Effectors of the regulatory protein acting on liver glucokinase: a kinetic investigation

In the absence of fructose 6-phosphate, the regulatory protein of rat liver glucokinase (hexokinase IV or D) inhibited this enzyme, though with a much (15-fold) lower potency than in the presence of a saturating concentration of fructose 6-phosphate. Evidence is provided that this inhibition is not due to contaminating fructose 6-phosphate. In the presence of regulatory protein, sorbitol 6-phosphate, a potent analog of fructose 6-phosphate, exerted a hyperbolic, partial inhibition on glucokinase, the degree of which increased with the concentration of regulatory protein. Plots of the reciprocal of the difference between the rates in the absence and in the presence of sorbitol 6-phosphate versus 1/[sorbitol 6-phosphate] at various concentrations of regulatory protein were linear, and demonstrated that the apparent affinity for sorbitol 6-phosphate increased with the concentration of regulatory protein. Plots of the reciprocal of the difference between 1/v in the presence and in the absence of sorbitol 6-phosphate versus 1/[sorbitol 6-phosphate] were also linear and crossed the axis at a value independent of the concentration of regulatory protein. Fructose 1-phosphate released the inhibition exerted by the regulatory protein in a hyperbolic fashion. The concentration of this effector required for a half-maximal effect increased linearly with the concentrations of sorbitol 6-phosphate and of regulatory protein. These results are consistent with a model in which the regulatory protein exists under two conformations, one form which binds inhibitors and glucokinase, and the other which binds activators, although not glucokinase. Sorbitol 6-phosphate, 2-deoxysorbitol 6-phosphate and mannitol 1-phosphate, all analogs of the open-chain configuration of fructose 6-phosphate, inhibited glucokinase in the presence of regulatory protein at lower concentrations than fructose 6-phosphate, whereas fixed analogs of the furanose form of fructose 6-phosphate were inactive or behaved as activators. This indicated that fructose 6-phosphate in its open-chain configuration is recognized by the regulatory protein. A series of compounds exerted an activating effect. These included, in order of decreasing potency: fructose 1-phosphate, psicose 1-phosphate, ribitol 5-phosphate, analogs of fructose 1-phosphate and of ribitol 5-phosphate and, at much higher concentrations, inorganic phosphate.     

Mechanism of the stimulatory effect of a potassium-rich medium on the phosphorylation of glucose in isolated rat hepatocytes.

We have investigated the mechanism by which the replacement of a Na(+)-rich medium by a K(+)-rich medium causes an increase in the apparent affinity of glucokinase (hexokinase IV or D) for glucose in isolated hepatocytes [Bontemps, F., Hue, L. & Hers, H. G. (1978) Biochem. J. 174, 603-611]. The stimulatory effect of a K(+)-rich medium on the rate of glucose phosphorylation, as assessed by the release of tritium from [2-3H]glucose, was only partially additive with the effect of fructose, suggesting that it was also due to a decrease in the inhibition exerted on glucokinase by its regulatory protein. Measurements of metabolites indicated that the effect of the K(+)-rich medium was neither due to the formation of fructose 1-phosphate, nor to changes in the concentrations of fructose 6-phosphate or Pi, two other effectors of the regulatory protein. Replacement of Na+ by K+ in the medium resulted in a time-dependent and dose-dependent increase in cell volume that paralleled the changes in the rate of detritiation observed at 5 mM glucose. The water and chloride contents, estimated using radiolabelled compounds, were threefold and tenfold higher, respectively, in K+ cells than in Na+ cells, and the intracellular Cl- concentration about threefold higher (94 versus 29 meq/l). The effects of the K(+)-rich medium on cell volume, Cl- concentration and rate of detritiation were greatly reduced by including 80 mM trehalose or sucrose in the medium at the start of the incubation. Addition of trehalose to cells incubated for 45-50 min in the K(+)-rich medium caused an immediate decrease in cell volume whereas the rate of detritiation and the Cl- concentration underwent a transient increase followed by a decrease. Replacement of KCl by KBr, potassium acetate or potassium trichloroacetate in the K(+)-rich medium resulted in different relationships between cell volume and the rate of detritiation, in agreement with the differential effect of these salts on the activity of purified glucokinase assayed in the presence of regulatory protein. From these results we conclude that the increase in the activity of glucokinase induced by a KCl-rich medium is at least partly due to an increase in the concentration of Cl-, which relieves the inhibition exerted by the regulatory protein on purified glucokinase.     

