The enantiomeric separation of 4,5-disubstituted imidazoles by HPLC and CE using cyclodextrin-based chiral selectors

Article abstract of DOI:10.1080/10610278.2010.506545

The enantiomeric separation of a series of fifteen racemic 4,5-disubstituted imidazole compounds was examined by high performance liquid chromatography (HPLC) and capillary electrophoresis (CE). These alkaloid analytes are important precursors for the total synthesis of the natural product calcaridine A and other Leucetta-derived alkaloids. Therefore, the enantiomeric analysis of these analytes is not only important in the production of enantiomerically pure calcaridine A, but also for a better understanding of the stereochemistry involved in related biosynthetic pathways. Several bonded cyclodextrin (both native and derivatised) stationary phases were evaluated for their ability to separate these racemates via HPLC. Likewise, several cyclodextrin derivatives were evaluated for their ability to separate this set of compounds via CE. Using HPLC, 14 of the 15 racemic compounds were separated. Eight of the analytes were separated using CE with resolutions up to 7.0. Using both HPLC and CE approaches, the entire set of analytes was separated. The optimisation of these separations was discussed and a comparison between the chiral selectors used was made. Lastly, the similarity of the 15 analytes allowed for insight into the mechanism of chiral recognition.

Full text of DOI:10.1080/10610278.2010.506545

This article was downloaded by: [New York University]
On: 01 June 2015, At: 14:49
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
Supramolecular Chemistry
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/gsch20
The enantiomeric separation of 4,5-disubstituted
imidazoles by HPLC and CE using cyclodextrin-based
chiral selectors
Zachary S. Breitbach a , Qing Feng a , Panduka B. Koswatta a , Edra Dodbiba a , Carl J. Lovely
a & Daniel W. Armstrong a
a Department of Chemistry and Biochemistry , The University of Texas at Arlington ,
Arlington, Texas, 76019, USA
Published online: 24 Aug 2010.
To cite this article: Zachary S. Breitbach , Qing Feng , Panduka B. Koswatta , Edra Dodbiba , Carl J. Lovely & Daniel W.
Armstrong (2010) The enantiomeric separation of 4,5-disubstituted imidazoles by HPLC and CE using cyclodextrin-based chiral
selectors, Supramolecular Chemistry, 22:11-12, 758-767, DOI: 10.1080/10610278.2010.506545
To link to this article: http://dx.doi.org/10.1080/10610278.2010.506545
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

Supramolecular Chemistry
Vol. 22, Nos. 11–12, November–December 2010, 758–767
The enantiomeric separation of 4,5-disubstituted imidazoles by HPLC and CE using
cyclodextrin-based chiral selectors
Zachary S. Breitbach, Qing Feng, Panduka B. Koswatta, Edra Dodbiba, Carl J. Lovely and Daniel W. Armstrong*
Department of Chemistry and Biochemistry, The University of Texas at Arlington, Arlington, Texas 76019, USA
(Received 13 May 2010; final version received 29 June 2010)
The enantiomeric separation of a series of fifteen racemic 4,5-disubstituted imidazole compounds was examined by high
performance liquid chromatography (HPLC) and capillary electrophoresis (CE). These alkaloid analytes are important
precursors for the total synthesis of the natural product calcaridine A and other Leucetta-derived alkaloids. Therefore, the
enantiomeric analysis of these analytes is not only important in the production of enantiomerically pure calcaridine A, but
also for a better understanding of the stereochemistry involved in related biosynthetic pathways. Several bonded
cyclodextrin (both native and derivatised) stationary phases were evaluated for their ability to separate these racemates via
HPLC. Likewise, several cyclodextrin derivatives were evaluated for their ability to separate this set of compounds via CE.
Using HPLC, 14 of the 15 racemic compounds were separated. Eight of the analytes were separated using CE with
resolutions up to 7.0. Using both HPLC and CE approaches, the entire set of analytes was separated. The optimisation of
these separations was discussed and a comparison between the chiral selectors used was made. Lastly, the similarity of the
15 analytes allowed for insight into the mechanism of chiral recognition.
Keywords: enantiomeric separations; chiral separations; high performance liquid chromatography; capillary
electrophoresis; cyclodextrin; calcaridine A
Introduction
biosynthesis of spiroleucettadine. Recently, calcaridine A
has been successfully synthesised through a multi-step
process that concluded with an oxidative rearrangement
of a 4,5-disubstituted imidazole (16–17). Throughout
this synthetic pathway, several newly synthesised chiral
alkaloids were produced. These chiral precursors were
synthesised and used in their racemic form, such that,
when the oxidative rearrangement step was performed,
the synthetic material was formed as a racemic mixture
of calcaridine A. In addition to calcaridine A, the
C14-epimer was formed in the oxidative rearrangement.
The enantiomeric separation and purification of any of
these racemic precursors would allow for the production of
enantiomerically pure calcaridine A. Furthermore, enan-
tioselective separation techniques that are applicable to
separate these types of compounds may be helpful in
understanding when and how proper stereochemistry is
introduced throughout the biosynthetic pathways invol-
ving these alkaloids. For these reasons, the stereoselective
separation of these precursors is of relevance.
