Molecular Microbiology (1992) 6(9), 1243-1252
The binding of Cellulomonas f/m/endoglucanase C (CenC) to cellulose and Sephadex is mediated by the A/-terminal repeats J. B. Coutinho, N. R. Gilkes,* R. A. J. Warren, D. G. Kilburn and R. C. Milter, Jr Department of Microbiology, University of British Coiumbia, 300-6174 University Bouievard, Vancouver, British Columbia, Canada V6T 1Z3. Summary Endoglucanase C (CenC) from Celluiomonas fimi binds to cellulose and to Sephadex. The enzyme has two contiguous 150-amino-acid repeats (N1 and N2) at its N-terminus and two unrelated contiguous 100amino-acjd repeats {C1 and C2) at its C-terminus. Polypeptides corresponding to N1, N1N2, C1, and C1C2 were produced by expression of appropriate cenC gene fragments in Escherichia coli. N1N2, but not N1 alone, binds to Sephadex; both polypeptides bind to Avicel, (a heterogeneous cellulose preparation containing both crystalline and non-crystalline components). Neither C1 nor C1C2 binds to Avicel or Sephadex. N1N2 and N1 bind to regenerated (amorphous') cellulose but not to bacterial crystalline cellulose; the cellulose-binding domain of C. fimi exoglucanase Cex binds to both of these forms of cellulose. Amino acid sequence comparison reveais that N1 and N2 are distantiy reiated to the cellulosebinding domains of Cex and C. fimi endoglucanases A and B.
Introduction The amino acid sequences of more than 60 [i-1,4-g!ucanases have been deduced from the nucleofide sequences of their genes. Characterization of the enzymes or analysis of their amino acid sequences has shown that many of them comprise two or more functional domains (Beguin, 1990; Gilkes ef a/,, 1991). In some (perhaps ail) such enzymes, the domains function independently and retain their functions when separated by proteolysis (Calza et ai, 1985; Van Tilbeurgh et ai, 1986; Langsford etai. 1987; Tomme etai, 1988; Gilkes etai. Received 26 August, 1991; revised 26 December, 1991; accepted 14 January, 1992. 'For correspondence. Tel. (604) 822 6845; Fax (604) 822 6041.
1988; Ghangas and Witson, 1988; Stahlberg etai. 1988; McGavin and Forsberg, 1989). A frequent arrangement is a catalytic domain linked to a cellulose-binding domain (CBD) by a linker sequence rich in proline and/or hydroxyamino acids (Beguin, 1990; Gilkes et ai. 1991), The CBDs are at the N- or C-termini of different enzymes. The molecular mass of this type of (i-1,4-glucanase is in the range 30-70kDa (Beguin, 1990; Gilkes et ai, 1991). Some (i-1,4-glucanases are larger than this and may contain other domains besides a catalytic domain and a CBD. For example, repeated sequences 20-150 amino acids long occur in several (i-1,4-glucanases of widely different molecular masses (Beguin, 1990; Gilkes et ai., 1991). The functions of the repeated sequences are mostly unknown but in Ctostridium stercorarium GelZ they mediate binding to cellulose (Jauris et ai, 1990) and the duplicated regions carried by Clostridium thermocettum cellulases appear to anchor these enzymes to the ceilulosome (Tokatiidis etai, 1991). Endogiucanase C (CenC) from Cellutomonas fimi is one of the largest p-1,4-glucanases charactenzed to date. The mature polypeptide is 1069 amino acids long, with residues 300-880 forming the catalytic domain (Coutinho et at., 1991). It is rare amongst p-1,4-glucanases for fhe catalytic domain to be flanked at both ends by extended sequences of amino acids (Beguin, 1990; Gilkes et at.. 1991), CenC has a 150-amino-acid tandem repeat at its A/-terminus and an unrelated 100-amino-acid repeat at its C-terminus (Coutinho etai. 1991). CenC is also unusual in that it binds to Sephadex as well as to cellulose. It was purified from C. ffm/culture supernatants by virtue of its Sephadex-binding capacity (Moser et at.. 1989), and may be related to enzyme CI, a 118 kDa endoglucanase purified from the culture supernatant of Cettutomonas sp- llbc by Sephadex affinity chromatography (Beguin and Eisen. 1978), In C. fimi and Escherichia coti, a truncated form of the enzyme, CenC. is produced by proteolysis of CenC (Moser et ai,, 1989), probably by removal of the C-terminal 100-amino-acid repeat. CenC also binds to Sephadex and to cellulose (Coutinho etat., 1991). Identification of the region(s) in CenC which mediate binding to cellulose and/or Sephadex is most easily accomplished by examining the isolated domains. Proteolytic separation of domains has been used to dissect
1244 J. a Coutinhoe\3\. Sad
Apal
Apal*
Mlul
CenC (pTZ-JC2) 1
148
1
148
1
150
300
886 975
1069
N1N2(pTZ-JG3) 299
Fig. 1, Schematic representation of CenC and poiypeptides containing the W- and C-terminal repeats. The plasmid encoding CenC or the CenC polypeptide is shown in parenthesis, The locations o< the restriction sites in c&v.C which were used to construct the vanous plasmids are indicated. Slippled regions represent N terminal repeats; cross hatched regions represent Cterminal repeats. Numbers refer to amino acid residues, starting from the /V terminus of mature CenC (Coutinhoela/., 1991).
