Direct resolution of optically active isomers on chiral packings containing ergoline skeletons. 5. Enantioseparation of amino acid derivatives

Article abstract of DOI:10.1021/ac950698c

A new procedure for ergot alkaloid-based chiral stationary phase preparation is described. Synthesis is based on bonding the allyl derivative of terguride to mercaptopropylsilanized silica gel. The packing exhibits higher content of chiral selector, stability, reproducibility, and enantioselectivity toward amino acids compared to that previously studied. The chromatographic behavior of amino acids with different side chains and substituent groups is investigated in order to obtain a deeper insight into the enantiodiscriminative mechanism as well as to determine the limitations and strengths of terguride as a chiral selector for this class of compounds. A variety of factors, including mobile phase parameters such as pH, ionic strength, content and nature of organic modifier, and temperature, are examined. A new procedure for ergot alkaloid-based chiral stationary phase preparation is described. Synthesis is based on bonding the allyl derivative of terguride to mercaptopropylsilanized silica gel. The packing exhibits higher content of chiral selector, stability, reproducibility, and enantioselectivity toward amino acids compared to that previously studied. The chromatographic behavior of amino acids with different side chains and substituent groups is investigated in order to obtain a deeper insight into the enantiodiscriminative mechanism as well as to determine the limitations and strengths of terguride as a chiral selector for this class of compounds. A variety of factors, including mobile phase parameters such as pH, ionic strength, content and nature of organic modifier, and temperature, are examined.

Full text of DOI:10.1021/ac950698c

Anal. Chem. 1996, 68, 1191-1196
Direct Resolution of Optically Active Isomers on
Chiral Packings Containing Ergoline Skeletons. 5.
Enantioseparation of Amino Acid Derivatives
A. Messina,*,† A. M. Girelli,† M. Flieger,‡ M. Sinibaldi,§ P. Sedmera,‡ and L. Cvak
Dipartimento di Chimica, Universita` di Roma “La Sapienza”, Piazzale Aldo Moro 5, I-00185 Roma, Italy, Institute of
Microbiology, Academy of Sciences of the Czech Republic, Videnska` 1083, C-142 20 Prague 4, Czech Republic, Istituto di
Cromatografia, CNR-Area della Ricerca di Roma, P.O. Box 10, I-00016 Monterotondo Stazione, Roma, Italy, and Galena
Opava, Opava, Czech Republic
stationary phases (CSPs) and to extend the study to a series of
AAs having different side chains and derivatizing groups. This
study should also provide a deeper insight into the resolution
mechanism, since an additional, steric-type interaction is hypoth-
esized to determine the stability of the two diastereoisomeric
complexes. The limitations and strengths of terguride as a chiral
selector for this class of compounds are considered. The prepara-
tion of a stationary phase based on 1-allyl-(5R,8S,10R)-terguride
as chiral selector is described. On this sorbent, the chromato-
graphic behavior of a number of AAs with aliphatic (valine),
aromatic (tryptophan), acidic (aspartic and glutamic acids), and
polar (serine and threonine) side chains as their dansyl, N-(2,4-
dinitrophenyl), N-(3,5-dinitrobenzoyl), N-benzoyl, and â-naphthoyl
derivatives is examined.
A new procedure for ergot alkaloid-based chiral stationary
phase preparation is described. Synthesis is based on
bonding the allyl derivative of terguride to mercaptopro-
pylsilanized silica gel. The packing exhibits higher con-
tent of chiral selector, stability, reproducibility, and
enantioselectivity toward amino acids compared to that
previously studied. The chromatographic behavior of
amino acids with different side chains and substituent
groups is investigated in order to obtain a deeper insight
into the enantiodiscriminative mechanism as well as to
determine the limitations and strengths of terguride as a
chiral selector for this class of compounds. A variety of
factors, including mobile phase parameters such as pH,
ionic strength, content and nature of organic modifier, and
temperature, are examined.
