Application off classical gel electrophoresis to the chiral separation off milligram quantities off terbutaline

Article abstract of DOI:10.1021/ac970524q

Although the last couple of decades has seen tremendous progress in the ability to do chiral separations, the ability to do larger scale chiral separations has lagged somewhat behind the current analytical chiral separation state of the art. The potential of classical gel electrophoresis for chiral separation of milligram quantities of chiral material is examined. A protocol for the chiral separation of milligram quantities of terbutaline using sulfated cyclodextrin as a chiral additive is demonstrated. The possible advantages of the approach are discussed.

Full text of DOI:10.1021/ac970524q

Anal. Chem. 1998, 70, 144-148
Application of Classical Gel Electrophoresis to the
Chiral Separation of Milligram Quantities of
Terbutaline
Apryll M. Stalcup,* Kyung H. Gahm,† Samuel R. Gratz, and Richard M. C. Sutton
Department of Chemistry, University of Cincinnati, P.O. Box 210172, Cincinnati, Ohio 45221-0172
and centrifugal partition chromatography19 have all been used to
effect larger than analytical-scale chiral separations.
Although the last couple of decades has seen tremendous
progress in the ability to do chiral separations, the ability
to do larger scale chiral separations has lagged somewhat
behind the current analytical chiral separation state of the
art. The potential of classical gel electrophoresis for chiral
separation of milligram quantities of chiral material is
examined. A protocol for the chiral separation of mil-
ligram quantities of terbutaline using sulfated cyclodextrin
as a chiral additive is demonstrated. The possible ad-
vantages of the approach are discussed.
Prep-LC is often the method of choice for preparative chiral
separations, but as with any technique, several factors must be
considered. Prep-LC requires the availability of a suitable chiral
stationary phase, optimization of chromatographic conditions (e.g.,
mobile phase and temperature), determination of the adsorption
isotherms of both enantiomers on the stationary phase, and
evaluation of the robustness of the method. In addition, the mode
of chromatography (e.g., batch vs displacement vs recycling vs
simulated moving bed) must be chosen. Ultimately, the sample
capacity of chiral stationary phases may be circumscribed by the
absolute amount of chiral selector in the chiral stationary phases
(e.g., typically 0.2-0.3 µmol/ m2 for cyclodextrins)20 or the density
of chirally selective sites of immobilized chiral ligands such as
proteins. Further, the low density of chirally selective sites in
some chiral stationary phases diminishes the likelihood of coop-
erativity between adjacent ligands for enhanced chiral recognition.
Historically, many of the chiral selectors currently available
as chiral stationary phases for HPLC originated as chiral mobile
phase additives, particularly in thin-layer chromatography (TLC).21
More recently, chiral additives have been shown to be effective
for chiral separations by capillary electrophoresis (CE).22 Chiral
additives in CE have several advantages, some of which are
highlighted in Table 1. Chiral additives in free solution also offer
the potential for multiple complexation,23 thus possibly reinforcing
chiral recognition. Recently, we reported24,25 the application of
sulfated cyclodextrins (ds ∼7-10) as chiral additives in CE for
the enantioseparation of terbutaline and over 70 other compounds
of pharmaceutical interest. In this work, the general approach
was to use a background electrolyte at low pH, to minimize the
electroosmotic flow and to operate the system with the anode at
the detector end of the column. Under these conditions, neutral
and cationic species did not reach the detector in the absence of
the polyanionic sulfated cyclodextrin. The structural diversity
Separation of enantiomers is an important topic to the phar-
maceutical industry. Many of the drugs marketed in the United
States have at least one chiral center. Examples of chiral drugs
currently on the market include ibuprofen and propranolol. Of
the 528 synthetically derived chiral drugs, 88% are sold as the
racemic mixture.1
Despite the commercialization of a large number of different
types of high-performance liquid chromatographic chiral stationary
phases in the last decade, including the cyclodextrin phases2,3 the
macrocyclic antibiotic phases,4-6 the π-π interaction phases,7,8
and the protein phases,9-12 as well as the cellulosic and amylosic
phases13,14 and the chiral crown ether phases,15,16 the problem of
large-scale chiral separations remains largely unexplored. Pre-
parative liquid chromatography (prep-LC)17 and countercurrent18
† Present address: Department of Chemistry, 142 Schrenk Hall, University
of MissourisRolla, Rolla, MO 65401.
