Retention and selectivity of teicoplanin stationary phases after copper complexation and isotopic exchange

Article abstract of DOI:10.1021/ac010444t

Teicoplanin is a macrocyclic glycopeptide that is highly effective as a chiral selector for LC enantiomeric separations. Two possible interaction paths were investigated and related to solute retention and selectivity: (1) interactions with the only teicoplanin amine group and (2) role of hydrogen bonding interactions. Mobile phases containing 0.5 and 5 mM copper ions were used to try to block the amine group. In the presence of copper ions, it was found that the teicoplanin stationary phase has a decreased ability to separate most underivatized racemic amino acids. However, it maintained its ability to separate enantiomers that were not α - amino acids. It is established that there is little copper - teicoplanin complex formation. The effect of Cu²⁺ on the enantioseparation of some α - amino acids appears to be due to the fact that these solutes are good bidentate ligands and form complexes with copper ions in the mobile phase. Isotopic exchange with deuterium oxide was performed using acetonitrile - heavy water mobile phases. It was found that the retention times of all amino acids were lower with deuterated mobile phases. The retention times of polar or apolar molecules without amine groups were higher with deuterated mobiles phases. In all cases, the enantio-selectivity factors were unaffected by the deuterium exchange. It is proposed that the electrostatic interactions are decreased in the deuterated mobile phases and the solute-accessible stationary-phase volume is somewhat swollen by deuterium oxide. The balance of these effects is a decrease in the amino acid retention times and an increase in the apolar solute retention time. The enantio-selectivity factors of all of the molecules remain unchanged because all of the interactions are changed equally. We propose a new global quality criterion (the E factor) for comparing and evaluating enantiomeric separations.

Full text of DOI:10.1021/ac010444t

Anal. Chem. 2001, 73, 5499-5508
Retention and Selectivity of Teicoplanin Stationary
Phases after Copper Complexation and Isotopic
Exchange
Alain Berthod,*,† Alain Valleix,‡ Veronique Tizon,† Estelle Leonce,‡ Celine Caussignac,‡ and
Daniel W. Armstrong§
Laboratoire des Sciences Analytiques, CNRS, Universite´ de Lyon 1, 69622 Villeurbanne, France,
CEA-Saclay, SMM, Bat. 547, 91191 Gif-sur-Yvette, France, and Department of Chemistry, Gilman Hall,
Iowa State University, Ames, Iowa 50011-3111
science1,2 in the latter part of the 20th century. When liquid
chromatography is used, three basic experimental approaches are
possible: (i) the enantiomeric pair is derivatized with a pure
enantiomer to form diastereoisomers that can be separated by
classical chromatography, (ii) a chiral additive can be added to
the mobile phase to form chiral complexes that can be separated
by a classical stationary phase, and (iii) chiral stationary phases
(CSP) are prepared. This latest solution proved to be the most
useful. It is now the most common way to separate enantiomeric
pairs.
Teicoplanin is a macrocyclic glycopeptide that is highly
effective as a chiral selector for LC enantiomeric separa-
tions. Two possible interaction paths were investigated
and related to solute retention and selectivity: (1) interac-
tions with the only teicoplanin amine group and (2 ) role
of hydrogen bonding interactions. Mobile phases contain-
ing 0 .5 and 5 mM copper ions were used to try to block
the amine group. In the presence of copper ions, it was
found that the teicoplanin stationary phase has a de-
creased ability to separate most underivatized racemic
amino acids. However, it maintained its ability to separate
enantiomers that were not r-amino acids. It is estab-
lished that there is little copper-teicoplanin complex
formation. The effect of Cu2 + on the enantioseparation of
some r-amino acids appears to be due to the fact that
these solutes are good bidentate ligands and form com-
plexes with copper ions in the mobile phase. Isotopic
exchange with deuterium oxide was performed using
acetonitrile-heavy water mobile phases. It was found that
the retention times of all amino acids were lower with
deuterated mobile phases. The retention times of polar
or apolar molecules without amine groups were higher
with deuterated mobiles phases. In all cases, the enantio-
selectivity factors were unaffected by the deuterium
exchange. It is proposed that the electrostatic interactions
are decreased in the deuterated mobile phases and the
solute-accessible stationary-phase volume is somewhat
swollen by deuterium oxide. The balance of these effects
is a decrease in the amino acid retention times and an
increase in the apolar solute retention time. The enantio-
selectivity factors of all of the molecules remain un-
changed because all of the interactions are changed
equally. We propose a new global quality criterion (the E
factor) for comparing and evaluating enantiomeric separa-
tions.
Chiral stationary phases can be prepared using (i) naturally
occurring chiral molecules, (ii) semisynthetic chiral molecules
(i.e., natural molecules that have been synthetically altered), and
(iii) totally synthetic chiral molecules. In general, the enantio-
selective separation mechanisms are best understood for the
smaller chiral selectors of known structure and limited functional-
ity. This includes molecules from all three of the aforementioned
classes. For example, some natural amino acids, like proline, make
effective ligand-exchange CSPs, separating chiral molecules that
can act as bidentate ligands to Cu2+ 3,4
Amino acid derivatives,
.
alkaloids, and other derivatized small molecules are among the
more successful π-π-complex CSPs.5-9 In addition, there are
totally synthetic π-complex CSPs. Cyclodextrin-based CSPs can
consist either of the natural molecule or the derivatized macro-
cycle.1 Larger, often polymeric, chiral molecules often have
undetermined structures and large numbers of interaction sites.10-13
Frequently, they have broad enantioselevities, but a basic under-
standing of their chiral recognition mechanism is limited. These
(1) Armstrong, D. W. LC-GC 1 9 9 7 , (May) 20.
(2) Maier, W. M.; Franco, P.; Lindner, W. J. Chromatogr. A 2 0 0 1 , 906, 3.
(3) Davankov, V. A.; Semechkin, A. V. J. Chromatogr. 1 9 7 7 , 141, 313.
(4) Davankov, V. A. In Giddings, J. C., Cazes, J., Brown, P. J. (Eds.); Advances
in Chromatography; Marcel Dekker: New York, 1983, Vol. 22, pp 71-103.
(5) Klemm, L. H.; Reed, D. J. Chromatogr. 1 9 6 0 , 3, 364.
(6) Mikes, F.; Boshart, G. J. Chromatogr. 1 9 7 8 , 149, 455.
(7) Pirkle, W. H.; Hyun, M. H. J. Chromatogr. 1 9 8 5 , 322, 309.
(8) Oi, N.; Kitahara, H. J. Chromatogr. 1 9 8 3 , 265, 117.
