By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Due to essential maintenance the subscribe/renew pages will be unavailable on Wednesday 26 October between 02:00- 08:00 BST/ 09:00 – 15:00 SGT/ 21:00- 03:00 EDT. Apologies for the inconvenience.
Xyloglucan endotransglucosylase/hydrolases (XTHs) are cell wall-modifying enzymes that align within three or four distinct phylogenetic subgroups. One explanation for this grouping is association with different enzymic modes of action, as XTHs can have xyloglucan endotransglucosylase (XET) or endohydrolase (XEH) activities. While Group 1 and 2 XTHs predominantly exhibit XET activity, to date the activity of only one member of Group 3 has been reported: nasturtium TmXH1, which has a highly specialized function and hydrolyses seed-storage xyloglucan rather than modifying cell wall structure. Tomato fruit ripening was selected as a model to test the hypothesis that preferential XEH activity might be a defining characteristic of Group 3 XTHs, which would be expressed during processes where net xyloglucan depolymerization occurs. Database searches identified 25 tomato XTHs, and one gene (SlXTH5) was of particular interest as it aligned within Group 3 and was expressed abundantly during ripening. Recombinant SlXTH5 protein acted primarily as a transglucosylase in vitro and depolymerized xyloglucan more rapidly in the presence than in the absence of xyloglucan oligosaccharides (XGOs), indicative of XET activity. Thus, there is no correlation between the XTH phylogenetic grouping and the preferential enzymic activities (XET or XEH) of the proteins in those groups. Similar analyses of SlXTH2, a Group 2 tomato XTH, and nasturtium seed TmXTH1 revealed a spectrum of modes of action, suggesting that all XTHs have the capacity to function in both modes. The biomechanical properties of plant walls were unaffected by incubation with SlXTH5, with or without XGOs, suggesting that XTHs do not represent primary cell wall-loosening agents. The possible roles of SlXTH5 in vivo are discussed.
Xyloglucan is the most abundant hemicellulose in the primary cell walls of non-graminaceous plants, where it coats and cross-links adjacent cellulose microfibrils through non-covalent associations (Bauer et al., 1973; Hayashi, 1989; McCann et al., 1990; Rose and Bennett, 1999). It is still debatable whether these cross-links function principally as ‘tethers’, constraining movement of microfibrils relative to each other, or as ‘scaffolds’ to prevent inter-microfibril adhesion (Chanliaud et al., 2004; Cosgrove, 2005; O'Neill and York, 2003; Thompson, 2005). However, in either scenario, xyloglucan is ascribed a key structural role. Xyloglucan degradation is thus a central element in models of wall modification during transient wall loosening in expanding cells, or the terminal wall degradation that occurs during processes such as fruit ripening and organ abscission (Cosgrove, 2003; Rose and Bennett, 1999; Rose et al., 2003). Accordingly, enzymes associated with xyloglucan metabolism represent a potentially important mechanism for regulating wall strength and extensibility and tissue integrity. Xyloglucan is also used as a storage reserve in the seeds of some plant species, such as nasturtium (Tropaeolum majus), where it accumulates as large deposits on the inside of the cotyledon cell wall during seed development and is subsequently hydrolysed during germination (Buckeridge et al., 2000; Reid, 1985).
XTHs are present throughout the plant kingdom in essentially all cell types, and numerous studies have described their expression during cell growth and differentiation and their regulation by hormones or biotic and abiotic stimuli (Imoto et al., 2005; Lee et al., 2005; Nishitani, 1997; Rose et al., 2002). Monocotyledonous and dicotyledonous species have extensive XTH gene families, with 33 identified in Arabidopsis thaliana (Yokoyama and Nishitani, 2001) and 29 in Oryza sativa (Yokoyama et al., 2004). These can be clustered into three or four main phylogenetic groups or subfamilies (Campbell and Braam, 1999; Rose et al., 2002), although it is still not known whether this grouping has a physiological or biochemical basis. One obvious consideration is that members of a subfamily might have a common physiological function or expression profile. However, this is not supported by comprehensive real-time PCR or microarray analyses (Imoto et al., 2005; Lee et al., 2005; Yokoyama and Nishitani, 2001; Yokoyama et al., 2004) and while general trends can be discerned, such as the association of members of Group 1 with rapid cell elongation, there are often exceptions (Rose et al., 2002). An alternative explanation might be different modes of enzyme action: XTHs have been shown to have two distinct enzymic activities (Fry, 2005; Rose et al., 2002). Most can act as transglucosylases, catalysing the endo-cleavage of a xyloglucan polymer (donor) backbone and subsequent transfer of the newly generated reducing end to the non-reducing terminus of another xyloglucan (acceptor) molecule. Alternatively, some XTHs act preferentially as hydrolases, although some have been reported to perform transglucosylation in vitro in the presence of high acceptor substrate concentrations. These two activities, referred to as XET or xyloglucan endohydrolase (XEH), respectively, are likely to have distinct physiological functions as they could have entirely different effects on wall integrity, depending on the nature of the acceptor molecules (Rose et al., 2002).
Several native or recombinant XTH proteins from Groups 1 and 2 have been studied and reported to have exclusively XET activity in vitro (Rose et al., 2002), although a Group 1 isozyme was described as having both XET and XEH activities (Schröder et al., 1998). The situation is less clear with Group 3, as the only member of this class for which activity has been characterized to date is NXG1 (referred to hereafter as TmXTH1; de Silva et al., 1993), the specialized nasturtium XTH that catalyses the breakdown of seed-storage xyloglucan. TmXTH1 functions principally as a hydrolase, although it has XET activity in vitro at high acceptor substrate concentrations (Fanutti et al., 1993; Farkašet al., 1992). This raises the possibility that the defining characteristic of all Group 3 XTHs might be their primary action as xyloglucan hydrolases; a hypothesis that is addressed in the current study.
