One major component of plant cell walls is a diverse group of polysaccharides, the hemicelluloses. Hemicelluloses constitute roughly one-third of the wall biomass and encompass the heteromannans, xyloglucan, heteroxylans, and mixed-linkage glucan. The fine structure of these polysaccharides, particularly their substitution, varies depending on the plant species and tissue type. The hemicelluloses are used in numerous industrial applications such as food additives as well as in medicinal applications. Their abundance in lignocellulosic feedstocks should not be overlooked, if the utilization of this renewable resource for fuels and other commodity chemicals becomes a reality. Fortunately, our understanding of the biosynthesis of the various hemicelluloses in the plant has increased enormously in recent years mainly through genetic approaches. Taking advantage of this knowledge has led to plant mutants with altered hemicellulosic structures demonstrating the importance of the hemicelluloses in plant growth and development. However, while we are on a solid trajectory in identifying all necessary genes/proteins involved in hemicellulose biosynthesis, future research is required to combine these single components and assemble them to gain a holistic mechanistic understanding of the biosynthesis of this important class of plant cell wall polysaccharides.
KeywordsCell walls Polysaccharide Glycosyltransferase Hemicellulose Mannan Xylan Xyloglucan Mixed-linkage glucan
All cells of higher plants are encased in a wall, a composite material that consists of numerous polymers including the polysaccharides cellulose, non-cellulosic polysaccharides including pectic polysaccharides, structural glycoproteins and, in secondary walls, the polyphenol lignin (Carpita and Gibeaut 1993; Somerville et al. 2004). The polymer type, structure, and abundance can vary greatly depending on the plant species, tissue type, developmental stage, and the wall layer/location within a single plant cell (Pauly and Keegstra 2010). A common feature of the walls of higher plants is the presence of cellulose, which consists of β-1,4-linked glucan chains partially crystallized into microfibrils. These microfibrils are associated with or crosslinked by non-cellulosic polymers such as a class of structurally diverse polysaccharides consisting of β-1,4-linked glycans with various glycosyl substitutents. The β-1,4-linked backbone structure is believed to allow these polysaccharides to interact with themselves and with cellulose chains non-covalently via H-bonds (Scheller and Ulvskov 2010). However, unlike cellulose, where the glucan chains associate tightly to form microfibrils, this class of non-cellulosic polysaccharides contains heterogeneous substitutents or other linkages in their polymer-backbone which render them amorphous and mostly soluble in aqueous solutions. There is evidence that these polysaccharides can be covalently linked to lignin (Carpita 1996; Harris and Trethewey 2010) and to some of the pectic polysaccharides (Bauer et al. 1973; Tan et al. 2013).
Some non-cellulosic polysaccharides have been historically grouped into the class of hemicelluloses. Hemicellulose does not refer to specific structures in the wall but is often synonymous with polysaccharides that can only be extracted from walls by strong chaotropic agents such as alkali (Valent and Albersheim 1974) excluding the glucan chains of cellulose. However, most hemicelluloses are easily extractable with water when they occur outside of a cell wall, for example as an amyloid in seed endosperms (Kooiman 1961). In general, hemicelluloses represent around one-third of the dry mass of cell walls (in industrial circles termed lignocellulosics) depending on the plant feedstock used (Pauly and Keegstra 2008). Hence, utilizing lignocellulosic materials in a biorefinery for the sustainable and economically competitive production of biofuels and other commodity chemicals requires an optimized conversion process for hemicelluloses.
While this review will not attempt to solve the nomenclature issue surrounding the term hemicellulose we refer here to hemicelluloses as wall polysaccharides other than cellulose in which the dominant backbone linkage is represented by the β-1,4 glycosidic bond. By that definition β-1,4 linked galactan is not included, as it represents a pectic sidechain rather than a backbone and is synthesized by a distinct glycosyltransferase family different from the other polymers we discuss here (Liwanag et al. 2012). This review will encompass the biosynthesis of the heteromannans, xyloglucans (XyGs), heteroxylans, and mixed-linkage glucan (MLG).
The most ancient hemicellulose of the wall polysaccharides discussed here is thought to be the mannans as they have been found in the walls of some algae (Domozych et al. 2012). Mannans are also the major hemicellulose in the secondary cell wall of gymnosperms (Pauly and Keegstra 2008). They are less abundant in the walls of spermatophytes, but up to 5 % (w/w) can be detected in some species (Scheller and Ulvskov 2010; Rodriguez-Gacio et al. 2012). Mannans are used in the food industry as a stabilizer and gelling agent (Buckeridge 2010; Edwards et al. 1992). Their biological role in planta is related to imparting cell wall rigidity, functioning as a seed storage polymer (Buckeridge 2010; Rodriguez-Gacio et al. 2012), and other physiological roles including wall signaling, embryogenesis, and tissue differentiation (Benova-Kakosova et al. 2006; Goubet et al. 2009; Rodriguez-Gacio et al. 2012).
As with all wall polysaccharides heteromannans are synthesized from activated nucleotide sugars. For mannans these nucleotide sugars are GDP-mannose, GDP-glucose, and UDP-galactose (Liepman et al. 2005). While the necessary enzymes for the nucleotide sugar conversion from sucrose to GDP-mannose and UDP-galactose have been identified, the essential enzyme for the generation of GDP-glucose for polysaccharide biosynthesis in planta has not been identified (Bar-Peled and O’Neill 2011). The activated nucleotide sugars are then utilized by highly specific, Golgi-localized glycosyltransferases (GTs), which facilitate the formation of the specific linkage between the monomers and thus synthesize the polymer (Breton et al. 2001, 2006).
