Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The binding of an afimbrial bacterial surface adhesin to glycophorin using aqueous polymer two-phase… Jones, Andrew John Melvill 1987

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

Item Metadata


831-UBC_1987_A6_7 J66.pdf [ 4.85MB ]
JSON: 831-1.0060461.json
JSON-LD: 831-1.0060461-ld.json
RDF/XML (Pretty): 831-1.0060461-rdf.xml
RDF/JSON: 831-1.0060461-rdf.json
Turtle: 831-1.0060461-turtle.txt
N-Triples: 831-1.0060461-rdf-ntriples.txt
Original Record: 831-1.0060461-source.json
Full Text

Full Text

T H E B I N D I N G O F A N A F I M B R I A L B A C T E R I A L S U R F A C E A D H E S I N TO G L Y C O P H O R I N U S I N G A Q U E O U S P O L Y M E R T W O - P H A S E P A R T I T I O N I N G by A N D R E W J O H N M E L V I L L J O N E S A THESIS S U B M I T T E D I N P A R T I A L F U L F I L M E N T OF T H E R E Q U I R E M E N T S FOR T H E D E G R E E OF M A S T E R O F SCIENCE in Department of Chemistry T H E F A C U L T Y OF G R A D U A T E STUDIES Apri l 1987 We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F BRITISH C O L U M B I A 6 Apr i l 1987 ® Andrew John Melvill Jones, 1987 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at The University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Apri l 1987 The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date: 6 Apri l 1987 A B S T R A C T Colonisation by many bacteria and viruses is now thought to depend upon their ability to adhere to host cells via proteinacious surface appendages called adhesins. Information relevant to the prevention and cure of many diseases therefore is supplied by knowledge of this adhesive process, especially the chemistry of the binding and the structure of the binding molecules. A t this time, the structure of very few adhesin receptors is known. Similarly, the quaternary and primary structure of only a small number of adhesins is currently available. Those associated with Escherichia coli are known in some cases to be arranged as helical coils with repeating proteinacious subunits with molecular weights of 10-30 kDa, however there is conflicting information on the distribution along these coils of the polypeptide involved in adhesion. Thermodynamic binding studies have not yet been- used to clarify this problem because of the size of the receptors for the adhesins. This thesis presents a thermodynamic study of the binding between the adhesin from an F41+ E.coli and its receptor, glycophorin, from the human red blood cell membrane using an aqueous polymer two-phase system. The study shows that 2.4±0.8 glycophorin molecules bind to the predominant subunit of this adhesin, suggesting that this subunit has one binding site, since glycophorin dissolves as a dimer. It is proposed that the assay could be used, in addition, to obtain information on the chemical specificity and the thermodynamics of this particular reaction, in order to obtain a broader understanding of the colonisation and infection by this particular pathogen. ii T A B L E O F C O N T E N T S Abstract ii List of Tables vi List of Figures vii Acknowledgements ix I. Introduction 1 II. Background and Theory 7 A . Introduction 7 B. Cellular Adhesion 7 C. Bacterial Surface Adhesins 9 D. The Binding Protein and its Receptor 12 1. The Afimbrial Adhesin 12 2. Glycophorin 13 E . Therapeutic Implications of Adhesin Research 15 F. The Equilibrium Binding Isotherm 16 G. The Two-Phase Separation Technique 19 1. Introduction 19 2. A Statistical Mechanical Description of Phase Separation 21 3. A Thermodynamic Description of Solute Partition 24 III. Preliminary Experiments 27 A . Introduction 27 B. Protein Isolation 28 1. Glycophorin Isolation 28 a. Materials 28 b. Methods 29 2. Adhesin Isolation 30 3. Assessment of the Bioactivity of the Proteins 31 a. Methods 31 b. Results 32 4. Protein Electrophoresis 33 a. Materials 34 b. Methods 34 c. Results 36 C. Protein Separation 41 1. Preparation of Two-Phase Systems 42 a. Materials 42 b. Notation 42 c. Methods 42 2. The Detergent Used in the Two-Phase System 44 a. Preparation of the Fixed Red Blood Cells - 45 b. Preparation of E.coli cells 46 c. Detergents and Adhesin-Glycophorin Binding 47 d. Results 49 iii 3. The Two-Phase Assay used to Separate the Proteins 52 a. Introduction 52 b. Methods 53 c. Results 54 D. Radiolabelling the Proteins 55 1. Introduction 55 2. The Fluorescamine Assay 57 a. Materials 57 b. Methods 57 c. Results 58 3. Radiolabelling of Glycophorin 60 a. Materials 60 b. Methods 60 c. T C A Precipitation 61 d. Removal of Free Label 63 e. Results 63 4. Radiolabelling of the Adhesin 63 a. Materials 64 b. Methods 64 c. Results 65 IV. Equilibrium Binding Experiments 67 A . Introduction 67 B. Experimental 68 1. Methods 68 2. Relative Efficiency Determination 70 3. Calculations 71 4. Test Run 72 5. Results 72 6. Control Experiments 72 V . Results and Discussion 80 A . Results 80 B. Sources of Error 80 1. Introduction 80 2. Precision 81 3. Accuracy 83 a. The Effect of Radiolabelling on Glycophorin Partition ... 83 b. The Solubility of the Proteins 84 c. Protein Concentration 86 C. Evaluation of the Technique 87 D. Evaluation of the Results 89 E . Additional Applications of the Assay 90 F . Conclusions 91 Bibliography 92 Appendix 1 101 iv Appendix 2 104 Appendix 3 105 Appendix 4 107 Glossary of Symbols and Abbreviations 115 v L I S T O F T A B L E S Table Page # 1 Solutions used to make 3% SDS P A G E gels. 35 2 Molecular weight assignments for the standards used on the 40 3% SDS P A G E gels. 3 The composition and notation of the buffers used. 46 4 The partition of adhesin in the lower phase of (5,4) 54 two-phase systems with various buffers, determined from the native fluorescence of the adhesin. 5 Approximate lysine content of adhesin and of glycophorin 60 determined using the fluorescamine assay. 6 Relative counting efficiencies of the radiolabeled proteins in 70 the counters. 7 Summary of the data from control experiments 3 and 4. 76 8 The partition of adhesin in the lower phase of (5,4) two 76 phase systems with various bufferes, determined from the beta counts of the 1 4 C-adhesin. 9 Distribution of proteins in the phase systems. 79 10 Summary of calculations 1 to 5 of Appendix 5. 108 vi L I S T O F F I G U R E S F igure Page # 1 Schematic diagrams of bacterial adhesion, and of the 02 structure of bacterial adhesins. 2 Typical equilibrium binding isotherm, and a schematic of the 04 dialysis experiment used to obtain an isotherm. 3-1 Densitometry scans of adhesin and glycophorin on 3% SDS 37 P A G E Gels. 3-2 Molecular weight markers for 3% SDS P A G E gels. 38 3-3 Molecular weight calibration of 3% SDS P A G E gels. 39 4 The dependence of red blood cell partition on cell binding in 50 a (5,3.5)a phase system. 5 The effect of detergent on the partition of blood cells in the 51 upper phase of the (5,3.5)a phase system. 6 Beta spectra of carbon, hydrogen and iodine isotopes. 56 7 Calibration of adhesin and glycophorin using the 59 fluorescamine assay. 8 Chromatograms of the columns used to clean the 62 radiolabeled proteins. 9 3% SDS P A G E gels of reduced and unreduced radio labelled 66 adhesin. 10 Isotherms for the binding between glycophorin and adhesin. 73 11 Scatchard plots for the binding between glycophorin and 74 adhesin. 12 Comparison of glycophorin partition determined using the 78 1 2 51-glycophorin and the fluorescamine labelled glycophorin. 13 Native fluorescence of adhesin in the spectrofluorometer. 103 14 The plot used to derive the dependence of the partition 111 coefficient of glycophorin on the mass of glycophorin in the upper phase of the (5,4)f two-phase system when the glycophorin was incubated in the presence of Tween 80. vii The dependence of the partition coefficient of glycophorin on the mass of glycophorin in the upper phase of the (5,4)f two-phase system when the glycophorin was incubated in the presence of Tween 80. viii A C K N O W L E D G E M E N T S I should like to thank the following people for the skills, knowledge and time which they have contributed to make this thesis possible: Dr. D.E. Brooks, Dr J . Janzen, Dr. R. Snoek, John Cavanagh, Raymond Norris-Jones, and Charles Ramey. I am also indebted to the National Science and Engineering Research Council for their generous support of this work. ix I. I N T R O D U C T I O N A preliminary step of colonisation by bacteria which infect the gastrointestinal tract, is adhesion of the bacteria to the cells which line the gut. This adhesion involves binding between specific receptors on the surface of the host cell, and fibrous proteins called adhesins, on the surface of the bacteria (Fig 1A). Enterotoxigenic bacteria produce toxins as waste products which are harmful to the host, and so when the population of the bacteria is large enough, the host will begin to produce vomit and diarrhea. This is a very serious problem in developing countries, where the World Health Organisation estimates that one third of the deaths of all children under the age of 5 can be attributed to this type of infection. Since infection involves specific protein protein binding, a knowledge of the structure of the adhesin wil l help those developing a cure for the disease. There is great diversity in the structure of bacterial adhesins, and very little detailed knowledge of their quaternary structure. Two adhesins from Escherichia coli have been examined using Xray crystalography (Fig IB). They are both right handed helical coils of a repeating proteinacious subunit. The molecular weight of these 'polymers' can be in the millions of daltons, and the subunits are found to have molecular weights of 18 to 30 kDa. There is controversial information in the literature on where the binding site for the host receptor is located on adhesins. It could be located at either end of the 'polymer', on the predominant structural monomer, or on a minor polypeptide interspersed along the 'polymer', but not directly associated with the predominant structural monomer. This controversy is addressed in this thesis by studying the adhesion of the F41 E.coli strain to human red cells, which are used to model 1 F i g 1A: A s c h e m a t i c d i a g r a m o f c e l l a d h e s i o n , a p r e l i m i n a r y s t e p o f c o l o n i s a t i o n by E n t e r o t o x i g e n i c B a c t e r i a . 7 nm 2nm F i g I B : A s c h e m a t i c d i a g r a m o f 5-o two f i m b r i a l a d h e s i n s t a k e n f r o m §* ' B a c t e r i a l A d h e s i o n 1 , D . C . S a v a g e , o M . F l e t c h e r , 1 9 8 5 , P l e n u m P r e s s . -Introduction / 3 the host cell in the gutt. The receptor for the F 41 adhesin on the red cell surface is known to be glycophorin, whose structure and function have been studied extensively. The structure of the adhesin is not well known. When it is examined using sodiumdodecylsulfate polyacrylamide gel electrophoresis (SDS PAGE) , which separates proteins by molecular weight, two types of traces are found (Fig 9). If the adhesin is put onto the gel in SDS, it does not enter the gel because it has a molecular weight of 10 3 to 10° kDa. If the adhesin is boiled in a solution of SDS for 3 to 5 minutes, two peaks are found, one large peak at 29 kDa, and one very small peak at 18 kDa. This suggests that the adhesin used in this study is a large polymer with one major and one minor repeating subunit. There is an example in the literature of a bacterial adhesin with a polypeptide which is not required for the production of the adhesin, but which must be present for the adhesin to bind to its receptor. Therefore the binding site on the F41 adhesin could be located on either of the two monomers. The methods used to date to investigate the binding between bacterial adhesins and their receptors, use cell-cell binding to observe the effects of modifying either the adhesin or its receptor. This study presents the first experiment where the adhesin receptor binding is observed in the absence of the cells. This is done by creating an equilibrium binding isotherm of v, the moles of glycophorin bound per mole of the 29 kDa monomer of the adhesin, vs L , the concentration of unbound glycophorin (Fig 2A). This isotherm can be replotted as a Scatchard plot of vfL vs v. v/L = n/k - v/k t F 41 is an arbitrary designation indicating the presence of a particular surface structure or antigen. Introduction / 4 V = rt V F i g 2 : The e q u i l i b r i u m b i n d i n g i s o t h e r m o f Vvs L o b t a i n e d u s i n g a d i a l y s i s e x p e r i m e n t , w h e r e u i s t h e m o l e s o f l i g a n d b o u n d p e r m o l e o f s u b s t r a t e , L i s t h e m o l e s o f u n b o u n d l i g a n d , a n d S i s t h e t o t a l amount o f s u b s t r a t e a v a i l a b l e f o r b i n d i n g . Introduction / 5 where k is the microscopic dissociation constant of the binding site, L is the moles of unbound ligand in solution, and n is the number of binding sites, which is also the saturation value of u in Fig 2A. If the binding sites on the adhesin are independent and identical, The Scatchard plot will give a straight line with a slope of -1/k. As L becomes very large, vFL tends to zero, and so the v intercept is equal to n. The value obtained for k will be accurate if the binding sites are independent and identical, but the value obtained for n will be independent of these restrictions, because it is a saturation value. The value obtained for n from the Scatchard plot is used in this study to determin the location of the binding site on the F41 adhesin. The data is adjusted so than n corresponds to the number of binding sites on the 29 kDa protein, which is the predominant monomer of the adhesin. Therefore, if n is close to 1, the binding site will be on the predominant monomer, and if n is much less than 1, the binding site will be located on some minor polypeptide, perhaps the 18 kDa protein shown on the SDS P A G E of Fig 9. The objective of the work described in this thesis was to help clarify the controversy surrounding the location of receptor binding sites on bacterial adhesins, by using an equilibrium binding experiment to find the location of the binding site for glycophorin on the adhesin of the F 41 E.coli. Binding isotherms can be obtained using dialysis experiments. These involve an incubation chamber which has been divided up into two compartments using a membrane with a known pore size (Fig 2B). The ligand equilibrates across the membrane, but the substrate is too large to pass through the pores, and so it is contained in one compartment. It is therefore possible to find the amount of bound and unbound ligand, since the unbound ligand equilibrates across Introduction / 6 the membrane. Dialysis was not used in this study, because glycophorin, the ligand, is too large to equilibrate across any commercially available membrane. The adhesin and the glycophorin were separated instead using aqueous polymer two-phase separation. A phase system was developed in which 97% of the adhesin was in the lower phase, while glycophorin was distributed between the two phases with a ratio close to unity. Therefore the separation of glycophorin in the two-phase system resembeled separation in a dialysis experiment, and was used to obtain the equilibrium binding isotherm for the two proteins. To summarise, the experiment described in this thesis involved developing a two-phase system for glycophorin and the adhesin, which was similar to a dialysis experiment. The proteins were radio-labelled, and a method was developed for counting the radiation in the presence of the phase systems. A Scatchard plot for glycophorin-adhesin binding was then obtained, using the two-phase system, and the v intercept was used to find n, the number of binding sites for glycophorin on the 29 kDa monomer of the F 41 adhesin. II. B A C K G R O U N D A N D T H E O R Y A . I N T R O D U C T I O N One of the fundamental questions in biology today concerns the nature of the language of cell-to-cell communication. Since this must involve an initial reception of a message on the surface of one of the cells, it is important to understand the nature of specific cellular recognition mechanisms (Keusch, 1979). One aspect of this process which is accessible to chemical investigation is the nature of cell-cell adhesion, which is now considered to be a crucial part of morphogenesis (Edelman, 1984), embryology (Steinberg, 1970) and bacterial colonisation (Bartus et al., 1985). The focus of this study is on the thermodynamics of the binding of one of the myriads of specific bacterial adhesins which appear today to be essential for the binding of an organism to a specific host tissue cell receptor (Bartus et al., 1985). B . C E L L U L A R A D H E S I O N Interactions between organisms and their hosts are so complex, they are almost sociological in nature (Keusch, 1979). Nevertheless, it is known that certain pathogens have an affinity for specific tissues, and that infection cannot occur, in some cases, without the pathogen first adhering to the host (Keusch, 1979). In fact with the E.coli bacterium, which is one of the predominant intestinal pathogens of man (Uhlin et al., 1985), this adhesive process is known to be mediated by proteinaceous surface appendages called adhesins (Salit and Gotschlich, 1977; Leffler and Spanporg-Eden, 1980; Jones and Isaacson, 1983; Labigne-Roussel et al., 1984). The receptors for these particular adhesins have 7 Background and Theory / 8 been identified in some cases as glycolipids (Svanberg-Eden et al., 1982) and sialoglycoconjugates (Bartus et al., 1985). These molecules are also receptors for the adhesion of most bacterial surface receptors as well as for viruses and parasites (Lindberg et al., 1984). Indeed saccharides are so ubiquitous in cell binding that (Feize, 1985) states: "The possibility must be considered seriously that saccharides are the much-sought surface markers that distinguish immature from mature cells and tumour cells from normal counterparts; that carbohydrate sequences, and/or some additional structures associated with them, may have specific roles in normal cell growth and differentiation; that inappropriate biosynthesis or processing of carbohydrate structures may contribute to the disordered behaviour of tumour cells." In this thesis, the adhesion of bacteria to a host cell, via specific protein saccharide interactions, is examined using an F41 E.coli bacteria, and by modeling the host cell with the human red blood cell. The red cell was used because although hemagglutination (the mutual adhesion of red blood cells to other cells mediated by intervening cell types or adhesins) is by no means a fullproof method of assaying bacterial adherence (Duguid et al., 1978; Chabanon et al., 1982), it has been used extensively in the past for monitoring bacterial cultures and proteins for adhesive properties (Jones and Isaacson, 1983). In addition, red blood cells are much easier to obtain and maintain than are host cells from the gut, and results obtained are often applicable to other cell systems (Bartus et al., 1985). For example, a disaccharide on the P blood group is the receptor for attaching E.coli bearing P-type adhesins to human urinary tract epithelial cells (Leffler and Spanporg-Eden, 1980). In the case of the F41 E.coli used in these experiments, hemagglutination and binding assays confirmed that a Background and Theory / 9 protein purified from the F41 bacterial surface was an adhesin, and that its receptor on the red cell surface was a glycoprotein called glycophorin (Brooks et al., 1987). The adhesin-glycophorin binding assay described in this thesis is a refinement which goes beyond the hemagglutination assay, allowing the adhesin -glycophorin binding to be observed in the absence of the cells. The assay involved finding an equilibrium binding isotherm using an aqueous polymer two-phase system to separate the radio-labelled proteins. This isotherm was used to address the complex question of the structure of bacterial surface adhesion, by finding the polypeptide on the adhesin involved in the adhesin - glycophorin binding. C . B A C T E R I A L S U R F A C E A D H E S I N S It used to be thought that bacterial surface adhesins were associated exclusively with fimbriae (Duguid, 1959). Fimbriae, or pili, of gram negative bacteria are filamentous appendages observed on the surface of bacteria by electron microscopy. They are proteinaceous polymers composed of identical subunits, with molecular weights of 18-30 kDa, and which are known in some cases to be arranged as right handed helical cylinders with diameters of 2-10 nm, and with lengths of up to 4 (im (Jones and Isaacson, 1983). The mechanism of polymerization of the monomers is unknown, although it is not covalent since they can be broken down into monomers with SDS and heat (Duguid, 1959). These fimbriae are classified according to their carbohydrate receptors. Mannose-sensitive (MS) or Mannose-resistant (MR) adhesins are classified according to whether or not their adhesivness is competitively inhibited by a-D mannose (Duguid, 1959). The "P" fimbriae, which compose 90% of the Background and Theory / 10 M R adhesins, are known to bind to the disaccharide aDGalp-(l-4)-/3-DGalp which is part of the P blood group antigen (Lindberg et al., 1984). The P fimbriae, which are associated with the intestinal disease pyelonephritis, are called P A P fimbriae (Norgren et al., 1984). Adhesion