Reversal of optical induction in transamination by regioisomeric bifunctionalized cyclodextrins

Article abstract of DOI:10.1016/S0968-0896(98)00193-X

Two isomeric compounds have been synthesized carrying a pyridoxamine on C-6 of β-cyclodextrin and an imidazole unit on C-6 of the neighboring glucose residue. Each one stereoselectively transaminates phenylpyruvic acid to produce phenylalanine, and with opposite stereochemical preferences. Structure determinations by X-ray crystallography and NMR spectroscopy indicate that the imidazole units serve to block proton addition from their side, rather than acting to protonate the transamination intermediates. Related cyclodextrin-pyridoxamine compounds had been reported carrying ethylenediamine units instead of imidazoles, and high enantioselectivities in transamination were claimed. Our work indicates that these claims are incorrect, and that only poor selectivities are seen that are often unrelated to the position of the ethylenediamine units. Neither of these transaminating systems yet approaches the enantioselectivity seen with a pyridoxamine carrying a chirally mounted internal base group. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(98)00193-X

Bioorganic & Medicinal Chemistry 7 (1999) 709±714
Reversal of Optical Induction in Transamination By Regioisomeric
Bifunctionalized Cyclodextrins*
y
{
x
Elisabetta Fasella, Steven D. Dong and Ronald Breslow
Department of Chemistry, Columbia University, New York, New York 10027, USA
Received 29 June 1998; accepted 6 August 1998
AbstractÐTwo isomeric compounds have been synthesized carrying a pyridoxamine on C-6 of b-cyclodextrin and an imidazole unit
on C-6 of the neighboring glucose residue. Each one stereoselectively transaminates phenylpyruvic acid to produce phenylalanine,
and with opposite stereochemical preferences. Structure determinations by X-ray crystallography and NMR spectroscopy indicate
that the imidazole units serve to block proton addition from their side, rather than acting to protonate the transamination inter-
mediates. Related cyclodextrin±pyridoxamine compounds had been reported carrying ethylenediamine units instead of imidazoles,
and high enantioselectivities in transamination were claimed. Our work indicates that these claims are incorrect, and that only poor
selectivities are seen that are often unrelated to the position of the ethylenediamine units. Neither of these transaminating systems
yet approaches the enantioselectivity seen with a pyridoxamine carrying a chirally mounted internal base group. # 1999 Elsevier
Science Ltd. All rights reserved.
Introduction
and cyclodextrin have been investigated in a continuing
e€ort to reproduce in simple model systems the selec-
tivity for substrate, reaction selectivity, and enantios-
electivity of enzymatic transaminations. The ®rst
cyclodextrin±pyridoxamine conjugate examined, com-
The ability of cyclodextrins to form inclusion complexes
with a variety of hydrophobic guests in aqueous solu-
tion and the relative ease of functionalization of their
primary and secondary hydroxyl groups makes them
7
pound 2, displayed a 5 to 1 preference for formation of
l-Phe over d-Phe in the transamination of phe-
nylpyruvic acid (cf. Scheme 1). In an e€ort to repro-
1
versatile enzymatic binding site mimics. Although the
chiral discriminating properties of cyclodextrins have
long been recognized2 and amply exploited in chro-
8
,3
duce the catalytic role of the lysine residue at the active
16
4
matographic separations of enantiomeric species, there
are only a few examples of optical induction by cyclo-
dextrins in reactions at prochiral centers: in the cyano-
site of transaminase enzymes, basic groups have been
incorporated in enzyme mimics. Attachment of a basic
0
side chain at the 5 -C of pyridoxamine in compound 3
9
5
hydration of benzaldehyde, in the reduction of
6
resulted in an increased l-/d-Phe ratio of 6.8. How-
ever, compound 4, incorporating N,N-dimethylglycine
in the linker between cyclodextrin and the cofactor,
7±12
ketones, and in the transamination of a-keto acids.
1
3,14
In a few studies
detailed structural and modelling
17
data have given insight into the mechanism of the
molecular recognition process, but no unifying picture
has emerged to account for the observed chiral dis-
formed l-/d-Phe in only a modest 1.4/1 ratio. These
examples illustrate the diculty in designing e€ective
enantioselective transaminase models.
1
5
criminating properties of cyclodextrins.
A remarkable increase in enantioselectivity (>95% ee)
has been reported using the bifunctional cyclodextrin
derivatives 5 and 6, incorporating a pyridoxamine moi-
In the past three decades, a number of transaminase
mimics combining the pyridoxamine cofactor 1 (Fig. 1)
ety and an ethylenediamine chain on adjacent glucose
1,12
residues.1
The structural information was insucient
Key words: Amino acids and derivatives; biomimetic reactions; cyclo-
dextrins; stereospeci®city.
to directly assign the regiochemistry of substitution (the
clockwise versus anticlockwise arrangement of the
ethylenediamine and of the pyridoxamine substituents
on adjacent glucose residues). However, the substitution
pattern was inferred from the reported stereochemistry
of the product amino acids assuming participation of
the ethylenediamine arm in the proton transfers needed
*
ordinary chemist and role model.