(9) La¨mmerhofer, M.; Svet, F.; Fre´chet, J. M. J.; Lindner, W. Anal. Chem. 2 0 0 0 ,
72, 4623.
(10) Lindner, W.; Mannschreck, A. J. Chromatogr. 1 9 8 0 , 193, 308.
(11) Okamoto, Y.; Kawashima, M.; Yamamoto, K.; Hakada, H. Chem. Lett. 1 9 8 4 ,
739.
(12) Hernansson, J. J. Chromatogr. 1 9 8 3 , 269, 71.
(13) Allenmark, S. G.; Bomgren, B.; Boren, H. J. Chromatogr. 1 9 8 3 , 269, 63.
The development of broadly applicable, high-efficiency enan-
tiomeric separations is one of the great successes of separation
† Universite´ de Lyon.
‡ CEA-Saclay.
§ Iowa State University.
10.1021/ac010444t CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/11/2001
Analytical Chemistry, Vol. 73, No. 22, November 15, 2001 5499

Figure 1. The teicoplanin molecule with its 4 macrocycles, its 7 aromatic rings, 4 phenolic groups, 6 amido groups, 3 ether links, 2 chlorine
atoms, a free primary amine, a carboxylic acid group, and 3 sugar moieties with 10 primary or secondary alcoholic groups and 2 amido groups,
one of which bears a 9-carbon-atom alkyl chain. Twenty-five protons can be exchanged for deuterium atoms; they are noted as X’s on the
figure.
polymeric chiral selectors can be naturally occurring molecules,
such as proteins; semisynthetic molecules, such as derivatized
carbohydrates and cross-linked derivatized tartaric acid;10,14 or
totally synthetic polymers.15
bonding interactions that are thought to be essential for chiral
recognition. These two types of interactions were evaluated using
copper complexation19 and deuterium isotope exchange, respec-
tively.
The macrocyclic glycopeptide antibiotics have recently emerged
as one of the more useful classes of chiral selectors for HPLC
enantioseparations.16 Although only of intermediate size, these
molecules most closely rival proteins in their diversity of structure,
number, functional groups, and variety of possible interactions.
Yet the macrocyclic glycopeptides are sufficiently small that their
structure is known, and they can be thoroughly examined by most
conventional chemical diagnostic methods. The CSP based on
teicoplanin has found widespread use.17,18 The teicoplanin-contain-
ing columns were able to resolve all naturally occurring amino
acids in their native form with common methanol-water mobile
phases.17 They were also able to separate a wide variety of other
amino acids and small peptides and most other enantiomers
containing a carboxylic group attached to the stereogenic cen-
ter.1,17
EXPERIMENTAL SECTION
The Teicoplanin Chiral Stationary P hase. Teicoplanin,
C88H97Cl2N9O33, MW 1878, contains a heptapeptide aglycone with
seven aromatic rings, four phenolic groups, six amido groups,
three ether links, and two chlorine atoms. Three saccharide units
are attached to the heptapeptide. They bear 10 primary or
secondary hydroxyl groups and two amido groups, one of which
bears a nine (or ten)-carbon atom alkyl chain. Figure 1 shows
the structure of the prevalent teicoplanin isomer, the A2-2 form,
making up ∼85% of the natural product. The four other teicoplanin
isomers have the same aglycone core. They differ by the alkyl
chain on the acyl glucosamine unit. A2-2 bears an 8-methyl-
nonanoyl chain. 8-Methyldecanoyl, 9-methyldecanoyl, n-decanoyl,
and (Z)-4-decanoyl variations are found on the alkyl chain of the
minor teicoplanin isomers.20 Teicoplanin, as all macrocyclic
glycopeptides, is patented for chiral separations (agreement
between University of Missouri, D.W. Armstrong, and the Astec
Company, Whippany, NJ 07981). The teicoplanin coverage of the
silica particles is between 0.35 and 0.45 µmol/ m2.21 The average
teicoplanin content of a 25-cm column is ∼0.4 mole or 750 mg.
Two 25-cm Chirobiotic T columns (Astec), serial number (S/ N)
A155-6-C and A155-6-D, were used for the copper ion study, and
The goal of this work was to investigate two particular
interactions involved in the teicoplanin retention mechanism and
to evaluate the impact on the enantioselectivity factors of various
solutes. The first point focuses on the only primary amine group
of the teicoplanin molecule. The second part focuses on hydrogen
(14) Allenmark, S. G.; Anderson, S.; Moller, P.; Sanchez, D. Chirality 1 9 9 5 , 7,
248.
(15) Okamoto, Y.; Honda, S.; Okamoto, I.; Yuki, H.; Murata, S.; Naori, R.; Tayaka,
H. J. Am. Chem. Soc. 1 9 8 1 , 103, 6971.
(16) Armstrong, D. W.; Tang, Y.; Chen, S.; Zhou, Y.; Bagwill, C.; Chen, J. R.
Anal. Chem. 1 9 9 4 , 66, 1473.
(19) Nair, U.; Chang, S. C.; Armstrong, D. W.; Rawjee, Y. Y.; Eggleston, D. S.;
McArdle, J. V. Chirality 1 9 9 6 , 8, 590.
(17) Berthod, A.; Liu, Y.; Bagwill, C.; Armstrong, D. W. J. Chromatogr. A 1 9 9 6 ,
731, 123.
(18) Armstrong, D. W.; Liu, Y.; Ekborg-Ott, K. H. Chirality 1 9 9 5 , 7, 474.
(20) Product 9269. The Merck Index, 12th ed.: Merck & Co.: 1996, p 1559.
(21) Berthod, A.; Chen, X.; Kullman, J. P.; Armstrong, D. W.; Gasparrini, F.;
D’Aquarica, I.; Villani, C. Anal. Chem. 2 0 0 0 , 72, 1767.
5500 Analytical Chemistry, Vol. 73, No. 22, November 15, 2001

two other 25-cm columns, S/ N 111-21 and 134-33 were used for
the isotopic exchange study.
to work with a higher copper ion concentration because of the
UV absorbency of the mobile phase. The copper ion adsorption
seems reversible. When the column is washed with a copper-free
mobile phase, most of the adsorbed copper ions are removed from
the CSP.
Materials. Chemicals. Acetic acid, triethylamine and most of
the racemic amino acids and pure enantiomers were obtained from
the Sigma-Aldrich-Fluka Co. (Saint Quentin Fallavier, France).
