Tertiary amine appended derivatives of N-(3,5-dinitrobenzoyl)leucine as chiral selectors for enantiomer assays by electrospray ionization mass spectrometry

Article abstract of DOI:10.1021/ac050438n

Derivatives of the chiral selector N-(3,5-dinitrobenzoyl)leucine were prepared and used as chiral selectors for enantiomer discrimination in single-stage electrospray ionization mass spectrometric experiments. The chiral selectors were designed to remove

Full text of DOI:10.1021/ac050438n

Anal. Chem. 2005, 77, 5019-5027
Tertiary Amine Appended Derivatives of
N-(3,5-Dinitrobenzoyl)leucine as Chiral Selectors
for Enantiomer Assays by Electrospray Ionization
Mass Spectrometry
ChengLi Zu, Bobby N. Brewer, Beibei Wang, and Michael E. Koscho*
Department of Chemistry, Mississippi State University, Mississippi State, Mississippi 39762
atom bombardment,6-11 matrix-assisted laser desorption,12 and
electrospray ionization (ESI).6,13-19 Though the focus of the
majority of these reports was the observation of chiral recogni-
tion, recent progress has been made in the development of
quantitative enantiomer assays via single-stage mass spectrometric
experiments.18-22
By far, the greatest strides in developing methods for the
determination of enantiomeric composition have been by tandem
mass spectrometry (MS/MS). The tandem mass spectrometric
measurements rely on isolating a specific ion and allowing this
ion to react with another reagent or on observing the collision-
induced dissociation (CID) of the complex. The first type of
tandem measurement has mainly been applied to cyclodextrin-
Derivatives of the chiral selector N-(3,5-dinitrobenzoyl)-
leucine were prepared and used as chiral selectors for
enantiomer discrimination in single-stage electrospray
ionization mass spectrometric experiments. The chiral
selectors were designed to remove the ionization site from
the sites required for effective chiral recognition. Addition
of a chiral analyte to a solution of the two pseudoenan-
tiomeric chiral selectors, which differ in absolute stereo-
chemistry and the length of the tether connecting the
tertiary amine site used for ionization via protonation and
the rest of the chiral selector, affords selector-analyte
complexes in the electrospray ionization mass spectrum
where the ratio of these complexes is dependent on the
enantiomeric composition of the analyte. The relationship
between the ratio of the selector-analyte complexes in the
electrospray ionization mass spectrum and the enantio-
meric composition of the analyte can be used to relate the
extent of enantioselectivity that is being observed and for
quantitative enantiomeric composition determinations.
Investigations into the scope and limitations of this
method, plus a comparison to the enantioselectivities
observed by chiral HPLC using a N-(3,5-dinitrobenzoyl)-
leucine-derived chiral stationary phase, is presented.
(6) Liang, Y.; Bradshaw, J. S.; Izatt, R. M.; Pope, R. M.; Dearden, D. V. Int. J.
Mass Spectrom. 1999, 185, 977-988.
(7) Swada, M.; Okumura, Y.; Shizuma, M.; YoshioTakai; Hidaka, Y.; Yamada,
H.; Tanaka, T.; Kaneda, T.; Hirose, K.; Misumi, S.; Takahashi, S. J. Am.
Chem. Soc. 1993, 115, 7381-7388.
(8) Swada, M.; Takai, Y.; Yamada, H.; Hirayama, S.; Kaneda, T.; Tanaka, T.;
Kamada, K.; Mizooku, T.; Takeuchi, S.; Ueno, K.; Hirose, K.; Tobe, Y.;
Naemura, K. J. Am. Chem. Soc. 1995, 117, 7726-7736.
(9) Dobo, A.; Liptak, M.; Huszthy, P.; Vekey, K. Rapid Commun. Mass Spectrom.
1997, 11, 889-896.
(10) Pocsfalvi, G.; Liptak, M.; Huszthy, P.; Bradshaw, J. S.; Izatt, R. M.; Vekey,
K. Anal. Chem. 1996, 68, 792-795.
(11) Swada, M.; Takai, Y.; Yamada, H.; Nishida, J.; Kaneda, T.; Arakawa, R.;
Okamato, M.; Hirose, K.; Tanaka, T.; Naemura, K. J. Chem. Soc., Perkin
Trans. 2 1998, 701-710.
(12) So, M. P.; Wan, T. S. M.; Chan, T.-W. D. Rapid Commun. Mass Spectrom.
Interest in the development of methods that utilize only mass
spectrometry for investigating chiral recognition systems, and for
the determination of enantiomeric composition, has recently
increased dramatically.1-4 Since Fales and Wright first reported
on the observation of chiral recognition between protonated
diastereomeric dialkyl tartrate dimers in the chemical ionization
(CI) mass spectrum,5 a number of reports have illustrated mass
spectrometric chiral recognition with a myriad of host-guest
systems using a variety of ionization methods, including CI,5 fast-
2000, 14, 692-695.
(13) Nierengarten, H.; Leize, E.; Garcia, C.; Jeminet, G.; Dorsselaer, A. V. Analusis
2000, 28, 259-263.
(14) Czerwenka, C.; Maier, N. M.; Lindner, W. Anal. Bioanal. Chem. 2004, 379,
1039-1044.
(15) Sawada, M.; Takai, Y.; Yamada, H.; Yoshikawa, M.; Arakawa, R.; Tabuchi,
H.; Takada, M.; Tanaka, J.; Shizuma, M.; Yamaoka, H.; Hirose, K.; Fukuda,
K.; Tobe, Y. Eur. J. Mass Spectrom. 2004, 10, 27-37.
(16) Mehdizadeh, A.; Letzel, M. C.; Klaes, M.; Agena, C.; Mattay, J. Eur. J. Mass
Spectrom. 2004, 10, 649-655.
(17) Seymour, J. L.; Turecek, F.; Malkov, A. V.; Kocovsky, P. J. Mass Spectrom.
2004, 39, 1044-1052.
(18) Koscho, M. E.; Zu, C.; Brewer, B. N. Tetrahedron: Asymmetry 2005, 16,
801-807.
* To whom correspondence should be addressed. Tel.: 662-325-9500. Fax:
662-325-1618. E-mail: mek40@ra.msstate.edu.
