Resolution of tert-butyl-1-(2-methylnaphthyl)phosphine oxide using selectors identified from a chemical combinatorial library

Article abstract of DOI:10.1021/ac0203443

Resolution of racemic tert-butyl-1-(2-methylnaphthyl)phosphine oxide 1, a chiral phosphorus compound, was achieved using selectors developed from a small peptide library. Separation factors as high as 3.2 were observed. The library consists of 81 peptide-based potential chiral selectors on polymeric synthesis resins. The linker needed to immobilize the identified chiral selectors onto silica gel proved important in the chiral separation; a longer linker provided a significantly higher separation factor in this study.

Full text of DOI:10.1021/ac0203443

Anal. Chem. 2002, 74, 5212-5216
Re s o lu t io n o f
t e rt -Bu t yl-1 -(2 -Me t h yln a p h t h yl)p h o s p h in e Ox id e
Us in g S e le c t o rs Id e n t ifie d fro m a Ch e m ic a l
Co m b in a t o ria l Lib ra ry
†
J orda n Blodge tt, Ya n Wa ng, Tingyu Li,* Pra s a d L. Pola va ra pu, J oze f Dra bow ic z,
K. Mic ha l Pie trus ie w ic z,‡,§ a nd Krys tyna Zygo
§
Department of Chemistry, Box 351822-B, Vanderbilt University, Nashville, Tennessee 37235-1822, Center of Molecular and
Macromolecular Studies, Polish Academy of Sciences, 93-236 Ł o´ d z´ , Sienkiewicza 112, Poland, Department of Organic
Chemistry, Maria Curie-Skłodowska University, Lublin, and Institut of Organic Chemistry, Polish Academy of Sciences,
Warsaw, Poland.
Resolution of racemic ter t-butyl-1 -(2 -methylnaphthyl)-
phosphine oxide 1 , a chiral phosphorus compound, was
achieved using selectors developed from a small peptide
library. Separation factors as high as 3 .2 were observed.
The library consists of 8 1 peptide-based potential chiral
selectors on polymeric synthesis resins. The linker needed
to immobilize the identified chiral selectors onto silica gel
proved important in the chiral separation; a longer linker
Fig u re 1 . R and S enantiomers of racemic tert-butyl-1-(2-methyl-
naphthyl)phosphine oxide 1.
provided a significantly higher separation factor in this
study.
throughput method for chiral selector development. However, in
Because human enzymes and cell surface receptors are all of
chiral structures, enantiomers of a racemic compound may be
absorbed, activated, and degraded by them in different manners.
In some instances, this phenomenon causes two enantiomers of
a racemic drug to have different or even opposite pharmacological
activities. To acknowledge these differing effects, the biological
activity of each enantiomer often needs to be studied separately.
This and other factors within the pharmaceutical industry have
contributed significantly to the need for enantiomerically pure
compounds and, thus, the need for enantioselective chromatog-
raphy.1
many of these reported examples, significant enantioselective
separation of the chosen analytes or their close analogues had
already been achieved prior to the combinatorial library studies.
In this article, we would like to report an example for which the
enantioselective separation of the analyte has not been well-studied
previously.
The analyte of interest is tert-butyl-1-(2-methylnaphthyl)phos-
phine oxide 1 (Figure 1).3 In this compound, there are two
chirogenic elements corresponding to the configurationally stable
stereogenic phosphorus atom and the stereolabile aromatic-
phosphorus rotation axis. At room temperature, because of the
stereolabile aromatic-phosphorus rotation axis, oxide 1 exists as
two instead of four diastereomers. Unlike compounds with chiral
carbon atoms, far fewer synthetic methodologies are available to
prepare enantiomerically pure compounds with chiral phosphorus
To develop efficient enantioselective stationary phases, several
groups, including ours, have investigated the development of chiral
2
selectors from chemical combinatorial libraries. A number of
approaches using either mixture combinatorial libraries or parallel
combinatorial libraries (the high throughput method) have been
investigated. These reported examples have demonstrated the
feasibility of the mixture combinatorial library or the high-
atoms. Consequently, chromatographic resolution of such com-
pounds takes on greater importance.4
EXPERIMENTAL SECTION
*
To whom correspondence should be addressed. Phone/ Fax: 615 343 8466.
E-mail: tingyu.li@vanderbilt.edu.
(For a list of the chemical abbreviations, see ref 5.)
