Use of achiral/chiral SFC/MS for the profiling of isomeric cinnamonitrile/hydrocinnamonitrile products in chiral drug synthesis

Article abstract of DOI:10.1021/ac060326b

A directly coupled achiral/chiral SFC/MS method has been developed for the profiling of a three-step stereoselective synthesis of cinnamonitrile and hydrocinnamonitrile intermediates. Semipurified reaction mixtures were screened in one step to determine the diastereomeric/enantiomeric composition of the final product as well as to identify any remaining E/Z isomers present from the starting material. The coupled achiral/chiral column combination was found to significantly enhance the separation of both enantiomers and diastereomers, without adding significantly to the overall analysis time. This analytical technique should prove to be generally useful for the profiling of isomeric reaction products in chiral drug synthesis.

Full text of DOI:10.1021/ac060326b

Anal. Chem. 2006, 78, 3835-3838
Use of Achiral/Chiral SFC/MS for the Profiling of
Isomeric Cinnamonitrile/Hydrocinnamonitrile
Products in Chiral Drug Synthesis
A. J. Alexander*,† and A. Staab‡,§
Analytical Research & Development and Process Research & Development, Bristol Myers Squibb Company,
Pharmaceutical Research Institute, 5 Research Parkway, Wallingford, Connecticut 06492
isomers of the same compound. Thus, the focus of many of the
studies in this area has been on the chiral separation of relatively
simple mixtures,5,6 or more specifically, in the case of SFC/MS,
on the rapid chiral screening of lead compounds in the drug
discovery environment.7 Indirectly coupled achiral and chiral
columns are commonly employed for the chiral analysis of drugs
in complex biological matrixes.5 This form of multidimensional
chromatography is typically implemented on-line via the use of
integrated switching valves and uses the first column (achiral) to
remove interferences from endogenous compounds and to deliver
primarily the drug of interest to the chiral column.8 This is a very
powerful technique as the sample cleanup and analysis steps are
isolated, thus allowing a wide range of columns/mobile phases
to be explored to achieve the desired enantiomeric separation,
but it does require compatible mobile phases for the two
separations.8 Although multidimensional SFC has been described,9
it does not appear to have been widely adopted for complex
mixture analysis. A simpler technique, which potentially allows
for the simultaneous achiral/chiral separation of a range of
components in a complex mixture is the direct serial coupling of
achiral and chiral columns.10 In LC, this approach is limited by
the increased total system backpressure and by the fact that the
mobile phase requirements for the achiral and chiral separations
are generally different.11 In contrast, direct serial coupling in SFC
is easily accomplished because the mobile phase fluid (CO2 plus
modifier) has a significantly lower viscosity than a liquid and is
an effective eluent for both chiral and achiral stationary phases.12
Surprisingly, despite these advantages, no recent applications
using directly coupled achiral/chiral stationary phases in SFC have
been reported (as far as the authors are aware). In this work, the
utility of this technique combined with MS detection is initially
demonstrated for the screening of a three-step synthesis of
