Relative affinities of bilirubin for serum albumins from different species

Biochimica et Biophysica Acta, 492 (1977) 64-69

© Elsevier/North-Holland Biomedical Press BBA 37649 RELATIVE A F F I N I T I E S OF BILIRUBIN FOR SERUM A L B U M I N S FROM D I F F E R E N T SPECIES

G. BLAUER, E. LAVIE and J. SILFEN Department of Biological Chemistry, The Hebrew University of Jerusalem, Jerusalem (Israel)

(Received October 19th, 1976)

SUMMARY Relative equilibrium constants ("affinity ratios") of complexes of bilirubin with a molar excess of charcoal-treated serum albumins from different species (human, bovine, rabbit and chicken) in aqueous solution, were estimated by circular dichroism measurements in the visible region at 26-27 °C, pH 7.4, and in the presence of 0.1 M NaCI. By variation of the mol ratios of the components of pairs of different bilirubin. serum albumin complexes showing circular dichroic bands of opposite sign, the apparent association constants of complexes of bilirubin with either human or chicken albumin were found to be greater by factors between 6 and 17 than those of bovine or rabbit albumins. The usefulness in the determination of affinity ratios is illustrated by the evaluation of single equilibrium constants of systems of high-ligand affinity from those of relatively lower affinity, the latter of which are more readily amenable to direct experimental measurement.

INTRODUCTION Previously, CD and light absorption spectra of bilirubin.serum albumin complexes, involving albumins from different species (human, bovine, goat, porcine, rabbit and chicken) were compared under similar conditions in aqueous solution [1, 2]. Since two large CD bands of opposite sign are recorded in most cases at neutral pH and at excess albumin in the visible region (400-500 rim), and since the sign and magnitude of the CD bands also differ for several complexes involving albumins from different species, it appears that CD is a very sensitive method for differentiation among these complexes [3, 1, 2]. In the present investigation, use is made of these large differences in ellipticity among complexes in order to assess the relative affinities of bilirubin for serum albumins from various species. Pairs of albumins have been selected in order to give CD bands of opposite sign in a given wavelength region for the two bilirubin.albumin complexes.

Abbreviations: CD, circular dichroism; HSA, human serum albumin; BSA, bovine serum albumin; RSA, rabbit serum albumin; CSA, chicken serum albumin.

65 MATERIALS AND METHODS A crystalline preparation of bilirubin from Sigma (Lot 73C-11418) was used (for content and formation of isomers, see ref. 1). Human serum albumin, crystallized (Lot 33) and rabbit serum albumin, crystallized (Lot 13), were purchased from Pentex. Bovine serum albumin, crystallized and lyophilized (Lot 83C-8090), and chicken serum albumin, Fraction V, powder (Lot 22C-3580), were obtained from Sigma. Other details, including evaluation of the purity and of the concentrations of bilirubin and of the proteins, preparation of solutions and instrumentation, were as described in preceding publications [4, 1, 2]. Measurement of the final solutions of the complexes at 26-27 °C were started about 20 min after their preparation. The CD band extrema in the visible region were recorded again after an additional 20 min, but in no case were any significant changes with time observed. The data measured were therefore considered to represent equilibrium values. Also, by changing the order of addition of the complex components (e.g. forming the bilirubin complex with each of the albumins prior to addition of the other), practically the same ellipticity values were obtained in each case, demonstrating again conditions of reversibility. Except for the data of Fig. 1, cells for the reference baseline contained distilled water, since there was practically no difference in baseline between water and aqueous protein solutions under the conditions used and in the wavelength range investigated (400-550 nm). Unbound bilirubin should not contribute measurably to the observed ellipticity [4, 1]. RESULTS A N D DISCUSSION

Typical visible-range CD spectra of complexes of bilirubin at a molar excess of the four different species of serum albumin investigated in the present work, are summarized in Fig. 1 (see also refs. 1 and 2). The CD band positions and magnitudes vary to some degree with the preparation of the albumin used [1, 4]. Of the two main bands of opposite sign observed for each bilirubin complex, the longer wavelength bands in the range of 460-480 nm give the largest differences in ellipticities between pairs of albumin complexes, having opposite signs in their CD bands and thus providing a sensitive measure for the relative content of a bilirubin, albumin complex. For any such pair of serum albumins (each designated A and B, respectively) the fraction complex of one componentfA, based on total bilirubin, is estimated at various concentrations of both proteins according to OX obs _ _

fA

--

01B

O)A -- OaB

(1)

where 0x is the ellipticity at a given (optimal) wavelength 2 and constant optical path length. The contributions to the ellipticity of each complex in a mixture are considered to be additive. Under the conditions used (see Table I), only negligible amounts of free bilirubin should be present, since each of the association constants of bilirubin.

