UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Protein binding studies by diafiltration Palmer, Cecily M. 1972

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1972_A6_7 P34.pdf [ 9.11MB ]
Metadata
JSON: 831-1.0101347.json
JSON-LD: 831-1.0101347-ld.json
RDF/XML (Pretty): 831-1.0101347-rdf.xml
RDF/JSON: 831-1.0101347-rdf.json
Turtle: 831-1.0101347-turtle.txt
N-Triples: 831-1.0101347-rdf-ntriples.txt
Original Record: 831-1.0101347-source.json
Full Text
831-1.0101347-fulltext.txt
Citation
831-1.0101347.ris

Full Text

PROTEIN BINDING STUDIES BY DIAFILTRATION by CECILY ti. PALMER Sc. (1969), B. Pharm. (19.70),. university of Otago Dunedin, New Zealand A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN PHARMACY in the Division of Pharmaceutical Chemistry of the Faculty of Pharmaceutical Sciences We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA APRIL, 1972 In presenting t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I f u r t h e r agree t h a t permission f o r extensive copying o f t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the Head of my Department or by h i s r e p r e s e n t a t i v e s . It i s understood that copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l gain s h a l l not be allowed without my w r i t t e n permission. Sc Department of ' ngrmacgutiCQ' •~>C\e.r*ce.& The U n i v e r s i t y of B r i t i s h Columbia Vancouver 8. Canada D a t e 3 • 5 • ABSTRACT A d i a f i l t r a t i o n technique was used to study drug-protein interactions. Fraction V human serum albumin and plasma and two drugs (phenylbutazone and bishydroxycoumarin) with a high a f f i n i t y for these substances were used in this investigation. Preliminary experiments were carried out to check for release of foreign substances and for binding of drug to the Ami con d i a f i l t r a t i o n apparatus. A binding experiment, in the absence of drug, revealed release of a protein-like, ultraviolet absorbing substance from Fraction V human serum albumin. The most suitable method of p u r i f i -cation for albumin was by d i a f i l t r a t i o n with T r i s buffer. Binding curves for bishydroxycoumarin - human serum albumin, phenyl-butazone - human serum albumin, and bishydroxycoumarin - plasma interactions were obtained. The r and r/D^ values were calculated and binding parameters estimated by both graphical extrapolation and by a computer non-linear least squares f i t analysis. Binding curves were not independent of human serum albumin concentration, but the cause of this effect was not f u l l y resolved. Results showed the d i a f i l t r a t i o n technique can yield precise data, can be used over a wide macromolecule concentration range and produces a bind-ing curve, from one experiment, over a wide range of molar binding ratios. Use of the Ami con d i a f i l t r a t i o n apparatus in desbrption (washout) experiments and equilibrium or direct experiments was also investigated. Attempts were made to obtain binding data by centrifugation (ultra-f i l t r a t i o n ) and by a gel f i l t r a t i o n technique (Sephadex G-25 batch method). These methods yielded unsatisfactory results which could not be compared with those obtained by d i a f i l t r a t i o n . This abstract represents the true contents of the thesis submitted. Supervisor - i i i -TABLE OF CONTENTS •'• Page I. INTRODUCTION ............ 1 II. LITERATURE 3 (1) Theory 3 (2) Physical and Biological Properties of Bishydroxy- . • . coumarin 13 (3) Physical and Biological Properties of Phenylbutazone 19 (4) Physical and Biological Properties of Human Serum Albumin 22 (5) Methodology 28 III. EXPERIMENTAL 36 (1) Apparatus 36 (2) Chemicals and Reagents 36 (3) Preparation of Tris Buffer ....... 37 (4) Determination of Absorptivity of Bishydroxycoumarin 38 (5) Determination of Absorptivity of Phenylbutazone .... 40 (6) Analysis of Mixtures of Phenylbutazone and Bis-hydroxycoumarin by Absorbance Ratio Method 40 (7) The Diafiltration Apparatus 42 (8) General Procedure for the Determination of Drug -Human Serum Albumin Binding Using the Diafiltration Technique 47 (9) Phenylbutazone - Human Serum Albumin Binding Studies by the Desorption (Washout) Diafiltration Technique . 50 (10) Bishydroxycoumarin - Human Serum Albumin Binding Studies by the Equilibrium or Direct (Diafiltration) . Technique 50 - iv -Page (11) Drug Binding Studies by the Centrifugation (Ultra-f i l t r a t i o n ) Method 51 (12) Phenylbutazone - Human Serum Albumin Binding Studies by a Molecular Sieve Technique Using Sephadex G-25 52 IV. RESULTS AND DISCUSSION ....... 59 (1) Preliminary Checks on the Diafiltration Apparatus ..... .59 A) Release of Foreign Substances from the Apparatus . 59 B) Binding of Phenylbutazone and Bishydroxycoumarin : . ' . to the Apparatus Components ..........— 59 C) Binding of Bishydroxycoumarin to the Diafi l t r a t i o n • Cell . . 6 0 D) Binding of Bishydroxycoumarin to Polyethylene Tubing . . 62 E) Binding of Phenylbutazone and Bishydroxycoumarin to Teflon and Skimatco Tubing 62 F.) Binding of Bishydroxycoumarin and Phenylbutazone to Components of the Apparatus (Connected with Teflon Tubing) . . . 6 4 G) Membrane Binding and Rejection: Diafiltration of j Phenylbutazone and Bishydroxycoumarin in the Absence of Human Serum Albumin 65 H) Membrane Retention of Human Serum Albumin: Dia-f i l t r a t i o n of Human Serum Albumin in the Absence of Drug ...... — 73 i) Effect of Time on the Appearance of the Unknown Substance in the Filtr a t e ... 73 i i ) The Effect of the Unknown Substance in the Filtrate on Phenylbutazone and Bishydroxy-coumarin Analyses .74 i i i ) Purification of the Human Serum Albumin (Fraction V) ............. 77 - V -Paje (a) Sephadex G-25 . 77 (b) . Dialysis with Cellophane Membranes 77 (c) DEAE Cellulose 78, (d) The Diafiltration Method ...... 82 iv) Nature of the Unknown Substance .. 82 v) Purification of Human Serum Albumin and Plasma ." by the Diafiltration Method .83 I) Void Volume in the Diafiltration Apparatus ....... 87 J) Fluctuations in the Cell Volume During Diafiltration 87 (2) Drug Binding Studies by the Diafiltration Technique ... 88 A) Calculation of Results ........... ............ 88 B) Plotting of Results 97 C) Bishydroxycoumarin - Human Serum Albumin Binding Results . 102 D) Phenylbutazone - Human Serum Albumin Binding Results 105 E) Bishydroxycoumarin - Plasma Binding Results ...... 117 F) The Effect of Human Serum Albumin Concentration on Binding Results Obtained by the Diafiltration Procedure ........ 122 i) Literature Reports on the Effect of the Human Serum Albumin Concentration on Binding Results : 122 i i ) Theoretical Investigation of Altered Para-meters Involved in Drug - Macromolecule -'J'1' Binding and their Effect on the Scatchard Plot 125 i i i ) Discussion of the Human Serum Albumin Con- -centration in Terms of Problems Inherent in the Di a f i l t r a t i o n Apparatus 131 - vi -Page (a) Ligand Polarisation at the Membrane ... 133 (b) Ligand Binding and/or Rejection by the Diafiltration Apparatus 134 (c) Altered Membrane Properties ... .... 134 (d) Conformational Changes in the Human Serum Albumin Molecule During Diafiltration .. 135 (3) Phenylbutazone - Human Serum Albumin Binding Studies by the Desorption (Washout) Diafiltration Method 136 (4) Bishydroxycoumarin - Human Serum Albumin Binding Studies by the Equilibrium or Direct Method 138 (5) Drug Binding Studies by the Centrifugation ( U l t r a f i l t r a -tion) Method 140 (6) Phenylbutazone - Human Serum Albumin Binding Studies by a Molecular Sieve Technique Using Sephadex G-25 140 V. SUMMARY AND CONCLUSIONS. ................... 142 REFERENCES • 147 - v i i -LIST OF.TABLES Table . Page 1 ••' Analysis of Bishydroxycoumarin and Phenylbutazone in Solution by the Absorbance Ratio Method 43 2 Binding of Phenylbutazone to Apparatus after 75 Minutes of Diafiltration 61 3 Binding of Bishydroxycoumarin to Apparatus after 75 Minutes of Diafiltration 61 4 Binding of Bishydroxycoumarin to the Diafiltration Cell ... 63 5 Binding of Bishydroxycoumarin to Polyethylene Tubing 63 6 Binding of Phenylbutazone to Components of the Apparatus (Connected with Teflon Tubing) 66 7 Binding of Bixhydroxycoumarin to Components of the Apparatus (Connected with Teflon Tubing) 66 8 The.Effect of the Unknown Substance in the Filtrate on the Analysis of Bishydroxycoumarin and Phenylbutazone ..... 76 9 The "Drugfit" Program 89 10 Bishydroxycoumarin - Human Serum Albumin Binding Results .. 110 11 Literature Values for n and k for Bishydroxycoumarin -Human Serum Albumin Binding Studies 115 12 Literature Values for n and k for Phenylbutazone - Human Serum Albumin Binding Studies 115 13 Phenylbutazone - Human Serum Albumin Binding Results 116 14 Bishydroxycoumarin - Plasma Binding Results . 121 15 Changes in General Data Cards ... 126 - v i i i -LIST OF FIGURES Figure Page 1 Spectral Characteristics of Phenylbutazone and Bishydroxy-coumarin .....<r.. 39 2 The Diafiltration Apparatus 44 3 Ami con (Model 52) Diafiltration Cell 46 4 The Centriflo Centrifugation (Ultrafiltration) Apparatus .... 51 5 Calibration Curve for Blue Dextran 2000 55 6 Phenylbutazone Adsorption to Sephadex G-25 ............ .. 58 7 Theoretical and Experimental Dilution Curves for Phenylbutazone 67 8 Theoretical and Experimental Dilution Curves for Bishydroxy-coumarin ....... 68 9 Plotting of Phenylbutazone - Human Serum Albumin Binding Results in the Manner of Blatt, Robinson and Bixler (1968) .. . 70 10 Plotting of Bishydroxycoumarin - Human Serum Albumin Binding Results in the Manner of Blatt, Robinson and Bixler (1968) .. 71 11 Ultraviolet Absorption Spectra of a 3.49 x 10~6 M HSA . Solution and for the Unknown Substance in the Fil t r a t e ... 75 12 Binding Experiment in the Absence of Drug: Appearance of Unknown Substance in the Filtrate During Experiment 75 13 Calibration Curve for the Determination of Human Serum Albumin at 280 mu .80 14 Calibration Curve for Protein Determination by the Folin -Ciocalteau Method 84 15 Disappearance of Protein-Like Substance from Plasma and Fraction V HSA on Diafiltration with Tris Buffer ............ 86 16 Calcomp Scatchard Plot for Bishydroxycoumarin - Human Serum Albumin Binding 103 - ix -Figure Page 17 The Effect of Human Serum Albumin Concentration on Bis-hydroxycoumarin - Human Serum Albumin Binding 106 18 Relationship Between ki and Human Serum Albumin Concentra-tion for Bishydroxycoumarin - Human Serum Albumin Binding : and Bishydroxycoumarin - Plasma Binding .. 107 19 Relationship Between r\i and Human Serum Albumin Concentra-tion for Bishydroxycoumarin - Human Serum Albumin Binding and Bishydroxycoumarin - Plasma Binding 108 20 Double Reciprocal Plots for Bishydroxycoumarin - Human Serum Albumin Binding ..' 108 21 . Calcomp Scatchard Plot for Phenylbutazone - Human Serum . Albumin Binding ........... '109 22 The Effect of Human Serum Albumin Concentration on Phenyl-butazone - Human Serum Albumin Binding ,. 113 23 Relationship Between n1 and Human Serum Albumin Concentra-tion for Phenylbutazone - Human Serum Albumin Binding 114 24 Double Reciprocal Plots for Phenylbutazone - Human Serum Albumin Binding 114 25 Calcomp Scatchard Plot for Bishydroxycoumarin - Plasma Binding 119 26 The Effect of Plasma Dilution on Bishydroxycoumarin - Plasma Binding' . .....120 27 - 30 Scatchard Plots Indicating the Effect of Changing Parameters \ \ on-Bishydroxycoumarin - Plasma Albumin Binding:-Figure 27 127 Figure 28 i '129 Figure 29.... ^ 130 Figure 30 . . . . 1 3 2 31 Binding of Phenylbutazone by Human Serum Albumin by the Desorption Method and Diafiltration Method 137 32 Bishydroxycoumarin - Human Serum Albumin Binding by the Equilibrium (or Direct) Method .... 139 - X -ACKNOWLEDGEMENTS I wish to express my sincere gratitude and appreciation to:-Dr. M. Pernarowski for his encouragement and guidance during my studies. Dr. A. G. Mitchell for academic advice. Mr. Rolf Klinger (Electrical Engineering Department) for his valuable assistance in preparing computer programs. Phyllis Moore for her able typing pf this manuscript. I. INTRODUCTION Serum albumin is a unique molecule. It binds extensively with most substances. However, lack of specificity distinguishes i t s binding characteristics from those between an enzyme and its substrate where.a high degree of spec i f i c i t y exists. Goldstein, in 1949, indicated the significance of drug interactions with proteins and most of the theory in this area has been developed using serum albumin as the protein model. Numerous studies, both in vivo and in vitro, have been undertaken to more f u l l y understand the nature of the interactions of ligands with protein molecules. Generally, most of these investigations have been an attempt to quantitatively study how much of a ligand binds, the number of binding sites on the protein, and the free energy changes involved in the process. Qua!itatively,studies have emphasized the type of forces involved and the nature of the binding sites. Methodology, in such studies, has varied widely. In quantitative studies by far the most commonly used method is equilibrium dia l y s i s . The chief disadvantages of this method are that i t is time-consuming and limited to low protein concentrations. Furthermore, each experiment yields only one data point for a binding curve. In this investigation, membrane d i a f i l t r a t i o n has been studied with respect to i t s applicability in yielding information on protein binding. Membrane u l t r a f i l t r a t i o n i t s e l f is by no means a new method. However, d i a f i l t r a t i o n , combining the best features of dialysis and u l t r a f i l t r a t i o n , is a relatively new technique. A solution of fixed - 2 -ligand concentration is passed under pressure through a protein solution and the f i l t r a t e is continually sampled for free ligand. Thus, from a single experimental run i t is possible to produce an entire binding curve. If the method is to be useful, d i a f i l t r a t i o n should yi e l d values comparable with those obtained by more conventional methods for the parameters which quantitate the binding of the ligand to the macro-molecule. The method was investigated by using human serum albumin (HSA) and two commonly used drugs with a high a f f i n i t y for albumin, bishydroxycoumarin (BHC) and phenylbutazone (PBZ). The binding charac-t e r i s t i c s of these two drugs to HSA have been determined by other methods and comparisons can, therefore,be made. In addition to studying the interaction of BHC and PBZ with HSA by the d i a f i l t r a t i o n method, two other methods ( u l t r a f i l t r a t i o n using centrifugation and gel f i l t r a t i o n ) were investigated. These methods . yielded erratic results and, therefore, the emphasis in this thesis is on the d i a f i l t r a t i o n method. Theoretically the di a f i l t r a t i o n method could be used to investigate the simultaneous binding of two substances to HSA. It is well established that protein binding is of in vivo significance and that the extent of binding, in some instances, will change when two drugs are administered concurrently. Although a method of analysis was developed for the BHC in the presence of PBZ, no detailed \ investigations in competitive binding were carried out because of unexpected d i f f i c u l t i e s with the d i a f i l t r a t i o n apparatus. II. LITERATURE (1) Theory The importance of the interaction between ligands and plasma protein is now well established. Bennhold (1938) recognized that the plasma proteins are the special transport system for the regulated distribution of both naturally occurring and medicinal substances throughout the body. Goldstein (1949), in his classical review, a t t r i -butes much of the impetus in this area to the studies by Cohn (1946) who isolated and characterized plasma proteins. Much of the work in the early part of this century was empirical and lacked.any clearly formulated theoretical background. It was not until advances were made in drug therapy, and p e n i c i l l i n and sulphonamides were in use, that i t became clear that drug-protein interactions were important. Goldstein reviewed the ways in which data describing protein bind-ing can be quantitatively expressed. The Freundlich isotherm states that at a given temperature the amount of substance absorbed per weight of absorbant is inversely proportional to some power of the concentration of unabsorbed material. However, this is a purely empirical function and has no thermodynamic validity. On the other hand, the Langmuir isotherm can be rationally interpreted. This approach relates the amount of substance absorbed (x) per weight of absorbant (m) to the concentration (c) of unabsorbed material. The constants in the equation are a and b. x^  m = be (Eq. 1) _ 4 -As Langmuir noted, simply because an absorption isotherm is produced, i t does not necessarily follow that binding forces are physical rather a l l binding forces as being s t r i c t l y chemical in nature. This concept is similar to that expressed by the Law of Mass Action. Klotz (1946 a) developed his approach to protein binding from the Law of Mass Action. For this approach, several assumptions are necessary: (1) All groups on the protein molecule, capable of interacting, have identical a f f i n i t i e s for the ligand molecule. . (2) The a f f i n i t y of any group is unaffected by combination of ligand molecules with other groups. (3) Activities can be replaced by concentration terms. (4) The interaction between groups on the protein with ligand molecules is entirely reversible. The derivation here is that given by Edsall and Wyman (1958) and the . terminology has been adapted to express i t in terms of association, rather than dissociation. A protein molecule (M) combines with a ligand molecule (D), according to the Law of Mass Action, than chemical in nature. In his opinion, i t was better to consider M - D, b (Eq. 2) [M] [D f] 1 J T 1 (Eq. 3) - 5 -where [M] = total protein concentration in moles/1. [Df] - ' concentration of unbound 1igand molecule in moles/1.. [M - D^ ] = concentration of ligand bound to protein molecules ; in moles/1. k „ = the association constant = 1/K,. ass diss If '« is the fraction of unoccupied binding sites on the protein molecule, .'then .• ;' ~ - M (Eq. 4) [M] + [M-Db] [M] [M] + [D f] kass[M] 1 1 +^ka s s (Eq. 5) (Eq. 6) Let r = fraction Of occupied sites = [D^ -MlEM], Therefore, r = 1 - « = [Dh-MJ (Eq. 7) [Db-M] + M r = kass[ Df] (Eq. 8) . 1 + kasst°f3 ' ; : . . V : ^ - ; ; v < ; ' ; r, in the above equation denotes the average number of ligand molecules bound to a molecule of protein. Eq. 8 describes the situation where the protein molecule has one site capable of binding ligand. This equation can be extended from one to 'n' binding sites. • /.: - 6 -This approach has been considered in reviews by Klotz (1946 a), Edsall and Wyman (1958), Steinhardt and Reynolds (1969). The follow-ing derivation is general and is that described by Steinhardt and Reynolds (1969). Consider a number of uniform particles, M, in solution each with n indistinguishable and identical sites. The equilibrium between ligand and sites on the macromolecule, M, can be described by: [M] + n[D f] W = [M - D b n] (Eq. 9) or else by stepwise addition of ligand: [M] + [D f] [M - Db. ] : [M - D b J + [D f] ^ — ^ [M - D b ?] . (Eq. 10) [M - D b 2] +.[D f] [M - D b 3] [M - Db ] + [D f] ^ — ^ [M - D. ] i - i i [M - D ^ ] + [D f] [M - D b n] where at equilibrium: [M - D b >] k = (Eq. 11) 1 CM - Db ] [D f] i - i , •' . \ • • Assume that (1) the association is thermodynamically reversible, and that (2) [M - D b > ] , [D f], [M - Db ] refers to a c t i v i t i e s . The number of possible combinations of n binding sites taken exactly i at a time (oralternatively the number of equally probable - 7 -forms of the complex [M - D, ] ), i.e., PM- D = : (Eq- 12) \ i ! (n - i ) ! For example, there are n probable forms of M - but only one form of M - D b n. By definition, a l l sites are equal and i f present alone must have the same in t r i n s i c association constant, k .• If the number of equally probable forms of each complex are taken into account, the k. 's may be deferred in terms of a single equilibrium constant, k Q. Again, from probability theory, M contains n possible sites but M -D^ can release only one ligand molecule. Hence, k1 = nk . Similarly, M - D, ^ - 1 contains n - 1 possible sites but M - can release two ligand molecules. Therefore: K - N - 1 • \ : : • k i = n ' | f 1 • k 0 (Eq- 13) kn = " ' k° ' " V "•.•^ ^ Edsall and Wyman (1958) noted that these s t a t i s t i c a l factors are always present at equilibrium, whether or not sites are independent and equivalent and therefore binding constants should always be expressed corrected for these s t a t i s t i c a l factors; i.e., k xo = k]7n, etc. Hence k 2o = k 2o = ••= k^o and any observed differences between these values should indicate in t r i n s i c differences between affinities of the groups, or interactions between groups, or both. - 8 -The average number of bound ligand molecules may be expressed as (Eq. 14) CM -'Db,] + 2[M - Db.] + — --n[M - Dfa ] (M] + [M - D b ] + — — — — [ M - D. ] • 1 n As a result of systematically applying Equation 11 to Equation 14 and then substituting the relationships between and k' defined in Equation 13, the following relationship for r is obtained: n ' v> . n! i i m ni 1 (n-Ol i ! x ko x [ D f ] . r = : : : : : (Eq. 15) 1 +4 W I T x ko >< tD,]1 The denominator is the general expression of the binomial theory and the numerator is the f i r s t derivative thereof. Hence, after appropriate substitutions and substituting K for the product of combined activity coefficients and k. o n K D, 1 + KDf (Eq. 16) This equation, which describes the binding of ligand to n sites (identical in binding ability) on a protein molecule, may be further extended to a situation where there is more than one set of sites on the molecule. Assume that set one has n1 sites with i n t r i n s i c association constant Kx, and set two has n 2 sites with intrinsic association constant K2. Then Equation 16 is extended to:-niKiD f n 2K 2D f n nK nD f . . r = — — — — + — + • • • + (Eq. 17) 1 + K xD f 1 + K 2D f : 1 + K nD f : - 9 -Complete independence of sets of sites is assumed. If the values of. K. are separated by a factor of more than 1 0 \ then at any particular : value only one term of Equation 17 will depend on D^ , and in this case r will be determined only by one particular . Equation 8 can be extended to describe a situation where there is only one set of sites but two different ligand molecules, A and B, . competing for these sites, each with a different association constant, K • and Kn. : • r f l = M TA (n - r R ) =, A rA (Eg. 18) A — v" 'B' FA n 'A u 1B 1 + K A ° f n 1 + K A D f n + K B D f c n kB D f Similarly r g = B 1 + KR D f + Kfl Df B A Equation 16 is used most often in quantitative binding studies. This assumes a fixed number of binding sites. Frequently, when this equation is used to express binding data and results are plotted in the manner indicated by Scatchard (1950), non-linearity of the plot is observed;, Karush. (1950 a) found that this non-linearity could not be explained by electrostatic interaction (where binding of ligand molecule may cause an alteration in the net charge on the molecules) or by a Gaussian dis-tribution of free energy of binding. He concluded that n is not. necessarily the maximum number of binding sites per molecule.. Non conformity of data to Equation 16 could arise from one or more of several factors:-- 10 -(1) Steric interference between bound molecules. (2) When a charged molecule is bound, electrostatic interaction may arise from the increasing charge on the protein molecule as more ligand binds. (3) Interaction of the ligand with the molecule may induce con-formational charges. The f i r s t of the above factors, steric interference, has not often been discussed in the literature but could be considered on a structure-activity basis. Electrostatic interaction has been mentioned frequently. Various models have been proposed to describe the above. For example, Tanford (1965) assumed that both ligand and protein molecules were rigi d spheres. Similarly, Steinhardt and Reynolds (1969). concluded that there is an additional contribution to the free energy from electrostatic inter-actions when the ligand added to the protein is charged. This additional contribution is the difference in charging energy between MD. and . V l MD^  , i.e., the electrical work done in increasing the charge on the surface when one charged ligand molecule is added. . •"' A function f(r) is defined such that f(r) = 0 at r = 0 and thereby yields an equation which varies with the number of bound ligands V = K 0 e [ ~ f ( r ) ] f-,v (Eq. 19) AF1 = -RT In KQ + RT In f(r) (Eq. 20) n K e["f(r)3 x Df .'