Human Serum Albumin I. ON THE RELATIONSHIP OF FATTY ACID AND BILIRUBIN BINDING SITES AND THE NATURE OF FATTY ACID ALLOSTERIC EFFECTS-A MONOANIONIC SPIN LABEL STUDY*

specificity of the relationship between the sites and how fatty acid free spin label signal intensity in label discussion of the observed binding phenomena and their implica-tions.

This spin label is shown to bind to endogenous bilirubin binding sites of human serum albumin, as addition of bilirubin results in the increase of free spin label signal intensity and a concomitant decrease in bound spin label signal intensity.
The binding isotherm of the interaction of the spin label with human serum albumin in the form of a Scatchard plot shows that the binding of the first molar equivalent of bilirubin quantitatively displaces the spin label from its primary binding site. Graphical analysis of the binding isotherm shows that the high affinity bilirubin binding site has the capacity to bind 2 mol of spin label with different affinities.
This binding isotherm at 2 molar equivalents of bilirubin shows that the spin label also binds specifically to the secondary bilirubin binding sites. Previous studies (Soltys, B. J., and Hsia, J. C. (1977) J. Biol. Chem 252, 4043-4048) have shown that the binding of palmitate to serum albumin allosterically enhances the spin label binding. The present study indicates that the fatty acid effect is chain length-dependent.
Fatty acids of chain length of less than 10 carbon atoms competitively displace the binding of spin labels by serum albumin while fatty acids of longer chain length enhance the spin label binding allosterically.
The binding isotherm for the spin label albumin interaction in the presence of 1 to 4 molar equivalents of octanoate, laurate, and palmitate shows octanoate decreases the num-* This research was supported by Grant MT-4129 from the Medical Research Council of Canada, and by Grant PR-524 from the Ministry of Health of the Province of Ontario. This is the second report in a series aimed at the resolution of serum albumin structure and function.
The first report is Ref. 12 Fig. 2 in the form of a Scatchard plot.
The binding of GABA-DNB-SL to albumin has been revealed to be complex and a detailed quantitative evaluation of the number of specific GABA-DNB-SL binding sites, their respective affinity constants, and changes in these parameters upon the formation of complexes with other ligands is outside the scope of the present study.
Such an analysis will require adopting computerized fitting of binding data according to the stepwise equilibrium model (19,20). Nevertheless, qualitative information will be drawn from the binding isotherms presented here to advance the arguments presented. Fig. 2  3. Electron spin resonance spectra of the binding of 1 molar equivalent of GABA-DNB-SL to human plasma albumin in the absence of added octanoate (solid spectrum) and in the presence of a fatty acid:albumin ratio of 4 (dashed spectrum). The conditions were the same as in Fig. 1. A, normal spectral scan; 23, spectra run at high instrument sensitivity.
Serum Albumin Binding Specificity and Mechanisms line spectrum (Fig. 8A), representing an increase of free spin label. As in the case of bilirubin binding, the effect is accompanied by a decreased intensity of the outer broad line extrema of the immobilized spectrum (Fig. 33), consistent with a competitive inhibition of spin label binding. These spectra contrast with the effect of the long chain fatty acid, laurate, on the ESR spectra presented in Fig. 4. As in the ESR spectra with palmitate previously presented (121, spin label binding is enhanced (Fig. 4A) and this enhanced binding is accompanied by an increased intensity of and splitting between the outer broad line extrema of the immobilized spectrum (Fig. 4B). In that protein-bound spin label is sensitive to the motional freedom and polarity of the binding site microenvironment (24, 25), the latter changes in the immobilized spectrum are consistent with an allosteric effect by fatty acid. Fig. 5 presents the effect of 1 to 4 molar equivalents of fatty acids of varying chain length on the binding of the first molar equivalent of GABA-DNB-SL.
The biphasic curves indicate that fatty acids of chain length less than 10 carbon atoms inhibit GABA-DNB-SL binding, whereas longer chain length fatty acids exert an enhancement effect on spin label binding. The accompanying changes observed in the immobilized spectra are consistent with a competitive displacement of spin label by fatty acids of chain length less than 10 carbon atoms whereas, as in the spectra for palmitate (12) and for laurate, longer chain length fatty acids appear to allosterically perturb albumin binding sites (fatty acid chain length effect on the immobilized spectrum not presented). The effect of fatty acid unsa' uration with fatty acids of 18 carbon atoms chain length has also been examined, the results of which are presented in Fig. 6. The little apparent difference between these fatty acids suggests that the fatty acid confor- However, subtle changes in the immobilized spectra are seen (spectra not presented), suggesting differences at the molecular level.
The binding isotherms for the interaction of GABA-DNB-SL with albumin in the presence of 1 to 4 molar equivalents of octanoate, laurate, and palmitate are presented in Fig. 7 in the form of Scatchard plots. A few qualitative observations will be drawn from these data at the present time. First, the binding of 1 to 4 molar equivalents of octanoate (Fig. 7A)