THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 250, No. 18, Issue of September 25, pp. 7473-7480, 1975

Printed in U.S.A.

Mechanism of Action of Bovine Testicular Hyaluronidase

MAPPING OF THE ACTIVE SITE*

(Received for publication, February 24, 1975)

STEFAN HIGHSMITH,$ JAMES H. GARVIN, JR.,+ AND DAVID M. CHIPMAN

From the Department of Biology, Ben-Gurion University of the Negev, Beersheva, Israel

The reactions of purified, homogeneous bovine testicular hyaluronidase have been studied with radioactively labeled oligomers of hyalobiuronic acid, (GlcUA-GlcNAc),, as substrates and acceptors. Transglycosylation occurs by transfer of a glycosyl residue with retention of configuration from a leaving group to an acceptor. On the basis of detailed examination of cleavage and transglycosylation patterns for the trimer; comparison of trimer, tetramer, and polymer as substrates; comparison of acceptors; equilibrium binding; and other data, it is proposed that the enzyme’s active site consists of five subsites for hyalobiuronate residues. In the terminology of Schechter, I., and Berger, A. ((1966) Biochemistry 5, 3371), these are s~-s,-s’~-s’~-s~, where the reducing terminus is to the right, and cleavage occurs between s, and s’,. It is proposed that subsite s’~ has a high affinity for a substrate residue, while s, and s’, have low substrate affinity, and sZ and s’, are intermediate in affinity. This proposal is seen to have mechanistic implications.

The reactions of several substrates show similar bell-shaped pH dependences, with optima in the region of pH 5 to 5.5.

The classic work of K. Meyer, Weissmann, and others on testicular hyaluronidase (EC 3.2.1.35) in the 1950’s provided one of the first clear pictures of the complexities involved in the action of a glycanohydrolase (l-3). Transglycosylation plays a very important role in the action of this enzyme on both oligosaccharides and polymers (1, 2). The active site of the enzyme appears to be quite large, with its reactivity toward oligomers increasing with their size at least to the octasaccha- ride (GlcUA-GlcNAc),’ (1). The transfer of glycosyl residues also involves considerable specificity toward the acceptor (2). Such properties have since become a familiar characteristic of glycanohydrolases.

Detailed studies of other glycanohydrolases have led recently to interpretations of their reactions in terms of descriptions of their active sites on the molecular level (4-6). On the other hand, little has been added to our understanding of the mechanism of action of bovine testicular hyaluronidase since

*This research was supported by the United States National Institutes of Health (Grant AM-13590) and the Merck Company Foundation. This work was taken in part from the Ph.D. theses of S. Highsmith (Massachusetts Institute of Technology, 1972) and J. H. Garvin, Jr. (Massachusetts Institute of Technology, 1974). The major initial part of this research was carried out in the Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Mass.

$ Recipient of Predoctoral Fellowships from the National Institutes of Health from 1969 to 1972.

‘The abbreviations used are: GlcUA, o-glucuronic acid; GlcNAc, 2-acetamido-2-deoxy-o-glucose. Oligosaccharides are assumed to have the structures derived from hyaluronic acid, i.e. -GlcUA-j3(1+3)- GlcNAc-0(1+4)-GlcUA-. NFU, National Formulary Units (of hyalu- ronidase activity).

1960. A major barrier has been the difficulty in isolating the enzyme in a purified, well defined form. Borders and Raftery have described a procedure leading to essentially homogeneous enzyme (7), and the data now available concerning the molecular weight (7-9) and composition (7, 9) of the enzyme indicate that the protein molecule contains but a single active site (9). We report here detailed studies of the action of the purified enzyme on generally tritium-labeled oligosaccharides derived from hyaluronic acid (10). These experiments, and other data, provide further insight into the mechanism of action of the enzyme, and allow us to propose a tentative model for the structure and specificity of the active site of bovine testicular hyaluronidase.

METHODS

Bovine testicular hyaluronidase was purified essentially by the method of Borders and Raftery (7) from crude enzyme (type I) ob- tained from Sigma Chemical Co. Activity was assayed by the turbido- metric method of Di Ferrante (11) (using hexadecyltrimethylam- monium bromide), and expressed in National Formulary Units (NFU) by comparison with a reference activity standard from the American Pharmaceutical Association (Lot 60572). The details of the purifica- tion, assay, and examination of the homogeneity of the enzyme have been published elsewhere (9).

Hyaluronate oligosaccharides (GlcUA-GlcNAc). were obtained by partial enzymic digestion of hyaluronic acid (Sigma Chemical Co., Grade III-S from human umbilicus) and separation as described by Weissmann et al. (12), and further purified on Bio-Gel P-4 gel filtra- tion columns eluted with 25% acetic acid (13). Generally tritium- labeled oligomers were obtained by the digestion of hyaluronic acid which had been labeled by the tritium recoil labeling method (14), and were purified and analyzed as previously described (10). These

7473

by guest on June 20, 2018http://w

ww

.jbc.org/D

ownloaded from

7474

oligosaccharides have been shown to be at least 97% radiochemically pure, and evidence has been provided to indicate that the extent of labeling in each of the several -GlcUA-GlcNAc- moieties in an oligomer is equivalent (10).

Tritium-labeled GlcNAc-GlcUA-GlcNAc-GlcUA-GlcNAc was pre- pared from labeled (GlcUA-GlcNAc), by incubation with P-glucuroni- dase (15) (2 mg/ml, Sigma, 696 Fishman units/g) in 0.1 M ammonium acetate, pH 5.0, at 37” for 3 days, adding 2 mg/ml of fresh enzyme every 24 hours. The reaction mixture was then chromatographed on a Bio-Gel P-4 column (1.0 x 90 cm) eluted with 25% acetic acid (13). The oligosaccharide fraction was collected and lyophilized. Labeled GlcNAc-GlcUA-GlcNAc was prepared similarly from (GlcUA- GlcNAc) ?.

