Chloride as allosteric effector of yeast aminopeptidase I

Activation of yeast aminopeptidase I by chloride was studied by kinetic methods. Several effects contributed to overall activity enhancement: At low concentrations of Zn2+ (an essential component of aminopeptidase I) chloride increased the amounts of active enzyme by reducing the cooperativity of metal binding. In addition, substrate turnover was enhanced due to increased kcat and a moderate decrease of Km. At high concentrations of Zn2+ substrate saturation curves were sigmoidal. Under these conditions chloride activated by restoring Michaelis-Menten kinetics of substrate turnover. At the same time, reconstitution of active enzyme from apoprotein and Zn2+ was substantially accelerated and its inactivation due to loss of Zn2+ was retarded. Co2+-Substituted aminopeptidase I, although catalytically active, was much less sensitive to chloride activation. Apparent binding constants for chloride, as estimated from its effects on metal binding and catalysis, respectively, were different. This suggests that two independent activation mechanisms may be operative. Both appear to be mediated by conformational changes of the enzyme protein.     

Interaction of the (2S,3S)-isomer of bestatin with yeast aminopeptidase I. Kinetic and binding studies

Inhibition of yeast aminopeptidase I by N-[(2S,3S)-3-amino-2-hydroxyl-1-oxo-4-phenylbutyl]-L-leucine [(2S,3S)-Ahp-Leu];a stereoisomer of natural bestatin, is a slow process with half-times in the minute range. Action of the inhibitor is non-competitive with respect to the substrate. Up to 1 mol of (2S,3S)-Ahp-Leu is bound per mol of enzyme subunit. Inhibitor binding does not interfere with binding of essential metal ions but completely suppresses allosteric activation by chloride and high Zn(II)-concentrations. These and other findings suggest that (2S,3S)-Ahp-Leu inhibits aminopeptidase I by stabilizing a weakly active enzyme conformation.     

Spectral and kinetic studies of metal-substituted Aeromonas aminopeptidase: nonidentical, interacting metal-binding sites

Apoenzyme prepared by removal of the 2 mol of Zn2+/mol from Aeromonas aminopeptidase is inactive. Addition of Zn2+ reactivates it completely, and reconstitution with Co2+, Ni2+, or Cu2+ results in a 5.0-, 9.8-, and 10-fold more active enzyme than native aminopeptidase, respectively. Equilibrium dialysis and spectral titration experiments with Co2+ confirm the stoichiometry of 2 mol of metal/mol. The addition of only 1 mol of metal/mol completely restores activity characteristic of the particular metal. Interaction between the two sites, however, causes hyperactivation; thus, addition of 1 mol of Zn2+/mol subsequent to 1 mol of Co2+, Ni2+, or Cu2+ per mole increases activity 3.2-, 42-, or 59-fold, respectively. The cobalt absorption spectrum has a peak of 527 nm with a molar absorptivity of 53 M-1 cm-1 for 1 mol of cobalt/mol, which increases to 82 M-1 cm-1 for a second cobalt atom and is unchanged by further addition of Co2+. Circular dichroic (CD) and magnetic CD spectra indicate that the first Co2+ binding site is tetrahedral-like and that the second is octahedral-like. Stoichiometric quantities of 1-butylboronic acid, a transition-state analogue inhibitor of the enzyme [Baker, J. O., & Prescott, J. M. (1983) Biochemistry 22, 5322], profoundly affects absorption, CD, and MCD spectra, but n-valeramide, a substrate analogue inhibitor, has no effect. These findings suggest that the tetrahedral-like site is catalytic and the other octahedral-like site is regulatory or structural.     

Metal ion substitution in the catalytic site greatly affects the binding of sulfhydryl-containing compounds to leucyl aminopeptidase

