Elucidation of the mechanism of polysaccharide cleavage by chondroitin AC lyase from Flavobacterium heparinum

Article abstract of DOI:10.1021/ja020627c

Chondroitin AC lyase from Flavobacterium heparinum degrades chondroitin sulfate glycosaminoglycans via an elimination mechanism resulting in disaccharides or oligosaccharides with Δ4,5-unsaturated uronic acid residues at their nonreducing end. Mechanistic details concerning the ordering of the bond-breaking and -forming steps of this enzymatic reaction are nonexistent, mainly due to the inhomogeneous nature of the polymeric substrates. The creation of a new class of synthetic substrates for this enzyme has allowed the measurement of defined and reproducible k_cat and K_m values and has expanded the range of mechanistic studies that can be performed. The primary deuterium kinetic isotope effect upon k_cat/K_m for the abstraction of the proton α to the carboxylic acid was measured to be 1.67 ± 0.07, showing that deprotonation occurs in a rate-limiting step. Using substrates with leaving groups of differing reactivity, a flat linear free energy relationship was produced, indicating that the C4-O4 bond is not broken in a rate-determining step. Taken together, these results strongly suggest a stepwise mechanism. Consistent with this was the measurement of a secondary deuterium kinetic isotope effect upon k_cat/K_m of 1.01 ± 0.03 on a 4-{²H}-substrate, indicating that no sp² character is developed at C4 during the rate-limiting step, thereby ruling out a concerted syn-elimination.

Full text of DOI:10.1021/ja020627c

Published on Web 07/25/2002
Elucidation of the Mechanism of Polysaccharide Cleavage by
Chondroitin AC Lyase from Flavobacterium heparinum
Carl S. Rye and Stephen G. Withers*
Contribution from the Department of Chemistry, UniVersity of British Columbia,
2036 Main Mall, VancouVer, British Columbia, Canada, V6T 1Z1
Received May 1, 2002
Abstract: Chondroitin AC lyase from Flavobacterium heparinum degrades chondroitin sulfate glycosami-
noglycans via an elimination mechanism resulting in disaccharides or oligosaccharides with ∆4,5-unsaturated
uronic acid residues at their nonreducing end. Mechanistic details concerning the ordering of the bond-
breaking and -forming steps of this enzymatic reaction are nonexistent, mainly due to the inhomogeneous
nature of the polymeric substrates. The creation of a new class of synthetic substrates for this enzyme has
allowed the measurement of defined and reproducible kcat and Km values and has expanded the range of
mechanistic studies that can be performed. The primary deuterium kinetic isotope effect upon kcat/Km for
the abstraction of the proton R to the carboxylic acid was measured to be 1.67 ( 0.07, showing that
deprotonation occurs in a rate-limiting step. Using substrates with leaving groups of differing reactivity, a
flat linear free energy relationship was produced, indicating that the C4-O4 bond is not broken in a rate-
determining step. Taken together, these results strongly suggest a stepwise mechanism. Consistent with
this was the measurement of a secondary deuterium kinetic isotope effect upon kcat/Km of 1.01 ( 0.03 on
a 4-{2H}-substrate, indicating that no sp2 character is developed at C4 during the rate-limiting step, thereby
ruling out a concerted syn-elimination.
Introduction
GAG degradation is effected via two classes of enzymes: (i)
polysaccharide hydrolases (eukaryotic) and (ii) lyases (prokary-
otic). The hydrolases may be either inverting or retaining
glycosidases, whereas the lyases act via an elimination mech-
anism, resulting in disaccharides or oligosaccharides with ∆4,5-
unsaturated uronic acid residues at their nonreducing end (Figure
1). Although the substrate specificity and product compositions
for a number of lyases have been characterized,1,3-7 little is
known about the lyase mechanism at the molecular level. X-ray
crystal structures have been solved for several different types
of polysaccharide lyases including chondroitin AC lyase,8
chondroitin B lyase,9 hyaluronate lyase,10 alginate lyase,11 and
pectate lyase.12 Other representative structures of these lyases
can be found on the Web at http://afmb.cnrs-mrs.fr/∼cazy/
CAZY/index.html. Putative catalytic residues of these polysac-
charide lyases have been suggested by site-directed mutagenesis
studies, often in combination with X-ray structures of various
enzyme-oligosaccharide complexes.9,10,13-19 A wide variety of
Glycosaminoglycans (GAGs) are highly sulfated, unbranched
polysaccharide chains built of a sequence of repeating disac-
charide units consisting of hexosamine and uronic acid residues.
These markedly heterogeneous polymers play a variety of
important roles intracellularly (usually in secretory granules),
at the cell surface, or, more commonly, in the extracellular
matrix. These biological roles include simple mechanical
support, the lubrication and cushioning of joints, the modulation
of cell signals, cell adhesion, motility and proliferation, and
acting as reservoirs for the specific binding of proteins in order
to regulate or stabilize their activity.1,2 There are four main
classes of GAGs: chondroitin sulfate and dermatan sulfate;
hyaluronic acid; heparin and heparan sulfate; and keratan sulfate.
With the exception of hyaluronic acid, all GAGs are linked
covalently to a protein core, forming a proteoglycan. Of these
proteoglycans, chondroitin sulfates are the most common type
of GAG chain, consisting of three major classes: chondroitin
sulfate A (chondroitin 4-sulfate), chondroitin sulfate B (dermatan
sulfate), and chondroitin sulfate C (chondroitin 6-sulfate). The
chondroitin sulfates consist of an N-acetyl-D-galactosamine
residue (usually O-sulfated at C4 or C6) attached through a
â-(1,4) linkage to D-glucuronic acid or L-iduronic acid (dermatan
sulfate) that is in turn attached via a â-(1,3) linkage (R-(1,3)
for dermatan sulfate) to the next hexosamine residue.
(3) Gu, K.; Linhardt, R. J.; Laliberte, M.; Zimmermann, J. Biochem. J. 1995,
312, 569-577.
(4) Jandik, K. A.; Gu, K.; Linhardt, R. J. Glycobiology 1994, 4, 289-296.
(5) Michelacci, Y. M.; Dietrich, C. P. Biochim. Biophys. Acta 1976, 451, 436-
443.
(6) Baker, J. R.; Pritchard, D. G. Biochem. J. 2000, 348, 465-471.
(7) Sutherland, I. W. FEMS Microbiol. ReV. 1995, 16, 323-347.
(8) Fethiere, J.; Eggimann, B.; Cygler, M. J. Mol. Biol. 1999, 288, 635-647.
(9) Huang, W. J.; Matte, A.; Li, Y. G.; Kim, Y. S.; Linhardt, R. J.; Su, H. S.;
Cygler, M. J. Mol. Biol. 1999, 294, 1257-1269.
* To whom correspondence should be addressed. Telephone: (604) 822-
3402. Fax: (604) 822-8869. E-mail: withers@chem.ubc.ca.
(1) Ernst, S.; Langer, R.; Cooney, C. L.; Sasisekharan, R. Crit. ReV. Biochem.
Mol. Biol. 1995, 30, 387-444.
(10) Li, S.; Kelly, S. J.; Lamani, E.; Ferraroni, M.; Jedrzejas, M. J. EMBO J.
2000, 19, 1228-1240.
(11) Yoon, H. J.; Mikami, B.; Hashimoto, W.; Murata, K. J. Mol. Biol. 1999,
290, 505-514.
(2) Kjellen, L.; Lindahl, U. Annu. ReV. Biochem. 1991, 60, 443-475.
(12) Yoder, M. D.; Jurnak, F. Plant Physiol. 1995, 107, 349-364.
9
9756
J. AM. CHEM. SOC. 2002, 124, 9756-9767
10.1021/ja020627c CCC: $22.00 © 2002 American Chemical Society