N1 (pTZ-JC6)
C1C2(pTZJC7) 860
975
860
993
1069
C l (PTZ-JC14)
the Structural and functional organization of endoglucanase CenA (44 kDa) and exoglucanase Cex (47 kDa) from C. ftm/(Gilkesefa/., 1988; 1989) and of other fi-1,4glucanases (Tomme et al.. 1988: Ghangas and Wilson, 1988). However, for large, mulfidomain poiypeptides such as CenC, proteolytic separation o1 domains is difficult. It is easier to dissect the gene and to independently express the sequences encoding each domain. This paper describes the independent production of the repeated regions in CenC and analysis of their binding to bacterial crystalline cellulose and to regenerated ('amorphous') cellulose. Only the /V-terminal repeats bind to cellulose or Sephadex, and this binding is differenf from that of the CBD of Cex, an exoglucanase from C. fimi. There is no detectable binding of the C-terminal repeats.
molecular-mass band seen in cell extracts of E, coli JM101 pTZ-JC3 was also seen in the E. co//JM101 pTZ18R control cell extract and was considered to represent the non-specific interaction of the antiserum with an E. co//protein. The apparent molecular masses of the N1 and N1N2 (20.0 and 40.0 kDa, respectively) were significantly greater than those predicted from their corresponding DNA sequences (15.4 and 30.4, respectively). The predicted pi values of N1 and N1N2 (3,54 and 3.48, respectively) showed that both are highly acidic poiypeptides and this may account for the discrepancies between their predicted and apparent molecular masses, as reported
JC3 Results
ce
JC6 cs
ce
cs
I8R ce
Expression of fragments of the cenC gene The A/-terminai repeats of CenC were designated N1 (amino acids 1-148) and N2 (amino acids 149-296); the C-terminal repeats were designated Cl (amino acids 886-975) and C2 (amino acids 976-1069). Deletion mutants of the cenC gene were constructed which encoded the leader peptide plus either N1 (amino acids 1-150), N1N2 (amino acids 1-299), Cl (amino acids 860-993) or C1C2 (amino adds 860-1069) (Fig, 1), Poiypeptides corresponding to N1 and N1N2 were detected on Western blots of fractions from cultures of E. coli JM101 carrying plasmids pTZ-JC6 and pTZ-JC3, respectively, using anti-CenC serum (Fig. 2). The low
Fig. 2. Intraceliular and extracelluiar distribution of poiypeptides N1N2 and N1 synthesized m E co//JM10i. Poiypeptides from cell extracts (ce) and culture supernatants (cs) were anaiysed by SDS-PAGE foliowed by Western blotting and detection with polycional anti CenC serum. Plasmids pTZ-JC3 and pTZ-JC6 encode N1N2 and N1, respectively; pTZ18R is the control plasmid, without cenC DNA.
