UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Functional and mutational analysis of the cellulose binding domain CBDN1 from Cellulomonas fimi [Beta]-glucanase… Kormos, Jeffery Michael 1998

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

Item Metadata


831-ubc_1998-0256.pdf [ 17.19MB ]
JSON: 831-1.0088527.json
JSON-LD: 831-1.0088527-ld.json
RDF/XML (Pretty): 831-1.0088527-rdf.xml
RDF/JSON: 831-1.0088527-rdf.json
Turtle: 831-1.0088527-turtle.txt
N-Triples: 831-1.0088527-rdf-ntriples.txt
Original Record: 831-1.0088527-source.json
Full Text

Full Text

FUNCTIONAL AND MUTATIONAL ANALYSIS OF THE CELLULOSE BINDING DOMAIN CBDN1 FROM CELLULOMONAS FIMI p-GLUCANASE C (CENC) by  JEFFERY MICHAEL KORMOS B.Sc, Brock University, 1995  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE  FACULTY OF GRADUATE STUDIES (Department of Microbiology and Immunology/Biotechnology laboratory)  We accept this thesis as conforming to the reauired standard  THE UNIVERSITY OF BRITISH COLUMBIA April, 1998 © Jeffery Michael Kormos, 1998  In  presenting  degree freely  at  the  available  copying  of  department publication  this  of  in  partial  fulfilment  University  of  British  Columbia,  for  this or  thesis  reference  thesis by  this  for  his thesis  and  scholarly  or for  her  Department The University of British Columbia Vancouver, Canada  DE-6  (2/88)  I  I further  purposes  gain  the  shall  requirements  agree  that  agree  may  representatives.  financial  permission.  Date  study.  of  be  It not  that  the  Library  by  understood be  an  allowed  the  advanced  shall  permission for  granted  is  for  make  extensive  head  that  without  it  of  copying my  my or  written  Abstract Endoglucanase C (CenC), a (3,1-4 glucanase from the soil bacterium Cellulomonas fimi binds to cellulose via the cellulose binding domains N l (CBDN1), and N2 (CBDN2). In this thesis, the contribution of 10 amino acids within the binding cleft of CBDN1 was evaluated by alanine mutagenesis. Each mutant was analyzed for binding to an insoluble allomorph of cellulose (phosphoric acid swollen Avicel), and a soluble glucopyranoside polymer (barley (3-glucan). A l l ten mutations affected the binding affinity of CBDN1 to some extent; Y19 and Y85 were identified as the two residues most essential for tight binding to insoluble cellulose and barley pVglucan. The structural integrity and intermolecular contacts of Y19A and Y85A were assessed using twodimensional N M R spectrospcopy and U V resonance Raman spectroscopy (UVRRS). Both techniques indicated that Y85A is similar to wild-type CBDN1 in structure, but that Y19A exhibits some structural perturbations in the binding cleft. In addition, UVRRS data indicated that Y85 forms a direct hydrogen bond to cellulose upon binding, but Y19 does not. Taken together, these data imply that CBDN1 binding is mediated by two tyrosine residues, one of which interacts direcdy with cellulose, the other important for structural integrity of the binding cleft.  Table of Contents Abstract Table  ;  of  ii  Contents  iii  List  of Tables  vi  List  of Figures  vii  List  of  viii  Abbreviations  Acknowledgments  x  Chapter 1: Introduction  1  1.1 Cellulose and cellulose degradation  1  1.1.1  Structure of cellulose  1  1.1.2  Degradation of cellulose  4  1.1.3  Cellulases of Cellulomonas  fimi  1.2 CBD's  5  1.2.1  General classification and properties  5  1.2.2  Function of CBDs  7  1.2.3  Structure/function studies on CBDs  9  1.2.4  CBDN1  1.3 Thesis Objectives  Chapter 2:  5  Materials and Methods  ...12 13  14  2.1  Materials  14  2.2  Bacterial strains, vectors, and growth conditions  14  2.3  DNA isolation and manipulation  16  2.4  Polymerase chain reactions  16  2.5  Expression and purification of CBDN1 and mutants from E. coli JM101  periplasms  19  2.6  20  Calculation of affinity constants on insoluble cellulose  2.7  Chapter 3:  Calculation of affinity constants on soluble glucan polymers  Expression and Purification of C B D N 1  21  24  3.1  Background  24  3.2  Initial considerations for protocol development  24  3.3  Purification of CBDN1 from E. coli supernatants  25  3.4  Purification of CBDN1 from E. coli periplasms  27  3.5  Summary  29  Chapter 4:  Binding Studies on Wild-type C B D N 1  30  4.1  Background  30  4.2  Demonstration of reversible binding to PAS A  30  4.3  Binding of CBDN1 to allomorphs of insoluble cellulose  33  4.4  Temperature dependence of CBDN1 binding  36  4.5  Use of CBDs to characterize cellulose  38  Chapter 5:  Characterization of C B D N 1 Binding by Site-directed Mutagenesis 42  5.1  Choice of mutation  42  5.2  Structural characterization of CBDN1 alanine mutants  44  5.2.1  Fundamentals of  nuclear magnetic resonance spectroscopy  5.2.2  Mutant characterization by 1-D *H NMR spectroscopy  44 45  5.3  Binding of CBDN1 alanine mutants to PAS A  49  5.4  Binding of CBDN1 mutants to barley (3-glucan  53  5.5  Comparison of mutant binding on PASA and barley (3-glucan  58  5.6  Identification of the residues critical for binding  59  5.7  Use of mutagenesis to study mechanistic differences in binding  60  V  Chapter 6:  Additional studies on Y19A and Y85A  63  6.1  Structural studies on Y19A and Y85A by H- N  6.2  Analysis of Y19A and Y85A binding to soluble oligosaccharides by NMR  6.3  Determination of the roles of Y19 and Y85 by UV resonance Raman  l  15  HSQC  spectroscopy  Chapter 7:  63 70  71  6.3.1  Overview of UVRRS theory  71  6.3.2  UVRRS  73  6.3.3  Evidence for structural involvement of Y19 but not Y85 in CBDN1.... 73  6.3.4  Evidence for direct interaction of Y85 but not Y19 with cellopentaose.. 78  protocol  Conclusions  81  7.1  Current view of CBDN1 binding  81  7.2  Carbohydrate binding proteins  83  Chapter 8:  7.2.1  Classification of carbohydrate binding proteins  83  7.2.2  Forces involved in carbohydrate binding  83  7.2.3  Mutational studies on carbohydrate binding sites  86  7.2.3  Comparison of CBDN1 to other carbohydrate binding proteins  88  References  90  List of Tables Table 1.1.  CBD families and their properties  6  Table 4.1. Allomorph Specificity of CBDN1  35  Table 4.2.  Temperature dependence of CBDN1 binding to PASA  36  Table 4.3.  Binding characteristics of CBDN1 and CBDCex on PASA  39  Table 4.4.  Summary of the binding data for CBDCex and CBDN1  40  Table 5.1. Binding data for CBDN1 and alanine mutants on PASA  50  Table 5.2. Binding data for CBDN1 and alanine mutants on barley pVglucan  55  Table 6.1.  Summary of the HSQC data for wild-type CBDN1, Y19A, and Y85A  67  Table 6.2.  Summary of the HSQC chemical shift analysis for wild-type CBDN1, Y19A, and  Y85A  67  Table 6.3.  NMR binding data for CBDN1, Y19A, and Y85A  71  Table 6.4.  Summary of internuclear distances in Y19, W16, and W137  76  Table 7.1.  Summary of residue involvement in carbohydrate binding proteins  87  vii  List of Figures  Figure 1.1.  Structure of cellulose  2  Figure 1.2.  Known CBD structures  10  Figure 2.1.  Restriction map of pTugKNl  15  Figure 2.2.  Protocol for PCR/site-directed mutagenesis  18  Figure 3.1.  Expression and purification of CBDN1 from E. coli supernatants  26  Figure 3.2.  Expression and purification of CBDN1 from E. coli periplasms  28  Figure 4.1.  Reversibility of CBDN1 binding to PASA  31  Figure 4.2.  Isotherms of wild-type CBDN1 on different allomorphs of cellulose  34  Figure 4.3.  Temperature dependence of CBDN1 binding to PASA  37  Figure 5.1.  Summary of the CBDN1 residues targeted for mutation  43  Figure 5.2.  1-D NMR spectra of CBDN1 and alanine mutants  47  Figure 5.3.  1-D NMR spectra of CBDN1 and alanine mutants  48  Figure 5.4.  Isotherm data for the binding of CBDN1 and mutants to PASA  51  Figure 5.5.  Isotherm data for the binding of CBDN1 and mutants to PASA  52  Figure 5.6.  Affinity electrophoresis profiles of CBDN1 and mutants  54  Figure 5.7.  Analysis of affinity electrophoresis data for CBDN1 and mutants  56  Figure 5.8.  Analysis of affinity electrophoresis data for CBDN1 and mutants  57  Figure 5.9.  Affinity electrophoresis profiles on HEC and barley 0-glucan  62  Figure 6.1.  Labeled ! H - N HSQC spectrum of CBDN1  65  Figure 6.2.  ! H - N HSQC spectra of wild-type CBDN1, Y19A, and Y85A  66  Figure 6.3.  Chemical shift perturbations in Y19A and Y85A  69  Figure 6.4.  Raman spectra of CBDN1, Y19A, and Y85A  75  Figure 6.5.  Proximity of perturbed chemical shifts to W16 and W137  77  Figure 6.6.  V8a/V8b region  79  1 5  1 5  of the Raman spectra for CBDN1, Y19A, and Y85A  List of Abbreviations  CBD  Cellulose binding domain  PASA  Phosphoric acid swollen Avicel  BMCC  Bacterial microcrystalline cellulose  CMC  Carboxymethylcellulose  HEC  Hydroxyethylcellulose  HPMC  Hydroxypropylmethylcellulose  DP  Degree of polymerization  bp  Base pair  kbp  Kilobase pair  K  Kilodalton  Ka  Affinity constant  Kd  Dissociation constant  [B]  Concentration of bound protein  [F]  Concentration of free protein  No  Binding saturation constant  G  Garbage constant  P  Protein  L  Ligand  [LJF  Free ligand concentration  [L]T  Total ligand concentration  rm  Relative mobility  LB  Luria broth  TY  Tryptone, yeast extract medium  TB  Terrific broth  Kan  Kanamycin  Kan  r  Kanamycin resistance  lacK  lac repressor protein under constitutive expression  Lx  Leader peptide from the cellobiohydrolase, Cex  ori  Origin of replication  Ptac  tac promoter  IPTG  Isopropyl-P-D-thiogalactoside  SDM  Site-directed mutagenesis  DTT  Dithiothreitol  PCR  Polymerase chain reaction  DMSO  Dimethyl sulfoxide  KP  Phosphate buffer  pi  Isoelectric point  wt  Wild-type  rpm  Revolutions per minute  xg  Multiples of the earth's gravitational force  r  Correlation coefficient  2  X  wavelength  UVRRS  Ultraviolet resonance Raman spectroscopy  v  Nuclear vibrational frequency  SDS-PAGE  Sodium dodecyl sulfate - polyacrylamide gel electroph  BSA  Bovine serum albumin  lH NMR  Proton nuclear magnetic resonance  COSY  Correlated spectroscopy  NOESY  Nuclear Overhauser enhancement spectroscopy  HSQC  Heteronuclear single quantum coherence  ppm  Parts per million  ITC  Isothermaltitrationcalorimetry  Acknowledgments I would first like to thank my supervisors Doug Kilburn, Charles Haynes and Peter Tomme for their support during my tenure at UBC. Not only was their guidance instrumental in the completion of this work, but they also provided me with the autonomy to develop my abilities as a scientist and my confidence as an independent thinker. I am especially grateful to the cellulase lab (Helen, Emily Kwan, Emily Akow, Diane, Greg, Linda, Hans-Olaf, Brad, Dominik, Mustafa, and Al) for their knowledge and friendship. Kudos to Emily Kwan for the FPLC tutorial and for generously supplying a stock of PASA, Al Boraston for his brilliant insights and programming skills, Brad McLean for the constant help with computers, and Peter Tomme for aiding in the cloning of several mutants. I would also like to thank Phil Johnson and Shane Greek for their collaborations and technical advice in the fields of NMR and UVRRS respectively. Financially, I would like to thank NSERC for two years of monetary support and the "bank of Mom and Dad" for ensuring the presence of food on the table and beer in the fridge. Special thanks to Darlene Crowe and the UBC biotechnology laboratory for helping me out in a time of need. My time in BC has provided me with a second home. Numerous friends have injected much needed fun into my life from both BC and my other home in Fonthill, Ontario. I would like to thank those friends for all of their support and friendship over the years; and hey guys, its finally time to join you all in the real world!!!! Last, but certainly not least, I am especially grateful to my family whose love and encouragement has been a constant in my life; although seldom acknowledged, that support has always been cherished.  Chapter 1: Introduction  1.1  1.1.1  Cellulose and cellulose degradation  Structure of cellulose  The structure of insoluble cellulose can be considered at three basic levels (Attala, 1993). The first level is concerned simply with covalent bonding patterns while the second and third levels of organization include the spatial relationship of the repeat units to one another and the association of molecules into ordered aggregates. At this time, the first and second levels of cellulose structure are well understood but the tertiary structure is still largely unknown. The chemical structure of cellulose is well established. Cellulose is composed of glucose residues linked together by (3-1,4 glycosidic bonds to form a repeating polymer (Fig. 1.1a). Each glucopyranoside residue is rotated 180 degrees with respect to adjacent glucopyranoside neighbours, creating a twisted structure that is stabilized by hydrogen bonding between the OH-3 hydroxyl of one sugar and the heterocyclic oxygen atom of an adjacent monomeric unit (Nardin and Vincendon, 1988). Consequently, the repeating unit in cellulose is cellobiose and not glucose. Native cellulose isolated from plant or bacterial sources is synthesized as glucan polymers with a degree of polymerization (DP) ranging from 100 to 1000 (Cannon and Anderson, 1990; Beguin and Aubert, 1994). At these DPs, P-1,4 linked glucan polymers are not water soluble, and coalesce into crystalline microfibrils (Pereira et al., 1988; Schmid, 1988). Native cellulose produces an x-ray diffraction pattern characteristic of a single crystalline allomorph, cellulose I (Sarko, 1986). Upon various treatments, at least four other crystalline allomorphs (celluloses II, IIIT, TV\, and IVn) can be prepared (Kolpak et al., 1978; Kolpak and Blackwell, 1978; Sarko, 1986; Hayashi, 1989). Cellulose II, produced from sodium hydroxide treatment of cellulose I (Kolpak et al., 1978) may also be present in native cellulose to some extent (Atalla and Vanderhart, 1989).  Cellobiose  Glucose  Linkage  B Crystalline region I  Amorphous region \  Linkage  F i g . 1.1. A ) Primary structure of cellulose with the monomer (glucose), repeating dimer (cellobiose) and linkage indicated. B) Putative tertiary structure of insoluble cellulose illustrating differences in chain conformation between crystalline and amorphous regions. C ) Primary structure of the mixed (3-1,3 (3-1,4 linked glucan polymer, barley (3-glucan.  X-ray diffraction has been used to classify cellulose allomorphs by differences in bulk "order" or "crystallinity" (Kulshreshtha and Dweltz, 1973; Hatakeyama, 1989; Hsu and Penner, 1989; Hayashi, 1989).  The crystallinity scale ranges from near 0% for regenerated cellulose  (i.e. cellulose initially dissolved and then regenerated back to its insoluble form) to 100% for cellulose isolated from Valonia macrophysa (Kulshreshtha and Dweltz, 1973). Other cellulose preparations fall within those extremes. Bacterial microcrystalline cellulose (BMCC) and Avicel exhibit crystallinities of approximately 76% and 50% with respect to Valonia (Kulshreshtha and Dweltz, 1973; Wood, 1988). Both BMCC and Avicel are described as "crystalline" allomorphs of cellulose in literature. In contrast, phosphoric acid swollen Avicel (PASA) has a crystallinity between 0 and 5% (Lee and Kim, 1983), and is described as "amorphous" in structure. Despite the convenience of a crystallinity index for insoluble cellulose, X-ray diffraction is limited in its abilities to resolve microdomains of crystallinity within cellulose (Attala, 1993). For instance, in figure 1.1b, the chain conformations of cellulose allomorphs such as PASA may be represented by the "amorphous" region of the cellulose. This can be referred to as the "spaghetti" analogy to chain conformation. However, this may not be an accurate portrayal of the true chain conformation since pockets of crystallinity may exist in PASA which are too small to be detected by X-ray diffraction (Attala, 1993). Furthermore, X-ray diffraction cannot distinguish between the two main crystalline forms present in insoluble cellulose, cellulose I and cellulose Ip (Attala a  and VanderHart, 1989). At this level of resolution, the techniques of Raman spectroscopy and solid state C NMR are more useful for defining the intricacies of cellulose structure (Attala and 1 3  VanderHart, 1989; Attala, 1990; 1993). While native cellulose is always in the form of insoluble microfibrils, glucan chains can also be prepared as water soluble molecules (Pereira et al., 1988; Schmid, 1988; Hofman, 1993). The solubility of a glucan polymer is dependent on two characteristics:  the degree of  polymerization (DP) and glucan derivatization (Sau and Landoll, 1989). For underivatized (3-1,4 glucan polymers, solubility is a function of chain length. Above seven to eight glucan units, increasing noncovalent interactions cause the aggregation of single chains to form insoluble matrices.  Derivatization of the glucopyranoside units within a glucan chain can increase the  solubility by preventing the close contact of glucan chains along their lengths. As a consequence, many derivatized [3-1,4 glucans remain water soluble at DPs above 100.  Some examples of  water-soluble cellulose derivatives include carboxymethylcellulose (CMC), hydroxyethylcellulose (HEC), and hydroxypropylmethyl-cellulose (HPMC). Another glucan polymer, barley (3-glucan, remains soluble at DPs near 100 despite being nonderivatized (Fig. 1.1c). In this case, the mixed (3-1,3 (3-1,4 linkage nature of the glucan is believed to create kinks in the polymer, preventing close assembly and subsequent aggregation of the glucan chains (Perlin and Suzuki, 1962).  1.1.2  Degradation of cellulose  Numerous bacteria and fungi can degrade insoluble cellulose. Cellulolytic activity has also been discovered in protozoa (Coleman, 1978), slime molds (Jones et al., 1979) and plants (Tucker et al., 1987). The natural degradation of cellulose involves mixed cultures of bacteria and fungi, and can occur in both aerobic and anaerobic environments. In aerobic systems, the complete degradation of cellulose yields carbon dioxide and water. Anaerobic systems typically convert cellulose to carbon dioxide, water and methane (Beguin et al., 1987). Cellulose is degraded to cellobiose units by the synergistic action of several enzymes, collectively termed as cellulases (Woodward et al., 1988). Aerobic organisms typically use batteries of secreted cellulases which act to synergistically degrade cellulose, while anaerobic organisms utilize large multienzyme complexes called cellulosomes. Cellulosomes are localized on the cell wall and operate at the cell-substrate interface to degrade cellulose. In both systems, two types of cellulases are present. "Endoglucanases" attack internal (3-1,4 glycosidic links in the glucan chain releasing free ends. The newly formed termini are then attacked from either the reducing or nonreducing end by "cellobiohydrolases" to release cellobiose units. Cellobiose is the major product of the cellulolytic cascade, along with smaller amounts of cellodextrins (Mandels, 1982; Wakarchuk et al., 1984).  1.1.3  Cellulases of Cellulomonas fimi Cellulomonasfimiis a gram positive soil bacterium which can utilize cellulose as its sole  carbon source (Vladut-Talor et al., 1986). C.fimi produces at least four endoglucanases (CenA, CenB, CenC, CenD) and two cellobiohydrolases (CbhA, and CbhB). One additional enzyme; Cex, exhibits both endoglucanase and cellobiohydrolase activity (Whittle et al., 1982; Gilkes et al., 1984; O'Neill et al., 1986; Wong et al., 1986; Owalabi et al., 1988; Moser et al., 1989; Coutinho et al., 1991; Meinke et al., 1993; Shen et al., 1995a, b). All of the cellulases produced by C.fimi are modular proteins composed of two or more domains. Each cellulase contains a cellulose binding domain (CBD) at one terminus attached to a catalytic domain through either additional domains, short linker sequences or both.  1.2 1.2.1  CBD's General classification and properties  To date, more than 180 putative cellulose binding domains (CBDs) have been organized into 13 families by Tomme et al. (1995, 1998). In Tomme et al.'s system, CBD sequences are classified by sequence homology to known CBDs which demonstrate an affinity for cellulose. Secondary emphasis is placed on the functional properties of each CBD (Gilkes et al., 1991; Tomme et al., 1995).  Table 1.1. CBD families and their properties (adapted from Tomme et al., 1998).  Family  Size  Characteristics  I  33-36  Exclusively CBDs from fungal enzymes > 40 members NMR structure for CBDCBHI (Trichoderma reseei) (Kraulis et al., 1989) Some bind to chitin  II  ~100  Two sub-families (Ila and lib) Ua contains two chitin binding domains -40 members NMR structure for CBDCex (Cellulomonasfimi)(Xu et al., 1995) Some Ua CBDs bind chitin, Ub CBDs show affinity for xylan  III  130-170  Two sub-families (IHa and IHb) -25 members X-ray structure for CBDCip (Clostridium thermocellum) (Tormo et al., 1996) Some bind to chitin  IV  125-170  5-6 members NMR structure for CBDN1 (Cellulomonasfimi)(Johnson et al., 1996a) Does not bind crystalline cellulose  V  63  VI  85-90  CBDEGZ from Erwinia chrysanthemi only NMR structure solved (Emmanual Brun, personal communication) 6 members Low affinity for crystalline cellulose  Vm.  152  IX  170-180  CBDCelAfromDictostelium discoidem only 9 members,fromthermostable xylanases CBDs occur mainly as tandem repeats  X  50-55  XI  120-180  XII  ~50  XIII  40-45  7 members almost exclusivelyfromPseudomonas fluorescens 4 members, mainly from Clostridium and Fibrobacter >10 members, mainlyfromBacilli endoglucanses and chitinases Contains triple repeated CBDsfromxylanases and lectin like domains with different specificities  CBDs are usually found in a single copy at either terminus of the protein, or internally sandwiched between adjacent domains. In some cases, more than one CBD is present, either individually (Meinke et al., 1991; Black et al., 1995; Mill ward-Saddler et al., 1995) or as tandem repeats (Coutinho et al., 1992; Winterhalter et al., 1995; Karlsson et al., 1997). The majority of CBD sequences are either bacterial or fungal in origin, and range in size from 35 to 180 amino acids. Both size and phylogeny are well conserved properties within each CBD family. Fungal CBDs include the smallest CBDs and demonstrate enough sequence dissimilarity from bacterial CBDs to be classified into their own family, that being family I. The remaining families appear to be found in bacteria. Notable exceptions include families VI and VIII. These families each contain a single nonbacterial CBD representative from the fungus Microspora bispora and the slime mold Distyostelium discoidum, respectively. Table I includes a breakdown of the current CBD classification scheme, noting the various physical and functional characteristics pertaining to each family (Tomme et al., 1998). The majority of CBDs are from cellulase enzymes. CBDs have also been identified in 01,4 xylanases (Millward-Saddler et al., 1994; 1995; Black et al., 1995; Winterhalter at al., 1995; Clarke et al., 1996), other hemicellulases (Vincent et al., 1997; Black et al., 1997), and at least one mannanase (Stahlbrand et al., 1995). In addition, CBDs have been found in nonhydrolytic proteins that are components of cellulosome complexes (Poole et al., 1992; Shoseyov et al., 1992; Gerngross et al., 1993; Morag et al., 1995; Pages et al., 1997).  1.2.2  Function of C B D s  Despite the large number of CBD sequences that have been identified, only a handful of CBDs have been characterized. Consequently, the various properties of CBDs are still being elucidated. Thus far, members of a single CBD family appear to share similar functional properties, although subtle differences in binding specificity do occur. For example, CBDCBHI and CBDCBHII, two family I CBDs, both bind strongly to cellulose.  