Evidence for nonacetylcholinesterase targets of organophosphorus nerve agent: supersensitivity of acetylcholinesterase knockout mouse to VX lethality.

The possibility that organophosphate toxicity is due to inhibition of targets other than acetylcholinesterase (AChE, EC 3.1.1.7) was examined in AChE knockout mice. Mice (34-55 days old) were grouped for this study, after it was determined that AChE, butyrylcholinesterase (BChE), and carboxylesterase activities had reached stable values by this age. Mice with 0, 50, or 100% AChE activity were treated subcutaneously with the nerve agent VX. The LD50 for VX was 10 to 12 microg/kg in AChE-/-, 17 microg/kg in AChE+/-, and 24 microg/kg in AChE+/+ mice. The same cholinergic signs of toxicity were present in AChE-/- mice as in wild-type mice, even though AChE-/- mice have no AChE whose inhibition could lead to cholinergic signs. Wild-type mice, but not AChE-/- mice, were protected by pretreatment with atropine. Tissues were extracted from VX-treated and untreated animals and tested for AChE, BChE, and acylpeptide hydrolase activity. VX treatment inhibited 50% of the AChE activity in brain and muscle of AChE+/+ and +/- mice, 50% of the BChE activity in all three AChE genotypes, but did not significantly inhibit acylpeptide hydrolase activity. It was concluded that the toxicity of VX must be attributed to inhibition of nonacetylcholinesterase targets in the AChE-/- mouse. Organophosphorus ester toxicity in wild-type mice is probably due to inhibition or binding to several proteins, only one of which is AChE.

Predicted Michaelis-Menten complexes of cocaine-butyrylcholinesterase. Engineering effective butyrylcholinesterase mutants for cocaine detoxication.

Butyrylcholinesterase (BChE) is important in cocaine metabolism, but it hydrolyzes (-)-cocaine only one-two thousandth as fast as the unnatural (+)-stereoisomer. A starting point in engineering BChE mutants that rapidly clear cocaine from the bloodstream, for overdose treatment, is to elucidate structural factors underlying the stereochemical difference in catalysis. Here, we report two three-dimensional Michaelis-Menten complexes of BChE liganded with natural and unnatural cocaine molecules, respectively, that were derived from molecular modeling and supported by experimental studies. Such complexes revealed that the benzoic ester group of both cocaine stereoisomers must rotate toward the catalytic Ser(198) for hydrolysis. Rotation of (-)-cocaine appears to be hindered by interactions of its phenyl ring with Phe(329) and Trp(430). These interactions do not occur with (+)-cocaine. Because the rate of (-)-cocaine hydrolysis is predicted to be determined mainly by the re-orientation step, it should not be greatly influenced by pH. In fact, measured rates of this reaction were nearly constant over the pH range from 5.5 to 8.5, despite large rate changes in hydrolysis of (+)-cocaine. Our models can explain why BChE hydrolyzes (+)-cocaine faster than (-)-cocaine, and they suggest that mutations of certain residues in the catalytic site could greatly improve catalytic efficiency and the potential for detoxication.

The active site of human paraoxonase (PON1).

Ideally we would like to treat people exposed to nerve agents with an enzyme that rapidly destroys nerve agents. The enzymes considered for such a role include human butyrylcholinesterase (BChE), acetylcholinesterase (AChE), carboxylesterase and paraoxonase (PON1). Success has been achieved in endowing BChE with the ability to hydrolyze organophosphates. The G117H mutant of BCHE hydrolyzes sarin and VX, whereas the double mutant G117H/E197Q hydrolyzes soman (Millard et al. Biochemistry 1995; 34: 15925-15933; 1998; 37: 237-247). However, the rates of organophosphate hydrolysis are slow and a faster organophosphate hydrolase is being sought. Native PON1 hydrolyzes paraoxon with a catalytic efficiency, of 2.4 x 10(6) M(-1) x min(-1), and our goal is to improve the organophosphate hydrolase activity of PON1. To achieve this we need to identify the amino acids in the active site of PON1. Using site-directed mutagenesis and expression in human 293T cells, we have identified the following eight amino acids as being essential to PON1 activity: W280, H114, H133, H154, H242, H284, E52 and D53. Fluorescence of PON1 complexed to terbium ion shows that at least one tryptophan is close to the calcium binding site.