Chondroitin AC Lyase
A R T I C L E S
Figure 1. Proposed elimination mechanism of polysaccharide lyases.
Scheme 1a
general-base residues responsible for the abstraction of the
relatively acidic H5 proton of the uronic acid moiety have been
implicated from these studies including histidine, tyrosine,
glutamate, and cysteine. General-acid catalysis also plays a
major role in the lyase mechanism by neutralizing the negative
charge on the carboxylate as well as the developing charge of
the enolic intermediate and also by assisting the departing
leaving group saccharide moiety. Divalent metal ions have often
been found to play a part in the first two roles; however, not all
polysaccharide lyases (including chondroitin AC lyase) have
been shown to possess this cationic metal center at the active
site. Consequently, a variety of amino acids have also been
implicated in these roles, such as lysine, arginine, histidine,
tyrosine, and aspartic acid.
Aside from these structural details and implications of
important catalytic residues, no work has been published on
deciphering the chemical mechanism of the polysaccharide lyase
reaction. The inhomogeneous nature of the polymeric substrates
acted upon by these enzymes is a major reason for the lack of
mechanistic details. The introduction of defined chromogenic
substrates for glycosidases not only considerably simplified the
assay of these enzymes but also permitted a variety of
mechanistic studies via the accurate measurement of kinetic
isotope effects and linear free energy relationships produced
by measuring the rates of hydrolysis of a series of substrates of
differing reactivity. Through the design of defined substrates
for chondroitin AC lyase, the equivalent analyses of this
polysaccharide lyase are now possible.
Chondroitin AC lyase from the Gram-negative soil bacterium
FlaVobacterium heparinum (EC 4.2.2.5) cleaves GAG chains
in a random endolytic fashion. It cleaves a variety of GAGs,
including chondroitin sulfates A (chondroitin 4-sulfate) and C
(chondroitin 6-sulfate), as well as the unsulfated chondroitin,
and hyaluronic acid. The proposed mechanism, suggested by
Gacesa,20 is composed of three steps: (i) neutralization of the
negative charge on the carboxylate anion of the glucuronic acid
moiety, either by a divalent metal ion or a positively charged
amino acid; (ii) general-base-catalyzed abstraction of the C5
proton and finally; (iii) the â-elimination of the 4-O-glycosidic
bond. The means of lowering the pKa of the protons adjacent
a Conditions: (i) HBr/AcOH, CH2Cl2, 0 °C f room temperature. (ii)
Phenol, TBAHS, 1 M NaOH, CH2Cl2. (iii) NaOMe, MeOH. (iv) Benzyl
alcohol, Ag2CO3, I2, CH2Cl2. (v) Benzoyl chloride, pyridine, -40 °C f
room temperature. (vi) Methyl DAST, CH2Cl2, -30 °C f room temper-
ature. (vii) NaOMe, MeOH. (viii) For 17 and 16: TEMPO, NaOCl, NaBr,
TBAB, NaHCO3, EtOAc/H2O, 0 °C. For 15: TEMPO, t-BuOCl, H2O.
to acid moieties (pKa ) 22) so that they can be abstracted by
an active site base (pKa ∼ 7) is a matter of some contention.21-26
However, the intent of this report on chondroitin AC lyase is
not to determine the specific mode of stabilization of the
proposed enolic intermediate but to decipher the order and
importance to catalysis of the bond breaking and forming steps.
Results
I. Synthesis of the Fluorinated Substrates. Phenyl 4-Deoxy-
4-fluoro-â-D-glucopyranosiduronic Acid (15), Benzyl 4-Deoxy-
4-fluoro-â-D-glucopyranosiduronic Acid (16), and Methyl
4-Deoxy-4-fluoro-â-D-glucopyranosiduronic Acid (17) (Scheme
1). Three 4-fluoro monosaccharide substrates were prepared,
differing only in their anomeric substituent. Phenyl â-D-
galactopyranoside (4) was prepared in a fashion similar to that
reported in the literature.27 Briefly, this consisted of making
the R-bromide 127 from per-O-acetylated galactose, followed
by a modified Ko¨nigs-Knorr reaction in a biphasic system of
(13) Yoon, H. J.; Choi, Y. J.; Miyake, O.; Hashimoto, W.; Murata, K.; Mikami,
B. J. Microbiol. Biotechnol. 2001, 11, 118-123.
(14) Huang, W.; Boju, L.; Tkalec, L.; Su, H.; Yang, H. O.; Gunay, N. S.;
Linhardt, R. J.; Kim, Y. S.; Matte, A.; Cygler, M. Biochemistry 2001, 40,
2359-2372.
(15) Ponnuraj, K.; Jedrzejas, M. J. J. Mol. Biol. 2000, 299, 885-95.
(16) Yoon, H. J.; Hashimoto, W.; Miyake, O.; Murata, K.; Mikami, B. J. Mol.
Biol. 2001, 307, 9-16.
(21) Gerlt, J. A.; Gassman, P. G. J. Am. Chem. Soc. 1993, 115, 11552-11568.
(22) Gerlt, J. A.; Gassman, P. G. J. Am. Chem. Soc. 1992, 114, 5928-5934.
(23) Gerlt, J. A.; Kozarich, J. W.; Kenyon, G. L.; Gassman, P. G. J. Am. Chem.
Soc. 1991, 113, 9667-9669.
(17) Pritchard, D. G.; Trent, J. O.; Li, X.; Zhang, P.; Egan, M. L.; Baker, J. R.
Proteins 2000, 40, 126-134.
(18) Kelly, S. J.; Taylor, K. B.; Li, S. L.; Jedrzejas, M. J. Glycobiology 2001,
11, 297-304.
(24) Cleland, W. W.; Frey, P. A.; Gerlt, J. A. J. Biol. Chem. 1998, 273, 25529-
25532.
(19) Shriver, Z.; Hu, Y.; Pojasek, K.; Sasisekharan, R. J. Biol. Chem. 1998,
273, 22904-22912.
(25) Guthrie, J. P.; Kluger, R. J. Am. Chem. Soc. 1993, 115, 11569-11572.
(26) Guthrie, J. P. Chem. Biol. 1996, 3, 163-170.
(20) Gacesa, P. FEBS Lett. 1987, 212, 199-202.
(27) Sinnott, M. L.; Souchard, I. J. L. Biochem. J. 1973, 133, 99-104.
9
J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002 9757