The synthesis of natural products is of great interest
to organic and medicinal chemists. Bioactive molecules
and their metabolites, as well as any related chemical
derivatives or analogues, are an abundant resource for the
development of new pharmaceutical compounds (1–6).
The intricate structural constitution of many naturally
occurring molecules can make synthesis a challenging
undertaking. Furthermore, the stereochemical require-
ments often increase the burden of the task and may
require the use of asymmetric synthetic pathways or
stereoselective separation techniques. Considering the
potential pharmacological importance of natural products,
researchers are continually searching for new, undiscov-
ered molecules. Recognising that about 70% of the earth’s
surface is water, many of these researchers have intuitively
turned to the seas and marine organisms in search of these
compounds (7–18). Currently, one group of organisms
which has come under scrutiny is marine sponges (10–18).
The Fijian sponge, Leucetta sp., has been shown to be a
good source of some very interesting alkaloid imidazole
compounds, such as calcaridine A, spirocalcaridine A and
B, and spiroleucettadine (11–17). The latter, spiroleucet-
tadine, has been shown to possess antibacterial properties
that are comparable to those of vancomycin and penicillin
(15). Calcaridine A is thought to be a precursor in the
Currently, the most commonly used technique for
enantiomeric separations is high performance liquid
chromatography (HPLC) (19). One popular set of chiral
HPLC stationary phases is cyclodextrins and their
derivatives. Cyclodextrins have been used to separate
a vast number of chiral analytes and have shown
*Corresponding author. Email: sec4dwa@uta.edu
ISSN 1061-0278 print/ISSN 1029-0478 online
q 2010 Taylor & Francis
DOI: 10.1080/10610278.2010.506545
http://www.informaworld.com

Supramolecular Chemistry
759
to be extremely applicable towards the separation of
racemic analytes containing aromatic functionality
(20–25). Under reverse phase (RP) solvent conditions,
aromatic moieties often form inclusion complexes with the
cyclodextrins. Separation by HPLC offers the advantage of
preparative capabilities, which could aid in the production
of the pure enantiomers of calcaridine A in the long-term.
Additionally, capillary electrophoresis (CE) may offer
another means by which these racemic precursors could be
resolved. Cyclodextrins and their derivatives dominate the
field of chiral selectors used in CE (26–28). Though these
separations offer limited potential in the preparative
production of enantiomerically pure calcaridine A, they
could be useful for determining enantiomeric excesses of
asymmetric synthetic steps.
capillaries (50 mm i.d.) for CE were purchased from
Polymicro Technologies (Pheonix, AX, USA).
Methods
All HPLC analyses were performed on one of two
chromatographic systems, both of which were produced
by Shimadzu (Kyoto, Japan). The pumps used were
models LC-6A and LC-10A, the detectors were SPD-6A
and SPD-10A, and the data were recorded using a SPD-6A
chromatographic integrator. In both cases, the samples
were injected using a 6-port injector equipped with 10 mL
injection loops.
All CE experiments were performed on Beckman
Coulter P/ACE MDQ (Fullerton, CA, USA) CE system
equipped with an on-column UV–visible detector.
In this study, the enantiomeric separation of 15 newly
synthesised 4,5-disubstituted imidazoles was explored.
Both HPLC and CE were tested and compared for their
abilities to separate the enantiomers of these chiral
analytes. To our knowledge, none of these compounds
have previously been subjected to any chiral chromato-
graphic or electrophoretic technique. Also, the structural
similarity of this series of analytes may allow some insight
into their chiral recognition by cyclodextrins and provide a
basis for exploration of related systems.
HPLC analyses
All samples were dissolved in MeOH for analyses done in
RP and in ACN when the tests were performed in the polar
organic mode (PO). The wavelength of detection for all
HPLC experiments was 254 nm. H4NOAc buffers were
adjusted to a desired pH using HOAc. The mobile phase
compositions listed throughout this manuscript are in
volume to volume ratios. All mobile phases were degassed
by ultrasonication under vacuum for 5 min before use.
The mobile phase flow rate throughout the entire study was
1 ml/min.
Experimental
Materials
HPLC grade acetonitrile (ACN) and methanol (MeOH)
were purchased from VWR (West Chester, PA, USA).
Deionised water was produced by a Milli-Q system
(Billerica, MA, USA). Ammonium acetate (H4NOAc),
glacial acetic acid (HOAc), triethylamine (TEA), sodium
phosphate monobasic and dibasic, phosphoric acid and
sodium hydroxide (NaOH) were obtained from Sigma-
Aldrich (Milwaukee, WI, USA). All the chiral analytes
listed in Figure 1 were synthesised according to the
literature by Koswatta et al. (16–18). Note, compound 5
possesses an azide group and was stored away from light.
