UPDATE: 14 Nov. 2000

SUBFAMILY CHARACTERISTICS

----- CTER group -----

CALMODULIN (CAM)

CAM is widely distributed among eukaryotes and is inferred to exist in most tissues. It plays a major role in transducing the information in a pulse of messenger Ca2+ ions into a change in conformation and reactivity of a score of target enzymes and structural proteins. Sequences of eight, four of them in Homo, pseudogenes are available. In Arbacia, Arabidopsis, and Petunia two functional isoforms of CAM are expressed.

Preparation and purification

Several procedures have been described for preparing CAM from different tissues [5161]. CAM is heat stable. Heat treatment can be used to remove extraneous protein in tissue extracts. To avoid the heat denaturation step, ammonium sulfate fractionation near the isoelectric point of 4.2 is usually applied. These steps are followed by DEAE ion-exchange chromatography. Phenothiazine derivatives, W-7 or W-5, coupled to Sepharose are used as an affinity matrix in calcium-dependent affinity chromatography. Phenyl-Sepharose is also used for calcium-dependent affinity chromatography and is sometimes applied as the sole purification step following CAM expression in bacterial systems.

Physical characteristics

Vertebrate CAM is 148 residues long; it is N'-acetylated and is trimethylated on Lys115. Its isoelectric point is 4.2. It has eight Phe residues, two Tyr residues and no Trp. It and its many homologs have unique u.v. absorption spectra [107]. CAM has four similar calcium-binding sites, two in the N-terminal lobe and two of higher affinity in the C-terminal lobe. Calcium CAM is extended and dumbbell shaped in the crystal [6828, 6831, 6852, 6849b] and in solution [6839, 2586]. Apo-CAM, as studied by small angle X-ray scattering (SAXS), is a few Å(Angstrom) shorter but still elongated. Helix F2, linker 2,3 and helix E3 form a continuous eight-turn a-helix. The two turns of a-helix in the linker, the handle of the dumbbell, are exposed to solvent on all sides. The linker is easily bent and functions as a flexible tether. When CAM binds to an a-helix of a target protein the linker bends, thereby permitting hydrophobic patches on the N- and the C-lobes to contact hydrophobic residues of the basic, amphiphilic target helix [6840]. Physical and chemical characteristics of all families are listed in Table 2.

Covalent modifications

The N-termini of all CAMs are acetylated. Lysl15 is also trimethylated in all CAMs except those from Dictyostelium, Chlamydomonas, and Trypanosoma. The trimethylation of Lysl15 reduces NAD kinase activation but does not affect cyclic nucleotide phosphodiesterase activation by CAM. CAMs from Paramecium and Euglena have dimethyl-Lys13 and trimethyl-Lys148, respectively. Covalent modification of proteins in all families is compiled in Table 3.

Proteolytic cleavage sites

Metal ion-free CAM is trypsinolyzed at Arg106. Calcium-CAM is cleaved at Lys77 after longer incubation. Cleavage sites and products are listed for all families in Table 4.

Ligands

The numerous CAM-dependent enzymes are inferred to be activated by a common mechanism. A pseudosubstrate sequence near the CAM-binding sequence supposedly blocks the catalytic site. CAM relieves this self inhibition by interacting with the CAM-binding sites [5623, 5624] and removing the pseudosubstrate from the catalytic site. This may be a common regulation mechanism of CAM-dependent enzymes. Noncation ligands for all families of EF-hand proteins are given in Table 5.

TROPONIN C (TNC)

TNC is the calcium-receptor protein of striated and cardiac muscle and imparts calcium sensitivity to muscle. It, together with troponin I (TNI) and troponin T (TNT), form the heterotrimer, troponin. The thin filament of muscle consists of a double helical polymer of globular actin with a linear tropomyosin polymer lying in the two grooves of the actin helix. Troponin is bound in both grooves at the ends of each tropomyosin molecule in the polymer. There is one troponin per seven actin monomers. The conformational changes of TNC induced by the binding of calcium provide a signal that causes an increase in myosin cross-bridge attachment to actin, an increase in actomyosin ATPase, and an increase in muscle tension. TNC consists of 159 amino acids and has four EF-hand domains. In the dendrograms based on both protein and DNA the vertebrate skeletal and cardiac forms cluster separately indicating that the common precursor of all vertebrates already had these two isoforms. Halocynthia, a protochordate, has the most divergent TNC; its TNC clusters near the base of the cardiac branch. Most animals have several isoforms encoded at separate loci.

Preparation and purification

Troponin is purified from skeletal and cardiac muscle acetone powder by high-salt extraction and ammonium sulfate fractionation. TNC is separated from troponin by DEAE-Sephadex ion-exchange chromatography in 6 M urea [5277].

Physical characteristics

TNC has four EF-hand domains, two in the N-terminal lobe and two in the C-terminal lobe. The N-terminal lobe has two low-affinity sites, and the C-terminal lobe two high-affinity sites, to which Mg2+ ions are bound in the resting cell. The binding of two Ca2+ ions to the N-terminal lobe causes a change in the conformation of troponin C. This in turn alters the interaction of troponin with actin, thereby relieving its inhibition of the ATPase activity of myosin [168]. TNC has eleven residues in the linker; hence, the F2 a-helix, linker, E3 a-helix is eight turns long. TNC is dumbbell shaped in the apo form. It has been revealed in the crystal structure that the two lobes of TNC in the Ca4-form enfold the target peptide of TNI, residues 1-47, [PNAS 95:4847-52(98)] in a conformation similar to the CAM*target complex. However, there is no high-resolution structure for TnC/TnI/TnT ternary complex. From the data of neutron scattering of TnC/TnI binary or TnC/TnI/TnT ternary complex, a model was proposed in which TnC is dumbbell shaped similar to crystal structure of Ca2-form [Prot. Sci. 9:1312-26(00)].

Covalent modifications

The N-terminus is acetylated; however, no methylated lysine is found.

ESSENTIAL LIGHT CHAIN (ELC)

Myosin consists of two heavy chains and four light chains. Four types of light chain, L-l, L-2, L-3 and L-4, can be separated by electrophoresis or ion-exchange chromatography. L-1 and L-4 are dissociated at high pH from the heavy chain. These chains participate in myosin ATPase activity and in the binding of filamentous actin by myosin. L-1 and L-4 have been called alkaline, catalytic, enzymatic, or essential light chains [3281]. L-1 and L-4, for a given myosin, are encoded by the same gene; however, differential splicing following differential initiation of transcription produces two mRNAs that encode the same C-terminal 140 amino acids but different N-termini [6369]. Within the myosin hexamer one heavy chain has L-4 and one has L-1 bound to the ahelical tail near the globular head of the heavy chain [6849d]. ELC has four EF-hand domains and is about 200 residues long for L1 and about 150 residues long for L4. The pre-domain 1 sequence is about 50-60 residues long in the L1 form and about 10 in the L4 form. This reflects alternate initiation sites and differential splicing. The precursor to vertebrates ELC has at least four isoforms, with the skeletal form most divergent, then the forms found in smooth muscle and, most closely related, the cardiac atrial and ventricular forms. Molluscan and Drosophila ELCs are more similar to one another than to vertebrate forms. Molluscan ELC is inferred to bind calcium in its first EF-hand; vertebrate ELC's do not bind calcium [PNAS 92:7652(95)].

Preparation and purification

Myosin is usually prepared from muscle by extraction with 0.3 M KCl - 0.15 M KH(P04) pH 6.8 and purified by repeated cycles of dilution and precipitation.

ELC is dissociated from myosin in 5 M guanidineHCl, and fractionated by DEAE ion-exchange chromatography after removal of guanidine-HCl [4829].

Physical characteristics

ELC from vertebrate myosin is inferred not to bind calcium. ELC from scallop myosin has one specific calcium-binding site [2970]. ELC from Physarum myosin also binds calcium and functions as a calcium receptive subunit [2877]. Myosin activity in vertebrates is not regulated by the binding of calcium to myosin. In molluscs calcium binding to ELC activates myosin; in Physarum, calcium binding to ELC inhibits activity.

REGULATORY LIGHT CHAIN (RLC)

The general function and organization of RLCs is similar to those of ELC. RLC has four EF-hand domains and is about 170 residues long. In contrast to the dendrogram for ELC the vertebrate cardiac form(s) is more closely related to the skeletal form. The nonmuscle form from Drosophila is similar to the Homo and Rattus placental and arterial forms; while the muscle form from Drosophila is most divergent relative to molluscs and vertebrates.

Preparation and purification

RLC is dissociated from myosin by incubation in 0.5M NaCI, 25mM Tris-Cl pH 8.1, 10mM DTNB and 2.0mM EDTA, for 10 min at 0 oC. The myosin heavy chain is precipitated by dilution with water, the supernatant concentrated, and excess DTNB removed by gel filtration [5231]. RLC was previously called DTNB light chain.

