The 3-(3-pyridine)propionyl anchor group for protease-catalyzed resolutions: p-toluenesulfinamide and sterically hindered secondary alcohols

Article abstract of DOI:10.1002/adsc.200606040

Compared to an acetyl acyl group, the 3-(3-pyridine)propionyl group increases substrate binding to many proteases and substrate solubility in water, thereby increasing the rates of protease-catalyzed reactions. For example, proteases reacted up to six hundred-fold faster with the 3-(3-pyridine)propionyl ester of 1-phenylethanol than with the corresponding acetate ester. In addition, the 3-(3-pyridine)propionyl group enables a simple, mild acid extraction to separate the remaining starting material and product. To demonstrate the synthetic usefulness of this strategy, we resolved multi-gram quantities of (R)- and (S)-p-toluenesulfinamide with α-chymotrypsin and gram quantities of (R)- and (S)-2,2-dimethylcyclopentanol with subtilisin Carlsberg. The 3-(3-pyridyl)propionyl group was better for these resolutions than the corresponding acetate or dihydrocinnamate because it decreased the reaction time due to increased reactivity, decreased the reaction volume due to increased substrate solubility and enabled purification without chromatography. Molecular modeling suggests the enantioselectivity of α-chymotrypsin toward (R)-p-toluenesulfinamide is high (E = 52) because of a favorable hydrophobic interaction between the p-tolyl group of the fast-reacting (R)-enantiomer and leaving group pocket. The enantioselectivity of subtilisin Carlsberg toward (S)-2,2-dimethylcyclopentanol is high (E = 43) because the large substituent (the 2,2-dimethyl quaternary carbon) of the slow-reacting (R)-enantiomer cannot fit in the S₁′ leaving group pocket.

Full text of DOI:10.1002/adsc.200606040

FULL PAPERS
DOI: 10.1002/adsc.200606040
The 3-(3-Pyridine)propionyl Anchor Group for Protease-Cata-
lyzed Resolutions: p-Toluenesulfinamide and Sterically Hindered
Secondary Alcohols
Christopher K. Savilea, b and Romas J. Kazlauskasa,*
a
Department of Biochemistry, Molecular Biology & Biophysics and the Biotechnology Institute, University of Minnesota,
1479 Gortner Avenue, Saint Paul, Minnesota 55108, USA
Fax : (+1)-612-625-5780; e-mail: rjk@umn.edu
Department of Chemistry, McGill University, 801 Sherbrooke Street West, MontrØal, QuØbec, Canada, H3A 2K6
b
Received: February 3, 2006; Accepted: March 30, 2006
Abstract: Compared to an acetyl acyl group, the 3- because it decreased the reaction time due to in-
(3-pyridine)propionyl group increases substrate bind- creased reactivity, decreased the reaction volume
ing to many proteases and substrate solubility in due to increased substrate solubility and enabled pu-
water, thereby increasing the rates of protease-cata- rification without chromatography. Molecular model-
lyzed reactions. For example, proteases reacted up to ing suggests the enantioselectivity of a-chymotrypsin
six hundred-fold faster with the 3-(3-pyridine)pro- toward (R)-p-toluenesulfinamide is high (E=52) be-
pionyl ester of 1-phenylethanol than with the corre- cause of a favorable hydrophobic interaction be-
sponding acetate ester. In addition, the 3-(3-pyridi- tween the p-tolyl group of the fast-reacting (R)-enan-
ne)propionyl group enables a simple, mild acid ex- tiomer and leaving group pocket. The enantioselec-
traction to separate the remaining starting material tivity of subtilisin Carlsberg toward (S)-2,2-dimethyl-
and product. To demonstrate the synthetic usefulness cyclopentanol is high (E=43) because the large sub-
of this strategy, we resolved multi-gram quantities of stituent (the 2,2-dimethyl quaternary carbon) of the
(R)- and (S)-p-toluenesulfinamide with a-chymotryp- slow-reacting (R)-enantiomer cannot fit in the S1’
sin and gram quantities of (R)- and (S)-2,2-dimethyl- leaving group pocket.
cyclopentanol with subtilisin Carlsberg. The 3-(3-pyr-
idyl)propionyl group was better for these resolutions Keywords: biocatalysis; chiral auxiliary; enantiose-
than the corresponding acetate or dihydrocinnamate lective; protease; resolution
Introduction
strates.[10,11] For example, subtilisins and chymotrypsin
favor esters and amides of amino acids containing hy-
Lipases and proteases are both useful enantioselective drophobic side chains such as phenylalanine
catalysts for resolutions,[1–3] but their preferred sub- esters.[10–12] The dihydrocinnamoyl group mimics phe-
strates differ significantly. Lipases favor insoluble sub- nylalanine so dihydrocinnamoyl esters are often good
strates and bind them in deep hydrophobic pockets[4] substrates for proteases. a-Chymotrypsin binds dihy-
in a folded conformation.[5] Proteases, on the other drocinnamoyl esters as tightly as N-acetylphenylala-
hand, require water-soluble substrates and bind them nine esters and they react at least one hundred-fold
in a shallow active site in an extended conforma- faster than N-acetylglycine esters.[10] We previously
tion.[6] For example, during resolution of esters of sec- used the dihydrocinnamoyl group to extend subtilisin
ondary alcohols, lipases bind both substituents in hy- E to a new class of substrates, N-acyl arylsulfinamides,
drophobic pockets within the active site, but proteases to make enantiopure sulfinamides.[7b] Subtilisin E did
bind only one substituent leaving the other substitu- not catalyze the hydrolysis of N-acetyl arylsulfin-
ent in the solvent. The shallow active site allows pro- amides, but it did catalyze the hydrolysis of N-dihy-
teases to accept more sterically hindered sub- drocinnamoyl arylsulfinamides. In this paper, we in-
strates[7,8] than lipases and to accept polar substrates troduce another analogue of phenylalanine – the 3-(3-
since one substituent can remain in water.[9]
pyridine)propionyl group. This group not only an-
While the hydrophobic nature of insoluble sub- chors the substrate in the active site of the protease,
strates binds them to the active site of lipases, pro- but also increases the water solubility of the ester as
teases contain a specificity pocket to bind sub- compared to the dihydrocinnamoyl group and, be-
Adv. Synth. Catal. 2006, 348, 1183 – 1192
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1183

Christopher K. Savile and Romas J. Kazlauskas
FULL PAPERS
cause it contains a basic nitrogen atom, enables sepa- tomyces griseus protease-catalyzed hydrolyses pro-
ration of product and remaining starting material by ceeded 10- to 23-fold faster, subtilisin Carlsberg react-
mild acid extraction. We report efficient resolutions ed 40-fold faster, while a-chymotrypsin from bovine
of
N-3-(3-pyridine)propionyl-p-toluenesulfinamide pancreas reacted only 3-fold faster. Although dihy-
(1a) and two sterically hindered secondary alcohol drocinnamate 2c mimics phenylalanine, five of the six
esters, 2,2-dimethylcyclopentyl 3-(3-pyridine)propio- proteases reacted no faster with 2c than with acetate
nate (3a) and 1-(2-mesityl)ethyl 3-(3-pyridine)propio- 2b, likely because of the lower water solubility of 2c
nate (4a).