The first β-mannan synthase (ManS), a member of the cellulose synthase-like family A (CSLA) from GT family 2, was identified in guar seeds (CtManS in Cyamopsis tetragonoloba, a AtCSLA9 ortholog) including the demonstration of its in vitro ManS activity (Dhugga et al. 2004). Recently a deep sequencing approach was used to identify genes involved in galactomannan biosynthesis in developing Trigonella foenum-graecum (Fenugreek) endosperm (Wang et al. 2012). This work identified a CSLA family protein involved in mannan backbone synthesis and activity assays with the heterologously expressed protein showed a preference towards GDP-mannose as a donor substrate. Exclusive ManS acitivity has also been shown for a CSLA7 protein in Arabidopsisthaliana (Liepman et al. 2005). This finding together with genetic evidence suggests that this protein accepts only GDP-man (Goubet et al. 2009). Reverse genetics and biochemical approaches in the plant model species Arabidopsis have implicated the CSLA family in the synthesis of the glucomannan backbone (Liepman et al. 2005, 2007; Goubet et al. 2009 Suzuki et al. 2006). Using an RNAseq approach a CSLA3 ortholog (AkCSLA3) was identified in the developing corm tissue of Amorphophallus konjac, a tissue that produces large amounts of glucomannans as a storage polymer (Gille et al. 2011a). The heterologously expressed protein was able to utilize both GDP-mannose and GDP-glucose to synthesize glucomannan (Gille et al. 2011a). The in vitro activities of other glucoManS have also been demonstrated including AtCSLA2, AtCSLA3, AtCSLA9 (Liepman et al. 2005), and PtCSLA1 (Populus trichocarpa; Suzuki et al. 2006). Regarding protein structure and topology of the CSLAs, the AtCSLA9 protein was shown to have its catalytic domain facing the Golgi lumen suggesting that the nucleotide sugars (GDP-mannose and GDP-glucose) need to be imported into the Golgi (Davis et al. 2010).
There is some evidence that CSLD (AtCSLD2, 3 and 5) proteins also mediate mannan biosynthesis in Arabidopsis, particularly in tissues that involve tip-growth such as root hairs (Verhertbruggen et al. 2011). Additional auxiliary genes involved in mannan biosynthesis were discovered in Fenugreek. A Golgi-localized mannan synthesis-related (MSR) gene was found to be highly and specifically expressed in the fenugreek endosperm (Wang et al. 2012, 2013). The TfMSR protein contains a DUF246 conserved domain and is distantly related to GT family 65. In Arabidopsis the homologs AtMSR1 and AtMSR2 are highly co-expressed with the ManS of the CSLA family. Studies of AtMSR knock-out mutants revealed a significant decrease in stem glucomannan as well as reduced ManS activity in vitro (Wang et al. 2013). While the biochemical activity of MSR proteins remains unknown, hypotheses include a role in primer synthesis to initiate mannan biosynthesis, the synthesis of oligosaccharides linked to CSLA or promoting folding, stability or activity of a mannan synthase complex (Wang et al. 2013). The identification of an apoplastic mannan transglycosylase activity in plants suggests that once the polymer is deposited in the wall it can undergo further modification adding complexity to the biosynthesis process (Schroder et al. 2004).
While a mannan:galactosyltransferase (GalT) has been identified originally in Trigonella foenum-graecum (Edwards et al. 1999), an enzyme that facilitates mannan O-acetylation remains unknown. However, the recent discovery of the involvement of a large plant-specific family of Trichome birefringence-like (TBL) proteins in wall polymer O-acetylation as specific O-acetyltransferases suggests that this gene family encompasses a mannan O-acetyltransferase (Gille et al. 2011b). Indeed, the Amorphophallus konjac deep sequencing database (Gille et al. 2011a) contained a highly expressed (among the 10 most abundant ESTs) homolog of AtTBL25 suggesting that this protein or the closely related AtTBL26 could represent mannan O-acetyltransferase(s) in Arabidopsis (Gille et al. 2011b).
XyG is the most abundant hemicellulose in the primary cell wall of dicots making up to 20–30 % (w/w; Fry and Janice 1989; Scheller and Ulvskov 2010). XyG is thought to form cross-links between cellulose microfibrils, forming a strong but extensible XyG-cellulose network that might function as the main load-bearing component of the primary cell wall (Albersheim et al. 2010; Hayashi 1989; McCann and Roberts 1991; Pauly et al. 1999a, b; Somerville et al. 2004). In addition, XyG can also be covalently linked to pectin (Popper and Fry 2008). Several apoplastic proteins have been discovered that can act on XyG such as XyG endotransglycosylases (Nishitani 1992; Nishitani and Tominaga 1992; Smith and Fry 1991), various glycosyl hydrolases (Iglesias et al. 2006; Gunl and Pauly 2011; Gunl et al. 2011), and expansins (Cosgrove 1998). Hence, XyG metabolism and turnover is thought to play an important role in cell elongation (Takeda et al. 2002; Pauly et al. 2001b). Such a role would predict that a plant lacking XyG should have severe cellular and tissue defects. However, an Arabidopsis mutant lacking detectable XyG has relatively minor growth phenotypes, indicating that XyG is not as vital for wall structure and function as previously thought (Cavalier et al. 2008). Alternatively, XyG may have a role as a spacer-molecule by preventing the formation of microfibrilar cellulosic aggregates (Thompson 2005; Anderson et al. 2010) or an adapter-molecule, which enables cellulose to interface with other cell wall matrix components (Keegstra et al. 1973; Talbott and Ray 1992; Ha et al. 1997; Cavalier et al. 2008). Recent data derived from mechanical testing experiments on primary walls using XyG modifying enzymes point towards a limited but essential structural role of XyG in wall mechanics (Park and Cosgrove 2012a, b). Some plant species (such as Tamarind, Tropaeolum, or Hymenaea) utilize XyG as a seed storage polymer (Kooiman 1961; Meier and Reid 1982; Rao and Srivastava 1973). The polysaccharide is deposited in large quantities during seed development and upon germination is mobilized by glycosyl hydrolases (Edwards et al. 1988). XyGs are known to be used for various food, industrial, and pharmaceutical applications depending on the tissue source and plant species (Mishra and Malhotra 2009). XyG derived from Tamarind seed has been tested successfully as a carrier for specific drug delivery systems (Mishra 2007; Rolando and Valente 2007), the administration of antibiotics, and treatment of ulcers (Ghelardi et al. 2000).