However, only  CBDCBHII binds to chitin (Under and Teeri, 1996). Interestingly, many of the CBDs present in xylanases and hemicellulases bind exclusively to cellulose and show no affinity for xylan or other  hemicellulosics. In this case, CBDs are thought to localize the enzyme onto the plant cell wall (Black et al., 1997) where it has access to the hemicellulosics that are naturally intermixed with the cellulose (Attala, 1993). Most CBDs bind to both crystalline and amorphous cellulose although some CBDs are allomorph specific. For instance, CBDN1 (family IV), CBDN2 (family IV) and CBDXynA (family VI) preferentially bind to amorphous cellulose (Tomme et al., 1995; Sakka et al., 1996). In addition, the binding specificity of a CBD is not always limited to cellulosic compounds. Some CBDs from families I, U, III, and IV are capable of binding to chitin (Ong et al., 1993; Morag et al., 1995; Under and Teeri, 1996; Garda et al., 1997; Tomme et al., 1996), barley fJ-glucan (Tomme et al., 1996) or xylan (Black et al., 1995; Garda et al., 1997; Vincent et al., 1997). One well established characteristic of CBDs is their ability to increase the rate of cellulose hydrolysis by localizing cellulases onto cellulose. In fact, CBDs are essential for the efficient hydrolysis of natural cellulose. Isolated catalytic domains exhibit activities that are similar to the intact enzymes on soluble substrates, but are severely impaired on insoluble substrates (Van Tilbeurgh et al., 1986; Tomme et al., 1988; Gilkes et al., 1991; Coutinho et al., 1993; Reinikainen et al., 1995; Hall et al., 1995; Garda et al., 1997). CBDs can also function as independent polypeptides when separated from their catalytic domains (Gilkes et al., 1988, 1989; Tomme et al., 1988; Coutinho et al., 1992; Under and Teeri, 1996). This independent binding ability has led to some important practical applications of CBDs. Fusion proteins can be constructed which consist of a CBD fused to a heterologous domain. Usually, these fusion proteins retain the activity of both fusion partners, making them useful for a variety of practical applications including protein purification, enzyme immobilization, and diagnostics (Greenwood et al., 1989, 1992, 1994; Ong et al., 1989, 1991). Although it is still the subject of debate, one unusual property displayed by some CBDs is the ability to disrupt the surface of insoluble cellulose by nonhydrolytic means. This property has been illustrated by CBDCenA and CBDCex, both family II CBDs (Din et al., 1991; Ong et al., 1993). However, surface modification has not been seen by any other CBDs to date, including at  least two other family II CBDs for which this type of behaviour was investigated (Black et al., 1997; Linder and Teeri, 1997).  1.2.3  Structure/function studies on C B D s  In order to understand how a CBD interacts with cellulose, it is useful to determine the three dimensional structure of the CBD. A total of six CBD structures have been solved. These include the crystal structures of CBDCipB from Clostridium thermocellum (Tormo et al., 1996) and CBDE4 from Thermomonospora fusca (Sakon et al., 1997), as well as the solution structures of CBDCBHI from Trichoderma reseei (Kraulis et al., 1989), CBDCex (Xu et al., 1995) and CBDN1 (Johnson et al., 1996a) from Cellulomonasfimi,and CBDEGZ from Erwinia chrysanthemi (Brun et al., 1997). Five of these structures, representing members of CBD families I through V are included in figure 1.2. The sixth structure, that of CBDE4, is a second representative from family III. The structure of CBDE4 is very similar to that of CBDCipB shown in figure 1.2. In addition, the NMR structure for CBDN2 of C.fimi, a second family IV member, is nearing completion (Emmanual Brun, personal communication). This CBD appears to be similar in structure to the other family IV representative, CBDN1. Taken together, these data imply that the backbone fold of different CBDs within a given family is conserved. Such findings should not be surprising considering the high homology shared by members of each CBD family. A comparison of the six available CBD structures indicates two classes of carbohydrate binding site. The CBDs from CBHI, Cex, EGZ, E4, and CipB all exhibit a planar surface lined with solvent exposed aromatic residues. The role of these aromatic residues has been probed in the case of CBDCBHI and CBDCex. With CBDCBHI, the introduction of alanine at either Y5 or Y32 eliminated binding to BMCC (Linder et al., 1995). However, it should be noted that in the case of the Y5A mutant, conformational perturbations were identified by two dimensional NMR spectroscopy. Mutation of Y31 to alanine also resulted in a marked decrease in binding affinity (Linder et al., 1995).  Fig. 1.2. Illustration of the five known C B D structures with family classification shown in parentheses: A) C B D N 1 (family IV) B) CBDCipB (family III) C) C B D E G Z (family V) D) CBDCBHI (family I) E) CBDCex (family Ha). Putative binding site residues are highlighted in red or green.  The binding face of CBDCex contains three solvent exposed tryptophan residues (W17, W54, W72). Chemical oxidation of these tryptophans (Bray et al. 1996) demonstrated their importance in binding to BMCC. Mutational analysis is currently underway to determine the relative contribution of each tryptophan to binding.  Preliminary evidence indicates that the  alanine mutants of these tryptophans result in affinity decrements between 10 and 100-fold on BMCC (Brad McLean, personal communication). Mutational analysis of CBDCenA, another family II CBD, has shown that alanine mutation of W14 and W68 respectively, results in a 50 and 30-fold reduction in binding affinity on BMCC (Din et al., 1994). Although the structure of CBDCenA has not been solved, fluorescence data indicate that W14 and W68 are surface exposed residues (Din et al., 1994). In the case of CBDCBHI, CBDCex and CBDCenA, solvent exposed aromatic residues lining a planar region of the CBD constitute a carbohydrate binding site. Given the similarities in binding site structure between these CBDs and others for which a structure has been solved, it is likely that aromatic residues also mediate the binding of family III and family V CBDs to crystalline cellulose. Many of the aromatic residues in the binding sites of CBDCex and CBDCBHI are highly conserved within their respective families. Additional polar residues, in close proximity to the aromatic residues, are also highly conserved. In the case of CBDCBHI, one of those polar residues (Q34) was identified as a major contributor to binding affinity (Linder at al., 1995). Other conserved polar residues may also make important contributions to binding. One hypothesis is that they form intramolecular hydrogen bond networks, resulting in the presentation of a rigid binding face (Linder and Teeri, 1997). Another possibility is that the polar residues interact directly with cellulose through intermolecular hydrogen bonding and/or polar van der Waals contacts. At this time, it is impossible to rule out either possibility since none of the available CBD structures present a clear picture of binding site interactions with substrate. The primary reason for this is because none include bound cellulose in the solved CBD structure.  1.2.4  CBDN1  The structure of the family IV CBD, CBDN1, differs noticeably from the other CBD structures in figure 1.2. CBDN1 is the only CBD without a planar binding face. Instead, the three dimensional structure of CBDN1 shows a distinctive cleft (Johnson et al., 1996a). This cleft was identified as the carbohydrate binding site by NMR titration studies (Johnson et al., 1996b). Because CBDN1 has a binding cleft, it is not surprising that CBDN1 shows little affinity for crystalline cellulose. Instead, CBDN1 preferentially binds to amorphous forms of insoluble cellulose like PASA (Coutinho et al., 1992; Tomme et al., 1996). Furthermore, consistent with the presence of a cleft instead of a binding face, CBDN1 has also been shown to associate with a wide range of soluble cellulosics, including hydroxyethylcellulose (HEC), hydroxypropylmethylcellulose (HPMC), carboxymethylcellulose (CMC), oat (3-glucan, barley (3-glucan and cellooligosaccharides with DPs longer than three (Johnson et al., 1996b; Tomme et al., 1996). Binding affinities increase markedly as the cellooligosaccharide length is increased between cellotriose and cellopentaose. However, at chain lengths greater than five (i.e. cellohexaose), no increase in binding affinity is observed, indicating that five saccharide units span the CBDN1 binding site (Tomme et al., 1996). A detailed analysis of the thermodynamics of CBDN1 binding to several different cellulosics shows that the binding to PASA, cellooligosaccharides and soluble derivatized cellulosics is enthalpically driven with an unfavourable entropic contribution (Tomme et al., 1996). This contrasts sharply with the thermodynamic findings for CBDCex, the only other CBD for which a thermodynamic study has been attempted. In the case of CBDCex, binding to BMCC is entropically driven with dehydration at the binding interface believed to be the major driving force for binding (Creagh et al., 1996). Despite the wealth of structural and functional information already compiled on CBDN1, many questions regarding binding still remain unanswered. In the absence of a CBDN1 structure complexed with cellulose, binding site interactions are undefined. Specifically, it is not known which residues are involved in binding or what their relative contributions to binding event are. Nor can we identify any of the intermolecular forces (i.e. hydrogen bonding) mediating the  interaction of individual residues with cellulose.  Deciphering such interactions is important for a  detailed understanding how CBDN1 interacts with cellulose, and consequently for how to optimize the use of CBDN1 in practical applications.  1.3  Thesis Objectives The main objective of this thesis was to characterize the binding site of CBDN1 by site-  directed mutagenesis.  To achieve this, several secondary objectives had to be completed  including: i) the development of an expression/purification protocol capable of purifying suitable quantities of CBDN1 and mutants for quantitative analysis; ii) an investigation of the binding properties of wild-type CBDN1; and iii) the development of suitable systems to quantify binding on soluble and insoluble substrates.  Chapter 2: Materials and Methods 2.1 Materials Avicel PH101 was obtained from FMC international. Bacterial microcrystalline cellulose (BMCC) was harvested from cultures of Acetobacter xylinum, as described by Hestrin (1963). Barley [J-glucan and hydroxyethylcellulose were purchased from Sigma and Aldrich respectively. Phosphoric acid swollen Avicel (PASA) was prepared using a protocol from Wood (1988). First, 50 g of Avicel PH101 was added to 250 ml of 85% phosphoric acid. The mixture was blended briefly to partially dissolve or "swell" the cellulose, and incubated on ice for 60 min with occasional stirring. The suspension was then transferred into 4 1 of ice-cold water and allowed to precipitate overnight. Finally, the precipitate was washed once with cold water and resuspended in a convenient volume of 50 mM potassium phosphate buffer, pH 7. Regenerated Avicel was prepared by the same procedure as that used for PASA, but with the following exceptions. First, the Avicel was completely dissolved in 85% phosphoric acid during the preparation of regenerated cellulose. Second, the regenerated cellulose was precipitated with acetone instead of water.  2.2  Bacterial strains, vectors, and growth conditions Escherichia coli strains JM101 (supE, thi-1, A(lac-proAB), [F'traD36, proAB,  laclqZAMIS]) (Yanisch-Perron et al., 1985) and DH5ct (F-;, endAl, hsdR17(rk, mk ), +  supE44, thi-1, recAl, gyrA96, relAl, (argF-laczya)U169, 0lacZ M15) (Hanahan, 1983) were used as host strains for the maintenance of plasmids and the production of recombinant protein. E. coli DH5a was the strain of choice in preparing DNA for cloning and sequencing. E. coli JM101 was used exclusively for protein production. All cbdnl gene fragments were cloned into the pTugK vector (Fig. 2.1). Important components of this vector include the constituitive expression of both a kanamycin resistance marker (Kan ) and the lac repressor protein (lacl), a tac promoter (Ptac) and glO translational R  ft ori  pER322 ori  Fig. 2.1. Restriction map of pTugKNl, the plasmid used for all cloning and expression experiments with C B D N 1 and mutants. PCR products were cloned into the pTugK vector via two unique restriciton sites: a) Nhel b)HindIII  enhancer for the expression of cloned genes, and a cex leader sequence (Lx) for secretion of the gene product. The construction of pTug was described previously (Graham et al., 1995). Bacterial cultures of E. coli DH5ct and JM101 were prepared from single plated colonies or frozen cell stocks containing glycerol at a final concentration of 10%. These cultures were grown at 30°C or 37°C in liquid tryptone/yeast extract media (TY, 6 g tryptone, 16 g yeast extract, 5 g NaCl per litre) or terrific broth (TB, 12 g protein hydrolysate, 23.9 g yeast extract, 9.4 g K2HPO4,  2.2 g K H 2 P O 4 , 8 ml glycerol per litre).  All cultures were supplemented with  kanamycin to a final concentration of 100 (Ig/ml. Isopropyl-P-D-thiogalactoside (IPTG, Sigma), at a final concentration of 0.2 mM, was used to induce gene expression.  2.3  D N A isolation and manipulation  DNA was isolated using the Qiagen DNA isolation kit (Qiagen) and transformed into competent host cells by electroporation. Various molecular cloning techniques were performed by the appropriate enzymes, as directed by suppliers. For digestions, 0.5-1.5 u,g of DNA were incubated with 5-10 units of restriction enzyme for 2-16 h. DNA fragments were resolved by electrophoresis using 0.8% agarose gels. Desired fragments were recovered from agarose gels using the Qiagen Qiax II Gel Extraction Kit (Qiagen). Ligations were performed with 100-500 ng of total DNA at molar insert to vector ligation ratios of approximately 10:1. 10-20 units of T4 DNA ligase (NEB) were used per reaction in 25 to 40 ul total reaction volume. Incubations generally lasted 16 h at 4°C, or alternatively, for one hour at room temperature followed by two hours at 4°C. All ligation reactions were desalted by butanol precipitation prior to transformation.  2.4  Polymerase chain reactions  Oligonucleotide primers were designed using the Oligo software package (Nucleic Acid Protein Sequencing (NAPS) unit, University of British Columbia). Silent site mutations were designed into the mutagenic primers with the Reverse Translator program (Ron Trimbur,  University of British Columbia, Vancouver, Canada). All oligonucleotides were synthesized by the NAPS unit (University of British Columbia). Each polymerase chain reaction (PCR) mixture (50 ul total volume) contained 10-100 ng template DNA, 10-150 pmol primers, 10% DMSO, 0.2 mM 2'-deoxynucleotide 5'-triphosphates and 1 unit Vent polymerase (New England Biolabs) in 10 mM KC1, 10 mM (NrL^SCU, 20 mM Tris-Cl, 2 mM MgSC»4, and 0.1% Triton X-100. Twenty cycles were performed as follows: denaturation at 95°C for 30 s, annealing at 55°C for 30 s, and primer extension at 72°C for 30 s. All of the cbdnl variants were constructed by PCR/site-directed mutagenesis. This procedure requires four PCR primers and a total of three PCR reactions to construct the mutagenic gene fragment (see Fig. 2.2). In the first PCR reaction, one mutagenic primer and one flanking primer are used to amplify one segment of the cbdnl coding sequence.  In a second PCR  reaction, the complimentary mutagenic primer and the opposite flanking primer are used to amplify the remaining segment of cbdnl. The two PCR products are then pooled as a template for the third PCR reaction. In this case, the flanking primers are used to complete the amplification of the mutant cbdnl coding sequence. Products from the third PCR reaction were purified (Qiagen), digested by restriction enzymes (Nhel and Hindlll) within the flanking primer sequences, and ligated into the gel purified pTugK vector fragment. The constructs were then transformed into E. coli DH5a and screened for mutant plasmid by restriction enzyme analysis. All positive clones were verified by DNA sequencing (NAPS units, University of British Columbia) prior to expression. In the gene fragments encoding two of the CBDN1 mutants (N50A, D90A), the availability of convenient restriction sites permitted the use of a two primer system. In this case, one flanking primer was combined with a mutagenic primer extending into a unique cloning site. The PCR product of this single reaction could then be cloned directly into the pTugK vector.  P C R rpartinn -\  PCR reartinn ?  Mute 1  N137 N1236  Mute 2  20 Cvcies 95°C, 30 seconds 55°C, 30 seconds 72°C, 30 seconds  PCR reaction 3  20 Cycles 95°C, 30 seconds 55°C, 30 seconds 72°C, 30 seconds  1) Purify, dilute and combine fragments 2) Denature (95°C) and anneal (55°C)  1) Extend fragments (72°C, 54 seconds) 2) Add flanking primers  N137  N1236 20 Cycles 95°C, 30 seconds 55°C, 30 seconds 72°C, 30 seconds  Fig. 2.2. Protocol for PCR/site-directed mutagenesis using the four primer technique. The two mutagenic primers (Mute 1 and Mute 2) are included in bold and the flanking primers (N1236 and N137) are included in regular type text. Note that the entire pTugKNl plasmid was used as the template for PCR reactions one and two while the product of PCR reactions one and two was used as template for PCR reaction three.  2.5  Expression and purification of CBDN1 and mutants from E. coli JM101  periplasms A l l CBDN1 mutants (and wild-type CBDN1) were expressed and purified by the same procedure.  E. coli JM101 frozen cell stocks were used to inoculate 2 ml of T Y media  supplemented with kanamycin at 100 |ig/ml. These cultures were grown overnight at 30°C until the A600nm was between five and eight. 0.8 ml of the overnight culture was then used to inoculate 400 ml of T Y media containing 100 u.g/ml kanamycin. The large volume cultures were incubated at 30°C with shaking at 250 rpm, and monitored for growth. At mid-log phase (A600nm between 0.1 and 0.5), the cultures were induced with IPTG at a final concentration of 0.2 mM. After induction, the cells were incubated for 24 hours at 30°C with shaking at 250 rpm. Protein was recovered from the periplasm using a modified osmotic shock procedure. Cells were harvested by centrifugation (5000 xg for 10 min, 4°C), and osmotic shock buffer (30 mM Tris, pH 8, 1 mM E D T A , 20% sucrose) added to 8% total culture volume. The cells were then resuspended by vortexing, incubated for 15 min at room temperature with shaking, and pelleted by centrifugation again (11000 xg for 10 min, 4°C). The supernatant was discarded and the cell pellet resuspended in 5 mM MgSOa to 4% of the initial culture volume. After a 20 min incubation at room temperature, the cells were pelleted by a low speed centrifugation (11000 xg for 5 min, 4°C), and the supernatant collected. Lastly, the supernatant was subjected to a high speed centrifugation (31000 xg for 5 min, 4°C) to remove any fine particulate matter prior to chromatography. Proteins were resolved from the osmotic shock fractions by anion exchange chromatography. The osmotic shock fractions were applied to a XK16 (1.6 X 20 cm) column (Pharmacia) packed with 10 ml of macro Q strong anion exchange resin (BIO-RAD). Columns were equilibrated in 25 mM potassium phosphate/acetate buffer, pH 5.2; elution was performed with a linear gradient of NaCl (0-1 M , 90 min) in the same buffer at a flow rate of 1.25 ml/min. Purity was judged by sodium dodecyl sulfate - polyacrylamide gel electrophoresis (SDSPAGE, BIO-RAD).  20 ul samples were boiled with 10 ul SDS-PAGE loading buffer (4% SDS,  20% glycerol, 125 mM Tris, 0.0025% bromophenol blue, 10% fi-mercaptoethanol), and analyzed on 13% SDS-PAGE gels.  2.6  Calculation of affinity constants on insoluble cellulose Affinity constants on insoluble cellulose were determined from depletion isotherms. All  isotherms were composed of at least 20 data points, each data point corresponding to one binding reaction carried out in a 1.5 ml eppendorf tube.  Each binding reaction contained 1 mg of  cellulose in the presence of 1-25 uM CBD, with binding buffer (50 mM potassium phosphate, pH 7) added to a final volume of one ml. Binding reactions were performed in duplicate for each of the 10 different protein concentrations. Ten control tubes were also prepared. The control tubes consisted of protein and buffer but no cellulose, and were used to quantify the total protein concentrations. • The eppendorf tubes were rotated end-over-end at 4°C for 4 h. The cellulose was pelleted by two successive centrifugation steps (18000 xg, 10 minutes, 4°C). The concentration of unbound protein ( F ) was determined spectrophotometrically (A280nm) using an extinction coefficient of 0.016926 c n r ^ M " (Peter Tomme, personal communication). The quantity of 1  bound protein (B) was calculated from the difference in free and total protein concentrations. Total bound protein was then plotted versus free protein to generate an adsorption isotherm. The equilibrium association constant was determined by nonlinear regression (GraphPad Prism software) of the isotherm data with the following Langmuir-type binding model:  [  where K  B  ]  _  [ N o ] • K ( [ F ] - G)  "  1 + K a ( [ F ] - G)  a  fi  "  G  m (  1  )  is the association constant (M ), [B] is the concentration of bound protein (umol/g _1  a  cellulose), [ F ] is the concentration of free protein (|iM) and N  0  is the total concentration of  binding sites (|imol/g cellulose). The "G" variable is included as a control for both optical scattering and instrument imprecision (Refer to section 4.2 for a discussion regarding the "G" variable).  21 Solely for the purpose of graphing, both bound and free protein concentrations were corrected for the "G" variable by translating all [B] and [F] coordinates by the magnitude of "G". Curves depicting the modified data (which now pass through the origin) then follow the classic Langmuir binding equation (Eq. 2). Note that the regressed values of K and N from equation 2 a  0  are identical to those regressed using equation 1.  r R 1  [ O J  2.7  _ [No] • K , • TF1  "  1+  Ka  • [F]  w  Calculation of affinity constants on soluble glucan polymers Affinity constants on soluble polymers (i.e. barley (J-glucan) were determined by affinity  electrophoresis (Takeo and Nakamura, 1972). All electrophoresis experiments were performed using the Modular Mini-Protean II Electrophoresis System (BIO-RAD) with 0.75 mm spacers and combs. 8.8.  Separating gels were 5 cm in height and contained 13% polyacrylamide in Tris-HCl, pH  Stacking gels, 1 cm in height, were 4% acrylamide in Tris-HCl, pH 8.8.  Soluble  cellulosics were only added to the separating gels. Electrophoresis was carried out at 65 volts in an ice bath and continued until the bromophenol blue (BPB) dye front was less than 0.5 cm from the bottom of the separating gel. The binding temperature was taken as the temperature of the running buffer within the inner electrophoresis chamber. Electrophoresis was carried out on each mutant with at least eight polyacrylamide gels. Two gels devoid of polymer were used as controls; the other gels were set up in duplicate with three or more different concentrations of polymer. All solutions of polymer were prepared from powder immediately before use. Protein samples were composed of purified protein in loading buffer (18% glycerol, 18% bromophenol blue, 120 mM Tris-HCl, pH 8.8), and were loaded as 10 (il aliquots containing 3 u\g of protein. 3 |ig of acetylated bovine serum albumin (BSA) was included in each gel as a control. All gels were stained (2.5 g/1 Coomassie brilliant blue R-250 (Sigma) in 40% methanol,10% acetic acid) for 20 min and destained (40% methanol, 10% acetic acid) until bands  were clearly visible. After drying, the relative migration distance (rm) was calculated as the ratio of the distance migrated by protein to that of the tracking BPB band. Affinity constants were determined based on the theory of affinity electrophoresis developed previously (Takeo and Nakamura, 1972; Takeo et al, 1975; Takeo and Kabat, 1978). Application of the affinity electrophoresis theory requires that the concentration of the ligand is very large relative to the protein.  