sens. Actuators B
1990, 1, 244±248. [34] Lu, H., Arriaga, E., Chen, D. Y., Figeys, D., Dovichi, N. J.,
J. Chromatogr. 1994, 680, 503±510.
[2] Harrison, D. J., Manz, A., Fan, Z., Ludi, H., Widmer, H. M.,
Anal. Chem. 1992, 64, 1926±1932. [35] Gerstner, A., Csapo, Z., Sasvari-Szekely, M., Guttman, A.,
Electrophoresis 2000, 21, 834±840.
[3] Effenhauser, C. S., Paulus, A., Manz, A., Widmer, H. M.,
Anal.Chem. 1994, 66, 2949±2953. [36] Yan, X., Grace, W. K., Yoshida, T. M., Habbersett, R. C.,
Velappan, N., Jett, J. H., Keller, R. A., Marrone, B. L.,
[4] Wooley, A. T., Mathies, R. A., Proc. Natl. Acad. Sci. USA
Anal. Chem. 1999, 71, 5470±5480.
1994, 91, 11348±11352.
[37] Southern, E. M., Anal. Biochem. 1979, 22, 6186±6193.
[5] Jacobson, S. C., Ramsey, J. M., Anal. Chem. 1996, 68,
720±723. [38] Cantor, C. R., Smith, C. L., Mathew, M. K., Annu. Rev.
Biophys. Biophys. Chem. 1988, 17, 287±304.
[39] Freifelder, D., J. Mol. Biol. 1971, 60, 401±403.