Molecular Microbiology (1992) 6(7), 921-931
Studies on the structure and function of the A/-terminai domain of the pneumococcai murein hydrolases J. M. Sanz, E. Diaz and J. L. Garcia* Unidad de Genetica Bacteriana. Centro de Investigaciones Biologicas (CSIC), Velazquez 144. 28006 Madrid, Spain. Summary The structures of the choline-dependent pneumococcai murein hydrolases, LYTA amtdase and CPL1 Iysozyme, and the choline-independent CPL7 Iysozyme were analysed by controlled proteolytic digestions. The trypsin cleavage of the CPL1 and CPL7 iysozymes produced two resistant polypeptides, F1 and F7 respectively, corresponding to the Nterminal domain of the enzymes, whereas the amidase LYTA was completely hydrolysed by the protease. Interestingly, the F1 and F7 fragments showed a low, but significant, choline-independent Iysozyme activity. Choline reduced the rate of proteolytic hydrolysis of choline-dependent enzymes, suggesting that the C-terminal choline-binding domain adopts a more resistant conformation in the presence of the ligand. On the other hand, the regions encoding the /V-terminal domains of the three enzymes have been cloned and expressed in Escherichia coli, showing that these domains adopt an active conformation even in the absence of their C-terminal domains. The lower activity shown by the catalytic domains when compared with that of the complete enzymes suggests that the acquisition of a substratebinding domain represents a noticeable evolutionary advantage for enzymes that interact with polymeric substrates, allowing them to achieve a higher catalytic efficiency. These results strongly reinforce the hypothesis that the pneumococcai murein hydrolases have been originated by fusion of two structural and functional independent domains, and provide new experimental support to the theory of modular evolution of proteins.
Introduction The evidence that many genes (proteins) have evolved Received 16 October, 1991; revised and accepted 20 December, 1991. •For correspondence. Tel. (1) 5644570; Fax (1) 5627518.
by fusion of pre-existing sequences is increasing, in so far as more gene (protein) sequences are being determined (Gilbert, 1978). Modularization is the central concept of this particular evolutionary pathway which is known as the modular theory of evolution (DuBose and HartI, 1989). The length of the sequences (modules) that participate in the fusions varies according to the different levels of protein organization. Although the sequences encoding protein domains represent the clearest reference to settle the concept of module (Darnell and Doolittle. 1986; Gilbert, 1981),. those encoding short chains of amino acids exhibiting a specific secondary structure have also been included in this concept (DuBose and HartI, 1989; Ohno, 1987). In addition, according to the modular theory, it has been proposed that viral genomes have evolved by the fusion of different genes, or clusters of genes, which provide specific physiological functions (Botstein, 1980; Campbell, 1988). We have recently proposed that the system provided by the pneumococcai murein hydrolases constitutes an excellent experimental model for studying the postulates of the modular theory of evolution (Garcia et ai, 1990). The genes encoding five pneumococcai murein hydrolases have been cloned and sequenced (Garcia et ai, 1986; 1988; 1990; Romero etai. 1990a); two enzymes have been characterized as /V-acetylmuramoyl-L-alanine amidases (EC.3.5.1.28), i.e. the major pneumococcai autolysin (LYTA amidase) encoded by the lytA gene (Garcia e/a/., 1986), and the HBL amidase encoded by the hbl gene of the pneumococcai bacteriophage HB-3 (Romero etai, 1990a), The other three enzymes have been shown to be Iysozymes (EC 3.2.1.17), i.e. the CPL1, CPL7 and CPL9 Iysozymes, which are encoded by the genes cpli. cp/7and cp/9 of the pneumococcai bacteriophages Cp-1, Cp-7 and Cp-9, respectively (Garcia et ai, 1988; 1990). On the basis of sequence comparisons, we have postulated that these proteins have been built up by fusion of two independent domains (modules): the W-terminal catalytic domain and the C-terminal binding domain which recognizes specific components of the pneumococcai cell wall (Garcia etai. 1988; 1990). With the sole exception of the CPL7 Iysozyme, all of the pneumococcai murein hydrolases that have been biochemically characterized so far, including the second autolysin identified in pneumococcus, a glucosaminidase (Garcia et ai. 1989), as well as the amidase encoded by the pneumococcai
922 J. M. Sanz, E. Diaz andJ. L. Garcia
B
A kDa
A
B
C
D
E
F
G
H
I
92,0 66.0--— 43.0-WP 31 . 0 -
kDa
A
B
C
D
E
F
G
H
I
J
92.066.045,0-
— 31.0-
21. 5 -
m
14 . 3 -
•»
21.514.3-
Fig. 1. SDS-PAGE of the controlled digestions of CPL1 and CPL7 lysozymes by trypsin. The CPL1 (panel A) and CPL7 (panel B) purified enzymes were digestedbytrypsinat 37"C at a protease/protein ratio of 1/100 (w/w) and 1/25 [w/w). respectively, and the results were analysed on a 15% SDS-polyacrylamide gel. The molecular masses of the markers are expressed in ktlodaltons. A. Lane A, proiein standards: lanes B, C, D, E and F, samples from the digestion mixture cf the CPL1 lysozyme taken at 0, 5.10. 20 and 40 min, respectively; iane G. purified tragment F I ; lanes H and I. samples of a digestion carried out in the presence of 120 mM choline and taken at 10 and 40 min. B. Lane A, protein standards; lanes B,C, D.E, F, G, Hand I, sampies trom the digesticn mixture of the CPL7 lysozyme taken at 0, 2, 5, 10, 30. 60, 90 and 120 mm. respectively; lane J. purified fragment F7. Proteins were visualized with Coomassie Brilliant Blue.