A R T I C L E S
Rye and Withers
Scheme 2a
Scheme 3a
a Conditions: (i) Methyl DAST, CH2Cl2, -30 °C f room temperature.
(ii) HBr/AcOH, CH2Cl2, 0 °C f room temperature.
CH2Cl2 and 1 M NaOH with tetrabutylammonium hydrogen-
sulfate (TBAHS) as a phase-transfer catalyst to introduce the
phenyl aglycon 2, and finally a Zemple´n deprotection of the
acetates (39% from per-O-acetylated galactose). Benzyl â-D-
galactopyranoside (5) was synthesized in a similar sequence but
utilized benzyl alcohol and Ag2CO3 in the Ko¨nigs-Knorr
reaction (79% from per-O-acetylated galactose). The methyl
â-D-galactopyranoside was purchased from a commercial source.
Selective benzoylation28 of the galactosides at -40 °C afforded
629 (28%), 730 (40%), and 828 (52%), as white solids. Over-
benzoylated products could be deprotected and rebenzoylated
to give additional material. The 4-fluoro substituent was then
easily introduced with methyl DAST in CH2Cl2 to give 9, 11,
and 13,31in yields of 83%, 60%, and 54%, respectively. After
Zemple´n deprotection of the benzoates, the primary hydroxyl
was selectively oxidized to the carboxylic acid using 2,2,6,6-
tetramethyl-1-piperidinyloxy (TEMPO).32 The oxidation of the
benzyl and methyl glucosides proceeded smoothly using NaOCl
as the primary oxidant, giving 16 (93%) and 17 (74%) as white
solids. However, under the same conditions, the phenyl glyco-
side (10) underwent an undesired side reaction, resulting in the
para chlorination of the aromatic ring, and giving an inseparable
mixture of products. This was overcome by using tert-butyl
hypochlorite (t-BuOCl)33,34 as the primary oxidant, to give 15
as a white solid in a yield of 40%.
a Conditions: (i) Acetyl chloride. (ii) Benzyl alcohol, Ag2CO3, I2, CH2Cl2.
(iii) NaOMe, MeOH. (iv) Pivaloyl chloride, pyridine, CH2Cl2, 0 °C. (v)
Tf2O, pyridine, -15 °C. (vi) H2O, 90 °C. (vii) NaOMe, MeOH. (viii) 2,2-
Dimethoxypropane, p-TsOH, then acetic acid, MeOH:H2O (10:1), 45 °C.
(ix) Ac2O, pyridine. (x) Acetic acid:H2O (4:1), 50 °C. (xi) tert-Butyldi-
methylsilyl chloride, imidazole, DMF, 80 °C. (xii) Ac2O, pyridine, 4-(di-
methylamino)pyridine. (xiii) TBAF, THF.
employed (Scheme 3) that involved many more steps. However,
a lot of the steps were high yielding and the products easily
crystallized, thus eliminating the need for extensive chroma-
tography. Due to the enormous cost difference between glu-
cosamine and galactosamine sugars, the synthesis of the acceptor
molecule started with the less expensive gluco epimer and
involved a conversion to the desired galacto epimer as part of
the synthesis. 2-Acetamido-3,4,6-tri-O-acetyl-2-deoxy-R-D-glu-
copyranosyl chloride (20) was synthesized according to the
literature procedure37 and then reacted with benzyl alcohol and
Ag2CO3 in a Ko¨nigs-Knorr38 reaction to produce 21,39 followed
by a Zemple´n deprotection with sodium methoxide, affording
2239 as a white solid. The conversion of the glucopyranoside
derivative 22 into the galactopyranoside derivative 24 proceeded
smoothly as described in the literature.40 This involved the
selective protection of the 3- and 6-hydroxyls as the pivaloyl
esters to give 23.40 A triflate was then introduced at the
remaining 4-hydroxyl group, which was subsequently displaced
by water at 90 °C to give the product of inverted stereochem-
istry, followed by the deprotection of the pivaloyl esters with
sodium methoxide in methanol to afford the desired benzyl
2-acetamido-2-deoxy-â-D-galactopyranoside (24)40 as a white
solid. The protection of the 3- and 4-hydroxyls with an
isopropylidene group to give 25 proceeded smoothly as de-
scribed in the literature,40 followed by the protection of the
remaining 6-hydroxyl with an acetate to give 26 in an overall
60% yield from 20 (eight steps). Removal of the isopropylidene
moiety with aqueous acetic acid at 50 °C (80%) followed by
the selective protection of the more reactive equatorial hydroxyl
with tert-butyldimethylsilyl chloride and imidazole in DMF at
80 °C gave crystalline 28 (79%). Acetylation of the remaining
Benzyl O-(4-Deoxy-4-fluoro-â-D-glucopyranosiduronyl)-
(1f4)-2-acetamido-2-deoxy-â-D-galactopyranoside (34) (Schemes
2-4). 1,2,3,6-Tetra-O-benzoyl-R-D-galactopyranose was syn-
thesized according to the literature procedure.35,36 The 4-fluoro
substituent was then easily introduced with methyl DAST in
CH2Cl2 to give 18 as a white solid (63%), followed by treatment
with HBr in acetic acid to give the R-bromide donor 19 as a
white solid in a yield of 72% (Scheme 2).
The obvious protection strategy to provide an acceptor
N-acetylhexosamine sugar with the 3-hydroxyl free is through
the one-step benzylidenation of C4 and C6. However, the
glycosylation of this compound with the donor 19 proved to be
problematic due to the low solubility of the acceptor and the
resulting cleavage of the p-methoxybenzylidene ring under the
acidic coupling conditions (AgOTf), affording several products,
including trisaccharides. An alternative protection strategy was
(28) Garegg, P. J.; Oscarson, S. Carbohydr. Res. 1985, 137, 270-275.
(29) Wollwage, P. C.; Seib, P. A. J. Chem. Soc. 1971, 3143-3155.
(30) Leung, O. T.; Douglas, S. P.; Whitfield, D. M.; Pang, H. Y. S.; Krepinsky,
J. J. New. J. Chem. 1994, 18, 349-363.
(31) Petrakova, E.; Yeh, H. J. C.; Kovac, P.; Glaudemans, C. P. J. J. Carbohydr.
Chem. 1992, 11, 407-412.
(32) Davis, N. J.; Flitsch, S. L. Tetrahedron Lett. 1993, 34, 1181-1184.
(33) Mintz, M. J.; Walling, C. In Organic Syntheses; Noland, W. E., Ed.; John
Wiley & Sons: New York, 1973; Vol. 5, pp 184-187.
(37) Horton, D. In Methods in Carbohydrate Chemistry; Whistler, R. L.,
BeMiller, J. N., Eds.; Academic Press: New York, 1972; Vol. VI, pp 282-
285.
(38) Ko¨nigs, W.; Knorr, E. Ber. 1901, 34, 957.
(39) Gross, P. H.; Jeanloz, R. W. J. Org. Chem. 1967, 32, 2759-2763.
(40) Rochepeau-Jobron, L.; Jacquinet, J. C. Carbohydr. Res. 1998, 305, 181-
191.
(34) Melvin, F.; McNeill, A.; Henderson, P. J. F.; Herbert, R. B. Tetrahedron
Lett. 1999, 40, 1201-1202.
(35) Garegg, P. J.; Hultberg, H. Carbohydr. Res. 1982, 110, 261-266.
(36) Kovac, P.; Taylor, R. B. Carbohydr. Res. 1987, 167, 153-173.
9
9758 J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002

