doi:10.1016/S0022-2836(03)00881-7
J. Mol. Biol. (2003) 333, 817–829
Engineering the Exo-loop of Trichoderma reesei Cellobiohydrolase, Cel7A. A comparison with Phanerochaete chrysosporium Cel7D Ingemar von Ossowski1, Jerry Sta˚hlberg2, Anu Koivula1 Kathleen Piens3, Dieter Becker4, Harry Boer1, Raija Harle1, Mark Harris5 Christina Divne5, Sabah Mahdi2, Yongxin Zhao6, Hugues Driguez7 Marc Claeyssens3, Michael L. Sinnott4 and Tuula T. Teeri1* 1 VTT Biotechnology, P.O. Box 1500, FIN-02044 VTT Espoo Finland 2
Department of Molecular Biology, Swedish University of Agricultural Sciences, Box 590 SE-75124 Uppsala, Sweden 3 Laboratory of Biochemistry University of Ghent, K.L. Ledeganckstraat 35, B-9000 Ghent, Belgium 4
School of Applied Sciences University of Huddersfield Huddersfield HD1 3DH, UK 5
Department of Cell and Molecular Biology, Uppsala University, Box 596, SE-75124 Uppsala, Sweden 6
Department of Chemistry University of Illinois at Chicago, 845 West Taylor Street, Chicago, IL 60607-7061 USA
The exo-loop of Trichoderma reesei cellobiohydrolase Cel7A forms the roof of the active site tunnel at the catalytic centre. Mutants were designed to study the role of this loop in crystalline cellulose degradation. A hydrogen bond to substrate made by a tyrosine at the tip of the loop was removed by the Y247F mutation. The mobility of the loop was reduced by introducing a new disulphide bridge in the mutant D241C/D249C. The tip of the loop was deleted in mutant D(G245-Y252). No major structural disturbances were observed in the mutant enzymes, nor was the thermostability of the enzyme affected by the mutations. The Y247F mutation caused a slight kcat reduction on 4-nitrophenyl lactoside, but only a small effect on cellulose hydrolysis. Deletion of the tip of the loop increased both kcat and KM and gave reduced product inhibition. Increased activity was observed on amorphous cellulose, while only half the original activity remained on crystalline cellulose. Stabilisation of the exo-loop by the disulphide bridge enhanced the activity on both amorphous and crystalline cellulose. The ratio Glc2/(Glc3 þ Glc1) released from cellulose, which is indicative of processive action, was highest with Tr Cel7A wild-type enzyme and smallest with the deletion mutant on both substrates. Based on these data it seems that the exo-loop of Tr Cel7A has evolved to facilitate processive crystalline cellulose degradation, which does not require significant conformational changes of this loop. q 2003 Elsevier Ltd. All rights reserved.
7
Centre de Recherche´ sur les Macromolecules Vegetales CNRS, F-38041 Grenoble Cedex 9, France
*Corresponding author
Keywords: cellulose; crystal structure; glycoside hydrolase; processivity; product inhibition
Present address: C. Divne, Department of Biotechnology, AlbaNova University Centre, Royal Institute of Technology, SE-106 91 Stockholm, Sweden. Abbreviations used: asu, asymmetric unit; BMCC, bacterial microcrystalline cellulose; C-C, Cel7A D241C/D249C mutant; CNP-Lac, 2-chloro-4-nitrophenyl-lactoside; DEL, Cel7A D(G245-Y252) deletion mutant; HEC, hydroxyethyl cellulose; Pc Cel7D, Phanerochaete chrysosporium Cel7D; pNP-Lac, 4-nitrophenyl-lactoside; Tr, Cel7A, Trichoderma reesei Cel7A; WT, wild-type. E-mail address of the corresponding author:
[email protected] 0022-2836/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
818
Introduction Biodegradation of crystalline cellulose is a considerable challenge since the substrate is insoluble and poorly accessible to enzymes. Filamentous fungi produce enzymes capable of fast and efficient cellulose degradation. Two cellobiohydrolases, Cel7A and Cel6A, are the key cellulases of the soft rot fungus Trichoderma reesei. Both enzymes exhibit fast rate of solubilisation but very slow decrease in the degree of polymerisation (DP) of crystalline cellulose.1,2 This is different from classical endoglucanases, which cleave bonds in the middle of the polymeric chains thus reducing dramatically the DP of the substrate. The catalytic domain structures of both T. reesei cellobiohydrolases have been determined in complex with a number of oligosaccharides and ligands.3 – 6 Both enzymes have active sites buried in tunnels extending deep inside their catalytic domains. Similar to other microbial cellobiohydrolases,7 – 10 the roof of each active site tunnel is created by two (Cel6A) or four (Cel7A) surface loops. Endoglucanases in families GH6 and GH7 utilise the same catalytic mechanisms, and have similar overall structures, but are nevertheless inefficient in the degradation of crystalline substrates. This is generally attributed to the presence of significantly shorter or no active site loops, leading to more open active site architectures.11 It is intriguing that cellobiohydrolases are capable of efficient cellulose degradation in spite of an active site structure that seems to exclude any direct access to the glucan chains on the cellulose crystals. Based on extensive structural and biochemical data, we have proposed that a single glucan chain end enters the tunnel from one end (“entrance”), threads through the entire tunnel for bond cleavage in the far end with the products leaving from the opposite (“exit”) end of the tunnel.3,4,6,12 This model is supported by data indicating that removal of substrate binding interactions at the tunnel entrance of Tr Cel6A selectively reduces the enzyme activity on crystalline but not on soluble or amorphous substrates.13 Similar observations have been made for the Tr Cel7A (A.K. & T.T.T., unpublished results) and for other types of glycosyl hydrolases working on crystalline substrates.14,15 According to an alternative model, the initial contact between the enzyme and cellulose may occur by sporadic opening of the active site loops.16,17 This would allow the chain to be first bound along the length of the channel where after the loops would close and cleavage take place. The first hydrolytic event, be it exo- or endolytic, is followed by processive degradation of the glucan chain, i.e. catalysis of several consecutive bond cleavages before the dissociation of the enzyme – substrate complex.4 – 6,12,13 The special active site design of cellobiohydrolases is likely to be responsible for their processive action on crystalline cellulose. Since the cellulose chain is enclosed and held by numerous protein
Loop Engineering in Cel7A
interactions, the enzyme is less likely to dissociate after each hydrolytic step, and can thus compete more efficiently with the interactions driving the cellulose chain back onto the cellulose crystal. However, such active site architecture also requires sufficient width and flexibility to accommodate the various conformations of the cellulose chain on its voyage through the tunnel. In the case of family 6 cellobiohydrolases, alternate loop conformations have been observed that probably facilitate the cellulose chain gliding and may allow for occasional endo type of cleavages.6,12,17,18 Further, so called processive endoglucanases employ a single active site loop, which undergoes a large “loopflip” conformational change to enclose the active site in a short tunnel upon substrate binding.18 – 21 These enzymes carry out initial endolytic cleavages, followed by processive action subsequent to the loop closure. Unlike the cellobiohydrolase loops, the mobile loops of processive endoglucanases are generally disordered in enzyme structures determined in the absence of a ligand, and hydrolysis of all substrates, crystalline or otherwise, is similarly impaired by the absence of the interactions contributed by the loop. In the case of Tr Cel7A, four active site loops serve to exclude as many as ten glucose units of the bound sugar chain from the solvent. An analysis of a series of complex structures of Tr Cel7A with oligosaccharides of different lengths suggest that the gliding of the glucan chain in the active site tunnel is facilitated by alternating direct and water-mediated protein– sugar interactions.4 The sugar chain conformation seems to be dictated by these interactions such that an almost fully extended conformation favouring chain sliding is promoted at the tunnel entrance, followed by a dramatic chain twist and subsequent distorted conformation favouring bond cleavage at the catalytic centre.4 No obvious flexibility or conformational changes have been observed in three of the four active site loops of Cel7A. However, in the “exo”loop, which covers the catalytic centre (residues 243 –256; see Figure 1(a)), backbone displacements ˚ have been observed and a tyrosine, of up to 2 A Y247, at the tip of the loop can have two alternative conformations.22,23 The side-chain of this tyrosine is in van der Waals contact with another tyrosine, Y371, on the opposite side of the tunnel, effectively enclosing the cellulose chain within the protein at the site of catalysis. In addition, the Y247 and a neighbouring residue, T246, have direct interactions with the glucosyl residues at sites 2 2 and þ 1, respectively.23 These interactions thus seemed to form a link between the sugar residues on each side of the scissile bond, which—after the bond cleavage—would keep the leaving group and the remaining cellulose chain each associated with the enzyme. We thus hypothesise that, subsequent to the major local conformational changes occurring in and around the 2 1 site during catalysis, these interactions might have a role in moving the substrate and the leaving group in and out of the
819
Loop Engineering in Cel7A
were made to investigate the contribution of the tyrosine 247 at the tip of the loop. The mutation Y247F abolished the hydrogen bond to the glucosyl residue at the 2 2 subsite, and the mutation Y247A reduced the bulk of the side-chain thus preventing its interaction with the loop on the opposite side of the tunnel. Deletion of an exo-loop of a bacterial cellobiohydrolase has been previously shown to increase its endolytic activity.24 In order to generate a more open active site architecture for Tr Cel7A, its exo-loop was shortened by deleting eight residues from the tip of the loop (the mutant D(G245Y252)). Stability of the mutant proteins
Figure 1. The loop structure over the active site of Tr Cel7A. (a) Two loops (green) fold over the active site and close the tunnel by an interaction of the tyrosines Y247 in the “exo” -loop and Y371 in a short loop on the opposite side of the tunnel. Side-chains are drawn for the residues D241 and D249 in the exo-loop (blue carbons) and the active site residues, E212, D214 and E217 (magenta carbons). A model of a bound cellulose chain (yellow carbons) (from PDB entry 8CEL4) is also drawn. (b) Sequences of the exo-loop regions of the different variants of T. reesei (Tr) Cel7A, P. chrysosporium (Pc) Cel7D and the endoglucanase Tr Cel7B. In the D241C/D249C mutant a disulphide bridge is formed by the cysteins introduced at positions 241 and 249 while the deletion mutant lacks the residues G245-Y252. The broken lines indicate the length of the natural “deletions” in the exoloop of Pc Cel7D and Tr Cel7B.
