Functional studies of a membrane-anchored cellulase from poplar

Ulla Jonsson Rudsander

Royal Institute of Technology School of Biotechnology Department of Wood Biotechnology Stockholm, 2007

Copyright © Ulla J. Rudsander Stockholm, 2007 ISBN 978-91-7178-790-3 Royal Institute of Technology AlbaNova University Center School of Biotechnology SE – 106 91 Stockholm Sweden Printed at Universitetsservice US-AB Box 700 14 100 44 Stockholm Sweden Cover illustration: 2 weeks old Arabidopsis thaliana seedlings. From the left: wildtype, korrigan1-1 mutant, korrigan1-1 complemented with construct pUR19, korrigan1-1 complemented with construct pUR20, korrigan1-1 complemented with construct pUR21, see section 8 (courtesy of Junko Takahashi).

Ulla J. Rudsander (2007). Functional studies of a membrane-anchored cellulase from poplar. Doctoral thesis in Wood Biotechnology. School of Biotechnology, Department of Wood Biotechnology, Royal Institute of Technology, AlbaNova University Center, Stockholm, Sweden. ISBN 978-91-7178-790-3 ABSTRACT Cellulose in particular and wood in general are valuable biomaterials for humanity, and cellulose is now also in the spotlight as a starting material for the production of biofuel. Understanding the processes of wood formation and cellulose biosynthesis could therefore be rewarding, and genomics and proteomics approaches have been initiated to learn more about wood biology. For example, the genome of the tree Populus trichocarpa has been completed during 2006. A single-gene approach then has to follow, to elucidate specific patterns and enzymatic details. This thesis depicts how a gene encoding a membrane-anchored cellulase was isolated from Populus tremula x tremuloides Mich, how the corresponding protein was expressed in heterologous hosts, purified and characterized by substrate analysis using different techniques. The in vivo function and modularity of the membrane-anchored cellulase was also addressed using overexpression and complementation analysis in Arabidopsis thaliana. Among 9 genes found in the Populus EST database, encoding enzymes from glycosyl hydrolase family 9, two were expressed in the cambial tissue, and the membrane-anchored cellulase, PttCel9A1, was the most abundant transcript. PttCel9A1 was expressed in Pichia pastoris, and purified by affinity chromatography and ion exchange chromatography. The low yield of recombinant protein from shake flask experiments was improved by scaling up in the fermentor. PttCel9A1 was however highly heterogenous, both mannosylated and phosphorylated, which made the protein unsuitable for crystallization experiments and 3D X-ray structure determination. Instead, a homology model using a well-characterized, homologous bacterial enzyme was built. From the homology model, interesting point mutations in the active site cleft that would highlight the functional differences of the two proteins could be identified. The real-time cleavage patterns of cello-oligosaccharides by mutant bacterial enzymes, the wildtype bacterial enzyme and PttCel9A1 were studied by 1H NMR spectroscopy, and compared with results from HPAEC-PAD analysis. The inverting stereochemistry for the hydrolysis reaction of the membrane-anchored poplar cellulase was also determined by 1H NMR spectroscopy, and it was concluded that transglycosylation in vivo is not a possible scenario. The preferred in vitro polymeric substrates for PttCel9A1 were shown to be long, low-substituted cellulose derivatives, and the endo-1,4-β-glucanase activity was not extended to branched or mixed linkage substrates to detectable levels. This result indicates an in vivo function in the hydrolysis of “amorphous” regions of cellulose, either during polymerization or crystallization of cellulose. In addition, overexpressing PttCel9A1 in A. thaliana, demonstrated a correlation with decreased crystallinity of cellulose. The significance of the different putative modules of PttCel9A1 was investigated by the construction of hybrid proteins, that were introduced into a knock-out mutant of A. thaliana, and the potential complementation of the phenotype was examined. A type B plant cellulase catalytic domain could not substitute for a type A plant cellulase catalytic domain, although localization and interaction motifs were added to the N- and C-terminus. Keywords: Populus, cellulose biosynthesis, endo-1,4-β-glucanase, cellulase, Pichia pastoris, Arabidopsis thaliana, NMR spectroscopy, cleavage pattern, stereochemistry, transglycosylation, complementation, hybrid proteins, overexpression, substrate analysis, overglycosylation Copyright © Ulla J. Rudsander, 2007

Ulla Rudsander (2007). Funktionella studier av ett membranbundet cellulas från poppel. Skolan för Bioteknologi, Avdelningen för Träbioteknik, Kungliga Tekniska Högskolan, AlbaNova Universitetscenter, Stockholm, Sverige. ISBN 978-91-7178-790-3 SAMMANFATTNING-eller varför alla goda ting är trä Cellulosa i synnerhet och trädens ved i allmänhet är värdefulla biomaterial för mänskligheten. På senare år har även cellulosan som utgångsmaterial för produktion av biobränsle uppmärksammats. Att förstå hur vedbildningen och cellulosans biosyntes fungerar skulle därför kunna vara givande, och både genomik- och proteomik-studier har inletts för att lära mer om vedens biologi. Till exempel har genomet för trädet Populus trichocarpa stutligen kartlagts under 2006. En infallsvinkel som fokuserar på en enda gen måste sedan följa, så att specifika mönster och enzymatiska detaljer kan klargöras. Denna avhandling skildrar hur en gen, som kodar för ett membranbundet cellulas, isolerades från Populus tremula x tremuloides Mich, proteinet uttrycktes i heterologa värdar, renades och karaktäriserades vad gäller olika substrat och med hjälp av olika tekniker. Funktionen in vivo och modulariteten för det membranbundna cellulaset studerades också, med hjälp av överexpression och komplementeringsanalys i Arabidopsis thaliana. Bland de 9 generna i Populus EST databasen, som kodar för glykosyl hydrolaser från familj 9, var två stycken uttryckta i kambium, och det membranbundna enzymet PttCel9A1 var det vanligast förekommande transkriptet. PttCel9A1 uttrycktes i Pichia pastoris, och renades med affinitetskromatografi och jonbyteskromatografi. Det låga utbytet av rekombinant protein från skakflaska förbättrades i fermentorexperiment. PttCel9A1 var emellertid mycket heterogent, både fosforylerat och mannosylerat, vilket gjorde proteinet dåligt lämpat för kristallisationsexperiment och 3D strukturbestämning. Istället gjordes en homologimodell med hjälp av ett homologt, bakteriellt enzym. Från homologimodellen kunde några intressanta punktmutationer i aktiva sitet identifieras som skulle belysa skillnaderna mellan de två proteinerna. Klyvningsmönster för PttCel9A1, det bakteriella vildtypsenzymet och bakteriella mutantenzymer studerades med hjälp av 1H NMR-spektroskopi, och resultaten jämfördes med HPAEC-PAD analyser. Stereokemin för PttCel9A1s hydrolysreaktion bestämdes också med hjälp av 1H NMR-spektroskopi, och slutsaten drogs att transglykosylering in vivo inte är möjlig. Studierna visar att PttCel9A1 föredrar långa, låg-substituterade cellulosaderivat som substrat, och att endo-1,4-β−glukanasaktiviteten inte utökades till att gälla förgrenade eller blandade bindingar i mätbara nivåer. Detta resultat pekar på att PttCel9A1 hydrolyserar amorfa regioner av cellulosa in vivo, antingen under polymeriseringen eller kristalliseringen. Dessutom visade det sig att överuttryck av PttCel9A1 i A. thaliana korrelerade med minskad cellulosakristallinitet. Betydelsen av de olika potentiella modulerna i PttCel9A1 studerades med hjälp av hybridproteiner som introducerades i en knock-out mutant av A thaliana. Den eventuella komplementeringen av mutanten undersöktes. Den katalytiska domänen för ett typ B växtcellulas kunde inte substituera för en katalytisk domän från ett typ A växtcellulas, även fast lokalisering- och interaktionsmotiv hade fuserats till N- och C-terminalen. Nyckelord: Populus, cellulosabiosyntes, cellulas, endo-(1,4)-β-glukanas, Pichia pastoris, Arabidopsis thaliana, NMR spektroskopi, klyvningsmönster, stereokemi, transglykosylering, komplementering, hybridproteiner, överexpression, substratanalys, överglykosylering Copyright © Ulla J. Rudsander, 2007

Till Elvira o Edvard, allra mest dyrbara mutanter

LIST OF PUBLICATIONS This thesis is based on the following publications, which in the text are referred to by their roman numerals: Paper I. Rudsander, Ulla; Denman, Stuart; Raza, Sami; Teeri, Tuula T. 2003. Molecular features of family GH9 cellulases in hybrid aspen and the filamentous fungus Phanerochaete chrysosporium. Journal of Applied Glycoscience 50(2), 253-256. Paper II. Master ER, Rudsander UJ, Zhou W, Henriksson H, Divne C, Denman S, Wilson DB, Teeri TT. 2004. Recombinant Expression and Enzymatic characterization of PttCel9A, a KOR homologue from Populus tremula x tremuloides. Biochemistry 43:10080-9. Paper III. Rudsander UJ, Sandstrom Corine, Piens K, Master ER, Brumer III H, Kenne L, Teeri TT. 2007. Comparative NMR analysis of cello-oligosaccharide hydrolysis by family GH9 endoglucanases from Thermobifida fusca and Populus tremula x tremuloides Mich. Submitted for publication in Biochemistry. Paper IV. Takahashi J, Rudsander UJ, Hedenström M, Banasiak A, Harholt J, Amelot N, Ryden P, Endo S, Brumer H, del Campillo E, Master ER , Vibe Scheller H, Sundberg B, Teeri TT, and Mellerowicz, EJ. 2007. Expression of aspen homologue of KORRIGAN1, PttCel9A1, affects cellulose crystallinity in secondary growth. Manuscript. In addition, previously unpublished data is included.

TABLE OF CONTENTS

TABLE OF CONTENTS __________________________________________________________________ 9 INTRODUCTION________________________________________________________________________ 1

1 BIOTECHNOLOGY IN THE FOREST INDUSTRY ______________________________________________ 1 2 SUBSTRATES AND ENZYMES __________________________________________________________ 2

2.1 Cellulose ____________________________________________________________________ 2 2.2 Functional modules____________________________________________________________ 4 2.3 Mechanism of cellulases ________________________________________________________ 6 2.4 Classification of cellulases ______________________________________________________ 8 2.5 Microbial cellulases ___________________________________________________________ 9 2.6 Glycosyl hydrolase family 9 ____________________________________________________ 10 2.7 Classifications of plant cellulases ________________________________________________ 12

3 WOOD FORMATION ________________________________________________________________ 13 3.1 Wood anatomy_______________________________________________________________ 13 3.2 Wood cell differentiation_______________________________________________________ 14 3.3 Approaches to understanding wood formation ______________________________________ 16

4 BIOSYNTHESIS OF CELLULOSE ________________________________________________________ 17 4.1 Biosynthesis of cellulose in bacteria ______________________________________________ 18 4.2 Biosynthesis of cellulose in plants________________________________________________ 18 4.3 A membrane-anchored cellulase is an actor in plant cellulose biosynthesis _______________ 20

OBJECTIVES OF THE PRESENT INVESTIGATION________________________________________ 23 RESULTS AND DISCUSSION ____________________________________________________________ 25

5 RECOMBINANT EXPRESSION OF PTTCEL9A1 (PAPER II) ____________________________________ 25 6 GLYCOSYLATION ANALYSIS OF RECOMBINANT PTTCEL9A1 (PAPERS I, II AND UNPUBLISHED WORK) _ 26 7 SUBSTRATE ANALYSIS OF GH9 MICROBIAL AND PLANT CELLULASES (PAPERS I, II AND III) _________ 30 8 DOMAIN SWAPPING OF PLANT CELLULASES (UNPUBLISHED WORK)____________________________ 38 9 MECHANISM FOR A MEMBRANE-ANCHORED CELLULASE IN PLANT CELLULOSE BIOSYNTHESIS? (PAPER IV) _______________________________________________________________________________ 40

CONCLUDING REMARKS AND FUTURE PERSPECTIVES _________________________________ 45 LIST OF ABBREVIATIONS______________________________________________________________ 47 ACKNOWLEDGEMENTS _______________________________________________________________ 49 REFERENCES _________________________________________________________________________ 51

1

INTRODUCTION

1 Biotechnology in the forest industry Today paper is made out of wood, but before the mid 19th century, paper was made from rags of cotton and flax (Sjunnesson, 2006). However, there was a shortage of raw material when industrial methods were implemented, which is why mechanical pulping of wood was invented. Paper manufacturing was transformed from textile industry, into the forest industry that is successful in Sweden also in the 21st century. The interest in biotechnological treatment of pulp was raised in response to growing environmental awareness in the early 1980´s. The enzymatic applications in use so far include cellulases for pulp fibrillation and deinking, xylanases and laccases for bleaching and beating, and the use of lipases for depitching (Gutierrez et al., 2001; Bajpai, 2004; Sukumaran et al., 2005). The elevated cost for enzymatic manipulations is however problematic in the paper industry, where large volumes at high speed are required. The introduction of biotechnology into the forest industry has consequently been relatively slow, but instead there are new areas emerging for the utilization of cellulose fibers. Global warming has channeled research efforts towards a phossil fuel free society (Ragauskas et al., 2006). Plant biomass represents a renewable resource, and the most abundant biopolymer on earth is cellulose. The polymer cellulose is the main structural component of the cell walls of plants and algae, and it is also produced in smaller quantity by some bacteria, fungi, protists and marine invertebrates (Nakashima et al., 2004). Cellulases are key-players in the wood-to-ethanol process, since they can perform the first depolymerization step of cellulose to fermentable sugars. Production of cellulases in tanks is expensive, however, and so is the pre-treatment allowing the access of enzymes for cellulose deconstruction. Recent advances in plant biotechnology could reduce the costs by depolymerization in situ in the tree, using crops that, for example, self-produce efficient microbial cellulases in a controlled manner. Tailoring the cellulose fiber inside the tree by transgenic techniques might be the future for forest biotechnology, making the fiber more accessible for biofuel or attractive for biomaterials. In this context, it is crucial that the risks and strategies of introducing transgenic plants into natural forests are studied in depth. High performance biomaterials could replace materials relying on petrochemicals or realise the promises of medical tissue engineering (Teeri, 2007). The cellulose fiber is biodegradable, biocompatible and has good mechanical properties, but it is an unreactive compound. There are, however, ways to circumvent this characteristic. Biomimetic engineering is an innovative tool, where design from biology is transferred to technological disciplines. Biomimetic modification of cellulose surfaces relies on the natural affinity of other polymers or proteins present in the plant. For example, chemically or enzymatically modified xyloglucan can be employed to attach functional groups to cellulose (Brumer et al., 2004). This functionalized biomaterial can then be implanted in the body as artificial blood vessels, or it can be employed in the packaging industry. The wood biotechnology group at KTH is involved in basic and applied research about wood formation. There is a high potential for very interesting applications if we can understand how cellulose is biosynthesized. However it is also a difficult area to study, since the process is carried out by large protein complexes in the plasma membrane. A cellulase with a single transmembrane anchor and its catalytic domain solvent-exposed might then be a good starting point.

Ulla Rudsander

2

2 Substrates and enzymes The genesis of cellulase biochemistry was in 1951 when crosslinked dextran gel (Sephadex) made it possible to separate different cellulolytic enzyme components (Himmel et al., 1999). First, a single cellulase enzyme was thought to be responsible for cellulose degradation, but it soon became clear that the breakdown of cellulose involved a set of enzymes. In order to understand the action mode of all the different cellulases, it is worthwhile to first take a closer look at the substrate.

2.1 Cellulose The chemical definition of cellulose is deceptively simple. It is a polymer composed entirely of glucose with identical chemical linkages. However, as outlined below, cellulose can form three-dimensional, crystalline networks, and the many different forms have their own characteristics.

Cellulose can have a degree of polymerization (DP) of up to 15 000 D-glucose units. D-glucose molecules exist in different forms in aqueous solution (Fig. 1a). An equilibrium mixture of the aldose (open chain aldehyde) and five- or six-membered rings is produced. In solution, the cyclic hemiacetal forms will exist in a ratio of α and β anomers, a thermodynamic phenomenon called mutarotation. Upon polymerization of glucose units, o-glycosidic bonds are formed, whereby the glycosidic carbon (C1) is linked through an oxygen atom to a carbon atom (C4) on another monosaccharide residue, a reaction that is accompanied by the departure of a water molecule (Fig. 1b). Depending on whether the configuration at the anomeric C1 carbon of a residue is α or β, resulting glucosidic bond will be either an α-1,4-glycosidic or a β−1,4-glycosidic bond. The polymerization of many α-D-glucose units produces the polymer known as starch, whereas the polymerization β−anomers of D-glucose yields cellulose. In a cellulose chain every glucose unit is rotated 180° relative to one another thus making cellobiose the repeating unit (Fig. 1b). The glucosidic bond is in the right orientation for cleavage by cellulolytic enzymes after every other glucose residue. Therefore, the main product of the enzymatic hydrolysis of cellulose is often cellobiose.

The ends of a cellulose chain have different characteristics (Fig. 1). The hemiacetals at the reducing end exist in equilibrium in aqueous solution with relatively small but significant concentrations of noncyclic aldehydes. The noncyclic aldehydes (the aldose compound) can undergo oxidation, and thus be detected by different oxidizing reagents.

Starch, a polymer of glucose linked in α−1,4-glycosidic bonds, tends to coil into helical structures. Callose, a polymer of glucose linked in β-1,3-glycosidic bonds, also forms helical structures. In contrast, cellulose, linked in β−1,4-glycosidic bonds, adopts long, linear chains. The linear arrangement of cellulose presents a uniform distribution of –OH groups that can ‘zip’ the chains together by intra- and interchain hydrogen bonding. Cello-oligosaccharides become increasingly insoluble in water after the DP of 6 glucose units. Zipping many cellulose chains together produces a crystalline lattice in the cell wall. The chains are arranged in microfibrils of about 30 to 100 molecules lying side-by-side. Four polymorphs of crystalline cellulose are known: cellulose I, cellulose II, cellulose III and cellulose IV. Native cellulose always exists as cellulose I, although cellulose II is the more thermodynamically stable form. In cellulose I, the parallel glucan chains all run in the same direction (Chanzy and Henrissat, 1985), whereas the glucan chains in cellulose II are probably arranged in an antiparallell fashion. If cellulose I is dissolved and recrystallized, it is converted to cellulose

Substrates and enzymes

3

Figure 1. (A) Six-membered ring forms and the open-chain form of D-glucose are shown. A chiral center is formed during cyclization. The α-anomer has the C1 hydroxyl group in the axial position, whereas the hydroxyl in the β-anomer is in the equatorial position. The aldose compound (the open-chain aldehyde form) is able to reduce metal ions (designated Me), whereby the aldose itself becomes oxidized to give a carboxylic open-chain form. Hydrogen atoms at C2, C3 and C5 are not drawn in the picture. (B) The β-1,4-linkages of D-glucose units in cellulose. Every glucose unit is rotated 180° relative to one another thus making cellobiose the repeating unit. The end of the molecule where the ring can open at the C1 carbon is called the reducing end.

Ulla Rudsander

4

II. Cellulose III is obtained by treatment of cellulose II with alkali solution and vacuum drying. The swollen cellulose crystals hold ions and water molecules within the lattice that influence the packing (Porro et al., 2007). Cellulose IV is obtained by heating cellulose III in glycerol. Cellulose III and IV are metastable and they revert to the parent polymorph (I or II) on hydrothermal treatment. All native celluloses are a composite mixture of two crystalline allomorphs, designated Iα and Iβ. The main component in higher plants is Iβ and in bacteria, tunicates and algae, the dominating component is Iα (Sturcova et al., 2004). These two crystal allomorphs differ in the way the sheet of parallel chains is stacked above the next, and heating as well as shearing can convert Iα into Iβ. Iα has a triclinic unit cell (Nishiyama et al., 2003), and Iβ has a monoclinic unit cell (Nishiyama et al., 2002).

Depending on the source of the microfibrils and how they are prepared, the microfibrils are surrounded by paracrystalline or “amorphous” regions, containing both cellulose and other polysaccharides (Lynd et al., 2002). Purified celluloses used as model substrates for studies of hydrolysis vary in structural details. Bacterial microcrystalline cellulose, BMCC, synthesized by Gluconoacetobacter xylinus (formerly Acetobacter xylinum), is the most highly crystalline substrate, but it contains mainly allomorph Iα, which will have positive implications for the enzymatic utilization (Hayashi et al., 1997). Avicel and Sigmacell are also highly crystalline, nearly pure celluloses from plant origin prepared by removal of hemicelluloses and amorphous regions. The insoluble substrate is difficult to attack by enzymes since the individual molecules are packed so tightly that not even water molecules can enter the crystalline array. But there are amorphous regions as well as irregularities such as kinks, voids and micropores. Some areas are spacious enough to allow enzymes to start breaking open the crystal. Phosphoric acid swollen cellulose, PASC, is an amorphous substrate where the strong acid treatment has removed hydrogen bonds. The difficulty of working with insoluble substrates has led to the use of soluble, artifical derivatives, carboxymethylcellulose (CMC) and hydroxyethylcellulose (HEC). Many enzymes active on PASC, CMC or HEC cannot hydrolyze crystalline cellulose, but are still referred to as ´cellulases´. The cellulose derivatives can be regarded as model substrates that reflect the cellulase activity with amorphous or disordered polymer. The enzymes capable of cellulose degradation are consequently a heterogenous group, reflecting the poor accessibility and/or heterogeneity of the substrate (Lynd et al., 2002).

2.2 Functional modules Efficient cellulolytic enzymes commonly have modular structures. One or two catalytic modules linked to one or several carbohydrate-binding modules is a frequent arrangement in aerobic microbes, and – sometimes – modules of unknown function are also attached (Lynd et al., 2002). The most extreme modular arrangements are found in anaerobic microbes, which produce enzyme complexes called cellulosomes. Cellulosomes are usually anchored to the microbial cell wall, and contain a scaffolding module that is anchored to cellulose by a cellulose-binding module. Some enzyme complexes can carry a large number of different catalytic modules degrading cellulose and hemicelluloses (Bayer et al., 2004). In anaerobic bacterial cellulolytic systems, there are additional modules associated with cellulase catalytic domains, called dockerins and cohesins, which are needed to anchor the enzymes to the scaffoldin proteins in cellulosomes, and fibronectin- and Ig-like modules of so far unknown functions.

