Apoptosis is regulated by the VDAC1 N-terminal region and by VDAC oligomerization: release of cytochrome c,AIF and Smac/Diablo

Mitochondria, central to basic life functions due to their generation of cellular energy, also serve as the venue for cellular decisions leading to apoptosis. A key protein in mitochondria-mediated apoptosis is the voltage-dependent anion channel (VDAC), which also mediates the exchange of metabolites and energy between the cytosol and the mitochondria. In this study, the functions played by the N-terminal region of VDAC1 and by VDAC1 oligomerization in the release of cytochrome c, Smac/Diablo and apoptosis-inducing factor (AIF) and subsequent apoptosis were addressed. We demonstrate that cells undergoing apoptosis induced by STS or cisplatin and expressing N-terminally truncated VDAC1 do not release cytochrome c, Smac/Diablo or AIF. Ruthenium red (RuR), AzRu, DIDS and hexokinase-I (HK-I), all known to interact with VDAC, inhibited the release of cytochrome c, Smac/Diablo and AIF, while RuR-mediated inhibition was not observed in cells expressing RuR-insensitive E72Q-VDAC1. These findings suggest that VDAC1 is involved in the release of not only cytochrome c but also of Smac/Diablo and AIF. We also demonstrate that apoptosis induction is associated with VDAC oligomerization, as revealed by chemical cross-linking and monitoring in living cells using Bioluminescence Resonance Energy Transfer. Apoptosis induction by STS, H₂O₂ or selenite augmented the formation of VDAC oligomers several fold. The results show VDAC1 to be a component of the apoptosis machinery and offer new insight into the functions of VDAC1 oligomerization in apoptosis and of the VDAC1 N-terminal domain in the release of apoptogenic proteins as well as into regulation of VDAC by anti-apoptotic proteins, such as HK and Bcl2.     

Calcium and lanthanide affinity of the EF-loops from the C-terminal domain of calmodulin

To obtain site-specific information about individual EF-hand motifs, the EF-hand Ca(2+)-binding loops from site III and site IV of calmodulin (CaM) were inserted separately into a non-Ca(2+)-binding cell adhesion protein, domain 1 of CD2 (denoted as CaM-CD2-III-5G-52 and CaM-CD2-IV-5G-52). Structural analyses using various spectroscopic methods have shown that the host protein CD2 retains its native structure after the insertion of the 12-residue loops. The Tb(3+) fluorescence enhancement upon formation of a Tb(3+)-protein complex and the direct competition by La(3+) and Ca(2+) suggest that native Ca(2+)-binding pockets are formed in both engineered proteins. Moreover, as revealed by NMR, both Ca(2+) and La(3+) specifically interact with the residues at the grafted EF-loop. The CaM-CD2-III-5G-52 has stronger affinities to Ca(2+), Tb(3+) and La(3+) than CaM-CD2-IV-5G-52, indicating differential intrinsic metal-binding affinities of the EF-loops.     

Characterization of a Ca2+ binding and regulatory site in the Ca2+ release channel (ryanodine receptor) of rabbit skeletal muscle sarcoplasmic reticulum.

A region in the skeletal muscle ryanodine receptor between amino acids 4014 and 4765 was expressed as a trpE fusion protein. Overlay studies revealed that this region bound Ca2+ and ruthenium red, an indicator of Ca(2+)-binding sites. Ca2+ binding was mapped to subregion 13b between amino acids 4246 and 4377, encompassing a predicted high affinity Ca(2+)-binding site, and to subregion 13c between amino acids 4364 and 4529, encompassing two predicted high affinity Ca(2+)-binding sites. Ca2+ binding was then mapped to three shorter sequences, 22(13b1), 36(13c1), and 35(13c2), amino acids long, each encompassing one of the three predicted Ca(2+)-binding sites. Site-directed polyclonal antibodies were raised against these three short sequences and purified on antigen affinity columns. The antibody against sequence 13c2, lying between residues 4478 and 4512, specifically recognized both denatured and native forms of the ryanodine receptor, suggesting that at least part of the 35 amino acid sequence containing the Ca(2+)-binding site is surface-exposed. The affinity purified antibody increased the Ca2+ sensitivity of ryanodine receptor channels incorporated into planar lipid bilayers, resulting in increased open probability and opening time without altering channel conductance. The antibody-activated channel was still modulated by Ca2+, Mg2+, ATP, ryanodine, and ruthenium red. These observations suggest that sequence 13c2 may be involved in Ca(2+)-induced Ca2+ release.     

