(Received for publication, August 27, 1996, and in revised form, November 3, 1996)
From the Department of Biophysics and Biochemistry,
Graduate School of Science, University of Tokyo, Hongo-7,
Tokyo 113, Japan, the ¶ Institute of Biological Chemistry,
Academia Sinica, Nankang, Taipei 115, Taiwan, and the
Department of Medicinal Chemistry, Victorian College of
Pharmacy, Monash University, Parkville,
3052 Victoria, Australia
Deaminoneuraminic acid residue-cleaving enzyme
(KDNase Sm) is a new sialidase that has been induced and purified
from Sphingobacterium multivorum. Catalysis by this new
sialidase has been studied by enzyme kinetics and 1H NMR
spectroscopy. Vmax/Km
values determined for synthetic and natural substrates of
KDNase Sm reveal that 4-methylumbelliferyl-KDN (KDN2MeUmb,
Vmax/Km = 0.033 min
1) is the best substrate for this sialidase,
presumably because of its good leaving group properties. The transition
state analogue, 2,3-didehydro-2,3-dideoxy-D-galacto-D-glycero-nonulosonic
acid, is a strong competitive inhibitor of KDNase Sm
(Ki = 7.7 µM versus
Km = 42 µM for KDN
2MeUmb).
2-Deoxy-2,3-didehydro-N-acetylneuraminic acid and
2-deoxy-2,3-didehydro-N-glycolylneuraminic acid are known to be strong competitive inhibitors for bacterial sialidases such as Arthrobacter ureafaciens sialidase; however, KDNase Sm
activity is not significantly inhibited by these compounds. This
observation suggests that the hydroxyl group at C-5 is important for
recognition of the inhibitor by the enzyme.
Reversible addition of water molecule (or hydroxide ion) to the
reactive sialosyl cation, presumably formed at the catalytic site of
KDNase Sm, eventually gives rise to two different adducts, the - and
-anomers of free
3-deoxy-D-glycero-D-galacto-nonulosonic acid. 1H NMR spectroscopic studies clearly demonstrate that
the thermodynamically less stable
-form is preferentially formed as
the first product of the cleavage reaction and that isomerization
rapidly follows, leading to an equilibrium mixture of the two isomers,
the
-isomer being the major species at equilibrium. Therefore, we
propose that KDNase Sm catalysis proceeds via a mechanism common to the known exosialidases, but the recognition of the substituent at C-5 by
the enzyme differs.
3-Deoxy-D-galacto-D-glycero-nonulosonic
acid (KDN)1 is a relatively new member of
the sialic acid family in which the N-acyl group at C-5
position of N-acylneuraminic acid (Neu5Acyl) is replaced by
a hydroxyl group. Following the first report of KDN in 1986 (1), its
association with a diverse number of living organisms ranging from
bacterial capsular polysaccharide (2) to animal glycoproteins and
glycolipids (3-11) has been demonstrated. Interestingly, immunochemical techniques suggested the presence of an oligomeric form
of 2
8-linked KDN in mammalian tissues (10, 11), and the
occurrence of KDN residues was confirmed by biochemical and chemical
methods (12). A unique feature of KDN-containing glycoconjugates is
their complete resistance to the action of the known bacterial and
viral exosialidases (1, 5, 9, 13).
Recently, we demonstrated the existence of a new sialidase, which
specifically cleaves both synthetic and naturally occurring KDN--ketosides in a Gram-negative soil bacterium,
Sphingobacterium multivorum (14). This enzyme, designated
KDNase Sm, catalyzes the hydrolysis of a wide variety of
different KDN-ketosidic linkages, e.g. KDN
2
3Gal,
KDN
2
6GalNAc, and KDN
2
8KDN, in a diverse range of
oligosaccharides, glycoproteins, and glycolipids as well as the
synthetic substrate, 4-methylumbelliferyl KDN (KDN
2MeUmb) (14).
KDNase Sm is an inducible enzyme in the periplasm of the bacteria (15).
The purified enzyme consists of a single polypeptide chain with
molecular mass of 47.5 kDa (15). Significantly, KDNase Sm
completely lacks any N-acylneuraminidase activity that
releases Neu5Ac and Neu5Gc from a variety of Neu5Ac- and
Neu5Gc-containing glycoconjugates (14). Studies from our group and from
Li et al. (13) describe the occurrence of "KDN
sialidases" that cleave both KDN and N-acylneuraminyl
ketosides in fish ovary (16) and liver (13).
Recent kinetic and crystallographic studies have provided information about the molecular mechanism of exosialidase catalysis of sialyl linkages (17-22). This paper describes a quantitative investigation of the equilibria and kinetics involved in KDNase Sm-catalyzed hydrolysis by the steady-state enzyme kinetics and 1H NMR spectroscopy. This process is discussed in both kinetic and thermodynamic terms and compared with other known bacterial sialidases.
