(Received for publication, September 16, 1996, and in revised form, December 20, 1996)
From the Tokyo Research Institute, Crude enzyme obtained from chondroitin
sulfate-induced Proteus vulgaris NCTC 4636 has been
fractionated into 1) an endoeliminase capable of depolymerizing
chondroitin sulfate and related polysaccharides to produce, as
end products, a mixture of Chondroitin sulfate ABC lyase (EC 4.2.2.4) was first purified from
extracts of Proteus vulgaris NCTC 4636 adapted to
chondroitin 6-sulfate (1). It is believed to be an endoeliminase that
splits There is a commercially available chondroitin sulfate ABC lyase
("chondroitinase ABC" from P. vulgaris, the product of
Seikagaku Corp.) which has found wide applications including the
quantification of chondroitin sulfate and dermatan sulfate (12), the
structural analysis of the carbohydrate moiety of proteoglycans
(13-16), the preparation of core proteins from proteoglycans (13, 17), the formation of antigenic epitopes to prepare anti-proteoglycan monoclonal antibodies (18), and the medical use for intervertebral disc
protrusion (19, 20). Because of the increased use of the enzyme, more
attention has been directed in our laboratories toward understanding
its purity, specificity, and mode of action. A purification procedure
yielded a chondroitin sulfate-degrading enzyme with a higher specific
activity whose end products were tetrasaccharides in addition to
disaccharides (21, 22). In the meantime, Sugahara et al.
(23) reported that a highly purified chondroitinase ABC preparation
(commercially available as "protease-free chondroitinase ABC") was
unable to degrade the unsaturated tetrasaccharides irrespective of
their sulfation patterns. These findings suggested the possibility that
the conventional chondroitinase ABC preparation contained a
tetrasaccharide-degrading enzyme that may have been removed during the
process to prepare the protease-free enzyme.
We report here the isolation and crystallization from P. vulgaris of two distinct eliminases, one to catalyze the endolytic cleavage of chondroitin sulfate and related compounds and the other to
split off disaccharide residues from the nonreducing ends of both
polymeric chondroitin sulfates and their oligosaccharide fragments
produced by the endoeliminase.
The following compounds were the
products of Seikagaku Corp., Tokyo: hyaluronan from human umbilical
cords; chondroitin prepared by desulfation of chondroitin 4-sulfate;
chondroitin 4-sulfate from whale cartilage (75% 4-sulfate and 20%
6-sulfate); chondroitin 6-sulfate from shark cartilage
(Mr 20-80 × 103, 15%
4-sulfate, 75% 6-sulfate, and 10% 2,6-disulfate); chondroitin sulfate
D from shark cartilage (20% 2,6-disulfate, 25% 4-sulfate, and 45%
6-sulfate); chondroitin sulfate E from squid cartilage (60%
4,6-disulfate, 25% 4-sulfate, and 10% 6-sulfate); dermatan sulfate
from pig skin (90% 4-sulfate, 2% 6-sulfate, and 7% 2,4-disulfate); keratan sulfate from bovine cornea, and heparan sulfate from bovine kidney. Heparin from porcine intestinal mucosa and bovine serum albumin
were obtained from Sigma; POELE1 was from
Nikko Chemicals Co., Tokyo; and CM-Sepharose FF was from Pharmacia
Biotech Inc. Chondroitin sulfate tetrasaccharides and hexasaccharides
were prepared from chondroitin 6-sulfate by testicular hyaluronidase
digestion followed by Cellulofine GCL-90sf gel chromatography,
according to the method of Rodén et al. (24). Chondroitin sulfate proteoglycan (A1D1 fraction) was prepared from
bovine nasal septum cartilage by the procedure described by Oegema
et al. (25).