as compared to acetate 2b. The exception was a-chy-
motrypsin, which reacted 200-fold faster with dihydro-
cinnamate 2c than with acetate 2b. This exception is
consistent with the strong preference of a-chymotryp-
sin for phenylalanine as acyl group.[13]
Results and Discussion
Carbodiimide-promoted coupling of 1-phenethyl alco-
A different leaving group – p-toluenesulfinamide –
hol with 3-(3-pyridine)propionic acid yielded 1-phen- showed a similar trend. Both N-3-(3-pyridine)pro-
ethyl 3-(3-pyridine)propionate (2a). Compound 2a pionyl-p-toluenesulfinamide (1a) and N-dihydrocinna-
was stable in buffered solutions (pH 7.2) for>5 days, moyl-p-toluenesulfinamide (1c) were good substrates
like the corresponding acetate 2b. 1-Phenethyl 3-(3- for a-chymotrypsin and subtilisin BPN’, while N-
pyridine)propionate (2a) reacted with six commonly acetyl-p-toluenesulfinamide (1b) did not react
used proteases up to 600-fold faster than the acetate (Table 2).[14] a-Chymotrypsin and subtilisin BPN’
2b (Table 1). Subtilisin BL (EC 3.4.21.62), Aspergillus gave 21–42% conversion for 1a and 19–39% conver-
melleus protease (EC 3.4.21.63), subtilisin Carlsberg sion for 1c after 3 h, while acetate 1b showed no reac-
(EC 3.4.21.62) and Streptomyces griseus (EC tion after 24 h. In this case, both 1a and 1c showed
3.4.21.80) protease-catalyzed hydrolyses proceeded similar activity in contrast to 2a and 2c above where
15- to 70-fold faster, a-chymotrypsin (EC 3.4.21.21) the 3-(3-pyridine)propionate 2a reacted 3 to 40-fold
from bovine pancreas reacted 600-fold faster, while faster. A possible reason for this similarity is that
Aspergillus oryzae (EC 3.4.21.63) reacted at the same compound 1c already dissolves in water at a concen-
rate as acetate 2b.
tration near its KM value (solubility ca. 5 mMin 10%
The 3-(3-pyridine)propionate 2a also reacted 3- to DMF) and therefore the higher solubility of 1a (solu-
40-fold faster than the dihydrocinnamate 2c for these bility ca. 50 mMin 10% DMF) does not significantly
six proteases (Table 1), likely due to the increased sol- increase the fraction of protease containing bound
ubility of 2a over 2c. Subtilisin BL, Aspergillus mel-
leus protease, Aspergillus oryzae protease and Strep-
Table 2. Reactivity and enantioselectivity of a-chymotrypsin
and subtilisin BPN’ towards 1a, 1b and 1c.
Table 1. Specific activity of proteases toward secondary alco-
hol esters 2a, 2b and 2c.
Protease
Prot. conc.
[mg/mL][a]
Specific acti-
vity[b]
Protease
1a
1b
1c
%c
%c[a] E[b] %c
E
E
2a 2b 2c
a-chymotrypsin[c] 42[e]
52
125 n.r.
n.r.[f] n.d.[g] 39[e] 56
n.d. 19 75
Bacillus lentus subtili-
sin[c]
2.4
10 0.6 0.7
subtilisin BPN’[d]
21
[a]
% Conversion: amount of sulfinamide formed after 24 h
except where noted.
Aspergillus melleus[c]
Aspergillus oryzae[c]
subtilisin Carlsberg[d]
a-chymotrypsin[d]
Streptomyces griseus[d]
4.4
5.1
9.0
8.6
6.6
4.6 0.3 0.3
1.0 1.0 0.1
20 0.5 0.5
19 0.03 5.4
[b] Enantiomeric ratio: the enantiomeric ratio E measures
the relative rate of hydrolysis of the fast-reacting enantio-
mer as compared to the slow-reacting enantiomer as de-
fined by Sih.[44]
7
0.1 0.3
[a]
[c]
Protein concentration was determined by the method of
Bradford using bovine serum albumin (BSA) as standard.[43]
[b] Specific activity is calculated in mmol/minmgÀ1 protein.
Sigma–Aldrich (St. Louis, USA).
[d] For protein expression and purification details see ref.[9]
[e]
Reaction for 3 h.
[c]
Altus Biologics (Cambridge, USA).
[f] No reaction.
[d] Sigma–Aldrich (St. Louis, USA).
Not determined.
[g]
1184
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1183 – 1192

The 3-(3-Pyridine)propionyl Anchor Group for Protease-Catalyzed Resolutions
FULL PAPERS
substrate. Nevertheless, compound 1a is still preferred
because its higher solubility decreases the reaction
volume and simplifies separation of product and re-
maining starting material (see below).
The 3-(3-pyridine)propionyl derivatives showed
similar or better enantioselectivity than the dihydro-
cinnamoyl derivatives (Table 2). a-Chymotrypsin
showed an enantioselectivity of 52 favoring the (R)-
enantiomer in the hydrolysis of N-3-(3-pyridine)pro-
pionyl 1a versus 56 for N-dihydrocinnamoyl 1c. Subti-
lisin BPN’ showed an enantioselectivity of 125 favor-
ing the (R)-enantiomer in the hydrolysis of N-3-(3-
pyridine)propionyl 1a versus 75 for N-dihydrocinna-
Scheme 2. a-Chymotrypsin-catalyzed resolution of 1a.
moyl 1c. In previous work subtilisin E showed an
enantioselectivity>150 favoring the (R)-enantiomer
for N-dihydrocinnamoyl 1c.[7b,15] Subtilisin BPN’ or maximum yield is 50% in a resolution) with 87% ee
subtilisin E are the best proteases for preparative- at 51% conversion after 4 d (Scheme 2). The remain-
scale resolution of 1a. However, preparation of subti- ing starting material was conveniently separated from
lisin BPN’ or subtilisin E requires a fermentation,[16] the product via acid extraction (0.1 N HCl). Treating
and many organic chemistry laboratories lack access (S)-1a with hydrazine hydrate[19] gave (S)-1 (4.29 g,
to fermentation facilities. For this reason, we use the 40% yield) with 92% ee. Recrystallization from hex-
commercially available a-chymotrypsin as the large- anes/ethyl acetate gave (R)-1 (3.81 g, 35% yield) with
scale example in this work. The enantioselectivity of 98% ee and (S)-1 (3.58 g, 33% yield) with 98% ee.
commercial a-chymotrypsin toward sulfinamides The enantioselectivity (E=47) was slightly lower than
varied with different samples from E=21 to 63 for 1a with small-scale reactions.
and E=18 to 87 for 1c. If one has access to fermenta-
tion facilities, we recommend using subtilisin BPN’ or propionyl group for resolving hindered secondary al-
E as both show consistently high enantioselectivity. cohols, we resolved 2,2-dimethylcyclopentanol (3)
To demonstrate the usefulness of the 3-(3-pyridine)-
To prepare substrate 1a, we treated p-toluenesulfin- using subtilisin Carlsberg. 2,2-Dimethylcyclopentanol
ic acid with oxalyl chloride, followed by ammonolysis (3) can be prepared via asymmetric reduction using
to give the p-toluenesulfinamide in 86% yield chiral organoborane reagents, but it requires long re-
(Scheme 1).[17] Acylation using a mixed anhydride action times and low temperatures, and shows varia-
prepared from isobutyl chloroformate and 3-(3-pyridi- ble selectivity.[20–22] Subtilisin Carlsberg previously
ne)propionic acid gave<10% yield,[17] so we used the showed high enantioselectivity with a structurally re-
symmetrical anhydride of 3-(3-pyridine)propionic lated substrate, cis,cis-6-(2,2-dimethylpropamido)-
acid. This symmetrical anhydride was prepared by spiro
[4.4]nonan-1-ol.[18] Subtilisin Carlsberg also
G
KMnO4 oxidation[18] of 3-(3-pyridine)propanol and showed good enantioselectivity toward 2,2-dimethyl-
coupling of the resulting acid using carbodiimide.[17] cyclopentyl 3-(3-pyridine)propionoate (3a) (E=43).