One of the common modules in cellulolytic enzymes are the cellulose-binding modules (CBM). By definition, a cellulose-binding module is a contiguous amino acid sequence, with a discrete fold and independent binding activity. Most, if not all, cellulases effective against crystalline cellulose have a CBM. If the CBM is removed, the cellulase looses much of its

Substrates and enzymes

5

activity on insoluble substrates (Tomme et al., 1995). An ongoing debate concerns whether CBMs can influence the structure of cellulose when it binds, not only serving as an anchor but also to recruit accessible chains (Hildén, 2004). The idea is that the energy dissipated from the hydrolysis reaction can be channeled to the CBM, and used in disruption of the crystallite (Sinnott, 1990).

The catalytic modules break or make bonds in the enormous turnover of glucosyl residues within the biosphere. The overall topologies of the active sites of cellulases fall into two general classes, namely endoglucanses (EC 3.2.1.4) and cellobiohydrolases (EC 3.2.1.91). Endoglucanases have an open active site cleft (Fig. 2). They can bind randomly at any available point along a cellulose molecule, hydrolyzing one or a few bonds before dissociation. In contrast, cellobiohydrolases have evolved loops that form an active site buried in a tunnel. They initiate their action in an exo-type of attack, which means that they can mainly access the ends of a cellulose molecule. Cellobiohydrolases feed a single glucan chain into the tunnel from one end, followed by threading, bond cleavage, and then release of cellobiose from the far end of the tunnel (Teeri, 1997). Cellobiohydrolases act processively, so that several consecutive hydrolytic events occur without dissociation from the enzyme-substrate complex. The division of cellulases in endoglucanases and cellobiohydrolases is not absolute and many intermediate forms have been found. Within these two classes there is a continuum of active site topologies with varying numbers of loops of different lengths and mobilities. As a result, certain endoglucanases can bind the cellulose molecule in an endo-mode, and then, owing to changes in loop-conformations that enclose the substrate in a tunnel, proceed in a processive mode. Some cellobiohydrolases have mobile loops which allow a certain opening and closing of the tunnel and may lead to occasional endoglucanase activity on soluble and amorphous substrates (Varrot et al., 1999; Zou et al., 1999).

Figure 2. Computer graphic images of a homologous exoglucanase-endoglucanase pair (family 6) to illustrate their different active-site topologies. Owing to the presence of two surface loops, Hypocrea jecorina Cel6A (formerly Trichoderma reesei CBHII) (Rouvinen et al., 1990) has a tunnel-shaped active site. The absence or different conformation of these loops in the Thermobifida fusca endoglucanase E2 (Spezio et al., 1993) results in a more open active site, allowing easier access to internal glycosidic bonds.

Ulla Rudsander

6

The tunnel or cleft of cellulases creates binding sites for the glucosyl units of cellulose. The binding sites are numbered using the position of cleavage of the glycosidic bond at –1/+1 as the reference point (Biely et al., 1981; Davies and Henrissat, 1995). The sites are numbered by increasing negative integers towards the non-reducing end of a carbohydrate chain, and by increasing positive integers towards the reducing end. Conserved tryptophans, tyrosines, phenylalanines, or occasionally histidines mediate sugar binding through aromatic stacking interactions with the glucose ring (Fig. 3) (Divne et al., 1998). In fact, aromatic stacking interactions are a common feature when there is protein-sugar binding (Vyas, 1991). The secret behind aromatic stacking interactions is extensive nonpolar van der Waals contacts (≤ 4 Å). One of the tightest binding affinities among all protein/enzymes with carbohydrate ligands is observed in the chemotaxis receptor, Maltose-binding protein (MBP) (Quiocho et al., 1997). In this protein, there are sixteen aromatic residues located in or near the binding site groove, that form polar/nonpolar van der Waals contacts with the substrate. Aromatic stacking interactionss are quite flexible and allow sliding of the substrate by reducing energy-barriers between subsites. The indole ring can act like a hydrophobic platform and thus, discriminate between different carbohydrates. In addition, an intricate network of hydrogen bonds occurs between the polar groups on cellulose and amino acid side chains. The most common residues involved in hydrogen bonding are aspartate, asparagine, glutamate and arginine. Together, these interactions position the substrate in a required sugar conformation for bond cleavage, usually a sofa or half-chair conformation.

Figure 3. The binding of a single continuous glucan chain into the active site tunnel of Hypocrea jecorina Cel7A (formerly Trichoderma reesei CBHI) (numbered) (Divne et al., 1994). Sugar residues from the site of bond cleavage towards the nonreducing end of the chain are numbered –1 and –2 and those towards the reducing end of the chain are numbered +1, +2, etc. Only the stacking tryptophan side chains are shown. They twist the substrate in a required conformation.

2.3 Mechanism of cellulases All cellulases (EC 3.2.1.-) catalyze the cleavage of β-1,4-glucosidic bonds. The reaction takes place via general acid catalysis that requires two essential amino acid residues on the protein (Sinnott, 1990). In nearly all cases so far described these two residues are aspartates or glutamates (Davies and Henrissat, 1995). The hydrolysis occurs via two major mechanisms giving rise to either overall retention, or inversion, of the anomeric configuration of the substrate (Fig. 4). For enzymes that perform catalysis with inversion, one catalytic residue is a proton donor which protonates the glycosidic oxygen and thus promotes leaving group departure. The other catalytic residue acts as a general base to activate the nucleophilic water by deprotonation. An oxocarbenium-ion-like transition state is formed where electron density is distributed over no less than 8 atoms, and three bonds are broken and formed almost simultaneously (Withers, 2001). For enzymes that catalyze glycosidic bond cleavage with retention, these two residues consist of a catalytic nucleophile, which forms a covalent

Substrates and enzymes

7

glycosyl-enzyme intermediate, and a general acid/base residue. The general acid/base first protonates the leaving group then, following the formation of the covalent intermediate, activates the incoming nucleophile, often a water molecule. The glycosyl-enzyme intermediate is formed and hydrolysed via an oxocarbenium ion-like transition state. In addition to hydrolytic reactions, retaining enzymes can also perform transglycosylation reactions under low water content or high substrate concentrations. In a transglycosylation reaction, a glycosidic bond is broken and the glycosyl-enzyme intermediate is formed. However, in contrast to hydrolytic reactions, an acceptor molecule other than water is promoted to attack the glycosyl-enzyme intermediate to form a glycosidic bond. Inverting enzymes are not able to perform transglycosylation reactions, since this requires the enzyme to be anomerically indiscriminate. For example, if an acceptor such as a sugar attacked the glycosidic carbon in a reaction catalyzed by an inverting enzyme, the product would have a glycosidic bond with an inverted configuration. Since under thermodynamically favored conditions, any given reaction can run in the reverse direction, the inverting enzyme would have to accept both α and β configurations as substrate.

Although inverting hydrolytic enzymes are not able to perform transglycosylation, there is actually a type of hydrolytic enzymes active on oligo-saccharides that perform a transferase reaction with inversion of the anomeric configuration, GH-type phosphorylases, families GH65 and GH94 (Hidaka et al., 2006). Phosphorylases break glycosidic bonds, and generate glycosyl-phosphates as product, and in addition they can synthesize sugar chains using glycosyl-phosphate as sugar donors and glucose, for example, as acceptors. Notably, enzymic glycosyl transfer with inversion of the anomeric configuration can only take place when the acceptor is different from the donor (glucose is different from glycosyl-phosphate), so that anomeric discrimination still is valid (Fig. 19).

Figure 4. (A) General mechanism for an inverting glycosyl hydrolysis one-step mechanism. (B) General mechanism for a retaining glycosyl hydrolysis two-step mechanism (Withers, 2001).

Ulla Rudsander

8

2.4 Classification of cellulases In nature, carbohydrates are used as source of energy, structural building blocks, and also as recognition elements. All carbohydrate active enzymes have been classified into a number of different families based on amino acid sequence similarities, as initially detected by the hydrophobic cluster analysis (HCA) (Henrissat et al., 1989). So far there are more than 200 families displayed at the web page, www.CAZy.org, containing 110 glycosyl hydrolase (GH) families, 90 glycosyl transferase (GT) families, 18 polysaccharide lyase families and 15 carbohydrate esterase families. Comparative analysis has shown that enzymes in a given family have the same chemical reaction mechanism (retention or inversion of the anomeric configuration of the glucosidic carbon). Enzymes from the same family have the same general fold even though details of the structures can vary. In a given family, there are often many activities associated with that particular fold. The hierarchical classification according to the Enzyme Comission (EC) number is based on different levels of similar activity of an enzyme. Glycoside hydrolases catalyzing the glycosidic bond between two or more carbohydrates or between a carbohydrate and a non-carbohydrate moiety have EC number 3.2.1.-. Enzymes catalyzing the cleavage of the same type of bond but different substrates will have different EC numbers, as for example with xyloglucan-specific endo-1,4-β-glucanases (EC 3.2.1.151) and cellulose-specific endo-1,4-β-glucanases (EC 3.2.1.4). Also enzymes catalyzing the cleavage of the same substrate and the same bond but different type of enzymatic attack sometimes have different EC numbers, as with endo-1,4-β-glucanases (EC 3.2.1.4) and cellobiohydrolases (EC 3.2.1.91).

Enzymes with cellulolytic activities are found in 12 different CAZy families. Endoglucanases catalyse the hydrolysis of 1,4-β-linkages in the middle of the glucan chain, while cellobiohydrolases catalyse the hydrolysis of 1,4-β-linkages from the ends of the glucan chain. In two families (6 and 7) there are both endoglucanases and cellobiohydrolases and in one family (48) there are only cellobiohydrolases. In eight families (5, 8, 9, 12, 44, 45, 51 and 61) there are only endoglucanases. Families can be grouped into clans or superfamiles where the evolutionary relationship is more distant. A similar fold only states that the proteins share major secondary structures and the same topological connections. The structural similarities of proteins in the same fold category probably arise from the physics and chemistry of folding that favor certain packing arrangements. Proteins that belong to the same CAZy clan were shown to share structural and mechanistic similarity, with conserved catalytic machinery (www.cazy.org). The Structural Classification of Proteins (SCOP) was constructed from all proteins in the Brookhaven Protein Databank (PDB) by structural alignments and visual inspections to separate fold similarity and common ancestry (Lo Conte et al., 2000). Family GH 9 is not included in any CAZy clan, but it is included in a SCOP superfamily (named ‘Six-hairpin glycosidases’) containing all inverting glycosyl hydrolases with (α/α)6-barrel folds cleaving α− or β− glycosidic bonds.

Glycosyl hydrolases are traditionally named by the first two letters of the organism of origin, followed by the dominating biochemical activity, the GH family number and a letter (Henrissat et al., 1998). Since the genome-scale assignment of Arabidopsis and poplar CAZymes, there has been a nomenclature dilemma for plant biologists (Urbanowicz et al., 2007b). Plants contain so many CAZymes that the letters of the alphabet are sometimes not enough. Therefore, a new classification scheme for plant cellulases was recently proposed. According to this, the name of a gene contains an indication of the genus or species (e.g. At for Arabidopsis thaliana), the designated GH family (e.g. GH9), a letter indicating the modular structure (e.g. ‘A’ for a membrane-bound cellulase), and a number corresponding to a given isoenzyme (Table 1). Once the dominating biochemical activity has been determined,

Substrates and enzymes

9

the GH family number can be changed to a trivial name such as Cel or Xyn and the enzyme be given an appropriate EC number (e.g. EC 3.2.1.4 or EC 3.2.1.8, respectively). Table 1. Selected glycosyl hydrolase family 9 (GH9) proteins of Arabidopsis thaliana according to (Urbanowicz et al., 2007b) .

Chromosome Locus

GenBank ID

New Designation Featuresa Synonyms

At5g49720 U37702 AtGH9A1 CT, TM, GH9 Endo-1,4-β-glucanase 25, KORRIGANb, KOR1b

At1g65610 AC001229 AtGH9A2 CT, TM, GH9 Endo-1,4-β-glucanase 7, KOR2c At4g24260 AL078637 AtGH9A3 CT, TM, GH9 Endo-1,4-β-glucanase 21, KOR3c

At3g43860 AY072099 AtGH9A4 TM, GH9 Endo-1,4-β-glucanase 16 At1g70710 AY048283 AtGH9B1 SP, GH9 Endo-1,4-β-glucanase 8,

Cellulase 1d, AtCEL1d At1g02800 AF034573 AtGH9B2 SP, GH9 Endo-1,4-β-glucanase 1,

Cellulase 2e, AtCEL2e At1g71380 U17888.1 AtGH9B3 SP, GH9 Endo-1,4-β-glucanase 9,

Cellulase 3f, AtCEL3f At1g22880 AF000657 AtGH9B4 SP, GH9 Endo-1,4-β-glucanase 3,

Cellulase 5f, AtCEL5f At1g48930 AC016041 AtGH9C1 SP, GH9, CBM49 Endo-1,4-β-glucanase 5 At1g64390 AC066689 AtGH9C2 SP, GH9, CBM49 Endo-1,4-β-glucanase 6 At4g11050 AF080120 AtGH9C3 SP, GH9, CBM49 Endo-1,4-β-glucanase 19 aFeatures noted are cytosolic domain (CT), transmembrane domain (TM), signal peptide (SP), glycosyl hydrolase family 9 catalytic domain (GH9) and a family 49 carbohydrate binding module (CBM49). b(Nicol et al., 1998b) c(Molhoj et al., 2001) d(Shani et al., 2006) e(Yung et al., 1999) f(del Campillo et al., 2004)

2.5 Microbial cellulases Owing to its high energy content, wood is an attractive energy source for many microbes. Microbes occupy their own ecological niche, so that all different parts of the wood tissue are utilized by different organisms. Cellulolytic microbes, for example, are generally unable to use lipids or proteins as energy source, they are specialized for their task in breaking down the polysaccharides. The cellulose of plant cell walls is closely associated with other cell wall polymers, with diverse chemical characteristics. Lignified cellulose fibers are especially resisitant to biodegradation since the lignin structure is compact and complex. Before cellulolytic organisms can digest the fibre, cellulose has to be uncovered by microbes that carry ligninases, hemicellulases and pectinases. In addition, cellulose cannot enter the microbial cells, since it is a long, insoluble polymer. Therefore, complete degradation of cellulose always requires multi-enzyme systems, composed of a set of different endoglucanases and cellobiohydrolases, which are secreted to the culture medium of the cellulolytic microbe. These enzymes act synergistically, creating new sites for each other, so that the combined activity in a mixture is higher than the added activities of the same enzymes acting separately (Irwin et al., 1993). The oligosaccharides and cellobiose produced by the endoglucanases and cellobiohydrolases are then degraded to glucose by β-glucosidases (EC 4.2.1.21). Although the β-glucosidases do not operate on cellulose and thus can not be described as cellulases, they are nevertheless an essential component of complete cellulolytic enzymes systems.

Ulla Rudsander

10

Efficient cellulolytic enzyme systems are generally found among different bacteria and fungi. Aerobic and anaerobic microbes have different strategies to deal with the poor accessibility and high heterogeneity of the substrate. Aerobic organisms secrete a synergizing enzyme mixture to the medium while anaerobic organisms have multimeric complexes called cellulosomes that are attached to the outer membrane (Bayer et al., 2004). One possible explanation for the difference in cellulase organization is that anaerobic organisms are more energy limited than aerobic organisms. Thus it is more important for them to keep closely attached to the substrate and to ingest the cellulolytic products.

Cellulases from aerobic fungi have received more study than those of any other physiological group (Sukumaran et al., 2005). Fungal cellulases also dominate the industrial applications of cellulases. Among fungal aerobic organisms, Hypocrea jecorina (formely Trichoderma reesei) is a non wood-decaying filamentous fungus while Phanerochaete chrysosporium attacks wood. In particular, P. chrysosporium causes white rot by depolymerizing lignin (the brown color) before degrading cellulose, which leaves infected wood with a bleached appearance. In fungi, the induction of the entire cellulolytic enzyme system can be coordinated as in T. reesei or the enzymes can be individually induced as in P. chrysosporium. The anaerobic fungus Piromyces sp. from the digestive tract of herbivores produces cellulosomes similar to those isolated in C. thermocellum for the degradation of plant cell walls (Steenbakkers et al., 2002).

The best-studied species of aerobic cellulolytic soil bacteria belong to the genera Cellulomonas and Thermobifida. Cellulomonas species produce at least six endo-glucanases and one exoglucanase. Thermobifida fusca secretes three endo-glucanases, two exoglucanases and an unusual cellulase with processive endo-glucanase activity. Clostridium thermocellum is an example if anaerobic bacteria from rumen that produce cellulosomes, which contains a number of different catalytic modules and accessory proteins.

2.6 Glycosyl hydrolase family 9 Family GH9 (Cel9) is a taxonomically diverse, plentiful and ancient CAzy family containing cellobiohydrolases, endoglucanases and β-glucosidases (Davison and Blaxter, 2005). Glycosyl hydrolysis by this family is catalyzed with inversion of the anomeric configuration. There are representatives from plants, fungi, bacteria, amoebozoa, insects and marine animals. Previously, animals were not recognized as able to produce cellulases, and the cellulose digestion in insects and mollusks was explained by symbiotic organisms. A Cel9 enzyme from termite, Nasutitermes takasogoensis, has now been identified and its 3-D structure has been solved (Khademi et al., 2002).

Bacterial family 9 cellulases conform to four thematic modular arrangements (Fig. 2, Paper I) (Ding et al., 1999). The simplest theme is typical of many plant cellulases and consists of a catalytic domain only. A second theme is exemplified by C. thermocellum Cel9A which has an Ig-like module upstream of the catalytic domain. A third type also contains an Ig-like domain in the same position, but it includes an N-terminal family IV CBM. An example is Cel9B from T. fusca. The fourth modular arrangement is composed of a family IIIc CBM immediately downstream from its GH9 catalytic domain, as in Cel9A from T. fusca. The modular structure of family 9 cellulases is reflected in the sequence of their catalytic regions, which is apparent in a phylogenetic tree of the family. Fungal PcCel9A (Paper I) is most similar to TfCel9A (37%) among the full-length cloned family GH9 genes, but in the phylogenetic tree it seems to form its own branch, accompanied by a number of hypothetical proteins from Basidiomycetes. PcCel9A has a predicted C-terminal transmembrane helix.

Substrates and enzymes

11

To date, there are six cellulases from family GH9 that have crystallized and their 3D x-ray structures have been solved. The central domain is an (αα)6 barrel and a calcium site is highly conserved in the family. There are usually 6 glucosyl subsites in the active site, except in Cbh9A from C. thermocellum, which has additional loops that block off the binding cleft, to form 4 subsites (Schubot et al., 2004). CtCbh9A is the only example with exocellulase activity in in the family. An unusual endoprocessive activity is reported for three enzyme species. T. fusca Cel9A (Sakon et al., 1997) and Clostridium cellulyticum Cel9G both have C-terminal CBM III closely fused to their catalytic domain (Mandelman et al., 2003). C. fimi Cel9B (formerly CenC) has two N-terminal CBMs and two C-terminal domains of unknown function (Tomme et al., 1996). The processive endoactivity is manifested by synergy with both endo- and exocellulases. The endoprocessive attack is performed by cleavage at a random point along the glucan chain, and subsequent movement of the substrate through the active site cleft for the next hydrolysis event to occur, releasing cellodextrins ranging from G2 to G4.

T. fusca Cel9A, TfCel9A (formerly E4), is so far the best characterized cellulase from family GH9 (Sakon et al., 1997). The modular structure of TfCel9A is shown in Paper II, Fig 1. It contains two C-terminal CBMs and a C-terminal fibronectin-like domain. The four crystal structures of TfCel9A show the catalytic domain and cellulose-binding module, alone and also in complex with cellobiose, cellotriose and cellopentaose (Fig. 5). The catalytic module of TfCel9A contains 6 glucose binding subsites and aligns closely with a Family III cellulose-binding module containing further 6 subsites (Fig. 5). Together the two modules form a very long substrate binding site. Probably, TfCel9A binds with three parallel cellulose strands on the flat CBM surface, and the middle chain is fed into the active site. If the CBM is removed, the enzyme loses very much of its activity against both amorphous and crystalline cellulose (Irwin et al., 1998). The active site also has a blocking loop at the negative subsites, but its removal does not influence the activity profile of TfCel9A significantly (Zhou et al., 2004).

Figure 5. Computer graphic image of Tf Cel9A showing the CBM closely attached to the catalytic domain. The CBM is oriented at the back of the picture. In the binding channel, the proton donor (E) is colored in magenta and the nucleophile (D) is colored in red. The arrow points at the blocking loop (Sakon et al., 1997).

Ulla Rudsander

12

2.7 Classifications of plant cellulases Plants also produce cellulases, but for a somewhat different purpose than microorganisms. In plants these enzymes probably perform limited hydrolysis to permit cells to divide and differentiate. Some cellulases seem to participate in cellulose biosynthesis by a so far unknown mechanism (see section 4.3). Microorganisms typically have a battery of cellulases from many different families. In plant genomes, numerous cellulases from just one single family, GH9, are present (Henrissat et al., 2001). Some plants also have genes encoding family GH5 enzymes. However, these plant genes most likely encode mannanase homologues, judging from sequence alignments with genes of known enzyme activity (Geisler-Lee et al., 2006).

The first family GH 9 plant cellulase was isolated from avocado, by differential screening with cDNA probes of ripened and unripened fruit (Christoffersen et al., 1984). Since then, many new genes have been cloned and studied, preferentially from angiosperms (flowering plants). Plant Cel9s have been divided into three groups, based on phylogeny of DNA sequences and intron/exon organization (Fig. 6) (Libertini et al., 2004). The largest subclass, α, includes genes coding for secreted enzymes. There is some variation in sequence, and some genes encode a significantly longer C terminus of the corresponding protein. This C-terminal stretch has recently been shown to encode a family 49 cellulose-binding module in a tomato gene (Urbanowicz et al., 2007a). Enzymes with associated CBMs have been observed in cDNA fragments that encode cellulases from both angiosperms, gymnosperms, and mosses (Molhoj, 2002; Trainotti et al., 2006; Urbanowicz et al., 2007a). A distinct subgroup of the phylogenetic tree, γ, characterized by a number of deletions/insertions, is the membrane-anchored plant cellulases. They are considerably longer and bear a predicted single N-terminal membrane-spanning domain. Homologues have also here been found throughout the plant kingdom by Expressed Sequence Tags (ESTs). The third subclass, β, is very small and consists of genes encoding N-terminal signal peptide for protein secretion.