Functional implication with the metal-binding properties of KChIP1

To investigate the metal-binding properties of KChIP1, the interaction of KChIP1 and mutated KChIP1 with divalent cations (Mg(2+), Ca(2+), Sr(2+), and Ba(2+)) was explored by 8-anilinonaphthalene-1-sulfonate (ANS) fluorescence. It showed that KChIP1 possessed two types of Ca(2+)-binding sites, high-affinity and low-affinity Ca(2+)-binding sites. However, only low-affinity-binding site for Mg(2+), Sr(2+), and Ba(2+) was observed. The metal-binding properties of KChIP1 are not appreciably affected after removal of the N-terminal portion and EF-hand 1. Deleting the EF-hand 4 of KChIP1 abolishes its high-affinity Ca(2+)-binding site, but retains the intact low-affinity-binding site for metal ions. A decrease in the nonpolarity of ANS-binding site occurs with all mutants. However, the binding of ANS with KChIP1 is no longer observed after removal of EF-hands 3 and 4. Intermolecular interaction assessed by chemical cross-linking suggested that KChIP1 had a propensity to form dimer in the absence of metal ions, and a KChIP1 tetramer was pronouncedly produced in the presence of metal ions. Noticeably, the oligomerization state depends on the integrity of EF-hand 4. Taken together, our data suggest that EF-hand 4 is of structural importance as well as functional importance for fulfilling the physiological function of KChIP1.     

The voltage-dependent anion channel-1 modulates apoptotic cell death

The role of the voltage-dependent anion channel (VDAC) in cell death was investigated using the expression of native and mutated murine VDAC1 in U-937 cells and VDAC inhibitors. Glutamate 72 in VDAC1, shown previously to bind dicyclohexylcarbodiimide (DCCD), which inhibits hexokinase isoform I (HK-I) binding to mitochondria, was mutated to glutamine. Binding of HK-I to mitochondria expressing E72Q-mVDAC1, as compared to native VDAC1, was decreased by approximately 70% and rendered insensitive to DCCD. HK-I and ruthenium red (RuR) reduced the VDAC1 conductance but not that of E72Q-mVDAC1. Overexpression of native or E72Q-mVDAC1 in U-937 cells induced apoptotic cell death (80%). RuR or overexpression of HK-I prevented this apoptosis in cells expressing native but not E72Q-mVDAC1. Thus, a single amino-acid mutation in VDAC prevented HK-I- or RuR-mediated protection against apoptosis, suggesting the direct VDAC regulation of the mitochondria-mediated apoptotic pathway and that the protective effects of RuR and HK-I rely on their binding to VDAC.     

Estimation of parvalbumin Ca(2+)- and Mg(2+)-binding constants by global least-squares analysis of isothermal titration calorimetry data

The use of competitive isothermal titration calorimetry (ITC) to measure high-affinity binding constants has been largely restricted to systems with a single binding site or multiple identical sites. This study demonstrates the extension of this approach to proteins with two nonequivalent EF-hand Ca(2+)-binding sites--rat beta parvalbumin and the S55D/E59D variant of rat alpha parvalbumin. The method involves simultaneous (global) least-squares analysis of titrations with Ca(2+), with Mg(2+), with Ca(2+) in the presence of Mg(2+), and with Ca(2+) or Mg(2+) in the presence of a competitive chelator (EDTA or EGTA). The Ca(2+) and Mg(2+) binding constants obtained for rat beta agree well with estimates obtained by flow dialysis. Although the Ca(2+) affinity of alpha S55D/E59D is too high to measure by flow dialysis, it was amenable to analysis using the ITC-based approach. The combined S55D and E59D mutations increase the Ca(2+) and Mg(2+) affinities of the mutated binding site by factors of 14 and 26, respectively. This behavior is consistent with that seen previously for the rat beta S55D variant.     