EXPERIMENTAL SECTION
Apparatus. The liquid chromatographic system consisted of
Enantioseparation of amino acids (AAs) is of great interest in
many fields of study concerning the life sciences, including
biomedical research, food production, geochronological and
archaeological dating procedures, etc. While the literature
abounds with techniques for the resolution of free and derivatized
AAs, new procedures are welcome whenever advantages in terms
of selectivity, cost of the analysis, and reproducibility are offered
with respect to current methodology. Accordingly, in previous
work we dealt with the use of a semisynthetic ergot alkaloid
derivative, 1-(3-aminopropyl)-(5R,8S,10R)-terguride, which, bonded
to silica gel, showed good enantioselectivity for a number of
carboxylic group-containing racemates, including dansylamino
acid derivatives.1 Proton NMR spectrometry of the chiral agent
and an arylcarboxylic acid (naproxen) in solution allowed us to
identify two types of interactions, π-π and electrostatic, that are
thought to lead to the enantiodiscriminative process; we were also
able to exclude any contribution of the aminopropyl chain to the
formation of diastereoisomeric adducts.2
a Series 400 (Perkin Elmer, Norwalk, CT) solvent delivery pump
equipped with a Model 7125 (Rheodyne) injection valve and
connected to a Model 2550 (Varian, Walnut Creek, CA) UV
detector. The chromatograms were monitored by a Chromatopac
CR3A (Shimadzu, Kjoto, Japan) integrator.
Synthesis of the Chiral P acking. A solution of (5R,8S,10R)-
terguride (4 g, 11.7 mM) in CH2Cl2 (160 mL) was mixed with a
solution of 20% (v/ v) tetraethylammonium hydroxide (8 mL) in
24 mL of 50% (w/ v) NaOH, and allyl bromide (5 mL, 58.6 mM)
was added dropwise with vigorous swirling. After the addition
was completed, the swirling was continued for 5 min. The
separated organic layer was washed with water (2 × 200 mL) and
concentrated under reduced pressure. The residue was chro-
matographed on a silica gel column (40 g) with CH2Cl2 as eluent.
The fractions containing allylterguride were evaporated, and the
pure compound was crystallized from diethyl ether-petroleum
ether (yield 2.5 g of 1-allyl-(5R,8S,10R)-terguride).
The structure was verified by EI-MS and 1H and 13C NMR
These findings and the fact that the synthesis of aminopropyl-
terguride is a relatively complicated procedure led us to develop
a new method for the preparation of terguride-based chiral
spectroscopy (data are summarized in Table 1).
EI mass spectra were measured on a Finnigan MAT 90
(double-focusing, BE geometry) instrument under the following
conditions: ionization energy 70 eV; source temperature, 250 °C;
emission current, 1 mA; accelerating voltage, 5 kV; direct inlet,
DIP; temperature, 170 °C.
† Universita` di Roma “La Sapienza”.
‡ Academy of Sciences of the Czech Republic.
§ CNR-Area della Ricerca di Roma.
Galena Opava.
EI-MS data for 1-allyl-(5R,8S,10R)-terguride [m/ z (relative
intensity)]: 381 (10), 380 (39), 308 (6), 307 (15), 265 (6), 264 (24),
263 (36), 249 (8), 221 (5), 220 (5), 209 (7), 208 (21), 207 (100),
195 (12), 194 (18), 100 (4), 74 (3).
(1) Sinibaldi, M.; Flieger, M.; Cvak, L.; Messina, A.; Pichini, A. J. Chromatogr.
A 1 9 9 4 , 666, 471.
(2) Castellani, L.; Flieger, M.; Mannina, L.; Sedmera, P.; Segre, A. L.; Sinibaldi,
M. Chirality 1 9 9 4 , 6, 543.
0003-2700/96/0368-1191$12.00/0 © 1996 American Chemical Society
Analytical Chemistry, Vol. 68, No. 7, April 1, 1996 1191

Table 1. NMR Data for 1-Allyl-(5R,8S,10R)-terguride
(400 and 100 MHz, CDCl3, 25 °C)
Table 2. Mean Value of the Surface Coverage, Derived
from Elemental Analysis and Determined for Four
1-Propyl-(5R,8S,10R)-terguride Silica Gel (CSPII)
Materials, Which Were Prepared under the Same
Reaction Conditions (See Experimental Section)a
CSP
allyl-TER
a
mmol/ g (CV, %)b
k′ c (SD)
Rd
Ne
D
0.051 (4.8)
17.7 ((0.65)
1.29
5090
The chromatographic properties of the CSPII were obtained by a
series of injections of Dns DL valine. b Coefficient of variation. c Chro-
matographic conditions: column size: 150 mm × 4.6 mm i.d.; eluent,
0.02 M potassium phosphate buffer (pH 3.6)/ ACN (60:40 v/ v); flow
rate, 0.8 mL/ min; detection, UV, 254 nm; room temperature. kD′ ,
capacity factor of the more retained enantiomer. d Selectivity factor (kD′ /
kL′ ). e Plate number, calculated for the more retained enantiomer and
related to a column of 150 mm length.
vacuum at 50 °C. Elemental analysis (N, 0.31; C, 3.88; H, 0.60;
and S, 1.60) corresponded to 0.045 mmol/ g.