(1) Lin, B.; Zhu, X.; Koppenhoefer, B.; Epperlein, U. LC-GC 1 9 9 7 , 15, 40.
(2) Ward, T. J.; Armstrong, D. W. In Chromatographic Chiral Separations; Zief,
M., Crane, L. J., Eds.; Marcel Dekker, Inc.: New York, 1988; pp 131-163.
(3) Armstrong, D. W.; DeMond, W. J. Chromatogr. Sci. 1 9 8 4 , 22, 411-415.
(4) Armstrong, D. W.; Tang, Y.; Chen, S. Anal. Chem. 1 9 9 4 , 66, 1473.
(5) Berthod, A.; Liu, Y.; Bagwill, C.; Armstrong, D. W. J. Chromatogr. 1 9 9 6 ,
731, 123.
(6) Armstrong, D. W.; Liu, Y.; Ekborgott, K. H. Chirality 1 9 9 5 , 7, 474.
(7) Pirkle, W. H.; Welch, C. J. Tetrahedron: Asymmetry 1 9 9 4 , 5, 777-780.
(8) Pirkle, W. H.; Murray, P. G. J. Liq. Chromatogr. 1 9 9 0 , 13, 2123-2134.
(9) Hermansson, J.; Eriksson, M. J. Liq. Chromatogr. 1 9 8 6 , 9, 621-639.
(10) Schill, G.; Wainer, I. W.; Barkin, S. A. J. Liq. Chromatogr. 1 9 8 6 , 9, 641-
666.
(18) Ma, Y.; Ito, Y.; Foucault, A. J. Chromatogr. 1 9 9 5 , 704, 75-81.
(19) Oliveros, L.; Puertolas, P. F.; Minguillon, C.; Camacho-Frias, E.; Foucault,
A.; Le Goffic, F. J. Liq. Chromatogr. 1 9 9 4 , 17, 2301-2318.
(20) Stalcup, A. M.; Williams, K. L. J. Liq. Chromatogr. 1 9 9 2 , 15, 29-37.
(21) Armstrong, D. W.; Faulkner, J. R., Jr.; Han, S. M. J. Chromatogr. 1 9 8 8 ,
452, 323-330.
(11) Allenmark, S. J. Liq. Chromatogr. 1 9 8 6 , 9, 425-442.
(12) Okamoto, M.; Nakazawa, H. J. Chromatogr. 1 9 9 0 , 508, 217-219.
(13) Isaksson, R.; Erlandsson, P.; Hansson, L.; Holmberg, A.; Berner, S. J.
Chromatogr. 1 9 9 0 , 498, 257-280.
(14) Yashima, E.; Okamoto, Y. Bull. Chem. Soc. Jpn. 1 9 9 5 , 68, 3289-3307.
(15) Shinbo, T.; Yamaguchi, T.; Nishimura, K.; Sugiura, M. J. Chromatogr. 1 9 8 7 ,
405, 145-153.
(22) Ward, T. J. Anal. Chem. 1 9 9 4 , 66, 632A-640A.
(23) Armstrong, D. W. ; Nome, F.; Spino, L. A.; Golden, T. D. J. Am. Chem. Soc.
1 9 8 6 , 108, 1418-1421.
(16) Hilton, M.; Armstrong, D. W. J. Liq. Chromatogr. 1 9 9 1 , 14, 9-28.
(17) Ku¨ sters, E. Chim. Oggi 1 9 9 6 , 14, 39-43.
(24) Stalcup, A. M.; Gahm, K. H. Anal. Chem. 1 9 9 6 , 68, 1360-1368.
(25) Gahm, G. H.; Stalcup, A. M. Chirality 1 9 9 6 , 8, 316-324.