catalytic site, thus promoting the processive action of the enzyme. Here we have undertaken a protein engineering approach to vary the structure and mobility of the exo-loop in Cel7A, in order to examine its role and importance in crystalline cellulose degradation.
Results and Discussion Design of the mutations Our first goal was to reduce the observed mobility of the exo-loop of Cel7A by bridging it by a disulphide bond to the short loop on the opposite site of the active site cleft. However, a detailed structural analysis revealed that the distance between and the mutual geometry of the two loops do not allow for such an interaction without compromising sugar binding at the catalytic site. Instead, we found that a stabilising interaction between two aspartic acid residues within the exoloop could be replaced by a disulphide to generate a more stable loop conformation (the mutant D241C/D249C; Figure 1(a) and (b)). Two mutations
The mutation Y247A led to an unstable protein that persistently precipitated during purification in a variety of buffers. Even though the protein could not be further characterised, this observation implies that, in spite of its location on a surface loop, this tyrosine may be critical in maintaining the integrity of the protein structure. The effect of the mutations on protein stability was evaluated using fluorescence spectroscopy.25 Protein unfolding was monitored both at pH 5 and pH 8, by gradually heating Cel7A samples to 80 8C and measuring the change in intrinsic protein fluorescence. At pH 5 the unfolding transition appeared between 60 8C and 65 8C, as observed before for Cel7A wild-type,26 and at pH 8 between 35 8C and 40 8C. The Y247F and C-C mutants were very similar to wild-type, whereas the deletion mutant showed a marginally reduced thermostability with , 2 8C lower transition temperature at both pH 5 and 8 (data not shown). Further, comparison of the activities of the wild-type Tr Cel7A and the Tr Cel7A C-C mutant at 27 8C, 40 8C and 60 8C revealed no differences in the release of reducing sugar from BMCC (data not shown). The new disulphide bridge did thus not significantly influence the thermal stability of Tr Cel7A. Comparison of the activities on small soluble substrates Table 1 compares enzymatic properties for Tr Cel7A wild-type (WT), the three mutants as well as the cellobiohydrolase Cel7D from Phanerochaete chrysosporium (Pc Cel7D). The overall sequence and structure of Pc Cel7D are very similar to Tr Cel7A, but, similar to the deletion mutant, it has a natural deletion of six residues in the corresponding exo-loop (Figure 1(b)).27 No dramatic changes were observed in the kcat or KM values of the Y247F or the disulphide mutants on 4-nitrophenyl-lactoside (pNP-Lac), nor was the cellobiose inhibition affected. In contrast, deletion of the tip of the loop led to increased kcat, KM and Ki values and changed the mode of cellobiose inhibition. On cellotetraose, which is a better substrate to Cel7A than aryl lactosides, a similar trend was observed with the highest turnover number for DEL followed by Pc
820
Loop Engineering in Cel7A
Table 1. Kinetics on chromogenic lactosides, inhibition by cellobiose, hydrolysis of cellotetraose and relative efficiency on insoluble cellulose for wild-type and loop mutated Cel7A Glc4 hydrolysisc (s 1)
Kinetics with pNP-Laca Enzyme
kcat (s 1)
KM (mM)
kcat/KM (M 1 s 1)
Ki for cellobiose (mM)
Type of inhibition
0.093 0.082 0.061 0.19 0.10
0.41 0.44 0.40 1.8 1.3
230 190 150 110 80
24 22 22 300d 180d
Competitive Competitive Competitive Mixed, a ¼ 3.3d Mixed, a ¼ 5.7d
2
WT C-C Y247F DEL P.c. Cel7D
2
2
2
0.50 0.58 0.58 0.83 0.67
Relative efficiency on celluloseb Amorphous
BMCC
1 1.7 1.5 1.6 8
1 1.5 0.9 0.5 4.4
a Kinetic experiments were performed at pH 5.0 and 30 8C using 4-nitrophenyl-lactoside (pNP-Lac) as substrate and cellobiose as an inhibitor. Standard errors for the kinetic parameters, derived by non-linear regression, were 7–15% for Km and 3 –6% for kcat values. b Each progress curve of amorphous and BMCC hydrolysis (Figure 2) was shifted by multiplying the reaction times with a scale factor, which was adjusted until the seemingly best overlap was obtained with the progress curve for wild-type Cel7A. The final scale factor values, listed here, are measures of the relative efficiency of insoluble cellulose hydrolysis. The shifted progress curves overlap well, as seen in the inset in Figure 2(b). c Rate of turnover of cellotetraose (Glc4; 300 mM) at 27 8C, pH 5.0. The uncertainty was estimated to be 10–15% on the basis of duplicate experiments. d Mixed-type inhibition constants were estimated from inhibition experiments with CNP-Lac as substrate at pH 5.7 and 33 8C.
Cel7D before Y247F, C-C and WT Tr Cel7A (Table 1). The substrate specificities of the Tr Cel7A WT, DEL and Pc Cel7D were then measured using a series of 3,4-dinitrophenyl cello-oligosaccharide glycosides (3,4-DNP-Glc2 – 5). The kcat/KM for release of the chromophore was measured from continuously monitored initial rates at very low substrate concentration and at very low conversion (Table 2). This method is not affected by non-productive binding or holosidic cleavage; the same value of kcat/KM was obtained at two substrate concentrations differing by a factor of 2, indicating [S] . . KM. The kcat/KM of 460 M21 s21, here measured for Cel7A WT on the 3,4-DNP-Glc2 (Table 2), is similar to the value 350 M21 s21 previously calculated on the same substrate by determination of the individual kinetic parameters kcat and KM.28 As with the pNP-Lac (Table 1), somewhat lower kcat/KM values were observed for both the deletion mutant and Pc Cel7D than for the Tr Cel7A WT. When the substrate length was increased from two to three glucose residues a 100-fold increase in the kcat/KM was observed for the Tr Cel7A WT (Table 2). Less pronounced but
Table 2. Specificity constants, kcat/Km, of release of chromophore from 3,4-dinitrophenyl-cellooligosaccharides with 2 –5 glucose units Specificity constant, kcat/Km (s 1 M 1) 2
2
Enzyme
–Glc2
–Glc3
–Glc4
–Glc5
WT DEL Pc Cel7D
460 260 310
110,000 20,000 25,000
270,000 62,000 250,000
930,000 56,000 170,000
Initial rates were measured at 25 8C, pH 5.0, by continuousmonitoring at very low substrate concentration and at very low conversion, in order to avoid influence by non-productive binding and holosidic bond cleavage.