by another type, S fimbriae, is inhibited by sialic acid (Knapp et al., 1986). It is difficult to determine where the adhesive component of the fimbriae lies because it is hard to dissociate the monomer without irreversible loss of activity. There is an exception (Eschdat et al., 1981), but even then, only 25% of the reassociated dimers appeared to be adhesive. Another exception is the case of the gonnoccocal adhesin where a CNBr fragment from the monomer showed adhesive character (Gubish et al., 1982), and inhibited hemagglutination of the cells (Schoolnik et al., 1982). Also, sera raised against a synthetic peptide common to many gonnoccocal adhesin monomers inhibited binding of a heterologous strain to unidentified receptors on endometrial carcinoma cells (Rothbard et al., 1985). It has recently become evident that adhesion of E.coli to human cell lines does not always depend upon the presence of fimbriae, but may involve afimbrial adhesins (Salit and Gotschlich, 1977; Chabanon et al., 1982). One such adhesin has been isolated, purified and partially characterised (Sheladia et al., 1982). Another with P type activity has been genetically identified (Labigne-Roussel et al., 1984). " X adhesins" have been found which which do not exhibit P type hemagglutination (Labigne-Roussel et al., 1985), and surface proteins on gonnoccocal strains have been found with adhesive character (Lambden et al., 1979). The adhesive strategies of E.coli appear even more complex with the genetic evidence showing that in the case of P and S fimbriae, there are two Background and Theory / 11 distinct D N A stretches which code for either fimbriae formation, or for bacterial hemagglutination (Knapp et al., 1986). In addition, Type I and P A P strains were found which were as ineffective in bladder colonisation as were mutants lacking the entire fimbrial structure (Keith et al., 1986). It is known that with the P A P strains, the mutation which causes this change is not in the D N A sequence coding for the fimbriae's monomer, but is found instead in another gene common to several different bacterial strains (Lindberg et al., 1984; Uhlin et al., 1985; Van Die et al., 1986). It is suggested that this second gene containing the adhesive character may code for a protein which is transported outside the carbohydrate environment of the cell's surface by the fimbriae (Van Die et al., 1986), and that it may be associated with a third protein which could be either forming a second, so far undetected fimbriae, containing more of the adhesins than the first fimbriae, or acting as a carrier for the adhesin on the first fimbriae (Lindberg et al., 1986b) To summarise, it appears as though adhesion of bacterial cells is mediated by both fimbrial and afimbrial surface proteins. Fimbriae are made up of repeating monomers, which may act as, or be associated with adhesins, or they may be used to transport adhesins away from the surface of the bacterial cell wall. Since the location and frequency of occurence of the adhesin on these fibers will affect the nature of the binding of the bacteria to its host cell, understanding the quaternary structure of these macromolecules could help those devising techniques for curing infected humans and livestock. This thesis presents a thermodynamic study of the binding of an afimbrial adhesin with its receptor from the human red blood cell, in order to determine where and how often the binding site is found on this particular adhesin. Background and Theory / 12 D . T H E B I N D I N G P R O T E I N A N D ITS R E C E P T O R 1. The Af imbr i a l Adhes in The afimbrial adhesin studied in this thesis is from an E.coli strain isolated from piglets with diarrhea, which reacts with antisera designated F41. The F41 E.coli are reported to carry fimbriae with a diameter of 3.2 nm, an isoelectric point at pH 4.6, and a monomer molecular weight of 29.5 kDa (Gastra and deGraaf, 1982). The F41 strain used in this thesis is found to exhibit few if any fimbriae when examined under the electron microscope. The protein co-isolating with the agglutinating activity has a molecular weight in excess of 10 7 daltons, but can be broken down into monomers with a molecular weight of 29 kDa by boiling in a solution of SDS. Analysis using SDS polyacrylamide gel electrophoresis (PAGE) occasionally reveals an additional trace amount of an 18 kDa protein. Therefore the adhesive activity could be due to the predominant 29 kDa 'monomer', as in the case of the gonnoccocal adhesins, or to the trace amount of the 18 kDa protein, as in the case of the P A P adhesin. The binding study described in this thesis was designed to find which protein was responsible for the adhesiveness by determining the number of binding sites on the adhesin using a thermodynamic approach. If this number turned out to be equal to the number of monomers making up an adhesin, then this F41 strain would be the first E.coli to be found with the proteinacious adhesive activity on the monomer of the adhesin macromolecule. Background and Theory / 13 2. Glycophorin The receptor used in the binding assay of this thesis was glycophorin, which is an integral sialoglycoprotein t found in the human red blood cell membrane. Glycosilated receptors have been found for many adhesins, among them the S fimbriae (Jones and Isaacson, 1983), and the Type I, the P and the gonnoccocal fimbriae described above. Glycophorin is one of a number of transmembrane glycoproteins in the human red blood cell membrane which link the lipid bilayer to the proteinaceous cytoskeleton lying underneath it. The cytoskeleton is thought to give the red blood cell membrane its remarkable strength and elasticty, and to regulate the lateral mobility of the membrane proteins (Wise, 1984). The function of glycophorin remains unknown since genetic changes causing a decrease or lack of the protein appear to be of little biological consequence in humans (Furthmayr and Marchesi, 1983b). Glycophorin's primary amino acid structure (Tomita, 1978), and molecular weight (Furthmayr and Marchesi, 1976) are known. The protein can be divided up into 3 sections with a central hydrophobic segment lying in the lipid bilayer separating two hydrophylic sections, one inside and the other outside the bilayer. The latter section contains the amino terminal and is associated with the carbohydrates which makes up 65% of the glycoproteins^ mass, and whose tertiary structures are known (Springer et al., 1978). The carbohydrates which are associated with lipids and proteins are only found on the exterior of the red blood cell membrane. Those on glycophorin are very flexible (Egmond et al., 1979), and, since they contain sialic acid,, they could be creating a negatively charged layer which might then t Sialoglycoproteins have saccharides covalently bound to them which contain N-acetylneuraminic acid, otherwise known as sialic acid. Background and Theory / 14 transmit electrical or mechanical constraints on mechanical motion to the inside of the membrane. In addition the carboxyl end of the glycophorin is linked to the cytoskeleton (Anderson and Lovrien, 1984; Anderson and Marchesi, 1985), so the protein may help in maintaining cell shape, and assisting in communication between the exterior and the interior of the membrane (Egmond et al., 1979). It is also thought that the carboxyl end of the protein could be a haemoglobin binding sight (Rauenbuehler et al., 1982). It is known that the carbohydrates participate in the binding of the influenza virus, as well as in binding of certain plant and bacterial hemagglutinins (Steck, 1974). For example, hemagglutination induced by K99 E.coli isolated from pigs was shown to be inhibited by N-acetylgalactosamine which is found on both glycophorin and on the surface of the small intestine of pigs (Lindberg et al., 1986). The S fimbriated E.coli causing sepsis and meningitis in newborn infants was found to have hemagglutinating activty which was inhibited by glycophorin. This inhibition was abolished when sialic acid on glycophorin was removed, and was restored with the enzymatic replacement of sialic acid with a2-3Gal/3 l-3GalNac, which is one of the disaccharides on glycophorin (Parkkinen et al., 1986). Glycophorin was also identified as a receptor whose activity was weakened in the presence of serine, N H 2 terminal C N B r fragments of glycophorin, and by oligosaccharides from glycophorin (Jokinen et al., 1985). Glycophorin is therefore an integral transmembrane glycoprotein of the human red blood cell which has been the object of extensive structural and functional studies, and which appears to contain several binding sites for both bacterial and viral surface adhesins. The binding of glycophorin to the afimbrial F41+ bacterial surface adhesin is studied in this thesis to obtain a more Background and Theory / 15 detailed knowledge of the structural specificty of one of the crucial steps in the colonisation and infection by this particular pathogen. E . T H E R A P E U T I C I M P L I C A T I O N S O F A D H E S I N R E S E A R C H It has been proposed that therapeutic approaches be based upon the idea that adhesion is essential for colonisation and infection (Keusch, 1979). This theory is known to be true in the case of pyelonephritis (Labigne-Roussel et al., 1984) and a type I adhesin (Salit and Gotschlich, 1977). It is strongly supported by studies showing that attachment of enteroinvasive E.coli on cultured cells (Knutton et al., 1984), and uropathogenicity (Fuenfstueck et al., 1986) involve at least the bacterial glycocalyx, hemolysin formation, Type I fimbriae and M R hemagglutination. Devising therapeutic strategies is not trivial however because bacterial interactions are extremely complex. For example the polysaccharide receptors can be arranged in a vast number of distinct anomeric configurations. Some bacteria are known to produce several adhesins (Wadstrom and Trust, 1984; Jones and Isaacson, 1983). It is possible that some adhesins are synthesised after initial, temporary addsorbtion to a host cell (Jones, 1980). Adhesiveness and hemolysis could be genetically linked (Knapp et al., 1986), and evidence exists that genes coding for M R hemagglutination may reside on transmissible plasmids, allowing easy transfer of the gene from one bacteria to another (Hales and Amyes, 1986). Nevertheless, several therepeutic techniques based upon the adhesin-saccharide reaction have been proposed (Keush, 1979; Svanberg-Eden et al., 1982), and in some cases implemented with success (Contrepois et al., 1982; Jones and Isaacson, 1983; Neeser et al., 1986; Ekback et al., 1986). These studies also indicate that the binding strength of a Background and Theory / 16 pathogen's surface adhesin increases dramatically as the structure of a carbohydrate receptor approaches that of some specific oligosaccharide. Therapeutic techniques for diseased livestock and for humans will therefore become increasingly effective as our knowledge of specific adhesive mechanisms improves. This thesis presents one such study where, for the first time, the adhesive process of an E.coli adhesin is studied from a thermodynamic perspective. F . T H E E Q U I L I B R I U M B I N D I N G I S O T H E R M The experiment described in this thesis established an assay to provide the equilibrium binding isotherm for the reaction between the adhesive polymer from the afimbrial F41+ E.coli, and its receptor on the human red blood cell, glycophorin. The data was interpreted, using the model proposed by Scatchard, to relate n, the number of binding sites on the 29 kDa monomer, and k, the microscopic dissociation constant for the binding site, to L , the equilibrium concentration of unbound ligand, and v, average number of moles of ligand bound per mole of 29 kDa monomer in solution. This was done using equation 1, which is derived below (Cantor and Schimmel, 1980) assuming that the binding sites are identical and noninteracting. *»/L = n / k - v/k [ 1 ] If a macromolecule has n binding sites for a ligand, then equations 2a to 2c will relate the macroscopic dissociation constant, K . , to the concentration of macromolecule, M . , where i denotes the total number of ligand molecules that are bound. Background and Theory / 17 M 0 + L = M , ; K , = g ^ y M , = f& [2a] M , + L - M „ K 1 . ^ - ^ i , M 1 . ^ [2b] Equation 3 relates v to M . . M . = -f^ [2c] . n K . I i [ M . ] v = J ^ O L _ [ 3 ] . 1 [ M . ] 1 = 0 1 The goal is to express v in terms of the microscopic dissociation constant, k. which is written in terms of m., the concentration of any microscopic species bearing i ligands, in equation 4. l ( i - n + L - m i ? k i • ^ ( i - n ^ i ^ The microscopic species m., is the concentration of the substrate when i ligands are bound in a specific configuration. The macroscopic species M . , is the average concentration of all the microscopic species with i binding sites occupied by the ligand. If one assumes that each site is independent of every other site, and that there are enough species present to make each configuration of a given m. available at the same concentration, then equation 5a holds. I m i = ( M . ) / Q j [5a] Background and Theory / 18 The denominator of [5 a] is an expression for the number of different microscopic states which bind i ligands to n sites. 0_ • = ( n ! ) / ( n - i ) ! i ! [5b] n , I Equation 6 is derived from equtations 4 and 5a. K. = [ U . ) - ( i ) ] / [ n - i + 1] [6] Equation 8 is obtained from equations 2c, 6 and 7, if it is assumed that k. is the same for each site. n ( M ± l ) = [ n ! ] / [ ( n - i ) ! i ! ] [ 7 ] D = 1 D M . = [ M q L 1 ! ! ! ] / [ k 1 ( n - i ) ! i ! ] [ 8 ] Combining equation 7 and 3 gives equation 9. 9 n! , 1 M 1 . £ 1-7 r - r - r - r - ( - ) v = i = 1 ( n - i ) ! i ! k 1 * . 8 . ( i ) 1 1 9 1 i = 1 ( n - i ) ! 1 ! k The denominator of [9] is the binomial expansion of [ l + (L ) /k ] , and differentiation of this expansion with respect to (L/k), followed by multiplication by (L/k), gives the numerator of [9]. Therefore equation 9 can be expressed as equations 10a and 10b, which are identical to equation 1. ( n L / k ) ( 1 + L / k ) (1 + L / k ) n Background and Theory / 19 n-1 v = [10a] v = [ n L ] / [ k + L ] [10b] The 'Scatchard plot' of vfL vs v will then give a straight line for data which obey this model, with a slope of - l /k and a v-intercept of n. In practice, Scatchard plots may have some curvature due to varying degrees of cooperativity between identical binding sites, or due to cases where there are several independent classes of binding sites with different microscopic dissociation constants. However, the degree to which this curvature can be interpreted depends upon the amount of experimental error, and the extent to which the experimental points cover the saturation curve (Noerby et al., 1980). G . T H E T W O - P H A S E S E P A R A T I O N T E C H N I Q U E 1. Introduction The experimental method used to provide an equilibrium binding isotherm must allow the identification and estimation of bound and unbound reactants, while maintaining equilibrium conditions. The standard procedure with proteins has been to use sealed nitrocellulose dialysis membrane tubing with a pore size chosen to contain the protein while allowing the ligand to equilibrate across it. The concentration of the ligand outside the membrane is then used to obtain the concentration of unbound ligand inside the membrane, which is subtracted from the total ligand concentration inside the membrane, to give the amount of ligand bound to the protein. There is an upper limit, however, to the pore size of Background and Theory / 20 these membranes, and this dialysis technique is no longer an option when the ligand is the size of glycophorin. Since techniques involving chromatography, sedimentation, and filtration, do not respect the equilibrium requirements of binding experiments, very few possibilities remain. One option which has been gaining increasing favour is the separation of proteins in dilute aqueous two-phase systems containing dextran and poly(ethylene glycol) (PEG). Because the lower phase of these systems is rich in dextran, and the upper phase is rich in P E G , proteins can usually be separated if their surface properties are different. Moreover, the free energy of mixing of a protein with each phase can be changed by varying ionic strength and polymer concentration. Some salts, such as phosphate, distribute unequally in the systems, producing an electrostatic potential difference between the phases. Since the degree of ionization of a protein can be changed with pH, it is possible to fine tune a two-phase system in order to adjust the partition of different proteins. If a system can be found where one protein is almost entirely in one phase, while another is distributed between both phases with a known partition coefficient, then an isotherm can be obtained using the same approach as with dialysis as described above. One possible drawback of this technique is the interference that the two polymers may have on the protein binding of interest. It has been suggested, for example (Berglund et al., 1984), that association could increase due to the excluded volume effect described by Brooks et al. (Brooks et al., 1985). However previous experiments involving adenine nucleotide binding to formyltetrahydrofolate (Curthoys and Rubinowitz, 1971), estradol-receptor-DNA interactions (Albergu et al., 1976), and antigen antibody interaction (Urios et al., 1982) have shown that results obtained using dilute aqueous dextran-PEG Background and Theory / 21 two-phase systems do not differ from results obtained using dialysis or fluorescence polarisation techniques. Indeed, of the three techniques, two-phase separation had the greatest potential for sensitivity (Urios et al., 1982) and adaptability (Albergu et al., 1976). In fact aqueous polymer two-phase systems have been used successfully for the analysis of proteins, nucleic acids, bacteria, viruses and phages, and plant and mammalian cells, membranes and organelles, since the late 1950s (Brooks et al., 1985). 2. A Statistical Mechanical Description of Phase Separation The process of phase separation of aqueous polymer two-phase systems has been described using the statistical mechanical approach of Flory-Huggins theory (Brooks et al., 1985). The first step is to find an expression for AGm, the free energy of mixing associated with the formation of a polymer solution from pure components. AGm = AHm - TASm [11] In the Flory Huggins model, the solution is represented by a lattice of sites where each lattice site has z contacting faces with adjacent sites, and each lattice site contains either solvent or polymer, with the size of the lattice site being determined by the solvent. This imposes the assumption that no volume change can occur upon mixing, and it treats the polymer as P segments, each of which occupy a lattice site. In the case of a single type of polymer in a solvent, the change in enthalpy per lattice face when the two components are brought out of the pure Background and Theory / 22 state into the mixed state is AW, 2 = W, 2 - (W, , + W 2 2 ) / 2 [12] where W ^  j is the enthalpy change per contact for component i and j . Therefore AHm = (ZAW, 2 )X [13] where X is the fraction of the faces which have changed neighbors upon mixing. Approximately, X = n , 0 2 [14] where 6 2 is the fraction of lattice sites occupied by a polymer segment, and n , is the total number of solvent molecules on a lattice. Therefore AHm= k T x 1 2 n i 0 2 [15] where X 1 2 = ( Z A W 1 2 ) / k T [16] where k is the Boltzmann constant, and T is the temperature in degrees Kelvin. The entropy expression is derived from its classic form ASm = k l n W [17] where W is the number of distinguishable ways of arranging n 1 solvent molecules and n 2 polymer segments. This turns out to be described by equation 18. ASm = -kCrnlnfl, + n 2 l n 0 2 ) [18] Therefore combining equations 18, 15 and 11 gives equation 19a. AGm = k T d ^ l n f l , + n 2 l n 0 2 + X n t i i ^ ) [19a] Note that 0., a unit of concentration, is the volume fraction of component i . Therefore 8^ + 82 - 1, and l n # . is always negative. The sign of AGm, or the solubility of component 2, is therefore determined by X i 2 > which describes Background and Theory / 23 the energy resulting from the change of environment around the solvent molecule and the polymer segment when these components are taken from their pure state into the mixed state. This Flory-Huggins model can be extended to provide an expression for AGm for a three component system. AGm = k T [ n , l n 0 , + n 2 l n 0 2 + n 3 l n 0 3 + [19b] (n , + n 2 P 2 + n 3 P 3 ) ( 0 , 0 2 X i 2 + ^ ^ X n + ^ 2 ^ 3 X 2 3 ) ] where P. are relative molecular volumes as described above; n. is the number of 1 1 molecules of component i on the lattice; x^ j = zAW^ j / k T , i = 1,2 o r 3 ; AW. .=W. . - ( W . .