Comparative study of tyrosinases from different sources: relationship between halide inhibition and the enzyme active site

The inhibition of tyrosinases from frog epidermis (Rana esculenta ridibunda), mushroom (Agaricus bisporus) and Harding-Passey mouse melanoma by halides is compared. In all cases, the inhibition is pH dependent, increasing when the pH decreases. The order of inhibition is I- greater than Br- greater than Cl- much greater than F- for frog epidermis tyrosinase, F- greater than I- greater than Cl- greater than Br- for mushroom tyrosinase and F- greater than Cl- much greater than Br- greater than I- for the mouse melanoma enzyme. These results are discussed in terms of the active site accessibility to exogenous ligands. The activation energies of the enzyme-catalysed L-dopa oxidation were also calculated, being the values 6.86, 17.01 and 20.25 kcal/mol for frog epidermis, mushroom and Harding-Passey mouse melanoma, respectively. A relationship between these values and the evolutionary adaptation of these enzymes is proposed.     

Interactions of bromide, iodide, and fluoride with the pathways of chloride transport and diffusion in human neutrophils

Isolated human neutrophils possess three distinct pathways by which Cl- crosses the plasma membrane of steady state cells: anion exchange, active transport, and electrodiffusion. The purpose of the present work was to investigate the selectivity of each of these separate processes with respect to other external halide ions. (a) The bulk of total anion movements represents transport through an electrically silent anion-exchange mechanism that is insensitive to disulfonic stilbenes, but which can be competitively inhibited by alpha-cyano-4-hydroxycinnamate (CHC; Ki approximately 0.3 mM). The affinity of the external translocation site of the carrier for each of the different anions was determined (i) from substrate competition between Cl- and either Br-, F-, or I-, (ii) from trans stimulation of 36Cl- efflux as a function of the external concentrations of these anions, (iii) from changes in the apparent Ki for CHC depending on the nature of the replacement anion in the bathing medium, and (iv) from activation of 82Br- and 125I- influxes by their respective ions. Each was bound and transported at roughly similar rates (Vmax values all 1.0-1.4 meq/liter cell water.min); the order of decreasing affinities is Cl- greater than Br- greater than F- greater than I- (true Km values of 5, 9, 23, and 44 mM, respectively). These anions undergo 1:1 countertransport for internal Cl-. (b) There is a minor component of total Cl- influx that constitutes an active inward transport system for the intracellular accumulation of Cl- [( Cl-]i approximately 80 meq/liter cell water), fourfold higher than expected for passive distribution. This uptake is sensitive to intracellular ATP depletion by 2-deoxy-D-glucose and can be inhibited by furosemide, ethacrynic acid, and CHC, which also blocks anion exchange. This active Cl- uptake process binds and transports other members of the halide series in the sequence Cl- greater than Br- greater than I- greater than F- (Km values of 5, 8, 15, and 41 mM, respectively). (c) Electrodiffusive fluxes are small. CHC-resistant 82Br- and 125I- influxes behave as passive leak fluxes through low-conductance ion channels: they are nonsaturable and strongly voltage dependent. These anions permeate the putative Cl- channel in the sequence I- greater than Br- greater than Cl- with relative permeability ratios of 2.2:1.4:1, respectively, where PCl approximately 5 X 10(-9) cm/s.     

A protein from rat liver confers to glucokinase the property of being antagonistically regulated by fructose 6-phosphate and fructose 1-phosphate

At a concentration of 1 mM, fructose 1-phosphate stimulated about twofold, and glucose 6-phosphate inhibited by about 30%, the phosphorylation of 5 mM glucose in high-speed supernatants prepared from rat liver or from isolated hepatocytes, but did not affect, or barely so, the activity of a partially purified preparation of glucokinase. Anion-exchange chromatography of liver extracts separated glucokinase from a fructose-6-phosphate-sensitive and fructose-1-phosphate-sensitive inhibitor of that enzyme. This inhibitor could be further purified by chromatography on phospho-Ultrogel. It was destroyed by trypsin and was heat-labile. It inhibited glucokinase competitively with respect to glucose and its inhibitory effect was greatly reinforced by fructose 6-phosphate although not by glucose 6-phosphate. Fructose 1-phosphate relieved the enzyme of the inhibitory effect of the regulator and antagonised the effect of fructose 6-phosphate in a competitive manner. It is concluded that the regulator plays a role in the physiological control of the activity of glucokinase, particularly with respect to the stimulatory effect of fructose in isolated hepatocytes (see preceding paper in this journal).     