Kinetic mechanism of DIDS binding to band 3 (AE1) in human erythrocyte membranes.

Stilbenedisulfonates (S) are used widely in cell biology as competitive inhibitors of anion exchange, but the mechanism of competition is not resolved. Resolution requires understanding the detailed steps in the reaction of stilbenedisulfonates with various anion-exchange proteins. Studies on the reversible binding of DBDS (4,4'-dibenzamido-2,2'-stilbenedisulfonate) and H2DIDS (4,4'-diisothiocyanatodihydro-2,2'-stilbenedisulfonate) to erythrocyte band 3 (B) have shown biphasic kinetic time courses at 25 degrees C. Yet, results for the reversible binding of DIDS (4,4'-diisothiocyanato-2,2'-stilbenedisulfonate) are controversial. One recent report has shown monophasic kinetics, in experiments performed at 0 degrees C, and at a single, very low concentration of DIDS (0.1 microM). Studies are presented which attempt to reconcile these recent findings with the other kinetic data in the literature. We measure the kinetics of DIDS reversible binding to band 3, over a wide DIDS concentration range. In addition, the time course for DIDS binding to band 3 at 0 degrees C is compared with that at 25 degrees C. The results show biphasic binding kinetics at both 0 and 25 degrees C, and they are consistent with expectations for a two-step binding mechanism (S + B <--> SB <--> SB*). Furthermore, computer-assisted model simulation studies reveal that monophasic DIDS binding kinetics are generated by a two-step mechanism, when calculations are performed at 0.1 microM DIDS and 0 degrees C. Under these conditions the initial binding step in the two-step reaction becomes rate limiting. We conclude that the two-step binding mechanism best describes stilbenedisulfonate binding to band 3 and that the observation of monophasic kinetics at low concentrations of DIDS, while valid, is not mechanistically discriminating, since both one-step and two-step mechanisms can yield the same result.

Abundant tissue butyrylcholinesterase and its possible function in the acetylcholinesterase knockout mouse.

We have described recently an acetylcholinesterase (AChE) knockout mouse. While comparing the tissue distribution of AChE and butyrylcholinesterase (BChE), we found that extraction buffers containing Triton X-100 strongly inhibited mouse BChE activity. In contrast, buffers with Tween 20 caused no inhibition of BChE. Conventional techniques grossly underestimated BChE activity by up to 15-fold. In Tween 20 buffer, the intestine, serum, lung, liver, and heart had higher BChE than AChE activity. Only brain had higher AChE than BChE activity in AChE +/+ mice. These findings contradict the dogma, based mainly on observations in Triton X-100 extracts, that BChE is a minor cholinesterase in animal tissues. AChE +/- mice had 50% of normal AChE activity and AChE -/- mice had none, but all mice had similar levels of BChE activity. BChE was inhibited by Triton X-100 in all species tested, except rat and chicken. Inhibition was reversible and competitive with substrate binding. The active site of rat BChE was unique, having an arginine in place of leucine at position 286 (human BChE numbering) in the acyl-binding pocket of the active site, thus explaining the lack of inhibition of rat BChE by Triton X-100. The generally high levels of BChE activity in tissues, including the motor endplate, and the observation that mice live without AChE, suggest that BChE has an essential function in nullizygous mice and probably in wild-type mice as well.

Determination of the DNA sequences of acetylcholinesterase and butyrylcholinesterase from cat and demonstration of the existence of both in cat plasma.

Cat serum contains 0.5 mg/L of butyrylcholinesterase (BChE, EC 3.1.1. 8) and 0.3 mg/L of acetylcholinesterase (AChE, EC 3.1.1.7); this can be compared with 5 mg/mL and < 0.01 mg/L, respectively, in human serum. Cat BChE differed from human BChE in the steady-state turnover of butyrylthiocholine, having a 3-fold higher k(cat) and 2-fold higher K(m) and K(ss) values. Sequencing of the cat BCHE cDNA revealed 70 amino acid differences between cat and human BChE, three of which could account for these kinetic differences. These amino acids, which were located in the region of the active site, were Phe398Ile, Pro285Leu, and Ala277Leu (where the first amino acid was found in human and the second in cat). Sequencing genomic DNA for cat and human ACHE demonstrated that there were 33 amino acid differences between the cat and human AChE enzymes, but that there were no differences in the active site region. In addition, a polymorphism in intron 3 of the human ACHE gene was detected, as well as a silent polymorphism at Y116 of the cat ACHE gene.