The glucan backbone of XyG is believed to be synthesized by one or more members of the cellulose synthase-like family C (CSLC) genes. Using a transcriptional analysis of developing nasturtium (Tropaeolum majus) seeds, which produce large amounts of galactoXyG as a storage polymer, the TmCSLC4 gene was identified (Cocuron et al. 2007). Heterologous expression of this gene in the yeast Pichia pastoris produced a β-1,4 glucan only when co-expressed with a XyG:XylT, despite a lack of UDP-xylose in Pichia (Cocuron et al. 2007). Hence the protein XyG:XylT, but not its transferase activity, is required for glucan synthesis. The AtCSLC4 protein has been localized to the Golgi apparatus (Davis et al. 2010), the site of XyG biosynthesis (Moore and Staehelin 1988). Interestingly, immuno-electron microscopy and membrane fractionation experiments have shown one of the four barley CSLCs (HvCSLC2) to be located in the plasma membrane (Dwivany et al. 2009) rather than in the Golgi. For this reason it has been proposed that perhaps not all the CSLCs are involved in XyG biosynthesis. The apparent catalytic domain of the AtCSLC4 protein is believed to be located on the cytosolic side of the Golgi membrane utilizing cytosolic UDP-glucose. During XyG biosynthesis the nascent glucan chain is presumably channeled into the Golgi lumen by the multi-transmembrane domain containing CSLC where it can be acted upon by XylTs (Davis et al. 2010).
While it remains unclear what mechanism is responsible for patterning of xylosyl substitutions, several XyG:XylTs have been identified. Arabidopsis has seven genes in GT family 34, five of which appear to be XyG:XylTs (Faik et al. 2002; Vuttipongchaikij et al. 2012). Transfer of radiolabeled xylose has been detected from UDP-xylose onto cellohexaose acceptors for AtXXT1, AtXXT2, and AtXXT4. Analysis of single, double, and triple mutants for AtXXT1, AtXXT2, and AtXXT5 in Arabidopsis demonstrates that all three of these genes are involved in XyG biosynthesis (Zabotina et al. 2008, 2012; Cavalier et al. 2008). Overexpression of AtXXT3 was able to partially restore XyG content in some of these mutant lines, indicating that it is also likely a XyG:XylT. Double xxt1xxt2 knockout mutants lacked detectable XyG in their walls and exhibited abnormal root hairs and slow growth compared to wild-type plants (Cavalier et al. 2008). The observation that both AtCSLC4 and AtXXT2 are required for heterologous glucan synthase (GlcS) activity when expressed in Pichia suggests that these proteins form a complex as demonstrated by bimolecular fluorescence complementation and immune-coprecipitation (Chou et al. 2012).
The xylosyl substituents of XyG can be further substituted. Two genes, AtMUR3 and AtXLT2, have been identified from GT family 47 as β-1,2 galactosyltransferases (Madson et al. 2003; Jensen et al. 2012). The AtMUR3 gene was identified by screening chemically mutagenized Arabidopsis plants for aberrant cell wall composition (Reiter et al. 1997) and was subsequently shown to act as a XyG:GalT in vitro with specificity for transferring the galactosyl residue to the third xylosyl residue (Fig. 2; Madson et al. 2003). AtXLT2, a XyG:GalT with specificity for the second xylosyl residue, was identified by transcriptional analysis of developing nasturtium seeds and homology to other GTs (Jensen et al. 2012). The Arabidopsisxlt2 mutant lacks galactosylation at the second position and plants with mutations in both AtXLT2 and AtMUR3 have XyG that is nearly devoid of galactosylation. Despite this severe biochemical phenotype in terms of sidechain substitution, the mur3 xlt2 double mutant plants have relatively minor defects in plant morphology with only a slight dwarfism observed (Jensen et al. 2012). AtXUT1, a gene required for the presence of galacturonic acid in the XyG of Arabidopsis root hairs, was identified using a reverse genetics approach (Pena et al. 2012). Similar to MUR3 and XLT2, XUT is present in the same subclade of GT family 47 and can add a glycosyl substituent to the O-2 of a xylosyl residue. Other genes within the GT 47 family may thus be responsible for the diversity of decorations found at this position in various plant species.
A single XyG: α-1,2 fucosyltransferase (AtFUT1) from GT37 was shown to add a terminal fucosyl group to a galactosyl or galacturonosyl residue of XyG (Perrin et al. 1999; Faik et al. 2000; Pena et al. 2012). The fut1/mur2 mutation revealed a complete absence of fucosylated XyG in Arabidopsis suggesting that AtFUT1 is non-redundant (Vanzin et al. 2002; Perrin et al. 2003; Keegstra and David 2010; Zabotina et al. 2012).
XyG contains acetyl substituents at various positions on the polymers depending on the plant species (Gille and Pauly 2012). In ArabidopsisO-acetylation can occur specifically on the galactosyl units (Kiefer et al. 1989; York et al. 1988) and the transfer of the acetyl group is mediated by AtAXY4, a gene required for the specific O-acetylation of XyG (Gille et al. 2011b). AXY4 is a member of the TBL protein family (Bischoff et al. 2010a, b). In grasses (Gibeaut et al. 2005) and Solanales species (Sims et al. 1996) the glucosyl residues of the XyG backbone can be O-acetylated, but the specific genes responsible for this modification, likely members of the TBL gene family, have not been identified to date.