Analysis of the cooperativity of human beta-cell glucokinase through the stimulatory effect of glucose on fructose phosphorylation

Using overexpressed Escherichia coli sorbitol-6-phosphate dehydrogenase to monitor fructose 6-phosphate formation, we found that the stimulation of fructose phosphorylation by glucose was reduced in two human beta-cell glucokinase mutants with a low Hill coefficient or when the activity of wild type glucokinase was decreased by replacing ATP with poorer nucleotide substrates. Mutation of two other residues, neighboring glucose-binding residues in the catalytic site, also reduced the affinity for glucose as a stimulator of fructose phosphorylation. Among a series of glucose analogs, only 3, all substrates of glucokinase, stimulated fructose phosphorylation; other analogs were either inactive or inhibited glucokinase. Glucose increased the apparent affinity for inhibitors that are glucose analogs but not for the glucokinase regulatory protein or palmitoyl-CoA. These data indicate that the stimulatory effect of glucose on fructose phosphorylation reflects the positive cooperativity for glucose and is mediated by binding of glucose to the catalytic site. They support models involving the existence of two slowly interconverting conformations of glucokinase that differ through their affinity for glucose and for glucose analogs. We show by computer simulation that such a model can account for the kinetic properties of glucokinase, including the differential ability of mannoheptulose and N-acetylglucosamine to suppress cooperativity (Agius, L., and Stubbs, M. (2000) Biochem. J. 346, 413-421).     

The mechanism by which rat liver glucokinase is inhibited by the regulatory protein

The fructose-6-phosphate-sensitive and fructose-1-phosphate-sensitive protein that inhibits rat liver glucokinase [Van Schaftingen, E. (1989) Eur. J. Biochem. 179, 179-184] was purified close to homogeneity by a procedure involving poly(ethyleneglycol) precipitation, chromatography on anion-exchangers and hydroxylapatite, gel filtration and chromatography on Mono S, a cation exchanger. In the last chromatographic step, the regulatory protein coeluted with a 62 kDa peptide. From the elution volume of the gel-filtration column a molecular mass of 60 kDa was determined, allowing the conclusion that the regulator is a monomer. The decrease in the apparent affinity of glucokinase for glucose, which the regulator induced, disappeared upon separation of the two proteins. Furthermore, the regulator did not appear to catalyse the formation of a heat-stable or a trypsin-resistant inhibitor of glucokinase. Finally, the inhibition exerted by the regulatory protein reached a steady value 1-2 min after the addition of the regulator. These results indicate that the regulator does not act by causing a covalent modification of glucokinase nor by catalysing the formation of a low-molecular-mass inhibitor. Raising the concentration of glucokinase in the assay from 6 mU/ml to 120 mU/ml caused a 2.5-fold increase in the concentration of regulator required to half-maximally inhibit the enzyme. The apparent mass of glucokinase, as determined by centrifugation in isokinetic sucrose gradient, was 55 kDa, and this value was unaffected by the separate presence of fructose 6-phosphate or of the regulatory protein. However, the apparent mass of the enzyme increased to 105 kDa when glucokinase was centrifuged together with both fructose 6-phosphate and the regulatory protein, although not when fructose 1-phosphate was also present. Conversely, the presence of glucokinase increased the apparent molecular mass of the regulator in the presence of fructose 6-phosphate. From these results, it is concluded that the regulatory protein inhibits glucokinase by forming a complex with this enzyme in the presence of fructose 6-phosphate, and that fructose 1-phosphate antagonises this inhibition by preventing the formation of the complex.     

Fructose 1-phosphate and the regulation of glucokinase activity in isolated hepatocytes

Fructose 1-phosphate kinase was partially purified from Clostridium difficile and used to develop specific assays of fructose 1-phosphate and fructose. The concentration of fructose 1-phosphate was below the detection limit of the assay (25 pmol/mg protein) in hepatocytes incubated in the presence of glucose as sole carbohydrate. Addition of fructose (0.05-1 mM) caused a concentration-dependent and transient increase in the fructose 1-phosphate content. Glucagon (1 microM) and ethanol (10 mM) caused a severalfold decrease in the concentration of fructose 1-phosphate in cells incubated with fructose, whereas the addition of 0.1 microM vasopressin or 10 mM glycerone, or raising the concentration of glucose from 5 mM to 20 mM had the opposite effect. All these agents caused changes in the concentration of triose phosphates that almost paralleled those of the fructose 1-phosphate concentration. Sorbitol had a similar effect to fructose in causing the formation of fructose 1-phosphate. D-Glyceraldehyde was much less potent in this respect than the ketose and its effect disappeared earlier. The effect of D-glyceraldehyde was reinforced by an increase in the glucose concentration and decreased by glucagon. Both fructose and D-glyceraldehyde stimulated the phosphorylation of glucose as estimated by the release of 3H2O from [2-3H]glucose, but the triose was less potent in this respect than fructose and its effect disappeared earlier. Glucagon and ethanol antagonised the effect of low concentrations of fructose or D-glyceraldehyde on the detritiation of glucose. These results support the proposal that fructose 1-phosphate mediates the effects of fructose, D-glyceraldehyde and sorbitol by relieving the inhibition exerted on glucokinase by a regulatory protein.     