A directly coupled achiral/chiral SFC/MS method has
been developed for the profiling of a three-step stereose-
lective synthesis of cinnamonitrile and hydrocinnamoni-
trile intermediates. Semipurified reaction mixtures were
screened in one step to determine the diastereomeric/
enantiomeric composition of the final product as well as
to identify any remaining E/Z isomers present from the
starting material. The coupled achiral/chiral column
combination was found to significantly enhance the sepa-
ration of both enantiomers and diastereomers, without
adding significantly to the overall analysis time. This
analytical technique should prove to be generally useful
for the profiling of isomeric reaction products in chiral
drug synthesis.
Due, in part, to more stringent FDA guidelines for the
marketing of chiral drugs,1 stereoselective drug syntheses are
rapidly becoming the norm in the development of new drug
candidates. However, in the scale-up of synthetic methods, many
other products (such as structural isomers, or diastereoisomers)
are often formed, either in addition to or instead of the expected
enantiomer of interest. This, in-turn, requires the development of
analytical methods that can both separate and characterize
potentially unknown isomeric side products in crude or semipu-
rified reaction mixtures.
Both reversed-phase chiral liquid chromatography-mass spec-
trometry (LC/MS)2 and chiral supercritical fluid chromatogra-
phy-mass spectrometry (SFC/MS)3 have recently been shown
to be powerful techniques for the characterization of chiral
mixtures. Due to the low efficiency and poor achiral selectivity of
most chiral columns, MS detection is essential to distinguish the
enantiomers of interest from other achiral impurities,4 but
importantly, molecular ion data alone cannot be used to distinguish
between the presence of enantiomers, diastereomers, or structural
(5) Ward, T. J.; Hamburg, D.-M. Anal. Chem. 2004, 76, 4635-4644.
(6) Chester, T. L.; Pinkston, J. D. Anal. Chem. 2004, 76, 4606-4613.
(7) Zhao, Y.; Sandra, P.; Woo, G.; Thomas, S.; Gahm, K.; Semim, D. Pharm.
Discovery 2005, February, 30-41.
* To
whom
correspondence
should
be
addressed.
E-mail:
Anthony.Alexander@bms.com.
(8) Fried, K.; Wainer, I. W. J. Chromatogr., B 1997, 689, 91-104.
(9) Juvancz, Z.; Payne, K. M.; Markides, K. E.; Lee, M. L. Anal. Chem. 1990,
62, 1384-1388.
(10) Phinney, K. W.; Dander, L. C.; Wise, S. A. Anal. Chem. 1998, 70, 2331-
2335.
† Analytical Research & Development.
‡ Process Research & Development.
§ Current address: Neurogen Corporation, 35 Northeast Industrial Road,
Branford, CT 06405.
(1) www.fda.gov/cder/guidance/stereo.htm (accessed Feb 2006).
(2) Bakhtiar, R.; Ramos, L.; Tse, F. L. S. Chirality 2001, 13, 63-74.
(3) Garozotti, M.; Hamdan, M. J. Chromatogr., B 2002, 770, 53-61.
(4) Zhao, Y.; Woo, G.; Thomas, S.; Semin, D.; Sandra, P. J. Chromatogr., A
2003, 1003, 157-166.
(11) Snyder, L. R.; Kirkland, J. J.; Glajch, J. L. Practical HPLC Method
Deveolopment, 2nd ed.; Wiley: New York, 1997.
(12) Berger, T. A. Packed Column SFC; RSC Chromatography Monographs:
Cambridge, U.K., 1995. Williams, K. L.; Sander, L. C. J. Chromatogr., A
1997, 785, 149-158.
10.1021/ac060326b CCC: $33.50 © 2006 American Chemical Society
Published on Web 05/05/2006
Analytical Chemistry, Vol. 78, No. 11, June 1, 2006 3835