Recent evidence suggests that apoplastic glycosyl hydrolases involved in the maturation and incorporation of XyG into the wall have a major impact on XyG substitution patterning (Gunl et al. 2011). This process involves several XyG modifying enzymes including endotransglucosylase/hydrolases (XTHs; Rose et al. 2002), which possess XyG: endotransglucosylase (XET) or XyG:endohydrolase (XEH) activities (Rose et al. 2002). XTH activity is likely required for altering the length of the polysaccharide, incorporating it into the wall structure, and/or remodeling the network during cell elongation. Several plant XyG glycosidases have been identified including a xylosidase, AtXYL1, a galactosidase, AtBGAL10, and a fucosidase, AtAXY8 (Gunl et al. 2010, 2011; Sampedro et al. 2001, 2012). Arabidopsis mutant lines of xyl1, bgal10, or both bgal10 xyl1 showed a reduction in their respective glycosidase activities, altered XyG sidechain substitution pattern, and impacts the growth of certain tissues such as siliques (Sampedro et al. 2012). The fucosyl residue of XyG is removed by an α-fucosidase (Léonard et al. 2008; Gunl et al. 2011). The corresponding Arabidopsis mutant axy8 showed an expected increase in the abundance of fucosylated XyG oligosaccharides, as well as unusual XyG oligosaccharides fragments of partially degraded XyG re-incorporated into the polymer (Gunl et al. 2011). Interestingly, removing fucosidase and xylosidase activity in the axy8 xyl1 double mutant leads to a XyG with a very homogeneous sidechain substitution pattern.
Xylans are the major hemicellulose in secondary walls of dicots and all types of walls in commelinid monocots. Xylans are essential in the development of tissues with high content of secondary walls and strengthening these load-bearing tissues. Mutant plants with abnormal xylan generally have a collapsed xylem phenotype, since impaired vessel walls cannot withstand the high negative pressure generated by transpirational pull. This defect leads to a problem with water transport to photosynthetically active tissues and subsequently growth reduction and decrease in biomass accumulation (Brown et al. 2007, 2009; Lee et al. 2007a, b). Xylans can also be abundant in the endosperm walls of cereal grains, where they are considered a storage polymer (Fischer et al. 2004; Naran et al. 2008). For industrial purposes xylan is utilized in food and non-food applications (Ebringerová 2005). Xylan can be used as an additive to improve the flavor and baking performance of flour (Courtin and Delcour 2002). Due to beneficial health properties xylans are also used in medical applications: they are considered anti-carcinogenic and able to improve beneficial bacterial population growth in the colon (Fooks et al. 1999). In addition, xylans can cross-link to form films or gels with favorable physical properties and bioaffinities in industrial applications (Sarossy et al. 2012).
Intensive research using forward and reverse genetic approaches mainly in Arabidopsis combined with expression analysis has revealed numerous genes involved in xylan biosynthesis (Doering et al. 2012; Scheller and Ulvskov 2010). Unlike other hemicelluloses there is no evidence so far that a CSL plays a role in the backbone formation of xylan. Instead, several GTs from other GT families have been identified that are thought to be involved in xylan backbone elongation. IRX9 (GT family 43), IRX14 (GT family 43), and IRX10 (GT family 47) are named according to the irregular xylem phenotype of their corresponding mutants, all of which exhibit dwarfism (Brown et al. 2009; Keppler and Showalter 2010; Wu et al. 2009, 2010; Chiniquy et al. 2013; Hornblad et al. 2013). GX abundance and xylan chain length are affected in all of these mutants, and xylan synthase (XylS) activity measured in microsome preparations is lower in the irx14 mutant and irx10irx10L double mutant (Brown et al. 2009). However, XylT activity has only been demonstrated when coexpressing AtIRX9 and AtIRX14, or the respective putative orthologs from poplar (Populus trichocarpa) PtGT43B and PtGT43C, simultaneously in tobacco microsomes with exogenously fed xylooligomers (Lee et al. 2012c, d). Close homologs IRX9-L, IRX14-L, and IRX10-L have recently been proven to act as partially redundant genes. Mutating both genes of a pair (AtIRX9/AtIRX9-L, AtIRX10/AtIRX10-L, and AtIRX14/AtIRX14-L) results in severe growth defects compared with the relatively mild irregular xylem phenotype observed in single mutants. It was impossible to isolate xylan with an endo-xylanase or to detect with a xylan-specific antibody in stem material of the irx14 irx14-L and irx10 irx10-L double mutants emphasizing the essential role of these genes in xylan synthesis (Keppler and Showalter 2010; Lee et al. 2010; Wu et al. 2009, 2010).
After having been isolated from wood (Shimizu and Ishihara 1983; Johansson and Samuelson 1977) the tetrasaccharide unit Xyl-Rha-GalA-Xyl was re-discovered to be also present on the reducing end of xylan in Arabidopsis (Pena et al. 2007). Several putative glycosyltransferases, AtFRA8/IRX7 (GT family 47), AtF8H (GT family 47), AtIRX8, and PARVUS (GT8), are thought to play a role in forming this oligosaccharide in Arabidopsis, but hitherto no defined enzyme activity has been demonstrated (Brown et al. 2007; Lee et al. 2007b, 2009b; Pena et al. 2007; Wu et al. 2010; Persson et al. 2007). Several poplar genes, PdGATL1.1, PdGATL1.2, PoGT8E, and PoGT8F, have been identified as PARVUS orthologs since they are able to rescue the irregular xylem phenotype when expressed in Arabidopsis parvus mutant (Kong et al. 2009; Lee et al. 2009a). Similar to other xylan mutants, the fra8/irx7, irx8, and parvus mutants exhibit reduced growth and abnormal xylem elements. The mutants contain less xylan and lack the reducing end oligosaccharide sequence in their xylan but retain xylan backbone elongation activity. FRA8/IRX7 belongs to the inverting GT family 47 and is thus thought to transfer xylose to form the β-linked Xyl-Rha unit or to transfer rhamnose to form the α-linked Rha-GalA unit. As heterologously expressed AtFRA8/AtIRX7 exhibits xylosyltransferase activity to different monosaccharides, it is likely that it catalyzes the Xyl-Rha linkage. AtF8H is partially redundant with AtFRA8 as overexpression of AtF8H can genetically complement the fra8 phenotype, but the fra8 f8h double mutant showed additional defects (Wu et al. 2010). IRX8 and PARVUS belong to the retaining GT family 8. AtIRX8 (also named GAUT12) is a homolog of a GalA transferase GAUT1, and therefore is predicted to be a strong candidate forming the GalA-Xyl linkage (Sterling et al. 2006). PARVUS has a unique localization in both ER and Golgi while the other GTs seem to be only present in the Golgi (Lee et al. 2007b). For this reason it was proposed that PARVUS catalyzes the initiation of the reducing end sequence by transferring xylose to an acceptor in an earlier stage of xylan synthesis (Lee et al. 2009b). Considering the extreme retarded plant growth of some of the mutants described above, the reducing end tetrasaccharide sequence seems to play an indispensable role in xylan biosynthesis. It may act as a primer, if one assumes that xylan is elongated at the non-reducing end. An alternative hypothesis was brought forward that it acts as a terminator (York and O’Neill 2008) implying that xylans elongate at the reducing end and the tetrasaccharide sequence is added to terminate elongation after a certain size of the polymer is reached. Such a hypothesis would explain the existence of xylan polymers with heterodisperse chain length without a reducing end tetramer in the fra8/irx7 and irx8 mutants (Pena et al. 2007; York and O’Neill 2008). The tetrasaccharide sequence is not found in grass GAX raising the question, whether xylan synthesis in grasses utilizes a primer/terminator-free mechanism, or if there is an alternative hitherto undiscovered molecule involved. In summary, there is still a large gap in our knowledge of the mechanism of xylan backbone biosynthesis.