. r = ^ : L _ (Eq. 21) 1 + K e ^ ™ x D f 0- T - 11 -From the Debye - Hu'ckel Theory: e [ - f O ) ] = e[-2wz] where e 2 1 K w = - - • • . . 2DRT b 1 + K ' a A more r e a l i s t i c model for electrostatic interactions was proposed by Tanford (1957) and Tanford and Kirkwood (1957). This model takes into account the location of charge density on a non-conducting surface in terms of distance between the charged sites and the position of these sites relative to the interface between the binding surface and the medium in which i t is immersed. Induced conformational charges in the protein molecule have also been mentioned in the literature in an attempt to explain non-linearity of binding data. Configurational adaptability was f i r s t suggested as a possibility by Karush (1950). This suggestion was supported by the observations that on binding of small molecules and ions to macro-molecules, conformational changes occurred; i.e., conformational changes as manifested by changes in viscosity, optical properties and anomalies in the shapes of binding curves. Investigations in this f i e l d , in more recent years (Steinhardt and Reynolds [1969])., suggest that the forces responsible for bringing about these conformational charges may be one or more of the following:-(a) electrostatic repulsion between bound ligand molecules, includ-ing the net charge on the molecule. Tanford (1965), on investigating - 12 -the effect of increased net charge on the macromolecules and applying Equation 21, found that an increase in the distance between binding sites will decrease w and consequently the free energy of the system. Therefore, as the number of sites occupied the free energy of the system. (b) penetration of a hydrophobic t a i l into a polar region of the macromolecule with the resultant replacement of the conformation stabilizing segment-segment forces by 1igand-segment interactions. Kauzrnan (1959) suggested that hydrophobic bonds, hydrogen bonds and dipole moments are important for stabilizing protein structure. If such segment-segment interactions are replaced by 1igand-segment interactions on binding of ligand, the s t a b i l i t y of the protein may be decreased. There are no proven examples of this. (c) a ratio of the number of binding sites and their association constants in the native form to those in the unfolded form of the macromolecule which is favourable to unfolding. This factor was f i r s t suggested by Foster and Aoki (1958) when discussing the N ->• F (N = native form -»- F = unfolded form) transition which occurred at pH.4 with bovine serum albumin. Steinhardt and Reynolds (1969) described a binding isotherm when such a transition occurred. The complete expression i s : -by a charged ligand increase, the molecule will tend to decrease D f nKQ F(D>) m J r = (Eq. 22) 1 + FD, 1 + K D, f L o f - 13 -where n = number binding sites on native form m = number binding sites on unfolded form . KQ = in t r i n s i c a f f i n i t y constantfor native , form J i n t r i n s i c a f f i n i t y constant for unfolded form F(D^)= ratio of protein present in the two states. • Although the above and other authors have attempted to explain . deviations from the expected binding data, i t is unlikely that more progress will be made in this f i e l d until more is known about HSA at the molecular level. (2) Physical and Biological Properties of Bishydroxycoumarin Chemically known as 3,31-methylene-bis-(4-hydroxycoumarin), BHC is a white crystalline or amorphous powder with a molecular weight of 336.29 and a melting point of 287-293°C (Merck Index). Its structure i s : Cho (1970) suggested that the low water-solubility of the unionized molecule arises from the formation of intramolecular bonds as shown below: - 14 -On the basis of his data, pkai '= 4.44 and pka 2 = 7.83. After the f i r s t ionization has occurred, solubility'would increase only slightly because one intramolecular bond is s t i l l present. After the second ionization has occurred a considerable increase in sol u b i l i t y should occur, because intramolecular hydrogen bonds are no longer present to maintain the • molecule in a r i g i d position. Coumarin drugs gained popularity in anticoagulant therapy following the observation of the anticoagulant effect of spoiled sweet clover hay on cows. The active principle causing the anticoagulant effect was a coumarin type compound (Link [1944]). BHC, one of the coumarin drugs, has been used therapeutically for approximately thirty years. Weiner and others (1950) were the f i r s t to report the binding of BHC to HSA. At levels of 5-100 mg. BHC/1. BHC was 99% bound to albumin, . 20% to (3 and y globulins and 50% to « globulins. At therapeutic dosage levels, the free BHC concentration was only 20yg/ml. indicating i t s ' high a f f i n i t y for plasma and partially explaining the long persistence of the drug in the body. Only very small amounts of BHC were found in the brain, red blood cells and cerebrospinal f l u i d . O'Reilly and others (1952) also reported that BHC was 99% bound to albumin. The same authors (1964) -15 -found that for a l l doses of warfarin and high doses of BHC, their long elimination half-lives were due to their small volumes of distribution, which are very similar to that of HSA. However, Solomon and Schrogie (1967) showed that the magnitude and duration of-the biological response to BHC was not related to i t s h a l f - l i f e . They suggested that the variation in , response amongst patients was a result of variable absorption and metabolic rates, possibly being explained by a difference in a f f i n i t y of BHC for receptor sites. O'Reilly and Aggeler(1964b), however, had a different explanation for this variation in response to BHC. They said this dis-continuous response to the drug must be genetically determined:(i.e., the distribution of the response to anticoagulant therapy instead of being normal or unimodal was multimodal). The above studies have shown that BHC has a very high a f f i n i t y for plasma. Hence, protein binding must influence the distribution and metabolism of BHC. Levy (1970) concluded that the increased retention of BHC up to a point where saturation of binding sites occurred could be accounted for by a conformational change in HSA on i n i t i a l combination with BHC, thus making more sites available. Interaction of other drugs with anticoagulants has been extensively investigated. Clinical observations have shown that BHC blood levels •'. change when certain other drugs are administered.. Most literature references deal with warfarin but these interactions would be expected ; to apply to BHC. Examples given here will be restricted to interactions of warfarin or BHC with PBZ or to interactions of drugs with BHC. - 16 -Weiner and others (1950) showed a decreased disappearance rate for BHC from the plasma, on concurrent administration of oxyphenylbutazone (oxy-PBZ). They suggested that the resultant enhanced hypoprothrombinaemia was due to oxy-PBZ inhibition of the hepatic enzymes. Aggeler and O'Reilly (1967), noting that this result ( i . e . , of Weiner and others) was contrary to that observed with warfarin, said these results were probably incorrect as the BHC levels calculated were not corrected for oxy-PBZ present. Aggeler and others (1967) found that in man PBZ competes with anti-coagulants for binding sites on plasma albumin, resulting in an increased delivery of coumarins to the l i v e r . Schrogie and Solomon (1967) showed that for doses of D-thyroxine,\ clofibrate and norethandrolone which did not affect the metabolism of BHC nor the levels of Vitamin K dependent clotting factors, the anticoag-ulant response could be increased. This they attributed to an altered a f f i n i t y of the receptor sites by the drugs. O'Reilly and Aggeler (1968) gave results for warfarin and PBZ interactions. Results would be expected to be similar for BHC. They showed by fluorometric assay that PBZ increased the prothrombin time and decreased the warfarin h a l f - l i f e and plasma warfarin levels, lead-ing to an increased amount of drug being available for metabolism in the 1iver. ' ' •.• Welch and others (1969) found that in dogs and in man PBZ had an opposite effect on anticoagulant activity. O'Reilly and Aggeler (1968) explained such an . effect in man by PBZ stimulation of drug metabolism via the hepatic microsomal enzymes. This could lessen the hypopro- . thrombinaemic affect of anticoagulants in the dogs but enhance i t in man. - 17 -Another interaction investigated was that of heptobarbital by Levy and others (1970). Heptobarbital decreased the response to BHC in man. In oral doses, heptobarbital did decrease the absorption o f . BHC but irregardless of the route of administration, prothrombin time, plasma levels and h a l f - l i f e were a l l decreased. ..V < ' v . O'Reilly and Levy (1970), who studied warfarin and PBZ, gave a more probable explanation for such interactions. The interaction, they concluded, may be the net result of several effects and i t s duration and magnitude may be dependent on the dosage of PBZ and the duration of its administration. The increased elimination rate and potentiation of pharmacological action of warfarin would both seem to be due, primarily, to a change in distribution. More recent studies on BHC have been concerned with the nature of the binding of BHC to serum albumin using various techniques. Chignell (1970) investigated the binding of BHC to HSA by circular dichroism and found that after three moles of BHC had bound there was no further change in molar e l l i p t i c i t y ( i . e . , no further change in the difference spectra). Perrin and Idsvoog (1971) studied the binding of BHC to bovine serum albumin (BSA) by circular dichroism.and differential absorption spectroscopy and found no major conformational changes occured on . binding, suggesting l i t t l e or no departure from the ^-helical structure of BSA at pH 7.4. However, the changes observed in molar e l l i p t i c i t y as BHC binds to BSA do not parallel those observed by Chignell (1970), • suggesting that binding results with BSA cannot always be extended to HSA Cho and others (1971) used equilibrium dialysis, spectrophotometry and solubility analysis to study the binding of BHC to HSA in more detail Their results suggested that the unsaturated lactone structure on the BHC molecule is involved with binding. The main binding energy was shown by thermodynamic studies to be derived from non-ionic forces. The binding sites appeared to be hydrophobic regions on the albumin molecule whose van der Waals contours are selective for the BHC molecule. O'Reilly (1971) investigated the interaction of a number of coumarin compounds with HSA by equilibrium dialysis and made several, general observations about the effect of structure on binding. Precursor type of compounds showed a one to one maximum binding ratio, whereas mono-coumarins and dicoumarins showed a two to one ratio. The strength of binding to albumin was related to the hydrophobic nature of the sub-stituents introduced to the coumarin molecule. Thermodynamic studies suggested that coumarin drug binding to HSA involved both hydrogen bonds (from favourable enthalpy changes) and hydrophobic bonds. A decrease in the number of binding sites as the pH is changed from 7.4 to 10, suggested that binding sites are not rigid but rather are influenced by environmental factors. One other observation made was that there was l i t t l e difference in binding energies for (-) S-warfarin and (+) R-warfarin, and yet the (-) S form has far greater anticoagulant activity in man. As a result of this observation, O'Reilly suggested plasma albumin might be an inadequate model for studying binding to intracellular receptors for any drug, the inadequacy arising perhaps - 1 9 -from the two dimensional nature of plasma albumin as compared to the three dimensional nature of intracellular receptors,which can distinguish optical isomers. Chignell and Starkweather (1971) studied the interaction of PBZ, flufenamic acid and BHC with acetylated HSA, HSA which had been previously incubated at 37° for 24 hours, and untreated HSA, by equilibrium dialysis and circular dichroism. They found that with acetylated HSA and incubated HSA, the affinity of BHC was increased several fold. They concluded that, in the former case, the binding site was altered. With incubated HSA, they suggested that the HSA conformation had changed and thus much more BHC could be bound. (3) Physical and Biological Properties of Phenylbutazone Phenylbutazone is a white crystalline powder, insoluble in water and has a melting point of 104.5 - 106.5°C. Its molecular weight is ' , 308.37 (Merck Index). Decomposition in aqueous solutions at room temperature has been . observed. The kinetics of this decomposition has been investigated by Pawelcyzk and others (1968), (1969). Their results indicated that the decomposition of PBZ (in ammonium acetate buffer at pH 8.6) after 24 hours at 37°C would be insignificant. Solutions can be kept at room temperature for up to one week. However, i f stock PBZ solutions are stored at 4°C they can be kept for several weeks without significant decomposition. - 2 0 -PBZ has a pK value of 4.5 and, hence, at pH 7.4 i t would be a expected to exist in the enol form (von Rechenburg [1962]). Diketone form of PBZ Enol' form of PBZ The above anion will have definite hydrophilic properties. Burns and others (1953) showed that at plasma levels of 50-150.mg|1. PBZ is 98% bound but at levels greater than 250 mg|l, less than 88% is . bound. With increased doses a plateau in plasma levels was reached. ' Pharmacokinetic studies indicated the excretion of PBZ was not increased nor were the PBZ tissue levels increased when the plateau in plasma levels was reached. Their explanation for the lack of dependence of blood levels on dose (at high dosage levels) was an increased PBZ meta-bolism rate. Brodie (1956) explained this lack of relationship between ' blood levels and dose, at high dosage levels, on a basis of saturation of PBZ binding sties. Under these conditions the amount of free drug, which is rapidly metabolized, increases. Von Rechenburg (1953) explained this phenomena in a similar manner. Brodie, and others (1954), showed that PBZ was bound to plasma proteins, particularly the albumin and <*-globul in fractions. Pulver, and others (1956) found, by dialysis, that, for dilute solutions, PBZ was 99% bound to serum but for more... concentrated solutions was only 94% bound. Wunderly (1956) using e q u i l i -- 21 -brium di a l y s i s , showed that with dilute protein solutions, an equilibrium between bound and free drug exists and this equilibrium constantly changes with PBZ concentration. He considered PBZ was bound to HSA not by chemical forces but rather by being held within the hydration shell enveloping the protein molecule. Thorp (1964) and Solomon, and others (1968), the former using equilibrium dialysis and the latter an u l t r a f i l t r a t i o n technique, found similar values for n and k, . Chignell (1969) investigated as s the nature of the binding of PBZ by HSA by means of optical studies of the PBZ-HSA complex. Cn binding of PBZ, a positive e l l i p t i c a l band occurs at 287 my; in the circulardichroism spectrum. This he explained resulted . from perturbation of the n-ir* transition in the carbonyl chromophore of PBZ by an assymmetric locus at, or near the albumin binding site. As was shown by displacement of a fluorescent probe for hydrophobic regions, hydrophobic interactions are important for maintaining a rigid drug-protein complex. With the introduction of hydrophilic groups into PBZ, the magnitude of the induced optical a c t i v i t y is considerably reduced. Chignell and Starkweather (1970) found, by equilibrium dialysis, that HSA had an additional two binding sites of low affinity for PBZ. They also noted that the a f f i n i t y of PBZ. to acetylated HSA or incubated.HSA (incubated at 37° for 24 hours) was altered when compared to that for normal untreated HSA. The number of binding sites was increased and i t appeared that these treated HSA samples had altered conformation. Chignell and Starkweather (1971) gave the symmetry rules for the n-n* transition , in the carbonyl group. This could obey either the quadrant rule or the octant rule. However, at present these changes in extrinsic optical - 22 -activity cannot be interpreted in terms of specific alterations at their protein binding sites, although i t could be said that acetylation of HSA or incubation at 37° for 24 hours alters the nature of the binding sites. Rosen (1970) gives quantitative binding data from circular dichroism studies on the binding of HSA and PBZ. From the changes in molar e l l i p t i c i t y as HSA is titrated with PBZ, he could determine the amount of bound drug and free drug present. .Thus he calculated n and k values. . . . Studies on the c l i n i c a l interaction of PBZ with anticoagulants were . outlined in the previous section. (4) Physical and Biological Properties of Human Serum Albumin Plasma contains approximately 4% w/v HSA. Values reported in the literature are 46.2 g/r. (Geigy Scientific Tables), 4.62% w/v (Riva [I960]), 3.575% w/v (Hitzig [I960]), 3.5 - 4.5% w/v (Schultze and Heide [I960]). Its physical parameters are described.by Conn, and others (1947). It has a molecular weight of 69,000 and i t s molecular shape is e l l i p s o i d a l , being approximately 150A° in length and 38A° in diameter. Schultze, and ' others (1962) determined i t s amino acid composition and the terminal . groups are characterized but l i t t l e is known about i t s primary.structure. A knowledge of such structure would answer questions about the hetero-geneity of the HSA molecule. It is known, however, that the amino acid residues are linked by a single peptide chain which is crosslinked by disulphide bridges. - 23 -Rapid advances in protein binding studies were made after Conn (1946) produced his scheme for the fractionation of plasma. Fraction V contains 95% or more albumin and also small amounts of transferrin, o^-acid glycoprotein and <*i-antitrypsin. Commonly, a Cohn-type method, is used to produce HSA commercially. Recrystallized HSA is used in some binding studies, but there appears to be l i t t l e difference in binding characteristics between crystalline HSA and Fraction V (Kostenbauder, and others [1971]). Foster (1960) summarized the types of molecules bound by HSA. Divalent and polyvalent cations bind more strongly than monovalent, as would be expected because of the net positive charge on HSA. The trans-ition element metals bind more strongly but d i f f e r in that they are capable of forming coordinate covalent types of bonds. Anions are bound much more strongly than cations (Klotz and Ayers [1953]). Significant differences have also been observed in studies involving albumin-isomer interactions (Klotz and Ayers [1952]). Klotz (1947) studied the binding of dyes and found that generally their binding is greater than that of the anions. Surface active agents exhibit even higher a f f i n i t i e s for the protein. Binding is cooperative and a large number of ions are bound (Foster [I960]). Karush and ; Sonenburg (1949) suggested that cooperative binding was due to a struc-tural change in the protein and hence a change in the nature of binding occurred as ligand concentration increases. Yang and Foster (1953) electrophoretically investigated the binding of alkylbenzene sulfonate - 24 -to albumin. He observed a single boundary similar to that of the native protein, which corresponded to a distribution of the molecules in a s t a t i s t i c a l manner over a l l the available albumin molecules. However, there was also a separate, faster moving boundary. The authors explain this as an all-or-none reaction where the molecule partially unfolds; allowing more ligands to bind. There is a third phase to this binding process where the binding ratio apparently increases without limit but presumably a f f i n i t y to the sites is much lower. Klotz and Urquart (1949) investigated the significance of thermo-dynamic changes occurring during binding of organic molecules to albumin. The large positive entropy change could not be due to unfolding and dis-orientation of the protein molecules, since the change in entropy was small and negative. He explained this in terms of the ice-like structure of water surrounding the macromolecule; i.e., there are several polarized water groups attached to the carboxyl and quaternary ammonium groups'. When the ligand molecule binds to one of these groups, water molecules are . then displaced. Therefore, an increase in the number of molecular species occurs, causing an increase in entropy. This view is also supported by an increase in volume. Major contributions to free energy of binding must come from the release of solvent molecules from the complex. Lovrien (1963) studied the binding of detergent molecules to BSA . and showed that even at low Df values, conformational changes occurred, resulting in a decrease in hydrodynamic volume. Klotz and Ayers (1952) noted an increase in the binding capacity of BSA for anions as the pH increases. This is the reverse of what would be expected i f the charge - 25 -on BSA is taken into account. A change in the protein molecule occurred above pH 7. This they described as a definite change in the conformation of the molecule, i.e., a reversible swelling or unfolding produced by strong internal electrostatic repulsions. Klotz and Ayers (1953) suggested that, for neutral molecules, the very act of complexation makes more sites available. On the basis of the above, i t is obvious that the mechanism of binding of ligands to albumin has not yet been f u l l y resolved. Steinhardt and Reynolds (1969) state that HSA and HSA are micro-' heterogeneous and a solution of albumin, therefore, contains a mixture of components which are not spontaneously interconvertible. These com-ponents can be separated because of their varying degree of susceptibility to partial unfolding by acid (Peterson, and others [1965], Sogami and Foster [1968], Peterson and Foster [1965 a,b]). The effect of this heterogeneity on binding properties has not as yet been determined. Its implications in preventing any simple thermodynamic analysis of the 'N^=^-F' equilibrium are clear. Both HSA and BSA show anomalous hydrogen ion titration behaviour below the isoelectric point and beginning at about pH 4. Yang and Foster (1954) studying this effect and its. i n f l u -ence on optical rotation and viscosity, concluded that isotropic expansion of the molecule occurred in acid solution as the positively charged ammonium groups repulsed each other. . Tanford and others (1955 b, 1956) confirmed and suggested that this expansion occurs through an intermediary expandible "F" form. - 26 -Sophisticated methods of analysis (such as gel electrophoresis, ion exchange chromatography, etc.) have indicated that albumin may not be a single molecular species (e.g., Pederson [1962]). The apparent hetero-geneity of albumin may be caused by:-(i) . the binding of endogenous molecules to the macromolecule. Binding of anionic substances increases the electrophoretic mobility of albumin, as is observed when bromophenol blue dye is used as a trace for albumin. Similarly, the mobility of albumin is altered in the nephrosis, perhaps because of binding of fatty acids. ( i i ) the association with other globular proteins in the serum. ( i i i ) the polymerisation of albumin.' Hughes (1947) showed that only 70% of the HSA molecules would form mercaptalbumin : dimers. Pederson (1962) and others have shown that dimers and higher polymers are present in HSA solutions under various conditions. ' .. (iv) true microheterogeneity in the HSA molecule. This is sub-stantiated by the formation of mercaptalbumin dimers (Hughes [1947]). (v) bisalbuminaemia. In certain diseases, e.g., diabetes m e l l i t i s , tuberculosis, hepatitis, electrophoretic. analysis of the serum has occasionally shown two, rather than the one expected albumin peak. A more detailed account of possible factors giving rise to heterogeneity is given by Schultze and Heremans(1962). - 27 -In spite of the above factors, albumin is a suitable model for ... binding studies in that i t is less susceptible to denaturation than other similar proteins. Its denaturation is readily reversible because mole-cules aggregate whilst s