nevertheless clear increase was thereafter observed for up to five glucose units. Interestingly, this dependence of kcat/KM on substrate length was much less pronounced with the Tr Cel7A DEL mutant on substrates longer than 3,4-DNP-Glc3. Some contribution of the intrinsic binding energy from the 2 5 subsite thus appears to be manifested in catalysis with wild-type Tr Cel7A, but not the deletion mutant or Pc Cel7D. Since the exo-loop excision also compromises interactions at the þ 1 and þ 2 subsites, as shown by reduced product inhibition, tight binding of the substrate on both sides of the catalytic centre seem to be required in order to achieve a complete transition state stabilisation on longer substrates. The more complex behaviour observed for Pc Cel7D suggest that, during evolution, its active site has been redesigned to accommodate a shorter exo-loop. Compared to the contribution from the 2 3 subsite, the contribution of intrinsic binding energy at the 2 4 subsite (and 2 5 subsite in WT Tr Cel7A) is small (Table 2). Likewise, the intrinsic kcat/Km for the same soluble substrate changes over a comparatively small range for each of the three enzymes. The specificity constant reflects the first irreversible step after substrate binding. For each protein, kcat/Km varies with substrate chain-length, so a protein-only event (such as opening of the loop) cannot be rate determining. Therefore, the loops must either open or close very rapidly and reversibly, so that their motion is not rate limiting, or the glycosides must thread their way into the tunnel from either end at a rate fast compared to bond cleavage. It is perhaps significant that, most clearly for the TrCel7A DEL and Pc Cel7D, the value of kcat/Km reaches a plateau value below 106 M21 s21. This is one or two orders of magnitude slower than the diffusion-controlled binding of small substrates to open enzyme active sites. If the opening of the tunnel is rapid and reversible, the limiting value of kcat/Km could represent
Loop Engineering in Cel7A
821
Figure 2. Progress curves for hydrolysis of insoluble cellulose by Tr Cel7A wild-type and loop mutants and Pc Cel7D at 27 8C, pH 5.0. Soluble sugars were analysed and quantified by HPLC. Solubilisation was calculated as the sum of the mass of all the measured soluble sugars divided by the initial mass of cellulose. (a) Amorphous cellulose. The enzyme and substrate concentrations were 1.5 mM and 0.7 mg/ml for the 0 – 4 hour time points, and 1.4 mM and 1.2 mg/ml for the 8 – 48 hour time points. (b) Bacterial microcrystalline cellulose (BMCC). Enzyme and substrate concentrations were 1.5 mM and 0.7 mg/ml, respectively. The inserted log – log diagram in (b) shows how well the progress curves overlap if incubation times are multiplied by the “relative efficiency” factors given in Table 1. A key to the symbols used in both diagrams is shown in (a).
diffusion-controlled binding of substrate to the small fraction of enzyme in which the lid of the tunnel is open. Alternatively, if the tunnel does not open, the approach of these large substrates to the restricted area of the tunnel entrance only is likely to be productive, so 106 M21 s21 could represent enzyme –substrate diffusion with these constraints. Hydrolysis of polymeric substrates In order to evaluate the performance of the mutant proteins on more complex substrates, their activities were measured on hydroxyethyl cellulose (HEC), amorphous cellulose and bacterial microcrystalline cellulose (BMCC). None of the mutations caused major changes in the activity of the enzyme on HEC, which is a typical endoglucanase substrate (data not shown). The HEC activity of Pc Cel7D (0.050 s21) was higher than that of the Tr
Cel7A WT (# 0.010 s21), but still much lower than that of the classical endoglucanases, Tr Cel7B and Tr Cel5A (4 s21 and 18 s21, respectively). The release of soluble sugars from amorphous cellulose and bacterial microcrystalline cellulose (BMCC) was analysed by HPLC. In 48 hours, extensive solubilisation was achieved with all of the enzymes, varying between 50% and 90% on amorphous cellulose and 27 – 60% on BMCC. The progress curves, shown in Figure 2, reveal a nonlinear levelling off with time, which is typical for enzymatic cellulose degradation and makes the kinetic analysis difficult. Here we have overcome the non-linearity by shifting the progress curves of the mutant enzymes by adjusting the time-scale until a satisfactory overlap was obtained with the progress curve of the wild-type enzyme (see inset in Figure 2(b)). The factor by which the time-scale is multiplied thus approximates the relative efficiency of conversion of cellulose to soluble sugars
822
Loop Engineering in Cel7A
Structural interpretation of the activity data The X-ray structures of the mutants Y247F and D241C/D249C, and of D(G245-Y252) in complex with the tetrasaccharide Glc2-S-Glc2, were deter˚ , 1.6 A ˚ and 1.8 A ˚ resolution, respecmined at 1.9 A tively. No significant structural changes were observed in any of the mutant proteins in other than the mutated regions. Y247F mutant
Figure 3. Comparison of the structure of the exo-loop in the D241C/D249C disulphide mutant (pink) and in the complex of Cel7A E217Q (wheat) with cellohexaose and cellobiose (PDB entry 7CEL), together with a model of a cellulose chain from PDB entry 8CEL.4 Hydrogen bonds in the 7CEL structure are indicated with small white spheres.
(Table 1). On amorphous cellulose, all the three mutants had higher activities than Tr Cel7A WT (Figure 2(a)). On bacterial microcrystalline cellulose (BMCC) the disulphide mutant had slightly higher activity than the Tr Cel7A WT, while the deletion mutant and the Y247F mutant had significantly lower activities (Figure 2(b)). The Pc Cel7D enzyme was considerably faster than Tr Cel7A WT or any of the mutants, on both of these substrates (Figure 2 and Table 1). While cellobiose was the main product, small amounts of glucose and cellotriose were also produced by all the enzymes (data not shown). In the case of the DEL and the C-C mutants, traces of cellotetraose were also detected. With a processive enzyme initiating its hydrolytic process at a chain end, the production of glucose and/or cellotriose can only occur upon an initial attack on a chain end, while only cellobiose is produced upon processive hydrolysis. Therefore, the ratio of Glc2/ (Glc3 þ Glc1) has been used as a rough estimate of the degree of processivity.12 Consistent with this assumption, a relatively high value of about 14 is observed for Tr Cel7A WT while a much lower value of 1.2 is obtained for the homologous endoglucanase, Tr Cel7B, after 24 hours of hydrolysis of amorphous cellulose. The Y247F mutant exhibited a value unaltered from the Tr Cel7A WT, while a slightly decreased value of 10 was obtained for the disulphide mutant and Pc Cel7D. A somewhat larger decrease down to a value of 7 was observed with the deletion mutant. On BMCC the cellobiose to glucose ratios were higher than on amorphous cellulose with a value of 23 for Cel7A WT, but the trend was similar with a lowest ratio for the deletion mutant (14), which implies decreased processivity.
The electron density of this mutant clearly revealed the missing tyrosine hydroxyl group (data not shown). The removal of a hydrogen bond to OH6 of the hexose in site 2, by the Y247F mutation, maintained the wild-type enzyme structure but caused a selective, albeit small decrease in kcat (Table 1). This shows that the interaction makes a small contribution to transition-state stabilisation. However, since the reduction of the efficiency on BMCC was tiny (Figure 2(b)), this interaction is obviously not crucial for the processive action of Tr Cel7A on crystalline cellulose. Disulphide mutant D241C/D249C The electron density confirms formation of a disulphide bridge between the introduced cysteine residues (data not shown) and superposition of the mutant protein with previous structures of Tr Cel7A reveals that the exo-loop structure is maintained. However, the loop position in the mutant protein is shifted slightly away from the active site relative to the unliganded structure of Cel7A WT (APO; J. S., M. H. & T. A. Jones, unpublished results) and somewhat more relative to the structures with bound oligosaccharides (Figures 3 and 4(b)). The major effect is on Y247, which forms a ˚ hydrogen bond to OH6 of the glucosyl resi2.8 A due in site 2 2 in the oligosaccharide complex, ˚ while the corresponding distance would be 3.6 A in the mutant (Figure 3). The other key interactions with substrate of residues T246 and R251 in the loop are not likely to be impaired. The temperature factors (B values) for the exo-loop are lower in the disulphide mutant than in the unliganded wildtype enzyme (APO), indicating reduced mobility (Figure 4(a)). A similar drop in the B values is observed for the exo-loop in Tr Cel7A WT structures determined in complex with oligosaccharides.4 It seems thus that introduction of the disulphide bridge accomplishes a loop stabilisation comparable to that caused by substrate binding. The reduced mobility of the exo-loop in the disulphide mutant did not impair the catalytic competence of the enzyme on small soluble substrates (Table 1) but instead led to enhanced rate of hydrolysis on both amorphous and crystalline cellulose. This shows that the somewhat better mobility of the exo-loop in the Cel7A WT is not essential for its processive action.