66 I

ioo

I

/r\ /

/

/

I

t \ /

/

/ /

/ i E ~, ,-,

50

/

I / / .~

/

/

. .'... . .

.-"

\...)

~-/~///

.....

......~ ~ Human/

o ~ :

--

i\

Rabbit ,,

. ...... ",

,..$.

/

i

-

........",

~"-

......" / \ t

'

/,"

-5O ~

//

\/

'Ik,v/

iI

I qO0

i I

I I

\ Chickedk

-150

I

\ I

350

\~/

/I /

l

400

450

500

550

Wavelength [nm} Fig. 1. CD spectra of complexes of bilirubin with charcoal-treated [5] serum albumins from different species, as indicated in figure. Bilirubin, 2.5.10 -5 M; serum albumin, 5.0.10 -5 M; NaCI, 0.1 M; Tris.HCl buffer, 0.02 M; pH 7.4; temperature, 27.0 ± 0.5 °C. Reference solutions contained all components, except for the bilirubin. 0 is the observed ellipticity.

serum albumin complexes would probably be not smaller than about 106 M -1 (see below). From the ellipticity values observed at constant bilirubin concentration and at varying concentrations of each of the albumins, apparent relative association TABLE I AFFINITY RATIOS FOR PAIRS OF DIFFERENT BILIRUBIN--SERUM ALBUMIN COMPLEXES IN AQUEOUS SOLUTION Tris.HCl buffer, 0,02 M; pH 7.4 ± 0.1 ; NaCI, 0.1 M; temperature, 26.5 ± 0.5 °C. All albumins were charcoal treated [5]. Serum albumin A

B

Human Human Chicken Bovine

Bovine Chicken Rabbit Rabbit

* See Eqn. 2.

Average

,~(~Jc o)

K~,~*

(kcal/mol)

13 2.3 7.5 1.2

--1.5 --0.50 --1.2 --0.11

-4- 2 ± 0.3 ± 1 ± 0.15

67 constants KA.a, designated below as "affinity ratios", are calculated for each pair of albumins according to KA

fA ([B], -- [n]b) - - f s ([A]t - - [A]b)

KA'B - - ~

(2)

where the subscripts t and b refer to the concentration of total and bound protein, respectively. Eqn. 2 follows from a combination of simple mass laws for each complex having individual association constants Kt. Activity coefficients of unity are taken. Assuming 1:1 complexes between bilirubin and each albumin to predominate at an excess of the latter, the concentration of each bound albumin is taken to be equal to that of the bilirubin bound to it, e.g. [A]b ~fA[BR]t, where [BR]t is the total concentration of bilirubin. KA.B values obtained at pH 7.4 for the bilirubin complexes of four different pairs of serum albumins from various species, are summarized in Table I. The values obtained for each pair of albumins result as averages from 5 to 10 experiments, in which the mol ratios between bilirubin (either 2" 10 -6 o r 2-10 -s M) and the two albumins have been varied between 1:1:1, and 1:2:10 or 1:10:2, covering a range of about 0.1-0.8 in the fraction of the bilirubin complex of one component. It appears that with the charcoal-treated albumins, the apparent association constants of complexes involving either human or chicken serum albumin are greater by one order of magnitude (factors of about 7 and 13) than those involving either bovine or rabbit serum albumin. Analogous calculations based on the shorterwavelength CD band near 410--420 nm, lead to similar K values in most cases, although the errors are larger because of the smaller ellipticities of some of the systems. Sedimentation velocity experiments in the analytical ultracentrifuge of 1:1:1 complexes of all four systems investigated, indicated the absence of significant association between the proteins under these conditions. Considering the following scheme, in which the affinity ratios KHSA, BSA

BSA @ HSA KBSA, RSA

']["

q[,

(3)

KCSA,HSA

RSA ~ CSA KRSA, CSA

are arbitrarily defined in a clockwise direction, it can be shown that KHSA, BSA" KCSA, H S A " KRSA, CSA" KBSA, RSA :

1

(4)

Taking the appropriate average values for KA,R(see also Table I), the above products equal 1.0 for the charcoal-treated albumins. The corresponding products of the average values of the reciprocal ratios is 1.3. With all the errors and assumptions involved, the values obtained for the products may be considered to satisfy the conditions of Eqn. 4. The present data also allow an estimate of the affinity ratios KCSA,BS A and KHSA,RSA for pairs of complexes showing CD bands of equal sign. Using the experimentally determined values (Table I), it follows that KCSA,BSA= 6.2 i 1 and KrlSA, RSA = 17 ± 2.

