Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Characterization of double binding domain derivatives of CenA from Cellulomonas fimi Nordquist, David Allen 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_spring_nordquist_david_allen.pdf [ 1.22MB ]
JSON: 831-1.0086721.json
JSON-LD: 831-1.0086721-ld.json
RDF/XML (Pretty): 831-1.0086721-rdf.xml
RDF/JSON: 831-1.0086721-rdf.json
Turtle: 831-1.0086721-turtle.txt
N-Triples: 831-1.0086721-rdf-ntriples.txt
Original Record: 831-1.0086721-source.json
Full Text

Full Text

Characterization of Double Binding Domain Derivativesof CenA from CellulomonasfimiByDavid Allen NordquistB.Sc., The University of British Columbia, 1990A THESIS SUBMITI’ED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(Department of Microbiology)We accept this thesis as conformingto the required standardUTHE UNIVERSITY OF BRITISH COLUMBIAAugust 1992© David Allen Nordquist, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of___________________The University of British ColumbiaVancouver, CanadaDate 41C’T L/c? LDE-6 (2/88)UABSTRACTDerivatives of CenA and of CBD.PT from C. fimi with tandemly repeated bindingdomains were constructed. A sensitive assay using 14C-labeled proteins was developed tomeasure adsorption affinities for microcrystalline cellulose at low protein concentrations. Itwas found that the CBDs in the double binding domain derivatives could function in tandemto increase the overall adsorption affinity of CenA, provided that the catalytic domain waspresent and that the CBDs were separated by a full Pro-Thr linker. The catalytic domain wasfound to contribute to the overall adsorption affinity of the enzyme, as the affinity of theisolated binding domain was reduced compared to that of CenA. Furthermore, the nature ofthe linker separating domains in wild-type CenA was important for adsorption affinity of theenzyme.The activities of the double binding domain derivatives of CenA on microcrystallineand soluble substrates were unchanged compared to the activity of the wild type enzyme.Furthermore, double binding domain derivatives of CBD.PT released more small particlesfrom intact cotton fibres than did CBD.PT, and this effect was independent of the adsorptionaffinity of the proteins.111TABLE OF CONTENTSPageABSTRACT iiLIST OF TABLES vLIST OF FIGURES viACKNOWLEDGMENTS viii1. Introduction 11.1 Cellulose 11.2 Cellulases 11.3 CenA from Cellulomonasfimi 21.4 Objectives 32. Materials and Methods 72.1. Bacterial strains, plasmids, and phage 72.2. Media and buffers 82.3. Recombinant DNA techniques 82.4. Sequencing of DNA 92.5. Screening colonies for endoglucanase activity 92.6. Production and purification of proteins 92.6.1. Growth of cells and preparation of cell extracts 92.6.2. Cellulose affinity chromatography 102.6.3. Electrophoresis of proteins 112.7. Purification of radiolabeled proteins 112.7.1. Growth of cells 112.7.2. Batch purification of labeled proteins 12ivPage2.7.3. Autoradiography of labeled proteins 132.8. Determination of protein concentration 132.9. Cellulose adsorption assays 132.10. Enzymatic activity assays 142.11. Fragmentation of cotton fibres and release of small particles 153. Results 163.1. Construction of plasmids encoding double binding domainderivatives of Cen and CBDCenA 163.1.1. Site directed mutagenesis 163.1.2. Construction of plasmids 173.2. Purification of proteins by affinity chromatography on cellulose 183.3. Adsorption assays 193.3.1. High protein concentration adsorption assays 213.3.2. Low initial protein concentration binding assays 213.4. Enzymatic activities 243.5. Small particle release from cotton fibres 254. Discussion 554.1. Interpretation of adsorption data 564.2. Effects of altering the linker between the catalytic and bindingdomains of CenA on the adsorption affinity of the enzyme 624.3. Small particle release and enzymatic activity 655. References 67LIST OF TABLESTable Page2.1 Bacterial strains, phage, and p1asd .73.1 Adsorption parameters from double reciprocal plots 243.2 Specific activities of CenA and derivatives on various substrates 25VviLIST OF FIGURESFigure Page1.1 Synergistic model for the degradation of cellulose by cellulases 51.2 Structural features of CenA 63.1 Block diagrams of CenA and derivatives 273.2 Addition of an NheI site at the junction of the leader peptide andthe CBDCenA coding sequences 283.3 Addition of an NheI site at the junction of the Pro-Thr linker andthe CenA catalytic domain coding sequences 293.4 Deletion of 30 nucleotides of the 3’ end of the coding region of thePro-Thr linker of CenA, and the addition of an NheI site 303.5 Construction of pTZ1 8R-CenADBD 313.6 Construction of pTZ18R-CenADBD1.2 333.7 Construction of pTZ18R-2xCBDpt 353.8 Construction of pTZ18R-2xCBDptl.2 373.9 Construction of pTZ1 8R-CBDptNhe 393.10 SDS-Page analysis of purified, non-labeled derivatives of CenA 413.11 Autoradiograms of purified, 14C-labeled derivatives of CenA 423.12 SDS-Page analysis of purified, non-labeled derivatives of CBD.PT 433.13 Autoradiograms of purified, 14C-labeled derivatives of CBD.PT 443.14 Western blot of purified, non-labeled proteins 453.15 High protein concentration adsorption isotherms of CenA,CenADBD, and CenADBD1.2 463.16 High protein concentration adsorption isotherms of CBD.PT2xCBDpt, and 2xCBDptl.2 47vii3.17 Effects of adding BSA in increasing concentrations on theadsorption of CenA 483.18 Effects of increasing BSA concentrations on the stability of CenAunder the conditions of the adsorption assay 493.19 Low initial protein concentration adsorption isotherm for‘4C-labeled CenA and CBD.PT 503.20 Low initial protein concentration adsorption isotherm for14C-labeled CenA and CBD.PT 513.21 Low initial protein concentration adsorption isotherm for14C-labeled 2xCBDpt and 2xCBDptl.2 523.22 Low initial protein concentration adsorption isotherm for14C-labeleci CenAl.2 and CenAt\PT 533.23 Small particle production from dewaxed cotton 54viiiACKNOWLEDGMENTSI would like to thank Drs. D.G. Kilburn, R.C. Miller, Jr., and R.A.J. Warren for theirsupervision and enthusiasm for this project. I am especially grateful to Doug for giving methe opportunity to pursue a Master’s degree. Thanks to all the members of the CellulaseGroup who answered so many questions and offered so much helpful advice, especially Dr.Edgar Ong, Neena Din, and Dr. Neil Gilkes. I greatly appreciate Drs. Roger Graham, CeliaRamirez, Peter Tomme and Edgar Ong for taking the time to review this thesis and makemany helpful suggestions. Thanks to Dedreia Tull for assistance in preparing dewaxed cottonfibres, and to Eric Jervis for helpful discussions regarding the adsorption model. I would alsolike to thank Pat and Bob for the weekend at Bliss. Finally, thanks to Beth for constantsupport and for always being around to listen to both the good and the not-so-good. Idedicate this thesis to the memory of my father, John Raymond Nordquist.11. Introduction1.1 CelluloseCellulose molecules are linear polymers of up to 10,000 glucose molecules joined by13-1,4-glucosidic bonds (Beguin et at, 1985). In the cell walls of plants, cellulose chainshydrogen bond to each other in parallel to form fibrils (Blackwell, 1981) that contain regionsthat are highly ordered, or crystalline, as well as regions that more disordered, or amorphous.The fibrils associate with each other and with hemicellulose and lignin to form fibres (Fan etat., 1980). The structural features of cellulose in plant cell walls are shown in Figure 1.1. Inaddition to plants, some fungi, bacteria, invertebrates, and protists also synthesize cellulose(Richmond, 1991). The gram-negative bacterium Acetobacter xylinum, for instance,produces large amounts of cellulose (Richmond, 1991). Instead of being associated with thecell wall as it is in other cellulose-producing organisms, the cellulose produced by A. xytinumis secreted from the cell in a long, continuous ribbon (Richmond, 1991). Furthermore, thiscellulose is highly crystalline in structure (Gillces et al., 1992).1.2 CellulasesMany microorganisms produce enzymes capable hydrolyzing 3-1,4-glucosidic bonds(Beguin et at., 1987), but the mechanism by which cellulases degrade cellulose to glucose isnot well understood. More than 40 years ago, Reese (1950) proposed that the degradation ofcellulose involves at least two steps. In the first step, C1, native cellulose is converted to aform that is more accessible to attack by a hydrolytic enzyme in the second step, C,. Morerecently, a general model involving several enzymes acting synergistically in the degradationof cellulose has been developed (Beguin et at., 1987), as shown in Figure 1.1. In the first2step in this model, endoglucanases cleave the cellulose chain at the amorphous regions of thefibrils. Subsequently, exoglucanases remove cellobiose molecules from the newly producednon-reducing ends of the cellulose molecules. Finally, 3-glucosidases cleave the cellobiosemolecules, producing glucose as the product. Although recent data suggests that this model isover-simplistic, it is still thought to be essentially correct (Beguin et a!., 1987).Many cellulases are capable of adsorbing very tightly to cellulose (Gilkes et at.,1991b; Klyosov, 1990), but the mechanism of adsorption is not known. The crystal structuresof several periplasmic sugar binding proteins from gram-negative bacteria, includingreceptors for L-arabinose, D-ribose, D-glucose/D-galactose, and maltose, have beendetermined to high resolution (Vyas eta!., 1991; Spurlino eta!., 1991). Although theirprimary sequences show little homology, the tertiary structures of the proteins are similar.Furthermore, the geometry of the ligand-binding sites and the residues actually involved inbinding are highly conserved (Vyas eta!., 1991). The primary mode of ligand binding inthese proteins is by hydrogen-bonding, but there are also van der Waals contacts formed bythe stacking of aromatic side chains against the pyranose ring of the sugars (Vyas et a!.,1991). Similar types of interactions are believed to occur in cellulase-cellulose complexes(Gilkes eta!., 1991).1.3 CenA from Ce!lulomonasfimiThe gram positive bacterium Cel!ulomonasfimi produces a system of cellulases whengrown on microcrystalline cellulose (Beguin et a!., 1977). To date, the genes encoding fourendoglucanases: CenA, CenB, CenC (Gilkes et a!., 1984), and CenD (A. Meinke,unpublished) and one exoglucanase: Cex (Whittle et a!., 1982) have been cloned in E. coli.CenA is a 418 amino acid polypeptide composed of two discrete functional domainsseparated by a linker sequence (Figure 2.1). The carboxyl-terminal domain of 284 residuescontains the active site of the enzyme and is thus termed the ‘catalytic domain’ (Gilkes et a!.,31988). The linker sequence, termed the ‘Pro-Thr linker’ or ‘PT box’ is 23 amino acids inlength and consists of only proline and threonine residues (Figure 1.2 A). The amino-terminal domain is 111 residues in length and is responsible for the tight adsorption of theenzyme to cellulose. As a result, it is referred to as the ‘cellulose binding domain’, or CBD.Many cellulases have discrete CBDs that retain their function when they are removedby proteolysis and isolated (Gillces et at., 1991a). The known CBDs from bacterial cellulasesshare several common characteristics including low charge density, high contents ofhydroxyamino acids, and conserved glycine, asparagine, and tryptophan residues (Gilkes eta!., 1991a). The propensity of CBDs in cellulases from many different organism suggeststhat binding domains are important for cellulose degradation (Din et at., 1991). However, theprecise role of the binding domain in the hydrolysis of cellulose is not known. Upon removalof the CBD from CenA by proteolysis, the activity of the enzyme is reduced on highlycrystalline substrates, but increased on both amorphous and soluble substrates (Gilkes et aL,1988; Gilkes et at., 1992). Forsome cellulase systems, tight adsorption of endoglucanasesonto cellulose is believed to be important for the ability of the cellulases to degrade cellulose(Klyosov, 1990). Stâhlberg et at. (1991) proposed that the function of binding