Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The linker of endoglucanase A from cellulomonas fimi: an investigation of structure, function and application Sandercock, Linda Eileen 1999

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

Item Metadata

Download

Media
831-ubc_1999-389731.pdf [ 7.17MB ]
Metadata
JSON: 831-1.0089277.json
JSON-LD: 831-1.0089277-ld.json
RDF/XML (Pretty): 831-1.0089277-rdf.xml
RDF/JSON: 831-1.0089277-rdf.json
Turtle: 831-1.0089277-turtle.txt
N-Triples: 831-1.0089277-rdf-ntriples.txt
Original Record: 831-1.0089277-source.json
Full Text
831-1.0089277-fulltext.txt
Citation
831-1.0089277.ris

Full Text

The Linker of Endoglucanase A from Cellulomonas fimi: an investigation of structure, function and application by Linda Eileen Sandercock B.Sc. in Cell Biotechnology, University of Alberta, 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department of M ic rob io logy and Immunology / B iotechnology Laboratory) W e accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A • M a r c h 1999 © L i n d a Ei leen Sandercock, 1999 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r a n a d v a n c e d d e g r e e a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e L i b r a r y s h a l l m a k e it f r e e l y a v a i l a b l e f o r r e f e r e n c e a n d s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s m a y b e g r a n t e d b y t h e h e a d o f m y d e p a r t m e n t o r b y h i s o r h e r r e p r e s e n t a t i v e s . It is u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n o t b e a l l o w e d w i t h o u t m y w r i t t e n p e r m i s s i o n . D e p a r t m e n t o f MtCgQg>QLfl6y £ ' \IAMUAJOI OfiX,/ T h e U n i v e r s i t y o f B r i t i s h C o l u m b i a V a n c o u v e r , C a n a d a D E - 6 ( 2 / 8 8 ) 11 Abstract Endoglucanase A ( C e n A ) , a component o f the mult i -enzyme cellulase system o f the bacterium Cellulomonas fimi (C. fimi), is typical o f many cellulases, being a mult i -domain protein. C e n A has two domains which have been wel l characterized, a catalytic domain and a cel lulose-binding domain ( C B D ) separated by a proline, threonine rich l inker ( P T linker). T h i s study investigated the structure, function and possible applications o f this interdomain linker. Mod i f i ca t ion o f the P T l inker's size and composit ion gave insights into the effect o f the linker on hydrolysis o f a variety o f cel lulosic substrates and the nature o f its susceptibility to proteolysis. M o d i f y i n g the linker size and composi t ion had little effect on the hydrolysis rate. Partial deletion reduced activity on crystalline and amorphous celluloses, whereas doubl ing or replacing it with the fibronectin type III (Fn3) repeats f rom the endoglucanase B increased the rate o f hydrolysis o f crystalline cellulose. Al ter ing the linker d id not affect activity on 2 ' ,4 ' -dini trophenyl (3-D-cellobioside ( 2 , 4 - D N P C ) except for the Fn3 construct wh ich showed decreased activity. A sequence susceptible to papain and C. fimi protease was defined wh ich could be modi f ied to prevent proteolysis. Th is has potential implications for the design o f linkers for stable C B D - f u s i o n proteins. E x a m i n i n g the glycosylat ion o f C e n A produced by C. fimi and Streptomyces lividans (S. lividans) gave insights into eubacterial protein glycosylat ion. G lycosy la t ion was conf ined to the P T linker at sites conforming to an O-g lycosyla t ion consensus sequence for some mammal ian glycoproteins. T h e oligosaccharide composi t ion was dependent on the host, but the glycosylat ion mechanism is l ikely similar as they discriminated between the same protein features. Ol igosaccharide size and linkage information for biose units o f S. lividans produced proteins were determined, the latter by a novel application o f fluorophore-assisted carbohydrate electrophoresis. G lycosy la t ion slowed proteolysis and slightly reduced the activity on 2 , 4 - D N P C . The C e n A linker resembles the hinge o f human immunoglobul in A l ( IgA l ) . A hybr id protein, C e n A I g A l h - where the P T l inker was replaced b y the I g A l hinge- and C e n A were useful substrates for probing the specificity o f a number o f I g A l proteases produced by a variety o f pathogenic bacteria. The most striking f inding was that only one o f two I g A l proteases wh ich have h igh sequence identity and cleave the identical bond in I g A l , cleaved C e n A . These substrates w i l l be useful in future studies o f I g A l protease specificity. IV Table of contents Abstract ii Table of contents iv List of tables xi List of figures xiii Abbreviations xvii Statement of Contribution in IgAl Protease Collaboration xxi Acknowledgements xxii 1. Introduction 1 1.1. Cel lu lose 1 1.2. Cel lulases 2 1.2.1. General 2 1.2.2. Cel lulase systems 2 1.2.3. L inkers 5 1.2.4. Proteolysis 10 1.3. Glycoproteins 11 1.3.1. General 11 1.3.2. Protein-carbohydrate linkages 11 1.3.3. Sugars present in oligosaccharide 12 1.3.4. Sizes o f oligosaccharides 13 1.3.5. Determinants o f protein glycosylat ion 13 1.3.6. Prokaryotic glycoproteins 15 1.3.7. Glycosyla ted polysaccharidases 16 V 1.3.8. G lycosy la t ion o f C e n A and C e x 20 1.3.9. Funct ion o f glycosylat ion, with an emphasis on cellulases 20 1.3.10. Technologica l advances in carbohydrate analysis 21 1.4. I g A l and I g A l proteases 22 1.4.1. T h e structure and function o f I g A 22 1.4.2. I g A l proteases 26 1.5. Objectives 33 2. Methods and Materials 34 2.1. Chemica ls , buffers, substrates and enzymes 34 2.2. Bacterial strains and plasmids 35 2.3. M e d i a and growth conditions 37 2.4. Recombinant D N A techniques 38 2.4.1. Pr imer design for introducing silent mutations 39 2.4.2. Synthetic oligonucleotides and oligonucleotide primers 39 2.4.3. D N A sequencing 39 2.4.4. P C R 41 2.4.5. Construct ion o f mutant P T linkers 44 2.5. Screening for mutants and confirmation o f identity 46 2.5.1. E.coli 46 2.5.2. S. lividans 46 2.6. Gene expression and protein purif ication 47 2.6.1. E.coli 47 2.6.2. C.fimi 48 vi 2.6.3. S. lividans 48 2.7. Protein detection and quantification and storage 50 2.8. Western blotting and N-terminal sequencing o f proteins and peptides 50 2.9. M a s s analysis o f proteins 51 2.9.1. Predicted molecular weights 51 2.9.2. Relat ive mass 51 2.9.3. Electrospray mass spectrometry 51 2.9.4. Mat r ix -Ass is ted Laser-Desorption/Ionization T i m e - O f - F l i g h t ( M A L D I - T O F ) mass spectrometry 52 2.10. Stability 53 2.11. E n z y m e assays 53 2.11.1. P N P C hydrolysis 53 2.11.2. B M C C hydrolysis 53 2.11.3. P A S C hydrolysis 54 2.11.4. C M C hydrolysis 54 2.11.5. 2 , 4 - D N P C hydrolysis 54 2.11.6. p H profiles 55 2.11.7. Hydro lys is products 55 2.11.8. Statistical analysis o f hydrolysis data 55 2.12. Proteolysis 56 2.12.1. Papain hydrolysis 56 2.12.2. C. fimi protease hydrolysis 56 2.13. I g A l protease digestion 57 V l l 2.14. G lycosy la t ion determination 57 2.14.1. Periodic ac id /Schi f f reagent and C o n A - H R P treatment o f Western blots 57 2.14.2. Phenol sulfuric acid assay 58 2.14.3. Determination o f the monosaccharide composi t ion and l inkage positions between sugars for C e n A and C e x produced b y C. fimi 58 2.14.4. Determination o f the monosaccharide composi t ion o f C e n A mutants produced by S. lividans and o f C e n A produced by C.fimi using F A C E ® 60 2.14.5. Determination o f the oligosaccharide composi t ion and linkages between sugars for C e n A 2 P T produced by S. lividans using F A C E ® 60 2.14.6. Locat ion o f glycans on C e n A and C e x produced by C.fimi 61 3. Results 63 3.1. Generation o f cenA mutants 63 3.1.1. Def in i t ion o f the boundaries o f the P T linker o f C e n A 63 3.1.2. Pr imer design for introducing silent mutations 64 3.1.3. Generation o f p U C 1 8 - 1 . 5 c e n A A P X A S 64 3.1.4. Construct ion o f the c e n A mutants 65 3.1.5. Subcloning c e n A P T mutants into E. coli and S. lividans expression vectors .66 3.2. Gene expression and protein purif ication 72 3.2.1. E.coli : 72 3.2.2. S. lividans 74 3.2.3. C.fimi 75 3.3. Protein quantification 75 V l l l 3.4. Mo lecu la r masses o f proteins 77 3.5. G lycosy la t ion determination 79 3.5.1. Detection o f glycosylat ion on proteins produced in S. lividans and E. coli 79 3.5.2. M o l e s o f sugar per mole o f protein for C e n A and C e x produced in C. fimi and C e n A I g A l h produced in S. lividans 81 3.5.3. Locat ion o f carbohydrates on C e n A and C e x produced in C. fimi and C e n A constructs produced in S. lividans 83 3.5.4. T h e monosaccharide composi t ion and l inkage posit ions between sugars o f C e n A and C e x produced b y C. fimi 89 3.5.5. T h e monosaccharide composit ion o f the glycans o f C e n A mutants produced by S. lividans and C e n A produced b y C. fimi 94 3.5.6. Ol igosaccharide composit ion and linkage positions between sugars o f C e n A 2 P T produced in S. lividans 95 3.6. E n z y m e activity 98 3.6.1. Protein stability 98 3.6.2. Ef fect o f method o n B M C C hydrolysis b y C e n A 98 3.6.3. B M C C hydrolysis 99 3.6.4. P A S C hydrolysis 102 3.6.5. 2 , 4 - D N P C hydrolysis 104 3.6.6. p H profi les 104 3.6.7. E n z y m e assay products 107 3.7. Proteolysis 107 3.7.1. Papain hydrolysis o f C e n A constructs produced in E. coli 107 ix 3.7.2. C.fimi protease hydrolysis o f C e n A constructs produced in E. coli 111 3.7.3. Effect on papain hydrolysis o f changing valine 176 to prol ine 114 3.7.4. Ef fect o f glycosylat ion o f C e n A constructs on susceptibility to papain and C. fimi protease 117 3.8. I g A l protease digestion o f C e n A and C e n A I g A l h 121 3.9. A compar ison o f the glycosylat ion o f the native I g A l hinge and that produced by S. lividans 125 4. Discussion 127 4.1. T h e effect o f l inker length and composit ion on activity 127 4.1.1. T h e effect o f l inker length and composit ion on activity on P A S C 127 4.1.2. Dif f icult ies in comparing cellulase activity data between studies 127 4.1.3. T h e effect o f l inker length and composit ion on activity on B M C C 128 4.1.4. Measur ing hydrolysis o f an insoluble substrate 130 4.1.5. T h e effect o f linker length and composit ion on activity on 2 , 4 - D N P C 130 4.1.6. Summary o f activity studies 131 4.2. Effect o f l inker length and composit ion on sensitivity to proteases 131 4.3. G lycosy la t ion 133 4.3.1. Sites o f glycosylat ion 133 4.3.2. U s e o f mass spectrometry for quantification o f g lycosylat ion 136 4.3.3. Monosacchar ide composit ion, size o f oligosaccharides and l inkage positions between sugars o f oligosaccharides 137 4.3.4. Ro le o f glycosylat ion on C e n A and C e n A derivatives 139 4.3.5. Summary o f glycosylat ion studies 142 4.4. Hydrolysis of CenA and CenAIgAlh by IgAl proteases 143 4.5. Final Summary 145 Bibliography 147 xi List of tables Table 1.1: Representative l inker sequences 7 Table 1.2: A selection o f glycosylated polysaccharidases f rom bacterial and some fungal sources 17 Table 1.3: I g A l protease producing organisms implicated in disease 27 Table 2. 1: Bacterial strains 35 Table 2. 2: E. coli p lasmids 36 Table 2. 3: S. lividans/E. coli shuttle plasmids 37 Tab le 2. 4: Ol igonucleot ide primers used for P C R 40 Table 2. 5: Ol igonucleot ide primers used for sequencing 40 Table 2. 6: Synthetic oligonucleotides for constructing mutant P T linkers 41 Table 2. 7: H P L C program 62 Table 3 . 1 : Protein concentrations o f C e n A constructs produced in E. coli determined by a variety o f methods 76 Table 3 .2 : Mo lecu lar masses o f C e n A and C e n A mutants produced in E. coli and S. lividans determined by three different techniques 78 Tab le 3 . 3 : Summary o f agitation methods studied for the hydrolysis o f B M C C 99 Table 3 .4 : Summary o f the hydrolysis o f B M C C b y C e n A and C e n A constructs produced in E. coli and S. lividans 101 Table 3 . 5 : Summary o f the hydrolysis o f P A S C by C e n A and C e n A constructs produced in E. coli and S. lividans 103 X l l Table 3. 6: Susceptibility of CenA substrates to IgAl proteases 124 Xll l List of figures Figure 1.1: M o d u l a r organization o f cellulases f rom C. fimi 4 Figure 1.2: Computer generated model o f the C e n A P T linker between k n o w n structures o f a C.fimi cellulase catalytic domain and C B D 9 Figure 1.3: T h e structure o f an Fn3 module f rom fibronectin 9 Figure 1.4: Papain and C. fimi protease cleavage sites in C e n A produced in E. coli 11 Figure 1.5: Mot i fs proposed for O-glycosylat ion based on in vivo g lycosylat ion 14 F igure 1. 6: G lycosy la t ion o f the hinge region o f human m y e l o m a serum I g A l and human pooled serum I g A l 25 F igure 1.7: C leavage sites in the hinge region o f m y e l o m a serum I g A l o f some I g A l proteases 28 Figure 1. 8: Products o f cleavage o f I g A l b y I g A l proteases 30 Figure 1. 9: T h e sequence o f the I g A l hinge compared to the P T linker o f C e n A 32 Figure 2.1: Generating p U C 1 8 - 1 . 5 c e « A A P T A S 43 Figure 2.2: Construct ion o f c e « A 1 . 5 P T and c e « A 2 P T using synthetic ol igonucleotides.. .45 Figure 3 . 1 : T h e P T linker o f C e n A as defined in various studies 63 Figure 3. 2: Des ign o f silent mutation sites 64 Figure 3. 3: Construct ion o f p U C 1 8 - 1 . 5 c e « A mutants 68 Figure 3 .4 : Part 1 o f subcloning strategy to insert cenA mutants into expression vectors p T U G a n d p I J 6 8 0 69 xiv Figure 3 .5 : Part 2 o f subcloning strategy to insert cenA mutants into expression vectors p T U G and pIJ680 70 Figure 3 .6 : D N A agarose gel showing the Nhel/HindUI fragment from each p S L l 1 8 0 c e « A mutant used to subclone into p T U g and pIJ680 71 Figure 3 .7 : Congo- red stained agar plate showing halos where S. lividans colonies expressing C e n A A P X were growing 71 Figure 3 .8 : S D S - P A G E gel o f C e n A constructs produced in E. coli 72 Figure 3 . 9 : T h e domain structure and linker sequences o f the C e n A and C e n A P T linker constructs 73 Figure 3. 10: S D S - P A G E gel o f C e n A constructs produced in S. lividans 74 Figure 3 .11 : S D S - P A G E gel o f C e n A and C e x produced in C. fimi 75 Figure 3. 12: S D S - P A G E gel and C o n A - H R P treated blot o f C e n A constructs produced in S. lividans 80 Figure 3 .13 : S D S - P A G E gel and C o n A - H R P treated blot o f C e n A constructs produced in S. lividans and E. coli 80 Figure 3. 14: S D S - P A G E gel o f C e n A constructs produced in E. coli and S. lividans 81 Figure 3 .15 : M A L D I - T O F mass spectra o f C e n A and C e x produced b y C.fimi 82 F igure 3 .16: Reverse phase H P L C profiles o f a tryptic digest o f C e n A produced b y C. fimi 85 Figure 3 .17 : Reverse phase H P L C profiles o f a tryptic digest o f C e x produced by C.fimi 86 XV Figure 3 .18: S D S - P A G E gel and periodic ac id /Schi f f stained Western blot o f C e n A and C e x produced b y C. fimi, and treated with papain 88 Figure 3 .19 : G lycosy la t ion sites on the l inker o f C e n A and C e x produced b y C.fimi 88 Figure 3. 20: Determination o f carbohydrate composit ion b y G L C o f C e n A and C e x produced by C.fimi 91 F igure 3. 21: S D S - P A G E gel and periodic ac id /Schi f f stained Western blot o f C e n A and C e x produced by C.fimi, and treated with exo-glycosidases 92 Figure 3. 22: Western blot treated with C o n A - F £ R P to detect the presence o f mannose on C e x and C e n A produced by C.fimi, and treated with exo-glycosidases 94 Figure 3. 23: F A C E ® Monosacchar ide composi t ion gel o f C e n A produced b y C. fimi and a subset o f C e n A mutants produced by S. lividans 96 Figure 3. 24: F A C E ® Oligosaccharide analysis gel o f C e n A 2 P T produced by S. lividans. 97 Figure 3. 25: 2 , 4 - D N P C hydrolysis at p H 6.5 by C e n A linker constructs produced in E. coli and S. lividans 105 Figure 3. 26: Composi te figure o f 2 , 4 - D N P C hydrolysis at p H 6.5 by C e n A linker constructs produced in E. coli and S. lividans 106 Figure 3. 27: p H profiles o f 2 , 4 - D N P C hydrolysis by C e n A and C e n A constructs produced in E. coli and S. lividans 106 Figure 3. 28: Papain digestion o f C e n A and C e n A P T constructs produced in E. coli 108 Figure 3. 29: Expanded gel o f a papain digestion o f C e n A produced by E. coli 109 Figure 3. 30: Papain cleavage sites in C e n A and C e n A P T mutants made in E. coli 110 Figure 3 . 3 1 : C . fimi protease digestion o f C e n A and C e n A P T constructs made b y E. coli 112 xv i Figure 3. 32: Papain and C.fimi protease cleavage sites in C e n A and C e n A P T mutants made in is. coli 113 Figure 3. 33: C.fimi protease digestion o f I g A l 114 Figure 3. 34: Papain digestion o f C e n A 1 . 5 P T and C e n A 1 . 5 P T m o d made b y E. coli 115 Figure 3. 35: Papain and C.fimi protease cleavage sites in C e n A 1 . 5 P T and C e n A l . 5 P T m o d produced in E. coli 116 Figure 3. 36: Papain digestion o f C e n A and C e n A P T mutants made by E. coli and S. lividans 118 Figure 3. 37: C. fimi protease digestion o f C e n A and C e n A P T constructs made b y E. coli and S. lividans 119 Figure 3. 38: Papain and C. fimi protease cleavage sites in C e n A and C e n A P T mutants made in E. coli and S. lividans 120 Figure 3. 39: I g A l protease digests o f C e n A I g A l h and C e n A produced by E. coli and S. lividans 122 Figure 3. 40: G lycosy la t ion o f the hinge region o f human m y e l o m a serum I g A l and C e n A I g A l h produced in S. lividans 126 Abbreviations 2,4-DNPC 2',4'-dinitrophenyl p-D-cellobioside x g Centrifugal force relative to the gravitational force A n Absorbance at wavelength "n" AA Amino acid Amp Ampicillin Amp" Ampicillin resistance ANTS 8-aminonaphthalene-l,3,6-trisulfonic acid Ara Arabinose BMCC Bacterial micro-crystalline cellulose bp Base pairs BSA Bovine serum albumin C- Carboxyl terminus CBD Cellulose-binding domain CSD Cleavage-specificity domain CMC Carboxymethylcellulose COH Carbohydrate ConA Concanavilin A Da Dalton DMSO Dimethylsulfoxide DNA Deoxyribonucleic acid dNTP Deoxyribonucleic acid triphosphate xviii DTT Dithiothreitol EDTA Ethylenediaminetetraacetic acid FACE Fluorophore-assisted carbohydrate electrophoresis Fn3, FNIII Fibronectin type III sequence FPLC Fast protein liquid chromatography Fuc Fucose Gal Galactose GalNAc N-acetylgalactosamine GLC Gas liquid chromatography Glc Glucose GlcNAc N-acetylglucosamine GRASP Centre for gastroenterology research on absorptive and secretory processes HBAH p-Hydroxybenzoic acid hydrazide HPLC High pressure liquid chromatography HRP Horseradish peroxidase IgA Immunoglobulin A IgAl Immunoglobulin A l I Inositol IPTG Isopropyl-P-D-thiogalactoside Kan Kanamycin KanR Kanamycin resistance kDa k i loDal ton Michae l is constant LB L u r i a broth MALDI-TOF Matrix-assisted laser-desorption/ionization t ime-of-f l ight MALDI-TOF MS Matrix-assisted laser-desorption/ionization t ime-of-f l ight spectrometry Man Mannose M r Realt ive molecular mass MS M a s s spectrometry MW Mo lecu lar weight N- A m i n o terminus NA, na No t applicable NAPS U B C nucleic acid and protein synthesis unit NeuAc N-acetylneueramic acid, also called sialic acid No., no. N u m b e r OD Opt ical denisty PASC Phosphoric acid swollen cellulose PCR Polymerase chain reaction PMSF Phenylmethylsulfonylf luoride PNPC Para-nitrophenyl -P -D-cel lobioside PT linker Prol ine, threonine rich l inker PVDF Polyvinyl idene dif luoride rpm Revolut ions per minute RT R o o m temperature Periodic acid/Schiff Per iodic ac id /Schi f f reagent treatment stain S-layer Crystal l ine surface-layer SC Polymer ic immunoglobul in receptor SD Standard deviation SDS Sod ium dodecyl sulfate SDS-PAGE Sod ium dodecyl sulfate polyacrylamide gel electrophoresis SELDI Surface-enhanced laser desorption ionization Sinapinic acid 3,5-dimethoxy-4-hydroxycinnamic acid TPCK N- tosyl -L-phenyla lanylchloromethyl ketone TSB Trypt ic S o y Broth Tsr Thiostrepton TsrR Thiostrepton resistance U Uni ts UV Ultraviolet wavelength WT Wi ld - type xxi Statement of Contribution in IgAl Protease Collaboration T h e use o f glycosylated and non-glycosylated C e n A and a hybr id protein C e n A I g A l h containing the I g A l hinge region, as substrates for probing the specificity o f I g A l proteases was investigated in collaboration with the laboratory o f Dr . A n d r e w Plaut o f G R A S P at Tuf ts -New Eng land M e d i c a l Centre Hospi ta l , Boston . D r . A n d r e w Plaut and Jiazhou Qiu, a scientist in the laboratory o f Dr . Plaut, have agreed that the results o f our collaboration may be reported in m y thesis and therefore I report here the nature o f the contributions made by each party. Dr . Plaut was approached b y Dr . Warren about entering into this collaboration. Dr . Plaut and Jiazhou Qiu made the decisions to test the variety o f I g A proteases that were tested and had the expertise to carry out the protease tests and guide the interpretation. T h e substrates were produced by L i n d a Sandercock who also analyzed the glycosylat ion o f the substrates and the susceptibility o f I g A l to C.fimi protease. Regular consultations were held between the collaborators. XXll Acknowledgements I wou ld like to acknowledge and thank m y supervisors, Drs . D o u g K i l b u r n and T o n y Warren, and their families for their good-natured and multi-faceted support, guidance and encouragement. Sincere thanks to m y committee members, Drs . Jul ian Dav ies , Gera ld Weeks and Steve Withers, for taking time to provide me with their unique perspectives on m y project and getting me to stop! M a n y thanks to Dr . A n d r e w Plaut and Jiazhou Qiu for taking time to teach me and let me collaborate with them in their wor ld o f I g A l proteases. T h e Cel lu lase L a b has provided m e with a dynamic , very supportive and friendly environment to pursue more than s imply science over the years. My.grat i tude goes out to all members past and present who have made m y time in Vancouver r ich indeed: especially P T , himself , for innumerable things including permission to invade his bench and ask just one more question; G r e g for stimulating late night discussions and safe-walk service; A l B , m y partner in glycoprotein forays; D o m i n i k , whose h igh spirits helped keep m y own up; Brad , m y computer saviour; He len , our organizer in the lab and o n the h i l l ; E m i l y and E m i l y who patiently taught me techniques and lent their ears many, many times; Edgar who introduced me to g lycobio logy and continues to inspire me; H u a and Pat whose work provided a base for m y o w n and who fueled m e even from a distance; and Henr ik , Laurent and N e i l for their help. F o r technical assistance, I an indebted to many including Suzanne for all her protein sequencing; the Withers lab, especially D a v e for the use o f his bench etc. and Karen; and Dr . Eigendorf , L i n a and Marshal l in the mass spectrometry facility. F o r keeping us graduate students organized, I c o m m e n d the hard work o f the microbio logy graduate secretaries past and present and the department off ice staff. A special thanks to the microbio logy dessert w o m e n past and present for great food, wonderful friendships and chat! F ina l ly , I feel very fortunate and am extremely grateful for the support (from far and near!) o f m y husband, Mark , and our families for the duration o f this project. 1 1. Introduction 1.1. Cellulose Cel lu lose, a linear polymer o f glucose residues jo ined P-l ,4, is abundant on earth, being the major component o f plant cell walls. Despite the simplici ty o f the l inkage, cellulose may be complex and heterogeneous. Parameters such as size, shape, porosity, surface area, association with non-cel lulosic components, molecular conformation and crystallinity m a y differ ( T o m m e et al, 1995). A s a result there are many different forms o f cellulose. Cel lu loses used in research include ol igomers up to six glucose units w h i c h are soluble (Warren, 1996), and longer polymers made soluble by substitutions such as in carboxymethylcel lulose ( C M C ) . Insoluble celluloses include the crystall ine cellulose, bacterial micro-crystal l ine cellulose ( B M C C ) made by Acetobacter, and a more amorphous, or disordered cellulose, A v i c e l ™ . Phosphoric acid treatment can render this latter cellulose even more amorphous, producing phosphoric acid swol len cellulose ( P A S C ) . P A S C is considered to be the most homogeneous form o f amorphous cellulose (Schulein, 1998). A l l the insoluble cellulosic substrates are to some degree heterogeneous however. Th is heterogeneity results f rom imperfections in pack ing or mechanical damage, wh ich cause disorder in the cellulose structure. F o r a more detailed overview o f cellulose structures see T o m m e et al., 1995 ( T o m m e et al., 1995). 2 1.2. Cellulases 1.2.1. General Cellulases are produced by a number o f different organisms including bacteria, fungi (Teeri et al, 1992; T o m m e et al, 1995; Warren, 1996), plants and most recently nematodes and termites (D ing et al, 1998; Smant et al., 1998; Watanabe et al, 1998). Cellulases can be broadly d iv ided into two classes: exoglucanases or cel lobiohydrolases, wh ich cleave cel lobiose units f rom the termini o f cellulose molecules, and endoglucanases which cleave internally. T h e former generally have the highest apparent specif ic activity on crystalline cellulose (Teeri et al., 1998). M o s t cellulases are modular: composed o f a catalytic domain associated with other domain(s) such as cel lulose-binding domains ( C B D ) , often v ia a linker sequence ( T o m m e et al., 1995). Examples o f specif ic systems are given below. 1.2.2. Cellulose systems T o degrade such a complex and heterogeneous substrate as cel lulose, organisms have evolved systems consisting o f multiple cellulases, varying in substrate specificity and hydrolyt ic activity, w h i c h work in concert. W e l l studied bacterial cellulase systems include the non-complexed systems o f Cellulomonas fimi (C.fimi) and Thermomonospora fusca (T.fusca), and the complexed and cell-associated systems o f Clostridium thermocellum (C. thermocellum) and Clostridium cellulolyticum (C. cellulolyticum). T h e best characterized fungal system is that o f Trichoderma reesei (T. reesei). 3 C.fimi is a Gram-posi t ive , mesophi l ic , coryneform bacterium (Warren, 1996). T h i s aerobe has a mult icomponent cellulase system consisting o f secreted endoglucanases, an exoglucanase/xylanase and cellobiohydrolases (Gi lkes et al, 1991a). F o u r endoglucanases C e n A - D , one exoglucanase/xylanase C e x , and two cel lobiohydrolases have been reported (Figure 1.1) (Gi lkes et al, 1991b; Shen et al, 1995; Warren, 1996). C e n A is an endo P- l ,4-glucanase o f about 48 k D a and C e x is a p- l ,4 -xy lanase/ exo -P - l ,4-glucanase o f about 46-49 k D a (Gi lkes et al., 1989). T h e genes encoding these cellulases have been c loned and expressed in Escherichia, coli (E. coli) and the recombinant proteins characterized in some detail (Damude, 1995; D i n et al., 1994; D i n et al., 1991; M a c L e o d et al., 1994; White et al., 1994). B o t h C e n A and C e x comprise two domains: a catalytic domain and a cel lulose-binding domain ( C B D ) separated by a proline, threonine r ich l inker (PT linker), but the orientation o f their domains is reversed. C e n B is an endo P- l ,4-glucanase o f about 110 k D a which has six domains: a catalytic domain, a C B D , three fibronectin type III domains (Fn3) and a second C B D , separated by linkers r ich in prol ine and hydroxyamino acids (Meinke et al., 1991a; M e i n k e et al., 1991b). C e n B is approximately fifty times more active than C e n A on B M C C , with comparable activities on A v i c e l , and C M C , but C e n A has four-fold higher on P A S C ( T o m m e et al, 1996). T. fusca is a thermophil ic actinomycete w h i c h produces at least six cellulases, including four endoglucanases and two cellobiohydrolases (Warren, 1996; W i l s o n , 1992; Zhang et al, 1995). C . thermocellum is a Gram-posi t ive, anaerobic, thermophil ic bacterium, wh ich degrades crystalline cellulose with high eff iciency. It has a complexed, cell-associated cellulase system characterized by a mult ienzyme, ce l l -bound complex 4 CenA CenB CenC CenD CbhA CbhB 48 TTHT • • • • Cex 1 0 C B D 11 C A T A L Y T I C D O M A I N L I N K E R UHI Fn3 D O M A I N O T H E R D O M A I N Figure 1 . 1 : Modular organization of cellulases from C.fimi. The numbers in the boxes indicate the catalytic domain family. (Cellulase catalytic domains have been divided into families based on sequence similarities.) termed a cellulosome. At least twenty-one endoglucanases, three cellobiohydrolases, and a scaffolding protein have been described to date (Beguin and Alzari, 1998; Beguin and Lemaire, 1996; Bronnenmeier and Staudenbauer, 1993). C. cellulolyticum, like C. thermocellum, is a Gram-positive, anaerobic bacterium which produces a cellulosome. A scaffolding protein, several endoglucanases and a potential cellobiohydrolase have been reported. The domains are not separated by defined linkers (Bronnenmeier and Staudenbauer, 1993; Gal, 1997; Warren, 1996). T. reesei, a mesophilic fungus, has a non-complexed cellulose hydrolyzing system similar to that of C. fimi with at least three 5 endoglucanases and two cellobiohydrolases (Kubicek, 1992; Schi i le in, 1998; Warren, 1996). 