823
Loop Engineering in Cel7A
The deletion mutant D(G245-Y252)
Figure 4. Comparison of (a) temperature factors in the exo-loop region, and (b) Ca-distances between the structures of the D241C/D249C disulphide mutant and the unliganded Cel7A wild-type (APO) or the E217Q-cellohexaose complex (7CEL),4 respectively.
The structure shows well defined electron density for the Glc2-S-Glc2 ligand bound in subsites þ 1 to þ 4 and for the new, shorter loop (data not shown), while the rest of the structure is unaltered. The only minor change is a rotation of the aromatic side-chain of Y371 (Figure 5(a)) to an alternate position above the subsite þ 1, but this is also observed in two earlier structures of Tr Cel7A WT (1CEL and 1DY4).3,23 Owing to the deletion, the carbohydrate interactions of Y247, T246, R251 and Y252 are lost, the cleavage site (2 2/2 1) and the product binding subsites (þ 1/þ 2) become more exposed (Figure 5(b) and (c)), but all the other relevant protein– carbohydrate interactions are maintained. The binding of Glc2-S-Glc2 is similar to one of the two molecules of cellopentaose in complex with the Cel7A mutant E212Q (6CEL),4 with minor deviations apparent at sites þ 3 and þ 4 (Figure 5(a)). It differs, however, from the Cel7A E212Q complex with cellotetraose (5CEL) where one ligand molecule was found in sites 2 2 to þ 2, spanning the cleavage site, and a second molecule appeared at sites 2 7 to 2 4.4 In this case, no electron density is observed for a second ligand molecule binding in the other end of the substrate binding tunnel in spite of the fact that those subsites were not mutated. Figure 5(a) also shows the exo-loop structure in Pc Cel7D (1GPI),27 which is naturally shorter than
Figure 5. (a) Structural comparison of the deletion mutant and wild-type Cel7A. The backbone of the deletion mutant D(G245-Y252) is shown as a light-blue thin ribbon with the engineered exo-loop thicker and highlighted in darker blue. Residue side-chains have blue-green C atoms and the bound tetrasaccharide Glc2-S-Glc2 blue C atoms. The exo-loop, ligand and selected residues of the Cel7A E212Q/cellopentaose are shown with white C atoms (PDB entry 6CEL)4 and the exo-loop of Pc Cel7D with gold C atoms (PDB entry 1GPI).27 Green dotted lines indicate hydrogen bonds in the 6CEL structure. The loop deletion results in the loss of several carbohydrate-interacting residues, T246, Y247, R251 and Y252. Note that the equivalent of R251 is retained in Pc Cel7D. (b) and (c) Space-filling models showing the difference in solvent exposure of the carbohydrate ligands and the active site in the E212Q/cellopentaose (6CEL)4 and the D(G245-Y252)/Glc2-S-Glc2 complexes, respectively. The residues in the exo-loop and Y371 in the opposing loop are highlighted (green and blue, respectively). The ligands are shown with yellow C, red O and green S atoms. In all the panels the exo-loop is defined as the region between the anchoring cystein residues at positions 243 and 256.
824
in Tr Cel7A, but two residues longer than in the deletion mutant (Figure 1(b)). Not surprisingly, the kinetic parameters measured for the deletion mutant were more similar to Pc Cel7D than to Tr Cel7A WT (Table 1). The strong cellobiose inhibition of Tr Cel7A WT was clearly reduced in Pc Cel7D and even more so in the Tr Cel7A DEL. The product binding sites of the two enzymes are very similar with the exception of the arginine R251,27 which is removed in Tr Cel7A DEL, but preserved in Pc Cel7D. It seems thus that the hydrogen bonding interactions of this arginine to the glucosyl hydroxyls OH6 in site þ 1 and OH3 in site þ 2 contribute significantly to product inhibition (see Figure 5(a)). The release of the chromophore from aryl lactosides and cellobiosides requires binding of the disaccharide moiety at subsites 2 1 and 2 2, while the aglycone primarily occupies subsite þ 1. Owing to the strong sugar binding at the product sites in Tr Cel7A WT, the substrate may also bind non-productively with its disaccharide moiety in sites þ 1 and þ 2. Weakening of binding in sites þ 1/þ 2 may thus also contribute to the observed increase in kcat and Km on pNP-Lac through reduced non-productive binding. However, changes in non-productive binding alone would affect kcat and Km equally leaving kcat/Km unchanged. Here the decrease in kcat/Km shows that transition-state stabilisation is also affected. Finally, decreased product inhibition may also explain, at least in part, the faster rate of hydrolysis of cellotetraose by Tr Cel7A DEL (Table 1). Despite the generally faster kinetics on soluble and amorphous substrates, the deletion in the exoloop led to a significant loss of activity on crystalline cellulose. This loss of activity was accompanied by a reduction in the cellobiose/glucose ratio, which implies reduced processivity. On amorphous cellulose, the apparent processivity of the Tr Cel7A DEL was also decreased but the activity was increased. These data suggest that the processive character of Tr Cel7A is exclusively required for crystalline cellulose degradation.
Conclusions Our current data show that interactions contributed by the exo-loop of Tr Cel7A contribute little to the general structural stability or the transitionstate stabilisation on short soluble substrates. Reduced mobility of the loop, as evidenced by lower B-factors of the disulphide mutant, did not impair the enzyme activity indicating that large conformational changes of the exo-loop are not required for cellulose degradation. However, deletion of the exo-loop led to selective reduction of the enzyme activity on crystalline cellulose, while the activity on soluble and amorphous substrates was maintained or increased. These data corroborate earlier suggestions that crystalline cellulose hydrolysis has a different rate limiting step than
Loop Engineering in Cel7A
soluble and amorphous substrates.13,21 The two obvious factors that could contribute to such a difference are the extraction of the glucan chains from the crystalline cellulose surfaces and the processive degradation. While the chain extraction is likely to be governed by interactions at the tunnel entrance,13 our present data show that the exoloop contributes to the high processivity of Tr Cel7A. Our data on the deletion mutant also suggests that efficient degradation of longer substrates may be achieved by compromising between processivity and end-product inhibition. Accordingly, we propose that the tight binding of the cellulose chain over the subsites 2 7 to 2 2 must be balanced by strong binding of the reducing end of the cellulose chain in the product sites þ 1 and þ 2 in order to ensure that that the enzyme proceeds forward along the chain once hydrolysis has taken place. However, since enzymes acting on lignocellulosic substrates often exhibit subtle functional differences reflecting details of their substrate specificity, the conclusions drawn for one enzyme in a given family do not necessarily apply to another. Thus, Pc Cel7D is able to maintain high processivity, leading to efficient action on crystalline cellulose, in spite of its shorter active site loops and weaker binding in the product subsites þ 1/þ 2. These natural differences, and the fact that extensive alterations of the exo-loop can be made without disturbing the structural and catalytic integrity of the enzyme, suggest that the balance between product inhibition, non-productive binding, catalytic rate and processivity can be further optimised by protein engineering for enhanced performance on different cellulosic substrates.