domains intwo-domain cellulases is to promote hydrolysis of crystalline cellulose by adsorbing theenzymes to the surface of the cellulose and increasing the local concentration of catalyticdomains. However, recent data suggests that the binding domain of CenA may play more ofan active role in the hydrolysis of cellulose than simply adsorbing the enzyme to the substrate(Din et al., 1991).Little is known about the structure of CenA. The model obtained for the protein bysmall-angle X-ray scattering (SAXS) analysis (Figure 1.1.A) suggests that CenA is tadpole-shaped (Pilz et a!., 1990). The catalytic domain comprises the globular ‘heads of the tadpole,while the extended ‘tail’ region corresponds to the CBD. Similar models were derived for twoendoglucanases from the fungus Trichoderma reesei (Abuja et at., 1988a; Abuja et at.1988b). Unlike the T. Reesel enzymes, the longitudinal axes of the two domains of CenA are4predicted to be at an angle of 135° with respect to each other (Pilz eta!., 1990). Thefunctional significance of this conformation remains to be determined. The current model ofthe structure of the CBD to explain its elongated conformation is that it forms a ‘hairpin’structure that is stabilized by a disulfide bond (Shen eta!., 1991). A hypothetical model ofthe enzyme is shown in Figure 1.2.C.1.4 ObjectivesThe role of the cellulose binding domain of CenA in the hydrolysis of cellulose is notwell understood, but current evidence suggests that the CBD performs a more active functionthan simply adsorbing the enzyme to cellulose. The objective of this study was to investigatefurther the role of the CBD by constructing a series of derivatives of CenA and of CBDppjthat have tandemly repeated binding domains. The specific questions addressed were: (1) cantwo binding domains function in tandem to increase the overall adsorption affinity of CenAfor crystalline cellulose, and: (2) does the adsorption affinity of CenA influence its catalyticproperties on crystalline and soluble substrates.5Crystalline regionAmorphous regionasorpLIon or CeItuIase enzymesendoqlucanaseQexoglucnaseendogluca,e-8-glucosidse0Figure 1.1 Synergistic model of cellulose hydrolysis by cellulases. Shaded hexagonsrepresent the reducing end of the cellulose chain. (from Beguin et at., 1987)Figure 1.2 Structural features of CenA.A. Amino acid sequence of the Pro-Thr linker and flanking regions. *: start and endof the linker.B. Model of CenA derived from small-angle X-ray scattering analysis. (adaptedfrom Pilz et at., 1990)C. Hypothetical model of CenA. The shaded area corresponds to the catalyticdomain and the thick line represents the Pro-Tbx linker (adapted from Shen eta!., 1991)6A.catalyticCBD * * domainGTVPTFSPTPTP1YITFrPTPTFFPWFFrVTPQPTSG------..—--.----B.C.72. Materials and Methods2.1 Bacterial strains, plasmids, and phageThe bacterial strains, plasmids, and phage used in this study are listed in Table 1.Bacterial stocks were maintained at -20°C in LB medium containing 20% glycerol and at-70°C in LB medium containing 10% DMSO.Table 2.1. Bacterial strains, phage, and plasmidsBacterial Strain Genotype ReferenceE. coli JM1O1 supE thi ?x(lac-proAB) [F traD36 Yannish-Perron etproAB lacIqZAMl5] al.,1985E. coli RZ1032 HfrKL16 P0/45 [lysA(61-62)] dutl Kunkel et a!., 1987ung 1 thu relA 1 Zbd-279: :Tn lOsupE44Plasmid Genetic Characters ReferencepTZ18R bla plac flori PharmaciapUC18-1.6cenA bla plac ApR Guo et al., 1988pUC18-CBD.PT bla plac ApR Din et a!., 1991Phage Genetic Character ReferenceM13K07 KmR Vieira and Messing,198782.2 Media and buffers.Luria-Bertani (LB) medium was as follows: 10 g/L tryptone, 5 g/L yeast extract,lOg/L NaC1, supplemented with 100 jig/mI ampicillin. M9 minimal medium contained 6 g/LNa2HPO4,3 g/L KH2PO4,0.5 g/L NaC1, 1 g/L NH4C1, and, following autoclaving, 100 juLof a sterile filtered 5% thiamine solution, 5 mi/L of a sterile filtered solution containing 1MCaC12 and 2M MgSO47HO,supplemented with 100 jig/mI ampicillin. 2xYT mediumcontained 16 g/L tryptone, 10 g/L yeast extract, and 5 g/L Nacl. Solid media consisted of LBwith 1.5% agar. Buffers and solutions used in this study were prepared as describedpreviously (Sambrook et a!., 1989).2.3. Recombinant DNA techniquesMost DNA manipulations were done as described previously (Sambrook et at., 1989).Restriction Endonucleases, 17 DNA polymerase, and T4 DNA ligase were obtained fromBethesda Research Laboratories (BRL), Pharmacia Inc., New England Biolabs (NEB), orBoehringer-Mannheim. All enzymes were used as directed by the manufacturers in buffersprovided. Small scale preparations of plasmid DNA were done by the alkaline-lysis method(Sambrook et at., 1989). Restriction fragments were electrophoresed in TAE buffer through0.8%, 1.0%, or 1.2% agarose gels containing 0.5 jig/mi ethidium bromide and visualizedunder ultraviolet light. Fragments to be isolated were excised from the gels and recoveredusing the GeneCleanTM or MerMaidl’M kits (BiolOl, La Jolla, CA) according to theprotocols provided.Site specific mutagenesis was done as described previously (Kunkel et a!., 1987).Oligodeoxyribonucleotides were chemically synthesized (Atkinson and Smith, 1984) at theUBC Oligonucleotide Synthesis Facility (Vancouver, B.C.) on an Applied Biosystems 380A9DNA synthesizer. They were purified by reverse-phase chromatography on Sep-Pak C-18 columns (Milipore).2.4. Sequencing of DNA.DNA was sequenced essentially according to the modified dideoxy chain terminationmethod described previously (Tabor and Richardson, 1987), with dGTP in the primerextension reaction mix substituted by 7-deaza GTP in order to reduce the frequency ofcompressions. Single stranded template DNA was produced from pTZ derived plasmids(Vieira and Messing, 1987), and was purified by the method of Kristensen et a!. (1987),modified as described by Trimbur et a!. (1992).2.5 Screening colonies for endoglucanase activityColonies of E. coli JM1O1 harboring recombinant plasmids encoding an activeendoglucanse (CenA or derivative) were picked onto LB agar plates containing 100, 100 mM IPTG and 1% high viscosity CMC. Following incubation overnight at37°C, the colonies were washed away with water, and the plates were stained for 15 minuteswith 0.2% Congo red and destained for 15 minutes with 1M NaCl. Colonies that hadexpressed an active endoglucanase were visualized as clear halos on a red background.2.6 Production and purification of proteins2.6.1 Growth of cells and preparation of cell extractsCultures of E. coli 3M 101 clones harboring recombinant plasmids encoding theproteins to be purified were grown in 20L of LB media supplemented with 100 .tg/ml10ampicillin at 37°C in a 35 litre fermenter (Chemap, Switzerland). When the cultures reachedan optical density of 2.0 A600 units, IPTG was added to a final concentration of 0.1 mM toinduce expression of the recombinant proteins. Between 2 and 3 hours following induction,after growth of the cultures had ceased, the cells were harvested by ultracentrifugation at45,000 RPM in a Sharples centrifuge. The cell paste was resuspended in 250 ml of 50 mMpotassium phosphate buffer, pH 7, containing 3 mM EDTA and 0.02% NaN3,centrifuged at6,000 RPM in a Beckman JA1O rotor at 4°C for 10 mm, and resuspended in an equal volumeof the same potassium phosphate buffer. The cells were disrupted by two passes through aFrench pressure cell. The protease inhibitor PMSF was added to the cell extracts to a finalconcentration of 1 mM. Streptomycin sulfate was added to 1.5% to precipitate the nucleicacids, and the cell extracts were incubated overnight at 4°C with slow stirring. The celldebris and the bulk of the precipitated nucleic acids were removed by centrifugation at15,000 RPM in a Beckman JA2O rotor, and the cell extracts were again incubated overnight at4°C and subsequently clarified by ultracentrifugation at 40,000 RPM in a Beckman Ti50rotor.2.6.2 Cellulose affinity chromatographyFor each protein to be purified, 500g of CF1’rM cellulose was washed several timeswith dH2O, resuspended in 50 mM potassium phosphate buffer, pH 7, and packed into aPharmacia XK 50/30 column. The column was attached to an FPLC system (Pharmacia) andequilibrated with 200 ml of 50 mM potassium phosphate buffer, pH 7. The clarified cellextract was passed over the column at a flow rate of 1.0 mI/mm, washed with 1L of 50 mMpotassium phosphate buffer, pH 7, containing 1M NaCl at 1.5 mI/mm followed by 500 ml of50 mM potassium phosphate buffer, pH 7, at 2.0 mI/mm. The proteins were eluted using alinear 0-8 M gradient of guanidinium hydrochloride in 50 mM potassium phosphate bufferover 100 ml at 1.0 mI/mm, followed by 500 ml of 8 M guanidinium hydrochloride in 50 mM11potassium phosphate buffer at 1.0 mI/mm. The washes and eluants were collected in 5 mlfractions. Peak fractions were identified by on-line absorbance readings at 280 nm and werepooled. The eluted proteins were then exchanged into 50 mM potassium phosphate buffer,pH 7, containing 0.02% NaN3 by ultrafiltration in a 350 ml AmicOn pressure cell throughpolyethersulfone membranes with lkD or lOkD cutoffs (Amicon PM-i or PM-lU), until theguanidinium hydrochloride concentration was calculated to be below 50p.M.2.6.3 Electrophoresis of proteinsPurified proteins were resolved by SDS-polyacrylimide gel electrophoresis(SDS-PAGE) using a Bio-Rad Mini-PROTEAN1II apparatus as described previously(Laemmli, 1970; Schagger and von Jagow, 1987). Protein bands were visualized by stainingwith Coomassie blue (Merril, 1990), and their relative masses (Mi) were estimated bycomparison to Sigma-SDS 6H Mr standards. The gels were air dried between two sheets ofacetate, and the bands were quantified using a densitometer (Molecular Dynamics[mageQuant version 3.0, Sunnyville, CA) in order to asses purity.2.7 Purification of radiolabeled proteins2.7.1 Growth of cellsFor each protein to be labeled and purified, single colonies of E. coli IM1O1 clonescontaining recombinant plasmids from overnight LB agar plates were picked into 5 ml of M9media containing 100 IWm1 ampicillin in test tubes, and these cultures were incubated at370C, 200 RPM for 8-12 hrs until turbid. From these cultures, 2.5 ml was used to inoculate250 ml of M9 media supplemented with 100ig/m1 ampicillin in 2L shake flasks, which were12incubated at 37°C, 200 RPM. When the cultures reached an optical density of 1.0 A280units, 625 tl of[14C(U)]-glycine (New England Nuclear) was added for a final concentrationof 0.17 jtg/ml glycine (0.25 p.Ci/ml), IPTG was added to a final concentration of 100p.M. andthe cultures were incubated a further 48 hrs. The cells were then centrifuged at 6,000 RPMfor 10 mm in a Beckman JA1O rotor, the supernatant was decanted to an Erlenmeyer flask,and NaN3 was added to a final concentration of 0.02%. The periplasmic proteins werereleased from the cells by an osmotic shock procedure essentially as described previously(Osborne and Munson, 1974). To the osmotic shock preparation was then added 1/10 of thevolume of 500 mM potassium phosphate buffer, pH 7, 0.2% NaN3.2.7.2 Batch purification of labeled proteinsApproximately 5 g of CF1TM cellulose was washed extensively and was added toboth the osmotic shock preparation and culture supernatants, and they were incubatedovernight at 4°C with stirring. The cellulose was then allowed to settle and was washed oncewith 50 mM potassium phosphate buffer, pH7, twice with 1M NaCl in 50 mM potassiumphosphate buffer, pH 7, and once with 50 mM potassium phosphate buffer, pH7. After thefinal wash was decanted, 100 ml of 8 M guanidinium hydrochloride in 50 mM potassiumphosphate buffer was added and was incubated overnight with stirring in order to elute theproteins from the cellulose. The cellulose slurry was then filtered through a Whatman GFICfilter, and the filtrate containing the eluted proteins was exhanged into 50 mM potassiumphosphate buffer and concentrated as described in section Autoradiography of labeled proteinsThe purified proteins were resolved by SDS-PAGE as described in section 2.6.3. Thegels were incubated overnight in 40% methanol! 