Cellulomonas fimi endoglucanase C for other acidic proteins (Kaufman ef ai, 1984). Similar considerations apply to intact CenC (Coutinho er ai, 1991). Polypeptides Cl and C1C2 reacted poorly with anti-CenC serum (data not shown); however, they were detected by sodium dodecyl sulphate/polyacrylamide gel electrophoresis (SDS-PAGE) in appropriate E. coti cell extracts as unique bands following induction with IPTG (Fig. 3). The apparent molecular masses of C1 and Cl C2 (16.0 and 28.0 kDa, respectivety) agreed well with those predicted from their corresponding DNA sequences (16.2 and 26.4 kDa, respectively); their predicted pl values are 5,07 and 8.76, respectively. W-terminal sequence analyses revealed that N l . N1N2, C1, and C1C2 all had the same W-terminus (ASPIG), as expected (Fig. 1). Each of the potypeptides was exported by E. coli, with a varying proportion appearing in the culture supernatant (Figs 2 and 3). Multidomain (J-1,4-glucanases may be sensitive to proteolysis between their domains (Gilkes et ai. 1988;Tomme etai, 1988). However, E co//cells producing N1N2 did not contain a polypeptide of the size of N l . and cells producing C1C2 did not contain one of the size of C l . All four polypeptides appeared to be stable in £. coli This is in contrast to the production of CenC in E. coli, where proteolysis gives rise to CenC by removing a polypeptide of about the size of C2 from the C-terminus (Coutinho era/., 1991).
JC7 cs
kDa
1245
JC14 ce
cs
ce
116 66
36 29
Fig, 3. Intracellular and extracellular distribution ot polypeptidesClC2 and Cl synthesized in E. co/(JM10i. Polypeptides trom cell extracts (ce) and culture supernatants (cs) were analysed by SDS-PAGE, 6 h after induction with 0.1 mM IPTG. The left lane ol each pair corresponds to induction when the Af,oo of the culture reached 0 4; the right lane corresponds to induction when the A^no reached l 00 PolypeptidesCtC2 and C l . indicated by arrows, are encoded by pT2-JC7 and pTZ-JC14, respectively. Molecular mass markers are shown in the far left lane.
18R kDa
ce
JC3 cs
JC6
pp
cs
pp
Cultures expressing N1N2. N l , C1C2, and C1 were devoid of detectable carboxymethylcellulase activity (data not shown). Attempts to express the sequence encoding the leader polypeptide fused to the catalytic domain of CenC (amino acids 300-886 of CenC) were unsuccessful (data not shown).
Binding of the N - andC-terminal repeats to cetlulose and to Sephadex Polypeptides N1N2 (data not shown) and N1 (Fig. 4) both bound to Avicel PH101 (a heterogenous microcrystalline cellulose preparation containing both crystalline and noncrystalline regions; see the Discussion), but only N1N2 bound to Sephadex G-50 (Fig. 4). Neither C1C2 nor Cl bound to Avicel PH101 or to Sephadex G-50 (data not shown). Nl N2 and Nl were purified to virtual homogeneity by affinity chromatography on Avicel PHI01 or Sephadex G-50, respectively (Fig. 4B). The binding specificities of Nl and N1N2 for two different cellulose allomorphs were investigated by determining their adsorption to regenerated cellulose and to crystalline bacterial cellulose. The adsorption of CBDce, was also included for comparison. Adsorption of the three polypeptides to the two forms of cellulose is described by the isotherms ([B] versus [F]), shown in Fig. 5. The partition coefficient (the initial slope of the adsorption isotherm) was determined as a measure of the relative
kDa
JC6
JC3
97 66
B Fig. 4. Purification of N1N2 and Nt by affinity chromatography. A. Polypeptides trom E. coliM^O^ cell extracts (ce) and culture supernatants (cs). and the purified polypeptides (pp) analysed by SDS-PAGE. Plasmids pTZ-JC3 and pTZ-JC6 encode Nt N2 and N l , respectively, pTZ-18Ft is the control plasmid, without cenC DNA. N1N2 and Nl were puritied by aftinity chromatography on Sephadex G-50 and Avicel PH101. respectively, as described in the Experimental procedures. B. SDS-PAGE analysis ot 10, 20 and 40 M9 ot N1 purified from E coli JM101(pTZ-JC6),and 10 and 40 |.ig of N1N2 purified from E. coli JMtO1(pTZ-JC3). to evaluate levels of contamination with other polypeptides. The far left lane of each panel shows the molecular weight markers.
1246
J. B. Coutinho eta\.
20
40
60
80
Fig. 5. Adsorption isotherms for Nl N2. Nl and CBDce.- Adsorption of the various polypeptides to regenerated oel)u)ose or bacterial crystalline cellulose was determined as described in the Experimentai procedures. [F\ is the tree polypeptide concentralion l\iM) and [B] is the bound polypeptide concentraVion (\imo\ per g cellulose) at equilibrium.