Tryptophan residue(s) as major components of the human serum paraoxonase active site.

Serum paraoxonase (PON1, EC 3.1.8.1.) is a high density lipid- (HDL)-associated, calcium-dependent enzyme whose 3D structure, active site residues and physiological substrates are not known. The kinetic parameters k(cat) and Km (relative to k(cat) and Km of the wild-type), determined with four substrates (phenylacetate, paraoxon, diazoxon and chlorpyrifosoxon) were less than 1, and more than 100% for the W280A and W280F mutant enzymes, respectively. These results indicated that the aromatic/hydrophobic character of the amino acid in position 280 is essential for PON1 activity. In this study, we investigated whether this aromatic residue is in the PON1 active site. Group-specific labelling studies with N-bromosuccinimide, an oxidative agent of tryptophan, strongly suggested that one or several Trp could be in the active site of PON1 but we could not conclude either on the specificity of the labelling reaction or on the number of oxidized Trp. However, although PON activity was not altered by the hydrophilic tryptophan-modifying reagent 2-hydroxy-5-nitrobenzyl chloride (NBC), it was significantly reduced by the p-nitrophenylacetate analog 2-acetoxy-5-nitrobenzyl chloride (ANBC), whose hydrolysis by PON1 generated NBC in the active site. Moreover, since at least one calcium ion is present in the PON catalytic site, we attempted to probe the metal local environment using the calcium analog terbium. The luminescence spectrum of the PON terbium complex exhibited an emission peak at 545 nm characteristic of an aromatic residue (Trp and/or Tyr)-terbium interaction. In conclusion, both the results obtained with the mechanism-based inhibitor of PON1 (ANBC) and the calcium-binding site luminescent probe terbium support the hypothesis of the presence of at least one Trp residue in the PON1 active site. Trp residue(s) may be involved in the binding of aromatic substrates.

Identification of residues essential for human paraoxonase (PON1) arylesterase/organophosphatase activities.

Human serum paraoxonase (PON1) is a calcium-dependent organophosphatase. To identify residues essential for PON1 activity, we adopted complementary approaches based on chemical modification and site-directed mutagenesis. To detect 45Ca2+ binding to native and chemically modified PON1, we performed nondenaturating gel electrophoresis. The environment of calcium-binding sites was probed using the Ca2+ analogue, terbium. Tb3+ binds to calcium-binding sites as shown by displacement of 45Ca2+ by Tb3+. Binding of Tb3+ is accompanied by a complete loss of enzyme activity. PON1 chemical modification with the Trp-selective reagent, N-bromosuccinimide, and the Asp/Glu-selective, dicyclohexylcarbodiimide, established that Trp and Asp/Glu residues are components of the PON1 active center and calcium-binding sites. Additional evidence for the presence of a Trp residue in the PON1 calcium-binding sites was a characteristic fluorescence emission at 545 nm from the PON1-Tb3+ complex and abolishment of that fluorescence upon modification by N-bromosuccinimide. The importance of aromatic/hydrophobic character of the residue 280 was demonstrated by site-directed mutagenesis: the W280F mutant was fully active while the W280A and W280L mutants had markedly reduced activity. Twelve amino acids among conserved His and Asp/Glu residues were found essential for PON1 arylesterase and organophosphatase activities: H114, H133, H154, H242, H284, D53, D168, D182, D268, D278, E52, and E194. Finally, the cysteines constituting the PON1 disulfide bond (C41 and C352) were essential, but the glycan chains linked to Asn 252 and 323 were not essential for PON1 secretion and activity.

Spectroscopic and kinetic characterization of eosin-5-maleimide.