†
Center of Molecular and Macromolecular Studies, Polish Academy of
General Supplies and Equipment. Solid-phase synthesis
resins and amino acid derivatives were purchased from NovaBio-
chem (San Diego, CA). All other chemicals and solvents were
purchased from either Aldrich (Milwaukee, WI), Fluka (Ronkonko-
ma, NY), or Fisher Scientific (Pittsburgh, PA). HPLC grade
Sciences.
‡
Maria Curie-Skłodowska University
Institut of Organic Chemistry, Polish Academy of Sciences.
§
(
(
1) Stinson, S. C. Chem. Eng. News 1 9 9 5 , 73 (41), 44-74.
2) For examples, see the following papers and references cited within: (a)
Welch, C. J.; Protopopova, M. N.; Bhat, G. Enantiomer 1 9 9 8 , 3, 471-476.
(
b) Lewandowski, K.; Murer, P.; Svec, F.; Frechet, J. M. J. Chem. Commun.
1
9 9 8 , 2237. (c) Bluhm, L.; Wang, Y.; Li, T. Anal. Chem. 2 0 0 0 , 72, 5201-
(3) Gasparrini, F.; Lunazzi, L.; Mazzanti, A.; Pierini, M.; Pietrusiewicz, K. M.;
5
205.
Villani, C. J. Am. Chem. Soc. 2 0 0 0 , 122, 4776.
5212 Analytical Chemistry, Vol. 74, No. 20, October 15, 2002
10.1021/ac0203443 CCC: $22.00 © 2002 American Chemical Society
Published on Web 09/14/2002

hydroxide (1.06 g, 26.5 mmol) in water (100 mL) was added
gradually Fmoc-Osu (8.50 g, 25.2 mmol) in 100 mL of THF. After
being stirred at room temperature for 4 h, the reaction mixture
was adjusted to pH 1-3 by adding hydrochloric acid. The solution
was extracted with ethyl acetate (3 × 100 mL). The organic
extracts were washed with water (3 × 100 mL) and then dried
over anhydrous sodium sulfate. Evaporation of the solvent yielded
Fig u re 2 . The 81-member library. Dnb: 3,5-dinotrobenzoyl; Hyp:
trans-4-hydroxyproline; Abu: 4-aminobutyric acid (NH2(CH2)3CO2H);
AmPS: aminomethylated polystyrene resin. Unless indicated other-
wise, amino acids adopt the L configuration.
Allsphere silica gel (particle size, 5 µm; pore size, 80 Å; surface
2
area, 220 m / g) was purchased from Alltech (Deerfield, IL).
the crude product (9.5 g), which was recrystallized from ethyl
Selecto silica gel (32-63 µm) from Fisher Scientific was used for
flash column chromatographic purification of target compounds.
Thin-layer chromatography was completed using EM silica gel
1
acetate to give the pure product (9.3 g, 84%). H NMR (CDCl
3
):
δ 12 (s, 1H), 7.8 (d, J ) 7.2 Hz, 2H), 7.6 (d, J ) 7.2 Hz, 2H),
.2-7.4 (m, 5 H), 4.3 (d, J ) 6.9, 2H), 4.2 (t, J ) 6.9, 1H), 3.0 (t,
J ) 6 Hz, 2H), 2.2 (t, J ) 6 Hz, 2H), 1.0-1.6 (m, 16H).
P reparation of Dnb- -Asn-Thr-Aun-NH(CH silica Sta-
7
6
0 F-254 TLC plates (0.25 mm) (EMerck, Merck KGaA, 64271
Darmstadt, Germany). Elemental analyses were conducted by
Atlantic Microlab, Inc. (Norcross, GA). HPLC analyses were
completed with a Beckman analytical gradient system (System
Gold). UV spectra were obtained with a Shimadzu UV 201
spectrometer (cell volume, 3 mL; cell pass length, 10 mm). The
HI-TOP manual synthesizer required for parallel library synthesis
was from Whatman Polyfiltronic (Rockland, Ma, USA).
D
2 3
)
tionary P hase. Other stationary phases were prepared by similar
reactions. To Rink acid resin (3.00 g, 1.29 mmol/ g) preswelled
with DCM (20 mL, 30 min) were added Fmoc-Aun-OH (1.64 g,
3
.9 mmol), DIC (0.49 g, 3.9 mmol), DMAP (157 mg, 1.29 mmol),
and NMM (130 mg, 1.29 mmol) in DCM-DMF (2:1, 20 mL). After
the mixture was agitated for 5 h, the resin was collected by
filtration and washed with DMF, IPA, and DCM (20 mL × 2).