All HPLC columns used were of analytical dimensions
(25 cm £ 4.6 mm, 5 mm) and were obtained from
Advanced Separation Technologies (Whippany, NJ,
USA). The columns used were as follows: b-CD
(b-cyclodextrin), HP-RSP (high performance hydroxypro-
pyl ether b-cyclodextrin), RSP (hydroxypropyl ether
b-cyclodextrin), AC (acetylated b-cyclodextrin), RN and
SN (R- and S-naphthylethyl carbamate b-cyclodextrin),
DMP (dimethylphenyl carbamate b-cyclodextrin), DNP
(2,6-dinitro-4-trifluoromethylphenyl ether b-cyclodextrin)
and DM (dimethylated b-cyclodextrin). CE chiral selectors
were obtained from Sigma-Aldrich (Milwaukee, WI, USA)
and are as follows: SBCD (sulphated b-cyclodextrin),
DMBCD (dimethylated b-cyclodextrin) and HPBCD
(hydroxypropyl b-cyclodextrin). The untreated fused silica
CE analyses
Unmodified fused-silica capillary was thermostated at
258C, with the length being 30 cm to the end and 20 cm to
the detection window. UV detection was accomplished at
214 nm. Sodium phosphate, monobasic and dibasic (1:1),
was dissolved in deionised water and adjusted with
concentrated phosphoric acid to the desired pH. Buffer
additives (b-cyclodextrin derivatives) were added directly
into the buffer. All samples were dissolved in a solution
consisting of 20% methanol and 80% water. Normal
polarity (NP) and reversed polarity analyses used voltages
of þ 7 and 2 7 kV, respectively.
At the beginning of each series of experiments, the
capillary was rinsed with 1 M sodium hydroxide for 5 min
and water for 5 min for conditioning. Between each run,
the capillary was flushed with 1 M sodium hydroxide for
2 min, water for 2 min and buffer for 2 min. Then, the
sample solution was injected hydrodynamically at 0.5 psi
for 5 s.
Calculations
The HPLC retention factors (k10 ), selectivities
(a) and resolution (Rs) values were calculated as
follows: k10 ¼ (tr1 2 t0)/tr1, a ¼ (tr2 2 t0)/(tr1 2 t0) and

760
Z.S. Breitbach et al.
HO
A
CH3
O
N
H2N
N
O
O
H3C
H3C
Calcaridine A
NH2
B
NH2
H3C
H3C
CH3
O
CH3
H3C
N
N
N
N
N
N
N
N
O
O
OH
O
CH3
CH3
CH3
CH3
CH3
O
O
OH
O
CH3
O
OH
O CH3
4
CH3
1
3
2
N3
CH3
H3C
H3C
H3C
N
O
N
N
N
N
N
N
N
O
OH
O
OH
O
OH
O
CH3
CH3
CH3
O
O
O
O
CH3
O CH3
5
6
7
8
CH3
O
H3C
CH3
CH3
O
N
N
N
N
N
N
N
N
OH
O
O
CH3
CH3
CH3
CH3
I
H
I
O
CH3
O
CH3
O
10
9
12
11
CH3
H3C
N
O
CH3
OH
N
CH3
N
N
N
N
O
OH
I
I
CH3
CH3
Si
CH3
CH3
CH3
CH3
CH3
O
CH3
CH3
Si
CH3
O
O
CH3
CH3
13
14
15
Figure 1. Structure of calcaridine A (A) and the structures of 15 racemates that were subjected to enantioselective HPLC and CE
analyses in this study (B).
R ¼ 2 (tr2 2 tr1)/W1 þ W2), where tr1 and tr2 are the
In CE, the migration time of the first peak is represented
by tm1. The efficiencies (N), selectivities (a) and resolution
(Rs) values were calculated as follows: N ¼ 16(tm/W)2,
*
s
retention times of the first and second eluting peaks, t0 is
the dead time (found by calculating the void volume of the
column)andW1 andW2 arethebaselinewidthsofthefirstand
second eluting peaks.
a ¼ mapp1/mapp2, R ¼ 2 (t 2 tr1)/W1 þ W2), where
*
s
r2
mapp1 and mapp2 represent the apparent electrophoretic

Supramolecular Chemistry
761
mobilities of the first and second peaks. Dimethyl
sulphoxide was used as the electroosmotic flow (EOF)
marker. All other constants are the same as above.
value observed in this study (Rs ¼ 2.0 for compound 10). It
is important to note that just because a CSP does not
appear in Table 1, does not necessarily mean it did not
produce any separations, because Table 1 lists only the
best results.
Results and discussion
A more complete comparison between the nine CSPs
tested in this study can be seen in Figure 2, which depicts the
total number of partial (0.4 , Rs , 1.5) and baseline
(Rs . 1.5) separations obtained using each chiral selector.
The chiral stationary phase that produced the greatest
number of separated compounds was the HP-RSP/RSP
media. This chiral selector offered enantioselectivity for
66% of the compounds tested. Additionally, this CSP
produced four baseline separations; a number that was
matched only by the S-naphthylethyl carbamate (SN)
derivative of b-cyclodextrin. For this reason, it was
determined that this stationary phase should be the
recommended starting point when attempting to separate
similar analytes. The chiral selectors AC, b-CD, SN, RN
and DMP (Figure 2) separated only six to seven of the
analytes. Keeping in mind that the DMP phase was
previously acknowledged as the one that gave the greatest
number of ‘best’ separations, it is interesting to see that it
only ranks modestly compared to other CSPs in terms of
total separations. This is a result of the DMP phase being
complementary to the other phases tested, which allowed
for the separation of certain compounds that were not
separable when using the other chiral selectors. The CSPs
that performed poorly were the DNP and the DM, both of
which only showed partial separations for four analytes and
no baseline separations.