The sorbent, slurred in methanol (3 g/ 20 mL), was packed in
a stainless steel tube (150 mm × 4.6 mm i.d.), using methanol as
the pressurizing solvent.
Material and Samples. All solvents were of HPLC grade and
were obtained from Merck (Darmstadt, Germany). Other chemi-
cals of analytical grade were purchased from Carlo Erba (Milan,
Italy).
Phosphate buffers were prepared using phosphoric acid and
sodium hydroxide, monitoring the pH by means of a Crison 2000
pH micrometer. Water was filtered through Millipore (Bedford,
MA) type GS (0.22 µm) filter disks.
13C NMR
1H NMR
nH mult.a
atom
δ
mult.
δ
J (Hz)
2
3
4
121.73
110.84
26.91
d
s
t
6.774
1
d
1.7
2.667
3.379
2.205
2.485
2.874
4.282
1.636
2.796
3.052
1
1
1
1
1
1
1
1
1
ddd
dd
ddd
dd
ddd
m
ddd
dddd
m
14.6, 11.1, 1.7
14.6, 4.3
11.1, 9.7, 4.3
11.7, 2.6
5
7
67.61
61.89
d
t
11.7, 2.6, 2.4
8
9
44.95
32.55
d
t
13.2, 13.0, 3.3
13.2, 4.3, 2.6, 2.6
10
11
12
13
14
15
16
NsCH3
36.52
133.52
112.83
122.73
107.12
133.68
126.71
43.39
d
s
d
d
d
s
s
q
s
t
6.888
7.158
7.100
1
1
1
ddd
dd
ddd
6.9, 1.3, 0.9
8.3, 6.9
8.3, 0.9, 0.8
Free amino acids and their derivatives, where not otherwise
indicated, were purchased from Sigma (St. Louis, MO).
The following DL- and
L-amino acid derivatives were exam-
2.416
3
s
ined: dansyl (Dns) derivatives of valine (Val), serine (Ser), aspartic
acid (Asp), tryptophan (Trp), methionine (Met), threonine (Thr),
phenylalanine (Phen), leucine (Leu), proline (Pro), and alanine
(Ala); N-(3,5-dinitrobenzoyl) (DNB) derivatives of glutamic acid
(Glu), Val, Ala, and Trp; â-naphthoyl (NPT) derivatives of Ala,
Val, Ser, and Trp; benzoyl (BNZ) derivatives of Val, Ala, and Trp;
and N-(2,4-dinitrophenyl) (DNP) derivatives of Val, Trp, Glu, and
Ser.
NPT and BNZ derivatives of Trp and Ala,3 DNP derivatives of
Val and Ser,4 and DNB derivatives5 were prepared according to
the procedures reported in the literature. They were all of a purity
sufficient for the purposes of the work.
NsCdO 156.57
1′
2′
3′
48.91
133.87
117.00
4.674
5.990
5.124
5.188
3.250
3.347
1.152
5.572
2
1
1
1
1
1
3
1
ddd
ddt
ddt
ddt
dq
dq
t
5.5, 1.7, 1.5.
17.0, 10.2, 5.5
17.0, 1.3, 1.7
10.2, 1.3, 1.5
14.5, 7.1
14.5, 7.1
7.1
d
t
R
41.06
13.85
t
â
q
NsH
d
8.2
a
Abbreviations: s, singlet; d, doublet; t, triplet; q, quartet; m,
multiplet.
Sample solutions were prepared by dissolving the AA deriva-
tives in the eluent to obtain a concentration of ∼1 mg/ mL.
Mercaptopropyltrimethoxysilane (5 mL) (Serva, Heidelberg,
Germany) was coupled with 5 g of vacuum-dried Exsil-100 silica
gel (particle size, 5 µm; average pore diameter, 100 Å) (Scientific
Glass Engineering, Milton Keynes, UK) by swirling gently in 150
mL of anhydrous toluene at 100 °C for 12 h under dry nitrogen.
