• The hydrolytic plant enzymes of family 32 of glycoside hydrolases (GH32), including acid cell wall type invertases (EC 220.127.116.11), fructan 1-exohydrolases (1-FEH; EC 18.104.22.168) and fructan 6-exohydrolases (6-FEH; EC 22.214.171.124), are very similar at the molecular and structural levels, but are clearly functionally different. The work presented here aims at understanding the evolution of enzyme specificity and functional diversity in this family by means of site-directed mutagenesis.
• It is demonstrated for the first time that invertase activity can be introduced in an S101L mutant of chicory (Cichorium intybus) 1-FEH IIa by influencing the orientation of Trp 82. At high sucrose and enzyme concentrations, a shift is proposed from a stable inhibitor configuration to an unstable substrate configuration.
• In the same way, invertase activity was introduced in Beta vulgaris 6-FEH by introducing an acidic amino acid in the vicinity of the acid–base catalyst (F233D mutant), creating a β-fructofuranosidase type of enzyme with dual activity against sucrose and levan.
• As single amino acid substitutions can influence the donor substrate specificity of FEHs, it is predicted that plant invertases and FEHs may have diversified by introduction of a very limited number of mutations in the common ancestor.
Polysaccharides are important biomolecules in plants, playing a role as reserve carbohydrate or as a structural component. The so-called glucans, polymers of glucose, are the most common and best studied plant polysaccharides; cellulose is the main component of the plant cell wall, while starch is the most important storage carbohydrate. However, not only glucans, also fructans, soluble sucrose-based polymers of fructose, are widespread alternative reserve carbohydrates. Fructans have been reported in various prokaryotes, fungi and algae, and accumulate in 15% of the flowering plants, including the economically important orders Poales and Asterales (Hendry & Wallace, 1993).
In contrast to the uniform structure of microbial fructans, plant fructans show more structural diversity and are biosynthesized by different types of fructosyltransferases (Shiomi, 1989; Vijn & Smeekens, 1999; Lasseur et al., 2006). Fructans are synthesized by an initial fructosyl transfer to one of the primary hydroxyl groups of sucrose, followed by further chain elongation (Van Laere & Van den Ende, 2002). Different types of fructan molecules can be distinguished depending on the linkage type between the fructosyl residues (Lewis, 1993). Fructans in plants may be classified into two categories according to the position of the glucosyl residue. Fructans with a terminal glucosyl residue include the β(2-1) type fructans (inulin, principally occurring in dicots), and the linear β(2-6) (levan) or branched (graminan) β(2-6) type fructans (as occurring in bacteria and monocots). Fructans with an internal glucosyl residue include the neo-inulin and neo-levan type (as occurring in Lolium perenne). In contrast to microbial fructans, the degradation of plant fructans appears to be catalysed exclusively by fructan exohydrolases (FEHs) releasing one terminal fructose molecule at a time using water as acceptor. Until now, no plant fructan endohydrolases have been reported.
Plant FEHs are more closely related to cell wall invertases than to vacuolar invertases (Van den Ende et al., 2000). Cell wall invertases are key metabolic enzymes involved in the regulation of sucrose partitioning, developmental processes such as pollen and seed development and the response to wounding and pathogen infection (Koch, 1996). Plant FEHs can be classified into 1-FEHs (inulinases, EC 126.96.36.199) and 6-FEHs (levanases, EC 188.8.131.52), according to the linkage type they hydrolyse. Some forms are able to degrade both β(2-1) and β(2-6) linkages and are termed 6&1-FEHs (De Coninck et al., 2005; Kawakami et al., 2005). Rapid fructan breakdown by FEHs is widely recognized to be an essential process for sugar mobilization in fructan plants whenever carbon supply is needed by sink organs (Morvan-Bertrand et al., 2001; Asega and Carvalho, 2004). Different types of these FEHs have been fully characterized and cloned (Van den Ende et al., 2000, 2001, 2003a; Van Riet et al., 2006; Lothier et al., 2007). Interestingly, FEH activities were also detected in plants that do not accumulate fructans. Specific FEHs (lacking invertase activity) were cloned and/or purified from Beta vulgaris (6-FEH) and Arabidopsis thaliana (6-FEH and 6&1-FEH) (Van den Ende et al., 2003b; De Coninck et al., 2005). The question why FEHs are present in plants that do not accumulate fructan substrates is still unanswered. It remains to be elucidated whether FEHs might be involved in signalling or other regulatory processes (De Coninck et al., 2005).