(1) Speranza, M. Int. J. Mass Spectrom. 2004, 232, 277-317.
(2) Tao, W. A.; Cooks, R. G. Anal. Chem. 2003, 75, 25A-30A.
(3) Filippi, A.; Giardini, A.; Piccirillo, S.; Speranza, M. Int. J. Mass Spectrom.
2000, 198, 137-163.
(19) Brewer, B. N.; Zu, C.; Koscho, M. E. Chirality. In press.
(20) Shizuma, M.; Imamura, H.; Takai, Y.; Yamada, H.; Takeda, T.; Takahashi,
S.; Sawada, M. Int. J. Mass Spectrom. 2001, 210/211, 585-590.
(21) Sawada, M.; Yamaoka, H.; Takai, Y.; Kawai, Y.; Yamada, H.; Azuma, T.;
Fujioka, T.; Tanaka, T. Int. J. Mass Spectrom. 1999, 193, 123-130.
(22) Sawada, M.; Yamaoka, H.; Takai, Y.; Kawai, Y.; Yamada, H.; Azuma, T.;
Fujioka, T.; Tanaka, T. Chem. Commun. 1998, 1569-1570.
(4) Sawada, M. Mass Spectrom. Rev. 1997, 16, 73-90.
(5) Fales, H. M.; Wright, G. J. J. Am. Chem. Soc. 1977, 99, 2339-2340.
10.1021/ac050438n CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/28/2005
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5019

analyte complexes,23-27 and chiral crown ether-chiral ammonium
cation complexes.28,29 The rate at which the analyte exchanges
for an achiral reagent gas in the host-guest complex is used as
a metric for determining the stereochemical composition of the
analyte. For the other type of tandem experiments, higher order
complexes are mass-selected and allowed to undergo CID, and
the observed relative branching ratios are related to the enantio-
meric composition by the kinetic method of Cooks.30-49
In addition to the methods above, where in each case
diastereomeric noncovalent complex ions are produced in the
experiment, it is also possible to measure enantiomeric composi-
tion using preformed covalent diastereomers.50-52 Derivatization
of a chiral analyte with a mixture of mass-labeled, pseudo-
enantiomeric chiral reagents (where each pseudoenantiomer has
the opposite stereochemistry, but a slightly different mass due to
labeling of one enantiomer at a remote position) affords derivatives
that can be discriminated by mass spectrometry. As long as kinetic
resolution is observed in the derivatization step, the relative
amounts of the derivatives can be related back to the enantiomeric
composition of the analyte. Of course, one major drawback to this
method is the requirement for the preparation of derivatives before
the analysis.
One main impetus for developing mass spectrometric methods
for enantiomer assays is the potential for rapid analysis. Such a
method would definitely be a boon to the development of catalytic
enantioselective reactions, particularly by combinatorial asym-
metric catalysis, whereby libraries of potential asymmetric cata-
lysts are produced in parallel and each catalyst is screened for its
ability to produce a product of high enantiomeric purity.53-57
Typically, the size of combinatorial catalyst libraries has not been
limited by the number of potential catalysts that one can prepare,
but rather by the time needed to evaluate large libraries using
contemporary methods for enantiomer analysis. Such a method
could also be used for the discovery and optimization of chiral
selectors by combinatorial methods.58-64 Instead of measuring the
enantiomeric composition, one would instead be using mass
spectrometry to directly measure the relative binding of analyte
enantiomers to potential chiral selectors. In addition to the
potential for rapid analysis, other attributes such as a high
tolerance to impurities, broad analyte scope, and high sensitivity
are beneficial for the determination of the enantiomeric composi-
tion of samples originating from a variety of sources, particularly
samples of biological origin.
Previously, we have reported the use of soluble analogues of
Pirkle-type chiral stationary phases (CSPs) as chiral selectors for
enantiomer assays by electrospray ionization mass spectrom-
etry.18,19 In each case, a solution of pseudoenantiomeric chiral
selectors (where each enantiomer is mass-labeled at a remote
site), when mixed with a chiral analyte whose enantiomers are
known to be resolved on the corresponding CSP, was shown to
afford selector-analyte complexes in the mass spectrum where
the relative peak intensities of the complexes depend on the
enantiomeric composition of the analyte.
(23) Ramirez, J.; He, F.; Lebrilla, C. B. J. Am. Chem. Soc. 1998, 120, 7387-
7388.
(24) Ramirez, J.; Ahn, S.; Grigorean, G.; Lebrilla, C. B. J. Am. Chem. Soc. 2000,
122, 6884-6890.
(25) Grigorean, G.; Ramirez, J.; Ahn, S. H.; Lebrilla, C. B. Anal. Chem. 2000,
72, 4275-4281.
(26) Grigorean, G.; Lebrilla, C. B. Anal. Chem. 2001, 73, 1684-1691.
(27) Grigorean, G.; Cong, X.; Lebrilla, C. B. Int. J. Mass Spectrom. 2004, 234,
71-77.
(28) Dearden, D. V.; Dejsupa, C.; Liang, Y.; Bradshaw, J. S.; Izatt, R. M. J. Am.
Chem. Soc. 1997, 119, 353-359.
For example, mass-labeled, soluble analogues of the N-(3,5-
dinitrobenzoyl)leucine CSP (DNB-leucine CSP, Chart 1), such as
a 1:1 mixture of (S)-1 and (R)-2 which differ only by the length
of the N-alkyl chain of the amide, when mixed with analyte 10
(Chart 2) and excess lithium chloride afforded peaks in the
electrospray ionization mass spectrum that correspond to the
following complexes: [1 + 10 + Li]+ and [2 + 10 + Li]+. It
was found that a plot of the natural log of the relative peak
intensities of the selector-analyte complexes in the mass spec-
trum versus the enantiomeric composition of the analyte is linear.18
Similar results were found using soluble analogues of the N-(3,5-
dinitrobenzoyl)phenylglycine CSP as chiral selectors (e.g., (R)-3
(29) Chu, I.-H.; Dearden, D. V.; Bradshaw, J. S.; Huszthy, P.; Izatt, R. M. J. Am.
Chem. Soc. 1993, 115, 4318-4320.
(30) Bagheri, H.; Chen, H.; Cooks, R. G. Chem. Commun. 2004, 2740-2741.