phage Dp-1 (Garcia et ai. 1983), depend on the presence of choline in the teichoic acid of the cell wall for activity. The peculiar presence ot choline in the celi wall of pneumococcus has been suggested to act as an evolutionary element of selective pressure for the cholinedependent murein hydrolases, preserving the existence of the choline-binding domain {Garcia et ai. 1988). The choline-dependent enzymes are unable to degrade efficiently pneumococcal cell walls where choline has been replaced by the structural analogue ethanolamine (Giudicelli and Tomasz, 1984). This property is used as a rapid test to distinguish between choline-dependent and choline-independent enzymes. Ethanolamine-containing cell walls are only efficiently degraded by the CPL7 lysozyme, an enzyme that has a C-terminal domain which contains three identical repeated amino acid sequences, but presenting a primary structure completely different from that of the choline-dependent enzymes (Garcia ef ai, 1990). It is alsc interesting that simiiar amino acid repeats have also been found in the C-terminal region of a family of other ligand-binding proteins that appears to be of modular design (Wren, 1991). Although we have already presented experimental evidences supporting the modular organization of the pneumococcal murein hydrolases (Diaz et ai, 1990; 1991; Garcia etai, 1988; 1990), and even the sequence encoding the choline-binding domain had been expressed and studied (Sanchez-Puelles et ai. 1990), a direct demonstration that the A/-terminal domain of these enzymes might act as an independent and functional catalytic protein was missing. The results reported in this paper have allowed us to determine several structural and functional
properties of the A/-terminal modules of the LYTA amidase and the CPL1 and CPL7 lysozymes, demonstrating. by using two different experimental approaches, that these modules are folded as independent protein domains that can act as choline-independent pneumococcal cell wall hydrolases. This finding reinforces our hypothesis that the pneumoooccal murein hydrolases should have evolved by the fusion of pre-existing modules.
Results Controlled proteolytic digestions of pneumococcal murein hydrolases The two domain structure of the pneumococca! murein hydrolases. the LYTA amidase and the CPL1 and CPL7 lysozymes were analysed by controlled proteolytic digestions using different proteases. We found that whereas the LYTA amidase was completely digested by trypsin at a protease/protein ratio of 1/60 (w/w) after 60 min of incubation at 3 7 X . the hydrolysis of the CPLI and CPL7 lysozymes at a protease/protein ratio of 1/50 and 1/10 (w/w) respectively, produced, in both cases, after 120 min of incubation at 37"C, a unique resistant polypeptide of high molecular weight, named FI (M, =24000) (Fig. IA, lane G) and F7 (M, =25000) (Fig. 1B, lane J), respectively. The fragments, F1 and F7, were not further degraded even after three hours of incubation in the conditions mentioned. Moreover, these fragments could also be obtained using higher trypsin concentrations, i.e. at a
The catalytic domain of pneumococcal murein hydrolases
kDa
A
B
D
92.0 — 66.0 — 43.0 — 31 . 0 21,5 —
H
923
Fig. 2. SDS-PAGE of ttie controlled dtgesiions o( the CPLl lysozyme by ti-ctiymotrypsin The CPL1 tysozyme was digested by a-chymoifypsin al 37 C at a protease protetn ratio ot 1 30 (ww). and analysed on a 15% SDS-potyacrylamide gel The molecular masses of ttie marttefS are expressed m kilodartons. Lane A. protein standards: lanes B. C, D. E and F. samples taken at 0. 5. 15. 30 and 60 mm. respectively: lanes G, H. t and J. samples from a digestion camed out m the presence of 120 mM cholme and taken at 5.15. 30 and 60 min. respectively Proteins were visualized witti Coomassie Bnltiant Blue
14.3 —