Chondroitin AC Lyase
A R T I C L E S
Scheme 4a
a Conditions: (i) AgOTf, CH2Cl2. (ii) NaOMe, MeOH. (iii) p-Anisaldehyde dimethyl acetal, p-TsOH, DMF, 50 °C. (iv) TEMPO, NaOCl, NaBr, TBAB,
NaHCO3, EtOAc/H2O, 0 °C. (v) Acetic acid, MeOH.
Scheme 5a
a Conditions: (i) HBr/AcOH, CH2Cl2, 0 °C f room temperature. (ii) Benzyl alcohol, Ag2CO3, I2, CH2Cl2. (iii) NaOMe, MeOH. (iv) p-anisaldehyde
dimethyl acetal, camphorsulfonic acid, CHCl3, reflux. (v) Ac2O, pyridine. (vi) NaCNBH3, TFA, THF/CH2Cl2. (vii) Tf2O, pyridine, -15 °C f room temperature.
(viii) TBAF, THF. (ix) Acetyl chloride, MeOH, 4 °C. (x) TEMPO, NaOCl, NaBr, TBAB, NaHCO3, EtOAc/H2O, 0 °C.
4-hydroxyl group required a catalytic amount of 4-(dimethyl-
amino)pyridine (DMAP) in the reaction mixture to afford 29
as fine white crystals in a yield of 95% The silyl protecting
group was removed with tetrabutylammonium fluoride (TBAF)
in THF to give the acceptor 30 as a white crystalline solid in a
moderate 49% yield. The yield of this final step was low due
to the partial migration of the C4 acetate to C3.
benzyl alcohol and Ag2CO3 to give 36, and finally a Zemple´n
deprotection with sodium methoxide.41 A benzylidene protecting
group was installed across C4 and C6 using p-anisaldehyde
dimethyl acetal and a catalytic amount of camphorsulfonic acid
in refluxing CHCl3,42 followed by the protection of the
remaining hydroxyls with acetic anhydride and pyridine to give
39 as a white solid (57%). The benzylidene protecting group
was selectively opened to the C6 position using sodium
cyanoborohydride and trifluoroacetic acid (TFA) to give 40 as
a colorless syrup (91%). The DAST reagent was unsuccessful
at introducing an axial fluorine at C4 in order to produce the
4-fluorogalactopyranoside series of sugars. Consequently, a C4-
triflate was displaced with TBAF starting from the suitably
protected glucopyranoside 40 to give the desired 4-fluoroga-
lactopyranoside, 41, as a syrup (81% from 40). After a facile
global deprotection with acetyl chloride in methanol, the primary
hydroxyl group was selectively oxidized with TEMPO and
NaOCl to give 43 as a white solid (30% from 41).
Glycosylation of the donor 19 and the acceptor 30 was
effected with silver triflate in CH2Cl2 to give the disaccharide
31 as a white crystalline solid in a 32% yield (Scheme 4). After
a global Zemple´n deprotection to give 32 (84%), a p-methoxy-
benzylidene group was installed using p-anisaldehyde dimethyl
acetal and p-toluenesulfonic acid in DMF at 50 °C, affording
33 as a white crystalline solid (88%). The oxidation of the
remaining primary hydroxyl group was accomplished with
TEMPO and NaOCl under phase-transfer conditions with TBAB
in NaHCO3(aq) and EtOAc.32 Under the reaction and workup
conditions employed, the cleavage of the p-methoxybenzylidene
protecting group also occurred to some extent. The crude
material was then dissolved in methanol with a drop of acetic
acid to complete the deprotection, affording the desired disac-
charide substrate 34 as a white solid in a modest 25% yield.
Benzyl 4-Deoxy-4-fluoro-â-D-galactopyranosiduronic Acid
(43) (Scheme 5). Benzyl â-D-glucopyranoside (37) was synthe-
sized from peracetylated glucose in a yield of 67% by first
making the R-bromide 35 with HBr/AcOH, followed by
glycosylation using a modified Ko¨nigs-Knorr38 procedure with
Phenyl 4-Deoxy-4,4-difluoro-â-D-xylo-hexopyranosiduronic
Acid (49) (Scheme 6). Commercially available phenyl â-D-
glucopyranoside was reacted with p-anisaldehyde dimethyl
acetal and p-toluenesulfonic acid in DMF at 50 °C, followed
by the protection of the remaining hydroxyls as benzyl ethers
(41) Wolfrom, M. L.; Thompson, A. In Methods in Carbohydrate Chemistry;
Whistler, R. L., Wolfrom, M. L., BeMiller, J. N., Eds.; Academic Press:
New York, 1963; Vol. II, pp 215-220.
(42) Ferro, V.; M., M.; Stick, R. V.; Tilbrook, D. M. G. Aust. J. Chem. 1988,
41, 813-815.
9
J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002 9759

A R T I C L E S
Rye and Withers
Scheme 6a
a Conditions: (i) p-Anisaldehyde dimethyl acetal, p-TsOH, DMF, 50 °C. (ii) NaH, benzyl bromide, DMF. (iii) NaCNBH3, TFA, THF/CH2Cl2. (iv) Ac2O,
DMSO. (v) Methyl DAST, CH2Cl2, -30 °C f room temperature. (vi) H2, Pd-C, EtOAc/EtOH, followed by CAN, CH3CN/H2O. (vii) TEMPO, t-BuOCl,
H2O.
Scheme 7a
Scheme 8a
a Conditions: (i) 2,4-Dinitrofluorobenzene, DABCO, DMF. (ii) acetyl
chloride, MeOH, 4 °C. (iii) TEMPO, NaOCl, NaBr, TBAB, NaHCO3,
EtOAc/H2O, 0 °C.
using sodium hydride and benzyl bromide in DMF, to afford
44 as a white solid (39% isolated). After selectively opening
the p-methoxybenzylidene ring to leave the 4-OH free, the
oxidation of this hydroxyl with DMSO and acetic anhydride43
gave the 4-keto compound 46 as a crystalline solid (75%). The
introduction of two fluorines at C4 was accomplished by
reacting the desired 4-keto compound 46 with DAST to give
47 in an 88% yield. The benzyl ethers were then removed by
hydrogenation over Pd-C, and the removal of the p-methoxy-
benzyl ether was completed by a short treatment with ceric
ammonium nitrate (CAN) to yield 48 (63%). As with the other
substrates, the final step was the selective oxidation of the
remaining primary hydroxyl with TEMPO and t-BuOCl to give
49 as a white solid in a moderate 27% yield.
a Conditions: (i) Tf2O, pyridine, -15 °C f room temperature. (ii)
4-Chloro-2-nitrophenol (K+ salt), DMF, 80 °C. (iii) NaOMe, MeOH. (iv)
TEMPO, NaOCl, NaBr, TBAB, NaHCO3, EtOAc/H2O, 0 °C.
and pyridine) using the potassium salt of the desired phenol to
give 53, 58, and 59 in yields of 37%, 38%, and 37%,
respectively. This required an appropriately protected galacto-
side, and for comparison purposes, two protecting group
strategies were used in the synthesis of these chromogenic
substrates. Selective benzoylation of benzyl â-D-galactopyra-
noside (5)44 provided the appropriately protected intermediate
730 in one step in a moderate yield of 40%. Overbenzoylated
products could be easily deprotected and subjected to further
benzoylation to yield additional product. The alternative protec-
tion strategy included the installation of a p-methoxybenzylidene
group, acetylation of the remaining hydroxyls, and the selective
opening of the benzylidene to afford the C6 p-methoxybenzyl
ether using NaCNBH3 and TFA to give 56 in an overall yield
of 43% (Scheme 9). Chromatography was only necessary after
the final ring-opening step. Zemple´n deprotection of the
benzoates of 53 with NaOMe proved problematic, with the
phenol moiety migrating to the C3 position, leaving the desired
compound 54 in a yield of only 53%. A higher yield could have
been achieved using the acidic global deprotection strategy
(acetyl chloride in methanol) employed for 58 and 59, which
gave the desired 60 and 61 in yields of 82% and 84%,
respectively. The final step for all three compounds was the
selective oxidation of the primary hydroxyl with TEMPO and
NaOCl under phase-transfer conditions with TBAB in NaHCO3(aq)
and EtOAc to give 55, 62, and 63 in yields of 78%, 87%, and
53%, respectively.
II. Synthesis of the Chromogenic Substrates. Benzyl 4-O-
(2′,4′-Dinitrophenyl)-â-D-glucopyranosiduronic Acid (52) (Scheme
7). The installation of the dinitrophenol moiety at C4 was
accomplished via a nucleophilic aromatic substitution reaction
with 2,4-dinitrofluorobenzene and DABCO in DMF to afford
50 as a pale yellow foam (80%). After deprotection of the
acetates and the p-methoxybenzyl ether using acetyl chloride
in methanol (89%), the primary hydroxyl group was selectively
oxidized using TEMPO and NaOCl under phase-transfer condi-
tions with tetrabutylammonium bromide (TBAB) in NaHCO3(aq)
and EtOAc affording 52 as a white solid (52%).32
Benzyl 4-O-(4′-Chloro-2′-nitrophenyl)-â-D-glucopyranosidu-
ronic Acid (55) (Scheme 8), Benzyl 4-O-(3′,4′-Dinitrophenyl)-
â-D-glucopyranosiduronic Acid (62), and Benzyl 4-O-(2′-
Nitrophenyl)-â-D-glucopyranosiduronic Acid (63) (Scheme 9).
Coupling of the aryl moiety to the 4-position through a
nucleophilic aromatic substitution reaction proved impractical
for less electron deficient reagents than 2,4-dinitrofluorobenzene.
These other phenol moieties were therefore introduced via the
SN2 displacement of a C4-triflate (prepared with triflic anhydride
III. Synthesis of the Fluorogenic Substrate. Phenyl 4-Meth-
ylumbelliferyl-â-D-glucopyranosiduronic Acid (69) (Scheme 10).
The preparation of 65 from phenyl galactoside (4) proceeded
(43) Albright, J. D.; Goldman, L. J. Am. Chem. Soc. 1967, 89, 2416-2423.
(44) Stoffyn, A.; Stoffyn, P. J. Org. Chem. 1967, 32, 4001-4006.
9
9760 J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002