This paper is dedicated to the memory of Derek Barton, an extra-
y
Taken in part from the Ph.D. thesis of Elisabetta Fasella, Department
of Chemistry, Columbia University, 1997. Current address: Depart-
ment of Chemistry, University of Wisconsin, Madison Wisconsin.
{
Taken in part from the Ph.D. thesis of Steven D. Dong, Department
of Chemistry, Columbia University, 1998.
x
11
Corresponding author. Tel.: +1 212 854 2170; fax: +1 212 854 2755.
for tautomerization of intermediate A to B in Scheme 1.
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(98)00193-X

7
10
E. Fasella et al. / Bioorg. Med. Chem. 7 (1999) 709±714
In this paper we report on the reversal in optical induc-
tion in transaminations by the unambiguously char-
acterized regioisomeric bifunctionalized cyclodextrin
derivatives 9 and 10. The origin of the observed enan-
tioinduction can be related to the structural data. Also,
we have re-evaluated compounds 5 and 6. Their enan-
tioselectivities are modest at best, and often unrelated to
the position of the ethylenediamine unit.
separated by careful reverse phase chromatography.
The ®rst eluted isomer readily crystallized in water. X-
ray di€raction data (Fig. 2) allowed its identi®cation as
compound 7. NOESY experiments carried out on the
second eluted isomer revealed cross peaks between the
anomeric proton of the glucose ring carrying the iodo
substituent and the C-6 methylene hydrogens adjacent
to the imidazole substituent. Such an NOE is in agree-
ment with structure 8.
Reaction of the two separated regioisomers 7 and 8 with
5
Results and Discussion
0
presence of Cs CO a€orded transaminase mimics 9 and
9
-thiopyridoxamine dihydrobromide in DMF in the
The key precursors to transaminase mimics 9 and 10 are
the mono-imidazolyl-mono-iodo-b-cyclodextrin deriva-
tives 7 and 8 di€ering in the clockwise or anti-clockwise
arrangement of the C-6 substituents on adjacent glucose
residues. Compounds 7 and 8 were synthesized by reac-
2
3
10, respectively. Compounds 9 and 10 have been eval-
uated in transamination reactions of phenylpyruvic acid
and of indolepyruvic acid. After incubation, aliquots of
the reaction mixture were treated with an o-phthalalde-
hyde/N-boc-cysteine derivatizing mixture and the dia-
stereomeric isoindole derivatives analyzed by HPLC on
a C-18 reverse-phase silica gel column.18 Under the
variety of conditions indicated in Table 1, compound 9
preferentially a€orded l-Phe, and compound 10 pre-
ferentially yielded d-Phe. At most a 5.9/1 ratio of the
enantiomeric amino acids was observed. Similar results
were obtained in transamination of indolepyruvic acid
to a€ord tryptophan.
A
B
tion of 6 ,6 -diiodo-b-cyclodextrin with imidazole and
Our reinvestigation of the reactions of compound 5 and
1,12
6
under the previously reported1
K HPO , adjusted to pH 8) showed that both isomers
conditions (2 M
2
4
showed a preference for the formation of l- rather than
d-Phe, and with at most a 4/1 ratio, as illustrated in
Table 1. Furthermore, 5 and 6 both reversed their pre-
ference at pH 4 in water in favor of the d-enantiomer.
These results indicate that the position of the ethylene-
diamine substituent plays only a minor role in deter-
mining the chirality of the product amino acid,
outweighed by the bu€er, protonation state, and intrin-
sic chirality of the cyclodextrin cavity.
Figure 1. Pyridoxamine and pyridoxamine±cyclodextrin conjugates
(
and their precursors) evaluated as transaminase mimics. The lettering
of glucose residues is clockwise and sequential A,B when the cyclo-
dextrin is viewed from the primary face that carries the hydroxymethyl
groups.
Scheme 1. Transamination of a ketoacid by pyridoxamine. In the
2
compounds described in this paper, the CH OH group is replaced by a
2
CH -S-cyclodextrin group.

E. Fasella et al. / Bioorg. Med. Chem. 7 (1999) 709±714
711
Table 2. Rates of ketimine to aldimine conversion as measured by
monitoring the absorbance at 3845 nm. Transaminations were run in
50% MeOH/H
Zn(OAc)
2
O in the presence of 0.16 mM catalyst, 0.16 mM
ꢀ
2
and 1.6 mM phenylpyruvic acid at pH 7.3 and 30 C
obs (s 1)
�
Relative rate
Transaminase
k
�
�
� 4
4
5
1
2
4
1.0Â10
6.0Â10
4.4Â10
4.8Â10
1.7
1.0
7.3
8.0
a
�
4
9
aFrom ref 17.