protease/protein ratio of 1/1 (w/w), suggesting that these core potypeptides were highly resistant to trypsin digestion. The analysis ol the products obtained during the time course of the proteolysis indicated that the CPLl and CPL7 lysozymes were cleaved in different ways (Fig. 1). Whereas the appearance o( the Fl resistant fragment was not preceded by any other high-molecular-weight protein intermediates {Fig. 1A). cleavage of the CPL7 lysozyme resulted in the simultaneous appearance of three fragments, the F7 fragment being the smaller sized polypeptide (Fig 1 B). On the other hand, the digestion of the CPLl lysozyme with a-chymotrypsin at a protease/protein ratio of 1/30 (w/w) for 60 min at 37"C yielded a resistant fragment with a molecular weight similar to that observed for the F1 trypsin fragment (Fig. 2). This fragment was also remarkably resistant to (i-chymotrypsin, since a prolonged incubation in the above conditions or the use of a higher protease concentration {1/1. w/w) yielded identical results. However, the LYTA amidase and the CPL7 lysozyme were more susceptible to the cleavage with this protease than the CPLl lysozyme, since at a protease/protein ratio of 1/30 (wM). both enzymes were completely digested after 60 min of incubation at 37 C. The pneumococcal murein hydrolases were very susceptible to the treatment with proteinase K. since this protease, at a protease/protein concentration of 1/30 (w/w), completely hydrolysed the enzymes after 120 min of incubation at 37'C. However, using lower relative concentrations of proteinase K (1/100. w/w). we could obtain a resistant fragment by digesting the CPU lysozyme at 37''C for 60 min. Nevertheless, this fragment disappeared when the incubation was prolonged for 120 min (data not shown). Since it has been observed that choline induces a conformational change in the structure of the (^terminal domain of the choline-dependent enzymes (SanchezPuelles ef a/.. 1990). we decided to investigate rts influ-
ence on the protease sensitivity of the pneumococcal murein hydrolases. Figures 1 and 2 show that choline slowed the rate of hydrolysis of the CPLl lysozyme by trypsin and a-chymotrypsin. The protective effect of chotine cannot be ascribed to an increase in the ionic strength, since an equivalent concentration of NaCI did not produce any effect. Choline also slows the rate of proteolysis of the choline-dependent LYTA amidase whereas, as expected, the choline-independent CPL7 lysozyme was not protected by this compound (data not shown). Analysis of the Fl and F7 trypsin resistant fragments To determine which region of the CPLl and CPL7 lysozymes corresponded to the Fl and F7 trypsin-resistant fragments, these polypeptides were purified and their A/-terminal sequences were analysed by Edman degradation. The same amino acid sequence. Val-Lys-Lys-AsnAsp-Leu-Phe-Val. was obtained for both fragments. This sequence is identical to the /V-terminal sequence of the CPLl and CPL7 lysozymes (Garcia ef al.. 1988; 1990), indicating that both polypeptides correspond to the W-terminal half of the proteins. To establish the exact position of the trypsin cleavage site we attempted the analysis of the 0-terminat amino acids of both fragments using carboxypeptidase Y, but unfortunately they were resistant to the carboxypeptidase hydrolysis, even in the presence of 6 M urea. According to the previously published amino ackj sequence of the CPLl lysozyme. this enzyme contains three cysteine residues (Garcia et al.. 1988). Taking into account the molecular weight of the Fl fragment, estimated by sodium dodecyl sulphate/polyacrylamide gel electrophoresis (SDS-PAGE). one of the cysteine residues {Cys-239) should be located near the putative trypsin cleavage site. Therefore the analysis of the cysteine content of the Fl fragment with 5,5'-dithio-bis{2nitrobenzoic) acid (DTNB) could allow us to determine
924
J. M. Sanz. E. Diaz and J. L Garcia
lOOi
[-•
\
75 50-
o
I
\ \
\
\
25 0
1
•.
\
(
-a.
