- Email: [email protected]
The autolytic peptidoglycan hydrolases of Streptococcus faecium
Ann. Inst. Pasteur /Microbiol.
~) ELSEVIER Paris 1985
THE
1985, 136 A, 63-66
AUTOLYTIC OF
PEPTIDOGLYCAN
STREPTOCOCCUS
HYDROLASES
FAECIUM
by G. D. Shockman, T. Kawamura, J. F. Barrett and D. L. Dolinger
Department o/ Microbiology and Immunology, Temple University School o[ Medicine, Philadelphia, PA 19140 (USA)
SUMMARY.
Streptococcus ]aecium ATCC 9790 possesses two peptidoglycan hydrolase activities. The first enzyme, an N-acetylmuramoylhydrolase, has been purified and has been shown to be a glucoenzyme. Studies of hydrolysis of soluble, linear uncross-linked peptidoglycan chains showed that the enzyme bound strongly to the non-reducing ends of the chains and then sequentially (processively) hydrolysed susceptible bonds in that chain. The second peptidoglycan hydrolase does not appear to be a glycoprotein and differs from the first enzyme in substrate specificity and mechanism of hydrolysis. The presence of two partially redundant activities which may play different roles in surface growth and division could, at least in part, explain previous difficulties in obtaining mutants which completely lack autolytic activity. KEY-WORDS: Streptococcus /aecium, Peptidoglycan, Hydrolysis, Autolysis, Muramidase, Porin.
INTRODUGTION.
As a bacterial cell increases in volume and divides, it must maintain the osmotically protective properties of its cell wall while it expands its wall surface area to enclose increased cellular contents. It was hypothesized [7, 20] that insertion of new peptidoglycan (PG) precursors into a pre-existing, protective PG exoskeleton involves the selective and exquisitely regulated action of endogenous PG hydrolases. Activities capable of hydrolysing bonds in their own PG have been found in a wide variety of both Gram-positive and Gram-negative species ([7, 20, 23] for recent reviews). Generally, endogenous PG hydrolases have been considered to be synonymous with autolysins. However, it is now clear that not all PG hydrolases are capable of dissolving cells or isolated walls; and therefore, in the strict sense, may not be cc autolysins ~, (see [23] for a review). Non-autolytic PG hydrolases would include activities that hydrolyse only a limited number of bonds essential to the two- or three-dimensional PG network, perhaps not even resulting in the release of soluble products, as well as enzymes that hydrolyse bonds which are not essential to maintain the overall structure. Numerous potential functions of PG hydrolases have been proposed [7, 20], including roles in cell wall assembly, surface enlargement, morphogenesis and cell division. Clearly, all postulated roles require tightly coordinated and regulated systems. Manuscrit resu le 12 septembre 1984.
64
G.D.
l~ESULTS
AND
SHOCKMAN AND COLL.
DISCUSSION,
The peptidoglycan hydrolase system o/ S. faecium A TCC 9790 - - General considerations. Although many bacterial species contain more than one PG hydrolase, a variety of data indicate that S./aecium contains only a single specificity of PG hydrolase activity [8, 14, 21, 27], a ~l-4-N-acetylmuramoylhydrolase (EC 3.2.1.17, muramidase). Over 95~o of the autolysin activity recovered from disrupted cells is in the cell wall fraction [27], as both a latent, proteinase-activatable zymogen and in an active form. Recent experiments (D. L. Dolinger and G. D. Shockman, unpublished results) using a protein renaturation technique [9] showed the presence of the presumably unfolded (denatured) activated form of the enzyme in culture supernates. Several types of evidence strongly indicated that, in exponential-phase cells, the autolysin is concentrated in recently assembled wall [14, 26] at nascent crosswalls [10]. Evidence exists for a role for autolysin action in the cell cycle, consistent with its participation in surface enlargement and division in a closely regulated fashion ([11, 12, 13, 24, 25] and L. Daneo-Moore, unpublished data). This potentially suicidal activity appears to be regulated at several levels, including proteinase activation [19] topologically [10, 14] and via the inhibitory action of regulatory ligands such as lipoteichoic acids and certain lipids, notably cardiolipin [4, 5, 6]. The roles of these regulatory processes in surface growth and division remain to be established.
