the prokaryotes || cellulose-decomposing bacteria and their enzyme systems

CHAPTER 1.19Cellulose-decomposing Bacteria and Their Enzyme Systems

Cellulose-Decomposing Bacteria and Their Enzyme Systems

EDWARD A. BAYER, YUVAL SHOHAM AND RAPHAEL LAMED

IntroductionFrom an anthropocentric point of view, for mil-lennia, human culture has been intricatelyinvolved with cellulose, the major component ofthe plant cell wall. The development of the wood,paper and textile industries has served to incor-porate cellulosic materials into the fabric of oursociety. Within the past century, however, cellu-losic wastes, derived mainly from the sameindustries, have also become a major source ofenvironmental pollution. This chapter will con-centrate mainly on cellulose and the cellulolyticbacteria, in view of their importance to mankindand world ecology. Nevertheless, the true sub-strate of these bacteria—i.e., the complement ofplant cell wall polysaccharides in general—ismuch more complex than cellulose alone. Like-wise, the complement of enzymes—both thecellulolytic and the non-cellulolytic glycosylhydrolases—are produced concurrently in thesebacteria for the purpose of efficient synergisticdegradation of the complete substrate compositeas it appears in nature. Consequently, when wediscuss the cellulose-decomposing bacteria andtheir enzyme systems, we cannot ignore therelated noncellulolytic enzymes, and these willalso be treated, albeit secondarily, in the presentchapter.

It should also be noted that this chapter ofthe The Prokaryotes is a sequel to the previouschapter of the same title (authored by M.P.Coughlan and F. Mayer) from the second edi-tion of this treatise (Coughlan and Mayer,1992). The reader is cordially invited to consultthe earlier chapter (to be considered as Part A)as an excellent complement to our own (PartB).

The plant cell wall consists of an intricatemixture of polysaccharides (Carpita andGibeaut, 1993); cellulose, hemicellulose andlignin are its major constituents. These polymersare of a very robust nature. They both equip theplant with a stable structural framework andprotect the plant cell from the perils of itsenvironment. Despite its recalcitrant nature, inthe guise of dead or dying plant matter, the

polysaccharides of the plant cell wall provide anexceptional source of carbon and energy, anda multitude of different microorganisms hasevolved which are capable of degrading plantcell wall polysaccharides.

In any given ecosystem, the polysaccharide-degrading microbes are not alone, but rely on thecomplementary contribution of other bacterialand/or fungal species (Bayer and Lamed, 1992;Bayer et al., 1994; Ljungdahl and Eriksson,1985). The polymer-degrading strains play a pri-mary and crucial role in the ecosystem by con-verting the plant cell wall polysaccharides to therespective simple sugars and other degradationproducts (Fig. 1). They are assisted by satellitemicrobes, which cleanse the microenvironmentfrom the breakdown products, producing, in thefinal analysis methane and carbon dioxide.

In a given polysaccharide-degrading microor-ganism, the enzymes that catalyze the degrada-tion may occur either in the free state and/or indiscrete complexes with other similar types ofenzymes. The latter are called “cellulosomes.”Both the free enzymes and cellulosomal compo-nents are usually modular proteins, which con-tain a multiplicity of functional domains. The“free” enzymes comprise a single polypeptidechain, which contains a catalytic domain usuallyconnected to a cellulose-binding domain orCBD. Cellulosomes are exocellular macromo-lecular machines, designed for efficient degra-dation of cellulose and associated plant cell wallpolysaccharides (Bayer et al., 1998). In contrastto the free enzymes, the cellulosome complex iscomposed of a collection of subunits, each ofwhich comprises a set of interacting functionalmodules. Thus, one type of cellulosomal mod-ule, the CBD, is selective for binding to thesubstrate. Another family of modules, thecatalytic domains, is specialized for the hydroly-sis of the cellulose chains. Yet another comple-mentary pair of domains—the cohesins anddockerins—serves to integrate the enzymaticsubunits into the complex and the complex, inturn, into the cell surface. Multiple copies of thecohesins form an integrating subunit called“scaffoldin” to which the dockerin-containing

Prokaryotes (2006) 2:578–617DOI: 10.1007/0-387-30742-7_19

CHAPTER 1.19 Cellulose-Decomposing Bacteria and Their Enzyme Systems 579

enzymes are attached. This “Lego™”-likearrangement of the modular subunits generatesan intricate multicomponent complex, theenzymes of which are bound en bloc to theinsoluble substrate and act synergisticallytowards its complete digestion.

Inherent to the study of cellulases and relatedenzymes is their potential industrial application—particularly towards conversion of cellulosicbiomass. For reviews on the potential uses ofthese enzymes, the reader is referred to appro-priate reviews on the subject (Bhat, 2000;Himmel et al., 1999; Lynd et al., 1991).

Plant Cell Wall Polysaccharides

Plant cells produce a composite matrix of hardyand durable polysaccharides on the outer surfaceof the plasma membranes, called “the cell wall”(Carpita and Gibeaut, 1993). The cell wall con-fers a protective coating to the plant cell, provid-ing structure, turgidity and durability, whichrenders the cell resistant to the outer elements,including mechanical, chemical and microbialassault. Different types of plant cell tissuesexhibit different ratios of the three major typesof cell wall component; on the average, the cellwall contains roughly 40% cellulose, 30% hemi-

cellulose and 20% lignin, but the exact composi-tion of an individual type of plant varies greatly.The first two polymers are indeed polys-accharides. On the other hand, lignin is a heter-ogeneous, high-molecular-weight hydrophobicpolymer, which consists of nonrepeating aro-matic monomers connected via phenoxy link-ages (Higuchi, 1990; Lewis and Yamamoto,1990). Unlike cellulose and hemicellulose, whichare degraded aerobically or anerobically, lignindegradation requires oxygen and is limited tofilamentous prokaryotes (e.g., the ActinomycetesStreptomyces viridans) and fungi (e.g., Phanero-chaete chrysosporium, Bejerkendera adusta andPleurotus ostreatus), which produce a compli-cated set of enzymes that hydrolyze the polymer.In fact, the recalcitrant lignin interferes severelywith the access of enzymes to the cellulosecomponent, and is rate limiting for anaerobicdegradation of cellulose. In any case, the lignincomponent must be degraded or removed,before efficient degradation of cellulose can takeplace. Nevertheless, considering lignin is not apolysaccharide, it will not be discussed further inthis chapter.

Cellulose

Cellulose is the major constituent of plant matterand thus represents the most abundant organicpolymer on Earth. Cellulose is a remarkablystable homopolymer, consisting of a linear(unbranched) polymer of

b-1,4-linked glucoseunits. Chemically, the repeating unit is simplyglucose, but structurally, the repeating unit isthe disaccharide cellobiose, i.e., 4-O-(

b-D-glucopyranosyl)-D-glucopyranose, inasmuch aseach glucose residue is rotated 180

∞ relative toits neighbor (Fig. 2). The individual cellulosechains contain from about 100 to more than10,000 glucose units, packed tightly in parallelfashion into microfibrils by extensive inter- andintrachain hydrogen bonding interactions, whichaccount for the rigid structural stability of cellu-lose. The microfibrils exhibit variable amounts ofcrystalline and amorphous components, againdepending on the degree of polymerization, theextent of hydrogen bonding and, ultimately onthe source of the cellulose. The microfibrilsthemselves are further assembled into plant cellwalls, the tunic of some sea animals, pelliclesfrom bacterial origin, etc. Highly crystallineforms of cellulose include cotton, bacterial cellu-lose (from Acetobacter xylinum) and the cellu-lose from the algae, Valonia ventricosa, whichexhibit crystallinity levels of about 45%, 75%and 95%, respectively. The following reviews areavailable for more information on the structureof cellulose (Atalla, 1999; Atalla and Vander-Hart, 1984; Chanzy, 1990; O’Sullivan, 1997).

Fig. 1. Simplified schematic description of a typical ecosys-tem comprising degrading plant matter. Cellulolytic, xyl-anolytic and ligninolytic microbes combine to decompose themajor polysaccharide components to soluble sugars. “Satel-lite” microorganisms assimilate the excess sugars and othercellular end products, which are ultimately converted tomethane and carbon dioxide.

The Cellulose Ecosystem

CH4 CO2Cellularend products

Saccharolytic Microbe

Soluble

sugars

Lignin-degradingFungus

Cellulolytic Microbe

Xylanoilytic Microbe

Hemicel lu lose

C e l l u l o s eC e l l u l o s eL i g n i n

580 E.A. Bayer, Y. Shoham and R. Lamed CHAPTER 1.19

Hemicellulose

Hemicelluloses are relatively low-molecular-weight, branched heteropolysaccharides associ-ated with both cellulose and lignin and togetherbuild the plant cell wall material (Puls andSchuseil, 1993; Timell, 1967). The main backboneof hemicellulose is usually made of one or twosugars, which determines their classification.For example, the main backbone of xylan iscomposed of 1,4-linked-

b-D-xylopyranose units.Similarly, the backbone of galactoglucomannansis made of linear 1,4-linked

b-D-glucopyranoseand

b-D-mannopyranose units with a-1,6-linkedgalactose residues. Other common hemicellulo-ses include arabinogalactan, lichenins (mixed1,3-1,4-linked

b-D-glucans) and glucomannan.Most hemicellulases are based on a 1,4-

b-linkageand the main backbone is branched, whereas theindividual sugars may be acetylated or methy-lated. For example, the linear xylan backbone ishighly substituted with a variety of saccharideand nonsaccharide components (Fig. 3). In the

Fig. 2. Structure of cellulose. Three parallel chains that formthe 0, 1, 0 face are shown, and a glucose moiety and repeatingcellobiose unit are indicated. The model was built by Dr. JoséTormo, based on early crystallographic data. The diagram wasdrawn using RasMol 2.6.

HOCH2

CH2

O

OO

O

O O

O

CH2HO OH

OH OH

HO

HOHO

HOGlucose Cellobiose

The structure of cellulose

Fig. 3. Composition of a typical xylan component of hemicellulose. The xylobiose unit (

b-Xylp-

b-Xylp) is indicated by theblue-sided box, as are major substituents: MeaGlcA, methylglucuronic acid;

aAraf, arabinofuranosyl; OAc, acetyl group. Apresumed lignin attachment site to a feruloyl substituent of xylan is also illustrated. Sites of cleavage by selected hemicellulasesand carbohydrate esterases are also shown: Xyn, xylanase; Abn, arabinofuranosidase; Glr, glucuronidase; Axe, acetyl xylanesterase; Fae, ferulic acid esterase.

Structure and degradation sites of a typical xylan

Me–n–GleA

HOOC O

OO

O

OO

O CH – CHC – OH2C

OO

OO

OO

O

OO

O

O

O

OO

OO

O

O

HO

HO

HO

HO

HOHO

OH

OH

OH

OH

OH OH

HOOH

OH

OH

OH

OH

MeO

MeO

— —

HOH2C

A b n

G l r

Lignin

β Xylpβ Xylp

A x eOAC

Acety groupX y n

F a e

α-Arafα-Araf

Ferulayl group

OH


plant cell wall, xylan is closely associated withother wall components. The 4-O-methyl-

a-D-glucuronic acid residues can be ester-linked tothe hydroxyl groups of lignin, providing cross-links between the cell walls and lignin (Das et al.,1984). Similarly, feruloyl substituents serve ascrosslinking sites to either lignin or other xylanmolecules. Thus, the chemical complexity ofxylan is in direct contrast to the chemical simplic-ity of cellulose. Likewise, the structural diversityof the xylans is in contrast to the structural integ-rity of the cellulose microfibril. Consequently,unlike the crystalline-like character of cellulose,the hemicellulose component adopts a gel-likeconsistency, providing an amorphous matrix inwhich the rigid crystalline cellulose microfibrilsare embedded.

Cellulose-Degrading Bacteria

The cellulolytic microbes occupy a broad rangeof habitats. Some are free living and rid the envi-ronment of plant polysaccharides by convertingthem to the simple sugars, which they assimilate.Others are linked closely with cellulolytic ani-mals, residing in the digestive tracts of ruminantsand other grazers or in the guts of wood-degrading termites and worms (Haigler andWeimer, 1991). Cellulose-based ecosystemsinclude soils, swamps, marshes, rivers, lakes andseawater sediments, rotting grasses, leaves andwood, cotton bales, sewage sludge, silage, com-post heaps, muds and decaying vegetable matterin hot and volcanic springs, acid springs, andalkaline springs (Ljungdahl and Eriksson, 1985;Stutzenberger, 1990).

The cellulolytic microorganisms include pro-tozoa, fungi and bacteria and are ubiquitous innature. The cellulose-decomposing bacteriainclude aerobic, anaerobic, mesophilic and ther-mophilic strains, inhabiting a great variety ofenvironments, including the most extreme vis-à-vis temperature, pressure and pH. Cellulolyticbacteria also have been found in the gut of wood-eating worms, termites and vertebrate herbi-vores, all of which exploit anaerobic symbiontsfor the digestion of wood and fodder.

In nature, many cellulolytic species exist insymbiotic relationships with secondary microor-ganisms (Ljungdahl and Eriksson, 1985). The pri-mary microorganisms degrade cellulose directlyto cellobiose and glucose. Only part of the break-down products is assimilated by the polymerdegrading strain(s), and the rest is utilized by thesatellite microorganisms. Removal of the excessof sugars promotes further cellulose degradationby the primary species because cellobiose-induced inhibition of cellulase action and repres-sion of cellulase synthesis are precluded.

Modern interest in cellulolytic microorgan-isms was spawned by the decay of cotton fabricin army tents and military clothing in the SouthPacific jungles during World War II. The basicresearch program that resulted from this militaryproblem led to the establishment of the UnitedStates Army Natick Laboratories (Reese, 1976).The resultant research led to the discovery thatthe causative agent for the costly problem wasa cellulolytic fungi, Trichoderma viride (subse-quently renamed Trichoderma reesei). Subse-quent research, originally from the NatickLaboratories and later spreading to otherresearch institutes and universities, led to theidentification and classification of thousands ofdifferent strains of cellulolytic fungi and bacte-ria. Many of the major types of cellulolytic bac-teria have been listed in Part A of the secondedition of The Prokaryotes (Coughlan andMayer, 1992). Since the latter publication, themajor emphasis in the area has not concentratedon the discovery or description of new cellu-lolytic strains. Rather, research in the area duringthe past decade has centered on characterizingthe enzymes and enzyme systems from selectedbacteria that degrade cellulose in particular andplant cell wall polysaccharides in general.

Enzymes That Degrade Plant Cell Wall Polysaccharides

The chemical and structural intricacy of plantcell wall polysaccharides is matched by the diver-sity and complexity of the enzymes that degradethem. The cellulases and hemicellulases are fam-ily members of the broad group of glycosylhydrolases, which catalyze the hydrolysis of oli-gosaccharides and polysaccharides in general(Gilbert and Hazlewood, 1993; Kuhad et al.,1997; Ohmiya et al., 1997; Schülein, 1997;Tomme et al., 1995a; Viikari and Teeri, 1997;Warren, 1996; Wilson and Irwin, 1999).

Historically, the type of substrate and mannerin which a given enzyme interacted with its sub-strate were decisive in the classification of theglycosidases, as established first by the EnzymeCommission (EC) and later by the Nomencla-ture Committee of the International Union ofBiochemistry (IUB). Enzymes were usuallynamed and grouped according to the reactionsthey catalyzed. Thus, cellulases, xylanases, man-nanases and chitinases were grouped a priori indifferent categories. Moreover, enzymes thatcleave polysaccharide substrates in the middle ofthe chain (“endo”-acting enzymes) versus thosewhich clip at the chain ends (“exo”-actingenzymes) were also placed in different groups.


For example, in the case of cellulases, the endo-glucanases were grouped in EC 3.2.1.4, whereasthe exoglucanases (i.e., cellobiohydrolases) wereclassified as EC 3.2.1.91.

The historical division of enzymes is inappro-priate for classification of the cellulases and otherglycosyl hydrolases. Like other enzymes (e.g.,proteases, etc.), previous classification systems ofthe glycosyl hydrolases centered on the types ofsubstrates and the bonds cleaved by a givenenzyme. The problem with the glycosyl hydro-lases is that the polysaccharide substrates andparticularly the bonds they cleave are all quitesimilar, and classification of the different typesof enzymes according to conventional criteriaoften misses the mark. Consequently, alternativeapproaches were pursued. The recent trend is toclassify the different glycosyl hydrolases intogroups based on common structural fold andmechanistic themes (Davies and Henrissat, 1995;Henrissat, 1991; Henrissat and Bairoch, 1996;Henrissat and Davies, 1997; Henrissat et al.,1998). A comprehensive website that providesa catalog of the different glycosyl hydrolasefamilies is now available (Coutinho andHenrissat, 1999a; Coutinho and Henrissat,1999c; [Carbohydrate-Active Enzymes server(afmb.cnrs-mrs.frl)]. The website also providesexcellent introductory explanatory material, and

the interested reader is encouraged to use thissite extensively.

It is interesting that the distinction betweenendo- and exo-acting enzymes is also reflectedby the architecture of the respective class ofactive site, even within the same family ofenzymes (Fig. 4). The endoglucanases, forexample, are commonly characterized by agroove or cleft, into which any part of a linearcellulose chain can fit. On the other hand, theexoglucanases bear tunnel-like active sites,which can only accept a substrate chain via itsterminus. The exo-acting enzyme apparentlythreads the cellulose chain through the tunnel,wherein successive units (e.g., cellobiose) wouldbe cleaved in a sequential manner. The sequen-tial hydrolysis of a cellulose chain is a relativelynew notion of growing importance, which hasearned the term “processivity” (Davies andHenrissat, 1995), and processive enzymes areconsidered to be key components whichcontribute to the overall efficiency of a givencellulase system.

Though instructive, there is growing dissatis-faction with the endo/exo terminology. As ourunderstanding of the nature of catalysis by theseenzymes progresses, it has become clear thatsome enzymes are capable of both endo- andexo-action (Johnson et al., 1996; Morag et al.,

Fig. 4. Structures of a typical endoglucanase and exoglucanase. In each case, the structure is viewed from a perspective, whichdemonstrates the comparative architecture of the respective active site. Despite the sequence similarity of both enzymes andtheir classification as family-6 glycosyl hydrolases, their respective active-site architecture is different. The endoglucanase(endoglucanase E2 from the bacterium, Thermomonospora fusca, PDB code 1TML) is characterized by a deep cleft toaccommodate the cellulose chain at any point along its length, whereas the active site of the exoglucanase (cellobiohydrolaseCBHI from the cellulolytic fungus, Trichoderma reesei, PDB code 1CEL) bears an extended loop that forms a tunnel, throughwhich one of the termini of a cellulose chain can be threaded. The ribbon diagrams, showing the secondary structures (

a-helices and

b-strands) of the two enzymes, were drawn using RasMol 2.6.

Typical cellulase structures:"Endo" versus "Exo"

Endoglucanase Exoglucanase


1991; Reverbel-Leroy et al., 1997; Sakon et al.,1997). Moreover, some glycosyl hydrolase fami-lies include both endo- and exoenzymes, againindicating that the mode of cleavage can be inde-pendent of sequence homology and structuralfold. In this context, relatively minor changes inthe lengths of relevant loops in the general prox-imity of the active site, may dictate the endo- orexo-mode of action without significant differ-ences in the overall fold.

Owing to subtle but diverse chemical andstructural aspects of the substrates involved,plant cell wall degrading enzymes do not followthe same rules as common enzyme standards,such as simple proteases, DNAse, RNAse andlysozyme. In fact, the cellulases and hemicellu-lases are usually very large enzymes, whosemolecular masses often exceed those of pro-teases by factors of 2–5 and more. Their polypep-tide chains partition into a series of functionalmodules and linker segments (frequently glyco-sylated), which together determine their overallactivity characteristics and interaction with theirsubstrates and/or with other components of thecellulolytic and hemicellulolytic system.

Cellulases

The cellulases include the large number ofendo- and exoglucanases which hydrolyze

b-1,4-glucosidic bonds within the chains that comprisethe cellulose polymer (Béguin and Aubert, 1994;Haigler and Weimer, 1991; Tomme et al., 1995b).Thus, in principle, the degradation of celluloserequires the cleavage of a single type of bond.Nevertheless, in practice, we find that cellulolyticmicroorganisms produce a variety of comple-mentary cellulases of different specificities frommany different families.

It may seem somewhat surprising that thecombined effect of so many different enzymesare required to degrade such a chemically sim-plistic substrate. This complexity reflects the dif-ficulties an enzyme system encounters upondegrading such a highly crystalline substrate ascellulose. As described in the previous section,cellulases that degrade the cellulose chain can beeither “endo-acting” or “exo-acting.” Moreover,the degradation of crystalline cellulose should beviewed three-dimensionally and in situ, wherethe cellulose chains are packed within the micro-crystal, thus generating the remarkably stablephysical properties of the crystalline substrate.The enzymes have to bind to the cellulose sur-face, localize and isolate suitable chains, destinedfor degradation. It would seem logical thatamorphous regions or defects in the crystallineportions of the substrate would be favorable sites

for initiation of the process. The structural asopposed to chemical heterogeneity of the sub-strate dictates the synergistic action of a complexset of complementary enzymes towards its com-plete digestion.

Various models have been suggested toaccount for the observed synergy between andamong two or more different types of cellulases.For example, an endo-acting enzyme can pro-duce new chain ends in the internal portion of apolysaccharide backbone, and the two newlyexposed chains would then be available foraction of exo-acting enzymes. In addition, twodifferent types of exoglucanases may exhibit dif-ferent specificities by acting on a cellulose chainfrom opposite ends (i.e., the reducing versus thenonreducing end of the polymer). Likewise, anendoglucanase may be selective for only one ofthe two sterically distinct glucosidic bonds on thecellulosic surface. In addition, some cellulasesmay display high levels of activity at the begin-ning of the degradative process, i.e., on the highlycrystalline material, whereas others would beselective for newly exposed, partially degradedchains, otherwise embedded within the crystal.Still others would show very high levels of activ-ity after the degradative process has advanced,and cellulose chains that have been freed of thecrystalline setting would then be hydrolyzedquite rapidly. A collection of various enzymes,which exhibit complementary specificities andmodes of action, would account for the observedsynergistic action of the complete cellulase “sys-tem” in digesting the cellulosic substrate.