:
• • : • ,
h- —=^-^?—
50
fl—
100 150 [NaCi] (mM)
_
j
1
1
200
250
Rg. 3. Influence ol ionic strength on the activity of the CPLI, CPL7, FI and F7 proteins. The CPLI (O—O). CPL7 (O O), FI ( • - - • ) and F7 (A---A) purified proteins were assayed in 20 mM phosphate buffer, pH 6.0, upon hydrolysis of the choline-containmg cell walls in the presence of different concentrations of NaCI. Results are expressed in terms of percentage activity relative to the absence ot NaCI in the assay mixture.
the location of the cleavage site more precisely. We found that the CPLI lysozyme contains three free cysteines (3.0 cysteines/molecule), whereas the FI fragment has only two (1.7 cysteines/molecule). assuming a M^ of 24 000 for this fragment (Fig. 1A, lane G). These results indicated that the cysteines of the CPLI lysozyme did not form a disulphide bridge, but also suggested that the trypsin cleavage site should be located before Cys-239.
Assay of the murein hydrolase activity of the F1 and F7 trypsin-resistant fragments According to the proposed modular organization of the pneumococcal murein hydrolases, the F1 and F7 fragments should correspond to the postulated catalytic module of the CPLI and CPL7 lysozymes (Garcia et ai, 1988; 1990) and therefore it seemed likely that the F1 and F7 fragments could still retain a detectable enzymatic activity. The results shown in Table 1 demonstrated that when the hydrolytic specific activity of the purified F1 and F7 polypeptides was assayed on pneumococcal cell walls containing choline or ethanolamine, these fragments turned out still to be active, although their specific activities were lower than those of the native lysozymes. Interestingly, the ethanolamine-containing cell walls were hydrolysed by the F1 polypeptide with better efficiency than that of the choline-containing cell walls and only threefold less than that shown by the native enzyme on this type of cell wall (Table 1). These results rule out the possibility that, in spite of the purification steps, the activity of the F1 fragment could be due to contamination by a residual undigested native enzyme. It is very important to point out that the activity shown by the FI and F7 fragments cannot be ascribed to an unspecific hydrolytic effect, since the soluble cell wall degradation products
generated by these fragments remained included on a Sephadex G-75 column (data not shown), indicating that both polypeptides were lysozymes. as has been previously demonstrated by determining the liberation of characteristic functional groups (Garcia et ai. 1987a; Tomasz, 1981). Moreover, the hen egg-white lysozyme (HEWL) used as control did not show detectable activity, even when assayed at high concentrations (Table 1). Figure 3 shows that the activity of the FI and F7 polypeptides was inhibited at low ionic strength as already observed for the native CPL7 lysozyme (Garcia et ai. 1990). This finding contrasts with the behaviour of the CPLI lysozyme that was only partially inhibited at concentrations higher than 100 mM NaCI. Since circular dichroism analyses indicated that the ionic strength does not induce a significant alteration of the secondary structure of the Fi polypeptide (Fig. 4). we concluded that the salt inhibition was probably due to a reduction in the substrate-binding capacity, a problem that is overcome by the presence of the specific choline-binding domain in the CPLI lysozyme. Although choline reduced the activity of the FI and F7 fragments, this inhibition was identical to that ascribed to an unspecific effect of the ionic strength (data not shown). This result is in agreement with the previous observation that choline specifically inhibits only the choline-dependent enzymes (Garcia etai. 1990). On the other hand, the FI and F7 enzymes showed an optimal pH for activity of 5, which Is similar to that found for the respective parental lysozymes.
Cloning and expression of the ^-terminal domains of the pneumococcal murein hydrolases To completely eliminate the possibility that an undetectable contamination of the FI and F7 polypeptides
Table 1. Pneumococcal ceil wall lyiic activity ot fragments FI and F7. Activity on Cell Walls*"" Enzyme
choline
ethanolamine
CH/EA ratio"
CPLI F1 CPL7 F7 HEWL"^
33x10^ 277 3.0x10^ 366 ND''
1936 647 2.2x10^ 1233 ND''
17210 0.45 1.3 0.30 -.
a. Assays were carried oul using pneumococcal cell walls containing choline or ethanolamine. Activity is expressed in c.p.m. solubilized per min and per mg of protein. Results are the averages of Iwo independent experiments. The standard error of the assays was less than 15% of the mean values. b. Ratio obtained between activities on cnoline and ethanolaminecontaining cell walls. c. HEWL was assayed at a final concentration of 0.2 mg ml"'. d. Activity not detected above the background. Non-enzymatic hydrolysis (background) was followed by measuring radioactivity released by the cell walls in the absence of enzyme, at 0 min (190119 c.p.m.) and after 24 h of incubation at 37 "C (177±23 cp.m.).