The theory also assumes that an equilibrium is established  within the gel, where: PL H P + L  (3)  and Kd = [P][L]/[PL]  (4)  Kd is the equilibrium dissociation constant. In addition, affinity electrophoresis theory assumes that the protein-polymer complex has a mobility of zero. If the above conditions apply, the general relationship between relative migration distance, and protein and polymer concentration is as follows (Takeo and Nakamura, 1972; Takeo and Kabat, 1978): rmi _ [P] rmo - ] + [PL] [P  (c\  V>  rmi and rmo are the relative migration distances of protein in the presence and absence of polymer; [P] and [PL] are the concentrations of free protein and the protein-polymer complex respectively. Because the concentration of polymer is large relative to that of the protein, the total concentration of polymer, [L]T, is nearly equal to that of free polymer, [L]p. Thus, [L]p = [L]j - [PL] = [L]x, and equations 4 and 5 can be combined to give:  K K d  =  [LjT-rmi (rmo - rm)  .\  ff ( 6 )  This equation can be transformed into:  j . rrrii  =  _i_fi + i y i ] rm I Kd ' 0  ) ^ '  ( 7  so that a plot of l/rmj against [L]T yields a straight line with an x-intercept equivalent to -Kd, or -1/K . The units of this binding constant were converted from per molar substrate to per molar a  binding site considering that four molecules of CBDN1 bind to a single polymer of barley (3glucan at saturation (Tomme et al., 1996).  Chapter 3: Expression and Purification of CBDN1 3.1 Background In previous work, wild-type CBDN1 was expressed in E. coli JM101 and purified by affinity purification on Avicel (Coutinho et al., 1993; Tomme et al., 1996). However, this procedure is not applicable for mutants with an altered affinity to cellulose. This chapter presents work on the development of a novel expression/purification system for CBDNL Two fundamental strategies are investigated. The first involves the purification of protein from culture supernatants while the second involves purification of protein from the periplasm of E. coli. Both strategies have proven to be successful in the purification of numerous proteins from E. coli (Hogset et al, 1990; Shen et al., 1991; Saya et al., 1993; Chowdhury et al., 1994; Robbens et al., 1995; Yamamoto and Ferretti, 1995; Tews et al., 1996).  3.2  Initial considerations for protocol development The development of an expression/purification protocol begins with our understanding of  E. coli physiology. In E. coli, approximately 90% of the total protein is retained within the cytoplasm while less than 10% is targeted for secretion into the periplasmic space (Beacham, 1979; Holland et al., 1986). Consequently, by directing CBDN1 into E. coli's secretory pathway and harvesting only secreted fractions, a 10-fold increase in protein purity can be attained, thereby easing downstream processing requirements. This was done by using an expression vector which fused the C.fimi exoglucanase leader sequence to the N-terminus of CBDN1 (Coutinho et al., 1992). During translocation across the plasma membrane, the leader peptide is cleaved to regenerate the native N-terminus (Wickner et al., 1991). Correct processing of the leader sequence was verified by N-terminal amino acid sequencing (NAPS unit, UBC) of the purified CBDNL  3.3 Purification of CBDN1 from E. coli  supernatants  Initial attempts to develop a purification protocol utilized the inability of E. coli to prevent some periplasmic proteins from leaking into the surrounding supernatant (Guo et al., 1988). A pTug vector exhibiting kanamycin resistance (Fig. 2.1) was used in combination with E. coli JM101 for all studies on the optimization of protein production. Experiments were designed to determine which combinations of media, temperature, induction time, and inducer concentration resulted in maximal protein production in the culture supernatant.  All expression studies were  done in 125 ml Erlenmeyer flasks. Figure 3.1(a) illustrates the results of one such experiment. From the SDS-PAGE gel where individual lanes represent supernatant profiles, it can be seen that the combination of terrific broth (TB), a 32 h induction, and 30°C culture temperature resulted in maximal CBDN1 yield within the supernatant. Attempts to purify CBDN1 from concentrated supernatant by column chromatography resulted in low yields due to saturation of the resin (Macro Q strong anion exchange, BIO-RAD). Yields were increased by introducing an ammonium sulphate precipitation step prior to the chromatographic separation. First, unwanted proteins were removed from the supernatant by adding ammonium sulphate at a concentration of 45% saturation in water. Then, to precipitate CBDN1, additional ammonium sulphate was added to 75% saturation; CBDN1 was found to precipitate at an ammonium sulphate concentration of 55% saturation (data not shown). The final precipitate was collected, resuspended in column equilibration buffer (25 mM potassium phosphate/acetate buffer, pH 5.2), concentrated, buffer exchanged, and loaded onto a 2.6 x 20 cm column (Pharmacia) packed with 50 ml of macro Q strong anion exchange resin (BIO-RAD). The bound protein was eluted with a linear gradient of NaCl (0-500 mM, 130 min) in the same buffer at a flow rate of 3 ml/min. Fractions of a typical purification are included in figure 3.1(b). Based on Bradford assay and A280 measurements, approximately 250 mg of protein could be purified from one litre of culture.  Pooled CBDN1 eluent, as estimated from SDS-PAGE gels, was 90-95% pure.  However, all CBDN1 samples purified in this manner were contaminated with a yellow colour, resistant to both on column washing with equilibration buffer, and buffer exchange.  26  1  2  3  4  5  6  IS  CBDN1  B M1  31K — •  22K  r  2 3 4 5 6 7 8 9 10 11 12  *  *  CBDN1  Fig. 3.1. 16% SDS/PAGE gels of expression and purification results for CBDN1 from E. coli JM101 supernatants. Lane M is a molecular weight marker A) 16% SDS/PAGE gel of CBDN1 expression under different growth and induction conditions. Lanes 1-6 include: 1) Terrific Broth (TB), 30°C growth and induction with 0.2 m M IPTG (30°C), harvest at 23 h post induction (23h), 2) TB/37°C/23h, 3) Tryptone/yeast extract media (TY)/37°C/32h, 4) TB/30°C/32h, 5) TB/37°C/32h, 6) TY/30°C/32h. B) Post ion exchange fractions of CBDN1 purification from culture supernatants. Lanes 1-12 represent fractions from a linear salt gradient over 150-350 m M NaCl in 25 m M phosphate/acetate buffer, pH 5.2.  The protein had wild-type CBDN1 binding activity as determined by binding isotherms, but proved to be unacceptable for isothermal titration calorimetry (ITC), resulting in endotherms instead of the characteristic exotherms usually observed during titration of CBDN1 with substrate (Tomme et al., 1996).  3.4  Purification of C B D N 1 from E. coli  periplasms  Due to concerns about the purity of CBDN1 isolated from culture supernatants, an alternative expression/purification protocol was developed.  In this procedure, protein was  harvested from the periplasm of viable cells, thereby preventing interaction of the protein with any media components.  As a first step, expression conditions were reoptimized. In this case,  conditions which suppressed leakage of the protein were desirable. The combination of TY media and low culture temperatures (28 - 30°C) proved to be best at expression and subsequent retention of CBDN1. Temperature had the greatest impact on the localization of expressed CBDN1 (Fig. 3.2c). At temperatures of 30°C or below, the majority of protein was retained within the periplasm. At temperatures above 30°C, increasingly more CBDN1 leaked from the periplasmic space into the supernatant. This leakage phenomenon is poorly understood (Blight et al., 1994), but has been observed with other C.fimi cellulases in E. coli (Guo et al., 1988). Inducer concentration had little impact on expression (Fig. 3.2b). 0.1 and 1.0 mM concentrations of D?TG resulted in similar yields of recombinant protein. In order to purify CBDN1, periplasmic fractions were harvested via an osmotic shock procedure and purified by anion exchange chromatography (see section 2.5). All purified proteins were colourless and in excess of 95% purity (Fig. 3.2a). As a final step, each protein preparation was exchanged into 50 mM potassium phosphate buffer, pH 7 and stored at 4°C for subsequent analysis.  Correct processing of purified CBDN1 was verified by N-terminal  sequencing (NAPS unit, UBC).  28  A  B M l  2  3  4  5  6  M  1  2  3  4  m  —.  CBDN1  CBDN1  C 1 2 3 4 5 6  7 8 9  10 11  CBDN1  Fig. 3.2. 16% SDS/PAGE gels of A) 28 hour cultures of £ coli JM101 grown in T Y media at 30°C and induced for 24 hours. Lanes 1 and 2 include CBDN1 purified from the periplasm, lanes 3-6 include cells + supernatant, sucrose fraction, osmotic shock fraction, and supernatant fractions. Lane M is a molecular weight marker B) Influence of IPTG concentration on CBDN1 periplasmic yields. Lane M , 5ug of purified CBDN1. Lanes 1-4, osmotic shock fractions of 28 hour cultures grown at 30°C in T Y media and induced with 0.2 (lane 1), 0.5 (lane 2), 1 (lane 3) and 0.1 m M (lane 4) IPTG. C) Effect of temperature on CBDN1 localization. Lanes 1 and 2 include samples from the sucrose fraction (lane 1) and periplasmic fraction (lane 2) of an osmotic shock treatment on cultures grown and induced at 28°C. Lanes 3-5, the same samples (supernatant fraction is also included as lane 3) from a culture grown and induced at 30°C, lanes 6-8 from a culture grown at 32°C, and lanes 9-11 from a culture grown at 37°C.  3.5  Summary  A comparison of the two purification strategies provides for several interesting insights. Both protocols resulted in purities above 90% despite differences in protein yield. CBDN1 yields from the supernatant purification were considerably higher (250 mg/1) than those in the periplasm (40 mg/1). Despite the lower yield, periplasmic protein was the preferred choice for quantitative characterization. This choice was due primarily to the ITC results. In this case, the titration of supernatant purified CBDN1 with substrate yielded endotherms, not the characteristic exotherms observed before for the same titrations (Tomme et al., 1996).  Although this negative result was  not investigated, a potential explanation is that small molecules were displaced from the surface of the protein upon binding, resulting in a net increase in enthalpy upon association of CBDN1 with substrate. In retrospect, the purification of a polysaccharide binding protein from rich media containing numerous polysaccharide components (Difco) is not a good practice. Contamination of the final product with adsorbed sugars is a majorriskthat is difficult to test for, as compared to the identification of protein contaminants by SDS-PAGE.  And, while a more elaborate  purification protocol could probably remove any contaminating sugars, due to the sheer number of purifications involved in completing this work, elaborate purification schemes were undesirable. For these reasons, avoidance of the rich media altogether by purifying protein from the periplasm was deemed to be beneficial. Furthermore, in the case of this study, yields of 40 mg/1 were adequate to perform all the characterizations presented in this research.  30  Chapter 4: Binding Studies on Wild-type CBDN1 4.1 Background The binding specificity of CBDN1, as well as its pH dependence, salt dependence, and thermodynamics have been addressed in previous papers for CBDN1 (Coutinho et al., 1992; Tomme et al., 1996; Creagh et al., 1998). This work supplements and extends the existing data on CBDN1 binding in three areas; namely, binding reversibility, allomorph specificity, and the temperature dependence of binding. Affinity and saturation constants describing CBDN1 binding to insoluble cellulose will be presented, as will preliminary work on the use of CBDs to characterize cellulose structure.  4.2  Demonstration of reversible binding to PASA CBDN1 binds to PASA reversibly, establishing equilibrium in less than two hours (Fig.  4.1). The reversibility experiment was performed as follows. First, a depletion isotherm was acquired on PASA. At two different points along the curve, the Eppendorf samples constituting those data points were retained. Each sample was centrifuged to pellet the cellulose (and bound CBD), and 800 ul of solution (containing free protein) was decanted and replaced by 800 ul of fresh buffer. The Eppendorfs were then rotated end-over-end at 4°C for two hours to re-establish equilibrium. The concentration of bound and free CBD was determined spectrophotometrically (see section 2.6), and plotted with the ascending isotherm in figure 4.1. If CBDN1 had bound irreversibly to PASA, one would expect the amount of bound protein to remain constant, and the concentration of free protein to decrease in proportion to the dilution factor. This would create a data point at the same concentration of bound CBD, but displaced toward a lower free protein concentration in figure 4.1. However, this was not observed. Instead, the new bound and free protein concentrations produced a data point which was superimposed upon the original isotherm curve, indicating a re-establishment of equilibrium, and hence the reversible binding of CBDN1 to PASA.  Fig. 4.1. Reversibility of CBDN1 binding to PASA at 4°C. A n ascending isotherm was initially constructed, represented by (—). Desorption experiments were performed at two concentrations (7 u M where f denotes samples and 12 u M denoted by t). "t 0 " represents the theoretical bound and free concentrations upon substitution of fresh buffer for isotherm buffer. "t 2 " represents the experimentally determined bound and free protein concentrations after a two hour incubation subsequent to dilution. ( ) =  ( ) =  Reversible binding to insoluble cellulose is a property that only some CBDs demonstrate. For instance, CBDCex (family II) will not desorb from either BMCC or PASA in 50 mM phosphate buffer, pH 7 (Jervis et al., 1997; Brad McLean, personal communication). Conversely, the binding of CBDCBHI is completely reversible on BMCC in 50 mM acetate buffer, pH 5 (Linder and Teeri, 1996). At this time, it is not known whether the binding kinetics (i.e. reversibility/irreversibility) are preserved amongst members within a given CBD family or whether they are CBD-specific; more reversibility data on different CBDs is required. The binding kinetics of CBDN1 on PASA was important in choosing the Langmuir model to analyze the isotherm data. Additional support for the use of the Langmuir model comes from the calculation of correlation coefficients (r ). Correlation coefficients represent how well fitted 2  curves follow the experimental results (Medve et al., 1997). When the binding of CBDN1 to PASA is analyzed by the method presented in section 2.6, the correlation coefficients were consistently 0.985 - 0.999, indicating that the Langmuir model accurately describes the binding isotherm for CBDN1 on PASA. In previous analyses of CBDN1 binding to PASA (Tomme et al., 1996), depletion isotherms were analyzed by a Langmuir binding model as well, albeit a slightly modified version which recognized cellulose as a substrate with overlapping binding sites (Gilkes et al., 1992). In their analysis, the raw binding data were plotted in a double reciprocal form prior to the generation of binding constants. One problem with the analysis of Gilkes et al. (1992) involves the use of double reciprocal plots to analyze isotherm data. In general, linear transforms (i.e. double reciprocal plots) are sensitive to imprecise data points at low protein concentrations (Winzor and Sawyer, 1995). Confounding this problem is the fact that any isotherm data compiled by the method described in section 2.6 is most sensitive to error at low protein concentrations. To avoid the compilation of inaccurate binding parameters, linear transforms were not used to analyze the raw binding data in this work. Instead, the binding constants were derived from the raw isotherm data (see section 2.6). The method of analysis presented in section 2.6 differs slightly from the standard Langmuir analysis in that a "G" variable is included in the first  equation (section 2.6, equation 1). This "G" variable was included to correct for optical scattering caused by cellulose "fines", and instrument imprecision. Optical and instrument effects are most problematic at low protein concentrations where they can comprise a significant component of the recorded A28O measurement. This often caused the raw binding curves to intersect at points away from the origin. Application of the basic Langmuir equation to these curves forced them through the origin, distorting their shape and resulting in inaccurate binding parameters. The use of a Langmuir equation including the "G" variable (section 2.6, equation 1) avoided that problem. The initial fit was instead applied directly to the curve; its intersection with the x-axis being irrelevant. Binding constants could then be determined from the shape of the raw data and not from forcing curves through the origin. The end result was the generation of binding constants which more accurately reflected the binding curve. Furthermore, because "G" varied for different isotherms, use of the modified Langmuir equation (section 2.6, equation 1) improved the consistency of affinity constant determination as a whole.  4.3  Binding of C B D N 1 to allomorphs of insoluble cellulose  The allomorph specificity of CBDN1 was studied using the binding analysis presented in sections 2.6 and 4.2. Four allomorphs of insoluble cellulose were investigated: bacterial microcrystalline cellulose (BMCC), Avicel PH101, PASA, and regenerated Avicel. Regenerated Avicel differs from PASA in that the Avicel is fully dissolved in phosphoric acid and not merely swollen as it is during the preparation of PASA (see section 2.1).  Physically, regenerated  cellulose appears much more homogenous than PASA. Avicel and BMCC, like PASA are particulate in appearance with particle sizes ranging from one to 10 um. All samples of cellulose were prepared in 50 mM potassium phosphate buffer, pH 7.  34  7i  Free (LIM)  Fig. 4.2. Isotherms of wild-type CBDN1 on different allomorphs of insoluble cellulose, included in the figure legend. Isotherms were performed at 4°C; each curve was generated by nonlinear regression analysis of the binding data, using a modified Langmuir equation (see section 2.6).  The results clearly show allomorph specificity (Fig. 4.2). CBDN1 binds to PASA and regenerated cellulose with similar affinities but associates weakly with the crystalline allomorphs of cellulose. The same allomorph specificity has been described in previous studies (Coutinho et al., 1992; Tomme et al., 1996). However, in the present study, binding was described in terms of affinity constants (Ka), as opposed to partition coefficients.  Table 4.1. Allomorph specificity of CBDN1 in 50 mM potassium phosphate buffer, pH 7, determined by depletion isotherms at 4°C.  Cellulose allomorph  crystallinity  PASA - isotherm #1  -0-5%  4.06 (+0.28) x 10  7.03 (+0.23)  PASA - isotherm #2  -0-5%  4.04 (±0.13) x 10  6.95 (±0.091)  PASA - isotherm #3  -0-5%  3.60 (+0.15) x 10  7.31 (±0.13)  PASA - isotherm #4  -0-5%  3.45 (±0.19) x 10  6.13 (±0.13)  Regenerated Avicel  -0-5%  2.60 (±0.096) x 10  6.47 (±0.11)  Avicel  50  nq  nq  BMCC  76  nq  nq  1  K (M )  1  -1  N (|imol/g cellulose)  2  2  a  0  5  5  5  5  5  Crystallinity of the cellulose preparation relative to Valonia cellulose recorded in literature  (Kulshreshtha and Dweltz, 1973; Lee et al., 1983; Wood, 1988) 2  Single isotherms were used to generate all binding parameters with the error recorded being  the standard error for one particular isotherm. nq - binding too low to accurately quantify  The quantification of K and N on BMCC and Avicel was complicated by the inability to a  0  acquire a full isotherm. With incomplete data sets, isotherm analysis requires excessive extrapolation, resulting in less accurate estimates of the binding parameters. Consequently the binding data for BMCC and Avicel were not included in table 4.1. In table 4.1, the four isotherms of CBDN1 on PASA were performed under identical conditions but used different preparations of protein. These data are displayed for two reasons.  36 First, they illustrate the reproducibility of the depletion isotherm technique, and second, they illustrate the consistency of the purified protein preparations. The standard deviation between these four affinity constants is 2.69 x 10 M" or 7% of the mean, indicating that accurate binding 4  1  data can be extracted from a single depletion isotherm.  4.4  Temperature dependence of CBDN1 binding Depletion isotherms at three different temperatures (Fig. 4.3) indicate that the binding  affinity of CBDN1 to PASA is highly dependent on temperature. Binding is strongest at 4°C and decreases with increasing temperature (Table 4.2), indicating an exothermic binding reaction. The family I CBD, CBDCBHI, illustrates a similar temperature dependence profile as CBDN1, but on BMCC instead of PASA (Linder and Teeri, 1996). However, the affect of temperature on binding is not uniform among carbohydrate binding proteins. CBDCex and CBDCenD, both family II CBDs, exhibit much smaller reductions in binding affinity to BMCC when the temperature is raised from 4°C to 37 °C (Al Boraston and Brad McLean, personal communication). CBHI, the intact cellobiohydrolase from T. reseei, actually exhibits a slight increase in adsorption to BMCC as the temperature is increased from 4°C to 40°C (Medve et al., 1997).  Table 4.2. Temperature dependence of CBDN1 binding to PASA.  Temperature (°C)  K (M ) -1  N (umol/g cellulose)  1  1  a  0  4( ) 2  3.79 (±0.27) x 10  6.85 (±0.43)  23  2.00 (±0.13) x 10  4.96 (±0.15)  36  1.06 (±0.08) x 10  5.57 (±0.23)  1  5  5  5  Binding parameters calculated from a single depletion isotherm with standard error recorded  for that isotherm 2  Binding parameters recorded as the average of four different depletion isotherms with error  being the standard deviation between quadruplicate values as calculated by:  ^ ( X i  - X ean)]/4 m  37  7i  6^  Free (LLM)  Fig. 4.3. Temperature dependence of CBDN1 binding to PASA. Isotherms were acquired at three different temperatures, included in the figure legend. Each curve was generated by nonlinear regression analysis of the binding data, using a modified Langmuir equation (see section 2.6).  The observed temperature dependence of binding to PASA correlates with previous calorimetric investigations. Those studies showed that CBDN1 binding to soluble cellulosics is associated with a small and negative A C (AC = 3H/3T at constant pressure), indicating a p  p  temperature dependent binding to those substrates (Tomme et al., 1996; Creagh et al., 1998). Defining the temperature dependence of binding is important for CBDN1 since much of the binding data has been compiled at different temperatures. Depletion isotherms have been performed at 4°C (Coutinho et al., 1992; Tomme et al., 1996; current research), affinity electrophoresis at approximately 22°C (current research), NMR titrations at 35°C (Johnson et al., 1996b), and isothermal titration calorimetry at 25°C, 30°C, and 35°C (Tomme et al., 1996). Direct comparisons of the binding data from these techniques should therefore consider the results presented in this section.  