Physical characteristics

RLC can be phosphorylated at Ser19 (gizzard myosin) by myosin light chain kinase, thereby converting L-2 to L-3. Muscle contraction is initiated by calcium- and CAM-dependent phosphorylation of regulatory light chain of myosin by myosin light chain kinase [212]. This enhances the interaction between actin and myosin. In skeletal muscle, calcium binding to TNC releases the inhibition of actin by troponin I and initiates contraction. RLC has a single Ca2+/Mg2+ site. Due to the high concentration of Mg2+ in the cytoplasm, this site are thought to be occupied by Mg2+ in vivo. In scallop myosin, the removal of RLC causes the loss of the calcium-binding site and of calcium sensitivity of the actin-activated ATPase activity of myosin. The actin-activated ATPase is active in the absence of calcium in the myosin lacking RLC. RLC restores Ca2+ regulation to the actin-activated ATPase and the Ca2+ specific site on the scallop ELC [2970, JBC 270:6773-6778 (95)]. Comparison of the crystal structure of molluscan RLC with a Ca2+ ion bound in EF-hand 1 with both chicken ELC and RLC complexed with the S1 fragment of myosin has permitted Houdusse et al.[Structure 4:21 (96)] to present a model of ELC and RLC complexed with myosin in the active and inactive states.

Covalent modifications

Ser19 of RLC is phosphorylated by MLCK, which is activated by calcium and calmodulin. This is a trigger for smooth muscle contraction.

TROPONIN, NONVERTEBRATE (TPNV)

Troponin, nonvertebrate (TPNV) is found in various nonvertebrates; for example lobster has three isoforms [ABB 291:98 (91)]. It is a close homolog of TNC; however, its mode of function remains unknown and it clusters in a subfamily distinct from TNC.

Preparation and purification

Troponin from arthropod muscle is prepared by methods similar to those used in vertebrates. TPNV is separated by DEAE-Sephadex ion-exchange chromatography in the presence of urea [5178, 6288].

Physical characteristics

TPNV has four EF-hands but only domains 2 and 4 bind calcium [6164]. TPNV interacts with troponin I and troponin T and restores calcium sensitivity to skinned fast-twitch fibers from rabbits. The interdomain linkers of CAM, TNC, ELC, and RLC are bent when those proteins are bound to a target. In contrast, the 2,3 linker of TPNV is predicted to be strongly helical and hence not to bend.

CAM like leaf (CLAT)

Calmodulin-like leaf protein from Arabidopsis thaliana (CLAT) is congruent with CTER but differs from CAM in that it has a 47 residue domain of unknown homology at its C-terminus [Plant Mol. Biol. 22:207 (93)]. Both the 1,2 and the 3,4 linkers are seven residues long as are the 1, 2 and 3,4 linkers of CAM. However, the linker connecting lobe 1,2 and lobe 3,4 consists of only TYSEK; whereas CAM is MKDTDSEE.

SQUIDULIN (SQUD)

SQUD is a high-affinity calcium-binding protein purified from squid optic lobe [5211a]. It is found also in squid giant axon, but not in squid muscle.

Preparation and purification

SQUD is purified from homogenates of squid optic lobe by calcium-dependent affinity chromatography using fluphenazine-Sepharose [429].

Physical characteristics

SQUD binds four equivalents of calcium. Calcium binding induces a net exposure of tyrosine residues to an aqueous environment. Magnesium binding induces a net burial of tyrosine. SQUD binds one equivalent of chloropromazine at 10 _(micro)M chloropromazine and in the presence of 0.1 mM CaCl2. It binds the bee venom peptide melittin in a Ca2+-dependent manner; however, it does not activate CAM-dependent phosphodiesterase [429, 6613]. SQUD co-purifies with the 235 kDa myosin of squid axoplasm which is an unconventional myosin, and is inferred to function as a myosin light chain [PNAS 93:6064 (96)]. Squidulin (SQUD) is identical to vertebrate CAM in 101 of 149 positions; however, its Lys115 is not methylated; Lys115 of CAM is trimethylated. Further, it has a Pro at the second residue of the 2-3 linker, which is helical in CAM when it is not bound to a target.

CDC31 AND CALTRACTIN (CDC)

Caltractin (CDC) was initially cloned from the unicellular green alga, Chlamydomonas. It is a structural component of the basal body complex, the major microtubule organizing center in Chlamydomonas and the functional homolog of the centrosome in the animal cell [6256]. The CDC31 gene product, a yeast homolog of caltractin, interacts with the KAR1 product. KAR1 encodes an essential component of the yeast spindle pole body that is required for karyogamy and duplication of the spindle pole body [6124, JCB 125:843 (94)]. CDC31 is 161 residues long and caltractin is 169. Both are congruent with CTER. Chlamydomonas mutant vfl2, E101K (101 = 3.- 1) has structural defects in the nucleus-basal body connectors, the distal striated fibers, and the flagellar transition regions and is defective for basal body localization and/or segregation [4536].

Preparation and purification

Caltractin is purified from the isolated basal body complex of Chlamydomonas by calcium-dependent binding to phenyl-Sepharose CL-4B [6256].

Physical characteristics

Caltractin associates with a calcium-sensitive contractile fiber system of the basal body complex [4536].

CALl (CAL)

The cal-l gene was isolated from the nematode, Caenorhabditis elegans, by hybridization with a human CAM cDNA probe [6423]. There is one copy of the cal-l gene per haploid genome. The encoded protein contains four EF-hands in 161 residues; it is similar to, but distinct from, CAM and TNC [6423]. The protein has not been characterized. Jamieson et al. [6272] purified CAM and a TNC-like protein from Caenorhabditis. This CAM has trimethyl-lysine and no Cys; the TNC has Cys but no trimethyl-lysine. The call product has a single Cys and lacks Lys at the position that corresponds to trimethyl-lysine of CAM. The amino acid compositions and the expected sizes of the BrCN peptides are different between the TNC-like protein and the call product. CAL is congruent with CTER.

CAST (CAST)

The pCAST clone from potato DNA encodes a four EF-hand protein, preceded by 38 residues [Plant Physiol. 101:1405 (93)]. It may have evolved from the OD*EV precursor of CTER. Its 2,3 linker is 24 residues long. It is not in the same subfamily as PMAT but it is interesting to note that both contain (Ser)5 in their N-terminal domains. The functions of CAST are not known. A cDNA coding for a birch pollen allergen, Bet v III, shows high similarity to CAST. It preferentially expressed in mature pollen. The binding to allergc patients' IgE of the allergen depends on the native conformation of protein and requires protein-bound calcium [EMBO J. 13:3481-3486 (94)].

----- CPV group -----

CALCINEURIN B (CLNB)

Calcineurin B (CLNB) is the regulatory subunit and calcineurin A is the catalytic subunit of a protein phosphatase. Calmodulin also binds to calcineurin A; both CLNB and CAM in their calcium forms are required for catalytic activity. CLNB is about 170 residues long and has four EF-hand domains. The crystal structure of Ca4-CLNB, complexed with an _-helical fragment (350-370) of calcineurin A and with the drug complex, FKBP12-FK506, [Nature 378:641 (95), Cell 82:507 (95)] shows both lobe 1,2 and lobe 3,4 embracing the fragment in their grooves homologous to those of CAM and rotation axis. The seven subfamilies (CLNB, P22, VIS CALS, DREM, CMPK, and SOS3) are reasonably congruent.

Preparation and purification

Calcineurin is usually prepared from brain by CAM-Sepharose affinity chromatography. The A and B subunits of calcineurin are separated by DEAE-Sephacel ion-exchange chromatography in the presence of dithiothreitol and urea [5235, 6105].

Physical characteristics

Calcineurin is a major CAM-binding protein in the brain. The B subunit of calcineurin is very similar to CAM. However, neither protein can replace the other in the the reconstitution and activation of calcineurin [3348].

Covalent modification

The N-terminus of CLNB is N-myristoylated [6105].

P22 (P22)

P22 (p22) is required for constitutive, exocytotic membrane traffic in rat liver [JBC 271:10183(96)].

Physical characteristics

The N-myristoylate group and the calcium induced change in conformation are essential to the function of P22 [JBC 271:10183(96)]. It is reported that P22 inhibits the GTPase stimulated, Na+/H+ exchanger [PNAS 93:12631 (96)] and calcineurin phosphatase activity[JBC 274:36125-31 (99)]. p22 associates with microtubules, which is N-myristoylation-dependent of p22, but does not involve p22's Ca2+-binding activity [Mol. Biol. Cell 10:3473-3488 (99)].

Covalent modification

p22 is N-myristoylated.