cinnamonitrile and hydrocinnamonitrile intermediates, which
generates both enantiomeric and diastereomeric products.
EXPERIMENTAL SECTION
Reagents and Chemicals. Carbon dioxide (SFC grade) was
obtained from AirGas (Cheshire, CT). Methanol (HPLC grade)
was purchased from EMD Chemicals, Inc. (Gibbstown, NJ).
Trifluoroacetic acid (TFA), isopropylamine, 1,2,4-triazole, 2-chloro-
2′,4′-difluoroacetophenone, and diethyl(1-cyanoethyl)phosphonate
were purchased from Sigma-Aldrich (St. Louis, MO).
Equipment and Instrumentation. Separations were per-
formed using a Berger Analytical SFC system supplied by Mettler-
Toledo AutoChem, Inc. (Columbia, MD), equipped with an SFC
pump (pump A), a modifier pump (pump B), a column oven, and
a dual-wavelength UV detector. SFC grade liquid CO2 and HPLC
grade methanol were used as mobile phases A and B, respectively.
Mass spectrometric detection was performed using a Sciex
(Concord, Ontario, Canada) model API 4000 triple quadrupole
mass spectrometer utilizing a heated nebulizer (APCI) probe. The
chiral columns, a Diacel Chiralcel OD-H and a Diacel Chiralpak
AD-H (both 2.0 × 150 mm, 5 µm), were purchased from Chiral
Technologies, Inc. (West Chester, PA). The achiral column, a
Phenomenex Luna Silica 2 (4.6 × 150 mm, 5 µm), was purchased
from Phenomenex (Torrance, CA). The SFC was connected
directly to the mass spectrometer via a stainless steel T-piece with
a bore of 150 µm (VICI Valco Instruments, Houston, TX), which
was inserted between the UV detector and the back-pressure
regulator (BPR). The existing UV-to-BPR connection (1/16 o.d.
× 0.020-in. i.d. tubing) was removed, and the T-piece was
connected using two appropriate lengths of 1/16 i.d. × 0.005-in.
i.d. stainless steel tubing and fittings (VICI Valco Instruments).
The final connection to the APCI probe was made with a 5-ft length
of 65-µm i.d. × 1/16 o.d. PEEK tubing (Upchurch Scientific, WA)
using standard finger-tight PEEK fittings. The final split ratio was
not determined.
Figure 1. Synthetic chemistry steps 1 and 2. N-Alkylation of 1,2,4-
triazole. Horner-Wadsworth-Emmons olefination of triazoloketone
3.
Experimental Conditions. The SFC system was operated
isocratically under the following conditions: oven temperature 30
°C, UV detection wavelength 220 nm, and set outlet pressure 120
bar. Flow rates of 0.8 mL/min were used for chiral-only chroma-
tography and 1.0 mL/min for coupled achiral/chiral separations.
Methanol, 5-15 vol %, was used as the modifier for all separations.
The injection volume was 5 µL, and the total run time was 10
min. The mass spectrometer was operated in positive ion mode
under the following conditions: probe temperature 400 °C,
nebulizing gas pressure 50 psi, curtain gas flow rate 0.9 L/min,
discharge current 2 µA, and declustering potential (DP) 70 V. Due
to the low flow rates required for narrow-bore columns, no
nebulizer makeup gas flow or additional heating of the transfer
line was necessary. Data acquisition, 150-500 Da. scan range in
0.5 min, was initiated by a contact closure out signal from the
SFC controller.
Synthetic Chemistry. A three-step synthesis was used to
transform 1,2,4-triazole into a final mixture of hydrocinnamoni-
trile diastereomers and enantiomers, as outlined in Figures 1
and 2. In reaction 1, 1,2,4 -triazole (compound 1) was N-alkylated
with 2-chloro-2′,4′-difluoroacetophenone (compound 2) under
basic conditions to yield a mixture of triazoloketone positional
isomers (compounds 3 and 4). Isomer 3, which was produced
in ∼85% yield, was isolated using radial chromatography over silica
Figure 2. Synthetic chemistry step 3. Dissolving metal reduction
of derivatives 5 and 6.
gel. In reaction 2, ketone 3 was reacted with diethyl(1-cyano-
ethyl)phosphonate under basic conditions (Horner-Wadsworth-
Emmons olefination) to generate a roughly 1:1 mixture of the
E/Z isomers of cinnamonitrile derivatives 5 and 6. In reaction 3,
derivatives 5 and 6 were subjected to a dissolving metal reduc-
tion using a mixture of magnesium in methanol to yield the
expected final products (diastereomers/enantiomers 7-10). The
solution containing the final products was semipurified by filtra-
tion to remove the excess magnesium, followed by aqueous
workup to remove inorganic magnesium salts. Aliquots were
typically diluted 250:1 with methanol prior to chromatographic
analysis.
3836 Analytical Chemistry, Vol. 78, No. 11, June 1, 2006

