Nội dung

Eur. J. Biochem. 269, 3433–3441 (2002)
FEBS 2002 doi:10.1046/j.1432-1033.2002.03023.x Structural and biochemical characterization of calhepatin, an S100-like calcium-binding protein from the liver of lungfish (Lepidosiren paradoxa) Santiago M. Di Pietro and José A. Santomé Instituto de Quı´mica y Fisicoquı´mica Biológicas (IQUIFIB), Facultad de Farmacia y Bioquı´mica, Universidad de Buenos Aires, Argentina We report the biochemical characterization of calhepatin, a calcium-binding protein of the S100 family, isolated from lungfish (Lepidosiren paradoxa) liver. The primary structure, determined by Edman degradation and MS/MS, shows that the sequence identities with the other members of the family are lower than those between S100 proteins from different species. Calhepatin is composed of 75 residues and has a molecular mass of 8670 Da. It is smaller than calbindin D9k (78 residues), the smallest S100 described so far. Sequence analysis and molecular modelling predict the two EF-hand motifs characteristic of the S100 family. Metal-binding properties were studied by a direct 45Ca2+-binding assay and by fluorescence titration. Calhepatin binds Ca2+ and Cu2+ but not Zn2+. Cu2+ binding does not change the affinity of calhepatin for Ca2+. Calhepatin undergoes a conformational change upon Ca2+ binding as shown by the increase in its intrinsic fluorescence intensity and kmax, the decrease in the apo-calhepatin hydrodynamic volume, and the Ca2+dependent binding of the protein to phenyl-Superose. Like most S100 proteins, calhepatin tends to form noncovalently associated dimers. These data suggest that calhepatin is probably involved in Ca2+-signal transduction. Cytoplasmic Ca2+ is a ubiquitous second messenger. The rise in intracellular Ca2+ is a widely established signal controlling a variety of processes in eukaryotic cells, such as cell growth and differentiation, cell motility, muscle contraction, gene expression, secretion, nerve impulse transmission, and apoptosis. The signal is partly transduced into metabolic or mechanical responses by calcium-binding proteins (CaBPs) which interact with cellular effectors in a Ca2+-dependent fashion [1]. S100 is a multigenic family of small dimeric CaBPs (78–119 amino acid residues) of the EF-hand superfamily, comprising 16 known members from mammalian species (S100 A1 to S100 A13, S100 B, Calbindin D9k, S100 P) and two putative additional members identified in chicken and channel catfish (MRP126 and ictacalcin, respectively). They have two EF-hand Ca2+-binding motifs, the N-terminal one having an extended loop characteristic of the S100 family [1–5]. S100 proteins show tissue-specific and cell-specific expression [4]. Some members of the family also bind Zn2+ and/or Cu2+ [5]. Most S100 proteins can form noncovalent dimers by a symmetric homodimeric fold mediated by hydrophobic contacts not found in other CaBPs [5–7]. Some S100 proteins form disulfide cross-linked homodimers [5]. The location of target protein-binding sites on opposite sides of the S100 homodimers could allow an S100 dimer to cross-bridge two homologous or heterologous S100 target proteins [2]. Calbindin D9k is the only S100 protein identified so far that does not form dimers [2,5–8]. We previously identified S100 A8 and S100 A9 from pig granulocytes [9] and discovered another member of the S100 family, the S100 A12 [10]. Hofmann et al. [11] proved that S100 A12, and possibly other members of the S100 family, mediates the activation of a novel proinflammatory axis by binding to RAGE (receptor for advanced glycation end products), a cell surface receptor. In this work, we focus on the characterization of a CaBP from lungfish (Lepidosiren paradoxa) liver, an S100 protein that apparently does not belong to any known S100 member. Its isolation, primary structure, metal-binding properties, and tissue expression