Bovine lens leucyl aminopeptidase (blLAP), a homohexameric metallopeptidase preferring bulky and hydrophobic amino acids at the N-terminus of (di)peptides, contains two Zn(2+) ions per subunit that are essential for catalytic activity. They may be replaced by other divalent cations with different exchange kinetics. The protein readily exchangeable site (site 1) can be occupied by Zn(2+), Mn(2+), Mg(2+), or Co(2+), while the tight binding site (site 2) can be occupied by Zn(2+) or Co(2+). We recently reported that introduction of Mn(2+) into site 1 generates a novel activity of blLAP toward CysGly [Cappiello, M., et al. (2004) Biochem. J. 378, 35-44], which in contrast is not hydrolyzed by the (Zn/Zn) enzyme. This finding, while disclosing a potential specific role for blLAP in glutathione metabolism, raised a question about the features required for molecules to be a substrate for the enzyme. To clarify the interaction of the enzyme with sulfhydryl-containing derivatives, (Zn/Zn)- and (Mn/Zn)blLAP forms were prepared and functional-structural studies were undertaken. Thus, a kinetic analysis of various compounds with both enzyme forms was performed; the crystal structure of (Zn/Zn)blLAP in complex with the peptidomimetic derivative Zofenoprilat was determined, and a modeling study on the CysGly-(Zn/Zn)blLAP complex was carried out. This combined approach provided insight into the interaction of blLAP with sulfhydryl-containing derivatives, showing that the metal exchange in site 1 modulates binding to these molecules that may result in enzyme substrates or inhibitors, depending on the nature of the metal.     

Yeast methionine aminopeptidase I can utilize either Zn2+ or Co2+ as a cofactor: A case of mistaken identity?

Yeast methionine aminopeptidase I (MetAP I) is one of two enzymes in Saccharomyces cerevisiae that is responsible for cotranslational cleavage of initiator methionines. It has previously been classified as a Co2+ metalloprotease in all prokaryotic and eukaryotic forms studied. However, treatment of recombinant apo-MetAP I with 12.5 microM Zn2+ produces an enzyme that is as active as that reconstituted with 200 microM Co2+. In the presence of physiological concentrations of reduced glutathione (GSH), Co-MetAP I is inactive, while the activity of Zn-MetAP I is increased more than 1.7-fold over Zn-MetAP I assayed in the absence of GSH. Given that the in vivo concentration of Zn2+ is at least 1,000-fold higher than that of Co2+, and that Co2+ is insoluble in physiological concentrations of GSH, it is probable that yeast MetAP I is actually a Zn2+ metalloprotease. Furthermore, unless there are extraordinary conditions that insulate or sequester them from this reducing milieu, that have yet to be identified, there are not likely to be any cytoplasmic enzymes that use free Co2+.     

Leucine aminopeptidase (bovine lens). The relative binding of cobalt and zinc to leucine aminopeptidase and the effect of cobalt substitution on specific activity.

Prolonged incubation of zinc-zinc leucine aminopeptidase (bovine lens) (EC 3.4.1.1) with 0.05 M CoCl2 and M KCl in 0.2 M N-ethylmorpholine-HCl at pH 7.5 and 37 degrees yields an active enzyme in which 2 g atoms of Co2+ per 54,000 dalton subunit have replaced the Zn2+. Incubation of cobalt-cobalt leucine aminopeptidase with various AnCl2 concentrations or zinc-zinc leucine aminopeptidase with various CoCl2 concentrations in M KCl and 0.2 M N-ethylmorpholine-HCl at pH 7.5 and 37 degrees demonstrates that Co2+ and Zn2+ compete reversibly for two independent binding sites per subunit for which the ratio of the association constants for Zn2+ and Co2+ (1KZn:1KCo = 1KZn/Co; 2KZn:2KCo = 2KZn/Co) are 115 and 15.9 for sites 1 and 2, respectively. The specific activities of the various species of enzyme with 2 mM L-leucine p-nitroanilide as substrate in 0.2 M N-ethylmorpholine-HCl and 0.01 M NaHCO3 at pH 7.5 are estimated to be (in micromoles per min per mg) 0.043 for the zinc-zinc. 0.039 for the zinc-cobalt, 0.541 for the cobalt-zinc, and 0.536 for the cobalt-cobalt forms, which implies that activity is affected only when cobalt is substituted at site 1, the "activation site." The site, at which cobalt substitution has no effect on activity, is designated the "structural site." The value of Km for cobalt-cobalt leucine aminopeptidase with L-leucine p-nitroanilide as substrate in 0.2 M N-ethylmorpholine-HCl at pH 7.5 containing 0.01 M NaHCO3 at 30 degrees is 0.52 mM while Vmax is 0.90 mumol per min per mg. In the additional presence of 1 M KCl, Km is 0.19 mM while Vmax is 0.68 mumol per min per mg.     