Analytes
Figure 1(A) shows the structure of calcaridine A. Also, in
Figure 1(B), the chemical structures of the 15, racemic
synthetic intermediates are shown. As can be seen, this set
of intermediates offers systematic changes in functional-
ities, as well as placement of the stereogenic centre (being
alpha to either the 4-position or the 5-position of the
imidazole core).
HPLC analyses
A summary of the optimised HPLC separations data is
given in Table 1. The optimised conditions were considered
to be when the resolution (Rs) values were greatest and the
retention factors (k0) were satisfactorily small. Enantios-
electivity (a) was observed for 14 of the 15 compounds.
These a values ranged from 1.08 (compounds 12 and 13) to
1.29 (compound 6). Compound 15 was not separable under
the conditions used. The Rs values for these compounds
ranged from 0.6 (compound 13) to 2.0 (compound 10).
Seven of the compounds had Rs values greater than or equal
to 1.5 and were considered to be baseline separated.
As can be seen in Table 1, the 3,5-dimethylphenyl-
derivatised b-cyclodextrin (DMP) chiral stationary phase
(CSP) produced the most favourable HPLC separations for
5 of the 15 compounds. The hydroxypropyl-b-cyclodex-
trin derivatives (HP-RSP and RSP), as well as the native
b-cyclodextrin (b-CD) chiral stationary phase, each
accounted for three of the baseline separations. In fact, it
was the native b-CD phase which produced the greatest Rs
In this study, both RP and PO modes were tested.
The normal phase was also attempted, but the initial results
were very poor. In the RP mode, both ACN and methanol
were tested as organic modifiers. In general, using ACN as
the organic modifier resulted in sharper peaks with smaller
Table 1. Summary of the optimised HPLC enantiomeric separations.
Analyte no.
CSPa
k10
a
Rs
Mobile phase
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
HP-RSP
DMP
RN
4.46
3.53
11.81
4.55
3.45
2.04
0.78
1.55
7.97
3.73
4.71
4.95
1.10
5.00
–
1.15
1.13
1.06
1.10
1.09
1.29
1.17
1.16
1.19
1.27
1.14
1.08
1.08
1.09
–
1.6
1.6
1.3
1.6
1.1
1.8
1.3
1.1
1.6
2.0
1.3
1.3
0.6
1.7
–
85/15 20 mM H4OAc/ACN, pH ¼ 4.1
80/20 20 mM H4OAc/ACN, pH ¼ 4.1
75/25 20 mM H4OAc/ACN, pH ¼ 4.1
85/15 20 mM H4OAc/ACN, pH ¼ 4.1
85/15 20 mM H4OAc/ACN, pH ¼ 4.1
95/5 20 mM H4OAc/ACN, pH ¼ 4.1
99/1/.3/.2 ACN/MeOH/HOAc/TEA
80/20 20 mM H4OAc/ACN, pH ¼ 4.1
85/15 20 mM H4OAc/ACN, pH ¼ 4.1
75/25 20 mM H4OAc/ACN, pH ¼ 4.1
95/5 20 mM H4OAc/ACN, pH ¼ 4.1
80/20 20 mM H4OAc/ACN, pH ¼ 4.1
80/20 20 mM H4OAc/MeOH, pH ¼ 4.1
85/15 20 mM H4OAc/ACN, pH ¼ 4.1
–
SN
HP-RSP
DMP
DMP
b-CD
b-CD
b-CD
DMP
DMP
AC
RSP
N/A
a For a list of the defined CSP abbreviations, see the Experimental section.
Notes: N/A means no enantioseparation was observed for this compound under any condition. For the structures of the analytes, refer to Figure 1.

762
Z.S. Breitbach et al.
11
10
9
Partial
Baseline
8
7
6
5
4
3
2
1
0
CSP
Figure 2. A representation of the total number of partial and baseline separations obtained using HPLC with different chiral stationary
phases. Note, ‘in PO’ means the separations were obtained in the polar organic mode, whereas all other separations were obtained in the
RP mode. For a list of the defined CSP abbreviations, see the Experimental section.
retention factors, whereas using methanol yielded the
opposite, i.e. broader, longer retained peaks. For this
reason, ACN typically resulted in better (higher Rs)
separations. As indicated in Table 1, only once did methanol
produce a better separation than ACN (compound 13).
Considering that all these compounds have an imidazole
core, it was necessary to use a buffer (ammonium acetate in
this study). Both the pH and the buffer concentration were
examined in the optimisation process. The optimum pH and
buffer concentrations were determined to be 4.1 and
20 mM, respectively. Changing the ratio of the aqueous to
organic co-solvents was the most important tool in
manipulating the separations. When using cyclodextrin-
based CSPs in the RP, it is important to allow hydrophobic
inclusion complexation to occur (21, 23). The organic
modifier competes with the analytes for the non-polar
cavity, thus increases in the ACN (or methanol) percentage
resulted in smaller retention factors. The occurrence of
inclusion complexation with cyclodextrins has been
extensively studied (29). Throughout this work, optimum
aqueous/organic ratios ranged from 75/25 to 95/5.
just three and two analytes, respectively. However, when
compared to the separations obtained in the RP mode, the
separations found using the PO mode were of greater
efficiency and of lower retention. This is a typical feature
of the PO mode and arises from the faster adsorption–
desorption kinetics involved in the hydrogen-bonding
interactions that occur at the mouth of the cyclodextrin
cavity versus the slower kinetics involved with inclusion
complexation (31). A prime example of this can be seen in
Figure 3, where the separation of compound 7 on the DMP
CSP is shown. Figure 3(A) shows the separation in PO
mode and Figure 3(B) shows the separation in the RP.