CRYSTALLINE CELLULOSE
REGENERATED CELLULOSE
100
20
40
60
80
100
O zt
100
0
50
100
200
100 150 200 2S0 300
20
40
affinities of NT, N1N2, and CBDc^, for a particutar cettutose atlomorph (Table 1). Alt three polypeptides adsorbed to regenerated celtutose (Fig. 5). The partition coefficients for Nl and N1N2 were approximately threefold lower than the corresponding vatue for CBDce, (Tabte 1). Saturation of regenerated cettutose was not obtained with N l , N1N2 or CBDce, at the highest polypeptide concentration tested (Pig. 5). tn contrast to the adsorption of Nl and N1N2 to regenerated cellulose, there was no apparent adsorption of either polypeptide to bacterial crystalline cellulose; however, CBDce, adsorbed welt to this substrate (Pig. 5, Table 1). When adsorption data were plotted in the form [B]/[F1 versus [B] (Scatchard plot), non-linear isotherms were obtained in all cases (data not shown).
150
60
80
200
100
becoming clear that there are also families of CBDs (Gitkes et ai, 1991). At present, however, some of the famities contain onty one or two members. The CBDs within a family are quite uniform in size, but there are considerabte differences between the famities (Tabie 2). Since the catatytic-domain families are designated by tetters, a different designation should be used for the CBD famities; roman numerats are used in this paper. Nl and N2 of CenC, and the Trichoderma reese/Egltt sequence, appear to comprise a subgroup oi Family II (Pig. 6). Atthough there is tittte overatt sequence identity
Table 1. Partition coefficients for the adsorption of N1N2. N l . and CBDce. '0 regenerated cellulose and bacterial crystalline cellulose. Partition Coefficient (I g"
Sequence relatedness between N1, N2 and other CBDs Beta-1,4-gtucanases can be grouped into famities according to amino acid sequence relatedness within their catalytic domains (tHenrissat et ai. 1989; Beguin, 1990 Gitkes et ai. 1991). As more amino acid sequences are deduced from the nucieotide sequences of genes, it is
Polypeptide
regenerated cellulose
bactenal cellulose
N1N2
0.20 0,24 0.60
The binding of Cellulomonas f/m/endoglucanase C (CenC) to cellulose and Sephadex is mediated by the A/-terminal repeats J. B. Coutinho, N. R. Gilkes,* R. A. J. Warren, D. G. Kilburn and R. C. Milter, Jr Department of Microbiology, University of British Coiumbia, 300-6174 University Bouievard, Vancouver, British Columbia, Canada V6T 1Z3. Summary Endoglucanase C (CenC) from Celluiomonas fimi binds to cellulose and to Sephadex. The enzyme has two contiguous 150-amino-acid repeats (N1 and N2) at its N-terminus and two unrelated contiguous 100amino-acjd repeats {C1 and C2) at its C-terminus. Polypeptides corresponding to N1, N1N2, C1, and C1C2 were produced by expression of appropriate cenC gene fragments in Escherichia coli. N1N2, but not N1 alone, binds to Sephadex; both polypeptides bind to Avicel, (a heterogeneous cellulose preparation containing both crystalline and non-crystalline components). Neither C1 nor C1C2 binds to Avicel or Sephadex. N1N2 and N1 bind to regenerated (amorphous') cellulose but not to bacterial crystalline cellulose; the cellulose-binding domain of C. fimi exoglucanase Cex binds to both of these forms of cellulose. Amino acid sequence comparison reveais that N1 and N2 are distantiy reiated to the cellulosebinding domains of Cex and C. fimi endoglucanases A and B.
Introduction The amino acid sequences of more than 60 [i-1,4-g!ucanases have been deduced from the nucleofide sequences of their genes. Characterization of the enzymes or analysis of their amino acid sequences has shown that many of them comprise two or more functional domains (Beguin, 1990; Gilkes ef a/,, 1991). In some (perhaps ail) such enzymes, the domains function independently and retain their functions when separated by proteolysis (Calza et ai, 1985; Van Tilbeurgh et ai, 1986; Langsford etai. 1987; Tomme etai, 1988; Gilkes etai. Received 26 August, 1991; revised 26 December, 1991; accepted 14 January, 1992. 'For correspondence. Tel. (604) 822 6845; Fax (604) 822 6041.