Chemicals
KDN and Neu5Gc were prepared as described previously (23).
4-Methylumbelliferyl KDN (KDN2MeUmb) was kindly provided by Dr. T. G. Warner (Genentech, Inc.). Neu5Ac and 4-methylumbelliferyl Neu5Ac
(Neu5Ac
2MeUmb) were purchased from Nacalai (Japan). KDN dimer,
KDN
2
8KDN (hereafter (KDN)2), was prepared from
KDN-rich glycoprotein as described previously (24). KDN-lactose,
KDN
2
3Gal
1
4Glc, was prepared by digestion of
(KDN)GM3 (5) with Rhodococcus sp. ceramide
glycanase (Seikagaku Kogyo, Japan) according to the method previously
described (9). In brief, (KDN)GM3 (240 µg) was dissolved
with sonication in 250 µl of 0.4% Triton X-100, 10 mM
sodium acetate (pH 5.2), and incubated with ceramide glycanase (5 milliunits) at 37 °C for 18 h. To this was added 1.25 ml of chloroform/methanol (2:1, v/v) with mixing. The aqueous layer was
applied to a Sep-Pak C18 Cartridge equilibrated with methanol/water (1:1, v/v) and eluted with 2 ml of water. The eluate, containing KDN-lactose, was evaporated and used for kinetic experiments. The
KDN-lactose thus obtained was pure based on TLC analysis (Si-HPF-Silica Gel TLC plate (J. T. Baker Inc.), chloroform/methanol/0.2%
CaCl2 (55:45:10, v/v/v)).
2-Deoxy-2,3-didehydro-N-acetylneuraminic acid (Neu5Ac2en) was commercially available from Boehringer Mannheim (Germany). N-Acetylneuraminate pyruvate-lyase was obtained from Wako Chemicals, Osaka. 2,3-Didehydro-2,3-dideoxy-D-glycero-D-galacto-2-nonulopyranosonic acid (KDN2en), 2,3-didehydro-2,3,4-trideoxy-D-glycero-D-galacto-2-nonulopyranosonic acid (4-deoxy-KDN2en), and 2-deoxy-2,3-didehydro-N-glycolylneuraminic acid (Neu5Gc2en) were chemically synthesized according to literature procedures (25, 27).2
Enzymes
KDNase Sm was isolated and purified from S. multivorum as described previously (15). One unit of enzyme
activity is defined as the amount of enzyme required to catalyze the
hydrolysis of 1 nmol of KDN2MeUmb per min at 25 °C. Exosialidases
from Arthrobacter ureafaciens and Clostridium
perfringens were purchased from Nacalai Co. (Japan) and Sigma,
respectively. One unit of the exosialidase activity is defined as the
amount of enzyme that releases 1 µmol of Neu5Ac from sialyllactose
per min at 37 °C (28).
Kinetics
Initial rates were measured for each substrate in the presence and absence of inhibitors as follows.
KDNEach of KDN2MeUmb and
Neu5Ac
2MeUmb at concentrations ranging from 13.9 to 139 µM was incubated with 19 milliunits of KDNase Sm or 0.020 milliunits of A. ureafaciens or C. perfringens
exosialidase in 21.6 µl of reaction mixture (containing 0.1 M Tris/acetic acid buffer (pH 6.0), 0.1 M NaCl,
and 1 mg/ml BSA) at 25 °C for 30 min. After incubation, 3.0 ml of 85 mM glycine/carbonate buffer (pH 9.3) was added to the
reaction mixture. The released 4-methylumbelliferone was
fluorometrically determined as described previously (29). Control
incubation was carried out in an identical fashion but without
enzyme.
A reaction mixture (10 µl) containing 82.0-410 µM (KDN)2 or 89.6-766 µM KDN-lactose, 0.152 units of KDNase Sm, 100 mM Tris/acetic acid buffer (pH 6.0), 100 mM NaCl, and 1 mg/ml BSA was incubated at 25 °C for 30 min for (KDN)2 or 1 h for KDN-lactose. The reaction was terminated by adding 10 µl of cold ethanol and then dried using Speed Vac SC 110. To this was added 25 µl of 50 mM potassium phosphate buffer (pH 7.2) containing 0.025 units of N-acylneuraminate pyruvate lyase which was then incubated at 37 °C for 1 h to destroy any free KDN remaining. The amount of remaining (KDN)2 and KDN-lactose was determined by the thiobarbituric acid method as described previously (30, 31). Control incubation was carried out without enzyme.
Km and Vmax values were determined from Lineweaver-Burk plots. Ki values were estimated by Dixon plots.