A
Conversely, the PO mode does not permit inclusion
complexes to form as the high organic content in the
mobile phase completely occupies the cyclodextrin cavity
(30). Rather, the dominant interaction in this operation is
through hydrogen bonding. Thus, evaluation in the PO
mode was also performed on all the CSPs which possessed
free hydroxyl groups. Operating in the PO mode resulted in
very few separations for this set of analytes. As listed
in Table 1, only one of the optimised separations
(compound 7) was obtained using PO mobile phases.
Figure 2 shows all the separations obtained in the PO mode.
The only CSPs that showed any enantioselectivity were the
DMP and native b-CD stationary phases, which separated
B
Time (min)
Figure 3. HPLC enantiomeric separations of compound 7 in the
polar organic mode (A) and in RP (B). The CSP used was DMP.
The mobile phase conditions were: (A) 99/1/.3/.2;
ACN/MeOH/HOAc/TEA and (B) 75/25; 20 mM H4NOAc (pH
4.1)/ACN. For other chromatographic conditions see the
Experimental section. See Figure 1 for the analyte structure.

Supramolecular Chemistry
763
Notice the highly efficient separation and short analysis
time in the upper chromatogram, while the lower
chromatogram is of poorer efficiency and longer retention.
Keep in mind that for this set of analytes, this is only one of
a few examples for which the PO mode gave the superior
separation.
and 14. When these analytes interact with the cyclodextrins
via inclusion complexation, there is no competition as
to which hydrophobic portion of the analyte will occupy
the cyclodextrin’s cavity. Thus, the stereogenic centre is
always near to the included moiety (i.e. the aromatic ring).
Compounds that have hydrophobic aromatic moieties in
both the 4-position and 5-position will dynamically
compete for inclusion in the cyclodextrin cavity. This
being the case, the stereogenic centre may not always be
near to the chiral selector. Examples of analytes that posed
such difficulties were compounds 1, 2, 3, 4 and 9. Yet, there
were compounds (e.g. compounds 7 and 12) which
possessed only one hydrophobic substituent, but were still
difficult to separate. Since these two analytes share an
additional terminal benzyl moiety, it is reasonable that
this is the cause of these molecules being more challenging
to separate. It is evident that the benzyl groups will
most easily occupy the stationary phase cavity, which will
in turn place the stereogenic centre further from the chiral
selector.
Generally, the main cause for such poor results in the
PO mode was the lack of sufficient retention. Usually, the
compounds that respond well to being chromatographed in
the PO mode were those that had smaller retention factors
in reversed phase conditions. Conversely, compounds that
retained well in the RP mode were not well retained in the
PO mode. An example of this phenomenon is shown in
Figure 4, where the retention curves for two different
analytes (compounds 1 and 14) are shown. The left-hand
region of the plot shows the retention in RP and the
right-hand region shows the retention in the PO mode.
Note that the curves cross each other when the separation
mode is changed. This indicates that compound 14 relies
heavily on the formation of an inclusion complex for
retention, whereas compound 1 does not. In the PO mode,
when inclusion complexation is not available, compound 1
gives rise to longer retention. Apparently, compound 1 can
interact with the CSP through hydrogen bonding more
than compound 14 can. Comparing the structure of these
two analytes reveals that compound 1 has more possible
hydrogen-bonding sites than compound 14.
Compound 15 was the only analyte that was not
separated by any CSP in either mode of operation. In fact,
it was difficult in most cases to even get this molecule to
elute from the stationary phase. This tenacious retention is
surely an effect of the presence of the tert-butyldimethyl-
silane (TBDMS) functionality. However, compound 13
also possesses the TBDMS group and enantioselectivity
was observed for it. This finding also supports the previous
conclusion that analytes with two hydrophobic portions
may compete with each other for the cavity of the
cyclodextrin. Hence with compound 13, the addition of
the second substituent allowed for shorter retention and
allowed for enantioselectivity to be observed.
Perhaps the most interesting trends, which can lead to
insights to the chiral recognition mechanism, were noted in
the RP mode. First, it was observed that the analytes which
possess aromatic substituents in either the 4-position or 5-
position, but not both, were most easily separated.
Examples of such analytes include compounds 6, 10, 11
Compound 14
Compound 1
9
8
7
Reverse phase mode
Polar organic mode
6
5
4
3
2
1
0
–1
Mobile phase composition
Figure 4. Retention curves for two of the analytes chromatographed using HPLC. The mobile phase composition in the RP region is
20 mM H4OAc (pH 4.1)/ACN. The mobile phase composition in the polar organic region is ACN/MeOH/HOAc/TEA. The stationary
phase was native b-CD. See Figure 1 for analyte structures.