10% acetic acid in order to fix the proteinbands, and the gels were dried onto Whatman #1 filter paper on a gel drier (Bio-Rad). Thedried gels were then used to expose X-ray photographic film (Kodak XAR 5) for 15 hrs, andthe film was developed using an automatic developer (Kodak M35A X-OMAT).2.8 Determination of protein concentrationThe concentrations of both the labeled and non-labeled purified proteins wereestimated by UV absorption at 280 nm. The extinction coefficients(E205’mg/mi) at 280 nmwere predicted using the theoretical method of Cantor and Sneyd (1980). The extinctioncoefficients were then estimated by the method of Scopes (1974), and in all cases theestimated values were found to be within 5% of the predicted values. For routinemeasurements, the predicted extinction coefficients were used.2.9 Cellulose adsorption assaysFor alladsorption studies, the cellulose substrate used was bacterial microcrystallinecellulose (BMCC) prepared by Emily Kwan (UBC Department of Microbiology) using themethod reported by Gillces et. a!. (1992). The high initial protein concentration bindingassays ([P]0 between 50 and 700 jig/mI) were carried out according to the protocol describedby Gilkes (1992) using purified, non-labeled proteins. Low initial protein concentrationbinding assays ([P10 between 10 and 50 jig/mI) done using purified, radiolabeled proteins bythe following method. To pre-silicanized, 1.7 ml microfuge tubes was added 1 ml of 3 mg!mlbovine serum albumin (BSA) in 50 mM potassium phosphate buffer, pH7, and one 3 mm14diameter glass bead. The tubes were placed in a rack and rotated slowly on a tube roller at30°C for 3 hrs in order to pre-coat the inside of the tubes and the glass bead with BSA. TheBSA solution was then removed. The proteins to be assayed were diluted in 50 mMpotassium phosphate buffer, pH7, to the appropriate concentrations in 675 and 75 tl of a1 mg/mi solution of BSA in 50 mM potassium phosphate buffer, pH7, was added to give afinal concentration of 50 jig/mi BSA. To each tube, 750 p.1 of a 2 mg/mi solution of BMCCin 50 mM potassium phosphate buffer, pH7, was added to give a final concentration of 1mg/mi BMCC. At each protein concentration, 4-6 replicates were used. The tubes wereincubated at 30°C for 30 mm on a tube roller. The BMCC was then pelleted bycentrifugation, and 1.0 ml of the supernatant was removed to scintillation vials. 10 ml ofAqua MixTM scintillation cocktail (ICN) was added to the vials, and the counts per minute(CPM) in each sample was determined using a Beckman LS-2000 liquid scintillation counter.The CPM per amount of protein (CPM/p.g) was determined for each purified labeled proteinby diluting a known amount of the protein in 50 mM potassium phosphate buffer, pH7, to afinal volume of 1.0 ml, adding 10 ml of scintillate, and determining the CPM of the sample.This value was then used to calculate the concentrations of the free protein in the bindingassay supernatants. Blanks, consisting of 1.0 ml of 50 mM potassium phosphate buffer, pH7,added to 10 ml of scintillate, were prepared in triplicate and counted along with each set ofsamples, and the mean CPM of the blanks was subtracted from the CPM of each sample.2.10 Enzymatic activity assaysThe specific activities of several of the non-labeled proteins were determined onCarboxymethylcellulose (CMC), 2,4 dinitrophenylcellobioside(dNPC), and BMCC. TheCMCase activities were determined as described by Gilkes et. a!. (1988). Activity on dNPC(prepared by Dedreia Tull, UBC Department of Chemistry), was determined as follows.Enzymes were diluted to 25 jig/mi in 50 mM potassium phosphate buffer, pH 7, with 0.215mg/mi BSA and were pre-equilibrated to 370C. Substrate was diluted to variousconcentrations ranging from 30 to 1200 .tM in the same potassium phosphate buffer andpre-equilibrated to 37°C. 100 il of the substrate was added to 20 il of the diluted enzyme,and 100 .tl of the assay mix was immediately pipetted into spectrophotometer cells that hadbeen pre-equilibrated to 37°C. The rates of hydrolysis of the dNPC were followed by thechange in absorbance at 400 nm over 1 mm using an Hitachi U-2000 spectrophotometer. Theinitial velocities obtained at each initial substrate concentration were plotted on a Hanes plot,and the Km, and Kcat for each enzyme were calculated.The activity of the enzymes on the insoluble substrate BMCC was determined bydiluting 1 nmol of each protein in 750 jil of 50 mM sodium citrate buffer, pH 6.8, whichincluded 0.2 mg/mI BSA and 0.02% NaN3,then adding 750 jil of a 2 mg/mi solution ofBMCC in the same sodium citrate buffer. The assay mix was put into pre-silicanized glassscintilation vials, and the vials were incubated for 15 hrs at 37°C, 300 RPM (New BrunswickModel G25 shaker incubator). The BMCC was then removed by centrifugation, and thesupernatant was assayed for reducing sugars by the hydroxybenzoic acid hydrazide (HBAH)method (Langsford et. at., 1987).2.11 Fragmentation of cotton fibres and release of small particlesCarded cotton fibres were dewaxed (Wood, 1988) with the assistance of Dedreia Tull(UBC Department of Chemistry). The disruption of these fibres following incubation withthe cellulose binding domain of CenA and derivatives prepared for this thesis was monitoredby the release of small particles (Halliwell, 1965) by the protocol described by Din et. a!.(1991).163. Results3.1 Construction of plasmids encoding double binding domain derivatives of CenA andCBDCenA.Derivatives of both CenA and CBDCenA encoding tandemly repeated bindingdomains were constructed. In addition to derivatives that included a full-length, 23 residuePro-Thr linker separating the duplicated binding domains, derivatives of both polypeptidesbearing a shortened, 13 residue linker (the ‘½ Pro-Thr linker’) were also constructed. Blockdiagrams of the double binding domain proteins used in this study are shown in Figure Site directed mutagenesisUnique NheI restriction sites were incorporated at various locations in the wild typeCenA coding sequence by site directed mutagenesis (SDM). To facilitate this process, the1.6 kb EcoRl-HindilI fragment containing the coding sequence of CenA and the 3’ flankingregion was subcloned from pUC18-1.6CenA to the phagemid pTZ18R, resulting in theplasmid pTZ18R-CenAAs shown in Figure 3.2, an NheI site was introduced at the junction of the sequencesencoding the leader peptide and the CBD. The oligonucleotide primer DN2 was used toinsert three nucleotides encoding a serine residue between the coding sequences for Ala-96and Pro-97 creating an NheI site at the 5’ end of the coding sequence for the CBD andresulting in the plasmidpTZ18R-CenAPTNhe. Primer DN3 was used to add an NheI site atthe junction of the coding sequences for the Pro-Thr linker and the CenA catalytic domain(Figure 3.3), resulting in the plasmid pTZ18R-CenALNhe. In the third mutagenesis (Figure3.4), the oligonucleotide primer DN1 was used. This primer was originally designed to add17an NheI site at the 3’ end of the Pro-Thr linker, but it was found that the twelve nucleotides 5’from the NheI site hybridized to a complementary twelve nucleotides on the template at aregion thirty nucleotides upstream from the intended hybridization site. This resulted in thelooping-out of the thirty nucleotides at the 3’ end of the sixty-nine nucleotide Pro-Thr linkercoding sequence, creating a sequence that encodes the thirteen residue, ‘½ Pro-Tb linker’, andresulting in the plasmid pTZ18R-CenAl.2. This construct was retained and later used toinvestigate the effect of a shortened linker on the adsorption of CenA and double bindingdomain derivatives.Following each site directed mutagenesis, the reaction mixture was used to transformE.coli JM1O1. Colonies were screened for endoglucanase activity on CMC plates, andplasmid preparations were made from positive clones. The plasmids were checked for thepresence of an NheI site, and, in each case, several clones with added NheI sites were retainedfor sequencing. In order to avoid second-site mutations, a 0.6 kb fragment including themutated region was subcloned out of the mutated plasmid and into wild-type pTZ18R-CenA.The sequence of the EcoRI-MluI region was then verified.3.1.2 Construction of plasmidsThe scheme for the construction of plasmids expressing CenADBD, CenA DBD1.2,2xCBDpt, and 2xCBDptl.2 are given in Figures 3.5, 3.6, 3.7, and 3.8 respectively. Theconstruction 2xCBDpt and 2xCBDptl.2 involved an intermediate construct, CBDptNhe,which was made as shown in Figure 3.9. The constructs were confirmed to be correct byrestriction endonuclease digestion.183.2 Purification of proteins by affinity chromatography on celluloseAll four double binding domain proteins, both radiolabeled and non-labeled, werepurified by affinity chromatography on cellulose as described in Materials and Methods.Coomassie-stained gels of purified, non-labeled CenADBD and CenADBD1.2 are shown inFigure 3.10. These proteins were estimated to be >90% pure by densitometry of these gels.Wild-type CenA and CenAPTNhe, expressed from plasmids pTZ18R-CenA andpTZ18R-CenAPTNhe respectively, were also purified for use as controls in the kinetic andadsorption experiments, and are shown in the same figure. Autoradiograms of the purified,radiolabeled proteins are shown in Figure 3.11. Coomassie-stained gels of 2xCBDpt,2xCBDptl.2 (estimated >80% pure), and CBDpt are shown in Figure 3.12, andautoradiograms of the labeled proteins are shown in Figure 3.13. CenA 1.2, expressed fromplasmid pTZ18R-CenAl.2, was also labeled and purified, and is shown along with CenA inFigure 3.13(D). CenAAPT expressed from plasmid pUC18-1 .6cenAp0’(Shen eta!.,1991) was also labeled and purified (not shown). This protein is CenA in which the Pro-Thrlinker is replaced by a Val and an Asn residue that make up a HpaI restriction site in theconstruct (Shen et at., 1991). Western blots of the purified, non-labeled proteins are shownin Figure 3.14. These westerns show that some degradation products exist in thepreparations. The preparations of CenA and CenAPTNhe (A, lanes 1 and 2) have bands atabout the same Mr as CBD.PT (B, lane 2), and CenA DBD and CenADBD1.2 (A, lanes 3and 4) have contaminating bands with about the same Mrs as 2xCBDpt and 2xCBDptl.2 (B,lanes 3 and 4), respectively. Degradation products in the preparation of 2xCBDptl.2 do notseem to be present, although the coomasie-stained gel (Figure 3.12.B) shows somecontaminating bands.193.3 Adsorption assaysThe adsorption parameters of cellulases on cellulose are usually analyzed by fittingbinding data to the Langmuir adsorption isotherm (see Stuart and Ristroph, 1985; Steiner eta!., 1988 for reviews). This isotherm models the case in which a ligand binds to a singlebinding site with no negative or positive cooperation or binding site exclusion. Gilkes et a!.