Chondroitin AC Lyase
A R T I C L E S
Scheme 9a
a Conditions: (i) p-Anisaldehyde dimethyl acetal, p-TsOH, DMF, 50 °C. (ii) Ac2O, pyridine. (iii) NaCNBH3, TFA, THF/CH2Cl2. (iv) Tf2O, pyridine,
-15 °C f room temperature. (v) K+ salt of appropriate phenol, DMF, 80 °C. (vi) Acetyl chloride, MeOH, 4 °C. (vii) TEMPO, NaOCl, NaBr, TBAB,
NaHCO3, EtOAc/H2O, 0 °C.
Scheme 10a
a Conditions: (i) p-Anisaldehyde dimethyl acetal, p-TsOH, DMF, 50 °C. (ii) Ac2O, pyridine. (iii) NaCNBH3, TFA, THF/CH2Cl2. (iv) Tf2O, pyridine,
-15 °C f room temperature. (v) 7-Hydroxy-4-methylcoumarin (K+ salt), DMF, 80 °C. (vi) CAN, CH3CN:H2O (9:1). (vii) CrO3, H2SO4, acetone/H2O, 35
°C, sonication. (viii) NaOMe, MeOH.
Scheme 11a
in a manner similar to that described for 57 in the synthesis of
the chromogenic substrates 62 and 63. The C4 triflate 65 was
displaced with the potassium salt of 7-hydroxy-4-methylcou-
marin (4-methylumbelliferone) in DMF at 80 °C to produce 66
as a white foam (29%). Selective deprotection of the p-
methoxybenzyl group with CAN gave 67 (87%) followed by a
sonicated Jones oxidation45 with CrO3 and H2SO4 in acetone/
H2O to afford the acid 68 as glassy white solid. The Jones
oxidation was used in place of the TEMPO oxidation employed
in the other syntheses due to the fact that the methylumbelliferyl
group was not compatible with the conditions of the TEMPO
oxidation. The facile Zemple´n deprotection of the acetates
a Conditions: (i) Ac2O, pyridine. (ii) NBS, hν, CCl4. (iii) Tributyltin
deuteride, toluene, reflux. (iv) NaOMe-d3, CD3OD.
afforded the desired compound, 69, as a white solid (94%).
L-ido compound. Purification difficulties reduced the isolated
IV. Synthesis of the Deuterated Substrates. Phenyl 4-Deoxy-
yield of 72 to 5%, which was then deprotected with NaOMe-d3
4-fluoro-5-{2H}-â-D-glucopyranosiduronic Acid (73) (Scheme
in CD3OD to give the desired substrate 73 as a white solid
11). Compound 15 was acetylated with acetic anhydride and
sulfuric acid (79%) followed by radical bromination with
N-bromosuccinimide (NBS) in CCl4, directed by the adjacent
carbonyl, to give the 5-bromo compound, 71, as a white solid
(39%). Deuterium incorporation at C5 proceeded smoothly with
tributyltin deuteride in refluxing toluene to give both the
D-gluco- and L-ido-configured sugars in roughly equal amounts.
1H NMR analysis confirmed the expected 4C1 chair conforma-
(92%).
Phenyl 4-Deoxy-4-fluoro-4-{2H}-â-D-glucopyranosiduronic
Acid (77) (Scheme 12). Sodium borodeuteride in methanol was
used to reduce the ketone of 46 and introduce the deuterium
functionality at C4, giving the galactoside 74 as a white
crystalline solid (94%). The stereochemistry at C4 was con-
firmed by analysis of the 1H NMR signals arising from the protio
compound also present as a result of the less than 100%
deuterium incorporation. After the incorporation of fluorine at
C4 with DAST to give 75, global deprotection was effected by
hydrogenolysis over Pd-C in EtOAc/EtOH. However, the
1
tion for the desired D-gluco compound and C4 chair for the
(45) Allanson, N. M.; Liu, D. S.; Chi, F.; Jain, R. K.; Chen, A.; Ghosh, M.;
Hong, L. W.; Sofia, M. J. Tetrahedron Lett. 1998, 39, 1889-1892.
9
J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002 9761

A R T I C L E S
Rye and Withers
Scheme 12a
a Conditions: (i) Sodium borodeuteride, MeOH. (ii) Methyl DAST, CH2Cl2, -30 °C f room temperature. (iii) H2, Pd-C, EtOAc/EtOH, followed by
CAN, CH3CN/H2O. (iv) TEMPO, t-BuOCl, H2O.
Table 1. Kinetic Data for Fluorinated Substrates
Table 2. Kinetic Data for Chromogenic and Fluorogenic
Substratesa
substrate
(aglycon)
kcat
(s-1
Km
(mM)
k
cat/Km
)
(M-1 s-1
)
phenol
kcat
Km
(mM)
k
cat/Km
15 (Ph)
16 (Bn)
2.3
0.16
114
28
20
5.9
substrate
pK
a
(s-1
)
(M-1 s-1
)
17 (Me)
34 (GalNAc)
no saturation observed
0.011
0.22
0.85
52
62
55
63
69
3.96
5.36
6.45
7.22
7.8
0.019
0.019
0.035
0.027
0.0016
7.0
9.0
12.2
16.9
1.0
2.7
2.1
2.8
1.6
1.5
12.5
presence of trace amounts of sulfur from the previous DAST
reaction poisoned the hydrogenation catalyst and made the
deprotection problematic, despite extensive purification. The
p-methoxybenzyl ether was removed with CAN in CH3CN/H2O
to afford 76 as a white solid (58%). The final step was the
selective oxidation of the primary hydroxyl with TEMPO and
t-BuOCl to give the desired 77 as a white solid in a yield of
50%
a Note that analyses of 52, 55, 62, and 63 were done at pH 6.8, whereas
that for 69 was performed at pH 8.
V. Enzymology: Effect of Varying Substrate Structure
on Reactivity. The importance of the anomeric substituent of
the glucuronic acid moiety on kinetic parameters was first
analyzed using substrates with a fluoride leaving group, since
these are the most readily synthesized. The first substrates
studied were those with a methyl (17), benzyl (16), or phenyl
(15) group at the anomeric center: the kinetic data are
summarized in Table 1. The preference for an aromatic
substituent is clear from the fact that no saturation was observed
for the substrate with the methyl anomeric group. The kcat/Km
for this substrate is 91 and 27 times smaller, respectively, than
those of the substrates with a phenyl or benzyl group at the
anomeric center. The substrate with a phenyl group at the
anomeric center has a significantly larger kcat than did the
benzyl-containing substrate; however, its binding is poorer as
reflected in its larger Km value. To investigate the contribution
to catalysis of the sugar residue in the +2 binding site, a
disaccharide substrate (34) was synthesized that incorporated
an N-acetylgalactosamine moiety at the reducing end. As
expected, the binding was improved with this disaccharide
substrate, although not as much as had been hoped. Surprising
was the fact that the kcat value was reduced significantly from
that of the monosaccharide substrates with a phenyl or benzyl
group at the anomeric center, resulting in a much poorer
substrate as indicated by the lower kcat/Km value (Table 1).
VI. Chromogenic and Fluorogenic Substrates. Substrates
incorporating phenol leaving groups of differing reactivity were
synthesized in order to perform a Brønsted analysis of the
Figure 2. Linear free energy relationship for substrates with phenol leaving
groups of varying pKa values.
enzymatic turnover of these substrates. Four competent sub-
strates, 52, 55, 62, and 63, with nucleofugalities (leaving-group
abilities) defined by their pKas were synthesized and tested as
substrates. The resultant kinetic data are summarized in Table
2 and the linear free energy relationship is presented in Figure
2. The second-order rate constant (kcat/Km) is independent of
phenol pKa as shown by the flat linear free energy relationship.
This strongly suggests that the breaking of the C4-O4 bond
does not occur in a rate-limiting step and essentially rules out
a concerted syn-elimination.
A synthetic substrate with a fluorescent 4-methylumbelliferyl
leaving group (69) was synthesized in order to create a more
sensitive assay. The overall catalytic efficiency of this fluoro-
genic substrate as measured by kcat/Km (1.5 M-1 s-1) is similar
to that of the chromogenic substrates (Table 2). Indeed, the
slightly lower value can be explained by the higher pH used in
this assay to facilitate aglycon detection. An increase in pH has
been shown not only to decrease kcat but also to increase Km
(vide infra).
9
9762 J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002