Ca2+ regulation of ion transport in the na+/ca2+ exchanger

The binding of Ca(2+) to two adjacent Ca(2+)-binding domains, CBD1 and CBD2, regulates ion transport in the Na(+)/Ca(2+) exchanger. As sensors for intracellular Ca(2+), the CBDs form electrostatic switches that induce the conformational changes required to initiate and sustain Na(+)/Ca(2+) exchange. Depending on the presence of a few key residues in the Ca(2+)-binding sites, zero to four Ca(2+) ions can bind with affinities between 0.1 to 20 μm. Importantly, variability in CBD2 as a consequence of alternative splicing modulates not only the number and affinities of the Ca(2+)-binding sites in CBD2 but also the Ca(2+) affinities in CBD1.     

Conformational changes in sarcoplasmic reticulum Ca(2+)-ATPase mutants: effect of mutations either at Ca(2+)-binding site II or at tryptophan 552 in the cytosolic domain

By analyzing, after expression in yeast and purification, the intrinsic fluorescence properties of point mutants of rabbit Ca(2+)-ATPase (SERCA1a) with alterations to amino acid residues in Ca(2+)-binding site I (E(771)), site II (E(309)), in both sites (D(800)), or in the nucleotide-binding domain (W(552)), we were able to follow the conformational changes associated with various steps in the ATPase catalytic cycle. Whereas Ca(2+) binding to purified wild-type (WT) ATPase in the absence of ATP leads to the rise in Trp fluorescence expected for the so-called E2 --> E1Ca(2) transition, the Ca(2+)-induced fluorescence rise is dramatically reduced for the E(309)Q mutant. As this purified E(309)Q mutant retains the ability to bind Ca(2+) at site I (but not at site II), we tentatively conclude that the protein reorganization induced by Ca(2+) binding at site II makes the major contribution to the overall Trp fluorescence changes observed upon Ca(2+) binding to both sites. Judging from the fluorescence response of W(552)F, similar to that of WT, these changes appear to be primarily due to membranous tryptophans, not to W(552). The same holds for the fluorescence rise observed upon phosphorylation from P(i) (the so-called E2 --> E2P transition). As for WT ATPase, Mg(2+) binding in the absence of Ca(2+) affects the fluorescence of the E(309)Q mutant, suggesting that this Mg(2+)-dependent fluorescence rise does not reflect binding of Mg(2+) to Ca(2+) sites; instead, Mg(2+) probably binds close to the catalytic site, or perhaps near transmembrane span M3, at a location recently revealed by Fe(2+)-catalyzed oxidative cleavage. Mutation of W(552) hardly affects ATP-induced fluorescence changes in the absence of Ca(2+), which are therefore mostly due to membranous Trp residues, demonstrating long-range communication between the nucleotide-binding domain and the membranous domain.     

Conformational fluctuations of the Ca2+-ATPase in the native membrane environment. Effects of pH, temperature, catalytic substrates, and thapsigargin

Digestion with proteinase K or trypsin yields complementary information on conformational transitions of the Ca(2+)-ATPase (SERCA) in the native membrane environment. Distinct digestion patterns are obtained with proteinase K, revealing interconversion of E1 and E2 or E1 approximately P and E2-P states. The pH dependence of digestion patterns shows that, in the presence of Mg(2+), conversion of E2 to E1 pattern occurs (even when Ca(2+) is absent) as H(+) dissociates from acidic residues. Mutational analysis demonstrates that the Glu(309) and Glu(771) acidic residues (empty Ca(2+)-binding sites I and II) are required for stabilization of E2. Glu(309) ionization is most important to yield E1. However, a further transition produced by Ca(2+) binding to E1 (i.e. E1.2Ca(2+)) is still needed for catalytic activation. Following ATP utilization, H(+)/Ca(2+) exchange is involved in the transition from the E1 approximately P.2Ca(2+) to the E2-P pattern, whereby alkaline pH will limit this conformational transition. Complementary experiments on digestion with trypsin exhibit high temperature dependence, indicating that, in the E1 and E2 ground states, the ATPase conformation undergoes strong fluctuations related to internal protein dynamics. The fluctuations are tightly constrained by ATP binding and phosphoenzyme formation, and this constraint must be overcome by thermal activation and substrate-free energy to allow enzyme turnover. In fact, a substantial portion of ATP free energy is utilized for conformational work related to the E1 approximately P.2Ca(2+) to E2-P transition, thereby disrupting high affinity binding and allowing luminal diffusion of Ca(2+). The E2 state and luminal path closure follow removal of conformational constraint by phosphate.     