The labeled reduced oligosaccharides were produced from unlabeled saccharides by reduction with tritiated borohydride in 0.05 M sodium bicarbonate buffer at pH 7.5 (16). In each case, about 5 pmol of an oligosaccharide were stirred well with 0.9 mCi of tritiated sodium borohydride (New England Nuclear Corp., 30 Ci/mol) in 0.12 ml of buffer for 3 min and then 0.9 ml of freshly prepared 0.1 M NaBH, in buffer added. After 7 min more, the solution was acidified to pH 5 with 4 M acetic acid, and then 0.5 g of Dowex 50 W-X8 (H+ form) added. After 1 hour of stirring, the resin was filtered from the mixture, and the filtrate evaporated on a rotary evaporator under water pump vaccum. Small portions of methanol were added repeatedly and evaporated (to remove boric acid) until a yellow film remained. The residue was chromatographed on a Bio-Gel P-4 column (1.0 x 90 cm) eluted with 25% acetic acid, and l-ml fractions taken and tested for radioactivity. The center of the radioactive peak was pooled and lyophilized. The product in each case gave a negative reducing sugar test (17), under conditions where a 3% contamination with the original saccharide could have been detected easily. Paper chromatography (descending, l-butanol/acetic acid/water, 10/6/3 (v/v) on Whatman No. 1 paper) revealed a single radioactive spot in each case, with R, slightly greater than for the original unreduced saccharide. The reduced monomer, dimer, and trimer had specific activities of 80,000, 130,000 and 101,000 cpm/pmol, respectively, when counted in dioxane scintillation solution.

Ziinetic Studies-Kinetic studies were carried out as previously described (10). The generally labeled hyalobiuronate oligomers (about lo8 cpm/mg) were incubated with purified enzyme in 0.66 M ammo- nium acetate/O.02 M EDTA brought to pH 5.0 with HCl, or in 0.33 M

ammonium acetate/O.33 M ammonium carbonate/O.02 M EDTA at other pH values. All quantitative results reported here were obtained using the same stock solution of purified hyaluronidase (42,000

A sample of the transglycosylation-labeled trimer (0.9 mg, 1800 cpm) was incubated with P-glucuronidase under the same conditions as described above for the pentasaccharide preparation. At the end of the incubation, the hydrolysate was mixed with 0.1 ml of glacial acetic acid and 0.6 mg of unlabeled glucuronic acid, and chromatographed on a Bio-Gel P-4 column as above. No trimer remained, essentially all the radioactivity appeared in the GlcNAc-GlcUA-GlcNAc-GlcUA-GlcNAc peak, and the glucuronic acid peak contained less than 10 cpm above background. The labeled dimer was analyzed similarly (10).

pH Dependence of Polymer Hydrolysis-The pH dependence of polymer hydrolysis was studied by following the release of reducing end groups in the reaction of the enzyme with potassium hyaluronate. A stock solution of potassium hyaluronate (Sigma, type III-S, containing 95% of theoretical uranic acid content (19)) was diluted into appropri- ate buffers to yield solutions of 200 fig/ml of hyaluronate in 0.2 M

sodium acetate-acetic acid buffers, 0.15 M in NaCl, at pH values from 4 to 7. Two-milliliter reaction mixtures were incubated at 37” and 2 c(g of purified enzyme (42,000 NFU/mg) added in 10 pg of water. Aliquots of 50 rg were removed with a microsyringe, and assayed by the method of Park and Johnson (17) with glucose as standard. The initial rates were determined from the results at 0, 15, and 30 min, and converted into units of moles of end groups released by comparison with the color yield of GlcUA-GlcNAc and (GlcUA-GlcNAc), in the Park-Johnson assay. When this assay of activity is compared with the turbidimetric assay for hyaluronidase (ll), the results are found to be closely parallel.

Binding of Labled Oligosaccharides-The binding of the labeled oligosaccharides to the purified hyaluronidase was measured by equilibrium dialysis in micro cells using pretreated cellulose dialysis membranes (20) as previously reported (9). The measurements were carried out at 24” in a buffer of 0.2 M sodium acetate-O.15 M sodium chloride-O.005 M EDTA at pH 6.0. After times which had been shown to be sufficient for equilibrium, duplicate samples were withdrawn from each side of the cell for counting in a dioxane scintillation cocktail, and protein concentrations checked by the method of Lowry et al. (21). Binding constants were determined from calculated concentrations of free and enzyme-bound saccharide and free enzyme.

Inhibition of Enzyme-The inhibition of the enzyme by added saccharides was studied by using the method of Di Ferrante (11) to follow the reaction of the eniyme on 200 fig/ml of hyaluronic acid in the presence of added compounds.

RESULTS

NFU/mg) and the same stock solutions of oligomers. Aliquots of reaction mixtures were analyzed by paper chromatography (descend-

Kinetics of Hydrolysis-The reactions of generally tritium-

ing, 1-butanol/acetic acid/water, 10/3/7, on Whatman No. 1 paper) and labeled hyalobiuronate trimer (GlcUA-GlcNAc), and tetramer scintillation counting of sections of the chromatograms in dioxane (GlcUA-GlcNAc), with purified, homogeneous hyaluronidase scintillation solution (10). The other compounds whose preparations (42,000 NFU/mg), were studied at 37” in a buffer consisting of are given above were studied by similar methods.