(31) Wu, L.; Meurer, E. C.; Cooks, R. G. Anal. Chem. 2004, 76, 663-671.
(32) Yu, C.-T.; Guo, Y.-L.; Chen, G.-Q.; Zhong, Y.-W. J. Am. Soc. Mass Spectrom.
2004, 15, 795-802.
(33) Wu, L.; Clark, R. L.; Cooks, R. G. Chem. Commun. 2003, 136-137.
(34) Wu, L.; Cooks, R. G. Anal. Chem. 2003, 75, 678-684.
(35) Augusti, D. V.; Augusti, R.; Carazza, F.; Cooks, R. G. Chem. Commun. 2002,
2242-2243.
(36) Augusti, D. V.; Carazza, F.; Augusti, R.; Tao, W. A.; Cooks, R. G. Anal. Chem.
2002, 74, 3458-3462.
(37) Tao, W. A.; Clark, R. L.; Cooks, R. G. Anal. Chem. 2002, 74, 3783-3789.
(38) Fago, G.; Filippi, A.; Giardini, A.; Lagana, A.; Paladini, A.; Speranza, M.
Angew. Chem., Int. Ed. 2001, 40, 4051-4054.
(39) Paladini, A.; Calcagni, C.; Palma, T. D.; Speranza, M.; Lagana, A.; Fago, G.;
Filippi, A.; Satta, M.; Guidoni, A. G. Chirality 2001, 13, 707-711.
(40) Tao, W. A.; Cooks, R. G. Angew. Chem., Int. Ed. 2001, 40, 757-760.
(41) Tao, W. A.; Gozzo, F. C.; Cooks, R. G. Anal. Chem. 2001, 73, 1692-1698.
(42) Tao, W. A.; Wu, L.; Cooks, R. G. J. Med. Chem. 2001, 44, 3541-3544.
(43) Tao, W. A.; Wu, L.; Cooks, R. G. Chem. Commun. 2000, 2023-2024.
(44) Tao, W. A.; Zhang, D.; Nikolaev, E. N.; Cooks, R. G. J. Am. Chem. Soc. 2000,
122, 10598-10609.
(45) Yao, Z.-P.; Wan, T. S. M.; Kwong, K.-P.; Che, C.-T. Anal. Chem. 2000, 72,
5383-5393.
(46) Yao, Z.-P.; Wan, T. S. M.; Kwong, K.-P.; Che, C.-T. Anal. Chem. 2000, 72,
5394-5401.
(47) Yao, Z.-P.; Wan, T. S. M.; Kwong, K.-P.; Che, C.-T. Chem. Commun. 1999,
2119-2120.
(48) Cooks, R. G.; Wong, P. S. H. Acc. Chem. Res. 1998, 31, 379-386.
(49) Vekey, K.; Czira, G. Anal. Chem. 1997, 69, 1700-1705.
(50) Yao, S.; Meng, J.-C.; Siuzdak, G.; Finn, M. G. J. Org. Chem. 2003, 68, 2540-
2546.
(51) Diaz, D. D.; Yao, S.; Finn, M. G. Tetrahedron Lett. 2001, 42, 2617-2619.
(52) Guo, J.; Wu, J.; Finn, M. G. Angew. Chem., Int. Ed. Engl. 1999, 38, 1755-
1758.
(53) Finn, M. G. Chirality 2002, 14, 534-540.
(54) Reetz, M. T. Angew. Chem., Int. Ed. 2002, 41, 1335-1338.
(55) Reetz, M. T. Angew. Chem., Int. Ed. 2001, 40, 284-310.
(56) Reetz, M. T.; Kuhling, K. M.; Deege, A.; Hinrichs, H.; Belder, D. Angew.
Chem., Int. Ed. 2000, 39, 3891-3893.
(57) Schrader, W.; Eipper, A.; Pugh, D. J.; Reetz, M. T. Can. J. Chem. 2002, 80,
626-632.
(58) Wang, Y.; Blum, L. H.; Li, T. Anal. Chem. 2000, 72, 5459-5465.
(59) Bluhm, L. H.; Wang, Y.; Li, T. Anal. Chem. 2000, 72, 5201-5205.
(60) Brahmachary, E.; Ling, F. H.; Svec, F.; Frechet, J. M. J. J. Comb. Chem.
2003, 5, 441-450.
(61) Welch, C. J.; Pollard, S. D.; Mathre, D. J.; Reider, P. J. Org. Lett. 2001, 3,
95-98.
(62) Lewandowski, K.; Murer, P.; Svec, F.; Frechet, J. M. J. J. Comb. Chem. 1999,
1, 105-112.
(63) Welch, C. J.; Bhat, G.; Protopopova, M. N. J. Comb. Chem. 1999, 1, 364-
367.
(64) Weingarten, M. D.; Sekanina, K.; Still, W. C. J. Am. Chem. Soc. 1998, 120,
9112-9113.
5020 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Chart 1
Chart 2
and (S)-4). Overall, it was found that the selectivity was rather
low, compared to the enantioselectivity obtained by HPLC using
either of the N-(3,5-dinitrobenzoyl)amino acid CSPs, presumably
due to interference of the requisite selector-analyte hydrogen
bonds needed for effective chiral recognition by the lithium
cations.
Subsequently, mixtures of deprotonated N-(3,5-dinitrobenzoyl)-
amino acids were used as chiral selectors (e.g., binary combina-
tions of 5, 6, and 7).19 As before, it was found that a plot of the
natural log of the relative peak intensities of the selector-analyte
complexes in the negative ion ESI-mass spectrum versus the
enantiomeric composition of the analyte is linear. In these cases,
it was found that the selectivity was indeed comparable to the
selectivity obtained by chiral HPLC.
In both instances, it was found that the reproducibility of
enantiomeric composition determinations using the linear calibra-
tion plot was ∼(0.05 for the calculated mole fraction of either
enantiomer. In the first instance, with the lithiated selector-
analyte complexes, the poor reproducibility was due primarily to
a low selectivity where small differences in measured relative peak
intensities will afford substantial differences in the calculated
enantiomeric composition value. In the second instance, with the
deprotonated chiral selectors, the low reproducibility was due to
a small absolute value of the intensities of the selector-analyte
complexes in the ESI-mass spectrum. In fact, the ion counts were
∼2 orders of magnitude less for the complexes derived from the
deprotonated selectors compared to the lithiated selectors. The
variability was therefore a manifestation of a decreased signal with
respect to experimental noise.