VISININ AND RECOVERIN (VIS)

Visinin and recoverin are retinal cone cell-specific proteins. Guanyl cyclase activating protein (GCAP) regulates guanylate cyclase activity. In the dark, cationic channels are kept open by bound cGMP. The Ca2+ enters through this channel, and this influx is matched by efflux through a Na+-K+, Ca2+ exchanger. Light activates an enzyme cascade that stimulates cyclic GMP hydrolysis, leading to channel closure, which blocks calcium influx but not its efflux. Free calcium in the cell drops within 0.5 s of light stimulation. Restoration of the dark state requires cGMP resynthesis by guanylate cyclase. Guanylate cyclase activity increases when Ca2+ is lowered to less than 100nM. Recoverin and its homolog visinin and S-modulin are inferred to extend the life time of photo-activated rhodopsin by inhibiting rhodopsin kinase at higher calcium concentration., which results in longer life time of phosphodiesterase that hydrolyzed cGMP [TINS 19:547 (96)]. However, the precise function of recoverin is still controversial. VIS is 190-200 residues long and has four EF-hand domains. VIS and CMSE are both unusual in that domains 1 and 2 and domains 3 and 4 more closely resemble one another; this is not consistent with the simple model of duplication and fusion of a two-domain precursor. All three interdomains 1,2 and 2,3 and 3,4 are different from one another and from all other interdomains. Recoverin is N-myristylated. In the crystal structure [Cell 75:709 (93)] of nonmyristoylated recoverin a single Sm3+ ion is bound in place of Ca2+ in EF-hand 2. The hydrophobic patches of lobe 1,2 and of lobe 3,4 are exposed and orientated so that their two clefts could not enfold a target helix as seen in CAM, TNC, ELC, RLC and BM40. NMR studies [JBC 270:4526 (95)] showed that the myristoyl side chain is buried between the two lobes in the apo-form but is exposed following calcium binding to EF-hands 2 and 3.

Preparation and purification

Recoverin is purified from bovine retinas by low ionic strength extraction, DEAE-cellulose chromatography, immobilized rhodopsin column chromatography, gel chromatography on Superose 12, and ion-exchange chromatography on Mono-Q [6176].

Physical characteristics

GCAP differs from calcium-dependent activators such as CAM and TNC, which require calcium to activate their targets. GCAP binds and activates its target following release of its calcium ions [TINS 19:547 (96)].

Covalent modifications

The N-terminus of recoverin is acylated. The most abundant acyl group is myristoylate (14: 1), but 14: O, 14:2, and 12:0 acyl residues are also present [589].

CALSENILIN (CALS)

Calsenilin (CALS) binds both presenilin 1 (PS1) and presenilin 2 (PS2) in a calcium dependent manner. Most early onset familial Alzheimer cases are caused by mutations in PS1 or PS2. CALS may mediate the role of PS2 in apoptosis and _-amyloid formation [Nature Med. 4:1177 (98)]. CALS is 251 residues long and contains four EF-hands. No homolog of its first 80 residues has been identified. The sequences of CALS and of DREM are quite similar and all domains cluster together. They are tentatively placed in different subfamilies because of the apparent differences in function.

DRE antagonist modulator (DREM)

The downstream regulatory element (DRE) is known to silence transcription of the prodynorphin and c-fos genes. DRE-antagonistist modulator (DREAM) binds to DRE in the apo-form, probably as a homotetramer. "Upon stimulation by Ca2+, DREAM's ability to bind to the DRE and its repressor function are prevented." [Nature 398:80 (99)]. DREM is 284 residues long and contains four EF-hands. The 110 residues preceding the first EF-hand may be responsible for binding to DNA; however, no homolog is recognized. The first and second EF-hands have unusual distributions of potential Ca2+ binding ligands - X, Y, Z, -X, -Z Asn, Cys, Thr, Asp, Thr and Asp, Asp, Asn, His, Asp. In proteins Ca2+ is usually coordinated by O; the S of Cys and the N of His may coordinate water which in turn binds Ca2+. The third and fourth EF-hands appear to bind calcium.

CAM dependent protein kinase (CMPK)

The calcium/calmodulin-dependent protein kinase (CMPK) of Lilium sp. has 520 residues and consists of a serine/threonine protein kinase domain, a calmodulin binding region, and four EF-hands. CMPK is preferentially expressed in anthers [PNAS 92:4897 (95)]. The mechanism and role of its regulation by calcium is not known.

CALCIUM SENSOR HOMOLOG (SOS3)

Mutations in the sos3 gene make A. thaliana hypersensitive to increased Na+ and of lowered K+ in the bathing medium [Science 280:1943 (98)]. Although SOS3 is congruent to CLNB, P22, VIS, CALS, DREM, and CMPK, none of its four EF-hands are inferred to bind calcium.

----- Pairings -----

RETCULOCALBIN (RTC)

Reticulocalbin (RTC), also known as E6BP (E6 binding protein) or ERC-55, appears to bind E6, a transforming protein from human papilloma virus that interacts with tumor suppressor protein p53. RTC has 210 amino acids [Science 269:529 (95)]. It contains six EF-hands, several of which are most similar to the corresponding domains of SCF. RTC is inferred to be in the lumen of the endoplasmic reticulum because it has an N-terminal leader sequence and a glycosylation site. The C-terminal sequence is HDEL; this is similar to the KDEL that serves as a signal to retain proteins within the ER. Four of six EF-hands (1, 4, 5 & 6) bind calcium but 2 & 3 do not [JB 121:145 (97)].

DNA SUPERCOILING FACTOR (SCF)

DNA supercoiling factor (SCF) is capable of generating negative supercoils into a relaxed DNA in conjunction with eukaryotic DNA topoisomerase II. SCF has 322 residues and contains six EF-hands, five of which are predicted to bind calcium. SCF requires ~10 _(micro)M Ca2+ for activity. This is inferred to reflect an in vivo response to a calcium signal [JBC 270:15571 (95)]. Several domains of SCF and RTC are most closely related; however, there is no apparent relation in their functions nor homology in the domains at their N-termini.

CALPAIN (CALP)

Calpain (CALP) (EC 3.4.22.17) is an intracellular cysteine protease that requires calcium for activation. The large subunit is about 700 residues long and contains a domain homologous to sulfydryl protease and five EF-hand domains at the C-terminus. The small subunit is about 270 residues long and contains a Gly-rich domain at its N-terminus and five EF-hands. In both subunits domains 1 & 2 and 3 & 4 form two lobes. The fifth domains of large and small subunits pair, as usually found in adjacent EF-hands, to join the two in the heterodimer [Nature Struct. Biol. 3:67 (96)]. Several isoforms are known.

Preparation and purification

Calpain is usually purified from muscle extracts made in the presence of EDTA. Subsequent steps are DEAE-cellulose ion-exchange chromatography, phenyl-Sepharose hydrophobic interaction chromatography, and gel filtration [5223].

Physical characteristics

Calpain consists of a 80kDa catalytic subunit and a 30kDa regulatory subunit. The calcium sensitivities of both m- and _(micro)-calpain is reduced significantly by phospholipids. It is supposed that calpain is activated on cell membranes. Many cells contain sufficient amounts of endogenous inhibitor, calpastatin, to suppress calpain activity. Calpain cleaves cytoskeletal proteins, and enzymes such as protein kinase C and CAM-dependent protein kinase II from which regulator independent activated forms are produced. The function of this protease remains unknown [195].

Proteolytic cleavage sites

The autolysis of the large subunit of calpain at its N-terminus increases its calcium sensitivity.

Covalent modifications

The active site of calpain (Cys107 in the large subunit of Gallus CALP) reacts specifically with iodoacetic acid in a calcium-dependent manner [872a].

SORCIN (SORC)

Sorcin is a small phosphorylated cytosolic protein overproduced by many multidrug-resistant cells [6490]. Grancalcin is a homodimer with apparent mass of 55 kDa by gel filtration [4618]. It is abundant in neutrophils and monocytes. The function(s) of sorcin and grancalcin are unknown; they are placed in the calpain subfamily because of the inferred common origin of their five EF-hands. Sorcin and grancalcin are similar to calpain small subunit. Sorcin has Gly-rich sequence at N-terminus. It is not known whether sorcin or grancalcin associates with some sort of large subunit.

Preparation and purification

Sorcin is purified from the actinomycin D-resistant mouse cell-line QUA/ADj in which ~ 1% of the total soluble protein is sorcin. The extract is fractionated on a TSK G3000Sw column and sorcin is purified by reversed-phase liquid chromatography [6490]. Grancalcin is purified from neutrophils by ammonium sulfate fractionation, Mono-Q ion exchange chromatography, and Sephadex G-75 gel filtration [4618].

Physical characteristics

Sorcin and grancalcin are similar to the small subunit of calpain. A role for sorcin in the development of multidrug-resistant cells has not been established. Grancalcin shows a calcium-dependent translocation to the granules and membranes of neutrophils [6490, 6134, 4618]. Sorcin (SORC) or grancyclin is a homodimer, congruent with CALP, and inferred to dimerize via its fifth EF-hand. Its concentration is greatly increased in multidrug resistant cells. It has been shown to bind to the ryanodine receptor [JBC 272:25333(97)] and to the N-terminal domain of annexin VII (synexin) [JBC 272:22182(97)].