1.2.3. Linkers Linkers r ich in hydroxyamino acids and often proline separate the modules o f many cellulases and hemicellulases. Th is type o f l inker is not restricted to these families o f enzymes, but can also be found in a diverse range o f other proteins. S o m e examples are shown in Table 1.1. A more extensive table was publ ished in G i lkes et al, 199Id (Gi lkes et al., 199 Id). F r o m the table it is evident that the length o f l inkers can vary greatly -linkers o f up to three hundred amino acids have been reported (Black et al., 1996) - and that only in rare cases such as for C e n A and C e x , is there h igh sequence identity between linkers. T h e linkers can be loosely div ided into three classes: those with h igh (30% or greater) prol ine content; those with less proline content; and a class o f interdomain linkers called Q- l inkers, wh ich are r ich in arginine, glutamine, glutamate, and hydrophobic amino acids (Gi lkes et al., 1991d). Prol ine and threonine/serine r ich l inker sequences can form extended helices (Argos, 1990; Wi l l i amson , 1994). A computer-derived mode l o f the P T linker o f C e n A is shown in Figure 1.2. M a n y fungal cellulases have an O-glycosylated linker region, but the structure and flexibil ity o f the linker region is not k n o w n , although even the less prol ine-r ich O-glycosylated linker sequences o f C B H I and C B H I I f rom T. reesei are assumed to adopt an extended conformation (Schii lein, 1998; Teeri et al, 1998). Some bacterial cellulases also have O-glycosylated linkers, such as E 2 and E 3 f rom T. fusca (Ca lza et al., 1985; Zhang et al., 1995). G lycosy la t ion o f proteins is dealt with later in the Introduction (Section 1.3). 6 Another potential spacer module is the Fn3 module. These are found in a number o f different bacterial cellulases, often in tandem (Shen et al, 1995). T h e structure o f an Fn3 module o f fibronectin has been solved (Figure 1.3) as has a f ibronectin segment with four Fn3 modules in tandem has also been determined (Leahy et al., 1996; M a i n et al, 1992). In contrast to the P T linker, the Fn3 module has a defined tertiary structure consisting o f seven antiparallel P-strands forming two facing P-sheets; yet, when in tandem the Fn3 modules are reported to have non-specif ic interdomain interactions which provide some flexibil i ty and elasticity (Er ickson, 1994; Leahy et al., 1996). Despite the widespread presence o f linkers in cellulases and hemicel lulases, a single role for all linkers has not been established (Hazelwood and Gilbert , 1998). Investigation into the role o f these linkers has mainly been done by deletion. Linkerless mutants o f C e n A f rom C.fimi (Shen et al., 1991), xylanase X Y L A and arabinofuranosidase X Y L C from Pseudomonas fluorescens subsp. cellulosa (Black et al., 1997; B l a c k et al, 1996), E 2 f rom T.fusca (Wi lson et al, 1995) and C B H I from T. reesei (Srisodsuk et al, 1993) all had lower activity on insoluble substrates, suggesting a role for the linker in the efficient hydrolysis o f insoluble substrates. O n l y the C e n A mutant exhibited decreased activity against a soluble substrate. It was also more diff icult to desorb from A v i c e l ™ . These changes, for C e n A , were attributed to a change in the relative orientation o f the catalytic domain and C B D . Studies on other modular proteins suggested a role for linkers in protein stability. F o r example, an O-glycosylated l inker separating the starch-binding domain and catalytic domain o f a glucoamylase from Aspergillus awamori is important for thermostability, and resistance to degradation (Baker -L ibby et al, 1994). 7 Table 1.1: Representative linker sequences. A. representative linker sequences f rom cellulases, hemicellulases and a glucoamylase; B. two representative non-cellulase proline r ich linkers; C. a representative Q linker. T h e bo ld entries indicate linkers with 3 0 % or higher proline content. Un less otherwise noted, the sequence information is taken f rom Table 1 in Gi lkes et al, 1991d (Gi lkes et al, 1991d). 8 ORGANISM PROTEIN LINKER S E Q U E N C E no. A A no. P no. S/T A Cellulomonas fimi CenA PT2S(PT)4T(PT)7VTPQPT 33 14 17 Cex (PT)3T(PT)3T(PT)3S 21 9 12 CenB i PTGT3DT2P2T2PGTP ii T 2DT 2GETEP 2T 2PGTP 17 17 5 4 9 8 iii T 2A 2PVTVAPTVPGTP 17 4 5 iv S2PVTFT2LPVTSTPS 16 3 8 CenC i SLT 2SATP 3 10 3 5 ii PVPTAP 6 3 1 Thermomonospora *1 E l i DAGEPG 2E 2PGPGD 2ETP 2S 20 5 2 fusca ii TTEKEDETPTPSASC 15 2 6 *2 E2 (NP)4(TP)5P2GS2GA 25 11 7 E4 PE 2GE 2PG 3EGPG 3E 2PGEDVTP 2SAP 31 7 3 GS Streptomyces sp. CasA i PRT2(PT)2P 9 4 4 strain KSM-9 ii PA2TGA(SP)2AP2ASPAPSADS 22 7 6 Clostridium CelA PLSDLSGQPTP2SNPTPSLP2 21 8 6 thermocellum CelB TPSVT(PS)2ATPSPT2ITAP2T 22 7 11 CelE PLVS(PT)3LMPTPSPTVT 20 7 8 CelH i (PT)3WTSTP2S3P 16 6 9 ii PGTYPSYSKPSPTPRPTKP2VTP 24 10 7 Trichoderma reesei CBHI PG 2NRGT 4R 2PAT 3GS 2PGPTQS 26 4 11 CBHII PGA 2S 5TRA 2ST 2SRVSPT 2SRS 3ATP 3 GST 3RVP 2VG 44 7 22 Egll P 5AS 2T 2FSTTR 2S 2T 2S 3PSCTQT 29 5 16 Eg mi PGAT2IT2STRP2SGPT4RA(TS)2S2TP2 TS 2 34 6 21 Microbispora CelA P2TY(SP)2TPST(PS)3QSDPGS(PS)3 30 12 14 bispora Dictyostelium cellulase PS2TSVPT3PTVTET(PTET) 1 7 VT 99 24 52 discoideum 270-6 (PT)2VTPTETPS2 Aspergillus glucoamylase TG 2T 4ATPTGSGSVTSTSKT 2ATAS 37 1 26 awamori GA J KTSTSTS2TS B Human IgAl PVPSTP2TPSPSTP2T 16 8 7 Escherichia coli OmpA APV2(AP)4 12 5 0 C Bacillus subtilis SpoOA SGNAS 2VTHRAPS 2QS 2I 18 1 8 Additional references:*l (Jung et al, 1993),*2 author's interpretation from Genbank accession number M73321,*3 (Baker-Libby et al, 1994) 9 Catalytic Domain PT Linker CBD Cellulose Figure 1.2: Computer generated model of the CenA PT linker between known structures of a C. fimi cellulase catalytic domain and CBD. Tryptophans involved in binding are shown below the CBD. (figure courtesy of P. Tomme) Figure 1.3: The structure of an Fn3 module from fibronectin. (Main etal, 1992). 10 1.2.4. Proteolysis M a n y cel luloytic microorganisms partially degrade their cellulases releasing active catalytic domains (Hagspiel et al., 1989; L a o and W i l s o n , 1996; Sandercock et al., 1996; Warren, 1996). Partial degradation may also be seen in the recombinant production o f some cellulases (Fierobe et al., 1993; Fierobe et al., 1991) and m a y be forced using the natural proteases as wel l as non-specif ic proteases such as papain and chymotrypsin (Gi lkes et al., 1989; G i l kes et al, 1988; T o m m e et al., 1988). These degradation products gave insight into the domain structure o f cellulases as the protease-sensitive regions lay between domains. Serine proteases have been identified for both C. fimi and T. fusca (Langsford et al., 1984; L a o and W i l s o n , 1996). T h e protease from T. fusca has been cloned and characterized. Partial degradation o f the cellulases during hydrolysis on cellulose m a y be advantageous as the catalytic domains o f some cellulases are more active than the full enzymes on soluble substrates (Beguin and A l z a r i , 1998; F ierobe et al, 1993; Fierobe et al, 1991; Ghangas and W i l s o n , 1988; Reverbe l -Leroy et al, 1997). The separated domains may thus enhance hydrolysis o f soluble cellulose preventing its accumulation during hydrolysis (Sandercock et al, 1996). T h e protease sensitivity, however, is a shortcoming o f natural l inker sequences used in the design o f fusion proteins. F o r single-chain antibodies containing linkers derived from the T. reesei C B H I linker, length and amino acid composit ion affected linker stability in the hybr id construct. A proline placed after the proteolytic cleavage site reduced the sensitivity o f these constructs to proteolysis (Alf than et al, 1995). 11 S o m e sites in C e n A cleaved by papain and C. fimi protease have been reported previously (Gi lkes et al, 1989; Gi lkes et al, 1988; Langsford et al, 1984; Sandercock et al, 1996). F igure 1.4 indicates these sites. Figure 1. 4: Papain and C. fimi protease cleavage sites in C e n A produced in E. coli. A r r o w s point to cleavage sites; P, papain; C , C.fimi protease; the underl ined sequences belong to the domains adjacent to the linker. 1.3. Glycoproteins 1.3.1. General Glycoproteins are proteins with covalently attached carbohydrate moieties. The carbohydrates are added post-translationally or co-translationally. Glycoproteins are present in a diverse range o f organisms. There are many books and reviews on glycoproteins (Dwek, 1996; D w e k , 1995; Jenkins and Cur l ing , 1994; L i s and Sharon, 1993 and references therein) . A brief overview o f glycoproteins is provided here. 1.3.2. Protein-carbohydrate linkages There are two main types o f protein-carbohydrate l inkages, N - l inked and O- l inked . N- l inked refers to oligosaccharides attached to the protein through the amide o f asparagine CenA C CP T V P T T S P T P T P T P T T P T P T P T P T P T P T P T V T • CBD • catalytic domain 12 residues. O- l inked involves attachment o f sugars to the hydroxyamino acids, serine and threonine, v ia their hydroxy l groups (Beeley, 1987). Less c o m m o n l inkages, for example involv ing the hydroxy l groups o f hydroxylysine, hydroxyprol ine or tyrosine, occur in plants and prokaryotes (Moens and Vander leyden, 1997; Paul and Wie land , 1986; W i l s o n et al., 1991 and references therein). A new linkage class, phosphoglycosylat ion, where the oligosacharides are l inked to serine or threonine v ia phosphodiesters has been proposed recently by Haynes , 1998 (Haynes, 1998). 1.3.3. Sugars present in oligosaccharide T h e sugar c o m m o n l y found N- l inked to asparagine is N-acetylglucosamine ( G l c N A c ) and for O- l inkages, N-acetylgalactosamine ( G a l N A c ) , but there are deviations from this pattern. F o r example, rhamnose and G a l N A c appear N - l inked to bacterial glycoproteins (Messner and Sleytr, 1988; Paul and Wie land , 1986), and mannose (Man) , galactose (Gal ) , arabinose (Ara) , glucose (Glc) , G l c N A c , and fucose (Fuc) m a y be found in O- l inkages, pr imari ly in plant, lower eukaryotic, and prokaryotic glycoproteins (Gooley and Wi l l i ams , 1994; O n g et al., 1994; P lummer et al., 1995; W i l s o n et al., 1991). T h e sugars making up the remainder o f the oligosaccharide structures also vary. Sugars c o m m o n l y found in mammal ian glycoproteins include M a n , G a l , F u c , G l c N A c , G a l N A c , and N-acetylneuramic acid ( N e u A c ) (Raju et al., 1996). Sugars less c o m m o n l y found include G l c , and glucuronic acid (Plummer et al., 1995). N o v e l sugars occur in prokaryotic, fungal, plant and mammal ian glycoproteins (Gerwig et al., 1989; L i s and Sharon, 1993; M o e n s and Vander leyden, 1997; van den E i j n et al., 1995). In addition to the variety o f sugars, another degree o f complexi ty arises from the number o f possible 13 linkage positions between sugars (1-2,3,4 or 6) and the anomeric configuration o f that linkage (alpha or beta) (Raju et al, 1996). 1.3.4. Sizes of oligosaccharides T h e g lycan components o f glycoproteins can be large. T h e distance across a carbohydrate residue f rom O - l to 0 - 4 is 5.4 Angstroms. N- l inked oligosaccharides tend to be larger branching structures compared to the less complex O- l inked oligosaccharides which m a y be unbranched. A typical N - l inked oligosaccharide consists o f a pentasaccharide core with two or three outer branches o f three or four sugar units. U p to five branches m a y be present. O- l inked oligosaccharides range in size f rom one to twenty residues (Dwek, 1996; L i s and Sharon, 1993). The g lycan groups m a y have considerable freedom o f rotation around the glycosidic bond al lowing mot ion that cou ld result in large portions o f a protein being shielded; even by a relatively small ol igosaccharide (Dwek, 1996; L i s and Sharon, 1993). 1.3.5. Determinants of protein glycosylation Protein glycosylat ion is determined by the presence o f potential glycosylat ion sites, the structure o f the protein, and the expressing cel l type. T h e consensus sequence described for sites o f N- l inked glycosylat ion is N - X - T / S - X , where X is any amino acid except proline. Th is consensus sequence is c o m m o n to both prokaryotes and eukaryotes (Gooley and Wi l l i ams , 1994). T h e search for the O- l inked consensus sequence has not yielded a definitive motif , although there has been some success in predicting threonine and serine glycosylat ion sites in eukaryotic cells within a l imited sample (Figure 1.5) 14 (Gooley and Wi l l i ams , 1994 and references therein). Less research has been done on motifs for prokaryotic O-glycosylat ion. D - S * and D - T * - T , where X * represents a glycosylated amino acid, have been proposed as motifs for Chryseobacterium meningosepticum and V - Y * for Thermoanaerobacter kivui (Moens and Vander leyden, 1997). 1. X--p-- X -- X where one X = T (glycosylated) 2. X-- X --T-- X where one X = R / K and the T is glycosylated 3. s-- X - - X -- X where one X = S and the S is glycosylated 4. T-- X - - X -- X where one X = T (glycosylated) and the T is glycosylated 5. X - - X - -s - - X - - X -- X where one X = P and another X = D / E and the S is glycosylated 6. c- - X - -s - - X - -p--c where one X = P and another X = any amino acid and S is glycosylated 7. c- - X - -X - -G--G-- T - C - S where X = any amino acid and the T and S are glycosylated Figure 1.5: Mot i fs proposed for O-g lycosylat ion based on in vivo g lycosylat ion. Mot i fs 1-4 have been identified f rom some human and Dictyostelium proteins, motifs 5 and 6 were identified f rom epidermal growth factor regions o f some mul t idomain proteins (Gooley and Wi l l i ams , 1994; Pisano et al, 1994). No t all consensus sites or proposed motifs are glycosylated and some are on ly occasional ly glycosylated. There is also variation, microheterogeneity, in sugar composi t ion w h i c h can occur at a single site (Raju et al, 1996). A s a result, glycoproteins do not generally exist in a single form, but rather as a series o f species cal led glycoforms (Dwek, 1996). Th is heterogeneity led to the recognition o f the effect o f protein conformation on glycosylat ion. B o t h the overall conformation and the environment immediately surrounding a potential 15 glycosylat ion site may have an effect (Dwek, 1996). F ina l ly , each cel l type has its o w n set o f glycosyltransferases and glycosidases involved in the addition and processing o f the oligosaccharides. These can vary with culture conditions. G lycosy la t ion is therefore cell and cel l cycle specific. Examples o f this variation abound due to tissue specif ic studies and the use o f heterologous expression systems (Dwek, 1996; Jung and Wi l l i ams , 1997; Raju etal., 1996). 1.3.6. Prokaryotic glycoproteins A s alluded to above, glycoproteins are c o m m o n l y produced by eukaryotic cells. In fact the majority o f proteins secreted by mammal ian cells are glycoproteins (Goochee et al., 1991). Glycoproteins are found in organisms across the spectrum o f eukaryotic organisms and as a result, most o f what is known about glycosylat ion has come from the study o f eukaryotic systems. T h e presence o f carbohydrates o n bacterial cells was established with the description o f the crystalline cel l surface (S-layer) glycoprotein o f Halobacterium salinarium in 1976 (Mescher and Strominger, 1976). S- layer proteins have been studied since that time and have been reviewed recently (Messner et al., 1997b). Despite publications on non-S- layer bacterial glycoproteins appearing over the last twenty years, it was only recently that they have been recognized to be widespread. Th is is evidenced b y three minireviews appearing in the last four years (Messner, 1997a; M o e n s and Vander leyden, 1997; Sandercock et al., 1994). Bacterial glycoproteins are found in members o f both the archaebacteria and eubacteria and are generally d iv ided into S-layer glycoproteins, and non-S- layer glycoproteins. In the review by Messner (Messner, 1997a), the non-S- layer glycoproteins are further subdivided into surface-associated glycoproteins 16 in which flagell in and pi l i components are the major representatives, cellular glycoproteins, membrane glycoproteins, and secreted glycoproteins and exoenzymes. Thir ty- f ive genera are represented in reviews, o f which there are carbohydrate structures for glycoproteins from only seven: Neisseria meningitidis, Halobacterium halobium, Chryseobacterium meningosepticum, Bacteroides cellulosolvens, Clostridium thermocellum, Mycobacterium tuberculosis and Thermoplasma acidophilum (Messner, 1997a; M o e n s and Vander leyden, 1997; Sandercock et al, 1994). 1.3.7. Glycosylated polysaccharidases Polysaccharidases from a number o f bacteria and fungi are glycoproteins. Table 1.2 summarizes the characteristics o f a number o f bacterial polysaccharidases wh ich are glycoproteins and some proteins from the best characterized o f the cel lulolyt ic fungi , T. reesei. T h e structures o f some o f the carbohydrate moieties have been determined as indicated. N o t all polysaccharidases from a given organism are necessarily glycosylated. F o r example, o f the seven C. fimi cellulases described so far, only C e n A and C e x are glycoproteins (Langsford, 1988; Langsford et al, 1987). C e x is also glycosylated when expressed in Streptomyces lividans (S. lividans) ( M a c L e o d et al, 1992; O n g et al, 1994). Simi lar ly for T. fusca, only two o f six described cellulases are glycoproteins and they too are glycosylated when produced in S. lividans (Ghangas and W i l s o n , 1988; Z h a n g et al, 1995). In the case o f the fungal polysaccharidases, glycosylat ion appears to be c o m m o n . A l l the major cellulases o f T. reesei are glycoproteins (Teeri et al, 1992). 17 X o ^ u 1/1 u = * 1 ** If c 5 < = £ CQ A « .2 35 -4! t/i O O 3 —<" o " CN rn rn m »-H 00 7 3 I s| 1 * Ml cj M to O O O O 5 s 5 S 3 a a Q Q a u 1 § o >, ° s a o u H H U ca CQ u 00 60 60 a a CJ w w w 3 a 3 cct o 60 X X! to 8 s o x cj U s tu a If 8- © 2 © 5 o « | 6 E-s E~i >> o o o x> 1—1 © tfcl <B K ' m N u CO c C3 C C« o J 3 "60 O •o a u 60 O X ill . CN , m caw caw 2 i s S tS 0) s •— .o> OS o ^ u 01 s s c o u u o .2 « © "M -3 0J3 o e 18 ON « <-! in od •<d; in £ ^ (N V ! in Q 55 Q Q ti CQ I « — o O co :>, ° 3 ,o o Is 60 o § 2 § £ ^ 60 *^ 60 oa ca m O O O O a o Q c3 on s a o u 0> H U ca < CQ u 60 60 60 a a a w w W s te £ ^ 3 5 | CQ 1> CO B j Cd y o c a O cd c3 60 X X (3 I t<5 5 5P I 4 u a o U 5 3 X01 l l u o S a o o m CN w w 2 •a -S 2 o s te cl C3 o "eb o c u 60 O X i tN i m caw a w 2 s « S 19 -a CU s a * J a o fN JU 3 H ^ ^ c <*> — S ON ON o\ J2 o\ o\ o\ — — — > _ T _ r _ r -a cd cd cd c 2 M S ON ON O 1) DO cd N "cd U cd • 60 a a O O K 60 c — i t N r o ^ - i n v o r - o o O N ^ H 20 1.3.8. Glycosylation of CenA and Cex Langsford was the first to examine the glycosylat ion o f C e n A and C e x produced by C fimi (Langsford, 1988). There are N -glycosylat ion sites but no N -glycans were bel ieved to be present because the molecular weight o f C e n A or C e x were not mod i f i ed b y treatment with E n d o - H , an enzyme used to remove mammal ian N - l inked glycans. Carbohydrates could be released by alkali , an indication o f O- l inked glycosylat ion. Mannose was the only sugar constituent detected. It was hypothesized that the glycans were attached to the linker region o f the enzyme. O n g et al later determined, for C e x recombinantly produced in S. lividans, that glycosylat ion occurred on the P T linker. Mannose and galactose were detected (Ong et al, 1994). 1.3.9. Function of glycosylation, with an emphasis on cellulases A comprehensive review by D w e k looks at the function o f sugars o f glycoproteins (Dwek, 1996). There it is stressed that no one function can be ascribed to oligosaccharides. Possible roles include use as a recognit ion marker or modif icat ion o f properties such as susceptibility to proteases, stability, or quaternary structure. F o r cel lulosomes and secreted cellulases, no definitive role for g lycosylat ion has been identified. F o r each example o f a putative role, an exception can be found. A role for glycosylat ion in secretion o f cellulases o f T. reesei has been debated ( H e m m i n g , 1995), but it is unl ikely that glycosylat ion is required for secretion o f cellulases f rom either C . fimi or T.fusca because not all the secreted cellulases o f these systems are glycoproteins. It is possible that glycans m a y be involved in interactions with the substrate (Sandercock et al, 21 1994). Glycosyla ted P T C B D c e x , made in S. lividans, had greater affinity for cellulose than non-glycosylated P T C B D c e x produced in E. coli (Ong et al, 1994), but the ability o f C i p A , a scaffolding protein f rom C. thermocellum, to b ind its target proteins was not affected by glycosylat ion (Beguin et al., 1992). F o r C e n A and C e x from C.fimi and E 3 f rom T.fusca, the glycosylated and non-glycosylated proteins also showed no difference in their ability to b ind cellulose. The i r kinetic properties, thermostability, and p H stability were also unaffected (Langsford, 1988; Zhang et al., 1995). G lycosy la t ion has been reported to have an impact on sensitivity to proteolysis. F o r example C. fimi proteins glycosylated by C. fimi or, in the case o f C e x , also S. lividans, are protected against the action o f a C . fimi protease when adsorbed to cellulose but to a lesser degree in solution. T h e same cannot be said o f glycosylated E 3 f rom T. fusca, wh ich retained its sensitivity to proteolysis (Langsford, 1988; O n g et al., 1994; Zhang et al., 1995). T h u s , the role o f glycosylat ion may very wel l be peculiar to specif ic proteins, and settings. 1.3.10. Technological advances in carbohydrate analysis There is no single method for the characterization o f all ol igosaccharides so a combinat ion o f methods is required (Raju et al., 1996). W i th in the time frame o f m y research for this thesis, the technology for analyzing glycoproteins, and particularly carbohydrates, has significantly advanced. M o s t notably, there have been developments which make carbohydrate analysis faster and more accessible to the molecular biologist. These include the automation o f carbohydrate sequencing for N - l inked proteins (Dwek, 1996; R u d d et al., 1997), mass spectrometry (Burl ingame, 1996; Coste l lo , 1997; Harvey et 22 al, 1996) and fluorophore-assisted carbohydrate electrophoresis ( F A C E ® ) technology (Jackson, 1996; Ra ju et al, 1996). 1.4. IgAl and IgAl proteases 1.4.1. The structure and function of IgA M o r e immunoglobul in A (IgA) is produced in humans than all o f the other immunoglobul ins combined (Underdown and Mestecky , 1994). I g A is found in human serum and is, more importantly, the major antibody in secretions where it has an important role in mucosal immunity . T h e structure and function o f I g A has been reviewed previously (Kerr, 1990; L a m m et al, 1995). Electron microscopy conf i rmed b iochemica l analysis which proposed that monomel ic I g A has the c o m m o n four-chain structure o f other immunoglobul ins: two light chains and two heavy chains (Svehag and B l o t h , 1970). T h e latter are comprised o f four domains: a variable domain and three constant domains. A prol ine-r ich hinge region separates heavy constant domains 1 and 2. I g A has two isotypic forms, I g A l and IgA2. I g A l contains the full length hinge region o f twenty amino acids wh ich has an octapeptide direct repeat ( P - S - T - P - P - T - P - S ) (Figure 1.6), whereas IgA2 has two allotypes, both o f wh ich lack 13 amino acids o f this hinge region (Kerr , 1990). I g A has a variety o f molecular forms which are distributed in various bodi ly fluids in a defined pattern. Ninety percent o f I g A is in the I g A l form in serum. T h e percentage in the I g A l form in other body fluids varies between thirty-five and ninety-f ive percent depending on location (Delacroix et al, 1982; Kerr , 1990; Kett et al, 1986). Serum I g A l is primari ly monomel ic , a characteristic particular to humans (Vaerman, 1973). Secretory I g A l is a 23 product o f two cel l types, mucosal lymphocytes and epithelial cells. T w o molecules o f I g A l and the J chain glycopeptide, produced by mucosal lymphocytes, covalently associate. Th is complex passes through the epithelial cells l in ing the mucosal surfaces v ia a receptor-mediated secretion process. Dur ing this process a component o f the polymer ic immunoglobul in receptor, a glycopeptide referred to as S C , becomes covalently attached to the dimeric I g A l - J chain complex. Secretory I g A l is therefore quite different f rom serum I g A l being primari ly dimeric with two covalently attached glycopeptides (Brandtzaeg, 1995; Kerr , 1990; N e z l i n , 1998). I g A l has both N and O- l inked carbohydrates. There are two potential N - l inked and five potential O- l inked sites. The latter are all located in the hinge region. T h e N- l inked carbohydrates have been described previously and wi l l not be detailed here (Tomana et al, 1972). T h e O-g lycosyla t ion patterns differ between m y e l o m a and pooled serum I g A l . M y e l o m a I g A l , a monoclonal antibody, has min imal heterogeneity o f the g lycan component whereas serum I g A l is po lyc lona l and has greater heterogeneity (Mattu et al, 1998) (Figure 1.6 ). Secreted I g A l isolated f rom mi lk has more complex and heterogeneous oligosaccharides, wh ich may include N e u A c , F u c and G l c N A c ( M c G u i r e et al, 1989; Pierce-Cretel et al, 1981). The roles o f I g A are not completely understood, and l ikely differ between locations in the body. In secretions, the major, and most important role o f I g A is to inhibit the binding o f micro-organisms to mucosal surfaces. Th is is achieved through a number o f strategies, including agglutination. Other proposed functions o f secretory I g A include trapping antigens in the mucosal layer o f membranes, inhibit ing antigen penetration through these membranes, neutralizing bacterial toxins and enzymes, and enhancing non-24 specific antibacterial factors in secretions (K i l ian et al, 1996). Serum I g A is bel ieved to be involved in the removal o f antigenic material without triggering inf lammation (K i l ian et al., 1996). W h i l e the activation o f complement by I g A was controversial (Kerr, 1990), it is now thought that I g A does not activate complement by either the classical or alternative pathways (K i l i an and Russel l , 1994; N i k o l o v a et al, 1994). I g A has few effector functions. M o s t are dependent on the F c portion o f the antibody and thus require I g A to be intact (K i l ian et al., 1996). W h i l e secreted I g A is regarded as having an important role in immunity, the importance o f serum I g A has been questioned because I g A def iciency is not always associated with increased susceptibility to infection (Kerr, 1990). However , increased serum I g A concentration is associated with auto- immune diseases such as rheumatic diseases, some liver diseases and persistent infections such as A I D S (E lkon et al, 1983; K a l s i et al, 1983; Procaccia et al, 1987). 25 m - -• o -O~co CO I CL • CO I Q_ • CL • CL • S O -GL I I CO I CL • CL • CL • CL •*-» • mmm c 3 CO 1-o o < o CO 0 o < z to O T 3 O O 3 -a 53 - 3 o s o •a u s u 3 o crt O u o e 3 (U o Jc 8 « p tS "3 2 1 3 "2 u S u ft «-S lo C r t -3 u ft) CJ o 0 M , S ID CJ 35 CJ "O i g * o < z U cn o o o cx CJ 1 £ 6 3 3 U . 3 " i l l § | o ft, S U ex 3 SB'S ^ -a 1 I & crt crt a s O c3 13 J 3 w J3 O <+H 3 g I '&! I-I . 