Chondroitin AC Lyase
A R T I C L E S
Figure 3. pH activity profiles for chondroitin AC lyase. (A) Vmax vs pH for chondroitin 6-sulfate. An extinction coefficient of ꢀ ) 3800 M-1 cm-1 for the
unsaturated products was used to calculate the rates.3 (B) kcat/Km vs pH for the synthetic substrate 16. (C) kcat/Km vs pH for chondroitin 6-sulfate.
VII. Inhibitors. Two compounds were synthesized that
contain fluorine at C4 in the axial position in the hope that these
might serve as improved substrates by virtue of an anti-
elimination reaction mechanism (E2). Benzyl 4-deoxy-4-fluoro-
â-D-galactopyranosiduronic acid (43) was not turned over by
the enzyme, yet was found to be a competitive inhibitor with a
Ki value of 3 mM. Similarly, the 4,4-difluoro compound 49
showed no activity (kcat < 1 × 10-5 s-1) and was also a
competitive inhibitor, with a Ki value of 24 mM.
deuterium kinetic isotope effects (KIEs). Using substrate
concentrations well below Km, the primary deuterium KIE on
kcat/Km for abstraction of the proton R to the carbonyl was
measured to be 1.67 ( 0.07. The presence of a primary
deuterium KIE demonstrates that the C5-H5 bond is broken
in a rate-limiting step and also rules out an E1 elimination
mechanism. The secondary deuterium KIE on kcat/Km was
measured to be 1.01 ( 0.03. This supports a stepwise mecha-
nism where the bond to the C4 leaving group is not broken in
a rate-limiting step and is consistent with the flat linear free
energy relationship.
VIII. pH Dependence. pH profiles for the reaction catalyzed
by chondroitin AC lyase were constructed for the natural
substrate (chondroitin 6-sulfate) as well as for one of the
synthetic monosaccharide substrates, the phenyl 4-deoxy-4-
fluoroglucuronide (16). A simple analysis of Vmax versus pH
for chondroitin 6-sulfate revealed the rate to be maximal at pH
6.8 and seemingly dependent on two ionizations of pKa1 ) 5.3,
and pKa2 ) 7.9 (Figure 3A). A similar analysis of Vmax versus
pH was not possible with the synthetic substrate due to its high
Km value. Instead, a kcat/Km versus pH analysis was performed
by measurement of initial rates at low substrate concentrations
and dividing by enzyme concentration. This analysis showed
the second-order rate constant to be maximal at a pH of ∼6,
slightly below the optimum found for the Vmax profile for the
natural substrate (Figure 3B). The one apparent pKa value
obtainable from this curve is pKa2 ) 7.2. When the kcat/Km
versus pH analysis was performed with chondroitin 6-sulfate,
remarkable changes in Km were observed with changes in pH.
The Km value decreased by a factor of 175 when the pH was
changed from 8.8 to 5.0. This resulted in a steadily increasing
kcat/Km value with decreasing pH, as shown in Figure 3C.
1
X. Solvent Deuterium Exchange. The H NMR spectrum
of remaining 4-fluoro substrate 16 and product after partial
conversion by chondroitin AC lyase in D2O does not reveal
any incorporation of deuterium into the C5 position of the
substrate (Figure 4). While this result does not prove that
deuterium exchange is not occurring within the active site, which
may be inaccessible to bulk solvent, this result is however
consistent with a mechanism that involves the rapid elimination
of the C4 substituent after a rate-limiting proton abstraction
((E1cb)irr).
Discussion
Substrate Structure. The use of heterogeneous polymeric
substrates places immense limitations on the type of mechanistic
studies that can be utilized to elicit the details of an enzymatic
reaction. Chondroitin sulfates are commercially available;
however, the sulfation patterns are not consistent throughout
the polymer, and the molecular weight is not accurately known
and most likely varies between individual chains. Further, the
structures of these substrates cannot easily be manipulated to
allow the use of mechanistic probes such as kinetic isotope
effects. The synthetic substrates described herein have allowed
IX. Kinetic Isotope Effects. Substrates with deuterium
incorporated at C5 (73) and C4 (77) were synthesized in order
to probe the enzyme mechanism via primary and secondary
9
J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002 9763

A R T I C L E S
Rye and Withers
Figure 4. Deuterium exchange experiment. Partial 400-MHz 1H NMR spectra of (A) the monosaccharide substrate 16 and (B) its partial enzymatic digestion
to product by chondroitin AC lyase in D2O.
the measurement of defined and reproducible kcat and Km values,
previously unattainable with the natural polymeric substrates,
as summarized in Tables 1 and 2. The use of a variety of leaving
groups allows a range of different detection techniques to be
used in their analysis, as described elsewhere.46 Interestingly,
the chromogenic substrates bind more tightly than do the
fluoride-releasing substrates, presumably because of hydropho-
bic interactions between the aromatic leaving groups and
hydrophobic residues in the enzyme active site, since residues
such as tyrosine and tryptophan are very commonly found at
the active sites of carbohydrate-binding proteins. In fact, the
three-dimensional structure of various enzyme-oligosaccharide
complexes of chondroitin AC lyase reveals tyrosine, histidine,
phenylalanine, and several tryptophan residues arranged through-
out the active site.14 Of mechanistic importance is the finding
of higher kcat values for the fluoride-releasing substrates than
for substrates with oxygen-based leaving groups. This presum-
ably arises from the stabilization of the developing negative
charge by the highly electronegative fluorine atom in the rate-
limiting deprotonation step. This preference of the enzyme for
substrates with aromatic appendages is also reflected in the
results shown in Table 1 in which the structure of the anomeric
substituent was varied. Compromises must be made in the choice
of anomeric substituent: phenyl glycosides have the highest
turnover. Disappointingly, a disaccharide substrate with an
N-acetylgalactosamine residue at the reducing end (34) binds
only slightly tighter than does the monosaccharide substrate,
and with a kcat value at least 10-fold lower, resulting in a much
less efficient substrate as measured by the low kcat/Km value
(Table 1). Perhaps binding interactions toward the nonreducing
end of the active site are more important for binding and
catalysis. Unfortunately the addition of extra sugar units to
occupy the -1 and -2 enzyme-binding sites is not compatible
with the design of this class of substrates. Based upon the
findings, the phenyl aglycon was chosen to be used in all
substrates for further mechanistic analyses.
The finding of higher kcat values for the fluoride-releasing
substrates than for the chromogenic or fluorogenic substrates,
possibly due to the stabilization of the developing negative
charge by the highly electronegative fluorine atom, suggested
that a substrate bearing two fluorines at that position, a 4,4-
difluoro substrate, might prove a better substrate still, especially
given the trans relationship of the axial fluorine with the C5
proton. A fluorine substituent is only slightly larger than a
hydrogen; thus, it is reasonable to expect that this substrate might
bind. However, surprisingly, no cleavage whatsoever of this
compound (49) was observed, nor was the 4-fluoro-galactopy-
ranoside analogue (43) a substrate for the enzyme. A trivial
explanation for this otherwise surprising finding would be that
the compounds with an axial fluorine substituent do not bind
to the active site. However, both of these compounds were
shown to be competitive inhibitors of the enzyme with Ki values
of 24 and 3 mM for the 4,4-difluoro- and 4-fluorogalacto
compounds, respectively. Despite the small difference in size
of fluorine and hydrogen atoms, the introduction of the axial
k
cat values, but bind poorly, while the converse is true for benzyl
glycosides.
Since the natural substrates for chondroitin AC lyase are long
polysaccharide chains, one would expect that increasing the
number of sugar units in the synthetic substrate would increase
the binding interactions with the enzyme, thereby yielding faster
(46) Rye, C. S.; Withers, S. G. Anal. Biochem., in press.
9
9764 J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002