Octahedral metal coordination in the active site of glyoxalase I as evidenced by the properties of Co(II)-glyoxalase I

Co(II)-glyoxalase I has been prepared by reactivation of apoenzyme from human erythrocytes with Co2+. The visible absorption spectrum showed maxima at 493 and 515 nm and shoulders at 465 and 615 nm. The absorption coefficients at 493 and 515 nm were 35 and 33 M-1 cm-1/cobalt ion, respectively; i.e. 70 and 66 M-1 cm-1 for the dimeric metalloprotein. The product of the enzymatic reaction, S-D-lactoylglutathione, although binding to Co(II)-glyoxalase I, had no demonstrable effect on the visible absorption spectrum, indicating binding outside the first coordination sphere of the metal. The EPR spectrum at 3.9 K was characterized by g1 approximately 6.6, g2 approximately 3.0, and g3 approximately 2.5, and eight hyperfine lines with A1 = 0.025 cm-1. Binding of the strong competitive inhibitor S-p-bromobenzylglutathione to Co(II)-glyoxalase I gave three g values: 6.3, 3.4, and 2.5, indicating a conformational change affecting the environment of the metal ion. Both optical and EPR spectra strongly suggest a high spin Co2+ with octahedral coordination in the active site of the enzyme. The similarities in kinetic properties between native Zn(II)-glyoxalase I and enzyme substituted with Mg2+, Mn2+, or Co2+ is consistent with the view that these enzyme forms have the same metal coordination in the protein.     

Modified activity of Aeromonas aminopeptidase: metal ion substitutions and role of substrates

Aeromonas aminopeptidase contains two nonidentical metal binding sites that have been shown by both spectroscopy and kinetics to be capable of interacting with one another [Prescott, J.M., Wagner, F.W., Holmquist, B., & Vallee, B.L. (1985) Biochemistry 24, 5350-5356]. The effects of metal ion substitutions on the susceptibility of the p-nitroanilides of L-alanine, L-valine, and L-leucine--substrates that are hydrolyzed at widely differing rates by native Aeromonas aminopeptidase--were studied by determining values of kcat and Km for the 16 metalloenzymes that result from all possible combinations of Zn2+, Co2+, Ni2+, and Cu2+ in each of the two sites. The different combinations of metal ions and substrates yield a broad range in kinetic values; kcat varies by more than 1800-fold, Km by 3000-fold, and kcat/Km ratios by more than 10,000. L-Leucine-p-nitroanilide is by far the most susceptible of the three substrates, and the hyperactivation previously observed with aminopeptidase containing either Ni2+ or Cu2+ in the first binding site and Zn2+ in the second site occurs only with the two poorer substrates, L-alanine-p-nitroanilide and L-valine-p-nitroanilide. Although the enzyme with Zn2+ in both sites hydrolyzes the substrates with N-terminal alanine and valine poorly, it is extremely effective toward L-leucine-p-nitroanilide. Neither metal binding site can be identified as controlling either Km or kcat; both parameters are influenced by the identity of the metal ions, by the site each occupies, and, most strongly, by the substrate.(ABSTRACT TRUNCATED AT 250 WORDS)     

Physiologically relevant metal cofactor for methionine aminopeptidase-2 is manganese

The identity of the physiological metal cofactor for human methionine aminopeptidase-2 (MetAP2) has not been established. To examine this question, we first investigated the effect of eight divalent metal ions, including Ca(2+), Co(2+), Cu(2+), Fe(2+), Mg(2+), Mn(2+), Ni(2+), and Zn(2+), on recombinant human methionine aminopeptidase apoenzymes in releasing N-terminal methionine from three peptide substrates: MAS, MGAQFSKT, and (3)H-MASK(biotin)G. The activity of MetAP2 on either MAS or MGAQFSKT was enhanced 15-25-fold by Co(2+) or Mn(2+) metal ions in a broad concentration range (1-1000 microM). In the presence of reduced glutathione to mimic the cellular environment, Co(2+) and Mn(2+) were also the best stimulators (approximately 30-fold) for MetAP2 enzyme activity. To determine which metal ion is physiologically relevant, we then tested inhibition of intracellular MetAP2 with synthetic inhibitors selective for MetAP2 with different metal cofactors. A-310840 below 10 microM did not inhibit the activity of MetAP2-Mn(2+) but was very potent against MetAP2 with other metal ions including Co(2+), Fe(2+), Ni(2+), and Zn(2+) in the in vitro enzyme assays. In contrast, A-311263 inhibited MetAP2 with Mn(2+), as well as Co(2+), Fe(2+), Ni(2+), and Zn(2+). In cell culture assays, A-310840 did not inhibit intracellular MetAP2 enzyme activity and did not inhibit cell proliferation despite its ability to permeate and accumulate in cytosol, while A-311263 inhibited both intracellular MetAP2 and proliferation in a similar concentration range, indicating cellular MetAP2 is functioning as a manganese enzyme but not as a cobalt, zinc, iron, or nickel enzyme. We conclude that MetAP2 is a manganese enzyme and that therapeutic MetAP2 inhibitors should inhibit MetAP2-Mn(2+).     