In addition to endo- and exoglucanases,included in the overall group of cellulases are the

b-glucosidases (EC 3.2.1.21), which hydrolyzesterminal, nonreducing

b-D-glucose residuesfrom cello-oligodextrins. In particular, this typeof enzyme cleaves cellobiose—the major endproduct of cellulase digestion—to generate twomolecules of glucose. Some

b-glucosidases arespecific for cellobiose whereas others showbroad specificity for other

b-D-glycosides, e.g.,xylobiose. Often, the

b-glucosidases are associ-ated with the microbial cell surface and hydro-lyze cellobiose to glucose before, during or afterthe transport process.

Hemicellulases

Strictly speaking, hemicellulases are not the pre-cise subject of this chapter, since they do notdirectly sever the

b-1,4-glucosidic bond of cellu-lose. Nevertheless, in nature, they are essentialto the bacterial degradation of insoluble cellu-lose because the natural bacterial substrate—the plant cell wall—comprises an architecturally


cogent composite of cellulose and hemicellulose.In natural systems, the two types of polysaccha-rides cannot be easily separated, and microbialsystems have to deal simultaneously with both.The xylan component is particularly of interestfor several reasons: 1) xylan is a major hemicel-lulosic component of the plant cell wall, 2) thexylanases are well defined enzymes, closely asso-ciated with the cellulase and 3) the repeatingunits (both xylose and xylobiose) bear strikingstructural resemblance to their cellulosic coun-terparts (i.e., glucose and cellobiose).

In contrast to cellulose degradation, thedegradation of the hemicelluloses imposes asomewhat different challenge, since this groupof polysaccharides includes widely differenttypes of sugars or non-sugar constituents withdifferent types of bonds. Thus, the completedegradation of hemicellulose requires theaction of different types of enzymes. Theseenzymes, the hemicellulases, can differ in thechemical bond they cleave, or, as in the case ofthe cellulases, they may cleave a similar type ofbond but with different substrate- or productspecificity (Biely, 1985; Coughlan and Hazle-wood, 1993; Eriksson et al., 1990; Gilbert andHazlewood, 1993).

Hemicellulases can be divided into two maintypes, those that cleave the mainchain back-bone, i.e., xylanases or mannanases, and thosethat degrade sidechain substituents or short endproducts, such as arabinofuranosidase, glucu-ronidase, acetyl esterases and xylosidase. Likethe cellulases, hemicellulases can be of theendo- or exo-types. A schematic view of thetypes of bonds that would be hydrolyzed by dif-ferent types of hemicellulases is presented inFig. 3.

XYLAN-DEGRADING ENZYMES. Thexylanases are by far the most characterizedand studied of the hemicellulases and involvethe cleavage of a major mainchain backbone.Endoxylanases (1,4-

b-D-xylan xylanhydrolase,EC 3.2.1.8) hydrolyze the 1,4-

b-D-xylopyranosyllinkage of xylans, such as D-glucurono-D-xylansand L-arabino-D-xylan. These single-subunitenzymes from both fungi and bacteria exhibita broad range of physiochemical properties,whereby two main classes have been described:alkaline proteins of low Mr (

<30,000) and acidicproteins of high Mr. This general classificationscheme correlates with their assignment intoglycosyl hydrolase families 10 and 11, wherebythe former represents the high Mr xylanases andthe latter coincides with the low Mr enzymes. Thetwo families also differ in their catalytic proper-ties, such that the family 10 enzymes seem todisplay a greater versatility towards the substratethan that observed for those of family 11, and are

thus typically able to hydrolyze highly substi-tuted xylan more efficiently. The family 10 xyla-nases exhibit a (

b/

a)8 topology whereas thosefrom family 11 form a

b-jelly roll fold. Both fam-ilies show a retaining catalytic mechanism ofhydrolysis.

MANNAN-DEGRADING ENZYMES.Glucomannans and galactoglucomannans arebranched heteropolysaccharides found in hard-wood and softwood. The degradation of thesepolymers again involve many hydrolyticenzymes, including endo-1,4-

b-mannanase (EC3.2.1.78),

b-mannosidase (EC 3.2.1.25),

b-glucosidase (EC 3.2.1.21), and

a-galactosidase(EC 3.2.1.22). 1,4-

b-D-Mannanases hydrolyzemainchain linkages of D-mannans and D-galacto-D-mannans. These enzymes, both of theendo- or exo-types, are produced in variousmicroorganisms, including Bacillus subtilis,Aspergillus niger and intestinal and rumenbacteria and commonly occur in families 5 and26.

LICHENIN-DEGRADING ENZYMES.Lichenase (1,3-1,4-

b-D-glucan 4-glucanohydro-lase, EC 3.2.1.73) is a mixed linkage

b-glucanase,which cleaves the

b-1,4 linkages adjacent tothe

b-1,3 bonds of the lichenin substrate.According to [{afmb.cnrs-mrs.fr/~pedro/CAZY/db.html}{modern structure-based classification,lichenases can be members of families 8, 16 or17.

b-D-XYLOSIDASES. The 1,4-

b-D-xylosidases (1,4-

b-D-xylan xylohydrolase, EC3.2.1.37) hydrolyze xylo-oligosaccharides (i.e.,xylan breakdown products and mainly xylobi-ose) to xylose. These enzymes are either intra-cellular or extracellular components and areclosely associated with hemicellulolytic activi-ties. Monomeric, dimeric and tetrameric xylosi-dases have been found with Mrs of 26,000 to360,000. Many of the xylosidases act on a vari-ety of substrates. For example, Aspergillusniger produces an enzyme classified as a

b-xylosidase that can hydrolyze

b-galactosides,

b-glucosides and

a-arabinosides, in addition to

b-xylosides.SIDECHAIN-DEGRADING ENZYMES.

a-D-Glucuronidases (EC 3.2.1.39) catalyze thecleavage of the

a-1,2 glucosidic bond of 4-O-methyl-

a-D-glucuronic acid side chain. Thisbond has a stabilizing effect on the neighboringxylosidic bonds of the main chain. Several

a-glucuronidase genes have recently been clonedand sequenced and usually occupy family 67.

a-L-Arabinofuranosidases (

a-L-arabinofura-noside arabinofuranohydrolase, EC 3.2.1.55) isanother important enzyme that cleaves non-reducing terminal

a-L-arabinofuranosidic link-ages in arabinoxylan, L-arabinan, and other


L-arabinose containing polysaccharides. Theseenzymes are found either in the cell-associatedor extracellular form and can be members offamilies 43, 51 or 62.

1,4-

b-Mannosidases hydrolyze 1,4-linked

b-D-mannosyl groups from the nonreducing end.These enzymes (similar to

b-xylosidases) hydro-lyze mainly the end products of the mannanases,i.e. mannobiose and mannotriose.

Carbohydrate Esterases

The side chain substituents of xylan arecomposed not only of sugars but also of acidicresidues, such as acetic, ferulic (4-hydroxy-3-methoxycinnamic) or p-coumaric (4-hydroxycin-namic) acids. Carbohydrate esterases that cleavethese residues (see Fig. 3) are found in enzymepreparations from both hemicellulolytic and cel-lulolytic cultures (Borneman et al., 1993). Suchenzymes sometimes represent separate modules,separated by linker segments from other cellu-lolytic or hemicellulolytic catalytic modules inthe same polypeptide chain. Like the glycosylhydrolases, the carbohydrate esters are curr-ently classified according to sequence homologyand common structural fold.

Cellulases and Hemicellulases are Modular Enzymes

The initial contribution of biochemical methodsfor determining the characteristics of a given cel-lulase was extended immeasurably by the contri-bution of molecular biology and bioinformatics.By comparing the sequences of the cellulasesand related enzymes, an entirely new view ofthese enzymes emerged.

Cellulases and hemicellulases are composed ofa series of separate modules. This fact explainsthe very large size of some of these enzymes andgives us some insight into their complex mode ofaction. Each module or domain comprises a con-secutive portion of the polypeptide chain andforms an independently folding, structurallyand functionally distinct unit (Coutinho andHenrissat, 1999; Gilkes et al., 1991; Teeri et al.,1992). Each enzyme contains at least one cata-lytic module, which catalyzes the actual hydroly-sis of the glycosidic bond and provides the basisfor classification of the simple enzymes (i.e.,those containing a single catalytic module).Other accessory or “helper” domains assist ormodify the primary hydrolytic action of theenzyme, thus modulating the overall propertiesof the enzyme. Some of the different themesillustrating the modular compositions of the

cellulases and related enzymes are illustrated inFig. 5.

The Catalytic Modules—Families of Enzymes

The definitive component of a given enzyme isthe catalytic domain. Former EC-based classifi-cation schemes according to substrate specificityare now considered somewhat obsolete becausethey fail to take into account the structural fea-tures of the enzymes themselves. The catalyticdomains of glycosyl hydrolases are presentlycategorized into families according to amino acidsequence homology (Coutinho and Henrissat,1999; Henrissat, 1991; Henrissat and Bairoch,1996; Henrissat and Davies, 1997; Henrissatet al., 1998). For more information, see theCarbohydrate-Active Enzymes (CAZy), de-signed and maintained by Pedro Coutinho andBernard Henrissat.

The enzymes of a given glycosyl hydrolasefamily display the same topology, and the posi-tions of the catalytic residues are conserved withrespect to the common fold. In recent years,X-ray crystallography has provided a generaloverview of the structural themes of the glycosylhydrolases and their interaction with theirintriguing set of substrates (Bayer et al., 1998;Davies and Henrissat, 1995; Henrissat andDavies, 1997).

The mechanism of cellulose and hemicellulosehydrolysis occurs via general acid catalysis and isaccompanied by either an overall retention or aninversion of the configuration of the anomericcarbon (Davies and Henrissat, 1995; McCarterand Withers, 1994; White and Rose, 1997;Withers, 2001). In both cases, cleavage is cata-lyzed primarily by two active-site carboxylgroups. One of these acts as a proton donor andthe other as a nucleophile or base. Retainingenzymes function via a double-displacementmechanism, by which a transient covalentenzyme-substrate intermediate is formed (Fig.6A). In contrast, inverting enzymes employ asingle-step mechanism as shown schematically inFig. 6B. The distance between the acid catalystand the base represents the major structural dif-ference between the two mechanisms. In retain-ing enzymes, the distance between the twocatalytic residues is about 5.5 Å, whereas ininverting enzymes the distance is about 10 Å. Inthe inverting enzymes, additional space is pro-vided for a water molecule, involved directly inthe hydrolysis, and the resultant product exhibitsa stereochemistry opposite to that of the sub-strate. In all cases, the mechanism of hydrolysisis conserved within a given glycosyl hydrolase


family (Coutinho and Henrissat, 1999; Daviesand Henrissat, 1995; Henrissat and Davies,1997).

Cellulose-Binding Domains Versus Carbohydrate-Binding Modules

In addition to the catalytic module, free cellu-lases and hemicellulases usually contain at leastone cellulose-binding domain (CBD) as an inte-gral part of the polypeptide chain (Linder andTeeri, 1997; Tomme et al., 1995a). The CBDserves predominantly as a targeting agent todirect and attach the catalytic domain to theinsoluble crystalline substrate. Like the catalyticdomains, the CBDs are categorized into a seriesof families according to sequence homology andconsequent structural fold.

In some cases, the term “CBD” is deceptivebecause not all of the CBDs bind to crystallinecellulose. Some families (or subfamilies or fam-

ily members) bind either preferentially or addi-tionally to other insoluble polysaccharides, e.g.,xylan or chitin. For example, the family-5 CBDand some of the members of the family-3 CBDsbind to chitin as well as cellulose (Brun et al.,1997; Morag et al., 1995). Moreover, the family-2 CBDs can be divided into two subfamilies, oneof which indeed binds preferentially to insolublecellulose, but the other binds to xylan (Borastonet al., 1999). The molecular basis for this wasproposed to reflect the fact that in the first sub-family, 3 surface-exposed tryptophans contrib-ute to cellulose binding (Simpson et al., 1999b;Williamson et al., 1999). However, in the caseof the xylan-binding members, one of thesetryptophans is missing, whereas the other twoassume a different conformation, thereby allow-ing them to stack against the hydrophobic sur-faces of two xylose rings of a xylan substrate.Other types of CBD prefer less crystallinesubstrates (e.g., acid-swollen cellulose), single

Fig. 5. Scheme illustrating the diversity of the modular architecture of cellulases and other glycosyl hydrolases. The differentmodules are grouped into families according to conserved sequences as shown here symbolically. A. One of the most commontypes of cellulases consists of a catalytic module or domain, flanked by a cellulose-binding domain (CBD) at its N- or C-terminus. This particular enzyme shown in “A” comprises a catalytic domain from family 48 and a family-2 CBD.B. Cellulosomal enzymes are characterized by a “dockerin domain” attached to a catalytic domain. In this case, the sametype of enzyme as in “A,” carrying a family-48 catalytic module, harbors a dockerin domain instead of a CBD. C. Manycellulases contain “X domains,” i.e., domains of unknown (as yet undefined) function. D. Some enzymes have more than oneCBD or other type of carbohydrate-binding module (CBM). Often, one CBD, such as the family-3 CBD shown here, servesto bind the cellulase strongly to the flat surface of the insoluble substrate, whereas the other one (the family-3c CBD) actsin concert with the catalytic module by binding transiently to a single cellulose or hemicellulose chain. E. Some cellulosomalcellulases have a CBD or CBM together with a dockerin in the same polypeptide chain. F. Some cellulases have more thanone type of catalytic module, such as the family-5 and family-44 modules shown here, and the two probably work in concertedfashion to degrade the substrate efficiently.

Cellulases are Modular Proteins

Catalyticdomain

Catalyticdomain

Catalyticdomain

Catalyticdomain

Catalyticdomain

Catalyticdomain

Catalyticdomain

CBD CBD CBD

CBD

CBDCBD

Dockerin Dockerin

A

B

C F

E

D48

48

48 5

9

9

44

2

2X

X

3

3c

3c 3


cellulose chains and/or soluble oligosaccharides,e.g, laminarin (1,3-

b-glucan) and barley 1,3/1,4-

b-glucan (Tomme et al., 1996; Zverlov et al.,2001). Still others exhibit alternative accessoryfunction(s), a topic to be described below inmore detail. Moreover, the CBDs responsiblefor the primary binding event may further dis-rupt hydrogen bonding interactions betweenadjacent cellulose chains of the microfibril (Dinet al., 1994), thereby increasing their accessibil-ity to subsequent attack by the hydrolyticdomain.

Consequently, the concept of CBD has beenbroadened and redefined as “CBM” i.e.,carbohydrate-binding module (Boraston et al.,1999; Coutinho and Henrissat, 1999). To date(March 2001), 26 different CBM families have

been described. The structures of CBDs from anumber of families and subfamilies have beendetermined, and an understanding of their struc-tures has provided interesting informationregarding the mode of binding to cellulose.Those that bind to crystalline substrates, appearto do so via a similar type of mechanism. One ofthe surfaces of such CBDs is characteristicallyflat and appears to complement the flat surfaceof crystalline cellulose. A series of aromaticamino acid residues on this flat surface form aplanar strip (Mattinen et al., 1997; Simpson andBarras, 1999a; Tormo et al., 1996) that stackopposite the glucose rings of a single cellulosechain. In addition, to the planar aromatic strip,several polar amino acid residues on the samesurface appear to anchor the CBD to two adja-

Fig. 6. The two major catalytic mech-anisms of glycosidic bond hydrolysis.A. The retaining mechanism involvesinitial protonation of the glycosidicoxygen via the acid/base catalystwith concomitant formation of a glyc-osyl-enzyme intermediate throughthe nucleophile. Hydrolysis of theintermediate is then accomplished viaattack by a water molecule, resultingin a product that exhibits the same ste-reochemistry as that of the substrate.B. The inverting mechanism involvesthe single-step protonation of the gly-cosidic oxygen via the acid/base cata-lyst and concomitant attack of a watermolecule, activated by the nucleo-phile. The resultant product exhibits astereochemistry opposite to that ofthe substrate. The type of mechanismis conserved within a given glycosylhydrolase family and dictated by theactive-site architecture and atomicdistance between the acid/base andnucleophilic residues (aspartic and/orglutamic acids).

A. Retaining Mechanism

B. Inverting Mechanism

acid/base acid/base

nucleophilenucleophile

nucleophile nucleophile

acid/base acid/base

acidacid

basebase

OO

H

HOR

H OR

H

O O

O

O

OOO

O

OO

O O

O

OO R

O

10 A

5.5 A

R

O

OOO

OO OO

OOOO

H

HO

O

O

O

O H

HH

H

H

OH


cent cellulose chains. The binding of the CBD tocrystalline cellulose would thus involve preciselyoriented, contrasting hydrophobic and hydro-philic interactions between the reciprocally flatsurfaces of the protein and the carbohydratesubstrate. Together they provide a selectivebiological interaction, which contributes tothe specificity that a CBD exhibits towards itsstructure.

In contrast to the interaction with the crystal-line cellulose surface, other CBMs seem to inter-act with single cellulose chains. The family-3c andfamily-4 CBDs preferentially bind to noncrystal-line forms of cellulose and clearly have a differ-ent function in nature (Johnson et al., 1996;Sakon et al., 1997; Tomme et al., 1996). Forexample, the role of family-4 CBD may be torecognize, bind to and deliver an appropriatecatalytic module to a cellulose chain, which hasbeen loosened or liberated from a more orderedarrangement within the cellulose microfibril. Thebinding of the family-3c CBD to single cellulosechains and its remarkable role in cellulosehydrolysis will be discussed later (Fig. 9).

The Family-9 Cellulases: An Example

This section pertains to enzyme diversity andhow a single type of catalytic module can bemodified by the class of helper module(s) thatflank its C- or N-terminus. We are only at thebeginning in our understanding of how the mod-ular arrangement affects the overall activity andfunction of a given enzyme.

In its simplest form, an enzyme would presum-ably consist of a single catalytic domain, usuallywith a standard CBM, which would target theenzyme to the crystalline substrate. Indeed, thisis the norm for many individual glycosylhydrolase families. However, in others, e.g.,the family-9 cellulases, the catalytic domainscommonly occur in tandem with a number ofaccessory modules. Although the story is stillrather incomplete, we can discuss the currentlyavailable information regarding family 9 anddraw several interesting conclusions from the fewpublications on this currently developing subject.

Family-9 Theme and Variations. The crystalstructure of the family-9 catalytic module isknown and displays an (a/a)6-barrel fold andinverting catalytic machinery. However, few ofthe prokaryotic family-9 enzymes consist of asolitary catalytic module (Fig. 7A). Actually,there are numerous family-9 cellulases of plantorigin, the great majority of which are such lonecatalytic modules that lack accessory modules.Another type of eukaryotic family-9 cellulasethat lacks helper modules is produced by thetermite. The prokaryotic family-9 enzymes, how-

ever, are almost invariably decorated with a vari-ety of subsidiary modules that modulate theactivity of the catalytic module.

Microbial family-9 cellulases commonlyconform to one of the themes shown in Fig. 7. Inone of these, the catalytic module is followedimmediately downstream by a fused family-3cCBM (Fig. 7B). This particular type of CBMimparts special characteristics to the enzyme(see below). A second theme consists of animmunoglobulin-like (Ig) domain (of unknownfunction) immediately upstream to the catalyticdomain (Fig. 7C). A variation of the latter themeincludes a family-4 CBM at the N-terminus ofthe enzyme, followed by an Ig domain andfamily-9 catalytic domain (Fig. 7D). In additionto the above-described modular arrangement,each of the free prokaryotic enzyme systemsincludes a standard CBD that binds strongly tocrystalline cellulose.

Until very recently, there has been but oneexample in the prokaryotic world of a family-9enzyme that contains no helper domain. This isthe family-9 glycosyl hydrolase of the celluloso-mal scaffoldin from the cellulolytic anaerobicbacterium, Acetivibrio cellulolyticus (Ding et al.,1999). The A. cellulolyticus enzyme forms part ofa multimodular scaffoldin, but the catalytic mod-ule appears to be a functionally distinct entitythat lacks adjoining helper modules. The othermodules are conventional scaffoldin-associatedmodules, e.g., cohesins and a true cellulose-binding CBD. More recently, a dockerin-containing cellulosomal family-9 enzyme fromClostridium cellulovorans has been sequencedand also seems to lack adjoining helper modules(Tamaru et al., 2000b).

This thematic arrangement of the family-9cellulases is mirrored in the respectivesequences of the catalytic modules. The diver-gent sequences are reflected by the phylogeneticrelationship of the parent cellulases (Fig. 8).Thus, the simplest cellulases (the group Aeukaryotic cellulases from plants) that lack adja-cent helper modules are all phylogeneticallyrelated (theme A). Interestingly, the catalyticmodule of CipV from A. cellulolyticus is distinctfrom the other groups designated in Fig. 8, butclosest to the plant enzymes, as might be antici-pated from its lack of a helper module. In a sim-ilar manner, catalytic modules from cellulasesthat are fused to a family-3c CBD (group B), allmap within the same branch (theme B). On theother hand, the catalytic modules that bear anadjacent Ig-like domain all fall into a cluster onthe opposite side of the tree. Cellulases whichhave the Ig-like domain only (theme C) occupya small separate branch and those that alsoinclude a family-4 CBD (theme D) that developdistally to form a separate subcluster.


Theme A enzymes: CipV Acece, CipV scaffol-din from the cellulolytic bacterium, A. cel-lulolyticus (AF155197); and plant (eukaryotic)cellulases from Prunus persica (X96853), Popu-lus alba (D32166), Citrus sinensis (AF000135),Persea americana (M17634), Pinus radiata(X96853), Arabidopsis thaliana (X98543),Phaseolus vulgaris (M57400), Capsicum annuum(X97189), Lycopersicon esculentum (U20590).

Theme B enzymes: CelF Clotm, endogluca-nase F from Clostridium thermocellum (X60545);CelZ Closr, exoglucanase Z from Clostridiumstercorarium (X55299); CelA Calsa, cellulase Afrom Caldocellum saccharolyticum (L32742);CelG Cloce, endoglucanase G from Clostridiumcellulolyticum (M87018); CelI Clotm, endogluca-nase I from Clostridium thermocellum (L04735);

CelB Celfi, endoglucanase B from Cellulomonasfimi (M64644); E4 Thefu, endo/exoglucanase E4from Thermomonospora fusca (M73322).

