Kinetic Properties and Substrate Specificities of Two Cellulases from Auxin-treated Pea

Two cellulases purified from growing regions of auxin-treated peas (buffer-soluble and buffer-insoluble) hydrolyze cellulose powder, partially substituted carboxymethylcellulose (CM-cellulose), higher cellodextrins, and certain mixed linkage glucans (e.g. barley beta-glucan), at rates comparable to these reported for the most active fungal cellulases, and with kinetics and product formation characteristic of endohydrolase action. They are unable to cleave 1,3-linkages in beta-glucans, or 1,4-linkages in dextrins containing excessive substitution at C6, alpha configuration, alternating beta-1,3- and 1,4-linkages, or residues other than anhydroglucose. They are not active towards cellobiose or the 1,4-linkage adjacent to the reducing end of cellodextrin chains. It is concluded that buffer-soluble and buffer-insoluble cellulases are true beta-1,4-glucan 4-glucanohydrolases (EC 3.2.1.4). On a molar basis, Vmax values for buffer-insoluble are higher than buffer-soluble cellulase acting towards any of the substrates tested, but Km values towards CM-cellulose and cellohexaose are essentially identical. Both cellulases were inhibited by C12+, Hg2+, and sulfhydryl-binding reagents. Buffer-insoluble, but not buffer-soluble, cellulose was inactivated by reagents that bind serine and threonine, which reflects differences in their amino acid composition. No major qualitative differences have been detected in the mode of action of the two enzymes. Despite marked differences in their physical and immunological properties, close similarities between buffer-soluble and buffer-insoluble enzymic properties suggest that their active sites are the same.

On a molar basis, V,,, values for buffer-insoluble are higher than buffer-soluble cellulase acting towards any of the substrates tested, but K,,, values towards CM-cellulose and cellohexaose are essentially identical. Both cellulases were inhibited by Cu'+, Hg2+, and sulfhydryl-binding reagents. Buffer-insoluble, but not buffer-soluble, cellulase was inactivated by reagents that bind serine and threonine, which reflects differences in their amino acid composition. No major qualitative differences have been detected in the mode of action of the two enzymes. Despite marked differences in their physical and immunological properties, close similarities between buffer-soluble and buffer-insoluble enzymic properties suggest that their active sites are the same. Cellulase, assayed by measuring hydrolysis of CM-cellulose' or cellulose, occurs indigenously in growing regions of pea epicotyls (1) and is one of very few plant enzyme activities known to increase dramatically following treatment of tissue with the auxin type of growth regulator (2,3). After purification to homogeneity, it has been established (4,5) that this activity is due to two proteins with molecular weights of 20,000 (buffer-soluble) and 70,000 (buffer-insoluble  Initial rate of production of reducing power from CM-cellulose (0.1 to l.O%, w/v) was measured (11) using cellulase levels and conditions described in Fig. 1 mobility similar to 4-o-P-n-laminaribiosyl-D-glucose, but different from 3-@/?+cellobiosyl+glucose.
As expected from results of these tests for specificity, neither cellulase was inhibited by addition of 1% nojirimycin (5-amino&deoxy-n-glucopyranose), an antibiotic reported (27) to strongly inhibit p-glucosidase, but not endoglucanase activity. Both enzymes were completely inactive in the presence of several noncompetitive inhibitors, including heavy metal ions (5 mM CL?+ or Hg2+) and sulfhydryl reagents (2 mM p-chloromercuribenzoate, mercaptoethanol, or n-ethylmaleimide). The only notable difference observed (11) in responses of the two enzymes to a variety of potential inhibitors was the inactivation of buffer-insoluble but not buffer-soluble cellulase by reagents which bind serine and threonine residues (e.g. succinic anhydride, diisopropyl fluorophosphate).
This may be related to the fact that buffer-insoluble cellulase contains many more of these residues per molecule than buffer-soluble cellulase (4).