S100 (S100)

The members of the S100 subfamily have diverse characteristics. However, they all have two EF-hand domains the first of which shows a characteristic insertion of two residues. At least 16 groups have different names and perhaps different functions. We summarize the names of the genes, synonyms, and the new names, as proposed by Heizmann et al. (???personal communication) in Table 6. Genes S100A1 through S100A13 in humans are located on chromosome 1 at lq21; whether they are subject to coordinate expression has not been determined. SlOO_ is at 21q22.3; S100P(S100E) is at 4p16. S100A14 is at 7q22-q31.1. S100 subfamily proteins range from 90 to 110 residues long. They have two EF-hand domains. The two-domain precursor may have branched from OD*EV.

Preparation and purification

S100a (S100A1 homodimer) and S100b (S100A1/S100B heterodimer) are purified from brain by ammonium sulfate fractionation, DEAE-Sephadex ion-exchange chromatography, and Sephadex G-75 gel chromatography [5225].

Calpactin I (S100A10) is enriched by extraction from brush border-derived membrane vesicles with Triton X-100 in the presence of EDTA after extraction in the presence of calcium. Further purification is performed by DEAE-cellulose ion-exchange chromatography and gel filtration on a Sephadex G-I00 column. The light chain (p10) is separated from the heavy chain (p36) of calpactin by chromatography on DE-52 cellulose in the presence of 8 M urea [5512]. S100P (S100E) and S100L (S100A2) are purified by calcium dependent affinity chromatography on phenyl-Sepharose, followed by ion-exchange chromatography on a Mono-Q column [5179, 6220]. Calgizzarin (S100A11) is purified by calcium-dependent affinity chromatography on a W-7 column [6484].

MRP-14 (S100A9) and MRP-8 (S100A8) (macrophage migration inhibitory factor-resembling protein) are purified from bovine neutrophil cytosol by using DEAE Sephacel ion-exchange chromatography [6175].

Physical characteristics

Most have been demonstrated to be dimers. S100a (S100A1)2 regulates giant protein kinase [Nature 380:636 (96)]. The NMR solution structure of (S100B)2 free and complexed with the C-terminal peptide (367-388) from protein p53 shows the peptide interacting primarily with a-helix F2 and of one monomer and a-helix E1 of the other monomer [Biochem. 37:1951 (98)]. S100A9 (p11) forms a heterotetramer with annexin II; the crystal structure of (S100A9)2 complexed peptide 1-13 from annexin II shows a similar orientation of the N-terminal a-helix of the annexin II lying in the 1-2 linker, a-helix F2 groove of one subunit with the N-acetyl group against a-helix E1 of the other subunit [Nature Struct. Biol. 6:89 (99)]. Certain neurons are characterized by their contents of S100, as well as contents of parvalbumins. Several S100's function in culture as neurite extension factors [BioMetals 11:383 (98)]; whether S100 functions extracellularly under non-pathological conditions has yet to be established. S100 may bind to the cytoskeleton. It has also been reported to activate fructose-l,6-bisphosphate aldolase and to facilitate neurite outgrowth [6100, 2860].

INTESTINAL CALCIUM BINDING PROTEIN (ICBP)

Intestinal calcium binding protein (ICBP) is strongly congruent with S100 and has a similar structure. It is placed in a different subfamily because it is a monomer; its encoding gene is on the X chromosome; and it is inferred to function in transcellular transport of calcium [Cell Tissue Res. 285:477 (96)].

Preparation and purification

ICBP is purified from intestine by Sephadex G-100 gel filtration, and preparative slab gel electrophoresis [1575]. DEAE-Sephadex ion-exchange chromatography using EDTA as the first step and calcium as the second step is another effective procedure [5218].

Physical characteristics

ICBP is a monomer. C-terminal extension of ICBP is shorter than that of S100, by which interaction S100 dimer is stabilized.

TRICHOHYLALIN & PROFILAGGRIN (HYFL)

Profilaggrin has two EF-hands at its N-terminus. They are encoded by the second and third exons; the first is not translated. It is congruent with the S100 subfamily. The first domain is closely related to the unique first domains of the S100 subfamily and the second domain is most similar to the second domains of the S100 subfamily. Trichohyalin also contains two EF-hands at its N-terminus. Profilaggrin and trichohyalin are placed in a separate subfamily because they both have 2000 residues at their C-termini that associate with keratin intermediate filaments. Profilaggrin contains 10 to 12 tandemly repeated filaggrin units [Mol. Cell. Biol. 13:613 (93)]. Trichohyalin contains seven tandemly repeated domains that are homologous to involucrin [JBC 268:12164 (93)]. Although the elongated tandem repeats of profilaggrin and of trichohyalin are not inferred to be homologous, the two are placed in the same subfamily because of the similarity of their two N-terminal EF-hand domains.

DIACYLGLYCEROL KINASE (DGK)

Diacylglycerol kinase (DGK) (EC 2.7.1.107) regenerates phosphatidylinositol from diacylglycerol during signal transduction. DGK catalyses the reaction between ATP and 1,2-diacylglycerol to form ADP and 1,2-diacyl-sn-glycerol 3-phosphate (phosphatidic acid) that is converted to CDP diacylglycerol by phosphatidate cytidyltransferase. Inositol and CDP dacylglycerol react to form phosphatidylinositol, then it is phosphorylated to form phosphatidylinositol-4,5-biphosphate that is hydrolyzed by PI-specific phospholipase C to produce the two intracellular messengers, inositol 1,4,5-triphosphate (calcium release from intracellualr stores) and sn-1,2-diacylglycerol (Protein kinase C activation). The cycle is called phosphatidylinositol cycle. DGK is 735 residues long. Its two EF-hands span residues 114-202. The domain at the N-terminus is not similar to other known protein families. The domain at the C-terminus contains a putative ATP-binding sites and two cysteine-rich zinc-finger sequences similar to those found in protein kinase C. Isoforms a, b and g show varying levels of calcium sensitivity [Biochem. J. 321:59 (97)].

Preparation and purification

DGK is purified by DE-52 ion-exchange chromatography, ammonium sulfate fractionation, Sephadex G-150 gel chromatography, ATP-agarose column chromatography and hydroxyapatite column chromatography. ATP-agarose column chromatography is an especially effective step of purification [4113, 6428].

Physical characteristics

DGK has an Mr = 80,000. The catalytic domain is at the C-terminus. There are two EF-hand domains in the middle, and a short region at the N-terminus of unknown homology or function. DGK can be activated by phosphatidylserine, deoxycholate, and sphingosine. Pig DGK is activated by calcium only in the presence of activator such as deoxycholate or sphingosine, while human DGK is not calcium dependent [6421, 4113, 6428].

NUCLEOBINDIN & NEFA (NUBN)

Nucleobindin (NUBN) was initially identified in KML1-7 cells from MRL/1 mice, which are prone to develop lupus erythematosis and produce elevated levels of anti-dsDNA and anti-ssDNA antibodies [BBRC 187:375 (92)]. NUBN has 455 residues. There is a signal sequence at the N-terminus followed by a two hundred residue domain of unknown homology, two canonical EF-hands, then seven heptad repeats as found in a coiled coil (leucine zipper). The first EF-hands of NUBN and of CRGP are most closely related to one another. The nuclear EF-hand acidic (NEFA) gene product is closely related to NUBN; its function is unknown [BBA 1407:84 (98)].

CALMODULIN-RELATED GENE PRODUCT (CRGP)

CRGP is encoded by the gene T+, which contains a transposon-like human repeat element, THE 1, in the 3'-untranslated region of its message [6172]. CRGP is 79 residues long. It has not been characterized.

a-ACTININ (ACTN)

ACTN is an actin filament cross-linking protein. It is a homodimer whose subunits assemble in an antiparallel fashion to form a rod-like structure [JMB 230:196 (93)]. ACTN is found in a wide variety of organisms in both muscle and nonmuscle cells. The difference between nonmuscle and muscle forms is that the cross-linking activity of the former is completely inhibited by micromolar calcium, while the activity of the muscle form is not calcium sensitive. Both forms have two EF-hand domains at the C-terminus. The N-terminal domain binds actin. ACTN is 890 residues long.

Preparation and purification

ACTN is extracted in low ionic strength buffer. It precipitates in a high ionic strength solution of

NaCl or of KCI, as used in its preparation. ACTN from platelets is purified by DE-52 cellulose ion-exchange chromatography and hydroxyapatite chromatography [1992, 2991].

Physical characteristics

Nonmuscle type ACTN binds two equivalents of calcium per subunit. Muscle type ACTN is inferred, from its sequence, not to bind calcium [6110].

a-SPECTRIN (FDRN)

a-Spectrin (FDRN) was first described in the erythrocyte in which it reinforces the cell membrane by cross linking ankyrin and protein 4.1. Nonerythroid aII-spectrin and its close homolog, a-fodrin, form elongated double stranded filaments. Each strand consists of segments of triple stranded a-helices that are formed by the strand going forward, back, forward. The two EF-hands at the C-terminus impart a calcium sensitive [Biochem. 23:7199 (97)] interaction with actin and other cytosolic proteins [Cell 98:537 (99)]. The first domains of ACTN and FDRN are most closely related; their second EF-hands are also quite similar to one another. Portions of the non-EF-hand regions of ACTN and FDRN appear to be related; the two are placed in distinct subfamilies because of their different interactions with the cytoskeleton.