Plant invertases preferentially hydrolyse the α1,2 Glc-Fru linkage in sucrose, while plant FEHs have no invertase activity and only split terminal β Fru-Fru linkages in fructans. The term ‘β-fructofuranosidase’ is used for enzymes that can hydrolyse both sucrose and fructans. Together with fructan biosynthetic enzymes, all these kinds of enzymes group together in the glycoside hydrolase family 32 (GH32) in the carbohydrate active enzymes database (Henrissat & Davies, 1997; http://www.cazy.org). Family GH32 is combined with the related family GH68, harbouring bacterial invertases, levansucrases and inulosucrases, in the clan GH-J (Naumoff, 2001).
As both plant invertases and FEHs are believed to fulfil crucial functions during plant growth and development, our research focuses on fully explaining the difference in substrate specificity between these very related enzymes. This research is facilitated by elucidation of the three-dimensional structures of Cichorium intybus 1-FEH IIa (Verhaest et al., 2005, PDB code 1ST8) and A. thaliana cell wall invertase 1 (Verhaest et al., 2006, PDB code 2AC1). Both enzymes consist of an N-terminal five-fold β-propeller domain and a C-terminal domain formed by two β-sheets. A similar fivefold β-propeller domain is found in all other protein structures of the GH-J clan, as summarized by Verhaest et al. (2007). The active site is located in the β-propeller domain and is composed of three highly conserved acidic amino acids, one glutamate and two aspartate residues, acting as nucleophile, acid–base catalyst and transition-state stabilizer, respectively (Reddy & Maley, 1990, 1996; Fig. 1).
The three-dimensional structure of chicory 1-FEH IIa in complex with its inhibitor sucrose was determined (Verhaest et al., 2007). These authors demonstrated that the binding of sucrose in the ‘inhibitor configuration’ in chicory 1-FEH IIa (in which the glucose moiety shows another orientation as the one observed in the levansucrase–sucrose complex) is determined by the specific orientation of a single amino acid, Trp82 of the highly conserved WSGSAT region. Furthermore, recent site-directed mutagenesis studies showed that a D239A mutant of a cell wall invertase of A. thaliana lost its capacity to degrade sucrose efficiently, while its intrinsic FEH activity appeared unaffected (Le Roy et al., 2007b), leading to the hypothesis that the FEHs occurring in A. thaliana might be considered as ‘defective’ invertases. To test this hypothesis further, we were challenged to (re)build invertase activity in FEHs, a process expected to be much more difficult than destroying invertase activity. The experiments were performed on two different types of FEH: chicory 1-FEH IIa strongly binding sucrose as an inhibitor; and B. vulgaris 6-FEH, which is not inhibited by sucrose.
Materials and Methods
Cloning and site-directed mutagenesis
Cichorium intybus L. 1-FEH IIa (Ci1-FEHIIa) and B. vulgaris 6-FEH (Bv6-FEH) were cloned into the pPICZαA vector as described by Verhaest et al. (2004); Van den Ende et al. (2003b). Single amino acid substitutions were generated following the Quick Change protocol (Stratagene, La Jolla, CA, USA), using the pPICZαA-Ci1-FEHIIa and pPICZαA-Bv6-FEH constructs as a template. For site-directed mutagenesis, the following forward oligonucleotide primers (and complementary reverse primers) were used: 5′-CCTTAAATTAAGTTTGGACG ATACTCAATATGAG-3′ (Bv6-FEH/F233D) and 5′-GTA CACCGGCCTCGATTCAAG-3′ (Ci1-FEHIIa/S101L). The mutations are illustrated in bold. Further handlings were as described by Le Roy et al. (2007a).