Purificalion and some properties of the autolgtic muramidase o/ S. faecium. Recently, we purified the latent form of this muramidase to near homogeneity [15]. The latent form (130,000 d) can be hydrolysed to the active form (87,000 d) with trypsin. Both forms contain convalently linked monomeric and oligomeric glucose units. Thus, this enzyme is a glucoenzgme, and perhaps the first proven example of a bacterial glycoenzyme. The action of this muramidase on soluble, linear uncross-linked PG (s-PG) chains, produced by penicillin-inhibited cultures, was studied in some detail [1, 2, 3]. The use of these substrates, especially the s-PG produced by autolysis-defective strains of S. /aecium, which are about 45-disaccharide peptide (DSP) units long and are fully substituted with peptide side chains, permitted kinetic analyses which heretofore have not been possible. In contrast to the action of hen egg-white lysozyme, the S. ]aecium muramidase failed to catalyse transglycosylation. Hydrolysis of the s-PG of S. faecium resulted in the production of nearly only one product, the DSP monomers, even after only short periods of hydrolysis. A variety of data, including the results of studies of hydrolysis kinetics and substrate competition, suggested t h a t each enzyme molecule firmly bound to an s-PG chain, and hydrolysed all susceptible bonds at the rate of 91 bonds per minute before it was released, suggesting a specific enzyme binding site on each s-PG chain. Such a site and the direction of hydrolysis of chains were established by using s-PG chains labelled at reducing ends with 3H by reduction of the terminal MurNAc residue with 3H-NaBH4 and labelled at non-reducing ends by the addition of 14C-galactose via the action of galactosyltransferase [22]. At high enzyme to substrate ratios, 14C-labelled products were released immediately and 3H-labelled products were only released after a 1.5- to 2-min lag. DS1 :) = d i s a c c h a r i d e p e p t i d e . PG = peptidoglycan. s-PG = soluble (linear uncross-linked) PG.
PEPTIDOGLYCAN HYDROLASES OF S. F A E C I U M
65
Thus, the muramidase is a processive exodisaccharidase which binds to non-reducing ends of chains and then successively hydrolyses the bonds between MurNAc and GlcNAc until it reaches the end of t h a t chain. At low enzyme to substrate ratios, biphasic hydrolysis curves were observed. Rapid hydrolysis of most susceptible bonds in the DSP45 chains was followed by a lag (attributed to slow release of the enzyme from terminal residues) and then a second period of rapid hydrolysis. It will be of interest to determine exactly how such well organized activity may play a role in wall surface enlargement and division.
The second peplidoglycan hydrolase o/S. faecium. Recently we discovered and partially purified a second PG hydrolase in S./aecium which does not seem to be a glycoprotein. Unlike the enzyme activity described above, this second enzyme appears to have little ability to dissolve intact walls of S./aecium, suggesting t h a t it may not be a true (( autolysin )) and explaining why it had not been seen earlier. It rapidly dissolves walls of Micrococcus luleus and the PG fraction S./aecium. The distinct substrate specificities of the two enzymes provide differential assays. The ability of this enzyme to attack the PG fraction of S. /aecium suggests t h a t it can hydrolyse at least some bonds in intact walls, although it is apparently unable to dissolve such walls. This type of subtle and selective ~( nicking )) action could be of considerable importance in postincorporation modification and ((remodeling)) of walls. Furthermore, since this second PG hydrolase activity has been found in the culture medium, it could play a role in cell separation (D. L. Dolinger, unpublished data). Although the two activities may have different functions in cells, it seems possible t h a t a defect in one may be partially compensated by the presence of the other. ]'he presence of two PG hydrolases in S. [aeeium, at least in part, explains previous difficulties in obtaining mutants which completely lack autolytie activity. Such redundancy of PG hydrolase activity could be a feature common to many bacteria and account for the dearth of truly autolytically negative mutants. Evidence indicating the presence of two PG transglycosylases in Escherichia coli has also been reported [28]. The redundancy of PG hydrolase activities in S. /aecium and the proposed compensatory activity of one or the other hydrolase is, in part, similar to the presence of multiple penicillin-binding proteins [28] and porin proteins in E. eoli [18]. In the latter case, mutants which lack one or even two porin proteins synthesize and incorporate into their membrane a third protein which can function as a porin. MOTS-CLES : Streptococcus [aecium, Peptidoglycane, Hydrolyse, Autolyse; Muramidase, Porine. ACKNOWLEDGEMENT
This work was supported by US Public Health Service Research Grant A105044.