(1992) proposed a modified Langmuir-based model for the adsorption of cellulases tocellulose, which is briefly summarized as follows. Cellulases are assumed to bind to one ormore of the repeating cellobiosyl units on the surface of crystalline cellulose (Henrissat et a!.,1988). Since the dimensions of cellulases and their isolated binding domains exceed thedimensions of one cellobiosyl unit, a cellulose-binding protein bound on the cellulose surfacewill cover several cellobiosyl units at once. As a result, the surface of cellulose must beviewed as a lattice of overlapping potential binding sites, and the Langmuir assumption ofdiscrete binding sites is not valid. In the model, the concentration of free binding sites at anydegree of saturation must be determined by a probability function that accounts for not onlythe concentration of bound ligands, but also the configuration of the ligands on the cellulosesurface (Gilkes et a!., 1992). This complication can be avoided if binding studies are doneonly at very low protein concentrations, were the probability of any two ligands binding closeenough to each other to exclude any cellobiosyl units from being potential binding sites isinsignificant. The equation for the modified isotherm is as follows:[N0] Ka [ii[B]=— (1)1 + a Ka [E9Where [B] is the concentration of bound ligand following equilibrium binding(.tmol ligand/mg cellulose), [N0] is the concentration of potential binding sites (jimol bindingsite/mg cellulose), Ka is the equilibrium association constant (LIp.mol), [F] is the20concentration of free ligand in solution (tM), and ‘a’ is the number of lattice units that theprotein ‘covers’ when it is bound on the cellulose surface (jimol lattice units4tmol ligand).Equation (1) is converted to linear form by rearranging it into double reciprocal form asfollows:1 1 1 a— =____— + (2)[B] Ka [N0] [F] [N0]When l/[B] is plotted versus 1/[Fj, the reciprocal of the slope of the line gives Ka[Noj, whichis defined as Kr, or the relative affinity constant. [N0] is not known, so the absolute value forKa cannot be determined; however, since [N0] is assumed to be constant for a givenpreparation of cellulose, Kr is a useful parameter to compare the affinities of differentproteins relative to each other on the same preparation of cellulose (Gilkes et at. 1992).The y-intercept (aI[N01) of the double reciprocal plot is also potentially useful,provided that it can be determined to a reasonable degree of certainty. I will define this termas the ‘relative a-value’, or ‘ar’. Again, the absolute a-value cannot be determined preciselysince [N0] is unknown. However, ar could be a particularly interesting parameter for theproteins examitied in this study because, by comparing the results for a single binding domainprotein and a double binding domain derivative, some insight may be gained into the way inwhich the repeated binding domain is binding to the surface of the cellulose. Finally, ifAm (tmol protein/g cellulose) is defined as the theoretical maximum number of proteinsthat could be bound onto the cellulose surface at once in a monolayer given closest possiblepacking (i.e., the theoretical saturation level for a given preparation of cellulose), then,[N0]Ay a (3)21Substitution of Equation (3) into the term for the y-intercept from Equation (2) yields(a/{Am a)) for the intercept, which reduces to l/Am. Thus, by taking the reciprocalof the value obtained for the y-intercept, Am can be determined.3.3.1 High protein concentration adsorption assaysAdsorption studies at high initial protein concentrations (between 50 .tg/ml and700 tg/ml) were done using purified, non-labeled proteins. Concentrations of free proteinfollowing equilibrium binding were measured by absorbance at 280 nm. The adsorptionisotherms for CenADBD, CenADBD1.2, and CenA are shown in Figure 3.15, and theadsorption isotherms for 2xCBDpt, 2xCBDptl.2, and CBD.PT are shown in Figure 3.16.As was the case for CenA and CBD.PT (Offices et a!., 1992), the double binding domainderivatives approached but did not reach saturation of the BMCC at the highest proteinconcentrations used. CenA and CBD.PT approached saturation levels that were greater thanabout 7 .tmo1Jmg cellulose and about 8 j.tmollmg cellulose, respectively (Gilkes et al., 1992,Figure 4). CenADBD and CenADBD1.2 approached saturation levels that were greater thanat 4.5 imol/mg cellulose, while for 2xCBDpt and 2xCBDptl.2 the saturation levelsapproached were greater than 6.7 .tmol/mg cellulose.3.3.2 Low initial protein concentration binding assaysIn order to obtain Kr values for these proteins, adsorption studies were done at lowinitial protein concentrations (between 5 and 50 tg/m1) using 14C- labeled proteins. In theseassays, concentrations of free protein were determined by scintillation counting as describedin Materials and Methods. Initially, these studies were attempted using the same bindingassay as was used at the high initial protein concentrations. It was found that, for most of theproteins studied, when the adsorption data was fitted to the modified Langmuir isotherm,22double-reciprocal plots of bound versus free protein tended to appear straight at the higherprotein concentrations used, but curved at the very low initial protein concentrations. Thisresult would normally indicate negative cooperativity, or that more than one type of bindinginteraction was occurring.It is possible that the phenomenon could also be due to the proteins denaturing at thevery low concentrations used, aggregating, precipitating and coming out of suspension alongwith the cellulose during the centrifugation steps of the binding assay. In order to test thishypothesis, bovine serum albumin (BSA) was added to the adsorption assay mix at variousconcentrations. It was supposed that if protein denaturation was occurring, the BSA wouldstabilize the proteins in solution and the slopes of the double-reciprocal plots would be linearover the range of concentrations used. The results are shown in Figure 3.17. From thesedata, two obvious conclusions can be drawn. First, the presence of BSA in the assay mix atthe concentrations used greater than 20 tgJml straightens out the double reciprocal plot overthe entire range of initial protein concentrations used. Second, with increasing concentrationsof BSA, the slope of the line increases and the relative affinity constant calculated from thisslope decreases. It is not clear from these data, though, whether the BSA actually affects thestability of the protein, or whether it simply interferes with the interaction of the protein withthe cellulose and, as a result, decreases the observed relative affinity. In order to address thisquestion, a set of control experiments was done in which 750 .t1 of phosphate buffer wasadded to the assay mix in place of the BMCC (see materials and methods), and theconcentration of BSA in the assay mix was varied. In this experiment, the initial proteinconcentration was kept constant at 1.89 g/m1. This very low initial concentration waschosen because it is typical of the concentration of free protein following equilibrium bindingin these low concentration adsorption assays. The data are summarized in Figure 3.18. Thepresence of BSA did appear to have the effect of stabilizing the protein in solution, but, evenat the highest BSA concentration used, a significant portion of the labeled protein was stilllost. In order to locate the lost protein, the assay supernatant was removed from the tubes to23which no BSA had been added, and the tubes were dried overnight. The radioactivity in thetubes was determined and compared to standards that consisted of tubes to which 1.89 jig oflabeled protein was added directly. Approximately 80% of the lost protein was found driedon the tubes (data not shown). In order to minimize the loss of labeled protein, the no-BMCC‘adsorption’ assay was used to test various conditions, including (1) pre-coating the tubeswith BSA (see Materials and Methods), (2) using silicanized tubes, and (3) including BSA inthe assay mix at 50 jig/mi (data not shown). By using silicanized tubes, pre-coating the tubeswith BSA and including BSA at 50 jig/mi, the loss of labeled protein in the assay wasreduced to about 4.8% of the initial 1.89 jig/mi. Finally, the binding time was reduced from18 hrs to 30 mm, which was found to be adequate to establish adsorption equilibrium (datanot shown), and the loss was further reduced to about 3.1%. This was deemed to be anacceptable loss, and these conditions were subsequently used to obtain Kr values for theproteins used in this study.The low-concentration adsorption isotherms for CenA and CBD.PT are shown inFigure 3.19. The isotherms for CenADBD, CenADBDI.2, 2xCBDpt, and 2xCBDptl.2 areshown in Figures 3.20.A, 3.20.B, 3.21.A, and 3.21.B, respectively. The adsorption isothermsfor CenAl.2 and CenAAPT are shown in Figure 3.22. A doubly weighted least squaresanalysis was used to obtain the equations of the regression lines. The constants derived fromthe various isotherms are summarized in Table 3.1. In all cases, no-BMCC controls weredone at various initial concentrations. The loss of protein in these controls usually rangedfrom 0 to about 5 % of the initial concentration, and never exceeded 10 % of the initialprotein (data not shown). An adsorption assay was also attempted for CenAPTNhe as acontrol to ensure that the added alanine and serine residues that comprise the NheI site in thedouble binding domain proteins do not influence binding. It was found, however, that theloss of protein in the no-BMCC controls was greater than 63% of the initial concentrationover the whole range of initial concentrations used (data not shown). Thus, this protein wasassumed to be unstable under the conditions of the adsorption assay. It is interesting to note24that CenAl.2, which also has an added alanine and serine residue but differs fromCenAPTNhe in that it has a 1/2 Pro-Thr linker, was stable in the adsorption assay.Table 3.1 Adsorption parameters from double reciprocal plotsProtein Kr(L/g cellulose)CenA 27.5± 1.4CBD.PT 11.5 ±2.1CenADBD 80.6 ± 4.6CenADBD1.2 2.46 ± 0.232xCBDpt 0.65 ± 0.032xCBDptl.2 3.04 ± 0.34CenAzPT 1.07 ± 0.09CenAl.2 5.88 ± 0.06note: errors reported are 95% confidence limitsar(g cellulose/mol)0.3 15 ± 0.0630.210 ± 0.0480.309 ± 0.0800.274 ± 0.2180.269 ± 0.1510.200 ± 0.1270.198 ± 0.2180.100 ± 0.164Amax(moW g cellulose)3.17 ± 0.634.76 ± 1.093.24 ± 0.843.65 ± 2.903.72 ± 2.095.00 ± 3.185.05 ± 5.5610.0 ± 16.43.4 Enzymatic activitiesSeveral previous studies, mainly using crude cellulase complex preparations, havesuggested a direct correlation between adsorption affinity of cellulases and their ability tohydrolyze insoluble, crystalline cellulose (see Klyosov, 1990 for review). These studies ledKlyosov et a!. (1986) to propose the rule: ‘the better the adsorption, the better the catalysis’.In order to test the application of the rule to purified endoglucanases, Klyosov et a!. (1990)examined the hydrolysis activity of several purified ‘multiple forms’ of endoglucanases fromTrichoderma reesei that differed in adsorption affinity and found that increased ability to -adsorb to cellulose paralleled increased hydrolytic activity. Furthermore, the isolatedcatalytic domain of CenA, which has negligible adsorption affinity for BMCC, was shown tobe less active than the intact enzyme on BMCC (Gilkes et a!., 1992). Some exceptions to the25rule have been noted between endoglucanases from different organisms (Klyosov et a!.,1990). In this study, the correlation between affinity and activity was tested by determiningthe specific activities of CenADBD and CenADBD1.2 on both BMCC and soluble substratesand comparing these activities to those of CenA. These three proteins constitute a set ofderivatives of CenA which differ widely in their adsorption affinity for crystalline cellulose(see previous section). CenAPTNhe was also used as a control to ensure that the addedalanine and serine residues constituting the NheI site in the double binding domain proteinsdo not have an effect on activity. The results are summarized in Table 3.2. No significantdifferences were observed in the specific activities of the four proteins on any of thesubstrates tested.Table 3.2 Specific activities of CenA and derivatives on various substratesProtein CMC dNPC BMCC(U/.tmol)a (U/pmo1)b (U/mmol)aCenA 5.97±0.65 16.7 287 ±43CenADBD 6.65 ± 0.46 19.5 277 ±57CenADBD1.2 6.74 ± 0.88 19.9 282 ±23CenAPTNhe 5.93 ±0.74 19.0 325 ±39a. U: units, tmol glucose equivalents produced/mmb. U: units, .tmol cli-nitrophenyl produced/mmnote: errors reported are 95% confidence limits263.5 Small particle release from cotton fibresDin et at. (1991) showed previously that the treatment of dewaxed cotton fibres withCBD.PT led to the release of small particles. In this study, the small particle release fromthese fibres following treatment with 2xCBDpt and 2xCBDpt was examined and compared tothe results of CBD.PT. The results are shown in Figure 3.23. The three proteins were used inequivalent molar concentrations (10 tM). CBD.PT was also used in double the molarconcentration of 2xCBDpt and 2xCBDptl.2 (20 pM) for an equivalent concentration ofbinding domains. A reaction mix consisting of the protein in the appropriate concentrationand 25 mg of dewaxed cotton were used in a total volume of 5 ml of potassium phosphatebuffer. As was shown by Din et al. (1991), CBD.PT released small particles, and the effectwas concentration-dependent. The small