They were used as received. The test molecules are listed in Table
1, along with their structures and properties. Vancomycin was
obtained from Sigma Chemical Co. (St Louis, MO). Teicoplanin
was a gift donated by Marian Merrill Dow Research Institute
(Cincinnati, OH). Copper acetate was from Prolabo (Paris,
France). Methanol and acetonitrile were from SDS (Marseille-
Peypin, France). Deuterium oxide (or heavy water) and CH3OD
were products of Euroisotop (Saint Aubin, France), a subdivision
of the French Center for Atomic Energy (CEA, Gif sur Yvette,
France).
Apparatus. A Shimadzu LC6A chromatograph was used. It
included a LC-6A or a LC-10AS pump, a SPD-6A UV detector, and
a CR-5A integrator. A Hitachi U-2000 UV-vis spectrophotometer
was used for the study of the complexation of copper with
vancomycin or teicoplanin.
Protocol. For the copper ion experiments, methanol-water,
60-40% v/ v, was used at 1 mL/ min flow rate. copper acetate (0.5
or 5 mM) was added for teicoplanin complexation. The mobile
phases were buffered with 0.25%, v/ v, triethylamine (18 mM) and
the necessary amount of acetic acid to obtain a pH of 4.1 or 7.1.
For the isotopic exchange, different mobile phases were used:
acetonitrile-water, 80-20% v/ v or 15-85% v/ v, and methanol-
water, 20-80% v/ v. The corresponding mobile phases were
prepared with acetonitrile and heavy water and CH3OD and heavy
water. No additive was added to the deuterated or normal mobile
phases (no buffer, no salts).
Solute Retention and Enantioselectivity. Table 2 lists the
chromatographic results obtained with the methanol-water 60-
40%, v/ v, mobile phase both with and without copper ions.
Different behaviors were noted. The native R amino acids (such
as cysteine, tryptophan, or phenylalanine) usually showed a
significant increase in their retention volume accompanied by a
decrease in the corresponding enantioselectivity and resolution
factors. The separation of the enantiomers of cysteine was not
possible when copper ions were present in the mobile phase and
its retention volume was twice that of the most-retained peak with
the copper-free mobile phase (Table 2). Other compounds, such
as N-acetyl-3-fluorophenylalanine, carnitine, or warfarin, do not
show any change when copper ions are added to the mobile phase.
Some compounds show an intermediate behavior with a slightly
increased retention volume and a slightly reduced enantioselec-
tivity and resolution factor.
To have a single global parameter to compare the chromato-
graphic separation of enantiomers, we define an “E”, or effective-
ness, term as
E ) RRs/ k1
(1)
where R is the enantioselectivity factor, Rs is the enantio-
resolution, and k1 is the retention factor of the first eluted
enantiomer. E is an all-inclusive quality criterion. It is designed
for just a quick screening of chromatographic variations. Rs is by
itself a function of the k and R parameters. But often, changes in
R and k factors were associated with constant Rs factors because
of changes in efficiency. The E factor has no thermodynamic or
physical meaning other than a quality criterion. High values of E
indicate short analysis times with well-separated peaks. E values
are useful in comparing enantioseparations on different columns
or using different conditions with the same column. If the value
of E increases, then the chromatographic quality increases and
vice versa. Separations with higher E values are the most effective
separations.
Figure 2 shows the relative value of the E factor (i.e., the E
factor for the given mobile phase divided by the E factor of the
copper free mobile phase) for all of the compounds of Table 2.
Figure 2 clearly shows that the chromatographic figures of merit
of all amino acids but two are dramatically degraded by the
presence of copper ions in the mobile phase. The two exceptions
are aspartic acid and N-acetyl-3-fluorophenylalanine. The other
result shown in Figure 2 is that the non-amino acid compounds
are unaffected or much less affected by the copper ions in the
mobile phase. The 40% decrease observed with p-hydroxymandelic
acid seems significant: mandelic acid is an R-hydroxy carboxylic
acid that is as effective as R amino acids at chelating Cu2+ in the
mobile phase. However, the error on the very low retention factor,
k1 < 0.2, of this compound is high, which results in a high variation
of its E factor.
RESULTS AND DISCUSSION
Copper II Ion Complexation. Two Chirobiotic T columns
containing CSPs from the same synthetic batch were used in this
study. One was exposed to copper containing mobile phases, the
second one served as a reference. All mobile phases contained
60%, v/ v, methanol and were buffered at pH 4.1 or 7.1 by adding
0.25%, v/ v, triethylamine (18 mM TEAA) and adjusting the pH
by adding acetic acid dropwise.
Copper Ion Adsorption. The amount of copper II ion adsorbed
by the Chirobiotic T column was estimated following their 230-
nm UV absorbency. The column was first equilibrated with a
copper-free methanol-water mobile phase. Next, the pumping
system was disconnected from the column and rinsed with the
0.5 mM copper containing mobile phase. The column was
reconnected, and the recorder was started. The dead volume of
the column was measured to be 2.4 mL. The amount of copper
ion contained in the dead volume was deducted from the amount
of copper found to be adsorbed by the CSP. The corrected copper
amount adsorbed at pH 4.1 was only 9 µmol for the column. This
amount is very low; it corresponds to only one copper ion for 45
teicoplanin molecules. This result is very different from what was
obtain with the vancomycin chiral selector that formed a strong
1:1 complex with copper ions at pH 5.19 Later, the same experiment
was performed using the 5 mM Cu II pH 4.1 methanol-water,
60-40% v/ v, mobile phase. The amount of adsorbed copper ion
was ∼50 µmol. This corresponds to about one copper ion for eight
teicoplanin molecules. This amount is still low. It is not possible
Amine Groups and Copper Complexation. In a recent work, we
demonstrated the role of the amine group of vancomycin, another
Analytical Chemistry, Vol. 73, No. 22, November 15, 2001 5501

Table 1. Enantiomeric Pairs Used as Test Solutes
a
The general structure of amino acids is +NH3-CHR-COO- with the R group structure listed.