Transglycosylation Studies-Transglycosylation studies were also 0.66 M ammonium acetate and 0.02 M EDTA at pH 5.0. The

carried out by the same general method. In order to study the product activity of the enzyme toward oligosaccharides is much lower,

of transglycosylation with labeled disaccharide as acceptor, 12 mM and variable, in the absence of EDTA, but is independent of generally labeled (GlcUA-GlcNAc), (8 x lo5 cpm/mg) and 4 mM EDTA concentration in this range. EDTA is presumably unlabeled (GlcUA-GlcNAc), were incubated with 4.25 mg/ml of purified enzyme at pH 5.0 and 37”. After 385 min, the reaction mixture

required to chelate trace amounts of inhibitory metal ions (1,

was streaked on Whatman No. 3 paper and the chromatogram eluted 22). Aliquots of the reaction mixtures were taken at various

with 1-butanol/acetic acid/water, 10/3/7 (v/v). The chromatogram was times and analyzed by paper chromatography and liquid analyzed, and the region containing the presumed labeled (GlcUA- scintillation counting of sections of the chromatograms (10). GlcNAc), cut out and eluted with water. The eluted material was The initial stages of the reactions were studied at trimer and diluted with unlabeled authentic trimer to yield a solution 0.1 M in saccharide, with a specific activity of 2000 cpm/mg. A similar solution

tetramer concentrations of 0.1 to 2 mM. Fig. 1 shows typical

of authentic generally labeled compound was prepared, and the two results, for 0.80 mM (GlcUA-GlcNAc), with 4.25 mg/ml of

solutions compared in three systems: paper electrophoresis at pH 6.4 hyaluronidase. The production of (GlcUA-GlcNAc),, and the (pyridine/acetic acid/water, 25/l/225, v/v) for 4 hours at 28 volts/cm nearly complete absence of monomer, show clearly the imnor- field on Whatman No. 3MM paper; descending paper chromatography with l-butanol/acetic acid/water, 5/3/4, v/v; and thin layer chroma-

tan& of tr&isglycosylation in the reaction, and are in qualita-

tography on MN 300 cellulose layers with the same solvent. In each tive agreement with the results reported by Weissmann (1).

system, the radioactive peaks from the two solutions migrated iden- The initial rates of disappearance of the substrate and of tically and could not be resolved when a mixture of the two solutions appearance of various products could be estimated from the was spotted, and coincided with the saccharide spot detected with data for the first 20% of reaction. Tables I and II present the silver nitrate or chlorine and starch-iodide (18). The solution of trans- glycosylation-labeled trimer was also chromatographed on a Bio-Gel P4

results for series of experiments with the trimer and tetramer,

column (1.0 x 90 cm) eluted with 25% acetic acid, and l-ml fractions respectively, at various substrate concentrations. The enzyme

collected. The fractions were analyzed for tritium content by liquid concentration in experiments with tetramer was 0.22 mg/ml, scintillation counting and for uranic acid by the carbazole test (19). about one-twentieth that used in experiments with trimer.

by guest on June 20, 2018http://w

ww

.jbc.org/D

ownloaded from

lo-

6

i-2 3

4

TIME, MINUTES

FIG. 1. Initial stages of hydrolysis of (GlcUA-GlcNAc),. [aH](GlcUA-GlcNAc), (0.80 mM) was incubated with 4.25 mg/ml of hyaluronidase (42,000 NFU/mg) in 0.66 M ammonium acetate-O.02 M EDTA, pH 5.0, at 37”. Aliquots were analyzed as described in text at various times. 0, (GlcUA-GlcNAc),; A, (GlcUA-GlcNAc),; 0, (GlcUA-GlcNAc),; n , (GlcUA-GlcNAc),.

7475

When the reactions are followed over extended periods the longer oligomers disappear, and the major product of digestion is the dimer (GlcUA-GlcNAc),. When 80 to 90% of the substrate has been consumed, only 3 to 6% of the saccharide is present in the form of the monomer. A representative plot for the digestion of trimer is shown in Fig. 2. Qualitatively similar results are obtained at other concentrations. The tetramer

shows the same sort of behavior at longer times; for instance, when 0.57 mM tetramer is incubated with 0.22 mg/ml of hyaluronidase, the reaction mixture consists of about 0.1 mM

tetramer, 0.2 mM trimer, 0.6 mM dimer, and 0.1 mM monomer after about 370 min. Such results can only be explained by fairly complex reaction paths involving transglycosylation (see “Discussion”).

No hydrolysis of oligosaccharides smaller than the trimer is

TABLE I Initial rates of reaction of trimer

[8H](GlcUA-GlcNAc), at the indicated concentrations was incu- of reaction (two to three points). The compounds II, III, etc., are

bated at 37” with 4.25 mg/ml of purified hyaluronidase in 0.66 M (GlcUA-GlcNAc),, etc. Under these conditions, . . . . monomer could ammonium acetate-O.02 M EDTA, pH 5.0. Aliquots were analyzed by not be detected.