1988; Ghangas and Witson, 1988; Stahlberg etai. 1988; McGavin and Forsberg, 1989). A frequent arrangement is a catalytic domain linked to a cellulose-binding domain (CBD) by a linker sequence rich in proline and/or hydroxyamino acids (Beguin, 1990; Gilkes et ai. 1991), The CBDs are at the N- or C-termini of different enzymes. The molecular mass of this type of (i-1,4-glucanase is in the range 30-70kDa (Beguin, 1990; Gilkes et ai, 1991). Some (i-1,4-glucanases are larger than this and may contain other domains besides a catalytic domain and a CBD. For example, repeated sequences 20-150 amino acids long occur in several (i-1,4-glucanases of widely different molecular masses (Beguin, 1990; Gilkes et ai., 1991). The functions of the repeated sequences are mostly unknown but in Ctostridium stercorarium GelZ they mediate binding to cellulose (Jauris et ai, 1990) and the duplicated regions carried by Clostridium thermocettum cellulases appear to anchor these enzymes to the ceilulosome (Tokatiidis etai, 1991). Endogiucanase C (CenC) from Cellutomonas fimi is one of the largest p-1,4-glucanases charactenzed to date. The mature polypeptide is 1069 amino acids long, with residues 300-880 forming the catalytic domain (Coutinho et at., 1991). It is rare amongst p-1,4-glucanases for fhe catalytic domain to be flanked at both ends by extended sequences of amino acids (Beguin, 1990; Gilkes et at.. 1991), CenC has a 150-amino-acid tandem repeat at its A/-terminus and an unrelated 100-amino-acid repeat at its C-terminus (Coutinho etai. 1991). CenC is also unusual in that it binds to Sephadex as well as to cellulose. It was purified from C. ffm/culture supernatants by virtue of its Sephadex-binding capacity (Moser et at.. 1989), and may be related to enzyme CI, a 118 kDa endoglucanase purified from the culture supernatant of Cettutomonas sp- llbc by Sephadex affinity chromatography (Beguin and Eisen. 1978), In C. fimi and Escherichia coti, a truncated form of the enzyme, CenC. is produced by proteolysis of CenC (Moser et ai,, 1989), probably by removal of the C-terminal 100-amino-acid repeat. CenC also binds to Sephadex and to cellulose (Coutinho etat., 1991). Identification of the region(s) in CenC which mediate binding to cellulose and/or Sephadex is most easily accomplished by examining the isolated domains. Proteolytic separation of domains has been used to dissect
1244 J. a Coutinhoe\3\. Sad
Apal
Apal*
Mlul
CenC (pTZ-JC2) 1
148
1
148
1
150
300
886 975
1069
N1N2(pTZ-JG3) 299
Fig. 1, Schematic representation of CenC and poiypeptides containing the W- and C-terminal repeats. The plasmid encoding CenC or the CenC polypeptide is shown in parenthesis, The locations o< the restriction sites in c&v.C which were used to construct the vanous plasmids are indicated. Slippled regions represent N terminal repeats; cross hatched regions represent Cterminal repeats. Numbers refer to amino acid residues, starting from the /V terminus of mature CenC (Coutinhoela/., 1991).
N1 (pTZ-JC6)
C1C2(pTZJC7) 860
975
860
993
1069
C l (PTZ-JC14)
the Structural and functional organization of endoglucanase CenA (44 kDa) and exoglucanase Cex (47 kDa) from C. ftm/(Gilkesefa/., 1988; 1989) and of other fi-1,4glucanases (Tomme et al.. 1988: Ghangas and Wilson, 1988). However, for large, mulfidomain poiypeptides such as CenC, proteolytic separation o1 domains is difficult. It is easier to dissect the gene and to independently express the sequences encoding each domain. This paper describes the independent production of the repeated regions in CenC and analysis of their binding to bacterial crystalline cellulose and to regenerated ('amorphous') cellulose. Only the /V-terminal repeats bind to cellulose or Sephadex, and this binding is differenf from that of the CBD of Cex, an exoglucanase from C. fimi. There is no detectable binding of the C-terminal repeats.