+W . . ) / 2 , i , j = 1 , 2 o r 3 ; and n n n 33 8. = ^ i -1 n , + n 2 P 2 + n 3 P 3 This can then be used to obtain an expression for the critical conditions of phase separation applying the conditions that the first and second derivatives of the chemical potential, Mj, of component i with respect to 6^ are zero at the critical point: = M? + N a ( 9 A G m / 3 n . ) n . f T f P [20] 3 M i / 3 0 2 = 3 2 M i / 3 2 0 2 = 0 [21] For example, for the symmetrical case o f P 2 = P3> Xi 2 = X 1 3 1 #2 = ^3> solution of these equations gives for the critical values of X 2 3 a n < ^ 0 ^ : Background and Theory / 24 d 2,c = d 3 , c = 0 - 0 1 ,c ) /2 [22] X 2 3 , c = 1 / ( P 2 0 ) [23] where the subscript c implies the critical value of the relevant parameter. Recall that P 2 is the number of segments on a polymer which are the same size as the solvent molecule, and therefore parametrises molecular weight, and that X 2 3 parametrises the energy required to replace like polymer interactions with unlike polymer interactions. Equation 23 therefore states that when the Flory-Huggins model is adopted, phase separation occurs when x 2 3 is positive, or in other words, when interactions between unlike polymers are less favourable than interactions between like polymers. It also shows that as the molecular weight of the polymers increases, this difference needs to be less and less for phase separation to occur, and that phase separation is not dependent upon solvent polymer interactions. Since this turns out to be a good description of the general features of aqueous polymer two-phase systems which are not very dilute (Gustafsson and Wennerstrom, 1986), the Flory-Huggins model provides a useful framework for understanding phase separation in these systems. 3. A Thermodynamic Description of Solute Partition A treatment of partitioning of a solute in a two-phase system, independent of any model, can be provided by a thermodynamic approach. This is based upon the expression for the chemical potential of a neutral solute molecule in one phase of a two-phase system, described by [24]. Ui = M? + R T l n ( f i C i ) [24] Background and Theory / 25 Where R is the gas constant, f. is the activity coefficient, and C. is the concentration. At equilibrium, JI. is the same in both phases, therefore [25] and [26] are obtained using the superscripts t and b to indicate top and bottom phase: M? f c + R T l n ( C ^ f J ) = M ° b + R T l n ( C ^ f ^ ) [25] Therefore, l n ( c j / c £ > - U ? b " * f ) / R T - l n ( f £ / f £ ) [26] Since the partition coefficient K = C f c /d°, and since for dilute systems f. = 1, K can be described by [27]. K = exp [ -U? f c - uf3) / RT ] [27] Furthermore, since some salts, such as phosphate, distribute unequally in the systems, an electrostatic potential difference Ai|/ = - i / / ^ , can be produced between the two phases. The effect of K on Ai/f is described by Brooks et al. (1985), by rewriting [25] as [25-1]. M? f c + R T l n ( C ^ f J ) + z .Ftf* = uf3 + R T l n ( c W ) + z^F^ [25-1] where z. is the net charge of the protein; F the Faraday constant = N e; e I a the electron charge; and ^ and the electrostatic potential in the top and bottom phases. The dependence of the partition coefficient on Ai£ is then Background and Theory / 26 described by [27-1] below. K ( A t f ) = K ( 0 ) • exp[-z F A ^ / R T ] [ 2 7 - 1 ] P where K(0) is from [27]; z is the charge of the protein. P Hence in the case of the neutral solute, the partition coefficient depends exponentially upon the difference in the standard state chemical potential of the two solutes, in the two phases. This difference represents the difference in the interaction of the solute with the two phases, due largely to differences in solute-polymer interactions. These differences can be amplified in a given system to optimise the separation of different particles using the extensive experimental data available on the properties of phase systems which is summarised below for the system used in this experiment. The effect of an electrostatic potential between the two phases can be used to adjust the partition coefficient of the protein, since the charge of a protein can be changed by varying the pH of the buffers in the phase system. Furthermore, the polymer solutions which have been adopted produce interfacial tensions and aqueous environments which are compatible not only with proteins, but with plant, bacterial and mammalian cells. Separation in aqueous polymer two-phase systems is therefore a powerful and versatile technique, and in the present case, suitable for the separation and subsequent analysis of the binding between the afimbrial F41+ adhesin and glycophorin. III. PRELIMINARY EXPERIMENTS A. INTRODUCTION The purpose of the work described in this thesis was to develop an experimental procedure capable of providing the isotherm for the equilibrium binding between a surface adhesin from an F41+ E.coli strain, and a glycophorin from the human red blood cell membrane, and to determine from this whether the glycoprotein was binding to the repeating subunit, or to a minor polypeptide of the adhesin, or perhaps to the terminus of the adhesin structure. The strategy which was adopted was to create a boundry using an aqueous polymer two-phase system, in which the adhesin was localised in one phase but in which glycophorin distributed more evenly. In the system chosen, the lower phase contained about 50% of the glycophorin, 97% of the adhesin, and therefore about 97% of the adhesin glycophorin complex. The amount of glycophorin bound to the adhesin was obtained using the amount of glycophorin in the upper phase, and its partition coefficient between the two phases, to obtain the amount of unbound glycophorin in the lower phase, which was subtracted from the total glycophorin in the lower phase, to give the amount of bound glycophorin. The data for the binding isotherm was obtained by examining the amount of glycophorin bound as a function of glycophorin concentration, and manipulated using the model proposed by Scatchard, to give an estimate for n, the number of glycophorins bound per adhesin monomer. Before this strategy could be adopted, it was necessary to develop both a two-phase system with the required partition coefficients which would not aggregate the proteins or adversely effect their chemical activities, and a labelling 27 Preliminary Experiments / 28 technique with a low enough uncertainty to allow a meaningful interpretation of the Scatchard plot. The two-phase system and the labelling techniques are described in the chapter called Preliminary Experiments, and the method used to combine these two aspects of the binding experiment is described in the chapter called Isotherm Experiments. The results are discussed in the last chapter. B. PROTEIN ISOLATION The first step in any binding experiment is to isolate the ligand and the substrate. In this experiment, the ligand was glycophorin, which has the antigenic site for the M and N blood types, and the substrate was the afimbrial surface adhesin coisolating with the hemagglutinating activity of the F41+ E.coli strain 1593. 1. Glycophorin Isolation Glycophorin was prepared according to the method of Furthmayr and Marchesi (Furthmayr and Marchesi, 1983a) using ghosts prepared from M M typet red blood cells. Two batches were isolated. Batch A was isolated in July 1984, and batch B was isolated in January 1986. a. Materials LIS (Lithium diiodosalicylic acid) was from Sigma, catalogue number D3635 as was Tris-HCl (Tris(hydroxymethyl) aminomethane hydrochloride), Sigma catalogue # T-3253. Dialysis tubing was from Spectrapor, E G T A (ethylene bis t M M is an arbitrary designation indicating the presence of a specific antigenic site on the surface of the cell Preliminary Experiments / 29 (oxyethylenenitrilo) tetracetic acid) was from Eastmman Kodak Co. catalogue # 8276; outdated blood was generously donated by the Red Cross. b. Methods Unless otherwise stated, all solutions and the centrifuge (Dupont-Sorvall RC5) were kept at 4 ° C . Whole blood was spun at 2,000 xg for 10 min, the plasma was removed and the cells were resuspended in three volumes of PBS (table 3) and centrifuged for 10 min at 2,000 xg. After repeating this three times, the red cells were poured into 30 volumes of ice cold phosphate buffer ( lOmM pH 8.0), mixed and left to stand for 1.0 hour, centrifuged at 1,000 xg for 30 min and the pellet of cells was resuspended in phosphate buffer (lOmM pH 8.0) and washed repeatedly until it was a clear pearly white. These cells, now called ghosts, were finally suspended in distilled water and freeze dried. The ghost cells were added to a solution of 0.3M LIS in 50mM Tris-HCl, pH 7.5 (25mg protein/ml), the solution was stirred at room temperature for fifteen min, (light protected), two volumes of distilled water were added, and the mixture was stirred 10 min (light protected). An equal volume of fresh 50% w/v phenol/water was added to the supernatant, the mixture was stirred for 15 min, and centrifuged at 4,000 xg for one hour. The upper phase was then dialysed extensively against distilled water. After being lyophilized, the protein was suspended in 100% ethanol (O.lml/mg protein), stirred for 1-2 hr to remove excess LIS, and centrifuged at 4,000 xg for 45 min to collect the protein. The ethanol extraction was repeated once more, and then the precipitate was suspended in distilled water, dialysed overnight, centrifuged at 4,000 xg for 45 min, and then lyophilized. Glycophorin solutions were then prepared, and the concentrations determined by the weight of protein dissolved in buffer with an accuracy of 0.0003 g. Preliminary Experiments / 30 2. Adhesin Isolation The F41 adhesin solution was kindly prepared and donated by Dr. J Janzen. The adhesin was isolated by him using bacterial strain 1593 originating from Dr. R.E . Isaacson, University of Michagan, (Lindahl and Wadstrom, 1983), using the following technique. The F41 bacteria were grown using as the culture medium 39 x 26 cm stainless steel trays containing 250 ml of minca gel per tray. After culturing the bacteria for 21 hr at 37 °C , the 'bugcrit't of the surface fluid was determined and the 169.4 grams of cells which were removed from the trays were diluted to a 'bugcrit' of 4.67%. The samples were then sheared in 30 ml aliquots in 50 ml conical polypropylene tubes for 2 min using a vortex mixer. The E.coli were sedimented by centrifugation for 15 min at 3,670 xg at 4 ° C , and the supernatant was poured off and dialysed against ethanolamine-HCl (adjusted to pH 8.4 with NaOH) plus 10% ammonium sulphate, using 6000 -8000 molecular weight cut-off tubing which had been boiled for 10 min in 10 m M E G T A . The sample was then dialysed for three days with several changes of the dialysing buffer, partially concentrated with 20,000 molecular weight P E G to about 350 ml, and the dialysis continued against 2.0 1 of fresh pH 8.4 buffer. The dialysate was then centrifuged at 4 ° C for 20 min at 3,670 xg, the supernatants were pipetted off and the adhesin precipitate, which could be seen as a thin cohesive layer at the bottom of the tubes, was freed from the tube by gently pipetting off the supernatant fluid. The pellets were collected and resuspended in 6 ml of PBS buffer containing 3.0 m M sodium azide adjusted to tthe concentration of the cells in % v/v after 5 min of centrifugation at 9,000 xg Preliminary Experiments / 31 pH 7.3, and dialysed against 2 1 of the same buffer at 4 ° C using 6000 - 8000 molecular weight cut-off tubing. The dialysed adhesin sample was frozen at -70°C. The adhesin prepared in this way had to be concentrated. An absorbance spectrum of redissolved lypholised adhesin was found to have a broad absorbance band between 240 and 320 nanometers. Since this may have been due to scattering from large aggregates, adhesin was concentrated using the 8MC micro-ultrafiltration system by Amicon using a 30,000 molecular weight filter. Protein that was found to adhere to this Filter was redissolved by spinning the stirring bar in the concentration chamber of the Amicon filter. The concentration of the adhesin in the solution was then determined from its absorbance using a Hewlett Packard 8450 uv/vis spectrophotometer. The extinction coefficient of the adhesin was found by freeze drying 4.50 ml of adhesin which had been dialysed against distilled water and found to have an absorbance of 0.555 at 280 nm. The mass was found to be 1.9 ±0 .1 mg, and so the extinction coefficient was taken as 1.31 ± 0.07 ml/mg-cm. 3. Assessment of the Bioactivity of the Proteins a. Methods Microtitre assays were done to determine the activity of adhesin and glycophorin. The adhesin was used to induce agglutination of fresh red blood cells and glycophorin was used to inhibit this agglutination. A 2% hematocrit of fresh N N type blood cells was washed three times with PBS in an Eppendorf tube as described above. Serial dilutions of 0.050 ml of the adhesin sample in PBS were made in one row of a microtitre plate containing 12 columns of Preliminary Experiments / 32 0.2 ml wells, 0.050 ml of the washed red cell suspension were added to each well, and the mixture incubated for three hr at room temperature. During the incubation time, any red cells which had not agglutinated collected at the bottom of the conical microtitre well leaving a red spot at the center of the well, whereas the agglutinated red cells were spread over the entire surface of the bottom of the well. It was therefore possible by visual inspection to select the approximate concentration of adhesin which would agglutinate the red cells. The concentration of glycophorin which would inhibit the binding of the adhesin with the red cells was found using an inhibition assay. Serial dilutions of 0.050 ml of glycophorin in PBS were mixed in a microtitre plate with adhesin dissolved at twice the concentration which was found to agglutinate the red cell sample, and the mixture incubated for one hr. To these solutions were then added 0.050 ml of the N N type cells, and the mixture left to incubate for two hr. The activity of a glycophorin sample was estimated from the concentration below which it failed to inhibit the hemagglutination caused by the adhesin. b. Results In a 1 % hematocrit of type N N human blood, agglutination began when the adhesin concentration was 0.0025 mg/ml. Under these conditions glycophorin continued to inhibit agglutination until its concentration fell below 0.001 mg/ml. Preliminary Experiments / 33 4. Protein Electrophoresis The molecular weight of a protein can be found using SDS polyacrylamide electrophoresis (PAGE) (Ornstein, 1964; Davis, 1964). This involves cleaving sulfur-sulfur bonds with 2-mercaptoethanol and then denaturing the protein by boiling in SDS. The protein is then placed in a voltage gradient which forces it through an aqueous polyacrylamide gel. Since the denaturing solution dissolves the protein as a filament with a uniform SDS/protein weight ratio, the protein acquires a uniform charge density, and so the distance traveled by the protein in a given time is determined by its size-dependent seiving by the gel matrix. The molecular weight-mobility relationship is determined by running each gel with a reference dye, and by running simultaneously a gel containing the dye and a range of proteins with known molecular weights. A calibration curve of the distance travelled by the reference proteins relative to the distance travelled by the reference dye is then used to determine the mass of the sample. The technique requires that the reference proteins and the sample behave identically in the denaturing solution. Glycophorin, however, is 60% saccharide, and SDS does not coat it with the same charge/weight ratio as the reference proteins, and so its electrophoretic behaviour is anomalous (Silverberg, 1978). Therefore it can only be identified by comparing a sample with one known to contain glycophorin. The proteins can be visualised by staining techniques. Coomassie Blue is a cationic dye which binds largely to amines (Fazekas et al., 1963). Glycophorins do not stain well with this dye since they are low in amino acids with free amines. They are stained instead with Schiff, which reacts with aldehydes produced by periodate oxidation of carbohydrates with vicinal diols (Kiernan, 1981). Preliminary Experiments / 34 a. Materials SDS (sodium dodecyl sulfate), Acrylamide, BIS (N,N-methylene-bis-acrylamide), T E M E D (N,N,N ' ,N ' , tetramethylethylenediamine) and ammonium persulfate, were Electrophoresis Purity Reagents from Bio-Rad. Basic Fuchsin, Technical Grade, was from Fisher. Coomassie Brilliant Blue G-250, was from Kodak, lot #B9H, Pyronin Y , from Baker Chemical Co., Lot #923344. Photo-flo 200 solution, was from Kodak Canada Inc.; catalogue # 146-4510. b. Methods Fairbanks gels were prepared according to Thompson and Maddy (Thompson and Maddy, 1982). Gels were run in a rod gel electrophoresis apparatus from Bio-Rad. Proteins were dissolved in water to a concentration of 2mg/ml for non-glycosylated protein, and l lmg/ml for glycoprotein, and mixed 1:1 with the sample reagent. The samples which were reduced were boiled for 5 min in the sample reagent (Table 1) which also contained 0.014M 2-mercaptoethanol. The 3% gel stock solution was then poured into clean dry tubes, and the overlaying buffer was carefully added to the top of the gels. When the gels had set, 10 yd of adhesin and 20 / i l of glycophorin were applied to the top of the gel. A voltage was applied across the gels at a constant current of 5 ma at a power of 0.1 watts per tube, and after 0.5 hr, the power was increased to 0.8 watts per tube. When the tracer dye had come close to the bottom of the gels, they were removed from the tubes and stained as described below. Solutions used for staining: Coomassie Blue Stain: Coomassie Blue G250 0.200g, methanol 28 ml, glacial acetic acid 5 ml, perchloric acid (70%) 25 ml, water 442 ml, Preliminary Experiments / 35 Table 1 : Solutions used to make 3% SDS P A G E gels. FAIRBANKS SAMPLE REAGENT Tris .HCl 0.02M E D T A 0.002M SDS 2.0% w/w sucrose 14.0% Pyronin Y 0.004% 10X TRIS-HC1 BUFFER Tris .HCl 24.23g Sodium Acetate 13.61g E D T A 3.72g water up to 500 ml 3% GEL STOCK water 18 ml 4% SDS 1.5 ml 0.5% T E M E D 1.5 ml lOxTris .HCl 3 ml 10xAcrylamide:bis 3 ml 1.5% Ammonium 3 ml Persulfate RESERVOIR BUFFER 10X tris.HCl 200 ml 4% SDS 100 ml water 1700 ml OVERLAYING BUFFER This was the same as the 3% gel solution with the acrylamide:bis replaced with water. 10 x ACRYLAMIDErBIS Acrylamide 40 g Bis 1.5 g Water to 100 ml Methanol Destain: methanol/water/concentrated acetic acid, 6/13/1 by volume. Schiff Reagent: Schiff stain was prepared by dissolving 1 g of Basic Fuchsin in 200 ml of boiling distilled water, stirred for 5 min and cooled to 50°C. The mixture was filtered and 20 ml of 1 N HC1 added. After cooling to 25°C 1 g of sodium or potassium metabisulfite was added, and the solution left standing in the dark for 14 to 24 hr. Two grams of activated charcoal were then added, and the solution shaken for 1 min, filtered and stored at 4 ° C . Staining method: Staining with Comassie Blue: The gels were washed in the Coomassie Blue stain for 1 hour, the methanol destain 1 hour, then in 7% acetic acid overnight. The process was then repeated and finally the gels were stored Preliminary Experiments / 36 in distilled water. Staining with SchifT* stain: The gels were washed in 10% acetic acid overnight, and then incubated 2 hr in 0.5% periodic acid. This was followed by two washes in 0.5% sodium arsenite in 5% acetic acid for 30 min, and 3 washes in 5% acetic acid for 30 min. The gels were left overnight in the Schiff reagent, and then washed in 0.1% sodium metabisulfate in 0.01 M sodium chloride until the wash solution no longer turned pink with the addition of formaldehyde. The bands on the gels were recorded by scanning them on a Beckman Model 25 spectrophotometer with a rod gel scanning adaptor at 555 nm for samples stained with Schiff, and at 590 nm for samples stained with Coomassie Blue. c. Results The adhesin samples were stained using Coomasssie Blue. The glycophorin samples were stained twice, once with Schiff stain, and then again with Coomassie Blue stain. The results of the densitometric scans for these two proteins are shown in Fig 3-1. The two batches of glycophorin that were isolated produced similar gel patterns. The gels were calibrated using serum and reduced fibrinogen from human blood using the molecular weights assigned by Janzen (Janzen, 1985) and tabulated in Table 2. Gels of serum and reduced fibrinogen run in parallel with the adhesin and the glycophorin samples and stained with Coomassie Blue are shown in F ig 3-2. The calibration curve of the log of molecular weight vs the Rf value (the distance travelled by a protein divided by the distance travelled by the tracking dye) for the proteins in Table 2 is plotted in Fig 3-3. Preliminary Experiments / 37 Sclniff Stain S3 8 o.*. o.e o. R e l a t i v e Mob i l i t y O » I ! I I o 0.2 0.-4. o.e O . S 1 R e l a t i v e Mob i l i t y Fig3.1: Densitometric scans of adhesin and glycophorin on 3% SDS.PAGE gels. A: 40 micrograms of glycophorin stained with Schiff. B: 40 micrograms of glycophorin stained with Schiff and Coomassie Blue; 20 micrograms of adhesin stained with Coomassie Blue. The tracking dye was Pyronin Y. The densitometer gain for glycophorin sample in B is 8 times that of the other two gels. ' I • — ~ " ~ — — I ' 1 — 0 .2 0 .4 0 .6 0 . 