Theme C enzymes: CelJ Clotm, cellulase Jfrom Clostridium thermocellum (D83704); CelDClotm, endoglucanase D from Clostridium ther-mocellum (X04584); CelC Butfi, endoglucanaseC from Butyrivibrio fibrisolvens (X55732).

Theme D enzymes: CbhA Clotm, cellobio-hydrolase A from Clostridium thermocellum(X80993); CelA Psefl, endoglucanase A fromPseudomonas fluorescens (X12570); CelC Celfi,endoglucanase C from Cellulomonas fimi(X57858); CelI Strre, endoglucanase I fromStreptomyces reticuli (X65616); E1 Thefu, endo-glucanase E1 from Thermomonospora fusca(L20094).

Fig. 7. Theme and variations: schematic view of the modular arrangement of the family-9 glycosyl hydrolases. A. The solitarycatalytic domain. B. The catalytic domain and fused family-3c cellulose-binding domain (CBD). C. Immunoglobulin-like (Ig)domain, fused to the catalytic domain. D. Successive family-4 CBD, Ig and catalytic domains. The representations of thedifferent modules are based on their known structures and are presented sequentially, left-to-right, from the N- to C-terminus.Structures (Ribbon diagrams produced by RasMol 2.6) in “A” and “B” are derived from cellulase E4 from Thermomonosporafusca (PDB code, 1TF4), those in “C” and “D” are from the CelD endoglucanase of C. thermocellum (PDB code, 1CLC).The figure used for the family-4 CBD in “D” is derived from the nuclear magnetic resonance (NMR) structure of the N-terminal CBD of Cellulomonas fimi b-1,4-glucanase CenC (PDB code, 1ULO). The structures in “B” and “C” are authenticviews of the respective crystallized bi-domain protein components. The CBD in “D” has been placed manually to indicateits N-terminal position in the protein sequence, but its spatial position in the quaternary structure and the structure of thelinker segment remains unknown.

Theme and Variations:The Family–9 enzymes

A

C D

B

GH9 GH9

GH9GH9

CBM3c

CBM4Ig Ig


The analysis of the designated catalytic mod-ules was performed using GenBee, based on therespective GenBank sequences (accession codesin parentheses).

Family-9 Crystal Structures. Two crystal struc-tures of family-9 cellulases have been elucidated,representing two subtypes of this particularfamily of glycosyl hydrolase. These are cellulaseE4 from Thermomonospora fusca (recentlyreclassified as Thermobifida fusca; Sakon et al.,1997) and CelD from Clostridium thermocellum(Juy et al., 1992). These two examples are archi-tecturally distinct—the E4 cellulase being anexample of a theme B family-9 enzyme (see Figs.7B and 8) and the CelD cellulase being a themeC enzyme. Fortunately, in both cases, one of theneighboring modules co-crystallized with the cat-alytic module, thus providing primary insightinto their combined structures. In the case ofT. fusca E4, the catalytic domain and neighbor-ing family-3c CBM were found to be intercon-nected by a long, rigid linker sequence, whichenvelops about half of the catalytic domain untilit connects to the adjacent CBM (Fig. 9A). Incontrast, in the C. thermocellum CelD, the cata-

lytic domain is adjoined at its N-terminus by a 7-stranded immunoglobulin-like (Ig) domain ofunknown function. The comparison between theE4 and CelD cellulases indicates that a giventype of catalytic module can be structurally andfunctionally modulated by different types ofaccessory domains.

Helper Modules. The family-3c CBM isspecial. To date, this particular type of CBMhas been found in nature associated exclusivelywith the family-9 catalytic domain. Structurally,the CBM is homologous to the otherfamily-3 CBMs, but contains substitutions inmany important surface residues. The three-dimensional crystal structure of the E4 cellulaserevealed the close interrelationship between thefamily-9 catalytic domain and the family-3cCBM, thus suggesting a functional role as ahelper module. This CBM seems not to binddirectly to crystalline cellulose but appears toact in concert with the catalytic domain bybinding transiently to the incoming cellulosechain, which is then fed into the active-site cleftpending hydrolysis (Gal et al., 1997; Irwin et al.,1998; Sakon et al., 1997; Fig. 9B).

The information derived from the family-9enzymes suggests that the activity of catalyticdomains can be modulated by accessory mod-ules. The accessory modules can either supple-ment or otherwise alter the overall propertiesof an enzyme (Bayer et al., 1998). The recur-rent appearance in nature of a given type ofmodule adjacent to a specific type of neighbor-ing catalytic domain may indicate a functionallysignificant theme. These observations raise thepossibility of a more selective role for certaintypes of CBM and other modules, wherebytheir association with certain types of catalyticdomains could signify a “helper” role. Thehelper module would provide hydrolytic effi-ciency and alter the catalytic character of theenzyme.

New Developments in Cellulase Analysis

The biochemical characterization of cellulases isin many cases a difficult task owing to the largevariety of enzyme types and modes of action. Atfirst glance, it is an intriguing phenomenon thatfor such a simple reaction (i.e., the hydrolysis ofthe b-1,4-glucose linkage in a linear glucosechain), Nature has evolved so many types ofcellulases. The vast varieties of enzymes arefound not only among the different species ofcellulolytic bacteria but also within the sameorganism. The reason for this extensive diversitycomes from the insoluble nature of celluloseand the fact that, although the chemical com-

Fig. 8. Phylogenetic analysis of the N-terminal family-9 cat-alytic module of CipV and its relationship with other family-9 members. The various theme groupings roughly follow thegroups shown in Fig. 7. Theme A (group A) enzymes lackassociated helper modules. Theme B (group B) enzymescarry a fused family-3c cellulose-binding domain (CBD)downstream to the catalytic module. Theme C (group C) andtheme D (group D) enzymes carry an immunoglobulin-like(Ig) domain upstream to the catalytic module, the theme Denzymes having an additional N-terminal family-4 CBM.

Phylogenetics ofFamily-9 Glycosyl Hydrolases

Theme BGH9-CBM3c

ClotmClotm

ClotmClosrCelF

CelD

CelJ

CbhA Clotm

CelA Psefl

CelC Celfl

Cell Strre

E1 Thefu

Capsium

Lycopersicon

Pha seolus

ArabidopsisPersea

Plant enzymes

Citrus

Populus

Prunus

Pinus

ButfiCelC

CelZ

CelG Cloce

Cell Clotm

CelA Galsa

CelB Celfi

E4 Thefu

CipV Acece

CBM4-Ig-GH9

Ig-GH9

Theme C

Theme D

Theme A


position of the homopolymer is rather trivial,the physical and three-dimensional arrangementof the chains within the crystalline andamorphous regions of the microfibril can differsignificantly.

Regarding the enzymes that degrade the sub-strate, the modular nature of the cellulasescontributes additional degrees of complexity inour quest to characterize a given enzyme. Thus,the number, types, and arrangement of the acces-

sory modules vis-à-vis the catalytic domain areimportant structural features that modulate theoverall activity of the enzyme in question. Thisdescriptive information should always be definedfor a recombinant enzyme. Whenever possible, itis desirable to determine the relative contribu-tion of the individual accessory modules to theactivity of the enzyme. In this regard, the affilia-tion of a given module, e.g., CBM, into a definedfamily does not necessarily define its contribu-tion to enzyme activity, as different specificitiesand functions have been attributed to differentmembers of the same family of module. More-over, sequences for almost 70 different “X” mod-ules (i.e., modules for which the function remainsundefined) are currently available (Coutinhoand Henrissat, 1999b; Coutinho and Henrissat,1999c), most of which probably play a bindingor processing role in assisting the catalyticdomain(s) in its capacity to hydrolyze thesubstrate.

A decade ago, the range of cellulases andhemicellulases within a given species wasassessed mainly by biochemical techniques. Insome cases, individual enzymes were isolatedand their properties assessed using desiredinsoluble or soluble substrates. Anotherapproach involved electrophoretic separation ofcell-derived or cell-free extracts, and analysis ofdesired activities using zymograms. There areadvantages and disadvantages with each of thesestrategies, and the employment of combinedcomplementary approaches is always advisable.

More recently, molecular biology techniqueshave been used to reveal cellulase and hemicel-lulase genes, which can often be characterized onthe basis of sequence homology with related,known genes (Béguin, 1990; Hazlewood andGilbert, 1993). If further information is requiredon the structure or action of a given enzyme, thegene can then be expressed in an appropriatehost organism, and the properties of the productcan be characterized.

It is always instructive to compare the proper-ties of an expressed gene product with those ofthe same protein isolated from the original bac-terial culture. The results may be surprising;there are hazards inherent to both approaches.Expression of a gene may yield preparationswith reduced or altered enzymatic properties.In this context, the expressed gene product maynot have been folded properly. It is of courseassumed that the investigator has taken the timeand trouble to sequence the cloned gene toensure no mutations have occurred. Unlike agene expressed in a host cell environment, thenative counterpart may have undergone post-translational modifications (e.g., glycosylation,proteolytic truncation, etc.) that improve its

Fig. 9. Structural aspects of family-9 theme-B cellulase E4from Thermobifida fusca. A. “Side view” of the E4 molecule,drawn using RasMol in spacefill mode. Shown are thefamily-9 catalytic module (turquoise, at left), the family-3ccarbohydrate-binding module (CBM; in yellow, at right) andthe intermodular linker (dark blue strip). The presumed pathof a single cellulose chain, from the CBM to the catalyticdomain, is shown at the bottom of the structure (arrows). Theenzyme also possesses a fibronectin-like domain (FN3) anda cellulose-binding family-2 CBM (not shown). Note that thelinker appears to serve a defined structural role by which thefamily-3c CBM is clamped tightly to the catalytic domain.Selected surface residues on the catalytic domain along theinterface of both the linker and the CBM3c also serve tofasten both features tightly to the catalytic module. B.“Bottom view” of the E4 molecule (~90∞ rotation of “A”).From this perspective, the proposed catalytic residues (red),positioned in the active site cleft, are clearly visible. The pathof the cellulose chain (arrows) passes through a successionof polar residues (green) on the bottom surface of the CBM,which would conceivably bind to the incoming cellulosechain and serve to direct it towards the active-site acidicresidues of the catalytic domain.

Architecture of T. fuscaFamily–9 Cellulase E4

A

90oB

Linker

Catalytic Domain

E424

D55D58

Family–3c CBM


physicochemical properties. Moreover, sincethe cellulase system in the native environmentincludes numerous enzyme types, often exhi-biting similar molecular masses and otherphysical characteristics, the reputed purificationof a given extracellular cellulase may still in-clude contaminating enzymes that alter (usuallyincreasing greatly due to synergistic action oftwo or more enzymes) the true enzymatic prop-erties of the desired enzyme. The onus belongsto the conscience of the investigating scientistwhen publishing the properties of a givenenzyme. Too often, erroneous data that enter thescientific literature are taken as fact. One shouldparticularly be wary of comparing enzymaticactivities of the same or similar types of enzymes(e.g., members of the same family) that havebeen published at different times and by differ-ent laboratories.

The assessment of cellulase activity is indeeda complicated undertaking, and there is no clearor standard methodology for doing so. This pre-dicament apparently reflects a combination offactors, including the complex nature of the sub-strate, the multiplicity of enzymes and theirsynergistic action, and the variety of productsformed. The fact that cellulose is an insolublesubstrate converted to lower-order cello-oligosaccharide products is a further complica-tion. It must be noted that as the cello-oligomersincrease in length, they become less soluble, suchthat cello-octaose of 8 glucose units is no longersoluble in aqueous solutions. Moreover, theaccumulation of one (particularly cellobiose) ormore of the cellulose degradation products maybe inhibitory towards enzymatic activity.

Today, the study of cellulase action usuallyincludes, in addition to conventional biochemicalassays, the analysis of the primary structure andthe assignment of the various domains intoknown families. The catalytic domains can usu-ally be assigned into one of the known glycosidehydrolase families (Henrissat and Bairoch, 1996;Henrissat and Davies, 1997). Whenever thesequence of a known polysaccharide-degradingenzyme failed to match a known family, a newfamily of glycosyl hydrolase was established. Thisapproach was extensively developed in the lastdecade, owing to the increasing number of avail-able DNA sequences and bioinformatics analysistools. At the same time, an increasing number ofcrystal or solution structures of various catalyticand accessory domains were published that allowus to examine a new protein sequence in light ofits structure. Sometimes, the publication of thestructure of an accessory domain precedes deter-mination of its function.

We can divide the analysis of a newlydescribed prospective cellulase into severalstages, such that a variety of complementary

approaches are currently in use to classify theenzyme. Some of the questions one may ask are:

1) What is the primary structure (the aminoacid sequence) of the enzyme? What are thebinding residues and/or binding module(s) asso-ciated with the enzyme? What are its other acces-sory domains and their respective role(s) incatalysis or stability?

2) Is the enzyme a “true” cellulase, i.e., itspreferred substrate is cellulose or cellulose deg-radation products, or whether the enzyme canact alone on insoluble cellulose.

3) What is the mode of action? Does theenzyme act as an endoglucanase, an exogluca-nase or a processive enzyme?

4) What is the stereochemistry of the reaction?Does the enzyme exhibit an inverting or retain-ing mechanism?

5) What are the catalytic residues: the acid/base residue and the nucleophile that character-ize a glycosyl hydrolase?

In the past ten years, several extensive reviewsand book chapters dealing with different assaysof cellulose degradation have been published(Ghose, 1987; Wood and Kellogg, 1988). In thistreatise, we will briefly summarize the variousapproaches currently in use and direct the readerto the relevant literature.

While characterizing the activity of a newenzyme preparation, one has to bear in mindseveral secondary or indirect issues, such as thepurity of the protein preparation, the sensitivityof the assay used, and the crossreactivity of theexpected enzymatic activities. In some cases,only detailed kinetic analysis can provide appro-priate characterization of the enzyme. As formany other types of glycosyl hydrolases, cellu-lases can exhibit crossreactivity with substratesof similar structure. This is particularly true whenusing, for example, p-nitrophenyl derivatizedsubstrates that provide highly sensitive assays.However, in many cases such a soluble syntheticchromogenic substrate can fit the active-sitepocket of a related but atypical enzyme, whichcatalyzes its hydrolysis. For example, family-10glycosyl hydrolases are typically xylanases butcan readily hydrolyze p-nitrophenyl cellobioside,which is a typical cellulase substrate. Without adetailed comparative kinetic analysis (kcat/Km)using different substrates, the true specificity ofthe enzyme might be overlooked. Given theamino acid sequence of the protein, its assign-ment to a given glycosyl hydrolase family willin many cases provide a reasonable general indi-cation of its activity. The description of the mod-ular structure provides additional knowledgethat can imply how the catalytic function mightbe modulated, but this knowledge can also be


misleading. In the final analysis, there is nosubstitute for extensive biochemical andbiophysical characterization of the given protein(recombinant or native) and its catalyticproperties.

General procedures for assaying for cellulaseand hemicellulase activities are very well docu-mented in the Methods in Enzymology volume160 (Wood and Kellogg, 1988). Conventionalprocedures for cellulase assay have been definedprecisely by the International Union of Pure andApplied Chemists (IUPAC; Ghose, 1987). How-ever, owing to the complexity of the substrateand enzyme systems, these procedures can onlyprovide a starting point for understanding thetrue nature of the enzyme in question.

Since the publication of Part A of this treatise(Coughlan and Mayer, 1992), many of the previ-ously reported assays of cellulase activity are stillin common use. These include the use of soluble,derivatized forms of cellulose, e.g., carboxyme-thyl cellulose and hydroxymethyl cellulose asconventional substrates for determining endo-glucanase activity. In addition, a derivatized,colored form of insoluble cellulose, i.e., azurecellulose, is frequently used as an indication ofcellulase activity. Zymograms with such coloredembedded substrates are useful in detectingendoglucanase or xylanase activities (Béguin,1983). Individual soluble cello-oligomers (cel-lotetraose, cellopentaose, cellohexaose, etc.) arestill used as substrates for analyzing enzymeaction, but the reliance on these substrates asdeterminants for assessing cellulase activity is nolonger a definitive approach. In the past decadeor so, newly developed substrate analoguesand reagents include thioglycoside substrates(Driguez, 1997), fluoride-derivatized sugars(Williams and Withers, 2000), chromophoric andfluorescent cello-oligosaccharides (Claeyssensand Henrissat, 1992; O’Neill et al., 1989; vanTilbeurgh et al., 1985). Recently, an ultraviolet-spectrophotometric method and an enzyme-based biosensor have been described (Bach andSchollmeyer, 1992; Hilden et al., 2001). In addi-tion, a novel and intriguing bifunctionalized flu-orogenic tetrasaccharide has been developed asan effective reagent for measuring the kineticconstants of cellulases by resonance energytransfer (Armand et al., 1997).

The thio-oligosaccharides serve as competitiveinhibitors that mimic natural substrates but areenzyme resistant (Driguez, 1997). In this typeof oligosaccharide, the oxygen of a bond tobe cleaved is replaced by sulfur. The thio-oligodextrins are sometimes more soluble thanthe native cellodextrins and longer chains can besynthesized. The modified sugars can be used inbiochemical studies or crystallographic studies togain some information about the geometry of the

active site or determine the mechanism of actionof an enzyme.

Determination of “True” Cellulase Activity:Solubilization of Crystalline CelluloseSubstrates True cellulase activity is usuallydefined as the ability to solubilize to an appre-ciable degree insoluble, “crystalline” forms ofcellulose. The extent of hydrolysis can be evalu-ated by turbidity assays, weight loss of insolublematerial, generation of reducing power, andaccumulation of soluble sugars. It is important torealize that crystalline cellulose is not of uniformcomposition and therefore the rate of catalysis isin most cases not linear with time or enzymeconcentration. Notably, the different prepara-tions of crystalline cellulose contain varyinglevels of loosely associated loops and chains. Thelatter are readily accessible to hydrolysis by agiven enzyme and lead to relatively high initialrates of activity, which do not reflect the actualdegree of true cellulase activity. For example,such loose chains can be degraded by a relativelyineffectual enzyme, whereas the crystalline por-tions of the substrate will be immune to furtherhydrolysis by the same enzyme. To overcomethese difficulties, IUPAC suggests determiningthe amount of enzyme required to achievedigestion of 5.2% of the insoluble substrate (e.g.,filter paper) in 16 h (Ghose, 1987; Irwin et al.,1993).

Cellulose substrates commonly in use includeAvicel, filter paper, cotton, Solka Floc, and morerecently bacterial cellulose from Acetobacteraceti and algal cellulose prepared from Valonia.Consequently, these assays should be treated asa relative and not quantitative assessment. Thenature of the original substrate selected—espe-cially its extent of crystallinity—should always betaken into account. Proper controls and refer-ence substrates should always be used. Oneshould be wary about comparison among resultsreported by different laboratories and even bydifferent researchers in the same laboratory.Nevertheless, such assays give an excellent indi-cation of whether a given enzyme preparationexhibits substantial activity towards crystallinecellulose substrates.

Endoglucanase Versus Exoglucanase Activ-ity As discussed earlier in this chapter, thecellulases have traditionally been divided intoeither endoglucanases or exoglucanases (Fig. 4).The biochemical or enzymatic assays that dis-criminate between these two modes of actionusually involve soluble forms of cellulose, i.e.,carboxymethyl or hydroxymethyl derivativesof cellulose. The action of a given enzyme onthese substrates is followed by determining the


amount of reducing ends generated by theenzyme and the degree of polymerization (DP).The reducing power is usually determined eitherby using reagents such as 3,5-dinitrosalicylic acid(DNS; Miller et al., 1960), ferricyanide (Kidbyand Davidson, 1973), or copper-arseno molyb-date (Green et al., 1989; Marais et al., 1966).

Despite their traditional popularity, these twomethods are intrinsically disadvantageous, owingto interference by metal ions and certain buffers.Moreover, such assays are sensitive to the chainlength of the reducing end. A more recentapproach involves the use of disodium 2,2¢-bicinchoninate (BCA) for determination ofreducing sugar. This procedure is more sensitivethan the conventional methods and gives com-parable values of reducing sugars for cello-dextrins of different lengths (Doner and Irwin,1992; Garcia et al., 1993; Vlasenko et al., 1998;Waffenschmidt and Jaenicke, 1987).

Viscosity-based measurements represent themost common approach for assessing the degreeof polymerization. This approach is highly sensi-tive for internal bond cleavage, which leads tosignificant reduction of the average molecularweight of the substrate. The comparison betweenthe amount of reducing sugars generated and theaverage molecular weight (i.e., viscosity or fluid-ity of the soluble cellulose substrate) gives a verygood indication whether an enzyme is essentiallyexo- or endo-acting.

The average degree of polymerization also canbe evaluated by size-exclusion chromatographyeither alone (Srisodsuk et al., 1998; Teeri, 1997)or combined with multiangle laser light scatter-ing (Vlasenko et al., 1998). Mass spectometricprocedures also can be applied to determine theidentity and distribution of degradation productsfollowing hydrolysis of cellulosic substrates by anenzyme (Hurlbert and Preston, 2001; Rydlundand Dahlman, 1997). The mode of enzymaticaction also can be appraised by determining theincrease in reducing power associated with theinsoluble versus the soluble fraction of the sub-strate. Increase in the proportion of reducingsugars associated with the soluble fraction indi-cates an exo-type of activity whereas a relativelylarge increase in the insoluble fraction wouldsuggest an endo-type of activity (Barr et al.,1996).

Exocellulases can exhibit different specificitiesdepending on their preference for the reducingor nonreducing end of the cellulose chain (Barret al., 1996; Teeri, 1997). This feature of anexocellulase can be determined either by usingoligosaccharide substrates labeled by tritium or18O at the reducing end. Other proceduresinvolve NMR, HPLC and/or mass spectrometricanalysis of products released from native (unla-beled) cello-oligosaccharides. Within the past

decade, the 3-D structures of enzyme-substratecomplexes have been obtained, and the specific-ities of the enzyme can be interpreted directlyfrom the data (Davies and Henrissat, 1995;Davies et al., 1998; Divne et al., 1998; Juy et al.,1992; Notenboom et al., 1998; Parsiegla et al.,1998; Rouvinen et al., 1990; Sakon et al., 1997;Zou et al., 1999).