5502 Analytical Chemistry, Vol. 73, No. 22, November 15, 2001

Table 2. Chromatographic Results Obtained with the Copper II Containing Mobile Phasesa
MeOH-H2O 60-40 + copper 0.5 mM
+ copper 5 mM
solute
k1′
R12
Rs
E
k′
R12
Rs
E
k′
R12
Rs
E
1
1
alanine
asparagine
aspartic acid
cysteine
dopa
glutamic acid
leucine
methionine
phenylalanine
N-acetyl-3-fluorophenylalanine
tryptophan
m-tyrosine
valine
carnitine
coumachlor
5-methyl-5-phenyl hydantoin
p-hydroxy mandelic acid
warfarin
0.73
0.83
0.39
2.14
0.83
0.50
0.70
0.83
0.99
0.65
1.15
0.85
0.65
2.12
1.41
0.58
0.15
0.69
1.9
1.6
1.5
1.2
1.9
1.7
2.2
2.1
1.5
5.4
1.5
2.3
1.7
1.05
1.2
1.9
4.7
1.1
4.2
2.9
0.6
0.5
3.7
2.0
4.4
4.3
2.5
7.3
2.4
4.1
3.2
0.4
1.2
2.8
3.8
0.5
11.0
5.56
0.77
1.04
0.49
3.96
1.26
0.60
0.81
1.00
1.83
0.58
2.94
1.66
0.74
2.15
1.19
0.54
0.19
0.61
1.8
1.5
1.4
1.0
1.6
1.6
2.0
1.9
1.2
5.4
1.1
1.7
1.6
1.05
1.2
1.8
3.8
1.1
3.5
1.2
0.7
0.0
1.3
1.2
2.7
2.4
0.4
6.9
0.2
1.4
2.0
0.5
1.2
2.9
3.5
0.4
8.26
1.72
2.01
0.0
1.63
3.10
6.58
4.63
0.26
0.79
1.17
0.55
4.91
1.45
0.67
0.92
1.16
2.25
0.60
2.79
2.02
0.81
2.13
1.13
0.55
0.19
0.62
1.8
1.5
1.4
1.0
1.5
1.5
1.9
1.9
1.2
5.4
1.2
1.6
1.5
1.05
1.3
1.8
4.0
1.1
2.7
1.1
1.0
0.0
1.1
1.1
2.3
1.9
0.3
7.3
0.2
1.4
1.2
0.4
1.4
2.6
3.6
0.6
6.19
1.36
2.48
0.0
1.14
2.45
4.76
3.04
0.16
2.29
0.275
8.59
6.72
13.5
10.7
3.67
60.8
3.10
10.8
8.18
0.20
1.01
9.00
125
0.81
63.8
66.1
0.08
1.42
4.20
0.22
1.23
9.85
0.08
1.10
2.25
0.20
1.65
8.68
71.8
0.73
75.8
1.09
a
Column 1 ) Chirobiotic T (S/ N: A-155-6C) 25 cm, 4.6 mm i.d. All mobile phases contain 18 mM TEAA buffer at pH 4.1.
Figure 2. The E ratio () RRs/k′1) for the compounds of Table 2 showing the effect of copper ions on the chromatographic figures of merit
obtained with the Chirobiotic T column 1. Relative values, reference mobile phase: MeOH-H2O, 60-40% v/v, pH 4.1.
macrocyclic glycopeptide antibiotic bearing some similarities with
teicoplanin,22 by blocking it using copper II ions.19 The first marked
difference between vancomycin and teicoplanin is that the former
was able to establish a strong 1:1 complex with copper II ions.19
Teicoplanin does not appear to form such a complex. This was
checked by studying the copper complexation with the free
macrocyclic glycopeptides by UV-vis spectroscopy. Small amounts
of dimethyl sulfoxide (DMSO) had to be used, because the
glycopeptides were not soluble at high concentrations in pure
water. A 0.01 M Cu2+ ion, 0.01 M vancomycin solution in DMSO-
(22) Gasper, M. P.; Berthod, A.; Nair, U. B.; Armstrong, D. W. Anal. Chem. 1996,
68, 2501.
Analytical Chemistry, Vol. 73, No. 22, November 15, 2001 5503

Table 3. Effect of pH on the Chromatographic Resultsa
column 1, pH 4.1
column 1, pH 7.1
column 2, pH 7.1 Cu 5 mM
solute
k′
R12
Rs
E
k′
R12
Rs
E
k′
1
R12
Rs
E
1
1
alanine
0.76
0.87
0.40
2.41
0.51
0.73
0.86
1.03
0.74
1.19
0.90
0.65
2.11
1.51
0.58
0.15
0.65
1.9
1.6
1.5
1.1
1.7
2.2
2.1
1.5
5.4
1.5
2.2
1.7
1.05
1.2
1.9
5.1
1.1
4.2
2.8
1.1
0.6
1.7
4.6
4.6
2.3
7.8
2.1
4.3
2.9
0.5
1.7
2.8
3.0
0.6
10.3
5.07
1.20
1.66
0.35
2.0
1.0
1.4
1.0
1.5
1.0
1.0
1.0
6.4
1.0
2.1
1.8
1.1
1.2
2.0
5.1
1.2
1.3
0.0
0.2
0.0
0.6
0.0
0.0
0.0
6.9
0.0
0.9
1.9
0.6
1.2
3.9
3.8
0.4
2.12
0.0
0.80
0.0
1.87
0.0
0.0
0.0
72.4
0.0
0.71
1.88
0.39
0.90
11.7
118
1.86
19.4
1.65
1.33
>50
30.9
28.2
40.2
10.2
1.3
2.2
1.0
-
0.2
0.9
0.0
0.01
1.21
0.0
asparagine
aspartic acid
cysteine
glutamic acid
leucine
4.00
0.28
5.62
13.6
11.3
3.23
57.0
2.58
10.7
7.55
0.25
1.36
9.00
103
1.05
13.0
-
-
0.47
2.85
4.25
2.44
0.61
2.70
2.63
1.78
1.66
1.55
0.65
0.16
0.25
1.3
1.1
1.4
2.0
4.1
1.4
1.5
1.1
1.6
1.2
1.6
1.4
1.0
0.5
0.2
1.9
1.1
4.5
1.1
1.0
0.2
0.7
1.2
1.3
0.3
0.0
0.02
0.01
0.07
0.21
methionine
phenylalanine
N-acetyl-3-fluoro phenylalanine
tryptophan
m-tyrosine
valine
carnitine
coumachlor
5-methyl-5-phenyl hydantoin
p-hydroxy mandelic acid
warfarin
0.60
31.5
23.2
14.0
24.0
17.4
0.07
0.11
0.01
0.06
1.21
2.67
0.42
0.0
1.19
0.78
1.00
0.75
a
Mobile phase: methanol-water, 60-40% v/ v, with 18 mM TEAA buffer at indicated pH, 1 mL/ min. Column 1 ) Chirobiotic T, S/ N ) A-155-
6C, 25 cm, 4.6-mm i.d.; column 2 ) Chirobiotic T, S/ N ) A-155-6D, 25 cm, 4.6-mm i.d.
water, 20-80% v/ v, buffered at pH 4 with 0.02 M TEAA shows an
absorption band at 550 nm (purple color) at the side of the copper
ion absorption band (790 nm). The 550-nm band corresponds to
the vancomycin-copper complex.19 In 0.01 M Cu2+ and 0.02 M
teicoplanin at pH 4 with 0.02 M TEAA in 20%, v/ v, DMSO solution,
the teicoplanin solution shows only the copper ion absorption band
at 790 nm (blue color). Copper ion complexation with amino acids
was checked with glycine. A 0.01 M Cu2+ solution with 0.2 M
glycine (DMSO, 20% v/ v, pH 4 with 20 mM TEAA) shows a single
absorption band at 650 nm, which is for the Cu-glycine complex.