Figure 6. Phylogenetic tree of the full-length Populus and Arabidopsis cellulase gene family (adapted from Paper IV, Fig. 1). Type C GH 9 gene models are shaded in blue, type A GH 9 gene models are shaded in pink. Nine subfamilies by (Molhoj, 2002) denoted by roman numerals, three subgroups from (Libertini et al., 2004) denoted by the greek symbols α, β, and γ.

Substrates and enzymes

13

The phylogeny of GH 9 plant cellulases has also been investigated based on protein sequence (Molhoj, 2002) and functional studies. Nine different branches were identified in the dendrogram which includes all 25 deduced Arabidopsis amino acid sequences, and some full-length homologues in higher plant species (Fig. 6). The plant cellulases are expressed during leaf and flower abscission, the ripening of fruits, differentiation of reproductive structures and vascular tissue, and cell elongation. Enzymes involved in ripening and abscission as well as expansive growth appear to be differentially regulated by the plant hormones ethylene and auxin (Ferrarese et al., 1995; Wu et al., 1996; Catala et al., 1997; Ferrarese et al., 1998; Harpster et al., 1998; Catala et al., 2000).

According to the new recommendation for a standardized classification of the plant cellulases (Urbanowicz, 2007), there are three subclasses in family GH 9, designated by the letters A-C (Fig. 6). Subclass A contains the enzymes with a predicted N-terminal transmembrane domain, subclass B is composed of the secreted enzymes with a signal peptide, and subclass C enzymes have a C-terminal CBM. This nomenclature provides information about potential function, but does not suggest orthologous sequences that reflect the patterns of evolution of the catalytic domains. For example, the phylogenetic tree for family GH 9 enzymes based on DNA sequence and exons/introns by Libertini et al (Fig. 6) shows that there is a subgroup of type B secreted enzymes, β, with divergent evolution from the subgroup of enzymes denoted α, which contains both type B and type C enzymes.

3 Wood formation Forests cover roughly one third of the planet landscape (Plomion et al., 2001). From this immense resource, many products can be derived and wood products are the fifth most important trade in the world. Wood formation also represents a considerable sink for excess atmospheric CO2, thereby reducing a major contributor to global warming.

3.1 Wood anatomy Some regions in the plant, the meristems, are embryonic and continue to divide throughout the life of the plant. Meristematic tissues are found in the root and shoot of angiosperms, and also in stem tissue. The meristematic tissue in the stem is called the cambium, and it is a thin cell layer, giving rise to all different cell types in the vascular system of the plant. The vascular tissue conducts water and nutrients and allows different parts of the plant to specialize. It is composed of the xylem and the phloem. The xylem is the inner part of the tree stem, whereas the phloem is the outer part, just beneath the cortex and the cork. The xylem supplies the leaves and all other tissues with water and minerals from the root system, and the phloem transports food from the photosynthetic activities of the leaves. The xylem tissue consists of five different kinds of cells, parenchyma, ray cells, fibers, tracheids and vessels (Fig. 7) (Mellerowicz et al., 2001). Vessels form the vessel element, a linear file of cells that makes up a continuous tube where water flows. The ray cells are responsible for the radial transport of fluids in the stem. Xylem fibers and tracheids (sclerenchyma cells) provide mechanical support, and in hardwood species (angiosperms such as poplar and birch), they comprise the major component of the wood. Parenchyma cells are the only cells in the xylem that remain living at maturity, and they seem to have a role in lignification, providing lignin precursors. Phloem consists of four different cell types, parenchyma, fibers, sieve elements and companion cells. Phloem fibers and parenchyma are similar to the cell types with the same names in xylem. Sieve elements and companion cells derive from the same mother cell. The function of companion cells is largely unknown, and the function of sieve elements is to

Ulla Rudsander

14

conduct nutrients, mainly sucrose, through multicellular sieve tubes. The sieve elements also remain alive at maturity.

In softwood species (gymnosperms such as pine and spruce), vascular transport is less efficient and there is also a smaller number of differentiated cells in the tissue. Softwood xylem contains a small amount of ray cells and a lot of tracheids that transport the water and also provide mechanical support.

tracheids

Figure 7. Cross section through the stem of a woody dicot showing different cell types, according to Plant Growth and Development, Academic Press, by Donald E. Fosket.

3.2 Wood cell differentiation All cell types of the vascular tissue undergo secondary growth, which makes them rigid and able to support large plant bodies that maximize the exposure to light. There is a high proportion of cellulose in secondary growth, 40-90% depending on the species. When viewed using electron microscopy, the walls of most plant cells can be shown to consist of fibrillar structures embedded in amorphous material. The fibrillar structures are crystalline microfibrils of cellulose. The microfibrils are embedded in the matrix material, which is composed of hemicelluloses and pectins. In a typical xylem fiber, several wall layers are distinguishable, that have different microfibril orientations (Fig 8). The middle lamella is the border of the cell. Next to it is the primary wall, where the microfibrils are laid down in a disordered network. By contrast, in the three layers of the secondary cell wall, the microfibrils are highly ordered, in different angles relative to the cell axis. Three principal groups of plant cell wall polysaccharides can be distinguished on the basis of extractability: the pectins, the hemicelluloses and cellulose. The terms ‘hemicelluloses’ and ‘pectins’ encompass complex groups of polysaccharides with branched chains or chains arranged in networks. Hemicelluloses vary greatly in different species. In hardwood species, the principal hemicelluloses are xyloglucan, xylan and glucomannan. Xyloglucan is the major component of the hemicellulose fraction of primary cell walls in hardwood species, and it is believed to coat and form extensive cross-links with cellulose microfibrils. Pectic polysaccharides are made up of a group of molecules rich in galacturonic acid, rhamnose, arabinose and galactose.

Wood formation

15

Wood is synthesized in four different steps (Fig. 9), including cell division, cell expansion and elongation, cell wall thickening, and programmed cell death (Plomion et al., 2001). As the cell divides, vesicles accumulate and fuse to form the new wall, called the cell plate. Once the cell plate has fused with the pre-existing walls of the parent cell, material is laid down to form the primary wall, and the cellulose microfibrils are oriented in an irregular fashion. The primary wall is made up of a network of mainly cellulose, pectins and hemicelluloses. During cell expansion, the network is extended by weakening of noncovalent bonds between polysaccharides, and enzymatic cleavage of the polymer backbones and the crosslinks between polymers (Cosgrove, 1999). Many plant cells contain only primary walls so that wall synthesis stops after cell elongation. However in the xylem cell, when it has reached its definitive size, a new, strong layer is formed inside the primary cell wall. This secondary cell wall is divided into three different layers, S1, S2 and S3, which are composed of cellulose, hemicelluloses and lignin. The microfibril angle of the cellulose microfibrils in the thickest layer, S2, has a major impact on the elasticity of wood. Lignin is a phenolic polymer, almost always associated with secondary cell wall polysaccharides.

Figure 8. A diagrammatic representation of the different layers of the cell wall of a wood fiber, and cellulose microfibril orientation according to Plant Biochemistry, Academic Press, edited by P.M. Dey and J.B. Harborne. Soon after the completion of the wood cells, all of them die in the process of programmed cell death, except for the parenchyma cells. The nucleus and the cytoplasm disappear, and instead, thick, lignified cell walls remain. When the parenchyma cells eventually die after several years, heartwood is formed. This is the innermost part of the tree, and a specific role for heartwood has not yet been determined.

In response to non-vertical orientation, for example wind, growth on a slope or the weight of snow, a certain kind of abnormal wood can form. In hardwood species it forms in the upper side of a leaning stem, and is termed tension wood. The overall lignin content of tension wood is low and the cellulose content is higher. An extra layer, called the gelatinous or g layer, with an almost vertical microfibrillar angle, is deposited inside the fibers.

Ulla Rudsander

16

Figure 9. (A) Cross section of a hybrid aspen stem stained with Toluidine blue. Black bars indicate the location of the sampled tissues for the microarray experiments referred to in section 3.3. A, meristematic cells; B, early expansion; C, late expansion; D, early secondary wall formation; E, programmed cell death. (B) Schematic representation of different cell-types and stages during wood formation. Bars depict timing and extent of the different developmental stages and the appearance of the major cell wall components (Hertzberg et al., 2001).

3.3 Approaches to understanding wood formation When high-throughput DNA sequencing was established in the 1990´s, large EST (expressed sequence tags) data sets could be obtained, to study metabolic processes in the plant. The generation of thousands of sequences was necessary to be able to discover not only the most abundant transcripts. More than 5,000 ESTs were analyzed from the wood-forming tissues of poplar (Sterky et al., 1998), and since then the poplar database has been extended to over 100,000 genes, (http://poppel.fysbot.umu.se/ (Sterky et al., 2004)), covering all major tissues of the plant.

The introduction of shot-gun sequence assembly techniques made it possible to achieve complete genome sequences for many organisms. The genome of Arabidopsis was assembled in 2000 (ArabidopsisInitiative, 2000), and the genome of Populus trichcarpa was published in 2006 (Tuskan et al., 2006). The poplar genome contains about 1.6 times more CAZyme genes than Arabidopsis, which is more than any other species characterized to date (Geisler-Lee et al., 2006). The wood-forming tissues are the richest source of the various transcripts. Arabidopsis undergoes some secondary wall thickening, but phenomena such as extensive secondary growth, juvenile and mature phases in wood formation, and adaptations to frost and dormancy are unique to tree species (Mellerowicz et al., 2001). Comparative genomics will provide more insight into the genetic differences between an annual species such as Arabidopsis and a perennial, woody species.

Wood formation

17

The genus Populus comprises about 30 species native to the northern hemisphere, for example Europena aspen (Populus tremula), hybrid aspen (Populus tremula x tremuloides Mich) and black cottonwood (Populus trichocarpa). Populus offers several advantages as a model tree: it grows fast, it is easy to propagate and transform, it has a small genome (~500 Mbp), and it has also an economic importance in the forestry industry. Pine trees are particularly important in the Swedish forestry, but they are not as easy to handle in experimental research. Obtaining the full genome of pine is for example not easily feasible because of its very large genome size, ~15,000 Mbp. For comparison, the genome size of Arabidopsis is 125 Mbp, the genome size of P. trichocarpa is 550 Mbp, and the size of the human genome is 2,910 Mbp. A comparison of pine ESTs with Arabidopsis reveals that the genes responsible for wood formation are well conserved (Kirst et al., 2003). Arabidopsis and poplar are thus good models for the study of wood formation also when it comes to gymnosperms.

cDNA microarray analysis has the advantage that the mRNA expression of large data sets can be compared in a feasible way from different developmental stages, or in response to external stimuli. For example, the expression of 22,000 Arabidopsis known and unknown genes can be obtained from a web-based interface, www.genevestigator.ethz.ch, during selected stresses, growth stages and particular organs. Poplar microarrary analysis is also well-established. In 30µM thin tissue sections representing the four major steps of wood differentiation, the genes expressed during secondary wall deposition could be identified (Paper I, Fig. 1) (Hertzberg et al., 2001; Aspeborg et al., 2005). The genes involved in tension wood formation have also been studied (Andersson-Gunneras et al., 2006). Another example is the monitoring of CAZyme gene expression in different tissues, which was compared to the EST database (Geisler-Lee et al., 2006). Real- time PCR experiments can complement the results from high-through put methods, by carful single gene approach, where cross-hybridizations and biased library constructions are eliminated.

Just as many other plant genomes, the Populus genome has been subjected to a number of genome duplications (Adams and Wendel, 2005; Djerbi et al., 2005). The occurrence of many genes with functional redundancy can be a problem when studies are carried out by reverse genetics. Forward genetics is the creation of random mutatations in the genome of a plant, followed by screening for interesting phenotypes. Reverse genetics is the up- or down-regulation of a target gene by sense/antisense constructs, DNA insertional mutagenesis or RNAi (RNA interference) (Waterhouse and Helliwell, 2003). It might be better to use Arabidopsis as a model organism for reverse genetics, since the number of isozymes is somewhat lower, and the chance to find a clear phenotype is higher.

4 Biosynthesis of cellulose Cellulose production is usually considered to be mainly a plant activity, but there are a few other organisms that also synthesize cellulose, for example algae, protozoa and marine animals. Gluconoacetobacter xylinus (formerly Acetobacter xylinum) and Agrobacterium tumefaciens are two bacterial species that are well-known for their cellulose-producing abilities, and especially G. xylinus has served as a model organism.

Ulla Rudsander

18

4.1 Biosynthesis of cellulose in bacteria G. xylinus secretes cellulose Iα microfibrils from enzyme complexes of four subunits organized in linear rows in the plasma membrane. The microfibrils form bundles that in turn form a twisting ribbon. The bacterium uses the cellulose ribbon to float in the air-water interface. 14[C] -labelling experiments show that cellulose biosynthesis by G. xylinus occurs from uridine 5´-diphosphate-glucose (UDP-Glc) as substrate via lipid and protein-linked cellodextrins as intermediates (Swissa et al., 1980). A gene localized immediately upstream of the CesA operon, encoding a carboxymethyl-cellulose (CMC) active enzyme, was shown to be required for cellulose production (Standal et al., 1994). CMCax is a secreted enzyme and belongs to the inverting family GH 8. The x-ray 3D crystal structure of CMCax has been solved, and the overall fold can be described as an (α/α)6 barrel motif, although one helix, α11 is missing (Yasutake et al., 2006). This is similar to the barrel fold found in family GH 9, and these two families also belong to the same SCOP superfamily, suggesting a distant evolutionary relationship (Lo Conte et al., 2000). CMCax is active with cellopentaose or longer oligosaccharides. There are conserved stacking aromatic residues in the subsites -3 to +2.

There are interesting experiments carried out in G. xylinus where the assembly of the cellulose ribbon has been studied. Polymerization and crystallization was shown to be coupled processes (Benziman et al., 1980). If calcoflour white is added to the culture medium, the rate of polymerization decreases and the ribbon assembly is disturbed. Ribbon assembly is also loosened by the addition of CMC or xylan (Tokoh et al., 2002). Reorientation of the microfibrils have been observed during biosynthesis of cellulose in the presence of xyloglucan or pectin (Astley et al., 2003). Addition of the enzyme CMCax to the culture medium enhances cellulose production, and an active-site mutant enzyme, does not have this effect (Tonouchi et al., 1995). It is thus the endo-glucanase activity itself that is important for cellulose synthesis and crystallization, not the cellulose-binding ability.

Agrobacterium tumefaciens also produces Iα microfibrils, which the bacterium uses to attach to the plant host cell. In Agrobacterium tumefaciens, 14[C]-UDP-glucose was also shown to be a substrate by incorporation into cellulose. The CelC gene product is required for synthesis to occur. CelC encodes a membrane-anchored family GH 8 endoglucanase homologous to CMCax from G. xylinus.

4.2 Biosynthesis of cellulose in plants The non-cellulosic cell wall polysaccharides are synthesized in the Golgi apparatus in plants by membrane-bound glycosyltransferases (GT). They are packaged in vesicles that fuse with the plasma membrane to secrete their load (Brett and Waldron, 1996). In contrast, plant cellulose is synthesized at the plasma membrane. The growing microfibrils can be observed by freeze-fracture techniques, and attached to the tips of the microfibrils are the so called terminal complexes, TCs (Brown and Montezinos, 1976).

Cellulose synthase (CesA) genes from plants have attracted attention from scientists since the 1960´s. However it took a long time before any CesA genes could be cloned. In 1996, Pear et al reported the first cloning of plant CesA genes from cotton (Pear et al., 1996). Short stretches of highly conserved amino acids were discovered in hydrophobic cluster analysis of bacterial enzymes. These short sequences could then be used to probe plant homologs in a cDNA library from cotton fibers, at the onset of secondary wall formation.

Biosynthesis of cellulose

19

The cellulose synthases were shown to constitute the catalytic subunits of the terminal complexes, TCs, secreting cellulose microfibrils from the plasma membrane (Kimura et al., 1999). TCs occur as six-fold structures termed rosettes in plants (Fig. 10). A common model for rosette assemply is based on six globules within each rosette, and every globule is composed of three different CesA isoforms (Doblin et al., 2002). According to this model, one CesA makes one chain, and 36 glucan chains are synthesized simultaneously from every rosette, which requires a remarkable molecular coordination. These chains then form the crystalline network of the elementary cellulose microfibril. The genome sequence of Arabidopsis has provided a family of ten CesA genes (Somerville et al., 2004). The functions of six of them have been established by mutant complementation analysis. At least three appear to be necessary for primary cell wall development, while three others are associated with secondary cell wall deposition. In poplar, there are 18 CesA genes (Djerbi et al., 2005)and the expression patterns of some of them have been studied by microarray analysis and real-time PCR (Djerbi et al., 2004). PttCesA1, PttCesA2, PttCesA3-1, PttCesA3-2 and PttCesA4 seem to be transcribed during both primary and secondary wall synthesis, with upregulation of PttCesA1, PttCesA3 and PttCesA9 happening during intensive cellulose synthesis (Djerbi et al., 2004). PttCesA6 seems to be involved solely in primary wall synthesis.

β-1,4-glucan chain

cellulose microfibril

CesA Rosette subunit

Rosette

β

α1

α2

Figure 10. A model for the structure of the rosette (Doblin et al., 2002). (A) Each CesA subunit is shown to be involved in the synthesis of one β-1,4-glucan chain. Once the 36 chains merge from the rosette, they coalesce to form an elementary cellulose microfibril. (B) Six subunits, possibly containing six CesA polypeptides interact to form a rosette. Three different isoforms, α1, α2, and β are required for rosette assembly. All cellulose synthases belong to family GT 2, catalyzing nucleotide-diphospho α-linked sugar donors to form β-linked products such as cellulose or chitin (www.CAZy.org) (Fig. 11). GT2 is the largest CAZyme family at present, containing more than 3000 representatives. GT2s can either perform single-addition of sugar monomers (repetitive), or processive addition of multiple sugar monomers (polymerizing). CesAs are integral membrane proteins with eight putative transmembrane regions (Doblin et al., 2002). The glucan chain is supposed to be extruded through the membrane, maybe through a pore-like structure, into the wall. Polymerization probably occurs from the non-reducing end of the glucan chain (Koyama et al., 1997; Him et al., 2001). Different mechanisms of polymerization have been proposed, that account for the 180º flip of glucose monomers, cellobiose being the repeating unit of cellulose. The first model involved the simultaneous transfer of two UDP-Glc residues to the growing chain (Saxena et al., 1995; Koyama et al., 1997). Other models have proposed the synthesis and transfer of cellobiose (Albersheim et al., 1997; Kim et al., 2000). Structure analysis of a soluble, repetitive GT2 from Bacillus subtilis showed only a single substrate

Ulla Rudsander

20

binding site (Charnock et al., 2001). As a result, they proposed a model where two different CesA proteins may dimerize and form one catalytic center including two substrate subsites.

When cellulose is synthesized a continuous supply of UDP-glucose units from the cytoplasm is needed. The concentration of sucrose has been shown to affect the synthesizing activity (Haigler et al., 2001), which suggests that sucrose synthase, making UDP-glucose from sucrose, may be a key enzyme. A putative membrane-associated sucrose synthase has also been shown to coregulate on transcriptional level with cellulose synthases (Hertzberg et al., 2001).

Figure 11. The putative mechanism for an inverting nucleotide-sugar glycosyltransferase (Charnock et al., 2001). One carboxylate residue functions as a base to activate the acceptor sugar and at least one more carboxylate functions to coordinate a divalent metal ion associated with the UDP-sugar.

4.3 A membrane-anchored cellulase is an actor in plant cellulose biosynthesis

The first plant membrane-anchored endo-1,4-β-glucanase (cellulase) was isolated from a tomato fruit cDNA library using a degenerate oligonucleotide (Brummell et al., 1997). Only one other membrane-anchored cellulase had been reported at the time, the celC gene product required for cellulose synthesis in Agrobacterium tumefaciens. In a forward genetics screen for short hypocotyls mutants, the korrigan (kor) Arabidopsis mutant was found (Table 1 ) (Nicol et al., 1998a). The mutant displays defects in the primary cell wall, manifested as reduced cell elongation, irregular cell shapes and reduced cellulose content. Pectin composition is also altered, probably due to feedback mechanisms in the primary cell wall (His et al., 2001). The kor1-1 mutant has a T-DNA insertion in the promoter region of a membrane-bound endo-1,4-β-glucanase, the KOR1 protein. A more severe primary cell wall phenotype is kor1-2 (Zuo et al., 2000). In kor1-2, defective organogenesis was observed, along with cell plate deficiency, resulting in cell wall stubs and multinucleated cells. The cell elongation defect but not the cell plate defect, was rescued by a KOR1 protein being mutated in either of the two polarized targeting signal in the cytoplasmic tail of the protein. Kor1-2 has a deletion encompassing the whole promoter and 5´-UTR. Several point mutation alleles that affect the primary cell wall deposition have been identified, altered cell wall, acw1 (Sato et al., 2001) and radial swelling, rsw2-1,rsw 2-2, rsw2-3, rsw2-4 (Lane et al., 2001). These are temperature-sensitive mutants, which show abnormal cell expansion, increased pectin and reduced cellulose content at the non-premissive temperature 31ºC. Acw1 is identical to rsw2-1 and rsw2-2. All three alleles have point mutations, G429R, S183N and G344R, that are predicted to be on the surface of the KOR1 protein (Molhoj, 2002). Point mutations in the KOR1 protein have also been reported to lead to an irregular xylem phenotype, the irx2-1 and irx2-2 mutants, with cellulose deficiency in the secondary cell wall, but not in the primary cell

Biosynthesis of cellulose

21

wall (Szyjanowicz et al., 2004). In the irx2 mutants, secondary cell wall deposition occurs at the corners of the cells, which are the earliest sites of deposition. Also here, the mutations P250L and P553L are predicted to be on the surface of the protein (Molhoj, 2002). Collectively, these Arabidopsis mutants show that the membrane-anchored cellulase KOR1 is required for both primary and secondary cellulose synthesis and/or cellulose assembly in the plant.