The Fmoc group was then removed by treatment with 20%
piperidine in DMF (20 mL) for 30 min. The deprotected Aun-
ORink resin was collected and washed with DMF, IPA, and DCM
(20 mL × 2).
P reparation of the P arallel 81-Member Library. The library
(
Figure 2) was synthesized using the polyfiltronic HI-TOP manual
synthesizer. The experimental procedure for the synthesis of the
Dnb- -Asn-Thr member of the library is given below. Other library
D
members were prepared following a similar sequence. To 75 mg
6
(
0.030 mmol in Abu) of Abu-AmPS resin in one well of the 96-
well unifilter microplate, were added mixtures of Fmoc-Thr(tBu)-
OH (16.4 mg, 0.0528 mmol), PyBOP (28.0 mg, 0.0528 mmol), and
DIPEA (7.0 mg, 0.053 mmol) in 0.50 mL of DMF. After agitating
for 2 h, the resins were filtered and washed with DMF, DCM,
IPA, and DCM. The Fmoc protecting group was then removed
by treatment with 0.60 mL of 20% piperidine in DMF for 20 min,
To Aun-ORink resin prepared above were added Fmoc-Thr-
(tBu)-OH (1.54 g, 3.9 mmol), PyBOP (2.0 g, 3.9 mmol), HOBt
(174 mg, 1.29 mmol), and DIPEA (666 mg, 5.2 mmol) in DMF
(20 mL). After agitating for 2 h, the resin was filtered and washed
with DMF, IPA, and DCM (10 mL× 2). The Fmoc group was
then removed by the same procedures discussed in the previous
paragraph.
followed by washing with DMF. Fmoc-D-Asn(Trt)-OH and dini-
trobenzoic acid were then coupled to the resin following an
identical reaction sequence. After that, the side chain protecting
groups of Thr and D-Asn were removed by reacting with 0.6 mL
The Thr(tBu)-Aun-ORink resin was then coupled with Fmoc-
Asn(Trt)-OH by following exactly the same procedure as described
above. After removing the Fmoc group, dinitrobenzoic acid was
of 95% TFA (2.5% water and 2.5% triisopropyl silane in TFA) for 1
h. The resin was then washed with DMF, DCM, IPA, and DCM,
respectively, to yield the desired library member on the solid resin.
Screening of the P arallel Library with Chiral HP LC. To
each well that contained resin was added racemic oxide 1 (0.15
coupled to D-Asn(Trt)-Thr(tBu)-Aun-ORink resin. This coupling
reaction was repeated one more time to ensure complete reaction.
The reaction conditions are identical to those used for the coupling
of Fmoc-Thr(tBu)-OH to the resin.
The resin was then treated with 1% TFA in DCM (20 mL, 10
3
mg) in CHCl -heptane (2:8, 0.6 mL). After incubating the mixture
min) to release Dnb-D-Asn(Trt)-Thr(tBu)-Aun-OH from the resin.
for 3 h, the supernatants were filtered into a collection plate using
the Polyfiltronic filtration manifold. The enantiomeric ratios of the
supernatants were then analyzed by HPLC using a homemade
Dnb-Leu column (column size, 50 × 4.6 mm; mobile phase, 15%
IPA/ hexane; flow rate, 1.2 mL/ min.; UV detection at 230 nm).
P reparation of Fmoc-Aminoundecanoic Acid (Fmoc-Aun-
OH). To 11-aminoundecanoic acid (5.33 g, 26.5 mmol) and sodium
The crude product obtained was purified by flash column chro-
matography on silica gel (mobile phase, 10% MeOH in DCM) to
yield the desired Dnb-
D
-Asn(Trt)-Thr(tBu)-Aun-OH as a white
solid (1.90 g, 100%). H NMR: (DMSO-d ) δ 0.98 (3H, d, J ) 6
Hz), 1.05 (9H, s), 1.12-1.45 (16H, m), 2.1 (2H, t, J ) 7.2 Hz), 2.8
2H, m), 3.0 (2H, m), 3.9 (1H, q, J ) 3 Hz), 4.1 (1H, dd, J ) 6
1
6
(
Hz), 5.0 (1H, m), 7.1 (15H, s), 9.0 (1H, t, J ) 2.1 Hz), 9.1 (2H, d,
+
J ) 2.1 Hz). ES-MS (m/ z): 931.7 (M + Na) .