pattern are described. As it is expressed mainly in hepatic cells and no other S100 member has been reported in liver, this protein will be referred to as calhepatin. Correspondence to J. A. Santomé, IQUIFIB, Facultad de Farmacia y Bioquı́mica, UBA, Junı́n 956, Buenos Aires (1113), Argentina. Fax: + 54 11 4508 3652, Tel.: + 54 11 4508 3651, E-mail: santome@qb.ffyb.uba.ar Abbreviations: CaBP, calcium-binding protein; MALDI-TOF, matrix-assisted laser desorption ionization time-of-flight. Note: The amino-acid sequence reported in this work has been deposited in the SWISS-PROT databank under accession number P82978. (Received 21 March 2002, accepted 24 May 2002) Keywords: calcium-binding protein; EF-hand; liver; lungfish; S100. MATERIALS AND METHODS Materials 45 CaCl2 (12.89 CiÆg)1) was from Dupont NEN. Endoproteinases Glu-C and Lys-C were obtained from Promega. All other reagents were purchased from Sigma, Baker, Bio-Rad, Amersham Pharmacia Biotech and/or Applied Biosystems. 3434 S. M. Di Pietro and J. A. Santomé (Eur. J. Biochem. 269) Preparation of the lungfish liver cytosolic fraction and purification of calhepatin Livers were excised from L. paradoxa weighing 600–700 g, cut into small pieces, suspended in homogenization buffer (40 mM sodium phosphate, 150 mM KCl, 4 mM EDTA, pH 7.4), and disrupted in a glass/Teflon homogenizer. The homogenate was then centrifuged at 20 000 g for 15 min, and the resulting supernatant further centrifuged at 105 000 g for 90 min in a Beckman XL-90 ultracentrifuge. The entire procedure was carried out at 4 C. The supernatant (4 mL) was loaded on a Sephadex G-75 column (2.5 · 40 cm) equilibrated with 15 mM Tris/HCl (pH 9.0)/ 1 mM EDTA. Elution was performed at 4 C with the same buffer at a flow rate of 16 mLÆh)1. The 6–18-kDa fraction was applied to a DEAE-cellulose column (1.1 · 10 cm) equilibrated with 15 mM Tris/HCl (pH 9.0). The material bound to the column was subsequently eluted with 10, 20, 30, 40, 50 and 100 mM NaCl in the same buffer. A portion of each fraction was concentrated and changed into 50 mM Tris/HCl buffer (pH 7.4) by using a Centriprep concentrator (Amicon) and assayed for 45Ca2+-binding activity as described below. The 10 mM fraction containing the calhepatin was concentrated by using the above concentrator and loaded on a Mono Q HR 5/5 column (Pharmacia LKB) previously equilibrated with 15 mM Tris/HCl (pH 9.0). The column was developed on an FPLC system (Pharmacia LKB), at a flow rate of 0.8 mLÆmin)1, with a 0–100 mM linear gradient of NaCl concentration over 60 min. Calhepatin was eluted at
40 mM NaCl. Protein purity was checked by SDS/PAGE (16% gel), isoelectric focusing, and RP-HPLC in a Vydac C4 column (4.6 · 250 mm). Preparation of the cytosolic fraction from lungfish and rat tissues Tissues were cut into small pieces, suspended in 40 mM sodium phosphate (pH 7.4) containing 150 mM KCl, 4 mM EDTA and 4 mM dithiothreitol, and homogenized in a Teflon Potter homogenizer. Homogenates were then centrifuged at 20 000 g for 15 min, and the resulting supernatants further centrifuged at 105 000 g for 90 min in a Beckman XL-90 ultracentrifuge. Electrophoresis SDS/PAGE (16% gel) was carried out as described by Schägger & von Jagow [12]. Isoelectric focusing was performed in a Phast System (Pharmacia). Antiserum production Calhepatin (500 lg) was mixed with Freund’s adjuvant and injected subcutaneously into a rabbit (first immunization), followed by a 250-lg boost 3 weeks later (second immunization). After 3 weeks, the animal was bled from the marginal ear vein and the serum was obtained. The antibodies were purified using conventional methods involving ammonium sulfate precipitation, DEAE-cellulose chromatography, and gel filtration on a Superdex 200 column. They were then concentrated with the Centriprep concentrator up to a concentration of 10 mgÆmL)1 and stored at )40 C.