particle release by both of the double bindingdomain derivatives was comparable, and each produced more small particles than didCBD.PT at both concentration.27CenA(43.8 kDa)CenA DBD(57.6 kDa)CenA DBD1.2(56.6 kDa)1 31 143 166CBD frsT I CenA cat4491 584CBD IP/TI CBD II’’TI CenA cat1/2 Pro-Thrlinker 554CBD CBD II’/TI CenA cat2xCBDpt(2.7 kDa) I P/TI300CBD I P/TJ1/2 Pro-ThrlinkerTLFigure 3.1 Block diagrams of CenA and derivatives. CBD: cellulose binding domain ofCenA. PT box: Pro-Thr linker. CenA cat: catalytic domain of CenA. The cross-hatched boxrepresents the leader peptide of CenA. Numbers refer to amino acid residues.2xCBDptl .2(2.6 kDa)BD CBD27028Primer DN2: NheI5 -pGCCGCGCGGCGGCTAGCCCCGGC’IGCCGCGTC--3’+ single stranded pTZ1SR-CenA DNAAnneal, extend primer, ligate,transform E.coli JMlOlpTZ1 8R-CenALNhe:96T A A Q A*A S P G C R ProThrleader linker5’ ACC GCC GCG CAG GCG GCT AGC CCC GGC TGC CGC GTC 3’Nhe IFigure 3.2 Addition of an NheI site at the junction of the leader peptide and the CBDnAcoding sequences. Site directed mutagenesis was done on this template using primer DN2resulting in the plasmid pTZ18R-CenALNhe as shown. The nucleotide sequence of thecoding strand in the region of the alteration is shown with the amino acid sequence of theprotein product above it. Numbers shown above the sequence refer to amino acid positions.Added nucleotides and amino acids are shown in bold face. * : the usual site of cleaveage ofthe CenA leader peptide, between Ala-95 and Ala-96.29Primer DN3: NheI5’ -pCCCCACCCCCACGCCGACGGCTAGDGTCACGCCGCAGCCG-3’÷ single stranded pTZ18R-CenA DNAAnneal, extend primer, ligate,transform E.coli JM1O1pTZ18R-CenAPTNhe:165Pro-Thr T P T P T A S V T p Q p catalyticlinker domain5’ ACC CCC ACG CCG ACG GCT AGC GTC ACG CCG CAG CCG 3 INheIFigure 3.3 Addition of an NheI site at the junction of the Pro-Thr linker and the CenAcatalytic domain coding sequences. Site directed mutagenesis was done on this templateusing primer DN3 resulting in the plasmid pTZ18R-CenAPTNhe as shown. The nucleotidesequence of the coding strand in the region of the alteration is shown with the amino acidsequence of the protein product above it. Numbers shown above the sequence refer to aminoacid positions. Added nucleotides and amino acids are shown in bold face.30Primer DN1: N1-iel5 ‘ —PACCCCCACGCCGACGGCTAGCGTCACGCCGCAGCCG-3’+ single stranded pTZ18R-CenA DNAAnneal, extend primer, ligate,transform E.coli JM1O1pTZ18R-CenAl .2:155Pro-Thr T P T P T A S V T p Q p catalyticlinker domain5’ ACC CCC ACG CCG ACG GCT AGC GTC ACG CCG CAG CCG 3’NheIDeleted sequence:P T P T P T P T P T5’- CCG ACC CCG ACC CCC ACC CCC ACG CCG ACG -3’Figure 3.4 Deletion of 30 nucleotides of the 3’ end of the coding region of the Pro-Thrlinker of CenA, and the addition of an NheI site at the junction of the resulting ‘½ Pro-Thrlinker’ and the CenA catalytic domain coding sequences. Site directed mutagenesis was doneon this template using primer DN3 resulting in the plasmid pTZ18R-CenAPTNhe as shown.The nucleotide sequences of the coding strand in the region of the alteration and of thedeleted segment are shown with the amino acid sequence of the protein product above them.Numbers shown above the sequence refer to amino acid positions. Added nucleotides andamino acids are shown in bold face.31Figure 3.5 Construction of pTZ1 8R-CenADBD. pTZ1 8R-CenAPTNhe was digestedcompletely with NheI and HindIll, and the 3325 bp fragment containing the pTZ18R vectorsequence plus the CBDCenA coding sequence including the whole Pro-Thr linker sequencewas isolated. pTZ18R-CenALNhe was also digested completely with NheI and HindIll, andthe 1534 bp fragment containing the CenA coding sequence minus the leader sequence wasisolated. The two fragments were ligated together to give pTZ18R-CenADBD.Cut with Nhel and HindillIsolate 3325 bp fragmentII ICut with NheI and HindlIlIsolate 1534 bp fragmentNheI Hindill32HindIll NheIJ•.Ligate33Figure 3.6 Construction of pTZ18R-CenADBD1.2. pTZ18R-CenAl.2 was digestedcompletely with NheI and HindIll, and the 3295 bp fragment containing the pTZ18R vectorsequence plus the CBDCenA coding sequence including the ½ Pro-Thr linker sequence wasisolated. pTZ18R-CenALNhe was also digested completely with NheI and HindIll, and the1534 bp fragment containing the CenA coding sequence minus the leader sequence wasisolated. The two fragments were ligated together to give pTZ1 8R-CenADBD 1.2.34Cut with Nhel and HindillIsolate 3295 bp fragmentHindlil NheICut with Nhel and HindlilIsolate 1534 bp fragmentNheI Hindill•. Ligatenke35Figure 3.7 Construction of pTZ18R-2xCBDpt. pTZ18R-CenAPTNhe was digestedcompletely with NheI and HindIll, and the 3325 bp fragment containing the pTZ18R vectorsequence plus the CBDCenA coding sequence including the Pro-Thr linker sequence wasisolated. pTZ18R-CBDptNhe was also digested completely with NheI and Hindu, and the452 bp fragment containing the CBD.PT coding sequence minus the leader sequence wasisolated. The two fragments were ligated together to give pTZ18R-2xCBDpt.36Cut with NheI and HindlilIsolate 3325 bp fragmentHindill NhelCut with Nhel and HindlilIsolate 452 bp fragmentNheI HindlilLigate37Figure 3.8 Construction of pTZ18R-2xCBDptl.2. pTZ18R-CenAl.2 was digestedcompletely with NheI and HindIll, and the 3295 bp fragment containing the pTZ18R vectorsequence plus the CBDCenA coding sequence including the ½ Pro-Thr linker sequence wasisolated. pTZ18R-CBDptNhe was also digested completely with NheI and HindIll, and the452 bp fragment containing the CBD.PT coding sequence minus the leader sequence wasisolated. The two fragments were ligated together to give pTZ18R-2xCBDptl.2.381/2 Pro-ThrCut with Nhel and HindillIsolate 3295 bp fragmentHindlil NhelCut with NheI and HindillIsolate 452 bp fragmentNheI HindlilLigate39Figure 3.9 Construction of pTZ18R-CBDptNhe. pUC18-CBD.PT was digestedcompletely with BaniHI, and the 262 bp fragment containing the 3’ end of the CBD.PTcoding sequence was isolated. pTZ18R-CenALNhe was also digested completely withBamHI, and the 3107 bp fragment containing the pTZ18R vector sequence plus the 51 end ofthe CBD.PT coding sequenceincluding the NheI site at the 3’ end of the CenA leader sequncewas isolated. The 262 bp and 3107 bp fragments were ligated together to givepTZl 8R-CBDptNhe.Cut with BamHIIsolate 262 bp fragmentBamHI BamHlI IBamHlCut with BamHlIsolate 3107 bp fragmentI40r InkerBamHILigatelinker41A.‘C.kDa 1 2208..97.4_kDa I123B.D.kDa1 2kDa 1 297.4__67— -—29Figure 3.10 SDS-Page analysis of purified, non-labeled derivatives of CenAA. Purified CenADBDlane 1: Mr standards (2 p.g in each band)lane 2: 20 .tg of purified CenADBDB. Purified CenADBD1.2lane 1: Mr standards (2 p.g in each band)lane 2: 20 ig of purified CenADBD1.2C. Purified CenAlane 1: Mr standards (2 .tg in each band)lane 2: 20 ig of purified CenAD. Purified CenAptNhelane 1: Mr standards (2 ig in each band)lane 2: 20 jig of purified CenAptNhe42c.B.Figure 3.11 Autoradiograms of purified, 14C-labeled derivatives of CenAA. Purified CenADBDlane 1:’4C-Mr standards,0.05 mCi (BRL, Bethesda, MD.)lane 2: 9.5 tg of purified CenADBD, 902 cpm/p.gB. Purified CenADBD1.2lane 1:’4C-Mr standards,0.05 mCi (BRL, Bethesda, MD.)lane 2:2.2 .tg of purified CenADBD1.2, 1029 cpm/J.LgC. Purified CenAlane 1: 14C-M standards,0.05 mCi (BRL, Bethesda, MD.)lane 2: 3.8 p.g of purified CenA, 891 cpmlj.tgD. Purified CenAptNhelane 1:14C-Mr standards,0.05 mCi (BRL, Bethesda, MD.)lane 2: 7.1 p.g of purified CenAptNhe, 380 cpm4tgA.D.43A. B.kDal 2206.-12397.467...57.5.-53—44-C41—3629—CkDaj 2208.- -123-575— -36.-29.-Figure 3.12 SDS-Page analysis of purified, non-labeled derivatives of CBD.PTA. Purified 2xCBDptlane 1: Mr standards (2 ig in each band)lane 2: 20 jig of purified 2xCBDptB. Purified 2xCBDptl.2lane 1: Mr standards (2 pg in each band)lane 2: 20 of purified 2xCBDptl.2C. Purified CBD.PTlane 1: Mr standards (2 p.g in each band)lane 2: 20 jig of purified CBD.PT44A. B.kDaC. D.Figure 3.13 Autoradiograms of purified, 14C-labeled derivatives of CBD.PTA. Purified 2xCBDptlane 1:’4C-Mr standards,O.05 mCi (BRL, Bethesda, MD.)lane 2:7.5 p,g of purified 2xCBDpt, 214 cpm4tgB. Purified 2xCBDptl.2lane 1:’4C-Mr standards,O.05 mCi (BRL, Bethesda, MD.)lane 2: 2.1 .Lg of purified 2xCBDptl .2, 626 cpm/pgC. Purified CBD.PTlane 1: 14CMr standards,O.05 mCi (BRL, Bethesda, MD.)lane 2:4.5 jig of purified CBD.PT, 509 cpm/jigD. Purified CenAl.2lane 1:’4C-Mr standards,0.05 mCi (BRL, Bethesda, MD.)lane 2: 8.2 jig of purified CenA, 415 cpm/jiglane 3: 2.2 jig of purified CenA 1.2, 1711 cpm/jig45A. B.1 2 3 4 5 1 23449.5— 49.5—32.5— 32.5—27.5— 27.5—18.5—Figure 3.14 Western blot of purified, non-labeled proteinsA. Purified CenA and derivativeslane 1: pre-stained Mr standards, 5 mg in each band (BRL, Bethesda MD.)lane 2: 100 ng of purified CenAlane 3: 100 ng of purified CenADBDlane 4: 100 ng of purified CenADBD1.2lane 5: 100 ng of purified CenAPTNheThe primary antibody used was anti-CenAB. Purified CBD.PT and derivativeslane 1: pre-stained Mr standards, 5 mg in each band (BRL, Bethesda MD.)lane 2: 100 ng of purified CBD.PTlane 3: 100 ng of purified 2xCBDptlane 4: 100 ng of purified 2xCBD.ptl.2The primary antibody used was anti-CBD.PT46765•5 4C)20Figure 3.15 High protein concentration adsorption isotherms of CenA, CenADBD, andCenADBDL2. Initial protein concentrations ranged from 50-700 .tg/m1. F: free proteinfollowing equilibrium binding. B: bound protein. The CenA data was published previously(Gilkes et a!., 1992)0 2 4 6 8 10 12[F) (.LM)a)8a,(30)[F] (FM)47Figure 3.16 High protein concentration adsorption isotherms of CBD.PT, 2xCBDpt, and2xCBDptl.2. Initial protein concentrations ranged from 50-700 g/ml. F: free proteinfollowing equilibrium binding. B: bound protein. The CBD.PT data was publishedpreviously (Gillces et at., 1992).640 5 10 15 200Ea,C’,0Da,C)486. .00.0—0— noBSA—0— 66.7 ug/mI BSA-&- 20 ug/mI BSA100 ug/mI BSAG 333.3 ug/mI BSA0 25 50 75 100 125[F] (jiM)Figure 3.17 Effects of adding BSA in increasing concentrations on the adsorption of CenA.Details are given in section 3.3.2. F: free protein following equilibrium binding. B: boundprotein.492.0Initial [CenA] 0(1.89 gig/mI)1.5 o0EC)C01.0 --ö 0U)CC)CC0.5C00.0 I • I • •0 100 200 300 400 500 600 700[BSA] (tgImI)Figure 3.18 Effects of increasing BSA concentrations on the stability of CenA under theconditions of the adsorption assay. Details are given in section 3.19 Low initial protein concentration adsorption isotherm for14C-labelled CenAand CBD.PT. A. CenA, B. CBD.PT. Details are given in section 3.3.2. F: free proteinfollowing equilibrium binding. B: bound protein. The adsorption constants derived fromthese plots are given in Table 3.1.A.0 10 20 30 40 50 60 701/ [F] (tM Iigand1) 90 1001.0 2.0 3.0 4.0 5.01/[F] (1AM Iigand)6.051-o0)0Ea,U,0a,0)-oC0)0Ea,00a,00)A.0B.100 200 3001/[F] (tiM Iigand1)4005. 151I[F] (M Iigand)Figure 3.20 Low initial protein concentration adsorption isotherm for‘4C-labelledCenADBD and CenADBD1.2. A. CenADBD, B. CenADBD1.2. Details are given in section3.3.2. F: free protein following equilibrium binding. B: bound protein. The adsorptionconstants derived from these plots are given in Table 3.1.5 1052VC0)0ECl)00aA.B.150E.- 100Cl)00a—5m00.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.01I[F] (j.M)[F] (tiM Iigand1)Figure 3.21 Low initial protein concentration adsorption isotherm for14C-labelled2xCBDpt and 2xCBDptl.2. A. 2xCBDpt, B. 2xCBDptl.2. Details are given in section 3.3.2.F: free protein following equilibrium binding. B: bound protein. The adsorption constantsderived from these plots are given in Table 3.1.5 10 15 20 2553-DC0)0S0U,00C)C)mS..0C0)0S0U)00C)C)S..A. 5 10 15 20 25 301/[F] (jiM Iigand1) 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.01/[F] (jiM Iigand1)Figure 3.22 Low initial protein concentration adsorption isotherm for14C-labelled Cenl.2and CenAj.S.PT. A. CenAl.2, B. CenAAPT. Details are given in section 3.3.2. F: free proteinfollowing equilibrium binding. B: bound protein. The adsorption constants derived fromthese plots are given in Table 3.1.540.40.300020.10.0Figure 3.23 Small particle production from dewaxed cotton. Details are given in section3.3.2. The protein concentration in all cases was 10 jiM, except for CBD.PT (2X) which was20 pM.