8 Relative Mobility Fig 3.2 A densitomelric scan of human blood serum polypeptides run on 3% SDS PAGE gels stained with Coomassie Blue. The reduced fibrinogen sample contained 20 micrograms of protein, and the volume of serum was 10 microliters. The tracking dye was Pyronin Y. Preliminary Experiments / 1000-\ 100-X \ X \ \ A s«rum polypeptides X reduced fibrinogen 10- i ' ' ' i ' 1 ' 1 r-0.2 0.4 0.6 R e l a t i v e Mobility 0.8 Hg 3.3 Calibration curve of 3% gels using human blood serum and reduced fibrinogen as standards. Molecular weight information is tablulated in Table 3 (janzen, 1985). The tracking dye was Pyronin Y. Preliminary Experiments / 40 Table 2 : Molecular weight assignments for the standards used on the 3% SDS P A G E gels. Rf values are with respect to Pyronin Y and are taken from Fig 3-2. P E A K # PROTEIN MOLECULAR Rf WEIGHT xlO-3 1 I G M 950 0.10 2 Heptoglobin 2-2 400 .150 4 Fibrinogen 340 .213 5 Fibrinogen 320 6 Heptoglobin 2-1 200 .361 7 IgG 160 .430 8 Transferin 76 .594 9 Albumin 66 .660 10 Hemaglobin 17 .881 11 Fibrinogen A 64 .549 12 Fibrinogen B 56 .560 13 Fibrinogen C 48 .620 14 Pyronin Y 1.00 The adhesin sample was reduced prior to electrophoresis. The single peak in Fig 3-1 B shows that this process broke the adhesin down into a single protein whose molecular weight was 30 kDa, whereas Fig 9 shows that unreduced radiolabeled protein was too large to enter the gel. The glycophorin gel stained with Schiff in Fig 3-1 A , has three peaks with apparent molecular weights of 110K, 7 I K and 33K corresponding to Rf values of 0.437, 0.568 and 0.721. Silverberg and Marchesi (Silverberg and Marchesi, 1978) published Ferguson plots of Log Rf vs % acrylamide for samples of glycophorin using gels with the same acrylamide-bisacrylamide ratio, the same tris buffer system and using the same tracking dye as was used in these experiments. Extrapolation from their plot to 3% gels gives Rf values of 0.43, 0.55 and 0.71 for the three peaks which stained with Schiff stain. This confirms the identity of glycophorin in our preparation sample. The Schiff-Coomassie Blue stained gels in Fig 3-1 B, Preliminary Experiments / 41 shows two extra peaks at Rf values corresponding to molecular weights of 18K and 3 OK. The lower molecular weight peak probably represents hemaglobin polypeptide chains and the Schiff peak near the tracking dye likely corresponds to sialated lipids (Steck, 1974). The 30 kDa peak is unidentifiable with the amount of information available. It is not sialated because it only stains with Coomassie Blue. It is not likely to be a plasma protein due to the exhaustive washes in the preparation of the ghosts. It could be band 7 of the red blood cell, which is the major unsialated membrane polypeptide with a molecular weight near 30K on the red blood cell membrane (Fairbanks et al., 1971; Dzandu et al., 1984), but there seems to be no precedent for such a selective isolation of that protein over the other membrane proteins. It could also be a proteolytic digest of a membrane polypeptide, but nothing conclusive can be said about its identity. It is useful - to remember however that Schiff stained gels contain 4.5 times as much protein as the Coomassie Blue stained gels due to the lower sensitivity of the Schiff stain, that the maximum for the Schiff stain is not at 590 nm, and that the trace has been amplified eight times more than the other two gel scans. The height of the 30 kDa peak is therefore deceptively large when compared to the Schiff peak in the glycophorin sample. The 30 kDa peak is nevertheless present and unidentifiable. C . P R O T E I N S E P A R A T I O N A Two-phase system was used to localize one protein but not the other, so that the concentration of unbound protein could be determined without interfering with the binding equilibrium. The first step in developing this phase system was to find a non-ionic detergent which would help to dissolve the Preliminary Experiments / 42 proteins in the polymer solutions of phase systems. A phase system incorporating this detergent was then found with the appropriate partitioning properties. These two steps are explained below, after a description of the technique used to prepare two-phase systems. 1. Preparation of Two-Phase Systems a. Materials Sentry Grade poly(ethylene glycol) 8000 (abbreviated PEG) was from Union Carbide Corp, Piscataway N . J . . The number average molecular weight was 7,500 - 8,000 g/mole; it was stored at 4 ° C . Dextran T500 was from Pharmacia Fine Chemicals, Uppsala, Sweden. The weight average molecular weight was 494,000 g/mole, and the number average molecular weight 181,200 g/mole. 13 x 100 mm Borosilicate Glass Disposable Culture Tubes were from Fisher Scientific, catalogue number 14-962-IOC. b. Notation Phase systems are described using the notation (X,Y)Z, where X is the % w/w of dextran, Y is the % w/w of P E G and Z is one of the buffer systems described in Table 3. c. Methods Due to the high viscosity of the polymer solutions, all stock concentrations were made up by weight and not by volume. Because of the variable water content of the polymer samples, accurate determination of stock solutions was done after solutions of roughly 20% dextran and 30% P E G had been prepared. This was done by diluting the samples by weight and then measuring P E G Preliminary Experiments / 43 concentration on a Bausch and Lomb refractometer (precision ± 0.0005), calibrated using sucrose standards and data from the CRC Handbook of Chemistry and Physics, 59th edition, 1979. Dextran concentrations were measured on a polarimeter (Dr. Steeg and Reuter, Hamburg, FDR with a 20 cm tube, precision ± 0.05°) , assuming [a]pj = + 1 9 9 ° . Stock solutions of 0.287 M phosphate buffer (see Table 3) pH 7.4, and 0.6 M sodium chloride were also prepared. Phase systems were then made up by weighing stock solutions into a beaker to an accuracy of 0.01 g. These solutions were then mixed for 30-60 min at room temperature, and allowed to settle by standing overnight or by gentle centrifugation at 2000 xg for 20 min The phases were next separated using separating funnels or 10 ml plastic pipettes, discarding the interfacial region which contained dust or particles. Phase systems and stock solutions were stored up to a week at 4 °C . Unseparated phase systems were also stored at -20°C for up to two months. Experimental phase systems were prepared by mixing 2.0 ml of each phase at room temperature in new 13 by 100 mm glass test tubes. The volume of sample subsequently added to the system never exceeded 0.150 ml. When many different phase systems were being used, a stock phase system was prepared, frozen for storage, and then thawed and diluted with the appropriate buffer solutions to the desired concentrations. Preliminary Experiments / 44 2. The Detergent Used in the Two-Phase System It is possible that the adhesin may be forced to aggregate with itself in solutions containing dextran or P E G , if the adhesin-polymer interaction is too high, because at high concentrations, similar proteins from the K99+ E.coli are known to aggregate. (Jones and Isaacson, 1983). Glycophorin is known to exist in an aqueous environment as a dimer, however in the absence of an amphiphile, it aggregates into micelles containing 10 to 18 proteins (Silverberg et al., 1976; Anderson and Marchesi, 1985). A mild detergent was therefore added to the phase system in order to increase the solubility of the adhesin, and to reduce the number of glycophorin molecules in a micelle. The four non-ionic detergents considered were (a.) Ammonyx LO, which is used in columns that separate the three glycophorins isotypes of human red blood cell membranes; (Furthmayr and Marchesi, 1983; Furthmayr et al., 1975); (b.) Triton X 100, which has been used extensively in protein solvation without affecting bioactivity (Helenius and Simons, 1975; Tanford and Reynolds, 1976; Lutz et al., 1979a); (c.)Tween 20 and (d.)Tween 80, which have P E G chains as part of their structure, improving the likelihood of their being compatible with the P E G rich phase of the dextran-PEG two phase systems. The effect of these detergents on adhesin glycophorin binding was examined by following the E.coli - red blood cell binding, since the proteins come from the surfaces of these two cells. Initially this involved the microtitre hemagglutination assay described above, however since Ammonyx L O and Triton X 100 lysed fresh red blood cells, these were replaced with fixed red blood cells. Since the microtitre assay could not be run with fixed cells, the agglutination assay was replaced by a phase system experiment. Preliminary Experiments / 45 Previous work done in this laboratory had shown that F41 E.coli and red blood cells have different partition coefficients in the (5,3.5)a two-phase system described below, and that these coefficients change when the degree of cell binding changes. The partition coefficients of the two cell types were therefore used qualititively to assess the affect of the detergents on the adhesin -glycophorin binding. This was done by incubating detergent, red cells, and E.coli in PBS, separating the cells in the two phase system, and then determining cell concentration by counting the 1 fl C labelled E.coli cells on a scintillation counter, and the blood cells on a cell counter. The preparation and radiolabelling of the cells, and the evaluation of the detergents is described below. a. Preparation of the Fixed Red Blood Cells Type M M red blood cells were fixed by crosslinking the cells with glutaraldehyde. This was done by incubating washed red cells from M M type blood resuspended at 10 to 15 % hematocrit in a 1% glutaraldehyde solution in PBS with constant agitation overnight at room temperature. The suspensions were stored at room temperature in the incubation solutions. The cells were washed three times prior to use by centrifuging them at 9000 xg in a 1.5 ml Eppendorf tube for 1 min, and then resuspending them in PBS buffer, shearing the suspensions as much as possible to break up cell aggregates. A sample of the blood was then spun in a microhematocrit centrifuge for five min. This produced packed cells with a density of 8.4 x 10 7 cells/ml (determined using the cell counter as described below) which was defined as a 100% hematocrit. Dilutions were then prepared in PBS. Preliminary Experiments / 46 Table 3 : The composition and notation of the buffers used. Sodium phosphate (NaP) solutions had moles dibasic : moles monobasic phosphate = 3.16 : 1. BUFFER pH NaCl NaP % Tween80 concentration concentration w/w mM mM PBS 7.4 130.4 20.0 / a 7.4 / 143.5 / b 7.4 / 143.5 .0416 c 7.4 120.0 1.0 .0416 d 7.4 150.0 1.0 .0220 e 7.4 150.0 1.0 .0280 f 7.4 150.0 1.0 .0416 g 7.4 150.0 1.0 .0500 h 7.4 180.0 1.0 .0500 i 7.4 210.0 1.0 .0500 J 7.4 150.0 17.0 / k 7.4 150.0 17.0 .0416 b. Preparation of E.coli cells Materials: Agar Noble was from Difco Laboratories, control 701844; casein, acid hydrolysate from Sigma, catalogue #C9386; and the radioactive amino acid solution used was NEL-445 L-amino acid mixture, Lot number 1669-167, with 0.1 millicuries (4 x 10 Bq) in 1 ml of 0.1N HC1 solution. A l l chemicals were reagent grade or better. Method: Bacteria were grown and manipulated in a sterile hood. Solid objects were sterilised in an autoclave for 30 min at 270°C, and liquids were autoclaved for 30 min at 250 °C. Cells were stored in liquid nitrogen in a tryptic soy broth (USP revision XIX, Gibco Diagnostics dry-form dissolved to 30g with water) in 15% glycerol, and were supported and grown in a minca (minimal casein medium) broth prepared as follows: minca broth contained 1.36 g of Preliminary Experiments / 47 K H 2 P O a , 10.1 g of N a 2 H P O „ . 2 H 2 0 , 1 g of glucose, 1 g of casamino acids, and 1 ml of trace salts brought up to 1 1 with water. The trace salt solution was prepared from 10 g of M g S O „ . 7 H 2 0 , 1 g of M n C l 2 . 4 H 2 0 , 0.135 g of F e C l 3 . 6 H 2 0 , 0.4 g of C a C l 2 . 2 H 2 0 brought up to 1 1 with water. A minca agar gel was prepared from 14 g of Bacto-agar in 1 1 of minca broth. The F41 E.coli were labelled as follows, insuring at every stage that the bacterial suspensions were sheared as little as possible to minimise any loss of adhesin from the surface of the cell. The cells were first suspended in 5 ml of sterile broth, a 0.050 ml aliquot of 1 4 C amino acids was then added, the solution mixed and poured onto a sterile minca plate, and the cells cultured at 3 7 ° C for 18 hr. The cells were then resuspended in PBS, and washed three times. A 100% 'bugcrit' was then made by centrifuging the cells at 9000 xg for 5 min, and the cells were diluted in PBS. c. Detergents and Adhesin-Glycophorin Binding Materials: Diluid was from Baker instruments, catalogue # 49-011-640-070, and Atomlight was from Dupont-NEN Research products, catalogue # N E F 968. Method: Phase systems were prepared as described above. The samples were serially diluted in duplicate in Eppendorf tubes to give varying bugcrits in a constant hematocrit. The incubation system was always the same as the phase system used to separate the cells except that it did not contain dextran or P E G . After incubating the red blood cell-bacteria mixtures for 15 min at room temperature, 0.10 ml of each mixture was added to a phase system which was mixed by rotating the test tubes end for end 20 times, before allowing the Preliminary Experiments / 48 phases to separate for half an hour. Care was taken to insure that each experiment had the same incubation and separation times. The 1 ' C E.coli cells were counted by dissolving in duplicate 0.10 ml of phase system in 10 ml of Atomlight, and then counting the solution on a Philips PW 4700 Liquid Scintillation Counter in the 6.00-150.00 K e V window for 5 min. Red cell concentrations were determined from samples of 0.20 ml, which were taken in duplicate from each phase, added to 10 ml of Diluid (a commercial, filtered electrolyte solution) and then counted using a cell counter (Electrozone Celloscope Model 112CLTH/RWP). It was found that the counter did not count every cell when more than 12,000 cells were in 0.10 ml of Diluid, therefore they were kept at concentrations below 5000 cells in 0.10 ml of Diluid. "Load" samples were prepared for both cell types in order to determine the number of cells added to a phase system, by counting the number of cells in 0.10 ml of the 5% 'bugcrit', and in 0.100 ml of the 2% hematocrit. A "load control" was prepared for each phase system to observe the partition coefficient of the cells in the absence of the other cell type. This involved following the above procedure, from incubation to separation and cell counting, using just one cell type. Binding between the two cell types was monitored by following the partition coefficient, K . , of a given cell type, which was defined as the number of cells in phase i divided by the number of cells added to the phase system. The "load" was diluted prior to counting in the same way as the samples in the phase systems, so that K i was equal to the number of cells counted in the sample taken from the phase i , divided by the number of cells counted in the "load" sample. The partitioning of E.coli and M M red blood cells in the (5,3.5)a Preliminary Experiments / 49 two-phase system is illustrated in Fig. 4. The partition coefficient of the red cells in the lower phase was not recorded, since visual inspection showed that the red cells partitioned mainly between the upper phase and the interface between the two phases, and the red blood cells were shown to be aggregated in the lower phase by microscopy. Fig 4 does show that the E.coli went to the interface between the two phases regardless of the presence of red cells, whereas the red cells are drawn out of the upper phase by the addition of E.coli. Therefore the red cell partition in the upper phase was used to monitor the effect of different detergents on the binding between adhesin and glycophorin. Since inhibition of red cell - E.coli binding would be reflected in an increase in the number of blood cells in the upper phase, a detergent suitable for the glycophorin - adhesin binding assay would be one which did not affect the red cell partition in the presence of E.coli. If a detergent caused the blood to go to the interface in the absence of E.coli, it would not be possible to see if it had affected the cell - cell binding. The dependence of red blood cell partition on the concentration of detergent in the (5,3.5)a two-phase system was therefore determined both in the absence of, and in the presence of the E.coli cells. d. Results The results, plotted in Fig 5, show that Ammonyx L O pulls the red cells out of the upper phase in the absence of E.coli making it impossible to assess the effect of Ammonyx L O on the cell binding; that Triton X 100 has a variable effect on the red cell partition in the absence of E.coli, and causes a small increase in the number of red cells cells in the upper phase in the presence of E.coli, suggesting that it might be decreasing the number of red cells which are binding to the E.coli cells; and that Tween 20 and Tween 80 had T h e D e p e n d e n c e of R e d B l o o d Cel l P a r t i t i o n o n Cel l B i n d i n g in a ( 5 ,3 .5 )a P h a s e s y s t e m all i n c u b a t i o n s in a 1% h e m a t o c r i t 7 0 - i - —_ , , , , , . . . , , . . . 1 . • • •——i • • • • r •• • • i O 0 .5 1 1.5 2 2 .5 3 % bugcr i t in the i n c u b a t i o n s y s t e m Fig. 4: The effect of the binding of F41 E.coli to MM red blood cells on the partition coefficient of the cells in a (5,3.5)a two phase system. In the absence of red blood cells, the E.coli had an upper phase partition of 9%, and a lower phase partition of 14%. There was a 10% error in each point. Preliminary Experiments / 51 .22 samples incubated In a 1^3 hematocrit x-/ / / / / ^ .001 0.01 0.1 d « » t « r - s « r - i t Tn tr-ie i n c u b a t i o n s y s t « A T W E E N 2 0 X T R I T O N X IOO a T W E E N 8 0 O A M M O N Y X I.O iamples incubated in a T5=S hematocrit and a 1.S?3 bugcrit B T W C E N 2 0 T R I T O N X IOO T W E E N e O O . O O I O . O i O.I " d e t e r g e n t In T n e I n c u b a t i o n s y s t « Fig. 5: The effect of detergent on the partition of blood cells in the upper phase of the (5,3.5)a phase system. The same concentration of detergent was used in the incubation system and the phase system for a given experiment. In the absence of detergent, the average blood partition for A above was 67%± 4%, and for B was Z 3 % ± 0.3% Preliminary Experiments / 52 approximately the same effect in the presence of both cell types. Since Tween 80 had the least effect of these detergents, it was used in the phase system in which the glycophorin adhesin binding was studied. 