Processivity One of the major recent concep-tual advances in assessing the mode of enzy-matic action of a cellulase is the concept of“processivity.” Processive enzyme action can bedefined as the sequential cleavage of a cellulosechain by an enzyme. In effect, exoglucanases areby nature and structure processive enzymes.Their tunnel-like active site thus allows proces-sive action on the cellulose chain. Endogluca-nases, however, were thought to be intrinsicallynonprocessive. However, the traditional distinc-tion between exo- and endocellulases was modi-fied recently.

Experiments combining two or more purifiedcellulases have shown that synergism can evenbe detected upon mixing two different types ofexo-acting enzymes. Such experiments led to therecognition that the exo enzymes can operate onboth ends (i.e., the reducing and nonreducingends) of the cellulose chain. Some enzymes,however, exhibit both endo and exo activities,although in such cases, the endocellulase activityis usually very low. In attempts to explain thesephenomena, the concept of processivity was pro-posed, by which the activity of the enzyme ischaracterized by the sequential hydrolysis of thecellulose chain. Implicit in this concept is thenotion that the catalytic site of the enzymeremains in continual and intimate contact with agiven chain of the cellulose substrate.

A more complete mechanistic picture of theprocessive nature of such cellulases was revealedwith the advent of high-resolution 3-D struc-tures. It was thus demonstrated that the cellulosechain makes contact with the protein at multiplesites, either via a tunnel-shaped structural ele-ment (such as that observed in the family-48enzymes) or by a special type of CBM (such asthe family-9 theme B cellulases). These arrange-ments allow the threading of the cellulose chaininto the active site, and, following initial cleav-age at the end of the chain, the enzyme canmove along the chain and position itself for thenext cleavage. In addition to this processivenature of the active site, these enzymes also canmake classic endo cleavages thus generating newends.

Biochemically processive enzymes exhibitcharacteristics between endo- and exoenzymes.They have low but detectable endo activitytowards soluble derivatives of cellulose (i.e.,


CMC), and may or may not possess exo activityon such substrates. With insoluble substrates,they will generate reducing power with a ratiobetween the soluble to the insoluble fractions ofabout 7. Endocellulases usually give a ratio ofless than 2, whereas exocellulases produce a ratioof 12 to 23 (Irwin et al., 1998).

Once the processive nature of an enzymehas been indicated experimentally, molecularinsight into the mechanisms responsible for thisfeature can be gained by determining the 3-Dcrystal structure of the active site together withmodel cellodextrins. In the case of the cellu-lases, the crystal structure of the catalyticdomain together with the fused module, com-bined with accumulating enzymatic activitydata, allowed further postulation as to theaccessory role of the fused module. The fusedCBM presumably interacts with a single cellu-lose chain and feeds it into the active site. Inter-estingly, this domain does not bind crystallinecellulose, but is inferred to act in dynamic bind-ing of the single cellulose chain prior to itshydrolysis, thereby imparting the quality of pro-cessivity to the enzyme. Once such a property isassociated with a given type of enzyme, the pri-mary structure of the protein can now be usedas an indication for all such enzymes. In thecase of the family-9 theme B enzymes, it is nowpossible to identify the catalytic domain (e.g.,glycosyl hydrolase family 9) and the additionalaccessory domains (in this case, family-3cCBM). Thus, the primary structure may by itselfgive a strong indication of the nature of theenzyme itself. Of course, the ultimate identifica-tion as to the mechanism of enzyme activity willcome from the detailed 3-D structure of theenzyme-substrate complex.

An intriguing recent development in theanalysis of the cellulolytic action of a givencellulase or a mixture of cellulase is the directtransmission electron microscopic (TEM) obser-vation of the enzymatic action on bacterialcellulose ribbons. The approach provides infor-mation as to the endo or exo preference of theenzyme, the extent of processivity as well as thedirectionality of hydrolysis (i.e., from the reduc-ing to the nonreducing ends or vice versa). Thisstrategy has been used to study the hydrolysis ofbacterial cellulose ribbons by individual purifiedenzymes, mixtures of purified enzymes, andintact cellulosomes.

Mechanism of Catalysis The mechanism ofcatalysis of cellulases address issues such asstereochemistry, binding and active-site residuesand transition state intermediates. Excellentreviews have been published recently coveringmany of these (Ly and Withers, 1999; McCarterand Withers, 1994; Rye and Withers, 2000;

Sinnott, 1990; White and Rose, 1997; Withers,2001; Withers and Aebersold, 1995; Zechel andWithers, 2000). The fact that the stereochemistryand catalytic residues are conserved betweenmembers of the same family allows the putativeidentification of these elements if one memberof the given (glycosyl hydrolase) has been char-acterized biochemically (Henrissat and Bairoch,1996; Henrissat et al., 1995; Henrissat andDavies, 1997).

The sterochemistry of the reaction can in mostcases be determined by proton NMR spectros-copy or by using chromatography systems thatallow the resolution of anomeric species. In thecase of NMR, the reaction between the testenzyme and its substrate is carried out in deuter-ated water (D2O) and the appearance of the ano-meric proton can be easily detected. Thus, for thedegradation of cellulose, a retaining enzymewould produce a product in the b configurationwhereas an inverting enzyme would yield thea-sugar.

The catalytic residues can be identified by per-forming site-directed mutagenesis on conservedacidic residues and studying the catalytic proper-ties of the mutants with substrates bearing differ-ent leaving groups. Commonly used phenolsubstituents include the following, listed in orderof leaving group ability (pKa values shownparenthetically): 2,4-dinitro (3.96) > 2,5-dinitro(5.15) > 3,4-dinitro (5.36) > 2-chloro-4-nitro(5.45) > 4-nitro (7.18) > 2-nitro (7.22) > 3,5-dichloro (8.19) > 3-nitro (8.39) > 4-cyano (8.49)> 4-bromo (9.34; Tull and Withers, 1994). Inretaining enzymes, the nucleophilic residue canbe identified directly by trapping the intermedi-ate with an appropriate inhibitor. Such inhibitorsinclude model saccharides containing a fluorinesubstituent in the 2- or 5-position and a goodleaving group, such as fluoride or dinitropheno-late (Williams and Withers, 2000). The substi-tuted substrate forms a relatively stable covalentsubstrate-enzyme complex, involving the nucleo-phile residues. The complex is then subjected toproteolytic cleavage and sequencing of the glyc-osylated peptide. Recently, the use of protocolsinvolving combined liquid chromatography andmass spectrometry has facilitated the identifica-tion of the modified residues.

The acid-base residue in a retaining enzymecan be identified by a combination of kinetics-based methodologies. Mutation of this residue(usually to alanine) should affect the rate of bothchemical steps, i.e., glycosylation and deglycosy-lation, though the effect on each step should bedifferent. The effect on the glycosylation step willdepend strongly on the leaving group ability ofthe aglycon. Thus, rates of hydrolysis for sub-strates with a poor leaving group should beaffected much more strongly than those with a


good leaving group. The deglycosylation step,however, will be affected equally for all sub-strates carrying different leaving groups, becausethe same glycosyl enzyme intermediate is hydro-lyzed during this step. Thus, detailed kinetic anal-ysis (i.e., determination of kcat and Km) withsubstrates bearing different leaving groups canreveal whether the corresponding mutation isthe acid-base residue. It should be noted that thisapproach requires synthetic substrates that arenot necessarily recognized by all families ofenzymes and are not necessarily commerciallyavailable. For example, the family-11 xylanasesfail to hydrolyze p-nitrophenyl xylobioside,which is an excellent substrate for the family-10xylanases. The assignment of the acid-base cata-lyst can also be examined by use of externalnucleophilic anions, such as azide. In thisapproach, termed “azide rescue,” the small azideanion enters the vacant space created by alaninereplacement of the acidic amino acid residue.The azide reacts with the anomeric carboninstead of a water molecule to form the corre-sponding b-glycosyl azide product. In theabsence of an acid-base catalyst, which normallyprovides general base catalysis during the secondstep, the deglycosylation step is severelyaffected. Thus, the acceleration of the reaction bythe mutant enzymes in the presence of theseexternal anions (provided that the second step israte limiting) is a good indication that a mutantresidue is the acid-base catalyst. Finally, theassignment of the acid-base catalyst can betested by comparing the pH-dependence profilesfor the wild-type and mutant enzymes. The pro-file for the native enzyme would approximate aperfect bell shape curve, reflecting the ionizationof the two active site carboxylic acids, whereasthe no reduction of activity at high pH valueswould be observed for the mutant. This pHdependency approach is also applicable for iden-tifying the nucleophile residues and the catalyticresidues in inverting enzymes.

Prokaryotic Cellulase Systems

The cellulolytic bacteria produce a variety of dif-ferent cellulases and related enzymes, whichtogether convert the plant cell wall polysaccha-rides to simple soluble sugars that can sub-sequently be assimilated. The complement ofcellulases and hemicellulases that are synthe-sized by a given bacterium for this purpose isreferred to as its “cellulase system.” Differentbacteria exploit different strategies for the ulti-mate degradation of their substrates. The givenstrategy is reflected by the complement andtype(s) of enzymes produced by a given bacte-rium. The bacterial cellulase system may be char-

acterized by free enzymes, cell-bound enzymes,multifunctional enzymes, cellulosomes, or anycombination of the latter.

Cellulase enzyme systems are comprised ofseveral different types of components, each typemay exist in a multiplicity of forms. To add to thecomplexity, the same component may exist asfree individual entities in the culture fluid, asindividual entities bound to cellulose, or associ-ated with the cell surface. Alternatively, an indi-vidual component may be organized as part of amulticomponent cellulosome complex attachedto the cell surface, to the cellulose, to both, or asfree complexes in the culture fluid. Furthermore,the situation existing during growth under oneset of conditions (e.g., pH, temperature, distribu-tion of carbon source, etc.) may not exist underanother, or may change considerably during thecourse of cultivation. The bacterium reacts tothese changes and its production of cellulasesand/or cellulosomes may reflect the dynamics ofthe growth conditions.

Free Enzymes

As mentioned earlier in this chapter, the freeenzymes in their simplest form comprise a cata-lytic module alone with no accessory domainsor modules. Such enzymes often specialize indegrading soluble oligosaccharide breakdownproducts. Alternatively, such single-modularenzymes may rely on an intrinsic association withinsoluble polysaccharide substrate such as cellu-lose, perhaps related to the active site of theenzyme.

A higher order level of organization and activ-ity are free enzymes composed of a polypeptidechain that includes both a catalytic domaintogether with a CBM. This basic bi-modulararrangement can be further extended by theinclusion of additional types of modules orrepeating units of the same module, all of whichserve to modulate the activity of the catalyticdomain on the substrate. The intact free enzyme,however, remains unattached to other enzymesand can work in an independent manner on agiven substrate.

Cell-Bound Enzymes

Some enzymes are connected directly to the cellwall. In Gram-positive bacteria, this is frequentlyaccomplished via a specialized type of module,the SLH (S-layer homology) module, previouslyshown to be associated with the cell surface ofGram-positive bacteria (Lupas et al., 1994). Thisarrangement may have evolved to provide amore economic degradation of insoluble sub-strates and to reduce competition with other bac-


teria for the soluble products, subject to diffusionin the media. As opposed to free enzymes, diffu-sion of an attached enzyme would itself beprevented.

Examples of enzymes, which are bound to thecell surface via an SLH module include, a family-5 cellulase and family-13 amylase-pullulanasefrom Bacillus, a family-10 xylanase fromCaldicellulosiruptor (Saul et al., 1990), a family-5 endoglucanase from Clostridium josui, afamily-16 lichenase and family-10 xylanase fromClostridium thermocellum (Jung et al., 1998), anda variety of enzymes (family-10 xylanases, afamily-5 mannanase and a family-13 amylase-pullulanase) from different species of Ther-moanaerobacter (Matuschek et al., 1996). Themodular architecture of these enzymes may beparticularly complicated, containing several dif-ferent modules in a single polypeptide chain,thus forming extremely large enzymes some-times comprising over 2,000 amino acids (Fig.10).

Multifunctional Enzymes

Some cellulases exhibit a more complex architec-ture in that more than one catalytic domain and/or CBD may be included in the same protein.Examples of such enzymes are the very similarcellulases from Anaerocellum thermophilum(Zverlov et al., 1998) and Caldocellum saccharo-lyticum (Te’o et al., 1995), both of which containa family-9 and a family-48 catalytic domain.Other paired catalytic domains include thosefrom family 44 and either family 5 or 9. Such anarrangement might indicate a close cooperationbetween two particular catalytic domains, whichmay lead to synergistic action on the cellulosicsubstrate, thus portending on a smaller scale theadvent of cellulosomes.

Like the cellulases, xylanases also tend toexhibit a modular structure, being composed ofmultiple domains joined by linker sequences.Family-10 and -11 xylanases may be linked in thesame polypeptide chain either to each other, tocatalytic domains from families 5, 16 and 43 orto carbohydrate esterases (Flint et al., 1993;Laurie et al., 1997). One particularly interestingcombination of multifunctional catalytic mod-ules that appear in the same polypeptide chainis a typical xylanase together with a feruloylesterase. Such a combination would allow therapid cleavage of hemicellulose from the ligninin natural systems, i.e., the plant cell wall (seeFig. 3). In this manner, the xylan chain would besevered by the xylanase component (Xyn in Fig.3) and the lignin-xylan association would bedisconnected simultaneously by the feruloyl acidesterase (Fae in Fig. 3).

Indeed, some xylanases are extremely com-plex in their modular architecture (Fig. 11). Inaddition to multiple catalytic modules, theseenzymes often contain several different types ofCBMs. Why would such a xylanase contain sev-eral types of CBM? And why would a xylanasecontain a cellulose-specific CBD? Unlike thecase of various cellulases, for which the CBD isusually essential for degrading insoluble crystal-line cellulose, the CBMs of a hemicellulase donot necessarily bind the hemicellulose com-ponent (xylan). In some cases, its CBM is infact an authentic CBD that situates the hemi-cellulase on the insoluble plant cell wall mate-rial by utilizing the most abundant and moststable cell-wall component—cellulose. Indeed,the three family-3 CBDs (CBM3) shown in Fig.11 apparently bind to crystalline cellulose. Whywould this xylanase require three tandem cop-ies of the same type of CBD is yet anothermystery that should eventually be addressedexperimentally. At any rate, once bound via the

Fig. 10. A very large, cell-surfaceenzyme from Thermoanaerobacterthermosulfurogenes. The 1861-residueenzyme contains an SLH module,which is believed to mediate theattachment of the enzyme to the cellsurface in Gram-positive bacteria.The enzyme contains a multiplicityof modules, which apparently serveto regulate the hydrolytic action ofits single family-13 catalytic modulewith the complex substrate. SeveralX domains of unknown function mayeither represent as yet undescribedcatalytic functions, carbohydrate-binding activities or structural entities.

Cell-bound amylase-pullulanase from

Thermoanaerobacter thermosulfurogenes

X25-X25–X25–GH13–FN3-X31-FN3–CBM20–X32–X–SLH

X25 X25 X21X32 X SLH20fn

3

X31 fn313


cellulose component of the plant cell wall com-posite substrate, the immobilized enzyme thenacts on the accessible and appropriate hemicel-lulose components. Once thus situated on theplant cell wall, another type of CBM on thesame molecule would then assist in the bindingto the xylan (or mannan, etc.) component todirect the appropriate catalytic module to itstrue substrate. Hence, the modular proximity ofthe xylanase shown in Fig. 11 would presumablyindicate that the two CBM22s would modulatethe action of the family-10 catalytic module,and the C-terminal CBM6 would facilitate thecatalysis by the family-43 module. Together, thetwo catalytic modules would act synergisticallyto degrade susceptible plant cell wall compo-nents. In this context, the complex architectureof a xylanase would reflect the complex chemis-try of its substrate and the neighboring poly-mers of its immediate environment in the plantcell wall.

Cellulosomes

Cellulosomes are multienzyme complexes, whichbind to and catalyze the efficient degradation ofcellulosic substrates. The first cellulosome wasdiscovered while studying the anaerobic ther-mophilic bacterium, Clostridium thermocellum(Bayer et al., 1983; Lamed et al., 1983). Since itsinitial description in the literature, the cellulo-some concept has been subject to numerousreviews (Bayer et al., 1996; Béguin and Lemaire,1996; Belaich et al., 1997; Doi et al., 1994; Doiand Tamura 2001; Felix and Ljungdahl, 1993;Karita et al., 1997; Lamed and Bayer, 1988;Lamed and Bayer, 1991; Lamed and Bayer,1993; Lamed et al., 1983; Shoham et al., 1999).

Cellulosomes in C. thermocellum exist in bothcell-associated and extracellular forms, the cell-

associated form being associated with polycellu-losomal protuberance-like organelles on the cellsurface. Later, cellulosomes were detected inother cellulolytic organisms (Lamed et al., 1987;Mayer et al., 1987), including Acetivibrio cellu-lolyticus, Bacteroides cellulosolvens, Clostridiumcellulovorans and Ruminococcus albus, all ofwhich contained protuberance-like organelles ontheir surfaces (Bayer et al., 1994; Lamed andBayer, 1988; Fig. 12).

The cellulosomes contain numerous compo-nents, many of which were shown to displayenzymatic activity. They also contain a char-acteristic nonenzymatic high-molecular-weightcomponent. This component proved to behighly antigenic and glycosylated (Bayer et al.,1985). The cellulosomal enzymatic subunitsfrom this organism showed a broad range ofdifferent cellulolytic and xylanolytic activities(Morag et al., 1990). Ultrastructural evidenceindicated the multisubunit nature of the cellulo-some (Fig. 13).

Eventually, genetic engineering techniquesled to the sequencing of cellulosomal genes inC. thermocellum and several other bacteria,thus confirming the existence of cellulosomes asa major paradigm of prokaryotic degradationof cellulose and related plant cell wallpolysaccharides.

Clostridium Thermocellum Cellulosomal Subunits and Their Modules

A simplified schematic view of the cellulosomefrom C. thermocellum and its interaction with itssubstrate is shown in Fig. 14. The cellulosomalenzyme subunits were found to be united into acomplex by means of a unique class of non-enzymatic, multimodular polypeptide subunit,termed “scaffoldin” (Bayer et al., 1994). The

Fig. 11. A very large, multimodularxylanase from Caldicellulosiruptor.The 1,795-residue enzyme contains 8separate modules, including 2 catalyticmodules from families 10 (invariablya xylanase) and 43 (frequently anarabinofuranosidase). These are mod-ulated by numerous carbohydrate-binding modules, which include 3 fromfamily 3 (likely for binding to crystal-line cellulose), 2 from family 22(newly classified and shown to func-tion in xylan binding and one fromfamily 6.

Multi–modular Xylanase from Caldicellulosiruptor

22 22 10

CBM22-CBM22-GH10-CBM3-CBM3-CBM3-GH43-CBM6

433 3 3 6


Fig. 12. Scanning electron microscopy (SEM) of Acetivibrio cellulolyticus showing the presence of large characteristicprotuberance-like structures on the cell surface. Cells are shown in the free state (A) or bound to cellulose (B). Cellpreparations were treated with cationized ferritin before processing. Cationized ferritin has been shown to stabilize suchsurface structures, thus allowing their ultrastructural visualization (Lamed et al., 1987a; Lamed et al., 1987b). Withoutpretreatment with cationized ferritin, these structures are invisible. In (B), the cellulose-bound cells appear to be connectedto the substrate via structural extensions of the cell-surface protuberances. Such a mechanism was originally observed forother cellulolytic prokaryotes, e.g., C. thermocellum (Bayer and Lamed, 1986).

A B

Fig. 13. Comparison between negative staining (bottom) and cryo images (top) of the purified cellulosome from C. thermo-cellum, adsorbed on cellulose microcrystals from the algae, Valonia ventricosa. The images illustrate the diversity of shapesof the cellulosomes, which adopt either compact or loosely organized ultrastructure. In the cryo images, the subunits of thecellulosomes (i.e., the individual enzymatic components) are clearly visible. Micrographs courtesy of Claire Boisset and HenriChanzy (CNRS—CERMAV, Grenoble, France).

50 nm


scaffoldins usually contain a family-3 CBD thatprovides the cellulose-binding function. The scaf-foldins also contain multiple copies of a defini-tive type of module, called “the cohesin domain.”The cellulosomal enzyme subunits, on the otherhand, contain a complementary type of module,called “the dockerin domain.” The interactionbetween the cohesin and dockerin domains pro-vides the definitive molecular mechanism thatintegrates the enzyme subunits into the cellulo-some complex (Salamitou et al., 1994; Tokatlidiset al., 1991; Tokatlidis et al., 1993). Cohesin anddockerins are considered to be cellulosome “sig-nature sequences”—i.e., their presence is a goodindication of a cellulosome in a given bacterium(Bayer et al., 1998).

The major difference between free enzymesand cellulosomal enzymes is that the freeenzymes usually contain a CBD for guiding thecatalytic domain to the substrate, whereas thecellulosomal enzymes carry a dockerin domainthat incorporates the enzyme into the cellulo-some complex. Otherwise, both the free andcellulosomal enzymes contain very similar types

of catalytic domains. The cellulosomal enzymesrely on the Family-3a CBD of the scaffoldin sub-unit for collective binding to crystalline cellulose.

The incorporation of the multiplicity ofenzyme subunits into the cellulosome complex isa function of the repeated copies of the cohesinmodule borne by the scaffoldin subunit. For mostspecies of scaffoldin, the cohesins have been clas-sified as type-I on the basis of sequence homol-ogy. The cohesin module is composed of about150 amino acid residues. The basic structure ofthe cohesin is known and comprises a nine-stranded b sandwich with a jelly-roll topology(Shimon et al., 1997; Spinelli et al., 2000; Tavareset al., 1997).

The dockerin domain contains about 70 aminoacids and is distinguished by a 22-residue dupli-cated sequence (Chauvaux et al., 1990), whichbears similarity to the well-characterized EF-hand motif of various calcium-binding proteins(e.g., calmodulin and troponin C). Within thisrepeated sequence is a 12-residue calcium-binding loop, indicating that calcium-binding isan important characteristic of the dockerin

Fig. 14. Simplified schematic view of the molecular disposition of the cellulosome and one of the associated anchoringproteins on the cell surface of C. thermocellum. The key defines the symbols used for the modules, from which the differentcellulosomal proteins are fabricated. The progression of cell to anchoring protein to cellulosome to cellulose substrate isillustrated. The SLH module links the parent anchoring protein to the cell. The cellulosomal scaffoldin subunit performsthree separate functions, each mediated by its resident functional domains: 1) its multiple type-I cohesins integrate thecellulosomal enzymes into the complex via their resident type-I dockerins, 2) its family-IIIa CBD binds to the cellulosesurface, and 3) its type-II dockerin interacts with the type-II cohesin of the exocellular anchoring protein.