° CBD.PT-*-- 2xCBDpt2xCBDptl.2-— CBD.PT (2X)G Cotton only0 2 4 6 8 10 12Time (Hr)554. DiscussionThe modified Langmuir model of specific adsorption to cellulose and the method ofanalyzing adsorption data proposed by Gilkes et at. (1992) represents a significantimprovement over standard Langmuir analysis. The curved Scatchard plots of adsorptiondata, which have been attributed to high- and low-affinity binding, are readily accounted forby the probability of binding site exclusion (Gilkes et at., 1992). It is clear, as the modelpredicts, that in order to make any accurate conclusions about the affinities of cellulosebinding proteins for cellulose, adsorption studies must be done at low protein concentrations.It is also clear from previous adsorption studies using proteins with relatively high affinitiesfor cellulose (see Gilkes et a!., 1992, Figure 6), that absorbance at 280 nm (A280)is not asensitive enough method to quantify free protein in the concentration ranges required to asatisfactory degree of certainty Activity assays are a plausible alternative to A280 as amethod to quantify free protein, provided that the protein in question has enzymatic activity,and that a sufficiently sensitive assay of its activity exists. Extrinsically labeling the proteinwith either fluorescent or radioactive compounds is another alternative, but this methodleaves open the question of whether the added compounds affect the adsorptioncharacteristics of the labeled protein. Intrinsic 14C labeling avoids this complication whileallowing for very sensitive detection at low concentrations. Further, this method of detectioncan be used for any protein regardless of enzymatic activity.While allowing accurate determinations of relative affinities, low proteinconcentration adsorption assays have many drawbacks as well. As was shown in this study,protein loss due to instability or to sticking to the side of the tubes used in the assay is a majorproblem. Failing to consider this could lead to erroneous conclusions from adsorptionexperiments. For instance, the curved plots observed at very low protein concentrations inthe original adsorption assay using 14C-labeled proteins could have been interpreted as56evidence for high- and low-affinity binding sites. Furthermore, CenAPTNhe, of which asignificant amount was lost even in the modified adsorption assay, originally appeared tohave a considerably higher affinity than CenA (data not shown). This suggested that theadded alanine and serine residues in CenAPTNhe were responsible for increasing theadsorption affinity of the protein.In the development of the adsorption assay used in this study, the problem of proteinloss was minimized by pre-coating the tubes with BSA and including BSA in the assay mix.The added BSA creates other potential problems, though, as it was shown that as theconcentration of BSA was increased, the observed relative affinity of labeled CenAdecreased. Since the non-specific adsorption of BSA on the BMCC was found to benegligible, it is not likely that BSA competes for binding sites. Rather, the change inobserved affinity is probably due to the BSA interfering with binding by physically blockingbinding sites and, as a result, increasing the observed equilibrium concentration of non-bound protein. Fortunately, any effects of the added BSA on the observed affinities can beignored since the values reported are ‘relative affinities’ and are meaningful as long as all ofthe proteins to be compared are assayed identically.4.1 Interpretation of adsorption dataThe relative affinities of intact CenA and the isolated CBD.PT were previouslyreported to be roughly equivalent (Gilkes et a!., 1992). In this study, the relative affinity ofthe intact enzyme was found to be approximately 2.4 fold higher than the isolated bindingdomain. This suggests that the catalytic domain contributes to the overall adsorption affinityof the intact enzyme. It is not unreasonable that the catalytic domain could be directlyinvolved in cellulose binding, as it is a f3-glucosidase and must bind cellulose in the active sitein order to catalyze hydrolysis of the 3-1,4-g1ycosidic bonds (M.L. Sinnott, 1987).Furthermore, binding in the active site is believed to be a major factor in the overall57adsorption affinity of intact CBH II from T. reesei (Tomme et a!., 1988). Gillces et at. (1992)previously reported that the adsorption of the isolated CenA catalytic domain is negligible,but the binding affinity of this domain could be very small and still account for the differencein affinities for the CBD and the intact enzyme, as will be discussed later. It is also possiblethat the catalytic domain, whose molecular weight is about twice that of the binding domain,influences the conformation of the CBD in the intact enzyme. When the catalytic domain isnot present, the binding domain could adopt a different conformation which could result inlower binding affinity. Of course, this will remain speculation until considerably more isknown about the structure of the intact enzyme and the isolated domains. Until the crystalstructures of these proteins are determined, it may be beneficial to employ small-angle X-ray-scattering analysis to determine the overall shape of CBD.PT, as was done for intact CenAand the isolated catalytic domain (Pilz et at., 1990). This would give some indication ofwhether the conformation of the isolated CBD is grossly different from its conformation inthe intact enzyme.The relative affinities of the double binding domain derivatives of CenA wereconsiderably different from each other and from that of CenA. CenADBD adsorbed withabout 3-fold higher affinity than did CenA. This result suggests that both tandemly repeatedbinding domains in this protein are functional and contribute to the overall affinity of theprotein. However, the affinity of this protein is not as high as might have been expected.Based of the model of Crothers and Metzger (1972) to describe the binding of a bivalentantibody to antigenic determinants on a surface, we can make some theoretical predictions asto the expected affinities of double binding domain proteins. If the influence of the catalyticdomain on overall binding affinity is ignored, the binding reaction of a double bindingdomain protein can be divided into two steps, as was done for the bivalent antibody case, asshown in the following diagram (adapted from Crothers and Metzger, 1972):58K1 K2C C C CI I I I I Iwhere ‘C’ represents a binding site on the surface of the cellulose and K1 and K2 are theequilibrium constants of the two steps of the reaction. If we assume a completely flexiblelinkage between domains such that each of the two domains is able to bind to the surface ofthe cellulose independently of the other without influencing the binding of the other, theobserved affinity constant (Kobs) will be:Kobs=2*K1*K2 (4)as derived by Crothers and Metzger (1972). If we further assume that both binding domainsare in an identical conformation such that each domain has an identical adsorption affinityand ignore any proximity effects on K2, then Equation (4) becomes:Kobs=2*Ka2 (5)where Ka is the affinity constant for each CBD. Although Ka cannot be determined preciselybecause [N0] cannot be measured (see Results), Gilkes et a!. (1992) estimated [N0] to beabout 100 Lmol/g cellulose and used this value to derive a rough approximation of Ka. Usingthis value, the Ka of CenA can be estimated to be approximately M1. Thus, the Ka of adouble binding domain derivative of CenA, given the stated assumptions, could betheoretically predicted to be about 1014 M1,which would give a Kr value about nine orders• of magnitude greater than that of CenA. Of course, this model is over-simplistic. In a recentstudy, Fab’ fragments were cleaved from several intact IgG molecules, and their bindingaffinities for specific cell surface receptors were measured (Temponi et a!., 1992). Fragmentsfrom two of the antibodies showed 7- and 40- fold reductions in affinity compared to theaffinity of the intact antibody. This was reported as ‘qualitatively’ fitting the theory ofbivalent binding in that the isolated fragments showed a marked reduction in affinityC C59compared to the intact antibody (Temponi et a!., 1992). In the same study, an Fab’ fragmentfrom another antibody had only a 2-fold reduction in affmity. This was interpreted as anindication that the intact antibody bound univalently rather than bivalently to the cell surface.Although it is not valid to make predictions based on a simple model or analogy to unrelatedproteins, the preceding discussion demonstrates that an increase in affinity of 3-fold over theaffinity of CenA is not as large as might have been expected for CenADBD.The model also demonstrates how the difference in affinities of CenA and CBD.PTcould, theoretically, be due largely to substrate binding in the active site. Using the modeland a value of M4 for the Ka of CBD.PT in Equation (4), the affinity constant for thesubstrate in the active site, K5, would be about 1 M’, which would represent a Kr value inthe range of six orders of magnitude lower than CenA. Again, the estimated theoreticalvalues are given here are for demonstration purposes only.There are several possible reasons for the seemingly lower than expected affinity ofCenADBD. It could be that one or both of the repeated binding domains are not in theoptimal conformation. The outside domain has another CBD at its C-terminal instead of acatalytic domain as in wild type CenA, and the inside binding domain has another domainfused at its N-terminal unlike the CBD in the wild-type enzyme. Thus, it is possible that theconformation of both CBDs could be altered when they are fused in tandem. It is alsopossible that the two domains simply sterically hinder each other. Moreover, the tandembinding domains in CenADBD may not be oriented properly with respect to the surface of thecellulose, as will be discussed in more detail later.CenADBD1.2 adsorbed to cellulose with about 33-fold lower affinity than CenADBDand about 11-fold lower affinity than CenA. This result suggests that the nature of the linkerconnecting the two tandemly repeated domains is important for adsorption affinity. Since thelinker separating the binding domains in this protein is shortened by 10 residues, the distancebetween the domains is reduced and any steric hindrance considerations should be increased.The reduced distance between domains could also increase any strain on the conformation of60either or both of the CBDs. Furthermore, the two CBDs could be oriented differently inCenADBD1.2 than they are in CenADBD.Both 2xCBDpt and 2xCBDptl.2 had reduced binding affinity compared to CBD.PT,indicating that the tandemly repeated binding domains in these proteins do not function intandem to increase the affinity of the protein. The protein with the shortened linker,2xCBDptl.2 adsorbed with a higher affinity than did 2xCBDpt. This result is unexpectedgiven that the shortened linker of CenADBD1.2 resulted in reduced adsorption affinitycompared to CenADBD. This could be interpreted as an indication that 2xCBDpt may bedifficult to refold properly following removal of guanidine in the purification of the protein(see Materials and Methods) and that there could be a significant population of improperlyfolded, inactive or partially active proteins in the preparation. This is reinforced by the factthat one preparation of labeled 2xCBDpt did not adsorb at all (data not shown). However, ashas been shown in binding studies of purified antibodies, the presence of a percentage of‘dead’ protein in a preparation would lead to non-linear plots of adsorption data (VanRegenmortel, 1982). Furthermore, the results of the small particle release experimentssuggest that this protein is functional.2xCBDpt adsorbed with considerably reduced affinity than did CenADBD. Thisresult tends to reinforce the suggestion that the catalytic domain contributes to the overalladsorption affinity of these proteins either by