While the KOR1 gene is most often expressed throughout the plant, there are actually two other membrane-bound endo-1,4-β-glucanases in Arabidopsis that show different expression patterns by the GUS fusion staining method (Molhoj et al., 2001). KOR2 is expressed in root hairs and the proximal parts of leaves and floral organs, and KOR3 is expressed in trichome support cells and bundle sheath cells. Multiple KOR genes are naturally also found in poplar, and the most abundant transcript in the cambial region EST library is PttCel9A1 (Sterky et al., 1998). PttCel9A1 is the closest homologue to KOR1. cDNA microarray expression analysis has shown that PttCel9A1 is coregulated with CesA1 and CesA3 in the wood developmental region of secondary cell wall deposition (Aspeborg et al., 2005). High expression of PttCel9A1 has further been observed in the tension wood EST library (Geisler-Lee et al., 2006). In Populus tremuloides, the coexpression of KOR1 protein with three secondary wall specific cellulose synthases was observed in tension wood, by both transcriptional and translational methods (Bhandari et al., 2006), lending more support to the notion that PttCel9A1 is involved in secondary wall deposition. In global microarray analysis of transcripts upregulated during tension wood formation, PttCel9A1 transcipts were slightly decreased, but this only implies that the abundance of the transcript was not limiting (Andersson-Gunneras et al., 2006). The same was true for one other secondary cell wall specific CesA.

The reports of subcellular localisation of KOR1 protein have been ambiguous. If the enzyme is required for proper cellulose biosynthesis, it would be expected to localize to the plasma membrane in the presence of its substrate. Immunoblotting revealed KOR1 in plasma membrane fractions as opposed to endoplasmic reticulum and tonoplast, but no endosomal markers were used in these experiments (Brummell et al., 1997; Nicol et al., 1998b). An inducible C-terminal Green Fluorescent Protein (GFP)-KOR1 fusion was localized to the cell plate during cell division in tobacco cells (Zuo et al., 2000), but the fusion protein was not tested for the complementation of the korrigan phenotype. In Arabidopsis, the 35S-promotor from cauliflower mosaic virus was used for constitutive overexpression of GFP fused to the N-terminus of KOR1. The GFP-KOR1 fusion failed to repeat the results of labelling in the cell plate, and instead KOR1 coclocalized with endocytic markers (Robert et al., 2005). However, intracellular endosome localization can be an artifact of 35S overeexpression. Alternatively, it can be a step in the natural recycling of plasma membrane complexes. A Yellow Fluorescent Protein (YFP) fusion of AtCesA6 has been observed to accumulate in the Golgi apparatus and later arrive at the plasma membrane (Paredez et al., 2006). In moss, rosettes have a lifetime of 20 min in the plasma membrane, and are then endocytosed (Somerville, 2006). If few copies of KOR1 are needed for a functional CesA complex, they would be very hard to detect by current microscopy techniques.

The association of KOR1 with three CesA proteins involved in secondary cell wall deposition was further tested by co-purification using a His-affinity resin (Szyjanowicz et al., 2004), and found to be negative. However, the low signal can have been caused by poor antibody recognition. Thus, there is no conclusive evidence about the cellular localization or protein association of KOR1.

Ulla Rudsander

22

Table 2. Published korrigan Arabidopsis mutants. korrigan mutants Phenotype Mutation Strength of allele

kor1-1 a

Reduced cell elongation (dwarf), reduced cellulose deposition,altered pectin in primary walls

T-DNA insertion in promoter, 200 Bp upstream of ATG

Reduced mRNA levels

kor1-2 b

Cell plate deficiency: cell wall stubs, multinucleated cells, defective organogenesis

1 Kb deletion encompassing promotor and 5’-UTR

Almost undetectable levels of mRNA and protein

rsw2-1,2,3,4 c

Radial swelling, cytokinesis (extra cells) and cell expansion abnormalities, reduced cellulose content

Point mutations predicted on surface of protein in three different alleles: rsw2-1,2:Gly-429-Arg rsw2-3: Ser-183-Asn rsw2-4: Gly-344-Arg

Thermosensitive phenotype, non-permissive temperature at 31˚C

irx2-2 d

Irregular xylem, no apparent radial swelling or growth defect, slightly smaller hypocotyl, cellulose deficiency in secondary cell wall, no primary cell wall defect

Point mutation in highly conserved amino acids: irx2-1: Pro-250-Leu irx2-2: Pro-553-Leu (close the GH conserved signature 2)

Decreased levels of protein

acw1e

Reduced cellulose content, increase of pectin at non-permissive temperature, soft inflorescent stems, dwarfism

Point mutations predicted on surface of protein: Gly-429-Arg (same as rsw2-1,2)

Thermosensitive phenotype, non-permissive temperature at 31˚C

a. (Nicol et al., 1998a). b. (Zuo et al., 2000) c. (Lane et al., 2001). d. (Szyjanowicz et al., 2004) e. (Sato et al., 2001)

23

OBJECTIVES OF THE PRESENT INVESTIGATION

The general objective of the present investigation was to study the role and function of family 9 enzymes in hybrid aspen. The specific goals were to:

• Clone and sequence full-length genes involved in wood formation in the hybrid aspen, Populus tremula x tremuloides

• Develop expression and purification systems for recombinant poplar cellulases • Analyze the functions and expression patterns of these cellulases in vitro and in vivo

24

25

RESULTS AND DISCUSSION

5 Recombinant expression of PttCel9A1 (Paper II) Purification of plant proteins from native sources can be feasible if the targeted tissue is easily available and devoid of contaminating activity. This was the case for xyloglucan endotransglycosylase (XET) 16A from cauliflower florets, where high purity was achieved after only two chromatographic steps (Henriksson et al., 2003). The purification from native sources of family GH 9 enzymes has proven more tedious. A rice family GH 9 enzyme was purified from root tissues by 5-step chromatography (Yoshida and Komae, 2006). Also, for the purification of strawberry endo-1,4-β-glucanase from ripe fruits, three rounds of cellulose affinity column chromatography were needed (Woolley et al., 2001). Finally, wall-bound and extracellular endo-1,4-β-glucanases have been purified by 4-step chromatography from suspension-cultured Populus alba cells (Park et al., 1993; Ohmiya et al., 1995). In addition, purification of proteins from stem tissue involves scraping the stem after the bark is pealed off, and grinding of wood cells with thick and lignified cell walls. Consequently, recombinant production in heterologous hosts is an attractive option for the study of proteins from cambial tissues. When dealing with membrane-anchored enzymes, chances to retrieve pure catalytic domains in sufficient amount for characterization work are even smaller with current techniques.

Escherichia coli is by far the best studied expression system for production of recombinant proteins. If sufficient amounts of protein can be obtained, it is the best option for optimization experiments and large-scale cultivations since it grows relatively fast. Two type B endo-1,4-β-glucanases from Pinus radiata have been expressed intracellularly in E. coli fused to an N-terminal His-tag, and purified to at least 500 ng amounts by a single step nickel-affinity column (Loopstra et al., 1998). The proteins were active against CMC. The carbohydrate binding module (CBM) of an endo-1,4-β-glucanase from rice of type C has been expressed in the periplasm of E. coli fused to bacterial TfCel6A catalytic domain (CD) (Urbanowicz et al., 2007a). Expression of the full-length protein generated two polypeptides consistent with proteolytic cleavage at the linker region between the CD and the CBD. The authors do not specify if the catalytic domain was soluble and active from their recombinant expression experiment. Attempts to recombinantly express PttCel9A1 have been carried out using E. coli as a heterologous host. The catalytic domain of PttCel9A1 with N-terminal His-tag was included in a high-throughput screen, directing the protein to the cytoplasm. A band with the correct molecular weight was detected by coomassie-stained SDS-PAGE (Dr Tan Tien-Chye, personal communication). Scale-up experiments will be carried out, using strains with enhanced ability to form disulfide bonds in the cytoplasm, since there are three possible cysteine pairs from the PttCel9A1 homology model (Dr Christina Divne, personal communication).

Recombinant expression of higher eukaryotic plant proteins often requires post-translational modifications which can be achieved in the yeast Pichia pastoris. P. pastoris can perform glycosylation, disulfide bond formation and proteolytic processing. P. pastoris also has the advantage of secretion into the growth medium which simplifies purification. In our lab,

Ulla Rudsander

26

plant XET from family GH16 has been produced at considerable levels (Henriksson et al., 2003; Kallas et al., 2005; Baumann et al., 2007). Recombinant production of active family GH 9 plant enzymes in P. pastoris has been reported for the catalytic domains of two secreted pepper endo-1,4-β-glucanase (type B) (Ferrarese et al., 1998), the catalytic domain of one membrane-anchored endo-1,4-β-glucanase from oilseed rape (type A) (Molhoj et al., 2001) and the catalytic domain of a tomato endo-1,4-β-glucanase joined to a carbohydrate binding module (type C) (Urbanowicz et al., 2007a). The yield of purified membrane-anchored endo-1,4-β-glucanase from oilseed rape was reported at 0.22 mg/L. The catalytic domain of PttCel9A1 was produced and purified in P. pastoris at a yield of 1.1 mg/L (Paper II). Screening transformants by gel electrophoresis and coomassie blue staining of expressed proteins was time-consuming. In the rear mirror, it is clear that several complicating factors contributed to this scenario. The rather low level of expression in combination with quite a low activity with the substrate used for screening was an important factor. Also, PttCel9A1 and the AOX1 protein have approximately the same molecular weight, and the AOX1 protein is induced by methanol from the same promoter as the target gene. Finally, Pichia has endogenous cellulase activity, which creates a disturbing background signal when the level of heterologous protein is low. PttCel9A1 was successfully expressed in P. pastoris after antibodies generated against solvent-exposed peptides were obtained to screen a large number of Pichia transformants. Activity was followed during purification by an assay with CMC(4M). Analysis of the purified proteins by gel electrophoresis revealed two highly glycosylated species. Both species were however identified as PttCel9A1 by mass spectrometric analysis of tryptic peptides.

6 Glycosylation analysis of recombinant PttCel9A1 (Papers I, II and unpublished work) A condition known as hyperglycosylation of N-linked glycosylation sites can arise during heterologous protein production. This is the addition of mannose residues, 50-150, to the original N-linked oligosaccharide core, Man8GlcNAc2 (Man = Mannose, 162 Da, Glc=glucose, GlcNAc= N-acetylglucosamine, molecular weight 203 Da). Hyperglycosylation of heterolougous proteins is usually not as common in P.pastoris as in Saccharomyces cerevisiae (Cereghino and Cregg, 2000). However, for PttCel9A1 extensive glycosylation was unfortunately detected. O-linked glycosylation has been reported in some heterologous plant proteins produced in Pichia, for example barley α-amylase 1 and 2 (Juge et al., 1996). O-glycosylations are assembled onto the hydroxyls of serine and threonine but there is no consensus primary amino acid sequence for O-linked sites. N-linked glycosylation occurs more often than O-linked glycosylation in P. pastoris (O'Leary et al., 2004). The catalytic domain of PttCel9A1-1 has 8 predicted N-glycosylation sites, according to the recognition sequence Asn-X-Ser/Thr where the asparagine residue is modified.

Growing crystals for 3D x-ray structure determination usually requires milligram amounts of homogenous, enzymatically active and stable protein. In the fermentor, pH, aeration and carbon feed can be controlled. A fermentation can lead to ultra-high cell densities and increased amounts of protein, which can be further processed by various deglycosylation enzymes if needed. PttCel9A1 was cultivated in a 5 L fermentor in minimal medium at a yield of 2 mg/L (Fig. 12, previously unpublished data) compared to 1 mg/L in shake flasks. The purification followed the published scheme (Paper II). For comparison, the yield of PttXET16-34 was 12.4 mg/L from production in shake flasks (Kallas et al., 2005), and 54 mg/L in a fed-batch fermentation (Bollok et al., 2005). A reason for the much lower yield of PttCel9A1 could be the extensive glycosylation. Two smeared protein species were

Recombinant expression

27

distinguished in SDS-PAGE analysis, just as in shake flask experiments (Fig. 12 lane 1). The lower band appears at ~66 kDa (Fig 13, lane 3) and the higher band is larger than 97 kDa (Fig. 13, lane 5). The higher and the lower species were equally active with CMC(4M), and both species were converted to a fuzzy band at ~60 kDa after prolonged incubation (3 hours). Deglycosylation by Endo H of the high molecular weight species seemed to proceed via the lower species as an intermediate (Fig. 13). Endoglycosidase H cleaves the chitobiose core of high-mannose-type N-glycans to leave a single asparagine-linked GlcNAc residue (Trimble and Maley, 1984). The molecular weight of the PttCel9A1 catalytic domain is 58 kDa. Native sizes for GH9 membrane-anchored endo-1,4-β-glucanases have been reported at 53, 88, 70, 72 and 93 kDa in Arabidopsis, tomato and oilseed rape (Brummell et al., 1997; Nicol et al., 1998a; Molhoj et al., 2001), suggesting many native glycoforms. Endo H treatment of recombinant PttCel9A1 did not result in complete removal of N-glycan structures. This result indicates that there might be O-glycosylation present or N-linked glycosylation blocked by for example phosporylation. Unfortunately, the protein also lost over 90% of its activity after Endo H treatment, indicating unfolding or aggregation.

Figure 12. Fed-batch fermentation of PttCel9A1. Total protein content was quantified by the Bradford reagent. OD was measured at 600 nm at appropriate dilutions. Activity on Azo-CM-cellulose (Megazyme) was quantified as the absorbance change at 595 nm from 100 µl supernatant incubated at 30˚ C for 30 min.

Figure 13. Silverstained SDS-PAGE of 0.5 µg recombinant Δ 1-105 PttCel9A1 (lanes 1, 3, 5) purified from fermentation experiment. 5 mU Endoglycosidase H (Roche) was incubated with 0.5 µg recombinant Δ 1-105 PttCel9A1 at 30˚ C for 1 hour in 50 mM NaAcetate buffer pH 5.5 (lanes 2, 4, 6 ). Fraction containing low and high kDa species (lane 1 and 2), fraction containing solely low species (lanes 3 and 4), fraction containing solely high species (lanes 5 and 6). The molecular weight of Endo H is 29 kDa.

Ulla Rudsander

28

Deglycosylation by (1,2)-(1,6)-α-mannosidase from jack bean is a milder treatment since it will trim the α-(1,2)-(1,6) bonds of the glycosylated tree. Minor changes in the glycoform pattern were observed from treatment with α-mannosidase, taking into account the addition of mannosidase subunits at 66 kDa (figure 14, comparing lanes 1B and 4B). The original activity with CMC(4M) was fully retained after incubation with α-mannosidase at 30˚C for 3 hours. KOR1 protein was identified in a study about phosporylation sites from Arabidopsis plasma membrane proteins based on identification of phosphopeptides by mass spectrometry (Nuhse et al., 2004). Phosphorylation of the oligosaccharide tree of the protein can block the action of deglycosylation enzymes. Consecutive treatment with Calf Intestine Alkaline Phosphatase (CIP) and α-mannosidase, followed by phospho-staining, revealed phospomannosylation (Fig. 14, lanes 1A-4A). The stepwise decrease in molecular weight for the lower protein species (Fig. 14, lanes 1B-2B) results from the addition of enzymatic subunits at 65 kDa and 66 kDa, one after the other. Lane 3B shows a decreasing molecular weight, although in this case, only an extra 0.5 µg α-mannosidase was added. Thus, in lane 3B there is a visible change in migration for the lower species in the SDS-PAGE after consecutive deglycosyltion/dephosphatase treatment. However, the higher molecular weight species was unaffected.

A totally different strategy was then pursued in order to reduce the degree of glycosylation of PttCel9A1. The approach chosen was based on glyco-engineering of P. pastoris by introducing 1,2-α-D-mannosidase from H. jecorina into the genome of Pichia and/or knocking out endogenous mannosyltransferase (OCH1) (previously unpublished data). This approach has been previously used to produce 90% homogenous protein-linked oligosaccharides from hyperglycosylated H. jecorina mannosidase and the plasmids were obtained from the authors (Vervecken et al., 2004). Knocking out the mannosyltransferase (OCH1) of the Pichia transformant leads to protein glycosylation of the Man8GlcNAc2 type, referred to below as Glycoswitch 8. Strain engineering by knocking in 1,2-α-D-mannosidase from H. jecorina and knocking out mannosyltransferase (OCH1) leads to the synthesis of N-linked oligosaccharides of the Man5GlcNAc2 type, referred to below as Glycoswitch 5. The lower protein species purified from Glycoswitch 8-engineering of P. pastoris (Fig 15, lane 1B) had a decreased size, similar to dephosphatase/demannosylation treated sample (Fig 14, lane 3B), but unfortunately, the higher species was not affected. PttCel9A1 produced from Glycoswitch 5-engineering of P. pastoris (Fig 15, lane 2A and 2B) did not show any visible change in migration in the SDS-PAGE. The partial success with 8-Glycoswitched PttCel9A1 could be useful in large-scale fermentation experiments, where milligram amounts of the modified lower molecular weight species could be produced and purified.

Glycosylation analysis

29

Figure 14. Pro-Q Diamond (Molecular probes) Phosphoprotein gel stain (A) and SDS-PAGE (B) of 2 µg recombinant Δ 1-105 PttCel9A1 (lanes 4A, 4B), treated with 1 µg α-mannosidase from jack bean (Sigma-Aldrich) at 37˚C pH 5.5 for 1 hour (lane 1A, 1B), treated with 1 µg α-mannosidase at 37˚C pH 5.5 for 1 hour + 2 units Calf Intestine Alkaline Phosphatase (CIP, Fermentas) at 37˚C pH 7.5 for 1 hour (lanes 2A, 2B), treated with 1 µg α-mannosidase at 37˚C pH 5.5 for 1 hour + 2 units Calf Intestine Alkaline Phosphatase at 37˚C pH 7.5 for 1 hour + 1 µg α-mannosidase at 37˚C pH 5.5 for 1 hour (lanes 3A and 3B). Molecular weight of the two subunits of α-mannosidase = 44 kDa and 66 kDa. Molecular weight of CIP subunit = 65 kDa.

Figure 15. Western blot (A) and SDS-PAGE (B) of recombinant Δ 1-105 PttCel9A1 purified from engineered P. pastoris strain GS115. Lanes 1A and 1B contain 1 µg Δ 1-105 PttCel9A1 produced from Glycoswitch 8-engineering of Pichia, lanes 2A and 2B contain 600 ng Δ 1-6105 PttCel9A1-1 produced from Glycoswitch 5-engineering of Pichia. To evaluate whether glycosylation of some potential sites could directly influence the activity, a homology model was made using bacterial GH9 structures as templates. The putative N-glycosylation sites were mapped in the homology model. Among the 8 predicted N-glycosylation sites, 6 are located on surface-exposed loops of the protein (Fig. 16). Asparagine 408 is situated in an α-helix which makes it very unlikely to be glycosylated. The rotamer of asparagine 324 that is predicted by the model would be buried in the protein. An alignment with five membrane-anchored cellulases for P. trichocarpa and full-length sequences from Arabidsopsis, tomato, oilseed rape and barley, shows that among the remaining 6 potential N-glyc sites, only four are strictly conserved (shown in blue, Fig. 16). PtCel9A5 has a different modular arrangement than the rest of the proteins, and is therefore treated as an outlier.

Another plant enzyme, BobXET16A, containing only one N-linked glycosylation site, was purified from the native source (cauliflower) which allowed the fractionation of a single, pure glycoform, while heterologous production in Pichia produced a range of protein glycoforms (Henriksson et al., 2003). Deglycosylation of recombinant BobXET16A resulted in a substantial decrease in activity, showing that the glycosylation was necessary for the folding of the protein or involved in substrate binding. In future work, to reduce the heterogeneous

Ulla Rudsander

30

nature of PttCel9A1, each N-glycosylation site could be removed step by step by point mutations, to see which ones are necessary for folding and/or activity. If hyperglycosylated sites are necessary for folding, the best way to achieve PttCel9A1 suitable for crystallization would then be to switch to a plant expression system.

Figure 16. Ribbon drawing showing the eight putative N-glycosylation sites in a theoretical model of PttCel9A1. The protein backbone is colored yellow with α-helices shown as spirals, β-strands as arrows and loops as coils. A cello-nonasaccharide has been modeled in the protein substrate-binding canyon. Putative N-glycosylation sites refer solely to asparagine residues that satisfy the criterium of occurring in an N-glycosylation sequence pattern Asn-X-Thr/Ser where X≠Pro. N-glycosylation patterns that appear conserved are colored blue whereas those that are more diverse are in red. Assuming that there are no large discrepancies in conformation between the theoretical and true structure of PttCel9A1, asparagine residues 324 and 408 are incompatible with N-glycosylation since these Asn side chains are buried (courtesy of Dr Christina Divne).

7 Substrate analysis of GH9 microbial and plant cellulases (Papers I, II and III)

Hypotheses about substrate interactions and site-directed mutagenesis strategies can also be put forward from computer-generated models, in the absence of an experimental crystal structure. The protein alignment with known experimental structures is then crucial. The homology model of PttCel9A1 in Paper I was made using a CLUSTALW multiple alignment and the software Modeller 4.0 using the crystal structure of TfCel9A as the template. The refined model in Paper II was done by manual alignment and inspection to optimize possible bonding interactions and taking into account putative disulfide formation in addition to hinge regions. A few residues involved in substrate binding were different in the refined model, the most important one being the aromatic stacking partner in subsite -4, W256 in TfCel9A. In model I, tyrosine 370 was located in a position to stack with the -4 glucosyl moiety, and in model II, glycine 371 was found to occupy the corresponding coordinates. The loop conformations around the active site are often more variable than other parts of a protein and

Substrate analysis

31

are therefore hard to predict, especially if the subtrate binding loops are longer than the template or cannot be meaningfully aligned against structure sequence. For the PttCel9A1 model structure, there is a loop preceeding the -4 subsite. This was termed ´the blocking loop´ in the TfCel9A experimental 3D structure, because it looks like the loop blocks off the negative subsites of the active site cleft. The interpretation of the ´blocking loop´ can explain the differences between the automated and the refined model.

PttCel9A1 is 35 % identical to TfCel9A and protein sequences with over 30% identity to a known structure can often be predicted with quite good accuracy (Xiang, 2006). In PttCel9A1, the aromatic residues that are involved in substrate interactions are conserved at subsites +1 and +2, but surprisingly not at subsites -2 through -4. The effect of the amino acid substitutions W256A, W209S and W313G in TfCel9A subsites -2, -3 and -4 respectively, was investigated, to explore the role of these aromatic residues in family GH 9 microbial and plant enzymes. Most microbial cellulases can cleave β-1,4-glycosidic linkages in a variety of substrates such as cellulose with different degrees of crystallinity, CMC, (β−1,3), (β-1,4-)-D-glucan, xyloglucan and xylan (Table 3).