Fourier transform infrared spectroscopic study on the binding of Mg2+ to a mutant Akazara scallop troponin C (E142Q)

Troponin C (TnC) is the Ca(2+)-binding regulatory protein of the troponin complex in muscle tissue. Vertebrate fast skeletal muscle TnCs bind four Ca(2+), while Akazara scallop (Chlamys nipponensis akazara) striated adductor muscle TnC binds only one Ca(2+) at site IV, because all the other EF-hand motifs are short of critical residues for the coordination of Ca(2+). Fourier transform infrared (FTIR) spectroscopy was applied to study coordination structure of Mg(2+) bound in a mutant Akazara scallop TnC (E142Q) in D(2)O solution. The result showed that the side-chain COO(-) groups of Asp 131 and Asp 133 in the Ca(2+)-binding site of E142Q bind to Mg(2+) in the pseudo-bridging mode. Mg(2+) titration experiments for E142Q and the wild-type of Akazara scallop TnC were performed by monitoring the band at about 1600 cm(-1), which is due to the pseudo-bridging Asp COO(-) groups. As a result, the binding constants of them for Mg(2+) were the same value (about 6 mM). Therefore, it was concluded that the side-chain COO(-) group of Glu 142 of the wild type has no relation to the Mg(2+) ligation. The effect of Mg(2+) binding in E142Q was also investigated by CD and fluorescence spectroscopy. The on-off mechanism of the activation of Akazara scallop TnC is discussed on the basis of the coordination structures of Mg(2+) as well as Ca(2+).     

Kinetics of conformational changes induced by the binding of various metal ions to bovine alpha-lactalbumin.

The kinetic effects of the binding of various metal ions (Ca(2+), Cd(2+), Co(2+), Mg(2+), Mn(2+), Sr(2+) and Zn(2+)) to apo bovine alpha-lactalbumin has been monitored by means of stopped-flow fluorescence spectroscopy. Our results show that the measured rate constant for the binding of metal ions to the Ca(2+)-site increases with increasing binding constant. This is, however, not the case for metal ions binding to the Zn(2+)-site. The binding experiments performed at different temperatures allowed us to calculate the activation energy for the transition from the metal-free to the metal-loaded state of the protein. These values do not depend on the nature of the metal ion but are correlated with the type of binding site. As a result, we were able to demonstrate that Mg(2+), a metal ion which was thought to bind to the Ca(2+)-site, shows the same binding characteristics as Co(2+) and Zn(2+) and therefore most likely interacts with the residues belonging to the Zn(2+)-binding site.     

Point amino acid substitutions in the Ca2+-binding centers of recoverin. I. Mechanism of successive filling of Ca2+-binding centers

The molecule of photoreceptor Ca(2+)-binding protein recoverin contains four potential Ca(2+)-binding sites of the EF-hand type, but only two of them (the second and the third) can actually bind calcium ions. We studied the interaction of Ca2+ with recoverin and its mutant forms containing point amino acid substitutions at the working Ca(2+)-binding sites by measuring the intrinsic protein fluorescence and found that the substitution of Gln for Glu residues chelating Ca2+ in one (the second or the third) or simultaneously in both (the second and the third) Ca(2+)-binding sites changes the affinity of the protein to Ca2+ ions in different ways. The Gln for Glu121 substitution in the third site and the simultaneous Gln substitutions in the second (for Glu85) and in the third (for Glu121) sites result in the complete loss of the capability of recoverin for a strong binding of Ca(2+)-ions. On the other hand, the Gln for Glu85 substitution only in the second site moderately affects its affinity to the cation. Hence, we assumed that recoverin successively binds Ca(2+)-ions: the second site is filled with the cation only after the third site has been filled. The binding constants for the third and the second Ca(2+)-binding sites of recoverin determined by spectrofluorimetric titration are 3.7 x 10(6) and 3.1 x 10(5) M-1, respectively.     