Chondroitin AC Lyase
A R T I C L E S
C4-fluorine must either introduce enough additional steric bulk
or result in an unfavorable dipolar interaction of sufficient
magnitude to cause an unproductive mode of binding. An
alternative explanation for the lack of activity shown by the
difluoro 49 may rest at least partly with the greater than 10
kcal/mol ground-state stabilization of the carbon-fluorine bond
afforded by geminal fluorine atoms.47,48 The greater inhibition
by 43 is most likely due to the fact that monosaccharide
substrates with a benzyl aglycon have been shown to bind tighter
to the enzyme than do those with a phenyl aglycon moiety, such
as 49.
this group should not affect the results in the pH range of 5-9.
With the use of Gran plots,50-52 the pKa values of chondroitin
6-sulfate and the synthetic substrate 16 were measured to be
2.9 and 2.7, respectively. This small difference in the pKa values
of the carboxylic acid groups of these two compounds is not
sufficient to explain the different effects of pH on the Km.
Inspection of the three-dimensional structure of chondroitin AC
lyase reveals that the proposed binding cleft on the enzyme is
positively charged, presumably in order to bind this polyanionic
substrate.8,14 It therefore seems reasonable that a decrease in
pH may afford a more positively charged binding pocket on
the enzyme that may well lead to enhanced binding only of a
long polysaccharide chain such as the natural substrate, but not
the monosaccharide synthetic substrate. It is also conceivable
that chondroitin 6-sulfate undergoes a pH-dependent confor-
mational change that affects its binding to the enzyme active
site1,53 and thus might be expected to change the Km value.
pH Dependence. Previous studies3,49 of the pH dependence
of chondroitin AC lyase unfortunately contradict each other;
thus, further analyses of pH dependence were performed using
both the enzyme’s natural substrate (chondroitin 6-sulfate) and
a synthetic monosaccharide substrate (16). A simple analysis
of Vmax versus pH with chondroitin 6-sulfate (Figure 3A)
revealed the Vmax to be maximal at a pH of 6.8, and this was
the pH chosen for subsequent enzymatic analyses. Two pKas
can be extracted from this curve, 5.3 and 7.9, which may
represent two ionizable groups important for catalysis in the
enzyme such as the general-acid and general-base residues.
These pKas could also reflect the ionization of the substrate
within the enzyme-substrate complex. However, the pKa values
of the substrate carboxylic acid moieties have been shown to
be much lower than either of these values (vide infra), rendering
this interpretation less likely. Consistent with this interpretation
is the very similar pH dependence of kcat/Km for 16 (Figure 3B),
which is maximal at around pH 6. The one clear pKa value that
can be abstracted from this curve is 7.2, reasonably close to
that deduced from the study with the natural substrate and
implying the presence of an essential group that must remain
protonated for activity, possibly the acid catalyst. From X-ray
crystallographic data, a tyrosine residue has been implicated as
the acid catalyst.14,15 The pKa of a tyrosine residue is normally
around 10; however, a decrease of 2-3 pKa units is not
unprecedented for amino acids located at the active site of
enzymes. Perhaps the most surprising finding was the remark-
able change in Km for the natural substrate with changing pH.
The Km was measured to be 1.23 mg/mL at pH 8.8 and only
7.0 × 10-3 mg/mL at pH 5.0, corresponding to a 175-fold
change. Conversely, for the synthetic substrate 16, the Km
dropped by only one-third between pH 6.8 and pH 5.2. The
substantial decrease in Km observed with chondroitin 6-sulfate
resulted in a significantly different profile of kcat/Km versus pH
from that seen with the synthetic substrate or from the profile
of Vmax with the same substrate (Figure 3C and A). The origin
of this difference must lie in the differences between the
synthetic and natural substrates, which are threefold: (i) the
synthetic substrate has a fluoride leaving group that does not
require acid catalysis, (ii) the synthetic substrate is not sulfated,
and (iii) the synthetic substrate is a monosaccharide, whereas
the natural substrate is a large polysaccharide. The dependence
of the leaving group on acid catalysis is not expected to affect
the binding of the substrate to the enzyme and thus should not
result in a dependence of Km upon pH. The pKa of the sulfate
group is expected to be sufficiently low that the ionization of
Toward a Mechansim. To probe the identity and nature of
the rate-determining step(s) for the cleavage of these substrates,
a study of the dependence of the rate upon leaving-group ability
was performed. The resulting flat linear free energy relationship
(Figure 2) shows that there is no significant charge development
on the C4 oxygen at the transition state of the rate-limiting step.
There are two main possibilities to explain this result. The first
is that there is highly effective proton donation to the departing
phenolate by a general-acid catalyst at the transition state, which
neutralizes any charge development at the C4 oxygen. If this is
the case, it is impossible to tell whether the reaction is concerted
or stepwise. The second possibility is that the reaction is
stepwise with the breaking of the C4-O4 bond not occurring
in a rate-limiting step, thus ruling out a concerted syn-
elimination. This second explanation is preferred not only by
the rarity of concerted syn-eliminations but also by its agreement
with the near-unity secondary deuterium KIE that was measured
with a substrate containing deuterium at C4 (vide infra).
The primary deuterium KIE on kcat/Km for abstraction of the
proton R to the carbonyl was measured using 73 to be kH/kD )
1.67 ( 0.07. The presence of a primary deuterium KIE
demonstrates that the C5-H5 bond is broken in a rate-limiting
step, ruling out an E1 elimination mechanism. This result, along
with the flat linear free energy relationship, suggests a stepwise
mechanism with rate-limiting proton abstraction, conforming
to an (E1cb)irr reaction scheme. This is consistent with the fact
that proton transfers to and from carbon acids are usually much
slower than those to and from oxygen, nitrogen, and sulfur.54
This slow transfer has been attributed to the need for structural
reorganization accompanying the delocalization of the negative
charge, solvent reorganization, and from the poor hydrogen-
bonding capability of carbon acids and of the carbanionic
carbon.55 The Hammond postulate tells us that the transition
state leading to the formation of a high-energy species, such as
the enolic intermediate proposed for this elimination mechanism,
occurs late on the reaction coordinate and resembles the enolic
intermediate. Consequently, the extent of proton transfer from
(50) Gran, G. Anal. Chim. Acta 1988, 206, 111-123.
(51) Harris, D. C. QuantitatiVe Chemical Analysis, 3 ed.; W. H. Freeman and
Co.: New York, 1991.
(52) Rossotti, F. J. C.; Rossotti, H. J. Chem. Educ. 1965, 42, 375-378.
(47) Reed, A. E.; Schleyer, P. v. R. J. Am. Chem. Soc. 1987, 109, 7362-7373.
(48) Schleyer, P. v. R.; Jemmins, E. D.; Spitznagel, G. W. J. Am. Chem. Soc.
1985, 107, 6393-6394.
(53) Scott, J. E. FASEB J. 1992, 6, 2639-2645.
(54) Crooks, J. E. In Proton-Transfer Reactions; Caldin, E., Gold, V., Eds.;
Wiley: New York, 1975; pp 153-177.
(49) Michelacci, Y. M.; Dietrich, C. P. Biochem. J. 1975, 151, 121-129.
(55) Bernasconi, C. F.; Terrier, F. J. Am. Chem. Soc. 1987, 109, 7115-7121.
9
J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002 9765