In contrast, more progress has been made in recent years on the substitution of xylan. GlcA is added to the xylan backbone by AtGUX1 and AtGUX2 as xylan from the Arabidopsisgux1 gux2 double mutant shows decreased abundance of GlcA and much lower xylan:GlcAT activity (Mortimer et al. 2010). There are five GUX members in Arabidopsis. A gux1 gux2 gux3 triple mutant led to a complete loss of GlcA and MeGlcA sidechains on xylan and xylan:GlcAT activity was detected, when these three GUX proteins were expressed in tobacco cells (Lee et al. 2012a). Detailed analysis of the xylan in various gux mutants as well as corresponding complemented lines indicates that the various GUX enzymes lead to distinct differences in GlcA substitution patterns of xylan (Bromley et al. 2013). In in vitro experiments, AtGUX1 strongly favors xylohexaose as an acceptor molecule and GlcA is almost exclusively added to the fifth xylose residue from the non-reducing end. However, only AtGUX1, 2, and 4 of the GUXs exhibited in vitro xylan:GlcAT activity (Rennie et al. 2012). While the gux1 gux2 mutant plants display normal plant growth, the gux1 gux2 gux3 triple mutant exhibits reduced secondary wall thickening, collapsed vessels, and dwarfism (Lee et al. 2012a). There is evidence that a rice OsXAX1 is involved in adding xylosyl units to form the β-Xylp-1-2-α-Araf sidechain of xylan (Chiniquy et al. 2012). When expressed in tobacco microsomes OsXAX1 displayed xylosyltransferase activity onto endogenous acceptors and the Osxax1 mutant lacked this specific sidechain structure on xylan as determined by NMR. Interestingly, Osxax1 mutant plants are also deficient in ferulic and coumaric acid, though the reason for this is unknown (Chiniquy et al. 2012). XAX1 belongs to GT family 61 as does another group of enzymes, the XATs. Wheat TaXAT1 and TaXAT2 mediate arabinofuranosyl transfer onto xylan (Anders et al. 2012). Knocking-down TaXAT1 and TaXAT2 expression in wheat endosperm strongly decreases α-1,3-linked arabinosyl substitution of xylan. Moreover, heterologous expression of wheat and rice XATs in Arabidopsis leads to arabinosylation of xylan providing gain-of-function evidence for α-1,3-arabinosyltransferase activity (Anders et al. 2012). Another member of GT family 61, At5g55500, has been shown to harbor xylan:XylT activity (Strasser et al. 2000). Taken together these data indicate that members of the GT family 61 entail a variety of substrate specificities. Several other GTs from GT family 75 have been identified to be potentially involved in xylan biosynthesis (Zeng et al. 2010). GT family 75 is predicted to be required for GAX:GlcAT activity (Dhugga et al. 1997). However, the exact function of these enzymes remains unclear. A GX methyltransferase (GXMT/GMX3) has been identified in Arabidopsis. The enzyme transfers specifically a methyl group from S-adenosyl-methionine to the O-4 position of glucuronosyl residues linked to xylan (Urbanowicz et al. 2012). The enzyme contains a DUF579 domain, which has now been proposed to represent a cation-dependent polysaccharide specific O-methyltransferase (Lee et al. 2012b; Urbanowicz et al. 2012). Mutants with a defect in this enzyme show a 75 % reduction in O-methylated GlcA. Double mutants with a homologous genes, gmx2 gmx3, contain a xylan, where only 10 % of the GlcA residues are O-methylated (Lee et al. 2012b). Additional genes with a DUF579 domain thus representing a putative methyltransferase include also IRX15 and IRX15-L that had been identified based on the Arabidopsis irregular xylem screen (Brown et al. 2011; Jensen et al. 2011). While the irx15 irx15-L double mutant shows a moderate collapsed xylem phenotype, the gxm single or double mutant plants do not suggest that a low GlcA methylation does not necessarily affect xylan function in xylem vessels (Lee et al. 2012b). Recently the TBL family member TBL29 has been reported to be responsible for xylan O-acetylation on both O-2 and O-3 positions and hence represents a putative xylan O-acetyltransferase. Xylan O-acetylation is decreased by ~60 % in the tbl29 mutant showing a severe collapsed xylem phenotype hinting the importance of this substituent for proper xylan function (Xiong et al. 2013).