t i l l in their folded state, t h i s . i s in contrast to ovalbumin where disaggregation would result in a highly unfolded and disoriented molecule. As a result of i t s high a f f i n i t y for a wide range of molecules, problems may arise in binding studies. For example, Karush (1951) showed that the phosphate ion bound competitively with BSA. Thus even buffer ions may compete with the ligand being examined and care must be taken in choosing a solvent system. Another consequence of the high a f f i n i t y of albumin for ligands is that the macromolecule s t i l l has one or two equivalents of endogeneous fatty acids attached to one or more of i t s higher energy sites. Various techniques have been described in the literature for the removal of these fatty acids. Goodman (1958) passed the HSA solution through a mixed ion exchange bed and then treated the eluate with iso-octane and glacial acetic acid. Chen (1967) removed the fatty acid by using mild acid and charcoal. However, Sogami and Foster (1968) found small but definite differences in s t a b i l i t y and UV absorption characteristics between un-treated and fatty acid free albumin. These fatty acid residues could significantly alter binding results, particularly for fatty acid-like . neutral molecules. The fatty acid molecule might occupy one,or part of one of the binding sites for the molecule in question. Therefore, at low molal binding ratios, discrepancies in data would be l i k e l y to occur when different protein preparations are used. - 28 -(5) Methodology Any quantitative investigation of ligand binding to protein must yield numerical values of bound and free drug. Such studies must, therefore, depend on changes in the properties of the interacting mole-cule or in changed properties or behaviour of the macromolecule. Reviews On methodology have been given by Goldstein (1949), Meyer and Guttman (1968), and Steinhardt and Reynolds (1969). . Some methods measure changes in properties of the ligand molecule on binding to the macromolecule. Examples of methods mentioned here will be those which have been used '.' in PBZ or BHC binding studies. Cho (1971) measured the binding of BHC to HSA by the depression of the absorption maxima of the BHC spectra in the presence of macromolecule. Perrin and Idsvoog (1971) measured changes in extrinsic Cotton effects (expressed as differential e l l i p t i c i t y ) as increasing amounts of BHC were added to serum albumin. They also investigated the effects of the binding of BHC to HSA by differential absorption. . Chignell and Stark-weather (1971) also used this method and studied the binding of BHC and PBZ to HSA. Their investigations and those of Perrin and Idsvoog (1971) resulted in qualitative data only. However, Rosen (1970), assumed that differences in the circular dichroism spectra (as HSA was titrated with PBZ) were proportional to the amount of bound PBZ and thus reported quantitative binding data. Generally, methods which measure free ligand concentration give more precise results. The most commonly used approach is equilibrium dialysis. A membrane is used which is permeable to a l l components of - 29 -the system, except the macromolecule. Thus, at equilibrium, the chemical potential of all the components except the macromolecule are equal and hence the free ligand concentration can be determined. The free ligand concentration can be measured in a variety of ways. O'Reilly (1971) measured free coumarin drug concentrations spectrophotometrically. Chignell (1969) used a similar approach for the analysis of PBZ. Aggeler, and others (1967), measured free warfarin levels using radioisotopic l t fC-warfarin. Problems associated with the use of equilibrium dialysis for binding studies are outlined by Steinhardt and Reynolds (1969). These include membrane binding and Donnan effects. Corrections can be made for the former; the Donnan effect is avoided by using higher ionic strengths. Reproducible results can, therefore, be obtained. However, lower protein concentrations are necessary to avoid volume changes; i t is a slow method and a considerable time is necessary for the diffusion of ligand mole-cules across the membrane to achieve equilibrium. Other methods are analogous in principle to equilibrium dialysis. A useful method and one which is amenable to automation and computer, calculation is the dynamic dialysis technique (Meyer and Guttman [1968]). The rate of disappearance of a small molecule from the dialysis cell i s proportional to the free drug concentration. This method is more rapid. Cho (1970) used the dynamic dialysis to study BHC - PVP interactions. In u l t r a f i l t r a t i o n methods, the free ligand is f i l t e r e d through the membrane by force, either by use of pressure or centrifugation. This is a rapid method but i t s accuracy is less than that obtained by other methods - 30 -because the macromolecule concentration increases as u l t r a f i l t r a t i o n proceeds. Solomon, and others (1968)- measured the displacement of warfarin - 1 4C by PBZ - 1 4C from HSA by u l t r a f i l t r a t i o n using cellophane bags and centrifugal force. McQueen and Warden (1971) studied the effect of PBZ on the distribution of sulphadoxine in plasma using Visking dialysis tubing and pressure (5% C0 2 in oxygen at 1080 m bar pressure). It was noted that in control experiments with large plasma volumes, even when as much as 60% of the i n i t i a l plasma volume had been removed, no progressive change in sulphonamide concentration in the u l t r a f i l t r a t e could be observed. However, no change in the sulphonamide would be detected at high plasma sulphonamide levels, since changes would be small. All the above methods, except dynamic di a l y s i s , yield only one data . point per experiment. Many experiments, therefore, are necessary to graph a binding curve. Membrane u l t r a f i l t r a t i o n is by no means a new method. It was f i r s t introduced by Martin in 1896 and Grollman (1926) reviewed the use of this method using collodion membranes. He reported agreement for,the determination of binding of phenol red to albumin by u l t r a f i l t r a t i o n with results achieved by dialysis. Thus i t can be seen that the macromolecular surroundings of protein solutions can be altered with permiselective membranes either by using conventional dial y s i s , u l t r a f i l t r a t i o n , or, f i n a l l y , simultaneous dilution and u l t r a f i l t r a t i o n . This latter case can also be called d i a f i l t r a t i o n . Diafiltration and i t s use in protein binding studies was f i r s t mentioned in the literature by Blatt, Robinson and Bixler (1968). They - 31 -described an adaption of an Ami con u l t r a f i l t r a t i o n cell which permitted the d i a f i l t r a t i o n of different volumes of macromolecule solution against ligand solutions of varying composition. There are a number of advantages in the use of this method:-Fi r s t , a range of biologically inert polymeric membranes with varying molecular weight cut-offs are. available.. Because of the range of membranes available, i t is possible to attain faster f i l t r a t i o n rates without loss of macromolecule. These membranes have a much higher solvent permeability relative to regenerated cellophane membranes (the most commonly used membranes in eq u i l i - .;. brium d i a l y s i s ) . Secondly, use of pressure with this Amicon cell allows for microsolute exchange in a protein solution to be achieved much more rapidly. Thirdly, the volumes of macromolecule solutions can be maintained at a constant level. A description of the Amicon u l t r a f i l t r a t i o n eel l i s given in the experi-mental section. Ligand solution of a fixed chosen concentration from a reservoir tank passes at a constant rate (determined by applied pressure) into the Amicon ce l l containing a volume of macromolecule solution. The f i l t r a t e passes through the f i l t e r at the same constant rate and thus the cell volume remains constant. Blatt, Bixler and Robinson (1968) show that there are several different ways in which this method can be used. (a) wash-in Experiments: Here the macromolecule solution is placed in the cel l and the ligand solution in the reservoir. The instantaneous ligand concentration in the f i l t r a t e can be predicted as - 32 -a function of cumulative f i l t r a t i o n volume. Where the molecule of ligand can pass the membrane unhindered, this experiment can be expressed mathematically:-C f v - v 1 In — = (Eq. 23) C f - C % where = concentration in reservoir C = cell u l t r a f i l t r a t e ligand concentration v - = average cell volume during the experiment v = cumulative f i l t r a t e volume • v 1 = apparent void volume of system. (b) Wash-out Experiments: These can be similarly performed with solvent (e.g., buffer solutions) in the reservoir and macromolecule and bound ligand in the c e l l . The wash-out experiment can be expressed mathematically:-C v - v 1 in -± = • (Eq. 24) C •> . where C Q = i n i t i a l Digandjin the c e l l . For binding studies, these authors showed two possible ways in which experiments could be done. (i) Direct Method: A wash-in experiment is performed and run until d i a f i l t r a t i o n has been completed, i.e., equilibrium is reached ( i . e . , [ 1 i g a n d ] r e s e r v o i r = [1igand f r e e ] c e 1 1 = [ l i g a n d ] u l t r a f i l t r a t e > -- 33 -Hence, for this experiment, a value for r can be obtained r = (Eq, 25) total concentration of ligand in cell concentration, in reservoir free ligand concentration in cell macromolecule concentration. Such a method has no particular advantage over equilibrium dialysis except that i t is run under pressure and hence equilibrium is reached more rapidly. Each experimental run would yield only one point on a binding curve. D i f f i c u l t i e s may also arise when i t is necessary to measure ligand in the presence of protein. (ii.) Diafiltration Equilibrium: In some instances, plots of log (Cf / C f - C) versus (v - v 1) are linear. In this case, a differential material balance within the cell would yield the relationship r = K BC f (Eq. 26) where Kg is determined from the slope, the value of which is 0.4343/(1 + KBM)vQ. Therefore, . Db = KgMC (Eq. 27) where C + C f - 34 -This indicates the molar binding ratio, D^ /M ( i . e . , r) is directly proportional to C. However, this is only an approxi-mation, being true only at low D f values. Again the .dia-f i l t r a t i o n technique used in the above manner, has few advantages over equilibrium dialysis and, indeed, is possibly more restricted in the range (of ligand concentration) in which i t can be used. Furthermore, i t yields only one data point for a binding curve from an experiment. Blatt, Robinson and Bixler (1968) gave some experimental data for the binding of methyl orange to HSA and the binding of C a 2 + to HSA., but most of their data was achieved by the direct method. Their results did not agree well with other literature values. The authors did derive, for wash-in experiments, an equation based on the Law of Mass Action, but they did v - vJ 1 + kMn (kD f +1)2 In res C - D, res ., f kMn kC + 1 res kD, L kDf + 1 +' ln(kD. + L) (Eq. 28) not apply this to any ligand-macromolecule interaction. This equation is not amenable to simple linear plotting and computer analysis of data would be necessary. The experimental section of this thesis will show how the Ami con c e l l , run under d i a f i l t r a t i o n conditions, can yi e l d an entire binding - 35 -curve from one experiment. At any instant in time, the concentration of ligand in the f i l t r a t e will indicate the concentration of free ligand in the c e l l . A necessary assumption in this case is that the ligand binds in a reversible equilibrium with the macromolecule in the cell instan- : taneously. The validity of this assumption is not well substantiated in the literature (for the equilibrium occurs too rapidly to study by con-ventional methods used for binding studies). However, Robbins and others (1965) found, by a method involving the quenching of the fluorescence of albumin by interaction with thyroxine, that association is complete within 150 msec. The dissociation process occurs in two steps with a half l i f e of 0.1 sec and 7 sees, respectively. Froese (1962), by a temperature jump technique, found the rate constants of association of albumin for two azo dyes were 0.36 x 10 6 and 2.1 x 10 6 moles"1 sec" 1. ; -Ryan and Hanna (1971) studied steroid HSA interactions by d i a f i l t r a -tion. Their experimental arrangements were similar to those used in this study. Crawford and others (1972) also used this method to study the binding of bromosulphthalein to serum andalbumin. However, these authors in contrast to Ryan and Hanna (1972), gave no data to substantiate the usefulness or disadvantages of the apparatus. III. EXPERIMENTAL Apparatus (a) Beckman Du Spectrometer (b) Beckman Du-2 Spectrometer (c) Bausch and Lomb Spectronic 505 Recording Spectrophotometer (d) Hitachi 124 Coleman Double Beam Spectrophotometer with Hitachi 165 Recorder (e) Perkin-Elmer Double Beam Spectrophotometer (Coleman 124) with Perkin-Elmer Recorder (Coleman 165) (f) Fisher Digital pH meter (g) Haake R 2i Thermoregulator (h) Isco Fraction C o l l e c t o r - Model 326 (i) Amicon Diafiltration Apparatus (j) International Centrifuge, Size 1, Model SBV. Chemicals and Reagents (a) Bishydroxycoumarin, U.S.P. The melting point of this chemical was 288-289°C. The drug was obtained from Abbott Laboratories Ltd., Montreal, Quebec. (b) Phenylbutazone, U.S.P. The melting point of this drug was 104.5°C. The drug was obtained from Geigy (Canada) Ltd., Montreal, , Quebec. . (c) Human Serum Albumin (Cohn Fraction V,). The albumin was obtained from Pentex Inc., Kankanee, 111. No loss of weight on drying was found. The substance was stored at 4°C. -37 -(d) Tris (hydroxymethyl) aminomethane (Tris). Fisher Reagent Grade. (e) Sephadex G-25 (fine), Pharmacia, Uppsala. (f) Blue Dextran '2000', Pharmacia, Uppsala. (g) Diethylaminoethyl (DEAE) cellulose, Baker Reagent grade. The capacity of the anion exchangerwas 0.7 mg./g. . (h) Copper sulphate; 5H20, Reagent Grade; (h) sodium carbonate, Reagent Grade; potassium dihydrogen tartrate, Reagent Grade, (i) Folin Phenol Reagent 2M. (j) Plasma, TA-2 Transfer Pack, Fenwal Labs., 111. This was kept frozen. (k) Nitrogen, "G" Grade. . (3) Preparation of Tris Buffer. Titrate 150.0 ml. of IN HC1 solution to approximately pH 7 with 0.5 molar Tris solution. Dilute to almost one l i t r e with d i s t i l l e d water and adjust to pH 7.4 with 0.5 molar Tris solution. Make up to one l i t r e with d i s t i l l e d water. The pH of this buffer at 37°C was 7.1. This buffer was used in a l l experiments unless otherwise stated and was prepared daily. The 0.5 molar Tris solution was stored at room temperature, for not more than one week. Its ionic strength i s 0.15 and i t s buffer capacity 0.021, capacity being calculated by a method g i v e n " by Bates (1961). He claimed that the useful range of this buffer was from pH 7 to pH 9. - 38 -Since Tris has intramolecular bonds, i t should be relatively inert (Benesch and Benesch [1955]). Chloride ions could bind to HSA but Scatchard and others (1950) showed this ion had a very low a f f i n i t y for the macromolecule. Hence binding results would not be significantly^ affected by a buffer containing chloride ions. Few other buffers are available for studies at pH 7.4. Phosphate buffer covers this range but Klotz and Urquart (1949) showed that phosphate ions interfere with the binding of small molecules to protein. Phosphate buffer has also been reported to interfere with the spectrophotometric analysis of BHC. (4) Determination of Absorptivity Value of Bishydroxycoumarin Accurately weigh 100 mg. BHC and make up to 100.0 ml. with 0.1N sodium hydroxide solution. Dilute 5.0 ml. of this solution to 100.0 ml. with Tris buffer. Dilute 8.0, 10.0, 12.0, 16.0, and 20.0 ml. aliquots of the Tris solution to 100.0 ml. with Tris buffer. This gives a series of solutions of concentration 4, 5, 6, 8 and 10 mg. BHC/1., respectively. Read the absorbance of these solutions at 310 mu. On the basis of 25 , determinations, the absorptivity value for BHC was 59.47 ± 0.54. Preliminary spectral characteristics of a 5 mg./l. and 10 mg./l. solution of BHC were determined on a Spectronic 505 recording spectro-photometer. See Figure 1. An absorbance maxima occurs at 310 mu and this wavelength was chosen for a l l subsequent analyses. 350 330 310 ; 290 270 250 230 Wavelength my Figure 1. Spectral Characteristics of Phenylbutazone^ - - - ) and Bishydroxycoumarin ( ). - 40 -(5) Determination of Absorptivity Value of Phenylbutazone Accurately weigh 100 mg. PBZ and make up to 100.0 ml. with 0.1 N sodium hydroxide solutions. Make dilutions as in the procedure for BHC in (4) above. This gives a series of solutions of concentration 4, 5, 6, 8 and 10 mg./l. Read the absorbance of these solutions at 264 my. On the basis of 20 determinations, the absorptivity value for PBZ was 67.58 ± 0.78. Preliminary spectral characteristics of a 5 mg./l. and 10 mg./l. solution of PBZ were determined on a Spectronic 505 Recording Spectro-photometer. See Figure 1. An absorbance maxima occurred at 264 my and this was the wavelength chosen for a l l subsequent analyses. (6) Analysis of Mixtures of Bishydroxycoumarin and Phenylbutazone by.the Absorbance Ratio Method (a) Determination of Q:310:264 for BHC Prepare a series of solutions containing 4, 5, 6, 8 and 10 mg. BHC/1. as in (4) above. Read absorbances of these solutions at 310 mp and 264 my. On the basis of 25 determinations the Q:310:264 value for BHC was determined to be 1.95 + 0.06. (b) Analysis of Mixtures of Phenylbutazone and Bishydroxycoumarin.. in Solution Weigh 100 mg. BHC and make up to 100.0 ml. with 0.1 P! NaOH. Similarly, weigh 100.0 mg. PBZ and make up to 100 ml. with 0.1 N NaOH. Dilute 5.0 ml. aliquots of each of these solutions to 100.0 ml. with Tris buffer. Dilute aliquots of these stock-solutions to 100.0 ml. with Tris buffer in the manner indicated below: - 41 -Solution ml. PBZ Solution ml. BHC Solution 1 8.0 22.0 2 8.0 22.0 3 10.0 12.0 4 10.0 12.0 5 10.0 8.0 6 10.0 8.0 7 16.0 4.0 8 16.0 4.0 20.0 \ 1 0 20.0 - • Read the absorbance of each of these solutions at 264 my and 310 my and calculate the amount of PBZ and BHC present in these mixtures (Pernarowski [1969]). From the spectral characteristics shown in Figure 1, 264 my and 310 my were found to be convenient wavelengths for the analysis of each drug in the mixture. At 264 my, both PBZ and BHC contribute to the absorbance whereas at 310 my only BHC contributes. Thus, since A 3 1 0 is measured and a 3 1 0 has already v been determined in this study, the concentration of BHC in the mixture is readily determined. Q B H C = a 3 i o / a 2 6 4 = A 3 i o / A 2 6 i t - This is a constant and is equal to 1.95. At 264 my, AT = APBZ + ABHC - 42 -where Ay = total absorbance and Apg^ = absorbance due to PBZ, and A B H C = absorbance due to B H C . Therefore, APBZ = AT " ABHC •. .=: A t - (BHC) QBHC Thus, the absorbance at 264 mp due to PBZ may be calculated by measuring absorbance of the mixture at 264 my and absorbance at 310 mp (due to BHC only). Results of the analysis of solutions containing mixtures of PBZ and BHC are given in Table 1. Agreement between calculated and expected concentrations is satisfactory. Preliminary objectives include compe-t i t i v e binding. This method was developed in anticipation of such studies but, because of d i f f i c u l t i e s in the d i a f i l t r a t i o n technique, was not used. (7) The Diafiltration Apparatus A diagram of the apparatus is shown in Figure 2. The 50 l i t r e glass waterbath was maintained at 37°C (unless other-wise stated) by a Haake pump and thermostat. The Amicon. reservoir tank (12 l i t r e capacity), with a maximum pressure capacity of 100.psi, was made of stainless steel (epoxy coated). It was fi t t e d with a f i l l port pressure r e l i e f valve, and inlet and output connectors. The Amicon u l t r a f i l t r a t i o n cell (Model 52, 50 ml. capacity, 43mm. diameter) is shown in Figure 3. The cel l was seated in a water-jacketed Table 1. Analysis of Bishydroxycoumarin and Phenylbutazone in Solution by the Absorbance Ratio Method Solution Expected Expected A 3 1 0 A 2 6 t + Calculated Calculated PBZ Cone. BHC Cone. of mixture of mixture PBZ Cone. BHC Cone. mg/1• mg/1. mg/1. mg/1. 1 3.95 12.0 0.720 0.640 3.99 12.06 2 3.95 12.0 0.723 0.635 3.91 12.09 3 5.93 6.0 0.363 0.595 5.99 6.07 4 5.93 6.0 0.362 0.594 5.99 6.06 5 5.93 4.0 0.241 0.530 5.97 4.05 6 5.93 4.0 0.243 0.530 5.96. 4.04 7 7.90 2.0 0.124 0.610 7.98 2.07 8 7.90 2.0 0.123 0.608 7.58 2.06 9 9.88 0.0 0.005 0.695 9.97 10 9.88 0.0 0.006 0.699 9.98 C/D Ami con Selector Inlet Figure 2. The Diafiltration Apparatus - 45 -beaker maintained at 37°C and the contents stirred by use of a magnetic st i r r e r . The cell is assembled and clamped together with a snap metal clamp. The cylindrical sleeve is transparent and graduated. A stir r i n g assembly, a pressure r e l i e f valve (and f i l l port), and inlet port and outlet port are built into the c e l l . 