In this paper, we report the use of tertiary amine appended
N-(3,5-dinitrobenzoyl)leucine derivatives, 8 and 9, as pseudo-
enantiomeric chiral selectors for enantiomer analyses by electro-
spray ionization mass spectrometry. These chiral selectors were
designed to overcome the problems delineated above, via a
separation of the requisite chiral recognition sites from the
ionization site. Addition of an acid to the solution to be assayed
should protonate the tertiary amine, affording high ion counts for
the selector in the electrospray ionization mass spectrum. Ad-
ditionally, since the ionization site is removed from the sites
needed for chiral recognition, its interference with the formation
of selector-analyte complexes should be minimized. In order for
this to be a viable analytical method, one would like to be able to
measure the enantiomeric composition of the analyte at a variety
of concentrations, even without necessarily knowing the concen-
tration of the analyte. Additionally, one would like to be able to
use a variety of solvent compositions to accommodate samples of
many types. Herein is reported the scope and limitations of this
method, including the effect of analyte concentration and a survey
of solvent effects, using protonated pseudoenantiomeric chiral
selectors, (S)-8 and (R)-9, for enantiomer discrimination and for
the determination of enantiomeric composition by electrospray
ionization mass spectrometry.
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5021

Figure 1. Electrospray ionization mass spectrum of a solution of pseudoenantiomeric chiral selectors (S)-8 and (R)-9 (1.0 mM each), racemic
analyte 10 (1.0 mM), and ammonium chloride (10 mM) in methanol/water (1:1).
EXPERIMENTAL SECTION
rich C-terminal carbonyl group of leucine. Any chiral analyte that
possesses complementary interaction sites, (1) an electron-rich
aromatic ring, (2) a functional group that can accept a hydrogen
bond, and (3) a hydrogen bond donor is a candidate for enantio-
differentiation by this type of chiral selector. Analytes 10-18 all
fit within this rubric and were chosen for this study. The appended
tertiary amine of chiral selectors 8 and 9 is superfluous for chiral
recognition purposes, but has been added to allow ready ionization
of the chiral selectors and any complexes derived thereof via
protonation.
Electrospray ionization mass spectrometry of solutions contain-
ing the chiral selectors, (S)-8 and (R)-9, the analyte, and
ammonium chloride in methanol/water affords significant pro-
tonated selector-analyte complexes in the mass spectrum for each
analyte. The samples were introduced into the electrospray unit
via flow injection of a 10-µL sample loop at a flow rate of 200 µL/
min of water/methanol (1:1). The full positive ion spectrum was
recorded every 0.7 s. All scans for which a significant total ion
count was observed were averaged together to afford the final
spectrum, typically requiring ∼10 s/injection. The full mass
spectrum for the solution containing racemic analyte 10 is shown
in Figure 1. The protonated analyte, [10 + H]+, and chiral
selectors, [8 + H]+ and [9 + H]+, are observed at m/z 303, 424,
and 438, respectively. Analyte dimers and selector dimers are
observed at m/z 605 [102 + H]+, 622 [102 + NH4]+, 883 [(8 +
H)2 + Cl]+, 897 [(8 + H) + (9 + H) + Cl]+, and 911 [(9 + H)2
+ Cl]+. The selector-analyte complexes, which, in this instance,
are the largest peaks in the spectrum, are observed at m/z 726
[8 + 10 + H]+ and 740 [9 + 10 + H]+.
Figure 2 presents the portions of the mass spectra that contain
the selector-analyte complexes obtained at three different enan-
tiomeric compositions of analyte 10. The relative intensities of
the selector-analyte complexes vary regularly with the enantio-
meric composition of the analyte, clearly demonstrating chiral
recognition. As the amount of (R)-10 is increased in the sample,
the complex with the (R)-9 selector is increased relative to the
complex with the (S)-8 selector; likewise, as the amount of (S)-
10 is increased in the sample, the complex with the (S)-8 selector
is increased relative to the complex with the (R)-9 selector. The
sense of chiral recognition is consistent with what is observed
chromatographically, whereby the (S)-enantiomer of analyte 10
is more retained when using the (S)-DNB-leucine CSP. Addition-
Procedures for the preparation of chiral selectors (S)-8 and
(R)-9 and characterization data for all new compounds are
included in the Supporting Information. The chiral stationary
phase DNB-leucine (4.6 × 250 mm column) was prepared by
literature methods.65 All solvents used were HPLC grade and used
without further purification.
All mass spectra were obtained on a Micromass Quattro Micro
(Beverly, MA) triple quadrupole mass spectrometer with electro-
spray ionization running in the positive ion mode. Solutions were
flow injected into the electrospray ionization source through a 10-
µL injection loop with mobile phase running at a flow rate of 200
µL/min. The full positive ion spectrum was recorded every 0.7 s.
All scans for which a significant total ion count was observed were
averaged together to afford the final spectrum, typically requiring
∼10 s/injection. Spectrometer conditions are as follows: capillary
voltage, 3.0 kV; cone voltage, 12 V; extractor voltage, 1.0 V; rf
lens, 0.5 V; source temperature, 80 °C; desolvation temperature,
325 °C; cone gas flow, 62 L/h; desolvation gas flow, 608 L/h. For
data collected by direct infusion with a syringe pump (entries 6-11
in Table 4), scans collected over a 2-min period were averaged
together to afford the final spectrum.
RESULTS AND DISCUSSION
Chiral Recognition. The chiral selectors 8 and 9 were
designed to (1) remove the ionization site from the sites required
for chiral recognition and (2) act as enantiomers where both can
be detected in a single mass spectrometric experiment (i.e., mass-
labeled pseudoenantiomers). As has been demonstrated through
a number of studies,65-69 the primary interactions between the
DNB-leucine chiral selectors and analyte that are necessary for
effective chiral recognition are as follows: (1) a π-stacking
interaction with the aromatic ring of the electron-deficient di-
nitrobenzamide, (2) a hydrogen bond with the electron-deficient
benzamide proton, and (3) a hydrogen bond with the electron-
(65) Pirkle, W. H.; Welch, C. J. J. Org. Chem. 1984, 49, 138-140.