Thus, treating p-toluenesulfinamide with NaH, fol- Thus, we resolved 3a (4.95 g, 20 mmol) with subtilisin
lowed by addition of the symmetrical anhydride of 3- Carlsberg (4.5 g) to give (S)-3 (1.07 g, 48% yield)
(3-pyridine)propionic acid yielded the crude 1a. Tritu- with 87% ee at 51% conversion (Scheme 3). The
ration with hexanes/ethyl acetate removed unreacted
p-toluenesulfinamide 1 and gave 1a in 73% yield.
This synthetic strategy is simple, inexpensive and does
not require low temperatures, special equipment or
chromatography.
We resolved 1a (20.2 g, 70 mmol) with a-chymo-
trypsin (12 g) to give (R)-1 (4.48 g, 41% yield; the
Scheme 1. Synthesis of racemic N-3-(3-pyridine)propionyl-p-
toluenesulfinamide (1a).
Scheme 3. Subtilisin Carlsberg-catalyzed resolution of 3a.
Adv. Synth. Catal. 2006, 348, 1183 – 1192
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1185

Christopher K. Savile and Romas J. Kazlauskas
FULL PAPERS
product and remaining starting material were conven- enantiomer of 1a similarly to subtilisin E, but binds
iently separated via acid extraction (0.1 N HCl) of the the slow-reacting (S)-enantiomer differently.[7b] Mod-
remaining starting material. Subsequent hydrolysis of eling the first tetrahedral intermediate for the hydrol-
the remaining starting material gave (R)-3 (980 mg, ysis of 1a with a-chymotrypsin gave one catalytically
44% yield) with 89% ee (Scheme 3). Lipases from productive conformation for each enantiomer
Candida rugosa (CRL, EC 3.1.1.3) and Pseudomonas (Figure 1). The two other plausible conformations
cepacia (PCL, EC 3.1.1.3) did not catalyze the hydrol- lacked catalytically essential hydrogen bonds or en-
ysis of 3a; however, lipase B from Candida antarctica countered severe steric clashes with the protein.[25]
(CALB, EC 3.1.1.3) catalyzed the enantioselective hy- Both productive conformations bind the 3-(3-pyridi-
drolysis of 3a (E=108) to give the (R)-enantiomer. ne)propionyl group in the hydrophobic acyl pocket
This discovery provides an enantiocomplementary (specificity pocket)[26] of a-chymotrypsin, based on its
strategy for synthesizing 3. Although kinetic resolu- similarity to phenylalanine.[13] The productive confor-
tion yields both enantiomers, enantiocomplementary mation of (R)-1a has the p-tolyl group in the hydro-
enzymes can simplify the next synthetic steps by pro- phobic leaving group pocket (n-site)[26] and the hy-
viding the necessary enantiomer without further ma- drophilic sulfoxide oxygen in the solvent. The p-tolyl
nipulation.[8]
group fits tightly in the leaving group pocket: [Cips–Sg
We also resolved 1-(2-mesityl)ethanol (4), a useful (Cys42) distance=3.79 Š, Cortho–Sg (Cys42) dis-
chiral inductor for cycloaddition reactions,[23] with tance=3.79 Š, Cortho–Cd2 (His57) distance=3.67 Š,
moderate enantioselectivity on a milligram scale using
Cortho–Cd (Phe41) distance=4.42 Š, Cmeta–Ca (Cys58)
the 3-(3-pyridine)propionyl group. Lipases react distance=3.92 Š].[27] This tight fit suggests a favora-
slowly with esters of 4 because of steric hindrance of ble hydrophobic interaction – tendency of non-polar
the ortho-methyl substituents with active site resi- compounds to transfer from an aqueous phase to an
dues[24] and the acetate and dihydrocinnamate deriva- organic phase – between the p-tolyl group and leaving
tives react only slowly with proteases (<1% conver- group pocket residues.[28] This favorable hydrophobic
sion in 24 h). In contrast, all proteases tested cata- interaction may account for the faster reaction of this
lyzed the hydrolysis of 4a and Aspergillus melleus pro- enantiomer. In contrast, for the slow-reacting enantio-
tease showed the highest enantioselectivity (E=7.5). mer (S)-1a, steric interactions between the active site
We resolved 4a (1.48 g, 5 mmol) with Aspergillus mel- and the sulfinamide moiety prevent it from binding in
leus protease (ca. 1.5 g) to give (R)-4 (352 mg, 43% this pocket so it only sits on top of this region:
yield) with 61% ee at 49% conversion (Scheme 4). Met192 (Ssulfinamide–Sd distance=3.41 Š, Ssulfinamide–Ce
Acid extraction did not cleanly separate product and distance=3.60 Š and Cipso–Ce distance=3.65 Š). This
remaining starting material due to the poor solubility productive conformation differs from the productive
of the hydrophobic ester in acidic solutions, so we conformation of slow enantiomer (S)-1c bound to
used chromatography to separate the product and re- subtilisin E.[7b] With subtilisin E, the substituents at
maining starting material. Subsequent hydrolysis of stereocenter are exchanged, which put the sulfoxide
the remaining starting material gave (S)-4 (343 mg, oxygen in the S1’ pocket and the p-tolyl group in the
42% yield) with 60% ee.
solvent. With a-chymotrypsin, the entire sulfinamide
Molecular modeling showed that a-chymotrypsin moiety of (S)-1a remains outside of the active site
binds the sulfinamide moiety of the fast-reacting (R)- and exposed to the solvent and this lack of hydropho-
bic interaction may account for the slower reaction of
this enantiomer.
The enantiopreference of subtilisin toward (S)-2,2-
dimethylcyclopentanol [(S)-3] agrees with predictive
models for subtilisin.[9,29] The enantiopreference of
subtilisin toward secondary alcohol esters depends
both on relative substituent polarity[9] and relative
substituent size.[29] Subtilisin favors the (R)-enantio-
mer of secondary alcohols when the large substituent
is approximately the size of phenyl because the
phenyl-sized substituent makes a favorable hydropho-
bic interaction with the S1’ leaving-group pocket of
subtilisin, as described above. Subtilisin favors the
(S)-enantiomer of secondary alcohols when the large
substituent is polar (hydrophilic) or too large to bind
in the S1’ leaving-group pocket. For (S)-2,2-dimethyl-
cyclopentanol [(S)-3], the large substituent is the 2,2-
dimethyl quaternary carbon, which is too large for the
Scheme 4. Aspergillus melleus protease-catalyzed resolution
of 4a.
1186
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1183 – 1192

The 3-(3-Pyridine)propionyl Anchor Group for Protease-Catalyzed Resolutions
FULL PAPERS
Figure 1. Catalytically productive tetrahedral intermediates
for the a-chymotrypsin-catalyzed hydrolysis of (R)-1a (I)
and (S)-1a (II) as identified by molecular modeling. The im-
portant active site and substrate atoms (sticks) are colored
as follows: green (substrate carbon), grey (enzyme carbon),
red (oxygen), blue (nitrogen) and orange (sulfur). Surround-
ing atoms (space fill) of a-chymotrypsin are shown in blue.
For clarity, all hydrogen atoms and water molecules are
hidden and Ser219 is hidden to expose the acyl group
pocket. Both I and II maintain all catalytically essential hy-
drogen bonds and the 3-(3-pyridine)methyl moiety of the 3-
(3-pyridine)propionate group binds in the S1 pocket, as ex-
pected based on its similarity to phenylalanine.[13,26] In the
fast-reacting enantiomer, (R)-1a (I), the p-tolyl group binds
in the hydrophobic leaving group pocket and sulfoxide
oxygen is exposed to solvent water. In the slow-reacting en-
antiomer, (S)-1a (II), the sulfoxide oxygen is forced out of
the hydrophobic leaving group pocket because of steric in-
teraction between the sulfinamide moiety and active site
residues. This interaction places both the p-tolyl group and
sulfoxide oxygen in solvent where this is no favorable hydro-
phobic interaction with leaving group pocket residues. The
non-productive conformations of (R)-1 and (S)-1 (not
shown) encountered severe steric clash with the active site
residues and lacked catalytically essential hydrogen bonds.