REFERENCES [1] BARRETT, J. F., DOLINGER, D. L., SCHRAMM, V. L. & SHOCKMAN, G. D.,
J. biol. Chem., 1984, 259, 11818-11827. [2] BARRETT, J. F., SCHRAMM, V. L. & SHOCKMAN, G . D., J. Bact., 1984, 159,
520-526. [3] BARNETT, J. F. & SHOCKMAN, G. D., J. Bad., 1984, 159, 511-519. [4] CLEVELAND, R. F., DANEO-MOORE, L., WICKEN, A. J. & SHOCKMAN, G. D., d. Bad., 1976, 127, 1582-1584. Ann. 51icrobiol. (Inst. Pasteur), 136 A, n~ 1, 1985. 5
66
G. D. SHOCKMAN AND COLL.
[5] CLEVELAND, R. F., HOLTJE, J.-V., WICKEN, A. J., TOMASZ, A., DANEOMOORE, L. • SHOCKMAN,G. D., Biochem. biophys. Res. Commun., 1975, 67, 1128-1135. [6] CLEVELAND,R. F., ~r A. J., DANEO-MOoRE, L. & SHOfiKMAN, G. D., d. Bad., 1976, 126, 192-197. [7] DANEO-MooRE, L. & S~IOCKMAN,G. D., in (~Cell surface reviews ,,, 4 (p. 597). Elsevier/North Holland, Amsterdam, 1977. [8] DEZELEE, P. & SHOCKMAN, G. D., J. biol. Chem., 1975, 250, 6808-6816. [9] HAGER, D. A. & BURGESS, ]:{. R., Analgl. Biochem., 1980, 109, 76-86. [10] HIGGINS, M. L., POOLEY, H. M. r SHOCKMAN,G. D., d. Bact., 1970, 103, 504-512. [11] HINKS, R. P., DANEO-MooRE, L. & SHOCKMAN,G. D., J. Bact., 1978, 133, 822. [12] HINKS, R. P., DANEO-MooRE, L. ~r SHOCKMAN, G. D., d. Bacl., 1978, 134, 1074. [13] HINKS, R. P., DANEO-MooRE, L. & SHOCKMAN,G. D., J. Bad., 1978, 134, 1188. [14] JOSEPH, R. & SHOCKMAN,G. D., J. Bacl., 1976, 127, 1482-1493. [15] KAWAMURA, T. & SHOCKMAN,G. D., g. biol. Chem., 1983, 258, 9514-9521. [16] KAWAMURA, T. & SHOCKMAN,G. D., F E M S Microbiol. Lellers, 1983, 19, 65-69. [17] METT, H., KECK, W., FUNK, A. 8r SCHWARZ,U., J. Bad., 1980, 144, 45-52. [18] NIKAIDO, H., in (( Bacterial outer membranes: biogenesis and function ,,, (p. 361). John Wiley & Sons Ltd, Chichester, 1979. [19] POOLEY, H. M. & SHOCKMAN,G. D., J. Bad., 1970, 103, 457-466. [20] ROGERS, H. J., PERKINS, H. R. & WARD, J. B., Microbial cell walls and membranes. Chapman & Hall Ltd, London, 1980. [21] I~OSENTHAL,B. S., JUNGKIND, D., DANEO-MooRE, L. & SHOCKMAN, G. D., J. Bad., 1975, 124, 398-409. [22] SCHINDLER,M., MIRELMAN,D. & SCHWARZ,U., Europ. J. Biochem., 1976, 71, 131-134. [23] SHOCKMAN,G. D. ~ BARRETT, J. F., Ann. Rey. Microbiol., 1983, 37, 501-527. [24] SHOCKMAN, G. D., DANEO-MOORE,L. 8r HIGGINS, M. L., Ann. N. Y. Acad. Sci., 1974, 235, 161-197. [25] SHOCKMAN,G. D., DANEO-MooRE, L., McDOWELL, T. D. & WONG, W., in (( ~-Lactam antibiotics ,,, (p. 31). Academic Press, London, New York, 1981. [26] SHOCKMAN, G. D., POOLEY, H. M. 3r THOMPSON, J. S., J. Bacl., 1967, 94, 1525-1530. [27] SHOCKMAN,G. D., THOMPSON,J. S. ~r CONOVER,M. J., Biochemislrg, 1967, 6, 1054-1065. [28] SPRATT, B. 6., Phil. Trans. roy. Soc. London (Ser. B), 1980, 289, 273-283.