(66) Pirkle, W. H.; Murray, P. G.; Wilson, S. R. J. Org. Chem. 1996, 61, 4775-
4777.
(67) Pirkle, W. H.; Pochapsky, T. C. J. Am. Chem. Soc. 1986, 108, 5627-
5628.
(68) Pirkle, W. H.; III, J. A. B.; Wilson, S. R. J. Am. Chem. Soc. 1989, 111, 9222-
9223.
(69) Pirkle, W. H.; Tsipouras, A. Tetrahedron Lett. 1985, 26, 2989-2992.
5022 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

assays but it is also a measure of the enantioselectivity of the chiral
selectors toward the analyte enantiomers. The larger the slope of
this plot, the greater the selectivity. Presumably, in this experi-
ment, complexes that are formed in solution are being transferred
to the gas phase such that the relative amount of each complex
in solution is reported in the mass spectrum. This allows
determination of the extent of enantioselectivity in solution based
on the slope of the calibration plot derived from the mass
spectrometric data.
In general, given pseudoenantiomeric chiral selectors, ΨR and
ΨS, and analyte enantiomers, AR and AS, the enantioselectivity in
solution, R, will be given by any of the ratios of concentra-
tions shown in eq 1. This, of course, assumes that the pseudo-
enantiomeric chiral selectors are acting as enantiomers, and
[ΨR] ) [ΨS].
[ΨR‚AR] [ΨS‚AS] [ΨR‚AR] [ΨS‚AS]
R )
)
)
)
(1)
[ΨR‚AS] [ΨS‚AR] [ΨS‚AR] [ΨR‚AS]
Figure 2. Partial electrospray ionization mass spectrum of solutions
containing pseudoenantiomeric chiral selectors (S)-8 and (R)-9 (1.0
mM each), analyte 10 (1.0 mM), and ammonium chloride (10 mM) in
methanol/water (1:1). Spectra: (a) 79.6% enantiomeric excess (R)-
10, (b) racemic 10, and (c) 81.8% enantiomeric excess (S)-10.
Each spectrum is normalized to the intensity of the larger complex
peak.
In this instance, since the chiral selectors have a tertiary amine
appended and a 5-fold excess of ammonium chloride has been
added to the solution, the complexes are likely protonated.
Electrospray ionization of the solution will transfer these pro-
tonated complexes to the gas phase with an efficiency given by
the transfer coefficient: tX. Each of the four possible complexes
will have its own transfer coefficient, though given the similarity
of the complexes, and that the ionization site has been removed
from the sites presumed to be involved in chiral recognition, it
can be assumed that all four are equal. With this assumption, one
can then relate the relative ion counts of the selector-analyte
complexes observed in the mass spectrum with the relative
concentrations of complexes formed in solution, and hence the
enantioselectivity, as measured by mass spectrometry, RMS, since
the ratios of the transfer coefficients will be unity.
t
RR[ΨR‚AR‚H]+ tSS[ΨS‚AS‚H]+
RMS
)
)
)
tRS[ΨR‚AS‚H]+ tSR[ΨS‚AR‚H]+
Figure 3. Complex intensity fraction in the electrospray ionization
mass spectrum versus the mole fraction of analyte (R)-10 (200 µM)
in the solution, using pseudoenantiomeric chiral selectors (S)-8 (1.0
mM) and (R)-9 (1.0 mM) with added ammonium chloride (10 mM) in
methanol/water (1:1). Slope 0.260; intercept 0.302; r2 ) 0.999.
t
RR[ΨR‚AR‚H]+ tSS[ΨS‚AS‚H]+
)
(2)
tSR[ΨS‚AR‚H]+ tRS[ΨR‚AS‚H]+
From the mass spectrum, it is the CIF that is measured, which
is given by sum of ion counts of both analyte enantiomers with
one chiral selector divided by the sum of ion counts of all
selector-analyte complexes.
ally, chiral recognition was observed in the electrospray ionization
mass spectrum for each of the remaining analytes, 11-18, and
the sense of chiral recognition was consistent with the sense of
chromatographic chiral recognition in each case.
A plot of the complex intensity fraction (CIF; intensity of one
selector-analyte complex divided by the sum of the intensities
for both selector-analyte complexes) versus the enantiomeric
composition of analyte 10 is presented in Figure 3. The plot is
linear with a correlation coefficient of 0.999. This plot can be used
for subsequent enantiomer assays. Simply measuring the CIF of
the selector-analyte complexes in the mass spectrum and
applying this value to the calibration line affords the enantiomeric
composition.
CIF )
[ΨR‚AR‚H]+ + [ΨR‚AS‚H]+
[ΨR‚AR‚H]+ + [ΨR‚AS‚H]+ + [ΨS‚AR‚H]+ + [ΨS‚AS‚H]+
(3)
The relation of each of the ion counts to solution concen-
trations, followed by substitutions, using the appropriate equi-
librium expressions and eq 2, and rearrangement, affords eq 4.
This equation predicts that the plot of the CIF versus the
enantiomeric composition of the analyte (XR is mole fraction of
Enantioselectivity. The linear plot of the CIF versus enan-
tiomeric composition is not only useful for subsequent enantiomer
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5023

the (R)-analyte) is linear.