Eosin-5-maleimide (EM) is an increasingly important and widely used probe in the study of membrane protein structure and function. Yet little is known about its spectral properties in hydrophobic and hydrophilic environments. Furthermore, EM is hydrolyzed faster than the traditional N-ethylmaleimide. To offer a more solid foundation for the use of EM in studies of membrane protein structure and function, we have undertaken a detailed study of the absorbance and fluorescence spectra of EM, eosin-5-maleic acid, and the L-cysteine and beta-mercaptoethanol adducts of EM in aqueous and hydrophobic environments; we have studied the kinetics of hydrolysis of EM in various environments, and we have investigated the reaction kinetics of EM with L-cysteine.

Role of aspartate 70 and tryptophan 82 in binding of succinyldithiocholine to human butyrylcholinesterase.

The atypical variant of human butyrylcholinesterase has Gly in place of Asp 70. Patients with this D70G mutation respond abnormally to the muscle relaxant succinyldicholine, experiencing hours of apnea rather than the intended 3 min. Asp 70 is at the rim of the active site gorge 12 A from the active site Ser 198. An unanswered question in the literature is why the atypical variant has a 10-fold increase in Km for compounds with a single positive charge but a 100-fold increase in Km for compounds with two positive charges. We mutated residues Asp 70, Trp 82, Trp 231, Glu 197, and Tyr 332 and expressed mutant enzymes in mammalian cells. Steady-state kinetic parameters for hydrolysis of butyrylthiocholine, benzoylcholine, succinyldithiocholine, and o-nitrophenyl butyrate were determined. The wild type and the D70G mutant had identical k(cat) values for all substrates. Molecular modeling and molecular dynamics suggested that succinyldicholine could bind in two consecutive orientations in the active site gorge; formation of one complex caused a conformational change in the omega loop involving Asp 70 and Trp 82. We propose the formation of three enzyme-substrate intermediates preceding the acyl-enzyme intermediate; kinetic data support this contention. Substrates with a single positive charge interact with Asp 70 just once, whereas substrates with two positive charges, for example succinyldithiocholine, interact with Asp 70 in two complexes, thus explaining the 10- and 100-fold increases in Km in the D70G mutant.

Characterization of the stilbenedisulphonate binding site on band 3 Memphis variant II (Pro-854-->Leu).

Band 3 Memphis variant II is a mutant anion-exchange protein associated with the Diego a+ blood group antigen. There are two mutations in this transporter: Lys-56-->Glu within the cytoplasmic domain, and Pro-854-->Leu within the membrane-bound domain. The Pro-854 mutation, which is thought to give rise to the antigenicity, is located within the C-terminal subdomain of the membrane-bound domain. Yet, there is an apparent enhancement in the rate of covalent binding of H2DIDS (4,4'-di-isothiocyanatodihydro-2, 2'-stilbenedisulphonate) to 'lysine A' (Lys-539) in the N-terminal subdomain, suggesting widespread conformational changes. In this report, we have used various kinetic assays which differentiate between conformational changes in the two subdomains, to characterize the stilbenedisulphonate site on band 3 Memphis variant II. We have found a significantly higher H2DIDS (a C-terminal-sensitive inhibitor) affinity for band 3 Memphis variant II, due to a lower H2DIDS 'off' rate constant, but no difference was found between mutant and control when DBDS (4,4'-dibenzamido-2,2'-stilbenedisulphonate) (a C-terminal-insensitive inhibitor) 'off' rates were measured. Furthermore, there were no differences in the rates of covalent binding to lysine A, for either DIDS (4,4'-di-isothiocyanato-2,2'-stilbenedisulphonate) or H2DIDS. However, the rate of covalent intrasubunit cross-linking of Lys-539 and Lys-851 by H2DIDS was abnormally low for band 3 Memphis variant II. These results suggest that the Pro-854-->Leu mutation causes a localized conformational change in the C-terminal subdomain of band 3.

Interactions between mutant and wild-type band 3 subunits in hereditary Southeast Asian ovalocytic red blood cell membranes.

