Steric enhancement of imidazole basicity in cis-urocanic acid derivatives: Models for the action of chymotrypsin

Article abstract of DOI:10.1006/bioo.1998.1115

To test the hypothesis that substrate-induced steric compression between His 57 and Asp 102 at the active site of chymotrypsin can increase the basicity of His 57, we have synthesized the cis- and trans-isomers of 2- bromo-3-(N-tritylimidazole)-2-propenoic acid and 2-chloro-3(N- tritylimidazole)-2-propenoic acid and compared selected properties with those of cis- and trans-urocanic acids. The cis-isomers display low field ¹H NMR signals at 17 ppm in dimethylsulfoxide, similar to cis-urocanic acid; whereas the trans-isomers do not show strong hydrogen bonds. Increasing the size of the C2 substituent (H < Cl < Br) in the cis-isomers increases the pK(a) of the imidazolium group from 6.78 for H to 7.81 and 9.10 for Cl and Br, respectively; whereas the pK(a)s of the trans isomers are all 6.0 ± 0.1. The results indicate that the cis-urocanic acid derivatives with large substituents at C2 act as proton sponges in water, and they support the concept that steric compression in the catalytic triad of chymotrypsin can increase the basicity of His 57.

Full text of DOI:10.1006/bioo.1998.1115

BIOORGANIC CHEMISTRY 26, 323–333 (1998)
ARTICLE NO. BH981115
Steric Enhancement of Imidazole Basicity in cis-Urocanic Acid
Derivatives: Models for the Action of Chymotrypsin
Mary J. Cloninger and Perry A. Frey
Institute for Enzyme Research, The Graduate School, and Department of Biochemistry, College of
Agricultural and Life Sciences, University of Wisconsin-Madison, Madison, Wisconsin 53705
Received March 11, 1998
To test the hypothesis that substrate-induced steric compression between His 57 and Asp
102 at the active site of chymotrypsin can increase the basicity of His 57, we have synthesized
the cis- and trans-isomers of 2-bromo-3-(N-tritylimidazole)-2-propenoic acid and 2-chloro-3-
(N-tritylimidazole)-2-propenoic acid and compared selected properties with those of cis- and
trans-urocanic acids. The cis-isomers display low field 1H NMR signals at 17 ppm in
dimethylsulfoxide, similar to cis-urocanic acid; whereas the trans-isomers do not show strong
hydrogen bonds. Increasing the size of the C2 substituent (H Ͻ Cl Ͻ Br) in the cis-isomers
increases the pKa of the imidazolium group from 6.78 for H to 7.81 and 9.10 for Cl and Br,
respectively; whereas the pKas of the trans isomers are all 6.0 Ϯ 0.1. The results indicate that
the cis-urocanic acid derivatives with large substituents at C2 act as proton sponges in water,
and they support the concept that steric compression in the catalytic triad of chymotrypsin
can increase the basicity of His 57.
1998 Academic Press
A new concept for the mechanism of general acid–base catalysis in serine prote-
ases such as chymotrypsin has recently been proposed (1). The mechanism explains
the function of the low barrier hydrogen bond (LBHB) between His 57 and Asp
102 (2, 3). Chemical, spectroscopic, and X-ray crystallographic evidence indicate
that a substrate-induced conformational change introduces steric compression be-
1
tween Nͳ of His 57 and the ͱ-carboxyl group of Asp 102. Strain is relieved by
2
1
␧
proton transfer from Ser 195 to N of His 57 and LBHB formation between Nͳ
of His 57 and Asp 102. Relief of strain by the LBHB can occur only when His
57 is protonated, so that this mechanism increases the basicity of His 57 and its
effectiveness as a base in abstracating a proton from Ser 195. To test the hypothesis
that steric compression can increase the basicity of imidazole, we have synthesized
derivatives 1–4 of cis- and trans-urocanic acid, with chloro or bromo substituents
at C2 of the side chain, and studied their properties.
323
0045-2068/98 $25.00
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.