Materials and Methods Strains, plasmids and DNA manipulations Escherichia coli XL-1 Blue (Stratagene, La Jolla, USA) was used for all DNA constructions and Trichoderma reesei ALKO3414 (Primalco Ltd, Finland) (cel7A-, cel7B-), for the expression of the mutant proteins. The plasmid pEM-F5, containing the cel7A cDNA gene under its own promoter29 was used as the fungal expression vector, and the phagemid pKS- (Stratagene, La Jolla, USA) for subcloning and mutagenesis. The hygromycin selection plasmid pRLMex3030 was used in the co-transformation of T. reesei. Established procedures were employed in all DNA manipulations.31 Mutagenesis and expression The D(G245-Y252), D241C:D249C, and Y247F mutations were introduced into the cel7A cDNA by PCR as described.32 A 383 bp Eco RV– Bam HI fragment from pEM-F5 was cloned into pKS- and mutagenised using the following pairs of oligonucleotide primers: 50 GGCG GCACTTGCGATCCCGATGGC and 50 GCCGCACCCAT CACCCTCGCAGAT for D(G245-Y252), 50 CGCCGCACC CACAACCCTCGCAGA and 50 GAACTTACTCCTGTA ACAGATATG for D241C:D249C, and 50 CGGCGGAACT TTCTCCGATAACAG and 50 CACCCATCACCCTCGCA
825
Loop Engineering in Cel7A
GATCTCC for Y247F. The mutations were confirmed by DNA sequencing where after the mutated fragments were reinserted in pEM-F5. All of the constructions in pEM-F5 were co-transformed in T. reesei (ALKO3414) using the hygromycin selection plasmid pRLMex30 as described.33 Screening of the transformants was done by using the anti-Cel7A monoclonal antibody CI-26134 in dot-blotting and Western blotting experiments as described.35,36 The transformants with highest levels of extracellular protein were subjected to large-scale shake flask cultivation.35
crystals used for X-ray analysis were extracted directly from the crystallisation drop with a cryo-loop (Hampton Research, Laguna Niguel, CA, USA) and flash-cooled in liquid nitrogen. Diffraction data were recorded at 100 K, either on our in-house Rigaku/R-AXIS IIC rotating anode X-ray diffractometer Y247F and D(G245-Y252), or at the synchrotron beamline I711, MAX-Lab, Lund, Sweden (D241C/D249C). The data were indexed, processed and scaled with the HKL program package.41 Statistics are summarised in Table 3. Initial phases for the Y247F and D241C/D249C structures were taken from the refined protein coordinates of
Protein preparation The purification of mutated and wild-type Cel7A proteins was carried out essentially as described.22 Briefly, protein concentrates were loaded onto Q-Sepharose Fast Flow columns (Pharmacia and Upjohn, Sweden) and eluted by sodium acetate buffered gradients of 40 – 1000 mM (pH 4) and 50 – 300 mM (pH 3.5), respectively. The procedure was repeated once where after the proteins were adsorbed on Source 30-Q columns (Pharmacia and Upjohn, Sweden) and eluted by a gradient of 50 – 300 mM sodium acetate (pH 3.5). P. chrysosporium Cellobiohydrolase Cel7D, purified37 from the culture liquid38 of P. chrysosporium strain K3, was a kind gift from Dr Go¨ran Pettersson, Department Biochemistry, Uppsala University. A final purification on a Source 30-Q column was carried out as described above. In all cases, column fractions containing Cel7A protein were identified by SDS-PAGE (Phast gel; Pharmacia and Upjohn, Sweden) and dot-blotting analysis with the antibody CI-261,34 pooled, concentrated and equilibrated in the appropriate buffer using an Amicon cell with a 10 kDa molecular mass cut-off polysulfone membrane. Enzyme concentrations were estimated spectrophotometrically at 280 nm using the following extinction coefficients: Tr Cel7A WT 83,000, Y247F 81,510, D241C/D249C 83,125, D(G245-Y252) 80,020 and Pc Cel7D WT 78,290 M21 cm21. For crystallisation, the catalytic modules for the Cel7A mutants were obtained by limited proteolysis22 followed by a deglycosylation treatment with a-mannosidase and endoglycosidase F.39 The deglycosylated catalytic module was then purified on a Source 30-Q column using the same conditions described above. X-ray structure analysis Hanging-drop vapour diffusion crystallisation40 was performed at room temperature with the catalytic modules of the enzymes. The mutants Y247F and D241C/ D249C were crystallised at conditions previously developed for Cel7A.4 A 2 ml portion of 1 – 5 mg/ml protein in 10 mM sodium acetate (pH 5.0) was mixed with 2 ml reservoir solution containing 18 – 22% (w/v) monomethylether polyethylene glycol 5000 (mPEG 5000; Fluka) as precipitant, 0.1 M sodium morpholine ethane sulphonic acid buffer (pH 6.0) 10 mM CoCl2 and 12.5% (v/v) glycerol as cryo protectant. The crystals belong to the space group I 222 with one molecule per asymmetric ˚ , b ¼ 83 A ˚, unit and with cell dimensions of a ¼ 83 A ˚ . Crystals of the deletion mutant D(G245-Y252) c ¼ 110 A did not appear under the same conditions, but were obtained using a reservoir with 16% (w/v) mPEG 5000 in 10 mM Tris–HCl (pH 7.0) 20 mM CaCl2 and 10% (v/v) ethylene glycol. They have similar unit cell dimensions, ˚ , b ¼ 84 A ˚ , c ¼ 110 A ˚ , but belong to space group a ¼ 83 A P21212 with two molecules in the asymmetric unit. The
Table 3. Crystallographic data and refinement statistics Complex PDB code Data collection environment Space group Cell parameters ˚) (a, b, c; A Resolution ˚) rangea (A Unique reflections Completenessa (%) Average multiplicitya kI/s(I)l Rmerge (%)a,b Number of reflections in work setc Completeness of work set (%) Number of protein atoms Number of hetero atoms Number of solvent molecules Heterocompounds Mean B all protein ˚ 2) atoms (A Mean B solvent molecules ˚ 2) (A R value/Rfree valued (%) Rmsd from ideal bond lengths Ramachandran outliers a
D241C/D249C no ligand
D(G245-Y252) Glc2-S-Glc2
1Q2B Synchrotron beamline I711 MAX-lab, Mar345 detector, ˚ l ¼ 0.986 A I 222 (1 mol/asu) 83.2, 83.1, 110.0
1Q2E Rotating anode, Cu Ka Rigaku/R˚ Axis IIC, l ¼ 1.542 A
40– 1.60 (1.66–1.60)
40–1.75 (1.81–1.75)
46,521
78,914
92.0 (95.8)
100.0 (99.9)
4.0 (3.8)
4.1 (3.8)
19.2 5.3 (14.7) 45,625
21.3 4.6 (13.2) 76,597
90.0
97.1
3216
6312
16
121
358
604
NAG at Asn 270 10.0
NAG at Asn270; Glc2S-Glc2 in active site 10.3
12.7
17.5
20.7/22.3
22.6/23.9
0.005
0.005
0.8% (3 res)
1.3% (5 res/prot)
P 21212 (2 mol/asu) 83.3, 84.3, 110.4
Outer shell are given parentheses. Pstatistics P inP P Rmerge ¼ ½ hkl i lI 2 kIll= hkl i lIl100%: The entire resolution range was used for refinement with no sigma cut-off P on amplitudes. P d R ¼ kFo l 2 lFc k= lFo l; the final R-factor is given and the last recorded value of Rfree for 2% test reflections. b c
826
Tr Cel7A E217Q mutant in complex with cellohexaose and cellobiose (PDB accession code 7CEL)4 with tem˚ 2. The refinement program perature factors reset to 20 A 42 CNS, and modelling program O43 were then used to refine the model, essentially as described.28 An initial cycle of simulated annealing was followed by five cycles of energy minimisation and B-factor refinement. The Rfree value42,44 was calculated at all stages from 2% of the reflections, and rebuilding and solvent addition was judged by (2Fo 2 Fc, ac) and (Fo 2 Fc, ac) maps. The electron density maps of the Y247F mutant obtained after initial rigid-body minimisation and simulated annealing confirmed that the tyrosyl hydroxyl was no longer present at residue 247. No other significant differences from the structure of the Cel7A wild-type catalytic module without ligand (J. S., M. H. & T. A. Jones, unpublished results) were observed and the structure was not further refined. To crystals of the deletion mutant, the cellotetraose analogue Glc2-S-Glc2 (methyl 4-S-b-cellobiosyl-4-thio-bcellobioside; synthesised as described)45 was added prior to data collection. A small aliquot of ligand (20 mM in water) was added directly to a crystallisation drop with mature crystals to a final concentration of approximately 2 mM. After three days of soaking, dif˚ and 1.75 A ˚ resfraction data were collected between 20 A olution on our Rigaku/R-AXIS diffractometer. The structure was solved by molecular replacement46,47 as implemented in the program AMoRe.48 Statistics of the diffraction data and for the final models of the disulphide and deletion mutant structures are summarised in Table 3. Coordinates and corresponding structure factors have been deposited with the RCSB Protein Data Bank49 with entry codes 1Q2B for the disulphide mutant and 1Q2E for the deletion mutant. Temperature induced unfolding Unfolding studies based on monitoring the intrinsic tryptophan fluorescence of Cel7A wild-type and the mutant proteins were performed on a Shimadzu RF-5000 spectrofluorometer. Emission and excitation spectra were recorded with bandwidth of 5 nm on both monochromators. Unfolding was monitored by heating samples gradually up to 80 8C and measuring the fluorescence intensity.25 A thermostated cuvette holder connected to a water bath controlled the temperature of the sample solution. The temperature was monitored continuously using a Fluke 52 electronic thermometer equipped with K-type thermocouple that was immersed in the solution. Fluorescence emission spectra from 300 nm to 500 nm were collected and emission intensity at 340 nm was measured after every 1 8C in the transition areas. The excitation wavelength was 285 nm and the spectra were corrected for the buffer spectrum. The enzyme concentrations used in the measurements varied from 0.5 mM to 1 mM Cel7A. The buffers used in the experiments were 50 mM sodium acetate (pH 5) and 50 mM potassium phosphate (pH 8). Enzyme kinetics on lactoside substrates Kinetic experiments with 0.5 mM full-length enzyme of Tr Cel7A WT, Y247F, D241C/D249C, D(G245-Y252), and Pc Cel7D WT, were performed at 30 8C in 50 mM sodium acetate (pH 5.0), in 96-well microtiterplates using 0.05– 5 mM 4-nitrophenyl-b-D -lactoside (pNP-Lac) as substrate, without and with cellobiose as inhibitor at