Preparation and purification

Spectrin is extracted from hemoglobin-free erythrocyte membranes by incubating it at 37oC in 5-10 volume of extraction buffer (0.1mM EDTA, 0.5mM 2-mercaptoethanol, 0.2mM DFP, pH 9.0). The membranes are removed by centrifugation. The suprenatant includes primarily spectrin and actin. Spectrin hetrodimer (aplha and beta subunits) is purified by Sepharose CL-4B gel chromatography. Spectrin dimer dissociates into alpha and beta subunits at alkaline pH (> pH 10.5) or in the presence of 3M Urea. Both subunits can be separated by Mono-Q ion-exchange chromatography [JBC 267:14775-82 (92), Biochem. 37:272-80 (98)]

FAD-DEPENDENT, MITOCHONDRIAL GLYDEROL-3-PHOSPHATE DEHYDROGENASE (GPD)

FAD-dependent, mitochondrial glycerol-3-phosphate dehydrogenase (EC 1.1.99.5) is located in the outer surface of the inner mitochondrial membrane and exists as a single subunit. Mammalian enzyme shows strong homology to yeast and bacterial FAD-dependent glycerol phosphate dehydrogenase and has two EF-hands at its C-terminus unlike yeast and bacterial enzyme [JBC 269:14363 (94)].

ALLGRAFT INFLAMMATORY FACTOR (AIF1)

Allograft inflammatory factor-1 (AIF1) is "a cytokine inducible, tissue-specific transcript transiently expressed in response to vascular trauma." [BBRC 228:29 (96)]. It has 143 residues with two EF-hands. Homologs of the first fifty residues have not been identified. The first domains of AIF1 and of GPD are most similar. The arrangement of possible calcium binding ligands (S-G-G---S--D) of the second EF-hand is unusual.

OSTEONECTIN, SPARK (BM40)

BM-40 (BM40), also known as osteonectin or SPARC (secreted protein acidic and rich in cysteines), consists of a short stretch of acidic amino acids, a region homologous to follistatin and an extracelluar calcium (EC) binding module. At least four other extracellular glycoproteins -- testican, tsc, QR1 and SC1 -- all of unknown function, have homologous follistatin and EC modules. The latter half of the EC module, residues 209-286, consists of a pair of EF-hands (a-helices D, E, F & G of the EC module) as confirmed in the crystal structure [Nature Struct. Biol. 3:67 (96)]. The second EF-hand binds calcium in canonical coordination. The first EF-hand binds calcium in the third (after RLC of molluscs and S100) known variation on the canonical pattern of calcium binding. Since BM40 is extracellular, calcium binding supposedly imparts stability, not information, to it. The cleft between the two EF-hands is occupied by the first a-helix of the EC module, 140-160, in an orientation very similar to that of a target helix bound by calmodulin; this appears to be another example of stabilization, not information transduction.

Preparation and purification

Osteonectin was extracted from bone with 0.5M EDTA, then purified by using gel filtration and polyanion anion-exchange fast protein liquid chromatography [Biochem. J. 253:139-51 (88)]. BM40 was purified from basement-memebrane-producing mouse EHS tumor [EJB 161:405-64(86)]. SPARC is a major product of mouse embryo parietal endoderm [EMBO J. 5:1465-72 (86)]. SPARC wa purified from primary Sertoli cell enriched culture medium by sequential anion-exchange, gel-permeation, C4 reversed phase, and diphenyl reversed phase HPLC [BBRC 167:1393-9 (90)].

Physical characteristics

BM40 binds calcium and several types of collagen at the EF-hand and the alpha-helical domains in C-terminal half of the molecule. It also binds plasminogen. Am 80 kDa protein purified from bone cross-links to BM40 with disulfide [BJ 330:1423-31 (98)].

Covalent modification

N-linked glycosylation in BM40 is observed. There are two potential N-glycosylation site in human osteonectin but only one site (Asn99-X-Thr-101) is glycosylated [JBC 270:23212-7 (95)].

QR1 & SC1 (QR1)

QR1, named from the cDNA, QR1, cloned from the neuroretina of quail, Coturni. coturnix has 676 residues. Its C-terminal portion is similar to the extracellular glycoprotein, BM40. QR1 and BM40 are the only examples of (normally) extracellular EF-hands proteins. Transcription of QR1 takes place only during the late phase of retinal development and is shut off sharply at hatching. [PNAS 88:4503 (91)]].

----- SELF -----

NEMATODA CALCIUM BINDING PROTEIN (EF12)

A 542 residue protein of unknown function has been isolated from the nematode, Onchocerca volvulus, as well as its close homolog of 482 residues from Caenorhabditis elegans [Infec. Immun. 8:3491-501 (00)]. It has twelve EF-hands, eleven of which conform well to the consensus of EF-hand and appear to bind calcium. The fourth EF-hand is highly diverged; however, of all EF-hand domains it most closely resembles EF12 #2. The eleven interdomain spacings are all six residues; this pattern strengthens the interpretation that the 41 (6 + 29 + 6) residues between EF-hands 3 and 5 harbors an aberrant domain. In a dendrogram the six ODD domains cluster together and are congruent with the six EVEN domains. Further domains 1(2) & 7(8), 3(4) & 9(10), and 5(6) & 11(12) most closely resemble one another, reflecting four cycles of gene duplication and fusion. It is surely significant that the O. volvulus and the C. elegans proteins align well, with no deletions, over the 414 residues (12 * 29 + 11 * 6) residues encompassing the twelve EF-hands but show many deletions in the N-terminal and C-terminal domains. The six pairs of EF-hands might be arranged helically and undergo a cooperative change in conformation upon binding calcium.

LYTECHINUS PICTUS SPEC-RESEMBLING PROTEIN (LPS)

LPS consists of eight EF-hand domains. Domains 1 and 5, domains 2 and 6, domains 3 and 7, and domains 4 and 8 are similar. LPS evolved by a recent duplication of a four-domain precursor. These four pairs of domains are similar to the four domains of SPEC; however, they are not the closest neighbours. The postulate that LPS evolved by gene duplication and fusion from the ancestor of SPEC is strengthened by the similar distribution of introns in the two halves of LPS and in SPEC [6516]. The function of LPS is not known.

CALBINDIN 28 kDa AND CALRETININ (CLBN)

CLBN is found in several epithelial tissues, and in many neurons. Its synthesis is induced by vitamin D and is associated with calcium absorption in intestine. However, it is not congruent with ICBP, sometimes referred to as calbindin D9k. Calretinin is found mainly in a variety of neurons. It is also found in developing rat ovary. Calretinin is similar to CLBN and is placed in the same subfamily; however, the functions of both proteins remain unknown.

Domains 1, 3 & 5 and domains 2, 4 & 6 are more closely related. Interdomains 1,2 and 3,4 and 5,6 are similar and are inferred to be homologous. CLBN appears to have evolved by two duplications and fusions from a two domain precursor. The six domains, individually synthesized, aggregate in pairwise combination to approximate the spectral properties of the intact CLBN [Prot. Sci. 6:2385 (97)].

Preparation and purification

CLBN is characterized by its increased electrophoretic mobility in nondenaturating gels in the presence of EDTA. It is heat stable; however, heat treatment can be circumvented by ammonium sulfate precipitation. CLBN is further purified by the calcium-dependent elution on DEAE cellulose [1726, 5252].

Calretinin is partially purified from brain homogenate by precipitation in 60-80% saturated ammonium sulfate. Further purification is carried out by combination of ion-exchange chromatography and gel-filtration chromatography [6067].

Physical characteristics

CLBN and calretinin have six EF-hand domains. Both proteins can bind four equivalents of calcium [276].

EP15 (EP15)

The gene product of eps15 (EP15) encodes a protein of 897 residues. The first domain of 282 residues contains a candidate phosphorylation site as well as six putative EF-hands, two of which are inferred to bind calcium. EP15 is phosphorylated on tyrosine following EGFR activation by EGF in vivo, or directly by EGFR in vitro [Mol. Cell. Biol. 13:5814 (93)].

TETRAHYMENA CALCIUM BINDING PROTEIN (TCBP)

Ciliary reversal in Tetrahymena pyriformis is calcium-dependent. TCBP is inferred to be the calcium-binding protein that regulates ciliary reversal [5267]. TCBP-23 is 207 and TCBP-25 218 residues long. Although they are only 35% identical in amino acid sequence, their eDNA sequences are 49% identical [6472]. Domains 1 and 3 and domains 2 and 4 more closely resemble one another and interdomains 1,2 and 3,4 are similar. The origin of the inferred two-domain precursor is obscure.