Studies of cation binding in ZnCl2-regenerated bacteriorhodopsin by x-ray absorption fine structures: effects of removing water molecules and adding Cl- ions.

The binding of Zn2+ in Zn2+-regenerated bacteriorhodopsin (bR) was studied under various conditions by x-ray absorption fine structures (XAFS). The 0.9:1 and 2:1 Zn2+:bR samples gave similar XAFS spectra, suggesting that Zn2+ might have only one strong binding site in bR. It was found that in aqueous bR solution, Zn2+ has an average of six oxygen or nitrogen ligands. Upon drying, two ligands are lost, suggesting the existence of two weakly bound water ligands near the cation-binding site in bacteriorhodopsin. When excess Cl- ions were present before drying in the Zn2+-regenerated bR samples, it was found that two of the ligands were replaced by Cl- ions in the dried film, whereas two remain unchanged. The above observations suggest that Zn2+ has three types of ligands in regenerated bR (referred to as types I, II, and III). Type I ligands are strongly bound. These ligands cannot be removed by drying or by exchanging with Cl- ions. Type II ligands cannot be removed by drying, but can be replaced by Cl- ligands. Type III ligands are weakly bound to the metal cation and are most likely water molecules that can be removed by evaporation under vacuum or by drying with anhydrous CaSO4. The results are discussed in terms of the possible structure of the strongly binding site of Zn2+ in bR.     

Characterization of three aminopeptidases purified from maternal serum

The biochemical characteristics of aminopeptidase A (EC 3.4.11.7), oxytocinase (EC 3.4.11.3) and alanyl aminopeptidase (EC 3.4.11.2) purified from serum of pregnant women were compared. Aminopeptidase A hydrolysed only acidic amino acid derivatives, whereas oxytocinase and alanyl aminopeptidase had partially overlapping broad substrate specificities. Oxytocinase showed the highest Vmax value with LeuNA but the lowest Km value with ArgNA (Km 0.059 +/- 0.08 mmol/l). Alanyl aminopeptidase hydrolysed AlaNA most rapidly, but showed the highest affinity for LysNA (Km 0.054 +/- 0.006 mmol/l). The enzymes were sensitive to EDTA. Co2+, Ni2+ and Zn2+ were able to reactivate all suppressed enzymes, but Mn2+ reactivated only aminopeptidase A after EDTA inhibition. The alkaline earth metals were activators of aminopeptidase A, while Co2+ activated only alanyl aminopeptidase. This enzyme was the most sensitive to L-amino acids. Acidic amino acids inhibited aminopeptidase A but had no effect on the two other enzymes. Oxytocinase was most sensitive to thermal treatment. Amastatin did not inhibit oxytocinase, whereas aminopeptidase A was more resistant than alanyl aminopeptidase to this effector.     

Isolation of the leucine aminopeptidase gene from Aeromonas proteolytica. Evidence for an enzyme precursor.

The leucine aminopeptidase of Aeromonas proteolytica (EC 3.4.11.10) is a monomeric metalloenzyme having the capacity to bind two Zn2+ atoms in the active site. Structural information of this relatively small aminopeptidase that could illuminate the catalytic mechanism of the metal ions is lacking; hence, we have obtained sequences from the purified enzyme, cloned the corresponding gene, and expressed the recombinant protein in Escherichia coli. The deduced primary amino acid sequence of this secreted protease suggests a potential signal peptide at the NH2 terminus. Expression of the recombinant and native proteins in E. coli and in extracts of culture media of A. proteolytica indicates that the aminopeptidase is secreted as an active and thermosensitive 43-kDa protein that is rapidly transformed to thermostable forms of 30 and 32 kDa. Comparison of the deduced amino acid sequence of the A. proteolytica leucine aminopeptidase with other Zn(2+)-binding metalloenzymes failed to show homologies to the consensus binding sequence His-Glu-X-X-His for the metal ion.     