After cooling, the modified silica was collected by filtration, washed
exhaustively with toluene, n-hexane, 2-propanol, and diethyl ether,
and dried under vacuum. Surface coverage, determined by
elemental analysis (C, 2.60%; H, 0.50%; S, 1.69%), corresponded to
0.73 mmol/ g.
To the slurry of derivatized silica gel (5 g) in chloroform (50
mL) were added 1-allyl-(5R,8S,10R)-terguride (500 mg, 0.102
mmol) and R,R′-azobis(isobutyronitrile) (22 mg). The mixture
was refluxed with gentle stirring at 60 °C for 15 h. The resulting
material was filtered out, washed successively with chloroform,
methanol, acetone, 2-propanol, and diethyl ether, and dried under
RESULTS AND DISCUSSION
P acking P roperties. The new procedure for the preparation
of ergot alkaloid-based chiral stationary phases starting from
1-allyl-(5R,8S,10R)-terguride (CSPII, allyl-TER) makes it possible,
in comparison to the aminopropyl derivative (CSPI, AM-TER), to
increase the surface concentration of the chiral selector by a factor
of 1.875.6 Data obtained from repeated syntheses of CSPII and
the corresponding chromatographic properties are reported in
Table 2. This resulted in an overall improvement in selectivity
(3) Allenmark, S.; Bomgren, B.; Boren, H. J. Chromatogr. 1 9 8 3 , 264, 63.
(4) Schroeder, W. A.; Le Gotte, J. J. Am. Chem. Soc. 1 9 5 3 , 75, 4612.
(5) Rao, K. R.; Sober, H. A. J. Am. Chem. Soc. 1 9 5 4 , 76, 1328.
(6) Flieger, M.; Sinibaldi, M.; Cvak, L.; Castellani, L. Chirality 1 9 9 4 , 6, 549.
1192 Analytical Chemistry, Vol. 68, No. 7, April 1, 1996

Table 3. Chromatographic Data of Dns-Amino Acids on
CSPI and CSPIIa
AMP-TER
allyl-TER
Dns-AA
k′
D
R
k′
D
R
threonine
serine
valine
9.19
7.96b
8.06
1.21
1.17
1.12
1.06
1.07
1.12
1.06
1.14
1.30
10.9
11.6b
17.7
20.3
16.9
35.4
51.8
60.2
74.6
1.49
1.47
1.29
1.27
1.12
1.14
1.28
1.34
1.29
leucine
10.1
methionine
phenylalanine
glutamic acid
aspartic acid
tryptophan
11.6
16.8
25.5
38.1
23.4
a
Conditions: column size, 150 mm × 4.6 mm i.d. eluent, 0.02 M
potassium phosphate buffer (pH 3.6)/ AcCN (6:4 v/ v); flow rate, 0.8
mL/ min; detection, UV at 254 nm; room temperature. b CSPI, N )
10 544, Rs ) 6.6; CSPII, N ) 4050, Rs ) 9.4.
for dansylamino acids, with the exception of tryptophan. The
chromatographic data for a series of Dns-amino acid derivatives
on CSPI and CSPII are summarized in Table 3. Comparing the
results leads to some immediate conclusions. CSPI exhibits a
notably higher efficiency: the plate number (N) is doubled relative
to CSPII. This could be explained in terms of polymerization
occurring on the surface of the allylterguride-derived support.
Since radical reactions of monofunctional thiols with allyl com-
pounds do not lead to the formation of polymers,7 polymerization
of the silica surface has to be ascribed to the mercaptosilanization
of the sorbent, yielding a layer rich in terguride moieties. This
layer is probably responsible for the decreased selectivity found
for Dns-tryptophan. Because of the large size of the molecules,
the tryptophan enantiomers could be partially inhibited to interact
through stereoselective repulsions between the imidazole group
of the amino acid and the urea side chain of terguride,2 explaning
the low resolution of Dns-tryptophan enantiomers on CSPII.
However, the large improvement in the enantioselectivities
observed with the other compounds more than compensates for
the lower efficiencies of the packing, resulting in an overall
increase in the resolution (Rs) values (Table 3). Furthermore,
the packing exhibited a very high stability, even after 2000
injections. Comparative enantioseparation of Dns-DL-serine on the
two sorbents is shown in Figure 1.