Preparation and purification

TCBP is purified from Tetrahymena cilia extracts by heat treatment, ammonium sulfate fractionation, DEAE-cellulose ion-exchange chromatography, and hydroxyapatite column chromatography. TCBP is also purified by calcium-dependent binding to phenyl-Sepharose 4B [5267, 6472].

Physical characteristics

TCBP is inferred to function in ciliary reversal. It shows a calcium-dependent change in electrophoretic mobility shift [5267].

P26OLF (P26)

P26olf (P26) was prepared from the olfactory epithelium of Rana catesberiana [BBRC 251:860-7 (98)]. P26 has 217 residues. The two putative EF-hands in the first half of P26 are most similar to the two in the second half; as in TCBP this reflects a recent gene duplication. EF-hand 3 is separated from EF-hand 2 by 36 residues that are 60% identical to the C-terminal twenty residues of a rabbitt S100. The expectation value of 0.97 indicates that the alignment is not significant for the entire database but this similarity with another EF-hand protein is interesting. None of the EF-hands are inferred to bind calcium in canonical conformation. However, P26 is purified by elution from phenyl Sepharose (following initial binding in 1.0 mM [Ca2+]) with EGTA. It is possible that domains 1 and 3 have an unconventional coordination as found in domains 1 of S100, ICBP, and HYFL.

Preparation and purification

P26Olf was extracted from frog olfactory epithelium. It was purified by calcium-dependent binding to phenyl Sepahrose, DEAE Sepharose CL-4B column chromatography, gel filtration HPLC and Mono-Q ion-exchange column chromatography [BBRC 251:860-7 (98)].

Physical characteristics

P26Olf binds 4 atoms of calcium ions per molecule at the saturating condition (200 m(micro)M). An apparent Kd determined by the CD signal change at 222nm was 2.4 m(micro)M with the Hill coefficient of 1.5 [JBC 275:27245-9 (00)]. It is localized in the cilia layer of the frog olfactory epithelium. P26Olf binds beta-adrenergic receptor kinase with calcium dependent manner [JBC 275:27245-9 (00)].

PHOSPHOLIPASE C (PLC)

Phospholipase C (PLC) [EC 3.1.4.3] hydrolyzes phosphatidylinositol 4,5-bisphosphate to inositol 1,4,5-trisphosphate and diacylglycerol which activate IP3 receptor calcium channel and protein kinase C, respectively. It has an absolute requirement for calcium and is activated by guanine nucleotide binding proteins (PLC-beta), or by receptor tyrosine kinases (PLC-gamma), or by transglutaminase II (PLC-delta). All three vertebrate isoforms of PLC have a plekstrin homolog domain, followed by four EF-hands, a triosephosphate isomerase-like beta-barrel catalytic domain, and a C2 domain [56]. Although calcium is absolutely required for activity, it is bound to the catalytic domain. None of the four EF-hands have canonical calcium binding loops nor is Ca2+ bound in the crystal structure [Nature 380:595 (96)].

CBP1 & CBP2 (CBP)

Two similar proteins, CBP1 and CBP2, have been identified in Dictyostelium discoideum [FEBS Lett. 382:198 (96)]. Both contain four EF-hands and bind calcium in gel overlay experiments. Expression of CBP1 and CBP2 is increased during development.

----- miscellaneous -----

PLASMODIUM FALCIPARUM SURFACE PROTEIN (PFS)

PFS is membrane associated and the major calcium binding protein in Plasmodium falciparum [6409]. PFS is 343 residues long and has six EF-hands, all of which are inferred to bind calcium. The N-terminus has a secretory signal sequence (3-20). The C-terminal hydrophobic domain 345-359 may signal transfer to a glycosylphosphatidylinositol anchor. The C-terminal IDEL motif indicates that PFS is found within the endoplasmic reticulum [Mol. Biochem. Parasitol. 89:283 (97)].

Preparation and purification

PFS is purified by two-dimensional gel electrophoresis [6409].

CALCIUM-DEPENDENT PROTEIN KINASE (CDPK)

CDPK is CAM-independent. It is found in root and shoot tissues of soybean (Glycine max) and oats (Arena sativa). Immunocytological localization indicates association with the microfilament system. Oat CDPK associates with plasma membranes. The 328-residue kinase domain at N-terminus is homologous to CAM kinase II. The four EF-hands at the C-terminus are all inferred to bind calcium. A pseudosubstrate sequence and a CAM-binding motif are found at the junction region between the kinase domain and the CAM-like domain. A membrane ATPase is one of the substrates of CDPK [BBA 1350:109 (97)].

Preparation and purification

CDPK is purified from soybean by DEAE-cellulose ion-exchange chromatography, phenyl-Sepharose hydrophobic interaction chromatography, and Sephadex G-100 gel chromatography. The final purification is achieved by chromatography on Cibacron-blue Sepharose [5279].

Physical characteristics

CDPK is activated by calcium ions (Kd = 2 m(micro)M) and by lipids (0.1-4 mg/ml saturation). This activation is highly synergistic. Limited proteolytic digestion reduces stimulation by lipids; however, the limited-digested CDPK still exhibits calcium dependence [4181].

Proteolytic cleavage sites

CDPK from oats is normally present as a calcium and lipid-dependent kinase that is capable of associating with plasma membrane. By immunological analysis, CDPK is present as Mr = 61000 (major) and Mr=79000 (minor) bands. Both forms are calcium- and lipid-dependent and are associated with plasma membrane. Soluble calcium-dependent lipid-independent kinase is also purified from oats; it is supposed that this is a proteolytic degradation product generated during purification [4181].

PLASMODIUM FALCIPARUM PROTEIN KINASE (PFPK)

Plasmodium falciparum has a gene that encodes an inferred protein kinase. The C-terminus contains four EF-hands, all of which probably bind calcium. The DNA that encodes PFPK contains four introns: one between domains 1 and 2 and one each within domains 2, 3, and 4 [6527]. It has not been characterized.

PFPK is 524 long and is encoded by the PfCPK gene from P. falciparum. Residues 56-325 encode a region homologous to protein kinases and to CDPK, a protein kinase from plant, with which it shares ~40% sequence identity. The catalytic domains are homologous; however the four EF-hands of PFPK are not congruent with CDPK nor are they congruent with CTER, as is CDPK. The two chimeric kinases probably acquired their respective four EF-hands by different duplication and splicing events.

STRONYLOCENTROTUS PURPURATUS ECTODERMAL CALCIUM-BINDING PROTEIN (SPEC)

SPEC mRNAs begin to accumulate 20 h after fertilization in ectoderm cells of the sea urchin embryo [JMB 202:417 (88)]. A 15 kDa protein in eggs of sea urchin, Hemicentrotus pulcherrintus, also belongs to this subfamily. This protein is localized in nuclei of fertilized eggs and in the mitotic apparatus of dividing eggs. The antibody against this protein arrests cell division; hence, SPEC is inferred to play an important role in mitosis of sea urchin egg. SPEC is about 150 residues long and contains four EF-hands. .

Preparation and purification

The Hernicentrotus protein is purified from eggs by using DEAE-cellulose ion-exchange chromatography, hydroxyapatite chromatography, and Sephadex G-50 gel chromatography [5221].

Physical characteristics

The Hemicentrotus protein binds calcium, and shows calcium-dependent electrophoretic mobility change. SPEC does not activate cAMP-phosphodiesterase, and does not bind to fluperazine-Sepharose in the presence of calcium [5221].

p24 THYROID PROTEIN (TPP)

p24 thyroid protein (TPP), or calcyphosine, consists of four EF-hands. Domains 3 and 4 resemble one another most closely. This deviation from the ODD, EVEN pattern may reflect a recent duplication. The function of TPP is unknown. The synthesis and phosphorylation of p24 is upregulated by thyrotropin (TSH) and cyclic AMP agonists. Its synthesis is downregulated by dedifferentiating agents such as epidermal growth factor and 12-O-tetradecanoylphorbol-l3-acetate [6320]. TPP has 189 residues in four EF-hand domains. Another member of the subfamily has been found in lobster and crayfish muscle [EJB 227:97 (95)].

Preparation and purification

p24 was identified by two-dimensional gel electrophoresis of dog thyroid cell proteins [6320]. CCBP-23, a crayfish homolog of TPP, was purified from muscle tissue extract by using calcium-dependent hydrophobic -interaction chromatography on a pheny Sepharose CL-4B column. Further purification was performed by FPLC anion-exchange chromatography on a Mono-Q column [EJB 227:97 (95)].

Physical characteristics

p24 is phosphorylated in intact cells in response to TSH. The putative phosphorylation site is Ser40, which is in the first calcium-binding loop at position -Y. The fourth domain contains an inserted Glu after Ser at the Y vertex. The putative calcium coordination might be either the sequence SEDTE or DSTDE, similar to the calcium binding of domain 3, DSDTD. More unique is the insertion of 17 residues --- EFQDYYSGVSASMDTDE --- instead of the Glu at position -Z. Whether this insertion results in an F helix five turns longer and/or in altered calcium affinity is unknown [6320]. The affinity and stoichiometry of calcium binding has not been determined. pI of CCBP-23 is 5.4. A disulfide-linked homodimer was formed in the absence of reducing agents [EJB 227:97 (95)].