68 Analogous measurements carried out with the same untreated albumins resulted in some eases in considerable differences between charcoal-treated and untreated pairs of albumins, e.g. KSSA.RSA ---- 4.2 and KHSA,CSA= 1.5, as compared to 1.2 and 2.3, respectively, for the constants of the systems involving treated albumins. Moreover, the product of Eqn. 4 gave unreasonable values for the untreated systems. It therefore appears that the charcoal-treated albumins provide more defined reference systems for bilirubin binding than the untreated albumins. This may be reasonable in view of the possibility of transfer of fatty acids and other small molecules from one albumin to the other in the binary untreated protein systems, causing possible changes in the binding characteristics of the latter. A molar excess of proteins over the ligand was present in all experiments reported. Provided bilirubin at one specific binding site only on each of two proteins is contributing to the observed optical activity, the affinity ratio KA,B will give the ratio KA/KBof the single constants. By independent determination of either KA or KB, all such constants in a cycle (Eqn. 3) could be estimated. In case more than one binding site for bilirubin on a single protein is involved, an apparent relative equilibrium constant will be measured in each case, depending on the type of binding (dependent or independent sites), the relative magnitude of the relevant constants and the relative contributions of the ligand at each site to the effect measured. The determination of relative equilibrium constants KA,B may be particularly useful where the high affinity of a certain system or other causes preclude direct determination of the free ligand concentration. By combination of such a system with another of considerably lower affinity for the ligand, where a single equilibrium constant can be measured more conveniently, the single constant of the higheraffinity system can now be estimated. The association constant for the high-affinity site of bovine serum albumin for bilirubin was 2.2.107 M -a [6] by albumin fluorescence quenching under conditions similar to those reported presently (similar data were obtained also at 36 °C with untreated bovine serum albumin by continuousflow experiments and by use of the peroxidase method [7]; for other relevant data, see ref. 7). Taking a value of 2.10 7 M -I for the bilirubin.bovine serum albumin high-affinity constant, the following values for the apparent association constants of other albumin systems can readily be estimated according to the relation given in Eqn. 2, using the data given in Table I, as well as the averages of the corresponding reciprocal affinity ratios: KnSA ~ 2' 108 M-~; Kcs A ----- 1" 108 M - I ; KRS A ----- 1.7" 107 M -1. This value for KnsA is larger than that obtained for the high-affinity site by the peroxidase method (1.4.108 M -a) obtained under somewhat different conditions [8]. The constant is higher by a factor of about 30 than that derived from ellipticity data by Beaven et al. [9]. However, the latter data were measured at pH 8.5. Determination of KnsA by fluorescence quenching gave a value of only 7.0.107 M -1 [6], while KRS A = 1.2" 107 M - 1 [6] is in better agreement with the above estimate. It is assumed in all above cases that small differences in the preparation of the proteins or in other conditions do not appreciably affect the comparison among equilibrium constants. We have observed that similar affinity ratios are obtained for human serum albumin/bovine serum albumin in the presence of either phosphate buffer [6] or Tris. HC1 buffer used presently.

69 REFERENCES 1 Harmatz, D. and Blauer, G. (1975) Arch. Biochem. Biophys. 170, 375-383 2 Blauer, G. and Harmatz, D. (1973) Jerusalem Symp. Quantum Chem. Biochem. (Bergmann, E. D. and Pullman, B., eds.), Vol. 5, pp. 709-714, Isr. Acad. Sci. Human., Jerusalem 3 Blauer, G. (1974) Structure and Bonding 18, 69-129 4 Blauer, G., Harmatz, D. and Snir, J. (1972) Biochim. Biophys. Acta 278, 68-88 5 Chen, R. F. (1967) J. Biol. Chem. 242, 173-181 6 Chen, R. F. (1973) in Fluorescence Techniques in Cell Biology ('lhaer, A. A. and Sernetz, M., eds.), pp. 273-282, Springer Verlag, Berlin 7 Faerch, T. and Jacobsen, J. (1975) Arch. Biochem. Biophys. 168, 351-357 8 Jacobsen, J. (1969) FEBS Lett. 5, 112-114 9 Beaven, G. H., D'Albis, A. and Gratzer, W. B. (1973) Eur. J. Biochem. 33, 500-510