Allosteric modulation of Euphorbia peroxidase by nickel ions

A class III peroxidase, isolated and characterized from the latex of the perennial Mediterranean shrub Euphorbia characias, contains one ferric iron-protoporphyrin IX pentacoordinated with a histidine 'proximal' ligand as heme prosthetic group. In addition, the purified peroxidase contained 1 mole of endogenous Ca(2+) per mole of enzyme, and in the presence of excess Ca(2+), the catalytic efficiency was enhanced by three orders of magnitude. The incubation of the native enzyme with Ni(2+) causes reversible inhibition, whereas, in the presence of excess Ca(2+), Ni(2+) leads to an increase of the catalytic activity of Euphorbia peroxidase. UV/visible absorption spectra show that the heme iron remains in a quantum mechanically mixed-spin state as in the native enzyme after addition of Ni(2+), and only minor changes in the secondary or tertiary structure of the protein could be detected by fluorescence or CD measurements in the presence of Ni(2+). In the presence of H(2)O(2) and in the absence of a reducing agent, Ni(2+) decreases the catalase-like activity of Euphorbia peroxidase and accelerates another pathway in which the inactive stable species accumulates with a shoulder at 619 nm. Analysis of the kinetic measurements suggests that Ni(2+) affects the H(2)O(2)-binding site and inhibits the formation of compound I. In the presence of excess Ca(2+), Ni(2+) accelerates the reduction of compound I to the native enzyme. The reported results are compatible with the hypothesis that ELP has two Ni(2+)-binding sites with opposite functional effects.     

Interactions among the three structural motifs of the C-terminal region of human thrombospondin-2

The C-terminal regions of thrombospondins (TSPs) contain three elements, EGF-like modules (E), a series of Ca(2+)-binding repeats (Ca), and a C-terminal sequence (G). We have looked for interactions among these elements in four recombinant proteins based on human TSP-2: E3CaG-2, CaG-2, E3Ca-2, and Ca-2. When bound Ca(2+) was assayed by atomic absorption spectroscopy or an equilibrium dialysis protocol in which Ca(2+) was removed from the proteins prior to equilibrium dialysis, E3CaG-2 bound 22-27 Ca(2+), CaG-2 bound 17-20 Ca(2+), and E3Ca-2 and Ca-2 bound 14-20 Ca(2+). Approximately 10 of the bound Ca(2+) in E3CaG-2 were exchangeable. The far UV circular dichroism (CD) spectrum of Ca(2+)-replete E3CaG-2 contained a strong negative band at 203 nm attributable to Ca and a less intense negative band at 218 nm attributable to Ca and G. Chelation of Ca(2+) with EDTA shifted the 203 nm band of all four proteins and the 218 nm band of E3CaG-2 and CaG-2 to less negative positions. The apparent EC50 for the far UV CD transition was 0.22 mM Ca(2+) for all proteins, indicating that Ca(2+) binding to Ca is primarily responsible for the CD change. Near UV CD and intrinsic fluorescence revealed that the tryptophan residues in G are sensitive to changes in Ca(2+). Differential scanning calorimetry of the proteins in 2 mM Ca(2+) showed that E3CaG-2 melts with two transitions, 44-51 degrees C and 75-83 degrees C. The lower transition required G, while the higher transition required Ca. Both transitions were stabilized in constructs containing E3. These results indicate that E3, Ca, and G function as a complex structural unit, and that the structures of both Ca and G are influenced by the presence or absence of Ca(2+).     