Table 3. Endo-1,4-β- glucanase substrates. Barley β-glucan has longer stretches of β-1,4-glucan between the β-1,3-glucosidic bonds than lichenan Xyloglucan and xylan are branched polymers.

Polymer sugar backbone linkage CMC glucose β-1,4 Xyloglucan glucose, xylose branches β-1,4 Xylan xylose β-1,4 Lichenan glucose (β-1,3)(β−1,4) Glucomannan glucose, mannose β-1,4 barley /cereal β−glucan glucose (β-1,3)(β-1,4)

The substrate specificity is not determined by the catalytic residues but rather by the number and the nature of substrate-binding sites. Planar stacking aromatic amino acids often discriminate against branched oligosaccharides (Hilge et al., 1998). In Family 9, TfCel9A and TfCel9B primarily act on cellulose (Irwin et al., 1993). However, many other microbial enzymes in Family 9 have broader substrate specificities. Family 9 Cel9T from C. thermocellum was shown to have strong activity toward CMC and barley β-glucan (Kurokawa et al., 2002). Also, CfCel9B (CenC) from C. fimi was active on CMC, barley β-glucan, lichenan and, to a lesser extent, glucomannan (Tomme et al., 1996). Finally, CtCel9A (CelD) has been shown to catalyze the hydrolysis of cellulose, lichenan and cereal β-D-glucans (Joliff et al., 1986). The crystal strucutures of CfCel9B and CtCel9A have been solved, revealing a reduced number of tryptophans contributing to stacking interactions with sugars in the substrate (Table 4). However, mutation of W256, W209 and W313 did not cause major changes in the activity profile of TfCel9A with different polymeric substrates (Fig. 5, Paper II), suggesting that other interactions are also involved in determining the substrate specificity of this enzyme.

Ulla Rudsander

32

Table 4. Glycosyl subsite mapping in bacterial and plant Family 9 cellulases. . Sitea TfCel9A PttCel9A1 AtKOR1 PttCel9B3 PcCel9A CtCel9A CfCel9B TfCel9B

-4 D261 D D D D D G G D262 N N N E N S W

W256 G G – A A D G-3 W209 Sc S Y S V S N

Y318 Y Y Y P T Q A

-2 F205 I K F D E A P

Y206 F F Y A Y Y Y R317 Q Q Qd N G S S

W313 P P E D W W W

D428 A A Y D N N N

-1 D55 D D D D D D D

D58 D D D D D D D

E424 E E E E E E E

+1 H125 H H H N F T T

Y420 Y Y Y W Y W W

H376 H H H H H H H R378 R R R A R R R

BLb 245-255 + + – + / – – – – a The reducing end of a glycosyl chain is bound in subsite +1, and the non-reducing end in subsite -4. b Blocking loop in T. fusca Cel9A. cTyr219 in PttCel9A1 (Asn207 in TfCel9A) may be able to compensate this interaction. dNot the equivalent backbone position. Hydrophobic stacking aromatic partners in TfCel9A are shown in bold. Functional characterization of type A plant cellulases has been carried out, and the function in cellulose biosynthesis by un unknown mechanism is well documented (see section 4.3). Biochemical analysis indicates a role in hydrolysis of non-crystalline or paracrystalline cellulose (Paper II). Biochemical and functional characterization of type B enzymes has not been as extensive, but still there is some evidence that allow us to speculate about the in vivo function. The first plant cellulases that were cloned were type B enzymes, not showing any CBM attached to the catalytic domain (Brummell et al., 1994). This is why doubts were raised about their ability to hydrolyse cellulosic substrates since CBMs are such an important feature of bacterial and microbial cellulases. Indeed, overexpression of type B Populus alba PopCel1 and PopCel2 in Arabidopsis, resulted in effects on both cellulose microfibrils and xyloglucan network (Table 5). However, in vitro substrate specificities of type B PaPopCel1 and PaPopCel2 were shown to be mainly “amorphous” cellulose (Ohmiya et al., 1995). Since xyloglucans are believed to coat and form cross-links with the cellulose microfibrils, the rearrangement of the xyloglucan network could be explained by altered microfibril structures and subsequent decrease in xyloglucan cross-links. The function of type B enzymes would thus be in hydrolysis of exposed non-crystalline regions to increase cell wall plasticity during differentiation processes, mainly cell expansion.

Substrate analysis

33

Table 5. Characteristics of dicot plant cellulases from family GH9, ND=not determined.

Plant cellulase type

Substrate specificity of purified enzymes

Cell wall phenotypes

Expression in Ptt EST library

Number of Pt gene candidates

Potential subsite stacking aromatics

Overnight hydrolysis of cello-oligosacharides

A (membrane-anchored)

amorphous cellulosej

Transgenic overexpression: increase in amorphous cellulose, no change in crystalline cellulosei

cambium, tension wood, wood cell deathi

5i - ≥ G5j

B (secreted)

amorphous cellulose, lichenana

Transgenic overexpression: reorganizaion of celluloseb,c, decrease in cross-linked xyloglucanc,d

cambium, tension wood, shoot apical meristemi

18i -3 Y/F and/or -4 Y/Fk

ND

C (secreted and CBM-anchored)

barley β-glucan, arabinoxylane, amorphous cellulose, xyloglucan f

Upregulation during fruit ripening: disorganization of celluloseg, decline in glycan fractionh

flower bud, young leafi

3i -2 Wl ≥ G4e

a(Ohmiya et al., 1995), b(Ohmiya et al., 2003), c(Ohmiya et al., 2000), d(Park et al., 2003), e(Urbanowicz et al., 2007a), f(Woolley et al., 2001), g(Trainotti et al., 2006), h(Harpster et al., 2002a), i Fig. 1, Paper IV, j(Master et al., 2004),k see alignment Figure 17, l see alignment Paper III, Figure 9

Recently, two separate groups published the biochemical characterization of type C plant cellulases, from rice root tissue (Yoshida et al., 2005) and from tomato fruit (Urbanowicz et al., 2007a). Type C plant cellulases contain a CBM composed of approxinmately 130 C-terminal amino acids. As shown in Table 5, type C plant cellulases are found in poplar tissues undergoing irreversible cell wall disassembly (flower) (Bernier et al., 1993) or rapid cell elongation (young leaf) (Ohmiya et al., 2003). Cells in fruit and flower tissue do not in general contain secondary cell walls, which is why type C cellulases could be exclusively active on primary cell walls. It is therefore interesting to take a look at the composition of primary cell walls in dicot species. β-glucan and arabinoxylan occur in very low amounts in the primary cell wall, instead the hemicellulose fraction is composed mainly of xyloglucan (Hayashi, 1989). In type C enzymes, high activity against β-glucan and arabinoxylan together with low activity on xyloglucan is then a bit puzzling. However, the substrate analysis can be indicative of non-cellulosic substrate specificity in vivo since in vitro analyses never take into account the fact that xyloglucans may change their conformation in contact with other cell wall polymers. Free xyloglucan has a twisted backbone conformation wheras xyloglucan adsorbed onto or between cellulose microfibrils would be predicted to adopt extended as well as coiled conformations upon adsorption (Hanus and Mazeau, 2006). Another line of evidence comes from the study of chemical composition at the onset of type C transcripts (Table 5). During fruit ripening of strawberry, the induction of two family GH 9 genes have been reported, type B (faEG1) and type C (faEG3) transcripts (Trainotti et al., 2006). The cell wall

Ulla Rudsander

34

changes in both the cellulose and hemicellulose fractions during ripening could of course result from the combined action of at least these two enzymes. However, evidence that one of the ripening-specific type B enzyme was active on amorphous cellulose or matrix glycans (i. e. glucomannan) was presented in transgenic work from pepper fruit (Harpster et al., 2002a; Harpster et al., 2002b). Regardless, the physiological process of fruit ripening is complex, but type C enzymes and their CBM play an important role. Binding studies of the type C CBM revealed (Urbanowicz et al., 2007a) specificity towards both crystalline and amorphous cellulose, but the authors did not report if binding to xyloglucan was tested. Thus, whether the substrate specificity of the catalytic domain is towards non-crystalline cellulose or in muro xyloglucan remains to be determined. Still, an important conclusion regarding the activity of GH9 plant cellulases is that none so far characterized are able to degrade crystalline cellulose, and this holds true despite the recent characterization of CBM-anchored type C enzymes.

The tryptophan in the -2 subsite for type C plant cellulases is conserved among all animal and bacterial family GH9 enzymes (Watanabe and Tokuda, 2001; Mandelman et al., 2003). In contrast, all type A and type B plant cellulases lack an aromatic residue that can perform stacking interactions in the -2 subsite (Paper III, Fig.9). Detection of cellotetraose hydrolysis by type C enzymes reflects a higher catalytic turnover number that is probably going to apply also for longer substrates (Table 5). This line of evidence agrees well with our results on the catalytic efficiency of the W313G TfCel9A mutant enzyme, proving that this residue is very important for a reactive conformation in the active site cleft (Paper II, Table 4). The significance of aromatic residues in position for hydrophobic stacking interactions in subsites -3 and -4 of type B secreted enzymes is not yet clear (Fig. 17), but might be substantiated when more genes are recombinantly expressed and characterized. Tyrosines and/or phenylalanines in position for hydrophobic stacking interactions might enhance the catalytic efficiency as was observed for TfCel9A in comparison with TfCel9B (Paper II, Table 4). The difference in catalytic efficiencies would then be especially striking for PtGH9B4, showing tyrosines in both positions, and for example PtGH9B6, lacking aromatic residues in both these positions (Fig. 17).

To resolve uncertainties about product formation from the hydrolysis of short cello-oligosaccharides in Paper II, Table 2 (cellotetraose > cellobiose, cellotriose > cellobiose), an extended real-time study by NMR spectroscopy was initiated. Equimolar amounts from hydrolysis events were shown with new standard curves. Since the mechanism of hydrolytic cleavage is governed by the spatial orientation of catalytic residues and since structure is dictated by sequence, the stereochemical outcome of the reaction is conserved within any given GH family. No exception to this rule has been reported (Henrissat et al., 2001), and the catalytic residues from other GH9 enzymes are conserved in the plant enzymes. In the NMR study, the expected inverting configuration of the anomeric proton after hydrolysis was observed for PttCel9A1 (Fig. 18). The signal for anomeric H1 in α-configuration at 5.224 ppm was increasing at the expense of the anomeric H1 in β-configuration, which can only occur if the stereochemistry of the reaction is inverting. It has been repeatedly suggested that the role of a cellulase required for cellulose biosynthesis, is not in the hydrolytic cleavage of the polymer, but instead in the addition of monomers to cello-oligosaccharides or glucan chains through transglycosylation (Matthysse et al., 1995; Brummell et al., 1997; Ohmiya et al., 2003; Robert et al., 2004). However, the inverting mechanism from our study contradicts this possibility, since transglycosylation to an identical acceptor never takes place in inverting enzymes (Koshland, 1953).

35

320 330 340 350 360 370 380 390 400 410 420 | | | | | | | | | | | PtGH9A5 EARMYYDSSGYQDELIWGGTWLFFATGNTSYLGYATS-NFSAAEGE---------ETASELGVFYWNNKLTANAVLLTRLRYFHDLGYPY---EVGLGSSSNKIDLLICSYLSRE---IY PtGH9B1 -CPFYCSYSGYKDELLWGAAWLFRATNEMSYYNIFKS-----LGAD-----------D-QPDLFSWDNKYAGVHVLLSRRALLNNDK--------NFEQFEGEAESFMCRILPNSPYKTT PtGH9B10 -CPFYCDFNGYHDELIWGAAWLSKATQDPKYWDYVVK-NMATLG--------------GSIFEFGWDSKHSGINIIVSPIYFDPQKV-M----SSSGSSFITNADSFVCSLLPESPTKSV PtGH9B11 -CPFYCDFSGYEDELIWGAAWLYRATKAPNYWSYVVQ-NISNLEKNVAKHTDRVGYGGGSFAEFGWDTKNAGINILVSKLLLSSK--------TSDVGPFIPNADKFVCTVLPESPTVYV PtGH9B12 -CPFYCSFSGYNDELLWSATWLYKATTKPMYLKYIKE-EATS----------------AAVAEFSWDLKYAGAQVLLSKLYFEG---------VKDLESYKKDADSFICSVLPGSPFHQV PtGH9B13 -ATYYN-STGYGDELLWAASWLYHATGDKSYLQYVTGQNGKLFAQW-----------G-SPTWFSWDNKLAGAQVLLSRLTFLGNKD-TS---NSGLQMYRKTAEAVMCGLIPDSPTATK PtGH9B14 -STYYN-STGYGDELLWAASWLYHATGDKSYLQYVTGKNGEVFAQW-----------G-SATWFSWDNKLAGTQVLLSRLTFFGNKD-TS---NSGLQMYRKTAEAVMCSLIPDSPTATQ PtGH9B15 -QNYYN-STGYEDELLWAAAWLYHASKDLSYLKYVTELNGQQFANW-----------G-NPTWFSWDDKHAGTHVLLSRLNIFGAKG-MSSEENLDLQMYRKTSEAIMCELLPDSPTATS PtGH9B16 -CPFYCSFSGYQDELLWAATWLYKASGENKYLSYISS-NK---GWS-----------Q-AVSEFSWDNKFAGAQTLLAKEFYGGK---------QDFDKFKSDAESFVCALMPGSSSVQI PtGH9B17 -KSYYASVSGYKDELLWGALWLYKATDNEKYLEYVIN-NAHCFGGT-----------GWAMEEFSWDVKYAGLQIMAAKLLVEEKHR-EH---GDTLEQYRSKAEHYLCSCLNKNNGTNV PtGH9B18 -KGHYTSVSGYMDELLWAGLWLYKATGDEEKILLAGK-----------------------------------------------SHR-KH---QHILKEYRSRAEYYLCACLNKNNVTNV PtGH9B2 -CPFYCSYSGYEDELLWGAAWLHKATKNPTYLNYIQV-NGQTLGAA-----------Q-FDNTFGWDNKHVGARILLSKAFLVQKVQ--------SLHDYKDHADNFICSLIQGAPFSSA PtGH9B3 -CPFYCSYSGYEDELLWGAAWLHKATKNPTYLNYIQV-NGQNLGAA-----------Q-FDNTFGWDNKHVGARILLSKAFLVQKLQ--------SLHDYKGHADNFICSLIPGAPFSSA PtGH9B4 -CPFYCSYSGYQDELLWGASWIHRASQNGSYLTYIQS-NGHTMGSD-----------D-DDYSFSWDDKRPGTKILLSKEFLEKTTE--------EFQLYKSHSDNYICSLIPGTSSFQA PtGH9B5 -CPFYCSYSGYQDELLWGASWLHKASLNGTYLAYIQS-NGHTMGSD-----------D-DDYSFSWDDKRPGTKILLSKEFLEKTTE--------EFQLYKSHSDNYICSLIPGTSSFQA PtGH9B6 -CPFYCDVNGYQDELLWGAVWLHKASRRRRYREYVVK-NEVILHAG-----------D-TINEFGWDNKHAGINVLISKEVLMGRAE--------YFESFKQNADDFICSILPGISHPQV PtGH9B7 -CPFYCDVNGYQDELLWGAVWLHKASRRRRYREYIVK-NEVILHAG-----------D-TINEFGWDNKHAGINVLISKEVLMGRAE--------YFESFKHNADGFICSILPGISHSQV PtGH9B8 -CPFYCDFDGYQDELLWGAAWLRRASSDDTYLSYLQN-NGKTLGAD-----------D-NINEFGWDNKHAGLNVLVSKEVLEGNMY--------SLQSYKASADSFMCTLIPESSSSHI PtGH9B9 -CPFYCDFDGYQDELLWGAAWLRRASYDDTYLSFLQN-NGETLGAD-----------E-NINEFGWDNKHAGLNVLVSKEVLEGNMN--------SLQSYKESADSFMCTLIPESSSSHV PtGH9C1 -KQFYTS-SGYSDELLWAAAWLYRATDDEYFLKYVVD-NAVYMGGT-----------GWAVKEFSWDNKYAGVQILLSQILLEGRGG-AY---TSTLKQYQAKANYFACACLQKNDGYNI PtGH9C2 -QKYYRSVSGYNDELLWAAAWLYQATNNQYYLDYLGN-NGDSMGGT-----------GWGMTEFGWDVKYAGVQTLVAKFLMQGKAG-HH---APVFEKYQQKAEYFMCSCLGKG-TRNV PtGH9C3 -QKYYRSVSGYNDELLWAAAWMYQATNNQYYLNYLGN-NGDSMGGT-----------GWGMTEFGWDVKYAGVQTLVAKFLMQGKAG-HH---APVFEKYQQKAEYFMCSCLGKG-SRNV PtGH9A1 -AIFY-NSTSYWDEFIWGGAWLYYATGNNSYLQLATM-PGLAKHAG-------AFWGGPDYGVLSWDNKLAGAQLLLSRLRLFLSPGYPY---EEILSTFHNQTSIIMCSYLPI--FTKF PtGH9A2 -AIFY-NSTSYWDEFVWGGAWLYYATGNNSYLQLATN-PGIAKHAG-------AFWGGPDYGVLSWDNKLAGTQLLLSRLRLFLSPGYPY---EEILRTFHNQTSIIMCSYLPI--FTKF PtGH9A3 -EPFY-NSTGYYDEFIWGATWLYYATGNVNYIRWATE-PGFSKHSK-------ALYRISDLSVLSWDNKLPAAMLLLTRCRIFLNPGYPY---EEMLHMYHNKTELNMCSYLQQ--FNVF PtGH9A4 -QPYY-NSTGYYDEYIWGATWLYYATGNITYIKLATE-PGFSKHSK-------ALLSIPDLSVLSWDNKLPAAMLLLTRYRIFLNPGYPY---EEMLHMYHQKTELNMCSYFQQ--FDVF -3 subsite -4 subsite

Figure 17. ClustalW alignment of P. trichocarpa 26 full-length gene models, showing aromatic stacking interactions at subsites -3 and -4. Numbering according to PtGH9A1. Residues shaded in gray are from the homology models of PttCel9A1 and PttCel9B3, or are strictly conserved in homologous PtGH9 enzymes shown in the alignment. Putative stacking aromatic amino acid residues in these subsites are shaded in cyan.

Ulla Rudsander

36

Possible evolutionary relationships, more distant than CAZy clans, have been suggested for both retaining and inverting enzymes, as well as α- and β- glycosyl hydrolases (Fig. 19A and D) (Stam et al., 2005). The evolutionary interconversion relies on the geometrical orientation of the glycosidic bond in the reactive substrate. There is also a possible mechanistic relationship between glycosyl hydrolases acting on β- glycosidic bonds, and phosphorylase activity (Fig. 19C and D) (Stam et al., 2005). However, transferase activity to an acceptor different from the donor by PttCel9A1, such as a phosphate group, would have been detected in the NMR spectra as large shifts for H1 and H2.

Figure 18. 1H NMR spectra (anomeric region) of cellohexaose hydrolysis (2 mM) in the presence of PttCel9A1 (260 ng/µl) at 25ºC ( 50 mM phosphate buffer, pH 6, 30 mM CaCl2) after 10, 17 and 32 min.

Figure 19. Mechanistic similarity in various inverting families of glycosyl hydrolases and related enzymes sharing an (α/α)6-barrel structure. (A) Mechanism of α-acting glycosyl hydrolases from families GH15, GH37, GH63, GH65, GH78, GH92, and GH95. (B) Mechanism of α-acting disaccharide phosphorylases from family GH65. (C) Mechanism of β-acting disaccharide phosphorylases from family GH94. (D) Mechanism of β-acting glycosyl hydrolases from families GH8, GH9, and GH48 (Stam et al., 2005).

Substrate analysis

37

Cellulases have a low catalytic efficiency on their soluble substrates when compared to the fastest metabolic enzymes that have catalytic rates approaching diffusion rates (Table 6). Hypocrea strains (formerly Trichoderma) have long been considered as the most efficient degraders of cellulose. An explanation to the low efficiency of HjCel6A (formerly CBHII) could be that the enzyme is not under the selection pressure to improve efficiency upon which it hydrolyses soluble glucosides. A tryptophan residue, 272, in HjCel6A located at the + 4 subsite at the entrance of the active site tunnel, was mutated to alanine or aspartic acid, and the activity on crystalline cellulose was impaired, while it was retained on amorphous cellulose and it was enhanced on soluble oligosaccharides no longer than cellopentaose (Koivula et al., 1998). Thus the bottleneck is not in the catalytic efficiency but in the ability to access the glucan chain, particularly for crystalline cellulose. The same seems to hold true for TfCel9A. The catalytic efficiency of TfCel9A, which is at least two orders of magnitude lower than that of HjCBHII, was improved by mutation of tryptophan 256 to alanine in the -4 substrate binding subsite (Paper III, Table 3). The activity on soluble cello-oligosaccharides no longer than cellopentaose was improved but activity with crystalline or amorphous cellulose was reduced (Paper II, Fig. 5). The slowest of all these four enzymes is PttCel9A1 (Table 6). A simlar trend has been observed across GH families 5, 12, 16 and 74 endo-xyloglucanases, where the catalytic turnover numbers were significantly lower for the microbial counterparts (Baumann et al., 2007). It thus seems that plant glucanases are not under selection pressure to maximize catalytic efficiency. Instead, a lower rate of hydrolysis is optimal in the plant cell, for PttCel9A1 probably in combination with loops for substrate recruitement and release, as well as targeting signals. Table 6. Comparison of the catalytic turnover numbers of different enzymes. Kcat/KM can be compared with the diffusion-controlled encounter of the enzyme and the substrate at 109 s-1/M (Fersht, 1985).