ChS ABC lyase activity was measured based on
the increase in A232 essentially as described by
Yamagata et al. (1). A typical incubation mixture contained
a measured quantity of enzyme, 150 µg of poly- or oligosaccharide
substrate, 10 µmol of sodium acetate, and 5 µg of casein in 250 µl of 40 mM Tris-HCl, pH 8.0. Control medium contained
heat-inactivated enzyme in the above mixture. After incubation at
37 °C for 20 min, the reaction was stopped by heating for 1 min in a
boiling water bath. The reaction mixture was then diluted with 10 volumes of 50 mM HCl, and UV absorption of the resultant
solution was measured at 232 nm against the corresponding blank
mixture. One unit of enzyme is the amount required for the eliminative
cleavage of substrate, yielding UV-absorbing materials corresponding to
1 µmol of P. vulgaris NCTC 4636 was
routinely grown at 30 °C in the medium containing peptone, meat
extract, yeast extract, and NaCl, as described previously (1), with
chondroitin 6-sulfate (5 g/liter) as the inducer. The cells were
harvested by centrifugation (16,000 × g for 30 min at
4 °C) at the end of logarithmic phase. Approximately 150 g (wet
weight) of cells were obtained in a 20-liter fermentation. A 100-g
portion of collected cells was suspended in 400 ml of 5% (w/v) POELE
in 20 mM phosphate buffer, pH 7.0, and allowed to stand at
35 °C with gentle stirring for 2 h. The resultant suspension
was diluted with an equal volume of 20 mM phosphate buffer,
pH 7.0, and centrifuged at 16,000 × g for 45 min at
4 °C. Cold water was added to the supernatant fluid, bringing the
final concentration of POELE to 1.5% (w/v), and the solution was
applied to a CM-Sepharose column (3 × 15 cm) equilibrated with 10 mM phosphate buffer, pH 7.0. The column was washed
successively with each 100 ml of 0.5% (w/v) POELE in water and 40 mM phosphate buffer, pH 6.2. The column was then subjected
to elution with a linear gradient of 0-0.2 M NaCl in 40 mM phosphate buffer, pH 6.2. The total volume of gradient
elution was 320 ml. The flow rate was 1.6 ml per min, and 5-ml
fractions were collected. The fractions containing the major portion of
each enzyme peak were pooled and dialyzed against cold water for
desalting. To the resultant nondiffusible fraction was added 30% (w/v)
polyethylene glycol 4000 in 10 mM phosphate buffer, pH 7.0, until a slight turbidity appeared. It was then allowed to stand in a
cold room until crystals appeared.
Amino acid analyses were performed on a
Hitachi L-8500 amino acid analyzer. Samples were hydrolyzed at
110 °C for 24 h in gaseous 6 M HCl under nitrogen
vacuum. N-terminal amino acid residue was identified by Edman
degradation using Tosoh phenylthiohydantoin-derivative analytical HPLC
system. SDS-PAGE was carried out in a 10% (w/v) polyacrylamide gel
under reducing conditions. Before electrophoresis, sample solutions
were diluted with sample buffer (1.2% w/v SDS, 1.0% (w/v)
dithiothreitol, 50 mM Tris-HCl, pH 6.8, and 10% (w/v) glycerol) at 100 °C for 2 min. The molecular weight markers
consisted of myosin (Mr 212,000),
Isoelectric point was determined on a gel of PhastGel isoelectric
focusing pH 3-9 by PhastSystemTM (Pharmacia Biotech Inc.)
using a pI calibration kit 3-10 as a standard.
Immunochemical examination of enzymes was performed with rabbit
antiserum prepared against each purified enzyme, according to the
method of Ouchterlony (27). To prepare the antisera, white rabbits
(18-week-old males) were immunized with two injections of 1 mg of each
enzyme emulsified in complete Freund's adjuvant (Yatoron Co., Tokyo)
by an interval of 1 week. At three subsequent intervals, a booster
injection of approximately 0.5 mg of enzyme protein was given. The
animals were bled 7 days later, and the sera were pooled. To the
antiserum was added ammonium sulfate, and the precipitate emerging with
the range from 20 to 50% saturation was collected, dissolved in 50 mM sodium phosphate buffer, pH 7.2, and dialyzed for
20 h against three 2-liter changes of the same buffer.