3 60 <U 3 TD a § 60 OO CJ >-< CJ < z cn Ov 3 Os 0 —c 1 A S CJ -5 u. 3 l l 26 1.4.2. IgAl proteases I g A l proteases are putative virulence factors for some pathogenic bacteria and have been comprehensively reviewed (K i l ian et al, 1996; Lomhol t , 1996). T h e I g A l protease family has representatives f rom the metallo-, serine- and cysteine-protease families. Var ious bacteria implicated in diseases originating f rom mucosal surfaces constitutively produce I g A l proteases (Table 1.3). Proteolysis occurs in the hinge region at post-proline peptide bonds in human I g A l . IgA2 is not a substrate for most o f the proteases as it lacks 13 amino acids in the hinge. O n l y one other bacterial endopeptidase, that f rom Flavobacterium meningosepticum, is known to cleave post-proline bonds but it does not cleave I g A l (Lomhol t , 1996; Walter et al, 1980). T h e specif ic site o f proteolysis in the I g A l hinge varies between I g A l proteases but is wel l defined for each enzyme. M o s t o f the enzymes cleave within the sequence containing the direct octapeptide repeats. It is remarkable that the proteases cleave at only one o f many identical peptide bonds. Some bacteria have strains w h i c h produce proteases with different specificities (Figure 1.7). Cleavage between proline and serine (P/S) is designated type 1 and cleavage between proline and threonine (P/T) is designated type 2. T h e cleavage specificity o f Neisseria type 2 I g A l proteases has been exploited in vitro by using a consensus sequence as a processing site for recombinant proteins (Pohlner et al, 1992). T h e consensus sequence for Neisseria type 2 I g A l proteases has been defined as X , P * X 2 P where X , can be P or rarely P - A , P - G , or P - T ; X 2 can be T , S , or A ; and * represents the cleavage posit ion. 27 Table 1.3: I g A l protease producing organisms implicated in disease. The * indicates that there is evidence o f I g A l protease activity in vivo. T h e table is based on Table 1 in Lomhol t , 1996 (Lomhol t , 1996). DISEASE ORGANISM(S) IMPLICATED IN DISEASE Caries, endocarditis, atopic disease * Streptococcus sanguis, Streptococcus oralis, and Streptococcus mitis Periodontis * Capnocythophaga and Prevotella species Meningitidis, respiratory disease, * Neisseria meningitidis, Haemophilus sinusitis, otitis media influenzae, and Streptococcus pneumoniae Gonorrhea, cystitis * Neisseria gonorrhoeae and Ureaplasma urealyticum Conjunctivitis * H. influenzae and N. gonorrhoeae Brazilian purpuric fever H. influenzae b iogroup aegyptius Structural features beyond the linear sequence m a y influence cleavage o f proteins by proteases. Protein folding, length o f substrate, interaction with other domains, spacing between recognit ion and cleavage sites, and glycosylat ion all have the potential to impact the eff iciency o f cleavage (Lomhol t , 1996; Q i u et al., 1996). N o single set o f rules covers all I g A l proteases. It has been proposed that mature N. gonorrhoeae and H. influenzae proteases have a cleavage-specif icity domain ( C S D ) wh ich differs in length between strains. T h e size o f the C S D is proportional to the distance from the disulphide bridge 28 C3 < g 0 u ea > o .89 60 u GO 03 CU -LJ O Ui OH < CL) E o 60 It-o < 0 0 CD CO E _o CL) E o c o '5b <u i-C U 6 0 C cu 6 0 CL) U CO 03 > CL) CU c 3 CL) CJ _c3 03 OC 03 a o c 03 LO O 03 00 CD CJ 03 O < g 03 a CO C O > CU X ) < 03 CU a . cu u cu cu SZ 03 CO tu CJ I Co a- ON u OS CO — I O o «S E cu o 29 between the I g A l alpha chains to the specif ic bond cleaved in the hinge. A s these enzymes are thought to recognize and interact with parts o f the F c a region, the C S D m a y act as a spacer between the catalytic site and substrate b inding site. Alternatively, the conformation o f the C S D m a y cause interference with access to other potential cleavage sites (Lomhol t , 1996; Q i u et al., 1996). G lycosy la t ion o f I g A l affects the rate o f cleavage by the streptococcal I g A l proteases. De-g lycosylated I g A l was a poor substrate whi le I g A l lacking the terminal sialic acids was a superior substrate (Reinholdt et al., 1990). T h e functions o f I g A l proteases are not yet wel l understood. Studies have indicated that some bacterial I g A l proteases are active in vivo during infection or colonization, but a definite role has not been established due to lack o f an appropriate animal mode l (Lomhol t , 1996) (Table 1.3). A recent development was the report o f a male human challenge mode l for N. gonorrhoeae (Cannon, 1998). T h e major l imitation o f this model was the short t ime-line al lowed for disease wh ich d id not al low the question o f I g A l protease function in disease maintenance to be addressed. Indirect evidence for an important role for I g A l proteases in disease has been reviewed (K i l i an et al., 1996). In br ief the evidence is: I g A l protease activity has evolved convergently f rom at least three independent evolutionary lines as I g A l proteases represent a variety o f enzyme families; three species o f pathogenic bacteria wh ich cause meningitis all produce I g A l proteases, whi le c losely related non-pathogenic species do not, although this observation is not true o f all I g A protease producing bacteria; and I g A l is important for immuni ty o f the mucosal surfaces that are colonized by most o f the I g A l protease-producing bacteria. In vivo and in vitro hydrolysis o f I g A l results in two intact fragments, the C-terminus o f the heavy chains (Fc a) and the antigen binding fragment (Faba) (Fig 1.8). The Fc a portion is responsible for the effector properties conferred by IgAl thus IgAl proteases effectively separate the antigen-binding and effector functions of IgAl . Another possible function for IgAl proteases was put forth in 1987 (Kilian and Reinholdt, 1987) and later described (Kilian et al., 1996; Lomholt, 1996): IgAl proteases cleave cross-reactive antibodies present from previous exposure to similar antigens. The resulting Fab a fragments of these cross-reactive antibodies might coat the surface of the pathogen, protecting it from recognition by intact antibodies. Figure 1. 8: Products of cleavage of IgAl by IgAl proteases. IgAl proteases have very narrow substrate ranges which vary with the particular protease. Three substrates are cleaved efficiently by serine IgAl proteases from Neisseria and H. influenzae in vivo: human serum and secretory I g A l ; and importantly, the precursor molecule of the IgAl proteases themselves. The serine IgAl proteases are the prototype for autotransporter proteins which contain within their precursor the protein component 31 which mediates their transport. Autotransporter proteins have been reviewed recently (Henderson et al, 1998). The I g A l precursor contains at least f ive components: a leader peptide w h i c h targets the protein to the periplasmic space; the (3 domain w h i c h creates a pore structure through wh ich the remaining domains pass; a and y peptide domains wh ich have prol ine-r ich sequences; and the mature protease wh ich folds and cleaves in the a and y peptide domains. Cleavage releases the mature protein and the a and y peptides domains from the bacterial cel l outer membrane. Alternative substrates for I g A l proteases have been reported. In vivo cleavage has been demonstrated for I g A l from chimpanzees, goril las, and orangutans (Kerr, 1990; Q i u et ah, 1996). There are also reports o f cleavage o f n o n - I g A l substrates by I g A l proteases in vitro. It needs to be stressed that the alternative substrates were not cleaved nearly as efficiently as I g A l . Reported substrates include: granulocyte/macrophage co lony stimulating factor, wh ich has functions including the up-regulation o f F c a receptors on neutrophils (Lomhol t , 1996; Weisbart et al, 1988); L A M P 1 , a major integral membrane glycoprotein o f late endosomes and lysosomes bel ieved to help maintain the stability o f these cellular compartments ( L i n et al, 1997); synaptobrevin II, a protein essential for exocytosis in neurons and chromaff in cells, wh ich contains a putative I g A l protease cleavage site (Binscheck et al, 1995); and a hybr id protein, C e n A I g A l h , where the linker o f the endoglucanase C e n A has been replaced by the I g A l hinge region (Mi l le r et al, 1992). C e n A I g A l h , produced in E. coli, was cleaved by N. gonorrhoeae type 1 and 2 proteases. However , C e n A was not cleaved, despite the sequence similarity to the I g A l 32 hinge region (Figure 1.9) (Mi l ler et al, 1992). A P S T P P T P S P S T P P T P S P S C B P T T S P T P T P T P T T P T P T P T P T P T P T P T Figure 1. 9: T h e sequence o f the I g A l hinge compared to the P T linker o f C e n A Panel A , I g A l hinge; Panel B , C e n A P T linker. 33 1.5. Objectives T h e ma in objective o f m y thesis research was to further investigate the structure, function and possible applications o f the interdomain linker o f C e n A . M o r e specif ic objectives were: 1. M o d i f y the l inker region o f C e n A and examine the properties, including activity and stability, o f the linker mutant proteins expressed in E. coli and S. lividans. 2. Further characterize the glycosylat ion o f C e n A and C e x from C. fimi inc luding definit ively determining the location o f glycans, re-addressing monosaccharide composi t ion and quantity, and the role o f glycosylat ion. 3. Characterize the glycosylat ion o f C e n A linker mutants produced in S. lividans. 4. Ut i l i ze glycosylated and non-glycosylated C e n A and C e n A I g A l h as potential substrates for I g A l proteases to aid in determining the specif icity o f these enzymes. 34 2. Methods and Materials 2.1. Chemicals, buffers, substrates and enzymes. Unless otherwise noted, chemicals were o f analytical or h igh pressure l iquid chromatography ( H P L C ) grade and purchased f rom S i g m a (St. L o u i s , M O ) or B D H (Toronto, O N ) . Solutions and buffers were prepared as described by Sambrook et al, 1989 (Sambrook et al., 1989) and were sterilized by filtration or autoclaving. A v i c e l ™ P H I 0 1 (micro-crystall ine cellulose) was obtained f rom F M C International (Cork, Ireland); bacterial micro-crystal l ine cellulose ( B M C C ) was produced by E . K w a n , Department o f M i c r o b i o l o g y and Immunology, U B C by a method previously described (Hestrin, 1963); phosphoric acid swollen cellulose ( P A S C ) was made f rom A v i c e l ™ P H I 01 by E . K w a n , Department o f M i c r o b i o l o g y and Immunology, U B C as described previously ( W o o d , 1988); para-nitrophenyl p-D-cel lobioside ( P N P C ) and carboxymethylcel lulose were f rom S igma; 2 ' ,4 ' -dini trophenyl P -D-cel lobioside (2,4-D N P C ) was a gift f rom K . Rupi tz and S. Withers from the Department o f Chemistry , U B C Immobi l ized papain was purchased from Pierce (Rockford , IL); C . fimi protease was produced f rom C.fimi A T C C 484 as described previously (Gi lkes et al., 1988); N -tosyl -L-phenyla lanylchloromethyl ketone ( T P C K ) trypsin was from S igma. H u m a n serum I g A l m y e l o m a paraprotein f rom patient V i e ( I g A l ) (Q iu et al., 1996) and I g A l proteases were gifts, as part o f a collaboration, f rom A . Plaut and J . Q i u as described on page xx. 35 2.2. Bacterial strains and plasmids T h e E. coli, C. fimi, and S. lividans strains used in this study are listed in Table 2.1. Bacterial stocks were maintained at - 70 ° C in media with 7 % dimethylsulfoxide ( D M S O ) or 10 % glycerol . Bacterial cells harboring constructs were stored at - 7 0 ° C in med ia with 7 % D M S O . D N A was stored in Tris-HCl/ethylenediaminetetraacetic acid ( E D T A ) buffer or water at - 2 0 ° C . T h e E. coli plasmids and the E. coli/S. lividans shuttle plasmids used in this study are listed in Table 2.2. and Table 2.3 respectively. Table 2.1: Bacterial strains. S T R A I N G E N O T Y P E R E F E R E N C E or S O U R C E Escherichia coli DH5oc F" end Al, hsdRl 7 (rk, m k + ) , supE44, thi-l, recAl, (argF-/aczya)U169, tylacZ 15 JM101 supE thi A (lac-proAB)[F' traA36 pro AB / a ? I q Z A M 1 5 ] (Hanahan, 1983) (Yanisch-Perron C . et al, 1985) C. fimi w i l d type A T C C 484 S. lividans T K M 3 1 a natural mutation o f T K 2 4 , a basic strain o f S. lividans 66 from D . A . H o p w o o d , selected for low protease activity A gift from D . B . W i l s o n , Section o f B iochemistry , Mo lecu la r and C e l l B i o l o g y , Corne l l Universi ty , Ithaca, N e w Y o r k 36 Table 2. 2: E. coli plasmids. PLASMID RELEVANT CHARACTERISTICS REFERENCE OR SOURCE pUC18-1.5ce«AAPT AmpR, the PT linker of cenA has been deleted A gift from H. Shen, Department of Microbiology and Immunology, UBC (Shen, 1990) pUC18-1.5ce«Ajgalh AmpR, the igal hinge replaced the PT linker in cenA A gift from P. Miller, Department of Microbiology and Immunology, UBC (Miller et al., 1992) pUC18-1.5ce«AAPTAS pUC18-1.5cenAAPT with Agel and Spel sites flanking the deletion to allow the insertion of mutant linkers This study pTAL3 AmpR, pTZ18R/PTIS with cenB A gift from A. Meinke, Department of Microbiology and Immunology, UBC (Meinke et al., 1991a) pUC18-1.5ce«A1.5PT, pUC18-1.5ce«A2PT, pUC18-1.5cenAFn3 first intermediates in construction of pTUg and pIJ680 variants; derived from pUC18-1.5ce«AAPTAS This study pSL1180 plasmid with superpolylinker used in this study for subcloning Pharmacia, (Brosius, 1989) pSL1180-cenAAPT, pSL1180-cenAIgAlh, pSL1180-ce«A1.5PT, pSL1180-cenA2PT, pSL1180-ce«AFn3 intermediates in construction of pTUg and pIJ680 variants This study pTUgKRG-1.5ce«^.N KanR, laclq, tac promoter, cex leader peptide sequence followed by a sequence encoding a six histidine tail with and IEGR site and wild type cenA; high expression vector A gift from H. Damude, Department of Microbiology and Immunology, UBC (Damude, 1995) pTUgKRG-cenAAPT, pTUgKRG-cen Alg A1 h, pTUgKRG-ce«A1.5PT, pTUgKRG-cenA2PT, pTUgKRG-cenAFn3 pTUgKRG-ce«A1.5PTmod pTUgKRG-1.5cenv4.N where cenA was replaced with cenA mutants This study 37 Table 2. 3: S. lividans/E. coli shuttle plasmids. pIJ680 i-CBD, cex Amp^ , Tsr**-, an S. lividans IE. coli shuttle vector with aph A gift from E. Ong, Department of Microbiology and Immunology, UBC (Ong etal., 1994) promoter sequence, E. coli origin of replication, cex leader peptide sequence and the cexCBD gene; a cloning and expression vector for use in Streptomyces pIJ680 pIJ680 pIJ680 i-ce«AAPT, i-ce«AIgAlh, >-cenA pIJ680-CBDcex where CBD c e x was replaced with cenA mutants This study pIJ680 -ce«A2PT, pIJ680-ce«AFn3 2.3. Media and growth conditions E. coli strains were grown in L u r i a broth ( L B ) (Sigma) at 37° C for D N A manipulations and Terr i f ic broth® (Sigma) at 30° C for protein production. M e d i a were supplemented with either 50 p g kanamycin or 100 p g ampici l l in m L ' 1 and, for recombinant protein induction, isopropyl-(5-D-thiogalactoside ( IPTG) to a final concentration o f 0.1 m M . 1 .5% agar was added to make sol id media. C.fimi was grown in modi f ied Leatherwood med ium (1 g K 2 H P 0 4 , 0.5 g K C 1 , 0.5 g M g S 0 4 , 1 g N a N 0 3 , 5 g peptone and 5 g yeast extract per Litre; p H 7.2) (Stewart and Leatherwood, 1976) supplemented with 0.1 % (w/v) carbon source at 30° C and shaken at 250 revolutions minute" 1 (rpm). So l id med ium used for C.fimi was low salt L B (0.5 g N a C l L"1) with 1.5% agar. 38 S. lividans was grown in tryptic soy broth ( T S B ) ( B D H ) , p H 7.2, supplemented with 5 p g thiostrepton (Tsr) mL" 1 at 30° C in baff led flasks at 250 rpm to increase aeration. T o prevent foaming during protein production, 1 drop ( about 30 p L ) o f antifoam 289 f rom S i g m a was added to 500 m L . So l id med ium used for S. lividans was T S B with 50 p g T s r mL" 1 and 1.5 % agar. 2.4. Recombinant DNA techniques Recombinant deoxyribonucleic acid ( D N A ) techniques for E. coli were as described in Sambrook et al, 1989 (Sambrook et ah, 1989) and for S. lividans as described in H o p w o o d et al, 1985 (Hopwood et al, 1985) unless otherwise specif ied. Restriction endonucleases, ligases, polymerases and nucleotides were f rom N e w Eng land B i o L a b s (Beverly, M A ) , Boehringer M a n n h e i m Canada, or G i b c o B R L (Gaithersburg, M D ) , and used as recommended by the suppliers. D N A was extracted f rom agarose gels using either the Geneclean® II kit (B IO101, Inc., V is ta , C A ) or Qiaex® II kit (Qiagen,Chatsworth, C A ) (Qiagen Inc., 1995). E. coli plasmid D N A was prepared using the Qiagen system (Qiagen) (Crowe, 1992). S. lividans plasmid D N A was prepared by a modif icat ion o f the E. coli alkaline lysis method (B i rnboim and D o l y , 1979), developed by the Jensen Laboratory (University o f Alberta , A B ) ; where 1.5 m L o f culture is spun to recover the myce l ia wh ich are then resuspended in 100 p L o f 50 m M glucose, 25 m M T r i s - H C l p H 8.0, 10 m M E D T A and 2 m g lysozyme m L ' 1 ; dispersed with a vortex mixer and incubated at 37° C for 30 minutes prior to the alkaline lysis method. Ligat ions were performed with a D N A insert to vector ratio o f approximately 10:1 in a total vo lume o f 12 p L and incubated at 1 6 ° C overnight. L igated 39 D N A was desalted by butanol precipitation (Thomas, 1994). Electroporat ion, with a GenePulser electroporator ( B i o - R a d , Miss issauga, O N ) , was used to transform E. coli. E. coli D H 5 a was used for D N A manipulations and JM101 for gene expression. S. lividans protoplasts were transformed by polyethylene glycol-mediated D N A uptake. 2.4.1. Primer design for introducing silent mutations The program Reverse Translator® written by D . Tr imbur , Department o f M i c r o b i o l o g y and Immunology, U B C was used to determine the silent mutations that could be introduced f lanking the P T linker o f C e n A . 2.4.2. Synthetic oligonucleotides and oligonucleotide primers Synthetic ol igodeoxyribonucleotides and ol igodeoxyribonucleotide primers were synthesized b y the U B C Nuc le ic A c i d and Protein Synthesis Un i t ( N A P S ) with an A p p l i e d B iosystem D N A synthesizer and purif ied by precipitation with n-butanol (Sawadogo and van D y k e , 1991). T h e oligonucleotides used in this study are described in Tables 2.4, 2.5 and 2.6. 2.4.3. DNA sequencing T h e A m p l i T a q polymerase dye termination protocol with the addit ion o f 5% D M S O was used to sequence D N A because o f the h igh G C content o f C. fimi D N A . T h e sequencing service was provided by the U B C N A P S unit using an A p p l i e d B iosystem D N A sequencer M o d e l 377 (Perkin-Elmer, Norwalk , C T ) . 40 Table 2. 4: Oligonucleotide primers used for PCR. NAME OLIGONUCLEOTIDE SEQUENCE (5'-3') SOURCE AND IMPORTANT CHARACTERISTICS LSMUTA GGG TCG ACG TAG AAG CCA CTA GTC GGC TGC GGC GTG ACG TTA ACG CTG GTC GTC GGC ACG GTA CCG GTG CAG GTG GTG CCG TT This study; introduces silent mutations forming Agel and Spel sites flanking the PT linker LSMUTC AAC GGC ACC ACC TGC ACC GGT ACC GTG CCG ACG ACC AGC GTT AAC GTC ACG CCG CAG CCG ACT AGT GGC TTC TAC GTC GAC CC This study; introduces silent mutations forming Agel and Spel sites flanking the PT linker SAENH AGG TCT ACT AGT CCC GGC TGC CGC GTC GAC A gift from P. Tomme, Department of Microbiology and Immunology, UBC CenARev TCA CCA CCT GGC GTT GCG CGC CAT A gift from H. Damude, Department of Microbiology and Immunology, UBC FNIIIDN GAA GCC ACT AGT CGG CTG AGG TGT TAC CGA CGG GGT GCT CGT CAC CGG CAG GGT This study: introduces a Spel site just downstream of the CenBFn3 repeats FNIIIUP CCT GCA CCG GTA CTG TAC CGA CCG GGA CGA CGA CGG AC This study: introduces a Agel site just upstream of the CenBFn3 repeats PTCCP GAA GCC ACT AGT CGG CTG AGG TGT TGG CGT AG This study; introduces a P (bold and underlined) at the C terminus of the PT linker Table 2. 5: Oligonucleotide primers used for sequencing. NAME OLIGONUCLEOTIDE SEQUENCE (5'-3') SOURCE AND IMPORTANT CHARACTERISTICS CenAP3 GTC ACG ATC ACC AAC CT A gift from A. Meinke, Department of Microbiology and Immunology, UBC FNIIIDN as in Table 2.4 FNIIIUP as in Table 2.4 41 Table 2. 6: Synthetic oligonucleotides for constructing mutant PT linkers. NAME OLIGONUCLEOTIDE SEQUENCE (5'-3') SOURCE AND IMPORTANT CHARACTERISTICS 2PTN CCT GCA CCG GTA CTG TA This study 2PTNC GTC GGT ACA GTA CCG GTG CAG G This study: 5 'phosphorylated 2PT1 CCG ACT ACC TCA CCT ACA CCA ACT CCT ACG CCG ACA ACT This study: 5'phosphorylated 2PT1C GTC GGA GTT GTC GGC GTA GGA GTT GGT GTA GGT GAG GTA This study: 5'phosphorylated 2PT2 CCG ACG CCA ACT CCA ACC CCG ACA CCA ACA CCT ACC CCT ACG This study: 5 'phosphorylated 2PT2C GGG TAG GTG TTG GTG TCG GGG TTG GAG TTG GC This study: 5'phosphorylated 2PT3 CCT ACC ACT TCG CCG ACT CCT ACT CCG ACG CCG ACC ACG This study: 5'phosphorylated 2PT3C GTC GGC GTG GTC GGC GTC GGA GTA GGA GTC GGC GAA GTG GTA GGC GTA G This study: 5'phosphorylated 2PTC GTA ACA CCT CAG CCG ACT AGT GGCTTC This study: 5'phosphorylated 2PTCC GAA GCC ACT AGT CGG CTG AGG TGT TAC CGT AG This study 2.4.4. PCR The polymerase chain reaction (PCR) was performed in a Twinblock™ System Easycycler™ thermocycler (Ericomp, San Diego, CA). 1-10 ng of template DNA, 20-50 pmol of each primer and 5 u,L of DMSO were combined in a 0.5 mL Eppendorf tube in a total volume of 10 pL. The mixture was heated for 1 minute at 96° C and then, while still 42 at 96° C , 40 p L o f a mixture containing 0.5 u L V e n t ™ polymerase (1 Uni t ) , 5 p X o f lOx V e n t ™ polymerase reaction buffer, and 4 m M o f each deoxyr ibonucleic acid triphosphate ( d N T P ) ( d A T P , d C T P , d G T P and d T T P ) were added. T h e mixture was covered with one drop o f mineral o i l and cyc led twenty times through program 83 (1 minute at 96° C , 45 seconds at 55° C , 90 seconds at 72° C ) and once through program 84 (7 minutes at 72° C ) . The reactions were then kept at 2 1 ° C until removed. p U C 1 8 - 1 . 5 c e « A A P X A S was constructed by using the four pr imer P C R method described in Figure 2.1. T w o P C R reactions using the template p U C 1 8 - 1 . 5 c e « A A P T and primer pairs L S M U T A , S A E N H and L U S M U T C , C e n A R e v , respectively were carried out as described above. 1 p L o f each o f the products o f these reactions were then combined and mixed with 5 p L D M S O in a 0.5 m L Eppendor f tube in a total vo lume o f 10 p L . The mixture was heated for 1 minute at 96° C . W h i l e still hot, 38 p L o f a mixture containing 0.5 p L V e n t ™ polymerase (1 Unit ) , 5 p L o f lOx V e n t ™ polymerase reaction buffer, and 4 m M o f each d N T P ( d A T P , d C T P , d G T P and d T T P ) were added. T h e mixture was then cycled once through program 83 (1 minute at 96° C , 45 seconds at 55° C , 90 seconds at 72° C ) and heated to 1 96° C for 1 minute. W h i l e still at 9 6 ° C , 2 p L o f a mixture containing 50 p m o l each o f primers S A E N H and C e n A R e v were introduced and the mixture was covered with one drop o f mineral o i l and cyc led twenty times through program 83 (1 minute at 96° C , 45 seconds at 55° C , 90 seconds at 72° C ) and once through program 84 (7 minutes at 72° C ) . T h e reaction was kept at 2 1 ° C until removed. E n z y m e and unincorporated nucleotides f rom the P C R were separated f rom the P C R products by phenohchoroform extraction and ethanol precipitation. T h e final P C R 43 product and vector p U C 1 8 - 1 . 5 c e « A A P T were digested with MM and Pflml, and the P C R fragment was ligated into the vector forming p U C 1 8 - 1 . 5 c e « A A P T A S . S A E N H L S M U T C A S L S M U T A P C R 1 20 cycles 96° C 60 sec. 55° C 45 sec. 72° C 90 sec. C e n A R e v P C R 2 20 cycles 96° C 60 sec. 55° C 45 sec. 72° C 90 sec. A P C R 3 1. combine products 2. denature (96° C, 60 sec.) X & anneal (55° C, 45 sec.) A S A A S 4. add primers (96 0 C) 3. extend fragments (72 0 C 90 sec.) S A E N H 5. 20 cycles 96" C 60 sec. 55°C 45 sec. 72° C 90 sec. C e n A R e v Figure 2.1: Generating p U C l 8 - 1 . 5 c e « A A P T A S . The four primer P C R method was used to engineer in silent Agel (A ) and Spel (S) sites f lanking the deleted P T linker site (A) to generate p U C 1 8 - 1 . 5 c e « A A P T A S . / / indicates D N A sequence that has been omitted for clarity. 44 2.4.5. Construction of mutant PT linkers cenAA?T and cenAigalh were obtained from p U C 1 8 - 1 . 5 c e « A A P T and p U C 1 8 -l.5cenAigalh, respectively. c e « A F n 3 and c e « A 1 . 5 P T m o d were constructed by P C R using the F N I I I D W and FNI I IUP primer set with the p T A L 3 template, and the P T C C P and S A E N H primer set with the p T U g K R G - c e « A 1 . 5 P T template, respectively. B o t h products encoded silent Agel and Spel sites wh ich al lowed insertion into p U C 1 8 - 1 . 5 c e « A A P T A S and p T U g K R G - c e « A F n 3 , respectively. c e « A 1 . 5 P T and c e « A 2 P T were constructed by hybr id iz ing and ligating synthetic oligonucleotides (Table 2.6) together as shown (Figure 2.2). E a c h pair o f ol igonucleotides ( 2 P T N and 2 P T N C = N , 2PT1 and 2 P T 1 C = 1, 2 P T 2 and 2 P T 2 C = 2, 2 P T 3 and 2 P T 3 C = 3, 2 P T C and 2 P T C C = C ) were first hybr idized together to form a double stranded piece o f D N A . T h e short double stranded D N A pieces were precipitated with ethanol in the presence o f 0.01 M M g C l 2 and recovered by ultracentrifugation at 27000 rpm for 1-2 hours at 4° C in a B e c k m a n T L - 1 0 0 Ultracentrifuge with T L A 45 rotor (Sambrook et al., 1989). Selected double stranded pieces were then ligated overnight using the ratios indicated in the brackets ( N and 1 (5:1), 2 and 3 (5:1)). The products were separated and the correct sized pieces isolated f rom a 2 % agarose gel. T o form 1.5PT, pieces N , 2-3-2, and C were ligated; and to form 2 P T , pieces N - l , 2-3-2, and C were ligated and concentrated by ethanol precipitation as above. T h e N - and C-terminus ol igonucleotide pairs encoded silent Agel and Spel sites wh ich al lowed insertion into p U C 1 8 - 1 . 5 c e « A A P X A S . + + + lc" + ~3C~ + 1. Hybridize 2. Concentrate by ethanol precipitation ligate & resolve on 2% gel ligate & resolve on 2% gel ligate ligate 1.5PT Figure 2.2: Construct ion o f cenkX .5PT and c e n A 2 P T using synthetic ol igonucleotides 46 F o r gene expression in E. coli and S. lividans, the cenA constructs were first subcloned, using Mlul and Pflml sites, f rom p U C 1 8 into p S L l 1 8 0 c e « A (created by subcloning an NhellHindlll fragment f rom p T U g K R G - 1 . 5 c e n A . N into p S L l 180). T h e n an Nhel/HindUI fragment f rom each p S L l 180cenA mutant was subcloned into either p T U g K R G - 1 . 5 c e « A . N to create p T U g K R G - c e « A mutants for expression in E. coli or p I J 6 8 0 - C B D c e x producing p I J 6 8 0 - c e « A mutants for expression in S. lividans. 2.5. Screening for mutants and confirmation of identity 2.5.1. E. coli Antibiotic-resistant colonies were screened by restriction endonuclease digestions o f p lasmid D N A preparations. Posit ive clones were conf i rmed by D N A sequencing. 2.5.2. S. lividans T o obtain stable clones, antibiotic-resistant colonies were placed in 1 m L T S B supplemented with 5 p g T s r mL" 1 at 30° C , 250 rpm. Those that grew were plated onto T S B agarose plates with high viscosity C M C (1%) and grown for 3 days. T h e clones were tested for C M C a s e activity using the C o n g o red staining procedure for carbohydrates (Teather and W o o d , 1982). Posit ive clones were conf i rmed by either product ion o f protein or sequencing o f plasmids. P lasmids isolated f rom S. lividans were diff icult to sequence so were transformed into E. coli D H 5 a and re-isolated for sequencing. 47 2.6. Gene expression and protein purification 2.6.1. E. coli C e n A was a gift from E . K w a n , Department o f M i c r o b i o l o g y and Immunology, U B C . Th is protein had been purif ied by affinity chromatography on C F 1 ™ cellulose (fibrous, long cellulose), (Sigma) with elution with 6 M guanidinium hydrochlor ide ( I C N Biomedica ls Inc, Aurora , O H ) and subsequent buffer exchange into 50 m M potassium phosphate buffer, p H 7.0. A 10 m g mL" 1 stock solution was made by di lut ing C e n A into 50 m M sodium citrate buffer, p H 7.0. Th is stock was used to make dilutions for the enzymatic assays and for use in proteolysis and other experiments. p30 was derived from C e n A by papain hydrolysis as described in Gi lkes et al, 1989 (Gi lkes et al., 1989) fo l lowed by gel filtration chromatography on a Superose 12 co lumn (Pharmacia, Uppsa la , Sweden) in 150 m M N a C l , 100 m M T r i s - H C l pH8 .0 in H P L C grade H 2 0 at a f low rate o f 0.3 m L min" 1 . F o r expression, plasmids were electroporated into E. coli J M 1 0 1 . E. coli J M 1 0 1 , carrying cenA mutant plasmids, were grown for 24 - 48 hours at 30° C and 250 rpm to an optical density at 600 n m ( O D 6 0 0 ) o f 5-6 in Terr i f ic broth® m e d i u m , supplemented with 50 m g kanamycin m L " 1 and 0.1 m M I P T G at the time o f inoculation. Ce l ls were separated from the supernatant by centrifugation. Phenylmethylsulfonylf luoride ( P M S F ) , sodium azide and E D T A were added to the supernatant to respective concentrations o f 0.1 m M , 0.02 % , and 3 m M . T h e supernatant was concentrated in an U l t rase t te™ tangential f low concentrator (3 k D a cut off, F i l t ron, Northborough, M A ) and then buffer was added to a final concentration o f 50 m M potassium phosphate, p H 7.0. T h e concentrate was clarif ied 48 by a 45 minute spin at 23700 x g and then loaded onto a C F 1 ™ cellulose co lumn and eluted and exchanged as for C e n A . C e n A mutants were further purif ied b y nickel affinity chromatography using the FfisBind® resin and protocol f rom N o v a g e n (Mi lwaukee, MI ) (Novagen, 1994). T h e proteins were exchanged into 50 m M sodium citrate buffer, p H 7 using an A m i c o n stirred cell concentrator (5 k D a cutoff), A m i c o n , (Bever ly , M A ) . A final centrifugation step at 40 000 rpm for 45 minutes in a B e c k m a n T L - 1 0 0 Ultracentrifuge with T L A 45 rotor was done to remove any protein wh ich had precipitated during the process. 2.6.2. C.fimi Native C e n A and C e x were purif ied from the supernatant o f C . fimi grown on modi f ied Leatherwood m e d i u m supplemented with 0.1 % (w/v) aspen holocel lulose (a gift f rom K . W o n g and J . Saddler o f the Department o f W o o d Science, Facul ty o f Forestry, Universi ty o f Br i t ish Columbia) . Cel lu lose-binding proteins were recovered by batch adsorption to A v i c e l ™ P H I 0 1 and elution with 8 M guanidinium hydrochloride. Glycoproteins were recovered by b inding to C o n A Sepharose (Pharmacia, Uppsa la , Sweden). C e n A and C e x were then resolved by anion exchange chromatography (Langsford, 1988). 2.6.3. S. lividans Methods were basical ly as provided by D. B . W i l s o n , Section o f B iochemistry , Mo lecu lar and C e l l B i o l o g y , Corne l l Universi ty , Ithaca, N e w Y o r k , w h i c h are modif icat ions o f publ ished methods (Hopwood et al., 1985). 49 Product ion o f the C e n A mutants: 500 u L o f an S. lividans T K M 3 1 D M S O stock f rom - 7 0 ° C carrying one o f the pIJ680cenA constructs was diluted into 20 m L o f T S B broth ( B D H ) supplemented with 5 p g T s r mL" 1 and grown at 30° C in baff led flasks to increase aeration until the cultures were dense but not pink (approximately 24 hours). 