Figure 3. Chiral SFC/MS separation of semipurified final product.
Upper trace: total ion chromatogram (TIC). Lower trace (dashed):
product [M + H]+ EIC (m/z 263) diastermers/enantiomers 7-10.
Lower trace (solid): starting material [M + H]+ EIC (m/z 261) E/Z
isomers 5, 6.
Figure 5. Coupled achiral and chiral SFC/MS separation of
semipurified final product. Upper trace: TIC. Lower trace (dashed):
product [M + H]+ EIC (m/z 263) diastereomers/enantiomers 7-10.
Lower trace (solid): starting material [M + H]+ EIC (m/z 261) E/Z
isomers 5, 6.
chromatogram clearly shows two well-resolved components of
equal abundance with the expected mass of the product ([M +
H]+ m/z 263), which is consistent with the presence of the two
expected diastereomers 7(10) and 8(9). The achiral/chiral SFC
separation of the same mixture, achieved simply by the serial
coupling of the two columns, is shown in Figure 5. This was
obtained under the conditions optimized for the achiral separation,
that is, at a flow rate of 1 mL/min and with 15% methanol. As can
be seen, a dramatic improvement in resolution has been achieved,
and all components are now baseline-separated. Note also that
this has been achieved without a significant increase in analysis
time (9 min versus the original 7 min). In addition, the relative
intensities of the four products (two enantiomeric and two
diastereomeric species) are now close to being equal, as would
be expected from the mechanism of the dissolving metal reduction
reaction. With regard to peak identity, the combined achiral/chiral
separation would be expected to yield one set of enantiomers,
followed by the other, that is, either 7(10) followed by 8(9), or
8(9) followed by 7(10). However, because the four peaks are
all of nearly equal intensity, this is difficult to confirm without
the use of a chiral detector. To investigate this, an artificial final
product mixture was generated by isolating the two diastereomers
(using achiral normal phase silica gel chromatography) and
recombining them in a ratio 1:4 instead of the natural ratio of 1:1.
This artificial mixture was then subjected to the same achiral/
chiral separation as shown in Figure 5 (data not shown). In this
case, the four previously observed chromatographic peaks were
replaced by two consecutive sets of two peaks (doublets). Each
peak within a given doublet was of equal intensity, and the
intensity ratio between the first and second sets of doublets was
∼1:4. This confirmed the expected elution order of consecutive
sets of enantiomeric pairs.
Figure 4. Achiral SFC/MS separation of semipurified product. Upper
trace: TIC. Lower trace: product [M + H]+ EIC (m/z 263).
RESULTS AND DISCUSSION
On the basis of previous experience, two small-bore (2.0 ×
150 mm, 5 µm) polysaccharide chiral columns, Chiralcel OD-H
and Chiralpak AD-H, were evaluated for this study (data not
shown). The most effective SFC separation of the semipurified
final product mixture (from reaction 3 after ∼2 h) is shown in
Figure 3. This was obtained using the Chiralcel OD-H column at
a flow rate of 0.8 mL/min and with 5% methanol as additive. The
addition of small amounts (e0.1% v/v) of either acid (TFA), or
base (isopropylamine), were not found to improve the peak
symmetry. Note that, although the resolution is poor, the extracted
ion chromatograms (EIC) for starting material (E/Z isomers 5
and 6, [M + H]+ m/z 261) and product (species 7-10, [M +
H]+ m/z 263) can be used to distinguish the partially resolved
components. Interestingly, the four expected product peaks are
present, indicating that this column does exhibit some selectivity
for both the separation of stereoisomers, as well as enantiomers.
The isocratic achiral SFC separation of the same final product
mixture, obtained using an analytical scale silica column (4.6 ×
150 mm), is shown in Figure 4. This was obtained at a flow rate
of 1 mL/min and with 15% methanol, conditions that were not
too different from those used for the chiral separation. The
CONCLUSION
Direct serial coupling of achiral and chiral columns in SFC
has previously been shown to improve the limited achiral selectiv-
ity of chiral stationary phases for the separation and measurement
of enantiomers in drug mixtures.10 It was also expected to be
useful for the analysis of compounds having multiple chiral
Analytical Chemistry, Vol. 78, No. 11, June 1, 2006 3837

centers, when both enantiomers and diastereomers may be
present in a complex sample. We have shown this to be the case
if directly coupled achiral/chiral SFC is further combined with
the specificity of mass spectrometric detection. Specifically, in this
study, we have demonstrated that semipurified reaction mix-
tures of cinnamonitrile and hydrocinnamonitrile intermediates
can be profiled in one step to determine the diastereomeric/
enantiomeric composition of the final product, as well as to identify
any remaining E/Z isomers present from the starting material.
The coupled achiral/chiral column combination was found to
significantly enhance the separation of both enantiomers and
diastereomers without adding significantly to the overall analysis
time.
Note Added after ASAP Publication. Due to a production
error, this paper was posted on the Web on May 5, 2006, before
a correction to the first sentence of the Equipment and Instru-
mentation paragraph had been made. The correction was made,
and the paper was reposted on May 5, 2006.
Received for review February 22, 2006. Accepted April
3, 2006.
AC060326B
3838 Analytical Chemistry, Vol. 78, No. 11, June 1, 2006