To measure the change in viscosity and to determine the size
distribution of oligosaccharide intermediates produced during enzymatic
digestion, assay solutions were prepared that contained 10 mg of
glycosaminoglycan per ml of 50 mM Tris-HCl buffer
containing 50 mM sodium acetate, pH 8.0, and allowed to
equilibrate for 10 min in 37 °C water bath. Sufficient enzyme was
then added so that there was 0.015 unit of enzyme per mg of substrate,
and incubation was continued at 37 °C. At specific times, an aliquot
was withdrawn and heated in a boiling water bath for 1 min. Then the
aliquot was subjected to the following three measurements. The
viscosity of the aliquot diluted with an equal volume of 0.4 M NaCl was measured in an Ubbelohde-type automatic
viscometer VMC-052 (Rigosha, Tokyo) at 30 °C. Each of the heated
aliquots was diluted with 50 volumes of 50 mM HCl and 40 volumes of 0.2 M NaCl, respectively. The former was used
for A232 measurement and the latter for gel permeation HPLC to determine the size distribution of oligosaccharides generated at the specific times. The analytical HPLC columns used were
TSKgel G3000PWXL and G2500PWXL connected in
this order. Gel permeation HPLC was performed with a flow rate of 0.5 ml per min with 0.2 M NaCl at 35 °C. The elution was
monitored with Tosoh UV (232 nm) monitor UV8010 and Tosoh differential
refractometer RI8010 in this order.
In initial experiments we noticed
that ChS ABC lyases could efficiently be extracted from intact P. vulgaris cells at 35 °C with 5% (w/v) POELE, a nonionic
surfactant, in 20 mM phosphate buffer, pH 7.0 (see Table
I, Footnote a). Fig. 1 shows
the fractionation of the 16,000 × g supernatant of
POELE extracts by CM-Sepharose column chromatography. A large portion
of the contaminant proteins including possible proteolytic enzymes and
sulfatases were passed through and removed by washing the column with
0.5% (w/v) POELE in H2O and then with 40 mM
phosphate buffer, pH 6.2. Little or no chondroitin sulfate lyase
activity was detected in the washings (data not shown). Subsequent
elution of the washed column with a linear NaCl gradient yielded two
major protein peaks (I and II) that accounted for ~70 and ~25%,
respectively, of the total protein eluates. When aliquots of each
fraction were checked for lyase activity on chondroitin 6-sulfate and
on the tetrasaccharide prepared from it, the elution curves were
obtained which suggest that the protein peaks I and II may actually
represent different lyases. The peak I enzyme acted on the chondroitin
6-sulfate substrate but not on the tetrasaccharide substrate. In
contrast, the peak II enzyme acted on the tetrasaccharide substrate at
about 5 times higher initial rate than on the polysaccharide substrate.
The fractions collected in peak II displayed similar ratios of activity against polysaccharide and tetrasaccharide substrates, suggesting that
both reactions are catalyzed by a single enzyme. The fractions containing the major portion of each enzyme peak were pooled for further analyses; hereinafter the peak I and peak II enzyme will be
referred to as ChS ABC lyase I and II, respectively. A summary of the
purification procedure is shown in Table I. SDS-PAGE under reduced
conditions illustrated that the peak I and peak II proteins were
apparently homogeneous (Fig. 2). On the basis of their
mobilities relative to those of molecular weight
(Mr) standards, the molecular weight of ChS ABC
lyase I and II were estimated to be 100 × 103 and
105 × 103, respectively. Both peak I and peak II
enzyme proteins were crystallized from solutions of polyethylene glycol
(Fig. 3). No significant reduction was observed in their
specific activity by repeating recrystallization, suggesting that the
enzyme preparations are at a high state of purity. However, isoelectric
focusing of the two enzymes on PhastGel IEF with a pH gradient of
3.50-9.30 showed that ChS ABC lyase II migrated as a single band,
pI = 8.45, whereas ChS lyase I migrated as a doublet of bands,
pI = 8.25 and 8.50. Although the results suggest the presence of
two isoforms of ChS lyase I, further work is needed to understand this
phenomenon.
Purification of ChS ABC lyase I and II
On double immunodiffusion analysis using anti-lyase I antiserum, a
precipitin line was detected against ChS ABC lyase I but not against
ChS ABC lyase II (data not shown). Using anti-lyase II antiserum, on
the other hand, a precipitin line was detected against ChS ABC lyase II
but not against ChS ABC lyase I. The results indicate that the two
enzymes are immunochemically distinct proteins.
Amino acid analyses of the two enzymes are shown in Table
II. The lyase I and lyase II proteins are similar in
overall compositional features, e.g. both have high Asx,
Glx, and Leu contents. The N-terminal amino acids of the lyase I and
lyase II protein were determined to be alanine and leucine,
respectively.