This enzyme-catalyzed route to enantiopure p-tol-
uenesulfinamide 1 using the 3-(3-pyridine)propionyl
group compares well with current chemical routes to
sulfinamides. One chemical route to enantiopure p-
toluenesulfinamide starts with menthyl p-toluenesulfi-
nate (Andersenꢁs reagent) prepared from menthol.[34]
The natural (À)-enantiomer of menthol is readily
available, but the unnatural (+)-enantiomer of men-
thol is ca. six times more expensive. The other syn-
thetic route to enantiopure p-toluenesulfinamide is a
double displacement strategy using a chiral auxiliary
derived from indanol.[35] This route includes compli-
cated steps requiring low temperatures and moisture-
and air-sensitive reagents. Our enzymatic resolution is
simple, convenient, avoids the use of costly auxiliaries
and allows facile separation of product and remaining
starting material.
S1’ pocket. This poor fit destabilizes reaction of the
Proteases are also useful for resolving sterically hin-
(R)-enantiomer and thus, the smaller methylene dered alcohols, such as 2,2-dimethylcyclopentanol (3)
group binds in the S1’ pocket, favoring reaction of the and 1-(2-mesityl)ethanol (4), that cannot be easily
(S)-enantiomer.
prepared through chemical routes and react slowly
Sulfinamides, such as p-toluenesulfinamide, are with lipases. Proteases show higher reactivity toward
useful chiral auxiliaries for synthesis of amines.[30] these large alcohols when 3-(3-pyridine)propionic
When condensed with an aldehyde or ketone to give acid is used as acyl group and the enantioselectivity
the sulfinimine, the N-sulfinyl group directs nucleo- can be high when there is a large difference in the rel-
philic addition across the C=N bond. This addition ative size[29] or polarity[9] of the alcohol substituents.
yields the N-alkylsulfinamide, which upon hydrolysis
The 3-(3-pyridine)propionyl group maintains reac-
À
of the S N link yields an amine. Enantioselective syn- tivity and enantioselectivity with subtilisin and a-chy-
theses using p-toluenesulfinimine include preparations motrypsin by mimicking phenylalanine and anchoring
of amines,[31] a- and b-amino acids,[30] aziridines[32] substrate to the active site. The pyridine moiety
and aminophosphonic acids.[33]
allows mild acid extraction to separate product and
Adv. Synth. Catal. 2006, 348, 1183 – 1192
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1187

Christopher K. Savile and Romas J. Kazlauskas
FULL PAPERS
remaining starting material and thus avoid chromato- Racemic Toluenesulfinamide (Æ)-1
graphic separation. Researchers have also tested
Oxalyl chloride (67.2 g, 530 mmol) was added drop-wise to
other acyl groups that simplify separation of product
and starting material. For example, Fukui and co-
workers[36,37] used the succinyl monoester acyl group
in a lipase-catalyzed resolution of menthol. While the
succinate group allowed base extraction to separate
product and remaining starting material, the charged
succinate esters were poor substrates for lipases,
which favor non-polar, insoluble substrates.[2] We pre-
viously tested the monomethyl ester succinyl acyl
group for a subtilisin E-catalyzed resolution of p-tol-
uenesulfinamide, but it showed lower enantioselectivi-
ty (E=52) than the dihydrocinnamoyl derivative (E=
>150).[7b]
p-toluenesulfinic acid sodium salt (89.0 g, 500 mmol) in tolu-
ene at 08C.[17] After 1 h at room temperature, the reaction
mixture was added to a biphasic mixture of NH4OH
(500 mL) and EtOAc (500 mL) at 08C and then stirred at
room temperature. After 1 h, the reaction mixture was dilut-
ed with EtOAc (500 mL) and the two layers were separated.
The aqueous layer was extracted with EtOAc (2”500 mL).
The combined EtOAc layers were washed with saturated
NaCl (500 mL) and dried over Na2SO4. The aqueous layer
was then extracted with CH2Cl2 (2”500 mL). The combined
CH2Cl2 layers were washed with saturated NaCl (500 mL)
and dried over Na2SO4. The combined organic layers were
concentrated under vacuum to give a white powder; yield:
1
66.7 g (86%); mp 115–1178C (lit.[17] 117–1188C); H NM R:
d=2.42 (s, 3H, PhCH3), 4.30 (br s, 2H, NH2), 7.30 (d, J=
8.1, 2H, phenyl), 7.62 (d, J=8.1, 2H, phenyl); 13C NM R:
d=21.7 (PhCH3), 125.6, 129.7, 141.5, 143.6 (phenyl).
The enantiomers were separated using HPLC [Chiralcel
OD-H column, 85:15 hexanes/EtOH, 0.75 mL/min, 254 nm;
(R)-1, tR =9.8 min; (S)-1, tR =11.1 min].
Conclusions
The 3-(3-pyridine)propionyl group increases substrate
binding to proteases, increases substrate solubility in
aqueous solutions, and provides a facile route for sep-
aration of remaining starting material and product.
This is a cost-effective route for large-scale prepara-
tion of the useful chiral auxiliary, p-toluenesulfina-
mide 1, and sterically hindered secondary alcohols
that are poor substrates for other hydrolases.
3-(3-Pyridine)propionic Acid
Solid KMnO4 (256 g, 1.62 mol) was added in portions over
30 min to a solution of 3-(3-pyridine)propanol (200 g, 1.46
mol) in 3 N H2SO4 (1.5 L) at 08C.[18] The reaction mixture
was stirred at room temperature. After 24 h, the reaction
mixture was adjusted to pH 6 by addition of solid KOH. In-
soluble MnO2 was pelleted by centrifugation and the clear
solution was decanted. Water was removed under vacuum
to afford 3-(3-pyridine)propionic acid;[38] yield: 153 g
Experimental Section
1
(70%); H NM R:d=2.72 [t, J=7.2, 2H, C(O)CH2], 3.03 (t,
General Remarks
J=7.5, 2H, CH2Ph), 7.34 (m, 1H, pyridyl), 7.68 (m, 1H, pyr-
idyl), 8.52 (m, 2H, pyridyl).
1H and 13C NMR spectra were obtained as CDCl3 solutions
at 300 MHz and 75 MHz, respectively. Chemical shifts are
expressed in ppm (d) and are referenced to tetramethylsi-
lane or solvent signal. Coupling constants are reported in
Hertz (Hz). GC analyses were performed on a 25 m ”
0.25 mm Chrompack CP-Chiralsil-Dex CB column (Varian
Inc., Palo Alto, USA) with He as carrier gas using one of
the two following temperature programs: A 17.5 psi, 508C,
58C minÀ1, 1508C held for 5 min, 2.58C minÀ1; 1758C held
for 5 min, 58C minÀ1, 2008C held for 30 min; B 17.5 psi,
1008C, held for 15 min, 258C minÀ1; 2008C held for 21 min).