Hydrolysis of Cellodextrins -The initial reaction rates of buffer-soluble and buffer-insoluble cellulases acting on nonlimiting concentrations of W-cellodextrins were assayed by measuring the amount of labeled substrate which disappeared during the early linear phase of digestion. The results (Fig. 3) show that neither enzyme hydrolyzes cellobiose (G2), but both degrade higher cellodextrins (G3 to G,) at rates which increase exponentially with D.P. Progress curves for production of degradation products from cellodextrins (GZ to G,) are shown in Fig. 4. The two cellulases display similar action patterns while cleaving each of these substrates and the dextrins derived from them. The following generalizations apply to activities of both enzymes. Cellotriose is slowly degraded to equimolar concentrations of G, and G2 indicating cleavage of one or both of the two glucosidic linkages. Cellotetraose is hydrolyzed more rapidly in early stages mainly to GZ with G, and G3 generated in smaller but equal amounts. It cannot be determined from these data whether the internal linkage of cellotetraose is most susceptible to hydrolysis with both terminal linkages cleaved at about half the rate, or whether one terminal linkage is resistant to hydrolysis while the other two linkages are cleaved at similar rates. Cellopentaose is degraded rapidly to equal amounts of Gt and G3, with G, and G4 accumulating relatively slowly. Evidently one or both of the internal linkages are cleaved more readily than either terminal linkage. The predominant final product is GB, presumably because G3 and G4 are subject to further degradation.
Cellohexaose degrades very rapidly (Fig. 3) to ,-01 X W d 3 Errzymic Properties of Pea Cellulases form G2 and G, as the main products, with transitory amounts of Gu, Gq, and G5 detectable in early stages of the decay (11). There was no indication in these tests that transglucosylation occurred in the mixtures e.g. cellodextrins of higher D.P. than the substrates were never detected on chromatographs. Tests for transglucosylation, using YScellodextrins as potential donors and Y-cellodextrins or cellulose as acceptors under a variety of reaction conditions, yielded no evidence that pea cellulases could form P-1,4-linkages.

Hydrolysis
of Reduced Cellodextrins -In order to assess the relative susceptibilities of linkages in cellodextrins to hydrolysis, the terminal reducing glucosyl residues (G, to G,) were reduced (tritiated) and used as asymetrically labeled substrates (GZH to G,,). Fig. 5 shows progress curves for the generation of labeled products. As with other substrates (Tables I and II), buffer-insoluble cellulase was more active towards reduced "H-cellodextrins than buffer-soluble cellulase on a molar basis but the mode of action of the two enzymes was indistinguishable.
None of these substrates gave rise to labeled sorbitol, indicating that the linkage adjacent to the reduced terminal glucose is not cleaved. The other terminal linkage is readily hydrolyzed, however, as shown by the rapid initial rate of production of labeled products containing one less glucose unit than the substrates used. During early stages of hydrolysis of G4s, equimolar amounts of G,, and GBH are produced i.e. the central and susceptible terminal linkage are cleaved at comparable rates. These same products are generated at equal rates initially from G5s, with G4s produced relatively slowly, suggesting a moderate preference for hydrolysis of the more central linkages in this substrate. There is clear evidence also that G, s and G, s formed in these preparations are subject to further degradation. Eventually, GPH is the only labeled product remaining from all substrates.

DISCUSSION
Evidence has been presented (Tables I and II) that pea cellulase activities display specificity for poly-or oligosaccharide chains containing /3-1,4-linkages between anhydroglucose residues. They do not possess P-glucosidase activity and they cannot hydrolyze cellobiose (Fig. 3). They are also inactive towards the linkage adjacent to a terminal glucose residue in cellodextrins when that unit is reduced (Fig. 5). A similar restricted action pattern has been observed (28) with purified Myrothecium cellulase. Kinetics of hydrolysis of higher soluble cellodextrins (Fig. 4) are consistent with the conclusion that the terminal linkage at the reducing end of these chains is also resistant to cleavage, whereas other linkages are hydrolyzed in an essentially random manner. Unsubstituted cellulose is hydrolyzed at a substantial rate (Table I) in such a way that none of the reducing groups introduced in early stages of the reaction are released as glucose, cellobiose, or soluble cellodextrins i.e. the enzymes have no exohydrolase activity. The action patterns of cellulases towards CM-cellulose (Fig.  11, and the exponential relationship observed (Fig. 3) between rate of cellodextrin hydrolysis and D.P., confirm that these enzymes are true endohydrolases.