144 Analytical Chemistry, Vol. 70, No. 1, January 1, 1998
S0003-2700(97)00524-6 CCC: $14.00 © 1997 American Chemical Society
Published on Web 01/01/1998

quent to the electrophoretic separation, an important consideration
for costly chiral additives. The presence of the chiral selector in
free solution in the electrophoresis system may allow for multiple
complexation as well as increased capacity compared to the
analogous chromatographic bed where, as noted previously, the
absolute amount of available chiral selector is limited by the
physical and chemical properties of the substrate and ligand.
Terbutaline, which was readily enantioresolved in our earlier
work, is a short-acting â2-adrenergic agonist which may be of
Table 1. Advantages of Chiral Additives in CE
additive can be readily changed
variety of chiral selectors available
rapid screening of chiral selectors
small amounts of BGE, chiral additive required
rapid screening of conditions
no pre-equilibration
multiple complexation possible
rapid screening of analytes
present in solutes successfully enantioresolved and the high
resolution (e.g., Rs ∼26 for 5-cyclobutyl-5-phenylhydantoin) easily
achieved under nonoptimized conditions for some compounds
established the versatility of this chiral additive. Unfortunately,
CE is generally more suited to analytical separations than to
preparative- or semipreparative-scale separations.
Classical gel electrophoresis, a mature separation method used
extensively for protein and nucleic acid purification and charac-
terization,26,27 may complement CE and offer a viable alternative
to chromatographic methods for preparative- or semipreparative-
scale chiral separations. However, its application to small mol-
ecule separations, other than for polypeptides, has been somewhat
limited, presumably because the solute begins to diffuse away from
the band center as soon as the applied voltage is removed.
Although detection is usually accomplished off-line in electro-
phoretic and thin-layer chromatographic methods, the affinity of
the analyte for the chromatographic bed and the immediate
removal of the mobile phase following the chromatographic run
minimizes the problem of solute diffusivity in TLC. In contrast,
in gel electrophoresis, the gel matrix serves primarily as an
anticonvective and/ or molecular sieving medium designed to
minimize interactions with the solute. Hence, there is no
mechanism to localize the analyte immediately postrun unless the
gel is “fixed”. The presence of bulk liquid in the postrun gel no
doubt contributes to the solute diffusivity problem which degrades
the separation and complicates detection. However, complexation
between the solute and a bulky additive (e.g., sulfated cyclodex-
interest because of its growth-promoting properties and potential
illegal use in livestock.30 Terbutaline is also approved in the
United States for the treatment of asthma31 and for the inhibition
of premature labor. The symptoms of asthma are thought to be
the result of an imbalance between the bronchodilator and
bronchoconstrictive regulatory systems. Of particular concern in
the treatment of asthma with â2-adrenergic agonists like terbuta-
line, administered to mediate bronchodilation and exclusively
marketed as the racemate, is that one enantiomer may actually
function as a bronchodilator while the other acts as a bronchoc-
onstrictor.
The purpose of this work was to explore the potential of
classical gel electrophoresis using sulfated cyclodextrin as an
additive for the chiral separation of milligram quantities of
terbutaline. To the best of our knowledge, this is the first report
of classical gel electrophoresis used for the chiral separation of a
racemic compound. Preliminary work using classical gel elec-
trophoresis for chiral separations employed a horizontal column
electrophoresis configuration. Subsequent work was adapted to
a vertical Bio-Rad Mini-prep continuous-elution electrophoresis
apparatus.
trin) with reasonably large binding constants (∼103 M-1
)
28 might
effectively localize the solute.