The cataiytic domain of pneumccoccal murein hydrolases
Studies on the structure and function of the A/-terminai domain of the pneumococcai murein hydrolases J. M. Sanz, E. Diaz and J. L. Garcia* Unidad de Genetica Bacteriana. Centro de Investigaciones Biologicas (CSIC), Velazquez 144. 28006 Madrid, Spain. Summary The structures of the choline-dependent pneumococcai murein hydrolases, LYTA amtdase and CPL1 Iysozyme, and the choline-independent CPL7 Iysozyme were analysed by controlled proteolytic digestions. The trypsin cleavage of the CPL1 and CPL7 iysozymes produced two resistant polypeptides, F1 and F7 respectively, corresponding to the Nterminal domain of the enzymes, whereas the amidase LYTA was completely hydrolysed by the protease. Interestingly, the F1 and F7 fragments showed a low, but significant, choline-independent Iysozyme activity. Choline reduced the rate of proteolytic hydrolysis of choline-dependent enzymes, suggesting that the C-terminal choline-binding domain adopts a more resistant conformation in the presence of the ligand. On the other hand, the regions encoding the /V-terminal domains of the three enzymes have been cloned and expressed in Escherichia coli, showing that these domains adopt an active conformation even in the absence of their C-terminal domains. The lower activity shown by the catalytic domains when compared with that of the complete enzymes suggests that the acquisition of a substratebinding domain represents a noticeable evolutionary advantage for enzymes that interact with polymeric substrates, allowing them to achieve a higher catalytic efficiency. These results strongly reinforce the hypothesis that the pneumococcai murein hydrolases have been originated by fusion of two structural and functional independent domains, and provide new experimental support to the theory of modular evolution of proteins.
Introduction The evidence that many genes (proteins) have evolved Received 16 October, 1991; revised and accepted 20 December, 1991. •For correspondence. Tel. (1) 5644570; Fax (1) 5627518.
by fusion of pre-existing sequences is increasing, in so far as more gene (protein) sequences are being determined (Gilbert, 1978). Modularization is the central concept of this particular evolutionary pathway which is known as the modular theory of evolution (DuBose and HartI, 1989). The length of the sequences (modules) that participate in the fusions varies according to the different levels of protein organization. Although the sequences encoding protein domains represent the clearest reference to settle the concept of module (Darnell and Doolittle. 1986; Gilbert, 1981),. those encoding short chains of amino acids exhibiting a specific secondary structure have also been included in this concept (DuBose and HartI, 1989; Ohno, 1987). In addition, according to the modular theory, it has been proposed that viral genomes have evolved by the fusion of different genes, or clusters of genes, which provide specific physiological functions (Botstein, 1980; Campbell, 1988). We have recently proposed that the system provided by the pneumococcai murein hydrolases constitutes an excellent experimental model for studying the postulates of the modular theory of evolution (Garcia et ai, 1990). The genes encoding five pneumococcai murein hydrolases have been cloned and sequenced (Garcia et ai, 1986; 1988; 1990; Romero etai. 1990a); two enzymes have been characterized as /V-acetylmuramoyl-L-alanine amidases (EC.3.5.1.28), i.e. the major pneumococcai autolysin (LYTA amidase) encoded by the lytA gene (Garcia e/a/., 1986), and the HBL amidase encoded by the hbl gene of the pneumococcai bacteriophage HB-3 (Romero etai, 1990a), The other three enzymes have been shown to be Iysozymes (EC 3.2.1.17), i.e. the CPL1, CPL7 and CPL9 Iysozymes, which are encoded by the genes cpli. cp/7and cp/9 of the pneumococcai bacteriophages Cp-1, Cp-7 and Cp-9, respectively (Garcia et ai, 1988; 1990). On the basis of sequence comparisons, we have postulated that these proteins have been built up by fusion of two independent domains (modules): the W-terminal catalytic domain and the C-terminal binding domain which recognizes specific components of the pneumococcai cell wall (Garcia etai. 1988; 1990). With the sole exception of the CPL7 Iysozyme, all of the pneumococcai murein hydrolases that have been biochemically characterized so far, including the second autolysin identified in pneumococcus, a glucosaminidase (Garcia et ai. 1989), as well as the amidase encoded by the pneumococcai
922 J. M. Sanz, E. Diaz andJ. L. Garcia
B
A kDa
A
B
C
D
E
F
G
H
I
92,0 66.0--— 43.0-WP 31 . 0 -
kDa
A
B
C
D
E
F
G
H
I
J
92.066.045,0-
— 31.0-
21. 5 -
m
14 . 3 -
•»
21.514.3-
Fig. 1. SDS-PAGE of the controlled digestions of CPL1 and CPL7 lysozymes by trypsin. The CPL1 (panel A) and CPL7 (panel B) purified enzymes were digestedbytrypsinat 37"C at a protease/protein ratio of 1/100 (w/w) and 1/25 [w/w). respectively, and the results were analysed on a 15% SDS-polyacrylamide gel. The molecular masses of the markers are expressed in ktlodaltons. A. Lane A, proiein standards: lanes B, C, D, E and F, samples from the digestion mixture cf the CPL1 lysozyme taken at 0, 5.10. 20 and 40 min, respectively; iane G. purified tragment F I ; lanes H and I. samples of a digestion carried out in the presence of 120 mM choline and taken at 10 and 40 min. B. Lane A, protein standards; lanes B,C, D.E, F, G, Hand I, sampies trom the digesticn mixture of the CPL7 lysozyme taken at 0, 2, 5, 10, 30. 60, 90 and 120 mm. respectively; lane J. purified fragment F7. Proteins were visualized with Coomassie Brilliant Blue.