Fruit development of tomato (Solanum lycopersicum, formerly Lycopersicon esculentum) provides an excellent model system in which to study XTH function. The cell elongation that accompanies fruit growth involves substantial wall synthesis and reorganization, and high levels of XET activity are found in expanding tomato fruit (Faik et al., 1998; de Silva et al., 1994). Subsequently, during fruit ripening wall polysaccharides undergo irreversible depolymerization and solubilization from the wall, as has been well documented in the case of xyloglucan (Brummell and Harpster, 2001; Cheng and Huber, 1997; Maclachlan and Brady, 1994; Sakurai and Nevins, 1993). The pattern of xyloglucan-degrading enzyme activity in ripening tomato fruit is apparently complex (Maclachlan and Brady, 1994), and may reflect a combination of hydrolases, transglucosylases and/or enzymes with both activities. However, despite their potentially crucial contribution to fruit texture, the genes responsible for these activities have not yet been identified or the proteins purified, although it might be anticipated that an XTH with XEH activity would play a significant role.
To date, the activities of only three tomato XTHs have been assessed (SlXTH1, Cataláet al., 2001; de Silva et al., 1994; SlXTH2; and SlXTH10, Arrowsmith and de Silva, 1995; Chanliaud et al., 2004), and these show XET but no detectable XEH activity. Importantly, these XTHs belong to either Group 1 (SlXTH1) or Group 2 (SlXTH2 and SlXTH10). Indeed, no Group 3 XTH isozyme from the primary walls of any plant species has been characterized. Given the substantial xyloglucan depolymerization that occurs during fruit ripening, we hypothesized the existence of a fruit ripening-related Group 3 XTH with predominant XEH activity. To address this possibility, The Institute for Genomic Research (TIGR) Tomato Gene Index database (http://www.tigr.org) and the SOL Genomics Network (SGN) database (http://www.sgn.cornell.edu) were searched for tomato XTH-related sequences. This paper describes the subsequent characterization of XTH gene expression during tomato fruit development, and reports the activity of a newly identified tomato Group 3 XTH (SlXTH5) that is abundantly expressed in ripening fruit, coincident with xyloglucan depolymerization.
Sequence analysis of tomato XTHs
A search of the TIGR Tomato Gene Index database revealed 14 full-length tomato XTH-related (SlXTH) sequences (SlXTH1-13 and SlXTH23; Table 1), derived from the 31 838 tomato unigene set. These are shown in Table 1 together with their relative abundance in 27 non-normalized, non-subtracted tomato cDNA libraries (van der Hoeven et al., 2002). Six of the 14 XTHs (SlXTH1, SlXTH2, SlXTH4, SlXTH10–12) have already been reported and were renamed with permission from the original authors, according to the recently proposed unified XTH nomenclature (Rose et al., 2002). In addition to the 14 full-length genes, eight XTH-related sequences were deduced from expressed sequence tags (ESTs), the assembly of which did not produce a full-length transcript, and three were represented by only one partial-length EST.
Table 1. Tomato XTHs
Representation in libraries (#ESTs)
agb, Genbank accession no.; TC, The Institute for Genomic Research Assembly number; EST, The Institute for Genomic Research EST identifier; SGN, Solanaceae Genomics Network Unigene build number.
bcDNAs amplified and sequenced by the authors using sequence information from the corresponding TCs in The Institute for Genomic Research database.
Analysis of the predicted SlXTH proteins showed that all the full-length sequences have typical structural features that are conserved among XTHs (Campbell and Braam, 1999), including a predicted N-terminal signal peptide for entry into the secretory pathway, although two (SlXTH13 and SlXTH18) were predicted to be membrane-anchored (http://www.cbs.dtu.dk/services/SignalP and http://psort.nibb.ac.jp). Most also contain the sequence DEIDFEFLG, which is proposed to be the active site and corresponds to the active site motif ExDxE, which is conserved among glycosyl hydrolase family 16 (GH16) enzymes (Campbell and Braam, 1999; Johansson et al., 2004). Most contain a potential N-linked glycosylation site adjacent to the putative active site motif, a feature characteristic of XTHs (Campbell and Braam, 1999). However, the glycosylation motif in SlXTH5, SlXTH8 and SlXTH25 is shifted towards the carboxy terminus and is absent in two (SlXTH6 and SlXTH14). These changes in the position or absence of glycosylation motifs are characteristic of Group 3 XTHs (Campbell and Braam, 1999).
The coding sequences of the 14 full-length tomato XTHs were subjected to phylogenetic analysis together with the XTH family from Arabidopsis, and the resulting tree revealed the presence of four well supported major subgroups (Figure 1). Groups 1 and 2 contained proteins with a higher level of sequence identity, while Group 3 represents a more divergent group of XTHs, as has been reported previously (Yokoyama et al., 2004). Four Arabidopsis XTHs (AtXTH1-3 and AtXTH11) that were previously reported as belonging to Group 1 (Rose et al., 2002; Yokoyama et al., 2004) were placed within a separate clade (Group 4). Depending on the parameters used for these analyses, this fourth group is not always evident; however the other three groups are invariably present (Campbell and Braam, 1999; Rose et al., 2002). Most of the tomato XTHs were associated with Groups 1 and 2, but SlXTH5, SlXTH6 and SlXTH8 aligned within Group 3 and are more closely related to the nasturtium seed TmXTH1 (Figure 1).