Considering the modest enantioselectivity displayed by
and 10, and the modest rate enhancement with respect
9
to compound 2, it does not appear that the imidazole is
acting as an e€ective base catalyst for tautomerization
in transamination. This is so even though most of the
reactions were performed without external bu€ers, to
maximize the chance that proton transfers would be
performed by the imidazole rings. They may be cata-
lyzing the formation or hydrolysis of the Schi€ base
intermediates, but cannot be catalyzing the proton
transfers involved in the conversion of intermediate A to
B (Scheme 1). Examination of CPK-molecular models
indicates clearly that the observed preference of 9 for l-
Phe formation and of 10 for d-Phe formation shows
that the proton is being added to the new amino acid on
the face of B away from the imidazole group. The
observed stereochemical preference probably involves
steric blocking of one face of the imine intermediate by
the imidazole ring so that protonation by solvent occurs
on the other face. Similarly, 4 displayed an optical
induction opposite to that predicted if the side arm were
mediating the proton transfer.18
Figure 2. The crystal structure of compound 7. The iodine is on the
A position according to the convention and structure shown in
Figure 1.
6
To ensure that no undesired racemization occurred in
our assay, control reactions were analyzed in which the
transamination mixtures were doped with known
amounts of l-Phe. The ratio of l-/d-Phe observed after
the standard incubation time re¯ected the addition of l-
Phe in the transamination mixture with no evidence of
racemization. The extraordinary results reported in refs
11 and 12, which we do not con®rm, may have been
caused by analytical artifacts.
Kinetic data on transamination by 1, 2, 4, and 9 are
reported in Table 2. They were obtained by monitoring
the ketimine to aldimine conversion spectro-
photometrically using the method developed by Mat-
High enantioselectivity in transamination, as much as
92% e.e., has been achieved previously by a rigid pyr-
idoxamine 11 carrying a chirally mounted base group.
That base indeed transfers protons on a single face of
intermediates like A, but the compound lacks a cyclo-
dextrin binding group.
2
1
19
sushima and Martell and modi®ed by Zimmerman et
2
0
al. Compound 9, incorporating an imidazole group,
displays an eightfold rate acceleration compared with
compound 2 in the transamination of phenylpyruvic
acid. This modest rate acceleration is comparable to
that observed for transaminase mimic 4 incorporating
an N,N-dimethylglycine moiety.17
Table 1. l-/d-ratio of amino acid products in transamination by 5, 6,
ꢀ
a,b
9
, and 10 after 4 h incubation at 25 C in water
Transaminase,
Amino Acid
K
2
HPO
4
pH9 pH8 pH7 pH6 pH5 pH4
2 M pH8
9
1
9
1
5
6
,Phe
0,Phe
,Trp
0,Trp
,Phe
,Phe
3.9/1
1/1.8
1.8/1
1/2
3.7/1
1.4/1
2.1/1 2.9/1 3.5/1 4.3/1
1/5.9 1/1.8 1/3.6
2.9/1 3.4/1 3.6/1 4/1
5/1
1/5
Conclusion
1/2
1/4
Our transamination studies with compounds 5, 6, 9, and
10 show that the enantioselectivity is highly dependent
on the pattern of substitution on the cyclodextrin rim,
but the absolute optical induction is still modest and
previous reports of high optical induction by bifunctio-
nalized cyclodextrins cannot be con®rmed. Considering
what has been achieved with compound 11, it should be
2.1/1
1/2
1/4.8
1/1.2
a
The absolute regiochemistry of 5 and 6 has not been determined.
The pH of the 2 M K HPO was adjusted to 8 with H PO . The other
2 4 3 4
b
solutions were unbu€ered except by the reagents. Their pHs were
adjusted with NaOH or HCl. Measurements indicated that the pH did
not change appreciably during the runs.

7
12
E. Fasella et al. / Bioorg. Med. Chem. 7 (1999) 709±714
possible to create full mimics of transaminase enzymes
with high enantioselectivity.
1H, C-5 imidazole), 7.09 (s, 1H, C-4 imidazole), 7.59 (s,
1H, C-2 imidazole); C NMR (75.4 MHz, DMSO) d
1
3
0
0
glucose ring), 58.90, 59.81, 59.91, 59.99, 69.78, 70.02,
9
.16 (C-6 iodo±glucose ring), 47.49 (C-6 imidazolyl±
7
8
1.95, 72.08, 72.19, 72.38, 72.93, 73.06, 73.26, 80.48,
1.09, 81.44, 81.86, 83.84, 86.11, 100.77, 101.75, 102.01,
Experimental
All reactions were performed in oven-dried glassware
under an Ar atmosphere.