Cellulosome

Key

CBD

Type-I cohesin domain

Type-II cohesin domain

SLH module

Cellulose chain

Type-II dockerin domains

Type-I dockerin domains

Catalytic domain

Cellulose

Cell

Anchoring protein

Enzymatic subunits

Scaffoldin subunit

The cellulosome of C.thermocellum


domain. This assumption was eventually con-firmed experimentally (Yaron et al., 1995). Thespecificity characteristics of the cohesin-dockerininteraction also have been investigated. Theresults showed that four suspected residues mayserve as recognition codes for interaction withthe cohesin domain (Mechaly et al., 2000;Mechaly et al., 2001; Pagès et al., 1997). Thethree-dimensional solution structure of the 69-residue dockerin domain of a Clostridium ther-mocellum cellulosomal cellulase subunit wasrecently determined (Lytle et al., 2001). As pre-dicted earlier (Bayer et al., 1998; Lytle et al.,2000; Pagès et al., 1997), the structure consists oftwo Ca2+-binding loop-helix motifs connected bya linker; the E helices entering each loop of theclassical EF-hand motif are absent from thedockerin domain.

The scaffoldin of C. thermocellum also con-tains a special type of dockerin domain. Thisdockerin failed to bind to the cohesins from thesame scaffoldin subunit, but instead interactedwith a different type of cohesin—termed “type-II cohesins”—identified on the basis of sequencehomology (Salamitou et al., 1994). These cohes-ins are somewhat different than those of type I,having an additional segment and diversity in thelatter half of the sequence. The type-II cohesinswere discovered as component parts of a groupof noncatalytic cell-surface “anchoring” proteinson C. thermocellum (Leibovitz and Béguin, 1996;Leibovitz et al., 1997; Lemaire et al., 1995;Salamitou et al., 1994). The three known anchor-ing proteins in C. thermocellum contain differentcopy-numbers of the type-II cohesins as illus-trated in Fig. 15. Each of these anchoring pro-teins also contains an S-layer homology (SLH)module, analogous to those of the cell-boundenzymes mentioned above. The interveningsequences, however, between the cohesins andSLH domains are different. In any case, the type-II cohesins selectively bind the type-II dockerins,and the cellulosome (i.e., the scaffoldin subunittogether with all of its enzyme subunits) isthereby incorporated into the cell surface of C.thermocellum.

Similarity and Diversity of Scaffoldins from Different Species

The modular architecture of the known scaffol-dins and their comparison to that of Clostridiumthermocellum is presented in Fig. 16. Two newscaffoldins have recently been described for Ace-tivibrio cellulolyticus and Bacteroides cellulosol-vens that, like C. thermocellum, carry dockerindomains at their C terminus (Ding et al., 1999;Ding et al., 2000). The A. cellulolyticus genomealso includes a gene (immediately downstreamof the scaffoldin gene) that contains type-II

cohesins that may represent an anchoring pro-tein. It thus seems that the arrangement of thecellulosome on the cell surface of these latterstrains may be analogous to that of C. thermocel-lum. It is interesting to note that the cohesinsof the Bacteroides cellulosolvens scaffoldin areclearly type-II cohesins and not of type I. Thisinfers that there is not a clear linkage betweenthe type-II cohesins and anchoring proteins.

The scaffoldins from the other clostridial spe-cies thus far described all lack “type-II dockerin”domains, the inference being that cells of C. cel-lulovorans, for example, would apparently notbear anchoring proteins that contain type-IIcohesins. It thus follows that either their cellulo-somes are not surface bound or, if indeed theyare surface components, then their anchoringthereto is accomplished via an alternative molec-ular mechanism. Recently (Doi and Tamura,2001; Tamaru and Doi, 1999a; Tamaru et al.,1999b), a cell-surface binding function has beenproposed for a domain of unknown function,designated “X2” (Coutinho and Henrissat,1999b; Coutinho and Henrissat, 1999c) of thescaffoldin from C. cellulovorans. On the basis ofsequence alignment of a few conserved identicalamino acids with S-layer proteins from Myco-plasma hyorhinis and Plasmodium reichenowi,the authors consider that this domain may berecognized as an SLH domain. The four X2domains of the C. cellulovorans scaffoldin arevery similar in sequence to the X-domains fromthe scaffoldins of C. cellulolyticum and C. josui,

Fig. 15. Schematic representation of the known anchoringproteins of the C. thermocellum cell surface. Each proteinbears an SLH domain that connects the protein to the cellsurface via yet undefined surface components. The differentproteins carry different numbers of type-II cohesins. SdbAhas one cohesin, Orf2p has 2 and OlpB has 4, presumablyallowing the corresponding number of scaffoldins (i.e.,cellulosomes) to be attached to the given protein.

Outer-layer proteins of C. thermocelluminvolved in anchoring the cellulosome

onto the cell surface

SdbA Orl2p OlpB

KeyType II cohesin domain

SLH module


which contain only two and one copies of thisdomain, respectively. If this domain functions inattaching the scaffoldin with its complement ofenzymes to the cell surface, it is unclear whythere would be different copy numbers of thedomain in the different scaffoldins. Likewise,one of the C. cellulovorans cellulosomal enzymecomponents (EngE) also contains a triplicatedsegment of unknown function, designated “X48”(Coutinho and Henrissat, 1999b; Coutinho andHenrissat, 1999c) that the authors consider to beinvolved in cell-surface attachment (Tamaru andDoi, 1999a). In any case, final proof of the func-tion of the X2 and X48 domains awaits biochem-ical examination, as has been clearly achieved forthe SLH domain of the C. thermocellum anchor-ing proteins (Chauvaux et al., 1999; Lemaire etal., 1998).

Finally, two new scaffoldins have recently beensequenced from the rumen bacterium, Rumino-coccus flavefaciens (Ding et al., 2001). Althougheach of the two proteins contains multiple cohes-ins, their sequences indicate that they are neither

of type-I or type-II, but occupy their own phylo-genetic branch. Interestingly, the ruminococcalscaffoldins lack a known type of CBD. Both havedissimilar X domains of unknown function, thesequences of which bear no resemblance to anyother known module. Both X domains wereexpressed, but the resultant proteins failed tobind to cellulose. The lack of a scaffoldin CBDraises the question as to how the ruminococcalcellulosome(s) and/or the bacterium bind to thesubstrate. Perhaps it does so like another closelyrelated species, R. albus, which binds cellulosevia a noncellulosomal cell-surface protein(Pegden et al., 1998).

Schematic Comparison of Prokaryotic Cellulase Systems

In this section, we will describe schematically thesimilarity and diversity of representative enzymesystems, demonstrating different strategies, from

Fig. 16. Schematic view of the modular similarity and diversity of scaffoldins from different cellulosome species. Class-Iscaffoldins feature an internal CBD and a C-terminal type-II dockerin domain. Class-II scaffoldins exhibit an N-terminalCBD and lack a dockerin domain. The newly described scaffoldins from Ruminococcus flavefaciens lack a defined CBD. Thefunctional role of the two different X domains in the two R. flavefaciens scaffoldins is currently unknown. All of thesescaffoldins contain multiple copies of cohesin domains.

Classification of scaffoldins

from different species

Class I Scaffoldins

Clastridum thermacellum

Acetiwbrio cellulolyticus

Bocteroides cellulosalvens

Clastridum cellulovorans

Ruminocacus flavefaciers

Clastridum cellulolytum

Clastridum josul

Class I Scaffoldins Ruminococcal Scaffoldins


different plant cell wall degrading bacteria. It isemphasized that the accumulating information isbased on what is known currently from biochem-ical data combined with gene sequencing andbioinformatics. The information is still rathersketchy but quite revealing when comparedamong different bacteria. As time progresses andthe entire genomes of cellulolytic microorgan-isms become known, the data concerning thecomplement of enzymes produced by a givenbacterium will be complete, and we will be ableto speculate with heightened certainty how thevarious cellulase systems might have evolved.A survey of genes, however, does not informus how a given bacterial system is regulatedand what role(s) the bacterium and its enzymesystem may play in nature. The explosivedevelopment of molecular biology techniques,however revealing, cannot supplant the funda-mental contribution of biochemical and ecolo-gical approaches to the study of microbialdegradation of cellulose and other plant cell wallpolysaccharides.

Free Enzyme Systems

Many cellulolytic microorganisms show a verysimilar pattern in the types of enzymes that com-prise the complement of their cellulase system.For the purposes of this discussion, the conceptof “cellulase system” will include the comple-ment of all plant cell wall hydolyzing enzymesand other glycosyl hydrolases, including the dif-ferent cellulases per se, the hemicellulases (e.g.,xylanases and mannanases), etc.

The cellulase system of the mesophilic cellu-lolytic aerobe, Cellulomonas fimi, is one of thefirst studied, and has since been one of the moststudied bacterial cellulase systems (O’Neill et al.,1986; Shen et al., 1995; Whittle et al., 1982). Theenzymes of this bacterium are essentially freeenzymes, which allowed their early isolation andcharacterization. Moreover, the genes of the cel-lulases from this bacterium were of the earliestto have been sequenced. To date, about 10 glyc-osyl hydrolases have been sequenced from Cel-lulomonas fimi. Their modular composition andfamily associations are shown symbolically inFig. 17. As an example of a free enzyme system,most of the enzymes bear a substrate-targetingCBM—in this bacterium, most of the CBMs arefrom family 2. Several of the enzymes have mul-tiple copies of the fibronectin 3 (FN3) domain,the function of which is still unknown.

The Cellulomonas system includes two family-6 enzymes—an endoglucanase and an exogluca-nase (cellobiohydrolase) of the types describedin Fig. 4. The modularity of the endoglucanase is

very simple, having the family-6 catalytic moduletogether with a family-2 CBM. The cellobiohy-drolase is a bit more complex with three addi-tional FN3 domains that separate the same twotypes of modules. Another cellobiohydrolase(that exhibits processive cleavage of the sub-strate) is from family 48. Its general modulararchitecture is similar to that of the family-6cellobiohydrolase with the substitution of thecatalytic module from a different family. The cel-lulase system from this organism also includestwo family-9 cellulases with modular themes Band D, familiar to us from the earlier description(Fig. 7). In addition, a simple family-5 cellulaseand an interesting cell-borne family-26 mannan-ase are components of the system. The fact thatan enzyme bears an SLH domain and is presum-ably cell-associated would underscore its impor-tance to the cell. Finally, three xylanases arecurrently known for Cellulomonas fimi. One ofthese xylanases is a simple enzyme consisting ofa family-10 catalytic domain connected to afamily-2 CBM. The other two are more compli-cated, each containing two catalytic domains—either a family-10 or -11 domain and a carbohy-drate esterase (in both cases, probably an acetylxylan esterase; Fig. 3)—plus several CBMs. Thisrather complex system is probably not nearlycomplete, and more enzymes will inevitably bedescribed in the future.

A second example of a free enzyme system,from the aerobic thermophilic bacteriumThermobifida fusca (formerly classified as Ther-momonospora fusca), has also been studiedextensively (Wilson, 1992; Wilson and Irwin,1999). A brief comparison of its known enzymecomponents (Fig. 18) shows a striking resem-blance to those of Cellulomonas (compare Figs.17 and 18). According to known data, both spe-cies produce similar types of cellulases from fam-ilies 5, 6, 9 and 48 plus xylanases from families10 and 11. Nevertheless, the modular repertoireof the corresponding enzyme in T. fusca is gen-erally somewhat simpler. For example, two of theT. fusca cellulases include single FN3 domains,whereas several Cellulomonas cellulases harbormultiple copies of the same domain. SomeT. fusca enzymes lack accessory modules otherthan a cellulose-binding CBM, whereas the cor-responding Cellulomonas enzyme is elaboratedby multiple copies of accessory modules. In somecases though, the respective CBMs appear onopposite termini of the polypeptide chain (i.e.,the family-48 and family-5 cellulases).

The complement of enzymes and their modu-lar content of the free enzyme systems from Cel-lulomonas and T. fusca are not necessarilysimilar in other free enzyme systems. Many freeenzyme systems, such as those of Butyrivibriofibrisolvens, Pseudomonas fluorescens, Fibro-


bacter succinogenes, various species of Strepto-myces, Erwinia and Thermatoga, appear to haveseveral cellulases, xylanases and mannanasesfrom the common families, together with otherglycosyl hydrolases, e.g., arabinosidases, lichena-ses, amylases, pullulanases, galactanases, polyga-lacturonase, glucuronidases and pectate lyases.In many of these bacterial enzymes, the family-2CBM appears to predominate as a common

cellulose-binding domain, but in others (e.g.,Erwinia) relevant enzymes usually bear acellulose-binding CBM from family-3. Neverthe-less, in many of the free systems, many enzymesare characterized by CBMs from other familiesas well as other noncatalytic domains ofunknown function (X domains). Once again,until the genome sequences of cellulolyticprokaryotes are widely available, we are still lim-

Fig. 18. Thermobifida fusca cellulasesystem. A cell-free enzyme system.The modular content of the enzymesis shown from (left to right) the N-terminus to the C-terminus of thepolypeptide chain. Compare with theCellulomonas system (Fig. 17). Key tosymbols: GH, glycosyl hydrolase (e.g.,cellulase, xylanase and mannanase);CBM, carbohydrate-binding module(e.g., CBD, cellulose-binding domain);FN3, fibronectin-3 (domain); Ig,immunoglobulin-like domain; and X,domain of unknown function.

Known Enzymes from Thermobifida fusca

CBM2-GH6

CBM2-GH5

GH9-CBM3c-FN3-CBM2 CBM2-X-X-GH48

GH6-CBM2

GH10-CBM2

GH11-CBM2

6

48

10

5

9

11

62 2

2

22

24

3c

2 2

X X

CBM4-Ig-GH9-FN3-CBM2

Cel6A(E2)

Cel48A(E6)

Cel6B(E3)

Cel9A(E4)

Cel9B(E1)

Cel5A(E5)

Xyn10A

Xyn11A

fn3

fn3

Ig

Fig. 17. Cellulomonas fimi cellulasesystem: Symbolic view of the enzymecomponents and their modular archi-tecture. An example of a cell-freeenzyme system. The modular contentof the enzymes in this and subsequentfigures is shown from (left to right) theN-terminus to the C-terminus of thepolypeptide chain. The family num-bers of the given domains are enumer-ated, the catalytic modules given inred. Key to symbols: GH, glycosylhydrolase (e.g., cellulase, xylanaseand mannanase); CE, carbohydrateesterase (e.g., acetyl xylan esteraseand ferulic acid esterase); CBM,carbohydrate-binding module (e.g.,CBD, cellulose-binding domain);SLH, S-layer homology (domain);FN3, fibronectin-3 (domain); Ig,immunoglobulin-like domain; and X,domain of unknown function.

Known Enzymes from Cellulomonas fimi

2

3c fn3

fn3

fn3

I9

fn3

fn3

fn3

fn3

fn3

fn3

fn3

fn32

2

2

22

2 2

9 9 X

4 4

23 X

2

Cel6A

Cel9A

Cel9B

Cel5A

ManA

Cel6B

Cel48A

Xyn10A(Cex)

Xyn10B(Xync)

Xyn11A

GH11-CBM2-CE4-CBM2

GH10-CBM2

GH48-FN3-FN3-FN3-CBM2

GH6-FN3-FN3-FN3-CBM2CBM2-GH6

6 6

48

10

10

11

Ce4

Ce4

9

9

5

26

GH9-CBM3c-FN3-FN3-FN3-CBM2

CBM4-CBM4-Ig-GH9-X-X

GH5-FN3-FN3-CBM2

GH26-CBM23-SLH-X

CE4-CBM22-GH10-CBM9-CBM9-X

X X

SLH


ited in our capacity to compare among theenzyme systems because our knowledge of theirenzyme sequences is incomplete.

Multifunctional Enzyme Systems

In an extremely thermophilic bacteria, classifiedas Caldicellulosiruptor, the enzymes currentlycharacterized in this system also appear to befree enzymes, but their modular organization isof a higher order (Daniel et al., 1996; Gibbset al., 2000; Reeves et al., 2000). Many of theenzymes of this system are bifunctional in thatthey contain two separate catalytic modules inthe same polypeptide chain (Fig. 19). As men-tioned earlier, the appearance of two catalyticmodules in the same enzyme would infer a dis-tinctive synergistic action between the two. Thus,in CelA, the family-9 and -48 catalytic moduleswould be expected to work in concerted fashionon crystalline cellulose. In another type ofenzyme, the family-10 xylanase and family-5 cel-lulase would likely be most effective on regionsof the plant cell wall that are characterized bycellulose-xylan junctions. The diversity in themodular architecture of the family-10 xylanasesis particularly striking, and the various com-binations of this type of catalytic module areapparently important to the sustenance of thebacterium in its environment. One of these xyla-nases appears to be attached to the cell surfacevia SLH domains. In contrast to the Cellulomo-nas and T. fusca enzymes that often harbor afamily-2 CBM, the module responsible for bind-ing to cellulosic substrates in Caldicellulosiruptor

enzymes is usually one or more copies of afamily-3 CBM.

Other bacterial strains that include at least onefree bifunctional enzyme in their enzyme sys-tems are Anaerocellum thermophilum, Bacillusstearothermophilus, Fibrobacter succinogenes,Prevotella ruminicola, Ruminococcus albus,Ruminococcus flavefaciens, Streptomyces chatta-noogensis and the thermophilic anaerobe NA10.Unlike the Caldicellulosiruptor system, most ofthe free bifunctional enzymes in the latter strainsappear to be isolated cases in the given system,rather than being a common character of theirenzymes.

Cellulosomal Systems

The inclusion of enzymes into a cellulosome viathe noncatalytic scaffoldin subunit represents ahigher level of organization. The association ofcomplementary enzymes into a complex is con-sidered to contribute sterically to their synergis-tic action on cellulose and other plant cell wallpolysaccharides. As mentioned earlier, in thecase of Clostridium thermocellum, Acetivibriocellulolyticus and Bacteroides cellulosolvens, thecellulosomes appear to be attached to the cellsurface. The cellulosomes of C. cellulolyticum, C.cellulovorans and C. josui may also be cell-associated, but if so, the lack of a scaffoldin-borne dockerin and reciprocal anchoring proteinwould suggest an alternative mechanism.

The cellulosomes of C. cellulolyticum, C. cellu-lovorans and C. josui are very similar. The genesencoding for many or most of the enzymes in all

Fig. 19. Caldicellulosiruptor enzymesystem. An example of a cell-freeenzyme system that includes severalmultifunctional enzymes. The modularcontent of the enzymes is shown from(left to right) the N-terminus to theC-terminus of the polypeptide chain.Key to symbols: GH, glycosyl hydro-lase (e.g., cellulase, xylanase and man-nanase); CBM, carbohydrate-bindingmodule (e.g., CBD, cellulose-bindingdomain); and SLH, S-layer homology(domain).

Known Enzymes from Caldicellulosiruptor

GH9-CBM3c-CBM3-CBM3-GH48

CBM22-CBM22-GH10

CBM22-CBM22-GH10-CBM9-CBM9-X-SLH

GH11-CBM9GH43-X-GH43

GH10-CBM3-GH5

GH5-CBM3-CBM3-GH44

GH10-CBM3-CBM3-CBM3-GH5 CBM22-CBM22-GH10-CBM3-CBM3-CBM3-GH43-CBM6

GH10

XynA

XynB

XynF (Abn)

CelB

Cel

CelA

Man-EG’ase

XynC

XynD

XynE & XynI

9

5

5

5

10

10

10

10

1043

11

10

48

44

43 43

3c 3 3

3

3

22 22

22 22

22 22 3 3 3 6

9 9 X SLH

3 3

X

3

9

3


three cellulosomal systems are arranged in alarge cluster on the chromosome. Some of thecellulosomal genes, however, are located outsideof the cluster in other regions of the chromo-some. The majority of the cellulosome gene clus-ters from C. cellulolyticum and C. cellulovoranshave been sequenced (Bagnara-Tardif et al.,1992; Belaich et al., 1999; Tamaru et al., 2000b).In contrast, the cellulosomal genes from C. ther-mocellum are generally scattered over a largeportion of the chromosome (Guglielmi andBéguin, 1998). A few small clusters of celluloso-mal genes are apparent in the genome, includinga scaffoldin-containing cluster that also containsseveral cell-surface anchoring proteins (Fujinoet al., 1993). The following descriptive analysisserves to compare the cellulosomal system ofthese three microorganisms.

Cellulosomal components from Clostridiumcellulolyticum. All of the sequenced enzymesfrom this organism are relatively common cellu-lases (Belaich et al., 1999). None of the knowncellulosomal enzymes yet described for this spe-cies contains more than one catalytic module(Fig. 20). The largest one, CelE (estimated at94 kDa), is a theme-D family-9 cellulase(Gaudin et al., 2000). The critical family-48 cel-lulase (CelF) is also a major cellulosome compo-nent (Reverbel-Leroy et al., 1997). Interestingly,the gene cluster of C. cellulolyticum containsthree copies of other family-9 cellulases (CelG,CelH and CelJ), all of which contain the theme-B fused family-3c CBM (Belaich et al., 1998; Fig.8). The currently known cellulosome system inthis bacterium also contains two family-5 cellu-lases (CelA and CelD), a family-5 mannanase(ManK, which bears an N-terminal rather thanC-terminal dockerin) and a family-8 cellulase(CelC).

Biochemical characterization of the C.cellulolyticum cellulosome demonstrated bysodium dodecyl sulfate-polyacrylamide gelelectrophoresis (SDS-PAGE) a 160-kDa scaf-foldin band and up to 16 smaller bands, repre-senting putative enzyme subunits (Gal et al.,1997). Many of these were clearly identified asknown gene products. Only two cellulosomalcellulase genes are currently known to belocated outside of the gene cluster. Furtherwork on the enzyme system of this species mayyet provide more complicated multimodularenzymes and/or other types of enzymes, such ashemicellulases. In this context, recent bioche-mical evidence has suggested that xylanasesfrom C. cellulolyticum are also organized in acellulosome-like complex, but defined xylanasesequences are still lacking from this organism(Mohand-Oussaid et al., 1999). The knownactivity of this organism on other plant cell wall

polysaccharides would indicate that numerousother enzymes, either cellulosomal or not,remain as yet undiscovered.