Red cell membranes from individuals with Southeast Asian ovalocytosis (SAO) contain approximately equal proportions of wild-type band 3 and a mutant SAO band 3 which lacks residues 400-408. It is known that the Vmax for anion exchange in SAO cells is reduced by about 50%, that SAO band 3 does not transport anions when expressed alone in a cellular expression system, that SAO band 3 does not bind stilbenedisulfonates, and that about 50% of the band 3 exists as wild-type/SAO heterodimers. In this report, we show that the kinetics of H2DIDS (4,4'-diisothiocyanatodihydro-2,2'-stilbenedisulfonate) release from the wild-type band 3 in SAO membranes is biphasic. The two phases were present in about equal proportions, with rate constants differing by about 5-fold. In contrast; control cells showed monophasic, exponential kinetics with a rate constant comparable to that of the fast phase of SAO membranes. We assign the fast phase in SAO membranes to H2DIDS release from wild-type subunits within homodimers and the slow phase to H2DIDS release from the wild-type subunit within the heterodimer. No differences were observed in kinetic studies of H2DIDS binding. These results suggest that the mutant band 3 subunit alters the conformation of its neighboring wild-type subunit within the heterodimer, resulting in about a 4-fold higher H2DIDS affinity. Additional evidence suggesting that the interactions in the heterodimer may be confined to a region of the wild-type subunit containing the C-terminal subdomain is presented. The relationship of these subunit interactions to the observation of a reduced cellular anion transport function is discussed.

Differential sensitivity of stilbenedisulfonates in their reaction with band 3 HT (Pro-868-->Leu).

Band 3 HT (Pro-868-->Leu) is a mutant anion exchange protein which has several phenotypic characteristics, including a 2- to 3-fold larger Vmax, and reduced covalent binding of the anion transport inhibitor 4,4'-diisothiocyanodihydrostilbene-2,2'-disulfonate (H2DIDS). We have used fluorescence kinetic methods to study inhibitor binding to band 3 to determine if the point mutation in band 3 HT produces localized or wide-spread conformational changes within the membrane-bound domain of this transporter. Our results show that covalent binding of H2DIDS by band 3 HT is slower by a factor of 10 to 20 compared with the wild-type protein. In contrast, no such difference in the kinetics was observed for covalent binding of 4,4'-diisothiocyanostilbene-2,2'-disulfonate (DIDS). In addition, the kinetics of H2DIDS release from band 3 HT was abnormal, while the kinetics of 4,4'-dibenzamidostilbene-2,2'-disulfonate (DBDS) release showed no difference when compared with the wild-type protein. We conclude that substitution of leucine for proline at position 868 does not perturb the structure of "lysine A" in the membrane-bound domain of band 3 but rather produces an apparently localized conformational change in the C-terminal subdomain of the protein which alters H2DIDS affinity. When combined with the observation of an increased Vmax, these results suggest that protein structural changes at position 868 influence a turnover step in the transport cycle.

Characterization of the stilbenedisulfonate binding site on band 3.

Stilbenedisulfonates are potent inhibitors of Band 3 mediated anion exchange. They bind tightly to the protein and form a 1-to-1 reversible complex. Those stilbenedisulfonates which contain isothocyanato groups such as DIDS (4,4'-diisothiocyanato-2,2'-stilbenedisulfonate) and H2DIDS (4,4'-diisothiocyanatodihydrostilbene-2,2'-disulfonate) can also react rapidly with lysine residues within the binding pocket to yield an irreversible covalent adduct. The reactive lysine residue is known as lysine-A, and is thought to have an unusually low pKa. In this report, we characterize the kinetics of DIDS adduct formation with respect to the effect of substrate anions, competitive inhibitory anions, and pH on the rate of covalent adduct formation. We investigate the following: (a) whether stilbenedisulfonates bind to or block access of substrate anions to the transport site; (b) whether the rapidity of the covalent reaction of DIDS at neutral pH is due to a low pKa for lysine-A within the binding pocket; and (c) whether once bound, DIDS and H2DIDS isothiocyanato groups are accessible to reagents. For this latter experiment, we have utilized a newly discovered reaction of the DIDS isothiocyanato groups with azide to test for accessibility. Our results show that substrate anions, DIDS, and Band 3 form a ternary complex. Significantly, the binding of large substrate anions, such as iodide, is not weakened by DIDS to any greater extent than is the binding of smaller substrates such as chloride or fluoride. These results are not consistent with a "partial blockade" hypothesis for the relationship between the stilbenedisulfonate and transport sites. Rather, they support an allosteric site-site interaction hypothesis. Our pH dependence results show that the apparent pKa for the DIDS/lysine-A reaction is greater than 9.26. This is consistent with typical lysine pKa values, and indicates that lysine-A does not have an unusually low pKa. Finally, we show that azide can react with the isothiocyanato groups of DIDS and H2DIDS within their Band 3 complexes, indicating that the stilbenedisulfonate binding site is accessible to solute. These results support a view which suggests that the stilbenedisulfonate site is a superficial inhibitory site on Band 3 which inhibits transport by allosteric interactions within the protein, rather than by either direct or partial blockade of the transport site.