Amino acid composition of ChS ABC lyase I and II
Each analysis is expressed as residues/1,000. No corrections were made
for the loss of threonine or serine or the incomplete release of valine
during the acid hydrolysis. Tryptophan was not determined.
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENT
REFERENCES
4-unsaturated tetra- and
disaccharides and 2) an exoeliminase preferentially acting on
chondroitin sulfate tetra- and hexasaccharides to yield the
respective disaccharides. Isolation of the two enzymes was achieved by
a simple two-step procedure: extracting the enzymes from intact
P. vulgaris cells with a buffer solution of nonionic surfactant and then treating the extract by cation-exchange
chromatography. Each of the enzymes thus prepared was apparently
homogeneous as assessed by SDS-polyacrylamide gel electrophoresis and
readily crystallized from polyethylene glycol solutions. Both
enzymes acted on various substrates such as chondroitin sulfate,
chondroitin sulfate proteoglycan, and dermatan sulfate at high, but
significantly different, initial rates. They also attacked hyaluronan
but at far lower rates and were inactive to keratan sulfate, heparan sulfate, and heparin. Our results show that the known ability of the
conventional enzyme called "chondroitinase ABC" to catalyze the
complete depolymerization of chondroitin sulfates to unsaturated disaccharides may actually result from the combination reactions by
endoeliminase (chondroitin sulfate ABC endolyase) and
exoeliminase (chondroitin sulfate ABC exolyase).
1,4-galactosaminidic bonds between
N-acetylgalactosamine and either D-glucuronic
acid or L-iduronic acid and degrades, therefore, a variety
of glycosaminoglycans of the chondroitin sulfate and dermatan sulfate
type to the respective unsaturated disaccharides. Using this enzyme,
both chondroitin sulfates and dermatan sulfates were shown to contain,
in addition to predominant 4- or 6-sulfated disaccharide residues, a
smaller proportion of nonsulfated or disulfated disaccharide residues,
or both (2). A variety of studies have since shown that the proportion
of these disaccharide residues varies greatly with species and
anatomical sites, during development and aging, and in pathology (3-11
among others). Many of these data suggest that variation in the
carbohydrate sequence and sulfation pattern may be used to specify the
functional properties of chondroitin sulfate and dermatan sulfate
proteoglycans.
Substrates and Materials
4-hexuronate residues per min, as calculated
with a value of 5,500 for its molar absorption coefficient. Protein
concentrations were determined by the Lowry method (26) using a bovine
serum albumin as a standard.
2-macroglobulin (Mr 170,000),
-galactosidase (Mr 116,000), transferrin
(Mr 76,000), and glutamic dehydrogenase (Mr 53,000). After electrophoresis, the proteins
in gel were detected by Coomassie Blue staining.
Purification of the Enzymes
Preparation
Total proteina
On
chondroitin 6-sulfate
On tetrasaccharide
Total
activitya
Specific activity
Total
activitya
Specific activity
mg
units
units/mg
units
units/mg
POELE extract
16,000 × g
supernatant
8,200
43,900
5.4
8,850
1.1
ChS ABC lyase
I
CM-Sepharose, peak
I
115
37,500
326
NDb
NDb
After double
crystallization
76
25,700
338
NDb
NDb
ChS ABC lyase II
CM-Sepharose, peak
II
40
1,240
31
5,880
147
After double
crystallization
32
1,010
32
4,760
149
a
The data obtained with 100 g of wet cells are
shown. When the insoluble fraction obtained after POELE treatment was
disrupted with a sonicater at pH 7, the resultant extract showed total
activities of approximately 1,000 units on chondroitin 6-sulfate, the
values that are 40-50 times lower than those of the POELE extract.
b
ND, not detected (below level of detection).
Fig. 1.
Elution pattern from a CM-Sepharose column
(3 × 15 cm). Aliquots were assayed as described under
"Experimental Procedures" using two different substrates. The
activity (units/ml) toward chondroitin 6-sulfate () and activity
toward the tetrasaccharide from chondroitin 6-sulfate (
) are shown.
The broken line denotes the protein (×). The
bars designated as I and II represent
the fractions that were pooled for crystallization of ChS ABC lyase I
and II, respectively.
[View Larger Version of this Image (19K GIF file)]
Fig. 2.
SDS-PAGE of peak I and peak II protein under
reducing conditions. POELE extract from P. vulgaris
(lane a), peak I (lane b), peak II (lane
c), and molecular weight standards (lane d) were
applied.