Heterologous expression and purification
The methylotrophic yeast Pichia pastoris was used for extracellular gene expression as described by De Coninck et al. (2005). Purification of the recombinant Ci1-FEHIIa and Bv6-FEH wild-type and mutant proteins was as described by Le Roy et al. (2007a). Enzyme concentrations were measured using the Bradford method with bovine serum albumin as a standard (Sedmak & Grossberg, 1977).
Appropriate aliquots of purified enzyme were mixed with different substrates (with a final concentration ranging from 1.0 mm to 1.0 m) in 50-mm sodium acetate buffer pH 5.0 also containing 0.02% (w/v) sodium azide. Chicory inulin (mean DP 30, Sigma-Aldrich, St Louis, MO, USA), 1-kestose (TCI Europe nv, Zwijndrecht, Belgium) and sucrose are commercially available substrates, bacterial levan (mean DP 72) and 6-kestose were generous gifts from Dr Iizuka (Iizuka et al., 1993). Reaction mixtures were incubated at 30°C. For the reactions with sucrose, time points were 30 min, 1 and 2 h. For the other substrates, time points were 10, 20 and 30 min, respectively. Total enzyme activity was determined by measuring the amount of released fructose by high-pressure anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD, Dionex, Sunnyvale, CA, USA) (Van den Ende & Van Laere, 1996). Only data from the linear range were used and experiments were repeated three times with consistent results.
Results and Discussion
Sucrose binding in FEHs: inhibitor vs substrate configuration
The three-dimensional structure of chicory 1-FEH IIa in complex with its inhibitor sucrose revealed that sucrose binds in a different configuration compared with the substrate configuration, as observed in B. subtilis levansucrase (Meng & Fütterer, 2003; PDB code 1OYG). The terminal fructosyl unit is positioned in a very similar way at the −1 subsite (Fig. 1), but the configuration of the glucosyl moiety of sucrose bound in chicory 1-FEH IIa at the +1 subsite is clearly different. Remarkably, Trp82 occupies a different orientation in 1-FEH IIa (Fig. 1c) compared with its structural homologues in B. subtilis levansucrase (Trp163; Fig. 1a) and AtcwINV1 (Trp82; Fig. 1b). Recently, site-directed mutagenesis studies clearly demonstrated that this specific orientation of Trp82 in chicory 1-FEH IIa is caused by the presence of the small, nonstacking Ser101 residue, allowing sucrose to bind in an alternative inhibitor configuration in the active site. Sucrose cannot be hydrolysed when it binds in the inhibitor configuration (Verhaest et al., 2007). Strikingly, most FEHs that are (very) weakly inhibited by sucrose are characterized by the presence of a hydrophobic residue (leucine, isoleucine, valine) at this Ser101 position, stacking with the Trp82 homologue and probably preventing sucrose from binding as an inhibitor in their active sites. By contrast, the FEHs that are strongly inhibited by sucrose contain a small glycine or serine residue, probably resulting in a different orientation of their Trp82 homologues (Fig. 2).