0- ring gaskets prevent,leakage. Materials used in the cell are Delrin, Teflon or other inert plastics. The sleeve is made of polycarbonate plastic. A polyethylene porous disc acts as a support for the membrane. Polyethylene tubing was provided with this c e l l but, for reasons given later, was replaced with Teflon tubing in a l l parts of the system. The membrane used in these studies was Diaflo PM-10 (43 mm diameter) as recommended by the manufacturers for binding studies with HSA. These membranes are made from biologically inert polymeric material and prevent passage of molecules of molecular weight greater than 10,000. This anisotropic diffusive membrane consists of an extremely thin (0.1 - 10y) layer of dense polymer supported on a much thicker structure of porous open cell structure. This membrane is resistant to most common chemicals and solvents. The u l t r a f i l t r a t e from the cell outlet port is passed via Teflon tubing to an Isco Fraction Collector (Model 326). An Amicon C/D (Concentration/Dialysis) Selector is connected as shown in the diagram. This allows gas or liquid to flow from the reservoir to the cell as desired and thus the solution in the cell will be concentrated or dialysed respectively. Operation of this apparatus is explained later. -46 -Figure 3. Amicon (Model.52) Diafiltration Cell - 47 -(8) General Procedure for the Determination of Drug - Human Serum Albumin Binding using the Diafiltration Technique (a) . Prepare an HSA solution of the desired concentration by dissolving the macromolecule in Tris buffer. Concentrations used are given in the latter part of this section. (b) Prepare solution of the drug in Tris buffer from a stock solution containing 100 mg. drug / 100.0 ml. of 0.1 N NaOH. Place this solution in the reservoir tank and allow to equilibrate at 37°C. (c) Prior to apparatus assembly, soak the Diaflo membrane in 100 ml. d i s t i l l e d water for 30 minutes. Discard the water and repeat this process with a further 100 ml. water. (d) Place a known volume of purified HSA solution (see section on purifications of HSA) in the d i a f i l t r a t i o n c e l l and allow to e q u i l i -brate at 37°C. (e) Connect the ce l l to the reservoir tank as shown in Figure 2. Turn the s t i r r e r on. The s t i r r i n g rate used was that suggested by the manufacturer, i.e., the vortex should not be more than one third the total depth of liquid in the c e l l . (f) i) For operation without the C/D selector: Close pressure r e l i e f valves on the reservoir tank and the c e l l . Open the pressure r e l i e f valve on the nitrogen cylinder until the desired pressure (on pressure gauge on the cylinder) is achieved - 48 -i i ) For operation with the C/D selector: Put the C/D selector in the 'GAS' position, thus permitting gas to pass from the reservoir tank to the c e l l . With the pressure r e l i e f valves on the tank and the cell both in the open position, open the pressure valve on the nitrogen cylinder slowly. Shut both pressure valves and open the cylinder pressure valve until the desired pressure is achieved on the cylinder gauge. Allow two minutes for pressure to equalize in the tank and the c e l l . Shift the C/D selector switch to the 'LIQUID' position. Very . slightly open the pressure r e l i e f on the cell to allow the solution to begin to flow into the tubing connected to the c e l l . If this valve is opened too much, solution flow will be exces-sive and a volume increase will occur in the c e l l . (g) Sample Collection: During preliminary experiments samples were collected manually. In later experiments samples were collected automatically by an Isco Fraction Collector. Samples can be collected on a time basis (i.e., x minutes per fraction). This approach was used in preliminary studies, but was discarded because flow rates of the f i l t r a t e were not constant. In subsequent experiments, samples were collected by volume ( i . e . , 10 ml. fractions). This procedure involved the use of a volume collecting device but checks on volumes delivered indicated that the device tended to be inaccurate. Hence, al l fractions collected were checked independently for volume. - 49 -(h) Record the volumes of successive fractions and their absorbance readings until equilibrium is reached or until approximately 500 ml. of f i l t r a t e has been collected. (This takes approximately 8 hours, depending on HSA concentration.) (i) Record the final volume in the c e l l . Determine the concentra-tion of drug in the reservoir tank. [Note: The above mentioned stock solutions of BHC and PBZ were stored in the dark at 4°C for not more than one week. There have been reports on the decomposition of PBZ solutions by oxidation and by hydrolysis (see Section II, 3). The extent of this decomposition occurring in solutions was investigated using methods given in a paper by Beckstead and Kaistha (1968). Solutions containing 100 mg. PBZ / 100 ml. of Tris buffer and 0.1 N NaOH, respectively, were pre-pared. These solutions were placed in a water-bath at 37°C in the presence of light. After 4 days, traces of n-(2-carboxy-2-hydroxy-caproyl-hydrazobenzene) and 4-hydroxyphenylbutazone were found by a thin layer chromatography technique. Qual i t a t i v e l y , the PBZ solution: in 0.T N NaOH contained more decomposition products. By the quanti-tative simultaneous equation method of Beckstead and Kaistha, after .4 days at 37°C, i t was shown that neither solution contained appreciable amounts of decomposition products. After 4 weeks at room temperature, in the presence of ligh t , a solution of PBZ in 0.1 N NaOH contained 9% of decomposition products. Thus, i t was concluded that in drug - 50 -binding studies where a d i a f i l t r a t i o n experiment is completed within 8 hours, decomposition of PBZ solutions would be insignificant. In addition, solutions stored at 4°C in the absence of lig h t , for not more than a week, would not decompose significantly.] (9) Phenylbutazone - Human Serum Albumin Binding Studies by the Desorpti (Washout) Diafiltration Technique Complete a PBZ-HSA binding experiment in the normal manner. Remove ligand solution from the reservoir tank and after rinsing, Tris buffer was placed in the reservoir tank. Rinse tubing to and from the ce l l and reconnect. Run the experiment in the same manner as for the wash-in dia-f i l t r a t i o n procedure. Analyse fractions of the f i l t r a t e spectrophoto-metrically at 264 my. (10) Bishydroxycoumarin - Human Serum Albumin Binding Studies by the ; Equilibrium or Direct (Diafiltration) Technique Place 30 ml. of an HSA solution in the d i a f i l t r a t i o n c e l l and solutions (500 ml.) of various BHC concentrations in the reservoir tank. Run the experiment at 25 psi in the same manner as for d i a f i l t r a t i o n . ; experiments. Analyse fractions of the f i l t r a t e until the A 3 1 0 value for the f i l t r a t e equals that of the reservoir solution. At this point, final equilibrium is considered to have been reached. Record the final volume in the cell and correct the HSA concentration for any volume increase. Determine the total BHC concentration in the cell by measuring the c e l l contents spectrophotometrically at 310 my. . - 51 -(11) Drug Binding Studies by the Centrifugation (Ultrafiltration) Method These studies were carried out using Amicon Centriflo membrane cones (CF 50), together with Centriflo supports and tubes (see Figure 4) and an International Centrifuge (Model SBV, Size 1). In principle this is similar to membrane u l t r a f i l t r a t i o n , with centrifugal force instead of. pressure being used to push the drug through the membrane. Since the forces involved are greater, the f i l t r a t e is collected rapidly. . Centriflo CF 50 membrane cone f i l t e r Centriflo membrane cone support HSA + drug Centrifuge tube Drug .• Figure 4. The Centriflo Centrifugation (Ultrafiltration) Apparatus i) Calibration of the Speed of the Centrifuge to Relative Centri-fugal Force (r.c.f.) The relationship between r.c.f. and r.p.m. is as follows: r.c.f. ••= 0.00001118 x r x S 2 where r = the rotating radius in cm. and S = the rotating speed in - 52 -r.p.m. The r.c.f. value for the Centriflo apparatus at 1100 r.p.m. was found to be 108.9 x G (r = 9 cm. and S = 1100 r.p.m.). ' i i ) Binding Studies ' . Soak the membrane cones for two hours in d i s t i l l e d water. Spin the empty cones in the centrifuge at 2000 r.p.m. for 10 minutes to remove residual water. Failure to remove residual water would result in an inaccurate determination of drug in the u l t r a f i l t r a t e . , The maximum amount of solution which could be placed in the cones was 3.5 ml. Studies indicated that several washings of the Centriflo apparatus were required to remove a foreign substance which absorbed ultra-violet energy at 264 my, the analytical wavelength for PBZ. Experiments with 'purified' HSA solutions (1.2 x 10~3 M) at 500, 1100, and 1500 r.p.m. indicated that substantial amounts of a protein-like substance passed through the membrane cone. The ultra-violet spectra of the u l t r a f i l t r a t e showed that this substance was the same as that found during the d i a f i l t r a t i o n process. (12) Phenylbutazone - Human Serum Albumin Binding Studies by a Molecular Sieve Technique using Sephadex G-25 This technique involves the molecular sieving effect which occurs when swollen gel particles of certain polymers come into contact with a mixture of molecularspecies dissolved in the swelling medium. The solute molecules partition between the solvent external to the gel particles and the solvent imbibed by the gel, as described by the distribution coefficient - 53 -(K D) which is equal to. C^/C , where and C Q are the solute concentrations in the internal and external compartments, respectively. The pore size of the gel should be such that for protein molecules, = 0, and for 1igand molecules, = 1. Thus protein and protein-ligand complexes can be separated from free ligand. Wood and Cooper (1970) reviewed the ways in which gel f i l t r a t i o n can be applied to the protein binding studies. The simplest approach is to use the Batch method. A known volume of sample solution of known protein and 1igand concentrations is added to a known weight of gel. The mixture is allowed to equilibrate and the gel and external phases are separated by centrifugation or f i l t r a t i o n . Protein and total ligand in the solution are determined. This method has been frequently used in HSA and plasma binding studies and its applications have been reviewed by Wood and Cooper. Scholtan (1965), however, was the only author to compare his results (on the binding of sulphonamides to serum albumin) with another method. Chromatographic gel f i l t r a t i o n techniques can also be used in protein-drug binding studies. The most commonly used chromatographic technique is zonal separation, but results from this procedure are often d i f f i c u l t to interpret quantitatively when binding is reversible. A technique avoiding this problem of dissociation of the protein-ligand complex was devised by Hummel and Dreyer (1962). They equilibrated the column with the same concentration of ligand solution as was present in the protein-ligand mixture applied to the column. Frontal analysis has been used i more recently and appears to give results comparable with those obtained - 54 -by other methods (Cooper and Wood [1968], Burke [1969]). The general validity of this method remains to be demonstrated but i t is expected to be applicable to the binding of a wide range of ligands to HSA and plasma. In these studies, PBZ interactions with HSA are studied by the Batch method. Sephadex G-25 (fine grade) has a nominal water regain of 2.5 g. water per g. dry gel. To prepare the swollen gel, add 4.0 g. of dry gel to 15.0 ml. Tris buffer so that the total volume of liquid in the gel is 25.0 ml. Equ i l i -brate the system by shaking for one hour at 37°C. Remove a sample from the external phase and centrifuge the sample to precipitate any gel particles. Analyse the supernatant f l u i d . Determination of the external volume of the gel: Sephadex G-25 has a nominal v/ater uptake of 2.5 g. ± 0.2 g. per g. dry gel but this value does not give the external volume accurately. Using Blue Dextran, the external volume can be determined with greater precision. Blue Dextran 2000 has an average molecular weight of 2,000,000 and remains in the external phase. Prepare a 0.4% w/v solution of Blue Dextran 2000 in Tris buffer. From this solution, make a series of solutions by dilution and record their A 6 2o values. The resultant calibration curve is shown in Figure 5. Add 5.0 ml. of an 0.2% Blue Dextran solution to the swollen gel. Add sufficient Tris buffer to make a total volume of 25.0 ml. of solution added to the gel. Equilibrate the gel for one hour. Remove a sample from the external phase and analyse at 620 my for Blue Dextran. The Blue Dextran concentration could be found from the calibration curve. - 55 -Figure 5. Calibration Curve for Dextran Blue 2000 - 56 -The mean external phase volume (on a basis of six determinations) was found to be 15.32 ml. Protein Adsorption to the Gel: Prepare a 3.16% w/v HSA solution. Dilute this solution 1 in 2.5, 1 in 5, 1 in 10, 1 in 20 and 1 in 40, ••., respectively. Add 5.0 ml. of each of these HSA solutions to swollen gel samples. Make the total volume of solution added to the gel up to 25.0 ml. with Tris buffer. Equilibrate the gel for one hour at 37°C, with shaking. Remove the entire external phase with gentle suction in a Buchner funnel. Wash the gel well with Tris buffer and analyse the washings for HSA at 280 my. ; - \ " ! The results showed no HSA was adsorbed onto the gel. Adsorption of Phenylbutazone onto the Gel: Add 5.0 ml. of a range of PBZ solutions (12.81, 20.04, 29.99, 60.70, and 155.30 mg./l.) to the swollen gel. Make the total volume of liquid added to the gel up to 25.0 ml. with Tris buffer. Equilibrate the gel for one hour at 37°C, with shaking. Allow the gel to settle and remove a sample from the external phase. Centrifuge the sample. Analyse the supernatant at 264 my for PBZ. The concentration of PBZ in the external phase is known. Thus, since the total amount of drug i n i t i a l l y added to the gel is known, the amount of PBZ associated with one g. of gel can be calculated. (The drug associated with the gel is the sum of adsorbed drug and of drug in solution in the internal phase. It is unnecessary to distinguish between these two effects or to know the internal volume.) . The results are - 57 -plotted in Figure 6. It can be seen, however, that PBZ is not linearly adsorbed onto the gel. Linearity is preferable for easy determination of binding parameters. Thus calculations of binding data, on the basis of data in Figure 6, may be inaccurate. PBZ-HSA binding using the Batch method: Add 5.0 ml. of an HSA solution and 5.0 ml. of a PBZ solution to the swollen gel. A range of HSA solutions and PBZ solutions were added. Equilibrate the gel for one hour at 37°C, with shaking. Sample and analyse as above for PBZ. Binding results are given in the Results and Discussion Section. 0.10 5 10 15 20 25 Concentration in external phase (mg. PBZ / 1.) Figure 6. Phenylbutazone Adsorption to Sephadex G-25 IV. RESULTS & DISCUSSION Preliminary Check on Diafiltration Apparatus A) Release of Foreign Substances from Apparatus D i s t i l l e d water was run through the apparatus and samples of the f i l t r a t e were collected every hour for three hours. Spectra of these samples were run on. the Bausch and Lomb 505 Recording Spectro-photometer. These spectra indicated that the apparatus released a substance with maximum absorbance at 241 my. To determine the source of this substance, the various parts of the apparatus were soaked overnight in d i s t i l l e d water. Spectra of these solutions were run on the Perkin-Elmer Recording Spectrophotometer. The yellow latex tubing (used for collecting the f i l t r a t e ) was found to be the source of this substance. The polyethylene tubing used in other parts of the apparatus, on the other hand, yielded no foreign substance. Hence the latex tubing was replaced by polyethylene tubing. B) Binding of Phenylbutazone and Bishydroxycoumarin to the Apparatu Components Various concentrations of solutions of PBZ and BHC, as indicated in Tables 2 and 3, respectively, were prepared. Twenty-five ml. of each solution was placed in the Amicon ce l l and 500 ml. of the same solution was added to the reservoir tank. The experiment was run for each concentration of drug, for 75 minutes. After this period of time, the concentration of drug in the f i l t r a t e was determined spectrophotometrically. In a l l experiments there was a lower con-centration in the f i l t r a t e than in the original solution in the. c e l l . Thus, the amount of drug bound can be calculated. Percent drug bound at various concentration levels for PBZ and for BHC are given in Tables 2 and 3, respectively. The results given in these tables indicate that for PBZ only a small percentage of the drug is •; • bound and that this percentage decreases as PBZ concentration increases. For BHC, on the other hand, the percent binding is higher and more variable. Since BHC binding values "were high, further investigations on the binding of BHC to the apparatus were carried out. . " \ C) Binding of Bishydroxycoumarin to the Diaf i l t r a t i o n Cell Two d i a f i l t r a t i o n cells with membrane supports in place, were f i l l e d with 50.0 ml. of a 5.60 x 10~5 M BHC solution, and the openings, were sealed. The cells were rotated on a tumbler at 30 r.p.m. in a water-bath at 37°C. The cell contents were sampled at 2, 4, 8 and 24 hours and the BHC concentration determined spectrophotometrically. A decrease in the BHC concentration was noted with respect to time. The loss of BHC from the solution was expressed as percent BHC bound. The results are given in Table 4. Up to 8 hours, binding, though significant, is s t i l l with a reasonable range. At 24 hours, however, the binding of BHC to the cell more than doubled, but since experi-- 61 -Table 2. Binding of Phenylbutazone to Apparatus after 75 minutes of Diafiltration Concentration of PBZ in Reservoir Tank moles/1. % Drug Bound to Apparatus * 3.319 x 10"5 4.30 % 6.505 x 10~5 3.13 % 1.191 x 10_tt 1.76 % 1.869 x 10-1+ 1.63 % 3.070 x 10_t* 1.53 % . 4.310 x 10"4 1.27 % 6.040 x 10_1+ 1.34 % Each value is the average of two determinations Table 3. Binding of Bishydroxycoumarin to Apparatus after 75 minutes of Diafiltration Concentration of BHC in Reservoir Tank moles/1. % BHC Bound at this Con-centration of BHC 5.068 x 10"6 10.79 % 1.111 X 10~5 6.04 % 2.781 x.l0"5 10.22 % 2.405 x 10"5 10.49 % 4.120 x 10"5 8.37 % 5.070 x 10"5 11.19 % - 62 -ments were not expected to run more than 8 hours, this is not a significant factor in studies of this type. D) Binding of Bishydroxycoumarin to Polyethylene Tubing A 40 cm. length of polyethylene tubing was placed in a glass flask containing 100.0 ml. of 5.60 x 10"5 M BHC solution. The flask was sealed and rotated in the water bath as in C) above. The flask contents were analysed spectrophotometrically at 2, 4, 8 and 24 hours. Percentages of BHC bound to polyethylene tubing with respect to time are shown in Table 5. These results show that binding of BHC to the tubing is very high even after only 8 hours of exposure to the drug. Therefore, other types of tubing had to be investigated. E) Binding of Phenylbutazone and Bishydroxycoumarin to Teflon and Skimatco Tubing Teflon and Skimatco tubing were exposed to PBZ and BHC as in D) above. After 12 hours at 37°C, the flask contents were analysed spectrophotometrically. Flasks, containing drug solutions only were used as controls. The results in these experiments showed:-(a) Skimatco tubing released a substance which absorbed ultra-violet radiation significantly at 264 mu, but not at 310 mu. This, therefore, would not.be a satisfactory tubing in that i t would interfere with PBZ analysis. - 63 Table 4. Binding of Bishydroxycoumarin to the Diafiltration Cell Time in hours % BHC Bound to Cell Components * 2 5.45 4 5.66 .. ' 8 6.42 •-24 14.02 * Results given are the. average of results from the two d i a f i l t r a t i o n c e l l s . Init i a l BHC concentration in c e l l s was 5.60 x 10~5 moles/1. Table 5. Binding of Bishydroxycoumarin to Polyethylene Tubing Time in hours % BHC Bound to Polyethylene Tubing 2 2.55 4 5.19 8 19.95 24 •;• ' 37.35 Results given are the average obtained from two : 40 cm. sections of tubing ( V diam.) Initial BHC concentration in the flasks was 5.60 x 10"5 moles/1 - 64 -(b) Teflon tubing, on the other hand, did not bind either BHC or PBZ and did not release any UV absorbing substances. As"a result of these investigations, Teflon tubing appears to be the tubing of choice for the d i a f i l t r a t i o n apparatus when PBZ and BHC are being used in studies. F) Binding of Bishydroxycoumarin and Phenylbutazone to Components of the Apparatus (Connected with Teflon Tubing) Binding studies were now repeated as in 1 '"(B) above, except that the experiments were continued for two hours. Approximately. 450 ml. f i l t r a t e were collected. The f i l t r a t e , in this series of experiments, was sampled as indicated in Tables 6 and 7. PBZ and BHC concentrations are also given in these tables. These results show that with both BHC and PBZ, binding to components of the apparatus occurs. The percent drug bound decreases as f i l t r a t i o n volume increases, and also as drug reservoir concentration increases. Because of binding v a r i a b i l i t y , corrections are d i f f i c u l t to make during a d i a f i l t r a t i o n experiment. At best, an average value would have to be used. However, such values may be in error, because results obtained in these experiments may not be due to binding but rather to membrane rejection. (See IV, ( 1), G). It should be noted : that these binding experiments were continued until a volume of f i l t r a t e similar to that in drug - HSA binding studies had been collected. Drug - HSA binding experiments, however, took about 8 hours for complete d i a f i l t r a t i o n . - 65 -G) Membrane Binding and Rejection: Diafiltration of Phenylbutazone and Bishydroxycoumarin in the Absence of Human Serum Albumin An alternative method (to those mentioned in IV (1), F), can be used to study binding and/or rejection. This is based on a ' comparison of theoretical and experimental dilution curves in the absence of HSA. The following experiments were carried out: For PBZ, a 8.77 x 10"s M solution of PBZ was placed in the reservoir tank and diafiltered through 30.0 ml. of Tris buffer in the d i a f i l t r a t i o n c e l l . Twelve ml. fractions were collected and analysed spectrophotometrically until d i a f i l t r a t i o n was essentially complete. For BHC, a 2.046 x 10 _ l t M BHC solution was similarly diafiltered through 25.0 ml. Tris buffer. Ten ml. fractions were collected and analysed spectrophotometrically until diafiltration.was essentially complete. The experimental dilution curves are shown in Figures 7 and 8. Theoretically dilution curves can be plotted and are shown in these figures. A simple exponential expression can relate the f i l t r a t e concentration to the reservoir tank concentration (Ryan and Hanna [1971]): [s n] = [s o] [ i - (i - « )n] : where [S ] and [S ] are the concentrations in moles/1, of the f i l t r a t e and reservoir solutions, respectively, n is the number of fractions collected and <* is the ratio of the volume of the . fraction to cell volume. As shown in Figures Zand 8, experimental dilution curves for both BHC and PBZ di f f e r from theoretical ; Table 6. Binding of Phenylbutazone to Components of the Apparatus .. (Connected with Teflon Tubing) Volume of Filtrate % Binding of PBZ During Experiment Reservoir Concentration 3.40 x 10"5 moles/1. Reservoir Concentration 6.91 x 10"5 moles/1. Reservoir •''. Concentration 1.15 x 10"4 moles/1. Reservoir Concentration 4.18 x 10~k moles/1. 60 ml. : 7.43 % 7.55 % - 5.93 % • • • • • • . - -120 ml. 6.66 % 6.56 % 4.64 % 180 ml. 6.09 % 6.56 % 3.64 % 3.42 % 300 ml. 5.44 % 5.96 % 450 ml. 5.44 % 4.55 % 3.64 % 3.87 % Table 7. Binding of Bishydroxycoumarin to Components of the Apparatus (Connected with Teflon Tubing) Volume of Filtrate % Binding of BHC During Experiment Reservoir Concentration 1.49 x 10"5 moles/1 Reservoir Concentration 3.08 x 10'5 moles/1. Reservoir Concentration 1.78 x lO"4 moles/1. Reservoir Concentration 4.18 x 10"', moles/1. 60 ml. 7.75 % . _ _ — • '• 120 ml. — 8.37 % — 4.09 % 180 ml. 7.82 % 6.50 % 3.40 % 300 ml. 6.35 % 5.99 % 6.00 % 1.60 % 450 ml. 5.37 % 5.30 % 6.00 % - 67 -20 4 0 60 80 Filtrate Volume (ml.) 100 120 (Reservoir Cone - 8,77 x 10~s M P Bz) - 68 -Moles BHC x l O ' V l . i n f i l t r a t e . - 69 -dilution curves. This difference is more marked for BHC. This, indicates that, with both BHC and PBZ, binding and/or rejection of the drug must be occurring in the apparatus. Blatt, Robinson and Bixler (1968) showed that plots of log C^/(C^ - C) versus v (seeEq. 23), can give more useful information than the dilution curves shown in Figures 7 and 8. Data from these experimental dilution curves is plotted in the manner indicated by Blatt, Robinson and Bixler in Figures 9 and 10. As predicted, a linear relationship exists (in the absence of HSA). The slope of these plots gives 0.4343 / Vq and thus the apparent cell volume ,v . can be calculated. For phenylbutazone, in the absence of HSA, (Figure 9) the apparent cell volume is 34.6 ml., which is larger than the experimentally determined value of 30.0 ml. For bishydroxy-coumarin, in the absence of HSA, (Figure 10) the apparent cell volume is 33.1 ml., also larger than the experimentally determined . value of 25.0 ml. Blatt, Robinson and Bixler suggest such apparent increases in ce l l volume arise from solute rejection. However, neither the experimental dilution curves, nor the plotting method indicated by Blatt, Robinson and Bixler, enable a distinction to be made between membrane (or apparatus) binding or rejection. In both cases, the concentration of-ligand',in the f i l t r a t e at any point during d i a f i l t r a t i o n is lower than that theoretically calculated. In the case of membrane rejection, how-ever, the concentration of ligand in the ce l l is higher than that in the f i l t r a t e at any instant of time. A distinction could be - 70 -Filtrate Volume (ml.) Figure 9. Plotting of Phenylbutazone - Human Serum Albumin Binding Results in the Manner of Blatt, Robinson and Bixler (1968). (o & ) = 9.77 x 10~5 M PBZ reservoir concentration in the absence of HSA, ( A A ) = 9.19 x 10~5 M PBZ : reservoir concentration, where HSA concentration = ' 1.385 x 10_1+ M, •(••') = 9.09 x 10~5 M PBZ reservoir concentration where HSA concentration = 4.772 x 10_lt - 71 -Filtrate Volume (ml.) Figure 10. Plotting of Bishydroxycoumarin - Human Serum Albumin Binding Results in the manner of Blatt, Robinson and Bixler (1968). -( o ® ) = 2.046 x 10 h M BHC reservoir concentration in the absence of HSA, ( A A ) = 2.876 x 10".5 M BHC reservoir concentration where HSA concentration = 0.420 x 10" 4 M, ( a a ) = 2.885 x 10~4 M BHC reservoir concentration and HSA concentration where HSA concen-tration = 6.129 x l O - 1 * M. HSA. - 72 -made, however, i f cell and f i l t r a t e were sampled simultaneously. Ryan and Hanna (1971), thus reported UM-TO membrane rejection of testosterone solutions. Testosterone concentrations in the f i l t r a t e rose steadily towards the reservoir concentration as d i a f i l t r a t i o n proceeded, but simultaneous sampling of eel 1 contents showed that cell testosterone levels rose above the reservoir concentration. Such experiments were not carried out in this work and therefore i t can only be concluded that membrane binding and/or rejection is occurring. „• Blatt, Robinson and Bixler (1968) also suggested that changing the membrane area exposed to the ligand solutions, whilst keeping the sample volume constant could be another method of distinguish-ing between membrane binding or rejection. These authors defined a reflection coefficient (a): o = 1 - [C/C 1] where C 1 is solute concentration in the c e l l . For a small molecule, ideally a should be one. Equations 23 and 24 can be modified, by multiplying the right hand side by a.. Thus, these experimental dilution curves could be corrected for membrane rejection. However, whether membrane rejection occurs in the same .. manner in the presence of macromolecule is not known and could not easily be resolved. Saturation binding to the membrane could also take place.