HPLC analyses were performed on a 4.6”250 mm Daicel
Chiralcel OD column (Chiral Technologies, Exton, USA)
and monitored at 254 nm. Subtilisin Carlsberg was pur-
chased from Sigma–Aldrich (St. Louis, USA) and a-chymo-
trypsin was purchased from Sigma–Aldrich (St. Louis, USA)
or Amresco (Solon, USA). Other hydrolases were a gener-
ous gift from Altus Biologics (Cambridge, USA). All re-
agents, buffers, starting materials and anhydrous solvents
were purchased from Sigma–Aldrich (Milwaukee, USA)
and used without purification. All air- and moisture-sensi-
tive reactions were performed under argon. Substrates 1b,
1c, 2c and 4c were available from previous studies.[7b,9]
3-(3-Pyridine)propionic Acid Anhydride
N-(3-Dimethylaminopropyl)-N’-ethylcarbodiimide
hydro-
chloride (25.2 g, 132 mmol) was added to a solution of 3-(3-
pyridine)propionic acid (40.0 g, 265 mmol) in CH2Cl2
(750 mL) and NEt3 (37 mL) at 08C and stirred at room tem-
perature.[17] After 24 h, the reaction was washed with ice-
cold saturated NaHCO3 (3”500 mL), dried over MgSO4
and concentrated under vacuum to give a pale yellow oil;
yield: 34.1 g (91%); 1H NM R: d=2.72 [t, J=7.2, 2H,
C(O)CH2], 3.03 (t, J=7.5, 2H, CH2Ph), 7.24 (m, 1H, pyrid-
yl), 7.58 (m, 1H, pyridyl), 8.50 (m, 2H, pyridyl).
Racemic N-3-(3-Pyridine)propionyl-p-toluene-
sulfinamide [(Æ)-1a]
Sodium hydride (60% dispersion in oil; 12.0 g, 300 mmol)
was added portion-wise over 15 min to a solution of p-tolu-
1188
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1183 – 1192

The 3-(3-Pyridine)propionyl Anchor Group for Protease-Catalyzed Resolutions
FULL PAPERS
enesulfinamide (15.5 g, 100 mmol) in THF (750 mL) at 08C.
The symmetric anhydride of 3-(3-pyridine)propionic acid
(32.1 g, 113 mmol) was added drop-wise over 15 min at 08C
and the reaction mixture was then stirred at room tempera-
ture for 3 h.[17] The reaction mixture was diluted with
EtOAc (400 mL) and saturated NaHCO3 (400 mL) was
added slowly. The layers were separated and the aqueous
layer was extracted with EtOAc (3”250 mL). The combined
EtOAc layers were washed with saturated NaHCO3
(500 mL) and dried over MgSO4. The aqueous layer was ex-
tracted with CH2Cl2 (2”250 mL). The combined CH2Cl2
layers were washed with NaHCO3 (250 mL) and dried over
MgSO4. The combined organic layers were concentrated
under vacuum to give a pale yellow solid. Trituration with
hexane/ethyl acetate gave a white powder; yield: 21.1 g
to 75:25 hexanes/acetone) gave a clear liquid; yield: 1.05 g
(86%); H NM R:d=0.90 (s, 3H, CH3), 0.93 (s, 3H, CH3),
1
1.43 (m, 2H, CH2), 1.52 (m, 2H, CH2), 1.67 (m, 2H, CH2),
2.08 (m, 2H, CH2), 4.73 (m, 1H, CH), 7.27 (m, 1H, phenyl),
7.61 (m, 1H, phenyl), 8.49 (m, 2H, phenyl); 13C NM R:d=
20.6 (CH2), 22.2 (CH2), 26.6 (CH2), 28.2 (CH2Pyr), 30.5
(CH2), 35.6 [C(O)CH2], 38.0 (CH2), 41.9 (C), 123.4, 135.9,
147.7, 149.8 (pyridyl), 172.1 (C=O); HR-MS: m/z=
248.1650, calcd for C15H21NO2 [M+H]+: 248.1655. The
enantiomers could not be separated using GC.
Racemic 3-(3-Pyridine)propionic acid 2-mesityl-ethyl
ester [(Æ)-4a]: Purification on silica gel (100% hexanes to
60:40 hexanes/acetone) gave a clear liquid; yield: 1.27 g
1
(86%); H NM R:d=1.53 (d, J=6.9, 3H, CH3), 2.26 (s, 3H,
p-CH3), 2.40 (s, 6H, o-CH3), 2.68 [m, 2H, C(O)CH2], 2.97
(t, J=7.5, 2H, CH2Pyr), 6.28 (q, J=6.9, 1H, CH), 6.83 (s,
2H, phenyl), 7.24–7.28 (m, 1H, pyridyl), 7.56–7.60 (m, 1H,
pyridyl), 8.46–8.50 (m, 2H, pyridyl); 13C NM R: d=19.6
(CH3), 20.5 (CH3), 20.9 (CH3), 28.1 (CH2Pyr), 35.6
(C(O)CH2), 69.9 (CH), 123.5, 130.1, 134.2, 135.9, 136.0,
136.1, 137.2, 147.7, 149.8 (phenyl or pyridyl), 171.7 (C=O);
HR-MS: m/z=298.1813, calcd. for C17H24NO2 [M+H]+:
The enantiomers could not be separated using GC.
1
(73%); mp 161–1638C; H NM R: d=2.39 (s, 3H, PhCH3),
2.72 [m, 2H, C(O)CH2], 3.01 (t, J=7.2, 2H, CH2Pyr), 4.78
(br s, 1H, NH), 7.25–7.48 (m, 3H, phenyl or pyridyl), 7.47
(m, 2H, phenyl or pyridyl), 7.68 (m, 1H, phenyl or pyridyl),
8.25 (m, 2H, pyridyl); 13C NMR ([DMSO-d6): d=21.4
(PhCH3), 27.9 (CH2Pyr), 36.9 [C(O)CH2], 124.0, 125.4,
130.2, 136.4, 136.6, 140.9, 142.1, 147.9, 150.2 (phenyl or pyr-
idyl), 173.8 (C=O); HR-MS: m/z=289.0989, calcd. for
C15H17N2O2S [M+H]+: 289.1010.
The enantiomers were separated using HPLC [Chiralcel
OD-H column, 85:15 hexanes/EtOH, 0.75 mL/min, 254 nm;
(R)-1a, tR =20.0 min; (S)-1a, tR =22.5 min].
Racemic Acetic Acid 1-Phenylethyl Ester [(Æ)-2b]
Acetyl chloride (15 mmol) was added drop-wise to a stirred
solution of 1-phenethyl alcohol (10 mmol) and pyridine
(15 mmol) in CH2Cl2 (50 mL) at 08C. The ice bath was re-
moved and the mixture stirred until the reaction was com-
plete by TLC. The reaction was quenched with the addition
of saturated NaHCO3 (25 mL). The layers were separated
and the aqueous layer was extracted with CH2Cl2 (2”
25 mL). The combined organic layers were washed with 1 N
HCl (2”25 mL), saturated NaHCO3 (2”25 mL), saturated
NaCl (25 mL) and dried over Na2SO4. The organic layer
was concentrated under vacuum to give the crude ester. Pu-
rification on silica gel (100% hexanes to 95:5 hexanes/
EtOAc) gave the 2b as a clear liquid; yield: 1.80 g (73%);
1H NM R: d=1.55 (d, J=6.6, 3H, CH3), 2.09 [m, 3H,
C(O)CH3], 5.90 (q, J=6.6, 1H, CH), 7.36–7.38 (m, 5H,
phenyl); 13C NM R: d=21.4 (CH3), 22.3 (CH3), 72.4 (CH),
126.2, 128.0, 128.6, 141.8 (phenyl), 170.4 (C=O). The enan-
tiomers were separated using GC [program B; (S)-2b, tR =
9.7 min; (R)-2b, tR =11.9 min].
General Procedure for the Synthesis of Racemic
Esters 2a–4a
N-(3-Dimethylaminopropyl)-N’-ethylcarbodiimide
hydro-
chloride (7.5 mmol) was added portion-wise to a stirred so-
lution of a secondary alcohol (5 mmol), 3-(3-pyridine)pro-
pionic
acid
(7.5 mmol),
4-(dimethylamino)-pyridine
(0.5 mmol) and NEt3 (7.5 mmol) in CH2Cl2 (25 mL) at 08C.