[View Larger Version of this Image (65K GIF file)]
Fig. 3.
Microphotographs of crystalline ChS ABC lyase
I and II. The microphotographs were taken by illumination with
transmitted light Nomarski differential interference contrast
attachment.
[View Larger Version of this Image (143K GIF file)]
Amino acid
ChS ABC lyase I
ChS ABC lyase II
Asx
135
125
Thr
64
75
Ser
74
58
Glx
109
125
Pro
44
41
Gly
66
61
Ala
69
86
1/2Cys
0
2
Val
45
41
Met
23
25
Ile
57
50
Leu
98
106
Tyr
42
41
Phe
42
38
Lys
71
62
His
24
29
Arg
37
35
To measure the lyase activity, chondroitin 6-sulfate and the tetrasaccharide prepared from it were chosen as substrate (for relative activities on other substrates, see below). Unless otherwise indicated, an incubation time of 20 min was selected.
The pH optima determined at 37 °C in 40 mM Tris-HCl
buffer were around 8.0 for both ChS ABC lyase I and II (Fig.
4).
The optimum temperatures for activity determined at pH 8.0 were 37 °C for lyase I and 40 °C for lyase II, respectively. The small temperature difference between ABC lyase I and II may reflect, at least in part, their temperature stability optima at pH 8.0. Thus, the activity losses at 37 and 40 °C in 30 min were 39 and 47%, respectively, for lyase I and 10 and 23%, respectively, for lyase II (data not shown).
Both ChS ABC lyase I and II were almost completely inhibited by Zn2+ at 1 mM. At the same cation concentration, Ni2+, Fe2+, and Cu2+ produced 80-90% inhibition of lyase I, whereas they had little or no effect on lyase II. Ca2+, Mg2+, Ba2+, and Mn2+ showed no effect on both lyase I and II.
Both enzymatic reactions of lyase I and II to chondroitin 6-sulfate exhibited typical saturation behavior. Analysis of their Lineweaver-Burk plots (not shown) exhibited an apparent Km of 66 µM and a Vmax of 310 µmol/min/mg of protein for lyase I and an apparent Km of 80 µM and a Vmax of 34 µmol/min/mg of protein for lyase II, based on hexuronate residues. In a similar way, an apparent Km and a Vmax value for the lyase II acting on tetrasaccharide were shown to be 33 µM and 155 µmol/min/mg of protein, respectively, based on hexuronate residues.
Substrate SpecificityThe specificity of ChS ABC lyase I and II was determined using various polysaccharide substrates (Table III). The chondroitin sulfate/dermatan sulfate samples were degraded at high, but significantly different, initial rates. In contrast, hyaluronan was degraded at very low rates with both lyase I and II. Even on the same glycosaminoglycan substrate, a significant difference in activity was observed between ChS lyase I and II, e.g. lyase I degraded both chondroitin and chondroitin sulfate E at one-third to one-forth the rate for chondroitin 6-sulfate, whereas lyase II degraded these three glycosaminoglycans at comparable rates. Keratan sulfate, heparan sulfate, and heparin were not substrates for lyase I or II. The activity of the enzymes was greatly influenced by the size of chondroitin sulfate chain. The specific activities of ChS ABC lyase I on the poly-, hexa-, and tetrasaccharide substrates were 330, 75, and 0 units/mg of protein, respectively. In contrast, the activity of ChS ABC lyase II increased with the decrease of chain size, 30, 110, and 150 units/mg of protein on the poly-, hexa-, and tetrasaccharide substrates, respectively.
|
The lyase-catalyzed breakdown of chondroitin
6-sulfate was followed by using both UV and viscosimetric assays. While
the change in A232 reflects the number of sites
cleaved, the change in viscosity is an indicator of the change in
average polysaccharide chain length. As Fig. 5,
a and b shows, a marked difference between ChS ABC
lyase I and II in the rates of viscosity decrease was observed. Thus,
lyase I caused a rapid drop in viscosity even before any extensive
cleavage has occurred (a). Lyase II, in contrast, showed a
slow viscosity change that was approximately proportional to the
increase of A232 (b). The results
indicate that ChS ABC lyase I acts endolytically on chondroitin
6-sulfate, whereas ChS ABC lyase II acts on the same substrate in an
exolytic fashion.