Loop Engineering in Cel7A
three concentrations:10, 20, 50 mM with Tr Cel7A WT, Y247F and D241C/D249C; 50, 200, 1000 mM with Tr Cel7A D(G245-Y252) and Pc Cel7D WT. The reaction (200 ml) was stopped after 45 min incubation by adding 150 ml of 0.5 M sodium carbonate and the absorbance was measured at 414 nm in a Multiscan MCC/340 microtiterplate reader (Lab Systems Ltd, Helsinki, Finland). Controls were included to compensate for possible background absorbance of the enzyme, substrate and inhibitor solutions. The background corrections ranged from 0.011 to 0.035 absorbance units. The amount of 4-nitrophenol released from pNP-Lac was calculated using an extinction coefficient of 16590 M21 cm21. Kinetic parameters KM and kcat were derived by non-linear regression with the program Ultrafit (Biosoft, UK) using the program-defined equations for Michaelis – Menten Kinetics, Robust Weighting and Statistical Weighting. Standard errors were 7 – 15% for KM and 3 – 6% for kcat values. Competitive inhibitor constants Ki for cellobiose could be derived for Tr Cel7A WT, Y247F and D241C/D249C. Since the deletion mutant D(G245-Y252) and Pc Cel7D WT displayed mixed inhibition by cellobiose, additional experiments were performed using 2-chloro-4-nitrophenyl-b-D -lactoside (CNP-Lac). All experiments were performed in microtiterplates at 33 8C. Substrate was prepared in 70 mM phosphate buffer (pH 5.7) and preincubated at 33 8C for five minutes. Cellobiose was used as inhibitor and was prepared in milliQ-water. Enzyme was diluted in the phosphate buffer. For all experiments 10 ml of enzyme, 10 ml of inhibitor and 180 ml of substrate were added in a microtiter plate and the release of CNP was measured continuously at 405 nm for ten minutes using a Benchmark microtiter plate reader. The final concentrations were: 750, 1690, 2440, 3190 and 3940 mM substrate, 100, 200, 400 and 600 mM inhibitor, 1.3 mM enzyme with mutant D(G245-Y252) and 2.3 mM enzyme with Pc Cel7D WT. The inhibition constant was estimated from a Dixon plot (1/reaction rate versus inhibitor concentration) and the corresponding a-value for mixed type inhibition from a plot of 1/nmax versus inhibitor concentration.
3,4-Dinitrophenyl cellooligosaccharide synthesis and kinetics Synthesis of 3,4-dinitrophenyl cellooligosaccharide b-glycosides A mixture of peracetylated cellooligosaccharide bromides (12.8 g), produced essentially as described,50 was dissolved in dry THF (20 ml). To this was added 4 mm molecular sieve, mercuric cyanide (2.0 g), anhydrous potassium carbonate (2.0 g) and a solution of previously dried 3,4-dinitrophenol (2.5 g) in dry THF (20 ml). The mixture was stirred overnight in the dark at room temperature, and then filtered. The solution was evaporated in vacuo, yielding the mixed 3,4-dinitrophenyl glycosides (14.8 g). A portion (10.8 g) was chromatographed on silica gel, using a gradient elution of acetonitrile/ dichloromethane from 1:50 to 1:2. Recrystallisation from dichloromethane/ethanol/hexane afforded the 3,4-dinitrophenyl glucoside (1.06 g), cellobioside (1.43 g), cellotrioside (0.98 g), cellotetraoside (0.67 g), cellopentaoside (0.33 g) and a mixture of cellopentaoside, cellohexaoside, and heptaoside (1.13 g). The purified fully acetylated glycosides were subjected to Zemple´n de-O-acetylation. The properties of 3,4-dinitrophenyl b-cellobioside, -cellotrioside, and -cellotetraoside prepared in this way, and
827
Loop Engineering in Cel7A
the parent acetyl derivatives, corresponded to those reported.51 Properties of 3,4-dinitrophenyl b-cellopentaoside 1 Mp 248– 254 8C (dec), [a]25 D 210.28 (c, 0.061, H2O), H NMR (2H2O) d 8.20 (d, 1H, J 9.2, arom H-5); 7.67 (s, 1H, arom H-2); 7.50 (d, 1H, J 9.2, arom H-6); 5.33 (d, 1H, J 7.6, anomeric (21)); 4.52 (apparent m, 4H, internal anomerics); 3.99–3.31 (36H). Anal. Calcd. For C36H54O30N2: C, 43.46; H, 5.47; N, 2.82. Found, C, 43.05; H, 5.31; N, 2.58. Hexadecaacetate mp 176–178 8C.
Kinetics on 3,4-dinitrohenyl-cellooligosaccharides Concentrations of stock solutions of 3,4-dinitrophenylcellooligosaccharides were determined by alkaline hydrolysis with 5 M NaOH and incubation at 37 8C until complete hydrolysis. The extinction coefficient for 3,4-dinitrophenolate is 13,700 M21 cm21 at 400 nm. In order to avoid non-productive binding influencing kcat/KM, turnover with very low substrate concentrations was measured. The substrate (G3 – 5) and enzyme concentrations (E) were as follows (in mM): with Tr Cel7A: G3 0.15 and 0.30, E 0.015; G4 0.07 and 0.14, E 0.015; G5 0.10 and 0.20, E 0.003; with DEL: G3 1.53 and 3.06, E 0.046; G4 0.07 and 0.14, E 0.092; G5 0.10 and 0.20, E 0.046; with Pc Cel7D: G3 0.30 and 0.60, E 0.031; G4 0.14 and 0.28, E 0.009; G5 0.20 and 0.40, E 0.009. Enzyme kinetics were followed at 400 nm with a Perkin – Elmer Lambda 19 UV/VIS spectrophotometer, which allows the measurement of very small absorbance changes with very high precision. Total volume in the 10 mm path length cuvettes was 700 ml. All experiments were performed at 25 8C, in a cell-block controlled by a Peltier element. The buffer solution was 0.1 M acetate (pH 5.0), in which a value of De of 5040 M21 cm21 at 400 nm was obtained. Initial rates were determined with the fitting routine of the Lambda 19 control software. The values for kcat/KM were determined and visualised using FigP for Windows. Activity assay on cellotetraose and hydroxyethylcellulose (HEC) Hydrolysis experiments on 300 mM cellotetraose were performed in 50 mM sodium acetate buffer (pH 5.0) at 27 8C. Samples were taken at intervals, stopped and analysed by HPLC (Waters Millipore, Milford, MA) equipped with a refractive index detector as described.52,53 The column used for the separation of the oligosaccharides was Aminex HPX-42A (Bio-Rad) and deashing cartridges (Micro-Guard, Bio-Rad) were used as pre-columns. Endoglucanase activity was measured on 0.9% (w/v) hydroxyethyl cellulose (HEC) (Fluka, Switzerland) in 50 mM sodium acetate buffer (pH 5.0) at 30 8C. Samples were taken at different time points and the reducing sugars released were determined by the DNS method using D -glucose as a standard.54 Hydrolysis of amorphous and crystalline cellulose Bacterial microcrystalline cellulose (BMCC) was prepared from Nata de Coco (Reyssons Food Processing, The Philippines) as described.55 The amorphous cellulose, regenerated from SO2-amine solvent, was prepared from Avicel (Fluka AG, Switzerland) as described.56 The enzymatic activities on BMCC were determined by
shaking the intact enzymes (final concentration of 1.5 mM) and substrate (0.7 mg/ml) at 27 8C in 40 mM sodium acetate (pH 5.0). The substrate concentration of amorphous cellulose was either 0.7 mg/ml (time points 0 – 4 hours) or 1.2 mg/ml (time points 8 – 48 hours) and the enzyme concentration 1.5 mM or 1.4 mM, respectively. Samples were taken at designated time points and the reaction was stopped by adding half the reaction volume of a stop-reagent containing 9 vol. of 94% ethanol and 1 volume of 1 M glycine (pH 11). The soluble sugars were analysed by HPLC after filtering the samples through Millex GV 0.22 mm units (Millipore) and using soluble oligosaccharides as standards as described.13 The effect of the temperature on the activity of the disulfide mutant and WT Tr Cel7A was determined by incubating 1.4 mM protein with 0.7 mg/ml BMCC in 40 mM NaOAc buffer (pH 5) at 27, 40 and 60 8C. To account for potential endoglucanase contamination control reactions with 1 nM Tr Cel5A added to the reaction mixtures were performed. The reactions were stopped at different time points (0–5 hours) by adding 60 mM NaOH. The amount of released cellobiose was determined by high performance anion exchange chromatography (HPAEC) and pulsed amperometric detection (PAD) using the conditions specified by the manufacture (Dionex, Sunnyvale, USA).