CIF )
Table 1. Observed Enantioselectivity as a Function of
the Concentration of Analyte 10 at a Constant
Concentration of Pseudoenantiomeric Chiral Selectors
(S)-8 and (R)-9
R - 1
R + 1
1
XR +
(4)
(
)
(
)
R + 1
selector:analytea,b
RMSc (σ)d
Additionally, any differences in the transfer coefficients of the
complexes, differences in enantioselectivity of the pseudo-
enantiomeric chiral selectors, or differences in the concentrations
of chiral selectors, as long as the differences remain constant
throughout, should manifest itself in the intercept of this plot. In
fact, because of this, rarely do the RMS values determined from
the slope and the intercept of this plot agree. For the pure
20:1
10:1
5:1
1.61 (0.02)
1.66 (0.03)
1.64 (0.02)
1.71 (0.04)
1.54 (0.02)
1.42 (0.02)
1.33 (0.02)
1:1
1:2
1:5
1:10
a Concentration of selectors (S)-8 and (R)-9 is 1.0 mM in all cases
in methanol/water (1:1). b Concentration of NH4Cl is 10 mM in all
cases. c Observed MS enantioselectivity. d σ is the standard deviation
obtained from the calibration curve for three replicate measurements
at three different enantiomeric compositions.
enantiomers (XR ) 1, and XR ) 0), the CIF values will be RMS
/
(RMS + 1) and 1/(RMS + 1). The ratio of these CIF values is RMS
and affords an additional measure of the enantioselectivity. In
practice, using the slope of the calibration plot is more convenient
since one only needs to measure the CIF at two different
enantiomeric compositions to construct the plot, whereas in order
to determine the CIF values for both pure enantiomers, one needs
access to both pure analyte enantiomers. All enantioselectivities
reported herein were calculated from the slope of the calibration
plots.
Table 2. Observed Enantioselectivity as a Function of
the Concentration of Pseudoenantiomeric Chiral
Selectors (S)-8 and (R)-9 and Analyte 10 at a Constant
Selector/Analyte Ratioa,b
[(S)-8]/[(R)-9], µM
[10], µM
RMSc (σ)d
In our previous reports,18,19 we plotted the natural log of the
relative peak intensities of the selector-analyte complexes versus
the enantiomeric composition, where the slope of this plot is then
two times the natural log of the enantioselectivity (2 ln RMS). The
RMS values calculated by this method agree, within experimental
error, with the values determined as described above. In fact, the
logarithmic function is an approximation of the hyperbolic function
above. The logarithmic function will deviate from linearity as RMS
increases, though eq 4 will still hold, and therefore, eq 4 has been
4000
2000
1000
500
2000
1000
500
1.76 (0.03)
1.69 (0.04)
1.67 (0.02)
1.69 (0.02)
1.67 (0.02)
1.62 (0.03)
1.65 (0.02)
1.63 (0.02)
250
250
125
125
62.5
31.2
62.5
31.2
15.6
a Concentration of NH4Cl is 10 times [10]. b Solvent is methanol/
water (1:1) in all cases. c Observed MS enantioselectivity. d σ is the
standard deviation obtained from the calibration curve for three
replicate measurements at three different enantiomeric compositions.
used throughout for the determination of RMS
.
Enantioselectivity as a Function of Selector/Analyte Con-
centration. In order for this method to be practicable for
quantitative enantiomer assays, the results of any assay should
be independent of the absolute concentrations of the analyte as
much as possible. One can readily control the concentrations of
the selectors and additives and the solvent composition by
preparing a stock solution, where this same solution is used for
the construction of the calibration curve and for subsequent
enantiomer assays. Ideally, one would like to add a sample of the
analyte to a small aliquot of this stock solution and record the
electrospray ionization mass spectrum, without measuring the
amount of analyte.
Table 1 presents the enantioselectivities obtained at a variety
of different concentrations of analyte 10. In each case, the
concentrations of the chiral selectors, ammonium chloride, and
solvent composition were kept constant. The RMS values are
relatively invariant, as long as the chiral selectors are in excess
of the analyte. The RMS values for the entries at a selector/analyte
ratio of 20:1, 10:1, 5:1, and 1:1 are all within 3% of the average RMS
(1.66). As the concentration of the analyte is decreased, the
intensity of the selector-analyte complexes in the mass spectrum
also decreases, compared to the intensity of the monomeric ions.
Since the measured enantioselectivity only depends on the relative
amounts of the (pseudo)diastereomeric complexes formed in
solution, the ratio of the complexes should not change in the limit
of low analyte concentration. In essence, each pseudoenantiomer
(via the pseudodiastereomeric selector-analyte complexes) is
acting as an internal standard for the other. Increasing the analyte
concentration beyond the concentration of the chiral selectors
results in saturation, where further increases in the analyte
concentration will further reduce the observed enantioselectivity.
The onset of this diminution of enantioselectivity was observed
at a 2:1 analyte/selector ratio for analyte 10.
Table 2 presents the observed enantioselectivities at a number
of different absolute concentrations of selectors, analyte, and
ammonium chloride. The relative concentrations of all the
components were kept constant throughout. Relatively little
variation of the RMS values was observed throughout this concen-
tration range. All but one of the RMS values in Table 2 are within
3% of the average RMS (1.67) for this data set, and this value also
agrees very well with the average RMS value for the data in Table
1. Again, since it is the ratio of (pseudo)diastereomeric selector-
analyte complexes that determines the selectivity, the relative
amounts of these complexes should not be affected by dilution.
At lower concentrations, the intensities of the selector-analyte
complexes in the mass spectra are diminished, which limits the
extent to which the solutions can be diluted by the signal-to-noise
ratio that one can obtain experimentally.
These two data sets have important implications for the use
of this method for quantitative enantiomer assays: (1) as long as
the concentrations of the selectors are in excess of the analyte,
the calibration line and subsequent enantiomer assays can be done
5024 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Solvent Survey. Electrospray ionization mass spectra of
solutions of chiral selectors (S)-8 and (R)-9 and analyte 10 with
added acid, in a variety of solvents, were obtained in order to
determine the scope of solvents for which effective chiral recogni-
tion would be observed. The observed mass spectrometric
enantioselectivities for a number of these are presented in Table
4. In addition to the solvent combinations shown, hexanes, toluene,
THF, dichloromethane, acetonitrile, methanol, ethanol, and 2-pro-
panol were each tested individually as solvents. For solubility
purposes, acetic acid was used instead of ammonium chloride as
the protonation source. In each case, either no selector-analyte
complexes were observed in the mass spectrum, or when the
complexes were observed, chiral recognition was minimal. It is
apparent from these data that water is a necessary component of
the solutions to be assayed if effective chiral recognition is to be
observed.
Entries 1-5 demonstrate the effect that the amount of water
in the solvent (using binary water/methanol solvent composi-
tions) has on the extent of the observed enantioselectivity. The
highest RMS value was observed at a composition of 10% water,
while the enantioselectivity decreased regularly with an in-
crease in water composition. Comparing entries 3, 6, and 12, or
entries 2 and 7, clearly demonstrates that using acetic acid as the
proton source affords higher enantioselectivities than ammonium
chloride.