4.5  Use of C B D s to characterize cellulose  Thus far in CBD research, cellulose has been used as a ligand to elucidate the binding characteristics of CBDs. However, as a greater understanding is gained about the binding properties of different CBDs, the utilization of CBDs to characterize cellulose may become possible. Despite nearly half a century of intensive work on cellulose, much remains unclear regarding the tertiary structure. For instance, while PASA is described as an amorphous form of cellulose with exceedingly low crystallinity (Lee and Kim, 1983), crystalline microdomains may be present that are too small to be detected as ordered structure by x-ray diffraction (Atalla, 1993). In this section, the existence of crystalline microdomains in PASA is investigated by a binding study involving both CBDCex and CBDNL As indicated in chapter 1, CBDCex and CBDN1 differ both structurally and functionally. Structurally, CBDCex has a planar binding face in contrast to the binding cleft of CBDNL Functionally, CBDCex recognizes all forms of insoluble cellulose, both crystalline and amorphous, while CBDN1 is specific for amorphous forms of insoluble cellulose. CBDN1 also binds soluble cellulosics with higher affinity than CBDCex (Peter Tomme, personal communication). The binding properties of CBDCex and CBDN1 on PASA are summarized in table 4.3.  39 Table 4.3. Binding data for CBDN1 and CBDCex on PASA, determined from depletion isotherms analyzed with the modified Langmuir method of analysis (section 2.6).  Binding parameter  CBDN1  Affinity constant (M ) _1  2  3.79 (±0.27) x 10  1.51 (±0.10) x 10  5.60 (±0.43)  14.29 (±0.38)  yes  no  5  Capacity at 25|iM (p:mol/g cellulose) Reversible binding 1  CBDCex  1  6  Binding parameters recorded as the average of four different depletion isotherms with error  being the standard deviation between quadruplicate values as calculated by: ^ ( X i  2  - X ea„)]/4 m  Binding parameters calculated from a single depletion isotherm with standard error recorded  for that isotherm (Brad McLean and Peter Tomme, personal communication)  The differences in active site structure and binding properties between CBDN1 and CBDCex suggest that each recognizes different binding sites within insoluble cellulose. If true, then binding to saturation of one CBD should not prevent the second from associating with the cellulose; there should be little competition for binding sites. This concept was tested in the following experiment. Making use of its irreversible binding nature, CBDCex was added to 1 mg of PASA at a final concentration of 25 |iM in one ml of 50 mM potassium phosphate buffer, pH 7.0. Under those conditions CBDCex should saturate all available binding sites on the cellulose. The cellulose was then washed four times withfreshbuffer. A28O measurements were taken after each wash to determine how much CBD had desorbed from the cellulose. After the fourth wash, CBDN1 was added to the Eppendorf tubes at a concentration of 25 |J.M, again near saturation conditions for the CBD. Finally, the amount of CBDN1 bound to the PASA was quantified. Duplicate samples were analyzed on PASA; the same procedure was also employed using BMCC as the substrate. The results are shown in table 4.4.  40  Table 4.4. Summary of the binding data for CBDCex and CBDN1 to PASA and BMCC. Concentration of bound CBD (umol/g cellulose) Sample  CBDCex - pre wash  CBDCex - post wash  CBDN1  PASA-#1  16.5  10.2  5.6  PASA-#2  15.7  9.6  4.8  BMCC  100  52  (10  1  1  Bound concentration determined from the spent buffer collected after each of four successive  washes  Based on the results presented in table 4.4, only 60% of CBDCex remains bound to the PASA. This finding is surprising since CBDCex binds irreversibly to PASA; the apparent desorption of CBDCex is probably due to the washing technique utilized. Multiple washings of the cellulose (and bound CBD) were essential to reduce the background A28O enough such that an accurate measure of bound CBDN1 could be obtained. Unfortunately, this washing technique was flawed for two reasons. First, upon decanting spent buffer (containing unbound CBD) during each wash, it was impossible to completely dry the cellulose. Therefore, unbound protein was carried over in each washing step where it was effectively measured more than once, leading to an overestimation of desorbed CBD. Second, CBDCex has been shown to adsorb to the walls of the plastic Eppendorf tubes (Brad McLean, personal communication). In the dilution phase of the washing procedure, some of the CBD could desorb from the plastic, again contributing to the amount of free protein measured, and hence an overestimation of the CBD desorbed. Considering these errors, table 4.4 provides strong evidence that the binding of CBDCex and CBDN1 to PASA is additive. After saturating the majority of CBDCex binding sites, CBDN1 binding is virtually unaffected. This implies that CBDN1 recognizes different binding sites in PASA than CBDCex.  Taken together with the different structural and functional  properties of the two CBDs, these data are a strong indication that CBDCex recognizes  rnicrocrystalline domains within PASA while CBDN1 probably associates with "single stranded" or amorphous regions (see section 5.4 for more data concerning this statement). Furthermore, because CBDCex binds irreversibly and with a similar affinity constant to both PASA and B M C C (Brad McLean, personal communication), evidence for the existence of microcrystalline domains within P A S A becomes even stronger. In the future, similar experiments using CBDCex and CBDN1 to determine the proportion of microcrystalline and amorphous regions may be an alternative way to assess the accessible surface area within different samples of cellulose.  Chapter 5: Characterization of CBDN1 Binding by Site-directed Mutagenesis 5.1 Choice of mutation In order to better understand the binding activity of CBDN1, ten alanine mutants were constructed and characterized for binding to both PASA and barley |3-glucan. Targeting of these residues was based on both their location within the putative binding cleft of CBDN1 (Johnson et al., 1996a) and calorimetry results which indicated a binding activity mediated by a balance of polar and aromatic residues (Tomme et al., 1996). For this reason, two charged residues (D90, R75), six polar residues (Q80, Q124, Q128, N50, N81, T87), and two tyrosines (Y19, Y85), all with solvent exposed side chains were targeted for mutation (Fig 5.1). Alanine was used as the replacement residue for several reasons. Alanine is uncharged, nonpolar, and small by comparison to most other amino acids. Consequently, alanine cannot participate in many of the noncovalent interactions which mediate protein binding. In addition, alanine has the smallest side chain for which typical <|),\|/ bond angles are preserved (Ramachandran and Sasisekharan, 1968; Zubay, 1983), thus preventing gratuitous bond angles and consequent backbone perturbations that mutation to smaller residues (i.e. glycine) can cause. For these reasons, alanine is an excellent reference to assess the contribution of a given residue's biochemical and steric properties for binding (Wells, 1991). All CBDN1 mutants were constructed by PCR and sequenced to ensure that no additional mutations were incorporated into the mutants (NAPS unit, UBC). Mutant proteins were expressed and purified using the protocol presented in section 2.5. After purification, the alanine mutants were first characterized for structural integrity by 1-D H NMR. The mutants were then J  characterized by binding isotherms on insoluble PASA and affinity electrophoresis on barley pglucan.  183(1  ]  Q124  Fig. 5.1. Colour coded summary of the binding site residues targeted for mutation to alanine. Colour coding is based on amino acid properties. Charged residues (D90 and R75) are highlighted in red. Polar residues (Q80, Q124, Q128, N81, N50, T87) are coloured in green and hydrophobic/polar residues (Y19 and Y85) are highlighted in cyan.  5.2 5.2.1  Structural characterization of CBDN1 alanine mutants Fundamentals of H nuclear magnetic resonance spectroscopy 1  Nuclear magnetic resonance spectroscopy (NMR) is used to elucidate the carbon/hydrogen framework of a molecule in aqueous solution. The nuclei of many atoms behave as though they are spinning on an axis. , Because these nuclei are charged, their movement creates a magnetic field with an associated magnetic moment. In field free space, these magnetic moments are oriented in random fashion. However, in the presence of an external magnetic field, only certain orientations are allowed. For some nuclei such as *H, C , and F , only two nuclear spin states 1 3  1 9  occur: the magnetic moments will align either with or against the applied field. These spin states differ in energy; those nuclei that align with the field are in a lower energy state than those that align against the field. By irradiation with radio waves of the proper frequency, a nucleus in the lower energy spin state can absorb energy and convert to the high energy spin state, a phenomenon known as "spin flipping". This "spin flipping" or "resonance" is the basis for NMR spectroscopy, and is responsible for producing absorption peaks in an NMR spectra (Streitwieser et al., 1992). Each absorption peak is referred to by its position in the spectrum, called a "chemical shift". The magnetic field experienced by a particular nucleus is affected by the electron clouds of nearby atoms. Electrons move in such a way that they induce their own magnetic field, opposed to the magnetic field applied by an external magnet. This results in a lower net field at the nucleus, a phenomenon known as "shielding". Protons in different electronic environments experience different amounts of shielding; therefore, the resonance of those protons will occur at different frequencies resulting in distinct chemical shifts in an NMR spectrum. Chemical shifts are sensitive not only to the proximity of nearby atoms but also to the identity and molecular geometry of nearby atoms since all of these factors will affect the electronic environment around one particular nucleus (Sternlicht and Wilson, 1967; Markley et al., 1967; Tigelar and Flygare, 1972; Wishart et al., 1991; Streitwieser et al., 1992). NMR spectroscopy can be performed at several different levels of resolution. At the lowest resolution, NMR spectra are acquired in one dimension. These spectra are useful to  analyze the carbon/hydrogen framework of low molecular weight compounds, although spectral overlap complicates analysis of the 1-D spectra for larger molecules (i.e. proteins). The second level of resolution involves the acquisition of an NMR spectrum in two dimensions. In this case, resolution of the peaks which overlap in 1-D spectra is possible, and most chemical shifts can be assigned. To fully assign the 3-D structure of a protein, multiple 2-D experiments must be performed. These 2-D experiments consist of two basic types.  In correlated spectroscopy  (COSY)-type experiments, "through bond" relationships are resolved for protons that are separated by not more than three covalent bonds. This allows for the assignment of amino acids along a peptide chain. In nuclear Overhauser enhancement spectroscopy (NOESY), "through space" relationships for proton pairs separated by less than five Angstroms are elucidated. In practice, by using a suitable selection of three to six COSY-type experiments in conjunction with a NOESY-type experiment followed by mathematical analysis of the data, model construction and then model refinement, the proton connectivities within a protein can be delineated in three dimensions (Wuthrich, 1989; 1990).  5.2.2  Mutant characterization by 1-D * H N M R spectroscopy  1-D H NMR spectra are sometimes used in mutagenic studies to assess the tertiary fold of A  a protein (Mayo et al., 1995). These spectra are useful since each folded protein will have an individualized spectrum with trademark peak dispersions and chemical shifts. Such features can be used to assess both the identity and structural integrity of a protein. In collaboration with Phil Johnson (UBC), 1-D H NMR spectra with water presaturation l  were acquired on samples of wild-type CBDN1 and mutants Q128A, Q124A, N81A, T87A, Y85A (Fig. 5.2). The samples of R75A, Q80A, D90A, N50A, Y19A had their 1-D !H spectra recorded using a jump-and-return pulse sequence to avoid excitation of the water (Fig. 5.3). Wild-type CBDN1 spectra was also recorded via a jump and return pulse sequence in figure 5.3. The 1-D *H NMR spectrum of unfolded CBDN1 was obtained by reducing the protein with a 100-fold excess of dithiothreitol (DTT). All spectra were recorded with 600 ul samples on a Varian Unity 500 MHz spectrometer at 30 °C. The buffer conditions were 50 mM phosphate  buffer, pH 7, 0.05% sodium azide, 90% H.20/10% D2O. Protein samples were prepared in the buffer at 0.5 mM final concentration. From a comparison of the spectra of wild-type, mutant and denatured CBDN1, it is clear that all CBDN1 mutants are folded proteins. None of the mutants show a collapsed 1-D spectrum indicative of fully unfolded protein, as seen for CBDN1 denatured by reduction of its disulphide bond. In addition, a comparison of the spectra from the folded mutants indicate key signature features of the wild-type protein. First, the conservation of peak dispersion is a strong indicator that all protein samples are similar in both primary sequence and fold. Second, the peak at 10.2 ppm is a signature peak for CBDN1 resulting from a W137 proton resonance (Johnson et al., 1996b). This peak is also present at the same chemical shift in the spectra of all alanine mutants in figures 5.2 and 5.3, providing proof that all proteins are variants of CBDN1, and folded similar to the wild-type protein. General differences between the spectra in figure 5.2 versus those in figure 5.3 are the result of different spectrum recording conditions. All spectra in figure 5.2 were recorded under water presaturation conditions while those spectra in figure 5.3 were recorded using a jump and return pulse sequence which tends to amplify the downfield signal. Subtle differences do exist between the various spectra recorded within figures 5.2 and 5.3. Because 1-D NMR is an extremely sensitive technique, these differences are difficult to interpret at a structural level. Some differences are due to the changes in primary structure upon mutation, but spectral differences could also occur indirectly, since mutation can cause subde backbone perturbations as the protein accommodates changes in spatial and biochemical properties (Kline and Mueller, 1992; Linder et al., 1995; Naghora et al., 1995; Siebert et al., 1997). In figure 5.2, it should also be noted that the wild-type CBDN1 spectrum was not obtained at the same time, nor was the protein purified by the same protocol as the mutants. This could account for some of the spectral differences observed in comparing mutant spectra with the wild-type CBDN1 spectrum. A good example is the collection of sharp peaks observed at 8.0 ppm, present in each mutant sample but not in the wild-type CBDN1 sample.  47  i o.o  9T0 p p m  sTo  F i g . 5.2. Downfield portion of the 1-D NMR spectra of various CBDN1 mutants, wild-type CBDN1 and denatured CBDN1 (+ DTT spectrum). Data was acquired on a Varian Unity 500 MHz spectrometer at 30°C. Protein samples were 0.5 mM, in 50 mM phosphate buffer, pH 7 supplemented with 10 % D 0. Raw data was analyzed using Felix version 2.3 software to generate the above plots. 2  10.0  9.0  8.0  p p m  Fig. 5.3. Downfield portion of the 1-D N M R spectra of various CBDN1 mutants, wild-type CBDN1 and denatured CBDN1 (+ D T T spectrum). Data was acquired on a Varian Unity 500 MHz spectrometer at 30°C. Protein samples were 0.5 mM, in 50 m M phosphate buffer, pH 7 supplemented with 10 % D 0. Raw data was analyzed using Felix version 2.3 software to generate the above plots. 2  5.3  Binding of C B D N 1 alanine mutants to PASA In section 4.3 it was shown that among the insoluble allomorphs of cellulose, CBDN1  binds mosdy strongly to PASA. Consequently, PASA was used to screen the alanine mutants for relative binding strength. All CBDN1 alanine mutants were characterized on PASA using depletion isotherms in combination with the Langmuir method of analysis described in section 2.6. Raw data and the Langmuir modeling results are included in figures 5.4 and 5.5. Figure 5.4 illustrates the isotherms for those mutated residues positioned on the left side of the binding cleft (see figure 5.1), while figure 5.5 contains the mutated residues positioned on the right side of the cleft. Affinity constants (Ka) derived from the isotherms are included in table 5.1 in order of decreasing affinity. All mutations affected the binding of CBDN1 to PASA. Alanine replacement at N50 resulted in the lowest measured affinity constant, although in figures 5.4 and 5.5, it is clear that alanine replacement at Y19 and Y85 had the largest affect on CBD binding. The affinity constants for Y19A and Y85A are not included in table 5.1. Both data sets are incomplete; they include only the initial, linear segment of the binding curve. As in section 4.2, these incomplete curves complicated the model's ability to predict a saturation value for binding, resulting in standard errors for the regressed affinity constants that were larger than the magnitude of the affinity constants themselves. For all other mutants, the data sets were more complete and showed some convergence on a saturation value. These data sets were more suitable for analysis, resulting in accurate estimates of the binding parameters. Attempts were made to acquire complete isotherms for Y19A and Y85A by generating binding data with less cellulose and more protein. However, reducing the quantity of cellulose resulted in decreased sensitivities for detecting bound CBD and poor isotherm data. Similarly, increasing the concentration of protein resulted in excess scattering (A350 values) and the need for sample dilution, both of which made contiguous curves difficult to produce. For the isotherms of Y19A and Y85A, an estimate of the affinity constants can be made considering that the same sample of PASA was used for all isotherms. In this case, the total number of binding sites per gram of PASA (i.e. N ) should be constant, and irrespective of the 0  50 mutant, provided that the mutants recognize the same binding site. Assuming Y19A and Y85A recognize the same binding site as CBDN1, affinity constants for these mutants can be regressed by setting N at 6.85 umol/g PASA (average value of N calculated from four isotherms of wild0  0  type CBDN1, see section 4.2) during isotherm analysis. Calculated this way, the affinity constants for Y19A and Y85A respectively, are 3.3 x 10 and 3.0 x 10 M . 3  3  _ 1  Table 5.1. Binding data determined from depletion isotherms of CBDN1 and alanine mutants on PASA in 50 mM potassium phosphate, pH 7 at 4°C.  Cleft position  Mutant  Fold reduction versus wt  K (M-l)l a  CBDN1 wt CBDN1  -  3.79 (±0.27) x 10  1  T87A  middle  1.74 (±0.12) x 10  2.2  Q80A  left  1.38 (±0.065) x 10  2.7  D90A  right/middle  1.27 (±0.057) x 10  2.9  N81A  left  8.11 (±0.72) x 10  4.7  R75A  right  6.33 (±0.38) x 10  6.0  Q128A  right  5.76 (±0.79) x 10  6.6  Q124A  left  4.99 (±0.34) x 10  7.6  N50A  left  1.60 (±0.6) x 10  24  Y19A  right  nq  nq  Y85A  left  nq  nq  1  2  5  5  5  5  4  4  4  4  4  3  Binding parameters calculated from a single depletion isotherm, and standard error  recorded for that isotherm 2  Binding parameters recorded as the average of four different depletion isotherms with  error being the standard deviation between quadruplicate values as calculated by: ^ ( X i 3  - X ean)]/4 m  Refers to the reduction in affinity relative to wild-type CBDN1  nq - binding too low to accurately quantify  51  7i  6H  Free (LIM)  Fig 5.4. Isotherm data for the binding of CBDN1 and alanine mutants (whose residue position is on the left side of the binding cleft in Fig. 5.1) to PASA at 4°C. Each curve was generated by nonlinear regression analysis of the binding data, using a modified Langmuir equation (see section 2.6).  52  7i  Free (|aM)  Fig 5.5. Isotherm data for the binding of CBDN1 and alanine mutants (whose residue position is on the right side of the binding cleft in Fig. 5.1) to P A S A at 4°C. Each curve was generated by nonlinear regression analysis of the binding data, using a modified Langmuir equation (see section 2.6).  5.4  Binding of C B D N 1 mutants to barley (3-gIucan  To complement the binding analysis on PASA, the mutants were also characterized on barley P-glucan. This polymer, comprised of repeating units of two or three P-1,4 linkages and a P-1,3 linkage alternate (Parrish et al., 1960), was used for several reasons. First, barley P-glucan is a soluble polymer as opposed to PASA. Second, the binding of wild-type CBDN1 to barley Pglucan has been well characterized by calorimetry (Tomme et al., 1996). In that work, CBDN1 demonstrated a high affinity for barley P-glucan, making it an attractive soluble substrate to screen the mutants for binding affinity. A final reason for using barley P-glucan instead of derivatized forms of cellulose is because barley P-glucan has a defined structure with a known DP (average DP is 96; Peter Tomme, personal communication). Consequently, substrate concentrations can be converted from a gram per volume basis to molarity. In addition, it has been determined that each polymer of barley P-glucan binds approximately four molecules of CBDN1 at saturation (Tomme et al., 1996), permitting the calculation of affinity constants per mole of binding site, the same units used in the previous analysis on PASA. While mutants were characterized on PASA using depletion isotherms, the soluble nature of barley P-glucan made characterization by depletion isotherms impossible.  Instead, the  technique of affinity electrophoresis was used to generate binding constants. An explanation of how affinity constants can be generated by affinity electrophoresis is included in section 2.7. In affinity electrophoresis, interactions between protein and polymer retard the movement of the protein through the gel. This retardation phenomenon is illustrated in figure 5.6. Different mutants exhibit unique retardation profiles at one concentration of barley P-glucan. For example, Y19A and Y85A show little retardation in the presence of 33 uM barley P-glucan whereas T87A and D90A are considerably retarded in their progress through the gel (see Fig. 5.6). Figures 5.7 and 5.8 include the results of the binding analysis where retardation (1/rm) is plotted against barley P-glucan concentration. Each data point represents the binding data from a single gel at one barley P-glucan concentration. For all mutants, retardation increases linearly with barley P-glucan concentration, indicating that the mutants interact with the polysaccharide.  A C  1  2  3  4  5  6  7  8  9  10 11  12  B C  1  2  3  4  5  6  7  8  9  10  11 12  Fig 5.6. Electrophoresis profiles of CBDN1 and alanine mutants on 13% native gels without (A) or supplemented with (B) 0.025% barley (3-glucan. In both gels, lane C includes bovine serum albumin, lanes 1-12 include: wild-type CBDN1 (lane 1), T87A (lane 2), D90A (lane 3), Q80A (lane 4), Q124A (lane 5), Q128A (lane 6), Y85A (lane 7), Y19A (lane 8), R75A (lane 9), N50A (lane 10), N81A (lane 11), and wild-type CBDN1 (lane 12).  The strength of this interaction varies widely for each mutant, and is proportional to the slope of the linear regression lines presented in figures 5.7 and 5.8. Mutants such as T87A and D90A bind more tightly than Y85A and Y19A, and consequently show much steeper slopes. BSA retardation shows a near zero slope, indicating that it does not interact with barley P-glucan. Table 5.2 includes a summary of the binding data generated on barley P-glucan. The range of barley P-glucan concentrations used in the binding analysis for each mutant are included in the table. Dissociation constants were determined by extrapolating x-intercepts from the linear regression lines generated for each mutant in figures 5.7 and 5.8. The association constants included in table 5.2 are given by 1/Kd.  