phage Dp-1 (Garcia et ai. 1983), depend on the presence of choline in the teichoic acid of the cell wall for activity. The peculiar presence ot choline in the celi wall of pneumococcus has been suggested to act as an evolutionary element of selective pressure for the cholinedependent murein hydrolases, preserving the existence of the choline-binding domain {Garcia et ai. 1988). The choline-dependent enzymes are unable to degrade efficiently pneumococcal cell walls where choline has been replaced by the structural analogue ethanolamine (Giudicelli and Tomasz, 1984). This property is used as a rapid test to distinguish between choline-dependent and choline-independent enzymes. Ethanolamine-containing cell walls are only efficiently degraded by the CPL7 lysozyme, an enzyme that has a C-terminal domain which contains three identical repeated amino acid sequences, but presenting a primary structure completely different from that of the choline-dependent enzymes (Garcia ef ai, 1990). It is alsc interesting that simiiar amino acid repeats have also been found in the C-terminal region of a family of other ligand-binding proteins that appears to be of modular design (Wren, 1991). Although we have already presented experimental evidences supporting the modular organization of the pneumococcal murein hydrolases (Diaz et ai, 1990; 1991; Garcia etai, 1988; 1990), and even the sequence encoding the choline-binding domain had been expressed and studied (Sanchez-Puelles et ai. 1990), a direct demonstration that the A/-terminal domain of these enzymes might act as an independent and functional catalytic protein was missing. The results reported in this paper have allowed us to determine several structural and functional
properties of the A/-terminal modules of the LYTA amidase and the CPL1 and CPL7 lysozymes, demonstrating. by using two different experimental approaches, that these modules are folded as independent protein domains that can act as choline-independent pneumococcal cell wall hydrolases. This finding reinforces our hypothesis that the pneumoooccal murein hydrolases should have evolved by the fusion of pre-existing modules.
Results Controlled proteolytic digestions of pneumococcal murein hydrolases The two domain structure of the pneumococca! murein hydrolases. the LYTA amidase and the CPL1 and CPL7 lysozymes were analysed by controlled proteolytic digestions using different proteases. We found that whereas the LYTA amidase was completely digested by trypsin at a protease/protein ratio of 1/60 (w/w) after 60 min of incubation at 3 7 X . the hydrolysis of the CPLI and CPL7 lysozymes at a protease/protein ratio of 1/50 and 1/10 (w/w) respectively, produced, in both cases, after 120 min of incubation at 37"C, a unique resistant polypeptide of high molecular weight, named FI (M, =24000) (Fig. IA, lane G) and F7 (M, =25000) (Fig. 1B, lane J), respectively. The fragments, F1 and F7, were not further degraded even after three hours of incubation in the conditions mentioned. Moreover, these fragments could also be obtained using higher trypsin concentrations, i.e. at a
The catalytic domain of pneumococcal murein hydrolases
kDa
A
B
D
92.0 — 66.0 — 43.0 — 31 . 0 21,5 —
H
923
Fig. 2. SDS-PAGE of ttie controlled dtgesiions o( the CPLl lysozyme by ti-ctiymotrypsin The CPL1 tysozyme was digested by a-chymoifypsin al 37 C at a protease protetn ratio ot 1 30 (ww). and analysed on a 15% SDS-potyacrylamide gel The molecular masses of ttie marttefS are expressed m kilodartons. Lane A. protein standards: lanes B. C, D. E and F. samples taken at 0. 5. 15. 30 and 60 mm. respectively: lanes G, H. t and J. samples from a digestion camed out m the presence of 120 mM cholme and taken at 5.15. 30 and 60 min. respectively Proteins were visualized witti Coomassie Bnltiant Blue
14.3 —