Collectively, SlXTHs are expressed across the range of vegetative and reproductive tissues, suspension cells and callus that were used to create the various cDNA libraries (Table 1), and no correlation was apparent between representation in particular libraries and members of a particular phylogenetic subgroup. Some (e.g. SlXTH3 and SlXTH5) showed expression in many organs or tissues, but many were identified in only one library and were represented by a singleton. Eleven (SlXTH1-3, SlXTH5, SlXTH9–12, SlXTH14, SlXTH17 and SlXTH21) were expressed at some stage during fruit development, and EST representation suggested that expression is developmentally regulated. For example, SlXTH1 ESTs were well represented in the immature green (IG) fruit library and not detected in other fruit libraries, while it could be inferred that SlXTH5 is not expressed during fruit expansion, but at high levels during ripening. ESTs from the two previously reported ripening-related XTHs (SlXTH10 and SlXTH11; Arrowsmith and de Silva, 1995) were present in relatively low numbers among the fruit libraries, and only in the breaker stage. Of the fruit-related XTHs that were identified, SlXTH5 was of particular interest as EST abundance not only was high in the breaker and red ripe libraries but, importantly, belonged to Group 3, the target of this study. The other two members of Group 3 (SlXTH6 and SlXTH8) were not present in any of the fruit EST collections.
Expression of tomato XTH genes during fruit ontogeny
While EST representation provides a useful snapshot of gene expression, and has been successfully applied to gain insights into tomato fruit ripening (Fei et al., 2004), it has some limitations. For example, there is typically a bias towards moderately or highly expressed genes and, depending on the tissue or scope of the sequencing initiative, the inventory of ESTs may be incomplete (Fei et al., 2004). Northern blot analysis was therefore used for to provide a higher-resolution expression analysis of the fruit-associated XTHs, as identified in Table 1, and particularly to confirm the ripening-related expression of SlXTH5. Specific cDNA probes were designed to minimize cross-reactivity with multiple XTHs, and Southern blot analyses showed that they had relatively high specificity, generally cross-hybridized with one or two bands per lane, and showed banding patterns that were different from each other (Figure 2). SlXTH10 and SlXTH11 have an extremely high degree of nucleotide sequence identity (90% across the full-length cDNAs) and have short 3′ UTR regions, so it was not possible to generate specific cDNA probes. A cDNA probe was designed based on the SlXTH10 sequence, although it would certainly cross-hybridize with the SlXTH11 paralogue.
Figure 3(a) shows the expression patterns of six fruit-associated tomato XTHs during fruit ontogeny, from the ovary stage through fruit expansion (stages I–III), maturation and ripening. The expression of the fruit ripening-specific expansin LeExp1 (Rose et al., 1997) and tomato actin were also evaluated in the same blots as ripening-related and loading controls, respectively. Expression was similarly evaluated in vegetative tissues, flowers and suspension cultured cells (Figure 3b), where patterns of mRNA accumulation were reasonably similar to the expression patterns indicated in Table 1, such as the relatively high expression of SlXTH10 in roots. The patterns of mRNA accumulation and EST representation in the different fruit stages were also generally similar (Table 1). For example, SlXTH1 was expressed at high levels in expanding fruit, as previously reported (Cataláet al., 2000), and XTHs that were present in the breaker fruit libraries showed corresponding expression on Northern blots. A notable discrepancy was that, while the SlXTH3 transcripts were absent from ovaries, as assessed by Northern analysis (Figure 3a), its representation in the ovary library was by far the highest of any XTH in any of the cDNA libraries (Table 1). This may reflect an artefact of library construction involving heavy bias in favor of some genes, and underlines the value of validating digital expression. Most fruit-related XTHs were expressed in both pre-ripe and ripe stages, and showed more complicated biphasic expression patterns. SlXTH3 and SlXTH9 showed a dramatic reduction in mRNA levels immediately prior to the onset of ripening, around the mature green (MG) stage, and this was seen consistently in three replicate fruit-ripening series. A similar pattern was seen for XTHs at the onset of ripening in grape berries (Nunan et al., 2001) and melon fruit (J.K.C.R., unpublished data). The reason for this rapid dip in an otherwise linear trend of changing expression levels is unclear, but a recent microarray analysis of tomato fruit ripening identified large numbers of genes with similarly dramatic spikes or troughs of expression at the onset of ripening (Alba et al., 2005).
Of particular note for this study was that SlXTH5 was abundantly expressed at the onset of, and throughout, fruit ripening, and expression levels were higher than those of the other ripening-related XTHs. The expression pattern of this Group 3 XTH closely parallels the well characterized depolymerization of xyloglucan in ripening tomato fruit, so SlXTH5 was targeted for analysis of its biochemical activity.
Characterization of SlXTH5 activity
Recombinant SlXTH5 protein was produced by heterologous expression in Pichia pastoris. The native N-terminal signal peptide was used to target expression to the yeast culture medium, and SlXTH5 was purified following sequential anion-exchange and cation-exchange column chromatography (Figure 4). Column fractions were assayed for XET activity as described previously (Cataláet al., 2001). The final fraction (lane 3) showed a single major band that corresponded to the predicted size of the mature SlXTH5 polypeptide (37 kDa) and the identity of the protein was confirmed by peptide mass fingerprinting using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) (data not shown).