The ice bath was removed and the reaction mixture was stir-
red at room temperature for 48 h. The reaction was
quenched with the addition of saturated NaHCO3 (25 mL).
The layers were separated and the aqueous layer was ex-
tracted with CH2Cl2 (2”25 mL). The combined organic
layers were washed with saturated NaHCO3 (2”25 mL) and
dried over MgSO4. The organic layer was concentrated
under vacuum to give the crude ester. The relevant analyti-
cal data are given below:
Racemic 3-(3-Pyridine)propionic acid 1-phenyl-ethyl ester
[(Æ)-2a]: Purification on silica gel (100% hexanes to 60:40
hexanes/acetone) gave a clear liquid; yield: 1.10 g (87%);
1H NM R: d=1.51 (d, J=6.6, 3H, CH3), 2.70 [m, 2H,
C(O)CH2], 2.98 (t, J=7.5, 2H, CH2Pyr), 5.88 (q, J=6.6,
1H, CH), 7.32–7.39 (m, 6H, phenyl), 7.54–7.58 (m, 1H,
phenyl), 8.46–8.50 (m, 2H, pyridyl); 13C NM R: d=22.2
(CH3), 28.1 (CH2Pyr), 35.6 [C(O)CH2], 72.7 (CH), 123.5,
126.1, 128.0, 128.6, 135.9, 136.1, 141.5, 147.7, 149.8 (phenyl
or pyridyl), 171.6 (C=O); HR-MS: m/z=256.1336, calcd. for
C16H18NO2 [M+H]+: 256.1337. The enantiomers were sepa-
rated using GC [program A; (S)-2a, tR =52.3 min; (R)-2a,
tR =52.5 min].
Racemic 2,2-Dimethylcyclopentanol [(Æ)-3]
NaBH4 (3.42 g, 90 mmol) was added portion-wise to a stir-
red solution of 2,2-dimethylcyclopentanone (6.72 g,
60 mmol) in EtOH (100 mL) at 08C. The ice bath was re-
moved and the reaction solution was stirred for 3 h at room
temperature. The solution was cooled to 08C, quenched
with 1 N HCl (50 mL) and extracted with EtOAc (3”
50 mL). The combined organic layers were washed with sa-
turated NaHCO3 (50 mL), saturated NaCl (50 mL), dried
over MgSO4 and concentrated under vacuum to give a clear
Racemic 3-(3-Pyridine)propionic acid 2,2-dimethylcyclo-
pentanol [(Æ)-3a]: Purification on silica gel (100% hexanes
1
liquid; yield: 5.87 g (86%); H NM R:d=0.95 (s, 3H, CH3),
Adv. Synth. Catal. 2006, 348, 1183 – 1192
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1189

Christopher K. Savile and Romas J. Kazlauskas
FULL PAPERS
0.97 (s, 3H, CH3), 1.40 (m, 2H, CH2), 1.57 (m, 2H, CH2),
1.72 (m, 2H, CH2), 2.03 (m, 2H, CH2), 3.69 (t, J=6.3, 1H,
CH); 13C NM R:d=19.9 (CH2), 21.3 (CH2) 26.7 (CH2), 32.7
(50 mL), saturated NaCl (50 mL) and concentrated under
vacuum to give (S)-1; yield: 4.29 g (40%) with 92% ee. Re-
crystallization from hexanes/ethyl acetate gave (R)-1 (yield:
3.81 g, 35%) with 98% ee and (S)-1 (yield: 3.58 g, 33%)
with 98% ee.
(CH2), 37.5 (CH2), 42.1 [C(CH3)2], 81.4 (CHOH). The
A
enantiomers were separated using GC [program B; (R)-3,
tR =5.6 min; (S)-3, tR =5.8 min].
Subtilisin Carlsberg-Catalyzed Resolution of (Æ)-3a
Small-Scale Protease-Catalyzed Hydrolysis of 2a and
3a
Protease from Bacillus licheniformis (subtilisin Carlsberg;
50 mL of 92 mg/mL solution; 4.5 g) was added to BES
buffer (400 mL, 1 mm, pH 7.2) and stirred for 15 min. Sub-
strate 3a (4.95 g, 20 mmol) was dissolved in MeCN (50 mL)
and added to the enzyme solution. The rate of hydrolysis
was monitored by pH stat, which maintained the pH at 7.2
by automatic titration with 1 N NaOH. After 2 d, the reac-
tion was terminated by extraction of remaining starting ma-
terial and product with EtOAc (3”75 mL). The organic
layer was extracted with ice-cold 0.1 N HCl (3”50 mL). The
combined aqueous layers were then extracted with EtOAc
(25 mL). The combined EtOAc layers were dried over
MgSO4 and concentrated under vacuum to give (S)-3; yield:
1.07 g (48%) with 87% ee. The aqueous layer was neutral-
ized with solid Na2HCO3 and extracted with EtOAc (3”
50 mL). The combined organic layers were dried over
Na2SO4 and concentrated under vacuum to give (R)-3a,
which was subsequently hydrolyzed in ethanolic KOH (1 N,
1:1 ethanol/water) to give (R)-3; yield: 980 mg (44%) with
89% ee.
Bacillus lentus subtilisin (subtilisin BL), Aspergillus melleus
protease, Aspergillus oryzae protease, subtilisin Carlsberg,
a-chymotrypsin from bovine pancreas, Streptomyces griseus
protease [10 mg/mL; solution in 50 mm BES buffer (pH 7.2,
450 mL)] and substrate (100 mMin CH CN, 50 mL) contain-
3
1
ing 50 mm n-decane as internal standard were mixed in a =
2
dram glass vial. The reaction mixture was shaken at 308C
for 6 h with 2a and 24 h with 2b, 2c, 3a and 4a. For specific
activity measurements the reaction was stopped at a conver-
sion below 10%. The reaction was terminated with the addi-
tion of CH2Cl2 (500 mL). The phases were separated by cen-
trifugation, and the organic layers were collected. The aque-
ous phase was extracted with CH2Cl2 (2”500 mL) and the
combined organics were evaporated under a stream of air.
The residue was diluted with EtOAc (150 mL) and analyzed
by GC. The enantiomers were separated using the condi-
tions described above.
a-Chymotrypsin-Catalyzed Resolution of (Æ)-1a
Aspergillus melleus Protease-Catalyzed Resolution of
(Æ)-4a
a-Chymotrypsin (12 g) was added to a solution of BES
buffer (3.15 L, 1 mm, pH 7.2) and 100 mm KCl and stirred
for 15 min to ensure complete dissolution. Substrate 1a
(20.2 g, 70 mmol) was dissolved in dimethylformamide
(350 mL) and added drop-wise to the enzyme solution. The
rate of hydrolysis was monitored by pH stat, which main-
tained the pH at 7.2 by automatic titration with 1 N NaOH.