FEBS 2002 Western blotting and immunoprecipitation experiments Western blotting was carried out as described by Harlow & Lane [13]. Immunoprecipitation was performed at 4 C using protein A–Sepharose beads [13]. Chromatographic analysis Gel-filtration analysis of pure calhepatin was performed as described by Drohat et al. [7] by FPLC on a Superose 12 HR 10/30 column (Pharmacia) calibrated with standard proteins. The column was equilibrated and eluted with 50 mM Tris/HCl/120 mM KCl/0.1 mM EDTA (pH 7.4) for the apo-calhepatin or with 50 mM Tris/HCl/120 mM KCl/2 mM CaCl2 (pH 7.4) for the holo-protein, at a flow rate of 0.5 mLÆmin)1. Both protein forms (7 lM monomer concentration) were incubated in the corresponding equilibration buffer for 30 min before being loaded. The monomeric and dimeric fractions obtained from the Superose column were incubated at room temperature for 12 h and applied again to the gel-filtration column under the same buffer and flow rate conditions. Hydrophobic interaction chromatography was carried out on a phenyl-Superose HR 5/5 column (Pharmacia) equilibrated with 50 mM Tris/HCl/120 mM KCl/1 mM CaCl2 (pH 7.4). After injection of 30 lg pure protein on the equilibration buffer, the column was eluted with 4 column vol. of the same buffer and then with 4 column vol. of 50 mM Tris/HCl/120 mM KCl/5 mM EDTA (pH 7.4). Enzymatic digestion and peptide purification For Glu-C protease digestion, 250 lg calhepatin was incubated in 0.1 M Tris/HCl (pH 7.9)/2 M guanidine hydrochloride with 4 lg enzyme, at 20 C for 24 h. For Lys-C protease digestion, 250 lg calhepatin was incubated in 0.1 M Tris/HCl (pH 8.5)/2 M guanidine hydrochloride with 5 lg enzyme, at 20 C for 24 h. Peptides were separated by RP-HPLC (Pharmacia LKB) on a Vydac C18 column (4.6 · 250 mm) equilibrated with solvent A [0.1% (v/v) trifluoroacetic acid in water]. Elution was performed at a flow rate of 0.8 mLÆmin)1 with a 0–50% linear gradient of solvent B [80% (v/v) acetonitrile, 0.08% (v/v) trifluoroacetic acid] over 80 min. Amino-acid analysis and sequencing Peptide amino-acid analyses and automatic amino-acid sequence determination by Edman degradation were carried out in an Applied Biosystems 420A Amino Acid Analyzer and an Applied Biosystems 477A Protein Sequencer, respectively, at the LANAIS-PRO (National Protein Sequencing Facility, UBA-CONICET), Buenos Aires, Argentina. Amino-acid sequence determination of the N-terminal peptide by MS/MS was performed in an Electrospray Ionization Ion Trap (Finnigan LCQ) at the Harvard University Microchemistry Facility, Cambridge, MA, USA. In-gel digestion and peptide purification Isolated calhepatin from lungfish intestine was digested with sequencing-grade trypsin by the in-gel procedure of 
FEBS 2002 Hepatic S100 calcium-binding protein (Eur. J. Biochem. 269) 3435 Rosenfeld et al. [14], as modified by Hellman et al. [15]. The resulting peptides were recovered by passive elution and then separated by HPLC on a BrownleeTM Aquapore RP 300 C18 column (2.1 · 220 mm) by using a combination of linear gradients of acetonitrile in 0.1% aqueous trifluoroacetic acid. Sequence alignment Multiple sequence alignment was performed by using MUL(http://pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page= /NPSA/npsa_multalin.html) with a BLOSUM 62 matrix and parameter default values. TALIN Evolutionary tree It was constructed with CLUSTAL W (http:/www.ebi.ac.uk/ clustalw/) by the Neighbour Joining method and using Kimura’s distance correction procedure. Mass spectrometry The molecular mass of lungfish calhepatin was determined with a Bruker Biflex III matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometer. Molecular modelling The three-dimensional structure of lungfish calhepatin was modelled using SWISS-MODEL, an automated modelling package of the ExPASy Molecular Biology Server, available through the internet (http://www.expasy.ch/swissmod/ SWISS-MODEL.html) [16]. Calhepatin was modelled with the PROMOD program on the basis of its similarity to homologous structures existing in the Brookhaven Protein Data Bank. After primary modelling, the structure was energy-minimized using GROMOS. The model was checked with the WHAT-IF program which utilizes WHAT-CHECK verification routines. wavelength was set to 278 ± 5 nm. Each spectrum (285– 420 nm) represents an average of three scans. The fluorescence intensity was corrected for sample dilution, the latter never exceeding 4%. Curve fitting was performed as described by Dell’Angelica et al. [10]. Data were analyzed by the following equation where F0 is the fluorescence at zero ligand concentration, Fm is the maximum fluorescence change, T is the