~) ELSEVIER Paris 1985
THE
1985, 136 A, 63-66
AUTOLYTIC OF
PEPTIDOGLYCAN
STREPTOCOCCUS
HYDROLASES
FAECIUM
by G. D. Shockman, T. Kawamura, J. F. Barrett and D. L. Dolinger
Department o/ Microbiology and Immunology, Temple University School o[ Medicine, Philadelphia, PA 19140 (USA)
SUMMARY.
Streptococcus ]aecium ATCC 9790 possesses two peptidoglycan hydrolase activities. The first enzyme, an N-acetylmuramoylhydrolase, has been purified and has been shown to be a glucoenzyme. Studies of hydrolysis of soluble, linear uncross-linked peptidoglycan chains showed that the enzyme bound strongly to the non-reducing ends of the chains and then sequentially (processively) hydrolysed susceptible bonds in that chain. The second peptidoglycan hydrolase does not appear to be a glycoprotein and differs from the first enzyme in substrate specificity and mechanism of hydrolysis. The presence of two partially redundant activities which may play different roles in surface growth and division could, at least in part, explain previous difficulties in obtaining mutants which completely lack autolytic activity. KEY-WORDS: Streptococcus /aecium, Peptidoglycan, Hydrolysis, Autolysis, Muramidase, Porin.
INTRODUGTION.
As a bacterial cell increases in volume and divides, it must maintain the osmotically protective properties of its cell wall while it expands its wall surface area to enclose increased cellular contents. It was hypothesized [7, 20] that insertion of new peptidoglycan (PG) precursors into a pre-existing, protective PG exoskeleton involves the selective and exquisitely regulated action of endogenous PG hydrolases. Activities capable of hydrolysing bonds in their own PG have been found in a wide variety of both Gram-positive and Gram-negative species ([7, 20, 23] for recent reviews). Generally, endogenous PG hydrolases have been considered to be synonymous with autolysins. However, it is now clear that not all PG hydrolases are capable of dissolving cells or isolated walls; and therefore, in the strict sense, may not be cc autolysins ~, (see [23] for a review). Non-autolytic PG hydrolases would include activities that hydrolyse only a limited number of bonds essential to the two- or three-dimensional PG network, perhaps not even resulting in the release of soluble products, as well as enzymes that hydrolyse bonds which are not essential to maintain the overall structure. Numerous potential functions of PG hydrolases have been proposed [7, 20], including roles in cell wall assembly, surface enlargement, morphogenesis and cell division. Clearly, all postulated roles require tightly coordinated and regulated systems. Manuscrit resu le 12 septembre 1984.
64
G.D.
l~ESULTS
AND
SHOCKMAN AND COLL.
DISCUSSION,
The peptidoglycan hydrolase system o/ S. faecium A TCC 9790 - - General considerations. Although many bacterial species contain more than one PG hydrolase, a variety of data indicate that S./aecium contains only a single specificity of PG hydrolase activity [8, 14, 21, 27], a ~l-4-N-acetylmuramoylhydrolase (EC 3.2.1.17, muramidase). Over 95~o of the autolysin activity recovered from disrupted cells is in the cell wall fraction [27], as both a latent, proteinase-activatable zymogen and in an active form. Recent experiments (D. L. Dolinger and G. D. Shockman, unpublished results) using a protein renaturation technique [9] showed the presence of the presumably unfolded (denatured) activated form of the enzyme in culture supernates. Several types of evidence strongly indicated that, in exponential-phase cells, the autolysin is concentrated in recently assembled wall [14, 26] at nascent crosswalls [10]. Evidence exists for a role for autolysin action in the cell cycle, consistent with its participation in surface enlargement and division in a closely regulated fashion ([11, 12, 13, 24, 25] and L. Daneo-Moore, unpublished data). This potentially suicidal activity appears to be regulated at several levels, including proteinase activation [19] topologically [10, 14] and via the inhibitory action of regulatory ligands such as lipoteichoic acids and certain lipids, notably cardiolipin [4, 5, 6]. The roles of these regulatory processes in surface growth and division remain to be established.