(Blatt, Robinson and Bixler [1968]). When the ligand has a very high a f f i n i t y for the membrane, the membrane will become saturated at low ligand f i l t r a t e concentrations. This would give an apparent increase in •, :-the void volume of the system, but once the membrane has become - 73 -saturated by ligand, Equations 23 and 24 will be obeyed by the d i a f i l t r a t i o n process. This discussion, based on current literature, is not likely to significantly affect the binding results in these studies. In a later section [ 2., F, ( i i ) ] , i t will be shown that the percent binding found has l i t t l e effect on the final binding curve. H) Membrane Retention of Human Serum Albumin: Diafiltration of Human Serum Albumin in the Absence of Drug This experiment checks membrane retentivity of the macro-molecule. Twenty-five ml. of 2.9 x 10~6 M HSA were placed, in the di a f i l t r a t i o n cell and Tris buffer in the reservoir tank. The experiment was run at 25 psi. The f i l t r a t e was collected for one hour and i t s UV spectrum determined on a Perkin-Elmer recording spectrophotometer. This spectrum i s shown in Figure 8 and showed that a substance, with a maximum absorbance at 280 mp, was passing through the membrane. As shown also in this figure, this substance did not have the same spectrum as an HSA solution. i) Effect of Time on the Appearance of Unknown Substance in the Filt r a t e Tris buffer was run through the d i a f i l t r a t i o n c e l l , con-taining 25.0 ml. of a 5.8 x ~\0~h M HSA solution at 25 psi. Spectra of samples of the f i l t r a t e collected at 1,2, 3 and 4 hours were recorded on the Perkin-Elmer recording spectro-photometer. The spectra obtained were identical to that shown - 74 -in Figure 11. The Ao f i n , c ., , f , . u , 3 2 8 0 values of these samples of f i l t r a t e collected were plotted against time (Figure 12). L i t t l e of the substance was present in the f i l t r a t e after three hours. i i ) The Effect of the Unknown Substance in the F i l t r a t e on Phenylbutazone and Bishydroxycoumarin Analysis Tris buffer was run through 30.0 ml. of a 1.45 x 10 - 1 + M HSA solution in the d i a f i l t r a t i o n c e l l for 3 hours at 35 psi. The f i l t r a t e was collected in 1 hour fractions. Solutions containing 87.35 mg. BHC / 100 ml. 0.1 N NaOH and 85.6 mg. PBZ / 100 ml. 0.1 N NaOH, respectively, were prepared. These solutions were then diluted 1 in 100 with each hourly fraction, of f i l t r a t e . Control solutions were prepared by diluting the BHC and PBZ solutions 1 in 100 with Tris buffer. Concentrations of BHC and PBZ were determined spectrophotometrically. Results are shown in Table 8. This series of experiments was carried out to determine how much effect the unknown substance would have on the analysis of BHC and PBZ iri the f i l t r a t e . Results . show that the unknown substance affects the analysis of PBZ but does not affect the BHC analysis. At low BHC concentra-tions, however, which do occur at the beginning of a binding experiment, the substance could affect BHC analysis in the f i l t r a t e . It is necessary, therefore, to remove the unknown "impurity", prior to a binding experiment. A. 0.2 0.1 --75 240 250 260 270 Wavelength, my 280 290 300 Figure 11. Ultraviolet absorption spectra for a 3.49 x 10~6 M HSA solution ( ) and for the Unknown Substance in the Filtrate ( — — ) . 3.0 A 2 8 0 2.0 of f i l t r a t e 1.0 ., Sampling Time (in hours) Figure 12. Binding Experiment in the Absence of Drug: Appearance of Unknown Substance in Fi l t r a t e During Experiment. HSA concentration = 5.8 x 10"4:moles/1. Table 8. The Effect of the Unknown Substance in the Filtrate on the Analysis of Bishydroxycoumarin and Phenylbutazone Solution Diluent ^26^ ^310 Drug Concentration in moles/1. PBZ Control 0.578 0.002 2.78 x TO"5 PBZ 1 hr. f i l t r a t e 0.625 • 0.022 3.00 x 10"5 PBZ 2 hr. f i l t r a t e 0.608 0.012 2.92 x 10"5 PBZ 3 hr. f i l t r a t e 0.580 0.003 2.78 x 10"5 BHC Control 0.275 0.519 2.59 x 10"5 BHC 1 hr. f i l t r a t e 0.310 ; 0.519 2.59 x 10"5 . BHC 2 hr. f i l t r a t e 0.294 0.516 2.58 x 10" 5 .;•>.' BHC 3 nr. f i l t r a t e 0.280 - 0.519 2.59 x 10" 5 - 77 -i i i ) Purification of the Human Serum Albumin (Fraction V) (a) Purification by Sephadex G-25: A Sephadex G-25 column (1 cm. by 30 cm.) was pre-pared and 5 ml. of a 5.9 x 10~4 M HSA solution was applied to the column. The HSA solution was collected from the column and transferred to a d i a f i l t r a t i o n c e l l . Tris buffer was then run through the cell and the f i l t r a t e collected for one hour. A spectrum of the f i l t r a t e obtained on the Perkin-Elmer recording spectrophotometer showed that the same unknown substance was present in the f i l t r a t e . Sephadex G-25 does not, therefore, purify HSA solutions. (b) Purification by Dialysis with a Cellophane Membrane: Dialysis cells were set up as described by Cho (1970), one set having Visking cellophane as a membrane and another set having a Diaflo PM-10 membrane. In one side of each set of c e l l s , 10.0 ml. of 1.45 x 10~ 4 M HSA solution was placed and, on. the other side, 10.0 ml. of Tris buffer. The cells were then rotated at 30 r.p.m. in a 37°C water bath for 24 hours. Spectra of the contents of the compart-ment originally containing Tris buffer were then obtained on the Perkin-Elmer recording spectrophotometer. These showed that the unknown substance could pass through the Diaflo PM-10 f i l t e r but did not pass through the cellophane membrane. Hence, dialysis with cellophane membranes (either in dialysis cells or in a dialysis sac) could not be used to purify the HSA. (c) Purification by DEAE Cellulose: The column was prepared in the following manner. The exchanger (1 g. dry weight / 15 ml. HC1) was poured into 0.5 N HC1 with gentle s t i r r i n g and allowed to stand for 30 minutes. The supernatant f l u i d was decanted and the exchange gently washed in a Buchner funnel with d i s t i l l e d water until the pH of the effluent was approximately 4. The exchanger was then treated with 0.5 N NaOH as indicated above. It was washed with d i s t i l l e d water until the pH of the effluent was approximately 7. Equilibration was attained by pouring the exchanger into 0.5 M NaCl and titr a t i n g slowly to pH 7.4 with 0.5 M HC1. The supernatant f l u i d was removed and replaced several times with Tris buffer to ensure that the pH remained at 7.4. The exchanger was poured into Tris buffer to make a 5-10% w/v suspension and then poured into the column (1 cm. x 25 cm.). After use of this column, i t was washed with Tris buffer and re-equilibrated. Ten ml. of a 5.9 x 10~k M HSA solution were then applied to the column. The fraction containing HSA (as determined by A 2so values) was then collected and placed in the d i a f i l t r a t i o n c e l l . Tris buffer was then - 79 -run through the cell and the A 2 8 0 values of the f i l t r a t e measured as d i a f i l t r a t i o n proceeded. Ten ml. of untreated HSA solution (5.9 x 1C1"4 M) required a d i a f i l t r a t i o n volume of approximately 250 ml. Tris buffer to produce a zero A 28o value. The treated HSA solution only required about 75 ml. Therefore, the DEAE column partially purified the HSA solution and purification was completed by d i a f i l t r a t i o n of the treated HSA with Tris buffer. If this method of purifica-tion is used, some loss of the original HSA in the solution would be expected. Therefore a simple method for deter-mining the final HSA concentration was necessary. The absorbance of the purified HSA solution was measured at 280 my and concentration of the HSA was then read off a calibration curve prepared from standard solutions con-taining the macromolecule. Calibration Curve for Measuring Human Serum Albumin Con- . centrations: With proteins containing tyrosine and tryptophan, measurement of absorbance at 280 mu, (the UV absorbance maxima) can be used as a quantitative measure of protein concentration (Schultze and Heremans [1966]). Since the spectrum of HSA has an absorbance maxima at 280 my, this approach can be used to analyse solutions containing HSA (or protein, in general). -81 -the f i l t e r was not established. Several general observations could be made about i t s nature from the above experiments, (a) The ultraviolet spectrum of the substance in the f i l t r a t e was that characteristic of a protein containing tryptophan or tyrosine. The substance, however, was not albumin. (See Figure 11) - (b) Since the albumin solution could not be purified by use of the Sephadex G-25 (which separates molecules of molecular weight 5000 or less), this suggests the molecular weight of the substance must be over 5000. If the PM-10 membrane does indeed cut off at a molecular weight of 10,000;, as the manufacturers claim, the substance should have a molecular weight between 5000 and 10,000. However, i t is li k e l y that molecular shape determines the molecular weight cut-off level. (c) Two other tests for protein were carried out on the unknown substance to determine i f they were more sensitive than the A 2so spectrophotometric method:-The Biuret Test: This is based on the formation of a reddish-violet complex* copper ions from a dilute solution of CuSO^ complex with amide linkages. This test was unsatisfactory because Tris buffer interfered to form a dee-blue coloured cupri-ammonium complex with copper ion. The Folin Ciocalteau Test: This test is not as reproducible or specific as the Biuret Test. Any peptide The calibration curve was prepared by measuring the A 2 8 0 °f a series of standard solutions containing 10, 15, 25, 40 and 100 mg. of Fraction V HSA per 100 ml. of Tris buffer. The calibration curve is shown in Figure 13. Beer's Law is obeyed. Percent of Recovery of Human Serum Albumin Purified by This Method: Calculations showed that approximately 33% of the HSA was lost, either on the column or in the d i a f i l t r a t i o n procedure when 10 ml. of 5.9 x 10 _ l f M HSA solution was purified by this method. Because of these losses the procedure is not satisfactory. (d) Purification Using the Diafiltration C e l l : The HSA solution was placed in the d i a f i l t r a t i o n cell and Tris buffer diafiltered through the cell until the A 28o value for the f i l t r a t e was zero. The HSA concentration in the cell was then determined by use of the calibration curve shown in Figure 13. For a 5.9 x 10 _ l f M HSA solution, this d i a f i l t r a t i o n procedure took approximately 250 ml. Tris buffer and four hours at 25 psi. This method was sub-sequently used for purifying HSA prior to use in a l l binding experiments. Nature of the Unknown Substance The chemical structure of the substance passing through: bond will yield some colour. The procedure used was that described by Lowryand others (1951). Reagent A: Sodium Carbonate, 5 gm./250 ml. 0.1 N .NaOH Reagent B: Aqueous Cupric Sulphate. 5H20 Solution, 0.5% w/v ... Aqueous Potassium Dihydrogen Tartrate Solution, 1% w/v Reagent C: Mix 1 ml. of Reagent B with 50 ml. Reagent A (Discard after 1 day) Reagent E: Folin Phenol Reagent (2N), diluted 1 in 2. To 1.0 ml. of standard HSA solution or of fi l t r a t e , add 5 ml. Reagent C, mix well and allow to stand 10 minutes or longer, at room temperature. Add rapidly 0.5 ml. Reagent E and mix immediately on a vortex mixer. Allow to stand 30 minutes or longer at room temperature. Read the absorbance of the solutions at 500 mu. A standard curve was prepared using HSA solutions of known concentrations; 0.314, 0.0314, and 0.0031% w/v. A Tris buffer control was also prepared and used as a blank. The standard curve is shown in Figure 14. This test was not as sensitive or reproducible as that which depended on absorbance measurements at 280 mu. v) Purification of Fraction V Human Serum Albumin and Plasma by the Diafiltration Method Both a 25.0 ml. plasma sample (undiluted) and 25.0 ml. of 5.9 x 10-lt M HSA solution were diafiltered at 25 psi in the dia-filtration cell with Tris buffer until the solutions were - 84 -Figure 14. Calibration Curve for Protein Determination by the Folm Ciocalteau Method. - 85 -essentially purified. The f i l t r a t e was collected in 20 ml. fractions on the Isco Fraction Collector. The spectrum of each fraction was recorded on the Hitachi Recording Spectro-photometer. A 28o values were measured and plotted against d i a f i l t r a t i o n volume in Figure 15. This figure shows that con-siderably more protein was released from the Fraction V HSA than from the plasma. Also spectra of fractions of f i l t r a t e from plasma were not identical with those from Fraction V HSA. It would be expected that plasma would release more protein-like material than Fraction V HSA, since there are proteins of lower molecular weight than HSA in plasma.. Thus i t would seem l i k e l y that the protein-like substance released from Fraction V HSA is some impurity present in the HSA powder supplied by the manufacturers. Pentex HSA of a different lot number revealed this same impurity on d i a f i l t r a t i o n with Tris buffer. From these studies, i t seems unlikely that the "impurities" anise as a result of the d i a f i l t r a t i o n procedure i t s e l f . Figure 15 shows that similar d i a f i l t r a t i o n volumes are required for puri-fying plasma and the Fraction V HSA (5.9 x 10" 4 M). However, the time for d i a f i l t r a t i o n of the plasma sample was prolonged (7 hours) compared to that of the HSA (4 hours), because the flow rate through the plasma was slower. The increased protein present in the plasma must cause this decrease in flow rate. '.- ': - 86 -4.0 100 200 ' 300 Filtrate Volume (ml.) Figure 15. Disappearance of Protein-Like Substance from Plasma ( ) and Fraction V HSA (-—) on Diafiltration with Tris Buffer. - 87 -I) Void Volume in the Diafiltration Apparatus With the d i a f i l t r a t i o n apparatus used in these studies, void volume was negligible ( i . e . , the volume between the membrane and outlet port in the c e l l ) . After purification of an HSA sample in the d i a f i l t r a t i o n c e l l , the tubing collecting the f i l t r a t e was always emptied prior to a binding experiment. Thus, the correction for void volume described by Blatt, Robinson and Bixler (1968) and Ryan and Hanna (1971) was not necessary. ' J) Fluctuations in the Cell Volume during Diafiltration When the d i a f i l t r a t i o n procedure (see Experimental Section II, 8) without the C/D selector was followed, an i n i t i a l increase in volume of approximately 5 ml. occurred. This was due to different pressures in the reservoir tank and the c e l l . Once this i n i t i a l increase occurred, the cell volume remained constant. As a result, the HSA concentration in the cell would be decreased and HSA concentrations were always corrected for this dilution factor (by multiplying the HSA concentrations by V^/V^. When the d i a f i l t r a t i o n procedure using the C/D selector, was followed, increases in cell volume were rarely observed. However, in the case of an increase in volume, the HSA concentration was always corrected for this dilution factor. [Note: (1) The volume of the reservoir tank is very large compared to that of the cell and presents d i f f i c u l t i e s in obtaining equal pressures in both containers. (2) Ryan and - 88 -Hanna (1971) eliminated the problems of cell volume fluctuations by f i l l i n g the cell to capacity with macromolecule solution. . In this case, s t i r r i n g efficiency may be reduced.] 2. Drug Binding Studies by the Diafiltration Technique A) Calculation of Results Drug binding data was obtained in a manner similar to that of Ryan and Hanna (1971), though in this study i t was unnecessary to calculate areas under the curve (when drug f i l t r a t e concentration is plotted versus cumulative f i l t r a t e volume). Calculations were done by computer. A Fortran IV program "DRUGFIT" was prepared for calculating r and r/Df values (see Table 9). Plotting routines, and a modified non-linear least squares f i t (LQF) analysis (used by Meyer and Guttman [1968 b]) were included as subroutines. Thus plots of the f i t t e d line and the experimental data points could be obtained, and n i , k 1 } n 2 and k 2 values computed. For each set of experimental data, a series of data cards were typed. (a) Preparation of General Data Card:. Table 9. "DRUGFIT" Program XXXXX X XXXX XXXXXXXXXXXX XXXXXXXX XXXXXXXXXXXXXXXXXXXXX XXXX XXXXXXXXXX XX XXXXXXXX XXXXXXXXXXXXXXXXXXXXX XXXX XXXXX XXXX XXXXXX XX XXX XXX* XX XX XX XX kF«i nn. 01122? IIMIVFRSTTY OF H C. P. IMPUTING CENTRE MTSIAN192) 27:^H!31 THU APR H/72 _._ » ~ < i : : » » 9 : S : » E » » c . : ! i : » » > » c » S > u s 3 i s » 3 s : » » : 3 H 3 » « 3 l » » 3 S » t B » » = FILE FUR DELIVERY = F J LE FOR DELIVERY = FILE FOR DELIVERY = FILE FOR DELIVERY = FILE FUR DELIVERY « t r ======== = = c = = = = = : = = == = = = = = = = = = = == = = r = = == = = = = = = = = = = = = = = == = == = = ,= = = 3 = , » = = . = ==== = c c B = 3>>S=iS...3 = a<.a = c c . \ »*»o*»<.o»*»******«»4i PLEASE RETURN' TO ELECTRICAL ENGINEERING #***»*«*«**»#»*«»»»* Sb IG C'IPR **LAST blGNDNWAb: 22 :37 : 2 9 THU APR 13/72 n< c o iiruiirm ^ i r n c n nM AT n o u - n MM T u n ADD l * / 7 ? 4LIST LURUGF1T 1 C*»* CECILY PALMER 2 C*****PRUGRAM UKUGFIT 3 CUHMIIU/UEUUG/FLAG 4 LOGICAL FLAG Mt-AI PW I 1 V> 1 -Kkhc M Of! 1 . M. Al ?<>n 1 . VI 21101 . III1FI 1 SO 1 .R Fl 1 511 1 - P I 41 6 7 U 9 10 1 1 REAL W < 100') t El ( 4 ) , E2I4I, RFUF(IOO) -INTEGER MU ( 4 | 41 FORMAT ('1','R E b U L T S1,3X,I 2,////) 42 FORMAT('0',3X, 11 = ',UX,'V = ' ,HX,'A=',7X,'DX=1 ,7X,'DF=' ,7X, 1 ' F I =.' ,bX , 1 S-' ,7X ," 1)11= • ,6X ,' 1)1111=' , B X ,' R=' i 6X, 'RUF=' ,// ) 4 3 FORMAT (IX, I5.10E1O.3) • 12 13 14 15 1 6 1 7 44 FORMAT I'O'1, 'TOTAL V O LUME = ' , F 9 . 3, 2 X , ' ML ' , / / / ) ' 45 FORMAT ( ' 0', 'CA=' ,E 15.7,2X , ' tl=' , El 5.7,2X, 'C=' , El 5. 7, 2X, 'C0=' , 1 E15.7.2X, 'M=' ,E15.7,2X, 'L = ' ,2X,I4,2X, ' VX= • , F6 .2 ,2 X, ' PLOT 1 ///.) 4 6 FORMAT I '0 ' , 'CA= • ,E 15.7.2X, ' H= •, El 5.7, 2 X, 'C=' , El 5. 7,2 X, 'C0=' , 1 El 5 . 7 . 2 X , •/'!=• ,P15 .7,2X. 'L = ' ,?X. I 4.2X, • V X = ' . F6.2 .2X. 10 19 20 21 22 j < 1 'I'lUPLUT' ,///) 47 FORMAT ('U', 'EST I MAT El) PARAMETERS') • 48 FORMAT (' ',/, 'Kl = ' ,t 15 .'.,//,' Nl = ' , El 5.4,//, 'K2= ', El 5.4,//, 1 'N2=',E15.4,////) 49 FORMAT i'O', 3X, ' I =' , 1HX, 'R=', 17X, 'RF=' , 1 5X, ' R/OF=' , 14X, 1 RF/DF= 1, 1 7 X , • H<H 1 PF I'.r HIT 1 = I T / / 1 24 25 2 b 27 2tt 2 9 50 i-ORMA'll' ' , 1 5 , 4fc20.5,F20.2) 51 FORMAT ( 1 0 '., 'MUNI. I NEAR FITTING WITH THE FUNC T I UN' ,///, 10 X , 1 ' K F = N 1 M K 1 * U F / I 1+K1*UF ) + N2 * K2* UF/I 1 + K2 *D F) 1 , / / / , • -1 'FIRbT ' , 1 2 , ' POINTS NEGLECTED',//) 52 FORMAT ( ' ' , ' E R R (PERCENT) = ABS ( 1-R/R F )*1()0 • ,///) 1 1 FORMAT [ F 6 . 4 . F A . 3 . F 6 . C . F 6 . 4 . F 6 . 4 . TO. 1 2 . 4 X . F4.2 .ftX . I 1 . 9 X . 1 1 . 30 31 32 33 34 35 1 9X . I 2 ) 12 FORMAT (P3.3,F4.3,F3.3,F4.3,F3. 3, F4. 3, F3. 3, F4. 3, F3. 3, F4. 3, 1 F3.3,F4.3,F3.3,F4. 3.F3.3, F4.3, F3. 3, F4.3, F3.3, F4. 3, l i : i . : i , F 4 . :i) KK = 1 I COUNT = 0 3 6 37 3b 3 9 40 4 1 P ( 1 ) = . 411E 6 P ( 2 ) = 2 . 3 7 P ( 3 ) = .59E4 P (4 ) = 1 3 . 6 C A L L PLOTS i o RI-AI) (').ni r.A. H . r . r.n. M . I . ' V X , K, K F , K I I -42 43 44 45 4 6 L 7 FLAG = .FALSE. IF (KF.Euh) FLAG = .TRUE. ICOUNT = ICOUNT +1 CUUNT = FLOAT!ICOUNT) IF (K.ME.I) KK=0 tp ir A .i-n.fi. i fin Til VII . _ . _.. _ Table 9. (Cont'd.) 4 a H = FLOAT(L1/100. 49 CO = CO*l .E-5 50 H = Mw 11 ,-H)*l ,E~4 51. .. MVB = 1 52 •" • .... 1 3 MVb = flV-1 +. 10 ' • ' • \ P.F&n• IH.l ? 1 H U 1 . Al I 1 .I=NVR. NVFI ? 54 1)0 21 1 =NVB,NVE t s 55 ••' IFIVIIl.EU.O.) GO TO 31 , . 56 .21 CONT I NUE • .... 57 NVB = NVE +1 . . . 5a GO TO 13 ' '• " "M M - I - \ 60 WRITE I 6 ,'t 1 ) I COUNT 61 IF IK .ElJ.l ) GCJ TO 15 . • 62 14 KKITfc<6,45) CA, Bf C t CO, M, L, VX «>3 GO O 16 64 15 WKITE ( 6,'.6) CA, B, C, CO, H, L, VX A * 1 A l i b I T F ( f- , & ? ) 66 F l = 0 . 67 . VT = 0 . . 68 00 17 I =1 ,IM - . '. . 69 V I I ) = VII1*100. 70 IF ( A l l ) .EU.O. 1 All)=l.E-3 7 1 11/ = A 1 1 ) /T.A* 11 . +H ) 72 OF = UX/B 73 Fl = F l + UX*V(II 74 VT - VT+V( I ) 75 S = C*VT 76 OB = b-FI-VX*DX . .• -77 DUB = CO*()ft 78 K = 0BI1/M 79 k[)P = R/UF . . 80 WKITE 16,43) I , V I I ) , A( I ) , DX , DF ,F I , S , Dti , DBB ,R , KD F 8 1 RK( I ) = K b2 ODFII) = DF h 1 7 HkNF 111 = H n F b4 KDU = KO+1 85 00 23 I=KUD,N 8 6 J = I-KU ' a 7 RKIJ) = RR(I 1 aa ODFIJ) = DDF(I ) HO 2 •> » p ( i f | i ) = « » p c ( M • ' 90 M - N-KU - " 91 CALL HAKF1T IN, DDF, RR, RF, P, SSO) 92 WKITE I ) VT 93 WRITE 16,51 ) Kl) 94 C Kl = Mil) . 95 r. M1 = V 1 ? 1 96 c K2 = y (3) 97 c N2 = H(4 ) • 9d WRITE lb.'.71 99 WRITE(6,4a) (Ml1,1=1,4) . 1 00 WKITE I 6,52) 10 1 WRITE 16,491 102 DU 18 I=1,N 103 RFDFI I ) = RF (I)/DDF I 1 ) 104 DIFF = RRDFIIl-RFDFI I) 105 ERR = ABS(U1FF/RFDFI 1 I )*100. . ' •106 J = 1 + KU •• 107 WKITF I A.«in 1 .1. HR (I 1 . BFIl 1. KKIlFIll . B F I l f l l 1. F R H J i au ie a. ^OITC 'Q.) r . 108 KnDi-d ) = HH\lf ( 1 l » l . t : - 6 109 18 KFUFI1 ) = KFDF ( 1 ) * l . f c - 6 ' . • 110 IF (K.f.0.1 ) 00 T(l 10 '• ' . . . . i l l CALL bCALE IKK,N, 10. ,KMIN, OR , 1 > 112 CALL SCALE IKRDF ,14, 10. ,KDF MIN.UKD F, 1) 117 no ?? 1=1 M . • • 114 22 RFOFll) = IKFIlF (1 l-RUFMINI/UKUF \ 115 CALL AXIS (0 . ,0. , 1HR.-1 , 10. .0. ,RMI N, OR) 116 CALL AXIS (0. ,0. ,9HKDF*10**6,9 , 10. ,90. ,R0FM1 N,URDF) 117 CALL L I 0 E IKK , K F D F ,Nt 1 ) 11B 00 19 1=1 ,14 1 14 1 9 CAM SYI1H0I ( H K I I I . R R O F I I l , .07. ?. 0.0. -11 120 CALL SYH'JOL (4.5,8.5, .3,4HPL0T,0. , 41 . * 121 CALL , UUI1BEK (5 .7,8.5., .3 .COUNT, 0. ,-1.) 122 CALL PLOT(11.,0.,-3) 123 00 TO 10 124 20 IF (KK.EU.O) CALL PLOT00 1 25 STOP 126 EMU 12 7 •, SUBROUTINE PARE IT (AT, X X I , YY, CY, A, SSEI 128 DIMENSION X1I150), X2I150), Y ( 1 501 , CY1150), DV(IO), A ( I O ) , 129 1 S X X d O . l O ) , SXY(IO), UIAG(IO), UA(10),-S(10,10) 130 niHEIIllUN XX11150), YY1150) . 171 r ritc.Mdf. I IV , y l T Y ? , Y .u 7 f n i ,vwi T 132 CALCY 1 Al ,A2, A3.A4,X )=Al*A2*X/( l . + Al*X )+A 3*A4»X/( 1 .+A3*X) 1 33 1)1.1 10 1=1,150 134 XI (I ) = XXI(I 1 135 10 Y d ) = YYU ) 136 M=4 . . -1 7 7 MAMP = 1 . . 13b V=10. 139 C=l.E-2 140 ' TF.b'T=l .t-4 • , 141 ITER=0 142 1)0 70 1=1 , H 1 ',3 OA(I)=0 . 144 70 COOT 1 HUE 145 SSE=bSUL; (A,UA,N,M) 146 100 ITER =ITER+1 14 7 CALL i DHi (H,SXX,SXY,A,DIAG,M) K b K RI T = 1 1 4 9 TC=f./V 150 . I F ( C - l .E-8)121, 121,110 151 110 CALL SLVMISXX ,bXY,TC,OIAG,l)A,S, M) 152 TSiE = SSUE (A,DA,N,Ml 153 116 lFUbbE-bSE)20(), 130, 120 154 130 If ITC-l.£+8)120, 300,300 1 55 1 20 KKIT=KRIT+] 156 121 TC=TC*V . 157 GU TO 110 15b 200 C=TC 159 SSE = TSSE 160 K0UL=1 1 hi nn ??n I=i .M ' : 162 A 11 )=A(1 )+0A(I )*DAMP 1 63 OUUT=AUS (UA( I 1 )/( 1.E-3 + ABSI A d 1 )) 164 IFlUUUT-TEST1220,220,210 1 65 210 KU0L=2 166 220 COOT IiMUE, V 167 r;o TO (300,1 o o i . K o n i . . . . . . . ._ Table 9. (Cont'd.) c 1 6b . 300 0(1 5 i-l . f f 169 5 CY(I) = CALCYIAI 1) ,A(2) ,A( 3), Al 4) , X 1 U ) > 170 Kb TURN 171 E N D 1 72 SUBROUTINE SOMS(N,SXX.SXY,A,UlAG,M) ' ' ' 173 nn.iFWs 1  IM >; KV (i (i, i fi) , IKV 11 n), nvi, l n ), ii] nr.f ) n ) ,j 11 n| 174 CUI'.MON U.V ' ; ~ : : —\ .175 DO 10 I=1,H 176 SXYll )=o. : •• 177 DO 10 J=l,1 178 SXXI I ,J ) =U. . ' " 179 1 (i Tf|MT INHF 180 DO 30 K = 1 ,N ; 181 CALL UVFY IK , A,FY , M) 182 DO 20 1 =1 ,11 183 SXY ( I ) =SXYl I )+FY*DVI I ) • 184 DO 20 J = l ,1 ' 1 KS SXX 1 I ..1 l=SXX I I ..1 I+DVI I 1*DV I .1) 1 86 20 CONTINUE • ' .' . 87 30 CONTINUE 1 8b DO W 1=1,11 189 OIAGl I 1 = 1 ./SORT ISXX ( I , I ) ) 190 ' SXYU )=SXY(I )*DIAG(I) 191 nn 4o .1 = 1 . i 192 SXXU ,'J l=SXX(I ,J)*DIAG(I )*DIAG(J) 193 40 CONTINUE 194 RETURN . . . . 195 END " 196 SUBROUTINE SLVM(SXX,SXY,C,DIAG,OA,S,M) 1^7 n |«pm [ i IM •.YY(inlin) ISYY(in)rDIA'*-(1Q)inMlQ),S(lniln) 198 DO 20 I=1 ,M 199 DO 10 J = l , I ' 200 S I 1 ,J ) =S XX 1 I , J ) 20 1 S (J , I ) =SXX II , J) 202 10 CONTINUE 9(1 1 M I , i l =>. I l , I ) +r 204 20 CONTINUE 20 5 CALL CU(S ,M,SXY,0,DET) 206 DO 40 1=1,M 20 7 DAI 1 1=0. 