Slow-binding inhibition of the aminopeptidase from Aeromonas proteolytica by peptide thiols: synthesis and spectroscopic characterization

Peptide-derived thiols of the general structure N-mercaptoacyl-leucyl-p-nitroanilide (1a-c) were synthesized and found to be potent, slow-binding inhibitors of the aminopeptidase from Aeromonas proteolytica (AAP). The overall potencies (K(I)) of these inhibitors against AAP range from 2.5 to 57 nM exceeding that of the natural product bestatin and approaching that of amastatin. The corresponding alcohols (2a-b) are simple competitive inhibitors of much lower potencies (K(I) = 23 and 360 microM). These data suggest that the free thiols are involved in the formation of the E. I and E.I complexes, presumably serving as a metal ligand. To investigate the nature of the interaction of the thiol-based inhibitors with the dinuclear active site of AAP, we have recorded electronic absorption and EPR spectra of Co(II)Co(II)-, Co(II)Zn(II)-, and Zn(II)Co(II)-AAP in the presence of the strongest binding inhibitor, 1c. Both [CoZn(AAP)] and [ZnCo(AAP)], in the presence of 1c, exhibited an absorption band centered at 320 nm characteristic of an S --> Co(II) ligand-metal charge-transfer band. In addition, absorption spectra recorded between 400 and 700 nm showed changes characteristic of 1c interacting with each active-site metal ion. EPR spectra recorded at high temperature (19 K) and low power (2.5 mW) indicated that in a given enzyme molecule, 1c interacts weakly with one of the metal ions in the dinuclear site and that the crystallographically identified micro-OH(H) bridge, which has been shown to mediate electronic interaction of the Co(II) ions, is likely broken upon 1c binding. EPR spectra of [CoCo(AAP)]-1c, [ZnCo(AAP)]-1c, and [CoZn(AAP)]-1c were also recorded at lower temperature (3.5-4.0 K) and high microwave power (50-553 mW). The observed signals were unusual and appeared to contain, in addition to the incompletely saturated contributions from the signals characterized at 19 K, a very sharp feature at g(eff) approximately 6.8 that is characteristic of thiolate-Co(II) interactions. These data suggest that the thiolate moiety can bind to either of the metal ions in the dinuclear active site of AAP but does not bridge the dinuclear cluster. Compounds 1a-c are readily accessible by synthesis and thus provide a novel class of potent aminopeptidase inhibitors.     

New chloride-activated aminopeptidase from human erythrocytes

A new Cl- -activated aminopeptidase was purified from the cytosol of human erythrocytes as a single chain protein of an approx. Mr of 70,000 and pI of 5.1. The enzyme hydrolysed 2-naphthylamides of aliphatic, aromatic and basic L-amino acids, with a preference for the alanyl residue. It also hydrolysed di-, tri-, and some hydrophobic tetrapeptides. The inhibitors were bestatin, amastatin, Co2+, Zn2+, Mn2+, 4-hydroxymercuribenzoate and 1,10-phenanthroline. The activity of the enzyme, inhibited by 4-hydroxymercuribenzoate, was partially restored by the addition of sulfhydryl compounds. The presence of 0.2 M Cl- (Br-,F-) caused a several-fold increase in the isolated aminopeptidase activity.     

Activation of sphingomyelinase from Bacillus cereus by Zn2+ hitherto accepted as a strong inhibitor

Sphingomyelinase (SMase) from Bacillus cereus has been known to be activated by Mg2+, Mn2+, and Co2+, but strongly inhibited by Zn2+. In the present study, we investigated the effects of several kinds of metal ions on the catalytic activity of B. cereus SMase, and found that the activity was inhibited by Zn2+ at its higher concentrations or at higher pH values, but unexpectedly activated at lower Zn2+ concentrations or at lower pH values. This result indicates that SMase possesses at least two different binding sites for Zn2+ and that the Zn2+ binding to the high-affinity site can activate the enzyme, whereas the Zn2+ binding to the low-affinity site can inactivate it. We also found that the binding of substrate to the enzyme was independent of the Zn2+ binding to the high-affinity site, but was competitively inhibited by the Zn2+ binding to the low-affinity site. The binding affinity of the metal ions to the site for activating the enzyme was determined to be in the rank-order of Mg2+ = Co2+ < Mn2+ < Zn2+. It was also demonstrated that these four metal ions competed with each other for the same binding site on the enzyme molecule.     

Nuclear magnetic relaxation studies of the role of the metal ion in Mn2(+)-substituted aminoacylase I

Substitution of the essential Zn2+ ions of porcine kidney aminoacylase I (EC 3.5.1.14) by Mn2+ did not markedly affect the kinetic properties of the enzyme. Using Mn2+ as a paramagnetic probe, we were able to study the conformations of bound ligands by measuring the enhancement of ligand proton relaxation in 1H NMR. In addition, the effects of inhibitors on the paramagnetic enhancement of water proton relaxation rates were examined. The results of both approaches, in agreement with kinetic evidence, suggest that the metal center of aminoacylase I is too distant from the ligand binding site to allow direct participation of the metal in substrate binding or catalysis. We, therefore, propose that the metal ion of aminoacylase I plays a purely structural role.     