324
CLONINGER AND FREY
EXPERIMENTAL
Synthetic Protocols
N-Tritylurocanic acid, dibutylammonium salt (6). Urocanic acid 5 (10.0 g, 72.4
mmol) and trityl chloride (44.4 g, 159.3 mmol, 2.2 eq) were dissolved in 250 ml of
N,N-dimethylformamide (DMF). Et3N (54.0 ml, 39.2 g, 388 mmol, 5.4 eq) was
added and the reaction was stirred for 14 h at room temperature. Methanol (250
ml) was added and the reaction was stirred for 24 h at 65ЊC. The solution was
cooled to room temperature, diluted with 150 ml of ethyl acetate, and washed with
4 ϫ 125 ml of 10% citric acid. The organic layer was condensed to 80 ml of a brown
oil. Dibutylamine (13.0 ml, 10.0 g, 77.1 mmol, 1.1 eq) and 50 ml of ethylacetate
was added. A white precipitate formed over 24 h. The solid was isolated and washed
repeatedly with hexanes until pure by TLC (Rf ϭ 0.8, 70% hexane/ethyl acetate)
1
and dried under vacuum to give 36.7 g (quantitative yield) of white solid 6. H
NMR (300 MHz, CDCl3) 9.17 (br. s, 1H), 7.44 (d, J ϭ 1.0 Hz, 1H), 7.35–7.31 (m,
9H), 7.28 (half of alkene doublet: other half obscured by 9H multiplet, 0.5H),
7.15–7.12 (m, 6H), 6.91 (d, J ϭ 1.0 Hz, 1H), 6.49 (d, J ϭ 15.8 Hz, 1H), 2.80
(t, J ϭ 7.4 Hz, 4H), 1.67 (quintet, J ϭ 7.4 Hz, 4H), 1.32 (sextet, J ϭ 7.4 Hz, 4H),
0.88 (t, J ϭ 7.4 Hz, 6H). 13C NMR (75.5 MHz, CDCl3 , missing 1 quaternary C)
146.9, 143.9, 142.1, 129.7, 128.7, 128.1, 127.9, 127.7, 127.2, 126.9, 81.9, 47.6, 20.2,
13.7. LSIMS calculated for C33N3O2H39 M ϭ 509 not observed; Ph3C ϭ 243.1,
C6N2O2H5 ϭ 136.0, H2N ϩ Bu2 ϭ 130.2.
Ethyl N-tritylurocanate (7). 6 (2.0 g, 3.93 mmol, 1 eq) was dissolved in 100 ml
of CHCl3 . 2,4,6-Trichlorobenzoyl chloride (1.9 ml, 2.92 g, 12.0 mmol, 3 eq) and
Et3N (1.2 ml, 0.87 g, 8.6 mmol, 2.2 eq) were added and the reaction was stirred for
10 min. Ethanol (1.0 ml, 0.80 g, 17.4 mmol, 4.4 eq) and 4-dimethylaminopyridine
(1.27 g, 10.4 mmol, 2.6 eq) were added and the reaction was stirred for 12 h. The
CHCl3 was reduced to about 10 ml in vacuo and the crude reaction mixture was
chromatographed (SiO2 , 70% hexane/ethyl acetate) to give, after removal of solvent,
1
1.08 g (67% yield) of white solid. H NMR (300 MHz, CDCl3) 7.51 (d, J ϭ 15.6
Hz, 1H), 7.47 (s, 1H), 7.37–7.32 (m, 9H), 7.16–7.11 (m, 6H), 6.53 (d, J ϭ 15.6 Hz,
1 H), 4.21 (q, J ϭ 7.4 Hz, 2H), 1.29 (t, J ϭ 7.4 Hz, 3H). 13C NMR (75.5 MHz,
CDCl3) 167.4, 141.9, 140.3, 137.2, 136.1, 129.7, 128.3–128.2 (large signal, 2 trityl
C’s), 123.8, 116.3, 75.7, 60.1, 14.3. LRMS calculated for C27N2O2H24 M ϭ 408; found
M ϭ 408.
Ethyl 2,3-dibromo-3-(N-tritylimidazole) propionate (8). 7 (425 mg, 1.0 mmol)
was dissolved in 30 ml CCl4 and cooled to 0ЊC in an ice bath. Bromine (Br2 , 70 Ȑl,
217 mg, 1.4 mmol, 1.3 eq) was added and the reaction was stirred for 1 h at 0ЊC.
It was then warmed to room temperature and stirred for 1 h. The solvent was
removed in vacuo to give 600 Ȑl of pale yellow solid (quantitative yield). 1H NMR
(200 MHz, CDCl3) 7.62 (br s, 1H), 7.40–7.34 (m, 9H), 7.16–7.11 (m, 6H), 6.92 (d,
J ϭ 1.5 Hz, 1H), 5.38 (d, J ϭ 11.5 Hz, 1H), 5.14 (d, J ϭ 11.5 Hz, 1H), 4.32 (q, J ϭ
6.8 Hz, 2H), 1.34 (t, J ϭ 6.8 Hz, 3H). LSIMS and HRMS and LRMS: dibromide
decomposed rather than fragmenting in all cases. MS of bromoalkenes for which
this compound is a precursor are provided.