Acknowledgements Matti Siika-Aho (VTT) is thanked for purified T. reesei endoglucanases Cel5A (EGII) and Cel7B (EGI) used as control enzymes. Riitta Suihkonen (VTT) is thanked for excellent technical assistance and Nina Aro (VTT) for helpful advice. Primalco Ltd, Finland is thanked for the T. reesei strain ALKO3414. This work was mainly financed from the EU Biotechnology project BIO4-CT96-0580. Financial support is also gratefully acknowledged from the Swedish Foundation for Strategic Research via the Centre for Forest Biotechnology and Chemistry, Bo Rydins Foundation for Scientific Research, and the Swedish Council for Forestry and Agricultural Research (Uppsala). David Wilson is thanked for critical reading of the manuscript.
References 1. Kleman-Leyer, K., Siika-Aho, M., Teeri, T. T. & Kirk, T. K. (1996). The cellulases endoglucanase I and cellobiohydrolase II of Trichoderma reesei act synergistically to solubilise native cotton cellulose but not to decrease its molecular size. Appl. Environ. Microbiol. 62, 2883– 2887. 2. Srisodsuk, M., Kleman-Leyer, K., Kera¨nen, S., Kirk, T. K. & Teeri, T. T. (1998). Modes of action on cotton and bacterial cellulose of a homologous endoglucanase – exoglucanase pair from Trichoderma reesei. Eur. J. Biochem. 251, 885–892. 3. Divne, C., Sta˚hlberg, J., Reinikainen, T., Ruohonen, L., Pettersson, G., Knowles, J. K. et al. (1994). The three-dimensional crystal structure of the catalytic
828
4.
5.
6.
7.
8.
9. 10.
11. 12.
13.
14.
15.
16.
17.
18.
core of cellobiohydrolase I from Trichoderma reesei. Science, 265, 524– 528. Divne, C., Sta˚hlberg, J., Teeri, T. T. & Jones, T. A. (1998). High-resolution crystal structures reveal how ˚ long tunnel of a cellulose chain is bound in the 50 A cellobiohydrolase I from Trichoderma reesei. J. Mol. Biol. 275, 309– 325. Rouvinen, J., Bergfors, T., Teeri, T., Knowles, J. K. & Jones, T. A. (1990). Three-dimensional structure of cellobiohydrolase II from Trichoderma reesei (Erratum: Science (1990) Sep 21;249(4975):1359). Science, 249, 380– 386. Zou, J.-Y., Kleywegt, G. J., Sta˚hlberg, J., Driguez, H., Nerinckx, W., Claeyssens, M. et al. (1999). Crystallographic evidence for substrate ring distortion and protein conformational changes during catalysis in cellobiohydrolase Cel6A from Trichoderma reesei. Structure, 7, 1035– 1045. Meinke, A., Gilkes, N. R., Kwan, E., Kilburn, D. G., Warren, R. A. & Miller, R. C., Jr (1994). Cellobiohydrolase A (CbhA) from the cellulolytic bacterium Cellulomonas fimi is a beta-1,4-exocellobiohydrolase analogous to Trichoderma reesei CBH II. Mol. Microbiol. 12, 413– 422. Shen, H., Gilkes, N. R., Kilburn, D. G., Miller, R. C., Jr & Warren, R. A. (1995). Cellobiohydrolase B a second exo-cellobiohydrolase from the cellulolytic bacterium Cellulomonas fimi. Biochem. J. 311(1), 67 – 74. Zhang, S., Lao, G. & Wilson, D. B. (1995). Characterization of a Thermomonospora fusca exocellulase. Biochemistry, 34, 3386– 3395. Varrot, A., Hastrup, S., Schu¨lein, M. & Davies, G. J. (1999). Crystal structure of the catalytic core domain of the family 6 cellobiohydrolase II. Cel6A, from ˚ resolution. Biochem. J. Humicola insolens, at 1.92 A 337(2), 297– 304. Teeri, T. T. (1997). Crystalline cellulose degradation: new insights into the function of cellobiohydrolases. Trends Biotechnol. 15, 160– 167. Teeri, T. T., Koivula, A., Linder, M., Wohlfahrt, G., Divne, C. & Jones, T. A. (1998). Trichoderma reesei cellobiohydrolases: why so efficient on crystalline cellulose? Biochem. Soc. Trans. 26, 173– 178. Koivula, A., Kinnari, T., Harjunpa¨a¨, V., Ruohonen, L., Teleman, A., Drakenberg, T. et al. (1998). Tryptophan 272: an essential determinant of crystalline cellulose degradation by Trichoderma reesei cellobiohydrolase Cel6A (Erratum: FEBS Letters (1999) Mar 26;447(23):334). FEBS Letters, 429, 341– 346. Watanabe, T., Ishibashi, A., Ariga, Y., Hashimoto, M., Nikaidou, N., Sugiyama, J. et al. (2001). Trp122 and Trp134 on the surface of the catalytic domain are essential for crystalline chitin hydrolysis by Bacillus circulans chitinase A1. FEBS Letters, 494, 74 – 78. Uchiyama, T., Katouno, F., Nikaidou, N., Nonaka, T., Sugiyama, J. & Watanabe, T. (2001). Roles of the exposed aromatic residues in crystalline chitin hydolysis by chitinase A1 from Serratia marcescens 2170. J. Biol. Chem. 276, 41343– 41349. Armand, S., Drouillard, S., Schu¨lein, M., Henrissat, B. & Driguez, H. (1997). A bifunctionalized fluorogenic tetrasaccharide as a substrate to study cellulases. J. Biol. Chem. 272, 2709–2713. Varrot, A., Schu¨lein, M. & Davies, G. J. (1999). Structural changes of the active site tunnel of Humicola insolens cellobiohydrolase. Cel6A, upon oligosaccharide binding. Biochemistry, 38, 8884– 8891. Boisset, C., Fraschini, C., Schu¨lein, M., Henrissat, B. & Chanzy, H. (2000). Imaging the enzymatic diges-
Loop Engineering in Cel7A
19.
20.
21.
22.
23.
24.
25. 26.
27.
28.
29.
30.
31.