A R T I C L E S
Rye and Withers
the substrate at the transition state is large, and one might expect
a KIE much less than the maximal value, as was observed. In
addition, the chemical step in an enzymatic reaction may well
not be rate-limiting, or only partially so, as a result of
evolutionary pressure on the enzyme to accelerate chemical steps
until they are just slightly faster than the release of product.56,57
Small primary KIEs may also arise from a nonlinear arrange-
ment of the proton donor, proton, and the proton acceptor, in
the transition state for the transfer of the proton. Such an
occasion is conceivable if the catalytic base residue responsible
for the abstraction of the proton is the same residue responsible
for delivering a proton to the leaving group. This is postulated
to be the case with alginate lyase A1-III from Sphingomonas
species, an enzyme that degrades a poly-â-D-mannuronic acid
substrate. From the three-dimensional structure of the enzyme
complexed with a trisaccharide product, a tyrosine residue has
been implicated as both the catalytic base that abstracts the C5
proton and the general acid that protonates the leaving sugar
unit.16 Less than maximal primary deuterium KIEs have been
measured with other enzyme systems undergoing elimination
reactions. For enolase, the primary KIE for the abstraction of
the proton R to the carbonyl group is dependent on both the
pH and Mg2+ concentration and ranges from ∼1.2 to ∼3.3.58,59
The crotonase-catalyzed dehydration of 3-hydroxybutyrylpan-
tetheine shows a primary KIE of 1.6060 and that for o-
succinylbenzoate synthase has been measured to be 2.7.61 The
reaction catalyzed by UDP-N-acetylglucosamine 2-epimerase
also shows a small primary KIE of 1.8 for the C2 hydrogen.62
Although not an overall elimination process, there is strong
evidence for the elimination of the UDP moiety as part of the
proposed reaction mechanism. The primary KIE for an antibody-
catalyzed elimination of HF adjacent to a ketone has been
measured to be 2.35.63
Using 77 with a deuterium incorporated at C4, the secondary
deuterium KIE was measured to be 1.01 ( 0.03. A secondary
KIE value close to unity would be expected if the rate-limiting
step were solely the proton abstraction and formation of the
enolic intermediate since C4 does not acquire any sp2 character
during this step. On the other hand, a partially rate-limiting
departure of the leaving group where C4 takes on partial sp2
character at the transition state is expected to show a secondary
KIE greater than 1, with values of kH/kD ) 1.3-1.4 as maximal
values. The elimination of water by the non-carbohydrate-
degrading enzymes fumarase and crotonase has been shown to
involve the kinetically significant departure of the leaving group
and shows secondary deuterium kinetic isotope effects ranging
from 1.13 to 1.23, illuminating the significant sp2 character at
this center.60,64 The low value of the secondary deuterium KIE
measured for chondroitin AC lyase at pH 6.8 and 30 °C therefore
supports a stepwise mechanism in which the bond to the C4
leaving group is not broken in a rate-limiting step.
The primary and secondary deuterium KIEs, combined with
the flat linear free energy relationship, suggest that the rate-
limiting step is the abstraction of the C5 hydrogen and the
formation of the enolic intermediate. Once this intermediate is
formed, the C4 substituent is eliminated in a non-rate-limiting
step. Further evidence that is consistent with this proposal would
be the absence of deuterium exchange at the C5 position before
the elimination of the C4 substituent. Other enzymes, such as
enolase,58 show rapid exchange of the proton R to the carbonyl
group with solvent and are believed to use a stepwise mechanism
in which proton abstraction and leaving-group departure are both
partially rate-limiting. On the other hand, the crotonase-catalyzed
â-elimination,60 which is thought to be concerted, shows almost
no deuterium exchange () 3%). Figure 4 shows the partial 1H
NMR spectra of a monosaccharide substrate (16) and that of
the substrate and product mixture after partial conversion by
chondroitin AC lyase in D2O. If deuterium exchange at C5
occurred faster than the elimination of the 4-fluoro group, then
the integration of H5 in the starting material would decrease
relative to that of H1, H2, and H3. However, this is clearly not
the case, as shown by the integrations in Figure 4. This is
consistent with a mechanism that involves the rapid elimination
of the C4 substituent after a rate-limiting proton abstraction
((E1cb)irr). However, the absence of exchange does not constitute
proof that this process is not occurring since enzyme active sites
have long been proposed to be sequestered from bulk solvent,
and thus, the residues located there may not be accessible to
the D2O solvent as required for deuterium exchange. The use
of modified substrates also has its limitations in that it is not
absolutely required that the lyase-catalyzed elimination of the
polysaccharide substrate occur in the same manner as that
deduced with the synthetic substrates. However, these studies
do at least provide general mechanistic insights into turnover
of these artificial substrates by the enzyme. Further, it is
important to note that the mechanistic studies presented in this
paper would not be possible without the modified substrates.
Conclusions
The synthesis and kinetic evaluation of several simple
substrates for chondroitin AC lyase has allowed the determi-
nation of defined and reproducible kcat and Km values, previously
unattainable with the inhomogeneous and polymeric natural
substrates. Evidence has been gathered from linear free energy
relationships, primary and secondary deuterium KIEs, and
isotope exchange experiments in order to elucidate the catalytic
mechanism of this enzyme. The flat linear free energy relation-
ship produced using substrates with leaving groups of differing
reactivity combined with the low secondary deuterium KIE
strongly suggests that the breaking of the C4-leaving group bond
does not occur in a rate-limiting step, thus ruling out a concerted
mechanism. The primary deuterium KIE of 1.67 ( 0.07 shows
that the abstraction of the C5 proton occurs in a rate-limiting
step and that the transition state for this step is late on the
reaction coordinate, resembling the proposed enolic intermediate.
(56) Klinman, J. P. In Transition States of Biochemical Processes; Gandour, R.
D., Schowen, R. L., Eds.; Plenum Press: New York, 1978; pp 165-200.
(57) Northrop, D. B. Biochemistry 1975, 14, 2644-2651.
(58) Anderson, S. R.; Anderson, V. E.; Knowles, J. R. Biochemistry 1994, 33,
10545-10555.
(59) Shen, T. Y.; Westhead, E. W. Biochemistry 1973, 12, 3333-3337.
(60) Bahnson, B. J.; Anderson, V. E. Biochemistry 1991, 30, 5894-5906.
(61) Taylor, E. A.; Palmer, D. R. J.; Gerlt, J. A. J. Am. Chem. Soc. 2001, 123,
5824-5825.
(62) Morgan, P. M.; Sala, R. F.; Tanner, M. E. J. Am. Chem. Soc. 1997, 119,
10269-10277.
(63) Shokat, K.; Uno, T.; Schultz, P. G. J. Am. Chem. Soc. 1994, 116, 2261-
2270.
(64) Blanchard, J. S.; Cleland, W. W. Biochemistry 1980, 19, 4506-4513.
(65) Saunders: W. H. In Techniques of Chemistry. Part I, InVestigations of
Rates and Mechanisms of Reactions, 4 ed.; Bernasconi, C. F., Ed.; John
Wiley & Sons: New York, 1986; Vol. 6, pp 565-611.
(66) Kobayashi, Y.; Shiozaki, M. J. Org. Chem. 1995, 60, 2570-2580.
(67) Wolfrom, M. L.; Thompson, A. In Methods in Carbohydrate Chemistry;
Whistler, R. L., Wolfrom, M. L., BeMiller, J. N., Eds.; Academic Press:
New York, 1963; Vol. II, pp 211-215.
(68) Piel, E. V.; Purves, C. B. J. Am. Chem. Soc. 1939, 61, 2978-2979.
9
9766 J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002

Chondroitin AC Lyase
A R T I C L E S
Acknowledgment. We thank the Canadian Institutes of
Health Research, the Natural Sciences and Engineering Research
Council (NSERC) of Canada, Dr. Paul Trussell, and the British
Columbia Science Council for financial support, and IBEX
Technologies Inc. and Dr. Mirek Cygler for their generous
donation of chondroitin AC lyase.
Supporting Information Available: Experimental procedures
and characterization data for all compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
See any current masthead page for ordering information and
Web access instructions.
JA020627C
9
J. AM. CHEM. SOC. VOL. 124, NO. 33, 2002 9767