When acting on dermatan sulfate, lyase I again gave a concave curve of viscosity change (Fig. 5c), but its shape is significantly shallower than that for chondroitin 6-sulfate (a). This difference appears to be primarily the result of the difference of initial viscosity of each polymer, 0.22 and 0.85 for 0.5% (w/v) solution of dermatan sulfate and chondroitin 6-sulfate, respectively. ChS ABC lyase II, on the other hand, acted on dermatan sulfate with an ideal exolytic pattern (d). Nearly identical curves with an endolytic and an exolytic action pattern were given by lyase I and II, respectively, when acting on chondroitin 4-sulfate whose initial viscosity is 0.28 (0.5%, w/v) (data not shown). These results, coupled with the size distribution patterns of digest products (see below), show that lyase I acts on polysaccharides of the chondroitin sulfate/dermatan sulfate type in an endolytic fashion, while lyase II acts in an exolytic fashion, irrespective of their sulfation/5-epimerization pattern.
Size Distribution of Oligosaccharide ProductsThe size
distribution of oligosaccharides generated at different digestion
stages was determined by gel permeation HPLC monitored with both a
differential refractometer and UV (232 nm) monitor (Fig.
6).
Panels b-b and c-c
show that, in relatively
short-term incubation, ChS ABC lyase I acted on chondroitin 6-sulfate
to give rise to a wide range of differently sized
4-oligosaccharides and that their average size was
progressively reduced with time. After exhaustive digestion (5 h), two
end products corresponding in size to tetra- and disaccharides were
detected (d-d
). In contrast, the digest products by ChS ABC
lyase II were of disaccharide size throughout the digestion periods
(f-f
, g-g
, and h-h
).
Gel permeation HPLC was used to further confirm the action patterns of
ChS ABC lyase I and II on dermatan sulfate. As shown in Fig.
7, relatively short term digestion with lyase I resulted in the appearance of differently sized intermediates (b-b
and c-c
), whereas the digest products by lyase II were of
disaccharide size throughout the digestion time (f-f
,
g-g
, and h-h
). The results are nearly identical
with those obtained by digestion of chondroitin 6-sulfate (see above),
except that the intact dermatan sulfate sample chromatographed as a
smaller polysaccharide than did the intact chondroitin 6-sulfate sample
(Fig. 7, a and e; cf. Fig. 6,
a and e). Also to be noted is the appearance of a shoulder just before, but overlapping, tetrasaccharides even after a
5-h digestion with lyase I (Fig. 7, d-d
, retention time of ~ 29 min). This fraction was not studied further. Since, however, the
proportion of the reducing terminal portion to the interior portion
should be high in relatively short glycosaminoglycan chains, such as
the dermatan sulfate used here, it would be possible that the shoulder
represents oligosaccharides derived from the reducing terminal portion
of dermatan sulfate where some non-cleavable sites are present. When
digest products of chondroitin 4-sulfate were examined in a similar
way, their size distribution patterns on HPLC were nearly the same as
those of dermatan sulfate digest products shown in Fig. 7 (data not
shown).
Using the hexasaccharide substrate prepared from chondroitin 6-sulfate
by testicular hyaluronidase digestion, the products at 100% reaction
completion with ChS ABC lyase I and II were analyzed in a similar way
(Fig. 8). The lyase I digest of hexasaccharide was
composed of two fractions corresponding in size to tetra- and
disaccharide (b). In this case, the disaccharide fraction exhibited little A232 (b),
indicating that the enzyme removed saturated disaccharide residue from
the nonreducing end of hexasaccharide thereby leaving unsaturated
tetrasaccharide. Panels d-d
show that the digest products
by ChS ABC lyase II were exclusively disaccharides. In a separate
experiment, the chondroitin sulfate moiety of bovine nasal chondroitin
sulfate proteoglycan (where the reducing ends of polysaccharide chains
are all linked to the core protein) was extensively cleaved by
digestion with lyase II; over 95% of the total hexuronate residues
were converted to unsaturated disaccharides at an initial rate
comparable to the rates of degradation of free chondroitin sulfates
(see Table III).