Various GTs might be organized into protein complexes to synthesize xylan in a cooperative fashion (Faik 2010). This could be the reason that attempts to produce xylan synthase activity by heterologously expressing several xylansynthase candidate genes (see above) have failed. XylT activity was only reported in tobacco microsomes, which co-express IRX9 and IRX14 suggesting that these GTs need to act in concert to synthesize the xylan backbone (Lee et al. 2012c, d). Direct evidence supporting the existence of a xylan synthase complex was obtained in experiments with wheat in which a protein complex containing members of GT family 43, 47, and 75 was enriched using co-immunoprecipitation (Zeng et al. 2010). This work also provides evidence for a cooperative model of backbone formation and substitution. Both UDP-arabinose and UDP-glucuronic acid can stimulate XylT activity in a complex. Similarly, coordination of GlcAT and XylT activity is also observed in microsomes isolated from Arabidopsis and pea epicotyls (Baydoun et al. 1989). Clearly, more research is needed to understand the mechanism of how these GTs and other auxiliary enzymes are positioned and cooperate in producing xylan.
In higher plants MLG is restricted to the walls of grasses (Poaceae) although the polymer has also been discovered in more ancient plant lineages such as that of the genus the Equisetum (Fry et al. 2008b; Sorensen et al. 2008). In addition, MLG-like structures were found in liverworts (Popper and Fry 2003) and even algae such as green algae, red algae, and charophytes (Popper et al. 2011). Due to the absence of MLG from many Spermatophytes it is likely that the occurrence of MLG in the relatively young Poales has evolved independently. The abundance of MLG is developmentally regulated and changes during cell elongation. MLG accumulates up to 20 % dry mass in walls of rapidly growing maize coleoptile particularly in the mesophyll tissue (Carpita 1996; Carpita and McCann 2010). MLG is then degraded through hydrolysis by plant licheninases as well as β-glucosidases (Kim et al. 2000; Hrmova and Fincher 2001) concomitant with an increase in cell elongation. Due to this polymer turnover MLG has been proposed to be a storage polymer for energy needed for cell elongation (Carpita 1996). However, low levels of MLG remain in most, if not all, grass tissues (Gibeaut et al. 2005). MLG is also present in high abundance in mature stems of rice (Vega-Sanchez et al. 2013), where it may have a structural role (Vega-Sanchez et al. 2012). Due to MLG’s presence in many endosperms of cereals it is an integral part of the human diet. As it cannot be digested in the human gut it is considered a dietary fiber. Regular uptake of MLG is thought to reduce the risk of cardiovascular disease, type II diabetes, colorectal cancer, and lowers blood cholesterol levels (Truswell 2002; Dikeman and Fahey 2006; Cummings 1992).
Two gene classes have been identified in mediating MLG synthesis, CSLF and CSLH, both representing grass-specific branches of the CSL gene family (Scheller and Ulvskov 2010). CSLF emerged as a candidate due to comparative genomics and a QTL analysis of barley varieties with different abundances of MLG in their grain endosperms (Burton et al. 2006). cDNA clones of several rice paralogs of CSLF (OsCSLF2, OsCSLF4; Hazen et al. 2002) were overexpressed in Arabidopsis for functional analysis and the presence of MLG could be confirmed in the transformed Arabidopsis leaves using an MLG-specific antibody (Burton et al. 2006). When the leaf material was treated with a barley licheninase, which specifically hydrolyzes MLG, the antibody detection of MLG was abolished. It was estimated that the OsCSLF overexpressing plants contained less than 0.1 % MLG (w/w). Additional support for the involvement of CSLFs in MLG synthesis was obtained when a barley HvCSLF6 gene under the control of an endoperm-specific promotor was transformed into barley resulting in an 80 % increase of MLG in the grain of barley (Burton et al. 2011). Utilizing a constitutive expression promotor such as 35S also resulted in a three-to-fourfold higher content of MLG in the leaves of barley, but the plants were severely stunted and in some cases died. The role of OsCSLF6 in MLG biosynthesis was also confirmed in rice using two approaches (Vega-Sanchez et al. 2012). Overexpression of the OsCSLF6 gene in tobacco leads to the formation of MLG and a cslf6 rice mutant had a 97 % MLG reduction in coleoptiles. However, the cslf6 mutant exhibited a dwarfed growth phenotype, a reduced stem internode diameter, and a compromised anthesis expanding the potential functions of MLGs (Vega-Sanchez et al. 2013). MLG synthesis can also be monitored in microsomal preparations by feeding radioactive glucose and subjecting the resulting products to a licheninase treatment (Taketa et al. 2012). Such an assay was used to demonstrate MLG synthase activity for HvCSLF6. Moreover, in endosperm microsomal preparations of a betaglucanless (bgl) barley mutant, which contains a point-mutation in the HvCSLF6 gene, the MLG synthase activity was not observed (Taketa et al. 2012).
CSLH’s role in MLG biosynthesis was demonstrated using a similar approach: overexpression of a barley cDNA (HvCslH1) in Arabidopsis (Doblin et al. 2009). As was the case with CSLF this led to the occurrence of MLG in Arabidopsis walls. These experiments demonstrate that CSLF and CSLH are each independently sufficient for MLG biosynthesis, but as an in vitro mono-component assay has not been established it is not known what precise role those proteins have in MLG biosynthesis—are they involved in forming the β-1,4 bond, or the β-1,3 bond or both? It is likely that hitherto unknown proteins are recruited to participate in the synthesis of MLG. Functional equivalents of these proteins are apparently also present in a dicot such as Arabidopsis. Feeding assays with enriched Golgi fractions have shown that MLG biosynthesis occurs in the Golgi (Gibeaut et al. 2005) and CSLF and CSLH have both been shown to be localized in the Golgi (Doblin et al. 2009). However, while immunocytochemistry has been a powerful tool to detect MLG in the apoplast, it has failed to detect the polymer in the Golgi of barley coleoptiles and developing grain, but there is some evidence that MLG is present in the Golgi of maize coleoptiles (Carpita and McCann 2010). A two-step mechanism has been proposed for the synthesis of MLG (Burton and Fincher 2012; Burton et al. 2010). First, cellodextrins are synthesized in the Golgi followed by connecting them via a β-1,3 linkage by a CSLF or CSLH or by a transglycosylase. The possibility of a potential MLG transglycosylase has recently emerged with the identification of a transglycosylase activity that grafts MLG to XyG oligos in Equisetum (Fry et al. 2008a). However, experimental evidence for such a proposed two step MLG synthesis mechanism is currently lacking.