Inhibition of gp120-CD4 interaction and human immunodeficiency virus type 1 infection in vitro by pyridoxal 5'-phosphate.

Pyridoxal 5'-phosphate and related compounds were tested for their ability to inhibit gp120-CD4 interaction and human immunodeficiency virus infection in vitro. The results show that pyridoxal 5'-phosphate is a unique CD4 antagonist whose antiviral potency derives from the presence of both lysine-reactive and anionic substituents.

Kinetic evidence for ternary complex formation and allosteric interactions in chloride and stilbenedisulfonate binding to band 3.

The molecular basis for chloride and stilbenedisulfonate interaction with band 3 was investigated by measuring the kinetics of stilbenedisulfonate release from its complex with the transporter. We found that 150 mM NaCl accelerated the rate of release of DBDS (4,4'-dibenzamidostilbene-2,2'-dibenzamidostilbene-2,2'-disu lfonate) and H2DIDS (4,4'-diisothiocyanodihydrostilbene-2,2'-disulfonate) by more than 10-fold at constant ionic strength. The acceleration effect saturated as a function of chloride concentration. This is an indication of specific binding within a ternary complex involving stilbenedisulfonate, chloride, and band 3. To see if stilbenedisulfonates block an access channel to the transport site, we studied the effect of rapidly mixing DBDS-saturated resealed ghosts with chloride at constant ionic strength and osmotic pressure. Once again, we observe a large, uniform acceleration in the rate of DBDS release. These findings are not consistent with molecular models where stilbenedisulfonates are proposed to block access to a deeper transport site. We suggest that the intramonomeric stilbenedisulfonate site is not located on the chloride transport pathway but rather interacts with the transport site though heterotropic allosteric site-site interactions. On the basis of our kinetic evidence for ternary complex formation and on transport inhibition evidence in the literature showing a linear dependence of KI-app on substrate, we suggest that stilbenedisulfonates are linear mixed-type inhibitors of band 3 anion exchange, not pure competitive inhibitors as has been assumed on the basis of analysis of transport inhibition data alone.

Calorimetric evidence for allosteric subunit interactions associated with inhibitor binding to band 3 transporter.

A calorimetric endotherm occurring at 68 degrees C (the C-transition) has been assigned previously to the integral domain of band 3 and was shown to be shifted to 78 degrees C after covalent binding of 4,4'-diisothiocyanostilbene-2,2'-disulfonate (DIDS). In this study, we correlate the fractional appearance of the shifted C-transition with the fraction of DIDS bound to the band 3 monomer population. Our results show a distinctly nonlinear correlation plot with the appearance of the shifted C-transition lagging behind DIDS labeling of the band 3 monomer population. The lag suggests that both monomers of a band 3 dimer must be labeled by DIDS in order for the shifted C-transition to appear at 78 degrees C, implying that the thermal unfolding of the integral domain of band 3 is modulated by allosteric interactions between subunits. This is the first in situ structural evidence supporting ligand-mediated subunit interactions within a "carrier"-type transporter protein oligomer.

Kinetics of conformational changes associated with inhibitor binding to the purified band 3 transporter. Direct observation of allosteric subunit interactions.