(
4) Absolute configuration of this compound was unknown at the time when
this research was initiated but has been determined recently. See: Wang,
F.; Wang, Y.; Polavarapu, P. L.; Li, T.; Drabowicz, J.; Pietrusiewicz, K. M.;
Zygo, K. J. Org. Chem., in press.
A mixture of Dnb-D-Asn(Trt)-Thr(tBu)-Aun-OH (581 mg, 0.64
mmol), PyBOP (333 mg, 0.64 mmol), HOBt (43 mg, 0.32 mmol),
and DIPEA (124 mg, 0.96 mmol) in DMF (5 mL) was then added
to aminopropyl silica gel (0.70 g, 0.32 mmol in amino group). After
agitating for 3 h, the silica gel was collected by filtration and
washed with DMF, IPA, and DCM (5 mL × 2). The above
(
5) Abu: 4-aminobutyric acid; Ahe: 6-aminohexanoic acid; Aoc: 8-aminooctanoic
acid; Aun: 11-aminoundecanoic acid; AmPS: aminomethylated polystyrene;
NMM: N-methylmorpholine; PyBop: benzotriazolyloxy-tris[pyrrolidino]-
phosphonium hexafluorophosphate; Fmoc: 9-fluorenylmethoxycarbonyl;
DIC: diisopropylcarbodiimide; HOBt: 1-hydroxybenzotriazole; TFA: tri-
fluoroacetic acid; DMAP: 4-(dimethylamino)pyridine; DIPEA: N,N-diiso-
propylethylamine; DCM: dichloromethane; DMF: dimethyl formamide;
IPA: 2-propanol; Fmoc-Osu: 9-Fluorenylmethyloxycarbonyl-hydroxysuc-
cinimide; Dnb: 3,5-dinitrobenzoyl.
coupling cycle was repeated one more time with Dnb-D-Asn(trt)-
Thr(tBu)-Aun-OH (290 mg, 0.32 mmol), PyBOP (166 mg, 0.32
mmol), HOBt (22 mg, 0.16 mmol), and DIPEA (83 mg, 0.64
mmol). The unreacted free amine group on the silica gel was end-
(
6) Wang, Y.; Li, T. Anal. Chem. 1 9 9 9 , 71, 4178-4182.
Analytical Chemistry, Vol. 74, No. 20, October 15, 2002 5213

S c h e m e 1 . P re p a ra t io n o f t h e Dn b -D-As n -Th r Me m b e r o f t h e P a ra lle l Lib ra rya
a
Conditions: (a) Fmoc-Abu-OH, PyBop; (b) (1) piperidine, (2) Fmoc-Thr(tBu)-OH, PyBop; (c) (1) piperidine, (2) Fmoc-D-Asn(Trt)-
OH, PyBop; (d) (1) piperidine, (2) Dnb-OH, PyBop; and (e) TFA.
2 3
S c h e m e 2 . S yn t h e s is o f Dn b -D-As n -Th r-Ab u -(CH ) -S ilic a Ge l S t a t io n a ry P h a s e
(
a) Fmoc-Abu-OH, DIC; (b) (1) piperidine, (2) Fmoc-Thr(tBu)-OH, PyBop; (c) (1) piperidine, (2) Fmoc-D-Asn(Trt)-OH, PyBop; (d) (1)
piperidine, (2) Dnb-OH, PyBop; (e) AcOH; (f) aminopropylsilica gel, PyBop; and (g) TFA.