8 m L o f these pre-cultures were used to inoculate each o f four 500 m L preparations o f T S B , 5 p g T s r mL" 1 and 1 drop (about 30 p L ) o f antifoam in 2 L baff led flasks. These were grown at 30° C and 250 rpm for 48 to 72 hours. A t this time the cells were harvested by centrifugation and the supernatant filtered through glass w o o l and then through a T y p e A / E glass fibre filter (Ge lman Sciences, A n n Arbor , MI ) to remove any remaining myce l ium. 150 m L (packed volume) o f C F 1 ™ ; potassium phosphate buffer, p H 7.0 (final concentration 5 0 m M ) ; P M S F (final concentration 0.1 m M ) ; and sodium azide (final concentration 0.02%) were added to the supernatant. Th is was put at 4° C overnight with stirring to al low binding to cellulose. Af ter 24 hours the C F 1 ™ was recovered by settling and packed into an fast l iquid chromatography ( F P L C ) co lumn. T h e proteins were washed, eluted, and buffer exchanged as described for the E. coli proteins except the proteins were exchanged into sodium citrate buffer. N o further purif ication procedures were done. A final centrifugation step at 40 000 rpm for 45 minutes in a B e c k m a n T L - 1 0 0 Ultracentrifuge with T L A 45 rotor removed any protein wh ich had precipitated during the process. 50 2.7. Protein detection and quantification and storage Protein purity was estimated f rom proteins resolved by sodium dodecyl sulfate-polyacrylamide gel elctrophoresis ( S D S - P A G E ) (12%) and visual ized with Coomass ie -B lue (Laemml i , 1970). Approx imate ly 5 p g o f protein were analyzed in this manner. T h e spectrophotometric methods o f M a c h , M i d d a u g h and L e w i s ( A2 8 0 /A320 /A350 ) ( M a c h et al, 1992) and Scopes (A280 /A205) (Scopes, 1974); the Bradford dye b inding assay (Bio-Rad) ( B i o - R a d , 1993; Bradford , 1976); and amino acid analysis by the U B C N A P S unit were used according to the referenced publications or the supplier 's directions to determine the most appropriate routine method to be used in this study. Spectrophotometry was done using a Hi tachi U 2 0 0 0 spectrophotometer (Toyko , Japan). Where appropriate, the theoretical masses were used in the equations. Puri f ied proteins were stored at 4° C . 2.8. Western blotting and N-terminal sequencing of proteins and peptides Western blotting was done using a B i o - R a d mini-protean trans-blot® apparatus (0.5 A , 20 minutes). Proteins were transferred f rom S D S - P A G E gels to polyv inyl idene difluoride ( P V D F ) membranes ( I m m o b i l o n ™ , M i l l ipore Corporation). Prestained standards f rom B i o - R a d were: phosphorylase B (106 kDa) , bovine serum albumin ( B S A ) (80 kDa) , ova lbumin (50 kDa) , carbonic anhydrase (33 kDa) , soybean trypsin inhibitor (28 kDa) , and lysozyme (19 kDa) . N-terminal amino acid sequencing was done b y S. Perry o f the U B C N A P S unit using automated E d m a n degradation for 6 or more cycles on an Perk in E l m e r A p p l i e d Biosystems 4 7 6 A gas-phase sequencer (Matsudaira, 1990) or b y S. K i e l l a n d o f the 51 Department o f Biochemistry and M i c r o b i o l o g y , Universi ty o f V ic to r ia using automated E d m a n degradation for 6 or more cycles on a Perkin E l m e r A p p l i e d Biosystems 4 7 0 A gas-phase sequencer. 2.9. Mass analysis of proteins 2.9.1. Predicted molecular weights T h e predicted molecular weights o f the proteins were calculated f rom D N A - d e r i v e d amino acid sequences with the computer program D N A Strider (version 2) ( C E A , France). 2.9.2. Relative mass T h e relative molecular masses (M r ) o f the proteins were estimated f rom Coomass ie -B lue stained S D S - P A G E separated samples using the fo l lowing molecular mass standards from B i o - R a d : m y o s i n (200 kDa) , (3-galactosidase (116 kDa) , phosphorylase B (97 kDa) , bovine serum albumin (66 kDa) , ovalbumin ( 45 kDa) , carbonic anhydrase (31 kDa) , soybean trypsin inhibitor (22 kDa) , lysozyme (14 kDa) and aprotinin (7 kDa) . 2.9.3. Electrospray mass spectrometry Masses o f non-glycosylated proteins and protease products were determined by S. H e , Department o f Chemistry, U B C , using electrospray mass spectrometry ( M S ) . M a s s spectra were recorded on a P E - S c i e x A P I 300 triple quadrupole mass spectrometer from Sciex (Thornhi l l , Ontario). 52 2.9.4. Matrix-AssistedLaser-Desorption/Ionization Time-Of-Flight (MALDI-TOF) mass spectrometry Masses o f glycosylated proteins were determined using two different mass spectrometers. T h e masses o f C e n A and C e x produced b y C.fimi were determined using a T o f S p e c ® (Fisons Instruments, M A ) matrix-assisted laser-desorption/ionization time-of-flight ( M A L D I - T O F ) mass spectrometer b y the demonstration staff o f F isons . Proteins were analyzed in a matrix consisting o f the protein in H P L C water and 50 p M sinapinic acid (3 ,5-dimethoxy-4-hydroxycinnamic acid) in water-acetonitrile-trifluoroacetic acid (50:50:0.1). Equa l volumes o f the sample (10 p m o l pL" 1 ) and the matrix were combined and 0.5 p L were applied to a stainless steel target. T h e sample was al lowed to dry at r o o m temperature before insertion into the mass spectrometer. T h e average number o f shots at the target was 50, and the laser intensity was set at 3.95. T h e signals from both the singly and doubly charged ions o f lysozyme were used for internal mass calibration. T h e mass o f C e n A I g A l h produced by S. lividans was determined using a surface-enhanced laser desorption ionization ( S E L D I ) mass analyzer from Ciphergen Biosystems (Palo A l t o , C A ) . 1 p L (10 pmol) o f the protein in 7 0 % acetonitrile, 0 .1% T F A was placed onto a S E L D I P r o t e i n C h i p ™ array cationic b inding target. T h e sample was air-dried before adding the matrix (sinapinic acid) in 7 0 % acetonitrile, 0 .1% T F A . T h i s was air-dried before the chip was put into the mass analyzer. The signals from both the singly and doubly charged ions o f horse radish peroxidase (43240 D ) and bovine serum albumin (66410 D ) were used for calibration. T h e mass spectra were obtained from the signal 53 averaging o f approximately 50 laser shots. M a s s calibration and conversion to mass/charge distributions were performed using software supplied by Ciphergen (Ciphergen Biosystems, 1997). 2.10. Stability Protein stability was assessed under assay conditions by compar ing the C M C a s e activity o f the enzymes sampled at 0 and 24 hours f rom a m o c k B M C C hydrolysis experiment where the assay was set up without the B M C C and the samples were not boi led. 2.11. Enzyme assays A l l enzyme assays were done in 50 m M citrate buffer, p H 7 in a total vo lume o f 1.5 m L at 37° C unless noted. G lucose was used to generate standard curves. 2.11.1. PNPC hydrolysis C. fimi C e x activity was fol lowed by a dot blot assay using P N P C . 10 p L o f 12.5 m M P N P C (in 50 m M potassium phosphate buffer, p H 7 and 0.02 % N a N 3 ) were pipetted onto a piece o f parafdm. 10 p L o f the enzyme containing solution were added and mixed. Development o f a yel low colour within a few minutes was indicative o f P N P C a s e activity. 2.11.2. BMCC hydrolysis F i v e trials were performed with samples set up in triplicate. 1 m g o f substrate and 0.15 nmole enzyme were combined, mixed for 5 seconds on a vortex mixer and incubated at 37° C without further agitation. Af ter 24 hours, the tubes were placed in boi l ing water 54 for 3 minutes and then spun twice at 16600 x g, saving the supernatant after each spin. T h e supernatants were then placed at - 20 ° C until the time the samples were analyzed. F o r analysis, the samples were al lowed to thaw on ice and then 200 to 500 p L aliquots were analyzed in duplicate for reducing sugar content using the p-hydroxybenzoic acid hydrazide ( H B A H ) reducing sugar assay (Lever, 1973). 2.11.3. PASC hydrolysis Th is was done as for B M C C hydrolysis, except three trials were performed with samples set up in triplicate, 10 m g o f substrate were used and 10 to 20 p L samples were analyzed in duplicate. 2.11.4. CMC hydrolysis Reactions were set up in duplicate. A n enzyme sample was m i x e d with pre-warmed 0.5 % C M C in 50 m M citrate buffer in a total vo lume o f 0.5 m L and incubated at 30° C without further agitation for 30 minutes. 1 m L o f H B A H reagent was added to stop the reaction and then the H B A H reducing sugar assay was carried out as for B M C C hydrolysis (Shen, 1990). 2.11.5. 2,4-DNPC hydrolysis 2 , 4 - D N P C hydrolysis was carried out as described previously (Damude, 1995). A concentration o f approximately 10 x Michae l is constant ( K J (1.5 m M ) 2 , 4 - D N P C was used in the assays. T h e assays were carried out at p H 7.0. 55 2.11.6. pH profiles p H profi les o f the enzymes were compi led f rom 2 , 4 - D N P C hydrolysis reactions over a p H range o f 4.5 to 8.0 at 0.5 p H unit intervals. The method was described previously (Damude, 1995). A concentration o f 5 x K,,, (840 p M ) 2 , 4 - D N P C was used in the assays. Absorbance (using the method o f M a c h et al., Section 2.7) was measured for diluted enzyme samples after each set was completed to get the most accurate protein concentration reading for calculations. 2.11.7. Hydrolysis products The products o f B M C C hydrolysis by C e n A and C e n A F n 3 were analyzed as described previously (Stalbrand et al., 1998). The samples analyzed were taken at 10 and 96 hours post enzyme addition. H P L C analysis was done by S. Mans f ie ld , Forest Product Biotechnology, Department o f W o o d Science, U B C . 2.11.8. Statistical analysis of hydrolysis data T h e hydrolysis data were expressed as a ratio o f l inker variant specif ic activity to w i ld type ( W T ) specif ic activity. A statistical method wh ich can be used to determine whether ratios differ significantly f rom each other is the paired t-test o f the logs o f the specific activities, because the log o f a ratio is equal to the difference between the logs o f the individual components (Equation 2.1). U s i n g the logarithms produces a distribution more l ikely to be Gaussian than s imply uti l iz ing the raw ratios (Motulsky, 1995). Equat ion 2.1 S A C e n A mutant log S A C e n A log (SACenAmutant ) - l o g ( S A C e n A ) 56 2.12. Proteolysis 2.12.1. Papain hydrolysis The papain reactions contained 108 pg protein, 6.36 x 106 units (U) papain mol"1 protein, 5 mM L-cysteine (made fresh), 2 mM EDTA, 100 mM Tris-HCl pH 7.0, and 0.02 % NaN 3 in a total volume of 330 pL. The reactions were incubated at 37° C with agitation. 30 pL (approximately 10 pg protein) samples were taken at 0 minutes, 10 minute intervals for an hour, then at 90, 120 minutes and 24 hours. The samples were immediately combined with 10 pL of 4x SDS loading buffer and stored at -20° C. 20 pL (approximately 5 pg) of the samples were analysed by SDS-PAGE (16% acrylamide). To determine the proteolysis sites, selected samples were prepared for N-terminal sequencing by Western blotting as described in section 2.8. The membranes were stained with 0.1% Coomassie-Blue in 50% HPLC grade MeOH and destained with 50% HPLC grade MeOH. Selected bands were excised and sequenced. 2.12.2. C. fimi protease hydrolysis C. fimi protease was prepared from Cellulomonas fimi A T C C 484 grown on glycerol medium, and assayed with hide powder azure as described previously (Gilkes et al, 1988). The C.fimi protease reactions contained 70 pg protein, 2.2 x 108 U C.fimi protease mol"1 protein, 50 mM citrate buffer pH 7.0 in a total volume of 225 pL. The reactions were incubated at 37° C with agitation. 30 pL (approximately 10 pg protein) samples were taken at 0, 1, 5 or 6, 24, 48, 72 and sometimes 96 hours. The samples were 57 immediately combined with 10 pL of 4x SDS loading buffer and stored at -20° C. 20 pL (approximately 5 pg) of the samples were analysed by SDS-PAGE (12% acrylamide). To determine the proteolysis sites, selected samples were prepared for N-terminal sequencing by Western blotting as described for papain hydrolysis (Section 2.12.1). 2.13. IgAl protease digestion These reactions were carried out by J. Qiu, GRASP at Tufts University, Boston, MA. The IgAl proteases used in this study were recombinant H. influenzae type 1 from H. influenzae Rd, and impure preparations ofN. meningitidis type 1 from clinical strain N. meningitidis S-8273, N. meningitidis type 2 from clinical strain N. meningitidis S-7941, N. gonorrhoeae type 2 from clinical strain N. gonorrhoeae MSI 1 and S. pneumoniae from S. pneumoniae ATCC 6314. The reaction mixtures contained 0.75 mg substrate (CenAIgAlh or CenA from E. coli or S. lividans.) mL"1; 15 U IgAl protease mL"1; 50 mM Tris-HCl, pH 7.5; and 0.01 % sodium azide (an IgAl protease unit equals the amount of enzyme which can cleave 1 pg of IgAl sec"1 at 37° C, pH 7.5). 5 mM CaCl2 and MgCl2 were added for the S. pneumoniae digests. The reaction mixtures were placed at 37° C for up to 7 days. 15 pg of the digested substrates were analyzed by 10% SDS-PAGE. 2.14. Glycosylation determination 2.14.1. Periodic acid/Schiff reagent and ConA-HRP treatment of Western blots Periodic acid/Schiff reagent treatment (periodic acid/Schiff stain) of Western blots of SDS-PAGE gels was done according to the procedure described previously (Stromqvist 58 and Gruffman, 1974) using periodic acid, acetic acid (Fisher), premixed Schiff reagent (Sigma) and sodium metabisulfite (BDH). Western blots were treated with Concanavalin A horseradish peroxidase (ConA-HRP) (Seikagaku, Tokyo, Japan) and bands visualized using the HRP development reagent from Bio-Rad (Hercules, CA) or the enhanced chemiluminescence (ECL) detection system from Amersham. 2.14.2. Phenol sulfuric acid assay The phenol sulfuric acid assay for quantification of sugars was performed as described previously (Chaplin, 1986). Glucose was used as a standard. 2.14.3. Determination of the monosaccharide composition and linkage positions between sugars for CenA and Cex produced by C. fimi GLC-MS: Monosaccharides were released from CenA and Cex produced by C. fimi by methanolysis using the methanolic HC1 kit (Alltech, Deerfield, IL), silver carbonate (Aldrich) and acetic anhydride, and P 20 5 (Aldrich). The dried monosaccharide mixture was derivatized by trimethylsilylation using Tri-Sil® (Pierce) and pyridine (BDH) at 60° C. Samples were analyzed by gas liquid chromatography (GLC) with a DB1701 30 metre microbore (0.252 mm diameter) column (J and W Scientific) on a Carlo Erba Series 4160 GLC (Kratos, New York). Conditions: injector temperature 250° C and detector temperature 280° C; program, 140° C start for 25 minutes then an increase of 3° C min"1 to 230° C followed by 40 minutes at 230° C; carrier gas was Helium (1 kg cm2 _1). Standards for GLC were D-mannose (Aldrich - a 99% mixture of anomers), a-D-glucose and inositol 59 (Aldrich), D-(+)-galactose (Sigma - Sigma grade), N-acetyl-D-glucosamine and N-acetyl-D-galactosamine (Sigma) and L-(-)-fucose (Sigma - 99%). Inositol was used as an internal standard. The GLC-mass spectrometry of CenA and Cex samples was done by L. Madilao, Department o f Chemistry, University of Brit ish Columbia. The G L C conditions were as above, the interphase temperature was 280° C , and the mass was determined using a MS80 R F A mass spectrometer with a DS-55 data acquisition system (Kratos, New York). Exo-glycosidase digestions of CenA and Cex produced by C.fimi: Exo-glycosidases used were a-mannosidase from Jack bean meal (specificity a 1-2,3,6), and a -galactosidase from green coffee beans (specificity a 1-3,4,6) both from Oxford GlycoSystems (Abingdon, U K ) . The enzymes were used according to the manufacturer's specifications. The reaction mixtures contained 0.25 mg of CenA or Cex mL" 1 , 5 U exo-glycosidase mL" 1 in buffers recommended by the suppliers, in a final volume of 0.4 mL. The mixtures were incubated at 37° C for 24 hours. The a-mannosidase was removed from the reaction mixtures by gel filtration (Superose 12, Pharmacia) to prevent a subunit with similar apparent M r as CenA and Cex from interfering with the interpretation of data from S D S - P A G E . A fraction o f the mannosdiase-free C e n A or Cex was subsequently digested with a-galactosidase. The proteins in the sample were concentrated by binding to Avicel™ followed by release from Avicel™ by boil ing in S D S - P A G E loading buffer. The supernatant was retained following centrifugation and analyzed by S D S - P A G E and periodic acid/Schiff staining as described in Section 2.14.1. 60 2.14.4. Determination of the monosaccharide composition of CenA mutants produced by S. lividans and of CenA produced by C.fimi using FACE® T h e fluorophore-assisted carbohydrate electrophoresis ( F A C E ® ) monosaccharide composit ion kit supplied by G l y k o (Novato, C A ) was used according to the manufacturer's specifications. Monosacchar ides were released by acid hydrolysis and then labeled using 8-aminonaphthalene-l ,3,6-tr isulfonic acid ( A N T S ) by reductive animation chemistry. The labeled carbohydrates were separated o n high-percent pre-cast polyacry lamide gels. Monosacchar ides separate on F A C E ® gels as a result o f the formation o f charged complexes o f borate with vicinal hydroxy l groups during electrophoresis and can be visualized by long wave U V (G lyko , 1997; G l y k o , 1995a). 2.14.5. Determination of the oligosaccharide composition and linkages between sugars for CenA2PTproduced by S. lividans using FACE® The F A C E ® O- l inked oligosaccharide prof i l ing kit from G l y k o (Novato, C A ) was used according to the manufacturer's specifications with the addition o f the use o f the fo l lowing standards: a 1-2 mannobiose, a 1-3 mannobiose, a 1-4 mannobiose, a 1-6 mannobiose, and a 1-3,1-6 mannotriose, all from Dextra Laboratories L t d . ( U K ) , and ct-lactose and D-(+)-cellobiose from S i g m a ( M O ) . Ol igosaccharides were released by hydrazinolysis and then labeled using A N T S as described for monosaccharide composit ion analysis in Section 2.14.4. T h e labeled carbohydrates were separated on high-percent (up to 40 %), pre-cast polyacrylamide gels. Separation o f oligosaccharides was affected by the charge/mass ratio o f the saccharide as wel l as its hydrodynamic vo lume. T h e separation o f 61 neutral oligosaccharides was based on size. The bands were visualized by long wave UV (Glyko, 1997; Glyko, 1995b). 2.14.6. Location of glycans on CenA and Cex produced by C.fimi 200 jug of CenA or Cex produced by C. fimi were combined with urea and ammonium carbonate to a final concentration of 6.4 M and 0.32 M, respectively. The mixture was heated at 50° C for 10 minutes to denature the protein. Distilled water and 80 pg TPCK trypsin in 10 mM Tris-HCl, pH 8.0 were then added to a final volume of 1032 pL. The mixture was incubated at 37° C with agitation for 17 hours. Digestion was stopped by the addition of PMSF to a final concentration of 0.1 mM. Reduction of the trypsinized protein was performed by the addition of 50 pL of a 50 mM dithiothreitol (DTT) solution to 500 pL of the digest and incubation at 50° C for 10 minutes. In order to separate out the glycosylated peptides, MnCl2»4 H 2 0 and CaCl2*2 H 2 0 were added to 1 mM final concentration and this mixture was added to a 100 pL ConA sepharose affinity column packed into an approximately 500 pL spin column. The sample was loaded by gravity, washed 2 times with 200 pL Buffer A (150 mM NaCI, 1 mM CaCl2 2 H 20, 1 mM MnCl2 4 H 20, 0.01 % NaN3 and 50 mM Tris-HCl, pH 7.3) and eluted with 200 pL Buffer B (0.5 M methyl-a-D-mannopyranoside in Buffer A). Buffer B was allowed to soak the column for at least an hour to facilitate elution. HPLC was performed with the assistance of P. Tomme, Department of Microbiology and Immunology, UBC using a Vydac 218TP54 4.6 x 25 cm C18 reverse phase column (300 A, 5 p support) with a resolve C18 guard pak column on a Shimadzu HPLC system with a SIL6B autoinjector, an RID6A refractive index detector and a CRS01 chromatopac integrator-plotter. 150 pL 62 (approximately 1.5 nmoles protein) were injected. T h e f low rate was 0.8 m L min" 1 . The solvents were all H P L C grade. Solvent A contained 0.08 % T F A in water and Solvent B comprised 0.072 % T F A and 80 % acetonitrile in water. The elution program used is shown in Table 2.7. T h e computer program P C gene was used to predict the elution times for the peptides. T h e peptides were detected by A 2 3 0 and collected by hand. Attenuation was 5 and the fullscale was 0.08 absorbance units for the integrator-plotter. G lycosyla ted peptides were attached to P V D F membranes in spin columns (Pro S p i n ™ sample preparation cartridge, A p p l i e d Biosystems, Foster C i ty , C A ) and N-terminal sequenced. H P L C was also performed on samples prior to the C o n A affinity chromatography step, and on tryptic digests o f C e n A and C e x produced by E. coli as controls. Table 2. 7: H P L C program. 0 5 5 65 75 75 100 80 100 90 5 91 stop 63 3. Results 3.1. Generation of cenA mutants. 3.1.1. Definition of the boundaries of the PT linker of CenA F o r this study the w i l d type ( W T ) P T linker was considered to be 27 amino acids ( P 1 0 8 - P 1 3 4 ) . In previous studies the P T linker was defined as 23 amino acids ( P U 2 - T 1 3 4 ) (Shen, 1990), and 33 amino acids ( P 1 0 g - S 1 4 1 ) (Gi lkes et al., 1991d). T h e linkers are shown in Figure 3.1. A s other cellulase P T r ich linkers contain serine (Gi lkes et al., 199Id), it was decided that the P T T S sequence should be included in the definit ion o f the linker. Va l ine 135 through threonine 140 was not included because the N-terminal amino acid o f a C.fimi protease product, p30, is valine 135 (Gilkes et al., 1989). A s p30 is stable against further proteolysis by the C.fimi protease, valine 135 is l ikely the beginning o f the catalytic domain. 108 134 ( A ) P T T S P T P T P T P T T P T P T P T P T P T P T P T 112 134 (B) P T P T P T P T T P T P T P T P T P T P T P T 108 (C) P T T S P T P T P T P T T P T P T P T P T P T P T P T V T 140 P Q P T Figure 3.1: T h e P T linker o f C e n A as defined in various studies. Th is study (Panel A ) ; previous studies (Shen, 1990) (Panel B ) , and (Gi lkes et al, 199Id) (Panel C ) . 64 3.1.2. Primer design for introducing silent mutations T h e program Reverse Translator®, as noted in Section 2.4.1, was used to determine appropriate silent mutations to flank the P T linker o f C e n A . T h e exact D N A bases changed to introduce the silent Agel and Spel sites are shown in F igure 3.2. 104 T G T V P T T S P T P T P T P ACC GGC ACC GTG CCG ACG ACC AGC CCC ACG CCG ACC CCG ACG CCG ACC GGT Agel T T P T P T P T P T P T P T P ACG ACC CCC ACG CCG ACG CCG ACC CCG A C C CCC A C C CCC ACG CCG 141 T V T P Q P T S ACG GTC ACG CCG CAG CCG ACC AGC ACT AGT Spel Figure 3. 2: D e s i g n o f silent mutation sites. The P T linker amino acid sequence is in bold . The base pair changes needed to introduce the silent mutations are also in bo ld and the restriction endonuclease recognit ion sites are underlined with the restriction endonuclease cited below. 3.1.3. Generation ofpUC18-1.5cenAAprAS In order to introduce silent mutations using only four primers as detailed in Section 2.4.4, pUC18-1.5ce / iAApT (Shen et al, 1991) was used as a template for P C R . T h e deletion in p U C 1 8 - 1 . 5 c e w A A P T from amino acids 108-134 made it possible to design a single set o f complementary mutagenic primers, shown in Table 2.4, w h i c h spanned the region where the P T linker variants would be inserted. Fewer P C R reactions were required 65 using the four primer method. This reduced the probability of introducing unwanted mutations. The product from the four primer PCR method was digested with Mlul and Pflml and subcloned into pUC18-1.5ce«AAPT to form pUC18-1.5ce«AAPTAS. 3.1.4. Construction of the cenA mutants Construction of pUC18-l.5cewAl.5PT and pUC18-1.5ce«A2PT : Oligodeoxyribonucleotides and oligodeoxyribonucleotide primers were synthesized as described in Section 2.4.2. The synthetic oligomers were optimized for two properties. First, the GC content was reduced because gene expression was to be done in E. coli, which has a lower GC content than C.fimi (48 - 52 % vs. 71 - 76 %) (0rskov, 1984; Stackebrandt and Keddie, 1984). The changes reduced the GC content of the linker region from 77% to 58%. Second, the sequences were manipulated to minimize erroneous oligomer hybridization using the overlap detection function of DNA Strider™. The synthetic oligonucleotides were hybridized and ligated together as outlined in Section 2.4.5. 1.5PT and 2PT oligonucleotides were digested with Agel and Spel and inserted into pUC18-1.5ce«AAPTAS to form pUC18-l.5cenAl.5PT and pUC18-1.5ce»A2PT (Figure 3.3). Construction of pUC18-l .5ce«AFn3: PCR was used to engineer Agel and Spel sites into the N- and C- terminus, respectively, of the linkers adjacent to the Fn3 repeats of CenB, as described in Section 2.4.5. This allowed replacement of the PT linker of CenA with the Fn3 repeats. Plasmid pTAL3, encoding CenB, was used as the Fn3 template in the PCR. The PCR product of 945 bp was digested with ^ 4gel and Spel and ligated into pUC18-1.5cenAAPTAS to form pUC18-1.5cenAFn3 (Figure 3.3). 66 p U C 1 8 - 1 . 5 c e « A A P X and p U C 1 8 - 1 . 5 c e « A z ' g a l h : these plasmids were gifts (Table 2.2) 3.1.5. Subcloning cenA PT mutants into E. coli and S. lividans expression vectors p T U g K R G c e « A mutants: p U C 1 8 - 1 . 5 c e « A A P T A S d id not have the unique Nhel and Hindlll sites needed for subcloning the cenA l inker constructs into the expression vectors p T U G for expression in E. coli (Graham et al., 1995) , and p I J 6 8 0 C B D c e x for expression in S. lividans ( O n g et al., 1994). Therefore, a subcloning strategy was required. T h e strategy that was used is outlined in Figures 3.4 and 3.5. A PflmllMlul D N A fragment f rom each p U C 1 8 - 1 . 5 c e « A mutant (sizes in base pairs: W T : 516; A P T : 453; 1.5PT: 555; igalh: 510; 2 P T : 597; F n 3 : 1350 ) was first subcloned into p S L l 1 8 0 c e « A , created by subcloning a 1.5 kb NheVHindlll fragment f rom p T U g K R G - 1 . 5 c e « A . N into p S L l 180. T h e n an Nhel/Hindlll fragment (Figure 3.6) f rom each p S L l 1 8 0 c e « A mutant was subcloned into p T U g K R G - 1 . 5 c e « A . N to create p T U g K R G c e w A mutants. p T U g K R G c e r c A mutants were then electroporated into E. coli D H 5 a . Posit ive clones were selected and conf i rmed as detailed in Sect ion 2.5.1. F o r expression, D N A constructs were electroporated into E. coli J M 1 0 1 . p I J 6 8 0 c e « A mutants: Nhel/Hindlll fragments f rom all the p S L l 1 8 0 c e « A mutants except p S L l 1 8 0 - c e « A 1 . 5 P T , were subcloned into p I J 6 8 0 C B D c e x to create p I J 6 8 0 c e « A mutants (Figure 3.5). p I J 6 8 0 c e « A mutants were then transformed into S. lividans T K M 3 1 and stable clones selected as described in Sections 2.4 and 2.5.2. T h e C M C a s e activity o f S. lividans colonies expressing C e n A A P T is shown in F igure 3.7. S. lividans colonies 67 wh ich d id not express a recombinant cellulase exhibited no C M C a s e activity (data not shown). Construct ion o f p T U g K R G - c e « A l . 5 P T m o d : Site-directed mutagenesis as described in Section 2.4.5 was used to change valine 176 to proline. Posit ive clones were selected and conf i rmed as detailed in Section 2.5.1. F o r expression, the p lasmid was electroporated into E. coli J M 1 0 1 . 68 Figure 3.3: Construct ion o f p U C l 8-1.5cenA mutants. Silent mutations were introduced in p U C 1 8 - 1 . 5 c e n A A p T to introduce Age I and Spe I sites f lanking the P T linker (1); P C R was used to generate cenB Fn3 repeats from p T A L 3 (2); the synthetic ol igonucleotides for 1.5 and 2 P T linkers and the P C R product encoding the Fn3 repeats from cenB were digested with Age I and Spe I and inserted into p U C 1 8 -1 .5cenAApTAS to form p U C 1 8 - 1 . 5 c e n A mutants (3). 69 P(tac) | a c l O r i M13 Ori Figure 3.4: Part 1 o f subcloning strategy to insert cenA mutants into expression vectors p T U G and pIJ680. cenA was subcloned f rom p T U g K R G - 1 . 5 c e n A . N into p S L l 180 us ing Nhe I and Hind III to form p S L l 1 8 0 c e « A (1); D N A fragments o f cenA containing the P T mutants were subcloned f rom p U C 18-1.5cenA into p S L 1 1 8 0 c e « A using Mlu I and Pfl M I to form p S L l 180cenA mutants (2). 70 Figure 3.5: Part 2 o f subcloning strategy to insert cenA mutants into expression vectors p T U G a n d p I J 6 8 0 . cenA mutants were subcloned into p T U g K R G - 1 . 5 c e « A . N and p I J 6 0 C B D c e x using Nhe I and Hind III f rom p S L l 180cewA mutants (1) to form p T U g K R G - 1 .ScenA mutants and pIJ680cenA mutants, (2) and (3), respectively. 71 1 2 3 4 5 6 Figure 3. 6: Agarose gel showing the NheVHindlll D N A fragment f rom each p S L l 1 8 0 c e « A mutant used to subclone into p T U g and pIJ680. cenAApT (Lane 1, 1470 bp), cenA (Lane 2 ,1533 bp), c e « A 1 . 5 P T (Lane 3, 1572 bp), c e « A 2 P T (Lane 4, 1614 bp), cenAFn3 (Lane 5 ,2367 bp), cenAigalh (Lane 6, 1527 bp). Figure 3. 7: C o n g o - r e d stained agar plate showing halos where S. lividans colonies expressing C e n A A P T were growing. T h e black marks inside the halos indicate positions o f colonies that were scraped off. 72 3.2. Gene expression and protein purification. 3.2.1. E. coli E. coli harbouring pTUgKRG-ce«A constructs were grown and proteins purified as outlined in Section 2.6.1. The yield was up to 5 mg purified protein L" 1 . Some slight yellow/brown colour was retained through the purification. A method o f protein measurement was chosen which accounted for the colour (Section 3.3). C e n A had been purified as described in Section 2.6.1. p30 was obtained by digesting C e n A with papain as described in Section 2.6.1., followed by size exclusion chromatography as described previously (Gilkes et al, 1991c). A l l proteins were purified to greater than 90% homogeneity as judged from a Coomassie-Blue stained S D S - P A G E gel (Figure 3.8). The modular structures of the mature proteins and the final linker sequences are detailed in Figure 3.9. kDa 1 2 3 4 5 6 7 116 97 66 45 31 Figure 3. 