Purificalion and some properties of the autolgtic muramidase o/ S. faecium. Recently, we purified the latent form of this muramidase to near homogeneity [15]. The latent form (130,000 d) can be hydrolysed to the active form (87,000 d) with trypsin. Both forms contain convalently linked monomeric and oligomeric glucose units. Thus, this enzyme is a glucoenzgme, and perhaps the first proven example of a bacterial glycoenzyme. The action of this muramidase on soluble, linear uncross-linked PG (s-PG) chains, produced by penicillin-inhibited cultures, was studied in some detail [1, 2, 3]. The use of these substrates, especially the s-PG produced by autolysis-defective strains of S. /aecium, which are about 45-disaccharide peptide (DSP) units long and are fully substituted with peptide side chains, permitted kinetic analyses which heretofore have not been possible. In contrast to the action of hen egg-white lysozyme, the S. ]aecium muramidase failed to catalyse transglycosylation. Hydrolysis of the s-PG of S. faecium resulted in the production of nearly only one product, the DSP monomers, even after only short periods of hydrolysis. A variety of data, including the results of studies of hydrolysis kinetics and substrate competition, suggested t h a t each enzyme molecule firmly bound to an s-PG chain, and hydrolysed all susceptible bonds at the rate of 91 bonds per minute before it was released, suggesting a specific enzyme binding site on each s-PG chain. Such a site and the direction of hydrolysis of chains were established by using s-PG chains labelled at reducing ends with 3H by reduction of the terminal MurNAc residue with 3H-NaBH4 and labelled at non-reducing ends by the addition of 14C-galactose via the action of galactosyltransferase [22]. At high enzyme to substrate ratios, 14C-labelled products were released immediately and 3H-labelled products were only released after a 1.5- to 2-min lag. DS1 :) = d i s a c c h a r i d e p e p t i d e . PG = peptidoglycan. s-PG = soluble (linear uncross-linked) PG.
PEPTIDOGLYCAN HYDROLASES OF S. F A E C I U M
65
Thus, the muramidase is a processive exodisaccharidase which binds to non-reducing ends of chains and then successively hydrolyses the bonds between MurNAc and GlcNAc until it reaches the end of t h a t chain. At low enzyme to substrate ratios, biphasic hydrolysis curves were observed. Rapid hydrolysis of most susceptible bonds in the DSP45 chains was followed by a lag (attributed to slow release of the enzyme from terminal residues) and then a second period of rapid hydrolysis. It will be of interest to determine exactly how such well organized activity may play a role in wall surface enlargement and division.
The second peplidoglycan hydrolase o/S. faecium. Recently we discovered and partially purified a second PG hydrolase in S./aecium which does not seem to be a glycoprotein. Unlike the enzyme activity described above, this second enzyme appears to have little ability to dissolve intact walls of S./aecium, suggesting t h a t it may not be a true (( autolysin )) and explaining why it had not been seen earlier. It rapidly dissolves walls of Micrococcus luleus and the PG fraction S./aecium. The distinct substrate specificities of the two enzymes provide differential assays. The ability of this enzyme to attack the PG fraction of S. /aecium suggests t h a t it can hydrolyse at least some bonds in intact walls, although it is apparently unable to dissolve such walls. This type of subtle and selective ~( nicking )) action could be of considerable importance in postincorporation modification and ((remodeling)) of walls. Furthermore, since this second PG hydrolase activity has been found in the culture medium, it could play a role in cell separation (D. L. Dolinger, unpublished data). Although the two activities may have different functions in cells, it seems possible t h a t a defect in one may be partially compensated by the presence of the other. ]'he presence of two PG hydrolases in S. [aeeium, at least in part, explains previous difficulties in obtaining mutants which completely lack autolytie activity. Such redundancy of PG hydrolase activity could be a feature common to many bacteria and account for the dearth of truly autolytically negative mutants. Evidence indicating the presence of two PG transglycosylases in Escherichia coli has also been reported [28]. The redundancy of PG hydrolase activities in S. /aecium and the proposed compensatory activity of one or the other hydrolase is, in part, similar to the presence of multiple penicillin-binding proteins [28] and porin proteins in E. eoli [18]. In the latter case, mutants which lack one or even two porin proteins synthesize and incorporate into their membrane a third protein which can function as a porin. MOTS-CLES : Streptococcus [aecium, Peptidoglycane, Hydrolyse, Autolyse; Muramidase, Porine. ACKNOWLEDGEMENT
This work was supported by US Public Health Service Research Grant A105044.