Enzyme Kcat (s-1) KM

(µM) Kcat/KM (s-1 /M). Substrate

TfCel9A a 16,9 470 3.6 x 104 cellohexaose HjCel6A b 14 14 1,0 x 106 cellohexaose PttCel9A a 2,3 x 10-2 1755 13 cellohexaose β-lactamase c 2 x 103 20 1 x 108 benzylpenicillin a. (Master et al., 2004) b. (Harjunpaa et al., 1996) c.(Fersht, 1985)

Ulla Rudsander

38

8 Domain swapping of plant cellulases (unpublished work) To address the role of the different domains in PttCel9A1, we have investigated complementation of the Arabidopsis thaliana kor1-1 mutant by PttCel9A1, PttCel9B3 and domain swapped hybrid proteins (Fig. 20). The kor1-1 mutant has an extreme dwarf constitution throughout the life cycle (Nicol et al., 1998b). Full-length PttCel9A1, (construct pEM4, Fig. 20), was shown to rescue kor1-1, thus proving its orthologue function (Fig. 4, Paper IV). While PttCel9B3, (construct pEM5, Fig. 18), did not rescue the phenotype (Fig. 4, Paper IV). PttCel9B3 belongs to a different subclass of plant cellulases, type B secreted enzymes with just a catalytic domain. Potential functional modules for type A plant cellulases are the cytoplasmic tail, the predicted N-terminal transmembrane helix, the conserved ‘blocking loop’ at the non-reducing end of the active site cleft and the proline-rich stretch at the C-terminus of the PttCel9A1 protein (Fig. 21). The proline-rich tail could be a possible target for protein partners, for example as in the interaction with SH3 domains (Agrawal and Kishan, 2002). SH3 domains generally bind to proline-rich domains that adopt a polyproine helix II conformation with a PXXP motif (where P is proline and X is any amino acid). In A. thaliana, an SH3-containing protein, AtSH3P1, has been localized to the plasma membrane and its associated vesicles, involved in clathrin-mediated vesicle trafficking (Lam et al., 2001). In the Populus EST database (Sterky et al., 2004), a highly homologous cluster, poplar.5971, is present in the active cambium EST library. A potential role in endosome recycling for the proline-rich tail would agree well with colocalization studies of a GFP-KOR1 fusion with endocytic markers (Robert et al., 2005). The cytoplasmic tail of PttCel9A1 contains two polarized targeting signals (Fig. 21). They have been shown to be essential for cell division, but not the cell elongation defect in the more severe kor1-2 mutant (Zuo et al., 2000). To investigate further the localization and interaction requirements of plant cellulase activity for normal cellulose biosynthesis, we created a number of constructs and introduced them into the kor1-1 mutant. The proline-rich tail, the transmembrane domain and the cytoplasmic tail of PttCel9A1 were added to the catalytic domain of PttCel9B3, (construct pUR19, Fig 20). The effect of removing the proline-tail from otherwise intact PttCel9A1 protein was also examined, (construct pUR21, Fig. 20), and the effect of removing the ‘blocking loop’ at the non-reducing end of the active site from PttCel9A1, (construct pUR20, Fig 20). As negative controls, the binary vector without any cDNA insert was introduced to the background ecotype WS and to kor1-1 mutant plants.

pEM4

pEM5

pUR19

pUR20

pUR21

Construct: Modular structure:

Figure 20. Domain swapping constructs for the complementation of the korrigan mutant by wildtype and hybrid poplar cellulases. PttCel9A1 domains: catalytic domain (green), blocking loop (black), cytoplasmic tail (pink), transmembrane helix (light blue), proline-rich tail (orange). PttCel9B3 domains: catalytic domain (blue), signal peptide (gray)

Domain swapping

39

Preliminary results show that the catalytic domain of PttCel9B3 flanked by a transmembrane domain, cytoplasmic tail and proline-rich tail could not rescue the kor1-1 dwarf phenotype, whereas the PttCel9A1 deletion constructs with no proline-tail or deleted ‘blocking loop’ were apparently able to restore the dwarf phenotype (Fig. 22A). The root lengths of complemented kor1-1 seedlings were measured and a similar result was obtained. The catalytic domains of PttCel9B3 and PttCel9A1 are 39 % identical at the amino acid level. The differences in the catalytic domains of PttCel9B3 and PttCel9A1 seem to play a decisive role for the function of PttCel9A1 in cellulose biosynthesis. A PttCel9A1 MYGRDPWGGPLEINAADSATDDDRSRNLNDLDRAALSRPLDETQQSWLLGPAEQKKKKKY 60 KOR1 MYGRDPWGGPLEINTADSATDDDRSRNLNDLDRAALSRPLDETQQSWLLGPTEQKKKK-Y 59 Pttcel9B3 ------------------------------MRRGAS--------FCLLFS---------- 12 PttCel9A1 VDLGCIIVSRKIFVWTVGSIVAAGLLVGLITLIVKTVPRHHHSHAPADNYTLALHKALMF 120 KOR1 VDLGCIIVSRKIFVWTVGTLVAAALLAGFITLIVKTVPRHHPKTPPPDNYTIALHKALKF 119 Pttcel9B3 --------------------LSLVLLG-----FVQAKP----------NYNEALAKSILF 37 B PttCel9A1 GAKTLFKFARDQRGRYS--AGSSEAAIFYNSTSYWDEFIWGGAWLYYATGNNSYLQLATM 354 KOR1 GAKVVYQFGRTRRGRYS--AGTAESSKFYNSSMYWDEFIWGGAWMYYATGNVTYLNLITQ 354 Pttcel9B3 TARKVFQFAMQYQGAYSDSLGSAVCPFYCSYSGYKDELLWGAAWLFRATNEMSYYNIFKS 260 PttCel9A1 PGLAKHAGAFWGGPDYGVLSWDNKLAGAQLLLSRLRLFLSPGYPYEEILSTFHNQTSIIM 414 KOR1 PTMAKHAGAFWGGPYYGVFSWDNKLAGAQLLLSRLRLFLSPGYPYEEILRTFHNQTSIVM 414 Pttcel9B3 LGADDQP---------DLFSWDNKYAGVHVLLSRRALLNN-----DKNFEQFEGEAESFM 306 C PttCel9A1 GGWK-WRDTSKPNPNTLVGAMVAGPDRHDGFHDVRTNYNYTEPTIAGNAGLVAALVALSG 588 KOR1 GGWK-WRDSKKPNPNTIEGAMVAGPDKRDGYRDVRMNYNYTEPTLAGNAGLVAALVALSG 588 Pttcel9B3 SGFEPFFHSANPNPNILTGAIVGGPNQNDGYPDERSDYSHSEPATYINAAMVGPLAYFAA 483 PttCel9A1 DK--TTGIDKNTIFSAVPPMFPTPPPPPAPWKP 619 KOR1 EEEATGKIDKNTIFSAVPPLFPTPPPPPAPWKP 621 Pttcel9B3 TLN------------------------------ 486

Figure 21. ClustalW protein alignments showing regions with characterisitic motifs. (A) cytoplasmic tail, tyrosine-based targeting signal (cyan), dileucine-based targeting signal (magenta) and transmembrane helix (grey). (B) Type A cellulase conserved blocking loop (red) at the non-reducing end of the active site. (C) Conserved proline-rich tail (yellow) at the C-terminus of type A cellulases. Further analysis will show if there is a statistically significant difference in height and root lengths for complementation of kor1-1 by construct pUR21, defective in the ‘blocking loop’, compared with construct pUR20, defective in the proline-rich tail. If the proline-rich tail is involved in endosome recycling, it might also be that the complementation of the more severe phenotype kor1-2 should be investigated, and that the cell elongation defect in the kor1-1 mutant is not enough to demonstrate the effect of its deletion.

The ‘blocking loop’ at the active site seems to be less significant for the function of PttCel9A1 in cellulose biosynthesis. A similar result has been observed for the removal of the “blocking loop” from TfCel9A, ΔT245-L251. Only slighlty improved filter paper activity and binding to BMCC was observed, but cellopentaose and cellohexaose hydrolysis profiles did not change. Deletion of the blocking loop in TfCel9a may thus provide better access to the cellulose chain. Further analysis will reveal if the ‘blocking loop’ has any influence on, for example, the crystallinity index.

Ulla Rudsander

40

(A) (B) Figure 22. Complementation of the kor1-1 mutant with hybrid poplar cellulases. cDNAs constructs pUR19, pUR20 and pUR21 were fused to CaMV 35S promoter and introduced to the homozygous kor1-1 mutant plants with the Agrobacterium-mediated vector transfer. (A) T1 seeds were selected on plates containing hygromycin as an antibiotic selection marker and then the plants were grown in soil for two and half weeks in LD conditions. A representative T1 line of several transgenic lines obtained is shown. (B) T1 selection. They were selected on 15 µg ml-1 hygromycin MS medium containing 1% sucrose. Grown under LD for ten days and moved to medium without hygromycin on vertical position and grown for four more days (figures courtesy of Junko Takahashi).

9 Mechanism for a membrane-anchored cellulase in plant cellulose biosynthesis? (Paper IV)

The occurrence of hydrolytic enzymes in polymer biosynthesis systems is commonly found in living organisms, for example DNA polymerase I for DNA synthesis, RNA as an enzyme for RNA splicing, and proteases for protein processing. Therefore, a cellulase as a necessary component in the cellulose biosynthesis machinery is not an odd circumstance. But exactly how does the slowly acting membrane-anchored cellulase become essential? Since polymerization and crystallization are coupled processes in the biosynthesis of cellulose, the function can lie in either of these two areas.

The first scenario could be that the membrane-anchored cellulase cleaves a primer or precursor molecule. Here it is important to remember that polymerization of cellulose by the CesAs is predicted to occur at the cytoplasmic side of the plasma membrane, via a supply of UDP-glucose as substrate (Doblin et al., 2002). The newly synthesized glucan chain would be extruded by CesA subunits or special pores. The catalytic domain of KOR protein is predicted to face the cell wall, which presents a topological problem when a role in primer or precursor hydrolysis is discussed (Fig. 23).

Mechanism in cellulose biosynthesis

41

Cell Wall

Plasma Membrane Figure 23. Hypothetical model of PttCel9A1 situated on the plasma membrane surface. The PttCel9A1 backbone model is drawn as beige coils (loops) and blue spirals (α-helices). The PttCel9A1 molecule is anchored in the plasma membrane through its transmembrane helix indicated in the lower right corner. Juxtapose to the PttCel9A1 molecule, a schematic model of one hexameric globule of the cellulose synthase rosette is shown as green cylinders traversing the membrane. It has been suggested that each rosette is a hexamer of six globules where each globule consists of six subunits (CesA), resulting in a complex of 36 subunits that accounts for the simultaneous syntheses of 36 β-glucan chains (Doblin et al., 2002). A growing cellulose chain that extends into, and fills, the cellulose-binding cleft of PttCel9A1 is shown as a ball-and-stick model (carbon atoms, yellow; oxygen atoms, red) (courtesy of Dr Christina Divne). An early hypothesis for the function of KOR was presented by Delmer and coworkers (Peng et al., 2002). The KOR protein would remove a sitosterol from sitosterol cellodextrins to allow further elongation by the CesA and incorporation of cellulose into the cell wall. The yield from plant extracts of sitosterol cellodextrins would be extremely low, for testing the hypothesis in vitro (Peng et al., 2002). The synthesis of such a molecule would be complex but possible (Harry Brumer, personal communication). Evaluating the hypothesis in vivo by analysis of metabolic engineering mutants has been performed. Sterols have been shown to affect cellulose synthesis in Arabidopsis mutants fackel and hydra1 (Schrick et al., 2000; Schrick et al., 2002). However the KOR gene exhibited normal transcript levels in fackel seedlings, and no sign of compensatory effects were observed (He 2003). Also, a significant decrease in sitosterol was not followed by a comparable decrease in cellulose levels by the mutants fackel and hydra1 (Schrick, 2004). Finally, Roberts et al studied sterol-glucosides in

Ulla Rudsander

42

Arabidopsis wild type and kor1-1 mutant to test the hypothesis (Robert et al., 2004). If the cellulose deficiency observed in kor1-1 would be a result of a change in recycling of the SG primer, a smaller amount of sitosterol-glucoside would be expected in the mutant. However, comparable amounts of sitosterol glucoside and sitosterol cellodextrin were visualized by autoradiography.

The second scenario includes many possibilities for the KOR protein in the crystallization process of the glucan chains. These alternative roles could be removal of misaligned glucan parts (Delmer, 1999; Molhoj, 2002; Szyjanowicz et al., 2004), release of CesAs from the growing microfibril, or unwinding tension in the helical microfibril, similar to DNA topoisomerase I (Molhoj, 2002). Often the KOR protein is proposed to remove misaligned parts. If the role of KOR protein was in the editing of glucan chains sticking out, downregulation would lead to decreased crystallinity index, which is not the case in our study (Paper IV). A role for the KOR protein in release of CesAs from the growing microfibrils would require many KOR proteins in the plasma membrane, one for each CesA, which so far has not been detected in colocalizaion and protein interaction studies (Szyjanowicz et al., 2004; Robert et al., 2005; Desprez et al., 2007).

Reduced crystallinity index from overexpression of PttCel9A1 in Arabidopsis (Paper IV) could be compatible with a role in the relaxation of tortional stresses around the microfibril axis (Molhoj, 2002), in analogy with the role of DNA topoisomerase IV during DNA replication. Twisted regions and regular periodicities have been observed by atomic force microscopy in microfibrils from green algae (Hanley et al., 1997), and if rosettes are not rotating in the membrane, unwinding might be needed. Loose ends created by hydrolytic relaxation would not be present in the kor1-1 mutant, and the cellulose would be more tightly packed. Overexpression of Pttcel9A1 on the other hand would leave many loose ends hanging and prevent proper packing.

A new model for cellulose biosynthesis was recently proposed from visualizations by atomic force microscopy in maize primary cell walls (Ding and Himmel, 2006). A number of elementary fibrils are synthesized at the cellulose synthase complex and they coalesce to form a macrofibril. The macrofibril is later dispersed to form microfibrils, which are coated with hemicelluloses and pectin. In this model, PttCel9A1 /KOR1 protein could function in the splitting of the macrofibril, entering para-crystalline regions to hydrolyze glucan chains that hydrogen bond between adjacent microfibrils, as outlined in Paper IV. Less KOR1 protein would then lead to the accumulation of tightly packed macrofibrils, manifested in increased crystallinity index and less cellulose being synthesized. More PttCel9A1 protein would lead to premature splitting or prevent formation of macrofibrils, and many exposed microfibrils would be present with more amorphous regions. These exposed microfibrils would then interact with hemicelluloses and pectins in a disordered fashion, and cell growth would be disturbed. Recently, the movement of the cellulose synthase complex was proposed to be driven by the polymerization and crystallization of the cellulose chains in a biophysical model (Diotallevi and Mulder, 2007). Polymer flexibility and membrane elasticity would be force transducers. If heavy macrofibrils accumulate due to deficient KOR1 protein, the movement of the rosettes could slow down, and the cellulose content would be decreased, as is the case in the kor1-1 mutant. Macrofibrils (microfibril aggregates) in plant cells have been described using a range of techniques including NMR and different microscopy methods (Donaldson, 2007). Also, macrofibrils have been observed in monocots as well as in dicots. In poplar, macrofibrils are present in all different wood tissue types examined, and their diameter seems to vary with the lignification degree. Both the secondary cell wall layers, S1-S3, and the primary cell wall have macrofibrillar structures. Some microscopy views have indicated that macrofibrils are not only surface features but also extend deep into the cell wall.

Mechanism in cellulose biosynthesis

43

When the genome of Populus trichocarpa was published (Tuskan et al., 2006), five full-length membrane-anchored cellulases were identified. The two highly similar pairs, PtGH9A1 and PtGH9A2, PtGH9A3 and PtGH9A4, probably originate from the duplication of the poplar genome (Djerbi et al., 2005; Tuskan et al., 2006). The fifth member of this group is an outlier in the branch of type A endoglucanases in the phylogenetic tree (Fig. 6). This putative enzyme, PtGH9A5, is lacking the following conserved features; cytoplasmic tail, polarized targeting signal, blocking loop at the non-reducing end of the active site cleft, and polyproline tail (Fig. 21). However, the transmembrane helix of PtGH9A5 is conserved and also the type II integral membrane protein topoplogy with the catalytic domain located outside the plasma membrane. There is no PtGH9A5 transcript in the populus EST database, harboring more than 100,000 transcripts from all major tissues of the plant (www.populus.db.umu.se (Sterky et al., 2004)). Future research will shed light on the unique function of PtGH9A5.

44

45

CONCLUDING REMARKS AND FUTURE PERSPECTIVES

A biotechnological approach was used here to study plant cellulases involved in wood formation in the tree P. tremula x tremuloides. The catalytic domain of the membrane-anchored cellulase was produced in P. pastoris but the protein was phosphomannosylated and it was retrieved at low yields. To achieve the goal of a 3D crystallographic structure of PttCel9A1, enhanced E.coli or plant expression systems could be used, or point mutations of selected N-linked glycosylation sites could be performed. The interaction between glycosyl residues and aromatic hydrophobic platforms is well-documented in the scientific literature. A homology model of PttCel9A1 showed that there are missing stacking aromatic platforms in glycosyl subsites -2 through -4. In particular, we have observed that the missing aromatic platform in the -2 subsite correlates with a low catalytic efficiency compared to microbial enzymes from family GH9. The missing aromatic platforms in subsites -2 to -4 make it likely that the low activity observed for the PttCel9A1 enzyme is true and that it is not an artifact of overglycosylation, although we cannot exclude that the extensive glycan structures decrease the activity somewhat. Still, it is peculiar that nature has created such a slow enzyme, whether it would be to protect the cell walls of the plant or whether it indicates a different in vivo substrate.

In the overexpression experiments of membrane-anchored cellulase in Arabidopsis and poplar, we were hoping to see positive effects on biomass production or cellulose composition, for example increased fiber cell length or thickness, or increased mechanical strength of fibres. Unfortunately, the phenotype of overexpressing PttCel9A1 lines was quite similar to the down-regulated phenotype in certain aspects and in certain aspects the overexpressing lines were very similar to wildtype plants. This suggests that an optimal level of PttCel9A1/KOR1 protein is required. Perhaps the simultaneous genetic manipulation of the three secondary cell wall specific cellulose synthases and the membrane-anchored cellulase could lead to enhanced fibre properties, thus balancing the effect of different proteins.

Other future perspectives for this project are in the production of ethanol from plant biomass. Hydrolysis of cellulose to glucose by cellulases is then the enzymatic pathway of interest. Transgenic plants that can produce their own hydrolytic enzymes induced after harvest would be an inexpensive alternative. However, plant cellulases are inefficient enzymes compared to cellulases from microbial organisms, so their catalytic machinery is not an attractive resource. Ancillary modules, such as CBDs, tails and transmembrane domains could however be used for targeting of efficient microbial catalytic domains to sites of interest. To learn more about the function of PttCel9A1 in cellulose biosynthesis, additional plant work is needed. With the actual visualization of the movement of YFP-CesA particles in the plasma membrane by spinning disk confocal microscopy (Paredez et al., 2006), the contradicting evidence for the cellular localization of PttCel9A1/KOR1 protein may be resolved by similar techniques. However, for all transgenic plant work the system in vivo is very complex with potential co-suppressions and co-expressions of duplicated genes, while protein characterization experiments in vitro simplify the process. A compromise might be possible once cellulose synthesis has been confirmed in vitro and the necessary ingredients are characterized. Then the direct role of PttCel9A1 can be addressed by making small changes to the system and analyze the nature of the synthesized cellulose.

46

47

LIST OF ABBREVIATIONS

AOX alcohol oxidase BMCC bacterial microcrystalline cellulose CAZyme carbohydrate active enzyme CaMV cauliflower mosaic virus CBD cellulose binding domain CBH cellobiohydrolase CBM carbohydrate binding module CesA cellulose synthase Cel cellulase cDNA complementary DNA CMC carboxy-methyl cellulose DNA deoxyribonucleic acid DP degree of polymerization ER endoplasmic reticulum EC enzyme commission EST expressed sequence tag G glucose monomer GFP green fluorescent protein GH glycosyl hydrolase GT glycosyl transferase Glc glucose GlcNAc N-acetyl glucosamine HCA hydrophobic cluster analysis HEC hydroxyl-ethyl cellulose HPLC high-performance liquid chromatography Ig-like Immunoglobulin-like KOR KORRIGAN Man mannose mRNA messenger RNA NMR nuclear magnetic resonance OCH1 gene for P. pastoris mannosyltransferase PASC Phosphoric acid swollen cellulose PCR polymerase chain reaction PDB Brookhaven Protein Databank Ptt Populus tremula x tremuloides Pt Populus trichocarpa SCOP Structural Classification of Proteins SDS-PAGE sodium dodecyl sulfate-polyacrylamide gel electrophoresis SH3 src homology domain Src Rous sarcoma virus transforming protein TC terminal complex for cellulose microfibril TLC thin layer chromatography RT PCR reverse transcriptase PCR RACE rapid amplification of cDNA ends

48

RNA ribonucleic acid RNAi RNA interference XET xyloglucan endotransglycosylase YFP yellow fluorescent protein

49

ACKNOWLEDGEMENTS

This thesis would not have been possible without the work and support from many people. I am very grateful to: Professor Tuula Teeri for guiding me all the way through this very interesting research project, and findings ways for me to survive and thrive after 2 maternity leaves. Dr Stuart Denman for initial supervision in the Pichia jungle, Dr Emma Master for high quality lab work and advice as well as good travelling company in Los Angeles, Dr Kathleen Piens for encouragement and support both when you were in Sweden and now in Belgium, Dr Corine Sandström for invaluable collaboration and good listening skills, and Dr Harry Brumer and Prof David Wilson for fruitful discussions and critical reading, Dr Christina Divne for fruitful discussions and model pictures, Dr Mehmedalija Jahic and Reza Salmalian for the unpublished fermentation work, Dr Dan Cullen and Amber Vanden Wymelenberg for a pleasant stay in Madison in the early PhD student years. Dr Ewa Mellerowicz, Prof Björn Sundberg, Junko Takahashi and Dr Satoshi Endo in Umeå for bringing together the transgenic plant stuff, and all other co-authors for your good work. I further want to thank: My former roommates Dr Åsa Kallas, Dr Soraya Djerbi and Dr Henrik Aspeborg for helpful interpretations of lab and life aspects during those years. The present roommates and colleagues of the wood club for making also the spurt an enjoyable time, especially Alex Rajangam for your Indian friendship. Lotta Rosenfeldt for administrative matters as well as jogging tours and entertaining DVDs, Marianne Wittrup-Larsen for good advice on how to deal with an uphill slope. Former colleagues Dr Joke Rotticci-Mulder and Dr Jenny Ottosson for lunches and skating activities. Former colleague Dr Nader Nourizad for his contagious positive attitude. Prof Sophia Hober, Dr Anna Ottosson and Dr Torkel Berglund for taking time to listen and encourage. All former and present people at XVE and floor 2 for creating such a nice working atmosphere. Min mentor Jan-Erik Nyström för samtal som fått mig att pröva mina vingar i tankarna. Uppsala/San Diego-vänner Carolina Wählby, Kia Strömberg, Anna Polgren, Karin Melen . Helene Andersson-Svahn, Maria Schultzberg, Sara Sjögren, Jenny af Ugglas och Niklas Mattsson för hejaropen och värmen. Gabriella Wiklund och Helena Hallgren, och det glada middagsgänget, för att ni hållit igång mitt sociala liv medan jag borrat mig ner i alla nördiga detaljer om enzymer och om bäbisar (det gjorde vi tillsammans).