Regulation of glucokinase by a fructose-1-phosphate-sensitive protein in pancreatic islets

In the post-microsomal supernatant of pancreatic islets, prepared from fasted or fed rats, D-fructose 1-phosphate increased the activity of glucokinase by 20-30% as measured in the presence of D-glucose 6-phosphate and D-fructose 6-phosphate. Such an activation was less marked than that found in liver extracts. The islet cytosol was also found to inhibit purified liver glucokinase, and this effect was antagonized by D-fructose 1-phosphate. In the presence of hexose 6-phosphates, partially purified islet glucokinase was inhibited by the hepatic glucokinase regulatory protein in a D-fructose-1-phosphate-sensitive manner. In intact islets, D-glyceraldehyde stimulated the generation of 14C-labelled D-fructose 1-phosphate from D-[U-14C]glucose and increased the production of 3H2O from D-[5-3H]glucose. These findings suggest that the activity of glucokinase in islet cells may be regulated by a protein mediating the antagonistic effects of D-fructose 6-phosphate and D-fructose 1-phosphate in a manner qualitatively similar to that operating in hepatocytes, but with lower efficiency.     

Inhibition of glucokinase translocation by AMP-activated protein kinase is associated with phosphorylation of both GKRP and 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase

The rate of glucose phosphorylation in hepatocytes is determined by the subcellular location of glucokinase and by its association with its regulatory protein (GKRP) in the nucleus. Elevated glucose concentrations and precursors of fructose 1-phosphate (e.g., sorbitol) cause dissociation of glucokinase from GKRP and translocation to the cytoplasm. In this study, we investigated the counter-regulation of substrate-induced translocation by AICAR (5-aminoimidazole-4-carboxamide-1-beta-D-ribofuranoside), which is metabolized by hepatocytes to an AMP analog, and causes activation of AMP-activated protein kinase (AMPK) and depletion of ATP. During incubation of hepatocytes with 25 mM glucose, AICAR concentrations below 200 microM activated AMPK without depleting ATP and inhibited glucose phosphorylation and glucokinase translocation with half-maximal effect at 100-140 microM. Glucose phosphorylation and glucokinase translocation correlated inversely with AMPK activity. AICAR also counteracted translocation induced by a glucokinase activator and partially counteracted translocation by sorbitol. However, AICAR did not block the reversal of translocation (from cytoplasm to nucleus) after substrate withdrawal. Inhibition of glucose-induced translocation by AICAR was greater than inhibition by glucagon and was associated with phosphorylation of both GKRP and the cytoplasmic glucokinase binding protein, 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (PFK2) on ser-32. Expression of a kinase-active PFK2 variant lacking ser-32 partially reversed the inhibition of translocation by AICAR. Phosphorylation of GKRP by AMPK partially counteracted its inhibitory effect on glucokinase activity, suggesting altered interaction of glucokinase and GKRP. In summary, mechanisms downstream of AMPK activation, involving phosphorylation of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase and GKRP are involved in the ATP-independent inhibition of glucose-induced glucokinase translocation by AICAR in hepatocytes.     

A novel glucokinase regulator in pancreatic beta cells: precursor of propionyl-CoA carboxylase beta subunit interacts with glucokinase and augments its activity

A glucokinase regulatory protein has been reported to exist in the liver, which suppresses enzyme activity in a complex with fructose 6-phosphate, whereas no corresponding protein has been found in pancreatic beta cells. To search for such a protein in pancreatic beta cells, we screened for a cDNA library of the HIT-T15 cell line with the cDNA of glucokinase from rat islet by the yeast two hybrid system. We detected a cDNA encoding the precursor of propionyl-CoA carboxylase beta subunit (pbetaPCCase), and glutathione S-transferase pull-down assay illustrated that pbetaPCCase interacted with recombinant rat islet glucokinase and with glucokinase in rat liver and islet extracts. Functional analysis indicated that pbetaPCCase decreased the K(m) value of recombinant islet glucokinase for glucose by 18% and increased V(max) value by 23%. We concluded that pbetaPCCase might be a novel activator of glucokinase in pancreatic beta cells.     