102.19, 102.38, 120.40, 127.85, 137.83. MS(FAB): 1295
(M+1).
A
B
B
A
B
0
6
6
6
-Monoimidazolyl-6 -monoiodo-ꢀ-cyclodextrin (7) and
A
6 -Monoimidazolyl-6 -(5 -thiopyridoxaminyl)-ꢀ-cyclo-
dextrin (9). Anhydrous DMF (10 mL) was added under
A
-Monoimidazolyl-6 -monoiodo-ꢀ-cyclodextrin (8). 6 ,
10
B
A
B
-Diiodo-b-cyclodextrin (2.8 g, 2.1 mmol) and imida-
Ar to a ¯ask containing 6 -monoimidazolyl-6 -mono-
iodo-b-cyclodextrin 7 (400 mg, 0.31 mmol), 5-thiopyr-
idoxamine dihydrobromide9 (315 mg, 0.91 mmol) and
Cs CO (900 mg, 2.7 mmol). The solution was stirred at
zole (1.4 g, 20.8 mmol) were stirred in 30 mL anhydrous
ꢀ
DMF under Ar at 85 C for 7 h. The solvent was
removed under reduced pressure and the residue sus-
pended in 50 mL acetone. The precipitate was recovered
by ®ltration, dissolved in a minimum amount of water
2
ꢀ
3
40 C for 3 h, and the solvent was removed under
reduced pressure. The residue was dissolved in a mini-
mum amount of water and the pH adjusted to 5 with
2 N HCl. The yellow mixture was loaded on a C-18
reverse-phase column (bed volume: 1.5Â13 cm) and
22
and loaded onto a C-18 reverse-phase silica gel col-
umn (5Â13 cm). Elution was performed with the fol-
lowing MeOH/H O gradient: 400 mL each of 0, 10, 15
2
and 20% MeOH/H O followed by varying amounts of
2
eluted with a 0 to 50% MeOH/H O gradient. The pro-
2
solvent (250±1200 mL) in 1% increments from 21 to
duct was eluted in the 30% MeOH fractions. Con-
centration and lyophilization of the 30% MeOH
fractions a€orded 9 (R 0.14, 180 mg, 0.13 mmol, 42%).
3
4
0% MeOH, and then in 2% increments from 32 to
0%, and ®nally in 5% increments from 40 to 60%
f
A
B
1
MeOH. 6 ,6 -Bisimidazolyl-b-cyclodextrin (R 0.1 on
H NMR (400 MHz, D O) d 2.17 (s, 3H), 3.2±3.9 (m,
f
2
silica TLC plates, eluent: i-PrOH/EtOAc/H O/NH OH,
7/7/3.5/3) was eluted in the 21±24% MeOH fractions,
followed by 7 (R 0.22) in the 27% MeOH fractions, and
40H), 4.0±4.1 (m, 4H), 4.34 (d, J=14, 1H), 4.8±5.0 (m,
7H), 6.74 (s, 1H), 6.97 (s, 1H), 7.27 (s, 1H), 7.48 (s, 1H);
2
4
13
C NMR (75.4 MHz, DMSO-d ) d 18.5, 31.4, 32.1,
f
6
8
in the 28±38% MeOH fractions (R 0.20). Finally
f
46.9, 59.9, 70±73, 81.3±83.6, 101.9±102.2, 120.2, 127.9,
128.8, 138.0, 138.8, 145.3, 152.7, 153.8. MS(FAB): 1351
(M+1).
unreacted starting material was eluted in the 45±60%
MeOH fractions. Concentration and lyophilization of
the appropriate fractions a€orded 7 (0.81 g, 0.62 mmol,
3
0%) and 8 (0.50 g, 0.38 mmol, 18.5%).
Compound 10 was prepared in an analogous manner by
reaction of 8 with 5-thiopyridoxamine dihydrobromide.