Table 3. Determination of the Enantiomeric
Composition of Five Different Samples of Analyte 10 by
Mass Spectrometry at Three Different Concentrations
Using a Single Calibration Linea,b
mole fraction of (R)-10
[10], µM
HPLCc,d
MS (σ)e
500
500
500
500
500
100
100
100
100
100
50
0.976
0.774
0.497f
0.240
0.044
0.976
0.774
0.497f
0.240
0.044
0.976
0.774
0.497f
0.240
0.044
0.979 (0.009)
0.782 (0.009)
0.502 (0.008)
0.240 (0.009)
0.041 (0.009)
0.960 (0.009)
0.782 (0.009)
0.482 (0.008)
0.239 (0.009)
0.056 (0.009)
0.966 (0.009)
0.779 (0.009)
0.495 (0.008)
0.235 (0.009)
0.075 (0.009)
50
50
50
50
a [8] ) 1.0 mM, [9] ) 1.0 mM, and [NH4Cl] ) 10 mM in methanol/
water (1:1). b Calibration curve constructed at [10] ) 200 µM. c (S,S)-
Whelk-O1, hexanes/2-propanol/methanol (60:38:2) at 2 mL/min.
d Repeated injections afforded standard deviations of <0.003 in all cases.
e σ is the standard deviation obtained from the calibration curve for
nine replicate measurements. f Racemic.
Comparing the enantioselectivities obtained for entries 7-11,
it is evident that solvent compositions containing substantial
amounts of organic solvents other than methanol, such as
acetonitrile, THF, dichloromethane, or toluene, can be used. This
has important implications for the use of this method in quantita-
tive enantiomer assays, namely, that the analyte does not have to
be dissolved in the same solvent as the selector and acid stock
solution. One can either construct a calibration line using the
appropriate final solvent composition or consider the additional
solvent as an interference and use the calibration curve con-
structed without this interference. To determine the extent that
a small amount of an additional solvent effects the enantio-
selectivity, a small amount (0.1 mL) of acetonitrile, THF, di-
chloromethane, and toluene was each added to solutions of chiral
selectors, analyte, and ammonium chloride in 1:1 water/methanol
(1.0 mL). For the additions of dichloromethane and toluene, a
two-phase mixture was formed, which added the partitioning
between the two phases as an additional complication. Though,
for the single-phase solutions obtained on addition of a small
amount of acetonitrile or THF, as is evident from entries 12-14,
the effect of this addition on the observed enantioselectivity is
small.
Comparison of Enantioselectivities by Mass Spectrometry
and Chiral HPLC. In addition to being used as a method for
quantitative enantiomeric composition determinations, this method
should also be useful for the discovery and optimization of chiral
selectors.58-64 Enantioselectivities of analytes 10-18, as mea-
sured by electrospray ionization mass spectrometry using chiral
selectors (S)-8 and (R)-9, and by HPLC using the (S)-DNB-leucine
CSP, under both normal-phase and reversed-phase conditions are
presented in Table 5. To facilitate comparisons among the data,
the HPLC conditions (as indicated in Table 5) were not optimized
for any of the analytes; rather they were held constant throughout
at any concentration of the selectors and analyte, and (2) valid
data can readily be obtained over at least 2 orders of mag-
nitude of analyte concentration (15.6 µM-2.0 mM for 10).70
Solubility will likely be the upper concentration bound, while
the magnitude of the signal (compared to the experimental
noise) from the selector-analyte complexes will be the lower
bound.
Quantitative Enantiomer Analysis. To test the validity of
this method to accurately determine enantiomeric composition
independent of analyte concentration, the enantiomeric composi-
tion of five different samples was determined at three different
concentrations, using a calibration line that was constructed at a
concentration different from each of these. The calibration curve
was constructed from 3 replicate measurements of the CIF values
in the electrospray ionization mass spectrum, where the sample
was introduced by flow injection, at 11 different enantiomeric
compositions. In each case, the concentrations of the selectors
were 1.0 mM, the concentration of ammonium chloride was 10
mM, and the analyte concentration was 200 µM. This one
calibration curve was used for all subsequent enantiomer deter-
minations. The analyte concentrations of the assayed solutions
span 1 order of magnitude (50-500 µM). To estimate the precision
of this method, each solution was injected a total of nine times,
the average and standard deviation of the calibrated enantiomeric
composition determinations are shown in Table 3. As can be seen
from the data, accurate enantiomeric composition values are
obtained over this data range, with observed standard deviations
that are less than 1.0% of the composition of either enantiomer,
for the nine replicate analyses for each sample.
(70) The analyte concentration, 15.6 µM, does not represent an actual limit of
detection; selector-analyte complexes can be detected at much lower
concentrations. Further 1:1 dilutions of the least concentrated sample in
Table 2 afforded greatly increased standard deviations.
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5025

Table 4. Observed Enantioselectivity of Pseudoenantiomeric Chiral Selectors (S)-8 and (R)-9 for Analyte 10
a
entry
solvent
[(S)-8]/[(R)-9], mM
[10], mM
acid
RMS
1b
MeOH/H2O (9:1)
0.25
0.25
0.25
0.25
0.25
1.0
0.125
0.125
0.125
0.125
0.125
0.50
NH4Cl (1.25mM)
NH4Cl (1.25mM)
NH4Cl (1.25mM)
NH4Cl (1.25mM)
NH4Cl (1.25mM)
HOAc (2.0mM)
HOAc (2.0mM)
HOAc (2.0mM)
HOAc (2.0mM)
HOAc (2.0mM)
HOAc (2.0mM)
NH4Cl (5.0mM)
NH4Cl (5.0mM)
NH4Cl (5.0mM)
1.98
1.83
1.70
1.66
1.60
2.04
2.15
2.14
2.23
2.12
2.08
1.75
1.79
1.77
2b
MeOH/H2O (3:1)
3b
MeOH/H2O (1:1)
4b
MeOH/H2O (1:3)
5b
MeOH/H2O (1:9)
6c
MeOH/H2O (1:1)
7c
MeOH/H2O (3:1)
1.0
0.50
8c
CH3CN/MeOH/H2O (2:1:1)
THF/MeOH/H2O (2:1:1)
CH2Cl2/MeOH/H2O (1:6:3)
toluene/MeOH/H2O (1:8:1)
MeOH/H2O (1:1)
1.0
0.50
9c
1.0
0.50
10c
11c
12b
13b
14b
1.0
0.50
1.0
0.50
1.0
0.50
CH3CN/MeOH/H2O (1:5:5)
THF/MeOH/H2O (1:5:5)
1.0
0.50
1.0
0.50
a Observed MS enantioselectivity. b Flow injection of a 10-µL sample at 200 µL/min. c Syringe pump infusion at 20 µL/min.