50

Mina föräldrarna Karin och Lars, mina systrar Åse och Frida, Per, Lovisa, Ebba, Håkan, som tar hand om mig och mina barn, och lånar ut sina kläder (systrarna). Och det största tacket går till min älskade Patrik. ♥ Utan dig hade jag varit som några poppelceller utan cellulosa, lite grön sörja som flyter omkring ute på öppet vatten. Nu ska vi ha mer tid för varandra, huset som Gud glömde och de två underbara mutanterna!! Till sist, tack till lilla Elvira för den fina massagen, och nu säger jag som Edvard: Jippiii!!!

51

REFERENCES

Adams, K.L., and Wendel, J.F. (2005). Polyploidy and genome evolution in plants. Current

Opinion in Plant Biology 8, 135-141. Agrawal, V., and Kishan, K.V.R. (2002). Promiscuous binding nature of SH3 domains to

their target proteins. Protein and Peptide Letters 9, 185-193. Albersheim, P., Darvill, A., Roberts, K., Staehelin, L.A., and Varner, J.E. (1997). Do the

structures of cell wall polysaccharides define their mode of synthesis? Plant Physiol. 113, 1-3.

Andersson-Gunneras, S., Mellerowicz, E.J., Love, J., Segerman, B., Ohmiya, Y., Coutinho, P.M., Nilsson, P., Henrissat, B., Moritz, T., and Sundberg, B. (2006). Biosynthesis of cellulose-enriched tension wood in Populus tremula: global analysis of transcripts and metabolites identifies biochemical and developmental regulators in secondary wall biosynthesis.(vol 45, pg 144, 2005). Plant Journal 46, 349-349.

ArabidopsisInitiative. (2000). Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 408, 796-815.

Aspeborg, H., Schrader, J., Coutinho, P.M., Stam, M., Kallas, A., Djerbi, S., Nilsson, P., Denman, S., Amini, B., Sterky, F., Master, E., Sandberg, G., Mellerowicz, E., Sundberg, B., Henrissat, B., and Teeri, T.T. (2005). Carbohydrate-active enzymes involved in the secondary cell wall biogenesis in hybrid aspen. Plant Physiol. 137, 983-997.

Astley, O.M., Chanliaud, E., Donald, A.M., and Gidley, M.J. (2003). Tensile deformation of bacterial cellulose composites. International Journal of Biological Macromolecules 32, 28-35.

Bajpai, P. (2004). Biological bleaching of chemical pulps. Critical Reviews in Biotechnology 24, 1-58.

Baumann, M.J., Eklof, J.M., Michel, G., Kallas, A.M., Teeri, T.T., Czjzek, M., and Brumer, H. (2007). Structural evidence for the evolution of xyloglucanase activity from xyloglucan endo-transglycosylases: Biological implications for cell wall metabolism. Plant Cell 19, 1947-1963.

Bayer, E.A., Belaich, J.P., Shoham, Y., and Lamed, R. (2004). The cellulosomes: Multienzyme machines for degradation of plant cell wall polysaccharides. Annual Review of Microbiology 58, 521-554.

Benziman, M., Haigler, C.H., Brown, R.M., White, A.R., and Cooper, K.M. (1980). Cellulose Biogenesis - Polymerization and Crystallization Are Coupled Processes in Acetobacter-Xylinum. Proceedings of the National Academy of Sciences of the United States of America-Biological Sciences 77, 6678-6682.

Bernier, G., Havelange, A., Houssa, C., Petitjean, A., and Lejeune, P. (1993). Physiological Signals That Induce Flowering. Plant Cell 5, 1147-1155.

Bhandari, S., Fujino, T., Thammanagowda, S., Zhang, D.Y., Xu, F.Y., and Joshi, C.P. (2006). Xylem-specific and tension stress-responsive coexpression of KORRIGAN endoglucanase and three secondary wall-associated cellulose synthase genes in aspen trees. Planta 224, 828-837.

Biely, P., Kratky, Z., and Vrsanska, M. (1981). Substrate-binding site of endo-1,4-beta-xylanase of the yeast Cryptococcus albidus. Eur J Biochem 119, 559-564.

52

Bollok, M., Henriksson, H., Kallas, A., Jahic, M., Teeri, T.T., and Enfors, S.O. (2005). Production of poplar xyloglucan endotransglycosylase using the methylotrophic yeast Pichia pastoris. Applied Biochemistry and Biotechnology 126, 61-77.

Brett, C., and Waldron, K. (1996). Physiology and Biochemistry of Plant Cell Walls. (London: Chapman & Hall).

Brown, R.M., and Montezinos, D. (1976). Cellulose Microfibrils - Visualization of Biosynthetic and Orienting Complexes in Association with Plasma-Membrane. Proc. Natl. Acad. Sci. U. S. A. 73, 143-147.

Brumer, H., Zhou, Q., Baumann, M.J., Carlsson, K., and Teeri, T.T. (2004). Activation of crystalline cellulose surfaces through the chemoenzymatic modification of xyloglucan. Journal of the American Chemical Society 126, 5715-5721.

Brummell, D.A., Lashbrook, C.C., and Bennett, A.B. (1994). Plant endo-1,4-β−glucanases: structures, properties and physiological function. American Chemical Society Symposium Ser. 566, 100-129.

Brummell, D.A., Catala, C., Lashbrook, C.C., and Bennett, A.B. (1997). A membrane-anchored E-type endo-1,4-beta-glucanase is localized on Golgi and plasma membranes of higher plants. Proc Natl Acad Sci U S A 94, 4794-4799.

Catala, C., Rose, J.K., and Bennett, A.B. (1997). Auxin regulation and spatial localization of an endo-1,4-beta-D- glucanase and a xyloglucan endotransglycosylase in expanding tomato hypocotyls. Plant J 12, 417-426.

Catala, C., Rose, J.K., and Bennett, A.B. (2000). Auxin-regulated genes encoding cell wall-modifying proteins are expressed during early tomato fruit growth. Plant Physiol 122, 527-534.

Cereghino, J.L., and Cregg, J.M. (2000). Heterologous protein expression in the methylotrophic yeast Pichia pastoris. Fems Microbiology Reviews 24, 45-66.

Chanzy, H., and Henrissat, B. (1985). Undirectional Degradation of Valonia Cellulose Microcrystals Subjected to Cellulase Action. Febs Letters 184, 285-288.

Charnock, S.J., Henrissat, B., and Davies, G.J. (2001). Three-dimensional structures of UDP-sugar glycosyltransferases illuminate the biosynthesis of plant polysaccharides. Plant Physiology 125, 527-531.

Christoffersen, R.E., Tucker, M.L., and Laties, G.G. (1984). Cellulase Gene-Expression in Ripening Avocado Fruit - the Accumulation of Cellulase Messenger-Rna and Protein as Demonstrated by Cdna Hybridization and Immunodetection. Plant Mol.Biol. 3, 385-391.

Cosgrove, D.J. (1999). Enzymes and other agents that enhance cell wall extensibility. Annual Review of Plant Physiology and Plant Molecular Biology 50, 391-417.

Davies, G., and Henrissat, B. (1995). Structures and Mechanisms of Glycosyl Hydrolases. Structure 3, 853-859.

Davison, A., and Blaxter, M. (2005). Ancient origin of glycosyl hydrolase family 9 cellulase genes. Molecular Biology and Evolution 22, 1273-1284.

del Campillo, E., Abdel-Aziz, A., Crawford, D., and Patterson, S.E. (2004). Root cap specific expression of an endo-beta-1,4-D-glucanase (cellulase): a new marker to study root development in Arabidopsis. Plant Mol.Biol. 56, 309-323.

Delmer, D.P. (1999). Cellulose biosynthesis: Exciting times for a difficult field of study. Annu. Rev. Plant Physiol. Plant Molec. Biol. 50, 245-276.

Desprez, T., Juraniec, M., Crowell, E.F., Jouy, H., Pochylova, Z., Parcy, F., Höfte, H., Gonneau, M., and Vernhettes, S. (2007). Organization of cellulose synthase complexes involved in primary cell wall synthesis in Arabidopsis thaliana. Proc. Natl. Acad. Sci. U. S. A. 104, 15572-15577.

53

Ding, S.Y., and Himmel, M.E. (2006). The maize primary cell wall microfibril: A new model derived from direct visualization. Journal of Agricultural and Food Chemistry 54, 597-606.

Ding, S.Y., Bayer, E.A., Steiner, D., Shoham, Y., and Lamed, R. (1999). A novel cellulosomal scaffoldin from acetivibrio cellulolyticus that contains a family 9 glycosyl hydrolase [In Process Citation]. J Bacteriol 181, 6720-6729.

Diotallevi, F., and Mulder, B. (2007). The cellulose synthase complex: A polymerization driven supramolecular motor. Biophysical Journal 92, 2666-2673.

Divne, C., Stahlberg, J., Teeri, T.T., and Jones, T.A. (1998). High-resolution crystal structures reveal how a cellulose chain is bound in the 50 A long tunnel of cellobiohydrolase I from Trichoderma reesei. J Mol Biol 275, 309-325.

Divne, C., Stahlberg, J., Reinikainen, T., Ruohonen, L., Pettersson, G., Knowles, J.K., Teeri, T.T., and Jones, T.A. (1994). The three-dimensional crystal structure of the catalytic core of cellobiohydrolase I from Trichoderma reesei. Science 265, 524-528.

Djerbi, S., Lindskog, M., Arvestad, L., Sterky, F., and Teeri, T.T. (2005). The genome sequence of black cottonwood (Populus trichocarpa) reveals 18 conserved cellulose synthase (CesA) genes. Planta 221, 739-746.

Djerbi, S., Aspeborg, H., Nilsson, P., Sundberg, B., Mellerowicz, E., Blomqvist, K., and Teeri, T.T. (2004). Identification and expression analysis of genes encoding putative cellulose synthases (CesA) in the hybrid aspen, Populus tremula (L.) x P-tremuloides (Michx.). Cellulose 11, 301-312.

Doblin, M.S., Kurek, I., Jacob-Wilk, D., and Delmer, D.P. (2002). Cellulose Biosynthesis in Plants: from Genes to Rosettes. Plant Cell Physiol 43, 1407-1420.

Donaldson, L. (2007). Cellulose microfibril aggregates and their size variation with cell wall type. Wood Science and Technology 41, 443-460.

Ferrarese, L., Trainotti, L., Gattolin, S., and Casadoro, G. (1998). Secretion, purification and activity of two recombinant pepper endo- beta-1,4-glucanases expressed in the yeast Pichia pastoris. FEBS Lett 422, 23-26.

Ferrarese, L., Trainotti, L., Moretto, P., Polverino de Laureto, P., Rascio, N., and Casadoro, G. (1995). Differential ethylene-inducible expression of cellulase in pepper plants. Plant Mol Biol 29, 735-747.

Fersht, A. (1985). Enzyme structure and mechanism. (New York: Freeman and Company). Geisler-Lee, J., Geisler, M., Coutinho, P.M., Segerman, B., Nishikubo, N., Takahashi, J.,

Aspeborg, H., Djerbi, S., Master, E., Andersson-Gunneras, S., Sundberg, B., Karpinski, S., Teeri, T.T., Kleczkowski, L.A., Henrissat, B., and Mellerowicz, E.J. (2006). Poplar carbohydrate-active enzymes. Gene identification and expression analyses. Plant Physiol. 140, 946-962.

Gutierrez, A., del Rio, J.C., Martinez, M.J., and Martinez, A.T. (2001). The biotechnological control of pitch in paper pulp manufacturing. Trends Biotechnol. 19, 340-348.

Haigler, C.H., Ivanova-Datcheva, M., Hogan, P.S., Salnikov, V.V., Hwang, S., Martin, K., and Delmer, D.P. (2001). Carbon partitioning to cellulose synthesis. Plant Molecular Biology 47, 29-51.

Hanley, S.J., Revol, J.F., Godbout, L., and Gray, D.G. (1997). Atomic force microscopy and transmission electron microscopy of cellulose from Micrasterias denticulata; evidence for a chiral helical microfibril twist. Cellulose 4, 209-220.

Hanus, J., and Mazeau, K. (2006). The xyloglucan-cellulose assembly at the atomic scale. Biopolymers 82, 59-73.

Harjunpaa, V., Teleman, A., Koivula, A., Ruohonen, L., Teeri, T.T., Teleman, O., and Drakenberg, T. (1996). Cello-oligosaccharide hydrolysis by cellobiohydrolase II

54

from Trichoderma reesei. Association and rate constants derived from an analysis of progress curves. Eur J Biochem 240, 584-591.

Harpster, M.H., Brummell, D.A., and Dunsmuir, P. (1998). Expression analysis of a ripening-specific, auxin-repressed endo-1, 4- beta-glucanase gene in strawberry. Plant Physiol 118, 1307-1316.

Harpster, M.H., Brummell, D.A., and Dunsmuir, P. (2002a). Suppression of a ripening-related endo-1,4-beta-glucanase in transgenic pepper fruit does not prevent depolymerization of cell wall polysaccharides during ripening. Plant Molecular Biology 50, 345-355.

Harpster, M.H., Dawson, D.M., Nevins, D.J., Dunsmuir, P., and Brummell, D.A. (2002b). Constitutive overexpression of a ripening-related pepper endo-1,4-beta-glucanase in transgenic tomato fruit does not increase xyloglucan depolymerization or fruit softening. Plant Molecular Biology 50, 357-369.

Hayashi, N., Sugiyama, J., Okano, T., and Ishihara, M. (1997). The enzymatic susceptibility of cellulose microfibrils of the algal-bacterial type and the cotton-ramie type. Carbohydrate Research 305, 261-269.

Hayashi, T. (1989). Xyloglucans in the Primary-Cell Wall. Annu. Rev. Plant Physiol. Plant Molec. Biol. 40, 139-168.

Henriksson, H., Denman, S.E., Campuzano, I.D.G., Ademark, P., Master, E.R., Teeri, T.T., and Brumer, H. (2003). N-linked glycosylation of native and recombinant cauliflower xyloglucan endotransglycosylase 16A. Biochem. J. 375, 61-73.

Henrissat, B., Teeri, T.T., and Warren, R.A. (1998). A scheme for designating enzymes that hydrolyse the polysaccharides in the cell walls of plants. FEBS Lett 425, 352-354.

Henrissat, B., Coutinho, P.M., and Davies, G.J. (2001). A census of carbohydrate-active enzymes in the genome of Arabidopsis thaliana. Plant Molecular Biology 47, 55-72.

Henrissat, B., Claeyssens, M., Tomme, P., Lemesle, L., and Mornon, J.P. (1989). Cellulase Families Revealed by Hydrophobic Cluster-Analysis. Gene 81, 83-95.

Hertzberg, M., Aspeborg, H., Schrader, J., Andersson, A., Erlandsson, R., Blomqvist, K., Bhalerao, R., Uhlen, M., Teeri, T.T., Lundeberg, J., Sundberg, B., Nilsson, P., and Sandberg, G. (2001). A transcriptional roadmap to wood formation. Proc Natl Acad Sci U S A 98, 14732-14737.

Hidaka, M., Kitaoka, M., Hayashi, K., Wakagi, T., Shoun, H., and Fushinobu, S. (2006). Structural dissection of the reaction mechanism of cellobiose phosphorylase. Biochem. J. 398, 37-43.

Hildén, L.J., G. (2004). Recent developments on cellulases and carbohydrate-binding modules with cellulose affinity. Biotechnology Letters 26, 1683-1693.

Hilge, M., Gloor, S.M., Rypniewski, W., Sauer, O., Heightman, T.D., Zimmermann, W., Winterhalter, K., and Piontek, K. (1998). High-resolution native and complex structures of thermostable beta- mannanase from Thermomonospora fusca - substrate specificity in glycosyl hydrolase family 5. Structure 6, 1433-1444.

Him, J.L.K., Pelosi, L., Chanzy, H., Putaux, J.L., and Bulone, V. (2001). Biosynthesis of (1 -> 3)-beta-D-glucan (callose) by detergent extracts of a microsomal fraction from Arabidopsis thaliana. European Journal of Biochemistry 268, 4628-4638.

Himmel, M.E., Ruth, M.F., and Wyman, C.E. (1999). Cellulase for commodity products from cellulosic biomass. Curr Opin Biotechnol 10, 358-364.

His, I., Driouich, A., Nicol, F., Jauneau, A., and Hofte, H. (2001). Altered pectin composition in primary cell walls of korrigan, a dwarf mutant of Arabidopsis deficient in a membrane-bound endo-1,4-beta- glucanase. Planta 212, 348-358.

55

Irwin, D., Shin, D.H., Zhang, S., Barr, B.K., Sakon, J., Karplus, P.A., and Wilson, D.B. (1998). Roles of the catalytic domain and two cellulose binding domains of Thermomonospora fusca E4 in cellulose hydrolysis. J Bacteriol 180, 1709-1714.

Irwin, D.C., Spezio, M., Walker, L.P., and Wilson, D.B. (1993). Activity Studies of 8 Purified Cellulases - Specificity, Synergism, and Binding Domain Effects. Biotechnology and Bioengineering 42, 1002-1013.

Joliff, G., Beguin, P., Juy, M., Millet, J., Ryter, A., Poljak, R., and Aubert, J.P. (1986). Isolation, Crystallization and Properties of a New Cellulase of Clostridium-Thermocellum Overproduced in Escherichia-Coli. Bio-Technology 4, 896-900.

Juge, N., Andersen, J.S., Tull, D., Roepstorff, P., and Svensson, B. (1996). Overexpression, purification, and characterization of recombinant barley alpha-amylases 1 and 2 secreted by the methylotrophic yeast Pichia pastoris. Protein Expression and Purification 8, 204-214.

Kallas, A.M., Piens, K., Denman, S.E., Henriksson, H., Faldt, J., Johansson, P., Brumer, H., and Teeri, T.T. (2005). Enzymatic properties of native and deglycosylated hybrid aspen (Populus tremula x tremuloides) xyloglucan endotransglycosylase 16A expressed in Pichia pastoris. Biochemical Journal 390, 105-113.

Khademi, S., Guarino, L.A., Watanabe, H., Tokuda, G., and Meyer, E.F. (2002). Structure of an endoglucanase from termite, Nasutitermes takasagoensis. Acta Crystallographica Section D-Biological Crystallography 58, 653-659.

Kim, J.B., Olek, A.T., and Carpita, N.C. (2000). Cell wall and membrane-associated exo-beta-D-glucanases from developing maize seedlings. Plant Physiol 123, 471-486.

Kimura, S., Laosinchai, W., Itoh, T., Cui, X.J., Linder, C.R., and Brown, R.M. (1999). Immunogold labeling of rosette terminal cellulose-synthesizing complexes in the vascular plant Vigna angularis. Plant Cell 11, 2075-2085.

Kirst, M., Johnson, A.F., Baucom, C., Ulrich, E., Hubbard, K., Staggs, R., Paule, C., Retzel, E., Whetten, R., and Sederoff, R. (2003). Apparent homology of expressed genes from wood-forming tissues of loblolly pine (Pinus taeda L.) with Arabidopsis thaliana. Proc. Natl. Acad. Sci. U. S. A. 100, 7383-7388.

Koivula, A., Kinnari, T., Harjunpaa, V., Ruohonen, L., Teleman, A., Drakenberg, T., Rouvinen, J., Jones, T.A., and Teeri, T.T. (1998). Tryptophan 272: an essential determinant of crystalline cellulose degradation by Trichoderma reesei cellobiohydrolase Cel6A [published erratum appears in FEBS Lett 1999 Mar 26;447(2-3):334]. FEBS Lett 429, 341-346.

Koshland, D.E. (1953). Biol.Rev 28, 416-436. Koyama, M., Helbert, W., Imai, T., Sugiyama, J., and Henrissat, B. (1997). Parallel-up

structure evidences the molecular directionality during biosynthesis of bacterial cellulose. Proceedings of the National Academy of Sciences of the United States of America 94, 9091-9095.

Kurokawa, J., Hemjinda, E., Arai, T., Kimura, T., Sakka, K., and Ohmiya, K. (2002). Clostridium thermocellum cellulase CelT, a family 9 endoglucanase without an Ig-like domain or family 3c carbohydrate-binding module. Appl Microbiol Biotechnol 59, 455-461.

Lam, B.C.H., Sage, T.L., Bianchi, F., and Blumwald, E. (2001). Role of SH3 domain-containing proteins in clathrin-mediated vesicle trafficking in Arabidopsis. Plant Cell 13, 2499-2512.

Lane, D.R., Wiedemeier, A., Peng, L., Hofte, H., Vernhettes, S., Desprez, T., Hocart, C.H., Birch, R.J., Baskin, T.I., Burn, J.E., Arioli, T., Betzner, A.S., and Williamson, R.E. (2001). Temperature-sensitive alleles of RSW2 link the

56

KORRIGAN endo-1,4-beta- glucanase to cellulose synthesis and cytokinesis in Arabidopsis. Plant Physiol 126, 278-288.

Libertini, E., Li, Y., and McQueen-Mason, S.J. (2004). Phylogenetic analysis of the plant endo-beta-1,4-glucanase gene family. Journal of Molecular Evolution 58, 506-515.

Lo Conte, L., Ailey, B., Hubbard, T.J.P., Brenner, S.E., Murzin, A.G., and Chothia, C. (2000). SCOP: a Structural Classification of Proteins database. Nucleic Acids Research 28, 257-259.

Loopstra, C.A., Mouradov, A., Vivian-Smith, A., Glassick, T.V., Gale, B.V., Southerton, S.G., Marshall, H., and Teasdale, R.D. (1998). Two pine endo-beta-1,4-glucanases are associated with rapidly growing reproductive structures. Plant Physiol 116, 959-967.