32.
tion of bacterial cellulose ribbons reveals the endo character of the cellobiohydrolase Cel6A from Humicola insolens and its mode of synergy with cellobiohydrolase Cel7A. Appl. Environ. Microbiol. 66, 1444– 1452. Davies, G. J., Tolley, S. P., Henrissat, B., Hjort, C. & Shu¨lein, M. (1995). Structures of oligosaccharidebound forms of the endoglucanase EGV from Humi˚ resolution. Biochemistry, 34, cola insolens at 1.9 A 16210– 16220. Varrot, A., Shu¨lein, M. & Davies, G. J. (2000). Insights into ligand-induced conformational change in Cel5A from Bacillus agaradhaerens revealed by a catalytically active crystal form. J. Mol. Biol. 297, 819– 828. Zhang, S., Irwing, D. C. & Wilson, D. B. (2000). Sitedirected mutation of non-catalytic resiudes of Thermobifida fusca exocellulase Cel6B. Eur. J. Biochem. 267, 3101– 3115. Sta˚hlberg, J., Divne, C., Koivula, A., Piens, K., Claeyssens, M., Teeri, T. T. & Jones, T. A. (1996). Activity studies and crystal structures of catalytically deficient mutants of cellobiohydrolase I from Trichoderma reesei. J. Mol. Biol. 264, 337– 349. Sta˚hlberg, J., Henriksson, H., Divne, C., Isaksson, R., Pettersson, G., Johansson, G. & Jones, T. A. (2001). Structural basis for enantiomer binding and separation of a common beta-blocker: crystal structure of cellobiohydrolase Cel7A with bound (S)-propranolol ˚ resolution. J. Mol. Biol. 305, 79 – 93. at 1.9 A Meinke, A., Damude, H. G., Tomme, P., Kwan, E., Kilburn, D. G., Miller, R. C., Jr et al. (1995). Enhancement of the endo-beta-1,4-glucanase activity of an exocellobiohydrolase by deletion of a surface loop. J. Biol. Chem. 270, 4383 –4386. Eftink, M. R. (1995). Use of multiple spectroscopic methods to monitor equilibrium unfolding of proteins. Methods Enzymol. 259, 487– 512. Boer, H., Teeri, T. T. & Koivula, A. (2000). Characterization of Trichoderma reesei cellobiohydrolase Cel7A secreted from Pichia pastoris using two different promoters. Biotechnol. Bioeng. 69, 486– 494. Munoz, I. G., Ubhayasekera, W., Henriksson, H., Szabo, I., Pettersson, G., Johansson, G. et al. (2001). Family 7 cellobiohydrolases from Phanerochaete chrysosporium: crystal structure of the catalytic module ˚ resolution and homology of Cel7D (CBH58) at 1.32 A models of the isozymes. J. Mol. Biol. 314, 1097– 1111. Becker, D., Braet, C., Brumer, H., III, Claeyssens, M., Divne, C., Fagerstro¨m, B. R. et al. (2001). Engineering of a glycosidase Family 7 cellobiohydrolase to more alkaline pH optimum: the pH behaviour of Trichoderma reesei Cel7A and its E223S/A224H/L225V/ T226A/D262G mutant. Biochem. J. 356 (1), 19 – 30. Margolles-Clark, E., Hayes, C. K., Harman, G. E. & Penttila¨, M. (1996). Improved production of Trichoderma harzianum endochitinase by expression in Trichoderma reesei. Appl. Environ. Microbiol. 62, 2145– 2151. Mach, R. L., Schindler, M. & Kubicek, C. (1994). Transformation of Trichoderma reesei based on hygromycin B resistance using homologous expressional signals. Curr. Genet. 6, 567– 570. Sambrook, J., Fritsch, E. & Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual, 2nd edit., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Imai, Y., Matsushima, Y., Sugimura, T. & Terada, M. (1991). A simple and rapid method for generating a deletion by PCR. Nucl. Acids Res. 19, 2785.
Loop Engineering in Cel7A
33. Penttila¨, M., Nevalainen, H., Ra¨tto¨, M., Salminen, E. & Knowles, J. (1987). A versatile transformation system for the cellulolytic filamentous fungus Trichoderma reesei. Gene, 61, 155– 164. 34. Aho, S., Olkkonen, V., Jalava, T., Paloheimo, M., Buhler, R., Niku-Paavola, M. L. et al. (1991). Monoclonal antibodies against core and cellulose-binding domains of Trichoderma reesei cellobiohydrolases I and II and endoglucanase I. Eur. J. Biochem. 200, 643–649. 35. Srisodsuk, M., Reinikainen, T., Penttila¨, M. & Teeri, T. T. (1993). Role of the interdomain linker peptide of Trichoderma reesei cellobiohydrolase I in its interaction with crystalline cellulose. J. Biol. Chem. 268, 20756– 20761. 36. Towbin, H., Staehelin, T. & Gordon, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl Acad. Sci. USA, 76, 4350–4354. 37. Uzcategui, E., Ruiz, A., Montesino, R., Johansson, G. & Pettersson, G. (1991). The 1,4-beta-D -glucan cellobiohydrolases from Phanerochaete chrysosporium. I. A system of synergistically acting enzymes homologous to Trichoderma reesei. J. Biotechnol. 19, 271– 285. 38. Szabo´, I. J., Johansson, G. & Pettersson, G. (1996). Optimized cellulase production by Phanerochaete chrysosporium: control of catabolite repression by fed-batch cultivation. J. Biotechnol. 48, 221– 230. 39. Garcia, R., Cremata, J., Quintero, O., Montesino, R., Benkestock, K. & Sta˚hlberg, J. (2001). Characterization of protein glycoforms with N-linked neutral and phosphorylated oligosaccharides: studies on the glycosylation of Endoglucanase 1 (Cel7B) from Trichoderma reesei. Biotechnol. Appl. Biochem. 33, 141–152. 40. McPherson, A. J. (1982). Preparation and Analysis of Protein Crystals, Wiley, New York. 41. Otwinowski, Z. & Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307– 326. 42. Bru¨nger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). new software system for macromolecular structure determination. Acta Crystallog. sect. D, 54, 905– 921. 43. Jones, T. A., Zou, J.-Y., Cowan, S. W. & Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallog. sect. A, 47, 110 – 119.
829
44. Kleywegt, G. & Bru¨nger, A. (1996). Checking your imagination: applications of the free R value. Structure, 4, 897– 904. 45. Reverbel-Leroy, C., Parsiegla, G., Moreau, V., Juy, M., Tardif, C., Driguez, H. et al. (1998). Crystallization of the catalytic domain of Clostridium cellulolyticum CeLF cellulase in the presence of a newly synthesized cellulase inhibitor. Acta Crystallog. sect D, 54, 114 – 118. 46. Rossmann, M. G. & Blow, D. M. (1962). The detection of sub-units within the crystallographic asymmetric unit. Acta Crystallog. 15, 24 – 31. 47. Rossmann, M. G. (1972). The Molecular Replacement Method, pp. 1 – 267 Gordon and Breach, New York. 48. Navaza, J. (1994). AMoRe: an automated package for molecular replacement. Acta Crystallog. sect A, 50, 157 –163. 49. Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H. et al. (2000). The Protein Data Bank. Nucl. Acids Res. 28, 235–242. 50. van Tilbeurgh, H., Claeyssens, M. & De Bruyne, C. K. (1982). The use of 4-methylumbelliferyl and other chromophoric glycosides in the study of cellulolytic enzymes. FEBS Letters, 149, 152– 156. 51. Capon, B. & Thompson, J. W. (1979). Studies on cellulolytic enzymes. I. The use of 3,4-dinitrophenyl glycosides as substrates. Bioorg. Chem. 8, 147– 173. 52. Koivula, A., Reinikainen, T., Ruohonen, L., Valkeaja¨rvi, A., Claeyssens, M., Teleman, O. et al. (1996). The active site of Trichoderma reesei cellobiohydrolase II: the role of tyrosine 169. Protein Eng. 9, 691– 699. 53. Teleman, A., Koivula, A., Reinikainen, T., Valkeaja¨rvi, A., Teeri, T. T., Drakenberg, T. & Teleman, O. (1995). Progress-curve analysis shows that glucose inhibits the cellotriose hydrolysis catalysed by cellobiohydrolase II from Trichoderma reesei. Eur. J. Biochem. 231, 250 –258. 54. Miller, G. L. (1959). Use of dinitrosalisylic acid reagent for determination of reducing sugar. Anal. Chem. 31, 426–428. 55. Gilkes, N. R., Jervis, E., Henrissat, B., Tekant, B., Miller, R. C., Jr, Warren, R. A. & Kilburn, D. G. (1992). The adsorption of a bacterial cellulase and its two isolated domains to crystalline cellulose. J. Biol. Chem. 267, 6743– 6749. 56. Isogai, A. (1990). Amorphous celluloses stable in aqueous media: regeneration from SO2 -amine solvent systems. J. Polym. Sci. A: Polym. Chem. 28, 113 – 119.
Edited by R. Huber (Received 29 June 2003; accepted 7 July 2003)