8: S D S - P A G E gel of CenA constructs produced in E. coli. CenA (Lane 1), C e n A A P T (Lane 2), C e n A I g A l h (Lane 3 ) , CenA1.5PT (Lane 4 ) , CenA2PT (lane 5), CenAFn3 (Lane 6) and p30 (Lane 7). The S D S - P A G E gel was stained with Coomassie-Blue. 73 Q H > l co Cc H CO H > Cu J H H fa > Cc CO oo 74 3.2.2. S. lividans Proteins were produced in and purified from S. lividans as described in Section 2.6.3. The yield was up to 10 mg purified protein L"1. Some slight yellow/brown colour was retained through the purification, as for proteins produced in E. coli. All proteins were purified to greater than 90% homogeneity as judged from a Coomassie-Blue-stained SDS-PAGE gel (Figure 3.10). To eliminate the possibility that the cellulose-binding proteins were of S. lividans origin, they were tested with an antibody against CenA and were all reactive (data not shown). kDa 1 2 3 4 5 -Figure 3 . 1 0 : SDS-PAGE gel of CenA constructs produced in S. lividans. CenAAPT (Lanel), CenAIgAlh (Lane2), CenA (Lane3), CenA2PT (Lane 4) and CenAFn3 (Lane 5). The SDS-PAGE gel is stained with Coomassie-Blue. 75 3.2.3. C.fimi CenA and Cex were purified from the supernatant of C. fimi as described in Section 2.6.2. 0.8 mg CenA L"1 and 0.3 mg Cex L"1 were purified. The proteins were purified to greater than 90% homogeneity as judged from an SDS-PAGE gel stained with Coomassie-Blue (Figure 3.11). kDa 1 2 9 7 6 6 45 Figure 3.11: SDS-PAGE gel of CenA and Cex produced in C.fimi. CenA (Lane 1) and Cex (Lane 2). The SDS-PAGE gel is stained with Coomassie-Blue. 3.3. Protein quantification A number of protein quantification methods was explored using a subset of the proteins produced in E. coli. Protein concentrations were determined by the spectrophotometric methods of Mach, Middaugh and Lewis (A2 8on m, A 3 2 0 n m , A 3 5 0 n m ) and Scopes (A 2 8 0 n m , A 2 0 5 n m ) , by the chemical method of Bradford, and by amino acid analysis, as described in Section 2.7. The results are shown in Table 3.1. It was assumed that the amino acid analysis was the most accurate assay but, as it was not practical for routine analysis, it was used as a benchmark against which the other methods were compared. The 76 Bradford dye-binding assay values were furthest from the amino acid analysis values for all but one sample. The method o f M a c h , M i d d a u g h and L e w i s was chosen for routine measurements over the Scopes method because it accounted for interference by absorbance due to colour retained during protein purif ication. Table 3.1: Protein concentrations o f C e n A constructs produced in E. coli determined by a variety o f methods. P R O T E I N D E T E C T I O N M E T H O D E. coli protein (concentration in mg mL"') C e n A C e n A A P T C e n A 2 P T C e n A F n 3 Bradford Dye-Binding Assay* 6.4 1.7 2.7 1.9 Scopes (A280,A205) 8.9 2.3 4.0 3.6 Mach (A280,A320,A350) 7.2 1.6 3.3 3.0 Amino Acid Analysis 8.2 1.9 3.6 3.5 used B S A as the protein standard 77 3.4. Molecular masses of proteins The sizes o f the proteins produced in E. coli and S. lividans were determined b y three methods as described in Section 2.9: predicted molecular weight, M r , and electrospray M S . The results are summarized in Table 3.2. T h e larger than expected M r results are due to the P T linker or the I g A l hinge. It has been shown previously that proteins containing P T r ich linkers or the I g A l hinge result i n aberrantly slow migrat ion dur ing S D S - P A G E (Mi l le r et al, 1992). T h e difference between the predicted molecular weight and the electrospray M S values for C e n A 1 . 5 P T construct l ikely reflects misprocessing by one amino acid at the N-terminus. T h e N-terminus was not sequenced to conf i rm this. 78 Table 3. 2: Mo lecu la r masses o f C e n A and C e n A mutants produced in E. coli and S. lividans determined by three different techniques. A l l masses are in D a . T h e electrospray M S values are accurate to 0.1%. MOLECULAR MASS Protein Organism used for protein production Predicted (DNA Strider) Relative Mobility Electrospray MS CenAAPT E. coli S. lividans 43276 50400 50400 43287 not determined CenAIgAlh E. coli S. lividans 45133 55800 56900 45149 not determined CenA E. coli S. lividans 43820 55800 59000 43826 not determined CenA1.5PT E. coli 46622 57200 46745 CenA1.5PTmod E. coli 46620 57200 not determined CenA2PT E. coli S. lividans 48009 63600 67900 48026 not determined CenAFn3 E. coli S. lividans 72466 75000 75000 72498 not determined p30 E. coli 30077 29500* 30078 * this was determined from a separate gel 79 3.5.Glycosylation determination 3.5.1. Detection of glycosylation on proteins produced in S. lividans and E. coli T h e proteins produced in E. coli and S. lividans were tested for the presence o f covalently attached sugars by the periodic ac id /Schi f f stain and C o n A - H R P as described in section 2.14.1. T h e proteins produced in E. coli were not glycosylated as they d id not react with the periodic ac id /Schi f f stain (data not shown). A l l but C e n A F n 3 were glycosylated by S. lividans (Figure 3.12). The lower bands in Lanes 3 and 4 (Figure 3.12) indicate a low level o f degradation has occurred. These degradation products are apparent due to the high sensitivity o f C o n A - H R P . T o conf i rm that C e n A F n 3 was not glycosylated by S. lividans, approximately 10 ug o f protein made by E. coli and S. lividans were blotted and tested with C o n A - H R P . Neither protein reacted (Figure 3.13). F igure 3.14 shows the molecular mass difference between the proteins produced by E. coli and S. lividans. The more diffuse bands seen for proteins produced by S. lividans are typical o f glycoproteins, which exist as a set o f g lycoforms varying by as little as one sugar unit. T h e lower molecular weight bands in Lane 1 (Figure 3.13) are degradation products o f an o ld preparation o f C e n A F n 3 and the lower molecular weight bands in Lanes 1 and 2 (Figure 3.14 B ) are degradation products evident from overloading these samples o f C e n A F n 3 . 80 Figure 3.12: SDS-PAGE gel and ConA-HRP treated blot of CenA constructs produced in S. lividans. SDS-PAGE gel stained with Coomassie-Blue (Panel A) and ConA-HRP treated blot (Panel B), CenAAPT (Lane 1), CenAIgAlh (Lane 2), CenA (Lane 3), CenA2PT (Lane 4) and CenAFn3 (Lane 5). Figure 3.13: SDS-PAGE gel and ConA-HRP treated blot of CenA constructs produced in S. lividans and E. coli. SDS-PAGE gel stained with Coomassie-Blue (Panel A) and ConA-HRP treated blot (Panel B ) , CenAFn3 produced by E. coli (Lanel), CenAFn3 (Lane2), CenAAPT (Lane3) and CenA (Lane 4) produced by S. lividans. The arrow indicates the size of CenAFn3 on the blot. 81 A kDa 1 2 3 4 5 6 7 8 Figure 3.14: S D S - P A G E gels o f C e n A constructs produced in E. coli and S. lividans. E. col i (odd lanes) S. lividans (even lanes), Panel A : C e n A A P T (Lanes 1 and 2), C e n A I g A l h (Lanes 3 and 4), C e n A (Lanes 5 and 6), and C e n A 2 P T (Lanes 7 and 8); Panel B : C e n A F n 3 (Lanes 1 and 2). T h e S D S - P A G E gels were stained with Coomass ie -B lue . 3.5.2. Moles of sugar per mole ofprotein for CenA and Cex produced in C. fimi and CenAIgAlh produced in S. lividans T h e masses o f C e n A and C e x produced by C.fimi were determined by M A L D I -T O F M S as detailed in Section 2.9.4. Masses o f 48519 and 47874 D a were determined in two runs for C e n A and a mass o f 49623 D a was obtained for C e x (Figure 3.15). T h i s information was used to determine the number o f moles o f sugars on the proteins by 82 comparing the masses o f the glycosylated proteins with their non-glycosylated recombinant counterparts. C e n A and C e x had 15-19 and 16 moles o f carbohydrate per mole o f protein, respectively. T h e phenol sulfuric acid assay, outl ined in Section 2.14.2, was also used to quantify sugars. T h e moles o f carbohydrate measured per mole o f protein were 29 and 30 for C e n A and C e x produced in C. fimi, respectively. T h e M A L D I - T O F M S data are considered the most accurate. The masses o f C e n A I g A l h produced by S. lividans and E. coli were determined by S E L D I M S , as described in Section 2.9.4, to be 46637 and 45024 D a , respectively (data not shown). T h i s indicated the presence o f approximately 7 moles o f sugar per mole o f protein. A a 4 8 5 1 9 B 4 9 6 2 3 b 4 7 8 7 4 Figure 3.15: M A L D I - T O F mass spectra o f C e n A and C e x produced by C . fimi. C e n A (Panel A : a and b) and C e x (Panel B ) . Masses are shown in D a . 83 3.5.3. Location of carbohydrates on CenA and Cex produced in C.fimi and CenA constructs produced in S. lividans. T o determine the location o f the glycans on C e n A and C e x produced in C. fimi, H P L C profiles o f tryptic digests o f native and recombinant cellulases were compared as detailed in Section 2.14.6. Potential trypsin cleavage sites indicated that the P T linker should be contained within a single peptide for both C e n A and C e x . T h e tryptic digests o f the native and recombinant cellulases gave very similar H P L C profiles (data not shown). F o r C e n A and C e x produced in C. fimi, one main peptide bound to C o n A (labeled 2 in Figure 3.16 and 1 in Figure 3.17). T h e peak labeled 2 in Figure 3.16 is undigested C e n A . The retention times o f these peptides matched those predicted by the computer program P C gene for the peptides containing the P T linker. Th is was conf i rmed by N-terminal sequencing o f the major retained peaks. M i n o r peaks in both profi les were not sequenced. F o r C e n A , the peptide containing the P T linker is 108 amino acids long and contains 19 amino acids o f the catalytic domain and 62 amino acids o f the C B D . T h e 19 amino acids o f the catalytic domain include 4 threonines and 1 serine residue as potential O -glycosylat ion sites and no potential N- l inked sites; the 62 amino acids o f the C B D include 10 threonines and 9 serines and 5 potential N - l inked sites. T h e remaining C B D sequences are found in peptides o f 5, 10 and 30 amino acids in size. The smaller ones m a y have been hard to detect, but the larger one should have been apparent i f glycosylated. F o r C e x , the peptide containing the P T linker is 60 amino acids long, consisting o f 15 amino acids o f the catalytic domain with no potential O - or N- l inked glycosylat ion sites, and 26 amino acids o f the C B D with 3 threonines residues as potential O- l inked sites and 1 potential N -84 l inked site. T h e remainder o f the C e x C B D is found on 2 peptides o f 40 and 43 amino acids, respectively. If these peptides were glycosylated, they wou ld have been concentrated on the C o n A co lumn to a similar level as the one containing the P T linker, but no other peaks concentrated to that level. 85 0.08 0.04 0 0.08 0.04 0 0 25 Retention Time (minutes) 50 60 30 60 30 Figure 3.16: Reverse phase H P L C profiles o f a tryptic digest o f C e n A produced by C. fimi. Acetonitr i le was used as Solvent B . T h e peptide profi le prior to C o n A affinity chromatography (Panel A ) and the peptides which were retained b y C o n A (Panel B ) . N -terminal sequences were determined for the numbered peaks. 86 0.08 0.04 GO 3 Q 0 0.08 0.04 0 60 30 60 PQ GO : 30 0 25 50 Retention Time (minutes) Figure 3.17: Reverse phase H P L C profiles o f a tryptic digest o f C e x produced by C. fimi. Acetonitr i le was used as Solvent B . T h e peptide profi le prior to C o n A affinity chromatography (Panel A ) and the peptides wh ich were retained by C o n A (Panel B ) . The N-terminal sequence was determined for the numbered peak. 87 Papain digests o f C e n A and C e x produced by C. fimi, as described in Section 2.12.1, fo l lowed b y periodic ac id /Schi f f staining conf i rmed that no carbohydrates were located on the C B D o f either glycoprotein (Figure 3.18). T h e proteins produced in C.fimi were not cleaved efficiently by papain. However , the C B D s were just distinguishable in the gel stained with Coomass ie -B lue (Figure 3.18 Panel A ) , but d id not react with the more sensitive periodic ac id /Schi f f stain (Figure 3.18 Panel B ) . N-terminal amino acid sequencing o f papain hydrolysis products and degradation products o f C e n A and C e x produced in C. fimi, respectively, gave an indication o f some glycosylated sites in the P T linker (Figure 3.19). F o r the proteins produced in S. lividans, it became apparent that S. lividans does not glycosylate the catalytic domain or C B D as C e n A F n 3 was not glycosylated (Figure 3.13). 88 kDa 1 2 3 4 5 Figure 3.18: S D S - P A G E gel and periodic ac id /Schi f f stained Western blot o f C e n A and C e x produced by C. fimi, and treated with papain. S D S - P A G E gel stained with Coomass ie -B lue (Panel A ) and per iodic ac id /Sch i f f stained Western blot (Panel B ) . C e n A untreated (Lane 1 ) , C e n A treated with papain (Lane 2), C B D c e x produced in E. coli (Lane 3), C e x treated with papain(Lane 4 ) and C e x untreated (Lane 5). CenA • T T T T _ T V P T T S P T P T P T P T T P T P T P T P T P T P T P T V T I Cex i | G A S P T P T P T T P T P T P T T P T P T P T S G P [ CBD ] catalytic domain * glycosylation site Figure 3.19: G lycosy la t ion sites on the linker o f C e n A and C e x produced by C.fimi. The underl ined sequences belong to the domains adjacent to the l inker and the bo ld sequence was N-terminal sequenced. 89 3.5.4. The monosaccharide composition and linkage positions between sugars of CenA and Cex produced by C. fimi T h e types o f monosaccharides o f the glycan component o f C e n A and C e x produced b y C . fimi were determined b y G L C - M S as described i n Sect ion 2.14.3. M a n n o s e was the major sugar detected. A smal l amount o f glucose was also detected but no galactose (Figure 3.20). G lucose is regarded as a c o m m o n contaminant in sugar analysis reactions. Therefore, mannose is the sole sugar on these glycoproteins. Th is was conf i rmed later by F A C E ® (Section 3.5.5). T o gain information about the anomeric configuration ( a , P) and the l inkage positions between sugars (1-2,3,4, or 6) o f native C e n A and C e x , treatment with exo-glycosidases was employed as described in Section 2.14.3: a -mannosidase (Jack bean meal) (specificity a-man-1-2,3,6) , and a-galactosidase (green coffee beans) (specificity a -gal-1-3,4,6) (Figure 3.21). Treatment with a-mannosidase reduced the M r o f C e n A to that o f recombinant C e n A produced in E. coli and abolished its reactivity with the periodic ac id /Schi f f stain. T h i s conf i rmed the presence o f a -D -mannose . A glycosylated C e n A species o f partially reduced M r remained that reacted with C o n A - H R P (Figure 3.22). The M r o f C e x was partially reduced by a-mannosidase. Th is product reacted with both the periodic ac id /Sch i f f stain and C o n A - H R P (Figures 3.21 and 3.22). a-galactosidase had no effect on the M r o f C e n A or C e x , even fo l lowing pretreatment with a-mannosidase. Deglycosylat ion o f papain hydrolysis products o f C e n A and C e x , w h i c h lack the C B D , was tried because the P T linker o f these peptides should be more accessible to the exo-glycosidases. Tota l deglycosylat ion was still not attained, however (data not shown). 90 Jack bean a-mannosidase cleaves mannose l inked a-1-2 ,3 , or 6. F r o m the digests and the information regarding the specificity o f the mannosidase it is evident that C e n A consists o f a heterogeneous population o f glycoproteins some o f wh ich contain only mannose l inked a-1-2,3,6. A l l the C e n A and C e x glycoforms contain terminal a -mannose. The protein-carbohydrate linkages in C e x and in some glycoforms o f C e n A remain to be elucidated. 91 Figure 3.20: Determination o f carbohydrate composit ion by G L C o f C e n A and C e x produced by C.fimi. Standards (Panel A ) ; fucose (F), mannose (M) , galactose (G) , glucose (Glc ) , Inositol (I), N - A c e t y l galactosamine ( G a l N A c ) , N-Acety lg lucosamine ( G l c N A c ) ; C e x (Panel B ) ; C e n A (Panel C ) . Retention time M M JUul Glc Glc ,1 .. 1 . JL Retention time 93 A 1 2 3 4 5 6 7 8 Figure 3.21: S D S - P A G E gel and periodic ac id /Schi f f stained Western blot o f C e n A and C e x produced by C. fimi, and treated with exo-glycosidases. Coomass ie -B lue stained S D S - P A G E gel (top) and periodic ac id /Sch i f f stained Western blot (bottom) o f C e n A (Panel A ) and C e x (Panel B ) produced by C . fimi and treated with exo-glycosidases: a-mannosidase (Jack bean meal) (Lane 4), a -mannosidase and then a -galactosidase (green coffee beans) (Lane 5), a-galactosidase (Lane 6). Controls: Mo lecu la r weight markers in k D a (Lane 1), a -mannosidase (Lane 2), non-glycosylated cellulase produced in E. coli (Lane 3), glycosylated C e n A (Panel A ) or C e x (Panel B ) produced by C.fimi (Lane 7), a-galactosidase (Lane 8). 94 1 2 3 4 5 6 7 8 9 Figure 3. 22: Western blot treated with C o n A - H R P to detect the presence o f mannose on C e x and C e n A produced by C. fimi, and treated with exo-glycosidases. C e x (Lanes 2-5) and C e n A (Lanes 6-9); exo-glycosidases: a -mannosidase f rom J a c k bean meal (Lane 3 and 8), and a-mannosidase and then a-galactosidase f rom green coffee beans (Lane 4 and 7). Controls: Molecu lar weight markers in k D a (Lane 1), untreated C e x and C e n A produced by C.fimi (Lanes 2 and 9) , untreated C e x and C e n A produced in E. coli (Lanes 5 and 6). 3.5.5. The monosaccharide composition of the glycans of CenA mutants produced by S. lividans and CenA produced by C. fimi T h e analysis was done using the F A C E ® Monosacchar ide composi t ion kit f rom G l y k o as described in Section 2.14.3. Figure 3.23 shows the F A C E ® Monosacchar ide composi t ion gel. G lucose was detected in the distil led water sample and in the sample o f C e n A produced in E. coli so was regarded as a contaminant in all cases. C e n A produced by C. fimi contained on ly mannose. The C e n A mutants produced b y S. lividans ( C e n A I g A l h , C e n A 2 P T , and C e n A A P T ) all had mannose l ike C e n A f rom C.fimi but also contained galactose. F r o m the intensity o f the bands there appeared to be more mannose than galactose. T h e intensity o f the signal increased with an increase o f potential glycosylat ion sites, with C e n A A P T < C e n A I g A l h < C e n A 2 P T . Quant i f icat ion is 95 theoretically possible using the kit but, due to the number o f steps invo lved , it was decided it was l ikely to be unreliable. 3.5.6. Oligosaccharide composition and linkage positions between sugars of CenA2PT produced in S. lividans The F A C E ® O- l inked oligosaccharide prof i l ing kit was uti l ized to determine the size and linkages between the sugars o f the oligosaccharides o f C e n A 2 P T produced in S. lividans as detailed in Section 2.14.5. F igure 3.24 shows the F A C E ® oligosaccharide analysis gels. T h e mannobiose standards, wh ich only differed in the l inkage between carbohydrate moieties, migrated to different positions in the gel (Figure 3.24 Panel A , Lanes 1, 3-6). However , substituting one neutral sugar, galactose, for another, glucose, did not affect migrat ion (Figure 3.24 Panel A , Lanes 7-8). The anomeric configuration also appears to affect migration in the case o f cel lobiose ( p i - 4 glucose) wh ich did not migrate at the same level as the starch ( a 1-4 glucose ) biose standard on the hydrolysis ladder (Figure 3.24 Panel A , Lanes 8-9). T h e oligosaccharide profi le o f C e n A 2 P T produced in S. lividans is shown in F igure 3.24 Panel B . U s i n g the relative mobi l i ty o f the mannobiose and triose standards to the starch hydrolysis ladder; a standard curve generated f rom the biose standards; and the relative mobi l i ty o f cellobiose to the starch hydrolysis ladder, the size o f the oligosaccharides and the type o f l inkage in the disaccharide were determined. C e n A 2 P T had single neutral sugars; biose units l inked most ly a 1-2, and some a 1-3; two triose units; and a tetrose unit o f undetermined linkage. S. lividans attached both mannose and galactose. Therefore, wh ich carbohydrates were involved in the l inkages could not be ascertained. Figure 3.23: F A C E ® Monosacchar ide composit ion gel o f C e n A produced b y C. fimi and a subset o f C e n A mutants produced by S. lividans. Panel A : C e n A produced by E. coli (Lane 1), C e n A produced b y C.fimi (Lane 2), and water (Lane3); Panel B : C e n A I g A l h (Lane 1), C e n A 2 P T (Lane2), and C e n A A P T (Lane 3) all produced in S. lividans. Monosacchar ide standards are indicated to the left o f the gel. 97 Figure 3.24: F A C E Oligosaccharide analysis gel o f C e n A 2 P T produced b y S. lividans. Panel A : Standards: a - 1 - 3 , 1 - 6 mannotriose (Lane l ) , a -1 -6 mannobiose (Lane3), a -1 -4 mannobiose (Lane4), a -1 -3 mannobiose (Lane5), a -1-2 mannobiose (Lane6), lactose (galactose-P-1-4 glucose) (Lane7), cellobiose (glucose-P-1-4 glucose) (Lane8), Panel B : Standards: a -1-2 mannobiose (Lane3), a-1-6 mannobiose ( L a n e 4 ) , S g C e n A 2 P T (Lane 2). Starch hydrolysis glucose standard ladder (Panel A : Lanes 2 and 9; Panel B : Lane 1). The number o f a -1 -4 l inked glucose units is indicated to the left o f the panels. 98 3.6 . Enzyme activity 3.6.1. Protein stability Protein stability was determined under assay conditions by compar ing the C M C a s e activity, as described in Section 2.11.4, o f the enzymes sampled at 0 and 24 hours from a m o c k B M C C hydrolysis experiment, where the assay was set up without the B M C C and the samples were not boi led. N o loss o f activity was detected (data not shown). 3.6.2. Effect of method on BMCC hydrolysis by CenA B M C C and P A S C are both insoluble substrates. T h e effect o f agitation on activity was analyzed for B M C C hydrolysis. Methods used included: no agitation, m i x i n g on a tube roller, shaking on platform shakers at 250 and 300 rpm, mix ing using an end over end l a b q u a k e ™ apparatus and agitation with a Pierce magnetic stirring apparatus. Earl ier studies ( data not shown) indicated that the presence o f B S A had no effect on the activity o f C e n A on C M C . Therefore, in this study reactions set up in parallel with and without B S A were compared as a means o f comparing the different agitation methods. F r o m Table 3.3, it is apparent that no agitation, mix ing on the tube roller and shaking at 300 rpm all gave comparable readings between samples with and without B S A . T h e other three methods had larger variation between samples. T h e method o f agitation also affected the absolute activity measured, with end over end mix ing g iv ing the lowest activity reading, and the Pierce magnetic stirring apparatus the greatest. N o stirring was chosen instead o f agitation for the activity studies because the method was simple and the substrate stayed well dispersed in suspension over the time per iod o f the assay. A l l the other agitation 99 methods resulted in the substrate aggregating into clumps. W h e n evaluating samples set up in triplicate later, this method proved reliable (data not shown). Table 3. 3: Summary o f agitation methods studied for the hydrolysis o f B M C C . A l l values are presented as pmole glucose pmole enzyme" 1 min" 1 . AGITATION METHOD CenA -BSA CenA + BSA no agitation .248 .214 tube roller .089 .095 shaker (250 rpm) .117 .161 shaker (300 rpm) .299 .306 end over end labquake^ M apparatus .084 .050 PIERCE magnetic stirring apparatus .496 .913 3.6.3. BMCC hydrolysis Trials were performed with samples set up in triplicate. Substrate (1 mg) and enzyme (0.15 nmole) were combined, m ixed by vortexing and placed at 37° C without further agitation for 24 hours. Sampl ing and analysis were as described in Section 2.11.2. C e n A produced in E. coli was always included as a reference. Results are presented in Table 3.4. T h e activities are presented as relative activities - ratios o f the activity o f C e n A mutant "x" to the activity o f C e n A produced in E. coli - because the ratios were reproducible but the absolute values varied between trials. Paired t-tests were performed on the ratios as described in Section 2.11.8. A P value o f less than 0.05 is generally taken 100 to be significant (Motulsky, 1995). The P value for this test reflects the p robab i l i t y that the means are the same. Thus a P value o f 0.004 for a ratio o f 1.22, as for C e n A F n 3 produced in E. coli, indicates a 0.4% chance that C e n A and C e n A F n 3 have the same activity. M o d i f y i n g the length o f the l inker o f non-glycosylated C e n A had a modest effect on its activity on B M C C . R e m o v i n g the linker and C B D , or shortening the linker decreased the activity: p30 had 4 8 % o f the activity o f w i l d type C e n A and C e n A A P T had 74 % . Increasing the length o f the linker by 150 % , as for C e n A 1 . 5 P T ; or modi fy ing the composit ion min ima l ly whi le retaining a similar length l inker as C e n A , exhibited by C e n A I g A l h , had no effect on activity. Doub l ing the linker length or replacing the linker with the Fn3 repeats increased activity: C e n A 2 P T had a 9 % increase, whi le C e n A F n 3 had a 2 0 % increase. T h e effect o f glycosylat ion o f C e n A and C e n A linker mutants on their activity on B M C C was inconsistent. T h e glycosylated variants o f C e n A and C e n A 2 P T had increased activities o f 26 % and 20 % , respectively. In contrast, the glycosylated variants o f C e n A A P T and C e n A I g A l h had decreased activities: 8 9 % and 9 1 % , respectively, o f their non-glycosylated counterparts. B o t h C e n A 2 P T and C e n A I g A l h had P values above 0.05. 101 Table 3. 4: Summary o f the hydrolysis o f B M C C by C e n A and C e n A constructs produced in E. coli and S. lividans. Act iv i ty is presented as a relative activity: the mean ratio (calculated for the number o f trials indicated in the table) o f the activity o f C e n A mutant "x" to the activity o f C e n A produced b y E. coli. E a c h trial consisted o f samples set up in triplicate. ( S D = standard deviation; na = not applicable) Protein Organism protein produced in Relative Activity (mean ratio) SD of mean ratio Paired t-test two tailed P-value Number of trials CenA E. coli 1.00 na na na S. lividans 1.26 0.006 0.01 2 CenAAPT E. coli 0.74 0.01 2 x 10-7 5 S. lividans 0.55 0.07 0.01 3 CenAIgAlh E. coli 0.95 0.17 0.50 4 S. lividans 0.84 0.10 0.12 3 CenA1.5PT E. coli 1.03 0.09 0.53 4 CenA2PT E. coli 1.09 0.06 0.03 5 S. lividans 1.29 0.22 0.12 3 CenAFn3 E. coli 1.22 0.06 0.004 4 S. lividans 1.20 0.06 0.03 3 p30 E. coli 0.48 na na 1 102 3.6.4. PASC hydrolysis Trials were performed with samples set up in triplicate. 10 m g o f substrate and 0.15 nmole enzyme were combined, mixed by vortexing and placed at 37° C without further agitation for 24 hours. Sampl ing and analysis were as described in Section 2.11.3. C e n A produced in E. coli was always included as a reference. Results are presented in Table 3.5. T h e activities are presented as ratios to C e n A produced by E. coli as for B M C C hydrolysis (Section 3.6.3). Statistical analysis was as described in Sections 2.11.8 and 3.6.3. C e n A A P T was the only mutant to show a significant difference f rom C e n A with a decrease o f 13%. p30 had 3 3 % more activity in the single trial performed. The other constructs had increases up to 7% but all had h igh P values. G lycosy la t ion had little or no effect on the activity o f the proteins, with the greatest difference being for C e n A I g A l h which showed a decrease o f 13% upon glycosylat ion, but wi th a h igh P value. Shortening the linker or removing the C B D were the only modif icat ions to have notable effects on the hydrolysis o f P A S C . 103 Table 3. 5: Summary o f the hydrolysis o f P A S C by C e n A and C e n A constructs produced in E. coli and S. lividans. Act iv i ty is presented as a relative activity: the mean ratio (calculated for the number o f trials indicated in the table) o f the activity o f C e n A mutant "x" to the activity o f C e n A produced by E. coli. E a c h trial consisted o f samples set up in triplicate. ( S D = standard deviation; na = not applicable) Protein Organism protein produced in Relative Activity (mean ratio) SD of mean ratio Paired t-test two tailed P value Number of trials CenA E. coli 1.00 na na na S. lividans 1.00 0.03 0.78 2 CenAAPT E. coli 0.87 0.05 0.02 4 S. lividans 0.92 0.03 0.05 3 CenAIgAlh E. coli 1.03 0.18 0.84 3 S. lividans 0.90 0.06 0.12 3 CenA1.5PT E. coli 1.07 0.14 0.37 4 CenA2PT E. coli 1.03 0.09 0.61 4 S. lividans 1.06 0.11 0.44 3 CenAFn3 E. coli 1.00 0.07 0.99 4 S. lividans 1.02 0.02 0.17 3 p30 E. coli 1.33 na na 1 104 3.6.5. 2,4-DNPC hydrolysis 2 , 4 - D N P C hydrolysis was carried out as described in Section 2.11.5. A concentration o f approximately 10 x K,„ for 2 , 4 - D N P C was used in the assays and the assays were done at p H 6.5. A single trial was carried out with triplicate samples. Protein concentration determination was performed on the diluted enzyme samples to get the most accurate protein concentration readings for calculations. T h e results are depicted graphically in Figures 3.25 and 3.26. C e n A F n 3 activity was 7 7 % that o f C e n A , while p30 was 117% that o f C e n A . A l l the other enzymes had activities similar to that o f C e n A . Glycosy la ted proteins all had decreased activity when compared to their non-glycosylated counterparts, although the level o f the decrease varied, being largest for C e n A I g A l h (45%). There was a negative correlation between the size o f the protein and the activity on 2 , 4 - D N P C . 