Comparison of the structural and dynamical properties of holo and apo bovine alpha-lactalbumin by NMR spectroscopy

In the presence of 0.5 M NaCl at pH 7.1, the Ca(2+)-free apo form of recombinant bovine alpha-lactalbumin (BLA) is sufficiently stabilised in its native state to give well-resolved NMR spectra at 20 degrees C. The (1)H and (15)N NMR resonances of native apo-BLA have been assigned, and the chemical-shifts compared with those of the native holo protein. Large changes observed between the two forms of BLA are mainly limited to the Ca(2+)-binding region of the protein. These data suggest that Na(+) stabilises the native apo state through the screening of repulsive negative charges, at the Ca(2+)-binding site or elsewhere, rather than by a specific interaction at the vacant Ca(2+)-binding site. The hydrogen exchange protection of residues in the Ca(2+)-binding loop and the C-helix is reduced in the apo form compared to that in the holo form. This indicates that the dynamic behaviour of this region of the protein is substantially increased in the absence of the bound Ca(2+). Real-time NMR experiments show that the rearrangements of the structure associated with the conversion of the holo to apo form of the protein do not involve the detectable population of partially unfolded intermediates. Rather, the conversion appears to involve local reorganisations of the structure in the vicinity of the Ca(2+)-binding site that are coupled to the intrinsic fluctuations in the protein structure.     

The Multidrug Transporter LmrP Protein Mediates Selective Calcium Efflux

LmrP is a major facilitator superfamily multidrug transporter from Lactococcus lactis that mediates the efflux of cationic amphiphilic substrates from the cell in a proton-motive force-dependent fashion. Interestingly, motif searches and docking studies suggested the presence of a putative Ca(2+)-binding site close to the interface between the two halves of inward facing LmrP. Binding experiments with radioactive (45)Ca(2+) demonstrated the presence of a high affinity Ca(2+)-binding site in purified LmrP, with an apparent K(d) of 7.2 μm, which is selective for Ca(2+) and Ba(2+) but not for Mn(2+), Mg(2+), or Co(2+). Consistent with our structure model and analogous to crystal structures of EF hand Ca(2+)-binding proteins, two carboxylates (Asp-235 and Glu-327) were found to be critical for (45)Ca(2+) binding. Using (45)Ca(2+) and a fluorescent Ca(2+)-selective probe, calcium transport measurements in intact cells, inside-out membrane vesicles, and proteoliposomes containing functionally reconstituted purified protein provided strong evidence for active efflux of Ca(2+) by LmrP with an apparent K(t) of 8.6 μm via electrogenic exchange with three or more protons. These observations demonstrate for the first time that LmrP mediates selective calcium/proton antiport and raise interesting questions about the functional and physiological links between this reaction and that of multidrug transport.     

Cation- and peptide-binding properties of human calmodulin-like skin protein

Human CLSP, a new Ca(2+)-binding protein specifically expressed in differentiated keratinocytes, is a 15.9 kDa, four EF-hand containing protein with 52% sequence identity to calmodulin (CaM). The protein binds four Ca(2+) ions at two pairs of sites with [Ca(2+)](0.5) values of 1.2 and 150 microM, respectively. Mg(2+) at millimolar concentrations strongly decreases the affinity for Ca(2+) of the two high-affinity sites, but has no effect on the low-affinity sites. The protein can also bind two Mg(2+) ([Mg(2+)](0.5) = 57 microM) at the sites of high Ca(2+) affinity. Thus, as fast skeletal muscle troponin C (TnC), CLSP possesses two high-affinity Ca(2+)-Mg(2+) mixed sites and two low-affinity Ca(2+)-specific sites. Studies on the isolated recombinant N- (N-CLSP) and C-terminal half domains of CLSP (C-CLSP) revealed that, in contrast to the case of TNC, the high-affinity Ca(2+)-Mg(2+) mixed sites reside in the N-terminal half. The binding of cations modifies the intrinsic fluorescence of the two Tyr residues. Upon Ca(2+) binding, hydrophobicity is exposed at the protein surface that can be monitored with a fluorescent probe. The Ca(2+)-dependency of the two conformational changes is biphasic in the absence of Mg(2+), but monophasic in the presence of 2 mM Mg(2+), both corresponding closely to direct binding of Ca(2+) to CLSP. In the presence of Ca(2+), human CLSP forms a high-affinity 1:1 complex with melittin, a natural peptide considered to be a model for the interaction of CaM with its targets. In the complex, CLSP binds Ca(2+) with high affinity to all four binding sites. Isolated N- and C-CLSP show only a weak interaction with melittin, which is enhanced when both halves are simultaneously presented to the model peptide.     