REFERENCES [1] BARRETT, J. F., DOLINGER, D. L., SCHRAMM, V. L. & SHOCKMAN, G. D.,
J. biol. Chem., 1984, 259, 11818-11827. [2] BARRETT, J. F., SCHRAMM, V. L. & SHOCKMAN, G . D., J. Bact., 1984, 159,
520-526. [3] BARNETT, J. F. & SHOCKMAN, G. D., J. Bad., 1984, 159, 511-519. [4] CLEVELAND, R. F., DANEO-MOORE, L., WICKEN, A. J. & SHOCKMAN, G. D., d. Bad., 1976, 127, 1582-1584. Ann. 51icrobiol. (Inst. Pasteur), 136 A, n~ 1, 1985. 5
66
G. D. SHOCKMAN AND COLL.
[5] CLEVELAND, R. F., HOLTJE, J.-V., WICKEN, A. J., TOMASZ, A., DANEOMOORE, L. • SHOCKMAN,G. D., Biochem. biophys. Res. Commun., 1975, 67, 1128-1135. [6] CLEVELAND,R. F., ~r A. J., DANEO-MOoRE, L. & SHOfiKMAN, G. D., d. Bad., 1976, 126, 192-197. [7] DANEO-MooRE, L. & S~IOCKMAN,G. D., in (~Cell surface reviews ,,, 4 (p. 597). Elsevier/North Holland, Amsterdam, 1977. [8] DEZELEE, P. & SHOCKMAN, G. D., J. biol. Chem., 1975, 250, 6808-6816. [9] HAGER, D. A. & BURGESS, ]:{. R., Analgl. Biochem., 1980, 109, 76-86. [10] HIGGINS, M. L., POOLEY, H. M. r SHOCKMAN,G. D., d. Bact., 1970, 103, 504-512. [11] HINKS, R. P., DANEO-MooRE, L. & SHOCKMAN,G. D., J. Bact., 1978, 133, 822. [12] HINKS, R. P., DANEO-MooRE, L. ~r SHOCKMAN, G. D., d. Bacl., 1978, 134, 1074. [13] HINKS, R. P., DANEO-MooRE, L. & SHOCKMAN,G. D., J. Bad., 1978, 134, 1188. [14] JOSEPH, R. & SHOCKMAN,G. D., J. Bacl., 1976, 127, 1482-1493. [15] KAWAMURA, T. & SHOCKMAN,G. D., g. biol. Chem., 1983, 258, 9514-9521. [16] KAWAMURA, T. & SHOCKMAN,G. D., F E M S Microbiol. Lellers, 1983, 19, 65-69. [17] METT, H., KECK, W., FUNK, A. 8r SCHWARZ,U., J. Bad., 1980, 144, 45-52. [18] NIKAIDO, H., in (( Bacterial outer membranes: biogenesis and function ,,, (p. 361). John Wiley & Sons Ltd, Chichester, 1979. [19] POOLEY, H. M. & SHOCKMAN,G. D., J. Bad., 1970, 103, 457-466. [20] ROGERS, H. J., PERKINS, H. R. & WARD, J. B., Microbial cell walls and membranes. Chapman & Hall Ltd, London, 1980. [21] I~OSENTHAL,B. S., JUNGKIND, D., DANEO-MooRE, L. & SHOCKMAN, G. D., J. Bad., 1975, 124, 398-409. [22] SCHINDLER,M., MIRELMAN,D. & SCHWARZ,U., Europ. J. Biochem., 1976, 71, 131-134. [23] SHOCKMAN,G. D. ~ BARRETT, J. F., Ann. Rey. Microbiol., 1983, 37, 501-527. [24] SHOCKMAN, G. D., DANEO-MOORE,L. 8r HIGGINS, M. L., Ann. N. Y. Acad. Sci., 1974, 235, 161-197. [25] SHOCKMAN,G. D., DANEO-MooRE, L., McDOWELL, T. D. & WONG, W., in (( ~-Lactam antibiotics ,,, (p. 31). Academic Press, London, New York, 1981. [26] SHOCKMAN, G. D., POOLEY, H. M. 3r THOMPSON, J. S., J. Bacl., 1967, 94, 1525-1530. [27] SHOCKMAN,G. D., THOMPSON,J. S. ~r CONOVER,M. J., Biochemislrg, 1967, 6, 1054-1065. [28] SPRATT, B. 6., Phil. Trans. roy. Soc. London (Ser. B), 1980, 289, 273-283.