MODELS FOR THE ACTION OF CHYMOTRYPSIN
325
Ethyl 2,3-dichloro-3-(N-tritylimidazole) propionate (11, 12). To condense Cl2 ,
a 15Љ syringe needle was attached to a lecture bottle and inserted into an NMR
tube in a dry ice/CH3CN bath. The chlorine gas was bubbled into the tube until
100 Ȑl of gas had been collected. The NMR tube was then rapidly inverted into 5
ml of precooled CHCl3 . Meanwhile, 7 (42 mg, 0.10 mmol) was dissolved in CHCl3
and cooled to Ϫ45ЊC in a dry ice/CH3CN bath. A stock Cl2/CHCl3 solution (480
Ȑl, 1.1 eq) was added and the reaction was stirred for 2 h at Ϫ45ЊC. It was warmed
to room temperature and solvent was removed in vacuo. A TLC (70% hexane/
EtOAc) of the crude reaction product revealed two spots. Chromatography (SiO2 ,
70% hexane/EtOAc) gave 8 mg of pure material (Rf ϭ 0.7), 14 mg of pure material
(Rf ϭ 0.6), and 10 mg of a mixture of the two compounds (65% combined yield).
The two products were assigned as the syn- and anti-diastereomers on the basis of
elimination experiments in refluxing CHCl3 . The higher Rf material gave a mixture
of products while the lower Rf material gave only the trans-urocanate derivative.
Thus, the products result from anti- and syn-addition of Cl2 , respectively. In general,
the diastereomeric mixture was carried on without separation. anti-Addition prod-
1
uct: H NMR (200 MHz, CDCl3) 7.45 (d, J ϭ 1.5 Hz, 1H), 7.37–7.32 (m, 9H),
7.15–7.10 (m, 6H), 6.91 (d, J ϭ 1.5 Hz, 1H), 5.20 (d, J ϭ 10.3 Hz, 1H), 4.89 (d,
J ϭ 10.3 Hz, 1H), 4.31 (q, J ϭ 6.8 Hz, 2H), 1.37 (t, J ϭ 6.8 Hz, 3H). Syn-addition
product: 1H NMR (200 MHz, CDCl3) 7.40 (d, J ϭ 1.5 Hz, 1H), 7.38–7.31 (m, 9H),
7.15–7.08 (m, 6H), 6.92 (br. s), 5.45 (d, J ϭ 6.5 Hz, 1H), 5.00 (d, J ϭ 6.5 Hz, 1H),
4.18 (q, J ϭ 7.5 Hz, 2H), 1.24 (t, J ϭ 7.5 Hz, 3H). Mixture of syn- and anti-dichlorides:
LSIMS calculated for C27N2O2Cl2H24 M ϭ 478; found M ϩ 1 ϭ 479.1.
Ethyl 2-bromo-3-(N-tritylimidazole)-2-propenate (9). E2 reaction conditions: 8
(115 mg, 0.20 mmol) was dissolved in 5.0 ml CH2Cl2 and cooled to Ϫ78ЊC in a dry
ice/acetone bath. DBU (61 Ȑl, 103 mg 0.68 mmol, 3.3 eq) was added and the
reaction was stirred at Ϫ78ЊC for 10 min, until TLC (70% hexane/ethyl acetate)
indicated no starting material remained. The reaction was warmed to room tempera-
ture and solvent was removed in vacuo to give 187 mg of brown oily solid which
was a 20 : 1 mixture of isomers, as determined by NMR. Chromatography (SiO2 ,
1
70% hexane/EtOAc) gave 75 mg of white solid (76% yield). H NMR (300 MHz,
1
CDCl3) 7.55 (br. s, 1H), 7.41 (d, J ϭ 1.1 Hz, 1H), 7.35–7.33 (m, 9H-trityl, H-
alkene), 7.15–7.12 (m, 6H), 4.16 (q, J ϭ 7.2 Hz, 2H), 1.20 (t, J ϭ 7.2 Hz, 3H). 13C
NMR (75.5 MHz, CDCl3) 164.0, 142.0, 138.9, 135.6, 134.8, 129.7, 128.2, 124.6, 107.5,
75.7, 61.9, 13.9. HRMS calculated for C27N2O2BRH23 M ϭ 486.094, 488.092; found
M ϭ 486.094, 488.090.