1H NMR ExperimentsThe hydrolysis of
KDN2MeUmb by KDNase Sm was monitored by 1H NMR
spectroscopy as described previously (21). The reaction conditions were
described in the legend for Fig. 3. A spectrum of KDN
2MeUmb, without
KDNase Sm, was also acquired under identical spectral conditions for
the purpose of providing a "zero time" spectrum.
Under identical conditions
to those used in the kinetic runs, KDN was released from all
substrates, linearly, within 1 h. Michaelis constants were
determined for KDN2MeUmb, (KDN)2, and KDN-lactose from
Lineweaver-Burk plots and are summarized in Table I. The
Vmax/Km value for
KDN
2MeUmb (0.033 min
1) was approximately 1.5 and 3.3 times greater than the values of (KDN)2 (0.023 min
1) and KDN-lactose (0.0099 min
1),
respectively, indicating that KDN
2MeUmb is the best substrate for
KDNase Sm.
|
KDN2en strongly inhibited the KDNase Sm activity
toward KDN2MeUmb; however, no inhibition was observed for either
Neu5Ac2en or Neu5Gc2en as shown in Fig. 1A.
Dixon plot analysis, Fig. 1C, showed that KDN2en was a
competitive inhibitor of KDNase Sm with a Ki value
of 7.7 µM. Notably, this value was smaller than the
Km value (42 µM) for KDN
2MeUmb.
When (KDN)2 is used as a substrate, KDN2en inhibited the
KDNase activity (Fig. 1B) competitively with a
Ki = 8.1 µM; however, neither Neu5Ac2en nor Neu5Gc2en had inhibitory effects at concentrations up to
100 µM (Fig. 1A). For both KDN
2MeUmb and
(KDN)2 substrates, no inhibition was observed with
4-deoxy-KDN2en at concentrations up to 100 µM (Fig.
1, A and B).
Effects of Free Sialic Acid on KDNase Sm Activity
Free KDN
had an inhibitory effect on KDNase Sm activity toward KDN2MeUmb,
whereas free Neu5Ac and Neu5Gc had no effect (data not shown). The
concentration required for 50% inhibition of KDNase Sm activity, using
the experimental conditions described, was 7.1 mM for KDN,
compared with 17 µM for KDN2en; therefore inhibition by
KDN was significantly weaker than KDN2en.
As previously reported, Neu5Ac2en was a strong
inhibitor for exosialidases from A. ureafaciens
(Fig. 2A) and C. perfringens (Fig.
2B) (32). Neu5Gc2en showed inhibitory effects on the
activity of both enzymes but to a lesser extent than Neu5Ac2en.
Interestingly, Neu5Gc2en displayed quite different effects on these two
exosialidase activities, with C. perfringens exosialidase
being less sensitive to Neu5Gc2en (Fig. 2, A and
B). No inhibition was observed for these bacterial enzymes
with KDN2en (Fig. 2, A and B) and 4-deoxy-KDN2en (data not shown) at concentrations up to 100 µM.
1H NMR Experiments
Fig. 3 shows a
time course reaction of the hydrolysis of the synthetic substrate
KDN2MeUmb catalyzed by KDNase Sm. The progress of the reaction is
conveniently monitored by 1H NMR chemical shifts of the H-3
methylene protons of KDN
2MeUmb and its hydrolysis products
-KDN
and
-KDN that resonate in a spectral region free of other KDN
resonances. The anomeric configuration of the hydrolysis products is
easily determined by the chemical shift of the H-3ax and
H-3eq protons. The first spectrum at t = 0 min shows the spectral region from 1.4 to 3.0 ppm of KDN
2MeUmb at
310 K without addition of KDNase Sm. In this region of the spectrum the
H-3eq and H-3ax protons of KDN
2MeUmb are
clearly seen at 2.84 and 1.98 ppm, respectively, and the methyl
resonance of the 4-methylumbelliferyl aglycon is at 2.49 ppm.
The remaining spectra show the hydrolysis reaction after the addition
of the KDNase Sm. At t = 5 min, resonances at 2.69 and 1.60 ppm are seen which are due to the H-3eq and
H-3ax protons, respectively, of the -anomer of KDN.
These signals increase in intensity as the time course proceeds, and at
the same time the intensity of the H-3 proton signals at 2.84 and 1.98 ppm decrease as KDN
2MeUmb is hydrolyzed. Eventually, after about 36 min hydrolysis of KDN
2MeUmb is complete, and the signals at 2.69 and
1.60 ppm completely disappear. The resonances at 2.69 and 1.60 ppm
steadily increase in intensity until 36 min and then decrease steadily after this time. This is due to conversion of the initial reaction product
-KDN to
-KDN. H-3eq and H-3ax
protons of
-KDN are visible at 2.18 and 1.81 ppm, respectively.