Mapping of three unique Ca(2+)-binding sites in human annexin II.

Site-directed mutagenesis was employed to map and characterize Ca(2+)-binding sites in annexin II, a member of the annexin family of Ca(2+)- and phospholipid-binding proteins which serves as a major cellular substrate for the tyrosine kinase encoded by the src oncogene. Several single amino acid substitutions were introduced in the human annexin II and the various mutant proteins were scored for their affinity towards Ca2+ in different assays. The data support our previous finding [Thiel, C., Weber, K. and Gerke V. (1991) J. Biol. Chem. 266, 14,732-14,739] that a Ca(2+)-binding site is present in the third of the four repeat segments which comprise the 33-kDa protein core of annexin II. In addition to Gly206 and Thr207, which are localized in the highly conserved endonexin fold of the third repeat, Glu246 is involved in the formation of this site. Thus the architecture of this Ca(2+)-binding site in solution is very similar, if not identical, to that of Ca2+ sites identified recently in annexin V crystals [Huber, R., Schneider, M., Mayr, I., R?misch, J. and Paques, E.-P. (1990) FEBS Lett. 275, 15-21]. In addition to the site in repeat 3, we have mapped sites of presumably similar architecture in repeats 2 and 4 of annexin II. Again, an acidic amino acid which is located 40 residues C-terminal to the conserved glycine at position 4 of the endonexin fold is indispensable for high-affinity Ca2+ binding: Asp161 in the second and Asp321 in the fourth repeat. In contrast, repeat 1 does not contain an acidic amino acid at a corresponding position and also shows deviations from the other repeats in the sequence surrounding the conserved glycine. These results on annexin II together with the crystallographic information on annexin V reveal that annexins can differ in the position of the Ca2+ sites. Ca(2+)-binding sites of similar structure are present in repeats 2, 3, and 4 of annexin II while in annexin V they occur in repeats 1, 2, and 4. We also synthesized an annexin II derivative with mutations in all three Ca2+ sites. This molecule shows a greatly reduced affinity for the divalent cation. However, it is still able to bind Ca2+, indicating the presence of (an) additional Ca2+ site(s) of presumably different architecture.     

Calcium-induced refolding of the calmodulin V136G mutant studied by NMR spectroscopy: evidence for interaction between the two globular domains

The Ca(2+) titration of the (15)N-labeled mutant V136G calmodulin has been monitored using (1)H-(15)N HSQC NMR spectra. Up to a [Ca(2+)]/[CaM] ratio of 2, the Ca(2+) ions bind predominantly to sites I and II on the N-domain in contrast with the behavior of the wild-type calmodulin where the C-terminal domain has the higher affinity for Ca(2+). Surprisingly, the Ca(2+)-binding affinity for the N-domain in the mutant calmodulin is greater than that for the N-domain in the wild-type protein. The mutated C-domain is observed as a mixture of unfolded, partially folded (site III occupied), and native-like folded (sites III and IV occupied) conformations, with relative populations dependent on the [Ca(2+)]/[CaM] ratio. The occupancy of site III independently of site IV in this mutant shows that the cooperativity of Ca(2+) binding in the C-domain is mediated by the integrity of the domain structure. Several NH signals from residues in the Ca(2+)-bound N-domain appear as two signals during the Ca(2+) titration indicating separate species in slow exchange, and it can be deduced that these result from the presence and absence of interdomain interactions in the mutant. It is proposed that an unfolded part of the mutated C-domain interacts with sites on the N-domain that normally bind to target proteins. This would also account for the increase in the Ca(2+) affinity for the N-domain in the mutant compared with the wild-type calmodulin. The results therefore show the wide-ranging effects of a point mutation in a single Ca(2+)-binding site, providing details of the involvement of individual residues in the calcium-induced folding reactions.