Ethyl 2-bromo-3-(N-tritylimidazole)-2-propenate (10). ͱ-Elimination reaction
conditions: 8 (100 mg, 0.18 mmol) was dissolved in 2.5 ml CHCl3 . Et3N (25 Ȑl, 18.2
mg. 0.18 mmol, 1 eq) was added and the reaction was refluxed for 14 h. The solvent
was removed in vacuo to give 100 mg of brown solid. Chromatography (SiO2 , 70%
hexane/EtOAc) gave 25 mg of cis bromoalkene 9 and 32 mg of trans bromoalkene
10 (65% yield). 1H NMR (300 MHz, CDCl3) 8.28 (s, 1H), 7.90 (d, J ϭ 0.9 Hz, 1H),
7.49 (d, J ϭ 0.9 Hz, 1H), 7.34–7.30 (m, 9H), 7.15–7.10 (m, 6H), 4.26 (q, J ϭ 7.3
Hz, 2H), 1.30 (t, J ϭ 7.3 Hz, 3H). 13C NMR (75.5 MHz, CDCl3) 163.3, 141.8, 139.3,
135.8, 135.4, 129.5, 128.2, 128.0, 125.6, 110.1, 76.0, 62.3, 14.1. HRMS calculated for

326
CLONINGER AND FREY
C27N2O2BrH23 M Ϫ 486.094, 488.090; found M Ϫ 486.094, 488.090. X-Ray crystal
structure was also obtained.
Ethyl 2-chloro-3-(N-tritylimidazole)-2-propenate (13, 14). A diastereomeric
mixture of dichlorides 11 and 12 (370 mg, 0.77 mmol) was dissolved in 20 ml CH2Cl2
and cooled to Ϫ78ЊC in a dry ice/acetone bath. DBU (231 Ȑl, 235 mg, 1.55 mmol,
2.0 eq) was added and the reaction was stirred for 2 h at Ϫ78ЊC. The reaction was
warmed to room temperature and solvent was removed in vacuo. Chromatography
(70% hexane/EtOAc) gave 65 mg (2% yield) of 13 and 200 mg (59% yield) of 14.
1
(13) H NMR (300 MHz, CDCl3) 7.74 (dd, J ϭ 1.5, 0.8 Hz, 1H), 7.44 (d, J ϭ 1.5
Hz, 1H), 7.37 Ϫ 7.32 (m, 9H), 7.25 (d, J ϭ 0.8 Hz, 1H), 7.17 Ϫ 7.13 (m, 6H), 4.17
(q, J ϭ 7.0 Hz, 2H), 1.20 (t, J ϭ 7.0 Hz, 3H). 13C NMR (75.5 MHz, CDCl3) 163.2,
1
142.0, 138.8, 134.5, 132.8, 129.8, 129.3, 128.4, 125.4, 118.7, 75.8, 61.7, 14.0. (14) H
NMR (300 MHz, CDCl3) 8.00 (s, 1H), 7.76 (d, J ϭ 1.5 Hz, 1H), 7.53 (d, J ϭ 1.5
Hz, 1H), 7.37–7.32 (m, 9H), 7.17–7.14 (m, 6H), 4.28 (q, J ϭ 7.0 Hz, 2H), 1.32
(t, J ϭ 7.0 Hz, 3H). 13C NMR (75.5 MHz, CDCl3) 163.0, 141.7, 139.3, 135.0, 131.8,
129.5, 128.0, 127.9, 125.8, 119.3, 75.9, 62.0, 14.1. HRMS of cis/trans mixture: Calcu-
lated for C27N2O2ClH23 M ϭ 442.1448; Found 442.1433.
Ester Hydrolysis
Esters of the four cis- and trans-urocanates were all synthesized by same proce-
dure, one example of which is given here.
2-Bromo-3-(N-tritylimidazole)-2-propenoic acid (1). 9 (20 mg, 0.04 mmol) was
dissolved in 1.4 ml THF. LiOH и H2O (5 mg, 0.12 mmol, 2.9 eq) was dissolved in
600 Ȑl of H2O and added to the THF solution. The reaction was stirred for 2 days
at room temperature until no more starting material was present (TLC, 70% hexane/
EtOAc). The THF was removed in vacuo and the aqueous solution was acidified
to pH 7.0 using 3 N HCl. White precipitate formed and was filtered and washed
with water. Drying for 24 h on a drying pistol with P2O5 (toluene) gave 18 mg (95%
1
yield) of product. H NMR (500 MHz, d6-DMSO) 16.82 (s, 1H, COOH), 7.77
(s, 1H), 7.64 (s, 1H), 7.37 (m, 10H), 7.09 (m, 6H). 13C NMR (125.8 MHz, d6-DMSO,
1C overlapped) 173.6, 152.4, 149.2, 140.1 (large signal, trityl), 139.4 (large signal,
trityl), 139.2 (large signal, trityl), 136.2, 117.2, 90.2, 86.7. HRMS calculated for
C25N2O2BrH19 M ϭ 459.3492 not observed: M-CO2 ϭ 414.0698 (calc 414.0687,
M-Br-Co2 ϭ 334.1470 (calc 334.1471), Trityl group ϭ 243.1121 (calc 243.1171).