If the chiral recognition of amino acids is dramatically
decreased when copper ions are present, and these ions do not
form a complex with teicoplanin, it must be the copper-amino
acid association in the mobile phase that is responsible for the
loss of enantioselectivity. Davankov3,4,23 used such copper ion
complexation to separate amino acids by ligand exchange since
the early 1970s. In our case, however, the amino acid copper ion
complexation can be detrimental for chiral recognition by the
teicoplanin selector. Analytes that can act as bidentate ligands to
Cu2+, are the most affected by the addition of copper ion to the
mobile phase. The solute N-acetyl-3-fluorophenylalanine cannot
form a complex with a copper ion, because its amine group is
acetylated. Thus, its enantioselective separation is not affected by
the copper ions (Table 2 and Figure 2). Aspartic acid has two
carboxylic acid groups. The amino group may not be available
for copper complexation, because it is blocked by the second
carboxylic acid group of the molecule. Alternatively, copper
complexation may occur through the amine and one carboxylic
acid group, leaving the other to associate with the CSP. Carnitine
is a γ-amino acid (betaine). In addition, the amine group of
carnitine is a quaternary ammonium group that is less capable of
forming a copper complex than the primary amine group of the
R-amino acids. This may explain why the enantiorecognition of
carnitine by the teicoplanin selector is not affected by copper ions.
Conversely, p-hydroxymandelic acid behaves more like the R
amino acids because, as an R-hydroxy acid, it also chelates copper
II.
pH Effect. The ionization state of both the teicoplanin chiral
selector and the amino acid solutes is dependent on the mobile
phase pH. Mobile phases of methanol-water, 60-40% v/ v, with
or without 5 mM copper nitrate were buffered at pH 7.1 with 0.25%,
v/ v, triethylamine and acetic acid (18 mM TEAA). A second
Chirobiotic T column was used to show the chromatographic
changes due only to pH variations. Table 3 lists the results
obtained at pHs 4.1 and 7.1. The results listed for pH 4.1 should
be compared with those listed in Table 2 and obtained using
column 1 in the same conditions. A less than 10% variation is
observed, so the results obtained on column 2, not exposed to
copper ions, can be used to compare with results obtained with
column 1 used with the copper containing mobile phases (Table
3). The retention times of most amino acids, obtained with a pH
7.1 mobile phase, were much higher than the corresponding times
obtained with the pH 4.1 mobile phases. For all amino acids except
N-acetyl-3-fluorophenylalanine, the pH 7.1 mobile phase produced
lower resolution than the pH 4.1 mobile phase. This effect is
known17,18, and working at pH 4.1 is recommended for optimum
resolution, and pHs lower than 3.8 or higher than 8 should be
avoided, because these extremes decrease column stability.24
At pH 7.1, 5 mM copper II ion concentration was added to the
60-40%, v/ v, methanol-water mobile phase. The chromato-
graphic results obtained with column 1 are listed in Table 3 and
are compared to the copper free mobile phase using our defined
E criterion (eq 1). In all cases, a dramatic loss of resolution was
observed when copper ions were present in the mobile phase at
pH 7.1. For many compounds, the copper ions induce an increase
in the retention factors of ∼1 order of magnitude. This occurs for
alanine, cysteine, glutamic acid, leucine, methionine, tryptophan,
valine, and carnitine (Table 3). The chelated analytes behave
differently and possibly more like a cationic species with regard
to an increasingly more effective ion-exchange-type stationary
(23) Davankov, V. A. In Chiral Separations by HPLC; Krstulovic, A. M., Ed.; Ellis
Horwood: Chichester, 1989; Chapter 15, pp 446-475.
(24) Chirobiotic Handbook, 2nd ed.; Astec: Whippany, NJ 1 9 9 7 , freely obtained
on request at www.astecusa.com.
5504 Analytical Chemistry, Vol. 73, No. 22, November 15, 2001

phase. The carboxylic groups of the amino acid solutes are
negatively charged at pH 7.1, and the amino group is neutral.
Copper ions and such negatively charged amino acid solutes may
form a complex with a positive charge. This complex could be
more retained on the teicoplanin stationary phase, which happens
to have a negative character at pH 7.1 (Figure 1). This increase
in the retention factors increases somewhat the enantioselectivity
factor, but the resolution is significantly decreased (Table 3)
because of poor efficiency. The change in the ionization state of
the teicoplanin stationary phase has an essential relevance to its
chiral recognition capability, because the enantioselectivity and
resolution factors of the nonionizable solutes, coumachlor, methyl
phenyl hydantoin, and warfarin are also affected, but to a much
lesser extent, by the 4.1 to 7.1 pH change (Table 3). An increase
of the ionic strength of the mobile phase would decrease the
retention times of the chelated analytes. This is not likely to
change the enantioresolution factors.
Table 4. Deuterium Oxide Compared to Water
Properties
deuterium
percent
change
property (unit)
MW (Da)
water
oxide
18.02
0.0
20.03
3.8
+11.20
+1.4a
+0.37a
+10.7
+0.4
+2.8
+4.5
+2.05
- 2.5
+6.9
- 67
+0.4
mp (°C)
bp (°C)
100.0
0.997
18.07
1.000
1.436
9.717
80
101.4
1.104
18.14
1.028
1.501
9.917
78
d25 (g cm-3
)
mol vol (cm3 mol-1
)
spec. heat liq. (cal g-1 °C-1
heat fusion (kcal mol-1
heat vap. (kcal mol-1
)
)
)
dielectric constant (@ 20 °C)
ionization constant (@ 25 °C)
[OH-] or [OD-] (M)
refractive index
viscosity (103 N s m-2 or cP)
15 °C
20 °C
30 °C
40 °C
14.0
14.96
10-7
1.333
3.3 10-8
1.338
1.139
1.002
0.797
0.719
1.40
1.25
1.03
0.81
+23.0
+24.7
+29.2
+12.7
Copper Adsorption Reversibility. After completion of the experi-
ment with copper containing mobile phases, column 1 was rinsed
with ∼10 column volumes of distilled water (30 min @ 1 mL/
min), followed by 10 column volumes of pure methanol and 12 h
(overnight @ 0.1 mL/ min) with the initial methanol-water, 60-
40% v/ v, 0.25% triethylamine adjusted to pH 4.1 by acetic acid.