3.6.6. pH profiles p H profiles o f the enzymes were compi led f rom 2-4 D N P C hydrolysis reactions over a p H range o f 4.5 - 8.0 at 0.5 p H unit intervals as described in Sect ion 2.11.6. A concentration o f 2 , 4 - D N P C o f 5 x K,,, was used. A l l the proteins showed the same profi le with m a x i m u m activity at p H 6.5(Figure 3.27). 105 5^ 0) O U a. "o '-B •a 2J o o 3 u < 0 o D . CO •4—* 0 •a < 0 O < a> < c 0 O Q. < 0 O CL 5! c 0 O c L L < c 0 O o CO B a "S O o 1- ° -a S o © o 3 < c 0 O < c 0 O O) co CL as 0 < c 0 O O) to < < c 0 O D) CO 5! c 0 O O) CO co c LL < c 0 O D) CO Figure 3.25: 2 , 4 - D N P C hydrolysis at p H 6.5 by C e n A linker constructs produced in E. coli and S. lividans. E. coli (Panel A ) and S. lividans (denoted by the prefix sg)(Panel B ) 106 5n O u a, © "3 o 3 2J < c O u -a < c u U U i l l S a l < c OJ O fts - i a. CL, < m (N 0 CL, < C <L> g OB o") | o U c u. I g 5b Figure 3.26: Compos i te figure o f 2 , 4 - D N P C hydrolysis at p H 6.5 b y C e n A linker constructs produced in E. coli and S. lividans. Proteins produced in S. lividans denoted by the prefix sg Figure 3.27: p H profi les o f 2 , 4 - D N P C hydrolysis b y C e n A and C e n A constructs produced in E. coli and S. lividans. C e n A (1); s g C e n A (2); C e n A A P T (3); C e n A A P T (4); C e n A I g A l h (5); s g C e n A I g A l h (6); C e n A 1 . 5 P T (7); C e n A 2 P T (8); s g C e n A 2 P T (9); C e n A F n 3 (10); s g C e n A F n 3 (11); p30 (12); proteins produced in S. lividans are denoted with the prefix sg. 107 3.6.7. Enzyme assay products T h e soluble products o f B M C C hydrolysis by C e n A and C e n A F n 3 produced in E. coli were analyzed as outlined in Section 2.11.7. These constructs were chosen because they had the greatest difference in activity. There was no difference in the soluble products formed (data not shown). 3.7. Proteolysis 3.7.1. Papain hydrolysis of CenA constructs produced in E. coli Papain was chosen for this study because it was known to cleave in the linker sequence o f C e n A and C e x produced by E. coli (Gi lkes et al., 1991c). Papain hydrolysis was carried out and analyzed as described in Section 2.12.1. Figures 3.28 and 3.29 show the S D S - P A G E gels o f papain digests o f C e n A constructs made by E. coli. Sites o f hydrolysis were determined by N-terminal sequencing o f selected products. These sites are detailed in Figure 3.30. C e n A F n 3 was most resistant to papain over a two hour time period, and the hinge region o f I g A l in the construct C e n A I g A l h was as susceptible as the P T linker. C e n A and the modi f ied C e n A - l i k e P T linkers yielded various products during the digestion. C e n A itself yielded three products. T w o transient products resulted f rom cleavage near the N-terminus o f the linker at the site P T * T S P , where * indicates the cleavage site, and from cleavage further into the linker, at the site P T * T P (confirmed by electrospray M S ) . A final stable product resulted f rom cleavage near the C-terminus o f the linker at the site P T V * T P . A l l the modi f ied C e n A - l i n k e r constructs were cleaved in these same sequences, even when they appeared multiple times as in C e n A 2 P T or when they 108 appeared in the middle o f the linker as in C e n A l . 5PT and C e n A 2 P T . W h e n the cleavage sites in C e n A I g A l h and C e n A F n 3 were examined, a pattern o f susceptible sites was discernible. T h e sequences cleaved in C e n A I g A l h were P T T S V S T P , P S * T P and P L Q * S N * V T P ; and in C e n A F n 3 were P T G T T T D T T P , V T F * T T D T T G E T E P , P V T S * T P and P S * V T P . T h e first two cleavage sites in C e n A F n 3 were determined f rom N-terminal sequencing o f the S. lividans produced protein. S. lividans does not glycosylate C e n A F n 3 so the sites were assumed to be the same for C e n A F n 3 produced in E. coli. In most o f the proteins, the cleavage sites were located in the consensus sequence P ( T / S / V ) , . 3 * ( T / S / V ) i . 2 P . T h e location o f the sequence d id not have to be near the termini o f what is def ined in this study as the P T linker. sample times sample times 0' 10' 20' 30' 40' SO' 60' 90' 120' 24h H 0' 10' 20' 30' 40' 50' 60' 90' 120' 24h Figure 3. 28: Papain digestion o f C e n A and C e n A P T constructs produced in E. coli. The numbered arrows indicate those products wh ich were N-terminal sequenced and are shown in Figure 3.30. T h e S D S - P A G E gels were stained with C o o m a s s i e - B l u e . 109 sample times 0' 10' 20' 30' 40' 50' 60' 90' 120' 24h Figure 3. 29: Expanded gel o f a papain digestion o f C e n A produced b y E. coli. The arrows indicate products which were N-terminal sequenced and are shown in F igure 3.30. T h e S D S - P A G E gel was stained with Coomass ie -B lue . 110 -• • CN >! H Cu H H CM EH Cu E H CM E H Cu E H ' E H Cu EH Cu £ H Cu E H > o CN • -p->\ > EH Cu H CU EH Cu E H Cu E H Cu H Cu EH EH Cu H • > EH Cu co E H " E H 04 E H Cu E H Cu H Cu E H Cu B CM E H Cu EH CM > | E H I EH ft IT) u -Si EH Cu E H CM EH CM EH CM EH CM EH CM Cu E - i Cu E H CM E H CM CO E H • EH CM EH Cu CU EH CU E H Cu H EH CM E H CM E H CM CO H " E H CM > | EH I • co E-i > CM H E H U J H > Cu CM EH o > fH 0 . rf > Q > < EH E H Cu EH UJ Cu E H E H Cu CU W H td O E H EH Q EH • > D » | • H U Cu EH E H Cu Cu H E H Q E H E H E H u E H Cu > | H I CN 1 • E H | CO Cu CO Cu E - i Cu Cu E H •oo E H Cu EH E H CM > | H d) U > u c cu 3 tr cu •H 4 o TJ a 10 e o T 3 10 U • Q CQ o 0 s CU -a 3 £ H OH u U U 00 CU CU 3° 3 U "o 3 OH © u S-6D CU CU ON CU C N j=i 2 ° ' O -4-> 3 c oo 8 CN cd cn ^ oo cd <U oo 3 | 00 3 '£ a .a % oo cu 13 -6 DO U c£ T3 CU u cu 0! 3 _o CU oo CU cu 3 cu 3 a 12 CU 3 c cu u •fl ^ 3 CU 3 "3 ju 3 . •rt 00 e Si i a 1 * 3 g cu 3 • S 00 l l cu > o -° ^ H "9 o u / — N 3 oo CJ .-3. g 00 •—I cu .2 3 ^ CU CU .5 <H-H oo © p I o CO I l l 3.7.2. C. fimi protease hydrolysis of CenA constructs produced in E. coli C fimi protease was used in this study because it cleaves in the l inker sequence o f C e n A and C e x produced by E. coli and C. fimi (Gi lkes et al., 1989). C . fimi protease hydrolysis was carried out and analyzed as described in Section 2.12.2. F igure 3.31 shows the S D S - P A G E gels o f the C e n A constructs made by E. coli digested with C.fimi protease. Sites o f hydrolysis were determined by N-terminal sequencing o f selected products. These sites are detailed in Figure 3.32. T h e papain sites are noted for comparison. T h e sites o f cleavage by C. fimi protease were very similar to those o f papain. C e n A F n 3 was most resistant to papain over a two hour time period but was as susceptible as the C e n A - l i k e linkers to C.fimi protease. The hinge region o f I g A l in C e n A I g A l h was as susceptible as the P T linker to C.fimi protease. However , the native I g A l hinge was not susceptible to C. fimi protease. Proteolysis or degradation d id occur at a location other than the hinge by 5 hours (Figure 3.33). Figure 3. 31: C. fimi protease digestion o f C e n A and C e n A P T constructs made by E. coli. The arrows point to sequenced products which are shown in F igure 3.32. T h e S D S -P A G E gels were stained with Coomass ie -B lue . 113 114 kDa 0 s a m p l e t i m e s ( h r ) 1 5 24 48 ! 96 45 31 6 6 Figure 3. 33: C. fimi protease digestion o f I g A l . The arrows indicate the approximate molecular weights o f the products i f the hinge region had been cleaved. The S D S - P A G E gel was stained with C o o m a s s i e - B l u e . 3.7.3. Effect on papain hydrolysis of changing valine 176 to proline T h e mutant C e n A 1 . 5 P T m o d was made to explore the possibi l i ty o f el iminating the C-terminal susceptible hydrolysis site in C e n A 1 . 5 P T b y changing val ine 135 to proline. The internal sites were maintained to test the activity o f the protease. Papain hydrolysis was carried out and analyzed as described in Section 2.12.1. F igure 3.34 shows the S D S -P A G E gels o f the C e n A constructs made by E. coli and digested with papain. Hydro lys is sites, detailed in Figure 3.35, were determined by N-terminal sequencing. T w o hydrolysis products were seen for C e n A l . 5 P T m o d and three for C e n A l .5PT. T h e first two products were identical for both constructs. The second cleavage site at the sequence P T * T P was conf irmed by electrospray M S . C e n A l . 5PTmod yielded no further products whereas C e n A l . 5PT was cleaved once more near the C-terminus o f the P T linker. Chang ing the sequence o f a site susceptible to cleavage by papain can abolish sensitivity to papain. T h e final C.fimi protease cleavage site was also abolished in C e n A l . 5 P T m o d (data not shown). sample times (hr) 0 1 5 24 48 B sample times 0' 10' 20' 30' 40' 50' 60' 90' 120' 24h Figure 3.34: C. fimi protease digestion (Panel A ) and papain digestion o f C e n A l .5PT and C e n A l . 5 P T m o d made by E. coli. T h e sequenced products are numbered and are shown in figure 3. 23 • • > l EH D J E-i Dj EH Oi E-i CU a, D J EH D J EH EH D J EH D J EH D J EH D J CO E H E H D J EH D J EH D J EH D J EH D J EH D J EH D J EH D J > l EH H Cu < CU U EH D J EH D J EH D J EH D J EH D J EH D J EH D J EH EH D J EH D J EH D J EH D J CO EH EH D J EH D J EH D J EH D J EH D J EH D J EH D J EH D J > l EH -a o S H IT) < a ( U U 4» > a a a cs S © >> "a CJ • o o T3 2 a. •a o a H PH i t ) U •a H PH u .a go ii Cu) 00 (D J-» s OH S T3 cd PH co co a o a T 3 <D X O ^ 3 <L> Xi H co « "I & i * ? o a <D O c3 ^ H <D PQ cu s-s o n 3 117 3.7.4. Effect of glycosylation of CenA constructs on susceptibility to papain and C. fimi protease Papain and C. fimi protease digests were carried out and analyzed as described in Section 2.12.1 and 2.12.2, respectively. Figure 3.36 and 3.37 show the SDS-PAGE gels of the CenA constructs made by E. coli and S. lividans digested with papain and C.fimi protease, respectively. Hydrolysis sites of selected products were determined by N-terminal sequencing and are detailed in Figure 3.38. Only a subset of the glycosylated products was analyzed because the patterns appeared to be the same as that seen for their non-glycosylated counterparts. Of the products sequenced, it became apparent that the cleavage of the glycosylated and non-glycosylated products occurred in the same sequences; although not always between the same amino acids. Digestion also took longer. Thus glycosylation serves to slow down the proteolysis by these two proteases. 118 sample times sample times 0' 10' 20' 30'40' 50' 60'90' 2h 24h 0' 10' 20' 30'40' 50' 60'90'2h 24h fi E.coli S. lividans E.coli S. lividans < fi U • 5. lividans Figure 3.36: Papain digestion of CenA and CenA PT mutants made by E. coli and S. lividans. Arrows point to the sequenced products which are shown in figure 3.38. The SDS-PAGE gels are stained with Coomassie-Blue. 119 Figure 3. 37: C.fimi protease digestion of CenA and CenA PT constructs made by E. coli and S. lividans. Arrows point to the sequenced products which are shown in Figure 3.38. The SDS-PAGE gels were stained with Coomassie-Blue. 120 • EH Cu H Cu H Cu H Cu EH CU EH Cu EH EH Cu EH CM EH • • cu H CU EH Cu EH DJ H CU EH Cu EH Cu EH Cu H H Cu EH cu EH CU EH EH I 0 4 H ID > CU > EH Q > Cu cu tt rt EH CM EH e> cu EH EH CU Cu a EH U O H H D EH EH EH| IU CO Cu H CM Cu EH ' c o Cu CO CU H CU EH H Cu > | EH I ns •—- •a •H > 0 •H u H CO —' a) • Oi rO rO > > 10 10 CO 0) H 1—1 O o 0) u to m H! m 0> 0 4-> 4 J o O u M a a •H •H E= E H •H CH t u cj O u >1 - rH a Cn tu Cn ro rO > > r0 rO cu OJ HH HH u u c c -H -H f0 ro a CC f0 rO Cu Cu — • • c •H rO e o •a t u o •a u •H • P >1 i—I r0 - P rO u • C •9 CO 13 o c tu T 3 i =3 OH U U I <D U _g CO UJ SP J U O u &o cd tu o 13 cd >, X ) -a tl) o d CO c 5 CO o 5J C5 tu td H cn X ) co •9 HI -a o CD d cu rd cd -rt <u cd tu 0) o Q CQ O OO v s-3 DJD 121 3.8. IgAl protease digestion of CenA and CenAIgAlh In a previous study, CenA and CenAIgAlh were tested as substrates for type 1 and type 2 IgAl proteases from N. gonorrhoeae (Miller et ah, 1992). Despite the sequence similarity between the IgAl hinge and the CenA PT linker, only the CenAIgAlh construct was cleaved. It was of interest to expand on this experiment by utilizing both non-glycosylated and glycosylated CenA and CenAIgAlh as substrates for a broader selection of IgAl proteases, as the native IgAl hinge is glycosylated. IgAl protease digestions were carried out and analyzed as detailed in Section 2.13. Figure 3.39 shows the SDS-PAGE gels of the digests of CenA and CenAIgAlh, produced by E. coli and S. lividans, by recombinant H. influenzae type 1, and impure preparations of N. meningitidis type 1, N. meningitidis type 2, N. gonorrhoeae type 2 and S. pneumoniae IgAl proteases. The results are summarized in Table 3.6. Glycosylated substrates were cleaved less efficiently than their non-glycosylated counterparts and all were cleaved much less efficiently than native IgAl. The H. influenzae type 1 IgAl protease had the broadest activity, slowly cleaving all substrates. N. meningitidis IgAl proteases cleaved both non-glycosylated substrates. The N. gonorrhoeae type 2 and S. pneumoniae IgAl proteases only cleaved non-glycosylated CenAIgAlh. The N. meningitidis type 1 enzyme cleaved the CenA PT linker though it doesn't have the sequence that is cleaved in IgAl. The CenA linker has no P-S bonds. Of the two type 2 enzymes, only the N. meningitidis enzyme cleaved CenA while both cleaved CenAIgAlh. 122 Figure 3.39: I g A l protease digests o f C e n A I g A l h and C e n A produced by E. coli and S. lividans. Digests o f C e n A I g A l h ( A Panels) and C e n A ( B Panels ) by the I g A l proteases noted above the gels. Lanes 1 , 2 , 5 and 6 show glycosylated proteins made in S. lividans and Lanes 3, 4, 7 and 8 show the non-glycosylated proteins produced in S. lividans. O d d lanes are time = 0 and even lanes are time = 4 days. H. influenzae A B 1 2 3 4 5 6 7 8 N. meningitidis type 1 N. meningitidis type 2 A B A B 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 N. gonorrhoeae type 2 A B 1 2 3 4 5 6 7 8 S. pneumoniae A B 1 2 3 4 5 6 7 8 124 Table 3. 6: Susceptibil i ty o f C e n A substrates to I g A l proteases. +++ = complete cleavage in 4 days; +/- = minor cleavage by 7 days; - = no cleavage. ENZYMES CenA g'yc SUBSTRATES CenA CenAIgAlh "g g'yc CenAIgAlh ng H. influenzae type 1 + + + + N. meningitidis type 1 - ++ +/- +++ N. meningitidis type 2 - ++ +/- +++ N. gonorrhoeae type 2 - - - ++ S. pneumoniae - - - + 125 3.9. A comparison of the glycosylation of the native IgAl hinge and that produced by S. lividans The monosaccharide composit ion o f the hinge region differed between C e n A I g A l h produced in S. lividans and the native hinge. C e n A I g A l h had mannose and galactose (Figure 3.23) whi le the human m y e l o m a serum I g A l hinge was reported to have G a l N A c and G a l (Baenziger and Korn fe ld , 1974). N-terminal sequencing o f a papain digestion product a l lowed the determination o f some glycosylat ion sites o f C e n A I g A l h produced in S. lividans (Figure 3.40). M o d i f i e d sites are indicated by the absence o f a signal in the amino acid sequence. A s the protein is known to be glycosylated, the unrecovered amino acids were assumed to be glycosylated amino acids. A l l the serines in the hinge region o f human m y e l o m a serum I g A l are ful ly occupied glycosylat ion sites (Baenziger and Korn fe ld , 1974). In contrast, the hinge region o f C e n A I g A l h , made in S. lividans, had ful ly occupied sites on all the serines and the last threonine and a partially occupied site on the second threonine. C e n A I g A l h produced in S. lividans has approximately 7 moles o f sugar per mole o f protein (Section 3.5.2) w h i c h is similar to that o f the human m y e l o m a I g A l hinge (Baenziger and Korn fe ld , 1974). 126 o I CO I CL • CO I CL Qi CO CD O 1 o c o • - o 5 CO CO I CL CL i H i CO • CL CD O 1 o c o CD CO N C CD 3 CD •2 c o 5 3 CD C a C O ! 5) •S c CD E CD <M CO </> .s c CD s CL • CO I CL • I-• CL • CL CO i CL • col CL I-• CL • CO CD CO CD O ! o c o • CNI — O ^ CD  t o c o CD CO N C CD CD •5 c o i CD C a co .12 .S c CD CD <N CO </) 0 ) .s CD S ro 3 5 75 CD "c 3 i o c 03 o "co O o < z o t o K CO c CD T 3 i s CD u < 00 C3 s cu I 4= o a o '5b a cu 00 s _3 o s o o o 5 cu C 3 OC O 3 C8 ^ -3 oj . . 3 <U • c o I 1 S -3 CJ W cu 3 il a y S a. I I B CJ 1 CJ •3 cj £ < -3 g 3 -2 3 1 •3 a § 1 a 5 1 1 c3 D , b 3 _ CJ < ° £ 6 V. S t CU OH <" CJ co G S 2 CU "5 3 ^ " o M _ cu •a I CU C 3 cu 3 ° 3 a a 127 4. Discussion 4.1. The effect of linker length and composition on activity 4.1.1. The effect of linker length and composition on activity on PASC O n l y the activities o f the linker deletion mutant, C e n A A P T , and the catalytic domain, p30, differed f rom wi ld type C e n A (Section 3.6.4). Th is study conf i rmed the findings o f Shen et al, 1991 that deleting most o f the l inker o f C e n A caused a slight decrease in activity on P A S C whereas the catalytic domain alone (p30) had increased activity. However , the extent o f the decrease reported varied between the two studies. T h e activities f rom Shen et al., 1991 were converted to ratios - activity o f l inker mutant to activity o f w i l d type C e n A - in order to compare the data. F o r C e n A A P T , ratios o f 0.87 and 0.52 were calculated for this study and Shen et al., respectively, and the ratios for p30 were 1.33 and 2.0. These differences are l ikely a reflection o f the heterogeneity o f the substrate, different substrate preparations and possibly different controls. If w i l d type C e n A was not included as a control for each activity measurement made by Shen et al., then the ratios calculated f rom their data are imprecise. Th is is a clear illustration o f the difficulties encountered when trying to compare data f rom separate studies in the cellulase field. 4.1.2. Difficulties in comparing cellulase activity data between studies Var iance in the absolute activity values for hydrolysis o f insoluble cel lulosic substrates makes it difficult to compare results f rom different studies. T h i s variance is 128 primarily due to significant differences in structure and composition of cellulosic substrates, even from similar sources such as BMCC, which are expected due to different preparation treatments by various suppliers and/or researchers. Reproducibility in dispensing the substrate for each new set of experiments adds to the difficulties. Activities in this study were reported as relative activities - ratios of the activity of the mutant to wild type - because it was the most consistent measurement and allowed meaningful comparison of activities of the different mutants. Ratios are not commonly reported in the cellulase literature and therefore only general trends can be compared. This drawback has been noted previously (Gal, 1997). More wide-spread use of a relative measure (ratio) in reporting activities would have the advantage of minimizing the variance due to substrate heterogeneity and preparation allowing for more consistent comparison of the activities of different cellulase enzymes within and between research groups. 4.1.3. The effect of linker length and composition on activity on BMCC Various modifications to the linker region of CenA had a modest impact on the hydrolysis of the more crystalline substrate, BMCC (Section 3.6.3). Both CenAAPT and p30 had decreased activity relative to CenA which confirms the importance of the CBD itself and its orientation in the hydrolysis of crystalline cellulose (Gilkes et al., 1988; Shen et al., 1991). In this thesis, it was hypothesized that the activity of CenA against crystalline cellulose could be increased by increasing the length of the linker or replacing the linker with the Fn3 repeats from CenB. A longer linker sequence might increase the number of glycosidic bonds available to the catalytic domain while tethered to the substrate by the CBD (Black et al, 1997) and the Fn3 module might confer properties which 129 enhance activity on crystalline cellulose (H . Stalbrand, manuscript in preparation). C e n B has fifty times the activity o f C e n A on B M C C ( T o m m e et al., 1996). Th is difference in activity is probably mainly due to differences in their catalytic domains (they are f rom different families), but there is evidence that the Fn3 modules o f C e n B m a y be important for hydrolysis o f the insoluble substrates P A S C and B M C C . W h e n the F n 3 repeats o f C e n B were deleted, the activity compared to W T - C e n B on P A S C and B M C C decreased six- fo ld and eighteen-fold, respectively (H . Stalbrand, manuscript in preparation). It has been suggested that the combinat ion o f the rigidity o f an individual Fn3 domain and the flexibil ity imparted by repeats may assist cooperativity between the C B D and catalytic domain by l imit ing unproductive binding (H . Stalbrand, manuscript in preparation). Doub l ing the length o f the linker and replacing the linker with the F n 3 repeats f rom C e n B resulted in only a modest gain in activity against B M C C : 10% and 2 2 % , respectively. Th is difference seen for B M C C hydrolysis is statistically significant, supported by the paired t-test (Section 3.6.3), but probably not b io logical ly significant since the activity o f the enhanced C e n A constructs on B M C C is still low in compar ison to other cellulases, such as C e n B . T h e smal l increase in activity observed for C e n A F n 3 m a y be due to the Fn3 domains assisting cooperativity between the C B D and catalytic domain as proposed above for C e n B . If the l inker length increased the range o f the tethered catalytic domain , one might expect there to be a change in the profi le o f the products released. T h e H P L C analysis o f the soluble products showed no such change (Section 3.6.7). A n a l y s i s o f the insoluble products was not performed. O n l y one other study on cellulases looked at the effect o f increasing the linker length. W i l s o n et al. found that doubl ing the length o f the 130 linker o f endoglucanase E 2 caused a slight increase in activity. T h e assay was done on filter paper and no values were reported (Wi lson et al., 1995). 4.1.4. Measuring hydrolysis of an insoluble substrate The approach I chose to evaluate the hydrolysis o f B M C C , that o f quantifying the soluble products released by either a reducing sugar assay or H P L C , has limitations. First, not all products o f hydrolysis o f B M C C by an endoglucanase are soluble; only oligosaccharides o f up to six sugar units. Hydro lys is products o f greater size are not measured by this method as they remain with the insoluble fraction. Second, an enzyme tethered to an insoluble substrate can potentially hydrolyze soluble products and thus alter their profi le (B lack et al., 1996). Therefore, the absence o f change in the H P L C profi le o f soluble products m a y not actually reflect that there has been no change. A n a l y s i s o f the insoluble products was not performed. One method for looking at the insoluble fraction is high performance size-exclusion chromatography where the insoluble fraction is derivatized to make it soluble. Th is method was not used because it is very labour intensive and requires a high percentage o f substrate to be hydrolyzed to get meaningful results (Stalbrand et al, 1998; K l e m a n - L e y e r et al., 1994). A direct, sensitive method for analyzing hydrolysis at the cellulose surface is needed to gain further insights. 4.1.5. The effect of linker length and composition on activity on 2,4-DNPC The activity on 2 , 4 - D N P C decreased only for the C e n A F n 3 construct. Th is is discussed further in the glycosylat ion discussion (Section 4.3). T h e p H profi les on soluble substrates d id not change with the modif ications to the linker length and composi t ion. 131 4.1.6. Summary of activity studies In summary, deleting most o f the linker decreases the activity on B M C C and P A S C whereas doubl ing the length or substituting Fn3 repeats for the l inker result in a modest increase in activity against B M C C . Advances in analyzing hydrolysis at the cellulose surface are needed to achieve a more complete understanding o f enzymatic cel lulose degradation. 4.2. Effect of linker length and composition on sensitivity to proteases A l l the nonglycosylated C e n A linker mutants were susceptible to proteolysis and al lowed insights into the nature o f this susceptibility (Section 3.7). C e n A F n 3 was the construct most resistant to papain during the first two hours o f incubation but it was eventually cleaved to p30. However , it was no more resistant to the C. fimi protease than the other constructs. Th is suggests that these proteases interact with C e n A F n 3 differently. It was k n o w n previously that that they attacked similar sequences in C e n A (Gi lkes et al, 1989) but these data urge caution in extrapolating data f rom papain hydrolysis to C . fimi protease hydrolysis. T h e other proteins were equally susceptible but were cleaved at different sites. In all cases, cleavage was in the linker sequences and not in the catalytic and cel lulose-binding domains. Sequence analysis o f the proteolysis products gave insight into the nature o f this susceptibility. Sequences near the boundaries o f the domains were previously shown to be susceptible to proteolysis (Gi lkes et al, 1989; Langsford et al, 1987). Sequencing o f the N-terminal region o f selected proteolytic fragments showed that it is not the structure formed at the junct ion that makes this region susceptible but rather the primary amino acid sequences wh ich form a susceptible structure. W h e n these 132 sequences were m o v e d from the N-terminus to a more central location in the l inker or when the C-terminal sequence was modi f ied (changing a valine to a proline), there was no hydrolysis at the junct ion, only at the susceptible sequences. Th is suggests that the boundaries o f the linker are truly defined by these susceptible sequences. T h e placement o f these sequences may not be a coincidence. Partial degradation o f the cellulases during hydrolysis o f cellulose may be advantageous since the catalytic domains o f some cellulases are more active than the intact enzymes on soluble substrates (Beguin and A l z a r i , 1998; Fierobe et al., 1993; Fierobe et al., 1991; Ghangas and W i l s o n , 1988; Reverbe l -Leroy et al., 1997). T h e separated domains may enhance hydrolysis o f soluble cel lulose which accumulates during hydrolysis (Sandercock et al., 1996). T h e results o f this study have implications in the design o f stable linkers based on natural cellulase linkers, for C B D - f u s i o n proteins. A l though each fusion protein is unique, this study suggests that using specif ic sequences m a y al low control over protease susceptibility. A s seen for the single-chain antibodies containing linkers derived from the T. reesei C B H I linker, replacing part o f a susceptible site wi th prol ine or adding proline adjacent to a susceptible site m a y reduce or eliminate proteolysis (Alf than et al, 1995). T h e effects o f higher proline content on the activity and independence o f the domains would have to be determined. There are not many bacterial proteases w h i c h cleave adjacent to proline; I g A l proteases and a protease from Flavobacterium meningosepticum are the only examples reported, wh ich m a y account for the stability o f sequences with alternating prolines (Lomhol t , 1996; Walter et al, 1980). Ana lys is o f the products o f proteolysis o f the various C e n A l inker constructs by C. fimi protease and papain gave insights into the specificity o f papain and C.fimi protease for 133 these substrates, the boundaries o f the natural C e n A linker, and strategies to prevent proteolysis o f hybr id proteins with proline r ich linkers. 