1
(2) H NMR (300 MHz, d6-DMSO) 17.2 (s, 1H, COOH), 7.60 (s, 1H), 7.37
(m, 10H), 7.07 (m, 6H). 13C NMR (125.8 MHz, D2O) 161.6, 141.2, 138.2, 132.2,
129.1, 128.4, 128.3, 127.5, 125.8, 79.1, 75.9.
1
(3) H NMR (500 MHz, d6-DMSO) 7.85 (s, 1H, COOH), 7.71 (s, 1H), 7.48
(s, 1H), 7.41 (m, 10H), 7.10 (m, 6H).
1
1
(4) H NMR (200 MHz, D2O) 7.50 (br. s, 2H), 7.40 (s, 1H), 6.95 (m, 15H). H
NMR (500 MHz, d6-DMSO) 7.50 (s, 1H), 7.42 (m, 10H), 7.12 (m, 6H). (C2H and
COOH exchanged out—not detected). 13C NMR (125.8 MHz, d6-DMSO) 166.0,
150.0, 147.8, 147.8, 143.1, 137.7, 137.5, 136.5, 134.3, 132.8, 83.7.
Evaluation of pKas. The imidazolium pKas of trans-urocanic acid and 1–4 were
evaluated from potentiometric titration data obtained at 25ЊC by use of an iterative

MODELS FOR THE ACTION OF CHYMOTRYPSIN
327
computer program written by W. W. Cleland. For 1, subtraction of pH increases
caused by NaOH addition to H2O was required. For 2–5, this subtraction was
unnecessary as the pKa-values were below the region where NaOH interfered.
RESULTS AND DISCUSSION
Synthesis of cis- and trans-2-halo-3-(N-tritylimidazole)-2-propenoic acids 1–4.
The synthesis of compounds 1–4 is shown in Scheme 1. N-Triphenylmethyl (trityl)
protection of commercially available trans-urocanic acid 5 following the procedure
described by Mutter and Hersperger gave 6 in quantitative yield (4).1 Yamaguchi
esterification via formation of the mixed anhydride using 2,4,6-trichlorobenzoyl
chloride gave ethyl ester 7 in 70% yield (5).2 Bromination of 7 afforded 8 (100%).
Elimination of HBr at low temperature (Ϫ78ЊC) with DBU afforded an 80% yield
of bromoalkenes 9 and 10, with 9 formed as the major product (Ͼ20 : 1).3 At higher
temperatures (refluxing CHCl3) with a weaker base (Et3N) (6), the ͱ-elimination
product 10 was the major product while E2 elimination to give 9 as the minor
product (3 : 2; 65% combined yield). Esters 9 and 10 were separated and indepen-
dently subjected to basic hydrolysis to give products 1 and 3 (95 and 100%;
Scheme 1c).
The chlorides were obtained from 7 in a similar fashion (Scheme 1b). Chlorination
of 7 gave a mixture of syn- and anti-addition products 11 and 12 (65%). Reaction
of this mixture with DBU at Ϫ78ЊC gave a 1 : 3 mixture of 13 and 14 in 80% yield.
E2 elimination of HCl from 11 produced 13, while E2 elimination from 12 and ͱ-
elimination from 11 both produced 14. Esters 13 and 14 were separated and hy-
drolyzed to afford 2 and 4 (both 80%; Scheme 1c).
1
1
Low field signals in the H NMR spectra of 1–4. The H NMR spectra of 1–4
were obtained in DMSO-d4. The chemical shift of the carboxylic acid proton in the
spectra of both the chloro- and the bromo-substituted derivatives of cis-urocanic
acid is about 17 ppm (1, 16.8 ppm; 2, 17.2 ppm in DMSO). Low field signals have
been used as a diagnostic of LBHBs (7), and such a signal was originally reported
1 N-Trityl protection was required as bromination of dimethyl-protected 3 occurred exclusively on
the imidazole ring.
2 Esterifications using dimethyl sulfate or methyl iodide were unsuccessful, presumably due to the
low solubility of 4 in appropriate solvents. The trityl ester was not suitable because exposure of the
derived dibromide to base caused cleavage of the trityl group followed by loss of CO2 :
3 X-ray crystal structures of 7, 8 and 12 were obtained to verify product assignments.
4 In organic solvents such as CDCl3 or d6-DMSO, the pKa of the carboxylic acid will be elevated
relative to the aqueous pKa , while the pKas of positively charged acids such as imidazolium ions will
be similar to their values in water. Thus, a low aqueous pKa for the carboxylic acid should create a
system which has been more highly optimized for LBHB formation. For further explanation, see reference
[8] (and references cited therein).