Table 5.2. Summary of the binding data for CBDN1 and alanine mutants on barley P-glucan, determined by affinity electrophoresis at approximately 22°C, pH 8.8. Mutant  Barley P-glucan  K (M-l)  Fold reduction  a  concentration range  versus wt CBDN1  (uM)l wt CBDN1  0-33  3.3 (±0.36) x 10  T87A  0-33  2.1 (±0.25) x 10  1.6  D90A  0-33  2.0 (±0.22) x 10  1.7  Q80A  0-33  1.9 (±0.27) x 10  1.7  N81A  0-66  7.0 (±0.24) x 103  4.7  R75A  0-66  5.2 (±0.39) x 10  6.3  Q128A  0-132  3.3 (±0.29) x 10  10  Q124A  0-132  1.9 (±0.11) x 10  17  N50A  0-662  9.2 (±0.35) x 10  35  Y19A  0-662  4.6 (±0.14) x 10  72  Y85A  0-662  4.2 (±0.17) x 10  79  1  4  4  4  4  3  3  3  2  2  2  1  Refers to the barley P-glucan concentration range over which data was collected for the  calculation of affinity constants included in the table. Refers to the reduction in affinity on barley P-glucan relative to wild-type CBDNL 2  2  56  L e g e n d + BSA ? wt CBDN1 • T87A • N81A  10  20  30  40  C o n c e n t r a t i o n o f p o l y m e r i n g e l (uM)  B L e g e n d + BSA v wt CBDN1 • O90A • R75A  20  30  40  C o n c e n t r a t i o n o f p o l y m e r i n g e l (uM)  Fig. 5.7. Analysis of the affinity electrophoresis data for CBDN1 and mutants. In A) and B), retardation is plotted as a function of barley P-glucan concentration. Bovine serum albumin (BSA) is included as a nonbinding control. Curves were generated by linear regression analysis of the raw data, and dissociation constants from the x-intercept of the linear regression lines.  57  L e g e n d + BSA * wt CBDN1 o Q80A • Q128A * Q124A  25  50  75  100  125  150  C o n c e n t r a t i o n o f p o l y m e r i n g e l (yiM)  B 10  L e g e n d BSA wt CBDN1 o N50A T Y19A V Y85A +  0  100  200  300  400  500  600  700  C o n c e n t r a t i o no f p o l y m e ri n g e l( j i M )  Fig. 5.8. Analysis of the affinity electrophoresis data for CBDN1 and mutants. In A) and B), retardation is plotted as a function of barley p-glucan concentration. Bovine serum albumin (BSA) is included as a nonbinding control. Curves were generated by linear regression analysis of the raw data, and dissociation constants from the x-intercept of the linear regression lines.  5.5 Comparison of mutant binding to PASA and barley p-glucan The binding data on barley pVglucan and PASA indicate that all ten mutations affect the binding affinity of CBDN1, and that the pattern of affect is similar on both substrates. For example, a mutant that bound weakly to PASA (i.e. Y85A) also bound weakly to barley P-glucan. And conversely, a mutant that bound strongly to PASA (i.e. D90A) also bound strongly to barley P-glucan (see tables 5.1 and 5.2). This implies that the mechanism of binding is similar on both substrates. Each residue in the binding cleft appears to recognize the same features in PASA and barley P-glucan, despite differences in the macroscopic properties of each substrate. This is a strong indication that CBDN1 interacts with single cellulose chains in PASA, despite the fact that PASA isan insoluble compound. A comparison of the affinities presented in tables 5.1 and 5.2 indicates that the binding of both wild-type CBDN1 and the mutants is roughly one order of magnitude lower on barley Pglucan than PASA. Because depletion isotherms are acquired at 4°C and affinity electrophoresis is performed at approximately 22°C, some of the affinity difference can be explained by temperature effects (see section 4.3).  According to the work presented in section 4.3, this temperature  dependence should only account for a two-fold reduction in affinity. However, in affinity electrophoresis, the temperature is measured from the buffer in the inner electrophoresis chamber. During electrophoresis, the gels could attain higher temperatures. In fact, the affinity constant determined by affinity electrophoresis for wild-type CBDN1 (Table 5.2) is in close agreement with that generated by ITC at 35°C (Tomme et al., 1996), implying that gel temperatures may be higher than 22°C during electrophoresis. Another explanation for the affinity difference is that other residue(s) may exist which interact with additional glucan chains in PASA. Because the mutational strategy was biased towards residues in the binding cleft, it is possible that other residues distal to the binding cleft could associate with secondary glucan chains. Alternatively, the mixed P,l-3 P,l-4 linkage between glucopyranoside units in barley P-glucan may account for some decrease in affinity. However, ITC results have shown that wild-type CBDN1 has similar affinities for barley pglucan and several soluble cellulosics (Tomme et al., 1996). Because these cellulosics only  contain (3,1-4 linkages, this would indicate that linkage is not responsible for the difference in affinities between PASA and barley P-glucan. Another possibility hinges on the undefined nature of PASA as a substrate. The current view suggests that phosphate used during the preparation of PASA may still associate with the cellulose (Peter Tomme, personal communication). Furthermore, positive counter ions are present in the isotherm buffer (50 mM potassium phosphate). While highly speculative, this ionic character may strengthen the binding affinity of CBDN1 for PASA, especially when considering the fact that CBDN1 has a pi of 3.5.  To support this claim, work done by this author  (unpublished data) has shown that the affinity of CBDN1 for PASA is three-fold lower in pure water, as compared to 50 mM potassium phosphate, pH 7. In addition, the binding capacity (No) of PASA is increased more than two-fold in the presence of 300 mM NaCl, MgCl2, or CaCl2These results provide experimental support that some degree of ionic character may be associated with PASA which could lead to an affect on CBD binding.  5.6  Identification of the residues critical for binding  Site-directed mutation identified Y19 and Y85 as two residues critical for binding to PASA and barley P-glucan. Tyrosine side chains can interact with substrate in two basic ways. First, the phenolic hydroxyl group is capable of accepting or donating hydrogen bonds, and second, the aromatic ring can form extensive nonpolar contacts in the form of van der Waals interactions. Both types of interaction agree with the calorimetry results (Tomme et al., 1996); an enthalpic driving force can be mediated by both hydrogen bonding and van der Waals interactions. The reduction in binding affinity ranged from less than two-fold for mutants such as D90A and T87A to greater than 70-fold in the case of Y85A and Y19A.  Based on their weak  contributions to binding, it is unlikely that residues such as D90, Q80 or T87 interact directly with substrate. However, it is interesting to note that their removal does result in modest decreases in binding affinity. Presumably, the binding site is a network of precisely positioned amino acids, the maintenance of which involves the correct placement of both binding site residues that interact directly with substrate and supporting residues that hold the binding site residues in place.  Removal of the supporting residues will affect on the geometry of covalent and noncovalent bond angles near the site of mutation. This affect may then propagate through the binding site network, ultimately leading to a weakening of the noncovalent forces between the substrate and those residues directly involved in binding (Quiocho, 1989; Spurlino et al., 1992; Naghora et al., 1995). For residues D90, Q80 and T87, the effect is minimal and their mutation results in minimal (but detectable) decrements in affinity. In contrast, the mutation of residues R75, N81, Q124 or Q128 has a significant affect on binding. These mutations may result in increased strain on the binding network, although direct interaction with substrate is also possible for some of the residues. For Y85, Y19, and perhaps N50, direct interactions are likely since the affinity of the CBD decreases substantially upon mutation.  5.7  Use of mutagenesis to study mechanistic differences in binding  Previous calorimetric studies have provided the thermodynamic parameters describing CBDN1 binding to substrates such as cellotetraose, cellopentaose, carboxymethylcellulose (CMC), hydroxyethylcellulose (HEC) and barley P-glucan (Tomme et al., 1996). Plots of enthalpy versus entropy for these substrates (compensation plots) resulted in a linear relationship, indicating a similar mechanism of binding to all substrates. Enthalpy/entropy compensation seems to be the rule rather than the exception when studying the binding of different substrates on the same protein (Lemieux et al., 1991; Lemieux, 1993; Bundle et al., 1994; Sigurskjold et al., 1994), or mutants of a protein on a single substrate (Brummell et al., 1993; Ito et al., 1993). To explain why information on binding mechanism can be gained from enthalpy/entropy plots, consider a linear compensation plot for several substrates on a single protein. Now consider a substrate which binds to a different binding site on that protein. In this case, different residues will be involved in binding; consequently, the net entropic and enthalpic contributions will be different, resulting in a data point removed from linearity on the original compensation plot. In this way, large differences in binding mechanism (i.e. different binding sites) can be discerned from compensation plots. However, in cases where enthalpy/entropy compensation is observed, small differences in binding site interactions may go unnoticed. If an enthalpic interaction is lost  (i.e. hydrogen bond or van der Waal's contact), an amount of entropy will be gained due to increased motional freedom. And similarly, if an enthalpic interaction is gained, there will be an entropic cost due to increased motional restrictions. This is the essence of enthalpy/entropy compensation. As was shown in section 5.5, mutagenic studies can provide insights into the binding mechanism. For instance, if one residue were to play a prominent role in the binding to one substrate but not another, this effect would be observed direcdy in binding studies, but not in compensation plots. A comparison of the mutant binding profiles on HEC and barley P-glucan is included in figure 5.9.  Most mutants have a similar retardation profile on both substrates, although Q124A,  Q128A, and R75A show different profiles in the presence of HEC as compared to barley Pglucan. The differences indicate that these residues could recognize chemical differences between the two substrates while the other mutated residues do not. Consequently, Q124, Q128 and R75 may be responsible for substrate recognition and specificity in CBDN1. Because of the undefined nature of HEC as a substrate (Bhattacharjee and Perlin, 1971), it is difficult to comment any further on the specific role of each of these residues in the recognition process.  Additional  screening of these mutants on other carbohydrate substrates is required to determine the role of Q124, Q128, and R75 in binding specificity. Although further studies are needed to determine the roles of Q124, Q128, and R75 in substrate recognition, the data does indicate a difference in binding site interactions between barley P-glucan and HEC that was not observed in the compensation plots (Tomme et al., 1996). Interestingly, the affinity of wild-type CBDN1 on those substrates is similar (Tomme et al., 1996). This illustrates the ability of CBDN1 to accommodate for chemical differences in potential substrates, a phenomenon often seen in protein/substrate interactions (Quiocho et al., 1997).  C  1  2  3  4  5  6  7 8  9  10 11  Fig 5.9. Electrophoresis profiles of CBDN1 and alanine mutants on 13% native gels supplemented with (A) 0.05% barley p-glucan or (B) 0.05% H E C . In both gels, lane C represents bovine serum albumin as a control and lanes 1-11 include: wild-type CBDN1 (lane 1), T87A (lane 2), D90A (lane 3), Q80A (lane 4), Q124A (lane 5), Q128A (lane 6), Y85A (lane 7), Y19A(lane8),R75A(lane9),N50A(lane 10), and N81A (lane 11).  Chapter 6: Additional studies on Y19A and Y85A In the previous chapter, Y19 and Y85 were identified as two residues essential for CBDN1 binding to both barley fi-glucan and PASA. The large reductions in affinity caused by the mutation of either Y19 or Y85, in combination with the structural data suggest that both tyrosines interact directly with substrate. However, no direct evidence was obtained from the binding studies presented. In this chapter, NMR and Raman spectroscopy will be used to refine the roles of Y19 and Y85, both in sugar binding and in maintaining the tertiary structure of the CBD.  6.1  Structural studies on Y19A and Y85A using H A  1 5  N Heteronuclear Single  Quantum Coherence (HSQC) spectroscopy  While 1-D !H NMR showed that Y19A or Y85A were folded variants of CBDN1 (section 5.2), the information obtained was limited by the low resolution inherent in the technique. Here, an attempt is made to improve the structural information by acquiring NMR spectra for Y19A and Y85A in two dimensions with  1 5  N labeled protein. In this way, resonances from the amide  groups of individual amino acids can be resolved, allowing us to determine where (if anywhere) within the mutants structural changes have occurred. Samples of Y19A and Y85A, isotopically labeled with N , were prepared by growing E. l 5  coli JM101 clones in minimal media (Maniatus et al., 1982) supplemented with 1 g/l of the N 1 5  labeled protein hydrolysate "cell tone" (Martek Biosciences Corporation) and 1 g/l of N labeled 1 5  (NH4)2S04 (Cambridge Isotopes). Cultures were grown in the minimal media under the same conditions as reported earlier for growth in TY media (section 2.5), but induced with 0.5 mM IPTG. Periplasmic protein yields were low (2-3 mg/1) compared to the supernatant yield (20-25 mg/1) for these cultures. Consequently, the periplasmic fractions were combined with the supernatant fractions for purification. All of the protein preparations were buffer exchanged into 50 mM potassium phosphate buffer and fractionated by anion exchange chromatography three times (see section 2.5) at pH 5.2, pH 7, and pH 6. Approximately 20 mg of N labeled protein 1 5  were purified from 1 litre of culture. A faint yellow colour remained associated with the protein through the entire purification but did not interfere with the acquisition of any 2-D NMR spectra. In collaboration with Phil Johnson (UBC), the ! H - N heteronuclear single quantum 1 5  coherence (HSQC) spectra for wild-type CBDN1, Y19A and Y85A were recorded on a Varian Unity 500 MHz spectrometer at 30°C. These spectra are included in figures 6.1 and 6.2. The buffer conditions were 50 mM phosphate buffer, pH 7.0, 0.05% sodium azide, 90% H20/10% D2O. Protein concentrations of 0.5 mM were used. HSQC experiments correlate the resonance of a *H with the resonance of a directly bonded 1 5  N nucleus. Consequently, the data points in figure 6.1 correspond to chemical shifts for  backbone amides and the functional groups of arginine, asparagine, glutamine, and tryptophan residues. Most backbone amides have been assigned in the ! H - N HSQC spectrum of wild-type 15  CBDN1 (Philip Johnson, personal communication), although spectral overlap does occur for some chemical shifts (see Fig. 6.1). HSQC spectral analysis was performed using Felix version 2.3 software, and involved the comparison of 122 chemical shifts common to wild-type CBDN1 and the two tyrosine mutants. Chemical shifts not analyzed included those that were poorly resolved in the ! H - N HSQC spectrum for wild-type CBDN1. Of the 122 chemical shifts I 5  analyzed, 102 remain unchanged in all three spectra. Unchanged chemical shifts were defined as the movement of a chemical shift by less than 0.1 ppm in the *H resonance dimension and 0.5 ppm in the N dimension, in a mutant spectrum as compared to wild-type. 1 5  Chemical shifts are a sensitive measure of molecular conformation, composition and environment within a protein. Each chemical shift is the product of many atomic effects including local electron distributions, bond hybridization states, proximity to polar and aromatic groups, main chain flexibility and hydrogen bond strength (Sternlicht and Wilson, 1967; Markley et al., 1967; Tigelar and Flygare, 1972; Wishart et al., 1991). The maintenance of a chemical shift within the spectra of CBDN1 and the two tyrosine mutants implies that the spatial properties (and hence structure) around those particular *H and N nuclei are similar in each protein. Because 1 5  80% of chemical shifts remain unchanged in both mutants, the data implies that the tertiary  65 i  *G123  @G>5 G44  G86  .J3130  m  G131 - > G 1 2  V88  SB G51  G20  @  V17®  @ ®  Dll/Yl? ^  Q124  0  .  -<©  ®  O I ® S97  D135  W  ^  A  o Q caffhr ° W©V150  ' I f ^ V 7 8 ^ efS> L 2 S T58/T,  V44/V47  U25/A136  JW  3  <®  Q101A66 3  A145  6  w  E55/  _  Y60/D147  V102  ^ ® T73 A94^72  I  5  4  ~W16 E151  7STES6  1  0  °  A  5  3  5  2  120  , F132  |- 122  O  A 3 1  /rtiAHO L89  124  1 [ h  126 128  190  L49  A108  L62  h 118  R 7 5  ©D71  V  ^  "  116  _<©&Y45I&T70  150 ® rA117©  V  V74  ©D120  ^%E149 ^ ^ 8 3E98  °  , 4 2  W137E1  114  JJ--F9/L129  D142 „n ^ %*P v  K  Q80e  180 S3»-  Q42e  QlOie  S40  ~@^D119 T27 T65 < ^ ^ S 9 2 A 1 1 4 S109 <tf* G 4 6 1  112  r  ©V64  Q  110  8 2  N508  G23  T96 T$l(0) G79 G57/T138 @>'  El  Til 5 ® °  °  G30 d@ S133  T29  V34 ©  108  @T87  130 A76  LI 3 9  L152  h 132  134 134  10 1  H(ppm)  Fig. 6.1. H - N HSQC spectrum of CBDN1; the assignments of both amide and side chain chemical shifts are included as text (Phil Johnson, UBC). J  15  66  Wild-type 0  • .  o - d  f  •  °. • •»  „0  ,  f  o » 0  0  •'.  • >•  •  o  i  E  0  a  o a •^ -o CM  -in  rs  10.0  9.0  8.0 HI  7.0  6.0  (ppm)  Fig. 6.2. H- N HSQC spectra of wild type CBDN1, Y85A, and Y19A. Data was acquired on a Varian Unity 500 MHz spectrometer at 30°C with N labeled samples in 50 mM phosphate buffer, pH 7, supplemented with 10% D 0. X  15  15  2  structures of Y19A and Y85A are largely  conserved.  Furthermore, many  of the  differences in chemical shift can be explained. An assessment of these changes is summarized in tables 6.1 and 6.2, and is discussed below.  Table 6.1. Summary of the H-^N HSQC data for wild-type CBDN1, Y19A, and Y85A. l  Total number of chemical shifts investigated  122  Total number of chemical shifts conserved in all three spectra  102  Table 6.2. Summary of the ! H - N HSQC chemical shift analysis for wild-type CBDN1, 15  Y19A, and Y85A.  Conservation of chemical shift Residue  CBDN1  Y19A  Y85A  Explanation for discrepancy in chemical shifts  D10, E14, V17, A66, A134,L141, A145 N81, G82, A83, T87 G20, Q42, G44, V45, G46, V48, L49, G51, G130  yes  no  no  yes  yes  no  yes  no  yes  differences in bound calcium changes in shielding due to mutation changes in shielding and possible structural affects due to mutation  Seven chemical shifts differ in the HSQC spectra of Y85A and Y19A, as compared to wild-type CBDN1. This can be explained by differences in calcium concentration within each sample. CBDN1 contains a calcium binding site (Phil Johnson, personal communication); consequently, certain chemical shifts are highly dependent on calcium concentration. These chemical shifts have been identified by calcium titration experiments, and include the seven chemical shifts mentioned above (Philip Johnson, personal communication). The position of  those seven chemical shifts in the CBDN1 spectrum indicates that calcium is present in the sample. For each mutant, the same calcium dependent chemical shifts have been displaced, indicating that lower concentrations of calcium are present in those samples. This finding is not surprising; the sample of CBDN1 used for analysis was prepared by a different purification procedure (Coutinho et al., 1993), which could have affected the calcium concentration in the purified product. An alternative explanation is that alanine mutation at Y19 or Y85 affects calcium binding of the CBD. This is unlikely for two reasons. First, the calcium binding site is opposite the binding cleft, and hence distant from either mutation site (Phil Johnson, personal communication). Second, even if calcium binding were affected by mutation, it is unlikely that both mutations would have the same affect on calcium binding. The remaining differences in chemical shift cannot be attributed to calcium; they are specific for one of the tyrosine mutants, and conserved within the spectra for wild-type CBDN1 and the second tyrosine mutant. Apartfromthe calcium dependent chemical shifts, the ! H - N HSQC spectrum of Y85A 15  differs from CBDN1 (and Y19A) at four chemical shifts for the amides of residues N81, G82, A83, and T87. These residues are all in close proximity to Y85 (Fig. 6.3a); this close proximity to the mutation site could cause the change in chemical shift, as opposed to a true structural perturbation. Removal of the phenol ring in Y85A will affect the level of shielding experienced by these *H and N nuclei. Consequently, the actual magnetic field at those nuclei will be different 1 5  when Y85 is replaced. The end result is a change in the chemical shift of those amides, but not necessarily in the structure of that region (McMurray, 1990). In summary, only 11 of 122 chemical shifts differed in the spectra of Y85A and wild-type CBDN1. Seven of those differences were ascribed to calcium concentration present in the sample, and four to shielding affects. Taken together, these data indicate that the structure of Y85A is virtually identical to CBDN1.  G46  G130  Fig. 6.3. Perturbed chemical shifts in the ' H - N HSQC spectra for (A) Y85A and (B) Y19A. The side chains of Y l 9 and Y85 are included for reference purposes. 15  Apart from calcium dependent chemical shifts, the ! H - N HSQC spectrum of Y19A 1 5  differs from CBDN1 at nine chemical shifts for the amides of residues G20, Q42, G44, V45, G46, V48, L49, G51, and G130. The position of these residues relative to Y19 is included in figure 6.3b. In this case, it is doubtful that shielding effects (caused by removal of Y19) are responsible for all nine chemical shift perturbations. Residues such as G20, G44, G46, and G48 could be affected by removal of the phenol ring, but L49 and G51 are quite distal to the mutation site. The solved structure of CBDN1 indicates that the perturbed residues belong to a specific microdomain in the protein. Six of the nine perturbed amides are found on a single beta sheet while the rest are part of two adjoining loop regions near Y19. It is probable that the replacement of Y19 is causing a localized structural perturbation. Similar structural effects, induced by mutation, have been verified by x-ray crystallography and NMR in other mutagenic studies (Kline and Mueller, 1992; Linder et al., 1995; Naghora et al., 1995; Siebert et al., 1997). Unfortunately, it is difficult to assess the structural perturbations in Y19A from this NMR study. Nonetheless, the HSQC data does indicate that some structural perturbation occurs upon the removal of Y19 which cannot be fully explained by changes in nuclear shielding induced by the mutation.  6.2  Analysis of Y19A and Y85A binding to soluble oligosaccharides by NMR To complement the structural information, binding constants for the association of mutant  CBDN1 proteins with cellotetraose and cellopentaose were determined by titration of those substrates into  1 5  N labeled samples of Y19A and Y85A. For each titration, chemical shift  perturbations were used to calculate a K value (Johnson et al., 1996b). The corresponding Ka a  values from 15-31 different *H and N nuclei representing 10-16 different residues were then 1 5  averaged to determine a Ka value for that protein ligand pair (Table 6.3).  