Collectively, the results in Figs. 6, 7, 8 are entirely consistent with the data of viscosity change (Fig. 5) and provide a sound basis for the characteristic reaction mode of lyase I and II action; the former acts endolytically on chondroitin sulfate/dermatan sulfate, whereas the latter functions as a typical exoeliminase capable of splitting off disaccharide residues from the nonreducing end of chondroitin sulfate/dermatan sulfate and related oligosaccharides.
The major finding in this report is that the extract of P. vulgaris contains two distinct enzymes that are both able to degrade chondroitin sulfate and dermatan sulfate. This finding was accomplished by introducing the new purification methodology including 1) the use of POELE buffer for extracting the enzymes from intact P. vulgaris cells and 2) the use of two different substrates, chondroitin 6-sulfate and the tetrasaccharide prepared from it, for monitoring the elution from CM-Sepharose (Fig. 1). Each of the resultant enzymes was apparently homogeneous as assessed by SDS-PAGE (Fig. 2) and readily crystallized from polyethylene glycol solutions (Fig. 3).
As shown by the high yields of enzymes (Table I, Footnote a), the POELE buffer extraction is a key step in the purification procedure. The POELE treatment of intact P. vulgaris cells had almost negligible effects on both morphology and total dry weight of the cells (data not shown), suggesting a high degree of selectivity of POELE buffer solution in releasing chondroitinase enzymes from the cells.
The failure in early studies to distinguish the two enzymes may be due, in part, to the molecular resemblance between the two enzyme proteins (see Fig. 2 and Table II). It is also pertinent that the early studies did not use sulfated tetra- or hexasaccharides as substrate for survey of chondroitin sulfate-degrading enzymes, whereas here these oligosaccharide substrates are most effective in distinguishing the exoeliminase from the endoeliminase (Fig. 1).
Evidence from early studies on P. vulgaris chondroitinase suggested that the depolymerization of both chondroitin sulfate and dermatan sulfate to the respective unsaturated disaccharides involves an endoeliminative attack by a single enzyme (28, 29). Using both UV and viscometric assays, Jandic et al. (29) showed that the conventional chondroitinase ABC preparation deviates from an ideal endolytic mode when acting on bovine trachea chondroitin 4-sulfate (product of Celsus Laboratories, Inc.; 60% 4-sulfate and 36% 6-sulfate). One interpretation by these investigators is that the shallowness of the concave curve of viscosity change may be attributable to a limited number of cleavable sites present in this polysaccharide or to a non-random distribution of cleavable sites within the polysaccharide chain. However, it is also possible that the failure of the chondroitin 4-sulfate sample to give a steep concave curve may be due to the relatively small size of the substrate molecules, since our data showed that the shallowness of the concave curve given by lyase I was influenced by the initial viscosity of each polysaccharide and that a decrease in the polysaccharide chain size caused an increase in the specific activity of lyase II concomitantly with a decrease in the lyase I activity. Whatever explanation is correct, these results suggest that more attention should be focused on the structural diversity of polysaccharides of the chondroitin sulfate family. On the basis of the data obtained in the present study, we concluded that the complete depolymerization to disaccharides cannot be achieved by the endoeliminase alone. Our results strongly suggest that it is the combination reactions by endoeliminase (ChS ABC lyase I) and exoeliminase (ChS ABC lyase II) that results in the complete depolymerization, irrespective of their sulfation/5-epimerization pattern (Table III and Figs. 5, 6, 7, 8).