The use of gel electrophoresis for preparative-scale chiral
separations may offer some of the same advantages listed in Table
1 for CE (e.g., choice and concentration of additive). Another
advantage of gel electrophoresis is the use of aqueous solvents
that are less hazardous than the hydroorganic or organic solvents
typically used in most preparative chromatographic-based separa-
tions.29 Also, the absolute volume of solvent required for a
classical electrophoretic separation may be much less than that
required for the analogous HPLC-, centrifugal partition-, or
countercurrent chromatographic-based methods, if one considers
the amount required to pre-equilibrate the chromatographic
systems. The chiral selector should be readily retrieved subse-
EXPERIMENTAL SECTION
Materials. Hydroxypropyl-â-cyclodextrin used in the original
gel work was obtained from Advanced Separation Technologies,
Inc. (Whippany, NJ) and functionalized in-house.32 The sulfated
â-cyclodextrin used in the Mini-prep system (13 sulfates/ cyclo-
dextrin) was generously donated by Cerestar, Inc. (Hammond,
IN). The terbutaline hemisulfate salt was obtained from Sigma
Chemical Co. (St. Louis, MO). The agarose (medium EEO,
electrophoretic grade) and all other buffer components were
obtained from Fisher Scientific (St. Louis, MO).
Apparatus. The instrument used for the CE experiments was
either a Quanta 4000 capillary zone electrophoresis (CZE) system
equipped with a UV detector (214 nm) interfaced to a Shimadzu
Chromatopac CR-501 data station or a Bio-Rad 3000 capillary
electrophoresis system equipped with a variable-wavelength UV
(26) Allen, R. C.; Budowle, B. Gel electrophoresis of proteins and nucleic acids:
selected techniques; Walter de Gruyter: Berlin, 1994.
(27) Bruno, T. J. Chromatographic and electrophoretic methods; Prentice Hall:
Englewood Cliffs, NJ, 1991; pp 141-169.
(28) Agyei, N. M. Heparin and dextran sulfate as novel chiral selectors for the
direct enantioseparation of racemic pharmaceutical drugs by capillary zone
electrophoresis. Master’s Thesis, University of Hawaii, 1994.
(29) Armstrong, D. W.; Bui, K. H.; Barry; R. M. J. Chem. Educ. 1 9 8 4 , 61, 457-
458.
(30) Doerge, D. R.; Churchwell, M. I.; Holder, C. L.; Rowe, L.; Bajic, S. Anal.
Chem. 1 9 9 6 , 68, 1918-1923.
(31) Stinson, S. C. Chem. Eng. News 1 9 9 7 , 75 (Jan 6), 25.
(32) Stalcup, A. M.; Gahm. K. H. Anal. Chem. 1 9 9 6 , 68, 1369-1374.
Analytical Chemistry, Vol. 70, No. 1, January 1, 1998 145

after ∼12 h of applied voltage (∼450 V). The gel was sliced and
extracted with methanol, and the individual extracts were sub-
jected to chiral CE analysis using 2% sulfated â-cyclodextrin, 10
mM phosphate buffer, pH 3.2. Individual slices were extracted
and analyzed by chiral CE to validate the separation. In the case
of the Mini-prep cell, the applied voltage was ∼70 V and run times
were ∼4-5 h. Fractions collected from the Mini-prep cell
identified from the UV trace as containing terbutaline were
subjected to chiral CE analysis to determine the enantiomeric
composition of the fractions. Most of the results have been
obtained loading ∼3-4 mg of terbutaline.
RESULTS AND DISCUSSION
Thus far, the classical electrophoretic experiments have
focused on disc electrophoresis using a sulfated cyclodextrin-
impregnated agarose-based gel. The polarity of the applied
voltage in the gel is directly analogous to the polarity in the CE
experiment, and the mobility of the terbutaline and the cyclodex-
trin are directly analogous to the CE experiment. Thus, optimiza-
tion of the CE separation has a direct relationship to optimization
of the gel protocol. For instance, as in the CE case, terbutaline
does not reach the end of the gel unless complexed with the
polyanionic cyclodextrin. Terbutaline and the cyclodextrin pro-
ceed in a “countercurrent” process through the gel, thus amplify-
ing the chiral discrimination inherent in the complex.
Figure 1. Schematic of original apparatus used for classical gel
electrophoresis.
detector set at 211 nm interfaced to a Gateway 2000 PC. Both
CZE systems were operated with the cathode at the injector end
of the capillary. The fused-silica capillary columns were either
75 µm i.d. with a column length of 60 cm (52.4 cm to detector
window) or 50 µm i.d. with a column length of 25 cm (20.4 cm to
detector window).