Table 6.3. Association constants of CBDN1, Y19A, and Y85A generated by H- N l  l5  HSQC  NMR titration experiments. Mutant  Cellotetraose (M" )  Cellopentaose (M )  wt CBDN1  4.2 x 10 (±0.7)  3.4 x 10 (±0.8)  Y19A  8.4 x 10 (±1.2)  2.7 x 10 (±0.3)  Y85A  5.9 x 10 (±0.8)  1.9 x 10 (±0.2)  1  3  1  1  _1  4  2  2  The affinity data for Y19A and Y85A on cellopentaose and cellotetraose are in close agreement with that reported earlier on barley P-glucan. Mutation of either Y19 or Y85 to alanine results in a substantial decrease in binding affinity to either substrate, indicating that both tyrosines are required for binding to cellopentaose and cellotetraose. The presence of the fifth sugar in cellopentaose leads to an order of magnitude increase in binding affinity to wild-type CBDN1 (Tomme et al., 1996; Johnson et al., 1996b). From table 6.3, it is clear that Y19A and Y85A are also sensitive to chain length, albeit less sensitive than wild-type CBDNL This sensitivity of K  a  to chain length, regardless of the residue type at positions 19 and 85, indicates that other residues must also interact with theterminalsugars in cellopentaose.  6.3  Determination of the roles of Y19 and Y85 by Ultraviolet resonance Raman  spectroscopy  (UVRRS)  It is not the purpose of this work to provide an extensive review of Raman spectroscopy, although a brief overview of the technique is presented in section 6.3.1. For more detailed reviews on the theory of Raman spectroscopy and its relation to other spectroscopic techniques, refer to Asher (1993a, b).  6.3.1  Overview of U V R R S theory  Raman spectroscopy measures the inelastic scattering of light as incident light interacts with matter. The technique utilizes a high intensity, monochromatic light beam which, in the case of UVRRS has a wavelength in the ultraviolet range of the electromagnetic spectrum. The  incident light beam has a well-defined frequency and polarization, and is used to illuminate the sample. Upon illumination, the incident light drives molecular vibrations at the incident frequency, v . In Raman scattering, the incident frequency v combines with other (natural) 0  0  nuclear vibrational frequencies (v ) in the molecule, giving rise to beat frequencies at v ± v n  0  n  which radiate Raman scattered light at the same frequencies. This scattered light is collected over a defined angle to determine changes in frequency, intensity, and polarization with respect to the incident light. Raman data is conventionally displayed as a plot of intensity versus Raman shift (Fig. 6.4). The Raman shift is defined as lA-iaser - lAscatter, where X refers to the wavelength of incident or scattered light. A Raman shift of zero indicates elastic scattering of the incident light, a phenomenon known as Rayleigh scattering. In Raman scattering, the Raman shift is always nonzero, and corresponds to different nuclear vibrations in the molecule. One of the main advantages to UVRRS is its resolution. For many spectroscopic techniques, a single spectrum is acquired that contains all of the relevant information, some of which is lost due to peak overlap. However, with UVRRS, problems associated with peak overlap can be minimized since certain peaks within the Raman spectrum can be magnified relative to other peaks. This is achieved by choosing incident light with a wavelength near those where natural electronic transitions can occur for the chromophore associated with the specific vibration to be studied. The end result is a large increase in Raman scattering associated with the chromophore of interest, as compared to other nuclear vibrations within the molecule (Asher et al., 1986). For tyrosine and tryptophan residues, electronic transitions are stimulated with incident wavelengths of 223 nm and 220 nm respectively (Asher, 1993b). By using an incident wavelength near 230 nm, an enhancement of the tyrosine and tryptophan vibrations will result, providing a Raman spectrum that is a "signature" of tyrosine and tryptophan vibrations (Harada and Takeuchi, 1986). UVRRS is relevant for protein biochemistry because it can be used to analyze the chemical environment and biochemical interactions of tryptophan, phenylalanine, histidine and tyrosine residues within proteins (Asher, 1993b; Jordon et al., 1995). Unfortunately, UVRRS by itself cannot focus on individual residues since the inelastic scattering of a common incident light will be  the same for all tyrosines within a molecule, or similarly for all tryptophans within a molecule. However, by combining mutagenesis with U V R R S , it is possible to focus on individual tryptophans or tyrosines by spectral comparison between the mutants and wild-type protein. In this section, preliminary data concerning the combination of mutagenesis and UVRRS to elucidate the roles of Y19 and Y85 will be discussed.  6.3.2  U V R R S protocol For specifics on the development of the UVRRS system used in this work, refer to the  Ph.D. thesis of Shane Greek (1997). In collaboration with Shane Greek (UBC), UVRRS spectra were acquired for pure protein in the presence and absence of 2.5 mM cellopentaose. Protein samples of wild-type CBDN1, Y19A and Y85A were diluted to 150 ug/ml in 40 mM MgSCU. Fiber-optic UVRRS data was collected from 1.5 ml samples using 12 minutes of integration time, and 500 uW of optical power at a wavelength of 230 nm. Each 12 minute integration was composed of eight independent 90 second acquisitions.  A l l samples were continuously stirred  using a 7 x 2 mm magnetic stir bar, and the entire UVRRS system was recalibrated with ethanol after each 12 minute integration. Additional control spectra were taken of water, 40 mM MgSC<4, and 2.5 m M cellopentaose in 40 m M MgSC»4. A l l spectral data were analyzed with C S M A software (Princeton Instruments) and origin software (Microcal). Processing of the spectral data (i.e. Fig. 6.4, 6.6) involved removal of the sloping background caused by Rayleigh scattering, subtraction of the water signal, and application of a five point low-pass fast fourier transform filter to reduce signal noise (Shane Greek, personal communication). The 40 m M MgSC»4 was used as an internal standard for peak intensity.  6.3.3  Evidence for structural involvement of Y19 but not Y85 within C B D N 1 Figure 6.4 includes the UVRRS spectra for wild-type CBDN1, Y19A and Y85A in the  presence and absence of cellopentaose.  Several Raman peaks have been labeled for reference  purposes and all are present at the expected Raman shift. The spectra of CBDN1, Y19A, and Y85A show certain differences in peak intensity. In the absence and presence of cellopentaose,  the spectra of CBDN1 and Y85A are similar with respect to the intensity of peaks associated with tryptophan and tyrosine residues, as compared to the sulphate standard. However, in the case of Y19A, all tyrosine and tryptophan peak intensities decrease in magnitude. For the tyrosine peaks, such affects are caused by reduced hydrogen bonding at the phenolic hydroxyl (Jordon et al., 1995). For tryptophan, these results can be interpreted as either a decrease in hydrophobicity of the indole side chain and/or a breaking of the side chain hydrogen bond (Jordon et al., 1995). The UVRRS data therefore suggest that the phenolic hydroxyl of Y19 forms an intramolecular hydrogen bond in CBDN1. This is supported by the ^ - ^ N HSQC data presented in section 6.1 which demonstrates that the region around Y19 experiences structural perturbations compared to wild-type CBDN1. Such perturbations in amide chemical shifts have been attributed to intramolecular hydrogen bonding properties in other proteins (Wishart et al., 1991). Furthermore, the Raman data indicate that the loss of hydrogen bond potential at Y19 appears to affect the structure of CBDN1 enough that one or both of the tryptophans in CBDN1 (W16, W137) also lose the ability to form a hydrogen bond, or move into a more surface exposed environment (Jordon et al., 1995). The Raman spectrum for Y19A shows similar reductions in intensity for tyrosine and tryptophan specific peaks when cellopentaose is present (see Fig. 6.4). This indicates that the structural perturbations observed in thefreestate also occur for Y19A in the bound state. Analysis of the distances between Y19 and either W16 or W137 indicate that a direct hydrogen bond between the side chains is highly unlikely (see table 6.4). Hydrogen bond donor and acceptor groups from either tryptophan to Y19 are at least 10 Angstroms apart, with the closest possible interresidue distance being 7.5 Angstroms. Hydrogen bonds typically range from two to four Angstroms (Creighton, 1993; Quiocho et al., 1997). A more likely explanation is that the mutation influences the intramolecular hydrogen bond network within Y19A, thereby affecting the solvent exposure or hydrogen bonding properties of one or both tryptophans. Alternatively, Y19 could interact indirectly with either W16 or W137 through a water molecule which serves as a hydrogen bond bridge (Quiocho and Vyas, 1984; Wright, 1990; Brummell et al., 1993; Quiocho et al., 1997).  W18  —I  1  1  1000  800  i  I  i  I  1200  1400  i  1  1600  R a m a n Shift (cm' ) 1  Fig. 6.4. Fiber-optic UVRRS data of wild-type CBDN1, Y19A, and Y85A in the absence (A, B, C ) and the presence (D, E, F) of cellopentaose. Important peaks are labeled: W18 and W16 are tryptophan specific peaks, v is a tyrosine specific region, and S0 " is an internal control for peak intensity. 2  ga/8b  4  Table 6.4. Summary of internuclear distances between various atoms in Y19, W16, and W137. Distance between Y19 and residue (Angstroms)  1  Atoms Involved  W16  W137  H-H closest distance  7.58  10.92  Y oxygen to nearest W hydrogen  8.64  10.86  W nitrogen to nearest Y hydrogen  11.4  12.3  1  Distances were calculated using a pdb file of the 3-D NMR structure of wild-type CBDN1  and Setor version 4.13.25 software  Figure 6.5 illustrates the position of W16 and W137 in relation to the chemical shifts that showed perturbations in the ^ - ^ N HSQC spectrum for Y19A. Both tryptophans are positioned near perturbed residues. Consequently, it is possible that either tryptophan, or both, could be affected by the mutation of Y19 to alanine. In summary, the UVRRS data, in combination with * H - N HSQC spectra indicate that 15  Y19 is involved in maintaining the tertiary structure of CBDN1. By mutation of Y19 to alanine in the Y19A mutant, the intramolecular hydrogen bond network in CBDN1 is perturbed, resulting in structural changes that were observed in NMR studies (see section 6.2). In contrast, the mutation of Y85A to alanine has little affect on the structure of the CBD, indicating that Y85 is not important for the structural integrity of CBDN1.  W16  Fig. 6.5. Residues within Y19A that exhibit chemical shift perturbations (coloured in blue) in the H - N HSQC spectrum, and their proximity with respect to Y19, W16 and W137. A) and B) represent two different views of CBDN1, caused by rotation of the image. ]  I5  6.3.4  Evidence for direct interaction of Y85 but not Y19 with cellopentaose Further information about Y85 and Y19 can be obtained from the v s a A ' s b region of the  Raman spectra in figure 6.4. This region is expanded in figure 6.6.  Reports in literature argue  that movement of the V8b peak reflect alterations in hydrogen bond strength at the phenolic hydroxyl of tyrosine, while the V8 peak is largely insensitive to changes in hydrogen bonding a  (Takeuchi et al., 1989; Rodgers et al., 1992; Jordon et al., 1995). Unfortunately, the vs and V8b a  regions are poorly resolved in the spectra presented in figure 6.6. Consequently, the sum total of these two peaks will be referred to as the V8a/V8b region in this work. Apart from a small shoulder on Y19A at 1599 cm , there are no significant differences -1  between the Y19A, Y85A and wild-type CBDN1 spectra in the absence of sugar. However, on the addition of cellopentaose, significant spectral changes are observed in the v s a / v s b region of wild-type CBDN1 and Y19A, but not for Y85A whose V8 /V8b region remains essentially a  unchanged. Because the V8 A'8b region is highly sensitive to hydrogen bonding (Rodgers et al., a  1992; Jordon et al., 1995), these data indicate that both wild-type CBDN1 and Y19A form hydrogen bonds to cellopentaose, but that no intermolecular hydrogen bonding is occurring at tyrosines in Y85A. This suggests that Y85 forms a hydrogen bond to cellopentaose, while Y19 does not. In figure 6.6, intensity of the shifted peak (on addition of cellopentaose) is greater for wild-type CBDN1 than Y19A. This is caused by the reduced affinity of Y19A for cellopentaose. In both spectra, identical concentrations of cellopentaose were used (2.5 mM). At that concentration, more than 99.5% of CBDN1 is bound to cellopentaose, compared to only 50% for either tyrosine mutant. Consequently, more hydrogen bonding will occur in the wild-type sample, leading to a more intense shift of the v s a / v s b region.  79  Fig. 6.6. Fiber-optic UVRRS spectra from figure 6.4 expanded to focus on the v / v region. "CP" and "free", respectivley, refer to the v / v region in the presence and abscence of 2.5 mM cellopentaose. 8 a  8 a  6 b  8 b  In conclusion, the Raman data suggest that Y85 forms a hydrogen bond to cellopentaose. The loss of this hydrogen bonding ability in the case of the alanine mutation at Y85 is likely to cause the observed reduction in binding affinity. An explanation of the affinity decrease in Y19A is somewhat more difficult. Y19 does not appear to be involved in a direct hydrogen bond with cellopentaose, but may be involved in significant van der Waal's interactions with the substrate. Such interactions, where aromatic residues stack against the pyranose rings of a carbohydrate substrate are important in the binding of many carbohydrate binding proteins (Quiocho, 1989). An alternative explanation for the decreased affinity of Y19A is that structural perturbations caused by Y19 replacement may displace another residue or residues that interact directly with cellopentaose.  Chapter 7: Conclusions 7.1  Current view of C B D N 1 binding  Previous work on CBDN1 has provided a three dimensional NMR structure, and a partial understanding of the binding properties (Coutinho, 1992; Johnson, 1996a,b; Tomme, 1996). It has been shown that CBDN1 binding is enthalpically driven, and largely unaffected by pH. In addition, a putative carbohydrate binding site was identified, and substrate specificity for a wide range of soluble and insoluble substrates has been established. In this work, the combination of site-directed mutagenesis with physico-chemical experiments has been used to characterize the binding site of CBDN1, and verify many of our hypotheses on how CBDN1 interacts with cellulose. For example, in previous work it had been demonstrated that CBDN1 associates with both soluble and insoluble substrates (Tomme et al., 1996). In this study, the similarities in binding kinetics (i.e. reversible binding, section 4.1) and mutant binding behaviour (section 5.5) on PASA and barley P-glucan argue that CBDN1 binds to single glucan chains in PASA. This argument is supported by the competition studies presented in section 4.6 which suggest that CBDN1 and CBDCex recognize different binding sites in PASA. Affinity electrophoresis studies imply that the active site of CBDN1 can accommodate for chemical differences in substrates, and still maintain a high affinity interaction. When the binding of CBDN1 mutants on HEC and barley P-glucan was compared (section 5.7), Q124, Q128, and R75 were shown to make different contributions to binding, although the affinity constant of CBDN1 is similar on both substrates (Tomme et al., 1996). The mutagenic study provided new information about the contribution of different binding site residues to binding. Ten alanine mutants representing two charged, six polar and two aromatic residues were constructed and characterized for binding affinity. The choice of mutation was based on both structural data (Johnson et al., 1996a) and NMR titration studies (Johnson et al., 1996b), the sum of which implied a potential involvement for each residue in binding. All mutations affected binding of the CBD to PASA and barley P-glucan. Reductions in binding affinity ranged from less than two-fold for mutants such as D90A and T87A to greater than 70-  fold in the case of Y85A and Y19A. Overall, these data suggest that the binding site of CBDN1 is a network of precisely positioned amino acids, involving residues that interact directly with substrate and supporting residues that hold the binding site residues in place. Maintenance of wild-type affinity requires the conservation of both spatial and biochemical properties throughout the binding site. Taken together with the binding data acquired on different substrates (Tomme et al., 1996), the CBDN1 binding site appears to be more tolerant to chemical differences in substrates than to chemical changes in binding site residues. Based on the alanine screening results, Y19 and Y85 were identified as two residues essential for high affinity binding. The results presented in chapter six suggest different roles for the tyrosines. cellopentaose.  First, the UVRRS data argue that Y85 forms a direct hydrogen bond with Additional UVRRS and NMR data indicate that Y85 is not involved in  maintaining the structural integrity of the CBD. Conversely, both NMR and UVRRS data suggest that Y19 is a structure-stabilizing component of CBDNL NMR experiments showed numerous chemical shift perturbations near Y19 in the Y19A mutant. These perturbations were most rationally explained by localized structural effects, caused by removal of the aromatic ring. The NMR findings are strengthened by UVRRS data which indicate perturbations in the intramolecular hydrogen bonding profile of the Y19A mutant. Furthermore, the UVRRS data suggest that Y19, unlike Y85, does not form a direct hydrogen bond with cellopentaose. The contribution of Y19 to binding must therefore be of a different nature, such as base stacking with the substrate or by maintaining the integrity of the binding site. Unfortunately, we do not (as of yet) have a structure of CBDN1 complexed with substrate. Thus, many of the interactions at the binding interface remain unknown. Such a structure would be extremely helpful in interpreting both the mutagenic data presented in this thesis, and the calorimetry data on CBDN1 presented elsewhere (Tomme et al., 1996).  7.2 7.2.1  Carbohydrate binding proteins Classification of carbohydrate binding proteins Carbohydrate binding proteins have been grouped into two classes based on the properties  of their sugar binding site (Quiocho et al., 1997). Group I proteins include periplasmic monosaccharide binding proteins, the phoshorylase active site, and the lad family of repressors. These proteins contain a deep binding site, typically consisting of two domains connected by a hinge region. The binding site is formed upon bending of the hinge region; this brings the two domains together to form an enclosed binding site. Consequently, the typical carbohydrate substrates for group I proteins are monosaccharides or short length oligosaccharides. Notable exceptions to the group I binding motif include cellobiohydrolase I and maltoporin which contain binding sites composed of a tunnel and pore respectively (Divne and Jones, 1994; Dutzler et al., 1996). Affinity constants for group I proteins are typically in the range of 10 per molar binding 6  site. Group II carbohydrate binding proteins contain a binding site on or near the surface of the protein and demonstrate affinity constants that are usually one to two orders of magnitude lower than the Group I carbohydrate binding proteins. The typical substrates for group II proteins include oligosaccharides of variable length and polysaccharides. Group II proteins include immunoglobulins, lectins, the phosphorylase storage site, starch binding domains and cellulose binding domains (Quiocho et al., 1997).  7.2.2  Forces involved in carbohydrate binding When reviewing binding by both classes of carbohydrate binding proteins, similar trends  emerge regarding the intermolecular forces involved in carbohydrate recognition. Hydrogen bonding and van der Waals interactions are the dominant forces involved in binding, at least for those proteins that bind soluble carbohydrates. Given the molecular nature of carbohydrates, it should not be surprising that these types of intermolecular forces dominate protein/carbohydrate interactions. Carbohydrates can be viewed as compounds which illustrate both polar and apolar characteristics. While the periphery of each sugar monomer clearly displays a high degree of  polarity, the sugar ring is quite hydrophobic. Consequently, one would expect proteins to use hydrogen bonds, polar and nonpolar van der Waals contacts to maximize their interaction with each saccharide unit of a carbohydrate substrate. X-ray crystallography of proteins complexed with their carbohydrate substrates has provided the best picture of sugar binding sites. A comparison of several crystal structures including those for the maltose binding protein (Spurlino et al., 1991; Quiocho et al., 1997), arabinose binding protein (Quiocho and Vyas, 1984), galactose binding protein (Vyas et al., 1988), various lectins (Cermakova et al., 1976; Wright, 1990) and antibodies (Lemieux et al., 1984; Lemieux, 1989; Cygler et al., 1991) indicate that hydrogen bonds are the main mediators of binding. This is due to the directional nature of hydrogen bonds, in addition to their strength relative to van der Waals interactions. Consequently, hydrogen bonds are believed to be responsible for the specificity involved in protein/carbohydrate interactions, while van der Waals forces are thought to stabilize the binding complex.  Water molecules are also found  protein/carbohydrate complexes, often occupying binding site voids and mediating hydrogen bonds between the sugar and protein. Hydrogen bonds are formed at sugar hydroxyl groups as well as acetal and hemiacetal oxygens, and often exhibit optimum or near optimum geometries (Quiocho and Vyas, 1984; Quiocho, 1989; 1993). Two basic types of hydrogen bonds are formed in protein/carbohydrate complexes: co-operative hydrogen bonds involving the simultaneous participation of a sugar hydroxyl as a hydrogen bond donor and acceptor, and bidentite hydrogen bonds involving the interaction of two adjacent hydroxyls with a different atom of the same polar side chain. In addition, binding sites often include both types of hydrogen bonds in close association, forming ' hydrogen bonded networks (Quiocho, 1989; 1993). In comparing group I and group II polysaccharide binding proteins, there are differences regarding the number of hydrogen bonds formed to their respective substrates. Complexes of the maltose binding protein (a group I protein) with maltose, maltotriose, and maltotetraose show 12, 14, and 14 direct hydrogen bonds to the sugar, respectively (Quiocho et al., 1997). For arabinose and galactose binding proteins, two other group I proteins, complexes with arabinose and  glucose, respectively, result in 10 and 12 direct hydrogen bonds (Quiocho and Vyas, 1984; Vyas et al., 1988). Interestingly, for the monosaccharide complexes involved, nearly every hydroxyl group is involved in a hydrogen bond. Because the type I proteins contain an enclosed binding site, this finding is hot entirely surprising since the energy required to isolate an hydroxyl group within a hydrophobic environment would normally be prohibitively high (Lemieux, 1989). It therefore makes energetic sense for sugar hydroxyls to find a polar partner and engage in a hydrogen bond. Furthermore, the formation of a hydrogen bond in a nonpolar environment tends to strengthen the bond itself. For group U proteins, fewer intermolecular hydrogen bonds are usually formed, despite the fact that their substrates are larger. The binding of methyl N-acetyl-|3-lactosimide by anti-I Ma involves only one key hydroxyl group (Lemieux et al., 1984) while the binding of the H-type 2 trisaccharide to a lectin from Galactia tenuiflora involves the recognition of a key polar grouping involving four hydroxyl groups (Lemieux, 1989). Exceptions do exist, however. For instance, the antibody Sel55-4 complexed with a trisaccharide involves seven intermolecular hydrogen bonds (Cygler et al., 1991; Brummell et al., 1993). It is probable that the formation of fewer hydrogen bonds in addition to the greater entropic cost associated with constricting longer chain carbohydrates is responsible for the decreased affinities typically observed in type U carbohydrate binding proteins. Van der Waals attractive forces are also important for ligand binding. These attractive forces arise when momentary polarizations in an atom's electron cloud cause an opposite polarization in a nearby electron cloud. The end result of these polarizations is a net attraction between the two oppositely charged induced dipoles. In crystallographic terms, van der Waals attractive forces are usually defined as contacts between two atoms that are less than four Angstroms apart. Depending on the identity of the atoms involved, van der Waal's interactions can be defined as polar/polar, polar/nonpolar, or nonpolar/nonpolar. A particular type of van der Waals interaction, important in protein/carbohydrate interactions is the stacking of aromatic residues against the faces of sugar rings. This stacking phenomenon has been found in many of the known stuctures of protein/carbohydrate complexes (Quiocho, 1993).  