capped by reacting with acetic anhydride (654 mg, 6.4 mmol) and
pyridine (505 mg, 6.4 mmol) in 10 mL of DCM for 2 h. The silica
gel was then collected by filtration. The protecting groups of Thr
Ser, Thr), aromatic stacking (Trp), or a combination of both (His,
Tyr). Structural diversity is achieved by the application of amino
acids with side chains that are significantly different from each
9
and
D
-Asn were then removed by reacting with 10 mL of 95% TFA
other, including a
D
-amino acid. In addition, the use of the electron
(
2.5% water and 2.5% triisopropyl silane in TFA) for 2 h. The
deficient dinitrobenzoyl (Dnb) group as the N-terminal end-
capping group of the peptides provide another potential interaction
site with the analyte.10
desired chiral stationary phase was collected and washed with
DCM, DMF, IPA, and methanol and then dried in vacuo. On the
basis of elemental analysis, the surface ligand concentration was
determined to be 0.17 mmol/ g. The silical gel was packed into a
7
The resin chosen for the synthesis of this library was ami-
nomethylated polystyrene (AmPS) resin, a derivative of the widely
used Merrifield resin. Unlike silica gel, such a resin allows the
reliable synthesis of the library. The synthesis of this library was
performed conveniently using a Hi-top filter plate manual synthe-
sizer.11 In this synthesizer, resin can be added to the wells of a
96-well filter plate. By adding individual amino acids into each
well, 96 different peptides can be synthesized in one run. This
Hi-top system allows for quick filtration between each synthetic
step without the need to remove any resin from the plate, and
the overall efficiency of this synthetic process is very high. In
terms of the chemistry involved, Fmoc solid phase synthesis was
chosen, and the detailed chemistry is illustrated in Scheme 2 with
5
0 × 4.6 mm HPLC column.
Chromatographic Measurements. The Retention factor (k)
equals (t
dead time. Dead time t
zene as the void volume marker according to the literature
procedure. The dead time was measured to be 0.43 min with a
flow rate of 1.2 mL/ min for the column prepared above using CH
Cl / heptane (7/ 3) as the mobile phase.
r
- t
0
)/ t
0
, in which t
r 0
is the retention time and t is the
0
was measured with 1,3,5-tri-tert-butylben-
8
2
-
2
RESULTS AND DISCUSSION
The library designed is a dipeptide library consisting of two
amino acid modules (Figure 2). The two modules are identical,
both contain the same nine amino acids. All of the possible
combinations of two modules yield a library containing a total of
the synthesis of the Dnb-
synthesis, commercially available, side-chain-protected forms of
Thr, the N-R-Fmoc-O-tert-butyl- -threonine (Fmoc-Thr(tBu)-OH)
D-Asn-Thr member of the library. In this
L
8
1 members. The amino acids in the peptide module are chosen
(
(
7) Yang, A.; Li, T. Anal. Chem. 1 9 9 8 , 70, 2827-2830.
8) Pirkle, W. H.; Welch, C. J. J. Liq. Chromatogr. A 1 9 9 1 , 14, 1-8.
on the basis of consideration of the molecular features of the target
analyte. The analyte contains a flat, aromatic naphthalene ring in
addition to a hydrogen bond acceptor group. The amino acids
chosen contain side chains that could interact with the analyte
(9) Billiot, E.; Warner, I. M. Anal. Chem. 2 0 0 0 , 72, 1740-1748.
10) (a) Pirkle, W. H. Chirality 1 9 9 7 , 9, 103-104. (b) Wainer, I. W.; Caldwell,
J. Chirality 1 9 9 7 , 9, 95-96.
(
(
11) The Hi-top system, Polyfiltronics, 100 Weymouth Street, Rockland, MA.
through either hydrogen bonding interaction (Asn,
D-Asn, Hyp,
02370.