The retention and enantioselectivity factors of the rinsed column
1 were within 10% of those of column 1 before exposure to copper
ion (Table 2) and column 2 (Table 3). However, the resolution
factors were ∼30% lower than the initial values as a result of a
significant decrease (about -50%) of peak efficiency. It can be
concluded that the copper adsorption is for the most part
reversible, because there is no strong teicoplanin-copper II
complexation. But it is not possible to say if the efficiency loss is
due to copper exposure or normal wearing of the column.
The conclusion of the first part of this work is that, contrary
to what was obtained with the vancomycin chiral stationary
phase,19 it is not possible to use copper II ions to evaluate the
role of the teicoplanin primary amine in the chiral recognition
mechanism of amino acids. The positive results are that teicoplanin
does not form strong complexes with Cu2+ and that analytes with
free carboxylic groups are highly recognizable on the CSP.
Making the COOH group unavailable via Cu complexation greatly
reduces enantioselectivity.
Role of Hydrogen Bonding Assessed by Deuterium Iso-
topic Exchange. In the second part of this study, we investigated
the role of hydrogen bond interactions in the retention and chiral
recognition process of teicoplanin. We thought that many polar
hydrogen atoms of the chiral selectors could be exchanged for
deuterium atoms (X’s in Figure 1). The separation of racemic pairs
on such an isotopically modified chiral stationary phase should
be enlightening. Heavy water (D2O) mobile phases were used in
HPLC to perform MS detection25,26 or NMR.27 RPLC mobile phases
containing water enriched with deuterium oxide were used to
study the hydratation state of C18 bonded phases.28-29
a
Percent change on absolute temperature in K.
Deuterium Oxide Properties. Heavy water (deuterium oxide)
has the interesting property of being able to exchange deuterons
for hydrogen atoms with any functional group containing a labile
hydrogen atom. These functional groups include the -OH in
alcohols, phenols, and carboxylic acids and the -NH- in amines
and amides in neutral conditions.30 Bonds to deuterium are slightly
stronger than bonds to hydrogen, resulting in small but significant
changes in reaction rates. Table 4 compares the properties of
heavy and normal water. The boiling point of D2O is 101.4 °C,
slightly higher than that of H2O, although the atomic weight of
the former is 12% higher (20.03). The energy needed to separate
the deuterium oxide molecules (heat of vaporization) is only 2%
higher that the corresponding energy for water, although the D2O
molecules are 12% heavier. These values show that the inter-
molecular hydrogen bonds are slightly stronger than the inter-
D2O molecule “deuterium” bonds.
The ionization constant of D2O is one unit higher than that of
water, corresponding to a concentration of DO- at equilibrium 3
times lower in D2O () 3.3 × 10-8 M) than in H2O ([OH-] ) 10-7
M). This shows that the oxygen-deuterium σ bond is slightly
stronger than the classical OH bond. It is clear that the dissocia-
tion constant of any ionizable group (phenol, carboxylic acid, or
amine) will be different in the hydrogen form from that in the
deuterium form. The fact that pKa values were higher in D2O than
in H2O was already mentioned in the literature.31 Because this
point is not the subject of this study, we chose neither to buffer
nor to add any salt to the mobile phases.
Table 4 also shows that the viscosity of heavy water is ∼25%
higher than that of normal water. Then, the observed column back
pressure will be 25% higher at equal flow rate. For this reason, a
flow rate of 0.8 mL/ min was used for all of the experiments with
deuterated mobile phases and respective comparison experiments
with aqueous mobile phase.
(25) Edmonds, C. G.; Pomeranz, S. C.; Hsu, F. F.; McClosky, J. A. Anal. Chem.
1 9 8 8 , 60, 2314.
(26) Olsen, M. A.; Cummings, P. G.; Kennedy-Gabb, S.; Wagner, B. M.; Nicol,
G. R.; Munson, B. Anal. Chem. 2 0 0 0 , 72, 5070.
(27) Fukuhara, T.; Komatu, K.; Yamamoto, S.; Takamatu, T. Bunseki Kagaku
1 9 9 6 , 45, 421.
Column Equilibration. Figure 1 shows the exchangeable
protons of the teicoplanin molecule as X’s instead of Hs. Residual
(28) Marchall, D. B.; McKenna, W. P. Anal. Chem. 1 9 8 4 , 56, 2090.
(29) Bliesner, D. M.; Sentell, K. Anal. Chem. 1 9 9 3 , 65, 1819.
(30) Dharmasiri, K.; Smith, D. L. Anal. Chem. 1 9 9 6 , 68, 2340.
(31) Okafo, G. N.; Camilleri, P. J. Chromatogr. 1 9 9 1 , 549, 404.
Analytical Chemistry, Vol. 73, No. 22, November 15, 2001 5505

Table 5. Chromatographic Results Obtained with the Deuterated Mobile Phasesa
aqueous mobile phase deuterated mobile phase
∆G (kJ/ mol)
solute
k′
R12
Rs
E
k′
R12
Rs
E
exch. Hb
enantiomer
isotopic
1
1
leucinec
3.22
3.81
4.01
3.49
2.86
0.76
0.33
6.61
6.63
3.12
1.29
1.18
1.16
1.20
1.25
1.43
3.37
1.08
1.12
1.21
2.4
2.3
1.5
2.2
2.3
3.5
4.7
1.3
1.4
1.6
0.96
0.71
0.43
0.75
0.83
6.55
3.1
3.5
1.29
1.20
1.18
1.20
1.23
1.45
3.28
1.08
1.09
1.20
2.5
2.3
1.5
2.2
2.4
3.3
4.7
1.3
1.4
1.6
1.04
0.79
0.49
0.79
0.73
5.47
3
3
4
4
1
2
3
0
0
0
0.62
0.42
0.38
0.44
0.52
0.89
2.93
0.19
0.25
0.45
0.09
0.19
0.26
phenylalaninec
tryptophanc
3.58
3.35
3.71
0.88
0.36
8.54
8.63
3.66
tyrosinec
0.10
coumachlorc
-0.62
-0.36
-0.17
-0.61
-0.61
-0.39
5-methyl-5-phenyl hydantoind
p-hydroxy mandelic acide
2-methyl-4-phenyl-indanonee
naphththyl butyrolactonef
phenyl phthalideg
48.2
43.2
0.21
0.24
0.62
0.16
0.18
0.53
a
Column, Chirobiotic T; S/ N ) 134-33; flow rate, 0.8 mL/ min; no buffer added. b exch. H ) number of hydrogen atoms exchanged by the
solute for deuterium atoms. c Acetonitrile (ACN)-H2O 80-20% v/ v and ACN-D2O 80-20% v/ v. d Acetonitrile (ACN)-H2O 20-80% v/ v and ACN-
D2O 20-80% v/ v. e Methanol-H2O 20-80% v/ v and CH3OD-D2O 20-80% v/ v. f Acetonitrile (ACN)-H2O 5-95% v/ v and ACN-D2O 5-95% v/ v.