7.2.3  Mutational studies on carbohydrate binding sites NMR and x-ray crystallography are among the best methods to identify amino acid side  chains that form contacts with substrate. Site-directed mutagenesis extends these structural studies by providing information about the contribution of those amino acids to binding. Point mutations in carbohydrate binding proteins have characteristically led to reductions in affinity ranging from nil to as much as 2500-fold (Bass et al., 1991). In some cases, mutational analysis cannot be performed due to apparent expression problems with the mutant protein. Such problems have been correlated with both reduced thermostability and increased periplasmic proteolysis of the mutant, resulting in a diminished yield of protein (Pakula et al., 1986; Castro and Anderson, 1996; Kwon et al., 1996). Often, diminished yields are associated with the mutation of core hydrophobic residues in proteins. The loss of such residues is believed to create energetically unfavourable pockets in the core of proteins, leading to destabilization (Wells, 1991; Blaber et al., 1995). In several cases, the mutation of surface exposed hydrophilic residues has also led to diminished yields. This usually occurs when the mutation affects electrostatic interactions on the surface of the protein (Akke and Forsen, 1990; Bass et al., 1991; Hammen et al., 1995; Mayo et al., 1995), when interior hydrophobic groups become more exposed to solvent (Hurley et al., 1995) or when glycine is the residue being replaced (Castro and Anderson, 1996). Table 7.1 summarizes the results of 24 site-directed mutagenesis studies on carbohydrate binding proteins. The inclusion of a residue in table 7.1 indicates that the mutation of that residue resulted in a significant reduction in affinity, and hence could be termed a key binding site residue. From table 7.1, it is apparent that a wide range of residues provide important binding functions. Because hydrogen bonding is so important in carbohydrate binding, it is not surprising that a number of polar residues are crucial for binding, especially those like asparagine and glutamic acid which can form bidentite hydrogen bonds with the substrate (Quiocho, 1989). Two of the most prevalent residues in protein/carbohydrate interactions are tryptophan and tyrosine. This should not be surprising since either residue can participate in both base stacking and hydrogen bonding. Tryptophan and tyrosine differ in the size of their n electron system. Because the IT electron system is larger in tryptophan than tyrosine, a greater degree of electronic  polarization is expected in the case of the tryptophan side chain. Hence, tryptophan side chains contribute more to the binding enthalpy than tyrosine upon base stacking (Hoffren et al., 1995).  Table 7.1. Identity of key active site residues from 24 different SDM studies on carbohydrate binding proteins and enzymes. Residue  Total number  Protein  Reference  Tyrosine  1 1 2 1 1 1 1 1 1  chicken hepatic lectin mannose receptor ECorL glycosyltransferase Griffornia lectin Ricinus lectin wheat germ lectin human lysozyme CBDCBHI  Burrows et al., 1997 Mullin et al., 1997 Moreno et al., 1997 Penninga et al., 1996 Zhu et al., 1996 Sphgris et al., 1995 Naghora et al., 1995 Muraki et al., 1992 Linder et al., 1995  Tryptophan  2 1 1 1 1 2 1  glycosyltransferase CBDXylA lac repressor CBDCenA animal lectin xylose isomerase HGH receptor  Penninga et al., 1996 Poole et al., 1993 Chang and Matthews, 1995 Din et al., 1994 Iobst and Drickamer, 1994 Lambeir et al., 1992 Bass et al., 1991  Asparagine  1 2 1 1  Griffonia lectin human galectin-1 animal lectin PHA-Z  Zhu et al., 1996 Hirabayashi and Kasai, 1994 Iobst et al., 1994 Mirkov and Chrispeels, 1993  Glutamic acid  1 1 1 1  mannitol permease lactose transport protein human galectin-1 animal lectin  Saraceni-Richards et al., 1997 Poolman et al., 1995 Hirabayashi and Kasai, 1994 Iobst et al., 1994  Aspartic acid  1 1 1  bark lectin Griffonia lectin lac repressor  Nishiguchi et al., 1997 Zhu et al., 1996 Chang and Matthews, 1995  Cysteine  1 4  lima bean lectin prolactin receptor  Jordon and Goldstein, 1995 Rozakis-Adcock et al., 1991  Arginine  1 2  sialoadhesins human galectin-1  Vinson et al., 1996 Hirabayashi and Kasai, 1994  Histidine  1 1  human galectin-1 animal lectin  Hirabayashi and Kosai, 1994 Iobst et al., 1994  Phenylalanine  1 1  bark lectin animal lectin  Nishiguchi et al., 1997 Iobst and Drickamer, 1994  Valine  1  chicken hepatic lectin  Burrows et al., 1997  7.2.3  Comparison of CBDN1 to other carbohydrate binding proteins CBDN1 can be discussed in terms of two classification systems. First, CBDN1 can be  assigned to the more general system of Quiocho et al. (1997) for classifying carbohydrate binding proteins. Or second, CBDN1 can be compared to other cellulose binding proteins in Tomme et al.'s (1995, 1997) C B D classification system. CBDN1 fits nicely into Quiocho et al.'s (1997) system as a group II carbohydrate binding protein. The C B D has a surface exposed binding site, and the magnitude of its demonstrated affinity for substrate is very similar to other group II proteins. Typical of most other carbohydrate binding proteins, CBDN1 also demonstrates an enthalpic driving force for binding (Tomme et al., 1996). In addition, residues with polar groups, most notably tyrosine, are essential for binding. This is again typical of the types of residues found in the binding sites of other carbohydrate binding proteins (see Table 7.1). In Tomme et al.'s (1995, 1997) system, CBDN1 is classified as a family IV C B D . CBDN1 is the first family IV C B D to be solved structurally (Johnson et al., 1996a), and have its binding site characterized by mutagenesis. Moreover, CBDN1 is only the third C B D , along with the family I C B D , CBDCBHI (Reinikainen et al., 1992; Linder et al., 1995) and the family II CBD, CBDCenA (Din et al., 1994) to be characterized by site-directed mutagenesis. Of these CBDs, only the structure of CBDCenA has not been solved; however, CBDCenA is probably similar in structure to CBDCex (see Fig. 1.2) (Din et al., 1994; X u et al., 1995). A comparison of the mutagenic results on these three CBDs is difficult to interpret based on the structural and functional differences between CBDN1 and the other two CBDs. First, CBDN1 does not bind to crystalline cellulose, unlike CBDCBHI and CBDCenA. Second, CBDN1 binds its substrates in a shallow cleft as opposed to a planar binding face in the case of C B D C B H I and (probably) CBDCenA. Despite many differences pertaining to both structure and function in CBDN1, CBDCenA, and CBDCBHI, it is interesting to note that aromatic residues are essential for the tight binding of all three to cellulose. Tryptophans have proven to be important for CBDCenA (Din et al., 1994), while tyrosines mediate the binding of CBDN1 (current research) and C B D C B H I (Linder et al.,  1995). In addition, both CBDN1 and CBDCBHI contain one tyrosine, present within a loop region, that is important for maintaining structural integrity of the binding domain (current research; Linder et al., 1995). Future work on CBDN1 should include improved efforts to determine the intermolecular forces involved in complexes of CBDN1 and cellulose. Areas to investigate include the modeling of different substrates into the CBDN1 binding site, efforts to produce a structure of the CBD complexed with cellulose, and conservative mutagenesis of Y19 and Y85. The mutation of Y19 and Y85 to phenylalanine, followed by analysis using the binding studies presented in this thesis, could depict whether hydrogen bonding or van der waal's interactions dominate the interaction of each tyrosine with cellulose. Calorimetric studies on other mutants of Y19 and Y85 may also decipher the thermodynamic contributions of each tyrosine to binding.  Chapter 8: References Akke, M., and Forsen, S. 1990. Protein Stability and Electrostatic Interaction Between Solvent Exposed Charged Side Chains. Proteins: Struct. Funct. Genet. 8:23-29. Asher, S. A., Ludwig, M., and Johnson, C. R. 1986. UV Resonance Raman Excitation Profiles of the Aromatic Amino Acids. J. Am. Chem. Soc. 108:3186-3197. Asher, S. A. 1993a. UV Resonance Raman Spectroscopy for Analytical, Physical and Biophysical Chemistry, Part 1. Anal. Chem. 65(2):59-66. Asher, S. A. 1993b. UV Resonance Raman Spectroscopy for Analytical, Physical and Biophysical Chemistry, Part 2. Anal. Chem. 65(4) :201-210. Attala, R. H., and VanderHart, D. L. 1989. Studies on the structure of cellulose using raman spectroscopy and solid state C NMR. In: Schuerch, C. (Ed.), Cellulose and Wood: Chemistry and Technology. Wiley International, New York, pp. 169-188. 1 3  Attala, R. H. 1990. The Structures of Cellulose. In: Attala. R. H. (Ed.). The Structures of Cellulose. ACS Svpm., Ser. 340. American Chemical Society, Washington DC, pp. 1-14. Attala, R. H. 1993. The Structures of Native Celluloses. In: Suominen, P., and Reinikainen, T. (Eds.), Trichoderma reset Cellulases and Other Hydrolases. Foundation for Biotechnical and Industrial Fermentation, Helsinki, pp. 25-39. Bass, S. H., Mulkerrin, M. G., and Wells, J. A. 1991. A systematic mutational analysis of hormone-binding dterminants in the human growth hormone receptor. Proc. Nad. Acad. Sci. USA. 88:4498-4502. Beacham, I. R. 1979. Periplasmic enzymes in gram-negative bacteria. Int. J. Biochem. 10:877-883. Beguin, P., Gilkes, N. R., Kilburn, D. G., Miller, R. C. Jr., O'Neil, G. P., and Warren, R. A. J. 1987. Cloning of Cellulase Genes. Crit. Rev. Biotechnol. 6(2):129-159. Beguin, P., and Aubert, J-P. 1994. The biological degradation of cellulose. FEMS Microbiol. Rev. 13:25-58. Bhattacharjee, S. S., and Perlin, A. S. 1971. Enzymatic degradation of carboxymethylcellulose and other cellulose derivatives. J. Polymer Sci. 36:509-521. Blaber, M., Baase, W. A., Gassner, N., and Mathhews, B. W. 1995. Alanine Scanning Mutagenesis of the a-Helix 115-123 of Phage T4 Lysozyme: Effects on Structure, Stability and the Binding of Solvent. J. Mol. Biol. 246:317-330. Black, G. W., Hazelwood, G. P., Millward-Saddler, S. J., Laurie, J. I., and Gilbert, H. J. 1995. A modular xylanase containing a novel non-catalytic xylan-specific binding domain. Biochem. J. 307:191-195. Black, G. W., Rixon, J. E., Clarke, J. H., Hazelwood, G. P., Ferreira, L. M. A., Bolam, D. N., and Gilbert, H. J. 1997. Cellulose binding domains and linker sequences potentiate the activity of hemicellulases against complex substrates. J. Biotechnol. 57:59-69.  Blight, M. A., Chervaux, C , and Holland, I. B. 1994. Protein secretion pathways in Escherichia coli. Curr. Opin. Biotechnol. 5:468-474. Bray, M. R., Johnson, P. E., Gilkes, N. R., Mcintosh, L. P., Kilburn, D. G., and Warren, R. A. J. 1996. Probing the role of tryptophan residues in a cellulose-binding domain by chemical modification. Protein Sci. 5:2311-2318. Brummell, D. A., Sharma, V. P., Anand, N. N., Bilous, D., Dubuc, G., Michniewicz, J., MacKenzie, C. R., Sadowska, J., Sigurskjold, B. W., Sinnott,B., Young, N. M., Bundle, D. R., and Narang, S. A. 1993. Probing the Combining Site of an Anti-Carbohydrate Antibody by Saturation-Mutagenesis: Role of the Heavy-Chain CDR3 Residues. Biochemistry. 32(4): 11801187. Brun, E., Moriaud, F., Gans, P., Blackledge, M. J., Barras, F., and Marion, D. 1997. Solution Structure of the Cellulose-Binding Domain of the Endoglucanase Z Secreted by Erwinia chrysanthemi. Biochemistry. 36:16074-16086. Bundle, D. R., Eichler, E., Gidney, M. A. J., Meldal, M., Ragauskas, A., Sigurskjold, B. W., Sinnott, B., Watson, D. C , Yaguchi, M., and Young, N. M. 1994. Molecular recognition of a Salmonella trisaccharide epitope by monoclonal antibody Sel55-4. Biochemistry. 33:51725182. Burrows, L., Iobst, S. T., and Drickamer, K. 1997. Selective binding of N-acetylglucosamine to the chicken hepatic lectin. Biochem. J. 324:673-680. Cannon, R. E., and Anderson, S. M. 1990. Biogenesis of Bacterial Cellulose. Curr. Rev. Microbiol. 17(6):435-446. Castro, M. J. M. and Anderson, S. 1996. Alanine Point-Mutations in the Reactive Region of Bovine Pancreatic Trypsin Inhibitor: Effects on the Kinetics and Thermodynamics of Binding to (3-trypsin and a-Chymotrypsin. Biochemistry. 35:11435-11446. Cermakova, M., Entlicher, G., and Kocourek, J. 1976. Studies on phytohemagglutinins. XXVU. A study of the pea lectin binding site. Biochimica et Biophysica Acta. 420:236-245. Chang, W. I., and Matthews, K. S. 1995. Role of Asp274 in lac repressor: diminished sugar binding and altered conformational effects in mutants. Biochemistry. 34(28):9227-9234. Chowdhury, P. S., Kushwaha, A., Abrol, S., and Chaudhary, V. K. 1994. An Expression System for Secretion and Purification of a Genetically Engineered Thermostable Chimera of Protein A and Alkaline Phosphatase. Protein Expr. Purif. 5:89-95. Clarke, J. H., Davidson, K., Gilbert, H. J., Fontes, C. M. G. A., and Hazelwood, G. P. 1996. A modular xylanasefrommesophilic Cellulomonasfimicontains the same cellulose-binding and thermostabilizing domains as xylanasesfromthermophilic bacteria. FEMS Microbiol. Lett. 139:27-35. Coleman, G.S. 1978. The metabolism of cellulose, glucose, and starch by the rumen ciliate protozoa Eudiplodinium maggii. J. Gen. Microbiol. 107:359-366. Coutinho, J. B., Moser, B., Kilburn, D. G., Warren, R. A. J., and Miller, R. C. Jr. 1991. Nucleotide sequence of the endoglucanse C gene (cenC) of Cellulomonasfimi,its high-level expression in Escherichia coli, and characterization of its products. Mol. Microbiol. 5:12211233.  Coutinho, J. B., Gilkes, N. R., Warren, R. A. J., Kilburn, D. G., and Miller, R. C. Jr. 1992. The binding of Cellulomonasfimiendoglucanse C (CenC) to cellulose and Sephadex is mediated by the N-terminal repeats. Mol. Microbiol. 6(9): 1243-1252. Coutinho, J. B. Gilkes, N. R., Kilburn, D. G., Warren, R. A. J., and Miller. R. C. Jr. 1993. The nature of the ceUulose-binding domain affects the activities of a bacterial endoglucanse on different forms of cellulose. FEMS Microbiol. Lett. 113:211-218. Creagh, A. L., Ong, E., Jervis, E., Kilburn, D. G., and Haynes, C. A. 1996. Binding of the cellulose-binding domain of exoglucanse Cex from Cellulomonasfimito insoluble microcrystalline cellulose is entropically driven. Proc. Natl. Acad. Sci. USA. 93:12229-12234. Creagh, A. L., Koska, J., Johnson, P. E., Tomme, P., Joshi, M. D., Mcintosh, L. P., Kilburn, D. G., and Haynes, C. A. 1998. Stability and Oligosaccharide Binding of the N l CelluloseBinding Domain of CellulomonasfimiEndoglucanase CenC. Biochemistry. 10:3529-3537. Creighton, T. E. 1993. Proteins, second edition. W. H. Freeman and Company. New York Cygler, M., Rose, D. R., and Bundle, D. R. 1991. Recognition of a Cell-Surface Oligosaccharide of Pathogenic Salmonella by an Antibody Fab Fragment. Science. 253:442445. Din, N., GUkes, N. R., Tekant, B., Miller, R. C. Jr., Warren, R. A. J., and Kilburn, D. G. 1991. Non-hydrolytic disruption of cellulose fibres by the binding domain of a bacterial cellulase. Biotechnology. 9:1096-1099. Din, N., Forsythe, I. J., Burtnick, L. D., Gilkes, N. R., Miller, R. C. Jr., Warren, R. A. J., and Kilburn, D. G. 1994. The cellulose-binding domain of endoglucanase A (CenA) from Cellulomonasfimi:evidence for the involvement of trytophan residues in binding. Mol. Microbiol. ll(4):747-755. Divne, C , and Jones, A. 1994. The three-dimensional sturcture of the catalytic core of cellobiohydrolase I from Trichoderma reesei. Science. 265:524-528. Dutzler, R., Wang, Y-F., Rizkallah, P. J., Rosenbusch, J. P., and Schirmer, T. 1996. Crystal structure of various maltooligosaccharides bound to maltoporin reveal a specific sugar translocation pathway. Structure. 4:127-134. Garda, A. L., Ferndndez-Abalos, J. M., Sanchez, P., Ruiz-Arribas, A., and Santamaria, R. I. 1997. Two genes encoding an endoglucanase and a cellulose-binding protein are clustered and co-regulated by a TTA codon in Streptomyces halstedii JM8. Biochem. J. 324:403-411. Gerngross, U. T., Romaniec, M. P. M., Kobayashi, T., Huskisson, N. S., and Demain, A. L. 1993. Sequencing of a Clostridium thermocellum gene (cipA) encoding the cellulosomal S L protein reveals an unusual degree of internal homology. Mol. Microbiol. 8:325-334. Gilkes, N. R., Kilburn, D. G., Langsford, M. L., Miller, R. C. Jr., Wakarchuk, W. W., Warren, R. A. J., Whittle, D. J., and Wong, W. K. R. 1984. Isolation and characterization of Eschericia coli clones expressing cellulase genes from Cellulomonasfimi.J. Gen. Microbiol. 130:1377-1384. Gilkes, N. R., Warren, R. A. J., Miller, R. C. Jr., and Kilburn, D. G. 1988. Precise excision of the cellulose binding domains from two Cellulomonasfimicellulases by a homologous protease and the effect on catalysis. J. Biol. Chem. 263:10401-10407.  Gilkes, N. R., Kilburn, D. G., Miller, R. C. Jr., and Warren, R. A. J. 1989. Structural and Functional Analysis of a Bacterial Cellulase by Proteolysis. J. Biol. Chem. 264(30): 1780217808. Gilkes, N. R., Henrissat, B., Kilburn, D. G., Miller, R. C. Jr., and Warren, R. A. J. 1991. Domains in Microbial P-1,4-Glycanases: Sequence Conservation, Function, and Enzyme Families. Microbiol. Rev. 55(2):303-305. Gilkes, N. R., Jervis, E., Henrissat, B., Tekant, B., Miller, R. C. Jr., Warren, R. A. J., and Kilburn, D. G. 1992. The Adsorption of a Bacterial Cellulase and Its Two Isolated Domains to Crystalline Cellulose. J. Biol. Chem. 267(10):6743-6749. Graham, R. W., Greenwood, J. M., Warren, R. A. J., Kilburn, D. G., and Trimbur, D. E. 1995. The pTugA and pTugAS vectors for high-level expression of cloned genes in Escherichia coli. Gene. 158:51-54. Greek, S. 1997. Development and characterization of an optical fiber based instrument for ultraviolet resonance Raman spectroscopy of biomolecules. Ph. D thesis. Department of Electrical and Computer Engineering, University of BC, Vancouver, BC, Canada. Greenwood, J. M., Gilkes, N. R., Kilburn, D. G., Miller, R. C. Jr. and Warren, R. A. J. 1989. Fusion to an endoglucanse allows alkaline phosphatase to bind to cellulose. FEBS Lett. 244(1):127-131. Greenwood, J. M., Ong, E., Gilkes, N. R., Warren, R. A. J., Miller, R. C. Jr., and Kilburn, D. G. 1992. Cellulose-binding domains: potential for purification of complex proteins. Protein Eng. 5(4):361-365. Greenwood, J. M., Gilkes, N. R., Miller, R. C. Jr., Kilburn, D. G., and Warren, R. A. J. 1994. Purification and Processing of Cellulose-Binding Domain-Alkaline Phospatase Fusion Proteins. Biotechnol. Bioeng. 44:1295-1305. Guo, Z., Arfman, N., Ong, E., Gilkes, N. R., Kilburn, D. G., Warren, R. A. J., and Miller, R. C. Jr. 1988. Leakage of Cellulomonasfimicellulases from Escherichia coli. FEMS Microbiol. Lett. 49:279-283. Hall, L, Black, G. W., Ferreira, L. M. A., Millward-Saddler, S. J., Ali, B. R. S., Hazlewood, G. P., and Gilbert, H. J. 1995. The non-catalytic cellulose binding domain of a novel cellulase from Pseudomonas fluorescens subsp. cellulosa is important for the efficient hydrolysis of Avicel. Biochem. J. 309:749-756. Hammen, P. K., Scholtz, J. M., Anderson, J. W., Waygood, E. B., and Klevit, R. E. 1995. Investigation of a side-chain-side-chain hydrogen bond by mutagenesis, thermodynamics, and NMR spectroscopy. Protein Sci. 4:936-944. Hanahan, D. 1983. Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166:557-580. Harada, I., and Takeuchi, H. 1986. Raman and ultraviolet resonance Raman spectra of proteins and related compounds. In R. J. H. Clark and R. E. Hester (Eds.), Spectroscopy of biological systems. John Wiley, New York, pp. 113-175. Hatakeyama, T. 1989. Structure and properties of the amorphous region of cellulose. In J. F. Kennedy, G. O. Phillips, and P. A. Williams (Eds.), Cellulose - Structural and Functional Aspects. Ellis Horwood Limited, West Sussex, pp. 45-55.  Hayashi, A. 1989. Supermolecular structure of cellulose - structural memory phenomena and its causes. In J. F. Kennedy, G. O. Phillips, and P. A. Williams (Eds.), Cellulose - Structural and Functional Aspects. Ellis Horwood Limited, West Sussex, pp. 35-43. Hestrin, S. 1963. Bacterial Cellulose. In R. L. Whistler (Eds). Methods in Carbohydrate Chemistry 3. Academic Press, New York, pp. 4. Hirabayashi, L, and Kasai, K. 1994. Further evidence by site-directed mutagenesis that conserved hydrophilic residues form a carbohydrate-binding site of human galectin-1. Glycoconj.  J. ll(5):437-442.  Hofman, H. 1993. Hydrophobically modified hydroxyethyl cellulose - a study of hydrophobe length and reagent reactivity. In: Kennedy, J. F., Phillips, G. O., and Williams, P. A. (Eds.). Cellulosics: Chemical. Biochemical and Material Aspects. Ellis Horwood Limited, West Sussex, pp. 305-311. Hoffren, A-M., Teeri, T., and Teleman, O. 1995. Molecular dynamics simulation of fungal cellulose-binding domains: differences in molecularrigiditybut a preserved cellulose bmding surface. Protein Eng. 8(5):443-450. Hogset, A., Blingsmo, O. R., Saether, O., Gautvik, V. T., Holmgren, E., Hartmanis, M., Josephson, S., Gabrielsen, O. S., Gordeladze, J. O., Alestrom, P., and Gautvik, K. M. 1990. Expression and Characterization of a Recombinant Human Parathyroid Hormone Secreted by Escherichia coli Employing and Staphylococcal Protein A Promoter and Signal Sequence. J. Biol. Chem. 265(13):7338-7344. Holland, I. B., Mackman, N., Nicaud, J. M. 1986. Secretion of proteins from bacteria. Biotechnology. 4:427-431. Hsu, J-C, and Penner, M. H. 1989. Influence of Cellulose Structure on Its Digestibility in the Rat. J. Nutrition. 119:872-878. Hurley, J. K., Coffrey, M. S., Markley, J. L., Cheng, H., Xia, B., Chae, Y. K., Holden, H. M., and Tollin, G. 1995. Mutations of surface residues in Anadaena vegatative and heterocyst ferredoxin that affect thermodynamic stability as detected by guanadine hydrochloride denaturation. Protein Sci. 4:58-64. Iobst, S. T., Wormald, M. R., Weis, W. I., Dwek, R. A., and Drickamer, K. 1994. Binding of sugar ligands to Ca(2+)-dependent animal lectins. I. Analysis of mannonse binding by sitedirected mutagenesis and NMR. J. Biol. Chem. 269(22): 15505-15511. Iobst, S. T., Drickamer, K. 1994. Binding of sugar ligands to Ca(2+)-dependent animal lectins. U. Generation of high-affinity galactose binding by site-directed mutagenesis. J. Biol. Chem. 269(22):15512-15519. Ito, W., Iba, Y., and Kurosawa, Y. 1993. Effects of Substitutions of closely Related Amino Acids at the Contact Surface in an Antigen-Antibody Complex on Thermodynamic Parameters. J. Biol. Chem. 268(22): 16639-16647. Jervis, E., Haynes, C. E., and Kilburn, D. G. 1997. Surface Diffusion of Cellulases and Their Isolated Binding Domains on Cellulose. J. Biol. Chem. 272(38): 24016-24028. Johnson, P. E., Joshi, M. D., Tomme, P., Kilburn, D. G., Mcintosh, L. P. 1996a. Structure of the N-terrninal cellulose-binding domain of CellulomonasfimiCenC determined by nuclear magnetic resonance spectroscopy. Biochemistry. 