At ca. 50% conversion (4 d), the reaction was terminated
by extraction of remaining starting material and product
with CH2Cl2 (3”500 mL). The combined organic layers
were washed with dH2O (3”500 mL), saturated NaCl (1”
500 mL), dried over MgSO4 and concentrated under
vacuum. The crude mixture was dissolved in EtOAc
(250 mL) and unreacted starting material was extracted with
ice-cold 0.1 N HCl (2”100 mL). The combined aqueous
layers were then back-extracted with EtOAc (50 mL). The
combined EtOAc layers were washed with saturated
NaHCO3 (100 mL), saturated NaCl (100 mL), dried over
MgSO4 and concentrated under vacuum to give (R)-1 as a
white solid; yield: 4.48 g (41%) with 87% ee. The combined
aqueous layers were neutralized with solid NaHCO3 and ex-
tracted with CH2Cl2 (2”200 mL). The combined CH2Cl2
layers were washed with sat. NaHCO3 (100 mL), sat. NaCl
(100 mL) and dried over MgSO4. The solution was concen-
trated under vacuum to give (S)-1a, which was subsequently
treated with hydrazine hydrate (35 mL).[19] After stirring for
3 h, the reaction solution was diluted with CH2Cl2 (100 mL)
and washed with 1 N HCl (50 mL), saturated NaHCO3
Aspergillus melleus protease (1.5 g) was added to BES
buffer (180 mL, 1 mm, pH 7.2) and stirred for 30 min to
ensure complete dissolution. Substrate 4a (1.49 g, 5 mmol)
was dissolved in DMF (20 mL) and added to the enzyme so-
lution. The rate of hydrolysis was monitored by pH stat,
which maintained the pH at 7.2 by automatic titration with
1 N NaOH. After 5 d, the reaction was terminated by ex-
traction of remaining starting material and product with
EtOAc (3”50 mL). Product and substrate were difficult to
separate via acid extraction. Thus, product and remaining
starting material were separated using column chromatogra-
phy (100% hexanes to 4:1 hexanes/acetone) to give (R)-4
(yield: 352 mg, 43%) with 61% ee and (S)-4a, which was
subsequently hydrolyzed in ethanolic KOH (1 N, 1:1: etha-
nol/water) to give (S)-4 (yield: 343 mg, 42%) with 60% ee.
Absolute Configuration of 2,2-Dimethylcyclopentanol
(3)
The absolute configuration of alcohol 3 was assigned using
the configurational correlation model for the corresponding
(R)-MTPA derivatives, which were synthesized using (R)-
(+)-a-methoxy-a-(trifluoromethyl)-phenylacetyl chloride in
pyridine.[39] 1H NM R: d=0.91 (s, 3H, CH3, R-enantiomer),
1190
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1183 – 1192

The 3-(3-Pyridine)propionyl Anchor Group for Protease-Catalyzed Resolutions
FULL PAPERS
0.93 (s, 3H, CH3, R-enantiomer), 0.98 (s, 3H, CH3, S-enan- References
tiomer), 1.00 (s, 3H, CH3, S-enantiomer), 1.46 (m, 2H,
CH2), 1.57 (m, 2H, CH2), 1.68 (m, 2H, CH2), 2.18 (m, 2H,
CH2), 3.54 (s, 3H, OCH3, S-enantiomer), 3.57 (s, 3H,
OCH3, R-enantiomer), 4.90 (m, 1H, CH), 7.42 (m, 3H,
phenyl), 7.52 (m, 2H, phenyl).
[1] a) K. Faber, Biotransformations in Organic Chemistry,
5th edn., Springer-Verlag, Berlin, 2004; b) K. Drauz, H.
Waldman (Eds. ), Enzyme Catalysis in Organic Synthe-
sis Wiley-VCH, Weinheim, 2002; c) C.-H. Wong, G. M.
Whitesides, Enzymes in Synthetic Organic Chemistry,
Pergamon, New York, 1994.
[2] U. T. Bornscheuer, R. J. Kazlauskas, Hydrolases in Or-
ganic Synthesis - Regio- and Stereoselective Biotransfor-
mations, Wiley-VCH, Weinheim, 1999, pp. 17–18.
[3] M. Breuer, K. Ditrich, T. Habicher, B. Hauer, M. Keß-
eler, R. Stürmer, T. Zelinski, Angew. Chem. Int. Ed.
2004, 43, 788–824.
[4] M. Cygler, P. Grochulski, R. J. Kazlauskas, J. D. Schrag,
F. Bouthillier, B. Rubin, A. N. Serreqi, A. K. Gupta, J.
Am. Chem. Soc. 1994, 116, 3180–3186.
Modeling of Tetrahedral Intermediates bound to
a-Chymotrypsin
All modeling was performed using Insight II 2000.1/Discover
(Accelrys, San Diego, USA) on a SGI Tezro UNIX worksta-
tion using the AMBER force field.[40] We used a non-
bonded cutoff distance of 8 Š, a distance-dependent dielec-
tric of 1.0 and scaled the 1–4 van der Waals interactions by
50%. Protein structures in Figure 1 were created using
PyMOL (Delano Scientific, San Carlos, CA, USA). The X-
ray crystal structure of a-chymotrypsin (entry 6CHA)[41] is
from the Protein Data Bank. The hydrogen atoms were
added to correspond to pH 7.0. Histidines were uncharged,
aspartates and glutamates were negatively charged, and ar-
ginines and lysines were positively charged. The catalytic
histidine (His64) was protonated. The positions of the water
hydrogens and then the enzyme hydrogens were optimized
using a consecutive series of short (1 ps) molecular dynamic
runs and energy minimizations.[42] This optimization was re-
peated until there was <2 kcalmolÀ1 in the energy of the
minimized structures. Thereafter, an iterative series of ge-
ometry optimizations were performed on the water hydro-
gens, enzyme hydrogens and full water molecules. Finally,
the whole system was geometry optimized.
[5] A. Mezzitti, J. Schrag, C. S. Cheong, R. J. Kazlauskas,
Chem. Biol. 2005, 12, 427–437.
[6] J. D. A. Tyndall, T. Tessa Nall, D. P. Fairlie, Chem. Rev.
2005, 105, 973–1000.
[7] a) D. C. Muchmore, US Patent 5,215,918, 1993; b) C. K.
Savile, V. P. Magloire, R. J. Kazlauskas, J. Am. Chem.
Soc. 2005, 127, 2104–2113.
[8] P. F. Mugford, S. M. Lait, B. A. Keay, R. J. Kazlauskas,
ChemBioChem, 2004, 5, 980–987.
[9] C. K. Savile, R. J. Kazlauskas, J. Am. Chem. Soc. 2005,
127, 12228–12229.
[10] a) Y. Y. Lin, D. N. Palmer, J. B. Jones, Can. J. Chem.
1974, 52, 469–476.
[11] a) D. A. Estell, T. P. Graycar, J. V. Miller, D. B. Powers,
J. P. Burnier, P. G. Ng, J. A. Wells, Science 1986, 233,
659–663; b) J. A. Wells, D. B. Powers, R. R. Bott, T. P.
Graycar, D. A. Estell, Proc. Natl. Acad. Sci. USA 1987,
84, 1219–1223.
[12] T. Pohl, H. Waldmann, Tetrahedron Lett. 1995, 36,
2963–2966.
[13] N. Kundu, S. Roy, F. Maenza, Eur. J. Biochem. 1972,
28, 311–315.
The tetrahedral intermediates were built manually and co-
valently linked to Ser195. Non-standard partial charges were
calculated using a formal charge of À1 for the substrate oxy-
anion. Energy minimization proceeded in three stages. First,
minimization of substrate with only the protein constrained
(25 kcalmolÀ1
Š
À2); second, minimization with only the pro-
tein backbone constrained (25 kcalmolÀ1
Š
À2) and for the
[14] Based on a limit of detection of 0.1% conversion.
[15] P. F. Mugford, V. P. Magloire, R. J. Kazlauskas, J. Am.
Chem. Soc. 2005, 127, 6536–6537.
final stage the minimization was continued without con-
straints until the rms value was less than 0.0005 kcal
molÀ1 À1. A catalytically productive complex required all
Š
five hydrogen bonds within the catalytic machinery. We set
generous limits for a hydrogen bond: a donor to acceptor
atom distance of less than 3.1 Š with a nearly linear ar-
rangement (>1208 angle) of donor atom, hydrogen, and ac-
ceptor atom. Structures lacking any of the five catalytically
relevant hydrogen bonds or encountering severe steric clash
with enzyme were deemed non-productive.
[16] a) C. R. Harwood, S. M. Cutting, Molecular Biological
Methods for Bacillus, John Wiley & Sons, Chichester,
U. K. , 1990, pp. 33–35 and 391–402; b) S.-J. Cho, S.-H.
Oh, R. D. Pridmore, M. A. Juillerat, C.-H. Lee, Agric.