208 DO 30 J = 1 , M 209 DA 1 I 1 =I)A I 1 ) +<, I 1 ..1 X-SXYI .11 2 10 30 C NTINUE 21 1 II)=DA(1>*D AG(I  212 40 . CONTINOE . • . 213 RETURN 214 END 21 5 FUNCTION SS0E1A.DA.M.M) 216 DIMENSION A I 10 ) ,UA( 10) , TAI 10) 217 DIMENSION XI1150),X2(150),Y(150),DV(10) 218 COMMON 0V,X1,X2,Y,R ZERO,YNIT 219 CALCY(A1,A2,A3,A4,X)=A1*A2*X/1 1. + A 1*X )+A 3»A 4«X/I l.+A3*X) 220 00 10 I=1 ,M •>-> 1 TA1n = A 1 I )+nA 11 ) 222 10 CONTINOE 22 3 SSOE=0. • 224 DO 20 K=1,N 225 FY=YIK l-CALCY (TAI 1) ,TA(2) , TAI 3), TAI 4) , XKK) ) 226 SSOE=SSUE+FY*FY 22 7 9(1 r ( I M T i MI I P I O U I C 3 , ^uonfa.) c 22 tf RETURN 22V EMU 230 SUBROUTINE DVFYIK,A,FY,M) 231 1)1 MENS I U N DV 1 1 0 ) , Al 10) ,X 1 ( 150) ,X2( 1 50) , Yt 1 50) 232 COMMON DV,X1,X2,Y,R/.ER0,YINIT s 2 33 CAI.CYIA! .A2.A3.A4.X)=Al*A2*X/( l.+Al#X)+A3*A4*X/(l . + A3*X) ? • 234 01) 5 1 = 1 ,!•'., 2 \ 235 TEKM=1.+A(I )*X1 (K 1 236 OVI I ) = TEKIi*A 1 1 + 1 )*Xl I K)-A( I )*A( 1 + 1 )*X 1IK )*X1(K ) 237 . nV(I)=UV(I )/(TEKH*TEKM) 238 5 • nVlI+1)=AI I )*X1IK)/TERM ; F Y =Y 1 X 1 - r A l r Y 1 A 1 1 1 , A 1 ? ) , A 1 7) , A l 4 ) , V 1 I K ) 1 240 RETURrJ 241 E N D . • • 242 SUBROUTINE ClilA.N.Il.M.DET) 243 . DIHENS ION A ( 10, 10) , B ( 10) ,IPVOT( 10) , I NOEX( 10,2) ,PI VOT< 10) 244 EQUIVALENCE I 1 ROW,JKOW),<ICOL,JCOL) ?4t> 57 I1FT =1 . 246 D O 17 J=l,0 24 7 17 IPVOT(J 1=0 24b DO 135 1 = 1 ,N . 2 4V 1=0. 2 5 0 D O 9 J=l,0 2 01 IF I IPVOT I J 1-1)13.9.13 252 13 D O 23 K=l,N 253 IF IIPVOTIK 1-1)43,23,81 254 43 IF (ABS IT J-ABS IA I J , K) ) ) 8 3, 23,23 ,.• 2 5 5 M3 I K(IW = J 256 ICOL = K . . . . . . . . . 7 ',1 ' T= A 1 ;l .K I • 25H 23 • (.•;..• i 1 HUE 2 5V V com i H U E • 260 IPVOTIICOLl=IPVOT(ICOLl+l 2 61 IF IIkOW-lCOL173, 109,73 2 6 2 73 I ) E T = - U I : T 7 i;>. nil 1 2 1 =1 ,M 2 6 4 T = A I I K O W , L ) 2 I.'J A I I k l IW , L ) = A I ICOL , L ) 2 6 6 12 AIICIIL,L) = T 2 67 IFIID10V,1U9,33 ' " ' 268 ' 33 T=BIIhOW) 7 l/l 11 1 1 I'Hl: 1 sit 1 i r.ni i 270 • B(ICOL)=T 2 7 1 10V IODEX11, )=IROW 272 INDEX I 1,2) = ICOL . •. 2 73 PIVUTII)=A(ICOL,ICOL) 274 DET=OET*PIVOT ( I ) ? 7 5 A( I C O L ,ICOL ) = 1 . 2 7 6 IJII 20'j L = l ,11 2 77 205 A 1ICOL,L)-h1ICOL,L)IV IV0TI1) 278 I F 1 ••'•).)'• 7 ,347, 52 2 7V 52 B ( ICOL ) =11 < ICOL >/P IVOTI I ) . . . 2 8 0 347 D O 135 L I =1 ,11 . ?hl I F (1 I - I C O I 121 .1 3S .7 1 282 21 T=AIL1,1C0L) 283 A (LI,ICUL) =0. 284 D O B9 L = l 2 85 89 A (LI ,L )=A IL I ,L ) - A ( ICOL,L)*T 286 IF(M)135,135,68 2 H 7 P.ll I UP.H I l-IU TTfU I'M J lame y. (Cont'd.) 2bb 135 r.uoTliiUE 2b9 222 DO 3 1=1,N 2 90 L= M-I + 1 291 IF( IHUEXU.l 1-1NDEXIL,2> ) 19, 3, 19 292 19 JRUW = IliDEX (L, 1 ) Jia . ,K:OL= i o o F x l L . 2 1 ; ' 2 9 4 .00 549 K = 1,N 295 T = A (K , JkUW ) 296 AU,JKUW)=A(K,JC0L) . 29 7 A(K,JCOL)=T 2 9b 549 CONTINUE r r ^ i T T fji it- ; •  300 . b l , RETURN 30 1 . . . END ENO OF FILE - 95 -CA B C CO M I VX K KF KD / / piA?' 53fcd9 OO^id 11 Oil O O D I D k? 1 1—2—3—4—5—6—7—8—9-10-1—12-3-14-5—16-7—18—19—20—2f—2-3—24—25—26-7—28—29—30—31—32—3—34—35—3G—37—38-9^40-1—42-3-4-15-4S-47-3-49-50-t—52-3-54—5-6-57—58-9-60 GENERAL DATA CARD 61 62 6  64 65 6  67 68 9 70 71 72 73 74 75 76 7  78 79 FROM U.B.C. COMPUTING CENTRE OflflOOOOOOOQOUUOOOOOflP 0 0 ijUO 0 0 0 0 0 fl 0 0 fl fl 0 0 fl 0 0 0 0 0 0 fl 0 0 0 0 0 fl 0 0 0 fl 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 3 4 5 6 7 8 9 10 1  12 13 14 15 16 17 18 19 20 21 2  2  24 25 26 27 28 29 30 31 32 3  34 35 35 37 3  39 40 41 42 434  45 46 47 48 4950 51 52 53 54 5  56 57 5B 59 IjO 61 62 63 64 65 6  67 68 6 9 70 71 72 73 74 75 76 7  78 79 80 t i n 111111111111111 DOi Gi n 11 Q11 n i n i l 111111111111Q1111111111111111111T1111111111 222222222 IJ2 22222 Q2 2222222222 2 2222222222 CZ 222222222222222222222222222222222222222 3 3 3 3 3 3 DC 3 3 3 3 3 3 3 03 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 44Q4444444444444444444 Q4 4 4444444444444444444444444444444444444444444444444444444 3 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 0 5 Q 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 66'B 6 6 G 6 6 D6 6 6 6 6 6 6 6 6 6 6 6 6 E 6 6 6 6 6 6 6 6 S 8 6 6 6 6 6 6 6 6 6 6 6 G 6 6 B 6 6 6 6 6 G 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 777Q777777777777777777777777777777777777077777777777 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 8 8 3 3 8 3 8 8 8 8 8 8 3 3 8 8 8 8 8 3 8 8 3 8 8 8 3 8 6 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 90999999990999099999999999999999999989999999999999999999999999999999999999999999 1 2 3 4 5 6 7 8 9 10 I  12 13 14 13 16 17 13 19 20 21 2  23 24 25 26 27 28 29 30 31 32 3  34 35 36 37 38 39 40 41 42 43 4  45 46 47 4649 50 51 52 53 54 5  56 57 58 59 60 61 62 63 64 65 6  67 68 69 70 71 72 73 74 75 76 7  7! 79 80 010 04 CS ' , The parameters CA, B, C, CO, M, and VX are printed onto the card. CA = absorptivity of the drug (e.g., 59.47 for BHC) B = molecular weight of the drug (e.g., 336.29 for BHC) C = concentration of drug in reservoir solution (e.g., 9.32 x 10"3 g./l.) CO = the reciprocal of the product of VX (the final volume in the cell) and the molecular weight of the drug M = the concentration of HSA solution in moles/1. : (e.g., 5.15 x 10~5 moles/1.) L .= • a constant normally left blank in binding studies. The use of this constant for varying various para-meters of the binding system will be explained in a later section. , VX = the volume in the cell, in ml. (e.g., 27.0 ml.) \ K = a constant. When K = 1, no printer plot will be plotted out. If K t 1, a printer plot will be given. KF = a constant. If KF = 1, the debug will be on. If KF f 1, the debug will be off. / - 96 -KD = in the case of inaccurate i n i t i a l data, the number of i n i t i a l r and D f values to be deleted f o r the non-linear least square f i t . All r and r/D f values are given on the read-out. Calcomp plots may be obtained instead of printer plots. (b) Preparation of Experimental Data Card: etc. i . ' b u c b d u 9 b u c o 1 U ? b Uc ?o' U9 -1—2—3—4—5—S—7—8—9—10-—12—13-4—15—16—1J-?. 14 096 0327 096 0341 09603550360367 09603790960339096 04 0 —20—21—2—23-24—25—26-27—28-29—30—31—32—3-34—35—38—37—38-3-40-41—42-4J~l4—45-46-'!7849-50-5l—52-53-54-5-56—57-58-59-G0-GENERAL DATA CARD FROM U.B.C. COMPUTING CENTRE El 62 63 64 65 6  67 68 59 70 71 72 73 74 75 76 7  78 79 ] o o C O O O L O O Q o o o Do o L'O o o Do o Co o o uo o uo o O L j O O L j O o o u o o i - j o o D L J o o u o o ouooyoooQooDoooQooOoLjoooo 1 2 3 4 5 6 7 8 9 10 1  12 13 14 15 16 17 18 19 20 21 2  23 24 25 26 27 28 29 30 31 32 3  34 35 36 37 3! 39 4 0 41 42 43 4  45 46 47 48 49 50 51 52 53 54 5  56 57 58 59 6 0 61 C2 63 64 6 5 6 6 67 6 6 69 70 71 72 73 74 75 75 7  70 79 80 1111111111Q111111111111011111111111111 Di 11111111 n i i 11111111111 i i n i i 11 i i 11 i i 2222G2222220222222G22222222222222Q222222222222222'2 22222222222222222222222222222 2 3 3 3 3 3 3 0 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 0 3 3 3 3 3 3 0 3 3 3 3 3 3 D 3 3 3 3 3 3 Q 3 3 3 3 3 3 D 3 3 3 3 3 3 D 3 3 3 3 3 3 Q 3 3 3 3 3 3 3 3 3 3 3 3 44444444444444444444444444404444444444440444444444444444444444444444444444044444 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 DD5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 &6.D66D6660666666D6666660666666Q666666066666606S6666D66D66606B6666D666666D6666666 7777777777777777777777777777777777077777 77777777777777707777 70777777777777 8 8 8 8 8 8 8 8 8 8 8 8 G 8 8 8 8 8 8 8 08 8 8 8 8 8 S 8 3 8 8 0 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 0 8 8 8 8 8 9G999999G9 9 ' 99990S9nG99D9999990999999099999909S9999D99999909999D939999Q9Q99999999 1 2 3 4 5 6 7 1 9 10 1  12 t3 14 15 16 17 18 19 20 21 2  23 24 a 26 27 2829 30 31 32 3  34 35 35 37 38 39 40 41 42 43 4 ! 454C4J 4849 50 51 52 53 54 5  56 57 50 59 6O61 62 G3 64 65 6  67 6  69 70 71 72 73 74 7576 17 7879 60 01.0 04 CS • Where V = volume of fraction of f i l t r a t e collected (e.g., 9.6 ml.) and A = absorbance reading of this fraction (e.g., 0.263). Experimental data is typed onto data cards f i l l i n g each card up to column 77 and then beginning on subsequent cards until A a l l the data is entered. The card deck is read into the computer, and the results are printed out in 10 columns with one row for each given data point. The ten columns are: V, A, DX, DF, FI, S, DB, DBB, R, RDF. - 97 -Where: . V = volume of fraction collected in ml. A = corresponding absorbance reading for this fraction DX = concentration of drug in f i l t r a t e fraction in g . / l . (Emg./ml.) i.e., concentration of free drug DF = concentration of drug in f i l t r a t e fraction in : moles/1. FI = cumulative sum of the amount of drug in the f i l t r a t e in mg. = FI Q .+ [(V) x (DX)] = T. [(V) x (DX)] S = the total amount of drug in the cell and f i l t r a t e in mg. = C x VT (i . e . , total f i l t r a t e volume is VT at this point) DB = S - FI - [(DX) x (VX)] = total bound drug in mg. in the ce l l where (DX) x (VX) is the amount of free drug in the cell in mg. DBB = the amount of bound drug in moles/1. = DB x CO R = DBB / M RDF = R / DF ni,.k 1,'n 2, k 2 values are printed out. The percent errors of the experimental data points as compared to the f i t t e d line are . given. B) Plotting of Results There are several ways of presenting binding data. The method of choice is the one which characterizes binding over as much of the - 98 -binding isotherm as possible. Plots vary in their a b i l i t y .to•show difference in binding behaviour and to give accurate values of n and k. Plotting methods have been described by Goldstein (1949) and Meyer and Guttman (1968). These can be derived from various rearrange-ments of Equation 8. An adsorption type of plot, where r (or some parameter propor-tional to Dfa) is plotted against Df, gives values of k which are subject to considerable error. It is d i f f i c u l t to ascertain i f data obeys the Law of Mass Action. A double reciprocal plot is frequently used [e.g., Ryan and Hanna (1971), Crawford and others (1972), Solomon and others (1968)]. . Here 1/r is plotted against 1/D^ , where the intercept on the 1/r axis gives 1/n and the slope of the line gives 1/nk. In this plot there i s disproportionate weighing of points at high and low values. Hence, a few low values will outweigh many high values. Nichol and others (1967), showed that double reciprocal plots could, theoretically, distinguish between the various combinations . of polymerization of the macromolecule and ligand binding. The double reciprocal plot could assume six different possible forms:-i) A straight line, with slope and intercept independent of protein concentration, i.e., a single protein or series of polymers with polymerization and binding occurring independently, a l l sites being equivalent, or non-competitive isomerization with one form being inactive. -- 99 -i i ) A family of straight 1ines with different intercepts on the 1/r axis but a common intercept on the 1/D^  axis; i.e., non-competitive polymerization and ligand binding, with one form of polymer binding no 1igand. i i i ) A curve convex to the 1/Df axis, independent of protein concentration; i.e., non-competitive isomerization and binding, . with one form having less a f f i n i t y for ligand than the other, or a single protein with two sets of sites or two non-interacting proteins with different binding a f f i n i t i e s . . . . * iv) A curve convex to the 1/Df axis with a common 1/r intercept; i.e., non-competitive polymerization and binding with polymeric species having less a f f i n i t y for ligand than.the monomer. v) A curve i n i t i a l l y concave to the 1/Df axis, independent of protein concentration; i.e., competitive isomerization and ligand binding. . ; vi) A series of curves i n i t i a l l y concave to the 1/Df axis with a common 1/r intercept; i.e., competitive polymerization and ligand binding. v;'/... (In (v) and ( v i ) , the curves may subsequently become convex to the l/Df axis). - Although these results are theoretical, the above effects may ^ be causing curvature of the plots, rather than deviations from the Law of Mass Action. (See Literature Section on Theory II (1)). - 100 -With albumin, isomerization occurs at low pH (Sogami and Foster [1968]). Various techniques have been used to show that polymerization occurs in HSA solutions. (Schultze and Heremans [1966]). Pederson (1962) showed that a 25% w/v aqueous solution of HSA stored for 24 . hours at 4°C was mostly in the monomeric form. Thus polymerization and isomerization could be factors influencing HSA-1igand binding and the shape of binding curves. Crawford and others (1972) used the classification of Nichol and others (1967). Double reciprocal plots of bromosulphthalein-HSA binding data were indicative of a competition between polymeriza-tion of the HSA and binding of the ligand. In a Scatchard plot, (Scatchard [1949]), r/D^ is plotted against r. The intercept oh the r axis (obtained by extrapolating the i n i t i a l linear portion of the curve) gives ni and extrapolation to the i"/Df axis gives n i k l v The slope of the i n i t i a l straight line portion gives -k i . In this plot, points are more evenly weighted. Frequently when this plot is used to express protein binding data, a curve is obtained. It is possible to resolve this curve graphically to give n and k values for two sets of sites. Accuracy in data decreases when more than three sets of binding sites are considered (Rosenthal [1967]). Relevant papers using Scatchard plots to obtain binding parameters are by O'Reilly (1971), Rosen (1970), Cho (1970) and Chignell (1969). From the above discussion, the Scatchard plot appears to be preferable for expressing binding data. Binding data from these . - 101 -studies was, therefore, expressed in Scatchard plots. nj_ and ki_ were obtained by graphical extrapolation of the i n i t i a l straight line portion of the curve. No attempt was made to evaluate n 2 and k2-Recently papers have been published on the use of computer techniques to f i t macromolecule-ligand binding data to a mathematical model, avoiding the subjective bias of plotting techniques. (Fletcher and Spector [1968] , Meyer and Guttman [1968 b]). In these studies, n l 5 k l 5 n 2, and k 2 values were also obtained from the computer program. A modified version of a non-linear least square f i t analysis (Meyer and Guttman [1968 b]) was included in subroutines in the computer program/Binding data was assumed to f i t to a relationship which assumes two classes of binding sites exist on the HSA molecule, i.e., n 1k 1D f n 2k 2D f r = — + — 1 + k xD f 1 + k 2D f r is the dependent variable and the independent variable. Approxi-mate estimates for n l 3 k l s n 2, k 2 were put into the program. (See Table 9). The more accurate these estimates are, the fewer the iterations required for the computer calculations. In contrast to graphical extrapolation methods, this LQF method has no subjective bias. - 102 -C) Bishydroxycoumarin - Human Serum Albumin Binding Results These results are shown in Table 10. ' i) Precision of Results: Figure 16 shows a typical calcomp plot for these binding results. The experimental values deviate l i t t l e from the fi t t e d l i n e , except at low r values. This plot shows that the d i a f i l t r a -tion technique yields precise results ( i . e . , a smooth binding curve with l i t t l e scatter of data points) except at low molar binding ratios. At low r values, the limits of accuracy of . spectrophotometry analysis are exceeded. A more sensitive analytical technique should eliminate this problem. i i ) Reproducibility of Results: Experiments 4 and 5 (Table 10) indicate that the d i a f i l t r a -tion technique can produce reproducible results. -. i i i ) Effect of Reservoir Bishydroxycoumarin Concentration: Experiment 3 indicates that the ligand reservoir concen-tration of BHC can influence binding results. Here the LQF ki value is low compared to results from other experiments in Table 10. When reservoir ligand concentration i s high compared to the HSA concentration, fewer.data points are obtained at low r values and thus the LQF value is lower. To obtain a wide range of data points, the ratio of HSA concentration reservoir ligand concentration should be unity or less. - 103 -8.0 6.4 4.8 1.6 3.2 4.8 6.4 8.0 r = Qh/\A : •%^ v'-;^ :^ -vv:-'V- '-y Figure 16. Calcomp Scatchard Plot for Bishydroxycoumarin - Human Serum Albumin Binding. & = experimental data points and solid line is non-1inear_ least square f i t t e d line. HSA concentration_= 3.685 x 10 _ 1 + M and BHC reservoir concentration = 2.821 x-'lO"1* M. (Experiment 6 in Table 10). Calcomp plot reduced in size from original plot. - 104 -iv) Effect of Pressure: The results in Table 10 suggest that pressure does not have a significant effect at 15 psi or above. Operation at the very low pressure in experiments 6 and 7 indicates that high n^x values are obtained. v) Comparison of n and k values obtained by Graphical Extra-polation and by LQF Method: Table 10 reveals that, in most cases, ki values are higher and rii values are lower, by the LQF method than by the graphical extrapolation method. This difference could be expected. , In the graphical extrapolation method, the i n i t i a l f i r s t few points v/ere ignored because of considerable scatter. The percentage error of these f i r s t few experimental data points compared to the f i t t e d ..line-was high. However, when the LQF method was used, al l experimental data points were included. vi) Comparison with Literature Values: Results given in Table 10 do not agree well with the literature values given in Table 11, and are consistently higher. However, results by the d i a f i l t r a t i o n technique cover a wider range of r values than do methods commonly used for studying binding. By d i a f i l t r a t i o n , binding data can be obtained at much lower r values than can be measured by equilibrium dialysis or by u l t r a f i l t r a t i o n . Thus mki values would be expected to be higher. Comparison with literature values is d i f f i c u l t and of l i t t l e value unless experimental conditions and the range of molar binding ratios investigated are the same. - 105 -vi i ) Effect of Human Serum Albumin Dilution: Figure 17 shows that HSA dilution has a very noticeable effect on BHC-HSA binding curves. As HSA concentration decreases the binding curves move toward higher r and r/D f values (Figure 17) > and ni and ki values increase (Table 10). However, there is no simple relationship between n x or k x and the HSA concentra-tion (Figures 18 and 19). The simple Law of Mass Action approach to protein binding studies suggests that BHC-HSA binding should be independent of HSA concentration. This deviation from the expected behaviour will be further discussed in Section IV, (2), F. ' ,; v i i i ) Double Reciprocal Plots: Binding data from Experiments 2, 8, 9 are plotted as double reciprocal plots in Figure 20. The series of curves are i n i t i a l l y convex to the 1/D^  axis, with a common intercept on the 1/r axis. Using the classification of Nichol and others (1967) discussed in IV, (2), B, this suggests that noncompetitive polmerization of HSA and binding of BHC is occurring, with the polymer having less a f f i n i t y for BHC than the monomer. D) Phenylbutazone - Human Serum Albumin Binding Results These results are shown in Table 13. i) Precision of Results: Figure 21 shows a typical calcomp plot for PBZ-HSA binding. Again, as. in IV, (2), C, ( i ) , the di a f i l t r a t i o n technique yields 20 - 106 -15 10 LO o Q Figure .17, • • • .. r = D b / M ' The Effect of Human Serum Albumin Concentration on Bishydroxycoumarin-Human Serum Albumin Binding. A = 0.420 x 10 _ l f M HSA and 2.875 x 10" 5 M BHC in reservoir, i = 0.724 x \0~k M HSA and 5.750 x 1G~-5 M BHC in reservoir, © = 6.123 x 10" 4 M HSA and 2.882 x 10 _ 1 + M BHC in reservoir. (Experiments 9, 8 and 1 in Table 10) - 107 -20 1 2 3 4 5 6 HSA Concentration, moles x 1 0 ~ V l . Figure 18. Relationship Between kj and Human Serum Albumin Concentration for Bishydroxycoumarin - Human Serum Albumin Binding and Bishydroxycoumarin - Plasma Binding. {-®—®-) = Fraction V HSA and (-*-—*-) =. Plasma. (Data from Figure 17 and Figure 26.) 2 3 4 HSA Concentration, moles x 1 0 ~ V l . Figure 19. Relationship between ni and Human Serum Albumin Concentration for Bishydroxycoumarin - Human Serum Albumin Binding, x = data from HSA (Fraction V) (from Figure 17). 0.8 Figure 20. Double Reciprocal Plots for Bishydroxycoumarin-Human Serum Albumin Binding. Symbols and concentrations are as for Figure 17. (Experiments 1, 8 and 9 in Table 10) 1/Df x 10 5 Figure 21 Calcomp Scatchard. Plot for Phenylbutazone - Human Serum Albumin Binding. J =• experimental data points and the solid line is the non-linear least square f i t t e d line. HSA concentration = 3.585 x 10_l> M and PBZ reservoir concentration =9.59 x 10"5 M Table 10. Bishydroxycoumarin- Human Serum Albumin Binding Results. Experiment HSA Cone, in moles/1. Reservoir BHC Cone, in moles/1. Pressure psi Extra-polated nj Extra-polated kj in l./M LQF1] LQF kj in l./M LQF ."2 LQG k2 in l./M Extra-polated nikj in l./M LQF nikj in l./M LQF n2k2 in l./M. 1 6.123 x 10"11 2.882 X 10"" 15 3.0 5.00 X 10s 2.4 4.81 x 10s 13.6 5.95 x 103 1.50 x 106 1.13 X 106 8.10 x 104 . 2 4.257 x 10"4 2.901 X 10"4 15 2.9 5.24 X 10s 2.4 5.38 X 10s 12.2 6.50 x 103 1.52 X 106 1.29 X 106 7.92 x 10\ • '3 1.802 x 10"4 2.805 X 10"4 25 3.0 5.39 X 10s 2.7 3.24 x 10s 17.9 5.92 x 103 1.62 X 106 8.90 X 105 1.05 x 10s 4 2.884 X lO"" 1.007 X 10"4 25 2.8 5.92 X 105 1.7 1.71 x 106 11.4 9.73 x 103 1.66 X 106 2.86 X 106 1.11 x 105 5 2.815 X 10"1* 1.185 X 10"4 25 3.0 5.84 X 10s 1.9 1.26 x 105 268.0 3.92 x 102 1.75 X 106 2.41 X 106 1.05 x 10s 6 3.685 X 10"4 2.821 X 10~4 7 3.5 5.78 X 105 2.3 1.46 x 106 28.4 3.66 x 103 2.02 X 106 3.44 X 106 1.04x10s 7 0.810 X 10"4 2.438 X 10"5 4 2.8 8.32 X 10S 2.0 2.83 x 106 39.7 1.59 x 103 2.33 X 10G 5.55 X lO6 6.32 x 105 8 0.724 X 10*4 5.749 X 10"5 15 4.5 8.22 X 105 3.1 1.37 x 106 8.4 2.30 x 104 3.70 X 106 4.24 X 106 1.94 x 10s 9 0.420 X 10~4 2.830 X 10*5 15 5.2 1.70 X 1Q6 3.9 1.53 x 106 13.0 1.10 x 104 8173 X 106 6.00 X 106 1.45 x 105 Note: LQF = values from non-linear least square f i t method. - i n -precise results, except at low molar binding ratios (where wider scattering of data points occurs). i i ) Effect of Reservoir Phenylbutazone Concentration: Experiments 11 and 15 (Table 13) show that reservoir PBZ concentration influences binding results. This effect is more marked than that with BHC, because n]_ is smaller. Again when the. ratio of PBZ reservoir concentration: HSA concentration is greater than unity, no data points are obtained at low molar binding ratios. In experiments 11 and 15 i t was impossibl to determine n and k values by graphical extrapolation or.by the LQF method. i i i ) Effect of Pressure: Experiments 10, 12, 13 and 14 (Table 13) indicate that pressure has l i t t l e . e f f e c t on the PBZ-HSA binding results. iv) Comparison of n and k Values Obtained by Graphical Extrapolation and by the LQF Method: Again, as in IV, (2), C, (v), Table 12 shows n x values are lower and kj_ values higher by the LQF method than by . : graphical extrapolation. However, as indicated in Table 13, poor non-linear least square f i t s were obtained for several experiments. In these experiments fewer data points are obtained for the i n i t i a l straight line portion of the Scatchard plot. If these few points are widely scattered, then the LQF . will be poor. But elimination of these data points; from the LQF analyses,, made i t impossible - to analyse data by LQF method. - 112 -v) Comparison with Literature Values: Results in Table 13 are higher than the literature values given in Table 12. In the literature, a variety of methods and experimental conditions were used. As explained in IV, (2), C, ( v i ) , comparison of literature values i s only valid i f the same range of r values are investigated under the same . experimental conditions. -vi) Effect of Human Serum Albumin Dilution on PBZ-HSA Binding Results: Figure 22 shows that as HSA concentration decreases, there is a shift of PBZ-HSA binding curves to higher r and r/D f values. Table 12 shows n 2 increases as HSA concentration . decreases (both by graphical extrapolation or by LQF method). k l 5 however, by graphical extrapolation appears to remain con-stant, whereas by the LQF method, i t increases except for the lowest HSA concentration used in Experiment 17. Figure 23 shows there is no simple relationship between n and HSA con-centration. The effect of HSA dilution on binding results will be discussed in IV, (2), F. . v ' : v : v i i ) Double Reciprocal Plots: Binding data from experiments 12 and 17 is plotted as double reciprocal plots in Figure 24. The curves are i n i t i a l l y convex to the 1/D.p axis, with a common intercept on the 1/r axis Using the classification of Nichol and others (1967), (see IV, 1 2 3 4 r = Db/M Figure 22. The Effect of Human Serum Albumin Concentration on Phenylbutazone - Human Serum Albumin Binding. = 0.420 x 10 _ t t M HSA and 3.359 x 10"5 M PBZ in reservoir a = 0.778 x 10 _ l + M HSA and 6.415 x 10" 5 M PBZ in reservoir o = 3.8.92 x 10"u M HSA and 1.284 x 10 - 1 + M PBZ in reservoir (Experiments 12, 16 and 17 in Table 13) - 114 -2 3 4 5 HSA Concentration, moles x 1 0 ~ V l • Relationship Between n1 and Human-Serum Albumin Concentration for Phenylbutazone - Human Serum Albumin Binding. (Data from Figure 22.) Q, Figure 24. Double Reciprocal Plot for Phenylbutazone Humar .Serum Albumin Binding. . (Data from Experiments 12 anc 17 in Table 13.) - 115 -Table 11. Literature Values for n and k for Bishydroxycoumarin -Human Serum Albumin Binding Studies n . k(l./M) Temp pH [HSA] Method Reference 3 3.5 x 10 5 20° C 7.4 , (Tris Buffer) 0.1 -> 0.3% w/v Equilibrium Dialysis Cho (1970) ' 2.8 1.7 x 10 5 40°C 2.0 1.15 x 10 5 * 27°C 10 0.4% w/v Equilibrium Dialysis O'Reilly (1971) 3.2 7.5 x 10 5 ** 25°C 7.4 (Phosphate Buffer) 1 x 10~5M Equilibrium Dialysis Chignell & ;; Starkweather (1971) * O'Reilly reports k x value as 2.31 x 10 5, but this is actually an n ^ value. * Calculated from data in paper. Table 12. Literature Values for n and k for Phenylbutazone - Human Serum Albumin Binding Studies n k(l./M) Temp PH [HSA] Method Reference 1 1.17 x 10 5 25°C 7.4 1 x 10"^ M U l t r a f i l t r a -tion (by centrifugation) Solomon and others (1968) 1 1.25 x 10 5 Thorp (1964) 1.14 1.86 2.37 x 10 5 4.56 x 10 4 7.4 (Phosphate Buffer) 0.1% w/v Circular Dichroism Rosen (1970) 1 : 2 1 x 10 5 4 x l-O4 7.4 (Phosphate Buffer) 1 x 10"5M Equilibrium Dialysis Chignell .