paper chromatography, and initial rates calculated from the first 20%

Initial Substrate Concentration

mM 2.4 1.6 0.80 0.20 0.10

3?- 20 f 4 13 k 3 se1 1.5 t; 0.s 0.9 f 0.3

Initial Reaction Rate ql 14 * 2 952 6+1 1.5 * 0.4 0.9 2 0.2

M nin -1 , x 106

ZiIgl 721 6+1 2.5 k 0.4 0.4 t 0.2 0.2 t 0.2

ql 0.8 f 0.3 0.9 t 0.3 0.4 + 0.2 0.1 t 0.1

TABLE II

Initial rates of reaction of tetramer

[3H](Gl~UA-Gl~NAc), at the indicated concentration was incubated chromatography, and initial rates calculated for the first 20% of

at 37” with 0.22 mg/ml of purified hyaluronidase in 0.66 M ammonium reaction (one to two points). The compounds I, II, III, etc., are

acetate-O.02 M EDTA, pH 5.0. Aliquots were analyzed by paper (GlcUA-GlcNAc) etc. 1,2 . . .,

Initial Substrate Concentration

IBM 1.7 0.57 0.28 0.093

ip 37 t 7 14 f 3 721 4.0 f 0.8

Initial Reaction Rate q < 2 <l < 0.4 0.1 2 0.08

M min-l, x lo6

2igl 18 f 4 a+2 2.5 t 0.5 0.7 A 0.1

+!F 25 f s 10 f 2 5+1 1.6 t 0.3

3!gl 7+3 2t1 1.0 f 0.5 0.3 i 0.1

by guest on June 20, 2018http://w

ww

.jbc.org/D

ownloaded from

1476

FIG. 2 (left). Time course of hydrolysis of (GlcUA-GlcNAc),. [3H](GlcUA-GlcNAc), was incubated with hyaluronidase under conditions identical with those shown for Fig. 1. q , GlcUA-GlcNAc; 0, (GlcUA-GlcNAc),; A, (GlcUA-GlcNAc),; W, (GlcUA-GlcNAc),; 0, (GlcUA-GlcNAc),.

FIG. 3 (right). Time course of the hydrolysis of borohydride-reduced trimer. (GlcUA-GlcNAc)~-GlcUA-N-acetyl[l-aH]glucosaminitol (1.1 mM) was incubated with 4.25 mg/ml of hyaluronidase (42,000 NFU/mg) in the same buffer system as for Fig. 1, at 37”, and aliquots analyzed as in text. Only N-acetyl [l-SH]glucosaminitol-containing saccharides were detected. 0, reduced GlcUA-GlcNAc; 0, reduced (GlcUA-GlcNAc),; A, reduced (GlcUA-GlcNAc),; 0, reduced (GlcUA-GlcNAc),.

observed. (GlcUA-GlcNAc), and GlcNAc-GlcUA-GlcNAc- GlcUA-GlcNAc were each incubated at concentrations of about 11 mM for 6 hours with 4.2 mg/ml of purified enzyme, under the same conditions used for studies of trimer hydrolysis. In each case, more than 95% of the saccharide remained in the initial substrate peak, and no identifiable product peaks appeared. This implies that the rates of reaction of the enzyme with these substrates are at least 2 orders of magnit.ude lower than with the trimer.

Oligosaccharides with the “reducing terminal” GlcNAc unit reduced to N-acetylglucosaminitol by sodium [3H]borohydride were also examined. The “reduced dimer,” GlcUA-GlcNAc- GlcUA-N-acetyl[l-3H]glucosaminitol, is not a substrate under the conditions described above. On the other hand, contrary to Weissmann’s report (l), the “reduced trimer,” (GlcUA- GlcNAc) ,-GlcUA-N-acetyl [1-3H]glucosaminitol, is a sub- strate. Its hydrolysis under conditions similar to those used for the trimer itself is shown in Fig. 3. It is apparent from a comparison of Figs. 2 and 3 (drawn to the same scale) that the reduced trimer is a poorer substrate than the trimer; its initial rate of disappearance is some 2.5-fold less. An interesting aspect of this reaction is that the reduced monomer is produced in significant amounts. (The distribution of products not containing N-acetyl [l-SH]glucosaminitol, accounting for some 20% of total saccharide at the last point shown, unfortunately is not determined in such an experiment.)

Transglycosylation-Hyaluronidase-catalyzed reactions be- tween the trimer and dimer shed light on a number of questions about the transglycosylation process.Z Two parallel experi- ments were carried out, one with tritium-labeled dimer and unlabeled trimer, and the other with labeled trimer and unlabeled dimer. In each case, 12 mM dimer and 4 mM trimer were incubated with 4.25 mg/ml of purified enzyme at pH 5.0 for 6% hours at 37”. The results of analysis of aliquots of the reaction mixtures are shown in Fig. 4.

z Meyer reports in a review (Ref. 3) that the dimer is not an acceptor for transglycosylation by hyaluronidase. It is not clear what exper- imental evidence this is based on, as it is stated without reference.

0’ 0’ IO IO 20 20 30 s 30 s

CM from Origin CM from Origin

FIG. 4. Transglycosylation experiments with trimer and dimer. (GlcUA-GlcNAc), (4 mM) and (GlcUA-GlcNAc), (12 mM) were incubated with 4.25 mg/ml of hyaluronidase (42,000 NFU/mg) at 37’ in the same buffer system as for Fig. 1. After 385 min, the reaction mixtures were analyzed by paper chromatography with 1-butanol/ acetic acid/water. A, experiment with labeled dimer and unlabeled trimer. B, experiment with labeled trimer and unlabeled dimer. 1, 2, and 3 indicate the positions of monomer, dimer, and trimer, respectively, on paper chromatography.

The new labeled compound formed in the transglycosylation reaction with labeled dimer (Fig. 4A), which behaves like the trimer, was isolated by preparative paper chromatography. A solution of this compound, diluted with a large excess of unlabeled authentic trimer, was then studied further. Aliquots of this solution were examined by paper and thin layer chromatography, electrophoresis, and gel filtration, and in each case the single radioactive peak or spot was coincident with the (GlcUA-GlcNAc) s.