The XET activity of purified SlXTH5 was assayed using fluorescently labeled xyloglucan oligosaccharides or XGOs (XLLG; for nomenclature see Fry et al., 1993) and high-Mr tamarind (Tamarindus indica) xyloglucan as acceptor and donor substrates, respectively. Transglucosylation was monitored over a 2-h time course following fractionation on a size-exclusion column that allowed clear distinction between the fluorescently labeled oligosaccharide and polysaccharide (Figure 5a). Labeled polymer (indicated in Figure 5 by a horizontal bar) was detected after 15 min and transglucosylation continued during the 2 h reaction, with a corresponding decrease in the peak height of the oligosaccharide (labeled in Figure 5 with an asterisk), suggesting that SlXTH5 can act as a transglucosylase in vitro. The XET activity of TmXTH1 (a generous gift from Professor W. York) was assayed similarly, but only a very small amount of labeled polymer was detected and no major decrease in XGO peak height was seen, indicating a low level of transglucosylation or rapid degradation of any labeled transglucosylation products. Similar activity levels were observed when tomato xyloglucan was used as a substrate (data not shown), despite the differences in structure (York et al., 1996), so for convenience tamarind xyloglucan was used in these and subsequent assays.
The ability of SlXTH5 and TmXTH1 to depolymerize xyloglucan was assessed using a viscometric assay in the presence or absence of XGOs over a 5-h time course (Figure 6). Addition of 1 μg SlXTH5 resulted in a small reduction in xyloglucan viscosity after 1 h (corresponding to less than a 10% decrease in flow time) and a gradual decrease over the subsequent 4 h (Figure 6a). However, the addition of XGOs induced a rapid decrease in viscosity. In contrast, addition of 1 μg of TmXTH1 resulted in a substantial viscosity decrease and while the presence of XGOs further increased the viscosity loss, this was to a lesser extent than with SlXTH5. When 20 μg of SlXTH5 were used (Figure 6b), the decrease in viscosity was far greater, with an approximately 50% decrease in flow time after 1 h and an 80% reduction in the same time in the presence of XGOs. The activity of recombinant SlXTH2 (20 μg; prepared as described in Cataláet al., 2001), an XTH protein from Group 2 that has been shown to exhibit predominant XET activity (Cataláet al., 2001), was also assayed. While the initial rate of viscosity loss was similar to that of SlXTH5, the final decrease in viscosity was less than that catalysed by SlXTH5.
SlXTH5 effects on cell wall biomechanical properties
The ability of SlXTH5 to affect elasticity, plasticity, stress relaxation and in vitro extensibility (creep) of isolated cucumber hypocotyl cell walls was tested. This wall material has been used in a number of previous studies of the wall-loosening protein expansin and a fungal EGase (Cosgrove, 1989; McQueen-Mason and Cosgrove, 1995; McQueen-Mason et al., 1992; Yuan et al., 2001) and is sensitive to various treatments that alter wall structure. In creep assays (Figure 7), SlXTH5 was unable to induce wall extension in heat-inactivated walls. SlXTH5 incubations were carried out at both pH 5.5 and pH 7.5, where XET enzyme activity is high. Subsequent addition of an α-expansin at pH 4.5 demonstrated that the walls were able to extend in response to added α-expansin and that pre-treatment of the walls with SlXTH5 did not make the walls more sensitive to α-expansin action. Therefore SlXTH5 did not act synergistically with α-expansin to stimulate cell wall creep. Similarly, stress/strain tests of the walls using a custom-built Instron showed that SlXTH5 did not affect elastic or plastic extensibilities (inset in Figure 7). Finally, stress relaxation of isolated walls was unaffected by SlXTH5 (data not shown). The effect of SlXTH5 on wall extensibility in the presence of XGOs (XLLG) was also tested using both stress/strain assays and creep assays, but again no change was observed compared with controls without SlXTH5 (data not shown). It was concluded that SlXTH5 had no detectable effect on the mechanical properties of the wall.
Tomato XTH gene discovery
While the biochemical modes of action and enzymatic properties of several Group 1 and Group 2 XTHs have been described, to date the only equivalent information available for an XTH from Group 3 is from studies of TmXTH1. This enzyme has a highly specialized role in mobilizing storage xyloglucan, rather than primary wall modification and is generally reported as having predominantly XEH, rather than a XET activity (Chanliaud et al., 2004; Fanutti et al., 1993; Farkašet al., 1992; Steele et al., 2001; Sulováet al., 2003). It was hypothesized that preferential XEH activity might be a distinguishing characteristic of all Group 3 XTHs and that, if so, specific Group 3 XTHs would also be expressed in tissues where xyloglucan depolymerization is a major feature.
Tomato fruit ripening represents an attractive system in which to test this idea, as substantial xyloglucan degradation and solubilization occur coincident with fruit softening. Importantly, tomato also has an extensive DNA sequence database, including a large fruit-biased EST collection and an ongoing genome sequencing initiative (http://www.sgn.cornell.edu), and so could provide comprehensive coverage of fruit-related XTH sequences. Moreover, Maclachlan and Brady (1992, 1994) reported a variety of xyloglucanase and XET activities in ripening tomato fruit, although no proteins were identified or sequence information obtained.
A search of the TIGR database identified 25 distinct XTH-related tomato sequences (Table 1), which probably represents most expressed tomato XTHs as this is similar in size to Arabidopsis (33 genes; Yokoyama and Nishitani, 2001) and rice (29 genes; Yokoyama et al., 2004). The tomato XTHs were divided among three phylogenetic groups (Figure 1) and three (SlXTH5, SlXTH6 and SlXTH8) aligned within Group 3. Representation of XTH ESTs within various cDNA libraries from different tomato organs and tissues was used as a preliminary assessment of expression patterns (Table 1). This suggested that 11 XTHs are expressed at various stages of fruit development and the expression patterns of six were further evaluated by Northern blot analysis (Figure 3). SlXTH5, the only fruit-related Group 3 XTH, was the predominant ripening-related family member and was abundantly expressed at the breaker stage and throughout fruit ripening, which is coincident with xyloglucan degradation during fruit softening. The original hypothesis was that Group 3 XTHs catalyse xyloglucan hydrolysis and so recombinant SlXTH5 was generated to characterize its activity.