A clue to the nature of P. vulgaris chondroitinase ABC was advanced recently by Ryan et al. (30). These investigators reported the cloning of cDNAs encoding chondroitinase ABC subcomponents, 110- and 112-kDa proteins, which are copurified by extensive chromatographic procedures (following mechanical or enzymatic cell disruption) but are separated by SDS-PAGE under reducing conditions. The deduced amino acid sequences of the two components show that the 110-kDa component is composed of 997 amino acid residues with alanine at the N-terminal, whereas the 112-kDa component has a different sequence of 990 amino acid residues with leucine at the N-terminal. Independent of this work, Sato et al. (31) demonstrated by cDNA cloning the deduced amino acid sequence of a major protein component of P. vulgaris chondroitinase ABC which is closely related to the 110-kDa component as above. More recently, Khandke et al. (32) have reported that the 110- and 112-kDa components are separated by Macro Prep High S chromatography and that extensive digestion of chondroitin sulfate (source not shown) or proteoglycan (product of ICN Biochemical Costa Mesa) by the 110-kDa component generates three end products (di- and oligosaccharides), whereas the 112-kDa component by itself is not capable of cleaving the polymeric chondroitin sulfate/proteoglycan substrates. Since, however, a mixture of the two protein components digests all the material to the disaccharides, they suggested that depolymerization to the small oligosaccharide fragments is a prerequisite for the action of the 112-kDa component to yield disaccharides. Our demonstration that the POELE extract of P. vulgaris contained two separate chondroitin sulfate ABC lyases would seem to consist with this idea. However, in the present study, the specificity of ChS ABC lyase II (105 kDa) was found to be different from that of the 112-kDa component reported by Khandke's group. In contrast to the 112-kDa component, it was found that ChS ABC lyase II by itself was able to split off disaccharide residues directly from polymeric chondroitin sulfate/proteoglycan in an exolytic pattern. Since a lyase which acts only on small-sized oligosaccharide chains was not detected by the present methodology, additional studies are needed to fully understand the nature of chondroitin sulfate-degrading enzymes occurring in P. vulgaris cells.
Available evidence shows that all chondroitin sulfate chains are
polydisperse in molecular weight (10-350 × 103) and
that one chondroitin sulfate chain can often consist of various
proportions of mono- and disulfated disaccharide repeats, e.g. GlcAGalNAc (4- or 6-sulfate), GlcA
(2-sulfate)
GalNAc (6-sulfate), and GlcA
GalNAc(4, 6-disulfate)
(2). The identification of the specific sequences of those disaccharide
units that are responsible for their physiological activities is a most
interesting challenge that is developing in the chondroitin sulfate
field (6, 7, 10, 16, among others). It must be recalled in this respect
that it is difficult to obtain a pure chondroitin sulfate for sequence analyses. Therefore, detection of desired sequences will require the
partial cleavage of starting polysaccharide samples to obtain oligosaccharide fragments in which desired (bioactive) sequences will
be detected. The ability of ChS ABC lyase I to catalyze the endolytic
cleavage of chondroitin sulfate chains, coupled with its lack of
exolyase activity, will permit the use of this enzyme as an effective
reagent in preparing the tetra- and higher oligosaccharides for
structure-function studies. Recent analysis of the unsaturated tetrasaccharides derived from shark fin chondroitin sulfate (33) has
indeed shown that the profile of sulfation is much less random than has
previously assumed, e.g. the disaccharide units linked by
-1,4 to GlcA(2-SO4)
GalNAc(6-SO4) are
exclusively GlcA
GalNAc(4-SO4), and there are no
disulfated disaccharide units linking by
-1,4 to each other. The
highly purified lyase II preparation that splits off disaccharide
residues in an exolytic fashion from the nonreducing end may be used to
get more detailed sequence information.
The commonly used P. vulgaris chondroitinase enzymes are designated as "chondroitinase ABC" and "chondroitinase ABC, protease-free" by Seikagaku Corp. It is now clear that the the former is a mixture of ChS lyase I and II (and some proteases) and the latter a ChS ABC lyase I enzyme free of both lyase II and proteases. When cleaving proteoglycan samples by extensive digestion with these enzymes, we must consider the possibility that the resulting core preparations may be different in the size of chondroitin sulfate remnants attached. Hascall and Riolo (13) showed that limiting digestion of chondroitin sulfate proteoglycan (from bovine nasal cartilage) with the conventional chondroitinase ABC leaves a residual disaccharide on the chondroitin sulfate chains in the core protein preparation. It is possible then that the use of protease-free chondroitinase ABC in place of the conventional enzyme would result in the production of core preparations bearing tetra- or higher chondroitin sulfate oligosaccharide remnants. If this is so, the commercially available enzyme designated as "chondroitinase ABC, protease-free" may be used to prepare octa- or higher oligosaccharide chains (including the linkage tetrasaccharide) closest to core protein for structural analyses.
We propose that the endo- and exoeliminase are to be named chondroitin sulfate ABC endolyase and exolyase or, a shorter form, endochondroitinase ABC and exochondroitinase ABC, respectively.
We are grateful to Dr. Sanae Kato for her skillful technical assistance, to Dr. Keiichi Yoshida for his helpful suggestions and discussions, and to Keiko Morimoto for her help in the preparation of this manuscript.