The extent of tritium labeling in the nonreducing terminus *of this labeled trimer was then compared with that of the labeled dimer from which it was formed. When an aliquot of the above trimer solution was treated with /?-glucuronidase, the glucuronic acid released contained less than 0.5% of the total radioactivity. The generally labeled dimer, however, had 14.6% of its total radioactivity in the nonreducing terminal glucuronic acid. This means that less than 4% of the terminal glucuronic acid in the transglycosylation-labeled trimer is derived from the dimer, and shows that the incorporation of label into the trimer occurs by transfer of a glycosyl residue from unlabeled trimer to the nonreducing end of the labeled dimer.

A comparison of the reaction in which dimer was labeled initially with that in which trimer was labeled initially is interesting. Table III (derived from the data of Fig. 4) shows the distribution of the tritium label at the end of the two experiments. Since the chemical conditions of the experiments were identical, their results may be combined and analyzed as if they were a single “double-labeling” experiment. This analysis, given in Table IV, must be based in addition on the observations that (a) the dimer is not itself a substrate for hyaluronidase, and (b) the mechanism of transglycosylation is as shown above. A notable result of this analysis concerns the origin of the monomer, which is derived overwhelmingly from material originally introduced in the form of the trimer. Since, by the end of the reaction, the reducing terminus of the trimer

by guest on June 20, 2018http://w

ww

.jbc.org/D

ownloaded from

7477

TABLE III

Results of transglycosylation experiments with trimer and dimer

For 385 min, 4 mM (GlcUA-GlcNAc), and 12 mM (GlcUA-GlcNAc), were incubated with 4.25 mg/ml of hyaluronidase at pH 5.0, 37”, and reaction mixtures analyzed as in Fig. 5.

-

% of ‘H label in saccharide Experiment

Gl.aA-GlCNAC

-!-

(GlcUA-GlcNAc)? (GlcUA-GlcNAc),

A: labelled dimer

R: labelled trimer

-

TABLE IV

Calculated composition of transglycosylation reaction mixture The data of Table III are analyzed as a “double label” experiment,

with additional assumptions as described in the text. The symbol *

represents a -GlcUA-GlcNAc- unit originating in the trimer, and # represents a unit originating in the dimer.

-_

Concentration. in ml. ! Oligosnccharide component

initial final

._*_. 4.0 0.87

1-11 12.0 10.8

*-e-x 0 1.1

.-• 0 3.9

(‘3 + ‘-‘-8) 0 ! O.la

f 0 0.35

” 0 I 0.1

a) Based on the assumption that formation of either of these would have to

have led also to X.

has incorporated large amounts of material originating in the

dimer, one can conclude that the monomer is formed from the nonreducing terminus of the trimer.

Under conditions similar to those used above (i.e. 4 mM (GlcUA-GlcNAc),, 10 to 20 mM labeled acceptor, and 4.24 mg/ml of hyaluronidase incubated at pH 5.0 and 37” for 6 hours) neither GlcUA-GlcNAc nor GlcUA-N-acetyl- glucosaminitol are incorporated into new compounds by transglycosylation. The “reduced dimer,” GlcUA-GlcNAc- GlcUA-N-acetyl [l-SH]glucosaminitol, is about as good an acceptor for transglycosylation as the dimer itself; about half the radioactivity originally in the reduced dimer appears in larger oligosaccharides after 6 hours.

pH Dependence-The initial rates of hydrolysis of the oligomers were determined at 37” at various pH values in ammonium carbonate-ammonium acetate buffers containing 0.02 M EDTA by the chromatographic technique. The trimer hydrolysis was studied with 0.40 M substrate and 4.25 mg/ml of hyaluronidase, and the tetramer with 0.64 M substrate and 0.22 mg/ml of enzyme. Initial rates were converted into apparent turnover numbers (moles of substrate consumed per min per mol of enzyme) by assuming an effective molecular weight of 61,000 for the purified (42,000 NFU/mg) enzyme (9). The results are shown in the lower two curves of Fig. 5.

The initial rates of hydrolysis of polymer at various pH

40

20

0. 2

I

0 0.0 I5

0.010

0005

0

(GlcUA-WC

P (GICUA-GIG

A, 4 5 6 7

PH

FIG. 5. pH dependence of hyaluronidase hydrolysis at 37”. For- trimer and tetramer (two lower curues), 0.4 and 0.64 mM labeled oligomer, respectively, were incubated with enzyme and initial rates determined by chromatographic analysis (see text). For polymer (upper curue), potassium hyaluronate (0.46 mM in hyalobiuronate units) was incubated with enzyme and aliquots assayed for reducing sugar (as GlcUA-GlcNAc) by the Park-Johnson method.

values were determined by incubation of 200 wg/ml of potas- sium hyaluronate (about 0.46 mM in -GlcUA-GlcNAc- units) with 1 kg/ml of hyaluronidase in sodium acetate buffers at 37”, and assaying aliquots of the reaction mixture at 15 and 30 min for reducing sugar (17). GlcUA-GlcNAc and its dimer were used as standards for the reducing power of end groups in hyaluronate oligosaccharides, and the rates calculated as moles of new end groups released per min per mol of enzyme. The results are shown in the upper curve of Fig. 5.

The pH dependence of the reaction of each of the substrates examined can be fit by the assumption that a basic group of pK, about 4.5, and an acidic group of pK, between 6 and 7 are required for activity. The apparent pK, of the acidic group is not well determined, and may not be the same for the polymer and for the smaller oligomers. These results are reminiscent of those found for other glycanohydrolases (5, 6).