It is proposed that the orientation of the bulky Trp82 might determine the sucrose configuration in the active site of FEHs (Verhaest et al., 2007). Therefore we tried to change the orientation of Trp82 by mutating its neighbouring amino acid serine into a larger apolar amino acid (leucine) in order to prevent sucrose from binding in the inhibitor configuration. Would the change in Trp82 orientation result in a shift from the inhibitor to the substrate configuration? To resolve this tempting issue, the sucrose-hydrolysing activities of the 1-FEH IIa wild-type and S101L mutant enzymes were compared. Fig. 3a clearly demonstrates that a S101L mutant is characterized by strongly increased sucrose-hydrolysing activity compared with the wild-type enzyme. However, it should be noted that high amounts of sucrose and enzyme were necessary to observe this invertase activity. At a very low substrate concentration (5 mm), the introduced sucrose-hydrolysing activity should be considered as small compared with the activities towards the preferential 1-FEH IIa substrates 1-kestose and inulin (Fig. 3b). It was demonstrated previously that the S101L mutant cannot bind sucrose in the inhibitor configuration (Verhaest et al., 2007). Here we demonstrate that sucrose can bind in the substrate configuration, although in a rather inefficient way. Taken together, the data show that sucrose can bind in two distinct configurations: a stable inhibitor configuration (Verhaest et al., 2007) and an unstable substrate configuration (this work). No (or extremely low) catalysis occurs in the stable inhibitor configuration (all sucrose inhibited FEHs), while only very limited catalysis is observed in the unstable substrate configuration. It has been postulated that, when bound in the inhibitor configuration, sucrose cannot be hydrolysed because of the presence of an H-linkage between the acid–base catalyst and the differentially oriented glucosyl moiety of sucrose, hindering the proton donor in its proton donation to the glycosidic oxygen (Verhaest et al., 2007). It is an intriguing question how the unstable substrate configuration (FEHs) can be further converted into a real substrate configuration, as observed in typical plant invertases showing much higher specific activities towards sucrose. For optimal sucrose binding in the substrate configuration and subsequent hydrolysis, it can be assumed that an adapted active site is required, containing some additional and/or specific amino acids. Amino acid sequence comparisons, together with examination of the three-dimensional protein structures of both types of enzyme, could help to elucidate why sucrose is efficiently hydrolysed by invertases and not (or in a very limited way) by FEHs.
Introduction of sucrose-hydrolysing activity in B. vulgaris 6-FEH
A closer look at the three-dimensional structures of Arabidopsis cell wall invertase 1 and chicory 1-FEH IIa reveals some marked differences in their active-site regions that might be responsible for the relative (in)capability to bind and hydrolyse sucrose. In AtcwINV1, an additional acidic amino acid Asp239 is observed in the vicinity of the conserved acid–base catalyst Glu203 (Fig. 4a). By contrast, the Asp239 homologue (Glu234) in chicory 1-FEH IIa is twisted over 180° and oriented away from the active site (Fig. 4b). This different orientation of Glu234 is probably caused by the presence of a double deletion in the vicinity of this residue (Fig. 5). Strikingly, in all known FEHs, such an acidic residue is absent or flanked by a single or double deletion. By contrast, the presence of an Asp239 homologue is highly conserved among cell wall invertases (Fig. 5). It can be questioned whether the introduction of an Asp239 homologue in an FEH would lead to increased invertase activity. As sucrose binds in the inhibitor configuration in chicory 1-FEH IIa and, more importantly, the introduction of a structural and functional Asp239 homologue would be complicated due to the presence of a double deletion in this region, we decided to perform site-directed mutagenesis experiments on B. vulgaris 6-FEH. In this enzyme, sucrose does not bind in the inhibitor configuration (Van den Ende et al, 2003b), and contains no deletions in the vicinity of its Asp239 homologue (Fig. 5). Modelling studies showed that a Phe233Asp substitution in B. vulgaris 6-FEH results in the introduction of a well positioned structural and functional Asp239 homologue (Fig. 4c).
A F233D mutant was constructed and the sucrose-hydrolysing activities of both wild-type and F233D mutant 6-FEH proteins were tested by incubation with increasing amounts of sucrose (up to 1.0 m final concentration). Fig. 6a shows that the introduction of the aspartate clearly results in a strongly increased sucrose-hydrolysing activity. This result clearly demonstrates for the first time that the introduction of an extra acidic amino acid in the vicinity of the acid–base catalyst can effectively stimulate sucrose binding and catalysis. When compared with the activities against its preferential substrates levan and 6-kestose, an important shift in substrate specificity could be observed for the F233D mutant (Fig. 6b). The F233D mutant can be considered as a β-fructofuranosidase, showing activities against both sucrose and levan. Such enzymes are well known in microbes, but the existence of real β-fructofuranosidase enzymes in plants is still a matter of debate (Van den Ende et al., 2004). Interestingly, in addition to an introduced sucrose-hydrolysing activity, the F233D mutant enzyme is characterized by a general decrease in activity against levan and 6-kestose, suggesting that Phe233 might play an important role in the overall activity of B. vulgaris 6-FEH. It should be noticed that, because of its hydrophobic and aromatic character, this residue could be crucial to stabilize 6-kestose and higher DP levan substrate in the active site. It is well known that aromatic residues can play an important role by stabilizing the sugar–enzyme complex (Nagy et al., 1998; Van der Veen et al., 2001).