Table 5. Comparison of the Observed
method for the discovery and optimization of chiral selectors,
particularly by combinatorial methods if one considers the alacrity
with which one can obtain the enantioselectivity for a given
analyte.
Enantioselectivities of Analytes 10-18 by HPLC on
(S)-N-(3,5-dintrobenzoyl)Leucinea Derived CSP and the
Enantioselectivity Obtained by Mass Spectrometry
Using Pseudoenantiomeric Chiral Selectors (S)-8 and
(R)-9
normal
phaseb
higher
affinity
CONCLUSIONS
more
analyte
k1
Rd
Observations of chiral recognition in the electrospray ionization
mass spectra, using tertiary amine appended derivatives of N-(3,5-
dinitrobenzoyl)leucine as chiral selectors, have been demonstrated
for a number of chiral analytes. The complex intensity fraction
for either of the selector-analyte complexes, when using a quasi-
racemic mixture of mass-labeled pseudoenantiomeric chiral selec-
tors, in the electrospray ionization mass spectra varies linearly
with the enantiomeric composition of the analyte. This relationship
provides a measure of the extent of enantioselectivity and allows
quantitative determination of the enantiomeric composition. The
results are independent of analyte concentration, provided that
saturation of the association between the selector and analyte is
not observed.
As implemented, each sample requires ∼10 s for analysis,
indicating the potential for rapid/high-throughput analysis. Others
have demonstrated the use of electrospray ionization mass
spectrometry for high-throughput analyses.57 What is lacking from
these is a general method for the determination of enantiomeric
composition using these methods. The method demonstrated
herein should close this gap, whereby the need to derivatize or
isotopically label the analyte is obviated. Additionally, the observed
precision using this method is more than adequate for most high-
throughput analyses where one is often willing to trade some
precision for analysis time.
The chiral selectors used in this study were derived from the
well-established chiral stationary-phase DNB-leucine. Given the
correlation between the enantioselectivities observed chromato-
graphically and by mass spectrometry, one would expect the scope
of analytes that one can assay by this method should be
comparable to the scope of analytes that can be enantioresolved
on the corresponding chiral stationary phase. Given the large
number of known chiral stationary phases, and the scope of chiral
analytes for each, adaptation of these myriad chiral selectors into
appropriate mass spectrometric chiral selectors would expand the
10
11
12
13
14
15
16
17
18
5.12 2.97 1.71 1.89
1.15 1.42 1.11 1.30
1.39 1.14 0.90 1.12
0.65 1.80 1.62 1.36
1.37 1.56 2.38 1.36
0.69 1.20 1.32 1.15
2.77 3.47 5.18 1.88
1.29 1.58 1.14 1.11
5.45 1.21 1.15 1.08
(S)
(S)
(S)
(S)
(S)
(S)
(S)
(R,R)
(S)
1.67 (0.02)
1.47 (0.02)
1.28 (0.02)
1.75 (0.03)
1.53 (0.02)
1.40 (0.03)
2.01 (0.08)
1.16 (0.02)
1.29 (0.02)
(S)
(S)
(S)
(S)
(S)
(S)
(S)
(R,R)
(S)
a See Experimental Section for column details. b Mobile phase:
hexanes/2-propanol (90:10), 2 mL/min. c Mobile phase: methanol/
water (75:25), 1.2 mL/min. d Chromatographic separation factor (ratio
of retention factors for the analyte enantiomers). e [8] ) [9] ) 1.0
mM, [analyte] ) 0.50 mM, and [NH4Cl] ) 5.0 mM in methanol/water
(1:1). f Observed MS enantioselectivity. g σ is the standard deviation
derived from the calibration line.
for both the normal-phase and the reversed-phase conditions.
Additionally, for the solutions assayed by mass spectrometry, the
same solvent system and concentrations were used for all of the
analytes. A linear relationship between the CIF and the enantio-
meric composition of the analyte was observed in all cases. The
sense of observed chiral recognition was consistent between these
two methods in all cases. The analyte enantiomer selectively
retained on the (S)-DNB-leucine CSP complexes to the (S)-chiral
selector to a greater extent in every case.
There is a good correlation between the enantioselectivities
observed by chiral HPLC and mass spectrometry. Plotting RHPLC
versus RMS (not shown) affords a correlation coefficient of 0.67
for the normal-phase data and 0.73 for the reversed-phase data. A
similar correlation has recently been reported by Lindner and co-
workers using cinchona alkaloid chiral selectors.14 Often, there
are differences between selector-analyte solution equilibria and
separation factors obtained chromatographically, primarily due to
“tether effects”. Even so, this will undoubtedly be a valuable
5026 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

set of chiral analytes that can be assayed by this method nearly
congruously with the set of analytes that can be assayed by
enantioselective chromatography. Additionally, it is expected
that this method will be a useful tool for the discovery of new
chiral selectors, particularly by combinatorial methods, which can
then in turn be used for the preparation of new chiral stationary
phases.
Mississippi State Chemical Laboratory, for the use of instrumenta-
tion and assistance with the mass spectrometry experiments.
SUPPORTING INFORMATION AVAILABLE
Procedures for the preparation of chiral selectors 8 and 9
and characterization data for all new compounds. This
material is available free of charge via the Internet at
http://www.pubs.acs.org.
ACKNOWLEDGMENT
Received for review March 14, 2005. Accepted May 31,
2005.
The authors gratefully acknowledge financial support provided
by Mississippi State University. We also thank William E. Holmes,
Director of Mass Spectrometry and Advanced Instrumentation,
AC050438N
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5027