The reaction of polymer was also followed for longer times at pH 5.5. After 20 hours, the concentration of reducing end groups had leveled off at 0.09 mM, about 20% of the hyaluroni- dase-susceptible bonds theoretically present in the polymer. Under the same conditions, a mixture of chondroitin 4- and g-sulfates was hydrolyzed with an apparent turnover number only 5% that for hyaluronic acid, and the release of reducing groups leveled off when less than 1% of the theoretically available susceptible bonds had been cleaved.

Saccharide Binding-The binding of the dimer (GlcUA- GlcNAc), was studied by the equilibrium dialysis technique at pH 6.0 at 24”, over an enzyme concentration range of 5 x 10m5 to 3.5 x lo-’ M and a saccharide concentration range of 0.2 to 1.2 mM (22). The dissociation constants determined ranged from 0.30 to 0.47 mM, with an average of 0.40 s 0.05 mM. A Scatchard analysis (23) of the data indicates that 1 mol of saccharide is bound per mol of enzyme, with a dissociation constant of 0.38 f 0.03 mM (9). The dissociation constants for the binding of a number of saccharides to the enzyme are given in Table V.

Inhibition experiments gave results of a qualitative nature only. At concentrations of 0.02 M, N-acetylglucosamine, glucu- ronic acid, glucuronamide, and acetylsalicylic acid caused 20%, 35%, 25%, and 10% inhibition of hydrolysis of the

by guest on June 20, 2018http://w

ww

.jbc.org/D

ownloaded from

TABLE V without significant changes in the distribution of oligomer Dissociation constants for hyaluronidase-saccharide complexes concentrations. Thus, after long

Constants were determined by equilibrium dialysis at pH 6.0, as described in text.

Saccharide No. of

experiments Kd, d!

GlcNAc 2 St3

GlcUA-GlcNAc 5 3.2 + 0.6

GlcNAc-GlcUA-GlcNAc 1 1.2 f 0.2

GlcUA-GlcNAc-GlcUA-GlcNAc 5 0.40 A 0.05

GlcNAc-GlcUA-GlcNAc-GlcUA-GlcNAc 1 0.2 k 0.1

polymer, respectively. Of possible interest is the observation that the @(l-+4)-linked trisaccharide of GlcNAc, tri-N-acetyl- chitotriose, does not inhibit the enzyme at mM concentrations. 3

DISCUSSION

The hyaluronidase digestion of the trimer (GlcUA-GlcNAc) 3 is dominated overwhelmingly by transglycosylation in the concentration range examined (0.1 to 2.4 mM). During the early stages of the reaction the major products are the dimer and tetramer, and the monomer is not detectable (Fig. 1 and Table I). These results can be accounted for almost entirely by two transglycoslation processes (Equations 1 and 2, where S represents a -GluUA-GlcNAc- unit of the substrate), with hydrolysis (e.g. Equation 3) accounting for less than 10% of the observed reaction.

s-s-s + s-s-s 4 s-s + s-s-s-s (1)

s-s-s-s + s-s-s 4 s-s + s-s-s-s-s (2)

S-S-S-S + H,O + S-S + S-S (3)

When trimer and excess dimer are incubated together with the enzyme for long periods (Tables III and IV), the major pathways for the reaction are those of Equations 4 and 5. Reaction 4 is a transglycosylation following the now familiar

s,-s,-s, + s,-s,b,-s,-s, + s,-s, (4) k, S,-S,-S, + H,O -S, + S,-S, (5)

pathway (4, 5, 26): transfer of a glycosyl residue with retention of configuration from one aglycone (the “leaving group”) to the nonreducing terminus of another oligosaccharide (the “accep- tar”). Reaction 5 is the transfer of this same glycosyl residue to water, and it should be noted that both reactions involve cleavage at the same bond in the trimer: that between units S, and S,.

With this in mind, one can explain the progress of the reactions of the trimer at longer times (Fig. 2) as follows: As the transglycosylation Reactions 1 and 2 proceed, the concentra- tion of dimer builds up, and dimer competes with trimer as an acceptor. Under such conditions, further transglycosylation leads only to a reshuffling of saccharides (e.g. Equation 4 or 6)

s-s-s-s + s-s + s-s + s-s-s-s (6)

incubation times, the effect of the slow hydrolytic processes becomes significant, and leads to over-all net hydrolysis. One can estimate from our results with trimer that the ratio of rate constants for transglycosylation and hydrolysis (e.g. k,:lz,, for Equations 4 and 5) must be at least of the order of magnitude of 10’ M-I. For comparison, the transglycosylation to hydrolysis ratio for lysozyme with cell wall oligosaccharides as acceptors is 3 x lo3 M-’ (27).

One can now describe some of the properties of the enzyme active site. Using the notation of Schechter and Berger (26) (where s, represents a hyalobiuronate binding subsite on the glycosyl side of the point of cleavage and s’, a subsite on the aglycone side), we consider initially five possible subsites, as in Scheme 1. The important productive mode of binding of the trimer must be across subsites a,-~‘,-.?,, and reaction stemming from binding across s2-~,-a’, must be at least 40 times less favorable. Subsite sip must therefore be responsible for a large part of the productive binding interaction between the enzyme and the trimer. The importance of sI1 in productive binding is demonstrated even more strikingly by the observation that neither (GlcUA-GlcNAc), nor GlcNAc-GlcUA-GlcNAc- GlcUA-GlcNAc is digested by hyaluronidase at concentrations of 11 mM, despite the fact that they are bound to the enzyme with dissociation constants less than 1 mM.