328
CLONINGER AND FREY
SCHEME 1. (a) The synthesis of 1 and 3. (b) The synthesis of 2 and 4. (c) The last step (1–4).

MODELS FOR THE ACTION OF CHYMOTRYPSIN
329
as present in cis-urocanic acid (17.4 ppm) (2). The corresponding proton in trans-
urocanic acid derivative 3 is at 7.9 ppm. Inasmuch as 7.9 ppm is very close to the
chemical shift of the acidic proton in 1-methylimidazolium p-toluenesulfonate in
CDCl3 (7 ppm) (8), 3 exists as the zwitterion. Accordingly, 3 served as a control,
in which very little (if any) strong hydrogen bonding occurs. In 1 and 2, the hydrogen
bond between imidazole and the carboxylic acid is expected to be stronger, and
the 9 ppm difference (1 and 2 vs 3) can be attributed to the difference between a
strongly hydrogen bonded ion-pair and a zwitterionic structure.
The halogen substituents C2 of 1–4 are acid strengthening and can therefore be
expected to alter the strength of the hydrogen bond between the imidazole and
carboxylic acid groups of 1 and 2. The hydrogen bond is not optimal in the case
of cis-urocanic acid (17.4 ppm in DMSO), perhaps because the pKa of the carboxylic
acid group is 3, whereas the optimal pKa may be somewhat lower (8). We thought
that the chloro- or bromo-substituents on C2 might lower the pKa enough to
strengthen the hydrogen bond; however, the chemical shift values did not indicate
a stronger bond. The plot of chemical shift for the hydrogen bond against aqueous
pKa for analogous 1-methylimidazole/carboxylic acid complexes showed a maxi-
mum at pKa ϭ 2.1, with a positive slope at low pKas and a negative slope at higher
pKas (8). cis-Urocanic acid falls on the negative slope of this plot. If the chloro-
and bromo-substituents shift the pKa lower than 2.1, the values for 1 and 2 may
fall on the positive slope of the plot and coincidentally display proton chemical
shifts similar to cis-urocanic acid. The pKas of fumarate (3.02) and 2-bromofumarate
(1.7) and 2-chlorofumarate (1.7) indicate that halogen substituents on C2 signifi-
cantly increase the acidity of unsaturated acids (9), perhaps too much to optimize
the hydrogen bonds in 1 and 2. In addition, it is not clear that the structures of
these molecules allow the formation of linear LBHBs.
Basicities of the imidazole rings in cis- and trans-urocanic acid derivatives. In
the catalytic triad of the chymotrypsin active site, it has been proposed that His 57
abstracts a proton from Ser 195. Inasmuch as histidine is not normally basic enough
(pKa ϭ 6) to be optimally effective in deprotonating serine, the pKa of which is
Ͼ13 (10), an LBHB between His 57 and Asp 102 is thought to elevate the pKa of
His 57 (1). With this in mind, we obtained the imidazolium-pKas of our model com-
pounds.
The pKas of the imidazolium ring protons of compounds 1–4 and trans-urocanic
acid were determined by potentiometric titration with the results shown in Table
1. In the trans-urocanic acid series, the results show that the bromo- and chloro-
substituents have essentially no effect on the pKa-values, all of which are 6.0 Ϯ 0.1
(3, 4, and 5). Therefore, the inductive effects of the chloro- and bromo-substituents
in 3 and 4 are not base-weakening in this system. The presence of the trityl group
also has no significant effect on the pKa values in the trans series (3 and 4 vs 5).
Interesting results were obtained from the substituted cis-urocanic acids. The
pKa for cis-urocanic acid is 0.85 higher than that for the trans-urocanic acid 5,
presumably because of the proximity of the imidazole and carboxyl groups. The
pKa of chloro-substituted 2 is 1 unit higher and that of bromo-substituted 1 is nearly
2.5 units higher than that of cis-urocanic acid. The effect of bromine at C2 represents
a 3.2 kcal/mol difference in the ionization free energy between 1 and cis-urocanic

330
CLONINGER AND FREY
TABLE 1
Imidazole Ring pKas of cis- and trans-Urocanic Acids and Halogenated Derivatives
Compound
pKa
Compound
pKa
cis-Urocanic acid
cis-2-Chloro-3-(N-tritylimidazole)-
propenoic acid (2)
6.78a
7.81
trans-Urocanic acid (5)
trans-2-Chloro-3-(N-tritylimidazole)-
propenoic acid (4)
5.92b
6.02
cis-2-Bromo-3-(N-tritylimidazole)-
propenoic acid (1)
9.10
trans-2-Bromo-3-(N-tritylimidazole)-
propenoic acid (3)
5.99
a Value from Ref. 11.
b Ref. 11 value ϭ 5.89 Ϯ 0.03.
acid and a 4.3 kcal/mol difference between 1 and 3. That the molecules have
different solvation energies might account for part of the observed change in pKa ,
but it would not be expected to account for most of the difference.