binding substrate in the active site or bystabilizing the conformation of the CBD. However, the Kr of 2xCBDptl.2 was roughlyequivalent, if not slightly higher, than the Kr of CenADBD1.2. It is tempting to hypothesizethat in the absence of the bulky catalytic domain, the tandemly repeated CBDs in this proteinhave more freedom to adopt a conformation relative to each other that would allow for moreoptimal interaction of both domains with the cellulose, and that this would tend to partlycompensate for the loss of the influence of the catalytic domain. The results for 2xCBDptargue against this idea, however. It is clear that the adsorption characteristics of these61proteins cannot be easily explained in relation to the results of the analogous proteins thathave catalytic domains.As the sizes of the 95% confidence limits indicate, the errors in parameters derivedfrom the y-intercepts of the double reciprocal plots (ar and Am) are, for the most part, verylarge (see Table 3.1). For CenA, CBD.PT, and CenADBD, though, the errors in the values ofthese parameters are reasonable enough to make some qualitative comparisons between them.The Am values derived from the low concentration isotherms are considerably lower thanthe saturation levels that these proteins seem to approach in the high concentration adsorptionassays. According to the adsorption model, the theoretical saturation level should be largerthan the observed saturation level because of the probability of binding site exclusion (Gilkeset at., 1992). This discrepancy could be an indication of a limitation of the adsorption model.It could also be explained to some extent by the ‘dangling ligand’ hypothesis that states that asthe cellulose surface becomes saturated, some ligands bind by making only partial contact(Gillces et at, 1992). On the other hand, the apparently high levels of adsorption in the highconcentration binding assays could, at least in part, be due to protein loss due to instability orto sticking to the sides of the tubes.The ar value of CenA was higher than that of CBD.PT. This is expected since CenAis a larger protein than CBD.PT and should, therefore, cover more lattice units when bound tothe surface of the cellulose. The ar value for CenADBD was roughly equivalent to that ofCenA. This suggests that the tandemly repeated binding domains in this protein are notoriented relative to each other such that each CBD is able to interact optimally with thesurface of the cellulose.624.2 Effects of altering the linker between the catalytic and binding domains of CenA on theadsorption affinity of the enzyme.Proline-rich linkers separating functional domains in cellulases and other enzymes aiecommon (Gilkes et at., 1991a). Several examples of such linkers have been shown to haveextended rather than globular structures (Gilkes eta!., 1991a). They are also often assumedto form ‘flexible hinges’, similar to the hinge regions of immunoglobulins, between domains.The Kol IgG, for example, is known to have a high degree of conformational flexibilitybetween the Fab arms and the Fc domain within the hinge region. The flexibility does notoccur over the whole length of the region, but begins at a single residue that follows apolyproline helix (Huber et at., 1976). Huber and Bennet (1983) cite several other examplesof proteins in which flexibility begins a single residue, which is often a glycine. NMRinvestigations of a number of peptides from proline-rich linkers have suggested that thesepeptides are in a ‘disordered’ or. ‘random-coil’ conformation. From this data substantialmobility of the polypeptide backbone of the linkers has been inferred (Bhandari et a!., 1986).Bhandari et a!. (1986) argued that the NMR observations could result from uncoordinatedmovement of the peptide and give no insight into the actual conformation of the backbone. Intheir study of a flexible region of the LC1 light chain of skeletal muscle, Bhandari et a!.(1986) found the backbone of this region to be in an extended, stiffened conformation withmotion restricted to segments around ‘elbow joints’ consisting of alanine-proline peptidebonds.Many of the proline-rich linkers that are known to have a very high degree ofconformational flexibility, such as those of the E.coli proteins L12 (Busheuv et a!., 1989) andthe pyruvate dehycirogenase complex (Radford et at., 1989), are also very rich in alanineresidues and contain few, if any, charged or hydroxy-amino acids. The hinge region of IgAlis comprised entirely of proline, serine, and threonine residues with a single valine (Burton etat., 1989), and the similarity of this linker to the Pro-Thr linker of CenA has been noted63(Miller et a!., 1992). The IgAl hinge is also often assumed to be a flexible linker (Gilkes eta!., 1991a), and has been used as a general linker to link two subunits of superoxidedismutase (Hallewell et a!., 1989). The flexibility is largely based on NMR studies ontetrapeptides from the hinge region that suggested that some of the peptides adopted random-coil conformations (Siemion et a!., 1988; Burton et a!., 1989). The actual extent of mobilityallowed by this hinge or site(s) where flexibility occurs within the region has not beendemonstrated. Jeske and Capra (1984) hypothesized that differences in the hinge regions ofimmunoglobulins could reflect differences in conformational flexibility between domains ofthese proteins. In a recent database search of natural linkers, Argos (1989) concluded thatoptimal general linkers between domains should contain serine and threonine residues thatcould hydrogen bond to solvent molecules or to the main chain nitrogen molecules andachieve conformational stability, as well glycine residues to allow some flexibility. Thus,linkers such as the IgAl hinge and the Pro-Thr linker that contain only proline and hydroxyamino acid residues and lack residues with short side chains should be expected’ to be fairlyrigid. This is consistent with small-angle X-ray scattering analysis that suggested a rigidconformation for the Pro-Thr linker (Shen et at., 1991). It is interesting to note that in theflanking regions at either ends of the Pro-Thr linker are both glycine and valine residues thatcould provide some flexibility at these points.Recently, computer modeling of peptides spanning the Pro-Thr linker of CenA andthe ½ Pro-Thr linker was done by Dr. R. Parker at the University of Alberta Department ofBiochemistry. The modeling predicts an extended conformation for the Pro-Thr linker with ahigh probability of a kink at the (Pro-Thr-Thr-Pro) region. This fits well with the model ofCenA based on SAXS analysis that predicts the longitudinal axis of the catalytic domain to beat an angle with respect that of the binding domain. Shen et at., (1991) hypothesized that therole of the Pro-Thr linker in CenA is tomaintain both the separation and the relativeorientation of the two domains necessary for proper function. It would be interesting todetermine whether a kinked linker is important for catalysis or adsorption. The Pro-Thr64linker of CenA has been replaced by the IgAl hinge (Miller et a!., 1992). This linker issimilar in amino acid composition and size, but may not adopt a similar shape as the Pro-Thrlinker. SAXS-analysis would give an indication of the relative orientation of the twodomains in this protein, and catalytic and adsorption studies would lend some further insightinto the role of the Pro-Thr linker.Shortening or eliminating the Pro-Thr linker of CenA resulted in reduced adsorptionaffinity of the enzyme. CenAl.2 adsorbed with about 4-fold lower affinity than CenA.Adsorption of CenAAPT was previously reported to be unchanged from that of CenA (Shenet al., 1991). Investigation of the adsorption of this protein at low concentrations in thisstudy revealed a reduction in affinity of about 26-fold. These results are consistent with thehypothesis that the presence of the Pro-Thr linker is important for the correct functioning ofCenA. The reduced affinities of these proteins are probably largely due to increased sterichindrance between the domains as the distance between them is decreased. The possibilitythat the reductions in affinity could also be at least partly due to conformational changes inthe CBD induced by shortening or deletion of the linker cannot be ruled out, however.Comparison of the SAXS-generated model of CenAAPT with that of CenA (Shen et al.,1991, Figure 9) suggests that the CBD in CenAAPT adopts a shorter, stubbier conformationthan the CBD in CenA.The model of the Pro-Thr linker of CenA, then, is that it is rigid and adopts anextended, kinked conformation, and that this conformation is important for the correctspacing and orientation of the catalytic and binding domains with respect to each other.While this linker may be optimal for the functioning of CenA, it may be somewhat less thanideal as a general linker between tandemly repeated binding domains. Replacement of thePro-Thr linker between the binding domains of CenADBD with a more flexible linker maylead to increased affinity of the protein by allowing the two domains more freedom to adopt amore optimal configuration relative to each other. A poly-glycine stretch is a possibility, butArgos (1989) argued that such a peptide does not occur naturally and could be unstable. He65proposed the use of peptides containing only glycine, serine, and threonine as general linkersequences between domains since such peptides should give some degree of both flexibilityand conformational stability. Alternatively, an alanine- and proline-rich linker, such as theone from pyruvate dehydrogenase (Radford et at., 1989), could be used. In addition to thelack of flexibility between binding domains linked by the Pro-Thr linker, anotherconsideration could be the orientation of the two domains relative to the surface of thecellulose. Although the mechanism by which the CBD binds to cellulose is unknown, it isbelieved that certain residues of the CBD interact with the cellulose though hydrogen bondingand van der Waals contacts resulting from stacking of aromatic side chains (Vyas eta!., 1991;Gillces et a!., 1991a). Thus, one or both of the two binding domains could be oriented suchthat the face or faces of the protein that contain the residues that interact with the cellulosecould be at an angle relative to the cellulose surface. In the proteins with ½ Pro-Thr linkers,the orientation of the CBDs could be different from their orientation in proteins with a full-length linker. Of course, this will remain speculation until more specific details are knownabout the structures, of the these proteins and the ways in which they interact with cellulose.4.3 Small particle release and enzymatic activityDin et at. (1991) observed the release of small particles from intact cotton fibresfollowing incubation with CBD.PT. Such release has been cited as evidence of Reese’s C1factor (Wood et at., 1989), the non-hydrolytic component of the cellulose system that initiallydisrupts crystalline cellulose and renders it susceptible to hydrolysis (Reese, 1950). Thesmall particle data and other observations led Din et a!. (1991) to suggest that the CBD ofCenA plays a more active role in the hydrolysis of cellulose than simply adsorbing theenzyme to the cellulose surface. They concluded that the cellulose binding domain is the C1factor in Reese’s model of cellulose degradation.66Both of the double binding domain derivatives of CBD.PT released more smallparticles from cotton fibres than did CBD.PT. The proposed mechanism of small particlerelease involves hydrogen bonding of the enzyme with the cellulose resulting in disruption ofthe hydrogen bonding networks of the cellulose crystal structure (Wood et a!., 1989). Itwould follow that reduced adsorption affinity of the enzyme should result in less hydrogenbonding to the substrate and reduced small particle production. However, 2xCBDpt and2xCBDptl.2 both have reduced adsorption affinities and caused greater small particle release.The fact that both of these proteins released more small particles than did CBD.PT at doublethe concentration (i.e., at equivalent concentrations of binding domains) suggests that there issynergy between the two domains in the action of small particle release.No differences were observed in the abilities of CenADBD, CenADBD1.2, and CenAto hydrolyze crystalline cellulose. This result suggests that the correlation betweenadsorption affinity and hydrolytic activity does not apply to CenA. Removal of the CBDfrom CenA reduced, but did not eliminate, the hydrolytic activity of the enzyme on BMCC(Gilkes et at., 1992). This suggests that tight adsorption of CenA to crystalline cellulose isnot as important for the hydrolytic activity of this enzyme as it apparently is for otherendoglucanases (Klyosov, 1990; Stâhlberg et at., 1991). The results of the hydrolysisexperiments using the double binding domain proteins are consistent with this hypothesis.Furthermore, these results suggest that the presence of the CBD in the enzyme is a moreimportant consideration for hydrolytic activity than is the overall adsorption affinity of theenzyme. This reinforces the idea that the binding domain of CenA plays more of an activerole in the hydrolysis of cellulose than simply adsorbing the enzyme to the substrate (Din etat., 1991).67ReferencesAbuja, P.M., M. Schmuck, I. Pilz, P.Tomme, M. Claeyssens, and H. Esterbauer, 1988a.Structural and functional domains of cellobiohydrolase from Trichoderma reesei. A smallangle X-ray scattering study of the intact enzyme and its core. Eur. Biophys. J. 15: 339-342Abuja, P., I. Pilz, M. Claeyssens and P. Tomme. 