764
Z.S. Breitbach et al.
In fact, of the 12 total separations, only 4 of them were
CE analyses
obtained using a chiral selector other than pure SBCD and
only 1 was found without use of any SBCD. As can be seen,
HPBCD was unable to separate any analytes on its own,
rather it needed to be used in combination with SBCD to
obtain three separations. The only other cyclodextrin
derivative determined to be useful for these enantiomeric
separations was DMBCD, which separated just one analyte.
Another facet to Figure 5 is the distinction of the
separations found in the reverse polarity mode versus
those found in the NP mode. Totally 9 of the 12
separations were found in the NP mode, which, therefore,
appears to be the mode of choice. Also, when comparing
those separations obtained using just SBCD, the NP
conditions still give rise to a greater number of separations,
but by a narrower margin. In the NP mode, the EOF causes
the bulk solution to migrate in the direction of the cathode,
whereas the anionic chiral selectors (SBCD) migrate
towards the anode (away from the detection window),
giving rise to countercurrent interactions between the
chiral selector and the analytes. As noted earlier, the
optimum pH found in the NP mode was 8.0. This is
because higher pH conditions are required for a strong
EOF. At this pH, the analytes will migrate and elute as
neutral molecules. Conversely, the optimum pH found
when using the reverse polarity mode was 2.6. In this
mode, the EOF must be strongly suppressed, and the
electrophoretic movement of anionic chiral selectors will
be towards the anode (towards the detection window). For
these reasons, separations done in the NP mode were
generally faster, but the resolution was usually lower. This
effect is shown in Figure 6, where the separation of
compound 1 was performed with both reverse polarity
(Figure 6(A)) and NP (Figure 6(B)). Compound 1 was
baseline separated under both conditions, but the analysis
In addition to the evaluation by HPLC, all the analytes were
also subjected to enantioselective CE using derivatised
cyclodextrin additives as chiral selectors. A summary of
the optimised CE separations is given in Table 2. As can be
seen, 8 out of the 15 compounds were separated by CE.
Enantiomeric selectivity values (a) ranged from 1.02
(compound 7) to 1.34 (compound 11). Separation
efficiency (N) values were observed from 2000 (compound
11) to 27,000 (compound 7). Lastly, resolution values (Rs)
ranged from 0.8 (compound 7) to 7.0 (compound 4). Of the
eight compounds that were separated using CE, five of
them were baseline separated, with the worst baseline
separation still exhibiting a resolution greater than 2.0.
For the optimised separations shown in Table 2, the
background electrolyte composition and pH, as well as
the chiral selector used, are also listed. The concentration of
the background electrolyte was evaluated at 10 mM intervals
from 10 to 100 mM in order to determine the proper
concentration. As can be seen, typical optimum concen-
trations fell between 30 and 70mM. The pH of the
electrolyte was also adjusted for optimum separations.
As listed in Table 2, optimised pH values were found to be
approximately 3 and 8, depending onwhich (reversed or NP)
mode was being employed. This is discussed in detail below.
Almost certainly, the most effective means by which an
enantiomeric separation can be modified in CE is by
changing the chiral selector itself. Seven of the eight best
separations (Table 2) were obtained using SBCD as the
chiral selector. Throughout this work, several cyclodextrin-
based chiral selectors were evaluated. Figure 5 compares
the efficacy of all the selectors tested. SBCD again shows its
superiority by occupying two-thirds of the entire chart,
as well as being a ‘co-selector’ for three other separations.
Table 2. Summary of the optimised CE enantiomeric separations.
a
Analyte no.
tm1
N
a
Rs
Conditionb
1
2
3
9.20
12.97
N/A
19000
13000
–
1.15
1.21
–
4.8
5.5
–
30 mM SBCD 50 mM phosphate, pH ¼ 2.6
50 mM SBCD 50 mM phosphate, pH ¼ 8.0
–
4
5
19.33
N/A
21000
–
1.22
–
7.0
–
70 mM SBCD 50 mM phosphate, pH ¼ 8.0
–
6
7
8
12.85
10.14
N/A
5000
27000
–
1.08
1.02
–
1.3
0.8
–
60 mM SBCD 50 mM phosphate, pH ¼ 8.0
30 mM SBCD 50 mM phosphate, pH ¼ 3.0c
–
9
10.78
N/A
10.39
N/A
N/A
4000
–
2000
–
1.19
–
1.34
–
–
–
1.05
2.1
–
2.7
–
–
–
1.3
50 mM SBCD 50 mM phosphate, pH ¼ 2.6
10
11
12
13
14
15
–
45 mM SBCD 50 mM phosphate, pH ¼ 8.0
–
–
–
–
N/A
30.30
–
14000
30 mM DMBCD 50 mM phosphate, pH ¼ 3.6
a tm1 is listed in minutes.
b SBCD and DMBCD stand for sulphated b-cyclodextrin and dimethylated b-cyclodextrin, respectively.
c Sample was made and injected quickly to prevent hydrolysis of acetal and ketone formation.
Notes: N/A means no enantioseparation was observed for this compound under any condition. For the structures of the analytes, refer to Figure 1.