1
1
Compound 7: H NMR (400 MHz, D O) d 2.50 (m, 1H,
10: H NMR (400 MHz, D O) d 2.09 (s, 3H), 2.6±2.8
2
2
0
iodo±glucose ring), 3.08 (m, 1H, C-5 iodo±glucose
0
C-6 iodo-glucose ring), 2.84 (d, J=10.98, 1H, C-6
(m, 2H), 2.9±3.1 (m, 2H), 3.2±4.1 (m, 38H), 4.3±4.4 (m,
2H), 4.8±5.0 (m, 7H), 6.79 (s, 1H), 6.91 (s, 1H), 7.19 (s,
0
ring), 3.28 (t, J=9.43, 1H, C-4 imidazolyl±glucose
0
ring), 3.31 (t, J=9.89, 1H, C-4 iodo±glucose ring), 3.48
13
1H), 7.47 (s, 1H); C NMR (75.4 MHz, DMSO-d ) d
6
0
m, 12 H), 3.72 (m, 23H), 4.07 (dd, J=23.8, J=5.78,
18.5, 31.4, 33.2, 36.5, 47.1, 58.9, 69.8±73.1, 80.5±81.7,
83.9, 85.0, 100.8±102.2, 120.3, 127.4, 127.7, 128.8, 137.7,
138.6. MS(FAB): 1424 (M+74), 1351 (M+1).
(
0
imidazolyl±glucose ring), 4.83 (d, J=3.66, 1H, C-1 ,
1
6
H, C-6 imidazolyl±glucose ring), 4.48 (d, 14.28, 1H, C-
0
iodo±glucose ring), 4.91 (m, 4H, anomeric), 4.96 (d,
0
Transamination reactions and HPLC analysis
0
dazolyl±glucose ring), 6.91 (s, 1H, C-5 imidazole), 7.05
J=3.66, 1H, anomeric), 5.02 (d, J=3.66, 1H, C-1 imi-
In a typical experiment 50 mL of a 2 mM solution of
catalyst in the appropriate medium (pH, solvent) was
vortexed in a 100 mL vial with 50 mL of a 40 mM solu-
tion of phenylpyruvic acid in the appropriate medium
13
(
s, 1H, C-4 imidazole), 7.57 (s, 1H, C-2 imidazole);
C
0
ring), 47.06 (C-6 imidazolyl±glucose ring), 59.84, 59.98,
NMR (75.4 MHz, DMSO) d 11.02 (C-6 iodo±glucose
0
ꢀ
6
7
1
0.55, 67.98, 70.95, 71.78, 72.04, 72.38, 72.84, 72.97,
3.25, 81.33, 81.42, 81.67, 82.46, 83.73, 85.69, 101.70,
01.82, 102.00, 102.25, 120.29, 128.18, 137.89.
(pH, solvent). The solution was then incubated at 25 C.
Except in the case of the reaction with phosphate bu€er,
no external bu€er was added and the solutions were
simply titrated to the correct pH with NaOH or HCl.
The pH was observed not to change appreciably during
the reaction.
MS(FAB): 1295 (M+1). The crystal structure is shown
in Figure 2.
Compound 8: 1H NMR (400 MHz, D O) d 2.89 (d,
J=9.89, 1H), 3.15 (d, J=12.37, 1H, C-4 imidazolyl±
2
0
glucose ring), 3.38 (m, 15H), 3.57 (m, 1H, C-6 iodo±
At various time points 10 mL samples were removed,
derivatized in the injector loop with 0.7 mL of a solution
of 5.2 mg o-phthalaldehyde and 8.8 mg N-boc-cysteine
in 100 mL MeOH and injected on a C-18 Rainin Micro-
0
glucose ring), 3.74 (m, 21H), 3.95 (t, J=9.23, 1H, C-5
0
imidazolyl±glucose ring), 4.24 (dd, J=23.35, J=5.94,
0
H, C-6 imidazolyl±glucose ring), 4.85 (d, J=3.30, 1H,
2
1
1
H, C-6 imidazolyl±glucose ring), 4.42 (d, J=14.73,
0
sorb column (4.6 mmÂ25 cm, particle size 5 mm). Elu-
tion was performed at a ¯ow rate of 0.35 mL/min with
an isocratic mixture of 65% MeOH and 35% citrate
bu€er (citrate bu€er composition: 9 parts of a pH 4
0
.99 (d, J=3.37, 1H, C-1 iodo±glucose ring), 6.87 (s,
C-1 imidazolyl±glucose ring), 4.90 (m, 5H, anomeric),
0
4

E. Fasella et al. / Bioorg. Med. Chem. 7 (1999) 709±714
713
ꢀ
solution 20 mM in sodium hydrogen citrate sesqui-
hydrate and 6 mM in NaH PO and 1 part MeOH).
degassed, anhydrous ethylenediamine at 60 C for
90 min. Excess ethylenediamine was removed under
reduced pressure and 15 mL EtOH added. The suspen-
2
4
Detection was performed measuring the absorbance at
44 nm with a UV/vis diode array detector. The reten-
ꢀ
3
sion was cooled overnight at 5 C and the precipitate
tion times for the l- and d-Phe isoindole derivatives
were respectively 19.9 and 23.9 min.