molecular-mass band seen in cell extracts of E, coli JM101 pTZ-JC3 was also seen in the E. co//JM101 pTZ18R control cell extract and was considered to represent the non-specific interaction of the antiserum with an E. co//protein. The apparent molecular masses of the N1 and N1N2 (20.0 and 40.0 kDa, respectively) were significantly greater than those predicted from their corresponding DNA sequences (15.4 and 30.4, respectively). The predicted pi values of N1 and N1N2 (3,54 and 3.48, respectively) showed that both are highly acidic poiypeptides and this may account for the discrepancies between their predicted and apparent molecular masses, as reported
JC3 Results
ce
JC6 cs
ce
cs
I8R ce
Expression of fragments of the cenC gene The A/-terminai repeats of CenC were designated N1 (amino acids 1-148) and N2 (amino acids 149-296); the C-terminal repeats were designated Cl (amino acids 886-975) and C2 (amino acids 976-1069). Deletion mutants of the cenC gene were constructed which encoded the leader peptide plus either N1 (amino acids 1-150), N1N2 (amino acids 1-299), Cl (amino acids 860-993) or C1C2 (amino adds 860-1069) (Fig, 1), Poiypeptides corresponding to N1 and N1N2 were detected on Western blots of fractions from cultures of E. coli JM101 carrying plasmids pTZ-JC6 and pTZ-JC3, respectively, using anti-CenC serum (Fig. 2). The low
Fig. 2. Intraceliular and extracelluiar distribution of poiypeptides N1N2 and N1 synthesized m E co//JM10i. Poiypeptides from cell extracts (ce) and culture supernatants (cs) were anaiysed by SDS-PAGE foliowed by Western blotting and detection with polycional anti CenC serum. Plasmids pTZ-JC3 and pTZ-JC6 encode N1N2 and N1, respectively; pTZ18R is the control plasmid, without cenC DNA.
Cellulomonas fimi endoglucanase C for other acidic proteins (Kaufman ef ai, 1984). Similar considerations apply to intact CenC (Coutinho er ai, 1991). Polypeptides Cl and C1C2 reacted poorly with anti-CenC serum (data not shown); however, they were detected by sodium dodecyl sulphate/polyacrylamide gel electrophoresis (SDS-PAGE) in appropriate E. coti cell extracts as unique bands following induction with IPTG (Fig. 3). The apparent molecular masses of C1 and Cl C2 (16.0 and 28.0 kDa, respectivety) agreed well with those predicted from their corresponding DNA sequences (16.2 and 26.4 kDa, respectively); their predicted pl values are 5,07 and 8.76, respectively. W-terminal sequence analyses revealed that N l . N1N2, C1, and C1C2 all had the same W-terminus (ASPIG), as expected (Fig. 1). Each of the potypeptides was exported by E. coli, with a varying proportion appearing in the culture supernatant (Figs 2 and 3). Multidomain (J-1,4-glucanases may be sensitive to proteolysis between their domains (Gilkes et ai. 1988;Tomme etai, 1988). However, E co//cells producing N1N2 did not contain a polypeptide of the size of N l . and cells producing C1C2 did not contain one of the size of C l . All four polypeptides appeared to be stable in £. coli This is in contrast to the production of CenC in E. coli, where proteolysis gives rise to CenC by removing a polypeptide of about the size of C2 from the C-terminus (Coutinho era/., 1991).
JC7 cs
kDa
1245
JC14 ce
cs
ce
116 66
36 29
Fig, 3. Intracellular and extracellular distribution ot polypeptidesClC2 and Cl synthesized in E. co/(JM10i. Polypeptides trom cell extracts (ce) and culture supernatants (cs) were analysed by SDS-PAGE, 6 h after induction with 0.1 mM IPTG. The left lane ol each pair corresponds to induction when the Af,oo of the culture reached 0 4; the right lane corresponds to induction when the A^no reached l 00 PolypeptidesCtC2 and C l . indicated by arrows, are encoded by pT2-JC7 and pTZ-JC14, respectively. Molecular mass markers are shown in the far left lane.
18R kDa
ce
JC3 cs
JC6
pp
cs
pp
Cultures expressing N1N2. N l , C1C2, and C1 were devoid of detectable carboxymethylcellulase activity (data not shown). Attempts to express the sequence encoding the leader polypeptide fused to the catalytic domain of CenC (amino acids 300-886 of CenC) were unsuccessful (data not shown).