1988b. Domain structure ofcellobiohydrolase II as studied by small angle X-ray scattering: close resemblance tocellobiohydrolase I. Biochem. Biophys. Res. Commun. 156: 180-185Argos, P. 1989. An investigation of oligopeptides linking domains in protein tertiarystructures and possible candidates for general gene fusion. J. Mo!. Biol. 211: 943-958Atkinson, T. and M. Smith. 1984. Solid phase synthesis of oligodeoxyribonucleotides by thephosphite triester method. In N.J. Gait (ed.) Oligonucleotide Synthesis: A practicalapproach. IRL Press, Oxford, pp. 35-8 1Beguin, P., H. Eisen and A. Roupas. 1977. Free and cellulose-bound cellulases in aCellulomonas species. Journal of General Microbiology. 101:191-196Beguin, P., N.R. Gilkes, D.G. Kilburn, R.C. Miller, Jr., G.P. O’Neill and R.A.J. Warren.1987. Cloning of cellulase genes. CRC Critical Reviews in Biotechnology. 6: 129-162Blackwell, 3. 1981. The structure of cellulose and chitin. pp. 523-535. In R. Srinivasan(ed.) Biomolecular Structure, Conformation, Function and Evolution. Vol. 1: Diffraction andRelated Studies. Pergamon, New York.Burton, J., S.G. Wood, A. Pedyczak and I.Z. Siemion. 1989. Conformational preferences of‘sequential fragments of the hinge region of human IgAl immunoglobulin molecule: II.Biophysical Chemistry 33: 39-45Bushuev, V.N., A.T. Gudkov, A. Liljas and N.F. Sepetov. 1989. The flexible region ofprotein L12 from bacterial ribosomes studied by proton nuclear magnetic resonance. J. Biol.Chem. 264(8): 4498-4505Cantor, C.R. and P.R. Sneyd. 1980. Estimates of protein concentration from UV absorbance.Biophysical Chemistry, Vol. 2, pp. 380-381. W.H. Freeman, San Francisco.Crothers, D.M., and H. Metzger. 1972. The influence of polyvalency on the bindingproperties of antibodies. Immunochemistry 9: 341-357Din, N., N.R. Gifices, B. Tekant, R.C. Miller, Jr., R.A.J. Warren and D.G. Kilburn. 1991.Non-hydrolytic disruption of cellulose fibres by the binding domain of a bacterial cellulase.Bioltechnology 9: 1096-109968Fan, L.T., L. Yong-Hyun, and D.H. Beardmore. 1980. Major chemical and physical featuresof cellulosic materials as substrates for enzymatic hydrolysis. Advances in BiochemicalEngineering. 14: 101-107Gilkes, N.R., D.G. Kilburn, M.L. Langsford, R.C. Miller, Jr., W.W. Wakerchuk, R.A.J.Warren, D.J. Whittle and W.K.R. Wong. 1984. Isolation and characterization of Escherichiacoli clones expressing cellulase genes from Cellulomonasfimi. Journal of GeneralMicrobiology. 130: 1377-1384Gilkes, N.R., R.A.J. Warren, R.C. Miller, Jr., and D.G. Kilburn. 1988. Precise excision ofthe cellulose-binding domains from two Cellulomonasfimi cellulases by a homologousprotease and the effect on catalysis. J. Biol. Chem. 263: 10401-07Gilkes, N.R., B. Henrissat, D.G. Kilburn, R.C. Miller, Jr. and R.A.J. Warren. 1991a.Domains in microbial 3-1,4-glycanases: sequence conservation, function, and enzymefamilies. Microbiological Reviews. 55(2): 303-315Gilkes, N.R., D.G. Kilburn, R.C. Miller, Jr. and R.A.J. Warren. 1991b. Bacterial Cellulases.Bioresource Technology 36: 21-35Gillces, N.R, E. Jervis, B. Henrissat, B. Tekant, R.C. Miller, Jr., R.A.J. Warren and D.G.Kilburn. 1992. The adsorption of a bacterial cellulase and its two isolated domains tocrystalline cellulose. J. Biol. Chem. 267(10): 6743-6749Guo, Z.M., N. Arfman, E. Ong, N.R. Gillces, D.G. Kilburn, R.A.J. Warren and R.C. Miller,Jr. 1988. Leakage of Cellulomonasfimi cellulases from Escherichia coli. FEMS MicrobiolLett. 49: 279-283Hallewell, R.A., I. Laria, A. Tabrizi, G. Carlin, E.D. Getzoff, J.A. Tamer, L.S. Cousens andG.T. Mullenbach. 1989. Genetically engineered polymers of human CuZn superoxidedismutase. J. Biol. Chem. 264(9): 5260-5268Halliwell, G. 1965. Hydrolysis of fibrous cotton and reprecipitated cellulose by cellulolyticenzymes from soil micro-organisms. Biochem. J. 95: 270-28 1Henrissat, B., B. Vigny, A. Buleon and S. Perez. 1988. Possible adsorption sites ofcellulases on crystalline cellulose. FEBS letters 231: 177-182Huber, R., 3. Deisenhofer, P.M. Colman, M. Matushima and W. Palm. 1976.Crystallographic structure studies of an IgG molecule and an Fc fragment. Nature. 264: 415-420Huber, R. and W.S. Bennett, Jr. 1983. Functional significance of flexibility in proteins.Biopolymers. 22: 261-27969Jeske, D.J., and J.D. Capra. 1984. Immunoglubulins: structure and function, pp. 131-165. InW.E. Paul (ed.), Fundamental Immunology. Raven Press, New York.Klyosov, A.A., O.V. Mitkevich, and A.P. Sinitsyn. 1986. Role of activity and adsorption ofcellulases in the efficiency of the enzymatic hydrolysis of amorphous and crystallinecellulose. Biochemistry 25, 540-542Klyosov, A.A. 1990. Trends in Biochemistry and Enzymology of Cellulose Degradation.Biochemistry 29 (47): 10576-10585Kristensen, T., H. Voss, and W. Ansorge. 1987. A simple and rapid preparation of M13sequencing templates for manual and automatic dideoxy sequencing. Nucleic Acid Res. 15:5507-5516Kunkel, T.A., J.D. Roberts and R.A. Zakour. 1987. Rapid and efficient site-specificmutagenesis without phenotypic selection. Methods in Enzymology 154: 367-382Laemrnli, K. 1970. Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature 227: 680-685Langsford, M.L., N.R. Gilkes, B. Singh, B. Moser, R.C. Miller, Jr., R.A.J. Warren and D.G.Kilburn. 1987. Glycosylation of bacterial cellulase prevents proteolytic cleavage betweenfunctional domains. FEBS Letters. 225: 163-167Merril, C.R. 1990. Gel staining techniques. Methods. Enzymology. 182: 477-488Miller, P.B., H. Shen, N.R. Gillces, D.G. Kilburn, R.C. Miller, Jr., A.G. Plaut and R.A.J.Warren. 1992. Endoglucanase A from Cellulomonasfimi in which the hinge sequence ofhuman IgAl is substituted for the linker connecting its two domains is hydrolyzed by IgAproteases from Neisseria gonorrhoeae. FEMS Microbiology Letters 92: 199-204Osborne, M.J. and R. Munson. 1974. Seperation of the inner (cytoplasmic) and outermembranes of gram-negative bacteria. Methods in Enzymology. vol XXXI, Biomembranes,part A, p.p. 642-667Pilz, I., E. Schwarz, D.G. Kilburn, R.C. Miller, Jr., R.A.J. Warren, and N.R. Gillces. 1990.The tertiary structure of a bacterial cellulase determined by small-angle X-ray-scatteringanalysis. Biochem J. 271: 277-280Radford, S.E., E.D. Laue, R.N. Perham, S.R. Martin and E. Appella. 1989. Conformationalflexibility and folding of synthetic peptides representing an interdomain segment ofpolypeptide chain in the pyruvate dehydrogenase multienzyme complex of Escherichia coli.J. Biol. Chem. 264(2): 767-775Reese, E.T., R.G.H. Sui and H.S. Levinson. 1950. Biological degredation of solublecellulose derivatives. J. Bacteriol. 59: 485-49770Richmond, P.A. 1991. Occurrence and functions of native cellulose. pp. 5-23. In C.H.Haigler and P.J. Weimer (eds.) Biosythesis and Biodegradation of Cellulose. MarcelDekker, Inc., New York. VSambrook, 3., E.F. Fritsch and T. Maniatis. 1989. Molecular Cloning: A LaboratoryManual, 2nd ed. Cold Spring Harbor Laboratory Press, New York.Schagger, H. and G. von Jagow. 1987. Tricine-sodium dodecyl suiphate-polyacrylamide gelelectrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal. Biochem.166: 368-379Scopes, R.K. 1974. Measurement of protein by spectrophotometry at 205nm. Anal.Biochem. 59: 277-282Shen, H., M. Schmuck, I. Pilz, N.R. Gilkes, D.G. Kilburn, R.C. Miller, Jr. andR.A.J. Warren. 1991. Deletion of the linker connecting the catalytic and cellulose-bindingdomains of Endoglucanase A (CenA) of Cellulomonasfimi alters its conformation andcatalytic activity. J. Biol. Chem. 266(17): 11335-11340Siemion, J., A. Pedyczak and J. Burton. 1988. Conformational preferences of the hingeregion of sequential fragments of the hinge region of human IgAl immunoglobulin molecule.Biophys. Chem. 31: 35-41Sinnott, M.L. 1987. Glycosylgroup transfer, pp. 259-296. In M.I. Page and A. Williams(ed.), Enzyme mechanisms. Royal Society of Chemistry, London.Spurlino, J.C., G.Y. Lu and F.A. Quiocho. 1991. The 2.3-A resolution of the maltose- ormaltodextrin- binding protein, a primary receptor of bacterial active transport andchemotaxis. J. Biol. Chem. 266(8): 5202-5219StAhlberg, 3., G. Johansson and G. Pettersson. 1991. A new model for enzymatic hydrolysisof cellulose based on the two-domain structure of cellobiohydrolase I. Bioltechnology. 9:286-290Steiner, W., W. Sattler and H. Esterbauer. 1988. Asorption of Trichoderma reesi cellulaseon cellulose: experimental data and their analysis by different equations. Biotechnol. Bioeng.32: 853-865Stuart, J.Y. and D.L. Ristroph. 1985. Analysis of cellulose-cellulase adsorption data: afundamental approach. Biotechnol. Bioeng. 27: 1056-1059Tabor, S. and C.C. Richardson. 1987. DNA sequence analysis with a modifiedbacteriophage Ti DNA polymerase. Proc. Nati. Acad. Sci. USA. 84: 4767-477 171Temponi, M., A.M. Gold and S. Ferrone. 1992. Binding parameters and idiotypic profile ofwhole immunoglobulin and Fab’ fragments of murine monoclonal antibody to distinctdeterminants of the human high molecular weight-melanoma associated antigen. CancerResearch 52: 2497-2503Tomme, P., H. van Tilbeurgh, G. Pettersson, 3. Van Damme, 3. Vandekerckhove, 3. Knowis,T.T. Teeri, and M. Claeyssens. 1988. Studies of the cellulolytic system of Trichodermareesei QM9414. Analysis of domain function in two cellobiohydrolases by limitedproteolysis. Eur. J. Biochem. 170: 575-58 1Trimbur, D.E., .R.A.J. Warren, and S.G. Withers. 1992. Region-directed mutagenesis ofresidues surrounding the active site nucleophile in (3-glucosidase from Agrobacteriumfaecalis. J. Biol. Chem. 267: 10248-1025 1Van Regenmortal, M.H.V. 1982. Serology and immunochemistry ofplant viruses, pp. 55-72.Academic Press, Toronto.Vieira, 3. and J. Messing. 1987. Production of single-stranded plasmid DNA. Methods inEnzymology 153: 3-11Vyas, N.D., M.N. Vyas and F.A. Quiocho. 1991. Comparison of the periplasmic receptorsforL-Arabinose, D-Glucose/D-Galactose, and D-Ribose. J. Blot. Chem. 266(8): 5226-5237Whittle, D.J., D.G. Kilburn, R.A.J. Warren and R.C. Miller, Jr. 1982. Molecular cloning ofa Cellulomonasfimi gene in Escherichia coil. Gene 17: 139-145Wood, T.M. 1988. Preparation of crystalline, amorphous and dyed cellulose substrates.Methods in Enzymoi. 160: 19-21Wood, T.M. 1989. Mechanisms of cellulose degradation by enzymes from aerobic andanaerobic fungi, p. 12-35. In: Enzyme systemsfor lignocellulose degradation. Coughlan,M.P. (ed). Elsevier Applied Science, New York.Yanisch-Perron, C., J. Vieire, and 3. Messing. 1985. Improved M13 phage cloning vectorsand host strains: nucleotide sequences of the Ml3mp and pUC vectors. Gene 33: 103-119


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items