recovered by ®ltration. The precipitate was dissolved in
a minimum amount of water and the pH adjusted to 6
with 1 N HCl. The reaction mixture was ®rst puri®ed by
ion exchange chromatography on a Sephadex CM-25
column (1.8Â19 cm). The following NH HCO gradient
Synthesis of compounds 5 and 6
4
3
In our reinvestigation, compounds 5 and 6 were pre-
pared by a three step procedure involving: (1) reaction
was used: 150 mL H O (starting material was eluted in
2
these fractions), 100 mL 0.005 M NH HCO , 100 mL
4
3
A
B
0
0
of 6 ,6 -diiodo-b-cyclodextrin with 4 -N-boc-5 -thioa-
cetyl-pyridoxamine, (2) reaction of the resulting mixture
of monoiodo-monopyridoxaminyl derivatives with
ethylenediamine and chromatographic separation of the
two regioisomers, and (3) deprotection of the two iso-
lated regioisomers. The two compounds had NMR and
MS(FAB) in agreement with those reported for com-
0.01 M NH HCO , 200 mL 0.015 M NH HCO , 100 mL
4 3 4 3
0.02 M NH HCO , 100 mL 0.022 M NH HCO , 100 mL
4
3
4
3
0.03 M NH HCO , 100 mL 0.04 M NH HCO , 100 mL
4
4
3
3
0.052 M NH HCO , 100 mL 0.084 M NH HCO . A
4
3
4
3
mixture of the two regioisomers was eluted in the 0.02±
0.022 M NH HCO fractions (R 0.04 and 0.08 on silica
4
3
f
TLC plates, eluent: i-PrOH/EtOAc/H O/NH OH, 5/5/
2
4
1
1,12
pounds 5 and 6.
3.5/3; R 0.73 for the starting material under these con-
f
ditions). These fractions were concentrated and lyophi-
lized. The residue was redissolved in a minimum
amount of water and the pH adjusted to 5 with 1 N
HCl. The mixture of bifunctionalized cyclodextrins was
loaded on a reverse-phase silica gel column (1.5Â6 cm).
The two regioisomers were separated using the follow-
0
0
-N-Boc-5 -Thioacetyl-pyridoxamine. A solution of
PPh (0.64 g, 2.4 mmol) in 60 mL anhydrous THF was
4
3
cooled in an ice bath. Diisopropylazodicarboxylate
(
0.48 g, 2.4 mmol) was added under Ar. After 30 min a
solution of thioacetic acid (170 mL, 2.2 mmol) and 4 -N-
0
boc-pyridoxamine (0.33 g, 1.2 mmol) in 25 mL THF
was added via syringe. The solution was stirred at 0 C
2
4
ing gradient: 100 mL H O, 100 mL 4% MeOH, 100 mL
2
ꢀ
10% MeOH, 100 mL 20% MeOH, 100 mL 24% MeOH,
100 mL 28% MeOH, 300 mL 30% MeOH, 200 mL 35%
MeOH, 200 mL 38% MeOH, 200 mL 40% MeOH. The
®rst regioisomer was eluted in the 24±28% MeOH frac-
tions, the second regioisomer was eluted with 30%
MeOH. A few mixed fractions were collected in
between. Concentration and lyophilization of the
appropriate fractions a€orded the separated N-boc-
protected derivatives of 5 and 6 in 22% (24 mg,
0.017 mm) and 13% (15 mg, 0.01 mm) yield.
for 1 h and then at room temperature for 1 h. The sol-
vent was removed under reduced pressure, the residue
was dissolved in Et O, and the Ph PO precipitate was
2
3
removed by ®ltration. The solution was concentrated
and the residue puri®ed by silica gel chromatography
0 0
EtOAc/CH Cl , 3/7 eluent) to a€ord 4 -N-boc-5 -
2 2
(
thioacetylpyridoxamine (188 mg, 0.58 mmol, 48%, R
1
f
0
.42, eluent: EtOAc/CH Cl , 1/3). H NMR (400 MHz,
2 2
CDCl ) d 1.47 (s, 8H), 2.4 (s, 3H), 2.49 (s, 3H), 4.1±4.2
3
(
(
m, 2H), 4.27 (d, J=6.7 Hz, 2H), 5.5±5.6 (br t, 1H), 7.95
s, 1H), 9.53 (s, 1H). MS (CI, NH ): 32 7 (M+1).
First eluted isomer (the absolute regiochemistry of sub-
1
3
stitution has not been determined): H NMR (300 MHz,
D O) d 1.47 (s, 8H), 2.29 (s, 3H), 2.74±2.77 (m, 3H),
2.95±3.1 (m, 3H), 3.3±4.0 (m, 36H), 4.97±4.99 (m, 7H),
7.60 (s, 1H). FAB(MS): 1443 (M+1).