total ligand concentration, P is the protein monomer concentration, and Ka is the apparent association constant: F ¼ F0 þ Fm ð2PÞ1 n o 1=2 1 2 T þ P þ K1 a  ½ðT þ P þ Ka Þ  4PT  The Cu2+-binding-induced fluorescence change was corrected for nonspecific quenching by subtracting the values of a linear term obtained from the final portion of the Cu2+-binding curve, corresponding to fluorescence quenching before binding saturation. Uncorrected Cu2+ fluorescence quenching data were also analyzed by direct fitting to an equation containing an additional linear term, but the value obtained for the association constant was indistinguishable from that obtained with corrected data. RESULTS Purification of lungfish calhepatin The 105 000 g supernatant from lungfish liver was fractionated on a Sephadex G-75 column. The 6- to 18-kDa fraction containing the calhepatin was applied to a DEAEcellulose column, and the 10 mM NaCl fraction containing Ca2+-binding activity was further purified by anionexchange chromatography on a Mono Q column. The protein was eluted at 40 mM NaCl from the last column in a symmetric peak that was homogeneous as confirmed by SDS/PAGE (Fig. 1), isoelectric focusing and RP-HPLC (not shown). 45 Ca2+-Binding assay Apo-(lungfish calhepatin) was prepared by incubation of freshly purified protein with 2 mM EGTA and 2 mM EDTA and subsequent dialysis against 50 mM Tris/HCl (pH 7.4). 45 Ca2+ binding was determined by the method of Mani and Kay [77] using 15 lM apo-(lungfish calhepatin) as described by Dell’Angelica et al. [10]. Binding data were analyzed by nonlinear regression curve fitting using the following equation, where v is the number of moles of Ca2+ bound per mol of monomer, x is the free Ca2+ concentration, n is the number of binding sites per monomer, and Ka1 and Ka2 are macroscopic binding constants:
¼ ½ðn=2ÞKa1 x þ nKa1 Ka2 x2 =ð1 þ Ka1 x þ Ka1 Ka2 x2 Þ Fluorescence measurements The intrinsic fluorescence of 2 lM calhepatin was recorded at 20 C on a Jasco FP-770 spectrofluorimeter (Japan Spectroscopic Co., Hachioji City, Japan). The excitation Fig. 1. SDS/PAGE analysis of calhepatin-containing samples at different stages of purification. Lane 1, 105 000 g supernatant of lungfish liver homogenate; lane 2, after Sephadex G-75 gel filtration; lane 3, after DEAE-cellulose chromatography; lane 4, after Mono Q chromatography. 3436 S. M. Di Pietro and J. A. Santomé (Eur. J. Biochem. 269)
FEBS 2002 Biochemical properties of lungfish calhepatin The protein migrates on SDS/PAGE as a 6-kDa polypeptide (Fig. 1) showing an aberrant mobility, in the same way as other CaBPs [10,18–20]. The molecular mass as determined by MALDI-TOF MS is 8672 Da. Analysis of the apo-calhepatin, at 7 lM monomer concentration, by Superose 12 gel filtration showed two peaks of apparent molecular mass 6.1 ± 0.1 and 16.8 ± 0.1 kDa (mean ± SD), respectively, showing that the protein exists as both a monomer (15%) and a dimer (85%). Taking into account the chromatography time scale (
25 min) the existence of the two well-separated forms suggests a very slow monomer–dimer equilibrium. To confirm that there is an equilibrium between the two forms, the monomeric (0.3 lM monomer concentration) and dimeric (1.1 lM monomer concentration) fractions obtained from the Superose column were incubated for 12 h and applied again to the gel-filtration column. In both cases, the two forms were obtained again and the ratio of monomer to dimer fraction was
60 : 40 and 40 : 60, respectively, indicating that the dissociation constant is in the micromolar order. When the same determination was performed with the holo-calhepatin, at 7 lM monomer concentration, almost 100% of the protein was recovered as a dimer of apparent molecular mass 14.6 ± 0.2 kDa. The holodimeric fraction obtained from the Superose column (0.9 lM monomer concentration) was incubated for 12 h and applied again to the gel-filtration column. Almost 100% of the protein was again recovered as a dimer, indicating that the monomer-dimer dissociation constant for the holoprotein is in the submicromolar range. On the other hand, differences between the apo-calhepatin and holocalhepatin hydrodynamic volume (16.8 ± 0.1 kDa and 14.6 ± 0.2 kDa, respectively) suggest that calhepatin undergoes a conformational change on Ca2+ binding. Ca2+ binding affected the chromatographic behaviour of calhepatin on a phenyl-Superose column. The protein was completely bound to the column in the presence of 1 mM CaCl2 but could be eluted from the column with 5 mM EDTA (not shown). Fig. 2. RP-HPLC separation of calhepatin peptides generated by enzymatic digestion. (A) The peptide mixture obtained by Glu-C