4.3. Glycosylation 4.3.1. Sites of glycosylation T h e sites o f glycosylat ion on C e n A and C e x f rom C. fimi were examined (Section 3.5.3). Langsford et al. previously described the C. fimi proteins C e n A and C e x as glycoproteins and suggested, based on susceptibility to alkali and the resistance to E n d o - H , that the glycosylat ion was O- l inked and l ikely located on the P T linker (Langsford, 1988). In this study, analysis o f the sites o f glycosylat ion was carried further using carbohydrate detection methods on a variety o f proteolysis products to conf i rm that the glycosylat ion occurs solely on the P T linker for Cex . F o r C e n A , the possibi l i ty o f O-g lycosyla t ion on a small part o f the catalytic domain could not be ruled out. There are no potential N- l inkage sites in the glycosylated regions therefore glycosylat ion o f C e n A and C e x produced by C. fimi proteins is restricted to O-l inkages. Th is agrees with the results for C e x , recombinantly produced in S. lividans (Ong et al., 1994), but is in contrast to the fungal cellulases w h i c h show both N - and O- l inked glycosylat ion (Salovuori et al., 1987). B y taking advantage o f a glycosylated degradation product o f C e x and a proteolysis product o f C e n A , l imited N-terminal sequencing revealed that C.fimi glycosylates both serines and threonines in the P T linker. However , not all the serines and threonines o f the linker get glycosylated (Section 3.5.3). T h e sites o f glycosylat ion fit one o f a number o f predicted O-g lycosylat ion sequences, X P X X (where one X is a glycosylated T and the 134 others are any amino acid except C , F , H , M , Y , or W ) , put forth b y G o o l e y and Wi l l i ams, 1994 and Pisano et al, 1994 for eukaryotic protein glycosylat ion. S o m e potential sites were not occupied. The 45 k D a glycopeptide f rom Mycobacterium tuberculosis has glycans at sites wh ich fit the above motif , and as for the C. fimi proteins, some potential sites are unoccupied. T h e glycosylated sequence is D P E P A P P V P T T A , where the b o l d letter indicates a glycosylated amino acid and the underl ined T reflects a potential site wh ich is reported not to be glycosylated (Dobos et al, 1995). T h e linkers o f C B H I and C B H I I f rom T. reesei are r ich in hydroxyamino acids but do not have as many prolines as the linkers o f C e n A and C e x (Gi lkes et al, 199Id). Therefore, g lycosylat ion m a y not be restricted to the predictive sequence discussed for the C. fimi proteins. There are limits to the technique that was used to determine the specif ic sites. G lycosy la t ion makes the amino acid more polar and as a result, in the standard gas phase E d m a n chemistry used for sequencing, the modi f ied amino acids are not soluble in the solvents used to transfer the amino acid to the reaction chamber. Therefore, no amino acid is detected for that cycle i f it is glycosylated. U s i n g the knowledge o f the sequence, can determine what amino acid should have appeared and can infer that it was glycosylated. O n l y sites wh ich are modi f ied greater than 5 0 % o f the time show up as blanks. The less frequently glycosylated sites appear to be non-glycosylated. Pisano and colleagues (Gooley and Wi l l i ams , 1994; Pisano et al, 1994) developed sol id phase E d m a n chemistry to al low the use o f more polar solvents to enable recovery o f the glycosylated amino acids. Compar ing recovery o f the non-glycosylated and glycosylated variants can give an idea o f site occupancy. O u r data therefore reflect sites which are 50-100% occupied. B o t h 135 secondary and tertiary structure immediately surrounding the glycosylat ion site m a y affect its occupancy; possibly by affecting accessibil ity o f glycosyltransferases. F o r the proteins produced in S. lividans; the C e n A linker, C e n A l inker analogs and the I g A l hinge sequences were the only sites o f glycosylat ion. T h e absence o f glycosylat ion o n the C e n A F n 3 construct ruled out the possibi l i ty o f g lycosylat ion on the C B D and catalytic domains assuming that the Fn3 repeats d id not change the conformation o f the protein preventing the catalytic domain and/or the C B D f rom being glycosylated (Section 3,5.1). Th is conclusion is consistent with the data obtained for C e n A produced by C. fimi and C e x produced by C. fimi and S. lividans (Ong et al., 1994). Speci f ic amino acid occupancy was determined for the C e n A I g A l h construct and part o f the C e n A 2 P T construct (data not shown). A g a i n , the sites o f glycosylat ion agree with the predictive sequence discussed above. Interestingly, there are potential g lycosylat ion sites o f this type in the linkers separating the Fn3 repeats, yet neither C e n B produced by C. fimi nor C e n A F n 3 produced in S. lividans are glycosylated. T h e other C . fimi enzymes with Fn3 repeats, C e n D , C b b A and C b h B , are not glycoproteins either (Sandercock et al., 1996). The linker sequences o f these regions are more heterogeneous, contain less prolines, and do not fol low the pattern o f alternating prolines and threonines seen for C e n A and Cex . These differences, coupled with the location o f the linkers next to domains different than those wh ich are found in C e n A and C e x , suggest that it is the protein structure surrounding these sites wh ich prevents glycosylat ion o f this region. T h e linkers m a y be more flexible and/or the F n 3 domains m a y interact wi th each other when in tandem (Er ickson, 1994; Leahy et ai, 1996). Th is freedom o f motion might be enough to al low the surrounding domains to obscure potential glycosylat ion sites from the glycosylat ion machinery. E v e n 136 the shortest l inker, found in the C e n A P T construct, is glycosylated. Th is further strengthens the argument that the surrounding domains and flexibil i ty o f the linker sequence have an impact on glycosylat ion. The glycosylat ion pattern o f glycoproteins produced by C.fimi and S. lividans are similar and the two organisms discriminate between the same protein features for glycosylat ion. Therefore, it is l ikely that their glycosylat ion systems are similar. 4.3.2. Use of mass spectrometry for quantification of glycosylation Despite the advances in mass spectrometry as a tool for analyzing glycosylat ion, I was unable to ful ly take advantage o f this technique (Sections 1.3.10, 3.5.2). E v e n with these advances, large glycoproteins are still quite difficult to work with and thus only a subset o f the proteins were analyzed in this manner. T h e masses o f C e n A and C e x f rom C. fimi revealed the presence o f on average 15-19 (9%w/w) and 16 (5%w/w) moles o f sugar per mole o f protein, respectively. T h e range seen for C e n A reflects the results o f two different trials. T h e M A L D I - T O F spectra are actually composed o f a set o f peaks f rom all the glycoforms o f C e n A which have been smoothed by computer software. T h e average mass reported is the highest point on that smoothed peak. T h e 1.3% difference in mass falls within the standard error and within the h igh range o f the smoothed peaks. The difference is l ikely the result o f calibration or an artifact o f the software smoothing o f the peak. The percent carbohydrate determined by mass spectrometry is similar to that determined by G L C by Langsford , 10% and 8% for C e n A and C e x , respectively, and to that determined by mass spectrometry o f C e x recombinantly produced in S. lividans, 5.4% (Langsford, 1988; O n g et al., 1994). T h e values obtained by mass spectrometry were 137 about ha l f o f those determined by the phenol sulfuric acid sugar assay (Section 3.5.2). Th is discrepancy is due to two factors: the phenol sulfuric acid sugar assay detects non-covalently attached sugars including any contaminating glucose from the environment or that was carried through from the cellulose co lumn purif ication step, and glucose was used as a standard wh ich reacts differently than mannose in the phenol sulfuric acid sugar assay (A . Boraston, personal communicat ion). The mass o f C e n A I g A l h produced in S. lividans was also determined. A n average o f 7 moles o f sugar were detected (Section 3.5.2). M a s s spectrometry has the potential to be a powerful tool for more detailed analysis o f the carbohydrates o f these bacterial glycoproteins. 4.3.3. Monosaccharide composition, size of oligosaccharides and linkage positions between sugars of oligosaccharides T h e monosaccharide composit ion o f C e n A and C e x produced in C. fimi was re-addressed after O n g et al. found low amounts o f galactose present on C e x produced in S. lividans (Ong et al., 1994). Over loading o f the G L C sample showed definit ively that no galactose was present on C e n A and C e x produced by C. fimi. Th is was conf i rmed for C e n A using the F A C E ® technology. C e n A and the C e n A glycosylated analogs, including the I g A l h hinge, all contained galactose and mannose when produced in S. lividans. Th is agrees with the observations o f O n g et al. for C e x (Sections 3.5.4 and 3.5.5) (Ong et al., 1994). T h e difference in monosaccharide composit ion o f glycans from glycoproteins produced in C. fimi and S. lividans l ikely reflects the absence o f a particular galactosyltransferase in C. fimi. Th is is not surprising because glycosylat ion in bacteria appears to be more diverse than in eukaryotes (Moens and Vander leyden, 1997). 138 Ol igosaccharide size and some sugar linkage information for C e n A 2 P T produced in S. lividans was obtained using F A C E ® technology (Section 3.5.6). T h e oligosaccharides ranged in size f rom 1 to 4 sugar moieties. T h e l inkage positions for the disaccharides were determined to be mostly a 1-2 with some a 1-3. O n g et al., using exoglycosidases to probe the sugar l inkage positions, hypothesized the existence o f mainly a 1-6 l inkages on P T C B D c e x produced by Streptomyces (Ong et al, 1994). Th is interpretation was from the observation that the a 1-2 specif ic exo-mannosidase from Aspergillus saitoi (A. satoi) d id not affect the mobi l i ty o f the protein as judged from S D S - P A G E . T h e discrepancy between the two studies may be the result o f the specificity o f the a 1-2 mannosidase or may s imply reflect the location o f the a 1-2 linkages in these oligosaccharides. T h e A. saitoi enzyme is h ighly specific for non-reducing terminal mannose a 1-2 l inkages (Oxford Glycosystems catalogue). Information is not available as to whether the enzyme w i l l cleave this l inkage when it is in close proximity to the protein such as is found for a disaccharide. Therefore, it is possible that these terminal sugar l inkage positions would not be cleaved. E n o u g h sugars need to be removed to result in a sufficient decrease in M W to be able to detect a mobi l i ty shift (by S D S - P A G E ) for samples treated with exo-glycosidases when compared to untreated controls. A n undetectable shift w o u l d result i f the a 1-2 l inkages are not at the terminus or are not in h igh enough quantity to change the M W , and hence the mobi l i ty significantly. T h i s is the first t ime, to the best o f m y knowledge, that the configuration o f the linkage o f disaccharides has been deduced directly from F A C E ® gels. Th is was possible due to the observation that linkage and anomeric configurations, but not the type o f neutral 139 sugar in the l inkage affected the mobi l i ty in the F A C E ® gels o f disaccharides containing combinations o f mannose, galactose and glucose. Further trials with other neutral sugar combinations need to be done to determine whether this observation reflects a more general phenomenon. Previously, the anomeric and linkage configuration were determined using exo-glycosidases. These enzymes are l imited, however, by their inability to distinguish between certain l inkage positions between sugars. A n example is Jack Bean a -mannosidase wh ich cleaves a 1-2,3 and 6 l inked mannose. T h e F A C E ® system proved to be a fast and simple method. W i th the correct battery o f standards, l inkage data could be generated quick ly for simple O- l inked sugars from bacterial, yeast or fungal proteins. Structural analysis o f the carbohydrate moieties o f N - l inked proteins o f mammal ian or igin has been accompl ished using a series o f exo-glycosidase treated proteins fo l lowed by F A C E ® analysis. Standard mobi l i ty shifts reflecting the loss o f specif ic sugars al lowed the determination o f the sequence o f the oligosaccharide but not the l inkage positions between sugars. T h e l inkage information was inferred from k n o w n N- l inked structures previously characterized by other means (Kumar et al., 1996). W i th F A C E ® technology it is not possible to garner information for l inkage positions between sugars or structures not previously described (Raju et al., 1996). 4.3.4. Role of glycosylation on CenA and CenA derivatives Determination o f the role o f glycosylat ion on C e n A and C e x is chal lenging as glycosylat ion could potentially impact many aspects o f cellulose degradation. Previous results from Langsford for C e n A and C e x from C. fimi (g lycosylated) and E. coli (non-glycosylated) showed no difference in their kinetic properties, thermostability, p H stability 140 or ability to b ind cellulose (Langsford, 1988). It is also unl ikely that glycosylat ion is required for secretion because not all the secreted cellulases o f the C. fimi system are glycoproteins. However , when P T C B D c e x was produced in S. lividans it had enhanced b inding properties over the non-glycosylated variant produced in E. coli, and when C e x was produced in Saccharomyces cerevisiae, wh ich hyperglycosylates the enzyme, the thermostability was enhanced (Curry et al, 1988; O n g et al, 1994). T h u s the effects reported for glycosylat ion o f cellulases have been variable. The glycosylated C. fimi proteins are protected against the action o f a C. fimi protease when adsorbed to cellulose (Langsford, 1988). W h e n the enzymes are in solution, the rate o f proteolysis is merely slowed down. Proteolysis studies o f the glycosylated and non-glycosylated variants o f C e n A and C e n A derivatives using papain and C. fimi protease conf irmed the results o f Langsford , but also showed that, in solution, the amount o f protection provided by glycosylat ion varied with both the quantity o f g lycosylat ion and the linker composit ion. T h e partial deletion construct, C e n A A P T , w h i c h had the least glycosylat ion; and the C e n A I g A l h construct wh ich had glycosylat ion comparable to C e n A , but with a more variable amino acid composit ion, were degraded most rapidly (Section 3.7.4). T h e l imited protection o f these two constructs b y glycosylat ion is l ikely to be the result o f too few carbohydrates to effectively mask the protease sensitive sites in the case o f C e n A A P T , and greater f lexibil ity o f the linker due to less prolines in the case o f C e n A I g A l h . T h e increased movement o f the amino acids relative to the carbohydrates might al low the protease greater opportunity to access the cleavage sites. 141 W h e n the effect o f glycosylat ion on cel luloytic activity was addressed, a difference was found for one substrate, 2 , 4 - D N P C , a smal l , soluble, synthetic substrate (Sections 3.6.3 - 3.6.6). T h i s was surprising at first as no activity difference o n a soluble substrate had been previously ascribed to glycosylat ion o f C e n A . T h e catalytic domain o f C e n A alone (p30) was more active than the whole enzyme and the least active enzyme variant was C e n A F n 3 , the largest enzyme. Therefore, it m a y not be an effect o f glycosylat ion per se that influences activity, but o f the mass and vo lume that glycosylat ion adds to proteins. The observation that the catalytic domain is more active than the whole enzyme on soluble substrates has also been observed for C M C a s e activity o f T. fusca E 2 , made in S. lividans, and an endoglucanase from Sporotrichum pulverulentum (Ghangas and W i l s o n , 1988), and C. cellulolyticum C e l A , C e l C and C e l F (Fierobe et al, 1993; Fierobe et al., 1991; Reverbe l -Leroy et al., 1997). The catalytic domains o f T. reesei C B H I and C B H I I had no change in activity on soluble substrates ( T o m m e et al., 1988; Z h a n g et al, 1995). In addition, a hybr id enzyme consisting o f the T. fusca E 2 C B D and the Prevotella ruminicola C M C a s e produced an enzyme, ah endoglucanase, wh ich had higher specif ic activity on C M C than the C M C a s e alone (Magl ione et al, 1992). L i k e the docker in and C B D domains, the sugar moieties add mass and volume to the protein. T h e extended linker region and the carbohydrate moieties are thought to be dynamic and have some flexibil ity, and so could conceivably b lock portions o f the catalytic domain at t imes, reducing the activity. F lexibi l i ty cou ld be tested using fluorescence quenching o f labels o n the protein's extremities (i.e. on the catalytic domain and C B D ) . T h e possibi l i ty that there is a conformational change resulting in decreased catalytic activity associated with having the C B D and linker attached cannot be discounted as there are no structures k n o w n for intact 142 cellulases, just for their domains alone. Future determination o f the K,,, and V m a x values for pairs o f glycosylated and non-glycosylated cellulases m a y give further insight into these results. G lycosy la t ion o f the C e n A constructs d id not have a clear effect on the hydrolysis o f the insoluble substrates, B M C C and P A S C . Th is is l ikely due to the C B D binding the enzyme to the substrate wh ich conceivably wou ld limit the freedom o f mot ion o f the enzyme compared to when it is in solution, as for 2 , 4 - D N P C hydrolysis. A role o f glycosylat ion, i f any, beyond proteolysis protection for C e n A remains elusive. 4.3.5. Summary of glycosylation studies Th is study conf irms that C. fimi and S. lividans differ in the type o f sugars added to glycoproteins but suggests that they are l ikely to have a similar mechanism for glycosylat ion. T o date, no evidence o f N-g lycosylat ion by these organisms has been found. A m i n o acid sequencing showed that predictive sequences for O-g lycosyla t ion in eukaryotes may be applicable in some bacterial systems but that the structure o f the protein strongly influences the occupancy o f potential glycosylat ion sites. T h e role o f glycosylat ion on C e n A and C e n A derivatives produced in S. lividans was examined in this study, but results indicated no expanded role beyond the protective function described previously. Surprisingly, examination o f glycosylat ion o f C e n A and C e n A derivatives gave some insights into the observation that the catalytic domain b y itself has increased activity on soluble substrates. F ina l ly , this study introduced the F A C E ® technology as a fast, general means for determining the sugar linkage positions for disaccharides o f bacterial glycoproteins. 143 4.4. Hydrolysis of CenA and CenAIgAlh by IgAl proteases Th is study built on a previous study in which N. gonorrhoeae type 1 and 2 proteases were shown to cleave C e n A I g A l h but not C e n A (Mi l ler et al., 1992). Cleavage o f C e n A I g A l h was m u c h slower than for I g A l . Th is reduction in proteolysis rate was hypothesized to be due to the conformation within and on either side o f the target sequence. It cou ld also have been due to the lack o f glycosylat ion on C e n A I g A l h produced in E. coli as human I g A l has a glycosylated hinge region. T h i s study tested whether glycosylat ion o f the hinge o f C e n A I g A l h would increase the eff iciency o f cleavage. However , the substrates glycosylated by S. lividans were cleaved less efficiently than their non-glycosylated counterparts, except b y the H. influenzae protease (Section 3.8). G lycosy la t ion is organism- and cel l -specif ic, so it is reasonable that the glycosylat ion pattern is different for human I g A l and the hybr id protein produced in S. lividans. T h e percent carbohydrate on the linkers d id not differ greatly between C e n A I g A l h and m y e l o m a I g A l , and so it is l ikely the other differences in glycosylat ion, including sites and composit ion, m a y have hampered the proteolysis (Section 3.9). T h e glycosylat ion sites o f pooled serum I g A l more closely resemble those o f C e n A I g A l h produced in S. lividans, but this form o f I g A l was not tested in this study. Th is gives further evidence that the glycosylat ion pattern m a y be important for proteolysis (Reinholdt et al., 1990) . Results from this study support previous reports that features beyond amino acid sequence m a y be important determinants o f cleavage eff iciency for I g A l proteases (Lomholt , 1996; M i l l e r et al., 1992). Regardless o f glycosylat ion, the C e n A and C e n A I g A l h substrates were all cleaved m u c h less efficiently than human I g A l . Other 144 substrates, such as L A M P 1 and some ape I g A l molecules, were also poor substrates (L in etal, 1997; Q iu etal, 1996). Despite the poor cleavage rate, interesting insights were obtained using CenA and C e n A I g A l h as substrates. H. influenzae protease cleaved at two sites in CenA and C e n A I g A l h , but only at one site in I g A l . This is l ikely to be due to conformational differences of the domains surrounding the hinge or linker. The H. influenzae protease interacts with the F c a domain of I g A l and the site of cleavage correlates with the length of the cleavage specificity domain. Without the F c a domain, control over the cleavage specificity may be lost. Interestingly, the H. influenzae protease does not require a P/S bond for cleavage; demonstrated by cleavage of the CenA linker which lacks a P/S bond. Cleavage of bonds by specific I g A l proteases different from that cleaved in human I g A l has been demonstrated previously. This was the case for the C. ramosum protease cleavage of goril la I g A l , and the N. meningitidis type 1 and 2 proteases cleavage of I g A l from orangutans (Qiu et al, 1996). The N. meningitidis and N. gonorrhoeae type 2 I g A l proteases cleave the identical peptide bond in I g A l . Both cleaved C e n A I g A l h . However, the susceptibility of CenA to these enzymes reveals a difference. Only the N. meningitidis type 2 protease could cleave CenA. This is the first example of such differentiation and is particularly striking and worth following up as the two enzymes have high sequence identity. The I g A l proteases had different substrate profiles and therefore l ikely have unique substrate requirements. Further studies need to be done to elucidate the intricacies of I g A l protease cleavage, especially as the proteases are considered to be good vaccine candidates 1 4 5 (Lomholt, 1996). There are few alternative substrates to I g A l for studying I g A l proteases. CenA and C e n A I g A l h are good candidates for alternative substrates to help define properties of substrates and proteases responsible for the specificity of some I g A l proteases because the ability to cleave the substrates differed between the proteases tested, most importantly between some proteases with high sequence identity, and they are readily produced. Future experiments using CenA and C e n A I g A l h as substrates for naturally occurring and designed I g A l protease mutants w i l l l ikely yield further insights into their intricacies. In summary, this study has defined CenA as an I g A l protease substrate, demonstrated that an I g A l hinge homologue, C e n A I g A l h , was cleaved by a variety of I g A l proteases, and that some oligosaccharides may interfere with cleavage. CenA and C e n A I g A l h w i l l be useful in defining the properties beyond the cleaved region that contribute to specificity of I g A l proteases. 4.5. Final Summary This study has shown that modifying the size and composition o f the PT linker o f CenA had a limited effect on hydrolysis rates on cellulosic substrates. Changes in composition, however, did have an effect on the susceptibility to proteolysis by papain and C. fimi protease for both glycosylated and non-glycosylated proteins. This knowledge could be applied in the design of linkers for CBD-fus ion proteins to prevent cleavage, or toadd cleavage sites. The analysis of the glycan components of the linkers added to the knowledge base of the fledgling field of prokaryotic glycosylation and introduced the FACE® technology as a fast, general means for determining sugar linkage positions for 146 disaccharides o f bacterial glycoproteins. H a v i n g established some o f the l inkage positions o f disaccharides on glycoproteins made by S. lividans, it wou ld be valuable to identify the glycosyltransferases involved since glycosyltransferases have potential commercia l value. F ina l ly , this study has defined C e n A and C e n A I g A l h as useful substrates for future investigations into properties inf luencing the specificity o f I g A l proteases. 147 5. Bibliography Alf than, K . , Takk inen, K . , S izmann, D. , Sdderlund, H . , and Teer i , T . T . (1995). Properties o f a single-chain antibody containing different l inker peptides. Protein Engineer ing 5 ,725-731. A r g o s , P. (1990). A n investigation o f oligopeptides l inking domains in protein tertiary structures and possible candidates for general gene fusion. Journal o f Mo lecu la r B i o l o g y 211, 943-958. Baenziger, J . and Korn fe ld , S. (1974). Structure o f the carbohydrate units o f I g A l immunoglobul in . II. Structure o f the O-g lycos id ica l ly l inked ol igosaccharide units. Journal o f B io log ica l Chemistry 249, 7270-7281. B a k e r - L i b b y , C , Cornett, C . A . G . , Re i l ly , P .J . , and F o r d , C . (1994). Ef fect o f amino acid deletions in the (9-glycosylated region of Aspergillus awamori g lucoamylase. Protein Engineer ing 7, 1109-1114. Beeley, J . G . (1987). Glycoproteins and proteoglycans. In Glycoprote in and proteoglycan techniques. R . H . B u r d o n and P . H . van Knippenberg , eds. (New Y o r k : Elsevier) , pp. 5-28. Begu in , P. and A l z a r i , P . M . (1998). T h e cel lulosome o f Clostridium thermocellum. B i o c h e m i c a l Society Transactions 26, 178-185. Begu in , P. and Lemaire , M . (1996). T h e cel lulosome: an exocellular, mult iprotein complex special ized in cellulose degradation. Cr i t ical Reviews in B iochemist ry & Mo lecu la r B i o l o g y 31, 201-236. B e g u i n , P. , M i l l e t , J . , and Aubert , J .P . (1992). Cel lu lose degradation b y C lost r id ium thermocel lum: from manure to molecular b io logy. F E M S M i c r o b i o l o g y Letters 79, 523-528. B inscheck , T . , Bartels, F . , Bergel , H . , B iga lke , H . , Y a m a s a k i , S . , Hayash i , T . , N iemann , H . , and Pohlner, J . (1995). I g A protease from Neisseria gonorrhoeae inhibits exocytosis i n bov ine chromaff in cells l ike tetanus toxin. Journal o f B i o l o g i c a l Chemistry 270, \110-\11A. B i o - R a d . M a n u a l for the B i o - R a d protein dye-binding assay. 1993. B i r n b o i m , H . C . and D o l y , J . (1979). A rapid alkaline extraction procedure for screening recombinant p lasmid D N A . Nuc le ic A c i d s Research 7, 1513-1523. 148 B lack , G . W . , R i x o n , J . E . , C larke, J . H . , Haz lewood , G . P . , Ferreira, L . M . , B o l a m , D . N . , and Gilbert , H . J . (1997). Cel lu lose binding domains and l inker sequences potentiate the activity o f hemicellulases against complex substrates. Journal o f B io technology 57, 59-69. B lack , G . W . , R i x o n , J . E . , C larke, J . H . , Haz lewood , G . P . , Theodorou , M . K . , Mor r is , P., and Gilbert , H . J . (1996). Ev idence that linker sequences and cel lu lose-binding domains enhance the activity o f hemicellulases against complex substrates. B i o c h e m i c a l Journal 319, 515-520. Bradford, M . M . (1976). A rapid and sensitive method for the quantitation o f microgram quantities o f protein uti l iz ing the principle o f protein-dye b inding. Ana ly t ica l B iochemistry 72, 248-254. Brandtzaeg, P. (1995). Mo lecu lar and cellular aspects o f the secretory immunoglobul in system. A p m i s 103, 1-19. Bronnenmeier , K . and Staudenbauer, W . L . (1993). Mo lecu lar b io logy and genetics o f substrate util ization in Clostridia. In T h e Clostridia and Biotechnology. D .R . W o o d s , ed. (London: Butterworth-Heinemann), pp. 261-309. Bros ius , J . (1989). Superpolyl inkers in c loning and expression vectors. D N A 8, 759-777. Bur l ingame, A . L . (1996). Characterization o f protein glycosylat ion b y mass spectrometry. Current O p i n i o n in Biotechnology 7, 4-10. Ca lza , R . E . , Irwin, D . C . , and W i l s o n , D . B . (1985). Purif ication and characterization o f two P - l ,4-endoglucanases f rom Thermomonospora fusca. B iochemist ry 24, 7797-7804. Cannon , J . G . N e w developments in research using human challenge with Neisseria Gonorrhoeae. A m e r i c a n Society for M i c r o b i o l o g y 98th Genera l Meet ing . 1998. Chap l in , M . F . (1986). Monosacchar ides. In Carbohydrate analysis: a practical approach. M . F . C h a p l i n and J .F . Kennedy , eds. (Washington, D C : I R L Press), pp. 1-36. Ciphergen Biosystems. M a n u a l for S E L D I technology. 1997. Costel lo , C . E . (1997). T i m e , life... and mass spectrometry. N e w techniques to address b io logical questions. B iophys ica l Chemistry 68, 173-188. Crowe , J . Qiagen p lasmid handbook. 1992. Chatsworth, C A , Qiagen Inc. Curry , C , G i lkes , N .R . , M i l l e r , R .C .Jr . , and Skipper, N . (1988). Express ion and secretion o f a Cellulomonas fimi exoglucanase in Saccharomyces cerevisiae. A p p l i e d and Environmental M i c r o b i o l o g y 54, 476-484. 149 Damude , H . G . Endoglucanase A f rom Cellulomonas fimi: Determination o f the amino acids involved in catalysis and their function. P h . D . 