g Acetonitrile (ACN)-H2O 15-85% v/ v and ACN-D2O 15-85% v/ v.
silanols of the silica stationary-phase backbone also bear ex-
changeable protons. To equilibrate the column, the acetonitrile-
heavy water 80-20%, v/ v, mobile phase (or the selected deuter-
ated mobile phase) was percolated through the Chirobiotic T
column for 12 h (overnight) at 0.1 mL/ min (∼70 mL or ∼35
column volumes without recycling). The equilibration was con-
firmed by the constancy of the retention times of naphthyl
butyrolactone and the other enantiomers injected in triplicate. C18
columns were reported to be equilibrated in <15 min at 1 mL/
min.26 In this case, only silanols had to be isotopically exchanged.
Unfortunately, the equilibration time needed to return to the
hydrogenated state was never indicated. In our case, exchange
of deuterium atoms back to protons, it was found that the
equilibration time was approximately four times longer (2 days
with acetonitrile-water, 80-20% v/ v, at 0.1 mL/ min). Here again,
the retention time of naphthyl butyrolactone was used to follow
the isotopic exchange. These results also confirm that bonds to
deuterium are stronger than bonds to hydrogen, that is, with the
deuterium bonds forming faster than they are broken.
Solute Retention and Selectivity. Table 5 lists the results of the
separation of enantiomers performed with classical hydroorganic
mobile phases and the corresponding deuterated mobile phases
on the teicoplanin-bonded column. Three results were obtained:
(1) amino acids were eluted earlier with deuterated mobile phases
than with the corresponding aqueous mobile phases; (2) non-
amino acid compounds, either very polar (mandelic acid is
anionic), semipolar (coumachlor), or apolar (naphthyl butyro-
lactone), were eluted later with deuterated mobile phases; (3) in
all of cases that were studied, the enantioselectivity and resolution
factors were not affected by these changes in retention factors.
Figure 3 shows the E ratio (eq 1) for the compounds that were
studied. The controlling factor of the effectiveness parameter, E,
is clear in this case. The R, enantioselectivity, and Rs, resolution,
factors are not significantly changed in the deuterated experi-
ments, so the E ratio variations are only due to the k, retention
factor, changes. It is shown that the maximum retention factor
change is observed for naphthyl butyrolactone. This corresponds
to a 24% increase in retention factor when going from an aqueous
(5% acetonitrile, v/ v) mobile phase to the corresponding deuter-
ated mobile phase. A 12% decrease in retention factor is observed
for tryptophan when a deuterated mobile phase is used (Table 5
and Figure 3).
Amino Acids. Experiments show that amino acids are signifi-
cantly less retained in a deuterium environment than in the usual
hydrogen environment (Table 5). Okafo and Camilleri observed
this phenomenon on a C18 stationary phase.31 They found that
amino acids were less retained when a 2/ 1 ACN/ D2O mobile
phase was used instead of a 2/ 1 ACN/ H2O mobile phase. They
were using tetrabutylammonium hydroxide as an ion-pairing
agent. They explained their results by stating that the amino acid’s
ionization was lower (higher pKa’s). Edmonds et al. also observed
the decreased retention times with small peptides and nucleosides
on a C18 stationary phase with a 50/ 50 ACN/ D2O mobile phase.25
Several authors found that, when isotopomers are separated on
C18 columns using an aqueous mobile phase, the most deuterated
compound elutes first.32-35 All amino acids have at least three
exchangeable protons. A decrease in the retention factor between
0.3 and 1% per D atom was observed for different molecules.32-35
This effect will account for a part of the observed decreased
retention times of the amino acids. The additional difference in
our case is that, unlike the C18 stationary phase, the teicoplanin
stationary phase is deuterated itself to a significant extent. Its
ionization state is likely to decrease in a manner similar to that of
amino acids.31
The decrease in electrostatic interactions between the charged
groups of the amino acid solutes and those of the teicoplanin
selectors (ammonium and carboxylate groups) explain the major
part of the earlier elution of amino acids with deuterated mobile
phases. From a chiral mechanistic point of view, it is unfortunate
that we cannot obtain clear evidence of the hydrogen bond role
in the teicoplanin chiral recognition mechanism. Our results show
that the separation of enantiomers is not changed by altering the
magnitude of the hydrogen bonding and electrostatic interactions.
We have demonstrated that the interaction of the amine group of
the teicoplanin molecule (a cationic site) with the carboxylate
moiety of the amino acid (an anionic site) can be a primary step
of the chiral recognition mechanism.17,18,21,22 In the deuterated
medium, this electrostatic interaction is weakened by the higher
pKa values of both amine and carboxylic acid groups on both the
(32) Tanaka, N.; Thornton, E. R. J. Am. Chem. Soc. 1 9 7 6 , 98, 1617.
(33) Jinno, K.; Hirata, Y.; Hiyoshi, Y. J. High Resolut. Chromatogr. Chromatogr.
Commun. 1 9 8 2 , 5, 102.
(34) Shi, B.; Davis, B. H. J. Chromatogr. A 1 9 9 6 , 731, 351.
(35) El Tayar, N.; Van de Waterbeemd, H.; Markoulina, M.; Testa, B.; Testa, W.