35:14381-14394.  Johnson, P. E., Tomme, P., Joshi, M. D., and Mcintosh, L. P. 1996b. Interaction of Soluble Cellooligosaccharides with the N-Terminal Cellulose-Bindng Domain of Cellulomonas fimi CenC. 2. NMR and Ultraviolet Absorption Spectroscopy. Biochemistry. 35:13895-13906. Jones, T. H., de Renobales, M., and Pon, N. 1979. Cellulases released during the germination of Dictyostelium discoideum spores. J. Bacteriol. 137:752-757. Jordon, E. T., and Goldstein, I. J. 1995. Site-directed mutagenesis studies on the lima bean lectin. Altered carbohyo^ate-binding specificities result from single arnino acid substitutions. Eur. J. Biochem. 230(3):958-964. Jordon, T., Eads, J. C , and Spiro, T. G. 1995. Secondary and tertiary structure of the A-state of cytochrome c from resonance Raman spectroscopy. Protein Sci. 4:716-728. Karlsson, E. N., Roxa, E. B., and Hoist, O. 1997. Cloning and sequence of a thermostable multidomain xylanase from the bacterium Rhodothermus marinus. Biochimica et Biophysica Acta. 1353:118-124. Kline, T. P., and Mueller, L. 1992. NMR study on solution structure of the site-specific mutant Leu48-Ala transforming growth factor alpha. Int. J. Peptide Protein Res. 39:111-116. Kolpak, F. J., Weih, M., and Blackwell, J. 1978. Mercerization of cellulose: 1. Determination of the structure of Mercerized cotton. Polymer. 19:123-131. Kolpak, F. J., and Blackwell, J. 1978. Mercerization of cellulose: 2. The morphology of Mercerized cotton cellulose. Polymer. 19:123-131. Kraulis, P. J., Clore, G. M., Nilges, M., Jones, T. A., Pettersson, G., Knowles, J., and Gronenborn, A. 1989. Detenrunation of the three-dimensional solution structure of the Cterminal domain of cellobiohydrolase from Trichoderma reesei. A study using nuclear magnetic resonance and hybrid distance geometry-dynamical simulated annealing. Biochemistry. 28:7241-7257. Kulshreshtha, A. K., and Dweltz, N. E. 1973. Paracrystalline Lattice Disorder in Cellulose. I. Reappraisal of the Application of the Two-Phase Hypothesis to the Analysis of Powder X-Ray Diffractograms of Native and Hydrolyzed Cellulosic Materials. J. Polymer Sci. 11:487-497. Kwon, W. S., Da Silva, N. A., and Kellis, J. T. Jr. 1996. Relationship between thermal stability, degradaton rate and expression yield of barnase variants in the periplasm of Escherichia coli. Protein Eng. 9(12):1197-1202. Lambeir, A. M., Lauwereys, M., Stanssens, P., Mrabet, N. T., Snauwaert, J., van Tilbeurgh, H., Matthyssens, G., Lasters, I., De Maeyer, M., and Wodak, S. J. 1992. Protein engineering of xylose (glucose) isomerase from Actinoplanes missouriensis. 2. Site-directed mutagenesis of the xylose binding site. Biochemistry. 31(24):5459-5466. Lee, S. B., and Kim, I. H. 1983. Structural properties of cellulose and cellulase reaction mechanism. Biotechnol. Bioeng. 25:33-51. Lemieux, R. U., Wong, T. C , Liao, J., and Kabat, E. A. 1984. The combining site of anti-I Ma (Group 1). Mol. Immunol. 21(9):751-759. Lemieux, R. U. 1989. The Origin of the Specificity in the Recognition of Oligosaccharides by Proteins. Chem. Soc. Rev. 18:347-374.  Lemieux, R. U., Delbaere, L. T. L, Beierbeck, H., and Spohr, U. 1991. Involvement of water in host-guest interactions. Ciba Found. Symp. 158:231-248. Lemieux, R. U. 1993. How proteins recognize and bind oligosaccharides. ACS Symp. Ser. 519:5-18. Linder, M., Mattinen, M-L., Kontteli, M., Lindeberg, G., Stahlberg, J., Drakenberg, T., Reinkainen, T., Pettersson, G., and Annila, A. 1995. Identification of functionally important amino acids in the cellulose-binding domain of Trichoderma reesei cellobiohydrolase I. Protein Sci. 4:1056-1064. Linder, M., and Teeri, T. 1996. The cellulose-binding domain of the major cellobiohydolase of Trichoderma reesei exhibits true reversibility and a high exchange rate on crystalline cellulose. Proc. Natl. Acad. Sci. USA. 93:12251-12255. Linder, M., and Teeri, T. T. 1997. The roles and function of cellulose-binding domains. J. Biotechnol. 57:15-28. Mandels, M. 1982. Cellulases. Ann. Rep. Ferment. Proc. 5:35-78. Maniatus, T., Fritsch, E. F., and Sambrook, J. 1982. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Press, New York. Markley, J. L., Meadows, D. H., and Jardetzky, O. 1967. Nuclear magnetic resonance studies of of helix-core transitions in polyamino acids. J. Mol. Biol. 27:25-35. Mayo, K. H., Ilyina, E., Roongta, V., Dundas, M., Joseph, L, Lai, C. K., Maione, T., and Daly, T. J. 1995. Heparin binding to platelat factor-4. An NMR and site-directed mutagenic study: arginine residues are crucial for binding. Biochem J. 312:357-365. McMurry, J. (Ed.). 1990. Fundamentals of Organic Chemistry, second edition. Brooks Cole, Pacific Grove, pp. 124-154. Medve, J., Stahlberg, J., and Tjerneld, F. 1997. Isotherms for Adsorption of Cellobiohydrolase I and II from Trichoderma reseei on Microcrystalline Cellulose. Appl. Biochem. Biotech. 66:3956. Meinke, A., Braun, C , Gilkes, N. R., Kilburn, D. G., Miller, R. C. Jr., and Warren, R. A. J. 1991. Unusual sequence organization in CenB, an inverting endoglucanase from Cellulomonas fimi. J. Biochem. 173:308-314. Meinke, A., Gilkes, N. R., Kilburn, D. G., Miller, R. C. Jr., and Warren, R. A. J. 1993. Cellulose-Binding Polypeptides from Cellulomonasfimi:Endoglucanse D (CenD), a Family A p1,4-Glueanse. J. Bacterid. 175(7): 1910-1918. Millward-Saddler, S. J., Poole, D. M., Henrissat, B., Hazlewood, G. P., Clarke, J. H. and Gilbert, H. J. 1994. Evidence for a general role for high-affinity non-catalytic cellulose binding domains in microbial plant cell wall hydrolases. Mol. Microbiol. ll(2):375-382. Millward-Sadler, S. L , Davidson, K., Hazelwood, G. P., Black, G. W., Gilbert, H. L , and Clarke, J. H. 1995. Novel cellulose-binding domains, NodB homologues and conserved moduar architecture in xylanasesfromthe aerobic soil bacteria Pseudomonas fluorescens subsp. cellulosa and Cellvibrio mixtus. Biochem. J. 312:39-48.  Mirkov, T. E., and Chrispeels, M. J. 1993. Mutation of Asnl28 to Asp of Phaseolus vulgaris leucoagglutinin (PHA-L) eliminates carbohydrate-binding and biological activity. Glycobiology. 3(6):581-587. Morag, E., Lapidot, A., Govorko, D., Lamed, R., Wilchek, M., Bayer, E. A., and Shoham, Y. 1995. Espression, Purification, and Characterization of the Cellulose-Binding Domain of the Scaffoldin Subunit from the Cellulosome of Clostridium thermocellum. Appl. Env. Microbiol. 61(5):1980-1986. Moreno, E., Teneberg, S., Adar, R., Sharon, N., Karlsson, K. A., Angstrom, J. 1997. Redefinition of the carbohydrate specificity of Erythrina corallodendron lectin based on solidphase binding assays and molecular modeling of native and recombinant forms obtained by sitedirected mutagenesis. Biochemistry. 36(15):4429-4437. Moser, R., Gilkes, N. R., Kilburn, D. G., Warren, R. A. J., and Miller, R. C. Jr. 1989. Purification and characterization of endoglucanase C of Cellulomonasfimi,cloning the gene and analysis of in vivo transcripts of the gene. Appl. Env. Microbiol. 55:2480-2487. Mullin, N. P., Hitchen, P. G., and Taylor, M. E. 1997. Mechanism of C a and monosaccharide binding to a C-type carbohydrate-recognition domain of the macrophage mannose receptor. J. Biol. Chem. 272(9):5668-5681. 2 +  Muraki, M., Harata, K., and Jigami, Y. 1992. Dissection of the functional role of structural elements of tyrosine-63 in the catalytic action human lysozyme. Biochemistry. 31(38):92129219. Nagahora, H., Harata, K., Muraki, M., and Jigami, Y. 1995. Site-directed mutagenesis and sugar-binding properties of the wheat germ agglutinin mutants Tyr73Phe and Phel 16Tyr. Eur. J. Biochem. 233:27-34. Nardin, R., and Vincendron, M. 1988. Conformational analysis of f3(l-4)-linked polysaccharides in aprotic solvent solutions. Makromol. Chem. 189:153-162. Nishiguchi, M., Yoshida, K., Sumizono, T., and Tazaki, K. 1997. Studies by site-directed mutagenesis of the carbohydrate-binding properties of a bark lectin from Robina pseudoacacia. FEBS Lett. 403(3):294-298. O'Neill, G. P., Kilburn, D. G., Warren, R. A. J., and Miller, R. C. Jr. 1986. Overproduction from a Cellulase Gene with a High Guanosine-Plus-Cytosine Content in Escherichia coli. Appl. Env. Microbiol. 52(4):737-743. Ong, E., Greenwood, J. M., Gilkes, N. R., Kilburn, D. G., Miller, R. C. Jr., and Warren, R. A. J. 1989. The cellulose-binding domains of cellulases: tools for biotechnology. Trends Biotechnol. 7(9):239-243. Ong, E., Gilkes, N. R., Miller, R. C. Jr., Warren, R. A. J., and Kilburn, D. G. 1991. Enzyme immobilization using a cellulose-binding domain: Properties of a (3-glucosidase fusion protein. Enzyme Microb. Technol. 13:59-65. Ong, E., Gilkes, N. R., Miller, R. C. Jr., Warren, R. A. J., and Kilburn, D. G. 1993. The Cellulose-Binding Domain (CBDCex) of an Exoglucanse from Cellulomonasfimi:Production in Escherichia coli and Characterization of the Polypeptide. Biotechnol. Bioeng. 42:401-409.  Owalabi, J. B., Beguin, P., Kilburn, D. G., Miller, R. C. Jr., and Warren, R. A. J. 1988. Expression in Escherichia coli of the Cellulomonasfimistructural gene for endoglucanse B. Appl. Env. Microbiol. 54:518-523. Pages, S., Gal, L., Belaich, A., Gaudin, C , Tardif, C , and Belaich, J-P. 1997. Role of Scaffolding Protein CipC of Clostridium cellulolyticum in Cellulose Degradation. J. Bacteriol. 179(9):2810-2816. Pakula, A. A., Young, V. B., and Sauer, R. T. 1986. Bacteriophage X cro mutations: Effects on activity and intracellular degradation. Proc. Natl. Acad. Sci. USA. 83:8829-8833. Parrish, F. W., Perlin, A. S., and Reese, E. T. 1960. Selective enzymolysis of poly-P-Dglucans, and the structure of the polymers. Can. J. Chem. 38:2094-2104. Penninga, D., van der Veen, B. A., Knegtel, R. M., van Hijum, S. A., Rozeboom, H. J., Kalk, K. H., Dijkstra, B. W., and Dijkhuizen, L. 1996. The raw starch binding domain of cyclodextrin glycosyltransferase from Bacillus circulans strain 251. J. Biol. Chem. 271(51):32777-32784. Pereira, A. N., Mobedshahi, M., and Ladisch, M. R. 1988. Preparation of Cellodextrins. In: Wood, W. A., and Kellog, S. T. (Eds.), Methods in Enzymology. volume 160. Biomass. Part A. Cellulose and Hemicellulose. Acedemic Press, Toronto, pp. 26-38. Perlin, A. S., and Suzuki, S. 1962. The structure of lichenin: selective enzymolysis studies. Can. J. Chem. 40:50-56. Poole, D. M., Morag, E., Lamed, R., Bayer, E. A., Hazlewood, G. P., and Gilbert, H. J. 1992. Identification of the cellulose binding domain of the cellulosome subunit SI from Clostridium thermocellum. FEMS Microbiol. Lett. 99:181-186. Poole, D. M, Hazelwood, G. P., Huskisson, N., Virden, R., and Gilbert, H. J. 1993. The role of conserved tryptophan residues in the interaction of a bacterial cellulose binding domain with its ligand. FEMS Microbiol. Lett. 106:77-84. Poolman, B., Knol, J., and Lolkema, J. S. 1995. Kinetic analysis of lactose and proton coupling in Glu379 mutants of the lactose transport protein of Streptococcus thermophilus. J. Biol. Chem. 270(22): 12995-30003. Quiocho, F. A., and Vyas, N. K. 1984. Novel stereospecificity of the L-arabinose-binding protein. Nature. 310:381-386. Quiocho, F. A. 1989. Protein-carbohydrate interactions: Basic moleuclar features. Pure and Appl. Chem. 61(7):1293-1306. Quiocho, F. A. 1993. Probing the interactions between proteins and carbohydrates. Biochem. Soc. Trans. 21:442-448. Quiocho, F. A., Spurlino, J. C , and Rodseth, L. E. 1997. Extensive features of tight oligosaccharide binding revealed in high-resolution structures of the maltodextrin transport/chemosensory receptor. Structure. 5(6):997-1015. Ramachandran, G. N., and Sasisekharan, V. 1968. Conformation of polypeptides and proteins. Adv. Prot. Chem. 23:283-437.  Reinikainen, T., Ruohonen, L., Nevanen, T., Laaksonen, L., Kraulis, P., Jones, T. A., Knowles, J. K. C., and Teeri, T. T. 1992. Investigation of the function of mutated cellulosebinding domains of Trichoderma reesei cellobiohydrolase I. Proteins. 14:475-482. Reinikainen, T., Teleman, O., and Teeri, T. T. 1995. Effects of pH and High Ionic Strength on the Adsorption and Activity of native and Mutated Cellobiohydrolase I from Trichoderma reesei. Proteins: Struct. Funct. Genet. 22:392-403. Robbens, J., Raeymaekes, A., Steidler, L., Fiers, W., and Remaut, E. 1995. Production of Soluble and Active Recombinant Murine Interleukin-2 in Escherichia coli: High Level expression, Kil-Induced Release, and Purification. Protein Expr. Purif. 6:481-486. Rodgers, K., R., Su, C , Subramaniam, S., and Spiro, T. G. 1992. Hemoglobin R-T Structural Dynamics from Simultaneous Monitoring of Tyrosine and Tryptophan Time-Resolved UV Resonance Raman Signals. J. Am. Chem. Soc. 114:3697-3709. Rozakis-Adcock, M., and Kelly, P. A. 1991. Mutational analysis of the ligand-binding domain of the prolactin receptor. J. Biol. Chem. 266(25):16472-16477. Sakka, K., Takada, G., Karita, S., and Ohmiya, K. 1996. Identification and Characterization of Cellulose-Binding Domains in Xylanase A of Clostridium stercorarium. Ann. N.Y. Acad. Sci. 782:241-251. Sakon, J., Irwin, D., Wilson, D. B., and Karplus, P. A. 1997. Structure and mechanism of endo/exocellulase E4 from Thermomonosporafusca. Nature Struct. Biol. 4(10):810-817. Saraceni-Richards, C. A., and Jacobson, G. R. 1997. A conserved glutamate residue, Glu-257, is important for substrate binding and transport by the Eschericia coli mannitol permease. J. Bacteriol. 179(4):1135-1142. Sarko, A. 1986. Cellulose - How much do we know about its structure? In: Kennedy, J. F., Phillips, G. O., and Williams, P. A. (Eds.). Wood and cellulosics: industrial utilisation, biotechnology, structure and properties. Ellis Horwood, New York, pp. 55-69. Sau, A. C , and Landoll, L. M. 1989. Polymers in Aqueous Media. In: Glass, J. E. (Ed.). Advances in Chemistry Series 223. American Chemical Society, Washington, D. C , pp. 343364. Saya, H., Lee, P. S. Y., Nishi, T., Izawa, I., Nakajima, M., Gallick, G., and Levin, V. A. 1993. Bacterial expression of an active tryrosine kinase from a protein A/truncated c-src fusion protein. FEBS Lett. 327(2):224-230. Schmid, G. 1988. Preparation of Cellodextrins: Another Perspective. In: Wood, W. A., and Kellog, S. T. (Eds.), Methods in Enzymology. volume 160. Biomass. Part A. Cellulose and Hemicellulose. Acedemic Press, Toronto, pp. 38-45. Shen, H., Chretien, P., Bastien, L., and Slilaty, S. N. 1991. Primary Sequence of the Glucanase Gene from Oerskovia xanthineolytica. J. Biol. Chem. 266(2): 1058-1063. Shen, H., Gilkes, N. R., Kilburn, D. G., Miller, R. C. Jr., and Warren, R. A. J. 1995a. Cellobiohydrolase B, a second exo-cellobiohydrolase from the cellulolytic bacterium Cellulomonasfimi.Biochemical J. 311:67-74.  100 Shen, H., Meinke, A., Tomme, P., Damude, H., Kwan, E., Kilburn, D. G., Miller, R. C. Jr., Warren, R. A. J., and Gilkes, N. R. 1995b. Cellulomasfimicellobiohydrolases. In: Saddler, J. N. and Penner, M. (Eds.), Enzymatic Degradation of Insoluble Polysaccharides. American Chemical Society, Washington, DC, pp. 174-196. Shoseyov, O., Takagi, M., Goldstein, M. A., and Doi, R. H. 1992. Primary sequence analysis of Clostridium cellulovorans cellulose binding protein A. Proc. Natl. Acad. Sci. USA. 89:34833487. Siebert, H. C , Adar, R., Arango, R., Burchert, M., Kaltner, H., Kayser, G., Tajkhorshid, E., von der Lieth, C. W., Kaptein, R., Sharon, N., Vliegenthart, J. F., and Gabius, H. J. 1997. Involvement of laser photo-CIDNP (chemically induced dynamic nuclear polarization)-reactive amino acid side chains in ligand binding by galactoside-specific lectins in solution. Eur. J. Biochem. 249(l):27-38. Sigurskjold, B. W., Svensson, B., Williamson, G., and Driguez, H. 1994. Thermodynamics of ligand binding to the starch-bindng domain of glucoamylase from Aspergillus niger. Eur. J. Biochem. 225:133-141. Sphyris, N., Lord, J. M., Wales, R., and Roberts, L. M. 1995. Mutational analysis of the Ricinus lectin B-chains. Galactose-binding ability of the two gamma subdomain of Ricinus communis agglutinin p-chain. J. Biol. Chem. 270(35):20292-20297. Spurlino, J. C , Lu, G-Y., and Quiocho, F. A. 1991. The 2.3 angstrom resolution structure of the maltose- or maltodextrin-binding protein, a primary receptor of bacterial active transport and chemotaxis. J. Biol. Chem. 266:5202-5219. Spurlino, J. C , Rodseth, L. E., and Quiocho, F. A. 1992. Atomic Interactions in ProteinCarbohydrate Complexes - Tryptophan Residues in the Periplasmic Maltodextrin Receptor for Active Transport and Chemotaxis. J. Mol. Biol. 226:15-22. Stalbrand, FL, Saloheimo, A., Vehmaanpera, L, Henrissat, B., and Penttila, M. 1995. Cloning and expression in Saccharomyces cerevisiae of a P-mannanase gene containing a cellulose binding domain. Appl. Env. Microbiol. 61:1090-1097. Sternlicht, H., and Wilson, D. 1967. Magnetic resonance studies of macromolecules. I. Aromatic-methyl interactions and helical structure effects in lysozyme. Biochemistry. 6:28812892. Streitweiser A., Heathcock, C. H., and Kosower, E. M. (Eds.). 1992. Introduction to Organic Chemistry, fourth edition. Maxwell MacMillan, Toronto, pp. 325-369. Takeo, K., and Nakamura, S. 1972. Dissociation constant of glucan phosphorylases of rabbit tissues studied by polyacrylamide gel disc electrophoresis. Arch. Biochem. Biophys. 10:1684. Takeo, K., Kuwahara, A., Nakayama, H., and Nakamura, S. 1975. Affinity electrophoresis of phosphorylases and a-amylases. Protides Biol. Fluids. 23:645. Takeo, K., and Kabat, E. A. 1978. Binding constants of dextrans and isomaltose oligosaccharides to dextran-speciftc myeloma proteins determined by affinity electrophoresis. J. Immunol. 121(6) :2305-2311. Takeuchi, H., Watanabe, N., Satoh, Y., and Harada, I. 1989. Effects of Hydrogen Bonding on the Tyrosine Raman Bands in the 1300-1150 cnr Region. J. Raman Spectrosc. 20:233-237. 1  Tews, I., Vincentelli, R. and Vorgias, C. E. 1996. N-Acetylglucosaminidase (chitobiase) from Serratia marcescens: gene sequence, and protein production and purification in Escherichia coli. Gene. 170:63-67. Tigelar, H. L., and Flygare, W. H. 1972. Molecular Zeeman effect in formamide and the ocproton chemical shift in poly (L-alanine). J. Amer. Chem. Soc. 94:343-344. Tomme P., Van Tilbeurgh, H., Pettersson, G., Van Damme, J., Vandekerckhove, J., Knowles, J., Teeri, T., and Claeyssens, M. 1988. Studies of the cellulolytic system of Trichoderma reesei QM9414. Analysis of domain function in two cellobiohydrolases by limited proteolysis. Eur. J. Biochem. 170:575-581. Tomme, P., Warren, R. A. J., Miller, R. C. Jr., Kilburn, D. G., and Gilkes, N. R. 1995. Cellulose-Binding Domains: Classification and Properties. In: J. M. Saddler and M. Penner (Eds.). Enzymatic Degradation of Insoluble Polysaccharides. American Chemical Society, Washington D.C., pp. 142-161. Tomme, P., Creagh, L. A., Kilburn, D. G., and Haynes, C. A. 1996. Interaction of Polysaccharides with the N-Terminal Cellulose-Binding Domain of Cellulomonas fimi CenC. 1. Binding Specificity and Calorimetric Analysis. Biochemistry. 35:13885-13894. Tomme, P., Boraston, A., Creagh, A. L., Gilkes, N. R., Haynes, C. A., Kormos, J., McLean, B., Sturch, K., Warren, R. A. J., and Kilburn, D. G. 1998. J. Chromatography. In press. Tormo, J., Lamed, R., Chirino, A. J., Morag, E., Bayer, E. A., Shoham, Y., and Steitz, T. A. 1996. Crystal structure of a bacterial family DT cellulose-binding domain: a general mechanism for attachment to cellulose. EMBOJ. 15:5739-5751. Tucker, M . L., Durbin, M. L., Clegg, M. T., and Lewis, L. N. 1987. Avocado cellulase: nucleotide sequence of a putative full-length cDNA clone and evidence for a small gene family. Plant Mol. Biol. 9:197-203. Van Tilbeurgh, H., Tomme, P., Claeyssens, M., Bhikhabhai, R., Pettersson, G. 1986. Limited proteolysis of the cellobiohydrolase I from Tricoderma reesei, separation of functional domains. FEBSLett. 204:223-227. Vincent, P., Shareck, F., Dupont, C , Morosoli, R., and Kluepfel, D. 1997. New oc-Larabinofuranosidase produced by Streptomyces lividans: cloning and DNA sequence of the abfB gene and characterization of the enzyme. Biochem. J. 322:845-852. Vinson, M., van der Merwe, P. A., Kelm, S., May, A., Jones, E. Y., and Crocker, P. R. 1996. Characterization of the sialic acid-binding site in sialoadhesin by site-directed mutagenesis. J. Biol. Chem. 271(16):9267-9272. Vladut-Talor, M., Kauri, T., and Kushner, K. J. 1986. Effects of cellulose on growth, enzyme production, and ultrastructure of a Cellulomonas species. Arch. Microbiol. 144:191-195. Vyas, N. K., Vyas, M. N. and Quiocho, F. A. 1988. Sugar and signal-transducer binding sites ofthe Escherichia coli galactose chemoreceptor protein. Science. 242:1290-1295. Wakarchuk, W. W., Kilburn, D. G., Miller, R. C. Jr., and Warren, R. A. J. 1984. The Preliminary Characterization of the P-Glucosidases of Cellulomonasfimi.J. Gen. Microbiol. 130: 1385-1389.  102 Wells, J. A. 1991. Systematic Mutational Analyses of Protein-pTotein Interfaces. In: Wood, W. A., and Kellog, S. T. (Eds.), Methods in Enzvmology. volume 202. Acedemic Press, Toronto, pp. 390-411. Whittle, D. J., Kilburn, D, G., Warren, R. A. J., and Miller, R. C. Jr. 1982. Molecular cloning of a Cellulomonasfimicellulase gene in Escherichia coli. Gene. 17:139. Wickner, W., Driessen, A. J. M., and Haiti, F-U. 1991. The enzymology of protein translocation across the Eschericia coli plasma membrane. Annu. Rev. Biochem. 1991. 60:101-124. Winterhalter, C , Heinrich, P., Candussio, A., Wich, G., and Liebl, W. 1995. Identification of a novel cellulose-binding domain within the multidomain 120kDa xylanase XynA of the hyperthermophilic bacterium Thermotoga maritima. Mol. Microbiol. 15(3) :431-444. Winzor, D. J., Sawyer, W. H. (Eds.). 1995. Quantitative characterization of ligand binding. Wiley-Liss, Toronto, pp. 8. Wishart, D. S., Sykes, B. D., and Richards, R. M. 1991. Relationship between Nuclear Magnetic Resonance Chemical Shift and Protein Secondary Structure. J. Mol. Biol. 222:311333. Wong, W. K., Gerhard, B., Guo, Z. M., Kilburn, D. G., Warren, R. A. J., and Miller, R. C. Jr. 1986. Characterization and structure of an endoglucanse gene CenA of Cellulomonas fimi. Gene. 44:315-324. Wood, T. M. 1988. Preparation of crystalline, amorphous and dyed cellulase substrates. In: Wood, W. A., and Kellog, S. T. (Eds.), Methods in Enzymology. volume 160. Biomass. Part A. Cellulose and Hemicellulose. Acedemic Press, Toronto, pp. 19-26. Woodward, J., Lima, M., and Lee, N. E. 1988. The role of cellulase concentration in determining the degree of synergism in the hydrolysis of microcrystalline cellulose. Biochem. J. 255:895-899. Wright, C. S. 1990. 2.2 angstom resolution structure analysis of two refined Nacetylneuraminyl-lactose-wheat germ agglutinin isolectin II. J. Mol. Biol. 215:635-651. Wuthrich, K. 1989. Protein Structure Determination in Solution by Nuclear Magnetic Resonance Spectroscopy. Science. 243:45-49. Wuthrich, K. 1990. Protein Structure Determination in Solution by NMR Spectroscopy. J. Biol. Chem. 265(36):22059-22062. Xu, G. Y., Ong, E., Gilkes, N. R., Kilburn, D. G., Muhandiram, D. R., Harris-Brandts, M., Carver, J., Kay, L., and Harvey, T. 1995. Solution structure of a cellulose-binding domain from Cellulomonasfimiby nuclear magnetic resonance spectroscopy. Biochemistry. 34:69937009. Yamamoto, M., and Ferretti, J. J. 1995. High level expression of Streptococcus pyogenes erythrogenic toxin A (SPE A) in Eschericia coli and its rapid purification by HPLC. FEMS Microbiol. Lett. 132:209-213. Yanisch-Perron, C , Vieira, J., and Messing, J. 1985. Improved M13 phage cloning vectors and host strains: Nucleotide sequences of the M13mpl8 and pUC19 vectors. Gene. 33:103-119.  Zhu, K., Bressan, R. A., Hasegawa, P. M., and Murdock, L. L. 1996. Identification of Nacetylglucaosamine binding residues in Griffonia simplicifolia lectin U. FEBS Lett 390(3):271274. Zubay, G. 1983. Biochemistry - Volume one: Energy. Cells and Catalysis, third edition. Wm. C. Brown Publishers, Dubuque.  


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