6A-Monoiodo-6 -[4 -N-boc-5 -thiopyridoxaminyl]-ꢀ-cyclo-
B
0
0
2
B
A
0
aminyl]-ꢀ-cyclodextrin. To a ¯ask containing 6 ,6 -
0
dextrin and 6 -monoiodo-6 -[4 -N-boc-5 -thiopyridox-
B
A
0
0
diiodo-b-cyclodextrin (1.7 g, 1.25 mmol), 4 -N-boc-5 -
thioacetyl-pyridoxamine (0.44 g, 1.25 mmol) and
Cs CO (1.5 g, 4.6 mmol) were added 150 mL anhydrous
DMF under Ar. The mixture was stirred at room tem-
perature for 10 h, and the solvent was removed under
reduced pressure. The residue was dissolved in a mini-
mum amount of water, acidi®ed to pH 6 with 1 N HCl
and loaded on 140 g of C-18 reverse-phase silica gel. The
product (R 0.23 on reverse-phase silica plates, eluent
f
MeOH/H O, 1/1) was eluted by a MeOH/H O gradient.
2
Concentration and lyophilization of the 28% MeOH
fractions a€orded the products as a mixture of regio-
isomers (0.51 g, 0.29 mmol, 27% yield). H NMR
1
Second eluted isomer: H NMR (300 MHz, D O) d 1.46
2
(s, 8H), 2.28 (s, 3H), 2.5±3.0 (m, 6H), 3.3±4.0 (m, 40H),
4±4.3 (m, 3H), 5±5.1 (m, 7H), 7.75 (s, 1H). FAB(MS):
1443 (M+1), 1515 (M+73).
2
3
A
B
0
6 -Monoiodo-6 -[5 -thiopyridoxaminyl]-ꢀ-cyclodextrin
B
A
0
and 6 -Monoiodo-6 -[5 -thiopyridoxaminyl]-ꢀ-cyclo-
dextrin, 5 and 6. The ®rst eluted isomer mentioned
above (24 mg, 0.017 mm) was stirred in 1 mL of tri-
¯uoroacetic acid at room temperature for 45 min. Ethyl
ether (8 mL) was added and the suspension cooled on
ice. The precipitate was recovered by ®ltration, dis-
solved in a minimum amount of water and loaded on a
reverse phase silica gel column (1.5Â8 cm). The product
2
1
(
(
(
400 MHz, DMSO-d ) d 1.4 (s, 9H), 2.3 (s, 3H), 3.1±4.2
m, 40H), 4.25±4.6 (m, 6H), 4.8±4.9 (m, 7H), 5.5±5.9
br, 14H), 7.79±7.81 (br, 1H). MS(FAB): 1511 (M+1).
6
was eluted by a MeOH/H O gradient. Concentration
2
and lyophilization of the 20% MeOH fractions a€orded
5 (6 mg, 0.0044 mm, 26% yield).
A
N-Boc protected 5 and 6. A mixture of 6 -monoiodo-
B
0 0
-[4 -N-boc-5 -thiopyridoxaminyl]-b-cyclodextrin and
A
6
6
B
0
0
1
-monoiodo-6 -[4 -N-boc-5 -thiopyridoxaminyl]-b-cyclo-
5: H NMR (300 MHz, D O) d 2.46 (s, 3H), 2.61±2.70
2
dextrin (110 mg, 0.078 mmol) was stirred in 1.5 mL
(m, 1H), 3.04 (d, J=12.3 Hz, 1H), 3.2±4.1 (m, 45H), 4.3

7
14
E. Fasella et al. / Bioorg. Med. Chem. 7 (1999) 709±714
(
(
s, 2H), 5±5.1 (m, 8H), 7.72 (s, 1H). FAB(MS): 1343
M+1) and 1415 (M+73).
7. Breslow, R.; Hammond, M.; Lauer, M. J. Am. Chem. Soc.
1980, 102, 421.
8. Breslow, R.; Czarnik, A. W.; Lauer, M.; Leppkes, R.;
Winkler, J.; Zimmerman, S. J. Am. Chem. Soc. 1986, 108,
1969.
6
fashion from deprotection of the second isomer:
NMR (300 MHz, D O) d 2.46 (s, 3H), 2.63 (dd,
J=15.6 Hz, J=6 Hz, 1H), 2.83 (d, J=13.8 Hz, 1H), 3.2±
4.1 (m, 55H), 4.3 (s, 2H), 4.9±5.1 (m, 8H), 7.70 (s, 1H).
FAB(MS): 1343 (M+1) and 1415 (M+73).
: (3 mg, 0.0022 mm, 22%) was obtained in a similar
1
H
9. Breslow, R.; Chmielewski, J.; Foley, D.; Johnson, B.;
Kumabe, N.; Varney, M.; Mehra, R. Tetrahedron 1988, 44,
2
5
1
515.