Figure 1. Chromatographic resolution of DL-dansylserine on ami-
nopropyl- and allylterguride CSPs. Chromatographic conditions:
column size, 150 mm × 4.6 mm i.d.; eluent, 0.02 M potassium
phosphate buffer (pH 3.6)/acetonitrile (6:4 v/v); flow rate, 0.8 mL/
min; detection, UV at 254 nm; room temperature.
To study the enantioselectivity mechanism of the chiral
packing, the retention and resolution were examined for a series
of amino acid derivatives (Figure 2) as functions of the mobile
phase parameters (viz., pH, content and nature of organic modifier,
ionic strength of the buffer), the temperature, and the substituent
groups.
Effect of pH. Two equilibria are influenced by the pH of the
buffer: protonation of basic groups present in both the chiral
selector and the analytes and dissociation of the carboxylic groups
of the analytes.
The capacity factor (k′) and the enantioselectivity (R) values
as functions of the pH of phosphate buffer are reported in Figure
3, parts a and b, respectively. The plots show that k′ and R values
increase with increasing pH (until 4), corresponding to a higher
dissociation of the carboxylic functions and, consequently, to a
stronger stereoselective “ion pairing” effect exerted by the
Figure 2. Molecular structures of the examined amino acid deriva-
tives: I, dansyl; II, N-(3,5-dinitrobenzoyl); III, N-(2,4-dinitrophenyl);
IV, N-benzoyl; and V, â-naphthoyl.
selector. In the pH range between 4 and 5, the enantioselectivity
remains constant, although the retention rapidly decreases. At
pH above 5, a decrease in the enantioselectivity was also observed,
probably due to the lower protonation of the nitrogen groups
present in the selector. Figure 3b shows that better resolution is
achieved within the pH range 3.0-6.0, in agreement with the
(7) Trost, B. M., Fleming, I., Eds. Comprehensive Organic Synthesis; Pergamon
Press: Oxford, 1991; Vol. 4, p 770.
Analytical Chemistry, Vol. 68, No. 7, April 1, 1996 1193

Figure 3. Influence of pH on (a) retention (ln k′ ) and (b) enantio-
L
meric separation factors (R) of the following Dns-amino acid deriva-
tives: valine (1), serine (2), aspartic acid (9), and tryptophan (b).
Chromatographic conditions: column, 150 mm × 4.6 mm i.d., packed
with silica gel based with allylterguride; eluent, acetonitrile/0.05 M
potassium phosphate buffer (50:50 v/v); flow rate, 0.8 mL/min;
detection, UV at 254 nm; room temperature.
Figure 4. Influence of the organic modifier content (acetonitrile) on
(a) retention (k′) and (b) enantiomeric separation factors (R) of
tryptophan (9) and serine (b) dansyl derivatives. Eluent, 0.05 M
potassium phosphate at pH 3.0. Other chromatographic conditions
as in Figure 3.
previously reported findings.8
Organic Modifier Effect. Terguride-based stationary phases
act on a series of structurally similar compounds (Dns-AA) as a
reversed-phase chromatographic material. This behavior is il-
lustrated in Figure 4, which shows the retention sequence that
parallels the hydrophobicity of the compounds.9 Plots of k′ and
R values versus acetonitrile content show that, in the range 10-
90% acetonitrile, retention decreases nonlinearly as the amount
of organic modifier increases. This indicates that more than one
interaction contributes to the retention mechanism. Indeed, the
ergoline skeleton, having nonpolar moieties and positively charged
nitrogen atoms, retains the samples by a combination of hydro-
phobic and ionic exchange modes. The anomalous retention trend
observed at high acetonitrile contents (95%) may be ascribed to
a predominance of the electrostatic attraction in this region, which
is favored by decreasing of the aqueous polar solvation shell
around the interacting groups.10 It should be noted that high
organic content in the mobile phase can also induce structural
changes,11 giving rise to rearrangements of the diastereoisomeric
complexes. Accordingly, changes in retention and enantiodis-
crimination have to be taken into account. Thus, both the
electrostatic and the hydrophobic interactions determine a con-
stant enantioselectivity in a wide range of acetonitrile levels.
The effect of the nature of the organic modifier has also been
investigated. Solvents with different proton-acceptor capacity and
eluent strength,12 such as MeOH, THF, and ACN, have been
considered. Table 4 summarizes the capacity and selectivity
(10) Helferich, F., Ed. Ion Exchange; McGraw-Hill Book Co.: New York, 1962; p
(8) Castellani, L.; Flieger, M.; Sinibaldi, M. J. Liq. Chromatogr. 1 9 9 4 , 17, 3695.