3. The Two-Phase Assay used to Separate the Proteins a . Introduction Once Tween 80 had been chosen as an appropriate detergent, it was possible to find a two-phase system which partitioned one of the proteins predominantly into one phase while leaving the other evenly distributed between both phases. The strategy used to find this system was based upon the knowledge that the partition of glycophorin in a (5,4) system is about 1.0, is relatively independent of salt concentration (private communication), and that large proteins are more sensitive to changes in a phase system than are small proteins (Albertsson, 1971). The (5,4) two-phase system was therefore adopted and fine tuned by monitoring the partition of the high molecular weight adhesin. The phase systems which were then examined were chosen based upon the trends described by Albertsson. He states that when the salt concentration is below 0.2M, increasing the salt concentration will decrease the partition of negatively charged proteins while above 0.2M salt, increasing the salt concentration will increase the partition of negatively charged proteins. The adhesin and phosphate are negatively charged, and phosphate partitions predominantly into the lower phase of a (5,4) system. A high phosphate concentration would therefore increase the partition of the adhesin in the upper phase. Hence two possibilities exist. Either (a) the adhesin can be put into the upper phase with a salt concentration above 0.2M and a high phosphate Preliminary Experiments / 53 concentration, or (b) the adhesin can be put into the lower phase with a salt concentration just below 0.2M and with a low phosphate concentration. In a (5,4) system, the partition of glycophorin using conditions (b) is about 1.0, and the partition increases when the phosphate concentration is increased. Conditions (a) will therefore increase the partition of both proteins in the upper phase. Therefore conditions (b) were adopted. The salt concentration was kept below 0.2M and the phosphate concentration was kept as low as possible while maintaining a minimum adhesin partition, without compromising the buffering capability of the solution. Buffers b through h of Table 3 were studied by following the adhesin partition, using fluorescence of native tryptophan to detect the adhesin. Although this technique was not as accurate as required in the final assay, it was sensitive enough to detect the fractions of a /ig of adhesin present in the phase system. The fluorescence assay involved preparing (5,4) phase systems with different buffers, transfering two aliquots of 0.300 ml of each phase into 1.700 ml of 0.100M pH 8.0 phosphate buffer, and viewing the sample at an excitation wavelength of 290 nm and an emmission wavelength of 331 nm. b. Methods After washing the P E G , and calibrating the spectrofluorometer using the methods in appendix A , (5,4) phase systems with buffers b through h in Table 3 were prepared as described above. The adhesin partition in each system was found by adding 6 jig of adhesin in 0.150 ml of PBS to two 13x100 mm test tubes containing 2 ml of upper and lower phase. This was done in parallel with a blank containing 0.150 ml of PBS. After mixing and separating the systems, 0.300 ml of each phase were added to a new test tube containing Preliminary Experiments / 54 Table 4 : The partition of adhesin in the lower phase of (5,4) systems with various buffers, determined from the native fluorescence of the adhesin. The partition coefficient K a was defined as the percent of the adhesin which could be accounted for that was found in the lower phase of the system studied. The notation used for the buffers was taken from Table 3. B U F F E R Ka ± 3 % NaCl phosphate % Tween80 concentration concentration w/w mM mM c 87% 120.0 1.3 .0416 d 66% 150.0 1.3 .0220 e 75% 150.0 1.3 .0280 f 95% 150.0 1.3 .0416 g 80% 150.0 1.3 .0500 h 74% 180.0 1.3 .0500 i 71% 210.0 1.3 .0500 1.700 ml of the 0.100 M pH 8.0 phosphate buffer. A reference sample was then prepared for each phase by adding 1 jug of adhesin to 0.300 ml of phase, and bringing the volume up to 2.0 ml with phosphate buffer. After mixing, these samples were examined in the spectrofluorometer in a quartz cuvette at an emmission wavelength of 331 nm, and an excitation wavelength of 290 nm. The spectrofluorometer was first zeroed using the blank. The sensitivity was then adjusted with the reference sample so that the output was 0.055 V , these two steps were iterated until consistent, and then the samples were run. This procedure was repeated for each phase system. c. Results The results in Table 4 show that the adhesin has the greatest partition coefficient with the (5,4)f phase system. Comparison of buffers d, e, f and g, shows that increasing the Tween 80 concentration increases K a until the concentration exceeds 0.0416%. Buffers g, h and i show that in the presence of Preliminary Experiments / 55 Tween 80, and above 0.2M salt, K a decreases, and buffers f and c show that decreasing the salt concentration below 0.150M decreases K a . The two-phase system chosen to partition adhesin and glycophorin was therefore the (5,4)f system. D . R A D I O L A B E L L I N G T H E P R O T E I N S 1. Introduction The double labelled experiment required two compatible radiolabels which were identifiable in the presence of each other. Two possible labels are 3 H and 1 * C, which are beta emmitters. A third label is 1 2 5 1 , which is both a gamma and a beta emitter and can therefore be detected quantitatively in the presence of, but without any interference from, a beta emitter and was therefore used as one of the labels. The beta spectra of the three labels in the scintillation fluid Atomlight in Fig 6, show that there is a greater overlap between 3 H and 1 2 5 1 than there is between 1 4 C and 1 2 5 1 . Since protein concentration is obtained from the area under the curves, resolution of two different labels will be optimised when the overlap between two spectra is minimised. In addition, distortion of the spectra due to quenching is reduced at higher K e V . Carbon and iodine were therefore chosen as the two labels, and the counting windows were set at 52-125 K e V to minimise 1 2 5 1 cpm and to maximise 1 4 C cpm. The choice of which label goes on which protein depends upon the 1 4 C labelling technique because the 1 4 C has a lower specific activity than the 1 2 5 1 , and therefore determines the detection limits of the assay. The 1 4 C should 120 130 Kev Fig. 6: Beta spectra of 3 H, , 4 C , and 1 2 5 l . Samples were disolved in 0.5 ml of lower phase of (5,4)a phase system and in 10 ml of the scinlillant Atomlighl, and then counted on a Philips 4700 bela counter for 0.5 min in 2.0 Kev sleps. The counts were normalised so that the maximum cpm for each peak was reduced fo 4000 cpm. Preliminary Experiments / 57 therefore be attached to the protein which reacts most effectivly with the carbon label. The labelling technique which was chosen and which is described in detail below, involves the binding of 1 "C-HCHO to lysine. The detection limits of the assay will be highest therefore when 1 " C is bound to the protein with the greatest number of chemically modifiable lysines. This can be determined using the fluorescamine assay. Fluorescamine is a chemical which will bind to primary amines, and therefore to lysines. When it does, its optical properties change so that it becomes a fiuorophore. It should therefore be possible to get a rough estimate of the relative number of lysines on the three proteins by comparing the fluorescence of the fluorescamine bound to the adhesin and glycophorin with the fluorescence of bovine serum albumin (BSA) whose lysine content is known. This assumes that the chemical activity of H C H O and the fluorescamine are the same in the presence of the three proteins, and that the quenching of the fluorescamine will be similar when it is bound to the three proteins. 2. The Fluorescamine Assay a. Materials The fluorescamine was from Aldrich Chemical Company, lot E J 0223EJ. b. Methods The assay was taken from Bohlen et al. (Bohlen et al., 1973) and was run in the Turner model 430 spectrofluorometer, with the emission wavelength set to 475 nm, and the excitation wavelength set to 395 nm. Sample solutions were prepared by diluting 10 to 50 M! of protein solution in 0.300 ml of phase Preliminary Experiments / 58 system, and bringing this up to 1.500 ml with 0.100 M pH 8 phosphate buffer in new 13x100 mm glass test tubes. This was then vortexed while adding 0.500 ml of freshly prepared 0.30 mg/ml fluorescamine in B D H Omnisolv glass distilled acetone. After incubating the samples for 7 min at room temperature, the zero and the sensitivity were adjusted with a blank, and a sample containing 12 ng of gamma globulin, until they gave consistent readings, and the samples assayed no later than 15 min after the 7 min incubation period. A standard 0.40 mg/ml in gamma globulin was stored at -70 °C, and used for all of the fluorescamine experiments. c. Results Calibration curves for glycophorin and adhesin are shown in Fig 7. A calibration curve of B S A was found to be linear by Bohlen et al. (Bohlen et al., 1973) in the range 0.5 to 50 micrograms. The number of lysines on each protein was then calculated as shown below for glycophorin using the fact that B S A has 59 lysines (Putnam, 1960). The B S A sample was used to find x, where from Table 5 x = (# lysines)/[(relative fluorescence per microgram) (molecular weight of BSA)] x = (59 lysines)/[(184rfl/Mg) (66210 ug /Mmole)] = 4.84 x 10" 6 lysines-umole /rfl The number of lysines on the glycophorin was therefore (4.84 x 10" 6 lysines-Mmole BSA/rfl) x (12.3 fl/yg x 31000 ug I umole) = 1.8 lysines per molecule The results in Table 5 are inaccurate because it is known that there are five and not two lysines per glycophorin molecule. The assay does show however that there are probably more chemically modifiable lysines on the adhesin than on glycophorin. Therefore the adhesin was labelled with 1 4 C, and Preliminary Experiments / 59 adhesin .2. adh«stn o.e o. licrograr i 1.2 i.» i.e is of protein glycophorin B 1 4O. glycophorlr CD BSA 2 s a 10 micrograms o"f protein Fig. 7: Calibration of adhesin and glycophorin using the fluorescamine assay. With the glycophorin, samples were taken from the upper phase end the lower phase of a (5,4)f two—phase system. With the adhesin, samples were also taken from a 0.1 M pH 8 phosphate buffer. The BSA was sampled in 0.1 M pH 8 phosphate buffer. Relative fluorescence units = fluorescence for a sample normalised with the fluorescence from 12 micrograms of a gamma globulin standard Preliminary Experiments / 60 Table 5 : Approximate lysine content of adhesin and of glycophorin determined using the fluorescamine assay. Protein rf l" Molecular weight # of Lysines per molecule B S A Adhesin Glycophorin Gamma Globulin 184 43.8 12.27 7.36 66210 2 29000 31000 3 59 2 5.9 1.8 1 fluorescence units normalised by a 0.030ml sample of 0.40 mg/ml sample of gamma globulin. 2 (Putnam, 1960) 3 (Tomita and Marchesi, 1975) glycophorin was labelled with 1 2 5 1 . This arrangment optimises the detection limits of the assay, and labels the ligand (glycophorin) with the most sensitive and interference free label. 3. Radiolabelling of Glycophorin a. Materials Iodobeads were from Pierce; crystalline albumin bovine, control # 3404, from Nutritional Biochemical Corporation, trichloroacetic acid (TCA), from Eastman Kodak Co, Lot # E 5 X , Sephadex G 25 Fine from Pharmacia, N a 1 2 5 I was from Amersham. A n L K B 1282 Compugamma Universal Gamma Counter was used. b. Methods Glycophorin was iodinated using the method of Morrison (Morrison, 1970). To an incubation tube was added 0.3 to 1.0 ml of 1 mg/m! protein, 4 iodobeads, and 1 to 5 ftl of , 2 5 I (100 - 500 nCi), and the mixture incubated for 10 min. A T C A precipitation was done (see below for details) to assay for Preliminary Experiments / 61 free label as a preliminary check of the completeness of the reaction. Free label was separated from the protein on a Sephadex G 25 column equilibrated with 10 m M phosphate at pH 7.4, the effluent from the column divided into 40 fractions of 15 drops, and each fraction assayed in the gamma counter. The fractions containing the protein were then pooled and assayed for free label using the T C A precipitation described below, and stored at -70 °C. The counts are shown in the histogram in Fig 8. c. TCA Precipitation The trichloroacetic acid (TCA) precipitation was done by taking 0.005 ml of 1 2 5 1-glycophorin into an Eppendorf tube labelled "P" for precipitate. To this was added 0.195 ml of 10 mg/ml B S A and 0.200 ml of 20% T C A . The solution was mixed, centrifuged, 0.200 ml of supernatant placed in an Eppendorf tube labelled "S" for supernatant, and the Eppendorf tubes and counted in gamma tubes. The % free label was calculated from the following equations: % Free Label = 1 - % Bound % Bound = (P - S)/P Adhesin samples were dissolved in 4 ml of scintillant before being counted so the T C A precipitation was done in gamma tubes, since Eppendorf tubes were too small. It was difficult to remove the supernatant from the gamma vial without coeluting some of the precipitate, so adhesin samples were assayed for free label by a double T C A precipitation. This involved taking 30 ii\ of "S" into 0.170 ml of 10 mg/ml B S A in a gamma counting tube marked "P"' , and adding 0.200 ml of 20% T C A . After mixing and centrifuging, 0.200 ml of the Preliminary Experiments / 62 C2 X "5 ~S5 so • 4 Adhesir-i •to 20 so drops x 1/IS Glycophorin B ~35 _S a *3 I- -l O 20 S O drops x 1/15 R g . 8: Histograms of the radioactivity of the fract ions collected f rom the Sephadex G - 2 5 co lumns used to wash the radiolabeled proteins. Preliminary Experiments / 63 supernatant was transferred to a gamma tube marked S'. The four tubes P,S P ' and S' were then counted and the % bound was calculated using % Bound = (P - Z)/P , where Z = ( 1 - (P' - S')/P'))S d. Removal of Free Label It was found that when the glycophorin was stored for several weeks, free label began to come off the protein. When this occurred, the percent free label could be brought down to 2% using a Centricon microconcentration device made by Amicon. Unfortunately, this technique was not used until the second glycophorin batch had been prepared. e. Results The histogram of the samples taken from the Sephadex G-25 column in Fig 8, shows that the label and the protein were well separated. A T C A analysis of the samples pooled from fractions 10 to 12 of the column showed that 4% of the activity was from free label. A microtitre hemagglutination assay of the pooled sample showed that the bioactivity of the sample was not changed by the labelling procedure. It was estimated that one 1 2 5 1 molecule was attached to every 14.5 glycophorin molecules using the calculations in Appendix 3. 4. Radiolabelling of the Adhesin Adhesin was labelled with 1 4 C using a method adapted from Jentoft and Dearborn (Jenthoft and Dearborn, 1979). Preliminary Experiments / 64 a. Materials Sodium cyanoborohydride, lot # 9213 K H , was from Aldrich, and 1 ftC formaldehyde, from N E N Research Products, lot # 2212-029; 0.25mCi; 53.0 mCi/mmole. b. Methods Before the labelling could be done, the sodium cyanoborohydride used in the reaction had to be recrystallised (Jentoft, 1979), and the optimum concentration of formaldehyde had to be determined. The N a C N B H 3 was recrystallized one day prior to use by dissolving l . l g of N a C N B H 3 in 2.5 ml of acetonitrile, centrifuging to remove undissolved residue, adding 15 ml of C H 2 C 1 2 , recrystalising overnight at 4 ° C , and then filtering and drying the product over P 2 0 5 . Jentoft and Dearborn showed that 8.0 moles of H C H O had to be incubated per mole of lysine for a 3.4 mg/ml solution of B S A to have 1.5 H C H O bound per lysine. Table 5 shows that the fluorescamine assay gave an approximate value of 6 lysines per adhesin monomer (m.w. 29,000). The 0.365 mg of adhesin incubated therefore had 0.365mg x (lmmole/29000mg) x (6 lysines/adhesin) = 7.55 x 10"° mmoles of lysine. Therefore 10*il of the 2.1 mg/ml stock solution of H ' ^ C H O in 0.010 M pH 7.0 phosphate buffer was used in the incubation. The reductive methylation was carried out by incubating 0.5 ml of 0.73 mg/ml adhesin with 0.63 mg of recrystalised N a C N B H 3 and 0.01 ml of formaldehyde solution at 3 7 ° C for 2 hr, and then at 4 ° C for 2 hr. The Preliminary Experiments / 65 mixture was loaded on a Sephadex G-25 column equilibrated with 0.1M pH 8 sodium phosphate buffer, 40 fractions of 15 drops were collected, 30 nl of each fraction was added to 4.0 ml of scintillant and counted for 10 min. The fractions containing protein were pooled and assayed for free radiolabel using the double T C A precipitation method described above. The bioactivity of the adhesin was then tested using the microtitre hemagglutination assay. c. Results The profile of the effluent from the Sephadex G-25 column in Fig 8 shows a clean separation between the free label and the protein, and the double T C A precipitation of the pooled fraction 11 to 17 showed 3% free label. A 3% SDS P A G E gel of the labelled adhesin was done and the results of that gel are shown in Fig 9. This shows that only one protein was labelled, and that its molecular weight was about 30,000. The absorbance of the pooled fractions was found to be 0.287 using the H P 8450 uv-vis spectrophotometer. The bioactivity of the adhesin was found to be unchanged by the labelling procedure using the microtitre agglutination assay. It was estimated as shown in appendix 3 that 5 H 1 ft CHO molecules were bound to each adhesin monomer using the activity of the radiolabeled adhesin sample. 5 0 0 0 0 4 O 0 0 O O 3 0 0 0 0 -c o q . 20000 o 10000 1 L_. REDUCED UNREDUCED 10 2 0 3 0 —I 4 0 gel slice n u m b e r Fig. 9: Activity of radiolabelled adhesin on a 3% SDS PAGE gel. The gels were sliced into 40 sections, and counted on the beta counter in 10 ml of Atomlight. 30 microliters of sample was loaded on each gel. a P •3 w •a CD i 3 3 CT> 05 I V . E Q U I L I B R I U M B I N D I N G E X P E R I M E N T S A . I N T R O D U C T I O N The purpose of the experiment was to obtain the isotherm for the equilibrium binding between glycophorin and the F41 adhesin. This involves plotting v, the number of moles of glycophorin bound to the adhesin per mole of adhesin in the solution, vs Gf, the concentration of unbound glycophorin in the solution. The data for this plot was obtained by incubating 1 2 51-glycophorin and 1 "C-adhesin for 30 min in 0.15 ml of buffer, and then separating the proteins by adding to the incubation mixture, 4.00 ml of the (5,4)f two-phase system (Table 3). When the proteins were measured in isolation in this system, 60% to 70% of the glycophorin was in the upper phase, while 97% of the adhesin was in the lower phase. The lower phase therefore contained most of the bound glycophorin, which was obtained using equations 28 and 29 below. Gbl = Gtl - Gfl [28] where Gbl is the amount of glycophorin bound in the lower phase, Gtl is the total amount of glycophorin in the lower phase, and Gfl is the amount of unbound glycophorin in the lower phase. Gfl = Gu/Kg [29] where Gu is the amount of glycophorin in the upper phase and Kg = Gu/Gfl, and is found in a separate experiment run in the absence of adhesin. The protein concentrations were measured by counting in duplicate 0.45 ml of each phase in both the gamma counter and the beta counter. The gamma counts gave the amount of 1 2 5 1-glycophorin, but the beta counts were due to both 1 2 51-glycophorin and to 1 4 C-adhesin. Therefore the amount of adhesin 67 Equilibrium Binding Experiments / 68 was obtained using [30]. A = T - t(blg) [30] Where T is the beta counts of the sample, t is the gamma counts of the sample, A is the beta counts due to adhesin and big is the ratio of beta counts to gamma counts for the 1 2 51-glycophorin obtained in the absence of adhesin. The isotherm was then transformed into a Scatchard plot as described above, to give n, the number of binding sites per adhesin monomer, and k, the microscopic dissociation constant of the binding site. B . E X P E R I M E N T A L 1. Methods After assaying the proteins for free label, the adhesin was diluted 1:1 with distilled water, and the 1 2 51-glycophorin was diluted in a solution of unlabelled glycophorin, to give a gamma activity between 300 and 4,000 cpm, so that the beta spectrum would not be dominated by the 1 2 5 1 counts. The proteins were then pipetted into new 13 x 100 mm glass test tubes. A buffer described in detail below was added to bring the volume up to 0.15 ml, the test tubes were centrifuged (2,000 xg, 30 s), and the solution incubated for 30 min at room temparature. After the incubation, 2.00 ml of each phase of a (5,4)f phase system was added, the test tubes were sealed with parafilm and rotated end over end 20 times, the phases separated by centrifugation (2,000 xg, 2.0 min), and then approximately 0.45 ml of phase transferred to a tared liquid scintillation (LSC) insert vial (Fisher high density polyethylene 7 ml vials, catalogue number Equilibrium Binding Experiments / 69 3-337-20). After weighing, the samples were centrifuged (2,000 xg, 2.0 min) to normalise the counting geometry, then counted in the gamma counter for 30 min (60 - 120 KeV). The bottom of the L S C insert vials were about 2 cm above the bottom of the rack holder. Since gamma counting efficiency falls off rapidly if a sample is 3 cm or higher above the base of the rack holder, the scintillant was added to the L S C insert vial after counting the sample in the gamma counter. After measuring the gamma counts of a sample, 0.15 ml of 4% SDS was added to each vial, the cap screwed on, and the vial placed in boiling water for 5 min. If this was not done, the efficiency of the 1 " C-adhesin counting in the beta counter dropped by one order of magnitude. The vials were again centrifuged to pull condensation down from the walls of the vials, 5.5 ml of Atomlight added and the solution was mixed. Dextran was found to precipitate under these conditions, making the counting efficiency time dependent, so the vials were centrifuged one last time (2,000 xg, 10 min) before counting the samples in the 50 to 125 K e V window of the beta counter for 30 min. To avoid chemiluminescence, the samples were left in the dark for 2.5 hr by running five blanks ahead of the samples in the light-protected beta counter. The total amount of protein added to each phase system was monitored by running a "load control" in parallel with the phase samples during every experiment. This involved putting a protein aliquot into a L S C insert vial, and bringing the volume up to 0.450 ml with distilled water. The partition coefficient of the glycophorin, and the ratio of the gamma counts to beta counts for the glycophorin, was monitored by running in parallel with each sample, an experiment containing everything in the sample except the adhesin. Equilibrium Binding Experiments / 70 Table 6 : Relative Efficiencies of the Radiolabeled Proteins in the Counters. Phase Glycophorin in the gamma counter Gamma/Beta cpm for Glycophorin Beta counts for Adhesin Upper Lower Water 1.000 1.018 1.002 11.9 14.2 7.37 1.000 1.010 1.015 2. Relative Efficiency Determination The amount of protein in a sample was determined from the load control, and the relative counting efficiencies for the radiolabeled proteins in the upper phase, the lower phase, and in the aqueous load control. Relative efficiencies were determined as follows. 1 2 5 1 standards: A standard solution of N a 1 2 5 1 was prepared. Sample size was found not to effect the efficiencies of the two counters in the 0.400 to 0.500 ml range. Therefore 0.050 ml of N a l standard was added separately to 0.44 ml of water, upper phase, and lower phase, in a L S C insert vial, and counted using the procedure described above. The average of 5 samples was used for each phase. The ratio of the relative efficiency for the counting of the glycophorin in the gamma and beta counters was shown to be independent of the amount of glycophorin added over the range of sample concentrations used. The results were plotted for both phases as beta cpm vs gamma cpm (corrected for gamma decay) and found to be linear with an r 2 of 0.984. The slopes of the two plots are shown in Table 6. 