Modes of XTH action in vitro
Two different approaches were used to compare the activities of SlXTH5 and TmXTH1: monitoring the incorporation of fluorescently labeled XGOs into polymeric unlabeled xyloglucan (Figure 5) and a viscometric assay to measure xyloglucan depolymerization in the presence or absence of XGOs (Figure 6). SlXTH5 clearly exhibited strong XET activity, as evidenced by the accumulation of labeled high Mr xyloglucan after incubation with the labeled XGOs (Figure 5a). Moreover, the reduction in xyloglucan viscosity was far more rapid and extensive in the presence, than in the absence, of XGOs (Figure 6a). This is characteristic of transglucosylation, as stochastic chain cleavage and ligation with small XGO acceptors would cause a rapid decrease in xyloglucan Mr. In contrast, TmXTH1 showed the expected predominant XEH activity, with no substantial accumulation of labeled polymeric xyloglucan (Figure 5b) and minimal differences in xyloglucan depolymerizing activity in the presence or absence of XGOs (Figure 6b). Thus, these two closely related Group 3 XTHs show distinct differences in their modes of action and it can be concluded that predominant XEH activity is not a hallmark of this phylogenetic group.
It is generally thought that XTH activity takes the form of a double-displacement reaction (Fry, 2005; Rose et al., 2002): the enzyme cuts the xyloglucan backbone and forms a highly stable covalently linked enzyme–substrate intermediate and the complex then disassociates when an appropriate acceptor is found, in the form of a xyloglucan polymer, XGO or water. The xyloglucan substrate used in the viscometric assay initially has a low concentration of acceptor chain ends and so, in the absence of XGOs, the early decrease in viscosity results principally from hydrolysis. As the number of non-reducing chain ends increases, the opportunity for transglucosylation increases, outcompeting hydrolysis, and hence the XET : XEH activity ratio also increases (Fanutti et al., 1993). When a high concentration of SlXTH5 was used in the viscometric assay (Figure 6b) and compared with the action of the same concentration of SlXTH2 (a Group 2 XTH), both enzymes appeared to have similar actions in the early stages, during the phase when hydrolysis would predominate and glycosyl acceptors were being generated. However, as the reactions continued and transglucosylation predominated, the final steady-state viscosity levels were significantly different and achieved earlier by SlXTH2. This difference in kinetics is interpreted as reflecting different propensities for the two enzymes to act as hydrolases versus transglucosylases, with SlXTH2 showing a greater tendency to act in ‘transglucosylase mode’.
It seems likely that all XTHs from all phylogenetic groups have both XEH and XET activities, but to substantially different degrees. At one extreme, TmXTH1 has potent hydrolytic activity, although it can perform transglucosylation effectively with XGOs at high concentrations (Fanutti et al., 1993). At the other end of the scale, some XTHs have a marked preference for xyloglucan as an acceptor. It is probably inaccurate to describe any XTH as a ‘pure transglucosylase’, as has been suggested in the literature, as either of the two modes of action can be favored under different reaction conditions. For example, a significant decrease in xyloglucan viscosity was achieved with SlXTH2 only when high concentrations of enzyme were used. Likewise, an XTH from azuki bean epicotyls was described as having no XET activity, but the conditions used were such that transglucosylation might well have been undetectable (Tabuchi et al., 2001). Indeed, a fluorescently labeled XGO incorporation assay (Figure 7 of Tabuchi et al., 2001) appeared to show a very low level of transglucosylation, and the chromatogram is strikingly similar to that resulting from TmXTH1 activity reported here (Figure 5b).
XTH action and wall mechanical properties
As cellulose microfibrils and xyloglucans are thought to form a load-bearing network, one might expect an enzyme such as SlXTH5 to cause wall loosening and extension. However, contrary to this prediction, recombinant SlXTH5 was not able to alter the mechanical properties of the cucumber hypocotyl cell wall, as assayed by three different methods (stress/strain; stress relaxation; creep assays), nor did it act synergistically to enhance wall extension induced by α-expansin (Figure 7). Even the addition of XGOs, which promote rapid SlXTH5-mediated xyloglucan depolymerization in vitro, did not make the wall more extensible.
There is no reason to believe that the cucumber hypocotyl wall used in this analysis is unusual, as it has been used extensively in the study of expansin action and responds similarly to other walls from growing seedlings (Cosgrove, 1989; McQueen-Mason and Cosgrove, 1995; McQueen-Mason et al., 1992). Moreover, treatment of cucumber walls with a xyloglucan-degrading EGase from Trichoderma reesei (Cel12A) caused a large increase in wall elasticity and plasticity, and induced wall extension in creep assays (Yuan et al., 2001). Thus, it is clear that the cucumber wall can become more extensible on sufficient xyloglucan breakdown. These results suggest that SlXTH5 cannot gain access to any key load-bearing xyloglucan strands that are important for cell wall strength and that limit wall extensibility. These strands are evidently inaccessible to SlXTH5, perhaps because they assume a specific conformation in the wall that SlXTH5 cannot recognize, or perhaps because they are buried in a complex that SlXTH5 cannot penetrate. The latter idea is supported by the observation that, even when given at saturating amounts, CEL12A caused wall extension (creep) only after a minimum lag of 6 min (Yuan et al., 2001). Hence, our results are at odds with the idea that XTH enzymes function as primary wall-loosening agents that cause cell wall extension.