Kinetic study of the interaction between frog epidermis tyrosinase and chloride

The effect of halide ions on frog epidermis tyrosinase has been characterized with the trypsin-activated enzyme. At pH 7, the order of inhibition is I- greater than Br- greater than Cl- greater than F-. Chloride, the most extensively studied halide, shows a competitive pattern with respect to the substrate, L-DOPA. Inhibition is strongly pH-dependent, with a pKa of 6.12 for the responsible protonatable group. Other kinetic constants are also calculated using a novel approach. The mechanism of interaction between chloride and the enzyme is discussed, and a model is proposed in which chloride interferes the tyrosinase activity by displacing a catalytically important ligand, probably a histidine residue of the side-chain, from the copper at the enzyme-active site.     

Ionic selectivity of volume-sensitive currents in human epithelial cells.

The ion selectivity of swelling-activated Cl- currents has been investigated in three different human epithelial cell lines, two derived from the airway epithelium (9HTEo- and CFNPE9o-) and one from a colon carcinoma (T84). The relative permeability of volume-sensitive currents with respect to Cl- is: I- (1.19) greater than NO3- (1.07) approximately Br-(1.05) greater than Cl-(1.0) greater than F-(0.5) approximately HCO3-(0.48) greater than isethionate(0.28) greater than aspartate (0.14) approximately gluconate(0.13) approximately SO4(2-)(0.12). This type of ion selectivity is similar to that described for depolarization-activated outwardly rectifying Cl- channels found in epithelial cells.     

Halide permeation through 10 pS and 20 pS anion channels in human airway epithelial cells.

Halide permeability sequences were obtained from reversal potential measurements of single-channel currents through 10 pS and 20 pS anion channels in human airway epithelial cells. The sequences obtained were Cl- greater than I- greater than Br- greater than or equal to F- for the 10 pS channel and Cl- greater than I- greater than or equal to Br- greater than or equal to F- for the 20 pS channel. However, the permeability differences were not large, the greatest being 0.66 for the ratio of fluoride to chloride permeability in the 20 pS channel. Single-channel currents were also measured with solutions of constant halide concentration but varying ratios of chloride to fluoride ions. An anomalous mole fraction effect was observed for the 20 pS channel but not for the 10 pS channel, suggesting that the former is a multi-ion channel. Comparison of the halide permeability sequences of these two channels with those of whole-cell currents in other epithelial cells does not support their involvement in any of the known whole-cell epithelial currents.     

Effects of anions on the activation thermodynamics and fluorescence emission spectrum of alkaline phosphatase: evidence for enzyme hydration changes during catalysis

The effect of anions on the thermodynamic activation functions for a model enzyme, calf intestinal alkaline phosphatase (EC 3.1.3.1), have been studied in order to examine the role of protein hydration changes in establishing the energetics of enzyme catalysis. The influences of these anions on the activation volume (delta V) and activation free energy (delta G) reflected clear Hofmeister (lyotropic) series effects, in the order F- greater than Cl- greater than Br- greater than I- (order of increasing salting-out potential). A pronounced covariation was observed between the influences of these anions on Vmax, which is proportional to delta G, and on the negative activation volume of the reaction. Fluoride was able to counteract the influences of Br- and I- on both Vmax and delta V when combinations of these anions were employed. The effects of Br- and I- on Vmax and delta V were more pronounced at lower temperatures. The control delta V was increasingly negative at reduced temperatures. The effects of the neutral salts and propanol on delta V and delta G, as well as the effects of salting-in anions on the activation enthalpy and the negative activation entropy of the reaction, are consistent with a model which proposes that peptide groups or polar side chains on the native enzyme exergonically increase their exposure to solvent during the catalytic activation event. These conclusions are in accord with the known free energy, enthalpy, entropy, and volume changes which occur when model peptide groups are transferred between water and concentrated salt solutions. Consistent with the kinetic results, the fluorescence emission wavelength maximum of alkaline phosphatase increased in the presence of anions in the order F- greater than Cl- greater than Br- greater than I-. The salting-out ion (F-) and the salting-in ions (Br- and I-) shifted lambda max in different directions, and these lambda max shifts could be counterbalanced by using equimolar combinations of salting-in and salting-out anions. Control experiments with a model compound, N-acetyltryptophanamide, showed that the spectra shifts caused by the salts did not result solely from differential quenching by the anions of the solvent-exposed tryptophan(s) on the enzyme. Hofmeister additivity phenomena indicated that the solvent is at the basis of these salt-induced enzyme structural changes. It is concluded that changes in protein solvation during enzymic reactions contribute significantly to the thermodynamic activation parameters in both the native and the salt-perturbed enzyme.     