Subsite .s’~ is also very important in the interaction of transglycosylation acceptors with the glycosyl enzyme: the dimer, S-S, is a good acceptor, while the monomer, which would have to be bound in s’~ to act as an acceptor, is not incorporated into transglycosylation products to a detectable extent.

From further experiments, one can deduce that subsite s’, can accommodate an N-acetylglucosaminitol residue in place of GlcNAc. (GlcUA-GlcNAc),-GlcUA-N-acetylglucosaminitol is comparable to the normal trimer as a substrate, and GlcUA-GlcNAc-GlcUA-N-acetylglucosaminitol is comparable to the dimer as an acceptor.

There must be at least one subsite in addition to sl, s’~, and sf2 of importance in the reactions of the enzyme, since the tetramer (GlcUA-GlcNAc), is a far better substrate than the trimer. The most relevant parameter for comparing the in- teraction of two substrates with an enzyme is the apparent second order rate constant, k,,,lK,- in which any ambiguities

in hat and K, due to multiple nonproductive binding modes cancel out. From initial rates at low substrate concentrations (Tables I and II), one can calculate k,,,/K, for the trimer and tetramer of about 120 and 10,000 M-’ min-‘, respectively. (A peculiar aspect of the data is the high apparent K, values for the oligomers compared to the equilibrium dissociation con- stant of 0.4 mM for the dimer. Lineweaver-Burk analyses of the data of Tables I and II lead to K, = 2.2 mM and k,,, = 20 mine1 for the tetramer, and K, = 11 mM and k,,, = 1 min-’ for the

8 Kuettner and his co-workers have suggested a role for lysozyme in the metabolism of proteoglycans in preosseous cartilage (24). Their claim is supported in part by their observation of inhibition of certain processes by the known inhibitor of lysozyme, tri-N-acetylchitotriose, at millimolar concentrations (see Ref. 25). On the other hand, we have found that the hyaluronate oligosaccharides (GlcUA-GlcNAc), and (GlcUA-GlcNAc), are completely unaffected by long incubation with lysozyme (D. M. Chipman and M. Schindler, unpublished results).

s, s, s; s’2 s3

SCHEME 1

by guest on June 20, 2018http://w

ww

.jbc.org/D

ownloaded from

trimer. Since the reactions of the oligomers become increas- ingly complex at high substrate concentrations due to trans- glycosylation, and the calculations involve large extrapolations, these results should not be taken as true Michaelis-Menten parameters).

Several considerations lead us to the conclusion that the additional significant subsite for reactions of the tetramer is on the aglycone side of the cleavage point, that is, sfg: (a) At the lowest tetramer concentrations examined (where the effect of transglycosylation is expected to be minimized) the initial rate of appearance of trimer is at least twice that of dimer, and detectable amounts of monomer are formed early in the reaction. This result is expected if the major productive binding mode for the tetramer is across s,-s’,-sf2-s’,, but not if it is across .s2-s,-s’,-s’,. (b) The fact that the trimer is a poorer substrate than the tetramer even at concentrations at which it should saturate the enzyme suggests that there is some nonproductive binding mode for the trimer. This can be explained by the existence of a subsite s13 of rather high binding affinity. (c) GlcNAc-GlcUA-GlcNAc-GlcUA-GlcNAc is bound strongly to the enzyme (more so even than the dimer, in fact), but is not a substrate. It might be bound across half of s’~, and s’, and s’,. One can imagine rationalizations for the latter two observations in terms of alternative arrangements of subsites, but they would be totally inconsistent with the first point, the observed cleavage pattern for the tetramer.4

Polymeric hyaluronic acid appears to be a better substrate than the tetramer, with a turnover number of up to 50 min-’ (e.g. Fig. 5), so it would seem that the enzyme has at least one more significant subsite. We propose that a further subsite is on the glycosyl side of the cleavage point. that is, sz, although the evidence here is not rigorous. The initial phase of hydrol- ysis of tetramer does produce some significant amounts of dimer, which would be expected if an alternative productive binding mode were across s2-s,-s’,-s’,, but not if s’, were the fifth strong binding site. Weissmann reported (1) that the major initial products of the action of crude hyaluronidase on the pentamer were trimer and heptamer, in accord with productive binding across s*-s ,-s’ ,-s’ p-s’ 3.

These conclusions may be summarized as follows. Testicular hyaluronidase can interact simultaneously with at least 5 -GlcUA-GlcNAc- units of an oligosaccharide, using five sub- sites as pictured in Scheme 1. Such an active site would be some 40 to 50 A long. Binding across subsites s,-s’,-s’, is required for reaction to occur, but most of the affinity for the saccharide in these sites is contributed by subsite stZ. Binding in s,-s’~ alone is sufficiently unfavorable that the dimer is not a substrate and binds only nonproductively. In the presumed glycosyl enzyme intermediate, binding of an acceptor in s’, alone is not favorable, so the monomer does not serve as an acceptor. The affinity of subsite st3 is also greater than that of s,, so that nonproductive binding of the trimer is competitive with productive binding by more than an order of magnitude. Subsite .s2 is weaker than either sf2 or sf3, and is of importance chiefly for oligomers longer than the tetramer. We hope that this qualitative description of the enzyme’s properties can be

‘The latter two observations may also be explained without proposing nonproductive binding in the usual sense. The saccharides in question might primarily be bound across the “cleavage site,” but undergo reaction slowly or not at all because their interactions with the enzyme are insufficient to lead to the “strain” or “induced fit” required for efficient catalysis (29). This explanation leads to the same mechanistic implications as nonproductive binding (see below)

7479

corroborated by suitable computer modeling of the course of its reactions (27).