The pea cellulases can also depolymerize P-glucans that contain a limited number of /3-1,3linkages, e.g. barley pglucan (Table I). This substrate has regions of two or three contiguous 1,4-linkages separated by 1,3-linkages (25). The pea cellulases, like many fungal cellulases (29,30), generate several products from such substrates, the most prominent of which is the trisaccharide 4-G-/?-D-laminaribiosyl-D-glucose, i.e. P-1,4-linkages are cleaved. This observation, coupled with the inability of the pea cellulases to hydrolyze alternating 1,3and 1,4-linked P-glucan (Table II), indicates that these enzymes should be classed with cellulases of the Streptomyces type (251, rather than with the "mixed-linkage hydrolases" from Bacillus (31). It is concluded that buffer-soluble and buffer-insoluble cellulases merit designation as true P-1,4-glucan 4-glucanohydrolases (EC 3.2.1.4) and require a minimum of a cellobiosyl unit within a polymer in order to effect cleavage. Of the substrates tested, the enzymes degrade cellohexaose most effectively, as do many fungal cellulases (32,33). This has been interpreted (34) as indicating that such cellulases possess a binding site which accommodates at least 6 glucose units. Cellobiose must also be capable of binding to pea cellulases to a limited extent since it inhibits cellulase activity towards CM-cellulose (Fig.  2), but the disaccharide is not cleaved and evidently cannot span the active site. Reduced cellodextrins are hydrolyzed less rapidly than the parent dextrins (cf. Figs. 4 and 5), suggesting that opening of the terminal glucopyranose ring of these substrates intereferes with binding. In view of the high rate of hydrolysis of some mixed linkage glucans (Table II), the binding site apparently does not require that all contiguous glucose units be p-1,4-linked. Excessive substitution of 6-hydroxyls in glucan chains (e.g. as in commercial DEAE-cellulose) does not prevent binding but completely blocks hydrolysis by pea cellulases (4). Accordingly, highly branched p-1,4-glucans in cell walls (e.g. xyloglucan) would not be expected to act as good substrates, although fragments may be produced from them by cellulase action (35,36), presumably at points where no substitution occurs. It is also possible that the cellulases could degrade potential intermediates of cellulose synthesis, e.g. cellodextrin covalently linked to lipid (37, 38) or protein (38,39).
The pea cellulase activities per unit of protein or per mole of enzyme are at least as active towards cellulose, cellodextrins, and CM-cellulose as the most active fungal cellulases (cf. 38,40). Pea and fungal cellulases show many similarities in enzymic properties but differ in that thermal instability, pH optimum, and K,,l values are all generally higher for the pea enzymes. This may be related to conditions found in living plant cell walls where endogenous cellulase would have to act if it is to have any functional significance.
A requirement for auxin-induced cellulase activity in growth and development in plants is still a tentative proposition. Nevertheless, as effective endohydrolases, both buffersoluble and buffer-insoluble cellulases possess the basic property required to fulfill the potential functions which have been suggested for them namely to promote "loosening" (2, 10) and to enhance cellulose synthesis by generating primer chain ends (9, 41) within the continuous microfibrillar framework of young expanding cells.
The most striking feature of buffer-soluble and buffer-insoluble cellulases as observed here is that so many of their enzymic properties are identical, implying that they possess similar active sites. Buffer-insoluble cellulase has a higher turnover number than buffer-soluble cellulase (Tables I and  II), but the difference is hardly sufficient to explain why auxin treatment should generate two such enzymes in approximately equal amounts (4,5). Indeed these observations justify reopening the question of whether one cellulase is a precursor of the other. The original reasons for considering this unlikely (4) were based on differences in amino acid composition and the absence of cross-reactivity between the enzymes and antibodies raised to them. However, the apparent multiplicity observed in many cellulases is often due to modification of a single enzyme species by culture conditions, complex formation, or proteolysis (42). Close examination of the development of pea cellulase activities in early stages after hormone treatment shows (5) that the level of buffer-soluble cellulase increases first i.e. within a few hours, and that buffer-insoluble reaches or exceeds it only later i.e. after 1 day. It may be that the larger cellulase (buffer-insoluble) is a form modified from the smaller during the process of secretion from the cell, an event whch must occur if the enzyme is to reach its substrate in the wall. This interpretation is in accord with evidence that mRNA for buffer-soluble cellulase is synthesized de nouo by rough endoplasmic reticulum polysomes following hormone treatment, while a distinct messenger for buffer-insoluble cellulase could not be located in these preparations (71, as well as ultrastructural observations (6) that buffer-soluble cellulase is concentrated in endoplasmic reticulum vesicles whereas most buffer-insoluble cellulase is bound firmly to inner surfaces of growing cell walls.
Acknowledgments-We wish to thank our colleagues who kindly provided us with substrates and diagnostic enzyme preparations that are not commercially available. We are particularly grateful for helpful advice and discussion accorded to us by Professors B. A. Stone, A. S. Perlin, and D. P. S. Verma.