Covalent modification

N-terminal Ser of CCBP-23 is acetylated [EJB 227:97 (95)].

1F8 AND TB17 (1F8)

1F8 and TB17 (1F8) cDNAs were prepared from highly redundant mRNAs from Trypanosoma cruzei [JBC 264:18627 (89)] and T. brucei [Nucleic Acids Res. 18:4252 (90)]. 1F8, or calcimedin, [BJ 287:187 (92)] is located in the flagellum [6187].

1F8 is 211 residues long, TB17 is 233 residues long and both consist of four EF-hands, all of which are inferred to bind calcium.

Physical characteristics

1F8 has four EF-hand domains and binds about two equivalents of calcium with an apparent Kd ( 50 m(micro)M) [6187].

SARCOPLASMIC CALCIUM BINDING PROTEIN (SARC)

Sarcoplasm calcium binding protein (SARC) binds three Ca2+ or Mg2+ ions competitively. It is found both as a monomer and as a dimer; it may function as a calcium buffer. Neither the sequences of its four EF-hands nor of the three interdomain linkers closely resemble other EF-hands or linkers. Crystal structures are available of SARC from the sandworm, Nereis diversicolor, [JBC 266:652 (91)] and from amphioxus, Branchiostoma lanceolatum, [JMB 229:461 (93)]. Although the sequences share only 22% overall identity, the main chain traces of the two are nearly superimposable. All four domains have Ca2+ bound. The hydrophobic faces of lobes 1,2 and 3,4 contact one another.

Preparation and purification

SARC is extracted from muscle in the presence of calcium and protease inhibitors. The extract is fractionated by ammonium sulfate fractionation, Sephadex G-100 gel chromatography and DEAE-cellulose ion-exchange chromatography in the presence of calcium. Finally, it is purified by DEAE-cellulose ion-exchange chromatography in the presence of EDTA [287].

Physical characteristics

SARCs exist as both monomers and dimers. The relative and absolute affinities for both calcium and magnesium vary significantly with the type of SARC [224].

AEQUORIN AND LUCIFERINBINDING PROTEIN (AEQ)

AEQ is about 189 residues long; numerous isoforms are found in one organism. The easily recognized domains are numbered 1, 3, and 4. Residues 44-107 might correspond to a fourth EF-hand diverged beyond recognition. A similar ambiguity exists for PPTS. Aequorin (AEQ) (EC 1.13.12.5 renilla-luciferin 2-monooxygenase) is a calcium-dependent photoprotein that oxidizes luciferin (coelenterazine) and thereby produces the luminescence of the marine coelenterate, Aequoria victoria [Biochem. 26:1326 (87)]. In the anthozoan, Renilla reniformis, calcium induced bioluminescence involves two protein, luciferase and luciferin binding protein (LBP). LBP has four EF-hands but lacks the luciferase activity of aequorin [FEBS Lett. 268:287 (90)].

Preparation and purification

Aequorin is purified from Aequorea by ammonium sulfate fractionation (25-75% of saturation), Sephadex G-50 gel filtration, QAE-Sephadex A-50 ion-exchange chromatography, a second Sephadex G-50 gel filtration, and DEAE-Sephadex A-50 ion-exchange chromatography [5151]. The luciferin-binding protein is purified from Renilla by aluminum hydroxide gel adsorption, DEAE-cellulose column chromatography, Sephadex G-75 gel chromatography, DEAE-Biogel ion-exchange chromatography, p-benzyloxyaniline Sepharose column chromatography, and Sephadex G-75 gel chromatography [5160].

Physical characteristics

The luminescence reaction of luciferin (colenterazine) requires molecular oxgen. Aequorin binds both celeterazine and molecular oxygen as a chromophore. After the binding of three equivalents of calcium to aequorin, coelenterazine is oxidized to coelentramide by the bound oxygen. This yields light (l~ max = 470 nm), CO2 and a blue fluorescent protein that consists of apoaequorin and ceolenteramide [5291]. Celenterazine has been proposed to exist in peroxidized form. It was revealed by X-ray crystallography that aequorin is a globular molecule containing a hydrophobic core cavity that accommodates the ligand coelenterazine-2-hydroperoxiside. The overall conformation is most similar to SARC [Nature 405:372-376 (00)]. Several hydrogen bonds stabilizes the binding of celentrazine-2-hydroperoxiside including the Y184-H169-W173 triad which interacts the hydroperoxiside coupled to celetrazine. AEQ has only a few residues N- and C-terminal to the four EF-hand domains. The triad is in the C-terminal extention. When calcium binds at either or both EF-hands 1 and 4, helix orientation of these EF-hands would change which induces the disruption of hydrogen bond network including Y184-H169-W173. No longer stabilized, the peroxiside is free to initiate the light-emitting reaction [Nature 405:372-376 (00)]. Renilla contains a luciferase, which catalyses the oxidation of luciferin in the presence of dissolved oxygen producing oxiluciferin and C02. This is the classical luciferin-luciferase reaction. Luciferase purified in the presence of EDTA is not calcium sensitive; it can oxidize free luciferin. The calcium-triggered luciferin-binding protein restores calcium sensitivity to bioluminescence. The luciferin-binding protein, like aequorin, contains three calcium-binding EF-hand domains. A region of about 40 residues between EF-hands 1 and 2 is of uncertain evolutionary origin and may be involved in binding luciferin. In the presence of calcium this bound luciferin is presented to the luciferase in a reactive state [5160].

PROTEIN PHOSPHATASE (PPTS)

Protein phosphatase (PPTS) (EC 3.1.3.16) is encoded by the rdgC gene of D. melanogaster. PPTS is 661 residues long and has a phosphoprotein phosphatase domain at its N-terminus that is 30% identical with the catalytic domains of type 1, 2A and 2B serine/threonine protein phosphatases. Three EF-hands are easily recognized and indicated in table 1 as 1, 3 & 4 [6446]. Nominal EF-hand 2, occupying 56 residues between domains 1 and 3, may be a far diverged EF-hand paired with domain 1, another example of guilt by association.

HRA32 (H32)

The cDNA clone, Hra32, corresponds to a RNA transcript that accumulates in Phaseolus vulgaris (jack bean) during a hypersensitive reaction [Mol. Plant Microbe Interact. 12:712 (99)]. The encoded protein, H32, of 161 residues, has four EF-hands, all of which are predicted to bind calcium.

EFH5 (EFH5)

EFH5 is the product of the EF-hand 5 gene originally identified in Trypanosoma brucei (TbEFH5) and recently in T. cruzi (TcCUB) and in Leishmania tarentolae (LtEFH5). The inferred calcium-binding ability for the domains of the TbEFH5 product is -++-, and for both TcCUB and LtEFH5 Whether EFH5 performs the same function in the three organisms remains to be seen. The TbEFH5 gene is transcribed as a polycistronic precursor RNA. It follows the CAM gene C by 111 bp and precedes ubiquitin-EP52/l by 116 bp and is inferred to be under the control of a single distant upstream promoter [6513, 6106]. The TbEFH5 gene is transcribed as a polycistronic precursor RNA. It follows the CAM gene C by 111 bp and precedes ubiquitin-EP52/l by 116bp [6513, 6106]. EFH5 is inferred to be under the control of a single distant upstream promoter in T. brucei, T. cruzi, and L. tarentolae.

CALCIUM VECTOR PROTEIN (CVP)

CVP protein is inferred to be involved in stimulus-contraction coupling in Amphioxus; however, its exact function remains unknown.

CVP is 161 residues long and consists of four EF-hand domains. It is unusual among the EF-hand proteins, which are all found in the cytosol, in having a disulfide bond. Cys16 (1.2) and Cys78 (2.26) are near one another in the canonical lobe. In the model [Prot. Engineering 4:23 (90)] of CVP, based on the crystal structures of CAM and of TNC, a target (-helix is bound between the two lobes of CVP and the disulfide bond linkingalpha-helices E1 and F2 is accommodated with no steric interference.

Preparation and purification

CVP is purified from Amphioxus muscle by heat treatment, ion-exchange chromatography and gel chromatography [5170].

Physical characteristics

CVP has four EF-hand domains and binds two Ca2+ ions noncooperatively. CVP interacts with an endogenous 36 kDa protein in a calcium dependent manner [5170]. Domains 3 and 4 are inferred to bind calcium. There are N-trimethyllysines at positions 3.7 and 3.28.