In contrast, two published reports indicate that XTH enzymes may have effects on cell wall mechanics in some circumstances. Chanliaud et al. (2004) used artificial cellulose–xyloglucan composites made from Acetobacter xylinus pellicles to examine the mechanical effects of two recombinant XTH enzymes, one with predominantly XET activity and another primarily XEH activity. They observed significant, but distinctive, mechanical effects of the two enzymes: the hydrolase made the artificial composite stiffer (less extensible) and did not cause creep, whereas the reverse effects were observed with the transglucosylase, as creep was increased but stiffness was not affected. The distinctive effects noted by Chanliaud et al. (2004), compared with our current results, highlight the important differences in mechanical and structural properties between plant cell walls and Acetobacter pellicles. In particular, polysaccharide conformation, density and accessibility are likely to be very different in the two materials, and extrapolation from bacterial pellicles to plant cell walls requires caution. A second study (Kaku et al., 2002) reported that an azuki bean epicotyl fraction with XEH activity caused an increase in wall extensibility. However, very long incubation treatments were used (48 h at 37°C), so a concern is that the observed mechanical effects may have resulted from low levels of contamination by other cell wall enzymes. Furthermore, their treatment of the walls would not have inactivated endogenous expansins, so the effects they noted may require the combined action of both proteins.
Modes of SlXTH5 action in vivo
While it is useful to know whether a particular XTH isozyme predominantly exhibits XEH or XET activity in vitro, knowledge of its action in vivo is of prime importance in understanding physiological function. The original hypothesis that SlXTH5 (a Group 3 XTH) might catalyse xyloglucan hydrolysis during fruit softening now seems highly unlikely. However, its expression pattern and potent xyloglucan-degrading activity in the presence of XGOs is reminiscent of the major xyloglucan-depolymerizing activity that was observed in crude protein extracts from ripening tomato fruit (Maclachlan and Brady, 1994). The detection of multiple XTHs from different phylogenetic groups in ripening fruits (Table 1; Figure 3) raises the possibility that another XTH might have significant XEH activity, but the results of Maclachlan and Brady (1994) do not support this idea. Moreover, these data further suggest the possibility of synergistic disassembly of the cellulose–xyloglucan network, and that XGOs may play a critical role in promoting xyloglucan depolymerization. Based on the results of Fry (1986), it is often stated that XGO concentrations in the apoplast are low (approximately 0.4 μm). However, this study was based on the secretion of one XGO type into the medium of suspension-cultured cells, so concentrations in the apoplast of a real plant tissue might be substantially higher, particularly in a localized environment. One model of SlXTH5 synergistic action is the formation of a stable enzyme–xyloglucan donor complex in the walls of ripening fruit, where it would be primed to catalyse transglucosylation and consequent xyloglucan depolymerization on release of XGOs by other xyloglucanolytic enzymes, such as EGases. Current studies include the generation of transgenic tomatoes with altered levels of SlXTH5, to address its contribution to xyloglucan breakdown and fruit softening.
Tomato (S. lycopersicum cv. Ailsa Craig) plants were glasshouse grown under 16-h light, 8-h dark conditions. Pericarp tissue was isolated from fruit harvested at different developmental stages: ovaries, young expanding fruit (stages I–III corresponding to fruit diameters of 0.5–1, 2–3 and 4–6 cm, respectively), immature green (IG), mature green (MG), breaker (Br), turning (Tu), pink (Pi), light red (LR) and red ripe (RR). Ripening fruit were staged as described by Lashbrook et al. (1994). To characterize the rate of tomato fruit growth, flowers were tagged at anthesis and fruit diameters measured over a period of 45 days. Vegetative tissues and flowers were collected from the same plants, with the exception of roots and etiolated hypocotyls, which were growth in vermiculite under dark conditions. Suspension-cultured tomato cells were as described by York et al. (1996). All tissues were frozen immediately after collection in liquid nitrogen and stored at –80°C.
Database search and sequence analysis
The Tomato Gene Index database (release 10.1, comprising 162 621 tomato ESTs corresponding to 31 838 unique genes) at TIGR was searched for sequences with homology to conserved XTH domains. The retrieved sequences are listed in Table 1 with the identifier of the EST assembly as a virtual transcript (tentative consensus, TC number), or the EST identifier in the cases of single ESTs (singletons) that are not contained in a gene assembly. XTH sequences were also found in the SGN database in the form of unigene builds from tomato ESTs, and the identifier of the transcript assembly in SGN is also given in Table 1. The longest, most complete sequence (from TIGR or SGN) was used to perform sequence alignments or to design probes for Southern and Northern blot analyses. In some cases, the same XTH sequence with a few base-pair changes was represented by more than one TC or SGN assembly, perhaps reflecting sequencing errors or paralogous genes. Some XTH sequences (SlXTH3, SlXTH5, SlXTH6, SlXTH7 and SlXTH9) were amplified by RT–PCR as described below, using primers designed from the corresponding TC sequences and resequenced to clarify discrepancies between the TC and SGN sequences, or to confirm database sequences. Corrected sequences of the full length transcripts were submitted to Genbank (Table 1) and are listed at http://labs.plantbio.cornell.edu/XTH.
Cloning of tomato XTHs
Total RNA (2 μg) from either small green or red ripe tomato fruit was subjected to reverse transcription using SuperScript II RNase H− reverse transcriptase (Life Technologies, Rockville, MD, USA) and random hexamers, as described by Rose et al. (1996). The resulting cDNA was used as a template for PCR amplification (35 cycles of 95°C for 30 sec, 45°C for 30 sec and 72°C for 2 min) with ProofStart DNA polymerase (Qiagen, Valencia, CA, USA) in the presence of gene-specific primer pairs. The PCR products were purified by gel electrophoresis, cloned into the pSTBlue1 vector (Novagen, Madison, WI, USA) and sequenced.