protease/protein ratio of 1/1 (w/w), suggesting that these core potypeptides were highly resistant to trypsin digestion. The analysis ol the products obtained during the time course of the proteolysis indicated that the CPLl and CPL7 lysozymes were cleaved in different ways (Fig. 1). Whereas the appearance o( the Fl resistant fragment was not preceded by any other high-molecular-weight protein intermediates {Fig. 1A). cleavage of the CPL7 lysozyme resulted in the simultaneous appearance of three fragments, the F7 fragment being the smaller sized polypeptide (Fig 1 B). On the other hand, the digestion of the CPLl lysozyme with a-chymotrypsin at a protease/protein ratio of 1/30 (w/w) for 60 min at 37"C yielded a resistant fragment with a molecular weight similar to that observed for the F1 trypsin fragment (Fig. 2). This fragment was also remarkably resistant to (i-chymotrypsin, since a prolonged incubation in the above conditions or the use of a higher protease concentration {1/1. w/w) yielded identical results. However, the LYTA amidase and the CPL7 lysozyme were more susceptible to the cleavage with this protease than the CPLl lysozyme, since at a protease/protein ratio of 1/30 (wM). both enzymes were completely digested after 60 min of incubation at 37 C. The pneumococcal murein hydrolases were very susceptible to the treatment with proteinase K. since this protease, at a protease/protein concentration of 1/30 (w/w), completely hydrolysed the enzymes after 120 min of incubation at 37'C. However, using lower relative concentrations of proteinase K (1/100. w/w). we could obtain a resistant fragment by digesting the CPU lysozyme at 37''C for 60 min. Nevertheless, this fragment disappeared when the incubation was prolonged for 120 min (data not shown). Since it has been observed that choline induces a conformational change in the structure of the (^terminal domain of the choline-dependent enzymes (SanchezPuelles ef a/.. 1990). we decided to investigate rts influ-
ence on the protease sensitivity of the pneumococcal murein hydrolases. Figures 1 and 2 show that choline slowed the rate of hydrolysis of the CPLl lysozyme by trypsin and a-chymotrypsin. The protective effect of chotine cannot be ascribed to an increase in the ionic strength, since an equivalent concentration of NaCI did not produce any effect. Choline also slows the rate of proteolysis of the choline-dependent LYTA amidase whereas, as expected, the choline-independent CPL7 lysozyme was not protected by this compound (data not shown). Analysis of the Fl and F7 trypsin resistant fragments To determine which region of the CPLl and CPL7 lysozymes corresponded to the Fl and F7 trypsin-resistant fragments, these polypeptides were purified and their A/-terminal sequences were analysed by Edman degradation. The same amino acid sequence. Val-Lys-Lys-AsnAsp-Leu-Phe-Val. was obtained for both fragments. This sequence is identical to the /V-terminal sequence of the CPLl and CPL7 lysozymes (Garcia ef al.. 1988; 1990), indicating that both polypeptides correspond to the W-terminal half of the proteins. To establish the exact position of the trypsin cleavage site we attempted the analysis of the 0-terminat amino acids of both fragments using carboxypeptidase Y, but unfortunately they were resistant to the carboxypeptidase hydrolysis, even in the presence of 6 M urea. According to the previously published amino ackj sequence of the CPLl lysozyme. this enzyme contains three cysteine residues (Garcia et al.. 1988). Taking into account the molecular weight of the Fl fragment, estimated by sodium dodecyl sulphate/polyacrylamide gel electrophoresis (SDS-PAGE). one of the cysteine residues {Cys-239) should be located near the putative trypsin cleavage site. Therefore the analysis of the cysteine content of the Fl fragment with 5,5'-dithio-bis{2nitrobenzoic) acid (DTNB) could allow us to determine
924
J. M. Sanz. E. Diaz and J. L Garcia
lOOi
[-•
\
75 50-
o
I
\ \
\
\
25 0
1
•.
\
(
-a.