1 4 C standards: To a L S C insert vial were added separately 0.450 ml of upper Equilibrium Binding Experiments / 71 phase, lower phase, and water. Then 0.15 ml of 4% SDS and 3 ug of cold adhesin were added, and the samples boiled for 5 min. The vials were then tared and about 0.1 ml of 1 4 C toluene was added. After weighing the vials, 6 ml of scintillant was added and the samples centrifuged and counted as described above. The results are shown in Table 6. 3. Calculat ions The method used to obtain n, the number of binding sites per adhesin monomer, and k, the apparent microscopic dissociation constant of the binding site, is described in detail in Appendix 4. The calculations involved using the concentration of glycophorin in the upper phase, and its partition coefficient between the two phases, to obtain the concentration of unbound glycophorin in the lower phase, which was subtracted from the' total glycophorin concentration in the lower phase to give the amount of glycophorin bound to the adhesin. Two types of results were obtained based upon the buffer used during the 30 min incubation of the proteins. When Tween 80 was present in the incubation solution, the partition coefficient of glycophorin in the (5,4)f phase systems was a nonlinear function of the amount of glycophorin added to the phase system, but when Tween 80 was absent, glycophorin partition was unaffected by its concentration. The calculations used to obtain the results of these two experiments differed only when the partition coefficient of glycophorin, K g , was being calculated. This is done in step 10 of Appendix 4. Equilibrium Binding Experiments / 72 4. Test Run A n experiment was run to show that it was possible to find the correct concentration of isotope from a known mixture of the two isotopes using the experimental conditions of Section III B 1. This involved counting samples with between 300 and 4,000 cpm of 1 2 5 I in the gamma and beta counters to obtain a calibration curve of gamma cpm vs beta cpm. The experiment was then repeated adding 689 cpm of 1 " C to the 1 2 5 1 samples. The gamma cpm and the calibration curve were then used to obtain the beta cpm due to 1 2 5 1 , which was subtracted from the total beta cpm to give X , the hypothetical beta cpm due to 1 flC. The percent error was calculated as 100 x [(X - 689)/689]. With the four samples run, the errors were between 0. 4% and 5%, when the statistical uncertainty was between 0.5 and 2%. 5. Results The results obtained using the two different incubation buffers are plotted as isotherms in Fig 10, and as Scatchard plots in Fig 11. The results for the first experiment were obtained by incubating the adhesin and the the first batch of glycophorin together in an incubation buffer containing Tween 80. The results of the second experiment were obtained by incubating the samples in the absence of Tween 80 using the second batch of glycophorin. 6. Control Experiments 1. The assay to determine the values of K a in Table 4 was repeated using 1 'C-adhesin instead of native fuorescence. The results in Table 8, show that K a in the (5,4)f two-phase system is the same using the two Equilibrium Binding Experiments / 73 . = 3 . 3 -"8 CI 2 - 3 -2L o O c O TO 2 0 3 O «*0 S O Q O licrograms o"f unbound glycophorin B •s -5 9 C O o o o o - 6 5 o 10 2 0 microgram; 3 0 o so o o 7o of unbound glycophorin Fig. 10: Binding isotherms for the reaction between radiolabelled glycophorin and adhesin. Curve A was obtained from samples incubated in 0.042% Tween 80. Curve B was obtained by incubating a second batch of glycophorin in the absence of Tween 80. The highlighted points are the results of experiment A repeated with the second batch of glycophoria Equilibrium Binding Experiments / 74 A IS — o 13 -12 - 0 9 -e -3 -o -c 3 C J \ o O . S 1 1.3 2 2 . 3 3 3 - 3 • * -*.3 2 & 3 . 3 « CD 10 — IS - o \ o e -o \ . O O O . S 1 1-3 2 2 . 3 3 .3.3 A . 3 V 3 3 . 3 6 Fig. 11: Scatchard plots for the reaction between radiolabelled glycophorin and adhesin. Curve A was obtained from samples incubated in 0.042% Tween—80. Curve B was obtained by incubating a second batch of glycophorin in the absence of Tween 80. The highlighted points are the results of experiment A repeated with the second batch of glycophorin. v is the moles of glycophorin bound per mole of adhesin, and G is the concentration of unbound glycophorin in moles per liter x 10 . Equilibrium Binding Experiments / 75 sampling techniques. Buffers t and q show that in the absence of Tween 80, and with a low phosphate concentration, changing the concentration of NaCl from 10 m M to 1 m M reverses the partition of the adhesin. Buffers n through r show that with the same concentration of phosphate, varying the NaCl concentration from 150 m M to 10 m M increases K a from 92.3% to 95.7%. Buffers 1 to m show that when the N a C l concentration is 150 m M , varying the phosphate concentration from 4.3 to 0.3 m M increases K a from 91.7% to 92.8%. When the experiments were repeated using the second batch of glycophorin, the points highlighted in F ig 10 A and Fig 11 A were obtained. The data in Table 7, shows that the amount of glycophorin bound to the adhesin was not affected by the incubation time over the range of 5 to 40 min. The data in Table 7 also shows that v did not change when 3, 6, and 9 Mg of adhesin were added to the phase system. For a correct interpretation of the assay, it was necessary to know how the proteins were distributed between the two phases, and the interphase between the phases. This was done by adding the protein to 2.00 ml of lower phase, and then after time t, adding 2.00 ml of upper phase, mixing the solution, and then pipetting 0.770 ml of each phase into a LSC insert vial. The interface between the phases was then destroyed by bringing the volume up to 4.0 ml with buffer containing everything in the phase except the polymers. Another 0.770 ml sample was then removed, and all three samples were mixed with 0.268 ml of 4% SDS, and incubated at 100 °C for 5 min. Two 0.488 ml aliquots of each sample were then mixed with Equilibrium Binding Experiments / 76 Table 7 : Summary of data from control experiments 3 and 4. Incubation time Mg glycophorin /xg adhesin found (min.) found in the in the lower lower phase phase 1.8 2 18.1 1.30 1.9 20 19.2 1.30 1.7 30 18.0 1.30 1.9 40 20.3 1.20 1.7 30 18.1 2.44 1.5 30 17.1 3.90 Table 8 : The partition of adhesin in the lower phase of (5,4) systems with various buffers, determined from the beta counts of the 1 "C-adhesin. The partition coefficient K a was defined as the percent of the adhesin which could be accounted for that was found in the lower phase of the system studied. B U F F E R Ka ±0.2% NaCl phosphate % Tween80 concentration concentration w/w mM mM f 97.2% 150.0 1.3 4.2 x 10" 3 1 91.7% 150.0 4.3 0 m 92.1% 150.0 1.3 0 n 92.8% 150.0 0.3 0 0 95.0% 100.0 0.3 0 P 95.1% 80.0 0.3 0 q 95.7% 60.0 0.3 0 r 95.7% 40.0 0.3 0 s 93.2% 10.0 0.3 0 t 5.0% 1.0 0.3 0 4.5 ml of Atomlight, and counted for 30 min. The efficiency of each system was determined as described above. The results in Table 9 show that the incubation time does not affect the amount of protein lost at the interface, but that time does affect the amount of protein that cannot be accounted for. Table 9 also shows that the 1 2 51-glycophorin is easier to trace than the 1 9 C-adhesin, that it does not matter i f adhesin is incubated Equilibrium Binding Experiments / 77 in the upper phase or the lower phase of the (5,4)f system, and that in the absence of Tween 80, a much greater fraction of the adhesin stays at the interface between the two phases. 6. A n attempt was made to run an experiment with no Tween 80 in the phase system using the (5,4)q phase system in Table 8. It was found with this system however, that in the absence of glycophorin, 33% of the adhesin was at the interface. When 50 Mg of glycophorin was reacted with 3Mg of adhesin, 70% of the adhesin and 3% of the glycophorin were missing from the two phases, therefore the results of this experiment were rejected. 7. The effect that the label had on the partition coefficient of glycophorin, K g , was studied by finding K g using 1 2 51-glycophorin and glycophorin labelled with fluorescamine. The samples used to determine the K g in part 10 of the section on calculations of appendix 4 were therefore assayed using both labels. The results using the 1 2 51-glycophorin are shown in Fig 12 A , which is reproduced from Fig 14 of Appendix 4. These same samples were also assayed for glycophorin using the fluorescamine assay described above. The samples were run in each phase in parallel with 0.030 ml of the gamma globulin standard of 0.40 mg/ml. The fluorescence units (fl) for a given sample were then converted to relative fluorescence units, rfl, using rfl = 1/b x fl where b is the fl obtained with the gamma globulin standard which was run in parallel with the unknown sample. The rfl was then converted to Mg of glycophorin in 0.300 ml of phase system, using the calibration curves of F ig 7, and this was multiplied by (2.00 ml / 0.300 ml), to give Equilibrium Binding Experiments / 78 Fig. 12: The plot used to derive the dependence of the partition coefficient of glycophorin on the mass of glycophorin in the upper phase of the (5,4)f two—phase system when the 30 minute incubation of the protein was done in the presence of Tween—80. The data for A is f rom radiolabeled glycophorin and the data for B is from glycophorin labelled with fluorescamine. Notice that the error in B is much larger than in A. Equilibrium Binding Experiments / 79 Table 9 : Distribution of Proteins in the Phase Systems. Phase Protein Incubation K % at the % missing System time (min) upper/lower interface (5,4)f Adhesin 1 2.0 .029 10 3 20 .029 7 8 40 .029 13 10 70 .029 10 11 (5,4)f Glycophorin 2 1.4 5 .5 (5,4)f Adhesin 2 70 .028 11 8 (5,4)s Adhesin 3 30 .04 33 8 1 Protein incubated in the lower phase 2 Protein incubated in the upper phase 3 The two-phase system with no Tween 80 (see Table 8). Sample calculation using the sample with the 2.0 min incubation time. The rcpm added to the phase system was 35,626rcpm ± 3 % There were 331±6% rcpm in 0.770 ml of lower phase, and 11,550± 5% rcpm in 0.770 ml of the upper phase There were 22,55213% rcpm in 2.460 ml of the 1 phase system. K = 331/11,550 ± [ (6%) 2 +(5%) 2 ] 1 / 2 .= 0.029± 8% Total rcpm recovered = 331±20 + 11,550±577 + 22,552±675 = 34,4431888 rcpm at the interface = 34,443 - ((331 + 11550) x (2.00 ml / 0.770 ml)) = 3,582± [(888)2 + ( 5 7 7 ) 2 ] 1 7 2 = 3,58211058 % at the interface = (100) x ([3,582]/[35,626]) = 10% rcpm missing = (35,626+1069) - (34,442+1258) = 1,1841 1650 % missing = (100) x ([l,184]/[35,626]) = 3% the mass of glycophorin in 2.00 ml of a given phase. A plot was made of the ug of glycophorin in the upper phase of a sample, vs the rfl for a given sample, which is shown in F ig 12 B. The uncertainty was so great that it did not show the non-linearity that was found using the 1 2 5 I method shown in Fig 12 A . The value of K g obtained from this plot was 1.9±, which is within the range of the values of K g obtained from Fig 12 A in Fig 15, however, the uncertainty of the fluorescamine labelling was too high to show conclusively that the label did not affect the partition of glycophorin. V . R E S U L T S A N D D I S C U S S I O N A . R E S U L T S The results are shown in the two Scatchard plots of Fig 11. They were derived from experiments which used glycophorin isolated from M M + blood samples isolated from different donors. The plots have been made by transforming the data from the equilibrium binding isotherms of Fig 10, using the Scatchard model described in section II F, which gives a straight line with a slope equal to the negative inverse of the microscopic dissociation constant for the binding site, and an X intercept equal to n, the number of binding sites on the 29 kDa monomer of the adhesin, provided that all of the binding sites on the adhesin are identical and independent. The Scatchard plots of Fig 11 show dissociation constants of 3.8 x 10"^ and 2.8 x 10"^ moles per litre, which averages to 3.3 x 10"^ moles per litre, and their X intercepts give n values of 5.6 and 4.0, which averages to 4.8. B . S O U R C E S O F E R R O R 1. Introduction The purpose of the study described in this thesis was develop the first equilibrium binding isotherm of a bacterial adhesin with one of its receptors, in order to find whether the binding site for the receptor was located on the major structural monomer of the adhesin, or on a polypeptide found at low concentration somewhere along the adhesin (see Fig IB). The data was treated so that a value of n = l would indicate that there was one binding site for 80 Results and Discussion / 81 every 29 kDa of adhesin. Figure 9 shows that SDS P A G E only reveals two proteins when the adhesin is broken down into monomers using boiling SDS, and that the 18 kDa protein occurs at a much lower concentration than the 29 kDa protein. Therefore a value for n of 0.5 to 5.0 would be very good evidence that the binding site was on the 29 kDa monomer. A value for n which was several orders of magnitude less than 1 would indicate that a minor polypeptide was involved in receptor binding. 2. Precision An accurate knowledge of the precision in the value obtained for n would be obtained from the standard deviation of the v intercepts of the Scatchard plots, provided enough independent plots were available. Only two plots were obtained, and although the two experiments were done using different samples of glycophorin, they were both done using the same sample of adhesin. Therefore, a standard deviation of the v intercepts does not give an accurate estimate of the precision in the value of 4.8. The precision was therefore estimated from the range between the two intercepts. If the the Scatchard plot had an independent axis, an error analysis of the slope and the v intercept would give the precision of a single experiment. It would not reflect the contribution of errors in concentration, since a single aqueous sample of each protein was used in each experiment, but it would reflect pipetting errors, weighing errors, and radioactive counting errors, which developed from one incubation to another. Unfortunately there is no independent axis in a Scatchard plot, therefore this error must be estimated from the error in each point in the binding isotherm. A sample calculation is given in Results and Discussion / 82 Appendix 4, however, this only provides an estimate of the error, because the analysis assumes that error is propegated as a gaussian about a true value (Squires, 1968), whereas the data from step 7 onwards in Appendix 4 is the average of only 2 values obtained from Table 10. Nevertheless, this approach was used to try to understand the propegation of error within a single isotherm experiment. The greatest error in the calculation of the binding isotherm appears in step 13 of Appendix 4, where the amount of bound glycophorin is determined by subtracting the amount of unbound glycophorin in the lower phase from the total amount of glycophorin in the lower phase. This subtraction amplifies the error because the ratio of bound to unbound glycophorin is small. Since this ratio is dictated by the dissociation constant, a low uncertainty can only be obtained with a very high precision in the glycophorin concentration. The greatest error for this value comes from the 'load' calculation in step 8, which introduces a 10% error into the calculations at step 10. This error is due to the pipetting technique. A more accurate knowledge of the concentration of the proteins would decrease the error in the values of n and k, but it would not smooth the isotherm since a single stock solution was used for a given isotherm. The pipetting error can only be improved by adding larger volumes of protein to the phase system, but there is a limit to the amount of sample that can be added to 4.0 mis of phase system. Therefore the pipetting error could be reduced by increasing the volume of the phase system. Results and Discussion / 83 3. Accuracy o. The Effect of Radiolabelling on Glycophorin Partition Since only one glycophorin in every 14 was iodinated, the accuracy of n will decrease if the iodinated glycophorin has a different partition coefficient from the unlabelled glycophorin. Control experiment 6 had too much uncertainty to show that the 1 2 51-glycophorin label did not effect the partition of glycophorin. However, the isotherms suggest that the label did not affect the glycophorin partition. Each curve was prepared from glycophorin samples with two different ratios of hot to cold protein. This was done to keep the number of gamma counts between 300 and 4000 rcpm as described above. Controls were run in parallel with each sample in the experiments to monitor the partition of glycophorin in the absence of adhesin. When the samples were incubated in the buffer without Tween 80, K g was 0.840 ± 1%. When samples were incubated in the buffer containing Tween 80, the plot of K g vs the mass of glycophorin in the phase system in Fig 15 of Appendix 4 was smooth. The experimental value obtained for the number of moles of glycophorin bound to the adhesin is highly dependent upon the experimental value of K g . A change in K g will therefore result in a change in v. The isotherms of Fig 10 do not have a noticeable step in them where the ratio of labelled to unlabelled protein was changed. Therefore changing the ratio of labelled glycophorin to unlabelled glycophorin did not produce a noticeable change in the partition of the glycophorin sample. Results and Discussion / 84 b. The Solubility of the Proteins The accuracy of n will depend upon the solubility of the two proteins in the two-phase solutons. Glycophorin dissolves in aqueous solutions of neutral detergents as dimers in the absence of amphiphiles, aggregate into micelle-like structures containing 10 to 18 proteins (Silverberg et al., 1976). When the concentration of a micelle forming molecule exceeds some 'critical micelle concentration' (CMC), additional molecules aggregate into micelles and the free concentration of the molecule remains unchanged. There are therefore two populations of the molecule above its C M C . The addition of the amphiphile Tween 80 could redissolve glycophorin into dimers, or it could produce a third population of glycophorin involving Tween-glycophorin micelles. In addition, the presence of lipids in the glycophorin sample suggests a fourth possible population of glycophorin associate with lipid or lipid/Tween 80 micelles. It is possible that the adhesin was aggregating because fimbriae from K 9 9 + bacteria are known to aggregate in parallel bundles when dissolved in aqueous solutions (Jones and Isaacson, 1983). The absorbance spectrum of F41 + adhesin that has been freeze dried and redissolved in aqueous solutions has a broad absorbance peak at 280 nm, suggesting the presence of aggregates. It is difficult however to determine the presence of adhesin aggregates in the two-phase system used. Detection of aggregates by light scattering would be complicated by the Tween 80 micelles. Molecular weight analysis would be ambiguous because the high molecular weight dextran molecules could be the same weight as lower molecular weight aggregates of the adhesin fibers, since both substances are polydispersed. Nevertheless, certain interpretations can be made about the aggregation of the adhesin from the binding experiment. Results and Discussion / 85 The distribution of adhesin in the two-phase system in Table 9, shows that 97% of the protein added to the solution could be accounted for, and that 90% was to be found in one of the two phases. The protein was therefore suspended or in solution. Table 9 also shows that the adhesin did not partition 100% into one of the two phases, therefore an active partition was occuring and the protein was behaving as a particle in solution. Since large aggregates tend to partition to the interface between the phases, this also suggests that adhesin aggregates, if they do exist, are not very large. Another encouraging observation is that the isotherms were smooth and reproducible. In addition, the data for control experiment 4 in Table 7, shows that the number of binding sites per mass of adhesin was independent of the amount of adhesin found in the lower phase of the (5,4)f phase system. These two observations suggest that aggregation, if it existed, was independent of concentration since aggregation would be expected to sterically hinder glycophorin binding in proportion to the amount of aggregation. Since there are no known examples of more than one binding site on a monomer, the corrected value of n obtained from the Scatchard plot, implies that very few of the binding sites are in fact hidden and therefore that aggregation is probably minimal. Finally, control experiment 6 shows that in the absence of Tween 80, there is an increase in the amount of adhesin at the interface, suggesting an increase in aggregation, and the binding experiment gives uninterpretable results, suggesting a major change in some aspect of the behaviour of the two proteins. These observations suggest that the Tween 80 is keeping the adhesin from forming large aggregates in the lower phase of the (5,4)a two-phase system, and that in the presence of Tween 80, the degree of aggregation is constant and probably minimal. Results and Discussion / 86 c. Protein Concentration The accuracy of for the