PM129 CLONE (PMAT)

PM129 clone from A. thaliana (PMAT) cDNA, enriched for plasma membrane associated proteins, contains a single continuous open reading frame [Plant Physiol. 102:1059 (93)]. The 37 residues before the first of four EF-hand domains may comprise a distinct domain. Each of the linkers 1,2 2,3 and 3,4 is eight residues long. There are three Cys residues in the fourth EF-hand and one in the 3,4 linker; a disulfide bond could be formed without distorting the canonical EF-hand. It is strange that domains 1 and 2 of PMAT most closely resemble one another and that domain 4 of PMAT closely resembles domains 1 and 3 of CVP.

LAVI (LAV)

LAV1 cDNA was identified as a plasmodial stage specific mRNA of Physarum polycephalurn [3755]. The 220-residue domain at the N-terminus has no known homolog, which is alpha helical. The four EF-hands at the C-terminus are inferred to bind calcium. LAV1 cDNA encodes a 355 amino acid protein. There are four EF-hand domains at the Cterminus. LAV may not be congruent with any other EF-hand protein, but interdomains 1,2 and 3,4 resemble 1,2 and 3,4 of CTER. Linker 2,3 consists of a sole Leu. This places severe constraints on the spatial relationship between the N-terminal lobe and the C-terminal lobe. The overall conformation of LAV1 EF-hand domain is similar to CALP whose linker 2, 3 also consists of only one amino acid. The function of this protein is unknown and no homolog has been identified for the residues at the N-terminus.

Physical characteristics

Recombinant LAV1 protein (CBP40) binds 4 atoms of calcium ions per molecule with pCa (1/2) = 6.5. When N-terminal alpha helical domain was removed, the affinity for calcium decreased to pCa (1/2) = 4.6 [Biochem. 39:3827-34 (00)].

SACCHAROPOLYSPORA ERYTHRAEA CALCIUM-BINDING PROTEIN (CMSE)

CMSE is the only known EF-hand protein from prokaryotes [6455]. It has not been characterized. CMSE is 177 residues long and has four EF-hand domains.

MSV097 (MSV)

The entire genome of the poxvirus (Entomopoxvirinae) of the North American migratory grasshopper, Melanoplus sanguinipes, is 236 kbp and contains 267 open reading frames, of which 107 are similar to previously described genes [J. Virol. 73:533 (99)]. One gene encodes MSV, 140 residues. MSV has four EF-hands, the first two of which are inferred to bind calcium. It is the only known EF-hand protein from a virus.

PARVALBUMIN (PARV)

PARV contains parvalbumins a and b, oncomodulin, and avian thymic hormone. It has been suggested to function as a cytoplasmic calcium buffer and to facilitate muscle relaxation. Its function(s) in other tissues is unknown. Oncomodulin is a small acidic calcium-binding protein normally found only in extra-embryonic tissues such as the placenta, but not in normal embryonic or adult tissues. It was originally isolated from a rat tumor, and is found in significant amounts in a number of chemically induced rat hepatomas and in virus-transformed rat kidney cells. It is found in a wide variety of human tumors in lower amounts. Avian thymic hormone promotes immunological maturation of bone marrow cells in culture. This protein is very similar to parvalbumin, but appears to have a different function. All members of the PARV subfamily contain three EF-hand domains, the second and third of which bind calcium. PARV is 108 residues long and has three EF-hand domains, the first of which does not bind calcium and covers the (potential) hydrophobic patch of the lobe formed by the C-terminal two domains. The EF-hands are numbered 2, 3, & 4 because they cluster sequentially with EVEN, ODD, EVEN. The first domain is inferred to have been deleted. EF-hands 3 & 4 of PARV have high affinity for calcium and for magnesium. In the resting muscle cell PARV is in the two magnesium form. Following a pulse of messenger calcium the bound Mg2+ ions dissociate slowly and the excess Ca2+ ions are bound by parvalbumin thereby permitting relaxation of skeletal muscle [J. Muscl. Res. Cell. Motil. 3:377 (82)]. Whether PARV also performs this relaxing function in other tissues is not known.

Preparation and purification

PARV is purified by a three-step procedure involving ammonium sulfate fractionation, Sephadex G-75 gel filtration, and DEAE-cellulose ion-exchange chromatography [5274]. Oncomodulin is purified by heat treatment, ammonium sulfate fractionation and DEAE cellulose ion-exchange chromatography [3167].

Avian thymic hormone is purified from chicken thymus glands by heat treatment, acid precipitation, Sephadex chromatography and DEAE- or antibody affinity chromatography [6136].

Physical characteristics

Parvalbumin has two high-affinity Ca2+-Mg2+ sites. It binds two Mg2+ ions in resting muscle. After stimulation the substitution of Mg2+ by Ca2+ ions occurs more slowly than does the binding of calcium to the low-affinity sites of TNC. This explains why calcium can trigger contraction in the presence of a high concentration of high affinity Ca2+-Mg2+ sites of PARV in muscle cells. Relaxation occurs when Ca2+ diffuses from the low-affinity site of TNC to the high-affinity Ca2+-Mg2+ sites of PARV [311].

Oncomodulin is similar to parvalbumin in its amino acid sequence, but has different metal ion binding properties. It has one high-affinity Ca2+-Mg2+ site and one low-affinity Ca2+-specific site [286]. It has Asp at -Z domain 3; all other PARVs have Glu.

Avian thymic hormone has similar cation binding properties to those of parvalbumin [427].

Covalent modifications

The N-terminus of PARV is acetylated.

FIMBRIN (FIMB)

Fimbrin (FIMB) is an actin filament bundling protein from chick. It has two EF-hands at its N-terminus and four, 125 residue actin binding domains, homologous to those of a-actinin, at its C-terminus [81]. The human homologs are I, L, and T-plastin whose genes are on chromosomes 3, 13, and X [Mol. Cell Biol. 14:2457 (94)]. Eventhough FIMB bears some similarities to ACTN, their pairs of EF-hands are not closely related.

RAS GUANYL NUCLEOTIDE-RELEASING PROTEIN (GRP)

Ras guanyl nucleotide-releasing protein (GRP) has 795 residues and contains a homolog of a REM domain, a CDC25 domain, two EF-hands, and a diacylglycerol binding domain [Science 280:1082 (98)]. The two EF-hand loops are canonical and are expected to bind calcium. Helices E1 and F2 appear conventional. However, there are only 15 residues for helices F1 and E2. One or both of these helices is truncated and the relationship between the two EF-hands is unclear. Rat fibroblasts expressing the RasGRP encoding gene vector are transformed. Addition of a DAG analog causes sustained activation of Ras-Erk signaling and changes in cell morphology.

PKD2 GENE PRODUCT (PKD)

The PKD2 gene product (PKD) has 968 residues. Mutations in this gene in humans segregate with polycystic kidney disease [Science 272:1339 (96)]. The homolog, PKD1, constitutes a voltage-activated calcium (and sodium) channel in Caenorhabditis elegans. PKD has six membrane spanning helices. Two EF-hands follow the sixth transmembrane helix, a hundred residues from the C-terminus.

RYANODINE RECEPTOE PROTEIN (RYR)

The ryanodine receptor (RYR) functions as a channel for release of calcium from endoplasmic reticulum of muscle and brain. The isoforms are ~5000 residues long and contain two EF-hands, the second of which is inferred to bind calcium. The channels open in the presence of ~10-6 M Ca2+ (calcium induced calcium release) and closed at ~10-3 M Ca2+. The EF-hands have low calcium affinity and have been suggested to function in closing the channels [Biochem. 37:4804 (98)].

CBL (CBL)

CBL binds to phosphorylated tyrosine residues and functions as a negative regulator of many signaling pathways. The crystal structure [Nature 398:84 (99)] of CBL (minus residues 1 - 24) alone and in complex with a phosphotyrosine decapeptide that represents its binding site in ZAP-70 shows the pair of EF-hands wedged between a four helix bundle (4HB) and a divergent SH2 domain. The coordination of Ca2+ is unique as indicated in table 1 by "d". Although the residue at -X is Ser, the -X oxygen ligand comes from the carboxylate group of a Glu from the fourth helix of the 4HB. Although calcium is required for phosphopeptide binding, it interacts with the SH2 domain, not the pair of EF-hands. The pair of EF-hands may position the 4HB relative to the SH2 domain.

CALCIUM- & INTEGRIN-BINDING PROTEIN (CIB)

Calcium and integrin binding protein (CIB) binds specifically to the cytoplasmic domain of integrin (IIb of platelets [JBC 272:4651 (97)]. It has 191 residues and contains two EF-hands. No homolog has been identified for the first hundred residues.

CALSENSIN (SENS)

Calsensin (SENS) is expressed in some the peripheral neurons of leech, Haemopsis sp., that fasciculate in a single axon tract. SENS has 83 residues and both EF-hands bind calcium [JCB 129:1355 (95)]. It co-purifies with a 200kDa protein, not yet characterized.

GROOVIN (GRV)

Groovin (GRV) (identical to kakapo) is encoded by grv (kak) and is expressed in tendon cells of D. melanogaster. GRV is essential for muscle dependent, tendon cell differentiation [JCB 143:1259 (98)]. GRV contains two EF-hands; both are inferred to bind calcium.