0. Breslow, R.; Canary, J. W.; Varney, M.; Waddell, S. T.;
Yang, D. J. Am. Chem. Soc. 1990, 112, 5212.
1. (a) Tabushi, I.; Kuroda, Y.; Yamada, M.; Higashimura,
1
H.; Breslow, R. J. Am. Chem. Soc. 1985, 107, 5545; (b)
Tabushi, I. Pure Appl. Chem. 1986, 58, 1529. R. Breslow
was included as an author in reference 11(a) as a courtesy,
since he had suggested the general approach used in this
work.
Acknowledgements
We thank Dr J. Decatur, Columbia University, for
assistance with the NMR experiments and Professor G.
Parkin, Columbia University, for the crystal structure
analysis. We acknowledge the support of the Kanagawa
Academy of Science and Technology, the NIH, and the
NSF, and a Merck graduate fellowship to S.D.D. We
are grateful to U. Mayer, Columbia University, for
helpful discussions and to Dr M. Yamada for kindly
providing a copy of his Ph.D. thesis.
1
1
1
1
1
2. Yamada, M. Ph.D. Thesis, Kyoto University (Japan),
985.
3. (a) Trainor, G. L.; Breslow, R. J. Am. Chem. Soc. 1981,
03, 154; (b) Breslow, R.; Trainor, G.; Ueno, A. ibid. 1983,
05, 2739; (c) Da€e, V.; Fastrez, J. J. Chem. Soc. Perkin
Trans. II 1983, 789.
4. (a) Brown, S. E.; Coates, J. H.; Duckworth, P. A.; Lin-
1
coln, S. F.; Easton, C. J.; May, B. L. J. Chem. Soc. Faraday
Trans. 1993, 89, 1035; (b) Lipkowitz, K. B.; Raghothama, S.;
Yang, J. J. Am. Chem. Soc. 1992, 114, 1554.
15. (a) Kano, K. J. Phys. Org. Chem. 1997, 10, 286. (b)
Topiol, S.; Sabio, M. Enantiomer 1996, 1, 251.
16. John, R. A. Biochim. Biophys. Acta 1995, 1248, 81.
17. Zimmerman, S. C. Ph.D. Thesis, Columbia University
(USA), 1984.
References
1
. (a) Tabushi, I. Acc. Chem. Res. 1982, 15, 66; (b) Breslow,
R. Acc. Chem. Res. 1995, 28, 146; (c) Komiyama, M. In
Comprehensive Supramolecular Chemistry; Atwood, J. L.;
Davies, J. E. D.; MacNicol, D. D.; Vogtle, F.; Suslick, K. S.,
Eds.; Pergamon: Oxford, 1996: Vol. 3, pp 401±422.
2
3
4
. Cramer, F.; Dietsche, W. Chem. Ber. 1959, 92, 378.
. Easton, C. J.; Lincoln, S. F. Chem. Soc. Rev. 1996, 25, 163.
. Snopek, J.; Smolkova-Keulemansova, E.; Cserhati, T.;
18. Kuang, H.; Brown, M. L.; Davies, R. R.; Young, E. C.;
Distefano, M. D. J. Am. Chem. Soc. 1996, 118, 10702.
19. Matsushima, Y.; Martell, A. E. J. Am. Chem. Soc. 1967,
89, 1331.
20. Zimmerman, S. C.; Czarnik, A. W.; Breslow, R. J. Am.
Chem. Soc. 1983, 105, 1694.
Gahm, K. H.; Stalcup, A. M. In Comprehensive Supramole-
cular Chemistry; Atwood, J. L.; Davies, J. E. D.; MacNicol,
D. D.; Vogtle, F.; Suslick, K. S., Eds.; Pergamon: Oxford,
1
5
996: Vol. 3, pp 515±572.
. (a) Cramer, F.; Dietsche, W. Chem. Ber. 1959, 92, 1739; (b)
21. Zimmerman, S. C.; Breslow, R. J. Am. Chem. Soc. 1984,
106, 1490.
Gountzos, H.; Jackson, W. R.; Harrington, K. J. Aust. J.
Chem. 1986, 39, 1135.
È
22. Kuhler, T. C.; Lindsten, G. R. J. Org. Chem. 1983, 48,
3589.
6
. (a) Fornasier, R.; Reniero, F.; Scrimin, P.; Tonellato, U. J.
23. (a) Chmielewski, J. A. Ph.D. Thesis, Columbia University
(USA), 1988; (b) Kumabe, N. Ph.D. Thesis, Columbia Uni-
versity (USA), 1990.
Org. Chem. 1985, 50, 3209; (b) Sakuraba, H.; Inomata, N.;
Tanaka,Y. J. Org. Chem. 1989, 54, 3482.