5214 Analytical Chemistry, Vol. 74, No. 20, October 15, 2002

Ta b le 1 . Ra t io s o f P e a k He ig h t s Ob t a in e d fo r t h e S u p e rn a t a n t s fo r t h e 8 1 Me m b e rs o f t h e Lib ra rya
His-His
.168
D-Asn-His
1.292
Hyp-His
1.135
Trp-His
1.191
Ser-His
1.218
Thr-His
1.279
Asn-His
1.121
Gln-His
1.152
Tyr-His
1.137
1
His-D-Asn
.135
D-Asn- D-Asn
1.489
Hyp-D-Asn
1.145
Trp-D-Asn
1.145
Ser-D-Asn
1.203
Thr-D-Asn
1.371
Asn-D-Asn
1.065
Gln-D-Asn
1.224
Tyr-D-Asn
1.131
1
His-Hyp
.157
D-Asn-Hyp
1.212
Hyp-Hyp
1.149
Trp-Hyp
1.209
Ser-Hyp
1.230
Thr-Hyp
1.334
Asn-Hyp
1.081
Gln-Hyp
1.177
Tyr-Hyp
1.173
1
His-Trp
.155
D-Asn-Trp
1.506
Hyp-Trp
Trp-Trp
1.177
Ser-Trp
1.269
Thr-Trp
1.300
Asn-Trp
0.968
Gln-Trp
1.151
Tyr-Trp
1.143
1
His-Ser
.128
D-Asn-Ser
1.380
Hyp-Ser
1.142
Trp-Ser
1.122
Ser-Ser
1.221
Thr-Ser
1.308
Asn-Ser
1.047
Gln-Ser
1.162
Tyr-Ser
1.128
1
His-Thr
.158
D-Asn-Thr
1.718
Hyp-Thr
1.135
Trp-Thr
1.152
Ser-Thr
1.197
Thr-Thr
1.269
Asn-Thr
1.014
Gln-Thr
1.151
Tyr-Thr
1.096
1
His-Asn
.162
D-Asn-Asn
1.329
Hyp-Asn
1.142
Trp-Asn
1.129
Ser-Asn
1.303
Thr-Asn
1.162
Asn-Asn
0.976
Gln-Asn
1.091
Tyr-Asn
0.997
1
His-Gln
.200
D-Asn-Gln
1.326
Hyp-Gln
1.155
Trp-Gln
1.219
Ser-Gln
1.259
Thr-Gln
1.303
Asn-Gln
1.085
Gln-Gln
1.166
Tyr-Gln
1.201
1
His-Tyr
.131
D-Asn-Tyr
1.240
Hyp-Tyr
1.124
Trp-Tyr
1.154
Ser-Tyr
1.252
Thr-Tyr
1.365
Asn-Tyr
1.021
Gln-Tyr
1.115
Tyr-Tyr
1.155
1
a
Ratio of the peak heights of racemic mixture 1 is 1.193. Library template, Dnb-Aa1-Aa2-Abu-AmPS.
Ta b le 2 . Ch ira l Re s o lu t io n o f Ra c e m ic
t e rt -Bu t yl-1 -(2 -m e t h yln a p h t h yl)p h o s p h in e Ox id e 1
and of
D
-Asn, N-R-Fmoc-N-â-trityl-D-asparagine (Fmoc-D-Asn(Trt)-
OH) are needed. Other library members were synthesized
following similar reactions.
surface
linker
length
loading
For screening purpose, an equal amount of the racemic analyte
kra
CSPs
R
(mmol/ g)
(
(
0.15 mg, 0.00061 mmol) in a mixture of chloroform and heptane
2:8, 0.60 mL) was added to each of the wells that contained 0.03
Dnb-D-Asn-Thr-Abu-NH(CH2)3silica (CH2)3 58
1.02
0.16
0.17
0.17
0.18
0.18
Dnb-D-Asn-Thr-Ahe-NH(CH2)3silica (CH2)5
Dnb-D-Asn-Thr-Aoc-NH(CH ) silica (CH2)7
3.8 1.9
1.8 2.3
mmol of selector. After equilibration, the supernatants were
collected into a collection plate. Enantiomeric ratio of the
supernatant was analyzed using a Dnb-Leu column prepared
previously in the laboratory, which yielded a separation factor of
2
3
Dnb-D-Asn-Thr-Aun-NH(CH2)3silica (CH2)10 5.9 3.2
Dnb-Asn-Asn-Aun-NH(CH2)3silica (CH2)10 3.2 1.7
a k is the retention factor of the least retained enantiomer. Mobile
r
phase, 7/ 3 CH2Cl2/ heptane.
1
.2. The ratio of the two peak heights was used as the relative
measure of the enantiomeric purity. For the racemic analyte 1 ,
the ratio was measured to be 1.193. Ratios of peak heights
obtained for the supernatants for all 81 wells of the library are
summarized in Table 1.
As seen from Table 1, several promising leads were identified.
Some of the ratios are higher than the ratio for racemic analyte
and some others are lower, indicating selectors with opposite
selectivities exist in the library. It is also noted that several
stationary phase was then packed into a column using a standard
slurry packing method.12
Surprisingly, the resulting stationary phase provided a separa-
tion factor of only 1.02. The lack of correlation could be attributed
to the differences of the solid supports used in these experiments.
For resin equilibration experiments, the selector was attached to
a polymer resin suitable for solid-phase synthesis, whereas
chromatographic experiments were performed on silica gel. Unlike
polymeric resin, silica gel support is hydrophilic and is covered
with silanol groups. Potential nonspecific interaction of these
surface polar silanol groups with the analyte could lead to a
reduction in chiral selectivity.