3′ RACE (Frohman et al., 1988) was used to amplify a fragment containing the 3′ UTR of SlXTH5. cDNA was generated from 2 μg total RNA from red ripe fruit using M-MuLV reverse transcriptase (Promega, Madison, WI, USA) and the oligonucleotide primer 5′-ACTATAGGGCACGCGTGG-(dT)17-3′ (AP2). After 1 h incubation at 37°C, the reaction was subjected to PCR (as above) with Pfu DNA polymerase (Stratagene, La Jolla, CA, USA) using 5′-CAACATTTTAAAGAGACTGGAAG-3′ and the complementary sequence of AP2 as the 5′- and 3′-end PCR primers, respectively. After 35 PCR cycles, a 385-bp band was amplified, subcloned and sequenced.
All available full-length XTH protein sequences from tomato (14) and Arabidopsis (33), as well as TmXTH1 (formerly NXG1; de Silva et al., 1993) from nasturtium, were aligned using the ClustalX algorithm (pairwise alignment with a gap opening penalty of 35 and a gap extension penalty of 0.75; multiple alignment parameters with a gap opening penalty of 15, a gap extension penalty of 0.3 and delay divergent sequences set at 25%) with Gonnet residue weights (Thompson et al., 1997). A neighbor-joining tree (Saitou and Nei, 1987) was constructed based on the alignment and a bootstrap analysis with 2000 replicates was used to assess the statistical reliability of the tree topology. The consensus tree was drawn using the treeview program (Page, 1996).
DNA gel-blot analysis
Genomic DNA was extracted from young tomato leaves as described by Dellaporta et al. (1983). Aliquots (10 μg) were digested with the appropriate restriction enzymes (EcoRI, EcoRV, HindIII and BamHI), fractionated on 0.8% (w/v) agarose gels and transferred to Hybond-N membranes (Amersham Biosciences, Piscataway, NJ, USA). The membranes were hybridized at 42°C in 50% formamide, 6 × SSPE, 0.5% (w/v) SDS, 5 × Denhardt's solution and 100 mg ml−1 sheared salmon sperm DNA, with radiolabeled cDNA probes corresponding to specific XTH genes. The cDNA probes for SlXTH1 and SlXTH2 were as described previously (Cataláet al., 1997, 2001, respectively) cDNAs corresponding to gene-specific regions of other tomato fruit-related XTHs were amplified by PCR using the conditions specified above (for PCR primers see Table S1). 32P-labeled cDNA probes were synthesized with the Ready-To-Go DNA Labeling Beads (−dCTP) Kit (Amersham Biosciences) and purified with ProbeQuant G-50 Micro Columns (Amersham Biosciences). Following hybridization, the membranes were washed three times in 5 × SSC, 1% (w/v) SDS at 42°C for 15 min, followed by three washes in 0.2 × SSC 0.5% (w/v) SDS at 65°C for 20 min, then exposed to film.
RNA gel-blot analysis
Total RNA was extracted from tomato fruit pericarp, vegetative tissues and flowers as described by Wan and Wilkins (1994). RNA from ovary, etiolated hypocotyls and roots was isolated using the RNeasy Plant Total RNA kit (Qiagen) according to the manufacturer's instructions. Poly(A) RNA was isolated from 15 μg total RNA from each fruit stage using the Oligotex mRNA kit (Qiagen). Poly(A) RNA (1 μg from fruit tissues) or total RNA (15 μg from vegetative tissues, suspension cells and flowers) was subjected to electrophoresis on 1.2% (w/v) agarose, 10% (w/v) formaldehyde gels and transferred to Hybond-N membranes (Amersham Biosciences). Blots were hybridized with the gene-specific probes and washed as described above, except for the high-stringency washes at 65°C in 0.5 × SSC, 0.5% (w/v) SDS and exposed to film. Hybridization to a tomato actin cDNA fragment was used as a loading control of RNA samples from different fruit development stages, as described by Cataláet al. (2001). Hybridization to tomato expansin LeExp1 cDNA fragment (Rose et al., 1997) was used as a control gene expressed during ripening.
Recombinant SlXTH5 production and purification
Recombinant SlXTH5 was generated using the P. pastoris expression system (Invitrogen, Carlsbad, CA, USA). The SlXTH5 coding sequence, including the native putative N-terminal signal peptide and stop codon, was amplified by PCR using the primers 5′-AGGATCTGGATCCCATCAGAATGA-3′ and 5′-GTAAACAAAATCTAGAATCACTCT-3′ to introduce BamHI and XbaI restriction sites at the 5′ and 3′ termini, respectively, with SlXTH5 cDNA as a template. After digestion with BamHI and XbaI, the PCR product was cloned into the pPIC3.5K expression vector (Invitrogen). The SlXTH5 expression vector was linearized with SacI and transformed into P. pastoris (strain GS115) according to the Invitrogen Expression System Manual. Recombinant yeast colonies were used to inoculate 10 ml buffered glycerol complex medium (Invitrogen) in 50-ml plastic tubes. After overnight growth at 30°C, 250 r.p.m., 5 ml culture were used to inoculate 500 ml of the same media in 2.8-l flasks and the cultures were shaken at 30°C, 250 r.p.m. until a culture OD600 2–6 was reached. The yeast cells and culture media were separated by centrifugation at 2500 g for 5 min and the pelleted cells resuspended in 50 ml buffered methanol complex medium (Invitrogen) in 1-l flasks and shaken at 16°C, 250 r.p.m. for 3 days. Methanol was added to the cultures every 24 h to a final concentration of 0.5% (w/v).