Enormous advances have been made in recent years in identifying genes that are involved in the biosynthesis of the various structurally complex hemicelluloses. In the future, applying cutting edge techniques and tools such as comparative genomics for the identification of novel glycosyltransferases (Schultink et al. personal communication), the synthesis of branched oligosaccharides through automated solid-phase-based systems for providing new, defined substrates for glycosyltransferases (Seeberger and Werz 2007), and metabolic labeling via “click”-technology for giving cellular insights into the dynamics of hemicellulose biosynthesis (Anderson et al. 2012) should enhance our knowledge even further. With the advent of sophisticated genetic approaches we have also begun to define functions for these hemicelluloses and more specifically that of their various substituents. Utilizing this knowledge we are now able to modify or tailor hemicellulose structures in planta, which might lead to emerging novel applications. The major challenge in the future will be to understand the molecular mechanism of biosynthesis by demonstrating enzyme activity and specificity of all components involved, by proving to what extent protein complexes play a role, by defining the involvement of auxiliary proteins such as hydrolases and transglycosylases, and by ultimately obtaining the crystal structure of the GTs and auxiliary enzymes. Taken together this should allow us 1 day to produce an entire hemicellulosic polymer heterologously in for example a yeast, bacteria or cell-free system.
We would like to mention the funding sources that supported the authors. S.G., G.X., A. de S., N.M. were supported by a grant from the Energy Biosciences Institute; A.S. by the Dickinsen Chair for wood science and technology for M.P.; L.L. by the Department of Energy grant: ER65037-1036816.
- Albersheim P, Darvill A, Roberts K, Sederoff R, Staehelin A (2010) Plant cell walls: from chemistry to biology. Garland Science, New YorkGoogle Scholar
- Anders N, Wilkinson MD, Lovegrove A, Freeman J, Tryfona T, Pellny TK, Weimar T, Mortimer JC, Stott K, Baker JM, Defoin-Platel M, Shewry PR, Dupree P, Mitchell RA (2012) Glycosyl transferases in family 61 mediate arabinofuranosyl transfer onto xylan in grasses. Proc Natl Acad Sci USA 109:989–993PubMedGoogle Scholar
- Benova-Kakosova A, Digonnet C, Goubet F, Ranocha P, Jauneau A, Pesquet E, Barbier O, Zhang Z, Capek P, Dupree P, Liskova D, Goffner D (2006) Galactoglucomannans increase cell population density and alter the protoxylem/metaxylem tracheary element ratio in xylogenic cultures of Zinnia. Plant Physiol 142:696–709PubMedGoogle Scholar
- Brown D, Wightman R, Zhang Z, Gomez LD, Atanassov I, Bukowski JP, Tryfona T, McQueen-Mason SJ, Dupree P, Turner S (2011) Arabidopsis genes IRREGULAR XYLEM (IRX15) and IRX15L encode DUF579-containing proteins that are essential for normal xylan deposition in the secondary cell wall. Plant J 66:401–413PubMedGoogle Scholar
- Burton RA, Collins HM, Kibble NA, Smith JA, Shirley NJ, Jobling SA, Henderson M, Singh RR, Pettolino F, Wilson SM, Bird AR, Topping DL, Bacic A, Fincher GB (2011) Over-expression of specific HvCslF cellulose synthase-like genes in transgenic barley increases the levels of cell wall (1,3;1,4)-beta-d-glucans and alters their fine structure. Plant Biotechnol J 9:117–135PubMedGoogle Scholar
- Cavalier DM, Lerouxel O, Neumetzler L, Yamauchi K, Reinecke A, Freshour G, Zabotina OA, Hahn MG, Burgert I, Pauly M, Raikhel NV, Keegstra K (2008) Disrupting two Arabidopsis thaliana xylosyltransferase genes results in plants deficient in xyloglucan, a major primary cell wall component. Plant Cell 20:1519–1537PubMedGoogle Scholar
- Courtin CM, Delcour JA (2002) Arabinoxylans and endoxylanases in wheat flour bread-making. J Cereal Sci 35:225–243Google Scholar
- Cummings JH (1992) The effect of dietary fiber on fecal weight and composition. In: Spiller GA (ed) Dietary fibre in human nutrition. CRC Press, Boca Raton, pp 263–349Google Scholar
- Dea ICM, Morrison A (1975) Chemistry and interactions of seed galactomannans. In: Tipson RS, Derek H (eds) Advances in carbohydrate chemistry and biochemistry, vol 31. Academic Press, London, pp 241–312. http://dx.doi.org/10.1016/S0065-2318(08)60298-X
- Ebringerová A (2005) Hemicellulose. Adv Polym Sci 186:1–67Google Scholar
- Edwards M, Scott C, Gidley MJ, Reid JSG (1992) Control of mannose galactose ratio during galactomannan formation in developing legume seeds. Planta 187:67–74Google Scholar
- Faik A, Bar-Peled M, DeRocher AE, Zeng WQ, Perrin RM, Wilkerson C, Raikhel NV, Keegstra K (2000) Biochemical characterization and molecular cloning of an alpha-1,2-fucosyltransferase that catalyzes the last step of cell wall xyloglucan biosynthesis in pea. J Biol Chem 275:15082–15089PubMedGoogle Scholar
- Fooks LJ, Fuller R, Gibson GR (1999) Prebiotics, probiotics and human gut microbiology. Int Dairy J 9:53–61Google Scholar
- Fry SC (1989) The structure and functions of xyloglucan. J Exp Bot 40:1–11Google Scholar
- Fry SC, Janice GM (1989) Toward a working model of the growing plant cell wall. Plant cell wall polymers. American Chemical Society, Washington, DC, pp 33–46Google Scholar