Isomerization of
-KDN continues until final equilibrium values for
the anomeric mixture of approximately 90%
- and 10%
-KDN (as
determined by integration of the H-3 signals) are established.
On the basis of the
Vmax/Km values (Table I), the
synthetic substrate (KDN2MeUmb) was demonstrated to be a better substrate for KDNase Sm than the natural substrates, KDN
2
8KDN and
KDN
2
3Gal
1
4Glc. This is presumably because the synthetic 4-methylumbelliferyl aglycon is a better leaving group than the naturally occurring leaving groups.
KDN2en is a potent competitive inhibitor for the KDNase Sm-catalyzed
reaction as revealed by Dixon plots (Fig. 1C). We propose that KDN2en is a transition-state analogue of KDNase Sm because it
fulfills the following criteria (33, 34). (i) A transition-state inhibitor binds more tightly to the enzyme than the substrate; (ii) the
structure of the inhibitor mimics that of the substrate in the putative
transition-state during the enzyme-catalyzed reaction. The
Ki values for KDN2en are significantly smaller than the Km values for the substrates used,
i.e. KDN2MeUmb (Ki = 7.7 µM versus Km = 42 µM)
and (KDN)2 (Ki = 8.1 µM
versus Km = 480 µM), indicating that
KDNase Sm binds more favorably to KDN2en than these substrates. By
analogy with 2-deoxy-2,3-didehydro-Neu5Acyl derivatives,
i.e. Neu5Acyl2en derivatives, for
N-acylneuraminidases, KDN2en can be referred to as a
transition-state analogue for KDNase Sm. Interestingly, KDNase Sm and
exosialidases were strongly inhibited only by their cognate
transition-state analogues, indicating that the binding site in each
enzyme is complementary and specifically accommodates the C-5
substituent of the substrates and inhibitors and is an important
determinant for enzyme specificity. Data supporting this view include
the following: (a) Neu5Gc2en had no inhibitory effect on
KDNase Sm; however, it exhibited a similar effect on other exosialidase
as Neu5Ac2en; and (b) hydrolysis of KDN
2MeUmb by KDNase
Sm was inhibited in the presence of KDN but not Neu5Ac or Neu5Gc. These
results suggest that the active site of KDNase Sm is unable to
accommodate an N-acyl group at C-5 of sialic acid residues.
Also our studies with the C-4 deoxygenated KDN2en derivative 4-deoxy-KDN2en, which did not inhibit the KDNase Sm activity, provide
evidence for the notion that a hydrogen bond donor or acceptor is
required at the C-4 position of KDN2en for enzyme recognition.
The stability of KDN and Neu5Acyl may be attributed to both electronic
and steric factors that provide a satisfactory interpretation of the
kinetic and equilibrium results. It is likely that the initial step in
KDNase- or sialidase-catalyzed hydrolysis of -ketosides involves
rapid removal of their aglycon moiety at C-2. An attack on the putative
sialosyl cation by a nucleophile such as a water molecule is
facilitated by the partial positive charge on C-2. It is therefore
convenient to discuss reaction in terms of sialosyl cation
(K+/enzyme) formation. From analysis of the experimental
results depicted in Fig. 3, it may be concluded that hydration of the unstable intermediate (K+/enzyme) takes place initially
from below the plane formed by -COO
group/C-6/Oring/C-2/C-3 to give the product
-anomer of
free KDN. Formation of the
-anomer proceeds rapidly
(k
> k
) indicating
that the free energy of activation is low compared with the
corresponding value associated with the
-anomer. This lower
activation energy may simply be due to steric hindrance around the
-face of the cation in the active site, providing a complete
stereofacial selectivity in favor of the
-face. Our mechanistic
studies with influenza virus sialidase demonstrated that formation of
the
-Neu5Ac was a consequence of the steric hindrance to the
-face of the sialosyl cation (19, 20).
Mutarotation of the -anomer then occurs leading to a final
equilibrium mixture of two species comprising of the
-isomer (~10%) and thermodynamically preferred
-isomer (~90%) of KDN. It is well known that the most thermodynamically stable isomer is not
always the initial product of a chemical reaction and that less stable
isomers may be formed first (26), which is the case for the KDNase
Sm-catalyzed hydrolysis. KDNase Sm- or sialidase-catalyzed reversible
hydrolysis is represented schematically in Fig. 4.
In this investigation we have highlighted the similarities and differences of a new class of sialidase (KDNase Sm) and other known bacterial exosialidases. To fully understand the catalytic mechanism of action of KDNase Sm, however, a crystal structure investigation of this sialidase is necessary, which is currently being pursued in our laboratories.