Subunit interaction effects were identified for isolated human erythrocyte band 3, the anion exchanger, by observing both static and stopped-flow kinetic protein fluorescence changes associated with inhibitor binding to the intramonomeric stilbenedisulfonate site. We measured the rate of conformational changes associated with reversible binding of H2DIDS (4,4'-diisothiocyanodihydrostilbene-2,2'-disulfonate). The rate of H2DIDS release was also measured. As a test for subunit interactions, we studied the effect of partial labeling of the band 3 monomer population with H2DIDS on the equilibrium and kinetics of H2DIDS reversible binding to the remaining monomers. The results showed biphasic kinetics for control band 3, with a pseudo-first-order ligand dependence for the fast phase followed by a slow ligand-independent relaxation. A second-order "on" rate constant for the fast phase was determined to be (1.2 +/- 0.1) x 10(7) M-1 s-1, while the associated "off" rate constant was found to be 1.1 +/- 0.5 s-1. From these kinetic constants, we calculated a Kd value of 95 +/- 50 nM, which is in excellent agreement with the Kd value determined at thermodynamic equilibrium (110 +/- 9 nM). Covalent labeling of 75% of the band 3 monomer population with H2DIDS changed the kinetics of the fast phase, slowing the apparent rate by changing the order of the reaction from pseudo-first-order to zero-order. Partial labeling did not affect the ligand-independent relaxation. Separate measurements of the H2DIDS "off" rate also showed a biphasic time course, with a 20-fold difference in apparent rate constants.(ABSTRACT TRUNCATED AT 250 WORDS)

Pyridoxal 5'-phosphate binds specifically to soluble CD4 protein, the HIV-1 receptor. Implications for AIDS therapy.

Considerable effort is being made to design anti-viral drugs for the human immunodeficiency virus type 1 (HIV-1) infection process. Some of this work has focused on CD4 protein, the HIV-1 receptor on T helper lymphocytes. One drug that binds to CD4 protein and inhibits both viral infection and growth is DIDS (4,4'-diisothiocyanato-2,2'-stilbenedisulfonate). DIDS is best known for its ability to inhibit erythrocyte band 3 anion exchange. Although the antiviral potency of DIDS is evident in vitro (IC50 approximately 30 microM), intravenous administration of DIDS should not be effective owing to the large number of band 3 molecules present on the red blood cell membrane (approximately 10(6)/cell), and to the very small Kd for DIDS binding to band 3 (approximately 30 nM). Therefore, we sought to identify other anion transport inhibitors that would bind weakly to band 3, but tightly to CD4 protein, and that could be administered to humans without significant toxic side effects. On the basis of our previous work with band 3 (Salhany, J. M., Rauenbuehler, P. B., and Sloan, R. L. (1987) J. Biol. Chem. 262, 15965-15973), we elected to study the binding of pyridoxal 5'-phosphate (PLP) to soluble CD4 protein. We have discovered that PLP binds surprisingly tightly to soluble CD4 protein (Kd = 45 microM), with a stoichiometry of about 1 mol of PLP/mol of protein. Furthermore, PLP binding was found to be competitive with DIDS for its binding site on soluble CD4 protein. These results suggest that PLP may be an effective anti-viral agent for the HIV-1 infection process.

Factors determining the conformation and quaternary structure of isolated human erythrocyte band 3 in detergent solution.

Fluorescence spectroscopy was used to follow the kinetics of covalent binding of DIDS (4,4'-diisothiocyanato-2,2'-stilbenedisulfonate) to isolated band 3 in C12E8. We have discovered a dilution-induced loss in the ability of band 3 monomer to form a covalent adduct with DIDS. The loss in DIDS reactivity with dilution followed a 50:50 biphasic time course despite the use of a homogeneous preparation of band 3 oligomers. The loss in reactivity generally correlated with the association of band 3 dimers and tetramers to higher oligomeric structures. The final aggregated product was capable of binding BADS (4-benzamido-4'-amino-2,2'-stilbenedisulfonate) reversibly, but with an affinity nearly 30-fold lower than that of the starting material. Removal of the cytoplasmic domain of band 3 slowed the conformational interconversion of the integral domain by about 5-fold and inhibited the aggregation process. The conformational interconversion was slowed in the presence of 150 mM chloride but not in 90 mM sulfate. Covalent binding of DIDS inhibited the aggregation of band 3. Addition of 250 microM lipid inhibited both the loss of DIDS reactivity and the protein aggregation process. While several types of lipid offer protection, phosphatidic acid accelerated the decay process by eliminating the biphasicity. We conclude that the conformation of the integral domain of band 3 can be modulated allosterically by the addition of ligands, including various lipids. The results offer direct evidence for cooperative interactions between band 3 subunits during loss of activity, and they show that the cytoplasmic domain participates in the control of this transition.