Supramolecular Chemistry
765
analyte listed in Table 1 that was separated with a neutral
chiral selector (DMBCD). When compound 15 was run
using SBCD and a basic buffer, the analyte was not
eluted. It can be concluded that the attraction between
SBCD and compound 15 was too strong, thus it was not
a suitable chiral selector. It is probably not coincidental
that compound 15 would not elute during HPLC
analyses in RP conditions. Fortunately, the use of
DMBCD at acidic pH values allowed for the
enantiomeric separation of this analyte. When the pH
was low enough to ionise the analyte, the electrophoretic
mobility of the positively charged compound 15 was able
to overcome the strength of the binding to the slowly
moving chiral selector which reaches the detector in a
reasonable time.
Number of compounds separated in CE using
different chiral selectors
DMBCD in NP
1
SBCD in NP
HPBCD/SBCD in NP
3
5
SBCD in RP
3
Figure 5. Pie chart representing the distribution of separation
among different chiral selectors. Note, the 12 separations
represented here are not necessarily unique. DMBCD, SBCD and
HPBCD/SBCD stand for dimethyl-b-CD, sulphated-b-CD, and
a mixture of hydroxypropyl-b-CD and sulphated-b-CD,
respectively. NP and RP denote normal polarity and reversed
polarity modes.
Comparison between HPLC and CE enantiomeric
separations
It is interesting to make a brief comparison of the results
obtained using HPLC to those obtained using CE for the
enantiomeric separation of these compounds. First, it
should be understood that the main importance of this work
is the ability to possibly purify one of these precursors for
use in the production of enantiomerically pure calcaridine
A. From this viewpoint, the HPLC results are more
valuable. However, the separations made by CE are not
without merit. For example, rapid separations of these
compounds may be useful in following the stereochemistry
as it is introduced and/or controlledthroughout biosynthetic
pathways. For such a study, the potentiality for preparative
separations is not necessary. Also, rapid enantiomeric
excess measurements may be desirable if these analytes are
the result of asymmetric syntheses.
Using HPLC, 14 of the 15 compounds were separated,
whereas using CE, only 8 were separated. Though HPLC
appears to be more broadly applicable to separating these
compounds, it should be noted that when a separation
was obtained by both separation techniques, the CE method
usually produced greater resolution. This is a common
observation in the general comparison between HPLC and
CE analyses, and this study was no exception. For example,
the greatest Rs found using HPLC was 2.0, whereas the best
CE separation had an Rs of 7.0. Figure 7 compares some
separations obtained using HPLC (Figure 7(A) and (B))
and CE (Figure 7(C) and (D)). The comparison made in the
separation of compound 2 by HPLC (Figure 7(A)) and by
CE (Figure 7(C)) shows the typical advantage of using CE.
However, there were some cases where the HPLC
separation was superior. One such example is shown in
the separation of compound 6 by HPLC (Figure 7(B))
and by CE (Figure 7(D)). These observations indicate
that both techniques have distinct advantages for this set
of analytes. Also, it is important to note that the one
compound that was not separated by HPLC was separated
Figure 6. A comparison of the separation of compound 1 by CE
in the reverse polarity mode (A) and in the normal polarity mode
(B). Conditions were 50 mM phosphate buffer (pH¼2.6(A) and
8.0(B)) and 30 mM SBCD. For other conditions, see
Experimental section. See Figure 1 for analyte structure.
time is shorter for the separation done in the NP mode.
However, it should be noted that for this analyte, the
efficiency and resolution were far greater for the
separation performed in the reverse polarity mode.
There was one exception to the pH rules outlined
above. Compound 15 was separated in the NP mode, yet
the optimum pH of the buffer was 3.6. This is the only

766
Z.S. Breitbach et al.
Figure 7. Enantioselective HPLC (A and B) and CE (C and D) analyses of compounds 2 (A and C) and 6 (B and D). The chiral selectors
were DMP and SBCD for HPLC and CE, respectively. The HPLC mobile phases were 80/20 (A) and 95/5 (B); 20 mMH4NOAc
(pH¼4.1)/ACN. The CE experiments were done in normal polarity using 50 mM phosphate buffer (pH¼8.0) with 50 mM and 60 mM
SBCD (C and D, respectively). See Experimental section for other details. See Figure 1 for analyte structures.
by CE (compound 15). Furthermore, by using both
techniques, the entire set of analytes was separated.
References
(1) Kumar, K.; Waldmann, H. Angew. Chem. Int. Ed. 2009, 48,
3224–3242.
(2) Newman, D.J.; Cragg, G.M. J. Nat. Prod. 2007, 70,
461–477.
(3) Koch, M.A.; Schuffenbauer, A.; Scheck, M.; Wetzel, S.;
Casaulta, M.; Odermatt, A.; Ertl, P.; Waldmann, H. Proc.
Natl. Acad. Sci. USA 2005, 102, 17272–17277.
(4) Wetzel, S.; Schuffenhauer, A.; Silvio, R.; Ertl, P.;
Waldmann, H. Chimia 2007, 61, 355–360.
(5) Breinbauer, R.; Vetter, I.R.; Waldmann, H. Angew. Chem.