Although the introduction of an additional acidic residue in the active site of B. vulgaris 6-FEH caused a significant increase in sucrose-hydrolysing activity in the F233D mutant, the preference for levan and 6-kestose as a substrate remained. This suggests that still some other amino acids might play a crucial role in the optimal binding and hydrolysis of sucrose. Recent structure–function work on AtcwINV1 shows an important role of the Trp47 residue in stabilizing sucrose in the active site (Le Roy et al., 2007b). The presence of a Trp47 homologue is highly conserved among cell wall invertases, in contrast to plant FEHs (Fig. 5). Interestingly, a Trp residue is also present in B. vulgaris 6-FEH, but modelling studies show that this Trp45 might be oriented away from the active site due to a tight stacking interaction with His46 (Fig. 4c). This could reasonably explain the observed suboptimal binding of sucrose in the active site of the F233D mutant. It should also be noticed that 1-FEH IIa contains a well oriented Phe46 homologue at this position, which could assist in stabilizing sucrose in the S101L 1-FEH IIa mutant, even in the absence of an Asp239 homologue (Fig. 4b).
Functional diversification and evolution of enzyme specificity
Taken together, the data presented by Le Roy et al. (2007b) and in this manuscript greatly support the idea that only a few amino acids determine the difference in substrate specificity between FEHs and invertases. It can be speculated that plant FEHs and invertases, each with their characteristic substrate specificities, were derived from an ancestral β-fructofuranosidase type of enzyme, as still present in microorganisms today. One of our mutant enzymes (F233D) behaved as a typical β-fructofuranosidase, but it is unclear whether such an ‘evolutionary intermediate’ exists in plants today. Whatever the exact process of enzyme speciation throughout evolution, apparently the uncoupling of sucrose and fructan metabolism turned out to be highly beneficial in plants, and in this respect the most important characteristic of FEHs in nonfructan plants might be their inability to degrade sucrose, putatively allowing them to fulfil a regulatory function in a high-sucrose background. It should be noted that the mutations leading to enzyme speciation might be situated not only in the active site region (e.g. removal of Asp239 homologue to create a FEH), but probably also in the cavity between the two structural domains in family 32 members. It was recently demonstrated by site-directed mutagenesis on FEHs that this cleft plays an important role in stabilizing higher DP fructans as a donor substrate in this cleft (Le Roy et al., 2007a).
For the first time, invertase activity was successfully introduced in two different FEH enzymes by specific site-directed mutagenesis. Our results strongly suggest that a correct orientation of Trp82 (avoiding binding of sucrose in the inhibitor configuration), and the presence of an Asp 239 homologue, are two essential requirements for stabilizing sucrose in the active site, although some additional amino acids (e.g. Trp45) are probably necessary to achieve optimal binding and catalysis of sucrose. Especially the absence or presence of an Asp239 homologue appears to be particularly important in understanding the functional difference between FEHs and invertases. This point of view is consistent with the fact that all functionally characterized FEHs lack an Asp239 homologue and are unable to degrade sucrose. On the other hand, an Asp239 homologue is found in all plant cell wall invertases, although few of them are functionally characterized. More research is needed to further test this hypothesis on a wider array of GH32 members, including microbial enzymes. As only a few mutations can affect donor substrate specificity, our results suggest that plant invertases in FEHs may have diversified by introduction of a very limited number of mutations in the common ancestor.
Wim Van den Ende is supported by a grant from the Fund for Scientific Research (FSR Flanders). The authors are very grateful to Rudy Vergauwen for his technical assistance.