Purification and partial characterization of glyoxalase I from bovine brain

Glyoxalase I was purified to homogeneity from bovine brain using affinity chromatography on S-hexylglutathione-Sepharose 6B with a yield of 22%. The enzyme was a dimer (44,000 Daltons) composed of, apparently, identical subunits (22,000 Daltons), as shown by SDS electrophoresis, and contained one mole of Zn2+/monomer. The active site metal ion, Zn2+, was removed by dialysis against EDTA, but the activity of the apoenzyme obtained was not completely restored after addition of Co2+ and Zn2+ (<25%), while a recovery of 50% was obtained after addition of Mg2+. The enzyme was inhibited by S-bromobenzylglutathione and S-p-nitrobenzylglutathione with a Ki value of 21 microM and 32 microM, respectively. The highest dissociation constant observed for the brain enzyme with respect to that reported for human erythrocytes, or other mammalian forms of enzyme could be related to a tissue-specific dependence of the glyoxalase I activity.     

Soluble and immobilized clostridial aminopeptidase and aminopeptidase P as metal-requiring enzymes

The dependence of enzymatic activity on Co2+ concentration was found to be bell-shaped for the soluble and immobilized clostridial aminopeptidase (alpha-aminoacyl-peptide hydrolase, EC 3.4.11.13) and aminopeptidase P (aminoacylpropyl-peptide hydrolase, EC 3.4.11.9), with maxima in the 3-18 microM range of Co2+ concentration. The Co2+-enzyme association constants derived from the activation of soluble, glass- and cellulose-bound clostridial aminopeptidase by Co2+ were KE-Co = 5.2 X 10(5), 4.5 X 10(6) and 2.0 X 10(5) M-1, respectively; for soluble and glass-bound aminopeptidase P, the KE-Co were 1.5 X 10(5) and 8.2 X 10(5) M-1, respectively. Kinetic measurements indicate the involvement of Co2+ in the enzyme-substrate binding. Cobalt-citrate (Co-cit) acted as a useful metallobuffer and protected both enzymes against inhibition by high concentrations of CoSO4. For association of citrate with Co2+ under the assay conditions, KCo-cit was determined as (5.3 +/- 1.4) X 10(3) M-1 by anodic stripping polarography. In contrast to the rapid association of Co2+ with soluble and glass-bound clostridial aminopeptidase (less than 1 min at 4 degrees C), the dissociation process was very slow (hours to days), being slower for the glass-bound than for the soluble and cellulose-bound enzyme. For aminopeptidase P, both processes were rapid. All the interactions were shown to be reversible.     

Metal ion-induced conformational changes in Escherichia coli alkaline phosphatase.

Ultraviolet difference spectra are produced by the binding of divalent metal ions to metal-free alkaline phosphatase (EC 3.1.3.1). The interaction of the apoprotein with Zn2+, Mn2+, Co2+ and Cd2+, which induce the tight binding of one phosphate ion per dimer, give distinctly different ultraviolet spectra changes from Ni2+ and Hg2+ which do not induce phosphate binding. Spectrophotometric titrations at alkaline pH of various metallo-enzymes reveal a smaller number of ionizable tyrosines and a greater stability towards alkaline denaturation in the Zn2+- and Mn2+-enzymes than in the Ni2+-, Hg2+- and apoenzymes. The Zn2+- and Mn2+-enzymes have CD spectra in the region of the aromatic transitions that are different from the CD spectra of the Ni2+-, Hg2+- and apoenzymes. Modifications of arginines with 2,3-butanedione show that a smaller number of arginine residues are modified in the Zn2+-enzyme than in the Hg2+-enzyme. The presented data indicate that alkaline phosphatase from Escherichia coli must have a well-defined conformation in order to bind phosphate. Some metal ions (i.e. Zn2+, Co2+, Mn2+ and Cd2+), when interacting with the apoenzyme, alter the conformation of the protein molecule in such a way that it is able to interact with substrate molecules, while other metal ions (i.e. Ni2+ and Hg2+) are incapable of inducing the appropriate conformational change of the apoenzyme. These findings suggest an important structural function of the first two tightly bound metal ions in enzyme.