The substituent-induced pKa elevations in 1 and 2 (relative to cis-urocanic acid)
are not caused by an electronic effect. Inductive electron withdrawal by halogens
would lower the basicity of the imidazole ring, but pKa-elevation is observed.
Furthermore, any electronic effect would be observed in the trans-isomers, all of
which display essentially the same imidazole pKa .
It appears that increased basicity in 1 and 2 relative to cis-urocanic acid is brought
about by the steric requirements of Cl and Br as substituents at carbon-2. The
atomic radii of chlorine (99 pm) and bromine (114 pm) are significantly larger than
that of hydrogen (32 pm). The halogens occupy considerably more space than
hydrogen, and this must force 1 and 2 to be in slightly different conformations than
that of cis-urocanic acid. Bromine is the larger substituent and causes the greater
increase in basicity. A nearer proximity of the carboxylic acid and the imidazole
ring, which shortens the internal hydrogen bond, is the change that would most
obviously account for the increased basicity of the imidazole ring with increased
substituent size.
Steric compression and low-barrier hydrogen bonding in chymotrypsin. The
LBHB-facilitated general base catalysis mechanism recently postulated for the ac-
tion of chymotrypsin in cleaving peptide bonds is illustrated in Fig. 1 (1, 16). The
mechanism was put forward to explain how His 57 can be basic enough to be an
optimally effective Bronsted acid–base catalyst in the acylation of chymotrypsin.
An optimal catalyst would be basic enough to abstract the proton from Ser 195 in
the transition state for tetrahedral adduct formation, and its conjugate acid would
be acidic enough to protonate the leaving group, the N-terminal amino group of a
peptide in the case of chymotrypsin. Thus, an ideal catalytic group would display
a pKa value between that of the 3-OH group of serine (about 13) and the N-terminal
ammonium group of a peptide (about 9). His 57 in peptidyl trifluoromethylketone
adducts of chymotrypsin displays pKa values between 10.6 and 12.1, depending on
the structure of the peptidyl moiety (1, 16). These values are in the range for effective
catalysis of acylation. Peptidyl trifluoromethylketone adducts of chymotrypsin are

MODELS FOR THE ACTION OF CHYMOTRYPSIN
331
FIG. 1. The mechanism of LBHB-facilitated general base catalysis postulated for the acylation of
chymotrypsin. In this mechanism, remote binding interactions between the peptidyl moiety and the
active center are postulated to induce a conformational change in the enzyme that results in steric
1
compression between His 57-Nͳ and Asp 102-OͲ. Protonation of His 57-Ne2 relieves the strain by
1
allowing LBHB-formation between His-57-Nͳ and Asp 102-OͲ, whose acidities can be matched only
when His 57 is in its conjugate acid form. Thus, the potential for LBHB-formation increases the basicity
of His 57 in the Michaelis complex, where there is compression, but not in the free enzyme, where
there is no steric compression. The deacylation mechanism should be in many respects analogous to
acylation, with water reacting in place of the leaving group for acylation.
regarded as close analogs of the transition state. Therefore, His 57 appears to
display nearly ideal acid–base properties for its function in catalysis.
It has been postulated that LBHB-formation between His 57 and Asp 102 facili-
tates general base catalysis by His 57 through stabilization of tetrahedral intermedi-
ates such as that shown in Fig. 1. The mechanism shown accounts for this effect as
follows. The binding of a substrate induces a conformational change in the enzyme
that leads to steric compression between His 57 and Asp 102. The compression
1
between His 57-Nͳ and Asp 102-OͲ would be relieved by LBHB-formation because
an LBHB is intrinsically short and would relieve the strain. However, LBHB-
2
␧
formation is possible only when His 57 is protonated on N . The strain of compres-
sion between His 57 and Asp 102 therefore potentiates the protonation of His
2
␧
57-N ; that is, it increases the basicity of His 57. The action of His 57 as a base is
directed toward Ser 195 and facilitates its nucleophilic addition to the peptide
carbonyl group. The resulting tetrahedral intermediate is stabilized by all of the
binding interactions between it and the active site, in addition to the stabilization
provided by the LBHB. The LBHB stabilizes the complex of the tetrahedral inter-
mediate with the active site, but the stabilization is not so great as to prevent
His 57 from protonating the leaving group, as shown by its pKa in the peptidyl
trifluoromethylketone adducts (1, 16), which are structurally similar to the tetrahe-
dral intermediate.