Int. Ed. 2002, 41, 2879–2890.
(6) Toure, B.B.; Hall, D.G. Chem. Rev. 2009, 109, 4439–4486.
(7) Morris, J.C.; Andrew, P.J. Marine Nat. Prod. Rep. 2009, 26,
245–265.
Conclusions
In this study, 15 chiral heterocycles were subjected to
enantioselective HPLC and CE analyses using b-cyclo-
dextrin and its derivatives. These compounds are
important precursors in the synthesis of calcaridine A.
The combination of HPLC and CE allowed for the
separation of all 15 analytes. HPLC seems to be more
broadly applicable for their separation, whereas the CE
separations obtained were generally of high resolution.
It was determined that in HPLC, RP separations using the
Cyclobond RSP and DMP phases resulted in the greatest
number of separations. Throughout the CE analyses,
SBCD proved to be the dominant chiral selector and the
NP mode produced more separations. In summary, using
cyclodextrin-based chiral selectors via HPLC and CE is a
viable means for separating this important set of analytes.
Furthermore, this work will likely prove useful in future
studies in the production of calcaridine A, as well as in the
study of other related natural products.
(8) Mancini, I.; Guella, G.; Defant, A. Mini-Rev. Med. Chem.
2008, 8, 1265–1284.
(9) Blunt, J.W.; Copp, B.R.; Munro, M.H.G.; Northcote, P.T.;
Prinsep, M.R. Nat. Prod. Rep. 2004, 21, 1–49.
(10) Akee, R.K.; Carroll, T.R.; Yoshida, W.Y.; Scheuer, P.J.;
Stout, T.J.; Clardy, J. J. Org. Chem. 1990, 55, 1944–1946.
(11) He, H.Y.; Faulkner, D.J.; Lee, A.Y.; Clardy, J. J. Org.
Chem. 1992, 57, 2176–2178.
(12) Crews, P.; Clark, D.P.; Tenney, K. J. Nat. Prod. 2003, 66,
177–182.
(13) Du, H.; He, Y.; Rasapalli, S.; Lovely, C.J. Synlett. 2006, 7,
965–992.
(14) Mancini, I.; Guella, G.; Debitus, C.; Pietra, F. Helv. Chim.
Acta 1995, 78, 1178–1184.
(15) Ralifo, P.; Crews, P. J. Org. Chem. 2004, 69, 9025–9029.
(16) Koswatta, P.B.; Sivappa, R.; Dias, H.V.R.; Lovely, C.J.
Org. Lett. 2008, 10, 5055–5058.
Acknowledgements
(17) Koswatta, P.B.; Sivappa, R.; Dias, H.V.R.; Lovely, C.J.
Synthesis 2009, 17, 2970–2982.
(18) Koswatta, P.B.; Lovely, C.J. Tetrahedron Lett. 2009, 50,
4998–5000.
We gratefully acknowledge the National Institute of Health
(NIH 5 R01-GM53825-13) and NIH GM065503) and the Welch
foundation (Welch Y-1362) for their financial support of this
work.

Supramolecular Chemistry
767
ˇ´
´
(19) Alves, G.; Fortuna, A.; Falcao, A. Trends Chromatogr.
2008, 4, 1–10.
(20) Armstrong, D.W.; DeMond, W.J. Chromatogr. Sci. 1984,
22, 411–415.
(26) Egger, M.D.; Liu, Y.; Sevcık, J.; Tesarova, E.; Rozhkov, R.;
Larock, R.C.; Armstrong, D.W. Electrophoresis 2003, 24,
2650–2656.
(27) Lantz, A.; Rozhkov, R.; Larock, R.C.; Armstrong, D.W.
Electrophoresis 2004, 25, 2727–2734.
(28) Jiang, C.; Armstrong, D.W.; Lantz, A.W.; Peter, A.; Toth,
G. J. Liq. Chromatogr. Relat. Technol. 2007, 30,
1421–1436.
(21) Armstrong, D.W.; Ward, T.J.; Armstrong, R.D.; Beesley,
T.E. Science 1986, 232, 1132–1135.
(22) Armstrong, D.W.; DeMond, W.; Czech, B.P. Anal. Chem.
1985, 57, 481–484.
´
´
(29) Bertrand, G.L.; Faulkner, J.R., Jr.; Han, S.M.; Armstrong,
D.W. J. Phys. Chem. 1989, 93, 6863–6867.
(30) Wang, C.; Jiang, C.; Armstrong, D.W. J. Sep. Sci. 2008, 31,
1980–1990.
´
´
(23) Sun, P.; Wang, C.; Armstrong, D.W.; Peter, A.; Forro, E.
J. Liq. Chromatogr. Relat. Technol. 2006, 29, 1847–1860.
(24) Armstrong, D.W.; Chang, C.D.; Lee, S.H. J. Chromatogr. A
1991, 539, 83–90.
(31) CyclobondTM Handbook. Bonded Cyclodextrin Phases for
Liquid Chromatography, 7th ed.; Advanced Separation
Technologies Inc., USA, 2005.
(25) Armstrong, D.W.; Faulkner, J.R., Jr.; Han, S.M.J.
Chromatogr. A 1988, 452, 323–330.