Cellulosomal components from Clostridiumcellulovorans. Like C. cellulolyticum, the cellu-lases from this organism are relatively simple(Fig. 21). In addition to the cellulosomalenzymes thus described, at least three non-cellulosomal endoglucanases have also beenpartially or totally sequenced (Doi et al., 1998;Tamaru et al., 1999b).

Several of the cellulosomal enzymes are archi-tecturally synonymous to those of the C. cellu-lolyticum system (compare Figs. 20 and 21). Thisincludes the critical family-48 cellulase (ExgS;Liu and Doi, 1998), two copies of the theme-Bfamily-9 cellulase (EngH and EngY), a family-5endoglucanase and a family-5 mannanase thatbears an N-terminal dockerin (Tamaru and Doi,2000a). Rather than a single theme-D family-9cellulase as in C. cellulolyticum, the C. cellulo-vorans system contains two such enzymes (EngKand EngM). The C. cellulovorans cellulosomealso appears to contain an unusual theme-A

Fig. 20. Clostridium cellulolyticum enzyme system. An exam-ple of a cellulosomal system. The modular content of theenzymes is shown from (left to right) the N-terminus to theC-terminus of the polypeptide chain. Key to symbols: GH,glycosyl hydrolase (e.g., cellulase, xylanase and mannanase);CBM, carbohydrate-binding module (e.g., CBD, cellulose-binding domain); and Doc, dockerin domain.

Known Enzymes from Clostridium cellulolyticum

Cellulosomal Enzymes

Scaffoldin

CBM4-Ig-GH9-Doc

GH5-Doc

GH5-Doc

Doc-GH5

GH8-Doc

GH48-Doc

GH9-CBM3c-Doc

GH9-CBM3c-Doc

GH9-CBM3c-Doc

CBM3a-X-7Coh-X-Coh

CelE

CelF

CelG

CelD

CelA

CelC

Mank

CelH

CelJ

4 Ig

3c

3c

3a

3c

9

5

5

5

8

9

9

9

48


family-9 cellulase (EngL) that lacks helperdomains. The remaining two known cellulosomalenzymes are thus far unique to C. cellulovorans.A dockerin-bearing pectate lyase (LyaA)infers that the bacterium would degrade pectin(Tamaru and Doi, 2001). Indeed, early evidence(Sleat et al., 1984) indicated that, in addition tocellulose, C. cellulovorans is capable of assimilat-ing a wide variety of other plant cell wall polysac-charides, including, xylans, pectins and mannans.As in the case of C. celluolyticum, it seems thatfuture work will yield new sequences of manyother types of cellulosomal and noncellulosomalenzymes.

More significant to the cellulosomal system ofC. cellulovorans, perhaps, is the large family-5enzyme that purportedly comprises both an N-terminal SLH domain and a C-terminal dockerin(Tamaru and Doi, 1999a). This arrangement mayimply that the entire cellulosome is bound to thecell surface via this enzyme. If this proves to bethe case, it is interesting to speculate whether theC. cellulolyticum and C. josui cellulosomes arealso connected to the cell surface by a similar,but as yet undiscovered enzyme that bears bothSLH and dockerin domains.

Cellulosomal components from Clostridiumthermocellum. Compared to the cellulosomal

Fig. 21. Clostridium cellulovorans: A second cellulosomal system. The modular content of the enzymes is shown from (leftto right) the N-terminus to the C-terminus of the polypeptide chain. Key to symbols: GH, glycosyl hydrolase (e.g., cellulase,xylanase and mannanase); CBM, carbohydrate-binding module (e.g., CBD, cellulose-binding domain); Doc, dockerin domain;SLH, S-layer homology (domain); Ig, immunoglobulin-like domain; and X, domain of unknown function.

EngE

EngK

EngM

EXgS

EngH

EngY

EngA

EngD

Cel5A(EngF)

GH5–CBM17

Lya4–Doc

GH48–Doc

CBM4–Ig–GH9–Doc

CBM4–Ig–GH9–Doc

3x48–GH5–X–Doc

GH9–CBM3c–Doc

CBM3a–X–2Coh–X–6Coh–2X–Coh

Doc–GH5

GH5–Doc

GH9–Doc

X17–GH9–CBM3c–Doc

GH5–CBM2

GH9–(?)EngL

EngB

ManA

LyaA

Known Enzymes from Clostridium cellulovorans

Cellulosomal EnzymesNon-Cellulosomal

Enzymes

5

5

5

5

5

9

9

99

94

4

3c9

48

2

3cx7

17

xxx

x

3a

Scaffoldin


systems of C. cellulovorans and C. cellulolyticum,the enzymes from C. thermocellum are relativelylarge proteins, ranging in molecular size fromabout 40–180 kDa (Bayer et al., 1998; Bayeret al., 2000; Béguin and Lemaire, 1996; Felixand Ljungdahl, 1993; Lamed and Bayer, 1988;Shoham et al., 1999). Examination of Fig. 22reveals why these enzymes are so big—many ofthe larger ones contain multiple types of catalyticdomains as well as other functional modules asan integral part of a single polypeptide chain(see Table I in Bayer et al., 1998, for a list ofrelevant references). In addition to thecellulosomal enzymes, several noncellulosomalenzymes have also been described from thisorganism (Morag et al., 1990). These include twofree enzymes (one of which lacks a CBM) andtwo cell-associated (SLH-containing) enzymes.Consequently, the potent cellulose- and plantcell wall-degrading activities of C. thermocellumare clearly reflected in its cellulase system, whichdisplays an exceptional wealth, diversity andintricacy of enzymatic components, thus repre-senting the premier cellulose-degrading organ-ism currently known.

Many of the C. thermocellum cellulosomalenzymes are cellulases, which include bothendo- and exo-acting b-glucanases. Some of theimportant exoglucanases and processive cellu-lases include CelS, CbhA, CelK and CelF. TheCelS subunit is a member of the family-48 glyc-osyl hydrolases, and this particular family is nowrecognized as a critical component of bacterialcellulosomes (Morag et al., 1991; Morag et al.,1993; Wang et al., 1993; Wang et al., 1994; Wuet al., 1988). Several other processive cellulasesare members of the family-9 glycosyl hydrolases.CelF and CelN are theme-B family-9 enzymes(Navarro et al., 1991; Fig. 7). The other two areremarkably similar theme-D enzymes, whichexhibit nearly 95% similarity along their com-mon regions (Kataeva et al., 1999a; Kataeva etal., 1999b; Zverlov et al., 1998; Zverlov et al.,1999). The main difference between CbhA andCelK is the presence in the former of three extramodules (a family-3 CBD and two modules ofunknown function). The functional significanceof these supplementary modules to the activityof CbhA has not been elucidated.

The fact that the cellulosome from this organ-ism contains many different types of cellulases is,of course, to be expected if we consider thatgrowth of C. thermocellum is restricted to cellu-lose and its breakdown products, particularly cel-lobiose. Consequently, it is surprising to discover,in addition to the cellulases, at least five classicxylanases, i.e., those belonging to glycosyl hydro-lase families 10 and 11. In addition, two of thelarger enzymes, CelH and CelJ, contain hemicel-lulase components, i.e., family-26 and -44 cata-

lytic modules (a mannanase and a xylanase,respectively), together with a standard cellulasemodule in the same polypeptide chain (Ahsan etal., 1996; Yagüe et al., 1990). It is also interestingto note the presence of carbohydrate esterasestogether with xylanase or cellulase modules insome of the enzyme subunits (i.e., XynU/A,XynY, XynZ and CelE), thus conferring thecapacity to hydrolyze acetyl or feruloyl groupsfrom hemicellulose substrates (Blum et al., 2000;Fernandes et al., 1999). Finally, the C. thermocel-lum cellulosome includes a typical family-16lichenase, a family-26 mannanase and a family-18 chitinase.

The non-cellulosomal enzymes includeanother theme-B family-9 cellulase (CelI), andcell-bound forms of a xylanase (XynX) and alichenase (LicA), both of which contain multipleCBMs adjacent to the catalytic module. In themidst of all this complexity, the C. thermocellumnon-cellulosomal cellulase system includes a sim-ple family-5 cellulase, CelC, which is completelydevoid of additional accessory modules.

Why does this bacterium—which subsistsexclusively on cellulosic substrates—need allthese hemicellulases? The inclusion of such animpressive array of non-cellulolytic enzymes ina strict cellulose-utilizing species would suggestthat their major purpose would be to collec-tively purge the unwanted polysaccharides fromthe milieu and to expose the preferred sub-strate—cellulose. The ferulic acid esterases, inconcert with the xylanase components of theparent enzymes, could grant the bacterium arelatively simple mechanism by which it coulddetach the lignin component from the cellulose-hemicellulose composite. The lichenase (LicB)and chitinase (ChiA) are also intriguing compo-nents of the cellulosome. The former wouldprovide the bacterium with added action oncell-wall b-glucan components from certaintypes of plant matter. It is not clear whether thepresence of the latter cellulosomal enzymewould reflect chitin-derived substrates fromthe exoskeletons of insects and/or from fungalcell walls. Whatever the source, the chitinbreakdown products, like those of the hemicel-luloses, would presumably not be utilized by thebacterium itself, but would be passed on toappropriate satellite bacteria for subsequentassimilation.

Phylogenetics of Cellulase and Cellulosomal Systems

Early in the history of the development andestablishment of the cellulosome concept, it wasnoted that the apparent occurrence of cellulo-


somes in different microorganisms tended tocross ecological, physiological and evolutionaryboundaries (Lamed et al., 1987). Initial biochem-ical and immunochemical evidence to this effecthas been supported by the accumulated molecu-lar biological studies.

Various lines of evidence indicate that themodular enzymes that degrade plant cell wallpolysaccharides have evolved from a restrictednumber of common ancestral sequences. Muchof the information in this direction remains as alegacy, inherently encoded in the sequences ofthe functional domains that comprise the differ-ent enzymes. By comparing sequences of the var-ious cellulosomal and noncellulosomal enzymes

within and among the different strains, we cangain insight into the evolutionary rationale of themultigene families that comprise the glycosylhydrolases.

Horizontal Gene Transfer

It is clear that very similar enzymes whichcomprise a given glycosyl hydrolase family areprevalent among a variety of different bacteriaand fungi, thus indicating that they were notinherited through conventional evolutionaryprocesses. The widespread occurrence of suchconserved enzymes among phylogenetically dif-ferent species argues that horizontal transfer of

Fig. 22. Clostridium thermocellum: A very complex cellulosomal system. The modular content of the enzymes is shown from(left to right) the N-terminus to the C-terminus of the polypeptide chain. Key to symbols: GH, glycosyl hydrolase (e.g.,cellulase, xylanase and mannanase); CE, carbohydrate esterase (e.g., acetyl xylan esterase and ferulic acid esterase); CBM,carbohydrate-binding module (e.g., CBD, cellulose-binding domain); Doc, dockerin domain; SLH, S-layer homology(domain); Ig, immunoglobulin-like domain; and X, domain of unknown function.

Known Enzymes from Clostridium Thermocellum

Anchoring Proteins

Scaffoldin

Cellulosomal Enzymes Non-Cellulosomal

EnzymesCelJ X7

4

X7-Ig-GH9-GH44-Doc-X

CBM4-Ig-GH9-2(X1)-CBM3-Doc

CBM22-GH10-CBM22-Doc-CE1

GH26-GH5-CBM11-Doc

CBM4-Ig-GH9-Doc

Ig-GH9-Doc

GH11-CBM6-Doc-CE4

GH9-CBM3c-Doc

GH9-CBM3c-Doc

CBM3-GH5-Doc

GH48-Doc

GH5-Doc

GH5-Doc

GH18-Doc

GH8-Doc

GH16-Doc

GH11-CBM6-Doc

CEI-CBM6-Doc-GH10

GH5-Doc-CE2CBMx4-GH9-Doc

2Coh-CBM3a-7Coh-X-Doc

CBM22-GH10-Doc

SLH-X-GH16-4(CBM4)

CBM22-GH10-2(CBM9)-SLH

GH9-CBM3c-CBM3

GH5

X

X1 X1 3

11

4

6

22 6

X 4

9

3c 3

922

4 4 4

6

3c

9

9

44

10

26

9

10

26 16

CelO

CelI

CelCCelB

CelG

ChiA

XynB(XynV)

CelA

LicB

LicA

10

9

9

11

3c

3

9

5

18

816

10

9

5

11

5

5

48

5x4

3a

5

Ig

Ig

Ig

Ig

CbhA

CelH

CelK

CelE

XynY

XynZ

CelS

CelF

CelN

XynA(XynU)

XynX

CelD

ManA

XynC

22 22


genes has been a major process by which a givenmicroorganism can acquire a desirable enzyme.Once such a transfer event has taken place, thenewly acquired gene would then be subjected toenvironmental pressures of its new surroundings,i.e., the genetic and physiological constitution ofthe cell itself. Following such selective pressure,the sequence of the gene would be adjusted tofit the host cell.

Gene Duplication

Sequence comparisons have also revealed thepresence of very similar genes within a genomethat may have very similar or even identicalfunctions. One striking example is the tandemappearance of cbhA and celK genes in the chro-mosome of Clostridium thermocellum. Otherexamples are xynA and xynB also of C. thermo-cellum and xynA of the anaerobic fungus Neo-callimastix patriciarum, which includes two verysimilar copies of family-11 catalytic moduleswithin the same polypeptide chain. These exam-ples imply a mechanism of gene duplication(Chen et al., 1998; Gilbert et al., 1992), wherebythe duplicated gene can serve as a template forsecondary modifications that could result in twovery similar enzymes with different properties,such as substrate and product specificities. Asimilar process could also account for the multi-plicity of other types of modules (i.e., CBDs,cohesins or helper modules) within a polypep-tide chain. Comparison of the modular architec-tures of similar genes from different specieswould suggest that individual modules canundergo a duplication process. This is exempli-fied by the multiple copies of FN3 in CelB fromCellulomonas fimi versus the single copy of thesame domain in cellulase E4 from Thermobifidafusca. But innumerable other examples areevident from the databases, whenever multiplecopies of the same modular type exist in thesame protein.

Domain Shuffling

Another observation from the genetic composi-tion of the glycosyl hydrolases argues for analternative type of process, which would propa-gate new or modified types of enzymes. It is clearthat many microbial enzyme systems containindividual hydrolases that carry very similarcatalytic domains but include different typesof accessory modules (Gilkes et al., 1991). Anexample that demonstrates this phenomenon isthe observed species preference of otherwisevery similar glycosyl hydrolases for a given fam-ily of crystalline cellulose-binding CBD, which isentirely independent of the type of catalyticmodule borne by the complete enzyme. In this

context, as we have seen above, the free enzymesof some bacteria, such as Cellulomonas fimi,Pseudomonas fluorescens and Thermomono-spora fusca, invariably include a family-2 CBD,irrespective of the type of catalytic domain. Incontrast, those of other bacteria, e.g., Bacillussubtilis, Caldocellum saccharolyticum, Erwiniacarotovora and various clostridia, appear to pre-fer family-3 CBDs. Moreover, the position ofthe CBD in the gene may be different for dif-ferent genes. For example, the CBD may occurupstream or downstream from the catalyticdomain; it may be positioned either internally(sandwiched between two other modules) or atone of the termini of the polypeptide chain. Thesame pattern is characteristic of several otherkinds of modules associated with the plant cellwall hydrolases. This is particularly evident infamily-9 cellulases and family-10 xylanases,where the number and types of accessory mod-ules may vary greatly within a given species. Itseems that individual domains can be transferreden bloc and incorporated independently intoappropriate enzymes. Once again, the modulararchitectures and sequence similarities betweenClostridium thermocellum cellulosomal enzymepairs (CbhA and CelK; XynA and XynB) areparticularly revealing: in both cases, following anapparent gene duplication event, one or moreadditional modules appear to have been incor-porated into the duplicated enzyme. Takentogether, the information suggests that domainshuffling is an important process by which theproperties of such enzymes can be modified andextended.

Proposed Mechanisms for Acquiring Cellulase and Cellulosomal Genes

Like the free enzyme systems, the phylogeny ofcellulosomal components seems to have beendriven by processes that include horizontal genetransfer, gene duplication and domain shuffling.In cellulolytic/hemicellulolytic ecosystems, theresident microorganisms are usually in close con-tact, often under difficult conditions and in com-petition or cooperation with one another towarda common goal: the rapid degradation of recal-citrant polysaccharides and assimilation of theirbreakdown products.

A possible scenario for the molecular evolu-tion of a cellulase/hemicellulase system in aprospective bacterium could involve the initialtransfer of genetic material from one microbe toanother in the same ecosystem. The size and typeof transferred material could vary, such as a geneor part of gene (e.g., selected functional mod-ules) or even all or part of a gene cluster. Theprocess could then be sustained by gene duplica-


tion, which would propagate the insertion ofrepeated modules, e.g., the multiple cohesindomains in the scaffoldins, or even smaller units,such as the linker sequences or the duplicatedcalcium-binding loop of the dockerin domain.Domain shuffling can account for the observedpermutations in the arrangement of domains inscaffoldin subunits from different species (Fig.16). Finally, conventional mutagenesis wouldthen render such products more suitable for thecellular environment or for interaction withother components of the cellulase system.

The available data suggest that there are no setof rules, which would, at this stage, enable us toanticipate the nature of a given cellulase systemfrom a given microorganism. It seems that phy-logenetically dissimilar organisms can possesssimilar types of cellulosomal or non-cellulosomalenzyme systems, whereas phylogeneticallyrelated organisms that inhabit similar niches maybe characterized by different types of enzymesystems. It is clear that to shed further light onthis apparent enigma, we require more informa-tion about more types of enzyme systems. Inaddition to more sequences and structures, wewill need more information—biochemical, phy-siological and ecological—to sharpen existingnotions regarding the enzymatic degradation ofplant cell wall polysaccharides or to formulatenew ones.

Acknowledgments. Grants from the IsraelScience Foundation (administered by the IsraelAcademy of Sciences and Humanities, Jerusa-lem) and the Minerva Foundation (Germany)are sincerely appreciated.

Literature Cited

Ahsan, M. M., T. Kimura, S. Karita, K. Sakka, and K. Ohm-iya. 1996. Cloning, DNA sequencing, and expression ofthe gene encoding Clostridium thermocellum cellulaseCelJ, the largest catalytic component of the cellulosome.J. Bacteriol. 178:5732–5740.

Armand, S., S. Drouillard, M. Schulein, B. Henrissat, and H.Driguez. 1997. A bifunctionalized fluorogenic tetrasac-charide as a substrate to study cellulases. J. Biol. Chem.272:2709–2713.

Atalla, R. H., and D. L. VanderHart. 1984. Native cellulose:A composite of two distinct crystalline forms. Science223:283–285.

Atalla, R. H. 1999. Celluloses. In: B. M. Pinto (Ed.) Com-prehensive Natural Products Chemistry. Elsevier.Cambridge, UK. 3:529–598.

Bach, E., and E. Schollmeyer. 1992. An ultraviolet-spectrophotometric method with 2-cyanoacetamide forthe determination of the enzymatic degradation ofreducing polysaccharides. Analyt. Biochem. 203(2):335–339.

Bagnara-Tardif, C., C. Gaudin, A. Belaich, P. Hoest, T. Citard,and J.-P. Belaich. 1992. Sequence analysis of a gene clus-ter encoding cellulases from Clostridium cellulolyticum.Gene 119:17–28.

Barr, B. K., Y.-L. Hsieh, B. Ganem, and D. B. Wilson. 1996.Identification of two functionally different classes ofexocellulases. Biochemistry 35:586–592.

Bayer, E. A., R. Kenig, and R. Lamed. 1983. Adherence ofClostridium thermocellum to cellulose. J. Bacteriol.156:818–827.

Bayer, E. A., E. Setter, and R. Lamed. 1985. Organizationand distribution of the cellulosome in Clostridium ther-mocellum. J. Bacteriol. 163:552–559.

Bayer, E. A., and R. Lamed. 1986. Ultrastructure of the cellsurface cellulosome of Clostridium thermocellum and itsinteraction with cellulose. J. Bacteriol. 167:828–836.

Bayer, E. A., and R. Lamed. 1992. The cellulose paradox:Pollutant par excellence and/or a reclaimable naturalresource?. Biodegradation 3:171–188.

Bayer, E. A., E. Morag, and R. Lamed. 1994. The cellulo-some—a treasure-trove for biotechnology. Trends Bio-technol. 12:378–386.

Bayer, E. A., E. Morag, Y. Shoham, J. Tormo, and R. Lamed.1996. The cellulosome: A cell-surface organelle for theadhesion to and degradation of cellulose. In: M. Fletcher(Ed.) Bacterial Adhesion: Molecular and EcologicalDiversity. Wiley-Liss. New York, NY. 155–182.

Bayer, E. A., H. Chanzy, R. Lamed, and Y. Shoham. 1998a.Cellulose, cellulases and cellulosomes. Curr. Opin.Struct. Biol. 8:548–557.

Bayer, E. A., E. Morag, R. Lamed, S. Yaron, and Y. Shoham.1998b. Cellulosome structure: Four-pronged attackusing biochemistry, molecular biology, crystallographyand bioinformatics. In: M. Claeyssens, W. Nerinckx, andK. Piens (Eds.) Carbohydrases from Trichoderma reeseiand Other Microorganisms. The Royal Society of Chem-istry. London, 39–67.

Bayer, E. A., L. J. W. Shimon, R. Lamed, and Y. Shoham.1998c. Cellulosomes: Structure and ultrastructure. J.Struct. Biol. 124:221–234.

Bayer, E. A., Y. Shoham, and R. Lamed. 2000. The cellulo-some—an exocellular organelle for degrading plant cellwall polysaccharides. In: R. J. Doyle (Ed.) Glycomicro-biology. Kluwer Academic/Plenum Publishers. NewYork, NY. 387–439.

Béguin, P. 1983. Detection of cellulase activity in polyacryla-mide gels using Congo red-stained agar replicas. Analyt.Biochem. 131:333–336.

Béguin, P. 1990. Molecular biology of cellulose degradation.Ann. Rev. Microbiol. 44:219–248.

Béguin, P., and J.-P. Aubert. 1994. The biological degradationof cellulose. FEMS Microbiol. Lett. 13:25–58.