Culture medium (200 ml) of P. pastoris cells expressing SlXTH5 was centrifuged at 3000 g for 10 min, and the supernatant was used as a source of the recombinant protein. The proteins in the supernatant were precipitated with ammonium sulfate (85% cut) and resuspended in 20 mm Tris–HCl buffer pH 8.5. After desalting using a PD-10 column (Amersham Biosciences), the sample was loaded onto a UNO Q column (Bio-Rad, Hercules, CA, USA) equilibrated with 20 mm Tris–HCl buffer pH 8.5. The flowthrough was collected, desalted and concentrated using an Amicon Ultra 30 centrifugal concentrator (Millipore, Billerica, MA, USA) and applied to a UNO S polishing column (Bio-Rad). The SlXTH5 protein was eluted with a 0 to 0.5 m NaCl gradient in 20 mm MES buffer pH 5.9. Fractions containing SlXTH5 protein were pooled and concentrated as described above. Protein purification was monitored by electrophoresis using 4–12% acrylamide (w/v) NuPAGE gels (Invitrogen) and staining with Coomassie Blue. Protein concentration was determined using the Bio-Rad protein assay.
XET and XEH activity assays
XET activity was assayed by measuring the transfer of non-labeled tamarind seed xyloglucan polymer (Megazyme, Bray, Ireland) to a fluorescently labeled XLLG-APTS acceptor molecule as described by Cataláet al. (2001). Reaction mixtures were analysed by gel-filtration chromatography on a Zorbax PSM 60 column (Agilent Technologies, Palo Alto, CA, USA) eluted with 20 mm sodium acetate buffer pH 5.2, at a flow rate of 1 ml min−1. Fluorescent xyloglucan-APTS derivatives were detected in the eluate using a spectrofluorometer (excitation, 424 nm; emission, 504 nm), as described by Cataláet al. (2001). Comparative studies were also performed with tomato xyloglucan from suspension-cultured cells, cv. Bonnie Best (isolated as described by York et al., 1996).
XEH activity was assayed by measuring changes in viscometric flow time (Edwards et al., 1985) of xyloglucan solutions following mixing with enzyme extracts, in a Cannon–Manning (State College, PA, USA) semimicro viscometer at 28°C. The reaction mixture contained 0.5 ml tamarind seed xyloglucan solution (5 mg ml−1) and 100 μl enzyme solution, both in 50 mm potassium phosphate buffer pH 6.0. A 4-mg ml−1 xyloglucan non asaccharide solution (XLLG, 10 μl; for nomenclature see Fry et al., 1993) were included in some reaction mixtures.
A partially purified preparation of TmXTH1 (a generous donation from Professor W. York, Complex Carbohydrate Research Center, University of Georgia, USA) was used in some of the XET and XEH assays.
Wall mechanical assays
Cucumber hypocotyl walls were prepared as described previously (Cosgrove, 1989; McQueen-Mason et al., 1992). Briefly, apical 1-cm hypocotyl segments from 3-day-old etiolated seedlings were stored frozen (−20°C) and subsequently thawed, abraded with carborundum, pressed to remove cell sap and inactivated with a 15-sec dip in boiling water.
For stress–strain analysis, 12 wall specimens were pre-incubated with 0.5 ml, 50 mm buffer (sodium acetate pH 5.5 or HEPES pH 7.5) with or without 60 μl purified recombinant SlXTH5 enzyme (60 ng μl−1) for 120 min at 30°C, with gentle agitation, then stored on ice. Wall specimens were clamped in a custom-built extensometer (Cosgrove, 1989) and extended in two cycles at 3 mm min−1 up to a maximum load of 20 g. Plastic and elastic compliances were calculated from the last part of the extension curves using second-order polynomial fitting, as described previously (Cosgrove, 1989). For stress-relaxation analyses, wall specimens were prepared as above and, following enzyme incubation for 120 min at 30°C, were stored on ice, then clamped in an extensometer and rapidly extended until a holding load of 25 g was reached, after which time the clamps were held steady and the decay in the holding force was recorded for 5 min. Typically 10–12 replicates were used for each treatment. Stress-relaxation spectra were obtained by calculating the rate of decay in the holding force on a log-time scale (Cosgrove, 1989). For creep (wall extension) analysis, wall specimens were clamped in an extensometer at a constant load of 20 g. Purified recombinant SlXTH5 (20 μl, 60 ng μl−1) was added to each cuvette containing approximately 180 μl buffer (pH 5.5, 50 mm sodium acetate or MES; or pH 7.5, 50 mm HEPES). For subsequent assays with expansins, the SlXTH5-containing buffer was replaced with 160 μl 50 mm sodium acetate pH 4.5, plus 40 μl α-expansin extract (ammonium sulfate-precipitated wall protein from cucumber hypocotyls containing 0.4 mg total protein, prepared as described by McQueen-Mason et al., 1992).
We thank Dan Durachko for technical assistance with the wall mechanical assays and Professor W. York (CCRC, University of Georgia) for providing xyloglucan oligosaccharides, TmXTH1 protein and useful discussion. We also thank colleagues for their permission to rename specific XTHs according to the new standardized nomenclature. Research support was provided to D.J.C. by DOE Grant DE-FG02-84ER13179 and to J.K.C.R. by the NSF award DBI-0431335, the CUAES Hatch Project, NYC-184485 and a grant from the New York State Office of Science, Technology and Academic Research (NYSTAR).