Stoichiometry of slow binding of palmitoyl-CoA to liver glucokinase

The interaction of palmitoyl-CoA with porcine glucokinase was studied by the gel permeation technique. The finding that glucokinase "bound" up to 60 molecules was unexpected from the specific inhibition of rat glucokinase by long chain acyl-CoA (Tippett & Neet, J. Biol. Chem. (1982) 287, 12839-12845). Sephacryl S-200 gel filtration in the presence of palmitoyl-CoA demonstrated a protein peak without enzyme activity that was eluted earlier than the active enzyme peak, indicating a large molecular weight shift for the inactivated enzyme form and confirming a large number (greater than or equal to 30) of associated palmitoyl-CoA molecules. The binding was also verified by analyzing the absorption characteristics of the inactivated enzyme peak. In the presence of glycerol, the size of the inactivated peak greatly decreased, but the separation between the two peaks remained unchanged. Therefore, the amphiphile bound predominantly to the inactive enzyme and not to the active form, suggesting that the rapid inhibitory interactions between palmitoyl-CoA and glucokinase previously observed are specific. Parallel enzyme activity studies showed that in the time range of the column experiments (4-20 h), both the rat and pig enzyme were greatly inactivated (greater than 90%) in the presence of palmitoyl-CoA (15 microM) in the absence of glycerol. This slow inactivation is different from the immediate specific inhibition previously reported and depends on both enzyme and palmitoyl-CoA concentrations. The presence of up to 20% glycerol slowed this inactivation process. These results demonstrated that even below the critical micelle concentration, partial inactivation of glucokinase occurs in the presence of palmitoyl-CoA over a long period of time.     

Anion binding properties of human serum albumin from halide ion quadrupole relaxation.

The nuclear magnetic quadrupole relaxation enhancement of 35Cl-, 81Br-, and 12I- anions on binding to human serum albumin has been studied under conditions of variable protein and anion concentration and also in the presence of simple inorganic, amphiphilic, and complex anions which compete with the halide ions for the protein anion binding sites. Two classes of anion binding sites with greatly different binding constans were identified. Experiments at variable halide ion concentration were employed to determin the Cl- and I- binding constants. By means of 35 Cl nuclear magnetic resonance (NMR) the relative affinity for different anions was determined by competition experiments for both the strong and the weak anion binding sites. Anion binding follows the sequence SO42- smaller than F- smaller than CH3COO- smaller than Ci- smaller Br- smaller than NO3- smaller than I- smaller than ClO4- smaller than SCN- smaller than Pt(CN)42- smaller than Au(CN)2- smaller than CH3(CH2)11OSO3- for the high affinity sites, and the sequence SO42- congruent to F- congruent to Cl- smaller CH3COO- smaller than NO3- smaller than Br- smaller than I- smaller than ClO4- smaller than SCN- for the low affinity sites. These series are nearly identical with the well-known lyotropic series. Consequently, those effects of anions on proteins described by the lyotropic series can be correlated with the affinities of the anions for binding to the protein. The data suggest that the physical nature of the interaction is the same for both types of biding sites, and that the differences in affinity between different binding sites must be explained in terms of tertiary structure. Analogous experiments performed using 127I- quadrupole relaxation gave results very similar to those obtained with 35Cl-. A comparison between the Cl-, Br- and I- ions revealed that, as a result of the increasing affinity for the weak anion binding sites in the series Cl- smaller than Br- smaller than I-, Cl- is much more useful as a probe for the specific anion binding sites than the other two halide ions. The findings with human serum albumin in this and other respects are probably of general relevance in studies of protein-anion interactions. In addition to competition experiments, the magnitude of the relaxation rate is also discussed. Line broadening not related to anion binding to the protein is found to be small. A comparison of transverse and longitudinal 35Cl relaxation rates gives a value for the quadrupole coupling constant of the high affinity sites in good agreement with a calculated coupling constant assuming anion binding to arginine.