Nonproductive binding of oligomeric substrates, as a result of the relatively unfavorable binding properties of the subsites directly around the catalytic site of an enzyme, has strong mechanistic implications (29). It suggests that these subsites are not optimally designed to complement the geometry of the substrate, but may rather be complementary to the geometry (or charge distribution, or both of these, etc.) of the transition state for the enzymic reaction. The strong preference of the enzyme for transglycosylation (to rather large acceptors at that) suggests, however, that the geometric changes in question are not localized, but involve contacts between the substrate and many subsites on the enzyme. The interactions between the presumed glycosyl-enzyme intermediate and an acceptor, which lead to transglycosylation, would also be expected to stabilize the formation of the glycosyl-enzyme-leaving group complex from the enzyme-substrate complex. These are cer- tainly not surprising suggestions, and are particularly reascna- ble for a glycanohydrolase, for which lysozyme serves as a precedent (4, 5, 30), and for which so few other catalytic strategies are available (31).

Acknowledgment-We are also happy to acknowledge help- ful discussions with Dr. J. E. Silbert, the earliest of which were instrumental in encouraging us to begin working on hyaluroni- dase.

1. 2.

Weissmann, B. (1955) J. Biol. Chem. 216, 783 Hoffman, P., Meyer, K., and Linker, A. (1956) J. Biol. Chem. 219,

653 3.

4. 5.

6.

Meyer, K. (1971) in The Enzymes (Boyer, P. D., ed) 3rd Ed, Vol. 5, pp. 307-320, Academic Press, New York

Chipman, D. M., and Sharon, N. (1969) Science 165,454 Imoto, T., Johnson, L. N., North, A. C. T., Phillips, D. C., and

Rupley, J. A. (1972) in The Enzymes (Boyer, P. D., ed) 3rd Ed, Vol. 7, pp. 666-868, Academic Press, New York

Thoma, J. A., Spradlin, J. E., and Dygert, S. (1971) in The Enzymes (Boyer, P. D., ed) 3rd Ed, Vol. 5, pp. 115-189, Academic Press, New York

7. Borders, C. L., Jr., and Raftery, M. A. (1968) J. Biol. Chem. 243, 3756

8. Khorlin, A. Y., Vikha, I. V., and Milishnikov, A. N. (1973) FE&S Lett. 31, 107

9. Garvin, J. H., Jr., and Chipman, D. M. (1973) FEBS Lett. 39,157 10. Highsmith, S., and Chipman, D. M. (1974) Anal. Biochem. 61,557 11. Di Ferrante, N. (1956) J. Biol. Chem. 220, 303 12. Weissmann, B., Meyer, K., Sampson, P., and Linker, A. (1954) J.

Biol. Chem. 208, 417 13.

14. 15.

Raftery, M. A., Rand-Meir, T., Dahlquist, F. W., Parsons, S. M., Borders, C. L.. Jr.. Wolcott. R. G.. Beranek. W.. Jr.. and Jao. L. (1969) Anal. Z&o&em. 30, 427 ’ ’

Wilzbach, K. E. (1957) J. Am. Chem. Sot. 79, 1013 Meyer, K., Fellig, J., and Fischer, E. H. (1951) J. Biol. Chem. 192,

225 16. 17. 18. 19. 20. 21.

Mayo, J. W., and Carlson, P. M. (1970) Carbohyd. Res. 15, 300 Park, J. T., and Johnson, M. J. (1949) J. Biol. Chem. 181, 149 Powning, R. F., and Irzkiewicz, H. (1965) J. Chromatogr. 17, 620 Dische, A. (1947) J. Biol. Chem. 167,189 Westhead, E. W., and McLain, G. (1964) J. Biol. Chem. 239,2464 Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J.

(1951) J. Biol. Chem. 193, 265 22. Garvin, J. H., Jr. (1974) Ph.D. thesis, Massachusetts Institute of

Technology 23. 24.

25.

Scatchard, G. (1949) Ann. N. Y. Acad. Sci. 51, 660 Pita, J. C., Cuervo, L. A., Madruga, J. E., Muller, F. J., and

Howell, D. S. (1970) J. C&n. Znuest. 49, 2188 Pita, J. C., Howell, D. S., and Kuettner, K. (1975) in Extracellular

Matrix Influences on Gene Expression (Slavin, H. C., and

REFERENCES

by guest on June 20, 2018http://w

ww

.jbc.org/D

ownloaded from

7480

Greulich, R. C., eds) pp. 721-726, Academic Press, New York 29. Jencks, W. P. (1969) Catalysis in Chemistry and Enzymology, pp. 26. Wallenfels, K., and Malhotra, 0. P. (1961) Adu. Carbohyd. Chem. 282-320, McGraw-Hill. New York

16, 239 30. Secemski,‘I. I., Lehrer, s. S., and Lienhard, G. E. (1972) J. Biol. 27. Chipman, D. M. (1971) Biochemistry 10, 1714 Chem. 247,474O 28. Schechter, I., and Berger, A. (1966) Biochemistry 5, 3371 31. Fife, T. H. (1972) AC&. Chem. Res. 5, 264

by guest on June 20, 2018http://w

ww

.jbc.org/D

ownloaded from

S Highsmith, J H Garvin, Jr and D M ChipmanMechanism of action of bovine testicular hyaluronidase. Mapping of the active site.

1975, 250:7473-7480.J. Biol. Chem. 

  http://www.jbc.org/content/250/18/7473Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/250/18/7473.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on June 20, 2018http://w

ww

.jbc.org/D

ownloaded from