(9) Melander, W. R.; Horva`th, Cs. In High Performance Liquid Chromatography;
Horva`th, Cs., Ed.; Academic Press: New York, 1980; Vol. 2, p 113.
161.
(11) Martin, D.; Weise, A.; Niclas, H. J. Angew. Chem., Int. Ed. Engl. 1 9 6 7 , 6,
318.
1194 Analytical Chemistry, Vol. 68, No. 7, April 1, 1996

Table 4. Capacity (k′) and Enantioselectivity (r)
Factors for a Series of Dansyl Derivatives as a
Function of the Organic Modifiera
phosphate
buffer/ ACN
phosphate
phosphate
buffer/ MetOHb
buffer/ THFb
Dns-AA
k′
D
R
k′
D
R
k′
D
R
Dns-Val
Dns-Ser
Dns-Trp
Dns-Asp
2.45
3.00
7.64
7.09
1.18
1.32
1.20
1.44
8.06
8.67
36.2
16.5
1.20
1.34
2.31
1.30
0.66
0.97
1.37
1.24
1.0
1.0
1.0
1.17
a
Chromatographic conditions: column, 150 mm × 4.6 mm i.d.;
buffer, 0.05 M potassium acetate (pH 3.0); flow rate, 0.8 mL/ min;
detection, UV at 254 nm; room temperature. b Values obtained with a
flow rate of 0.6 mL/ min.
factors for a number of Dns derivatives as a function of the organic
modifier. The eluent strength was THF < ACN < MeOH. A
significant difference in the selectivities was observed between
ACN and MeOH only in the case of Trp, while the use of THF
did not allow any resolution for Dns-DL-Val, Dns-DL-Ser, and Dns-
DL-Trp.
Ionic Strength of the Buffer. A study of the effect of ionic
strength on the retention and selectivity in the range 0.01-0.05
M at constant pH and organic solvent percentage (Figure 5) shows
that, at a certain buffer strength, there exists a minimal R value
exhibited by the CSP. The data may be interpreted as a net effect
of the buffer strength on the two main types on binding interac-
tions. The rapid decrease of retention with buffer concentration
up to 0.03 M should mainly reflect the competition of ion balancing
by the mobile phase with the electrostatic attraction, while the
slower decrease in k′ should be ascribed to the increasing role of
binding in terms of hydrophobic interaction, i.e., by a “salting out”
effect.13 Indeed, the plots demonstrate the role of this last
interaction in stabilizing the diastereoisomeric complexes.
Temperature Effect. High column temperatures are com-
monly used in HPLC to improve the rate of mass transfer and to
reduce the elution times. Since in the present system both
equilibria controlling the retention are temperature dependent,
the influence of this parameter was also investigated. The effect
of temperature on the retention and selectivity for some Dns-amino
acids with the same mobile phase composition was examined at
25, 35, 45, 55, and 65 °C (data not reported). As expected,14
increasing temperature brought about an overall linear decrease
in the retention and a reduction of the enantioselectivity, which
could probably be ascribed to changes of the structural conforma-
tion of the chiral selector and to a greater degree of freedom of
simultaneous bonding in the diastereomeric adsorbate.15
Solute Structural Effect. The structure of the Dns-amino
acids strongly affects both retention and selectivity, the elution
order paralleling the hydrophobicity of the side chain. Thus,
aromatic amino acids (Phe, Trp) are more strongly retained than
aliphatic ones (Val, Ala). A different behavior is evident for polar
amino acid derivatives (Glu, Asp, Ser, Thr) that can interact with
the ureo chain of the chiral selector by hydrogen bonding (Ser,
Figure 5. Influence of the buffer ionic strength on (a) retention (ln
L
k′ ) and (b) enantiomeric separation factors (R) of Dns derivatives of
valine (1), serine (2), aspartic acid (9), and tryptophan (b). Eluent,
acetonitrile/potassium phosphate buffer at pH 3.0 (50:50 v/v). Other
conditions as in Figure 3.
Thr) or by Coulombic attraction (Glu, Asp). These additional
interactions reflect the increased enantioselectivity and, at a low
concentration of the buffer, an inversion of the elution sequence
between Asp and Trp (see Figure 5). Two possible diastereo-
isomeric associations between terguride and the (I)
D
- (more
retained) and (II)
Figure 6.