Béguin, P., and M. Lemaire. 1996. The cellulosome: An exo-cellular, multiprotein complex specialized in cellulosedegradation. Crit. Rev. Biochem. Molec. Biol. 31:201–236.

Belaich, J.-P., C. Tardif, A. Belaich, and C. Gaudin. 1997. Thecellulolytic system of Clostridium cellulolyticum. J. Bio-technol. 57:3–14.

Belaich, A., J.-P. Belaich, H.-P. Fierobe, C. Gaudin, S. Pagès,C. Reverbel-Leroy, and C. Tardif. 1998. Cellulosomeanalysis and cellulases CelF and CelG from Clostridiumcellulolyticum. In: M. Claeyssens, W. Nerinckx, and K.Piens (Eds.) Carbohydrases from Trichoderma reeseiand Other Microorganisms. The Royal Society of Chem-istry. London, 73–86.


Belaich, J.-P., A. Belaich, H.-P. Fierobe, L. Gal, C. Gaudin, S.Pagès, C. Reverbel-Leroy, and C. Tardif. 1999. The cel-lulolytic system of Clostridium cellulolyticum. In: K.Ohmiya, K. Hayashi, K. Sakka, Y. Kobayashi, S. Karita,and T. Kimura (Ed.) Genetics, Biochemistry and Ecol-ogy of Cellulose Degradation. Uni Publishers. Tokyo,479–487.

Bhat, M. K. 2000. Cellulases and related enzymes in biotech-nology. Biotechnol. Adv. 18:355–383.

Biely, P. 1985. Microbial xylanolytic systems. Trends Biotech-nol. 3:285–290.

Blum, D. L., I. A. Kataeva, X. L. Li, and L. G. Ljungdahl.2000. Feruloyl esterase activity of the Clostridium ther-mocellum cellulosome can be attributed to previouslyunknown domains of XynY and XynZ. J Bacteriol182(5):1346–1351.

Boraston, A. B., B. W. McLean, J. M. Kormos, M. Alam,N. R. Gilkes, C. A. Haynes, P. Tomme, D. G. Kilburn,and R. A. Warren. 1999. Carbohydrate-binding mod-ules: Diversity of structure and function. In: H. J.Gilbert, G. J. Davies, B. Henrissat, and B. Svensson(Eds.) Recent Advances in Carbohydrate Bioengineer-ing. The Royal Society of Chemistry. Cambridge, 202–211.

Borneman, W. S., L. G. Ljungdahl, R. D. Hartley, and D. E.Akin. 1993. Feruloyl and p-coumaroyl esterases fromthe anaerobic fungus Neocallimastix strain MC-2 prop-erties and functions in plant cell wall degradation. In:M. P. Coughlan and G. P. Hazlewood (Eds.) Hemicellu-lose and Hemicellulases. Portland Press. London, 85–102.

Brun, E., F. Moriaud, P. Gans, M. Blackledge, F. Barras,and D. Marion. 1997. Solution structure of the cellu-lose-binding domain of the endoglucanase Z secretedby Erwinia chrysanthemi. Biochemistry 36:16074–16086.

Carpita, N. C., and D. M. Gibeaut. 1993. Structural modelsof primary cell walls in flowering plants: Consistency ofmolecular structure with the physical properties of thewalls during growth. Plant J. 3:1–30.

Chanzy, H. 1990. Aspects of cellulose structure. In: J. F.Kennedy, G. O. Philips, and P. A. Williams (Eds.) Cellu-lose Sources and Exploitation: Industrial Utilization,Biotechnology and Physico-chemical Properties. EllisHorwood. New York, NY. 3–12.

Chauvaux, S., P. Béguin, J.-P. Aubert, K. M. Bhat, L. A. Gow,T. M. Wood, and A. Bairoch. 1990. Calcium-bindingaffinity and calcium-enhanced activity of Clostridiumthermocellum endoglucanase D. Biochem. J. 265:261–265.

Chauvaux, S., M. Matuschek, and P. Béguin. 1999. Distinctaffinity of binding sites for S-layer homologous domainsin Clostridium thermocellum and Bacillus anthracis cellenvelopes. J Bacteriol. 181:2455–2458.

Chen, H., X. L. Li, D. L. Blum, and L. G. Ljungdahl. 1998.Two genes of the anaerobic fungus Orpinomyces sp.strain PC-2 encoding cellulases with endoglucanaseactivities may have arisen by gene duplication. FEMSMicrobiol. Lett. 159:63–68.

Claeyssens, M., and B. Henrissat. 1992. Specificity mappingof cellulolytic enzymes: Classification into families ofstructurally related proteins confirmed by biochemicalanalysis. Protein Sci. 1(10):1293–1297.

Coughlan, M. P., and F. Mayer. 1992. The cellulose-decomposing bacteria and their enzyme systems. In:A. Balows, H. G. Trüper, M. Dworkin, W. Harder, and

K.-H. Schleifer (Eds.) The Prokaryotes, 2nd ed.Springer. New York, NY. 1:459–516.

Coughlan, M. P., and G. P. Hazlewood. 1993. b-1,4-D-xylan-degrading enzyme systems: Biochemistry, molecularbiology and applications. Biotechnol. Appl. Biochem.17:259–289.

Coutinho, P. M., and B. Henrissat. 1999a. Carbohydrate-active enzymes: An integrated database approach. In:H. J. Gilbert, G. J. Davies, B. Henrissat, and B. Svensson(Eds.) Recent Advances in Carbohydrate Bioengineer-ing. The Royal Society of Chemistry. Cambridge, 3–12.

Coutinho, P. M., and B. Henrissat. 1999b. Carbohydrate-Active enZYmes and associated MODular Organi-zation server. CAZyModO Website afmb.cnrs-mrs.fr/.

Coutinho, P. M., and B. Henrissat. 1999c. The modular struc-ture of cellulases and other carbohydrate-activeenzymes: An integrated database approach. In: K. Ohm-iya, K. Hayashi, K. Sakka, Y. Kobayashi, S. Karita, andT. Kimura (Eds.) Genetics, Biochemistry and Ecology ofCellulose Degradation. Uni Publishers. Tokyo, 15–23.

Daniel, R. M., H. S. Toogood, and P. L. Bergquist. 1996.Thermostable proteases. Biotechnol. Genet. Engin. Rev.13:51–100.

Das, N. N., S. C. Das, and A. K. Mukerjee. 1984. On the esterlinkage between lignin and 4-O-methyl-D-glucurono-D-xylan in jute fiber (Corchorus capsularis). Carbohydr.Res. 127:345–348.

Davies, G., and B. Henrissat. 1995. Structures and mecha-nisms of glycosyl hydrolases. Structure 3:853–859.

Davies, G. J., L. Mackenzie, A. Varrot, M. Dauter, M. Brzo-zowski, M. Schülein, and S. G. Withers. 1998. Snapshotsalong an enzymatic reaction coordinate: Analysis of aretaining b-glycoside hydrolase. Biochemistry 37:11707–11713.

Ding, S.-Y., E. A. Bayer, D. Steiner, Y. Shoham, and R.Lamed. 1999. A novel cellulosomal scaffoldin from Ace-tivibrio cellulolyticus that contains a Family-9 glycosylhydrolase. J. Bacteriol. 181:6720–6729.

Ding, S.-Y., E. A. Bayer, D. Steiner, Y. Shoham, and R.Lamed. 2000. A scaffoldin of the Bacteroides cellulosol-vens cellulosome that contains 11 type II cohesins. J.Bacteriol. 182:4915–4925.

Ding, S.-Y., M. T. Rincon, R. Lamed, J. C. Martin, S. I.McCrae, V. Aurilia, Y. Shoham, E. A. Bayer, and H. J.Flint. 2001. Cellulosomal scaffoldin-like proteins fromRuminococcus flavefaciens. J. Bacteriol. 183:1945–1953.

Divne, C., J. Stahlberg, T. Teeri, and T. Jones. 1998. High-resolution crystal structures reveal how a cellulosechain is bound in the 50 Å long tunnel of cellobiohy-drolase I from Trichoderma reesei. J. Molec. Biol.275:309–325.

Doi, R. H., M. Goldstein, S. Hashida, J. S. Park, and M.Takagi. 1994. The Clostridium cellulovorans cellulo-some. Crit. Rev. Microbiol. 20:87–93.

Doi, R. H., J. S. Park, C. C. Liu, L. M. Malburg, Y. Tamaru,A. Ichiishi, and A. Ibrahim. 1998. Cellulosome and non-cellulosomal cellulases of Clostridium cellulovorans.Extremophiles 2:53–60.

Doi, R. H., and Y. Tamura. 2001. The Clostridium cellulo-vorans cellulosome: An enzyme complex with plant cellwall degrading activity. Chem. Rec. 1:24–32.

Doner, L. W., and P. L. Irwin. 1992. Assay of reducing end-groups in oligosaccaride homolgues with 2,2¢-bicinchon-inate. Analyt. Biochem. 202:50–53.


Driguez, H. 1997. Thiooligosaccharides in glycobiology.Topics Curr. Chem. 187:85–116.

Eriksson, K.-E. L., R. A. Blanchette, and P. Ander. 1990.Biodegradation of hemicelluloses. Microbial and Enzy-matic Degradation of Wood and Wood Components.Springer-Verlag. Heidelberg, 181–397.

Felix, C. R., and L. G. Ljungdahl. 1993. The cellulosome—the exocellular organelle of Clostridium. Ann. Rev.Microbiol. 47:791–819.

Fernandes, A. C., C. M. Fontes, H. J. Gilbert, G. P. Hazle-wood, T. H. Fernandes, and L. M. A. Ferreira. 1999.Homologous xylanases from Clostridium thermocellum:Evidence for bi-functional activity, synergism betweenxylanase catalytic modules and the presence of xylan-binding domains in enzyme complexes. Biochem. J.342:105–110.

Flint, H. J., J. Martin, C. A. McPherson, A. S. Daniel, andJ. X. Zhang. 1993. A bifunctional enzyme, with separatexylanase and b(1,3-1,4)-glucanase domains, encoded bythe xynD gene of Ruminococcus flavefaciens. J. Bacte-riol. 175:2943–2951.

Fujino, T., P. Béguin, and J.-P. Aubert. 1993. Organization ofa Clostridium thermocellum gene cluster encoding thecellulosomal scaffolding protein CipA and a proteinpossibly involved in attachment of the cellulosome tothe cell surface. J. Bacteriol. 175:1891–1899.

Gal, L., C. Gaudin, A. Belaich, S. Pagès, C. Tardif, and J.-P.Belaich. 1997a. CelG from Clostridium cellulolyticum:A multidomain endoglucanase acting efficiently on crys-talline cellulose. J. Bacteriol. 179:6595–6601.

Gal, L., S. Pagès, C. Gaudin, A. Belaich, C. Reverbel-Leroy,C. Tardif, and J.-P. Belaich. 1997b. Characterization ofthe cellulolytic complex (cellulosome) produced byClostridium cellulolyticum. Appl. Environ. Microbiol.63:903–909.

Garcia, E., D. Johnston, J. R. Whitaker, and S. P. Shoemaker.1993. Assessment of endo-1,4-beta-D-glucanase activityby a rapid colorimetric assay using disodium 2,2¢-bicinchoninate. J. Food Biochem. 17:135–145.

Gaudin, C., A. Belaich, S. Champ, and J. P. Belaich. 2000.CelE, a multidomain cellulase from Clostridium cellu-lolyticum: A key enzyme in the cellulosome? J. Bacte-riol. 182:1910–1915.

Ghose, T. K. 1987. Measurments of cellulase activity. PureAppl. Chem. 59:257–268.

Gibbs, M. D., R. A. Reeves, G. K. Farrington, P. Anderson,D. P. Williams, and P. L. Bergquist. 2000. Multidomainand multifunctional glycosyl hydrolases from theextreme thermophile Caldicellulosiruptor isolateTok7B.1. Curr. Microbiol. 40(5):333–40.

Gilbert, H. J., G. P. Hazlewood, J. I. Laurie, C. G. Orpin, andG. P. Xue. 1992. Homologous catalytic domains in arumen fungal xylanase: Evidence for gene duplicationand prokaryotic origin. Molec. Microbiol. 6:2065–2072.

Gilbert, H. J., and G. P. Hazlewood. 1993. Bacterial cellulasesand xylanases. J. Gen. Microbiol. 139:187–194.

Gilkes, N. R., B. Henrissat, D. G. Kilburn, R. C. J. Miller, andR. A. J. Warren. 1991. Domains in microbial b-1,4-glyca-nases: Sequence conservation, function, and enzymefamilies. Microbiol. Rev. 55(2):303–315.

Green 3rd, F., C. A. Clausen, and T. L. Highley. 1989. Adap-tation of the Nelson-Somogyi reducing-sugar assay to amicroassay using microtiter plates. Analyt. Biochem.182(2):197–9.

Guglielmi, G., and P. Béguin. 1998. Cellulase and hemicellu-lase genes of Clostridium thermocellum from five inde-

pendent collections contain few overlaps and are widelyscattered across the chromosome. FEMS Microbiol.Lett. 161:209–215.

Haigler, C. H., and P. J. Weimer. 1991. In: C. H. Haigler andP. J. Weimer (Eds.) Biosynthesis and Biodegradation ofCellulose. Marcel Dekker. New York, NY.

Hazlewood, G. P., and H. J. Gilbert. 1993. Molecular biologyof hemicellulases. In: M. P. Coughlan and G. P. Hazle-wood (Eds.) Hemicellulose and Hemicellulases. Port-land Press. London, 103–126.

Henrissat, B. 1991. A classification of glycosyl hydrolasesbased on amino acid sequence similarities. Biochem. J.280:309–316.

Henrissat, B., I. Callebaut, S. Fabrega, P. Lehn, J.-P. Mornon,and G. Davies. 1995. Conserved catalytic machinery andthe prediction of a common fold for several families ofglycosyl hydrolases. Proc. Natl. Acad. Sci. USA 92:7090–7094.

Henrissat, B., and A. Bairoch. 1996. Updating the sequence-based classification of glycosyl hydrolases. Biochem. J.316:695–696.

Henrissat, B., and G. Davies. 1997. Structural and sequence-based classification of glycoside hydrolases. Curr. Opin.Struct. Biol. 7:637–644.

Henrissat, B., T. T. Teeri, and R. A. J. Warren. 1998. Ascheme for designating enzymes that hydrolyse thepolysaccharides in the cell walls of plants. FEBS Lett.425:352–354.

Higuchi, T. 1990. Lignin biochemistry, biosynthesis and bio-degradation. Wood Sci. Technol. 24:23–63.

Hilden, L., L. Eng, G. Johansson, S. E. Lindqvist, and G.Pettersson. 2001. An amperometric cellobiose dehydro-genase-based biosensor can be used for measurement ofcellulase activity. Analyt. Biochem. 290(2):245–50.

Himmel, M. E., M. F. Ruth, and C. E. Wyman. 1999. Cellulasefor commodity products from cellulosic biomass. Curr.Opin. Biotechnol. 10(4):358–364.

Hurlbert, J. C., and J. F. Preston 3rd. 2001. Functional char-acterization of a novel xylanase from corn strain ofErwinia chrysanthemi. J. Bacteriol. 183:2093–2100.

Irwin, D., L. Walker, M. Spezio, and D. Wilson. 1993. Activitystudies of eight purified cellulases: Specificity, synergism,and binding domain effects. Biotech. Bioengin. 42:1002–1013.

Irwin, D., D.-H. Shin, S. Zhang, B. K. Barr, J. Sakon, P. A.Karplus, and D. B. Wilson. 1998. Roles of the catalyticdomain and two cellulose binding domains of Ther-momonospora fusca E4 in cellulose hydrolysis. J. Bacte-riol. 180:1709–1714.

Johnson, P., M. Joshi, P. Tomme, D. Kilburn, and L. McIntosh.1996. Structure of the N-terminal cellulose-bindingdomain of Cellulomonas fimi Cen C determined bynuclear magnetic resonance spectroscopy. Biochemistry35:14381–14394.

Jung, K. H., K. M. Lee, H. Kim, K. H. Yoon, S. H. Park, andM. Y. Pack. 1998. Cloning and expression of a Clostrid-ium thermocellum xylanase gene in Escherichia coli.Biochem. Molec. Biol. Intl. 44(2):283–292.

Juy, M., A. G. Amit, P. M. Alzari, R. J. Poljak, M. Claeyssens,P. Béguin, and J.-P. Aubert. 1992. Crystal structure ofa thermostable bacterial cellulose-degrading enzyme.Nature 357:39–41.

Karita, S., K. Sakka, and K. Ohmiya. 1997. Cellulosomes,cellulase complexes, of anaerobic microbes: Their struc-ture models and functions. In: R. Onodera, H. Itabashi,K. Ushida, H. Yano, and Y. Sasaki (Eds.) Rumen


Microbes and Digestive Physiology in Ruminants. Japa-nese Scientific Society Press and S. Karger. Tokyo/Basle,Germany. 14:47–57.

Kataeva, I., X.-L. Li, H. Chen, S. K. Choi, and L. G.Ljungdahl. 1999a. Cloning and sequence analysis of anew cellulase gene encoding CelK, a major cellulosomecomponent of Clostridium thermocellum: Evidencefor gene duplication and recombination. J. Bacteriol.181:5288–5295.

Kataeva, I., X.-L. Li, H. Chen, and L. G. Ljungdahl. 1999b.CelK—a new cellobiohydrolase from Clostridium ther-mocellum cellulosome: Role of N-terminal cellulose-binding domain. In: K. Ohmiya, K. Hayashi, K. Sakka,Y. Kobayashi, S. Karita, and T. Kimura (Eds.) Genetics,Biochemistry and Ecology of Cellulose Degradation.Uni Publishers. Tokyo, 454–460.

Kidby, D. K., and D. J. Davidson. 1973. A convenient ferri-cyanide estimation of reducing sugars in the nanomolerange. Analyt. Biochem. 55(1):321–325.

Kuhad, R. C., A. Singh, and K.-E. Eriksson. 1997. Microor-ganisms and enzymes involved in the degradation ofplant fiber cell walls. Adv. Biochem. Engin. 57:45–125.

Lamed, R., E. Setter, and E. A. Bayer. 1983a. Characteriza-tion of a cellulose-binding, cellulase-containing complexin Clostridium thermocellum. J. Bacteriol. 156:828–836.

Lamed, R., E. Setter, R. Kenig, and E. A. Bayer. 1983b.The cellulosome—a discrete cell surface organelle ofClostridium thermocellum which exhibits separate anti-genic, cellulose-binding and various cellulolytic activi-ties. Biotechnol. Bioeng. Symp. 13:163–181.

Lamed, E., J. Naimark, E. Morgenstern, and E. A. Bayer.1987a. Scanning electron microscopic delineation ofbacterial surface topology using cationized ferritin. J.Microbiol. Methods 7:233–240.

Lamed, R., J. Naimark, E. Morgenstern, and E. A. Bayer.1987b. Specialized cell surface structures in cellulolyticbacteria. J. Bacteriol. 169:3792–3800.

Lamed, R., and E. A. Bayer. 1988a. The cellulosome concept:Exocellular/extracellular enzyme reactor centers forefficient binding and cellulolysis. In: J.-P. Aubert, P.Beguin, and J. Millet (Eds.) Biochemistry and Geneticsof Cellulose Degradation. Academic Press. London,101–116.

Lamed, R., and E. A. Bayer. 1988b. The cellulosome ofClostridium thermocellum. Adv. Appl. Microbiol. 33:1–46.

Lamed, R., and E. A. Bayer. 1991. Cellulose degradation bythermophilic anaerobic bacteria. In: C. H. Haigler andP. J. Weimer (Eds.) Biosynthesis and Biodegradation ofCellulose and Cellulose Materials. Marcel Dekker. NewYork, NY. 377–410.

Lamed, R., and E. A. Bayer. 1993. The cellulosome con-cept—a decade later!. In: K. Shimada, S. Hoshino, K.Ohmiya, K. Sakka, Y. Kobayashi, and S. Karita (Eds.)Genetics, Biochemistry and Ecology of LignocelluloseDegradation. Uni Publishers. Tokyo, 1–12.

Laurie, J. I., J. H. Clarke, A. Ciruela, C. B. Faulds, G.Williamson, H. J. Gilbert, J. E. Rixon, J. Millward-Sadler, and G. P. Hazlewood. 1997. The NodBdomain of a multidomain xylanase from Cellulomo-nas fimi deacylates acetylxylan. FEMS Microbiol.Lett. 148:261–264.

Leibovitz, E., and P. Béguin. 1996. A new type of cohesindomain that specifically binds the dockerin domain ofthe Clostridium thermocellum cellulosome-integratingprotein CipA. J. Bacteriol. 178:3077–3084.

Leibovitz, E., H. Ohayon, P. Gounon, and P. Béguin. 1997.Characterization and subcellular localization of theClostridium thermocellum scaffoldin dockerin bindingprotein SdbA. J. Bacteriol. 179:2519–2523.

Lemaire, M., H. Ohayon, P. Gounon, T. Fujino, and P.Béguin. 1995. OlpB, a new outer layer protein ofClostridium thermocellum, and binding of its S-layer-like domains to components of the cell envelope. J. Bac-teriol. 177:2451–2459.

Lemaire, M., I. Miras, P. Gounon, and P. Béguin. 1998. Iden-tification of a region responsible for binding to the cellwall within the S-layer protein of Clostridium thermo-cellum. Microbiology 144:211–217.

Lewis, N. G., and E. Yamamoto. 1990. Lignin: Occurrence,biogenesis and biodegradation. Ann. Rev. Plant Physiol.Plant Molec. Biol. 41:455–496.

Linder, M., and T. T. Teeri. 1997. The roles and function ofcellulose-binding domains. J. Biotechnol. 57:15–28.

Liu, C. C., and R. H. Doi. 1998. Properties of exgS, a genefor a major subunit of the Clostridium cellulovoranscellulosome. Gene 211:39–47.

Ljungdahl, L. G., and K.-E. Eriksson. 1985. Ecology ofmicrobial cellulose degradation. Adv. Microb. Ecol.8:237–299.

Lupas, A., H. Engelhardt, J. Peters, U. Santarius, S. Volker,and W. Baumeister. 1994. Domain structure of the Ace-togenium kivui surface layer revealed by electron crys-tallography and sequence analysis. J. Bacteriol.176:1224–1233.