:
• • : • ,
h- —=^-^?—
50
fl—
100 150 [NaCi] (mM)
_
j
1
1
200
250
Rg. 3. Influence ol ionic strength on the activity of the CPLI, CPL7, FI and F7 proteins. The CPLI (O—O). CPL7 (O O), FI ( • - - • ) and F7 (A---A) purified proteins were assayed in 20 mM phosphate buffer, pH 6.0, upon hydrolysis of the choline-containmg cell walls in the presence of different concentrations of NaCI. Results are expressed in terms of percentage activity relative to the absence ot NaCI in the assay mixture.
the location of the cleavage site more precisely. We found that the CPLI lysozyme contains three free cysteines (3.0 cysteines/molecule), whereas the FI fragment has only two (1.7 cysteines/molecule). assuming a M^ of 24 000 for this fragment (Fig. 1A, lane G). These results indicated that the cysteines of the CPLI lysozyme did not form a disulphide bridge, but also suggested that the trypsin cleavage site should be located before Cys-239.
Assay of the murein hydrolase activity of the F1 and F7 trypsin-resistant fragments According to the proposed modular organization of the pneumococcal murein hydrolases, the F1 and F7 fragments should correspond to the postulated catalytic module of the CPLI and CPL7 lysozymes (Garcia et ai, 1988; 1990) and therefore it seemed likely that the F1 and F7 fragments could still retain a detectable enzymatic activity. The results shown in Table 1 demonstrated that when the hydrolytic specific activity of the purified F1 and F7 polypeptides was assayed on pneumococcal cell walls containing choline or ethanolamine, these fragments turned out still to be active, although their specific activities were lower than those of the native lysozymes. Interestingly, the ethanolamine-containing cell walls were hydrolysed by the F1 polypeptide with better efficiency than that of the choline-containing cell walls and only threefold less than that shown by the native enzyme on this type of cell wall (Table 1). These results rule out the possibility that, in spite of the purification steps, the activity of the F1 fragment could be due to contamination by a residual undigested native enzyme. It is very important to point out that the activity shown by the FI and F7 fragments cannot be ascribed to an unspecific hydrolytic effect, since the soluble cell wall degradation products
generated by these fragments remained included on a Sephadex G-75 column (data not shown), indicating that both polypeptides were lysozymes. as has been previously demonstrated by determining the liberation of characteristic functional groups (Garcia et ai. 1987a; Tomasz, 1981). Moreover, the hen egg-white lysozyme (HEWL) used as control did not show detectable activity, even when assayed at high concentrations (Table 1). Figure 3 shows that the activity of the FI and F7 polypeptides was inhibited at low ionic strength as already observed for the native CPL7 lysozyme (Garcia et ai. 1990). This finding contrasts with the behaviour of the CPLI lysozyme that was only partially inhibited at concentrations higher than 100 mM NaCI. Since circular dichroism analyses indicated that the ionic strength does not induce a significant alteration of the secondary structure of the Fi polypeptide (Fig. 4). we concluded that the salt inhibition was probably due to a reduction in the substrate-binding capacity, a problem that is overcome by the presence of the specific choline-binding domain in the CPLI lysozyme. Although choline reduced the activity of the FI and F7 fragments, this inhibition was identical to that ascribed to an unspecific effect of the ionic strength (data not shown). This result is in agreement with the previous observation that choline specifically inhibits only the choline-dependent enzymes (Garcia etai. 1990). On the other hand, the FI and F7 enzymes showed an optimal pH for activity of 5, which Is similar to that found for the respective parental lysozymes.
Cloning and expression of the ^-terminal domains of the pneumococcal murein hydrolases To completely eliminate the possibility that an undetectable contamination of the FI and F7 polypeptides
Table 1. Pneumococcal ceil wall lyiic activity ot fragments FI and F7. Activity on Cell Walls*"" Enzyme
choline
ethanolamine
CH/EA ratio"
CPLI F1 CPL7 F7 HEWL"^
33x10^ 277 3.0x10^ 366 ND''
1936 647 2.2x10^ 1233 ND''
17210 0.45 1.3 0.30 -.
a. Assays were carried oul using pneumococcal cell walls containing choline or ethanolamine. Activity is expressed in c.p.m. solubilized per min and per mg of protein. Results are the averages of Iwo independent experiments. The standard error of the assays was less than 15% of the mean values. b. Ratio obtained between activities on cnoline and ethanolaminecontaining cell walls. c. HEWL was assayed at a final concentration of 0.2 mg ml"'. d. Activity not detected above the background. Non-enzymatic hydrolysis (background) was followed by measuring radioactivity released by the cell walls in the absence of enzyme, at 0 min (190119 c.p.m.) and after 24 h of incubation at 37 "C (177±23 cp.m.).
The cataiytic domain of pneumccoccal murein hydrolases