Binding of the N - andC-terminal repeats to cetlulose and to Sephadex Polypeptides N1N2 (data not shown) and N1 (Fig. 4) both bound to Avicel PH101 (a heterogenous microcrystalline cellulose preparation containing both crystalline and noncrystalline regions; see the Discussion), but only N1N2 bound to Sephadex G-50 (Fig. 4). Neither C1C2 nor Cl bound to Avicel PH101 or to Sephadex G-50 (data not shown). Nl N2 and Nl were purified to virtual homogeneity by affinity chromatography on Avicel PHI01 or Sephadex G-50, respectively (Fig. 4B). The binding specificities of Nl and N1N2 for two different cellulose allomorphs were investigated by determining their adsorption to regenerated cellulose and to crystalline bacterial cellulose. The adsorption of CBDce, was also included for comparison. Adsorption of the three polypeptides to the two forms of cellulose is described by the isotherms ([B] versus [F]), shown in Fig. 5. The partition coefficient (the initial slope of the adsorption isotherm) was determined as a measure of the relative
kDa
JC6
JC3
97 66
B Fig. 4. Purification of N1N2 and Nt by affinity chromatography. A. Polypeptides trom E. coliM^O^ cell extracts (ce) and culture supernatants (cs). and the purified polypeptides (pp) analysed by SDS-PAGE. Plasmids pTZ-JC3 and pTZ-JC6 encode Nt N2 and N l , respectively, pTZ-18Ft is the control plasmid, without cenC DNA. N1N2 and Nl were puritied by aftinity chromatography on Sephadex G-50 and Avicel PH101. respectively, as described in the Experimental procedures. B. SDS-PAGE analysis ot 10, 20 and 40 M9 ot N1 purified from E coli JM101(pTZ-JC6),and 10 and 40 |.ig of N1N2 purified from E. coli JMtO1(pTZ-JC3). to evaluate levels of contamination with other polypeptides. The far left lane of each panel shows the molecular weight markers.
1246
J. B. Coutinho eta\.
20
40
60
80
Fig. 5. Adsorption isotherms for Nl N2. Nl and CBDce.- Adsorption of the various polypeptides to regenerated oel)u)ose or bacterial crystalline cellulose was determined as described in the Experimentai procedures. [F\ is the tree polypeptide concentralion l\iM) and [B] is the bound polypeptide concentraVion (\imo\ per g cellulose) at equilibrium.
CRYSTALLINE CELLULOSE
REGENERATED CELLULOSE
100
20
40
60
80
100
O zt
100
0
50
100
200
100 150 200 2S0 300
20
40
affinities of NT, N1N2, and CBDc^, for a particutar cettutose atlomorph (Table 1). Alt three polypeptides adsorbed to regenerated celtutose (Fig. 5). The partition coefficients for Nl and N1N2 were approximately threefold lower than the corresponding vatue for CBDce, (Tabte 1). Saturation of regenerated cettutose was not obtained with N l , N1N2 or CBDce, at the highest polypeptide concentration tested (Pig. 5). tn contrast to the adsorption of Nl and N1N2 to regenerated cellulose, there was no apparent adsorption of either polypeptide to bacterial crystalline cellulose; however, CBDce, adsorbed welt to this substrate (Pig. 5, Table 1). When adsorption data were plotted in the form [B]/[F1 versus [B] (Scatchard plot), non-linear isotherms were obtained in all cases (data not shown).
150
60
80
200
100
becoming clear that there are also families of CBDs (Gitkes et ai, 1991). At present, however, some of the famities contain onty one or two members. The CBDs within a family are quite uniform in size, but there are considerabte differences between the famities (Tabie 2). Since the catatytic-domain families are designated by tetters, a different designation should be used for the CBD famities; roman numerats are used in this paper. Nl and N2 of CenC, and the Trichoderma reese/Egltt sequence, appear to comprise a subgroup oi Family II (Pig. 6). Atthough there is tittte overatt sequence identity
Table 1. Partition coefficients for the adsorption of N1N2. N l . and CBDce. '0 regenerated cellulose and bacterial crystalline cellulose. Partition Coefficient (I g"
Sequence relatedness between N1, N2 and other CBDs Beta-1,4-gtucanases can be grouped into famities according to amino acid sequence relatedness within their catalytic domains (tHenrissat et ai. 1989; Beguin, 1990 Gitkes et ai. 1991). As more amino acid sequences are deduced from the nucieotide sequences of genes, it is
Polypeptide
regenerated cellulose
bactenal cellulose
N1N2
0.20 0,24 0.60