1995. Univers i ty o f Br i t ish C o l u m b i a , Vancouver , B . C . , Canada. Delacroix , D . L . , D i v e , C , Rambaud , J . C , and Vaerman, J P . (1982). I g A subclasses in various secretions and in serum. Immunology 47, 385 D i n , N . , Forsythe, I.J., Burtnick, L . D . , G i lkes , N .R . , M i l le r , R . C . J . , Warren, R . A . , and K i l b u r n , D . G . (1994). T h e cel lulose-binding domain o f endoglucanase A ( C e n A ) from Cellulomonas fimi: evidence for the involvement o f tryptophan residues in b inding. Molecu lar M i c r o b i o l o g y 11, 747-755. D i n , N . , G i lkes , N .R . , Tekant, B . , M i l le r , R . C . , Jr., Warren, R . A . J . , and K i l b u r n , D . G . (1991). Non-hydro lytic disruption o f cellulose fibres by the b inding domain o f a bacterial cellulase. B i o / T e c h n o l o g y 9, 1096-1099. D i n g , X . , Shields, J . , and Hussey , R . S . (1998). A secretory cel lu lose-binding protein c D N A cloned from the root-knot nematode (Meloidogyne incognita). Mo lecu la r Plant-M i c r o b e Interactions 11, 952 D o b o s , K . M . , Swiderek, K . M . , K h o o , K . H . , Brennan, P .J . , and Bel is le , J . T . (1995). Ev idence for glycosylat ion sites on the 45-ki lodalton glycoprotein o f Mycobacterium tuberculosis. Infection & Immunity 63, 2846-2853. D w e k , R . A . (1996). G lycob io logy : T o w a r d understanding the function o f sugars. C h e m i c a l Reviews 683-720. Dwek , R . A . (1995). G lycob io logy : "Towards understanding the function o f sugars". B iochemica l Society Transactions 23, 1-25. E l k o n , K . B . , Inman, R . D . , Culhane, L . , and Christ ian, C L . (1983). Induction o f polymer ic I g A rheumatoid factors in infective endocarditis. A m e r i c a n Journal o f M e d i c i n e 75, 785-789. Er ickson , H .P . (1994). Reversible unfolding o f fibronectin type III and immunoglobul in domains provides a structural basis for stretch and elasticity o f titin and fibronectin. Proceedings o f the Nat ional A c a d e m y o f Sciences o f the Un i ted States o f A m e r i c a 97,10114-10118. Fierobe, H . -P . , Bagnara-Tardi f , C , Gaud in , C , Guer lesquin, F . , Sauve, P., Be la ich , A . , and B e l a i c h , J . -P . (1993). Purif ication and characterization o f endoglucanase C from Clostridium cellulolyticum. Journal o f Bacter iology 217, 557-565. Fierobe, H . -P . , Gaud in , C , Be la ich , A . , Lout f i , M . , Faure, E . , Bagnara, C , Baty , D. , and Be la ich , J . -P . (1991). Characterization o f endoglucanase A from Clostridium cellulolyticum. Journal o f Bacter iology 173, 7956-7962. 150 G a l , L . Etude du cel lulosome de Clostridium cellulolyticum et de l'un de ses composants: la cellulase C e l G . P h . D . 1997. Universite de Provence, France. Gerwig , G . J . , Kamer l ing , J .P . , Vliegenthart, J . F . , M o r a g , E . , L a m e d , and Bayer , E . A . (1993). T h e nature o f the carbohydrate-peptide l inkage region in glycoproteins f rom the cel lulosomes o f Clostr id ium thermocel lum and Bacteroides cel lulosolvens. Journal o f B io log ica l Chemistry 268, 26956-26960. Gerwig , G . J . , Kamer l ing , J .P . , Vliegenthart, J . F . G . , M o r a g , E . , L a m e d , R., and Bayer , E . (1992). N o v e l oligosaccharide constituents o f the cellulase complex o f Bacteroides cellulosolvens. European Journal o f Biochemistry 205, 799-808. Gerwig , G . J . , Kamer l ing , J .P . , Vliegenthart, J . F . G . , M o r a g (Morgenstern) , E . , L a m e d , R., and Bayer , E . A . (1991). Pr imary structure o f O- l inked carbohydrate chains in the cel lulosome o f different Clostridium thermocellum strains. European Journal o f Biochemistry 196, 115-122. Gerwig , G . J . , de Waard , P., Kamer l ing , J .P . , Vliegenthart, J . F . , Morgenstern, E . , L a m e d , R., and Bayer , E . A . (1989). N o v e l O- l inked carbohydrate chains in the cellulase complex (cellulosome) o f Clostr id ium thermocellum. 3 - O - M e t h y l - N -acetylglucosamine as a constituent o f a glycoprotein. Journal o f B i o l o g i c a l Chemistry 264, 1027-1035. Ghangas, G . S . and W i l s o n , D . B . (1988). C l o n i n g o f the Thermomonospora fusca endoglucanase E 2 gene in Streptomyces lividans: A f f in i ty purif ication and functional domains o f the c loned gene product. A p p l i e d & Environmenta l M i c r o b i o l o g y 54, 2521 -2526. Gi lkes , N .R . , K i l b u r n , D . G . , M i l l e r , R .C .Jr . , and Warren, R . A . J . (1991a). Bacterial cellulases. Bioresearch Technology 36, 21-35. Gi lkes , N .R . , M e i n k e , A . , Cout inho, J . B . , O n g , E . , Greenwood, J . M . , M i l l e r , R .C .J r . , K i l b u r n , D . G . , and Warren, R . A . J . (1991b). Cellulomonasfimi |3-l ,4-glucanases. In E n z y m e s in B iomass Convers ion. G . F . Leatham and M . E . H i m m e l , eds. (pp. 349-360. G i lkes , N .R . , Claeyssens, M . , Aeberso ld , R., Henrissat, B . , M e i n k e , A . , M o r r i s o n , H . D . , K i l b u r n , D . G . , Warren, R . A . , and Mi l l e r , R . C . J . (1991c). Structural and functional relationships in two families o f beta- l ,4-glycanases. European Journal o f Biochemistry 202, 361-311. Gi lkes , N .R . , Henrissat, B . , K i l b u r n , D . G . , M i l l e r , R . C . J . , and Warren, R A . (1991d). D o m a i n s in microbia l beta- l ,4-glycanases: sequence conservation, function, and enzyme families. M ic rob io log ica l Reviews 55, 303-315. 151 Gi lkes , N .R . , K i l b u r n , D . G . , M i l l e r , R . C . J . , and Warren, R . A . (1989). Structural and functional analysis o f a bacterial cellulase by proteolysis. Journal o f B i o l o g i c a l Chemistry 264, 17802-17808. Gi lkes , N .R . , Warren, R . A . , M i l l e r , R . C . J . , and K i l b u r n , D . G . (1988). Precise excision o f the cellulose b inding domains f rom two Cellulomonas fimi cellulases by a homologous protease and the effect on catalysis. Journal o f B i o l o g i c a l Chemistry 263, 10401-10407. G l y k o . G l y k o catalog. 1997. G l y k o , Inc. F A C E monosaccharide composit ion kit manual . 1995. G l y k o , Inc. F A C E oligosaccharide analysis kit manual. 1995. Goochee , C . F . , Gramer, M . J . , Andersen, D . C . , Bahr , J . B . , and Rasmussen, J .R. (1991). The oligosaccharides o f glycoproteins: bioprocess factors affecting oligosaccharide structures and their effect on glycoprotein properties. B i o / T e c h n o l o g y 9, 1347-1355. Goo ley , A . A . and Wi l l i ams , K . L . (1994). Towards characterizing O-g lycans: the relative merits o f in v ivo and in vitro approaches in seeking peptide motifs specifying O -glycosylat ion sites. G l y c o b i o l o g y 4, 413-417. Graham, R . W . , Greenwood, J . M . , Warren, R . A . , K i l b u r n , D . G . , Tr imbur , and D E (1995). The p T u g A and p T u g A S vectors for high- level expression o f c loned genes in Escherichia coli. Gene 158, 51-54. Hagspie l , K . , Haab , D. , and K u b i c e k , C P . (1989). Protease activity and proteolytic modif icat ion o f cellulases f rom a Trichoderma reesei Q M 9414 selectant. A p p l i e d M i c r o b i o l o g y & Biotechnology 32, 61-67. Hanahan, D . (1983). Studies on transformation o f Escherichia coli wi th plasmids. Journal o f Mo lecu la r B i o l o g y 166, 557-580. Harvey, D . J . , N a v e n , T . J . , and Kuster, B . (1996). Identification o f ol igosaccharides by matrix-assisted laser desorption ionization and electrospray M S . B i o c h e m i c a l Society Transactions 24, 905-912. Haynes, P A . (1998). Phosphoglycosylat ion: a n e w structural class o f glycosylat ion?. G l y c o b i o l o g y 8, 1-5. Haze lwood , G . P . and Gilbert , H . J . (1998). Structure and function analysis o f Pseudomonas plant cel l wal l hydrolases. B iochemica l Society Transactions 26, 185-190. H e m m i n g , F . W . (1995). Expression and secretion o f glycoproteins b y hyphal fungi. B i o c h e m i c a l Society Transactions 23, 180-185. 152 Henderson, I.R., Navarro-Garc ia , F . , and Nataro, J .P . (1998). T h e great escape: structure and function o f the autotransporter proteins. Trends in M i c r o b i o l o g y 6, 370-377. Hestr in, S. (1963). Bacterial cellulose. Methods in Carbohydrate Chemist ry 3, 4-9. H o p w o o d , D . A . , B i b b , M . J . , Chater, K . F . , K ieser , T . , Bruton, C . J . , K ieser , H . M . , Lydiate , D . J . , Smith , C P . , W a r d , J . M . , and Schrempf, H . (1985). Genetic Manipulat ion o f Streptomyces, A Laboratory M a n u a l (Norwich, Eng land: T h e John Innes Foundation). H u , P., Chase, T .Jr . , and Eve le igh , D . E . (1993). C l o n i n g o f a Microbispora bispora cel lobiohydrolase gene in Streptomyces lividans. A p p l i e d M i c r o b i o l o g y & Bio technology 35, 631-637. Jackson, P. (1996). The analysis o f f luorophore-labeled carbohydrates b y polyacrylamide gel electrophoresis. Mo lecu lar B iotechnology 5, 101-123. Jenkins, N . and Cur l ing , E . M . A . (1994). Glycosyla t ion o f recombinant proteins: Problems and prospects. E n z y m e and Mic rob ia l Technology 16, 354-364. Jung, E . and Wi l l i ams , K . L . (1997). The production o f recombinant glycoproteins with special reference to simple eukaryotes including Dictyostelium discoideum. Biotechnology & A p p l i e d Biochemistry 25, 3-8. Jung, E . D . , L a o , G . , Irwin, D. , Barr , B . K . , Ben jamin , A . , and W i l s o n , D . B . (1993). D N A sequences and expression in Streptomyces lividans o f an exoglucanase gene and an endoglucanase gene from Thermomonospora fusca. A p p l i e d & Environmenta l M i c r o b i o l o g y 59, 3032-3043. K a l s i , J . , Delacroix , D . L . , and H o d g s o n , H . J . (1983). I g A in alcohol ic cirrhosis. C l in ica l & Experimental Immunology 52, 499-504. Kerr , M . A . (1990). T h e structure and function o f human IgA. B i o c h e m i c a l Journal 271, 285-296. Kett, K . , Brandtzaeg, P., R a d l , J . , and Haai jman, J .J . (1986). Different subcalss distribution o f IgA-produc ing cells in human lymphoid organs and various secretory tissues. Journal o f Immunology 136, 3631-3635. K i l i a n , M . , Reinholdt , J . , Lomhol t , H . , Poulsen, K . , and Frandsen, E . V . (1996). B io log ica l signif icance o f I g A l proteases in bacterial colonizat ion and pathogenesis: critical evaluation o f experimental evidence. A p m i s 104, 321-338. K i l i a n , M . and Russe l l , M . W . (1994). Funct ion o f mucosal immunoglobul ins . In Handbook o f M u c o s a l Immunity. P . L . Ogra , J . Mestecky, M . E . L a m m , W . Strober, J .R. M c G h e e , and J . Bienenstock, eds. (Orlando, F A : A c a d e m i c Press), pp. 127-137. 153 K i l i a n , M . and Reinholdt , J . (1987). A hypothetical mode l for the development o f invasive infection due to I g A l pro tease-producing bacteria. Advances in Experimental M e d i c i n e & B i o l o g y 216B, 1261-1269. K leman-Leyer , K . M . , G i lkes , N .R . , M i l l e r , R . C . J . , and K i r k , T . K . (1994). Changes in the molecular-size distribution o f insoluble celluloses by the action o f recombinant Cellulomonas fimi cellulases. B iochemica l Journal 302, 463-469. K u b i c e k , C P . (1992). The cellulase proteins o f Trichoderma reesei: Structure, mult ipl ici ty, mode o f action and regulation o f formation. Advances in B i o c h e m i c a l Engineer ing/Biotechnology 45, 1-27. K u m a r , H . P . M . , Hague , C , Haley , T . , Starr, C M . , B e s m a n , M . J . , L u n d l a n d , R . L . , and Baker , D . (1996). E lucidat ion o f N- l inked oligosaccharide structures o f recombinant human Factor VIII using fluorophore-assisted carbohydrate electrophoresis. B iotechnology & A p p l i e d Biochemistry 24, 207-216. L a e m m l i , U . K . (1970). Cleavage o f structural proteins during the assembly o f the head o f bacteriophage T 4 . Nature 227, 680-685. L a m m , M . E . , Nedrud , J . G . , Kaetzel , C . S . , and Mazanec , M . B . (1995). I g A and mucosal defense. A p m i s 103, 241-246. Langsford , M . L . T h e purif ication and characterization o f two cel lulose-binding, glycosylated cellulases f rom the bacterium Cellulomonas fimi. P h . D . 1988. Univers i ty o f Br i t ish C o l u m b i a , B . C . Langsford , M . L . , G i lkes , N .R. , S ingh, B . , M o s e r , B . , M i l l e r , R . C . J . , Warren, R . A . , and K i l b u r n , D . G . (1987). Glycosyla t ion o f bacterial cellulases prevents proteolytic cleavage between functional domains. F E B S Letters 225, 163-167. Langsford , M . L . , G i lkes , N .R . , Wakarchuk, W . W . , K i l b u r n , D . G . , M i l l e r , R .C .J r . , and Warren, R . A . J . (1984). The cellulase system ofCellulomonas fimi. Journal o f General M i c r o b i o l o g y 130, 1367-1376. L a o , G . and W i l s o n , D . B . (1996). C l o n i n g , sequencing, and expression o f a Thermomonospora fusca protease gene in Streptomyces lividans. A p p l i e d & Environmental M i c r o b i o l o g y 62, 4256-4259. Leahy , D . J . , A u k h i l , I., and Er ickson , H .P . (1996). A crystal structure o f a four-domain segment o f human fibronectin encompassing the R G D loop and synergy region. C e l l 84, 155-164. Lever , M . (1973). Color imetr ic and fluorometric carbohydrate determination with p-hydroxybenzoic acid hydrazide. B iochemica l Med ic ine 7, 274-281. 154 L i n , L . , A y a l a , P., La rson , J . , M u l k s , M . , Fukuda , M . , Car lsson, S .R . , E n n s , C , and So (1997). T h e Neisseria type 2 I g A l protease cleaves L A M P 1 and promotes survival o f bacteria within epithelial cells. Mo lecu la r M i c r o b i o l o g y 24, 1083-1094. L i s , H . and Sharon, N . (1993). Protein glycosylat ion: structural and functional aspects. European Journal o f Biochemistry 218, 1-27. Lomhol t , H . (1996). Mo lecu lar b io logy and vaccine aspects o f bacterial immunoglobul in A l proteases. A p m i s Supplementum. 62, 5-28. M a c h , H . , M i d d a u g h , C . R . , and L e w i s , R . V . (1992). Statistical determination o f the average values o f the extinction coefficients o f tryptophan and tyrosine in native proteins. Ana ly t ic Biochemistry 200, 74-80. M a c L e o d , A . M . , Lindhorst , T . , Withers, S . G . , and Warren, R . A . (1994). T h e acid/base catalyst in the exoglucanase/xylanase from Cellulomonas fimi is glutamic acid 127: evidence from detailed kinetic studies o f mutants. B iochemistry 33, 6371-6376. M a c L e o d , A . M . , G i lkes , N .R. , Escote-Car lson, L . , Warren, R . A . J . , K i l b u r n , D . G . , and Mi l l e r , R .C .J r . (1992). Streptomyces lividans glycosylates an exoglucanase (Cex) from Cellulomonas fimi. Gene 121, 143-147. Magl ione , G . , Matsushita, O . , Russel l , J . B . , and W i l s o n , D . B . (1992). Properties o f a genetically reconstructed Prevotella ruminicola endoglucanase. A p p l i e d and Environmental M i c r o b i o l o g y 58, 3593-3597. M a i n , A . L . , Harvey , T . S . , Baron , M . , B o y d , J . , and Campbe l l , I.D. (1992). T h e three-dimensional structure o f the tenth type III module o f f ibronectin: an insight into R G D - m e d i a t e d interactions. C e l l 71, 671-678. Matsudaira, P. (1990). L imi ted N-terminal sequence analysis. Methods In E n z y m o l o g y 182, 602-613. Mattu, T . S . , Pleass, R . J . , W i l l i s , A . C . , K i l i a n , M . , W o r m a l d , M . R . , L e l l o u c h , A . C . , R u d d , P M , W o o f , J . M . , and Dwek , R . A . (1998). T h e glycosylat ion and structure o f human serum I g A l , F a b , and F c regions and the role o f N-g lycosy la t ion on F c alpha receptor interactions. Journal o f B io log ica l Chemistry 273, 2260-2272. M c G u i r e , E . J . , K e r l i n , R., Cebra , J . J . , and Roth , S. (1989). A human mi lk galactosyltransferase is specific for secreted, but not p lasma, IgA. Journal o f Immunology 143, 2933-2938. M e i n k e , A . , Braun , C , G i lkes , N .R . , K i l b u r n , D . G . , M i l l e r , R . C . J . , and Warren, R . A . (1991a). Unusua l sequence organization in C e n B , an inverting endoglucanase from Cellulomonas fimi. Journal o f Bacter iology 173, 308-314. 155 M e i n k e , A . , G i lkes , N .R . , K i l b u r n , D . G . , M i l l e r , R .C .Jr . , and Warren, R . A . J . (1991b). Mul t ip le domains in Endoglucanase B (CenB) f rom Cellulomonas fimi: functions and relatedness to domains in other polypeptides. Journal o f Bacter io logy 173, 7126-7135. Mescher , M . F . and Strominger, J . L . (1976). Structuralm (shape-maintaining) role o f the cel l surface glycoprotein o f Halobacterium salinarium. Proceedings o f the Nat ional A c a d e m y o f Sciences o f the Uni ted States o f A m e r i c a 73, 2687-2691. Messner , P. (1997a). Bacterial glycoproteins. Glycoconjugate Journal 14, 3-11. Messner , P., A l lma ier , G . , Schaffer, C , Wugedi tsch, T . , Lorta l , S. , K o n i g , F L , N iemetz , R., and Dorner , M . (1997b). Biochemistry o f S-layers. F E M S M i c r o b i o l o g y Rev iews 20, 25-46. Messner , P. and Sleytr, U . B . (1988). Asparginyl - rhamnose: a novel type o f protein-carbohydrate l inkage in a eubacterial surface-layer glycoprotein. F E B S Letters 228, 317-320. Mi l le r , P . B . , Shen, F L , G i lkes , N .R . , K i l b u r n , D . G . , M i l l e r , R . C . , Jr., Plaut, A . G . , and Warren, R A . (1992). Endoglucanase A f rom Cellulomonas fimi in w h i c h the hinge sequence o f human I g A l is substituted for the l inker connecting its two domains is hydro lyzed by I g A proteases f rom Neisseria gonorrhoeae. F E M S M i c r o b i o l o g y Letters 71, 199-203. M o e n s , S. and Vander leyden, J . (1997). Glycoproteins in prokaryotes. Arch ives o f M i c r o b i o l o g y 168, 169-175. Motu lsky , H . (1995). Compar ing two paired groups: paired t and W i l c o x o n tests. In Intuitive Biostatistics. H . Motu lsky , ed. (New Y o r k : O x f o r d Univers i ty Press Inc.), pp. 225-229. Nez l in , R. (1998). T h e Immunoglobul ins: Structure and Funct ion (Boston, M T : A c a d e m i c Press). N i k o l o v a , E . B . , T o m a n a , M . , and Russe l l , M . W . (1994). T h e role o f the carbohydrate chains in complement (C3) fixation by sol id-phase-bound human IgA. Immunology 82, 321-327. Novagen . TB#5 . 4. 1994. O n g , E . , K i l b u r n , D . G . , M i l l e r , R .C .J r . , and Warren, R . A . J . (1994). Streptomyces lividans glycosylates the l inker region o f a P-l,4-glycanase f rom Cellulomonas fimi. Journal o f Bacter io logy 776, 999-1008. 156 0 r s k o v , F . (1984). Genus 1 Escherichia. In Bergey's M a n u a l o f Systematic Bacter iology. N . R . K r u g and J . G . Hol t , eds. (Balt imore: Wi l l i ams and Wi l l i ams) , pp. 420 Paul , G . F . L . and Wie land , F. (1986). Asparaginyl -N-acetylgalactosamine, l inkage unit o f halobacterial g lycosaminoglycan. Journal o f B i o l o g i c a l Chemist ry 30, 17615-17618. Pierce-Cretel , A . , Pamblanco, M . , Strecker, G . , Montreui l , J . , and Spik , G . (1981). Heterogeneity o f the glycans O-g lycos id ica l ly l inked to the hinge region o f secretory immunoglobul ins f rom human mi lk . European Journal o f Biochemistry 114, 169-178. Pisano, A . , Packer, N . H . , R e d m o n d , J . W . , Wi l l i ams , K . L . , and G o o l e y , A . A . (1994). Characterization o f O- l inked glycosylat ion motifs in the glycopeptide domain o f bovine kappa-casein. G l y c o b i o l o g y 4, 837-844. P lummer , T .H .J r . , Tarentino, A . L . , and Hauer, C R . (1995). N o v e l , specif ic O-glycosylat ion o f secreted Flavobacterium meningosepticum proteins. Journal o f B i o l o g i c a l Chemistry 270, 13192-13196. Pohlner, J . , Klauser , T . , Kuttler, E . , and Halter, R. (1992). Sequence-specif ic cleavage o f protein fusions using a recombinant Neisseria type 2 I g A protease. B i o / T e c h n o l o g y 10, 799-804. Procaccia , S . , Lazzar in , A . , C o l u c c i , A . , Gasparini , A . , Force l l in i , P., Lanzanova , F o p p a , C . U . , Novat i , R., and Zanussi , C . (1987). I g M , I g G and I g A rheumatoid factors and circulating immune complexes in patients with A I D S and AIDS-re la ted complex with serological abnormalities. C l in ica l & Experimental Immunology 67, 236-244. Qiagen Inc. Qiaex II handbook. 1995. Chatsworth, C A , Qiagen Inc. Q i u , J . , Brackee, G . P . , and Plaut, A . G . (1996). Ana lys is o f the specif icity o f bacterial immunoglobul in A (IgA) proteases by a comparative study o f ape serum IgAs as substrates. Infection & Immunity 64, 933-937. Raju , T . S . , Lerner, L . , and O 'Connor , J . V . (1996). G l y c o p i n i o n : b io logica l signif icance and methods for the analysis o f complex carbohydrates o f recombinant glycoproteins. B iotechnology & A p p l i e d Biochemistry 24, 191-194. Re inhold , B . B . , Hauer, C . R . , P lummer , T . H . , and Re inhold , V . N . (1995). Detai led structural analysis o f a novel , specific O- l inked glycan from the prokaryote Flavobacterium meningosepticum. Journal o f B i o l o g i c a l Chemist ry 270, 13197-13203. 157 Reinholdt , J . , T o m a n a , M . , Mortensen, S . B . , and K i l i a n , M . (1990). Mo lecu la r aspects o f immunoglobul in A l degradation by oral streptococci. Infection & Immunity 58, 1186-1194. Reverbel -Leroy , C , Pages, S. , Be la ich , A . , Be la ich , J . -P . , and Tardif , C . (1997). T h e processive endocellulase C e l F , a major component o f the Clostridium cellulolyticum cel lulosome: Purif ication and characterization o f the recombinant form. Journal o f Bacter io logy 179, 46-52. R u d d , P . M . , Gu i le , G . R . , Kuster, B . , Harvey, D .J . , Opdenakker, G . , and D w e k , R A . (1997). Ol igosaccharide sequencing technology. Nature 388, 205-207. Saloheimo, A . , Henrissat, B . , Hof f ren, A . , Te leman, O . , and Penttila, M . (1994). A novel , smal l endoglucanase gene, eg/5, f rom Trichoderma reesei isolated by expression in yeast. Mo lecu la r M i c r o b i o l o g y 13, 219-228. Salovuor i , I., Makarow, M . , Rauvala , H . , K n o w l e s , J . , and Kaar ia inen, L . (1987). L o w molecular weight high-mannose type glycans in a secreted protein o f the filamentous fungus Trichoderma reesei. B iotechnology 5, 152-156. Sambrook, J . , Fr i tsch, E . F . , and Maniat is , T . (1989). Mo lecu lar C l o n i n g : A Laboratory M a n u a l ( C o l d Spr ing Harbour, N Y : C o l d Spring Harbour Laboratory) . Sandercock, L . E . , M e i n k e , A . , G i lkes , N .R . , K i l b u r n , D . G . , and Warren, R . A . J . (1996). Degradation o f cellulases in cultures o f Cellulomonas fimi. F E M S M i c r o b i o l o g y Letters 143, 7-12. Sandercock, L . E . , M a c L e o d , A . M . , O n g , E . , and Warren, R . A . J . (1994). Non-S - layer glycoproteins in eubacteria. F E M S M i c r o b i o l o g y Letters 118, 1-8. Sawadogo, M . and van D y k e , M . W . (1991). A rapid method for the purif ication o f deprotected ol igodeoxynucleotides. Nuc le ic A c i d s Research 19, 61A Schii lein, M . (1998). Kinet ics o f fungal cellulases. B i o c h e m i c a l Society Transactions 26, 164-167. Scopes, R . K . (1974). Measurement o f protein by spectrophotometry at 205 n m . Analy t ica l B iochemistry 59, 277-282. Shen, H . , M e i n k e , A . , T o m m e , P., Damude , H . G . , K w a n , E . , K i l b u r n , D . G . , M i l l e r , R . C . J . , Warren, R . A . J . , and Gi lkes , N .R . (1995). Cellulomonas fimi cel lobiohydrolases. In Enzymat ic Degradation o f Insoluble Carbohydrates. J . N . Saddler and M . H . Penner, eds. (Washington, D C : A m e r i c a n C h e m i c a l Society) , pp. 174-196. Shen, H . , Schmuck , M . , P i l z , I., G i lkes , N .R . , K i l b u r n , D . G . , M i l l e r , R . C . J . , and Warren, R . A . (1991). Delet ion o f the linker connecting the catalytic and cel lulose-binding 158 domains o f endoglucanase A ( C e n A ) of Cellulomonas fimi alters its conformation and catalytic activity. Journal o f B io log ica l Chemistry 266, 11335-11340. Shen, H . T h e construction and characterization o f a P ro -Thr box deletion o f a Cellulomonas fimi Endoglucanase ( C e n A ) . M . S c . 1990. Univers i ty o f Br i t ish C o l u m b i a , B . C . , Canada. Smant, G . , Stokkermans, J .P . , Y a n , Y . , de Boer , J . M . , B a u m , T . J . , W a n g , X . , Hussey , R .S . , G o m m e r s , F . J . , Henrissat, B . , Dav is , E . L . , Helder , J . , Schots, A . , and Bakker , J . (1998). Endogenous cellulases in animals: isolation o f beta-1, 4-endoglucanase genes from two species o f plant-parasitic cyst nematodes. Proceedings o f the Nat ional A c a d e m y o f Sciences o f the Uni ted States o f A m e r i c a 95, 4906-4911. Srisodsuk, M . , Reinikainen, T . , Penttila, M . , and Teer i , T . T . (1993). R o l e o f the interdomain linker peptide o f Trichoderma reesei cel lobiohydrolase I in its interaction with crystalline cellulose. Journal o f B i o l o g i c a l Chemist ry 268, 20756-20761. Stackebrandt, E . and K e d d i e , R . M . (1984). Genus Cellulomonas. In Bergey's M a n u a l o f Systematic Bacter iology. N . R . K r u g and J . G . Hol t , eds. (Balt imore: Wi l l i ams and Wi l l iams) , pp. 1325 Stalbrand, H . , Mansf ie ld , S . D . , Saddler, J . N . , K i l b u r n , D . G . , Warren, R . A . J . , and Gi lkes , N . R . (1998). Ana lys is o f molecular size distributions o f cel lulose molecules during hydrolysis o f cellulose by recombinant Cellulomonas fimi P-l-4 glucanases. A p p l i e d and Environmental M i c r o b i o l o g y 64, 2374-2379. Stewart, B . J . and Leatherwood, J . M . (1976). Derepressed synthesis o f cellulase by Cellulomonas. Journal o f Bacter iology 128, 609-615. Stromqvist, M . and G m f f m a n , H . (1974). Periodic ac id /Sch i f f staining o f glycoproteins immobi l i zed on a blotting matrix. Biotechniques 13, 744-746. Svehag, S . E . and B lo th , B . (1970). Ultrastructure o f secretory and h igh-polymer serum immunoglobul in A o f human and rabbit origin. Science 168, 847-849. Teather, R . M . and W o o d , P.J. (1982). U s e o f congo red-polysaccharide interaction in enumeration and characterization o f cel lulolyt ic bacteria f rom bovine rumen. A p p l i e d and Environmental M i c r o b i o l o g y 43, 777-780. Teer i , T . T . , K o i v u l a , A . , L inder , M . , Wohlfahrt , G . , D i v n e , C , and Jones, T . A . (1998). Trichoderma reesei cel lobiohydrolases: w h y so efficient on crystalline cellulose? B iochemica l Society Transactions 26, 173-178. Teer i , T . T . , Penttila, M . , Keranen, S. , Nevala inen, H . , and K n o w l e s , J . K . C . (1992). Structure, function, and genetics o f cellulases. In B iotechnology o f Fi lamentous 159 F u n g i . D . B . Finkelstein and C . B a l l , eds. (London: Butterworth-Heinemann), pp. 417-445. Thomas , M . G . (1994). A procedure for second-round differential screening o f c D N A libraries. Biotechniques 16, 193-211. Tomana , M . , Mestecky , J . , and Niedermeier, W . (1972). Studies on human secretory immunoglobul in A . IV. Carbohydrate composit ion. Journal o f Immunology 108, 1631-1636. T o m m e , P., K w a n , E . , G i lkes , N .R . , K i l b u r n , D . G . , and Warren, R A . (1996). Characterization o f C e n C , an enzyme from Cellulomonas fimi with both endo- and exoglucanase activities. Journal o f Bacter io logy 178, 4216-4223. T o m m e , P., Warren, R . A . , and Gi lkes , N . R . (1995). Cel lu lose hydrolysis b y bacteria and fungi. Advances in M ic rob ia l Phys io logy 37:1-81, 1-81. T o m m e , P., van T i lbeurgh, F L , Pettersson, G . , V a n D a m m e , J . , Vandekerckhove, J . , K n o w l e s , J . , Teer i , T . , and Claeyssens, M . (1988). Studies o f the cel lulolyt ic system o f Trichoderma reesei Q M 9414. European Journal o f B iochemist ry 170, 575-581. Underdown, B . and Mestecky , J . (1994). M u c o s a l immunoglobul ins. In H a n d b o o k o f M u c o s a l Immunology. P . L . Ogra , J . Mestecky , M . E . L a m m , W . Strober, J .R. M c G h e e , and J . Bienenstock, eds. (Oralndo, F A : A c a d e m i c Press), pp . 79-97. Vaerman, J .P . (1973). Res. Immunochem. Immunobiol . 3, 91-99. van den E i j n , D . H . , Neeleman, A . P . , V a n der Knaap , W . P . , Bakker , H . , Agterberg, M . , and V a n D i e , I. (1995). N o v e l glycosylat ion routes for glycoproteins: the l a c d i N A c pathway. B i o c h e m i c a l Society Transactions 23, 175-179. Walter, R., S immons , W . H . , and Yosh imoto , T . (1980). Prol ine specif ic endo- and exopeptidases. Molecu lar & Cel lu lar Biochemistry 30, 111-127. Warren, R A . (1996). M ic rob ia l hydrolysis o f polysaccharides. A n n u a l R e v i e w o f M i c r o b i o l o g y 50:183-212, 183-212. Watanabe, W . , N o d a , H . , T o k u d a , G . , and L o , N . (1998). A cellulase gene o f termite origin. Nature 396, 330-331. Weisbart, R . H . , K a c e n a , A . , Schuh, A . , and G o l d e , D . W . (1988). G M - C S F induces human neutrophil IgA-mediated phagocytosis by an I g A F c receptor activation mechanism. Nature 332, 647-648. 160 White, A . , Withers, S . G . , G i lkes , N .R . , and Rose , D . R . (1994). Crystal structure o f the catalytic domain o f the beta- l ,4-glycanase cex f rom Cellulomonas fimi. Biochemistry 33 , 12546-12552. Wi l l i amson , M . P . (1994). T h e structure and function o f prol ine-r ich regions in proteins. B i o c h e m i c a l Journal 297, 249-260. W i l s o n , D . B . , Spez io , M . , Irwin, D. , Karp lus , A . , and Tay lor , J . (1995). C o m p a r i s o n o f enzymes catalyzing the hydrolysis o f insoluble polysaccharides. In Enzymat ic Degradation o f Insoluble Carbohydrates. J . N . Saddler and M . H . Penner, eds. (Washington, D C : A m e r i c a n C h e m i c a l Society), pp. 1-12. W i l s o n , D . B . (1992). Biochemistry and genetics o f actinomycete cellulases. Cr i t ical Rev iews in Biotechnology 12, 45-63. W i l s o n , I.B., G a v e l , Y . , and v o n Hei jne, G . (1991). A m i n o acid distributions around O -l inked glycosylat ion sites. B iochemica l Journal 275, 529-534. W o o d , T . M . (1988). Preparation o f crystalline, amorphous and dyed cellulase substrates. Methods In E n z y m o l o g y 160, 19-25. Yanisch-Perron C , V ie i ra , J . , and M e s s i n g , J . (1985). Improved M 1 3 phage c loning vectors and host strains: nucleotide sequences o f the M 1 3 m p and p U C vectors. Gene 33, 103-119. Zhang , S. , L a o , G . , and W i l s o n , D . B . (1995). Characterization o f a Thermomonospora fusca exocellulase. Biochemistry 34, 3386-3395. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0089277/manifest

Comment

Related Items