The apparatus used for the original gel electrophoretic work
is illustrated in Figure 1. A West condenser was used to supply
mechanical support for the electrophoretic bed and provided a
convenient method for cooling. Subsequent gel electrophoretic
work employed a Bio-Rad Mini-prep cell. The power supply for
both systems was a Bio-Rad PowerPac 3000 power supply.
In the Mini-prep cell, the eluent flushing the bottom end of
the electrophoretic bed (∼10 cm × 7 mm i.d.) was delivered by
a Bio-Rad Econo peristaltic pump to a Shimadzu SPD-6A UV
variable-wavelength detector interfaced to a Shimadzu Chromato-
pac CR-501 data station and then to an Isco Retriever II fraction
collector.
In this system, the agarose serves primarily as an anticonvec-
tive media. The decision to work with agarose was based on its
ready availability in high purity and low cost. Although agarose
gels are generally not used in a rod configuration because of poor
elasticity, in the present application, the gel seemed to be stabilized
by the presence of the additive and gel fragility was not
encountered in the work performed thus far.
A typical electropherogram of the individual slices of the gel
used in the West condenser can be seen in Figure 2. CE analysis
of the individual slices was used to generate a histogram (Figure
3) to show how far the individual enantiomers had migrated in
the gel and how well the system worked for resolving the
enantiomers. It should be noted that while the CE was used to
optimize conditions for the gel (maximum resolution), the CE
conditions to analyze the individual fractions were selected only
on the basis of baseline resolution within the shortest possible
analysis time. As can be seen from Figure 3, in some fractions,
near enantiomeric purity was achieved. Although there appears
to be some inconsistencies in the distribution of the enantiomers
in the different fractions, this may well be an artifact of the crudity
of the solute recovery process. In any case, the binding between
terbutaline and the sulfated cyclodextrin to some extent seemed
to mediate the expected solute diffusivity problems associated with
classical electrophoretic methods.
Methods. The buffer pH, ionic strength, and sulfated cyclo-
dextrin concentration for the gel electrophoretic run was derived
from the CE conditions.
To prepare the electrophoretic bed in the West condenser, the
condenser was positioned in an upright position, with the bottom
end capped with a ground glass stopper, and filled with a 2%
agarose gel (medium EEO) aqueous buffer solution (10 mM
phosphate, pH 3.2) incorporating 1-2% sulfated hydroxypropyl-
â-cyclodextrin. The gel bed length was ∼50 cm with a diameter
of ∼1 cm. A similar procedure was used for the Mini-prep system
using the sulfated â-cyclodextrin. In the Mini-prep cell, the gel
bed was ∼10 cm × 7 mm i.d.
Sample introduction in both systems required that the sample
be dissolved in a small amount of the pregelled agarose and
introduced to the top end of the column after the gel bed had
solidified (∼1 h). Standard electrophoretic methods of layering
the sample on top of the gel using ethylene glycol or glycerin
could not be used in the horizontal configuration and because of
competition between the solvent and terbutaline for the cyclo-
dextrin cavity.
The decision to adapt the overall approach to a Mini-prep
continuous-elution electrophoretic cell was predicated on reducing
the labor-intensive process required for solute recovery, which
also introduced potential opportunities for sample remixing. In
the Mini-prep system, the eluent is sent to an HPLC UV detector
and then to a fraction collector. Figure 4 illustrates a trace
obtained from the UV detector monitoring the eluent coming from
the end of the electrophoretic bed. It is important to note that
the blunted shape of the peaks probably arises from the fact that
the sample concentration is well beyond the linear range of the
In the investigation using the gel immobilized in the West
condenser, the slab gel was extruded from the column support
146 Analytical Chemistry, Vol. 70, No. 1, January 1, 1998

Figure 4. UV trace for ∼3 mg of terbutaline eluting from a 2%
sulfated cyclodextrin-impregnated agarose gel using 2% sulfated
cyclodextrin, 10 mM phosphate buffer (pH 3) as the elution buffer.