Lynd, L.R., Weimer, P.J., van Zyl, W.H., and Pretorius, I.S. (2002). Microbial cellulose utilization: Fundamentals and biotechnology. Microbiology and Molecular Biology Reviews 66, 506-+.

Mandelman, D., Belaich, A., Belaich, J.P., Aghajari, N., Driguez, H., and Haser, R. (2003). X-ray crystal structure of the multidomain endoglucanase Cel9G from Clostridium cellulolyticum complexed with natural and synthetic cello-oligosaccharides. Journal of Bacteriology 185, 4127-4135.

Master, E.R., Rudsander, U.J., Zhou, W.L., Henriksson, H., Divne, C., Denman, S., Wilson, D.B., and Teeri, T.T. (2004). Recombinant expression and enzymatic characterization of PttCel9A, a KOR homologue from Populus tremula x tremuloides. Biochemistry-Us 43, 10080-10089.

Matthysse, A.G., White, S., and Lightfoot, R. (1995). Genes Required for Cellulose Synthesis in Agrobacterium-Tumefaciens. J Bacteriol 177, 1069-1075.

Mellerowicz, E.J., Baucher, M., Sundberg, B., and Boerjan, W. (2001). Unravelling cell wall formation in the woody dicot stem. Plant Molecular Biology 47, 239-274.

Molhoj, M., Ulvskov, P., and Dal Degan, F. (2001). Characterization of a functional soluble form of a Brassica napus membrane-anchored endo-1,4-beta-glucanase heterologously expressed in Pichia pastoris. Plant Physiol 127, 674-684.

Molhoj, M.P., S. Hofte H. (2002). Towards understanding the role of membrane-bound endo-β-1,4-glucanases in cellulose biosynthesis. Plant Cell Physiol 43, 1399-1406.

Nakashima, K., Yamada, L., Satou, Y., Azuma, J., and Satoh, N. (2004). The evolutionary origin of animal cellulose synthase. Development Genes and Evolution 214, 81-88.

Nicol, F., His, I., Jauneau, A., Vernhettes, S., Canut, H., and Hofte, H. (1998a). A plasma membrane-bound putative endo-1,4-beta-D-glucanase is required for normal wall assembly and cell elongation in Arabidopsis. Embo J 17, 5563-5576.

Nicol, F., His, I., Jauneau, A., Vernhettes, S., Canut, H., and Hofte, H. (1998b). A plasma membrane-bound putative endo-1,4-beta-D-glucanase is required for normal wall assembly and cell elongation in Arabidopsis. Embo Journal 17, 5563-5576.

Nishiyama, Y., Langan, P., and Chanzy, H. (2002). Crystal structure and hydrogen-bonding system in cellulose 1 beta from synchrotron X-ray and neutron fiber diffraction. Journal of the American Chemical Society 124, 9074-9082.

Nishiyama, Y., Sugiyama, J., Chanzy, H., and Langan, P. (2003). Crystal structure and hydrogen bonding system in cellulose 1(alpha), from synchrotron X-ray and neutron fiber diffraction. Journal of the American Chemical Society 125, 14300-14306.

Nuhse, T.S., Stensballe, A., Jensen, O.N., and Peck, S.C. (2004). Phosphoproteomics of the arabidopsis plasma membrane and a new phosphorylation site database. Plant Cell 16, 2394-2405.

O'Leary, J., Radcliffe, C., Willis, A., Dwek, R., Rudd, P., and Downing, K. (2004). Identification and removal of O-linked and non-covalently linked sugars from

57

recombinant protein produced in Pichia pastoris. Protein Expression and Purification 38, 217-227.

Ohmiya, Y., Takeda, T., Nakamura, S., Sakai, F., and Hayashi, T. (1995). Purification and properties of wall-bound endo-1, 4-beta-glucanase from suspension-cultured poplar cells. Plant Cell Physiol 36, 607-614.

Ohmiya, Y., Samejima, M., Shiroishi, M., Amano, Y., Kanda, T., Sakai, F., and Hayashi, T. (2000). Evidence that endo-1,4-beta-glucanases act on cellulose in suspension- cultured poplar cells. Plant J 24, 147-158.

Ohmiya, Y., Nakai, T., Park, Y.W., Aoyama, T., Oka, A., Sakai, F., and Hayashi, T. (2003). The role of PopCel1 and PopCel2 in poplar leaf growth and cellulose biosynthesis. Plant Journal 33, 1087-1097.

Paredez, A.R., Somerville, C.R., and Ehrhardt, D.W. (2006). Visualization of cellulose synthase demonstrates functional association with microtubules. Science 312, 1491-1495.

Park, J.S., Nakamura, A., Horinouchi, S., and Beppu, T. (1993). Identification of the cellulose-binding domain of a Bacillus subtilis endoglucanase distinct from its catalytic domain. Biosci Biotechnol Biochem 57, 260-264.

Park, Y.W., Tominaga, R., Sugiyama, J., Furuta, Y., Tanimoto, E., Samejima, M., Sakai, F., and Hayashi, T. (2003). Enhancement of growth by expression of poplar cellulase in Arabidopsis thaliana. Plant Journal 33, 1099-1106.

Pear, J.R., Kawagoe, Y., Schreckengost, W.E., Delmer, D.P., and Stalker, D.M. (1996). Higher plants contain homologs of the bacterial celA genes encoding the catalytic subunit of cellulose synthase. Proc. Natl. Acad. Sci. U. S. A. 93, 12637-12642.

Peng, L., Kawagoe, Y., Hogan, P., and Delmer, D. (2002). Sitosterol-beta-glucoside as primer for cellulose synthesis in plants. Science 295, 147-150.

Plomion, C., Leprovost, G., and Stokes, A. (2001). Wood formation in trees. Plant Physiology 127, 1513-1523.

Porro, F., Bédué, O., Chanzy, H., and Heux, L. (2007). Solid-State 13C NMR Study of Na-Cellulose Complexes. Biomacromolecules 8, 2586-2593.

Quiocho, F.A., Spurlino, J.C., and Rodseth, L.E. (1997). Extensive features of tight oligosaccharide binding revealed in high-resolution structures of the maltodextrin transport chemosensory receptor. Structure 5, 997-1015.

Ragauskas, A.J., Williams, C.K., Davison, B.H., Britovsek, G., Cairney, J., Eckert, C.A., Frederick, W.J., Hallett, J.P., Leak, D.J., Liotta, C.L., Mielenz, J.R., Murphy, R., Templer, R., and Tschaplinski, T. (2006). The path forward for biofuels and biomaterials. Science 311, 484-489.

Robert, S., Mouille, G., and Hofte, H. (2004). The mechanism and regulation of cellulose synthesis in primary walls: lessons from cellulose-deficient Arabidopsis mutants. Cellulose 11, 351-364.

Robert, S., Bichet, A., Grandjean, O., Kierzkowski, D., Satiat-Jeunemaitre, B., Pelletier, S., Hauser, M.T., Hofte, H., and Vernhettes, S. (2005). An Arabidopsis endo-1,4-beta-D-glucanase involved in cellulose synthesis undergoes regulated intracellular cycling. Plant Cell 17, 3378-3389.

Rouvinen, J., Bergfors, T., Teeri, T., Knowles, J.K., and Jones, T.A. (1990). Three-dimensional structure of cellobiohydrolase II from Trichoderma reesei. Science 249, 380-386.

Sakon, J., Irwin, D., Wilson, D.B., and Karplus, P.A. (1997). Structure and mechanism of endo/exocellulase E4 from Thermomonospora fusca. Nat Struct Biol 4, 810-818.

Sato, S., Kato, T., Kakegawa, K., Ishii, T., Liu, Y.G., Awano, T., Takabe, K., Nishiyama, Y., Kuga, S., Nakamura, Y., Tabata, S., and Shibata, D. (2001). Role of the

58

putative membrane-bound endo-1,4-beta-glucanase KORRIGAN in cell elongation and cellulose synthesis in Arabidopsis thaliana. Plant Cell Physiol 42, 251-263.

Saxena, I.M., Brown, R.M., Jr., Fevre, M., Geremia, R.A., and Henrissat, B. (1995). Multidomain architecture of beta-glycosyl transferases: implications for mechanism of action. J Bacteriol 177, 1419-1424.

Schrick, K. (2004). A link between sterol biosynthesis, the cell-wall, and cellulose in Arabidopsis (vol 38, pg 227, 2004). Plant Journal 38, 562-562.

Schrick, K., Mayer, U., Martin, G., Bellini, C., Kuhnt, C., Schmidt, J., and Jurgens, G. (2002). Interactions between sterol biosynthesis genes in embryonic development of Arabidopsis. Plant Journal 31, 61-73.

Schrick, K., Mayer, U., Horrichs, A., Kuhnt, C., Bellini, C., Dangl, J., Schmidt, J., and Jurgens, G. (2000). FACKEL is a sterol C-14 reductase required for organized cell division and expansion in Arabidopsis embryogenesis. Genes & Development 14, 1471-1484.

Schubot, F.D., Kataeva, I.A., Chang, J., Shah, A.K., Ljungdahl, L.G., Rose, J.P., and Wang, B.C. (2004). Structural basis for the exocellulase activity of the cellobiohydrolase CbhA from Clostridium thermocellum. Biochemistry 43, 1163-1170.

Shani, Z., Dekel, M., Roiz, L., Horowitz, M., Kolosovski, N., Lapidot, S., Alkan, S., Koltai, H., Tsabary, G., Goren, R., and Shoseyov, O. (2006). Expression of endo-1,4-beta-glucanase (cel1) in Arabidopsis thaliana is associated with plant growth, xylem development and cell wall thickening. Plant Cell Reports 25, 1067-1074.

Sinnott, M.L. (1990). Catalytic Mechanisms of Enzymatic Glycosyl Transfer. Chemical Reviews 90, 1171-1202.

Sjunnesson, H. (2006). Tekniska museet, Stockholm. KTH&Co 3, 28. Somerville, C. (2006). Cellulose synthesis in higher plants. Annual Review of Cell and

Developmental Biology 22, 53-78. Somerville, C., Bauer, S., Brininstool, G., Facette, M., Hamann, T., Milne, J., Osborne,

E., Paredez, A., Persson, S., Raab, T., Vorwerk, S., and Youngs, H. (2004). Toward a systems approach to understanding plant-cell walls. Science 306, 2206-2211.

Spezio, M., Wilson, D.B., and Karplus, P.A. (1993). Crystal structure of the catalytic domain of a thermophilic endocellulase. Biochemistry 32, 9906-9916.

Stam, M.R., Blanc, E., Coutinho, P.M., and Henrissat, B. (2005). Evolutionary and mechanistic relationships between glycosidases acting on alpha- and beta-bonds. Carbohydrate Research 340, 2728-2734.

Standal, R., Iversen, T.G., Coucheron, D.H., Fjaervik, E., Blatny, J.M., and Valla, S. (1994). A New Gene Required for Cellulose Production and a Gene Encoding Cellulolytic Activity in Acetobacter-Xylinum Are Colocalized with the Bcs Operon. Journal of Bacteriology 176, 665-672.

Steenbakkers, P.J.M., Ubhayasekera, W., Goossen, H., van Lierop, E., van der Drift, C., Vogels, G.D., Mowbray, S.L., and den Camp, H. (2002). An intron-containing glycoside hydrolase family 9 cellulase gene encodes the dominant 90 kDa component of the cellulosome of the anaerobic fungus Piromyces sp strain E2. Biochem. J. 365, 193-204.

Sterky, F., Regan, S., Karlsson, J., Hertzberg, M., Rohde, A., Holmberg, A., Amini, B., Bhalerao, R., Larsson, M., Villarroel, R., Van Montagu, M., Sandberg, G., Olsson, O., Teeri, T.T., Boerjan, W., Gustafsson, P., Uhlen, M., Sundberg, B., and Lundeberg, J. (1998). Gene discovery in the wood-forming tissues of poplar: analysis

59

of 5, 692 expressed sequence tags, http:// www.biochem.kth.se/PopulusDB/. Proc Natl Acad Sci U S A 95, 13330-13335.

Sterky, F., Bhalerao, R.R., Unneberg, P., Segerman, B., Nilsson, P., Brunner, A.M., Charbonnel-Campaa, L., Lindvall, J.J., Tandre, K., Strauss, S.H., Sundberg, B., Gustafsson, P., Uhlen, M., Bhalerao, R.P., Nilsson, O., Sandberg, G., Karlsson, J., Lundeberg, J., and Jansson, S. (2004). A Populus EST resource for plant functional genomics. Proc. Natl. Acad. Sci. U. S. A. 101, 13951-13956.

Sturcova, A., His, I., Apperley, D.C., Sugiyama, J., and Jarvis, M.C. (2004). Structural details of crystalline cellulose from higher plants. Biomacromolecules 5, 1333-1339.

Sukumaran, R.K., Singhania, R.R., and Pandey, A. (2005). Microbial cellulases - Production, applications and challenges. Journal of Scientific & Industrial Research 64, 832-844.

Swissa, M., Aloni, Y., Weinhouse, H., and Benizman, M. (1980). Intermediary Steps in Acetobacter-Xylinum Cellulose Synthesis - Studies with Whole Cells and Cell-Free Preparations of the Wild-Type and a Cellulose-Less Mutant. Journal of Bacteriology 143, 1142-1150.

Szyjanowicz, P.M.J., McKinnon, I., Taylor, N.G., Gardiner, J., Jarvis, M.C., and Turner, S.R. (2004). The irregular xylem 2 mutant is an allele of korrigan that affects the secondary cell wall of Arabidopsis thaliana. Plant Journal 37, 730-740.

Teeri, T.T. (1997). Crystalline cellulose degradation: New insight into the function of cellobiohydrolases. Trends in Biotechnology 15, 160-167.

Teeri, T.T.B.I., H. Daniel, G. Gatenholm, P. (2007). Biomimetic engineering of cellulose-based materials. Trends in biotechnology 25, 299-306.

Tokoh, C., Takabe, K., Sugiyama, J., and Fujita, M. (2002). Cellulose synthesized by Acetobacter xylinum in the presence of plant cell wall polysaccharides. Cellulose 9, 65-74.

Tomme, P., Warren, R.A.J., Miller, R.C., Kilburn, D.G., and Gilkes, N.R. (1995). Cellulose-binding domains: Classification and properties. In Enzymatic Degradation of Insoluble Carbohydrates (Washington: Amer Chemical Soc), pp. 142-163.

Tomme, P., Kwan, E., Gilkes, N.R., Kilburn, D.G., and Warren, R.A. (1996). Characterization of CenC, an enzyme from Cellulomonas fimi with both endo- and exoglucanase activities. J Bacteriol 178, 4216-4223.

Tonouchi, N., Thara, N., Tsuchida, T., Yoshinaga, F., Beppu, T., and Horinouchi, S. (1995). Addition of a Small Amount of an Endoglucanase Enhances Cellulose Production by Acetobacter-Xylinum. Bioscience Biotechnology and Biochemistry 59, 805-808.

Trainotti, L., Pavanello, A., and Zanin, D. (2006). PpEG4 is a peach endo-beta-1,4-glucanase gene whose expression in climacteric peaches does not follow a climacteric pattern. Journal of Experimental Botany 57, 589-598.

Trimble, R.B., and Maley, F. (1984). Optimizing Hydrolysis of N-Linked High-Mannose Oligosaccharides by "Endo-Beta-N-Acetylglucosaminidase-H. Analytical Biochemistry 141, 515-522.

Tuskan, G.A., DiFazio, S., Jansson, S., Bohlmann, J., Grigoriev, I., Hellsten, U., Putnam, N., Ralph, S., Rombauts, S., Salamov, A., Schein, J., Sterck, L., Aerts, A., Bhalerao, R.R., Bhalerao, R.P., Blaudez, D., Boerjan, W., Brun, A., Brunner, A., Busov, V., Campbell, M., Carlson, J., Chalot, M., Chapman, J., Chen, G.L., Cooper, D., Coutinho, P.M., Couturier, J., Covert, S., Cronk, Q., Cunningham, R., Davis, J., Degroeve, S., Dejardin, A., Depamphilis, C., Detter, J., Dirks, B., Dubchak, I., Duplessis, S., Ehlting, J., Ellis, B., Gendler, K., Goodstein, D., Gribskov, M., Grimwood, J., Groover, A., Gunter, L., Hamberger, B., Heinze, B.,

60

Helariutta, Y., Henrissat, B., Holligan, D., Holt, R., Huang, W., Islam-Faridi, N., Jones, S., Jones-Rhoades, M., Jorgensen, R., Joshi, C., Kangasjarvi, J., Karlsson, J., Kelleher, C., Kirkpatrick, R., Kirst, M., Kohler, A., Kalluri, U., Larimer, F., Leebens-Mack, J., Leple, J.C., Locascio, P., Lou, Y., Lucas, S., Martin, F., Montanini, B., Napoli, C., Nelson, D.R., Nelson, C., Nieminen, K., Nilsson, O., Pereda, V., Peter, G., Philippe, R., Pilate, G., Poliakov, A., Razumovskaya, J., Richardson, P., Rinaldi, C., Ritland, K., Rouze, P., Ryaboy, D., Schmutz, J., Schrader, J., Segerman, B., Shin, H., Siddiqui, A., Sterky, F., Terry, A., Tsai, C.J., Uberbacher, E., Unneberg, P., Vahala, J., Wall, K., Wessler, S., Yang, G., Yin, T., Douglas, C., Marra, M., Sandberg, G., de Peer, Y.V., and Rokhsar, D. (2006). The genome of black cottonwood, Populus trichocarpa (Torr. & Gray). Science 313, 1596-1604.

Urbanowicz, B.R., Catala, C., Irwin, D., Wilson, D.B., Ripoll, D.R., and Rose, J.K.C. (2007a). A tomato endo-beta-1,4-glucanase, SlCel9C1, represents a distinct subclass with a new family of carbohydrate binding modules (CBM49). J Biol Chem 282, 12066-12074.

Urbanowicz, B.R., Bennett, A.B., del Campillo, E., Catala, C., Hayashi, T., Henrissat, B., Hofte, H., McQueen-Mason, S.J., Patterson, S.E., Shoseyov, O., Teeri, T.T., and Rose, J.K.C. (2007b). Structural organization and a standardized nomenclature for plant endo-1,4-beta-glucanases (Cellulases) of glycosyl hydrolase family 9. Plant Physiology 144, 1693-1696.

Urbanowicz, B.R.B., A. B.;del Campillo, E; Catalá, C; Hayashi, T.; Henrissat, B.; Höfte, H.; McQueen-Mason, S. J.; Patterson, S. E.; Shoseyov, O.; Teeri, T.T.; Rose, J. K.C. (2007). Structural Organization and a Standardized Nomenclature for Plant Endo-1,4-β−glucanases (´Cellulases´) of Glycosyl Hydrolase Family 9 (GH9). Plant Physiology accepted.

Varrot, A., Schulein, M., and Davies, G.J. (1999). Structural changes of the active site tunnel of Humicola insolens cellobiohydrolase, Cel6A, upon oligosaccharide binding. Biochemistry-Us 38, 8884-8891.

Watanabe, H., and Tokuda, G. (2001). Animal cellulases. Cell Mol Life Sci 58, 1167-1178. Waterhouse, P.M., and Helliwell, C.A. (2003). Exploring plant genomes by RNA-induced

gene silencing. Nature Reviews Genetics 4, 29-38. Vervecken, W., Kaigorodov, V., Callewaert, N., Geysens, S., De Vusser, K., and

Contreras, R. (2004). In vivo synthesis of mammalian-like, hybrid-type N-glycans in Pichia pastoris. Applied and Environmental Microbiology 70, 2639-2646.

Withers, S.G. (2001). Mechanisms of glycosyl transferases and hydrolases. Carbohydrate Polymers 44, 325-337.

Woolley, L.C., James, D.J., and Manning, K. (2001). Purification and properties of an endo-beta-1,4-glucanase from strawberry and down-regulation of the corresponding gene, cel1. Planta 214, 11-21.

Wu, S.C., Blumer, J.M., Darvill, A.G., and Albersheim, P. (1996). Characterization of an endo-beta-1,4-glucanase gene induced by auxin in elongating pea epicotyls. Plant Physiol 110, 163-170.

Vyas, N.K. (1991). Atomic features of protein-carbohydrate interactions. Current Opinion in Structural Biology 1, 732-740.

Xiang, Z. (2006). Advances in Homology Protein Structure Modeling. Current Protein and Peptide Science 7, 217-227.

Yasutake, Y., Kawano, S., Tajima, K., Yao, M., Satoh, Y., Munekata, M., and Tanaka, I. (2006). Structural characterization of the Acetobacter xylinum endo-beta-1,4-

61

glucanase CMCax required for cellulose biosynthesis. Proteins-Structure Function and Bioinformatics 64, 1069-1077.

Yoshida, K., and Komae, K. (2006). A rice family 9 glycoside hydrolase isozyme with broad substrate specificity for hemicelluloses in type II cell walls. Plant and Cell Physiology 47, 1541-1554.

Yoshida, K., Imaizumi, N., and Komae, K. (2005). Biochemical and molecular characterization of a rice endo-1,4-beta-glucanase from root tissues at the early stage of auxin-stimulated adventitious root formation. Plant and Cell Physiology 46, S205-S205.

Yung, M.H., Schaffer, R., and Putterill, J. (1999). Identification of genes expressed during early Arabidopsis carpel development by mRNA differential display: characterisation of ATCEL2, a novel endo-1,4-beta-D-glucanase gene. Plant J 17, 203-208.

Zhou, W.L., Irwin, D.C., Escovar-Kousen, J., and Wilson, D.B. (2004). Kinetic studies of Thermobifida fusca Cel9A active site mutant enzymes. Biochemistry-Us 43, 9655-9663.

Zou, J.-y., Kleywegt, G.J., Ståhlberg, J., Driguez, H., Nerinckx, W., Claeyssens, M., Koivula, A., Teeri, T.T., and Jones, T.A. (1999). Crystallographic evidence for substratering distortion and protein conformational changes during catalysis in cellobiohydrolase Cel6A from Trichoderma reesei. Structure 7, 1035-1045.

Zuo, J., Niu, Q.W., Nishizawa, N., Wu, Y., Kost, B., and Chua, N.H. (2000). KORRIGAN, an Arabidopsis endo-1,4-beta-glucanase, localizes to the cell plate by polarized targeting and is essential for cytokinesis. Plant Cell 12, 1137-1152.