value of n is highly dependent upon the methods used to obtain the concentration of the proteins. The concentration of the adhesin was obtained from its absorbance, and extinction coefficient. The extinction coefficient was obtained from the absorbance of a dialysed sample which was then freeze dried and weighed. This technique will give a negative bias to the extinction coefficient because the freeze dried protein could contain trace amounts of salt due to incomplete dialysis and unknown quantities of water, due to incomplete freeze drying, and to adsorbtion of water onto the protein during weighing. The concentration of the glycophorin was obtained from the mass of freeze dried protein and the volume of its solution. This method has been a standard practice in this laboratory, however it is not entirely accurate due to the fact that glycophorins A B and C are coisolated from the red cell membrane in the ratio 1.0:0.35:0.10 (Dahr et al., 1981), and only glycophorins A and B are associated with the M N blood type (Anstee, 1979) which is associated with the binding of this adhesin (Brooks et al., 1987). In addition, it is known that 46% of the mass of a sample of glycophorin isolated as described above is in fact glycosylated phospholipid. (Van Zoelen et al., 1977; Buckley, 1978). This value will no doubt vary from one isolation to another, however accurate determination of the concentration of glycophorin is quite complex. Absorbance at 280 nm is not linear with mass because of the light scattering due to glycophorin micelles. Standard protein assays such as the Lowry assay bind to a range of proteins sugars and lipids with unknown specificities. Analysis of the sugars on the protein is not effective because the oligosaccharides on the lipids are very similar Results and Discussion / 87 to those on the glycophorins (Rauvala and Finne, 1979). Sialic acid is an exception, (Lutz and Fehr Jorg, 1979), but different blood samples appear to have variable amounts of sialic acid. Determination of the lipid concentration using a phosphate assay was attempted, but it revealed more phosphate than could be accounted for by lipids alone, despite exhaustive dialysis of the proteins against phosphate free buffer. Therefore concentration determination was done by weight, and rough corrections were made using the literature values cited above. Since Scatchard plots were made using concentrations determined by weight, the presence of glycophorin C increased the value of n by 6%, and the presence of the lipids increased n by another 46%, bringing n down to 2.4 with a range of 0.8. C. E V A L U A T I O N O F T H E T E C H N I Q U E The assay is time consuming, but fairly routine once the phase system is known and the proteins have been isolated and labelled. The calculations are labour intensive and each point on the isotherm involves six hours of radioactive counting, but the experimental work for an isotherm of 14 points can be done in 4 days. The partitioning behavior of the phase system used was both predictable and adaptable. The information provided by Albertsson (Albertsson, 1971), and by colleagues, provided enough background knowledge to obtain the (5,4)f phase system from the seven experiments summarised in Table 4. Subsequent experiments, summarised in Table 8, show that minor changes in salt concentration have little effect on adhesin partition, where as substantial changes can actually reverse the partition. Results and Discussion / 88 The isotherms obtained using two slightly different techniques are shown in Fig 10. The uncertainty in these plots is large, but it was shown above that this was due primarly to pipetting errors, and that improvements could be made by increasing the volume of the phase systems used. Binding experiments must respect equilibrium conditions, however the technique which was adopted involved incubating the proteins in 0.150 ml of buffer, and then determining the protein concentration in 2.00 ml of the lower phase, in order to minimise possible polymer-protein interactions. Since the proteins had to go through this dilution, their chemical activity will have changed, and so the concentrations obtained may not be the equilibrium values. The results from control experiment 5 in Table 9, show that increasing the incubation of the adhesin in the lower phase from 20 to 70 minutes increased by a factor of three the amount of protein that could not be accounted for. This suggests that the polymers are having an affect on the adhesin, but that the affect is minimal for the first 20 minutes and then unaltered for at least the next 50 minutes. Control experiment 3 shows that changing the incubation time in the absence of phase system from 2 to 40 minutes has an insignificant affect on the value of n. These two experiments suggest that the proceedure could be improved by incubating the proteins in the phase system for short periods of time, provided the dependence of time on the binding and on the polymer-protein interactions was more fully understood. Results and Discussion / 89 D . E V A L U A T I O N O F T H E R E S U L T S The two parameters which can be obtained from the binding isotherm via the Scatchard plot are the microscopic dissociation constant k, and n, the number of binding sites on the adhesin monomer. The value obtained for k is only reliable if (a) the glycophorin-adhesin binding occurs independently from site to site, with no interference between bound ligands, (b) if the phase polymers and the detergent have no effect on the binding constant, and (c) if the uncertainty in the Scatchard plot is minimal, since errors can hide non-linearity caused by binding which does not follow the Scatchard model (Noerby, 1980). The value of k therefore may be unreliable, although it is likely to be on the order of 10*^ moles per liter. The value obtained for n is more reliable than the value obtained for k. It is a saturation value, and can be obtained directly from the binding isotherm at infinite glycophorin concentration. The Scatchard plot is used to extrapolate to this saturation value, but the value obtained is not affected either by the Scatchard model or by factors altering the strength of the binding, although interpretation of nonlinear Scatchard plots can be misleading (Noerby, 1980). The value of n is therefore correct if the concentration of both proteins is accurately known, if data is available for the full range of the binding isotherm, and if the nature of the solubility of the protein is well understood. The problems of determining the concentration of the proteins has been discussed above. Neither of the Scatchard plots in Fig 10 has a low enough uncertainty to give reliable information on the linearity of the plot. Only the isotherm corresponding to F ig 10 B approaches saturation, and the problems of determining the solubility of the proteins have been discussed above. However, Results and Discussion / 90 the value for n is the average of two experiments involving different incubation techniques, which were shown by control experiment 3 to give the same results with M M + glycophorin isolated from two different donnors. In addition, 2.4 ± 0.8 is a realistic value for n, because it is known that the glycophorin dimers which exist in aqueous solutions remain as dimers even in denaturing solutions of SDS (Silverberg et al, 1976). If the corrected value of n is correct, it suggests that dimers of glycophorin are binding to a single site situated on each 29 K d monomer of the adhesin, and that very little aggregation of the adhesin is occuring in the phase system. Another interpretation is that the micelles containing more than two molecules of glycophorin are binding to adhesin fibers which are aggregated and therefore have some of their sites hidden from the micelles. The situation might be clarified by running an experiment with glycophorin monomers prepared by carboxymethylation (Silverberg et al., 1976). If the Scatchard plot from such an experiment gave a value of n=1.0, it would lend strong support to the first interpretation. In either case however, the results show that glycophorin is binding to a repeating subunit and not a minor polypeptide of the fibrous adhesin macromolecule. E . A D D I T I O N A L A P P L I C A T I O N S OF T H E A S S A Y The assay described in this thesis could also be used to examine the chemical specificity of the reaction by studing the effect of polysaccharides, proteolytic fragmemts and synthetic polypeptides on the shape of the isotherm. If the methods discussed above could be used to reduce the uncertainty and improve the incubation technique, it should also be possible by varying pH, Results and Discussion / 91 temperature and ionic strength, to obtain thermodynamic information on the protein binding, using the dissociation constant obtained from the Scatchard plot. F . C O N C L U S I O N S The purpose of the experiments described in this thesis was to determine whether the binding site of a surface adhesin of an F41+ E.coli was on the 29 K monomer which is the major building block of the adhesin, or on a minor polypeptide interspersed along the adhesin. Since a receptor for this adhesin exists on glycophorin, an equilibrium binding study was done with these two proteins to determine n, the number of binding sites on the 29 kDa monomer. If n was less than 1, a minor polypeptide would probably be involved in the binding, however since n was 2.4±0.8, and since glycophorin exists as a dimer in solution, the results suggest that the binding site is on the 29 kDa monomer. B I B L I O G R A P H Y Albergu, A . , Ferrez M . , Baulieu E . E . , 1976. Estradiol-receptor-DNA interaction : liquid polymer phase partition, F E B S Lett. 6_1 p 223-226. Albertsson, P .A. 1971. Partition of Cell Particles and Macro Molecules 2nd. ed., published by Wiley-Interscience, New York Anderson, R.A. , Lovrien R . E . , 1984. Glycophorin is linked by band 4.1 protein to the human erythrocyte membrane skeleton, Nature 307 p 655-658. Anderson, R.A. , Marchesi V . T . 1985. Regulation of the association of membrane skeletal protein 4.1 with glycophorin by a polyphosphoinositide, Nature 318 p 295-298. Anstee, D . J . 1979. Blood group MNSs-active sialoglycoproteins of the human erythrocyte membrane. Immunobiology of the Erythrocyte p 67-98. Bartus, H . , Actor P., Snipes E . , Seldlock D., Zajac I., 1983. Indication that the erythrocyte receptor involved in enterotoxigenic Escherichia-Coli attachment is a sialoglycoconjugate. J . Clin. Microbiol. 21 p 951-954. Bartus.H.l , Actor P., Snipes E . , Sedlock D., Zajac E . , 19S5. Indications that the erythrocyte receptor envolved in enterotoxigenic Escherichia-Coli attachment is a sialoglycoconjugate, J . Clin. Microbiol. 21 p 951-954. Berglund, A . , Backman L . , Shanlibag V . P . . 1984. Calmodulin binding to human spectrin, F E B S Lett. 172 p 109-113. Bohlen, P., Stein S., Dairman W., Undenfriend S. 1973. Fluorometric assay of proteins in the nanogram range. Archives of Biochemistry and Biophysics, 155 p. 213-220 Branton, D., Cohen C M . , Tyler J . 1981. Interaction of cytoskeletal proteins on the human erythrocyte membrane, Cell 24 p 24-32 Brooks, D .E . , Cavanagh J . , Jayroe D. , Janzen J . , Snoek R., Trust T.J . , 1987. Involvement of the M N blood group antigen in shear-enhanced hemagglutination induced by a K99/F41 positive adhesin from Escherichia Coli , submitted to Infection and Immunity. Brooks, D.E. , Sharp K . A . , Fisher D. , 1985. Theoretical aspects of partitioning, from Partitioning in Aqueous Two-Phase Systems edited by Harry Walter, Donald E . Brooks, and Derek Fisher, Academic Press, inc., pp 11-78. Buckley, J.T., 1978. Coisolation of glycophorin A and polyphosphoinositides from human erythrocyte membranes, Can. J . Biochem. 56 p 349-351 92 PAGE 93 OMITTED I N PAGE NUMBERING. PAGE 93 OMISE DANS L A PAGINATION. / 94 Cantor, C.R., Schimmel P.R., 1980. Biophysical Chemistry Part E I : The Behavior of Biological Macromolecules, published by W . H Freeman and Company, San Fransisco, p. 849-856 Chabanon, G. , Hartley C.L. , and Richmond M . H . , 1982. In vitro attachment to a human cell line of Escherichia-Coli strain causing urinary tract infection: occurance of fimbriae (pili) in adhesive and nonadhesive strains. Ann. Microbiol. (inst. Pasteur) 133A p 357-369 Contrepois, M . , Girardeau J.P., Bouet Phi, Desmettre Ph., 1982. Vaccination contre l'antigene K99 et d'autres "adhesines" des Escherichia Coli bovins, Proceedings X l l t h World Congress on Diseases of Cattle The Netherlands, Volume I, p332 Curthoys, N.P . , Rubinowitz J .C . , 1971. Binding of A T P and related ligands determined by partition equilibrium, J . Biol. Chem. 246 p 6942-6952. Dahr, W., Beyreuther K . , Bause E. , Kordowicz M . , 1981. Structures and antigenic properties of human erythrocyte membrane sialoglycoproteins, Protides of the Biological Fluids, 29 p 57-62 Davis, B . J . , 1964. Disc Electrophoresis II, Method and application to human serum proteins, New York Academy of Sciences, 121 p 404 deGraaf, Frits K . , and Ingrid Roorda 1982. Production, purification and characterisation of the fimbrial adhesive antigen F41 isolated from calf enteropathogenic Escherichia-Coli strain B4117. Infect. Immun. 36 p 751-758. Duguid, J .P. 1959. fimbriae and adhesive properties in klebsiella strains, J . Gen. Microbiol. 21 p 271-286 Duguid, J.P., Cleeg S., Wilson M.I . , 1978. The fimbrial and non fimbrial hemagglutinins of Escherichia-Coli, J . Med. Microbiol. 12 p 213-227 Dzandu, J . K . , Mercy E.D. , Barrat D .L . , Wise B .E . 1984. Detection of erythrocyte membrane proteins, sialoglycoproteins, and lipids in the same polyacrylamide gel using a double-staining technique. Proc. Natl . Acad. Sci. U S A 81 pp 1733-1737 Edelman, G . M . 1984. Cell-adhesion molecules: a molecular basis for animal form, Scientific American 250 p 118-129 Egmond, M.R. , Williams R.J .P. , Welsh E . J . , Rees D.A. , 1979. 1 H-Nuclear-magnetic-resonance studies on glycophorin and its carbohydrate-containing tryptic peptides, Eur. J . Biochem. 97 p 73-83. Ekback, G. , Morner S., Lund B . , Normaker S., 1986. Correlation of genes in the P A P gene cluster to expression of globoside-specific adhesin by uropathogenic Escherichia-Coli., F E M S Microbiol Lett., 34 p 355-360. / 95 Eschdat, Y . , Silverplatt F . J . , and Sharon N . N . . 1981. Dissociation and reassembly of Escherichia-Coli type I pili, J . Bacterid., 148 p 308-314. Fairbanks, G., Steck T.L. , Wallach D . F . H . 1971. Electrophoretic analysis of the major polypeptides of the human erythrocyte membrane, Biochemistry 10 p. 2606 Fazekas, S., de St. Groth, Webster R.G. , Daylner A . , 1963. Two new Staining proceedures for quantitative estimation of proteins on electrophoretic strips, Biochimica et Biophysica Acta, 71 p 377-391 Feize, Ten 1985. Demonstration by monoclonal antibodies that carbohydrate structures of glycoproteins and glycolipids are onco-developmental antigens, Nature 314 p 53-57 Fuenfstueck, R., Tschaepe H . , Stein G. , Kunath H . , Bergner M . , Wessel G., 1986. Virulence properties of Escherichia-Coli strains in patients with chronic pyelonephritis, Infection _14 p 145-150. Furthmayr, H . , Marchesi V . T . 1983b. Glycophorins: isolation, orientation, and localization of specific domains, Methods in Enzymology 96 p 268-281. Furthmayr, H . , Marchesi V .T . , 1976. Subunit structure of human erythrocyte glycophorin A , Biochemistry _15 p 1137-1144. Furthmayr, H . , Marchesi, V . T . 1983a. Methods in Enzymology 96 p 268 Furthmayr, H . , Tomita M . , Marchesi V.T. , 1975. Fractionation of the major sialoglycopeptides of the human red blood cell membrane, Biochemical and Biophysical Research Communications, 65 p 113-121. Gaastra, W., deGraaf F . K . 1982. Host-specific fimbrial adhesins of noninvasive enterotoxigenic Escherichia coli strains, Microbiological Reviews, 46 p 129-161. Gubish, Jr. , E.R., Cheng K . C . S . , Buchanan T . M . , 1982. Attachment of gonococcal pili to lectin-resistant clones of Chinese hamster ovary eggs, Infect. Immun. 37 p 189-194. Gustafsson, A . , Wennerstrom H . , 1986. The nature of phase separation in aqueous two-polymer systems, Polymer 27 p 1768-1770. Hales, B . A . , Amyes S.G.B., 1986. The transfer of genes encoding production of mannose-resistant hemagglutinating fimbriae from uropathogenic enterobacteria. J . Gen. Microbiol. 132 p 2243-2248. Helenius, A . , Simons K . , 1975. Solubilization of membranes by detergents, Biochimica et Biophysica Acta, 415 p 29-79. / 96 Ip, S .M. , Crichton P.B. , Old D.C. and Duguid J.P., 1981. Mannose-resistant and eluting hemagglutinins and fimbriae in Escherichia-Coli J . Med. Microbiol. 12 p 213-227. Jackson, R., Segrest J.P., Kahane I., Marchesi V.T. , 1973. Studies on the major sialoglycoprotein of the human red cell membrane. Isolation and characterization of tryptic glycopepties. Biochemistry 12 p 3131-3138. Janzen, Johan 1985. Fibrinogen adsorbtion to red blood cells P H D Thesis, University of British Columbia, p. 78. Jentoft, Dearborn 1979. Labelling of proteins by reductive methylation using sodium cyanoborohydride, The Journal of Biological Chemistry, 254, No 11, pp 4359 - 4365 Jokinen, M . , Ehnholm C , Vaisanen-Rhen V . , Korhonen T., Pipkorn R., Kalkkinen N . , Gahmberg C.G. , 1985. Identification of the major human sialoglycoprotein from red cells, glycophorin A m as the receptor for Escherichia-Coli IH 11165 and characterisation of the receptor site. Eur. J . Biochem. 147 p 47-52. Jones, G.W. 1980. Some aspects of the interaction of microbes with the human body, from Contemporary Microbial Ecology, edited by Ellwood D . C , Hedger J . N . , Latham M . J . , Lynch J . M . , Slater J . H . , Academic Press, pp 253-282. Jones, G.W., Isaacson R .E . 1983. Proteinaceous bacterial adhesins and their receptors, CRC Critical Reviews in Microbiology, 10 p 229-260 Keith, B.R., Maurer L . , Spears P .A. , Orndorff P.E. 1986. Receptor-binding function of type I pili effects bladder colonization by a clinical isolate of Escherichia-Coli. Infect. Immun. 53 p 693-696 Keusch, G.T., 1979. Specific membrane receptors: pathogentic and therapeutic implications in infectious diseases, Reviews of Infectious Diseases _1_ p 517-529 Kiernan, J . A . 1981. Histological and Histochemical Methods: Theory and  Practise, Pergamon Press, p. 128. Knapp, S., Hacker J . , Jarchau T., Goebel W., 1986. Large unstable inserts in the chromosome affect virulence properties of uropathogenic Escherichia-Coli 0-6 strain 536. J . Bacterid. 168 p 22-30 Knutton, S., Williams P . H . , Lloyd D.R. Candy D.A. , McNeish A.S . , 1984. Ultrastructural studies of adherence to and penetration of cultured cells by two invasive Escherichia-Coli strains isolated from infants with enteritis, Infect. Immun 44 p 559-608. / 97 Labigne-Roussel A . F . , Alexander Schmidt M . , Waltraud Waly, Stanley Falkow 1985. Genetic organisation of the afimbrial adhesin operon and nucleotide sequence from a uropathogenic Escherichia-Coli gene encoding an Afimmbrial Adhesin. Journal of Bacteriology, p. 1285-1292. Labigne-Roussel, A . F . , La r l D. , Schoolnik G. , and Falkow S., 1984. Cloning and expression of an afimbrial adhesin (AFA-I) responsible for P blood group-independent, mannose-resistant hemagglutination from a pyelonephritic Escherichia-Coli strain. Infect. Immun. 46 p 251-259. Lambden, P.R., Heckels J .E . , James L.T. , Watt P .J . , 1979. J . Gen. Microbiol., 114 p 305-312. Leffler, H . , Spanporg-Eden C , 1980. Chemical identification of a glycosphingolipid receptor for Escherichia-Coli attaching to human urinary tract epithelial cells and agglutinating human erythrocytes. F E M S Microbiol. Let. _8 p 127-134. Lindahl, M . , Wadstrom T., 1983. Terminal N-acdtylgalactosamine and sialic acid residues are recognized by the K99 surface haemagglutinin of enterotoxigenic E.coli, IRCS Med. Sci., 11 p 790 Lindberg, F. , Lund B. , Normark S., 1986. Gene products specifying adhesion of uropathogenic Escherichia coli are minor components of pili, Proc Natl Acad Sci, U S A 83 p 1891-1895. Lindberg, F. , Lund, B . and Normark, S., 1984. Genes for pyelonephritigenic Escherichia-Coli required for digalactoside specific agglutination of human cells, E M B O Journal _3 p 1167-1173. Lutz, H . U . , Fehr Jorg 1979. Total sialic acid content of glycophorins during senescence of human red blood cells, The Journal of Biological Chemistry, 254 p 11177-11180 Lutz, H . U . , Von Daniken A . , Semenza G., Bachi T. 1979. Glycophorin-enriched vesicles obtained by a selective extraction of human erythrocyte membranes with a non-ionic detergent, Biochimica et Biophysica Acta 552 p 262-280. Madrid Concon, J . , Soltess D. , 1972. Rapid micro Kjeldahl digestion of cereal grains and other biological materials, Analytical Biochemistry 53 p 35-41. Morrison, 1970. Iodination of tyrosine: Isolation of lactoperoxidase (Bovine), Methods in Enzymology, 17A p. 653 Neeser Jr . , Doellreutter B . , Wuersch P., 1986. Oligomannoside-type glycopeptides inhibiting adhesion of Escherichia-Coli strains mediated by type I pili preparation of potent inhibitors from plant glycoproteins. Infect. Immun. 