Ly, H. D., and S. G. Withers. 1999. Mutagenesis of glycosi-dases. Ann. Rev. Biochem. 68:487–522.

Lynd, L. R., J. H. Cushman, R. J. Nichols, and C. E. Wyman.1991. Fuel ethanol from cellulosic biomass. Science251:1318–1323.

Lytle, B., B. F. Volkman, W. M. Westler, and J. H. D. Wu. 2000.Secondary structure and calcium-induced folding of theClostridium thermocellum dockerin domain determinedby NMR spectroscopy. Arch. Biochem. Biophys.379:237–244.

Lytle, B. L., B. F. Volkman, W. M. Westler, M. P. Heckman,and J. H. Wu. 2001. Solution structure of a type i dock-erin domain, a novel prokaryotic, extracellular calcium-binding domain. J. Molec. Biol. 307(3):745–753.

Marais, J. P., J. L. De Wit, and G. V. Quicke. 1966. A criticalexamination of the Nelson-Somogyi method for thedetermination of reducing sugars. Analyt. Biochem.15(3):373–381.

Mattinen, M.-L., M. Kontteli, J. Kerovuo, M. Linder, A.Annila, G. Lindeberg, T. Reinikainen, and T. Draken-berg. 1997. Three-dimensional structures of threeengineered cellulose-binding domains of cellobiohy-drolase I from Trichoderma reesei. Protein Sci. 6:294–303.

Matuschek, M., K. Sahm, A. Zibat, and H. Bahl. 1996. Char-acterization of genes from Thermoanaerobacteriumthermosulfurigenes EM1 that encode two glycosylhydrolases with conserved S-layer-like domains. Molec.Gen. Genet. 252(4):493–496.

Mayer, F., M. P. Coughlan, Y. Mori, and L. G. Ljungdahl.1987. Macromolecular organization of the cellulolyticenzyme complex of Clostridium thermocellum asrevealed by electron microscopy. Appl. Environ. Micro-biol. 53:2785–2792.

McCarter, J. D., and S. G. Withers. 1994. Mechanisms ofenzymatic glycoside hydrolysis. Curr. Opin. Struct. Biol.4:885–892.


Mechaly, A., S. Yaron, R. Lamed, H.-P. Fierobe, A. Belaich,J.-P. Belaich, Y. Shoham, and E. A. Bayer. 2000.Cohesin-dockerin recognition in cellulosome assembly:Experiment versus hypothesis. Proteins 39:170–177.

Mechaly, A., H.-P. Fierobe, A. Belaich, J.-P. Belaich, R.Lamed, Y. Shoham, and E. A. Bayer. 2001. Cohesin-dockerin interaction in cellulosome assembly: A singlehydroxyl group of a dockerin domain distinguishesbetween non-recognition and high-affinity recognition.J. Biol. Chem. 276:9883–9888.

Miller, G. L. R., W. E. Blum, and A. L. Burton. 1960. Mea-surements of carboxymethylcellulase activity. Analyt.Biochem. 2:127–132.

Mohand-Oussaid, O., S. Payot, E. Guedon, E. Gelhaye, A.Youyou, and H. Petitdemange. 1999. The extracellularxylan degradative system in Clostridium cellulolyticumcultivated on xylan: Evidence for cell-free cellulosomeproduction. J. Bacteriol. 181:4035–4040.

Morag, E., E. A. Bayer, and R. Lamed. 1990. Relationshipof cellulosomal and noncellulosomal xylanases ofClostridium thermocellum to cellulose-degradingenzymes. J. Bacteriol. 172:6098–6105.

Morag, E., I. Halevy, E. A. Bayer, and R. Lamed. 1991.Isolation and properties of a major cellobiohydrolasefrom the cellulosome of Clostridium thermocellum. J.Bacteriol. 173:4155–4162.

Morag, E., E. A. Bayer, G. P. Hazlewood, H. J. Gilbert, andR. Lamed. 1993. Cellulase SS (CelS) is synonymous withthe major cellobiohydrolase (subunit S8) from the cel-lulosome of Clostridium thermocellum. Appl. Biochem.Biotechnol. 43:147–151.

Morag, E., A. Lapidot, D. Govorko, R. Lamed, M. Wilchek,E. A. Bayer, and Y. Shoham. 1995. Expression, purifi-cation and characterization of the cellulose-bindingdomain of the scaffoldin subunit from the cellulosomeof Clostridium thermocellum. Appl. Environ. Microbiol.61:1980–1986.

Navarro, A., M.-C. Chebrou, P. Béguin, and J.-P. Aubert.1991. Nucleotide sequence of the cellulase gene celF ofClostridium thermocellum. Res. Microbiol. 142:927–936.

Notenboom, V., C. Birsan, R. Warren, S. Withers, and D.Rose. 1998. Exploring the cellulose/xylan specificity ofthe b-1,4-glycanase Cex from Cellulomonas fimi throughcrystallography and mutation. Biochemistry 37:4751–4758.

Ohmiya, K., K. Sakka, S. Karita, and T. Kimura. 1997. Struc-ture of cellulases and their applications. Biotechnol.Genet. Engin. Rev. 14:365–414.

O’Neill, G., S. H. Goh, R. A. Warren, D. G. Kilburn, andR. C. Miller. 1986. Structure of the gene encoding theexoglucanase of Cellulomonas fimi. Gene 44(2–3):325–30.

O’Neill, R. A., A. Darvill, and P. Albersheim. 1989. A fluo-rescence assay for enzymes that cleave glycosidic link-ages to produce reducing sugars. Analyt. Biochem.177(1):11–5.

O’Sullivan, A. C. 1997. Cellulose: The structure slowlyunravels. Cellulose 4:173–207.

Pagès, S., A. Belaich, J.-P. Belaich, E. Morag, R. Lamed, Y.Shoham, and E. A. Bayer. 1997. Species-specificity of thecohesin-dockerin interaction between Clostridium ther-mocellum and Clostridium cellulolyticum: Prediction ofspecificity determinants of the dockerin domain. Pro-teins 29:517–527.

Parsiegla, G., M. Juy, C. Reverbel-Leroy, C. Tardif, J. P.Belaich, H. Driguez, and R. Haser. 1998. The crystal

structure of the processive endocellulase CelF ofClostridium cellulolyticum in complex with a thiooli-gosaccharide inhibitor at 2.0 Å resolution. EMBO J.17:5551–5562.

Pegden, R. S., M. A. Larson, R. J. Grant, and M. Morrison.1998. Adherence of the Gram-positive bacterium Rumi-nococcus albus to cellulose and identification of a novelform of cellulose-binding protein which belongs to thePil family of proteins. J. Bacteriol. 180:5921–5927.

Puls, J., and J. Schuseil. 1993. Chemistry of hemicellulases:Relationship between hemicellulose structure andenzymes required for hydrolysis. In: M. P. Coughlan andG. P. Hazlewood (Eds.) Hemicellulose and Hemicellu-lases. Portland Press. London, 1–27.

Reese, R. T. 1976. History of the cellulase program at the U.S.Army Natick Development Center. Biotechnol. Bioeng.Symp. 6:9–20.

Reeves, R. A., M. D. Gibbs, D. D. Morris, K. R. Griffiths,D. J. Saul, and P. L. Bergquist. 2000. Sequencing andexpression of additional xylanase genes from the hyper-thermophile Thermotoga maritima FjSS3B.1. Appl.Environ. Microbiol. 66(4):1532–1537.

Reverbel-Leroy, C., S. Pagés, A. Belaich, J.-P. Belaich, and C.Tardif. 1997. The processive endocellulase CelF, a majorcomponent of the Clostridium cellulolyticum cellulo-some: Purification and characterization of the recombi-nant form. J. Bacteriol. 179:46–52.

Rouvinen, J., T. Bergfors, T. Teeri, J. K. C. Knowles, and T. A.Jones. 1990. Three-dimensional structure of cellobiohy-drolase II from Trichoderma reesei. Science 279:380–386.

Rydlund, A., and O. Dahlman. 1997. Oligosaccharidesobtained by enzymatic hydrolysis of birch kraft pulpxylan: Analysis by capillary zone electrophoresis andmass spectrometry. Carbohydr. Res. 300:95–102.

Rye, C. S., and S. G. Withers. 2000. Glycosidase mechanisms.Curr. Opin. Chem. Biol. 4:573–580.

Sakon, J., D. Irwin, D. B. Wilson, and P. A. Karplus. 1997.Structure and mechanism of endo/exocellulase E4 fromThermomonospora fusca. Nature Struct. Biol. 4:810–818.

Salamitou, S., M. Lemaire, T. Fujino, H. Ohayon, P. Gounon,P. Béguin, and J.-P. Aubert. 1994a. Subcellular localiza-tion of Clostridium thermocellum ORF3p, a protein car-rying a receptor for the docking sequence borne by thecatalytic components of the cellulosome. J. Bacteriol.176:2828–2834.

Salamitou, S., O. Raynaud, M. Lemaire, M. Coughlan, P.Béguin, and J.-P. Aubert. 1994b. Recognition specificityof the duplicated segments present in Clostridiumthermocellum endoglucanase CelD and in the cellulo-some-integrating protein CipA. J. Bacteriol. 176:2822–2827.

Saul D. J., L. C. Williams, R. A. Grayling, L. W. Chamley,D. R. Love, and P. L. Bergquist. 1990. celB, a gene codingfor a bifunctional cellulase from the extreme thermo-phile “Caldocellum saccharolyticum”. Appl. Environ.Microbiol. 56(10):3117–3124.

Schülein, M. 1997. Enzymatic properties of cellulases fromHumicola insolens. J. Biotechnol. 57:71–81.

Shen, H., A. Meinke, P. Tomme, H. G. Damude, E. Kwan,D. G. Kilburn, R. C. Miller Jr., R. A. J. Warren, andN. R. Gilkes. 1995. Cellulomonas fimi cellobiohy-drolases. In: J. N. Saddler and M. H. Penner (Eds.)Enzymatic Degradation of Insoluble Polysaccharides.American Chemical Society. Washington DC, 174–196.


Shimon, L. J. W., E. A. Bayer, E. Morag, R. Lamed, S.Yaron, Y. Shoham, and F. Frolow. 1997. The crystalstructure at 2.15 Å resolution of a cohesin domain ofthe cellulosome from Clostridium thermocellum. Struc-ture 5:381–390.

Shoham, Y., R. Lamed, and E. A. Bayer. 1999. The cellulo-some concept as an efficient microbial strategy for thedegradation of insoluble polysaccharides. Trends Micro-biol. 7:275–281.

Simpson, H. D., and F. Barras. 1999a. Functional analysis ofthe carbohydrate-binding domains of Erwinia chrysan-themi Cel5 (Endoglucanase Z) and an Escherichia coliputative chitinase. J Bacteriol. 181(15):4611–4616.

Simpson, P. J., D. N. Bolam, A. Cooper, A. Ciruela, G. P.Hazlewood, H. J. Gilbert, and M. P. Williamson. 1999b.A family IIb xylan-binding domain has a similar second-ary structure to a homologous family IIa cellulose-binding domain but different ligand specificity.Structure Fold. Des. 7:853–864.

Sinnott, M. L. 1990. Catalytic mechanisms of enzymic glyco-syl transfer. Chem. Rev. 90:1171–1202.

Sleat, R., R. A. Mah, and R. Robinson. 1984. Isolation andcharacterization of an anaerobic, cellulolytic bacterium,Clostridium cellulovorans, sp. nov. Appl. Environ.Microbiol. 48:88–93.

Spinelli, S., H. P. Fierobe, A. Belaich, J. P. Belaich, B.Henrissat, and C. Cambillau. 2000. Crystal structure ofa cohesin module from Clostridium cellulolyticum:Implications for dockerin recognition. J. Molec. Biol.304(2):189–200.

Srisodsuk, M., K. Kleman-Leyer, S. Keranen, T. K. Kirk, andT. T. Teeri. 1998. Modes of action on cotton and bacterialcellulose of a homologous endoglucanase-exoglucanasepair from Trichoderma reesei. Eur. J. Biochem.251(3):885–892.

Stutzenberger, F. 1990. Bacterial cellulases. In: W. M. Fogartyand C. T. Kelly (Eds.) Microbial Enzymes and Biotech-nology. Elsevier Applied Science. London, New York,37–70.

Tamaru, Y., and R. H. Doi. 1999a. Three surface layer homol-ogy domains at the N terminus of the Clostridium cellu-lovorans major cellulosomal subunit EngE. J. Bacteriol.181:3270–3276.

Tamaru, Y., C.-C. Liu, A. Ichi-ishi, L. Malburg, and R. H. Doi.1999b. The Clostridium cellulovorans cellulosome andnon-cellulosomal cellulases. In: K. Ohmiya, K. Hayashi,K. Sakka, Y. Kobayashi, S. Karita, and T. Kimura (Eds.)Genetics, Biochemistry and Ecology of Cellulose Deg-radation. Uni Publishers. Tokyo, 488–494.

Tamaru, Y., and R. H. Doi. 2000a. The engL gene cluster ofClostridium cellulovorans contains a gene for celluloso-mal ManA. J. Bacteriol. 182:244–247.

Tamaru, Y., S. Karita, A. Ibrahim, H. Chan, and R. H.Doi. 2000b. A large gene cluster for the Clostridiumcellulovorans cellulosome. J Bacteriol. 182(20):5906–5910.

Tamaru, Y., and R. H. Doi. 2001. Pectate lyase A, an enzy-matic subunit of the Clostridium cellulovorans cellulo-some. Proc. Natl. Acad. Sci. USA 20:20.

Tavares, G. A., P. Béguin, and P. M. Alzari. 1997. The crystalstructure of a type I cohesin domain at 1.7 Å resolution.J. Molec. Biol. 273:701–713.

Teeri, T. T., T. Reinikainen, L. Ruohonen, T. A. Jones, andJ. K. C. Knowles. 1992. Domain function in Trichodermareesei cellulases. J. Biotechnol. 24:169–176.

Teeri, T. T. 1997. Crystalline cellulose degradation: Newinsight into the function of cellobiohydrolases. TrendsBiotechnol. 15:160–167.

Te’o, V. S., D. J. Saul, and P. L. Bergquist. 1995. CelA, anothergene coding for a multidomain cellulase from theextreme thermophile Caldocellum saccharolyticum.Appl. Microbiol. Biotechnol. 43:291–296.

Timell, T. E. 1967. Recent progress in the chemistry of woodhemicelluloses. Wood Sci. Technol. 1:45–70.

Tokatlidis, K., S. Salamitou, P. Béguin, P. Dhurjati, and J.-P.Aubert. 1991. Interaction of the duplicated segmentcarried by Clostridium thermocellum cellulases withcellulosome components. FEBS Lett. 291:185–188.

Tokatlidis, K., P. Dhurjati, and P. Béguin. 1993. Propertiesconferred on Clostridium thermocellum endoglucanaseCelC by grafting the duplicated segment of endogluca-nase CelD. Protein Engin. 6(8):947–952.

Tomme, P., R. A. J. Warren, and N. R. Gilkes. 1995a. Cellulosehydrolysis by bacteria and fungi. Adv. Microb. Physiol.37:1–81.

Tomme, P., R. A. J. Warren, R. C. Miller, D. G. Kilburn, andN. R. Gilkes. 1995b. Cellulose-binding domains: Classi-fication and properties. In: J. M. Saddler and M. H.Penner (Eds.) Enzymatic Degradation of InsolublePolysaccharides. American Chemical Society. Washing-ton DC, 142–161.

Tomme, P., A. L. Creagh, D. G. Kilburn, and C. A. Haynes.1996. Interaction of polysaccharides with the N-terminalcellulose-binding domain of Cellulomonas fimi CenC.Biochemistry 35:13885–13894.

Tormo, J., R. Lamed, A. J. Chirino, E. Morag, E. A. Bayer,Y. Shoham, and T. A. Steitz. 1996. Crystal structure of abacterial family-III cellulose-binding domain: A generalmechanism for attachment to cellulose. EMBO J.15:5739–5751.

Tull, D., and S. G. Withers. 1994. Mechanisms of cellulasesand xylanases: A detailed kinetic study of the exo-beta-1,4-glycanase from Cellulomonas fimi. Biochemistry33(20):6363–6370.

van Tilbeurgh, H., G. Pettersson, R. Bhikabhai, H. DeBoeck, and M. Claeyssens. 1985. Studies of thecellulolytic system of Trichoderma reesei QM 9414.Reaction specificity and thermodynamics of interac-tions of small substrates and ligands with the 1,4-beta-glucan cellobiohydrolase II. Eur. J. Biochem.148(2):329–34.

Viikari, L., and T. Teeri. 1997. In: L. Viikari and T. Teeri (Eds.)Biochemistry and Genetics of Cellulases and Hemicel-lulases and Their Application.

Vlasenko, E. Y., A. I. Ryan, C. F. Shoemaker, and S. P. Shoe-maker. 1998. The use of capillary viscometry, reducingend-group analysis, and size exclusion chromatographycombined with multi-angle laser light scattering to char-acterize endo-1,4-b-D-glucanases on carboxymethylcel-lulose: A comparative evaluation of three methods.Enzyme Microb. Technol. 23:350–359.

Waffenschmidt, S., and L. Jaenicke. 1987. Assay of reducingsugars in the nanomole range with 2,2¢-bicinchoninate.Analyt. Biochem. 165:337–340.

Wang, W. K., K. Kruus, and J. H. D. Wu. 1993. Cloning andDNA sequence of the gene coding for Clostridium ther-mocellum cellulase SS (CelS), a major cellulosome com-ponent. J. Bacteriol. 175:1293–1302.

Wang, W. K., K. Kruus, and J. H. D. Wu. 1994. Cloning andexpression of the Clostridium thermocellum cellulase


celS gene in Escherichia coli. Appl. Microbiol. Biotech-nol. 42:346–352.

Warren, R. A. J. 1996. Microbial hydrolysis of polysaccha-rides. Ann. Rev. Microbiol. 50:183–212.

White, A., and D. R. Rose. 1997. Mechanism of catalysis byretaining b-glycosyl hydrolases. Curr. Opin. Struct. Biol.7:645–651.

Whittle, D. J., D. G. Kilburn, R. A. Warren, and R. C.Miller. 1982. Molecular cloning of a Cellulomonasfimi cellulose gene in Escherichia coli. Gene17(2):139–145.

Williams, S. J., and S. G. Withers. 2000. Glycosyl fluorides inenzymatic reactions. Carbohydr. Res. 327:27–46.

Williamson, M. P., P. J. Simpson, D. N. Bolam, G. P.Hazlewood, A. Ciruela, A. Cooper, and H. J. Gilbert.1999. How the N-terminal xylan-binding domain fromC. fimi xylanase D recognises xylan. In: H. J. Gilbert,G. J. Davies, B. Henrissat, and B. Svensson (Eds.) RecentAdvances in Carbohydrate Bioengineering. The RoyalSociety of Chemistry. Cambridge, 212–220.

Wilson, D. B. 1992. Biochemistry and genetics of actino-mycete cellulases. Crit. Rev. Biotechnol. 12:45–63.

Wilson, D. B., and D. C. Irwin. 1999. Genetics and propertiesof cellulases. Adv. Biochem. Engin. 65:1–21.

Withers, S. G., and R. Aebersold. 1995. Approaches to label-ing and identification of active site residues in glycosi-dases. Protein Sci. 4(3):361–372.

Withers, S. G. 2001. Mechanisms of glycosyl transferases andhydrolases. Carbohydr. Res. 44:325–337.

Wood, W. A., and S. T. Kellogg. 1988. In: W. A. Wood andS. T. Kellogg (Eds.) Biomass. Part A: Cellulose andHemicellulose. Academic Press. San Diego, CA. 160.

Wu, J. H. D., W. H. Orme-Johnson, and A. L. Demain. 1988.Two components of an extracellular protein aggregateof Clostridium thermocellum together degrade crystal-ine cellulose. Biochemistry 27:1703–1709.

Yagüe, E., P. Béguin, and J.-P. Aubert. 1990. Nucleotidesequence and deletion analysis of the cellulase-

encoding gene celH of Clostridium thermocellum.Gene 89:61–67.

Yaron, S., E. Morag, E. A. Bayer, R. Lamed, and Y. Shoham.1995. Expression, purification and subunit-binding prop-erties of cohesins 2 and 3 of the Clostridium thermocel-lum cellulosome. FEBS Lett. 360:121–124.

Zechel, D. L., and S. G. Withers. 2000. Glycosidase mecha-nisms: Anatomy of a finely tuned catalyst. Acc. Chem.Res. 33(1):11–18.

Zou, J., G. J. Kleywegt, J. Stahlberg, H. Driguez, W. Ner-inckx, M. Claeyssens, A. Koivula, T. T. Teeri, andT. A. Jones. 1999. Crystallographic evidence for sub-strate ring distortion and protein conformationalchanges during catalysis in cellobiohydrolase Ce16Afrom trichoderma reesei. Struct. Fold. Des. 7(9):1035–1045.

Zverlov, V. V., S. Mahr, K. Riedel, and K. Bronnenmeier.1998a. Properties and gene structure of a bifunctionalcellulolytic enzyme (CelA) from the extreme ther-mophile Anaerocellum thermophilum with separateglycosyl hydrolase family 9 and 48 catalytic domains.Microbiology 144:457–465.

Zverlov, V. V., G. V. Velikodvorskaya, W. H. Schwarz, K.Bronnenmeier, J. Kellermann, and W. L. Staudenbauer.1998b. Multidomain structure and cellulosomal localiza-tion of the Clostridium thermocellum cellobiohydrolaseCbhA. J. Bacteriol. 180:3091–3099.

Zverlov, V. V., G. V. Velikodvorskaya, W. H. Schwarz, J.Kellermann, and W. L. Staudenbauer. 1999. DuplicatedClostridium thermocellum cellobiohydrolase geneencoding cellulosomal subunits S3 and S5. Appl. Micro-biol. Biotechnol. 51:852–859.

Zverlov, V. V., I. Y. Volkov, G. A. Velikodvorskaya, andW. H. Schwarz. 2001. The binding pattern of twocarbohydrate-binding modules of laminarinase Lam16Afrom Thermotoga neapolitana: Differences in beta-glucan binding within family CBM4. Microbiology147(3):621–629.

the prokaryotes || cellulose-decomposing bacteria and their enzyme systems

Documents