(1969) Table 13. Phenylbutazone - Human Serum Albumin Binding Results. Experiment HSA Cone, in moles/1. Reservoir BHC Cone, in moles/1. Pressure psi Extra-polated Extra-polated ki in l./M LQF ni LQF ki in l./M LqF "2 :• LQG k2 in l./M Extra-polated n ^ i in l./M LQF nikj in l./M LQF n2k2 in l./M . . * 10 4.775 x lO"4 9.09 x 10"5 8 i.i 4.66 x 10s • - • 4.89 x 10s r * ** 11 4.398 x 10""* 6.260 x 10"4 15 • - • . • - V- • - •. - ' * 12 3.892 x lO"4 1.284 x 10"4 15 1.2 4.59 x 10s - - - 5.51 x 105 ••• -.' : 13 3.585 x 10"" 9.59 x 10" 5 10 1.3: 4.78 x 105 1.1 8.33 x TO5 29.5 5.97 x 102 6.46 x 105 • -* 14 1.809 x 10"4 7.240 x 10"5 25 1.1 4.77 x 105 • - - - . 5.25 x 105 • - •• . ** 15 1.495 x 10"4 4.530 x lO"4 25 - - - - ' - - - • -16 0.778 x 10"4 6.415 x 10"5 15 1.5 4.51 x 10s 1.1 2.86 ;c 106 3.59 9.12 x 103 6.81 x 10s 3.16 x 106 3.25 x 104 17 0.420 x 10"4 3.359 x 10"5 15 • 1.9 4.54 x 105 1.3 8.41 x 105 430 ? 6.44 ? 8.80 x 105 1.09 x 106 2.77 x lO4 Poor non-linear least square f i t . No non-linear least square fit possible. - 117 -(2), B), this suggests that noncompetitive, polymerization of HSA and binding of PBZ could be occurring, .with, the polymeric species having less a f f i n i t y for PBZ than the monomer. E) Bishydroxycoumarin - Plasma Albumin Binding Results These results are shown in Table 14. The pressures used in Experiments 18, 19 and 20 were chosen to give similar flow rates. With plasma, d i a f i l t r a t i o n experiments proceeded more slowly than with Fraction V HSA. i) Precision of Results: , Figure 25 shows a calcomp plot for plasma albumin - BHC binding. Again, the d i a f i l t r a t i o n technique yields precise results, except at low r values. ; i i ) Effect of Pressure: Experiments 20 and 21 (Table 14) show that the pressure applied does not appreciably affect BHC - plasma binding. i i i ) Comparison of n and k Values Obtained by Graphical Extra-polation and by LQF Method: As in the case of BHC - HSA and PBZ - HSA, the LQF method '. gave higher kx values and lower n x values than the graphical; extrapolation method. iv) Comparison of n and k Values with Literature Values: With BHC - plasma albumin binding, the d i a f i l t r a t i o n : technique again yields higher values for k x than given in the - 118 -literature (Table 11). Explanations for the higher values : have already been given in IV, 2, C, ( v i ) . v) Comparison of BHC - Plasma Albumin Binding and BHC - HSA Binding: Comparison of the results given in Tables 10 and 14, indicate that binding is similar, although plasma has a s l i g h t l y higher a f f i n i t y for BHC. Plasma proteins other than albumin, may also bind BHC. It has been shown that BHC does bind to other proteins such as «-,3- and y-globulins.; but with less a f f i n i t y : (Weiner and others [1950]). Another explanation could be that differences exist in the conformation of Fraction V HSA and plasma albumin. Endogenous substances may have been removed from Fraction V HSA during the fractionation procedure. It was noted also, in IV, (1), H, (v) that much more protein-like substance was lost into the f i l t r a t e on purification of 4% w/v HSA, than from plasma. vi) Effect of Plasma Dilution: Figure 26 indicates that shifts of BHC - plasma binding -. curves occurs to higher r and r/D f values as plasma dilution increases. By the graphical extrapolation method n : remains quite constant, although i t increases sl i g h t l y by the LQF method, as plasma dilution increases kl increases as plasma dilution increases (either by graphical extrapolation or by LQF method). There is no simple relationship between kx and the extent of plasma dilution. (See Figure 18). This plasma dilution effect on BHC - plasma binding will be discussed in IV, (2), F. -119 1.6 r = Db/M Figure 25. Calcomp Scatchard Plot for Bishydroxycoumarin - Plasma Binding. ..•A= experimental data points and the solid 1 ine is the non linear least square f i t t e d line. HSA concentration = 4.632 x 10 _ I + M :and BHC reservoir concentration — 2.185 x 10 h M. (Reduced in size from original.) 20 - 120 -15 + r = Db/M •Figure-26. The Effect of Plasma Dilution on Bishydroxycoumarin - Plasma Binding. 0.516 x 10" 4 M HSA and 2.768 x 10"5 M BHC in reservoir. 1.208 x lO" 4 M HSA and 5.100 x 10" 5 M BHC in reservoir. 4.632 x 10"'f M HSA and 2.185- x 1:0"^  M BHC in -reservoir. ' . A = m = o = Table 14. Bishydroxycoumarin - Plasma Binding Results. Experiment ' HSA Cone, in Plasma moles/1. Reservoir BHC Cone, in moles/1. Pressure psi Extra-polated • " i Extra- * polated kj in l./M LQF ni LQF ki in l./M LQF n2 • LQG k2 in l./M Extra-polated niki in l./M LQF njki 1n l./M LQF n2k2 in l./M : 4.632 x 10"" 2.185 x TO"4 45. 2.9 5.52 x 10s 1.9 9.863 x 105 10.6 6.22 x 103 1.49 x 106 1.87 x 106 6.64 x lO* 19 1.208 x 10"1* 5.10 x 10"5 15 . 3.1 1.22 x 106 2.7 1.45 x 106 381 1.75 x 102 3.78 x 106 3.88 x 106 6.67 x- 10" 20 0.516 x 10"4 2.768 x 10"5 10 3.0 1.92 x 106 2.5 5.58 ;:x 106 289 2.89 x 102 5.46 x 103 1.37 x 107 8.35 x 10" ;. 21 0.516 x 10"4 2.888 x 10"5 22 - - 2.6 5.94 x 105 7.1 1.34 x 10" - 1.57 x 107 9.50 x 10" CM F) The Effect of Human Serum Albumin Concentration on Binding Results Obtained by the Diafiltration Procedure On the basis of the binding results shown in the preceeding sections, i t can not be concluded at this stage that the d i a f i l t r a -tion procedure is a satisfactory technique for drug - HSA binding studies. The problem of dependence of binding curves on HSA con-centration has to be resolved. This problem is possibly inherent in the d i a f i l t r a t i o n procedure. However, i t is also possible that the simple Law of Mass Action does not adequately describe drug - HSA binding and that binding is not independent of macromolecule con-centration. This emphasizes the necessity of determining the effect of HSA concentration on binding curves in a l l binding studies. i) Literature Reports on the Effect of Human Serum Albumin . Concentration on Binding Results It is surprising to find that only in rare instances has the effect of protein concentration been studied. Meyer and Guttman (1968 b) when studying the binding of several drugs to BSA by dynamic dialysis, used two concentrations of BSA. They showed only a slight shift in binding curves at different BSA concentrations. The shifts,however, were in the opposite direction to those of BHC binding curves in these studies. A; small s h i f t like this could be explained by experimental errors in D f and/or M (see Theoretical Discussion of this in IV, (2), Cho (1971) using equilibrium dialysis and 0.1, 0.2,..0.4% w/v HSA solutions showed no significant effect of HSA concentration on BHC-HSA binding curves. However, Cho only investigated the effect - 123 -of HSA concentration over a very narrow range of HSA concentration. Crawford and others (1971) studied bromosulphthalein (BSP) -HSA binding by gel f i l t r a t i o n . In the 2.7% w/v - 5.5% w/v HSA range, the BSP binding capacity in moles/moles HSA decreased as albumin concentration was increased [according to the equation y = k + a'i + b + c 2 , where y = log (number of BSP molecules/molecule HSA), A X k^jand c are constants common to a l l groups of serum tested, a^ is characteristic of the serum used, and x is log of albumin concen-tration]. However, since they only used one equilibrium concentration of BSP, they could be effectively obtaining molar binding ratios at a section through a family of binding isotherms. .' '• To avoid this latter problem, Crawford and others (1972) investi-gated d i a f i l t r a t i o n as a method whereby an entire binding isotherm could be obtained. Thus, the effect of HSA concentration on BSP binding, both in serum and in isolated HSA could be examined. These authors reported that, in serum, there appeared to be a transition between molar BSP binding capacity increasing with increasing HSA concentration, and a decreasing capacity with increasing HSA con-, centrations, especially at equilibrium BSP concentrations of about 5 x 10~5 M. The molar binding capacities for albumin and serum are very similar in the 40 yM free BSP concentration range but, below this, serum has a lower capacity. At high D f values, serum has a higher capacity (this must be due to presence of endogenous compounds bound to HSA in serum with a relatively low a f f i n i t y ) . However, these - 124 -researchers worked only at 1.5%, 2.3 and 3.39% albumin concentrations. At the two lower HSA levels, BSP binding results (on double reciprocal plot) were very similar, whereas at 3.39% the results indicated an increased n value and a resultant decreased k value (they published ass no results for these binding parameters, however). It would have been of interest to see results at lower HSA concentration. It should be , noted that these authors give no indication of any preliminary studies on the d i a f i l t r a t i o n apparatus, e.g., binding studies to apparatus membrane binding or retention, etc., for these factors could indeed be present and affect results. They do not compare their results : quantitatively with other literature values. The effect of BSA concentration on testosterone - BSA binding over ; a wider BSA concentration range was studied by Ryan and Hanna (1971). Their results were similar to those shown in these studies. Their nk values were always consistently higher than the literature values although the n values were the same. They also encountered the problem of dependence of testosterone binding values on BSA concentration. Their explanations for the anomalous results are given below:- . (a) They suggested that the properties of the membrane may be ., influenced by the presence of protein. (b) The binding properties of the protein changes under the conditions of d i a f i l t r a t i o n . (c) Values of n and k obtained for steroid-binding by the classical 48 hours equilibrium dialysis are not true values. - 125 -There was, however, no evidence given to support these views. Thus, there is l i t t l e in the literature on this HSA concentration effect. When this subject has been investigated with HSA, studies have not been over a wide range of HSA con-centration. To determine whether binding is independent of HSA concentration, a wide range of HSA concentrations (i.e., from 0.1% - 4%. w/v) including physiological concentrations, should be used. i i ) Theoretical Investigation of Altered Parameters Involved in Drug-Macromolecule Binding and Their Effect on the Scatchard Plot The effects of changes in the parameters M and D f were examined theoretically. Experimental data previously obtained was used and general data cards were appropriately altered (Table 15). The computer program allowed for such changes. Situation 1: Assume that 10% of the original macromolecule. M, in the c e l l escaped through the membrane. This macromolecule interfered with measurement of absorbance of the drug in the f i l t r a t e by absorbing at the analytical wavelength. This increased the apparent concentration of drug in the f i l t r a t e , D^ , by 10%. From Figure 27, i t can be seen that, in this situation, the curve is moved slightly upwards and across to the right, i.e. for any point on the original curve, the r and r/D^ values are increased. This effect does simulate the shift of binding curves observed in experimental data. However, the experimental shift - 126 -Table 15. Changes in General Data Cards Situation Effect Under Investigation Column 34, 35 on new General Data Card [HSA] on Original * General Data Card 1 [HSA] h by 10% D f f by 10% 10 , Unchanged 2 [HSA] unaltered D f f by 10% 10 t by 10% 3 [HSA] sV by 10% D^ unaltered — I by 10% 4 [HSA] unaltered D f^ by 9% - 9 ," I by 9% * When a number, e.g., -9 is put in columns 34 and 35 this means that whilst D f is decreased by 9%, [HSA] will be increased by 9%. Hence, for the [HSA] to be maintained constant, i t is necessary to decrease [HSA] on original data card by 9% on new data card. Figure 27. Scatchard Plots Indicating the Effect of Changing Parameters on Bishydroxycoumarin - Plasma Albumin Binding ( © © ) = experimental binding curve where reservoir BHC concentration 2.860 x 10~5 M and HSA concentration = 0.515 x 10 _ l + M. . ( * * ) ' = theoretical binding curve where f i l t r a t e BHC concentration, is,* increased-by 10% and HSA concentration is decreased by 10%. - 128 -was much greater than that calculated. To achieve such results theoretically would require a much greater percent change in M ., and Df. Experimentally, this would be unlikely and this does not seem to be an explanation for the results obtained. Situation 2: Assume a 10% increase in D f but an unaltered HSA concentration. This plot (Figure 28) shows this does not . alter the Scatchard plot significantly. For a given value of r, the r/D^ . value is slightly decreased and the curve moved down-wards. Thus a 10% error in is unlikely to alter results, significantly. Situation 3: Assume a 10% decrease in HSA concentration but unaltered values. This could occur i f the protein became concentrated at the membrane. This might reduce the concentra-tion of HSA available for drug binding. This plot (Figure 29) moves significantly up and to the right. The effect is similar but of lesser magnitude than that observed experimentally. This implies that concentration of protein at the membrane could influence results. Situation 4: Assume a 9% decrease in but an unaltered HSA concentration. This could occur i f , as Blatt, Robinson and Bixler (1968) suggested, the drug became polarized at the membrane and hence was retained. The D f values.in the f i l t r a t e would be :, low. This effect is the reverse of Situation 2 and, the plot 20 - 129 -15-la-s'-r = Db/M . Figure 28. Scatchard Plots Indicating the Effects of Changed Parameters on Bishydroxycoumarin - Plasma Binding. ( @ o ) = experimental binding curve (data as in Figure 27), and ( B a ) - theoretical binding curve where f i l t r a t e BHC concentration •\ is increased by 10% and HSA concentration is unaltered. 1 : : 1 : : ( • — _ 2 4 6 r = Db/M Figure 29. Seatchard Plots Indicating the Effects of Changed Parameters on Bishydroxycoumarin - Plasma Binding Curves. (©.'••) = experimental binding curve (data as in Figure 27) and ( o o ) = theoretical binding curve where BHC f i l t r a t e concentration is unaltered and HSA concentration is decreased by 10%. - 131 -(Figure 30) shows that there is a slight movement upwards of the curve. The net result of this theoretical study was that:-(!) It is unlikely that loss of protein molecule, through the membrane could account for the observed HSA concentration effects. (2) High D f values would not affect results significantly. (3) Experimental error in the measurement of M (or an inaccurate HSA concentration due to volume changes) in the direction of a decreased macromolecule concentration is probably the most significant of these four effects. A decreased protein concentration could also be obtained by concentration of the protein at the membrane. None of these effects, however, completely explain the experimental data obtained at a range of macromolecule concen-trations . i i i ) Discussion of the Human Serum Albumin Concentration Effect in Terms of Problems Inherent in the Dia f i l t r a t i o n Technique Blatt, Robinson and Bixler (1968), expressed, their methyl : orange - HSA binding data (using PM-10 f i l t e r ) , in a plot of log C.p/C.p - C against f i l t r a t e volume. Theoretically, this should be a straight line relationship [see Literature Section , II, (5)]. Their results showed that as the HSA concentration • r = VM- : 'r:': Figure 30. Scatchard P l o t s I n d i c a t i n g the E f f e c t s of Changed Parameters on Bishydroxycoumarin - Plasma Binding Curves. ( a &) = experimental binding curve (data as i n Figure 27) and ( i\ ^  ) = t h e o r e t i c a l b i n d i n g curve where f i l t r a t e BHC c o n c e n t r a t i o n i s . decreased by 9% and HSA c o n c e n t r a t i o n i s u n a l t e r e d . . increased, the curvature of these lines increased and their slope decreased. When binding data from PBZ - HSA and BHC-HSA inter-actions was plotted in the above manner of Blatt, Robinson and Bixler, similar results were observed. (See Figures 9 and 10.) Thus, binding results of Blatt, Robinson and Bixler also.appeared to be dependent on albumin concentration, as were the binding results in this study and those of Ryan and Hanna (1971). All these studies used the d i a f i l t r a t i o n technique. Therefore, this, apparent HSA concentration effect could be an inherent problem ' with the diafiltration technique. Possible inherent problems will now be discussed:-(a) Ligand polarisation at the membrane: Blatt, Robinson and Bixler (1968) explained their results in terms of polarisation of the methyl orange dye at the membrane. To overcome this problem they suggested using the lowest possible pressures and HSA concentrations or use of : a larger parosity membrane. Work with PBZ and BHC in these studies, suggests that operation at very low pressures gives high k1 values. Also i t is preferable to use near physiological HSA concentrations, rather than low HSA con-centrations. A larger porosity membrane than the PM-10 . Diaflo would only increase the problem of interfering impurities in the Fraction V HSA. - 134 -(b) Ligand Binding and/or Rejection by the Diafiltration Apparatus: The d i f f i c u l t y of distinguishing between ligand binding and rejection was discussed in IV, (1), 6. The effect of binding and rejection on free drug concentrations in the c e l l and f i l t r a t e was also discussed here. In IV, (1), F, the binding of PBZ and BHC to the apparatus was shown to average approximately 5-6% (though a relationship existed between the percent drug bound and f i l t r a t e volume, and ligand reservoir concentration). The theoretical section investiga-tion in IV, (2), F, ( i i ) , however, indicates that even i f : 10% of the drug were bound, i t could not cause shifts of the curves observed in these studies when HSA concentration decreased. It could be possible also that, in the presence of HSA, ligand binding and/or rejection is altered. (c) Altered Membrane Properties: Ryan and Hanna (1971) suggested that membrane properties may change during d i a f i l t r a t i o n . However, i t would seem more li k e l y that apparent altered membrane properties would be caused by increasing HSA concentrations. As the HSA concentration increases, clogging of the membrane may occur and flow rates would, therefore, decrease. Decreased flow rates with increased HSA concentration were observed in these studies and also those of Ryan and Hanna (1971). Decrease in HSA concentration caused by clogging at the - 135 -at the membrane is considered in the theoretical investi-gations in IV, (2), F, ( i i ) . It was shown that this effect was unlikely to cause binding curve shifts of the magnitude observed experimentally on HSA dilution (although i t could have some influence on s h i f t s ) . (d) Conformational Changes in the Human Serum Albumin Molecule During D i a f i l t r a t i o n : Diafiltration i t s e l f could possibly alter the HSA molecule and change i t s binding parameters, either through conformational changes or by polymerization and/or denatura-tion. This could be caused by effects of pressure or sti r r i n g stress or s t i r r i n g rate on the HSA molecule. Such changes would also occur during the d i a f i l t r a t i o n purification procedure. Ryan and Hanna (1971) considered this possibility. They washed BSA in the d i a f i l t r a t i o n c e l l for 12 hours with buffer. A subsequent equilibrium dialysis experiment indica-ted that the washed BSA had a decreased a f f i n i t y for testosterone. Chignell and Starkweather (1971) incubated HSA solutions for 24 hours at 37°. Data for binding curves for the inter-actions of both BHC, and PBZ, with this incubated HSA, were obtained by equilibrium dialysis. Similar shifts of Scatchard binding curves occurred with incubated HSA as did with bind-ing curves at low HSA concentrations in the present studies. Chignell and Starkweather (1971) suggested this might be due - 136 -to conformational changes, since the shifts of binding curves with acetylated HSA were similar to those with incubated HSA. Thus, i t could be possible that shifts of binding curves at low HSA concentrations in the present studies are due to conformational changes. It would be d i f f i c u l t to show whether such changes were indeed occurring. (e) Accuracy of the Literature Values: As has already been discussed in Section IV, (2), C, and D, i t is d i f f i c u l t to compare literature values with values obtained by the d i a f i l t r a t i o n technique because of the wider range of molar binding ratio covered by the dia- . f i l t r a t i o n technique. 3. Phenylbutazone - Human Serum Albumin Binding Studies by Desorption (Washout) Diaf i l t r a t i o n Technique These desorption studies were carried out by using 3.585 x 10_t* M HSA and 0.959 x 10"5 M PBZ. PBZ was desorbed from HSA molecules as dia-f i l t r a t i o n proceeded. Hence a decrease of PBZ concentration could be followed in the f i l t r a t e . Results are shown in Figure 31. Binding and desorption patterns were different. Even after 500 ml. of f i l t r a t e had been collected in the wash-out experiment, the absorbance of the f i l t r a t e was s t i l l 0.258. The reason for the lack of agreement between these two curves is not clear. . One possible explanation might be that the length of time required for the TOO 200 300 400 500 ml. filtrate collected Figure 31. . Binding of Phenylbutazone by Human Serum Albumin by the Desorption Method (o © ) and the Diafiltration Method (• o «) [HSA] = 3.583 x 10*1* M, [PBZ] = 0.959 x 10"" M. - 138 -purification, wash-in and washout experiments may permit denaturation and/ or aggregation of the HSA molecule. This may alter binding properties of HSA.' • 4. Bishydroxycoumarin - Human Serum Albumin Binding Studies by the . Equilibrium or Direct Method In these experiments, the HSA concentration was approximately 0.72 x 10~k M and BHC concentration in the reservoir varied from 1.78 x 10~5 M to 1.34 x 10~4 M. Since Df, Dt and M are known, Db and r can be found by means of Equation 25. The results are shown on a Scatchard plot in Figure 32. It was assumed in these studies, that the unknown substance passing through the membrane (from Fraction V HSA) would not affect BHC analysis in the f i l t r a t e at equilibrium. Results obtained with this method di f f e r from those obtained by the d i a f i l t r a t i o n technique and are closer to literature values (Table 11).. The ni value was found to be 3.50 and the ki value 2.69 x 10 5 1/M. The results have not been corrected for:-i) Depression of BHC absorbance at 310 my by HSA. Hence, there is a slight inaccuracy in the D^ reading. The inaccuracy is not signi-ficant (Cho [1970]). i i ) BHC binding and/or rejection. : This method is equivalent to equilibrium dialysis but the time involved is shorter. Only one point on a binding curve is obtained from each experi-- 140 -ment. There i s , therefore, no real advantage in this procedure, although the results indicate that the binding parameters are similar to those obtained by other methods. (See Table 11). It is not possible to cover such a wide range of molar binding ratios as is possible with the dia-f i l t r a t i o n procedure. Data points are more scattered than those from d i a f i l t r a t i o n experiments. Thus the equilibrium or direct method does not appear to offer advantages over conventional methods for protein binding studies. 