L-forms of Dns-derivatives are illustrated in
Chiral resolution was obtained for all the examined derivatives,
with the exception of proline, in which the rigid ring present in
the molecule probably hinders the formation of differentiated
“stacked” adducts with the selector. The enantiomeric separations
of a mixture of four Dns derivatives by isocratic elution is shown
in Figure 7. However, elutions with pH or organic solvent gradient
could be used to improve the enantiomeric analysis of more
complex mixtures. The present method may be of interest in the
study of racemization reactions,16 because it offers the ability of
resolving aspartic acid (R ) 1.34) enantiomers in a column zone
clear from other constituents of proteins.
(12) Karger, B. L.; Gant, J. R.; Hartkopf, A.; Weiner, P. H. J. Chromatogr. 1 9 7 6 ,
128, 65.
(13) Hjerte`n, S. Methods Biochem. Anal. 1 9 8 1 , 27, 89.
(14) Majors, R. E. In High Performance Liquid Chromatography; Horva`th, Cs.,
Ed.; Academic Press: New York, 1980; Vol. 1, p 107.
(15) Pirkle, W. H.; Tsipouras, A. J. Chromatogr. 1 9 8 4 , 291, 291.
To obtain additional information on the recognition mechanism,
the influence of the derivatizing group on the retention and
(16) Elster, H.; Gil Av, E.; Weiner, S. J. Archaeol. Sci. 1 9 9 1 , 18, 605.
Analytical Chemistry, Vol. 68, No. 7, April 1, 1996 1195

Table 5. Enantioselectivity Factors (r ) k′ /k′ ) for a
D
Series of Amino Acids as a Function of the
Derivatizing Groupa
L
derivatizing group
amino acid
Dns
DNB
1.24
DNP
benzoyl
1.00
naphthoyl
Val
Ser
Trp
Ala
Glu
1.25
1.35
1.22
1.26
1.30
1.30
1.10
1.94
1.00
1.00
1.44
1.00
1.38
1.26
1.20
1.35
1.00
1.15
a
Mobile phase: 0.05 M phosphate buffer (pH 3.0)/ acetonitrile (1:1
v/ v). Other conditions as in Table 3.
Figure 6. Possible diastereoisomeric adducts between terguride
and the (I) D- and (II) L-forms of a Dns-amino acid derivative.
kB′ NZ < k′DNB < kN′ PT < k′Dns < k′DNP
It appears that resolution has to be ascribed mainly to the
ability of the compound to favor the formation of “stacking”
adducts upon the ergoline skeleton of the selector, involving
simultaneous Coloumbic and π-π interactions. The aromatic
moieties in the molecules of dansyl, N-benzoyl, â-naphthoyl, and
N-(3,5-dinitrobenzoyl) derivatives possess a lower freedom of
rotation around the bond axis than the N-(2,4-dinitrophenyl) one,
the amide bond providing two fixed directions for associations
with the selector. Consequently, the reciprocal position of the
π-electron acceptor groups in the sample and the imidazole (π-
electron donor) in the selector with respect to the chiral centers
seems to be critical for the resolution. Thus, N-benzoyl and
â-naphthoyl derivatives are not resolved, with the exception of
tryptophan, which can provide π-π interactions with its side chain.
Indeed, resolution of all the dinitro derivatives confirms that the
enantiodiscriminative process is in good agreement with the NMR
assumptions.2
ACKNOWLEDGMENT
This work was supported in part by the Project Chimica Fine
II (CNR, Rome). The authors are grateful to Ing. V. Havlicek
(Institute of Microbiology, Academy of Sciences of the Czech
Republic, Prague) for measuring and interpretating EI MS data
and to Mr F. Pezzotti for technical help. Thanks are also due to
Mr. F. Dianetti and L. Petrilli for the elemental analysis.
Figure 7. Enantiomer separation of a standard mixture of Dns-
amino acid derivatives.
Received for review July 13, 1995. Accepted December
18, 1995.X
selectivity has also been investigated. Table 5 summarizes the
selectivity (R) factors of five (Dns, DNB, DNP, BNZ, and NPT)
derivatives of a series of amino acids. The retention order of the
amino acids, determined on the basis of the derivatizing group,
was
AC950698C
X Abstract published in Advance ACS Abstracts, February 1, 1996.
1196 Analytical Chemistry, Vol. 68, No. 7, April 1, 1996