Figure 2. Typical electropherograms obtained for two individual
fractions (a, b) collected from the gel used in Figure 1 and the racemic
terbutaline (c) using 10 mM phosphate buffer (pH 3.0) with 2%
sulfated â-cyclodextrin.
Figure 5. Electropherograms of two individual fractions collected
from the Mini-prep electrophoresis cell. CE conditions: 2% sulfated
cyclodextrin, 10 mM phosphate buffer (pH 3).
enantiomer but is only included to illustrate the migration times
for both enantiomers. Figure 6 illustrates a typical histogram
showing the distribution of enantiomers in the individual fractions.
As can be seen from the histogram, there is a more uniform
distribution of the enantiomers in the fractions, thus supporting
the supposition that the inconsistencies in enantiomer distribution
in the original gel slices may have been an artifact of the recovery
process.
Figure 3. Histogram for terbutaline enantiomers generated from
CE analysis of the individual gel slices.
detector. Figure 5 illustrates a typical CE analysis of two
terbutaline fractions obtained from the Mini-prep cell. It should
be noted that the lower electropherogram in Figure 5 does not
represent the most complete separation achieved for the second
Routine applications of gel electrophoresis typically involve
separations of fairly complex, biologically derived samples. As a
result, the gel matrix is usually contaminated by denatured
Analytical Chemistry, Vol. 70, No. 1, January 1, 1998 147

CONCLUSIONS
The work presented here demonstrates that classical gel
electrophoresis is indeed viable as a method for the separation of
milligram quantities of chiral compounds although many questions
remain to be answered. It should be noted that the separations
obtained thus far in the gel are based on nonoptimized conditions.
In our initial investigation, it seemed likely that solutes most
amenable to this classical gel electrophoresis method should
contain an ionizable moiety that complemented the charge on the
chiral selector. It was thought that significant binding between
the analyte and the cyclodextrin was critical to circumvent the
solute diffusivity problems associated with classical electrophoretic
methods particularly when the gel had to be extensively manipu-
lated to recover the analyte. However, implementation of the
continuous-elution electrophoretic cell should, in principle, allow
even neutrals to be chirally resolved because the enantiomers are
recovered during instead of after the electrophoretic run. Thus,
this approach should be generally applicable to any of the over
90 structurally diverse compounds successfully enantioresolved
using the sulfated cyclodextrins and no doubt to other chiral
selector/ chiral analyte combinations as well.
Figure 6. Histogram for terbutaline enantiomers generated from
CE analysis of the individual fractions.
proteins or other sample components that fail to elute from the
gel, thereby compromising future experiments. However, our
samples are generally comprised only of the enantiomeric mixture
to be separated. This feature of the separations coupled with the
elimination of the need for slicing and extracting the gel by using
the continuous-elution cells allows the gel to be used repeatedly.
Indeed, gels have been used five times without any noticeable
deterioration (e.g., gel integrity, ability to maintain applied volt-
age). The ability to use the same gel with different analytes or
under slightly different experimental conditions (e.g., sample load,
run buffer, pH, ionic strength) allows for more meaningful com-
parisons of experimental results because the bed-to-bed variability
is minimized. Typical electrophoretic runs, with an applied voltage
of 70 V, take ∼4-5 h. Sulfated cyclodextrin is recovered by simply
adjusting the run buffer to pH 8, using NaOH solution, adding
ethanol.
ACKNOWLEDGMENT
The authors gratefully acknowledge the generous support of
the National Institutes of Health First Award (Grant 1R29
GM48180-04), BioRad Laboratories for the loan of the Mini-prep
electrophoresis system, Yung-Hae Kim for assistance with the
initial electrophoresis setup, and Cerestar, Inc. for the donation
of the sulfated cyclodextrin.
Received for review May 21, 1997. Accepted October 7,
1997.X
AC970524Q
X Abstract published in Advance ACS Abstracts, November 15, 1997.
148 Analytical Chemistry, Vol. 70, No. 1, January 1, 1998