Food Chem. 2003, 51, 7664–7670.
[17] B. J. Backes, D. R. Dragoli, J. A. Ellman, J. Org. Chem.
1999, 64, 5472–5478.
[18] L. Crombie, S. H. Harper, J. Am. Chem. Soc. 1950,
2685–2689.
[19] D. D. Keith, J. A. Tortora, R. Yang, J. Org. Chem. 1978,
43, 3711–3713.
[20] H. C. Brown, B. T. Cho, W. S. Park, J. Org. Chem. 1988,
53, 1231–1238.
[21] H. C. Brown, M. Srebnik, P. V. Ramachandran, J. Org.
Chem. 1989, 54, 1577–1583.
Acknowledgements
[22] B. T. Cho, Y. S. Chun, Tetrahedron: Asymmetry 1992, 3,
1539–1542.
[23] P. Delair, A. M. Kanazawa, M. B. M. de Azevedo, A. E.
Greene, Tetrahedron: Asymmetry 1996, 7, 2707–2710.
C.K.S. thanks McGill University for a fellowship. We thank
the University of Minnesota for financial support and the
Minnesota Supercomputing Institute for access to molecular
modeling computers and software.
Adv. Synth. Catal. 2006, 348, 1183 – 1192
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1191

Christopher K. Savile and Romas J. Kazlauskas
FULL PAPERS
[24] K. Naemura, M. Murata, R. Tanaka, M. Yano, K.
Hirose, Y. Tobe, Tetrahedron: Asymmetry 1996, 7,
3285–3294.
drophobic surface of the leaving group pocket, which
would increase the value, and neglects that the pocket
does not completely surround the p-tolyl group, which
would decrease the value. It is likely that these two
omissions cancel out. For details, see: a) A. Fersht,
Structure and Mechanism in Protein Science W. H.
Freeman and Company, New York, 1998, pp. 324–348;
b) A. Leo, C. Hansch, D. Elkins, Chem. Rev. 1971, 71,
525–616.
[25] The sulfoxide oxygen of non-productive (R)-1a
bumped catalytic His57 (O–Cd2 distance=3.04 Š) and
the p-tolyl group was hindered by Met192 (Cortho–Cg
distance=3.33 Š and Cmeta–Cg distance=3.73 Š),
Gly193 (Cpara–N distance=3.96 Š) and Phe41 (CH3–Ce
distance=3.75 Š). As well, the p-tolyl group forms an
unfavorable syn-pentane interaction with the oxyanion.
This intramolecular interaction and severe steric clash
between p-tolyl group and active site residues signifi-
cantly distorts the substrate and results in the loss of
two catalytically relevant hydrogen bonds (Nd2–Og dis-
tance=4.49 Š and Nd2–Nsulfinamide distance=5.13 Š).
The p-tolyl group of non-productive (S)-1a was hin-
dered by catalytic His57 (Cortho–Ne2 distance=3.38 Š,
Cortho–Cd2 distance=3.27 Š, Cmeta–Cd2 distance=
3.73 Š and Cmeta–Cg distance=3.90 Š) and the sulfina-
mide bumped the Met192 (Ssulfinamide–Sd distance=
3.33 Š and Ssulfinamide–Ce distance=3.52 Š). The steric
hindrance between p-tolyl group and active site resi-
dues results in the loss of one catalytically relevant hy-
drogen bond (Nd2–Nsulfinamide distance=3.20 Š).
[29] P. A. Fitzpatrick, A. M. Klibanov, J. Am. Chem. Soc.
1991, 113, 3166–3171.
[30] F. A. Davis, P. Zhou, B.-C. Chen, Chem. Soc., Rev.
1998, 27, 13–18.
[31] P. Moreau, M. Essiz, J.-Y. Merour, D. Bouzard, Tetrahe-
dron: Asymmetry 1997, 8, 591–598.
[32] F. A. Davis, P. Zhou, G. V. Reddy, J. Org. Chem. 1994,
59, 3243–3245.
[33] I. M. Lefebvre, S. A. Evans, J. Org. Chem. 1997, 62,
7532–7533.
[34] a) K. K. Anderson, in: The Chemistry of Sulphones and
Sulphoxides, (Eds.: S. Patai, Z. Rappoport, C. J. M. Stir-
ling), John Wiley & Sons, New York, 1988, pp. 55–94;
b) H. J. Phillips, J. Chem. Soc., 1925, 127, 2552–2587;
c) M. Hulce, J. P. Mallamo, L. L. Frye, T. P. Kogan,
G. A. Posner, Org. Synth. 1986, 64, 196–206.
[35] Z. Han, D. Krishnamurthy, P. Grover, Q. K. Fang, C. H.
Senanayake, J. Am. Chem. Soc. 2002, 124, 7880–7881.
[36] T. Omata, N. Iwamoto, T. Kimura, A. Tanaka, S. Fukui,
Eur. J. Appl. Microbiol. Biotechnol. 1981, 11, 199–204.
[37] S. Fukui, A. Tanaka, Methods Enzymol. 1987, 136,
293–302.
[38] F. A. Walker, M. Benson, J. Am. Chem. Soc. 1980, 102,
5530–5538.
[39] J. A. Dale, H. S. Mosher, J. Am. Chem. Soc. 1973, 95,
512–519.
[40] a) S. J. Weiner, P. A. Kollman, D. A. Case, U. C. Singh,
C. Ghio, G. Alagona, S. Profet, P. Weiner, J. Am.
Chem. Soc. 1984, 106, 765–784; b) S. J. Weiner, P. A.
Kollman, D. T. Nguyen, D. A. Case, J. Comp. Chem.
1986, 7, 230–252.
[26] S. G. Cohen, Trans. N. Y. Acad. Sci. 1969, 31, 705–719.
À
À
[27] The van der Waals distance for CH C=3.99 Š, CH
À
À
N=3.84 Š, CH S=4.09 Š and O HC=3.81 Š. These
were estimated from the van der Waals radii of carbon
(1.70 Š), nitrogen (1.55 Š), sulfur (1.80 Š) or oxygen
À
(1.52 Š) and hydrogen (1.20 Š) and the C H bond
length (1.09 Š) from: A. Bondi, J. Phys. Chem. 1964,
68, 441–451.
[28] The energy of a hydrophobic interaction is from the re-
gaining of entropy by water after it is removed from a
hydrophobic group and can be estimated using the in-
cremental Gibbs free energy of transfer. The incremen-
tal Gibbs free energy of transfer of the p-tolyl group
from n-octanol to water is estimated to be ca. +3.4 kcal
molÀ1, using DGtrans =2.303RTp, where p is the hydro-
phobicity constant of the p-tolyl group [p=2.5, using
ClogP (BioByte, Claremont, USA)] relative to hydro-
gen, whereas the sulfoxide oxygen is ca.
[41] A. Tulinsky, R. A. Blevins, J. Biol. Chem. 1987, 262,
7737–7743.
[42] S. Raza, L. Fransson, K. Hult, Protein Sci. 2001, 10,
329–338.
[43] M. M. Bradford, Anal. Biochem. 1976, 72, 248–254.
[44] C. S. Chen, Y. Fujimoto, G. Girdaukas, C. J. Sih, J. Am.
Chem. Soc. 1982, 104, 7294–7299.
À1.8 kcalmolÀ1
. Switching from (S)-1a to (R)-1a
places the p-tolyl group into the hydrophobic leaving
group pocket and removes the sulfoxide oxygen for a
total of ca. 5.2 kcalmolÀ1 in favor of placing the p-tolyl
group in the hydrophobic leaving group pocket when
in an aqueous medium. This estimate neglects the hy-
1192
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1183 – 1192