52 p 428-436. / 98 Noerby, J .G. , Ottolenshi P., Jonsen J . , 1980. Scatchard Plot: Common misinterpretation of binding experiments, Anal . Biochem., 102 p 318-320. Norgren, M . , Normark S., Lark, P., O'Hanley, Schoolnick, G. , Falkow, S., Svanborg-Eden, C , Bana, M . , and Uhlin, B . E , 1984. E M B O Journal _3 p 1159-1165 Normark, S., Lark, D., Hul l , R., Norgren, M . , Baga, M . O'Hanley, P., Schoolnick, G. , and Falkow, S., 1983. genetics of digalactoside-binding adhesin from a uropathogenic Escherichia-Coli Strain. Infect. Immun. 4_1 p 942-949 Ofek, I, Mirelman P. Sharon N . , 1977. Adherence of Escherichia-Coli to human mucosal cells mediated by mannose receptors. Nature 265 p 623-625. Olden, K . , Parent J .B. , White S.L. 1981. Carbohydrate moieties of glycoproteins a re-evaluation of their function. Biochimica et Biophysica Acta, 650 p 209-232. Ornstein, L . , 1964. Disc electrophoresis I, Background and theory, New York Academy of Sciences, 121 p 321 Parkkinen, J , Rogers G .N . , Korhonen T.,Dahr W., Finne J . , 1986. Identification of the o-linked siayloligosaccharides of glycophorin A as the erythrocyte receptors of S-fimbriated Escherichia-Coli. Infect. Immun. 54 p 37-42. Putnam, F.W., 1960. The Plasma Proteins, Sturcture, Function and Genetic Control, Second Edition, vol 1, New York Academic Press, p. 141, 147. Rauenbuehler, P .B. , Cordes K . A . , Salhany J . M . , 1982. Identifaction of the hemoglobin binding sites on the inner surface of the erythrocyte membrane, Biochimica et Biophysica Acta 692 p 361-370. Rauvala, H . , Finne J . , 1979. Structural similarity of the terminal carbohydrate sequences of glycoproteins and glycolipids, F E B S Letters, 97 p 1-8. Rothbard, J . , Fernnandez, M . , Wang, L . , Teng, W.N.H. , and Schoolnick, G.K. , 1985. antibodies to peptides corresponding to a conserved sequence of gonococcal pilins block bacterial adhesion, Proc. Natl. Adad. Sci. U S A 82 p 915-919 Salit, I .E., Gotschlich E .C . , 1977. Type I Escherichia-Coli pili: characterisation of binding to monkey kidney cells. J .Exp.Med. 146 p 1182-1194. Schoolnik, G.K. , Tai J . Y . , Gotschlich E .C . , 1982. The human erythroyte binding domain of gonoccocal pili, In: Weinstein L . , and Fields B . N . , (e.d.) Seminars in Infectious Disease, 4_ Bacterial Vaccines. Theime-Stratton, New York, pp 172-180. / 99 Scoog, W.,West D . M . . Principles of Instrumental Analysis, Second Edition, p. 288 Sheladia, V . L . , Chambers J .P. , Guevara J . Jr,and Evans D.J . , 1982. Isolation, purification and parial characterisation of type V A hemagglutinin from Escherichia-Coli GV-12 01:H", J . Bacteriol. 152 p 757-761 Silverberg, M . , Furthmayr H . , and Marchesi V.T. 1976. The effect of carboxymethylating a single methionine residue on the subunit interactions of glycophorin A , Biochemistry 15 p 1448-1454. Silverberg, M . , Marchesi V .T . 1978. 'The anomolous electrophoretic behavior of the major sialoglycoprotein from the Human Erythrocyte, J . Biol Chem. 253 #1 p. 95-98 Smit, H . , Gaastra W., Kamerling J .P. Vliegentthart J .F .G. , deGraaf F . K . , . Isolation and structural characterisation of the equine erythrocyte receptor for enterotoxigenicc Escherichia-Coli K99+ fimbrial adhesion, Infect. Immun. 46 p 578-584. Springer, G.F. , Yang H.J . , Desai P.R., 1978. Three-dimensional model of highly M-active NH 2 - te rmina l sialoglycopentapeptide from human blood group M M red cells, Naturwissenschaften 65 S. 547. Squires, G .L . 1968. Practical Physics, McGraw Hi l l , London, p 38-45. Steck, Theodore L . 1974. The organisation of proteins in the human red blood cell membrane. The Journal of Cell Biology 62, p 1-19 Steinberg, M.S . 1970. Does differential adhesion govern self-assembly processes in histogenesis? Equilibrium configurations and the emergence of a heirarchy among populations of embryonic cells, J Exp. Zool., 173 p 395-434 Svanberg-Eden, C , Freler, R., Haberg L . , Hull R., Hugg S., Leffler H . , Schoolnick G. , 1982. Inhibition of experimental ascending urinary tract infection by epithelial cell surface receeptor analogue Nature 298 p 560-562. Tanford, C , Reynolds J .A . 1976. Characterization of membrane proteins in detergent solutions, Biochimica et Biophysica Acta 457 p 133-170. Thompson, S., Maddy A . H . 1982. Red Cell Membranes, A^ Methodological Approach, edited by J .C. Ellory and J .D . Young, Academic Press, chapter 5 Tomita, M . , Furthmayr H . , Marchesi V . T . , 1978. Primary structure of human erythrocyte glycophorin A . Isolation and characterization of peptides and complete amino acid sequence, Journal of the American Chemical Society, Y7_ P 4757-4769 Tomita, M . , Marchesi V .T . 1975. Proc. Natl . Adad. ScL USA 72 p.2964 / 100 Uhlin, B . E . , Norgren, M . , Baga, M . , Normark, S., 1985. adhesion to human cells by Escherichia-Coli lacking the major subunit of a digalactoside-specific pilus-adhesin Proc. Natl . Acad. Sci. U S A 82 p 1800-1804 Urios, P., Rajkowsdi K . M . , Engler R., Cittanova N . , 1982. Investigation of human horionic somalomammotropin antihuman chorionic somatomammotropin antibody binding by two physiocochemical methods: phase partition and fluorescence polarization, Anal Biochem 119 p 253-260. Van Die, I., Van Megen I., Zuidweg E. , Hoekstra W., De Ree H . , Van Den Bosch H . , Bergmans H . , 1986. Functional relationship among the gene clusters encoding F-71 F-72 F-9 and F l l Fimbriae of human uropathogenic Escherichia-Coli, J . Bacterid 167 p 407-410. Van Zoelen, E . J . J . , Zwaal R . F . A . , Reuvers F . A . M . , DemelR.A., Van Deenen L . L . M . , 1977. Evidence for the preferential interaction of glycophorin with negatively charged phospholipids, Biochimica et Biophysica Acta, 464 p 482-492. Wadstrom, T., Trust T .J . 1984. Bacterial surface lectins, from Medical Microbiology volume 4, edited by Easmon C.S.F., Jeljaszewicz J . , Academic press, pp287-334. Wise, Gary E . 1984. Identification and function of transmembrane . glycoproteins-the red cell model, Tissue and Cell 16 p 665-679. A P P E N D I X 1 Puri f ica t ion of the P E G The P E G used in the phase systems was found to have a fluorescence peak in the U V range used in this experiment. Precipitation and re-dissolution of the P E G removed this background fluorescence. The P E G was precipitated by dissolving 30g of P E G in 200 ml of analytical grade acetone, in a warm water bath. The solution was transferred to an ice bath, and 50 ml of reagent grade anhydrous diethyl ether was added. After precipitation and suction filtration, the product was washed with 150 ml of ether. The precipitation was repeated three times, and the product was stored under vacuum for 24 hr. Cal ibra t ion of the Spectroflurometer The time constant of the spectroflurometer was first determined by exciting a fluorophor, closing the shutter, and determining the time taken by the apparatus to give a reading of 0.0. A noise analysis was then done on a blank by zeroing the instrument and reading the output of the machine every 5 seconds. The average of 20 readings gave a standard deviation of 2.0 millivolts. The detection limits were therefore taken as 4.0 mV. The linearity of the fluorescence for the adhesin was then shown by mixing 1.6 ug of adhesin in 0.050 ml of PBS buffer into 1.950 ml of 0.100 M pH 8.0 phosphate buffer. After recording the fluorescence of this sample, it was diluted by replacing 0.800 ml from the cuvette with 0.800 ml of the phosphate buffer. This new sample's fluorescence was recorded, the process of serial dilutions was repeated 3 more times, and the results plotted (Fig 13). The effect of the phase systems on the linearity of this calibration curve was then studied. Samples of (5,4)j and (5,4)k phase system containing 101 / 102 Tween 80 (Table 3) were prepared as described above, and each phase was diluted 6.7 fold in the phosphate buffer to mimic the dilution required prior to viewing a sample from the phase system in the spectroflurometer. The sensitivity of the spectrofluorometer was then adjusted by introducing 1.0 jug of adhesin in a sample of the diluted phase system and adjusting the sensitivity to give the reading expected from the calibration curve using phosphate buffer. The adhesin sample was then replaced by the diluted phase system and the zero was adjusted. This process was iterated until the readings were consistent. The experiment above was then repeated replacing the phosphate buffer with the diluted phase systems. The results of these calibration curves are shown in Fig 13. / 103 N a t i v e F l u o r e s c e n c e o f A d h e s i n O n t h e S p e c t r o f l u o r o m e t e r • / / j6 A / / / O /* / ? / A o H — i — i — i — i — • — i — ' — i — ' — i — • — i — ' — i — • — i — 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 Micrograms of Adhesin Legend & Buffer + Low«r Phaa« X Uppar Phas* O Low«r Phasa + Tw«»n • Uppar Pho»« * TW«en Fig 13: Calibration curve of adhesin fluorescence in a Turner Model 430 spectrofluorometer with excitation wavelength=290 nm, and emission wavelength=331 nm. The buffer was pH 8. 0.1 M sodium phosphate. Phases from a (5,4)j and k system were diluted 7 fold in this buffer to produce the buffers for the upper and lower phases. The linear regression is from all the points. APPENDIX 2 Ka Calculated from the Fluorescence Assay The phase systems studied are summarized in Table 4 along with the partition coefficient of the adhesin found in these phase systems using the fluorescence assay. The mass of adhesin in phase i was 6.7 times the mass found in the 0.300 ml sampled in the spectrofluorometer, because only 0.300 ml was sampled from a total of 2.00 ml of each phase. The mass of the adhesin in the 0.300 ml of sample was determined from the linear regression of the calibration curve of ug of adhesin vs relative fluorescence units in Fig 13. For a given buffer, the upper phase value of "y" mv and the lower phase value of "z" mv were converted to " Y " Mg and "Z"Mg- The upper phase therefore contained U = (6.67 x Y) Mg and the lower phase contained L = (6.67 x Z)Mg of adhesin. For Buffer f: y = 5.6 mv ± 2 mv, z = 52.1 mv ± 2 mv. Y = 0.039 Mg ± 36%, Z= 0.901 Mg ± 3.8%. U = 0.25 Mg ± 0.09, L = 6.01 Mg ± 0.23 (4%). Therefore the total adhesin accounted for in the system was 6.26 Mg, whereas 6.6Mg had been added to the phase system. Thus 95% of the protein was accounted for, and the partition coefficient of the adhesin in the lower phase was K a = (6.01)/(6.01 + 0.25) = 0.96 ± 5% 104 A P P E N D I X 3 The Degree of Radiolabel l ing Glycophor in The activity of the N a 1 2 5 I on Apri l 23, 1984 was 13 mCiJug. Using the molecular weight of N a l as 150g/mole, and 1 mCi = 2.2 x 10 9 dpm, it was found that on Apri l 23, 1984 the N a l sample had 4.33 x 10 1 8 dpm/mole. On July 6 1985, the activity of the glycophorin was 15020 cpm/iil. Using 60 days as the half life of 1 2 5 1 , on Apri l 23, the glycophorin would have had 1.92 x 10 s cpm//xl, or assuming an 80% efficiency in the gamma counter, 2.402 x 10 6 dpm/iil. The sample was about 0.25 mg/ml in glycophorin, therefore, using 31,000 as the molecular weight of the protein, its activity was about 2.979 x 10 ^dpm/mole of glycophorin. Dividing the two activities gives 14 moles of glycophorin per mole of N a 1 2 5 1 . Adhes in The activity of the labelled adhesin was 2,666 cpm/iil. In order to find the dpm for this sample, an estimate had to be made of the efficiency of the beta counter. The adhesin was counted in parallel with a blank, by diluting 0.015 ml of the adhesin sample in the lower phase of the (5,4)f phase system, and boiling it in SDS using the method described below. After boiling the solution 33,440 dpm of 1 0 C toluene was added by weight, Atomlight was added, and the samples were counted. The toluene sample had 15.0 x 10 3 cpm, therefore the efficiency of the system was about 44.9%. The adhesin activity was therefore about 5,938 dpm/iil. The absorbance of the adhesin sample was 0.287, therefore if the extinction coefficient was 1.3 ml/mg-cm, then the concentration of the solution was 0.221 mg/ml. Using a molecular weight for 105 / 106 the adhesin monomer of 29,000, the activity of the adhesin solution was therefore 7.8 x 1 0 1 0 dpm/mole of adhesin. The H 1 t t C H O had an activity of 53 mCi/mmole or, using the same calculations as were used for the N a 1 2 5 I sample above, 1.2 x 1 0 1 4 dpm/mole H C H O . Dividing the two activities gives 6.5 H i a C H O per mole of adhesin monomer. This corresponds well with the value of lysines/adhesin monomer found in Table 5 using the fluorescamine assay. A P P E N D I X 4 Isotherm Calculat ions 1. Subract the background counts for each sample. 2. Find the weight of the sample in each vial. 3. Find the volume of the sample in each vial using the density of the upper phase = 1.020 g/ml and the density of the lower phase = 1.054 g/ml. 4. Find the cpm/ml for each sample and multiply by 2.00 ml to to get G, the gamma cpm in 2.00 ml of phase, and Bt, the total beta cpm in 2.00 ml of phase. 5. Find the beta counts due to carbon. This involves correcting for the decay of the 1 2 5 1 using: X(t) = X(0)e" a t , a = 4.814 x 10 ' 4 h r ' 1 where X(t) is cpm at time t (hr), X(0) is the activity at t=0, and a is the time constant for 1 2 5 1 . This decay is linear if t is less than 12 days, therefore since the beta counts were counted less than 2 days after the gamma counts, and the samples were counted in the same order, each sample was corrected by the same amount using the equation above, where t is the number of hr between taking a vial out of the gamma counter, and putting it in the beta counter. The cpm due to carbon is therefore: Be = Bt - B i = Bt -GqC where Bc= beta cpm due to carbon, Bt= total beta cpm B i = beta cpm due to iodine, G = gamma cpm, q = e"a^ C = (beta cpm)/(gamma cpm) for iodine from Table 6. A sample calculation was done for one of the points in Fig 10 A. A 107 / 108 Table 10 : Summary of Calculations 1 to 5 with one of the points in Fig 10 A. Phase Upper Upper Lower Lower G 1 8888 8816 6222 6312 % N 2 0.3 0.3 0.4 0.4 G q C 3 725 719 423 429 Bt 1 1466 1446 6615 6890 % N 2 0.5 0.5 0.2 0.2 Be" 741 727 6192 6461 1 as calculated in Part 4. 2 % uncertainty is 100/(total counts) total counts = cpm x min counted; samples were counted for 30 min. 3 q, from Part 5, is 0.969. * as calculated in Part 5. summary of the first five steps of the calculations for this sample are shown in Table 10. 6. The two samples taken from each phase were averaged: For the example this gives, with the gamma counts, 8852 ±36 cpm for the upper phase and 6267 ±45 cpm for the lower phase; For the beta counts: 734 ±7 cpm for the upper phase and 6326±134 cpm for the lower phase. 7. Adjust cpm to relative cpm (repm) using the relative efficiencies in Table 6. For the example: For the gamma counter: Upper phase repm = (8852 ±36) x (1.000) = 8852±36 Lower phase repm = (6267 ±45) x (1.017±2%) = 6379±135 For the beta counter: Upper phase repm = (734 ±7) x (1.000) = 734±7 / 109 Lower phase rcpm = (63261134) x (1.01012%) = 6389+ 186 8. Calculate the "load", or the activity of the protein solutions. For the example, the glycophorin load had 4032 rcpm/0.030 ml and (13.2ug± 10%)/(4032rcpm+0.5%). For the adhesin sample, the load was ljxg/2016rcpm. 9. If the percent free label was high (ie. above 2%), it was necessary to correct for this. A better solution was to clean the glycophorin sample as described above. If this was not done, the counts due to the free label were subtracted from the rcpm to give the rcpm', where: rcpm' =rcpm - TX(i) rcpm' is the rcpm for a given phase corrected for free label, T is the total rcpm in the upper and lower phase, and X(i) is the fraction of the total label which was not attached to protein in phase i . Since K for free N a l is 1.06, X(upper) = (0.516) x (%free label), and X(lower) = (0.483) x (%free label) For the example, there was 18% 1 3% free label with the glycophorin. This sample had 0.130 ml of glycophorin stock, therefore, from 8 above, T was [(4032 rcpm)/.030 ml] x [.130 ml] = 17472 rcpm 1 100[(3/130)2 + (3/30) 2 + (40320) _ 1 ) ] 1 / 2 % = 17472rcpm 110% For glycophorin therefore, rcpm' upper = 8852136 - (.09313% x 17472110%) = 7227+[(36)2 + (1624{(.03)2 + ( . 1 0 ) 2 } 1 / 2 ) 2 ] 1 / 2 = 72271216 (or 3%) / 110 repm' lower = 6267 - (0.087)(17472) = 4746 repm' load = 4032 x 0.82 = 3306 ±0.5% 10. Determine the partition coefficient, K g , for glycophorin. When K g was dependent upon the concentration of glycophorin in the phase system, the concentration dependence was quantified as follows: the micrograms of glycophorin in the upper phase was determined from the repm' load, and from the repm' of glycophorin in the upper phase. A plot was then made of repm' in the upper and in the lower phase vs the mass of glycophorin in the upper phase. This is shown in Fig 14. A smooth curve through these points was used to make the plot of K g vs the mass of glycophorin in the upper phase shown in Fig 15. The mass of glycophorin in the upper phase was then found for a given experiment, and K g was found from this plot. For the example, the upper phase had 7227±3% repm' x (13.2±10% Mg / 3306±.5% repm') = 28.9 ±10% Mg Therefore, from Fig 15, K = 2.02±.04 11. Determine the mass of glycophorin free in the lower phase, Gfl . Gfl = Gu/K Gu is the mass of glycophorin in the upper phase For the example Gfl = 28.9±10% yg / 2.02±2% = 14.3 ± 10% Mg T h e D e p e n d e n c e of K g o n t he M a s s of G l y c o p h o r i n in t h e U p p e r P h a s e + u p p e r p h a s e X lower p h a s e m i c r o g r a m s of g l y c o p h o r i n In the u p p e r p h a s e Fig. 14: The plot used to derive the dependence of the partition coefficient of glycophorin on the mass of glycophorin in the upper phase of the (5,4)f two-phase system when the 30 minute incubation of the protein was done in the presence of Tween-80. The line for the upper phase is a straight line because the repm' in the upper phase was used with the 'load' to obtain the mass of glycophorin in the upper phase T h e D e p e n d e n c e of K g o n the M a s s of G l y c o p h o r i n in t h e U p p e r P h a s e of t h e (5,4) f T w o — P h a s e S y s t e m ~ t — ' — • — • — • — i — • — • — • — — i — • — • — • — • — i — • — • — • — • — i •—• • • i • * • i ' • • ' i 0 10 2 0 3 0 4 0 5 0 6 0 7 0 m i c r o g r a m s of g l y c o p h o r i n in the u p p e r p h a s e Fig. 15: The partition coefficient of glycophorin, the (5,4)f two-phase system when the protein was incubated for 30 minutes in a 0.0416% Tween 80 solution. The values for Kg were obtained from Fig. 14. / 113 12. Determine the total glycophorin in the lower phase, Gtl , from the load and the total gamma counts in the lower phase. Gtl = 4746 rcpm' x (13.2 ug I 3306 rcpm') = 18.95 ± 10%. 13. Determine the mass of glycophorin bound in the lower phase, Gbl. Gbl = Gtl - Gfl 18.95±1.9 - 14.3±1.4 = 4.65±2.4 14. Determine A l , the mass of adhesin in the lower phase from the rcpm of adhesin in the lower phase and the load in the beta counter. A l = 6389±4% rcpm x d u g / 2622±.5% rcpm) = 2.43±4%. 15. Determine the moles of glycophorin bound per mole of adhesin monomer. v = (/ng glycophorin / ug adhesin) x (molecular weight of adhesin) / (molecular weight of glycophorin) v = (4 .65±51% ug) / (2.43 ±4% ug x (29,000/31,000)) = 1.9± 51% 16. Determine the concentration of free glycophorin in the lower phase, [Gfl], in moles per liter. / 114 [Gfl] = {(Gfl Mg) x (1 mole / 31.0 x 1 0 9 itg)} / (0.002 liters) 17. When there was a considerable amount of adhesin in the upper phase, it was necessary to iterate the procedure to account for glycophorin bound to adhesin in the upper phase. v' = (Gtl -(Gfu - v(Al x u/l))/K)/Al Where u/1 is the ratio of repm' due to carbon in upper and the lower phase for a given experiment. Therefore (Al x u/1) is the moles of adhesin in the upper phase, v(Al x u/1) is the approximate number of moles of glycophorin bound in the upper phase, Gfu - v(Al x u/1) is the moles of glycophorin free in the upper phase, and v' is the new v as calculated in 12 and 14. Iterate this equation replacing v with v' until v = v'. 18. Plot the isotherm which is v vs Gfl. 19. Plot the Scatchard plot which is v/[Gfl] vs v. GLOSSARY OF SYMBOLS AND ABBREVIATIONS BIS N,N-methylene-bis-acrylamide. Bq Becquerel, l B q = 1 s"\ 2.7 x 1 0 " 1 1 C i . C i Curie, 3.7 x 1 0 ^ disintegrations per second. E G T A Ethylene bis (oxyethylenenitrilo) tetracetic acid F 41 A n arbitrary designation indicating the presence of a particular surface structure or antigen. k Microscopic dissociation constant of the adhesin binding site. K 99 A n antigenic designation similar to F 41. K a Partiton coeficient of adhesin: (percent adhesin in the lower phase). kDa Kilo daltons. K e V Kilo electron volts. K g Partition coefficient of glycophorin: (mass in the upper phase)/(mass in the lower phase). LIS Lithium diiodosalicylic acid. M M Glycophorins A and B have the antigenic site for M M and N N type blood. M R Mannose resistant hemagglutination. M S Mannose sensitive hemagglutination. N N See M M . P A G E Poly acrylamide gel electrophoresis. P E G poly(ethylene glycol) rfl Fluorescence for a sample normalised with the fluorescence from 12 micrograms of a gamma globulin standard. SDS Sodiumdodecylsulfate. T C A Trichloroacetic acid. T E M E D N , N , N ' , N \ tetramethylethylenediamine. 115 Tris Tris(hydroxymethyl) aminomethane, xg Times one gravity. (5,4) The notation used for Dextran-PEG two phase systems, described section III 1. b. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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


Related Items