Page 109 - Chiral Separation Techniques

3.7 Combinatorial Libraries of Selectors for HPLC 85

of two powerful selectors, DNB-L-leu-gly and DNB-L-ala-gly in agreement with the
results of the parallel screening [87].
These two selectors terminated with a glycine were then prepared on a larger
scale, their carboxyl groups reacted with 3-aminopropyltriethoxysilane, and the con-
jugate immobilized onto silica. Each CSP was packed into columns and used for the
separation of racemic (1-naphthyl)leucine ester 17. Separation factors of 6.9 and 8.0
were determined for the columns with DNB-ala-gly and DNB-leu-gly selector
respectively. These were somewhat lower than those found for similar CSPs using
the parallel synthesis and attached through a different tether [87].
The major weakness of this method appears to be the limited size of the library
that can be screened. Although the authors believe that their method is well suited to
screen medium-sized libraries with up to a few hundred members, the most impor-
tant limit is the requirement of having equal sets of both L and D building units. This
is easily achieved with amino acids that provide relatively large diversity at a rea-
sonable cost. However, this may be a serious problem with many other families of
chiral compounds. The other drawback of the current implementation is the large
number of sublibraries that must be synthesized and screened to specify the best
selector. In fact, the number of sublibraries in the published procedure equaled the
number of members of the original library thus making the expected acceleration
effect of this combinatorial approach questionable. However, the use of the mass
spectroscopic detection during the first two parallel screening separations of Fig. 3-
16 would afford molecular weights of the separated compounds that are specific for
each individual selector. If the retention of any of the injected compounds is the
same for both D- and L-libraries, no selectivity occurs. In contrast, different retention
times in both runs indicate selectivity and even allow an estimation of selectivity fac-
tors. Such an approach might totally avoid the tedious multistep deconvolution pro-
cess and accelerate the screening procedures.

3.7.4 Library-On-Bead

Our group also demonstrated another combinatorial approach in which a CSP carry-
ing a library of enantiomerically pure potential selectors was used directly to screen
for enantioselectivity in the HPLC separation of target analytes [93, 94]. The best
selector of the bound mixture for the desired separation was then identified in a few
deconvolution steps. As a result of the “parallelism advantage”, the number of columns
that had to be screened in this deconvolution process to identify the single most selec-
tive selector CSP was much smaller than the number of actual selectors in the library.
Our strategy consisted of the following steps: A mixture of potential chiral selec-
tors is immobilized on a solid support and packed to afford a “complete-library col-
umn”, which is tested in the resolution of targeted racemic compounds. If some sep-
aration is achieved, the column should be “deconvoluted” to identify the selector
possessing the highest selectivity. The deconvolution consisted in the stepwise
preparation of a series of “sublibrary columns” of lower diversity, each of which
constitute a CSP with a reduced number of library members.