- Harris P, Trethewey JK (2010) The distribution of ester-linked ferulic acid in the cell walls of angiosperms. Phytochem Rev 9:19–33Google Scholar
- Hayashi T (1989) Xyloglucans in the primary-cell wall. Annu Rev Plant Physiol Plant Molec Biol 40:139–168Google Scholar
- Johansson MH, Samuelson O (1977) Reducing end groups in birch xylan and their alkaline-degradation. Wood Sci Technol 11:251–263Google Scholar
- Kato Y, Matsuda K (1981) Occurrence of a soluble and low-molecular weight xyloglucan and its origin in etiolated mung bean hypocotyls. Agr Biol Chem Tokyo 45:1–8Google Scholar
- Kato Y, Ito J, Mitsuishi Y (2004) Further structural study of the barley and bamboo shoot xyloglucans by xyloglucanase. J Appl Glycosci 51:327–333Google Scholar
- Keegstra K, David C (2010) Glycosyltransferases of the GT34 and GT37 families. Annu Plant Rev Plant Polysacch Biosynth Bioeng 41:235–249Google Scholar
- Kiefer LL, York WS, Darvill AG, Albersheim P (1989) Xyloglucan isolated from suspension-cultured sycamore cell walls is O-acetylated. Phytochemistry 28:2105–2107Google Scholar
- Kooiman P (1961) The constitution of tamarindus-amyloid. Recl Trav Chim Pay B 80:849–865Google Scholar
- McCann MC, Roberts K (1991) Architecture of the primary cell wall. In: Lloyd CW (ed) The cytoskeletal basis of plant growth and form. Academic Press, London, pp 109–129Google Scholar
- Meier H, Reid JSG (1982) Reserve polysaccharides other than starch in higher plants. In: Loewus F, Tanner W (eds) Plant carbohydrates I, vol 13/A. Encyclopedia of plant physiology. Springer, Berlin, pp 418–471. doi:10.1007/978-3-642-68275-9_11
- Mishra MU, Khandare JN (2007) Tamarind seed polysaccharides: biodegradable polymer for colonic drug delivery. In: Second International Conference and Indo-Canadian Satellite symposium on pharmaceutical science, technology, practice and natural products, conference chronicle, DDSGoogle Scholar
- Mishra A, Malhotra AV (2009) Tamarind xyloglucan: a polysaccharide with versatile application potential. J Mater Chem 19:8528–8536Google Scholar
- Moore P, Staehelin LA (1988) Immunogold localization of the cell-wall-matrix polysaccharides rhamnogalacturonan I and xyloglucan during cell expansion and cytokinesis in Trifolium pratense L.; implication for secretory pathways. Planta 174:433–445Google Scholar
- Mortimer JC, Miles GP, Brown DM, Zhang Z, Segura MP, Weimar T, Yu X, Seffen KA, Stephens E, Turner SR, Dupree P (2010) Absence of branches from xylan in Arabidopsis gux mutants reveals potential for simplification of lignocellulosic biomass. Proc Natl Acad Sci USA 107:17409–17414PubMedGoogle Scholar
- Nishitani K (1992) Endo-xyloglucan transferase, a novel enzyme which mediates transglycosylation reaction between xyloglucan molecules (chapter). In: Plant cell walls as biopolymers with physiological functions. pp 263–268Google Scholar
- Rao PS, Srivastava HC (1973) Tamarind in industrial gums, 2nd edn. In: Whistler RL (ed) Academic Press, New York, pp 369–411Google Scholar
- Rodriguez-Gacio MDC, Iglesias-Fernandez R, Carbonero P, Matilla AJ (2012) Softening-up mannan-rich cell walls. J Exp Bot 63:3975–3988Google Scholar
- Sarossy Z, Blomfeldt TOJ, Hedenqvist MS, Koch CB, Ray SS, Plackett D (2012) Composite films of arabinoxylan and fibrous sepiolite: morphological, mechanical, and Baffler properties. Acs Appl Mater Inter 4:3378–3386Google Scholar
- Shimizu K, Ishihara M (1983) Hemicellulases of brown rotting fungus, Tyromyces-Palustris. 5. Isolation and characterization of oligosaccharides from the hydrolyzate of larch wood glucomannan with endo-beta-d-mannanase. Agr Biol Chem Tokyo 47:949–955Google Scholar
- Tan L, Eberhard S, Pattathil S, Warder C, Glushka J, Yuan C, Hao Z, Zhu X, Avci U, Miller JS, Baldwin D, Pham C, Orlando R, Darvill A, Hahn MG, Kieliszewski MJ, Mohnen D (2013) An Arabidopsis cell wall proteoglycan consists of pectin and arabinoxylan covalently linked to an arabinogalactan protein. Plant Cell 25:270–287PubMedGoogle Scholar
- Urbanowicz BR, Pena MJ, Ratnaparkhe S, Avci U, Backe J, Steet HF, Foston M, Li H, O’Neill MA, Ragauskas AJ, Darvill AG, Wyman C, Gilbert HJ, York WS (2012) 4-O-methylation of glucuronic acid in Arabidopsis glucuronoxylan is catalyzed by a domain of unknown function family 579 protein. Proc Natl Acad Sci USA 109:14253–14258PubMedGoogle Scholar
- Vega-Sanchez ME, Verhertbruggen Y, Christensen U, Chen X, Sharma V, Varanasi P, Jobling SA, Talbot M, White RG, Joo M, Singh S, Auer M, Scheller HV, Ronald PC (2012) Loss of Cellulose synthase-like F6 function affects mixed-linkage glucan deposition, cell wall mechanical properties, and defense responses in vegetative tissues of rice. Plant Physiol 159:56–69PubMedGoogle Scholar
- Wang Y, Mortimer JC, Davis J, Dupree P, Keegstra K (2013) Identification of an additional protein involved in mannan biosynthesis. Plant J 73:105–117Google Scholar
- Wu AM, Hornblad E, Voxeur A, Gerber L, Rihouey C, Lerouge P, Marchant A (2010) Analysis of the Arabidopsis IRX9/IRX9-L and IRX14/IRX14-L pairs of glycosyltransferase genes reveals critical contributions to biosynthesis of the hemicellulose glucuronoxylan. Plant Physiol 153:542–554PubMedGoogle Scholar