LBHB-formation can be observed in free chymotrypsin at low pHs (2, 17, 18),
but the LBHB does not lead to high basicity for His 57 in resting chymotrypsin.
This is because increased basicity is brought about by compression between His-
1
Nͳ and Asp 102-OͲ, which occurs only in the Michaelis complex when a specific
substrate is bound (1). Thus, by the mechanism in Fig. 1, the basicity of His 57 is

332
CLONINGER AND FREY
increased only when a specific substrate is bound at the active site. Correlations of
the basicity of His 57 and the strength of the LBHB with the structures of peptidyl
trifluoromethylketones and substrates indicate that the best substrates induce the
formation of the strongest LBHBs (16).
Conclusion. An efficient synthesis of bromo- and chloro-substituted derivatives
of cis- and trans-urocanic acids has enabled the discovery that small steric changes
can have dramatic effects on the basicity of the imidazole ring in the cis- but not
the trans-isomers of urocanic acid derivatives. The results discussed above indicate
that the halo substituents of 1 and 2 cause a change in geometrical structure,
which makes the internally hydrogen-bonded proton much more difficult to remove
relative to that in trans-urocanic acid. The simplest explanation is that the internal
hydrogen bond is shortened and thereby strengthened. Shortening and strengthen-
ing of hydrogen bonds has been suggested to explain certain aspects of enzymatic
mechanisms (1, 2, 12–15). Alternatively, the steric compression brought about by
the halo substituents may increase the probability of internal hydrogen bonding,
which stabilizes the protonated imidazole ring and increases the value of its pKa .
In either case, the steric effect increases the basicity of the imidazole ring in 1, 2,
and cis-urocanate relative to the corresponding trans-isomers. Compression in the
2-halo-cis-urocanic acids may serve as a chemical counterpart and model for com-
pression between the imidazole ring of His 57 and the carboxylic acid group of
Asp 102 of chymotrypsin, which has been postulated to strengthen the LBHB of
the incipient tetrahedral intermediate at the transition state for acylation of Ser
195 (1, 16). A similar stabilization of the transition state for deacylation of the acyl-
chymotrypsin intermediate may also make a contribution to catalysis.
ACKNOWLEDGMENTS
This research was supported by Grant No. GM 51806 from the National Institute of General Medical
Sciences. We thank D. R. Powell for X-ray crystal structures of compounds 9, 10, and 14.
REFERENCES
1. Cassidy, C. S., Lin, J., and Frey, P. A. (1997) Biochemistry 36, 4576–4584.
2. Frey, P. A., Whitt, S. A., and Tobin, J. B. (1994) Science 264, 1927–1930.
3. Golubev, N. S., Denisov, G. S., Gindin, V. A., Ligay, S. S., Limbach, H. H., and Smirnov, S. N.
(1994) J. Mol. Struct. 322, 83.
4. Mutter, M., and Hersperger, R. (1989) Synthesis, 198.
5. Inanaga, J., Hirata, K., Saeki, H., Katsuki, T. and Yamaguchi, M. (1979) Bull. Chem. Soc. Japan
52, 1989.
6. Marsura, A., Luu-Duc, C. and Gellon, G. (1985) Synthesis 537.
7. Hibbert, F., and Emsley, J. (1990) Adv. Phys. Org. Chem. 26, 255–379.
8. Tobin, J. B., Whitt, S. A., Cassidy, C. S., and Frey, P. A. (1995) Biochemistry 34, 6919–6924.
9. Jencks, W. P., and Regenstein, J. (1970) Handbook of Biochemistry and Molecular Biology (H. A.
Sober, Ed.), p. J186, Chem. Rub. Co., Cleveland.
10. Bruice, T. C., Fife, T. H., Bruno, J. J., and Brandon, N. E. (1962) Biochemistry 1, 7.
11. Zimmerman, S. C., Korthals, J. S., and Cramer, K. D. Tetrahedron 1991, 47, 2649.

MODELS FOR THE ACTION OF CHYMOTRYPSIN
12. Cleland, W. W. (1992) Biochemistry 31, 317–319.
333
13. Cleland, W. W., and Kreevoy, M. M. (1994) Science 264, 1887–1890.
14. Gerlt, J. A., and Gassman, P. G. (1993) Biochemistry 32, 11943–11952.
15. Gerlt, J. A., Kreevoy, M. M., Cleland, W. W., and Frey, P. A. (1997) Chem. Biol. 4, 259–267.
16. Lin, J., Cassidy, C. S., and Frey, P. A. (1998) Biochemistry 37, 11940–11948.
17. Markley, J. L. (1978) Biochemistry 17, 4648–4656.
18. Robillard, G., and Shulman, R. G. (1972) J. Mol. Biol. 71, 507–511.