We then immobilized the selector, using longer linkers, hoping
that the distance of separation between the chiral selector and
silica gel surface would minimize the impact of the polar surface.
Three new stationary phases were prepared, one with an amino-
hexanoic acid (Ahe) linker, one with an aminooctanoic acid (Aoc)
linker, and one with an aminoundecanoic (Aun) acid linker. They
were prepared according to essentially the same procedure as
described for the immobilization of selector with the short Abu
linker (Scheme 2).
promising leads contain
used previously in chiral micellar electrophoresis studies. The
D
-Asn, which has been investigated and
9
ratio of the peak heights for D-Asn-D-Asn is 1.489, 25% higher than
that of the racemic mixture; In contrast, the ratio of the peak
heights for Asn-Asn is 0.976, 22% lower than that of the racemic
mixture. These two results demonstrate the reliability of the
screening method.
The first attempt was to immobilize the most promising
selector
.44 as calculated from its peak height ratio in Table 1, using a
D-Asn-Thr, which corresponds to an enantiomer ratio of
1
short Abu linker. Immobilization of this selector onto silica gel
was achieved according to the chemistry described in Scheme 2.
The key in this synthetic scheme was the use of an acid-sensitive
resin (Rink acid resin) to prepare a peptide acid fragment in which
the side chains of Thr and D-Asn are still protected. The protected
peptide acid fragment was then coupled to aminopropylsilica gel.
After coupling to the solid support, the protecting groups of Thr
Separation factors obtained for these stationary phases are
summarized in Table 2. Excellent resolution of the phosphorus
and
D-Asn were removed by treatment with TFA, yielding the
(
12) Poole, C. F.; Poole, S. K. Chromatography today; Elsevier: New York, 1991;
pp 350-353.
desired stationary phase on the solid support. The resulting
Analytical Chemistry, Vol. 74, No. 20, October 15, 2002 5215

Fig u re 3 . Chiral resolution of tert-butyl-1-(2-methylnaphthyl)phosphine oxide 1 with Dnb-D-Asn-Thr-Aun-NH(CH2)3silica. First eluting peak
4
was assigned an R configuration.
analyte was achieved when aminoundecanoic acid was used as
the linker (Figure 3). Dependence of the separation factors on
linker length is apparent from Table 2.
chromatographic experiments are possible, because the chemical
properties of the synthesis resins are different from the chro-
matographic resins. However, a longer linker could be used to
minimize the impact of solid support in order to maintain good
correlation. Other approaches to avoid the possible discrepancy
are being explored. Such a difference in the properties of the solid
supports in these two types of experiments also prevents a
quantitative correlation at this time. With improved surface
modification procedure or different support for synthesis or for
chromatography, such a correlation may become feasible in the
future.
Another selector, Dnb-Asn-Asn, which yields the opposite
selectivity as the
D-Asn-Thr selector and which corresponds to
an enantiomer ratio of 0.82 as calculated from its peak height ratio
in Table 1, was subsequently immobilized with the amino-
undecanoic acid linker. Excellent resolution of the target racemic
analyte was also achieved with this selector. As expected, the
opposite elution order was observed with this stationary phase.
One can, of course, prepare a stationary phase using the opposite
enantiomer of
elution order.
D-Asn-Thr, Asn-D-Thr to provide such a reverse
ACKNOWLEDGMENT
CONCLUSIONS
Results of this study demonstrated the practical potential of
the library screening method using synthesis resins. Unlike
analytes used for our previous model studies, this target analyte
has not been well-studied previously in chiral separation. In
addition, resolution of chiral phosphorus compounds are of special
interest, because far fewer synthetic methods are capable of
preparing this class of compounds in enantiomerically pure form.
The excellent separation factors achieved enable the preparative
resolution of this racemic compound.
The financial support from NIH (GM 63812, GM 60637 for T.
Li) and NSF (CHE0092922 for P. L. Polavarapu) are greatly
appreciated. Studies in Ł o´ d z´ were supported in part by the State
Committee for Scientific Research (KBN Grant 4TO9A 105 22 for
J.D.). We also thank Dr. Louis H. Bluhm for assistance.
Received for review May 21, 2002. Accepted July 29,
2002.
It should be pointed out that discrepancies between the
outcomes of the screening experiment and the subsequent
AC0203443
5216 Analytical Chemistry, Vol. 74, No. 20, October 15, 2002