F. Int. J. Pharm. 1 9 8 4 , 19, 271.
5506 Analytical Chemistry, Vol. 73, No. 22, November 15, 2001

Figure 3. Effect of the deuterium isotopic exchange on the chromatographic figure of merit (E ratio, E ) RRs/k′) of the solutes of Table 5;
relative values, the aqueous phase is the reference; column, Chirobiotic T; experimental conditions, see Table 5.
amino acid solutes and the stationary phase. Because the enan-
tioselectivity factors are not altered, it can be concluded that the
other interactions, especially hydrogen bonds, necessary to obtain
chiral recognition are similarly attenuated by the deuterated
mobile phases.
deuterium oxide molecules. More studies outside the scope of
this work are needed to establish that more deuterium oxide
molecules are incorporated in the stationary phase or that D2O
molecules occupy a greater volume than H2O molecules in the
bonded organic layer of the stationary phase. The chiral mecha-
nism involved in the separation of isomers by the teicoplanin
bonded stationary phase is not affected by modifications of the
stationary phase volume. It is also not changed by altering the
magnitude of the hydrogen and electrostatic bonds. It is known
that enantioselectivity changes when a variation in the relative
number or magnitude of enantioselective versus nonenantio-
selective interactions occurs.38
Case of p-Hydroxymandelic Acid. This solute is a polar com-
pound that can exchange three hydrogen atoms. Its pKa value is
likely increased in the deuterated mobile phase. However, the
retention times of its two enantiomers are slightly higher in the
deuterated mobile phase than in hydrogenated mobile phase
(Table 5). p-Hydroxymandelic acid is sensitive to the increase in
stationary phase volume, but the K distribution constants of its
two enantiomers are similarly decreased by the change in
electrostatic interactions (higher pKa) and the isotopic exchange
of three hydrogen atoms. The electrostatic and isotopic effects
partly cancel the stationary-phase volume effect. Compared to
amino acids, p-mandelic acid lacks the amine group. This shows
the significance of this amine group in the overall retention of
amino acids by the teicoplanin stationary phase. For amino acids,
the electrostatic and isotopic effects overcome the stationary phase
volume effect.
Uncharged Solutes. Oppositely, experiments show that the
uncharged solutes without or with fewer exchangeable protons
are significantly more retained in a deuterium environment (Table
5). Similar results were obtained with C18 columns and various
solutes.26,36 Considering that these solutes are likely more retained
through partitioning than through electrostatic interactions, these
results can be explained by an increase in the solute accessible
stationary phase volume, VS, or by identical change of both
enantiomers’ distribution coefficient between phases, K. In coun-
tercurrent chromatography, changes in the stationary phase
volume, that is a liquid phase, are very common. These changes
do not affect the solute distribution coefficients, K, and the
selectivity factors, R.37 Indeed, in the partition mechanism, the
solute retention factor, k, depends on VS and K according to
k ) KVS/ VO
(2)
with VO, the mobile phase volume. A VS increase produces a k
increase. If the distribution constants do not change, the selec-
tivity factor, R ) k2/ k1, remains unchanged, as observed ex-
perimentally.26,36 The increase in stationary phase volume may be
due to a higher wetting of the organic bonded layer by the
(36) Valleix, A.; Leonce, E.; Caussignac, C.; Buisson, D.; LeGallo, E.; Tchapla,
A. Abstract A106, SEP01 Symposium, June 12-14, 2001, Paris, France.
(37) Berthod, A. J. Chromatogr. 1 9 9 1 , 550, 677.
(38) Boehm, R. E.; Martire, D. E.; Armstrong, D. W. Anal. Chem. 1 9 8 8 , 60,
522.
Analytical Chemistry, Vol. 73, No. 22, November 15, 2001 5507

Comparison of the Chiral and Isotopic Interactions. A convenient
way to evaluate the magnitude of the interactions involved in the
separation mechanism is to give a rough estimate of the Gibbs
enantioselectivity energy difference, ∆Genantio. This energy, ex-
pressed in kJ/ mol, corresponds to the difference in the energy
interaction of the two enantiomers and the teicoplanin stationary
phase. ∆Genantio is calculated using the enantioselectivity factor39
naphthyl butyrolactone (Table 5), corresponding to a ∼20%
increase in retention factors of both enantiomers. In the opposite
way, the ∼10% decrease in retention factors observed for tryp-
tophan gives a 0.26 kJ/ mol ∆Gisotop energy difference. Considering
the energy values listed in Table 5, it seems that the bigger
-∆Genantio values (hydroxy mandelic acid, methyl phenyl hydantoin
and leucine) correspond to the smaller -∆Gisotop value (0.17, 0.36,
and -0.09 kJ/ mol, respectively). A wider variety of compounds
should be studied to establish firmly this possible trend. It would
mean that, with teicoplanin chiral selectors, the lower the
deuterium isotopic effect, the higher the enantioselectivity factor.
-∆Genantio ) RT ln k′/ k′ ) RT ln R
(3)
2
1
Eq 3 shows that R and ∆Genantio are closely related. The values
are listed in Table 5. The highest R value, 3.3, was obtained for
mandelic acid. It corresponds to a -2.9 kJ/ mol ∆Genantio energy
difference in the teicoplanin interaction with the L and the D forms
of mandelic acid (Table 5) in hydrogenated or deuterated mobile
phases. Most other solutes have a ∆Genantio energy level between
-0.3 and -0.6 kJ/ mol (Table 5). In all cases studied, the
∆Genantio H O value is not significantly different from the ∆Genantio D O
CONCLUSION
The use of mobile phases containing copper ions did not allow
us to elucidate the chiral recognition role played by the primary
amine group of the teicoplanin molecule. It did prove, however,
that teicoplanin does not form complexes with copper ions in the
same manner as amino acids and vancomycin do. The use of fully
deuterated mobile phases showed that several chiral interactions
are simultaneously modified, resulting in the constancy of the
enantioselectivity factors. The chromatograms obtained with such
deuterated mobile phases are significantly different from the
chromatograms obtained with normal hydrogenated mobile
phases. These changes may be critical in the case of the coupling
of HPLC with MS or NMR.
2
2
value.
These energy levels between two enantiomers can be com-
pared to the energy difference for the same enantiomer when
isotopic exchange is performed. The parameter, ∆Gisotop, will
measure the energy difference in the interaction of an enantio-
meric form of a solute with the deuterated stationary phase and
with the normal, proton-containing, stationary phase.
ACKNOWLEDGMENT
Alain Valleix thanks the French Commissariat a` l’Energie
Atomique, Service des Mole´cules Marque´es for its continuous
support and generous supply of deuterium oxide solvent. Alain
Berthod thanks the French Centre National de la Recherche
Scientifique (UMR 5619, 2007, and FRE2394). Daniel Armstrong
gratefully acknowledges support of this work by the National
Institutes of Health, NIH RO1 GM53825-05.
-∆Gisotop ) RT ln k1′H O/ k′
) RT ln k2′H O/ k′
(4)
1D2O
2D2O
2
2
The + or - sign on the value obtained indicates that the
deuterated solute has less or more affinity for the deuterated
teicoplanin stationary phase, respectively. This affinity difference
may be as big as -0.61 kJ/ mol for methyl phenyl indanone and
Received for review April 17, 2001. Accepted September 4,
2001.
(39) Berthod, A.; Li, W.; Armstrong, D. W. Anal. Chem. 1 9 9 2 , 64, 874.
AC010444T
5508 Analytical Chemistry, Vol. 73, No. 22, November 15, 2001