5. Drug Binding Studies by the Centrifugation (Ultrafiltration) Method These experiments showed that Amicon Centriflo (CF 50) membrane cones cannot be used for binding studies involving Fraction V HSA. Further studies were not carried out, since membrane cones with a lower molecular weight cut-off were not available. It is interesting to note that this method is widely used in c l i n i c a l studies of drug binding to plasma. 6. Phenylbutazone - Human Serum Albumin Binding Studies by a Molecular Sieve Technique (Batch Method) Using Sephadex G-25 In PBZ - HSA binding studies, the total amount of PBZ (D b and Df) in the external phase is analysed spectrophotometrically. ; Since the total amount of PBZ i n i t i a l l y added to the gel is known, the amount (mg.) of PBZ associated with one g. of the gel can be calculated. . Thus, from Figure 6, the concentration of drug in the external phase can be deter-mined. Therefore, D^ is known and D, and r can be calculated. - 141 -For the PBZ - HSA interactions studied, erratic results were obtained. The free drug concentration was, in some experiments, higher than the i n i t i a l PBZ concentration in the external phase (in the absence of HSA). It was found that in control binding experiments where.Tris buffer only was added to the swollen gel, that a contaminant was present in the external phase. The external phase had a low absorbance reading at 264 my, the analytical wavelength for PBZ. This contaminant probably explains the erratic binding results obtained. Results for the determination of PBZ in the external phase were not corrected for depression of absorbance by HSA. However, this would probably have only a small effect on binding results. (Cho [1970] showed the depression of absorbance readings of BHC by HSA was most significant.) V. SUMMARY AND CONCLUSIONS The Amicon d i a f i l t r a t i o n apparatus was used to study BHC - Fraction ; V HSA, PBZ - Fraction V HSA, BHC - plasma interactions. • . , (a) The polyethylene tubing provided with the apparatus was found to bind BHC with a high a f f i n i t y . However, teflon tubing was found to release no UV absorbing substances, nor to bind either PBZ or BHC. Teflon tubing, was, therefore, used in the d i a f i l t r a t i o n apparatus. (b) Binding studies, with both PBZ and BHC, were carried out in the absence of HSA. Both PBZ and BHC bound to the d i a f i l t r a t i o n apparatus. The percent drug bound decreased with increase in f i l t r a t e volume, and with increase in the concentration of drug in the reservoir. (c) Experimental dilution curves ( i . e . , in the absence of HSA) deviated from theoretical dilution curves, for both PBZ and BHC. : This also suggested binding was occurring in the apparatus. (d) In these studies i t was not possible to distinguish between binding and/or rejection. ; •(e) Binding studies, in the absence of drug, revealed that a protei like substance, with maximum UV absorbance at 280 my, was passing through the Diaflo membrane. Its UV absorption characteristics and - 143 -a positive Folin test suggested the substance was protein-like. Its molecular weight appeared to be greater than 5000. (f) It was not possible to purify the HSA by dialysis with cello-phane membranes or by Sephadex G-25. The most satisfactory purifica tion procedure was by d i a f i l t r a t i o n for several hours, with Tris buffer, until no protein-like substance appeared in the f i l t r a t e . (g) Plasma also yielded protein-like material in the f i l t r a t e on d i a f i l t r a t i o n with Tris buffer. UV spectra revealed this substance differed from that released from HSA. (h) HSA released more protein-like material into the f i l t r a t e than did plasma. The above indicates that the d i a f i l t r a t i o n apparatus should always be checked for release of substances which may interfere with analysis. For each ligand used, binding and/or rejection in the apparatus should be determined. In protein binding studies, a membrane should be chosen with maximal retention for the protein and minimal binding or rejection : of the ligand. 2. Study of BHC - HSA, PBZ - HSA, BHC - plasma binding by the d i a f i l t r a tion technique had advantages over conventional methods used for protein binding studies:-(a) From one d i a f i l t r a t i o n experiment an entire binding curve can . . be obtained. - 144 (b) Except at low molar binding ratios, the binding data was precise (i. e . , there was l i t t l e scatter in data points). (c) An experiment can be completed in approximately 8 hours (the time is dependent on HSA concentration and ligand reservoir concentra-tion). Less time, therefore, is required for this method than for V"; the equilibrium dialysis technique. (d) Data for a binding curve was obtained over a wide range of molar binding ratios. Binding data could be obtained at lower molar binding ratios than can be obtained by conventional methods. (e) A wide range of protein concentrations could be used. On the other hand, equilibrium dialysis, circular dichroism, etc., studies are often restricted to low protein concentrations. : (f) The d i a f i l t r a t i o n technique is readily amenable to automation (i . e . , in-line analysis). (g) The Fraction V HSA can be purified by the d i a f i l t r a t i o n technique 3. Binding parameters obtained by graphical extrapolation differed from those obtained by the non-linear least square f i t computer method. The ni values were higher and ki values lower by the graphical extra-polation method. 4. Binding parameters for BHC - HSA, BHC - plasma and PBZ - HSA inter-actions obtained by either graphical extrapolation or by the LQF computer analysis, differed from the literature values, i.e., kj. values were higher and ni values lower. Binding curves for BHC - Fraction V HSA and BHC - plasma interactions were similar, but k1 values for BHC - plasma.interactions were slightly higher. Binding curves for BHC - HSA, BHC - plasma and PBZ - HSA interactions did not appear to be independent of macromolecule concentration. It is not clear whether the HSA concentration effect is a problem inherent in the d i a f i l t r a t i o n apparatus or whether the simple Law of Mass Action approach does not adequately describe such interactions. ; Although the d i a f i l t r a t i o n technique offers advantages over conventional methods, the problem of HSA concentration effect on binding curves must f i r s t be resolved before i t can be concluded that the d i a f i l t r a -tion technique is suitable for protein binding studies. The desorption (washout) d i a f i l t r a t i o n technique gave unsatisfactory results for PBZ - HSA interactions. The equilibrium or direct d i a f i l t r a t i o n method gave values for bind-ing parameters for BHC - HSA interactions which were close to literature values. This method, however, has no real advantage over the con-ventional methods used in protein binding studies.: Two other techniques, centrifugation (u l t r a f i l t r a t i o n ) and a Sephadex G-25 gel f i l t r a t i o n method (batch method) were also studied. These yielded unsatisfactory results and were not further investigated. REFERENCES Aggeler, P., O'Reilly, R.A., Kowitz, P., Leong, L. (1967). N. Engl. J . Med. 27_6, 496. Potentiation of the Anticoagulant Effect of Warfarin by Phenylbutazone. Aggeler, P., O'Reilly, R.A. (1969). J. Lab. Clin. Med. 74, 229. The Effect of Heptobarbital on the Response to Bishydroxycoumarin in Man.. Bates, R. (1961). N.Y. Acad. Sci. 92^ , 341. Physicochemical Properties of Amine Buffers. Amine Buffers for pH Control. Beckstead, H.D., Kaistha, K.K., Smith, S.J. (1968). J. Pharm. Sci. 57, 1952. Determination and Thin Layer Chromatography of Phenylbutazone .. in the Presence of Decomposition Products. Benesch, R.E., Benesch, R. (1955). J. Am. Chem. Soc. 77, 2749. The Stability of the Silver Complex of Tris-(hydroxymethyl)-aminomethane. Bennhold, H. (1938). Die Vehikelfunction der Bluteiweisserkbrper. Ch. in Die Eiweisskb'rper des Blutplasmas by H. Bennhold, E. Kylin and S. Rusznyak, Steinkopf, Leipzig, 1938. Blatt, W.F., Robinson, S.M., Bixler, H.J. (1968). Anal. Biochem. 26, 151. Membrane Ultrafiltration: The Diafiltration Technique and . its Application to Microsolute Exchange and Binding Phenomena. Blatt, W.F. (1969). American Laboratory, March 1969. Ultrafiltration: A preparative procedure for Protein Chemistry. Brodie, B.B., Lowman, E.W. (1954). J. Am. Med. Assoc.-, J_6, 181. Observa-tions on Antirheumatic and Physiological Effects of Phenylbutazone and some Comparisons with Cortisone. Brodie, B.B., Burns, J.J., Steele, J.M., Y i i , T.F., Guttman, A.M. (1956), in Proc. I l l European Rheum. Congress. Elsevier, Amsterdam, p. 600. Burke, C.W. (1969). Biochim. Biophys. Acta. 176, 403-413. Accurate Measurement of Steroid-Protein Binding by Steady-State Gel Filtration.: Burns, J.J., Rose, R.K., Chenkin, T., Goldman, T., Schuert, A., and Brodie, B.B. (1953). J. Pharm. 109_, 346. Physiological Deposition of Butazolidin in Man and a Method for its Estimation in Biological . Material. Chen, P.S. (1967). J. Biol. Chem. 242_, 173. The Removal of Fatty Acids from Serum Albumin by Charcoal Treatment. - 148 -Chignell, C F . (1969). Mol. Pharmacol. 5_, 244. Optical Studies of Drug-Protein Complexes II. Interaction of Phenylbutazone and i t s Analogues with Human Serum Albumin. Chignell, C F . (1970). Mol. Pharmacol. 6_, 1. Optical Studies of Drug-Protein Complexes IV. The Interaction of Warfarin and Dicoumarol with Human Serum Albumin. . Chignell, C F . and Starkweather, D.K. (1971). Mol. Pharmacol. 7_, 229. Optical Studies of Drug-Protein Complexes V. Interaction of Phenylbutazone, Flufenamic Acid and Bishydroxycoumarin with Acetyl- .. sa l i c y l i c Acid Treated Human Serum Albumin. Cho, M.J. (1970). M.Sc. Thesis, University of British Columbia, Vancouver , 8, B. C., Canada. Cho, M.J., Pernarowski, M., Mitchell, A.G. (1971). J. Pharm.'Sci. 60, 196. Interaction of Bishydroxycoumarin with Human. Serum Albumin. Cohn, E.J., Strong, L.E., Hughes, W.L., Mulford, J.J., Ashworth, J.N., Melin, M., Taylor, H.L. (1946). J. Amer. Chem. Soc. 68, 459. Preparation and Properties of Serum and Plasma Proteins IV. A System for the Separation into Fractions of the Protein and Lipo-protein Components of Biological Tissues and Fluids. Cohn, E.J., Hughes, W.L., Weare, J.H. (1947). J. Amer. Chem. Soc. 69_, 1753. Preparation and Properties of Serum and Plasma Proteins.. XIII Crystallisation of Serum Albumins from Ethanol-Water Mixtures. Cooper, P.F., and Wood, G.G. (1968). Protein-Binding of Small Molecules: New Gel Fil t r a t i o n Method. J. Pharm. Pharmacol. 20, Suppl, 1505-1565. Crawford, J.S.,Davies, P., Da vies, P. (1971). Brit. J. Anaesth.' 43,' 344. Further Studies of the Binding of Bromosulphthalein by Serum . Albumin. Crawford, J.S., Jones, R.L., Thompson, J.M., Wells, W.D.E. (1972). Brit. J. Pharmacol. 44, 80. Binding of Bromosulphthalein by Human Serum Albumin Using a Continuous Diafiltration Technique. Edsall, J.T., and Wyman, J. (1958). "Biophysical Chemistry". Academic Press Inc., New York,1958. Ch, 11, pp. 591-660., Foster, J.F., and Aoki, K. (1958). J. Amer. Chem. Soc. 80, 5215. Further Studies of the Isomerization of Bovine Plasma Albumin. The Effect of Detergent Ions at low pH and Preliminary Observations at high pH. Foster, J. (1970) in "The Plasma Proteins" Vol. I; Eds. F.W. Putnam. 1960 Academic Press, New York and London, Ch. 6. . Froese, A., Schon, A.H., Eigen, M. (1962). Can. J. Chem. 40, 1786. Kinetic Studies of Protein-Dye and Antibody-Hapten Interactions with the Temperature Jump Method. Goldstein, A. (1949). Pharmacol. Rev. 1_, 102. Drug-Protein Interactions. Goodman, D.S. (1958). ••!. Amer. Chem. Soc. 80_, 3892. Interaction of Human"Serum Albumin with Fatty Acid Anions. Grollman, A. (1926). J. Gen. Physiol., p. 813. U l t r a f i l t r a t i o n through Collodion Membranes. Herman (1959). Med. Exptal. 1_, 170. Uber den Stoffwechsel des Butazolidin Hitzig, W.H. (1960). Die Plasmaproteine in der Klinischen Medizin Springer Berlin, p. 110. Hughes, W.L. (1947). J. Amer. Chem. Soc. 69, 1836. An Albumin Fraction Isolated from Human Plasma as the Crystalline Mercuric Salt. Hummel, J.P. and Dreyer, W.J. (1962). Biochem. Biophys. Acta 63_, 530-2. Measurement of Protein Binding Phenomena by Gel F i l t r a t i o n . Karush, F. and Sonenburg, M. (1949). J. Amer. Chem. Soc. 71_, 1369. Inter-actions of Homologous ATkyl Sulfates with Bovine Serum Albumin. Karush, F. (1950). J. Amer. Chem. Soc. 72^ , 2705. . Heterogeneity of Binding Sites of Human Serum Albumin. Karush, F. (1951). Ibid 73, 1246. The Study of Protein Binding by . Partition Analysis. The Effect of Protein Charge. Kauzman, W. (1959). Adv. Prot. Chem. ^ 4, 1. Some Factors in the Inter-pretation of Protein Denaturation. Klotz, I.M. (1946 a). Arch. Biochem. 9_, 109. The Application of the Law of Mass Action to Binding by Proteins. Interactions with Calcium., Klotz, I.M., Walker, F.M., Pivan, R.B. (1946 b).J.Amer. Chem. Soc. 68, 1486. The Binding of Organic Ions by Proteins. Klotz,.I.M., Urquart, J.M. (1949). J. Phys. Coll. Chem. 53_, 100. Binding of Organic Ions by Proteins. Buffer Effects. Klotz, I.M., Ayers, J. (1952). J. Amer. Chem. Soc. 74, 6178. Interactions of Some Neutral Organic Molecules with Proteins. Klotz, I.M., Ayers, J. (1953). Disc. Farad. Soc. 13^ , 189. Protein Inter-actions with Organic Molecules. - 150 -Kostenbauder, H.B., Jawad, M.J., Perrin, J.H., Averhart (1971). J. Pharm. Sci. 6(3, 1658. Dialysis and Circular Dichroism Study of Binding of Sulphaethidi.o.l to Crystalline and Fraction V Human Serum Albumin. Lee, C C , Trevoy, L.W., Spinks, J.W.T., Jaques, L.B. (1950). Proc. Soc. Exp. Biol. Med. 74, 151. Mechanism of Bishydroxycoumarin Metabolism. Levy, G.,0'Reilly, R.A., Aggeler, P.., and Keech (1970). Clin. Pharmacol. Therap. 1J_, 372. Pharmacokinetic Analysis of the Effect of Barbituates on Anticoagulant Action of Warfarin in Man. Link, K.P. (1944). Harvey Lectures 39_, 162. The Anticoagulant from Spoiled Sweet Clover Hay. Lovrien, R. (1963). J. Amer. Chem. Soc. 85_, 3677. Interaction of Dodecyl. Sulfate Anions of Low Concentration with Bovine Serum Albumin. Lowry, O.H., Rosebrough, N.J., Farr, L., Randall, R.J. (1951). J. Biol. Chem. 193, 265. Protein Measurement with the Folin Phenol Reagent. McQueen, E.G., and Wardell, W.M. (1971). Brit. J. Pharmacol. 43_, 312. Drug Displacement from Protein Binding: Isolation of a Redistri-butional Drug Interaction in Vivo. V Meyer, M.C., and Guttman, D.A. (1968 a). J. Pharm. Sci. 5_7, 895. Review: The Binding of Drugs by Plasma Proteins. • ., Meyer, M.C, Guttman, D.E. (1968 b). J. Pharm.. Sci., 57., 1627. Dynamic Y Dialysis as a Method for Studying Protein Binding. Nichol, L.W., Jackson, W.J.H., Winzor, D.J. (1967). Biochemistry 6_, 2449. A Theoretical Study of Binding of Small Molecules to a Polymerising Protein System. A Model for Allosteric Effects. O'Reilly, R.A., Aggeler, P., Hoag, M.S., Leong, L.S. (1962). Thrombos-Diathes. Haemorrh (Stuttg) 8_, 32. Studies on Coumarin Anticoagulant Drugs: Assay of Warfarin and it s Biological Application.; -. O'Reilly, R.A., Aggeler, P., Leong, L.S. (1964). Ibid U_, 1. Pharmaco-dynamics of Dicumarol and Warfarin. O'Reilly, R.A., Aggeler, P., and others (1964 b). N. Engl. J. Med. 271, 809. Hereditary Transmission of Exceptional Resistance to Coumarin Drugs. -.. O'Reilly, R.A., Aggeler, P. (1968). Proc. Soc. Exph. Biol. Med. 128, 1080. Phenylbutazone Potentiation of Anticoagulant Effect. Fluoro-metric Assay of Warfarin. - 151 -O'Reilly, R. A., and Levy, G. (1970). J. Pharm. Sci. 59, 1258. Pharmaco-kinetic Analysis of the Potentiating Effect of Phenylbutazone'on the Anticoagulant Action of Warfarin in Man. O'Reilly, R.A. (1971). Mol. Pharmacol. 7, 209. Reaction of Several Coumarin Compounds with Human and Canine Serum Albumin. Pawelcyzk, E., Ewaryst and Wachowiak, R. (1969). Diss. Pharm. Pharmac. 21_, 491. Kinetics of Na - Phenylbutazone Hydrolysis.. -Pawelcyzk, E., Wachowiak, R. (1968). Diss. Pharm. Pharmac. 20_, 653. Chemical Characterisation of the Decomposition Products of Drugs: Identification of Some Decomposition Products of a Na - PBZ Solution. Pederson, K.O. (1962). Arch. Biochem. Biophys. Suppl. 1_, 157. Exclusion Chromatography. Pernarowski, M. (1969), in 'Pharmaceutical Chemistry". Chatten, L.G., ed., Dekker, New York and London, Vol. 2, Ch. 1, p. 1. Absorption Spectrophotometry. Perrin, J.H. and Idsvogg, P. (1971). J. Pharm. Sci. 60_, 602. Extrinsic Cotton Effects of Bishydroxycoumarin when Bound to Bovine Serum Albumin Peterson, H.A., Foster, J.F., Sogami, M.A., Leonard, W.J. (1965 a) J. Biol Chem. 240, 2495. The Microheterogeneity of Plasma Albumins I. Cri t i c a l Evidence for and Description of the Microheterogeneity model. Peterson, H.A., Foster, J.F. (1965 b,c). Ibid 240, 2503, 3858. The Microheterogeneity of Plasma Albumins II. Preparation and Solubility of Subfractions. The Microheterogeneity of Plasma Albumins III. Physiochemical Properties of Some Subfractions. Pulver and others (1956). Schweiz Med. Wschr. 8(5, 1080. Uber die Bge- . influssung enzymatische Reaktionen durch Phenylbutazon und die Uber agbarkeit fermentchemischer Befunde auf die Stoffwechselprozesse der Zelle. von Rechenburg, H.R. (1953). Schweiz Med. Wschr. 83, 159. . Rheumatherapie mit Butazolidin. von .Rechenburg, H.R. (1962) in "Phenylbutazone: Butazolidin". Ed. V. Rechenburg. Edward Arnold, London, 1962. Riva, G. (1960). "Das Serumweizild" 2nd Ed. Haben, Berne, p. 257. Robbins, J., Rail, J.E., Berman, M. (1965). Proc. Int. Thyr. Congress 5th, Rome 1965. Andreoli M. ed., p. 635. . : - 152 -Rosen, A. (1970). Biochem. Pharmacol. 1_9, 2039. Circular Dichroism Measurements of Binding Constants: Phenylbutazone and oxyphenyl-butazone. Rosenthal, H.E. (1967). Anal. Biochem. 20, 525. A Graphical Method for Determination and Presentation of Binding Parameters in a Complex System. Ryan, M.T., Hanna, N.S. (1971). Anal. Biochem. 40, 364. Investigation of Equilibrium Ultrafiltration as a Means of Measuring Steroid-Protein Binding Parameters. Scatchard, G. (1949). Ann. N.Y. Acad. Sci. 5J_, 660. The Attraction of Proteins for Small Molecules and Ions. Scatchard, G., Scheinburg, J.H., Armstrong, H. (1950). J . Amer. Chem. Soc. 72, 540. Physical Chemistry of Protein Solutions V. .The Combination of Human Serum Albumin with the Thiocyanate Ion. Schrogie, J .J ., Solomon, H.M. (1967). Clin. Pharmacol. Therap. 8, 70. The Anticoagulant Response to Bishydroxycoumarin. II The Effect of D-thyroxin, Clofibrate and Norethandrolone. Schultze, H.E., Heide, K. (1960). K. Fr. Bauer Ed. in "Medizinische Grundlagenforschung" Vol. Ill George Thieme, Stuttgart, p. 351. Schultze, H.E.Heimburger, W., Frank, G. (1962). Biochem. 2 336, 388. Die Aminosaurezusammensetzung des menschlin Prealbumins (Try) und die des Human-und Rinderserumalbumins. Schultze, H.E., Heremans, J.F. (1966), in "Molecular Biology of Human Proteins. Vol. I. Nature and Metabolism of Extracellular Proteins. Elsevier Co. Amsterdam, London, New York, 1966. Sogami, M., Foster, F. (1968). Biochemistry 7_, 2172. Isomerisation Reactions of Charcoal Defatted Human Serum Albumin. The N-F Transition and Acid Expansion. , Solomon, H.M,, Schrogie, J . (1967). Clin. Pharmacol. Therap. 8_, 65. The Anticoagulant Response to Bishydroxycoumarin. The Role of Individual Variation. Solomon, H.M.,.Schrogie, J . , Williams, D. (1968). Biochem. Pharmacol.17, 143. The Displacement of Phenylbutazone and Warfarin -;Clh from Human Albumin by Various Drugs and Fatty Acids. Steinhardt, J . , Reynolds, J.A. (1969). Eds. "Multiple Equilibria in Proteins" Molecular Biology Series. Academic Press, New York and London. Tanford, C, Buzell, J.G., Rands, D.G., Swanson, S.A. (1955). J . Amer. Chem. Soc. 7_7,. 6421. The, Reversible Expansion of Bishydroxycoumarin in Acid Solutions. - 153 -Tanford, C , Buzell, J.G. (1956). J. Phys. Chem. 60, 225. The Viscosity of Aqueous Solutions of Bovine Serum Albumin between pH 4.3 and 10.5. Tanford, C , Kirkwood (1957 a). J. Amer. Chem. Soc..79_, 5333. Theory of . Protein Titration Curves. I General Equations for Inpenetrable Spheres. Tanford, C. (1957 b). Ibid 79, 5340, 5348. (i) Theory of Protein Titration Curves. II Calculations for Simp! Models at Low Ionic Strength ( i i ) The Location of Electrostatic Charges in Kirkwood's Model for Organic Ions Binding. Tanford (1965) "Physical Chemistry of Macromolecules". Wileg and Sons, New York, London and Sydney. Ch. 8, p. 526. Thorp (1964) in "Absorption and Distribution of Drugs". Binns, T.B. Ed. Livingstone, London, p. 64. . von Scholtan, W. (1964). Arzneimittel-Forsch 1_5, 146. Vergleicherde quantitative Bestimmung der Eiweikbiridung von Chemotherapeutica mittels Sephadex und Dialyse. Weiner, M.S., Shapiro, Axelrod, M., Cooper, Brodie, B.B. (1950). J. A Pharmacol. Exptl. Therap. 99_, 409. The Physiological Disposition of Dicumarol in Man. Welch, R.M., Harrison, Y.E., Conway, A.H., Burns, J.J. (1969). Clin. Pharmacol. Therap. 1_0, 817. An Experimental Model in Dogs for Studying the Interaction of Drugs with Bishydroxycoumarin. Wood, G.C., and Cooper, P.F. (1970). Chromatographic Reviews 1_2, (1), 88. The Application of Gel Filtration to the Study of Protein-Binding : of Small Molecules. , ' Yang, J.T., Foster, J.F. (1954). J. Amer. Chem. Soc. 76, 1588. Changes in Intrinsic Viscosity and Optical Rotation of Bovine Plasma Albumin Associated with Acid Binding. Yang, J.T., Foster, J.F. (1953). J. Amer. Chem. Soc. 75_, 5560. The Bind-ing of Alkylbenzenesulfonate by Albumin. APPENDIX 1 Sample Calculation of Results by "DRUGFIT" Computer Program CA = 59.47, B = 336.29, C = 9.32 x 10"3 g . / l . , CO - 0.11014 x 10"3, M =0.515 x lO" 4 moles/1., L = 0, VX =27.00 ml., (K and KF as desired), KD = 0. (Results from plasma-BHC experiment). p\2\ v A R : • RDF / 9.5 .001 .188 3.76 x 106 9.7 .001 .381 7.62 x 106 9.7 .002 .573 5.73 x 106 9.4 .002 .759 7.59 x 106 9.2 ; .002 .942 9.42 x 106 9.6 .003 1.13 7.54 x 106-9.7 .003 1.32 8.82 x 106 9.7 .003 1.52 1.01 X 107 9.6 .006 1.70 5.67 x 106 9.7 .009 1.89 4.20 x 106 9.6 .016 2.07 2.59 x 106 9.6 .026 2.24 1.72 x 106 9.6 .041 2.40 1.17 x 106 9.6 .061 2.55 8.38 x 105 9.6 .085 2.69 6.34 x 105 9 . 6 .109 2.82 5.18 x 105 9.6 .134 2.94 4.40 x 105 9.6 .158 3.06 3.87 x 105 9.6 .183 3.16 3.46 x 105 9.6 .204 3.26 3.20 x 105 9.6 .226 3.35 2.97 x 105 9.6 .245 3.44 2.81 x 105 9.6 .263 3.53 2.68 x 105 9.6 .281 3.60 2.56 x 105 9.6 .298 3.67 .' 2.47 x 105 9.6 .314 3.74 2.38 x 105 9.6 .327 3.81 2.33 x 105 9.6 .341 3.87 2.27 x 105 9.6 .355 3.92 2.21 x 105 9.6 .367 3.98 2.17 x 105 APPENDIX 1 (Cont'd.) V A R . RDF 9.6 .379 4.03 2.12 x 10 5 9.6 .389 4.07 2.09 x 10 5 9.6 .400 4.12 2.06 x 10 5 9.6 .411 4.15 2.02 x 10 5 9.6 .421 4.19 1.99 x 10= 9.6 .429 4.23 1.97 x TO5 No data points eliminated from LQF analysis. ni = 2.5 n 2 = 85.7 ki = 5.57 x 10 5 k 2 = 971.2 APPENDIX 2 Results and Calculations for Phenylbutazone - Human Serum Albumin Binding by Equilibrium or Direct Method 1 2 3 Wt. of Sephadex (g.) i Ext. Volume . (ml.) HSA added to gel mg. mg. PBZ initially added to gel c Total PBZ (D. + Df) measured in external phase in mg. mg. PBZ associated with Ig. gel . Free PBZ in external phase in mg.. (Df from Figure 6) Bound PBZ in m.g. • 4.00,.. 15.32 161 •'. 1.553 0.894 0.165 0.515 0.279 4.00 15.32 161 0.776 0.282 0.T24 • 0.453 — 4.00 15.32 161 0.304 0.132 0.043 0.256 — 4.00 15.32 80.5 0.899 . 0.356 . . 0.133 0.469 . . . 4.00 , 15.32 33.4 0.T00 0.086 0.003 0.046 0.040 4.00 15.32 33.4 0.064 . . 0.045 .. 0.005 0.064 — Notes: 1 Total Volume =25 ml. 2 PBZ (mg.) associated with 1 g. gel = [PBZ (mg.) added initially - Total PBZ measured in external phase] x. wt. of gel. 3 Bound PBZ (mg.) = Total PBZ in external phase - Free PBZ in external phase. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0101347/manifest

Comment

Related Items