digestion was fractionated on a Vydac C18 column (4.6 · 250 mm) equilibrated with solvent A [0.1% (v/v) trifluoroacetic acid in water]. The column was eluted with a 0–50% linear gradient (dashed line) of solvent B [80% (v/v) acetonitrile, 0.08% (v/v) trifluoroacetic acid]. (B) The products of Lys-C digestion were separated as described for the Glu-C peptide mixture. Numbered peaks represent peptides submitted to sequencing and/or amino-acid analysis. This value is close to that obtained by MALDI-TOF MS (8672 Da). The calculated isoelectric point [21] (pI ¼ 5.12) agrees with the experimental value (pI ¼ 5.15 for both the holo and apo protein). Sequence comparison and evolutionary relationship Primary structure of lungfish calhepatin Purified protein (250 pmol) was subjected to four cycles of Edman degradation. No phenylthiohydantoin derivative could be identified, indicating that its N-terminal amino acid is blocked. Calhepatin fragments were generated by digestion with proteases Glu-C and Lys-C, fractionated by RP-HPLC (Fig. 2A,B) and submitted to Edman degradation and amino-acid analysis. Information on the complete amino-acid sequence except for the four N-terminal residues was obtained. According to the amino-acid determination and the blocked N-terminal residue of the peptide, peak 1 from Lys-C digestion corresponds to the N-terminal portion of lungfish calhepatin. The material corresponding to this peak was submitted to sequencing by MS/MS. A summary of the sequence analyses and the resulting primary structure of lungfish calhepatin is shown in Fig. 3. The protein is composed of 75 residues. From its aminoacid sequence, assuming that the N-terminus is an acetyl group, the molecular mass was calculated to be 8670 Da. The alignment of the amino-acid sequence of lungfish calhepatin with other members of the S100 family indicates that the number of amino-acid identities between calhepatin and other S100 proteins ranges from 12 to 21 (Fig. 3). These values are far lower than those between S100 proteins from different species. This is strong evidence that calhepatin is a novel S100 protein, as shown in the evolutionary tree of the S100 family in Fig. 4. Despite the above evolutionary relationships, according to the BLASTP program [22], the higher similarity when the calhepatin amino-acid sequence is compared with all database proteins corresponds to one or more segments of CaBPs with a higher molecular mass than that of S100 proteins. Identity can reach 41%. Structural modelling of lungfish calhepatin The three-dimensional structure of the lungfish calhepatin monomer was predicted by using the SWISS-MODEL modelling package [16] (Fig. 5). Molecular modelling of the 
FEBS 2002 Hepatic S100 calcium-binding protein (Eur. J. Biochem. 269) 3437 Fig. 3. Primary structure of calhepatin and its sequence alignment with S100 family members. Peptide fragments are indicated by solid arrows when determined by sequencing, and by a dotted line arrow for those inferred on the basis of their amino-acid analysis. Each peptide is labelled with a letter (E for peptides derived by Glu-C digestion and K for peptides obtained by Lys-C digestion) and a number that agrees with that of Fig. 2. Numbers above the sequence indicate residue positions in the protein. aaaaa correspond to predicted a helices. Underlined residues are equivalent to the residues of S100 A4 critical for S100 A4 dimerization [23]. The amino-acid sequence of calhepatin was aligned with those of the human form of each S100 protein, except for MRP126 and ictacalcin, which have only been isolated from chicken and catfish, respectively. The number of identities between each S100 protein and calhepatin is indicated after each amino-acid sequence. Both canonical (C) and noncanonical (NC) EF-hands are also indicated. sequence was conducted with the PROMOD program using the known three-dimensional structures of other members of the S100 family as templates. The lungfish calhepatin monomer model has the same overall conformation as that of other members of the S100 family containing four a helix segments (Fig. 5A). Calhepatin residues present in sequence positions equivalent to residues crucial for dimerization and monomer stability, in Sl00 A4 and other S100 proteins [23] (Fig. 3), are clustered between helix I and IV (Fig. 5B), in the same way as in Sl00 A4 and other S100 proteins [23]. This agrees with biochemical data showing that calhepatin is able to dimerize. Direct Ca2+-binding studies The 45Ca2+-binding isotherm of calhepatin at 20 C in 25 mM Tris/HCl (pH 7.4) is shown in Fig. 8. The binding constants determined are Ka1 ¼ (2.9 ± 0.3) · 105 M and Ka2 ¼ (6.0 ± 0.7) · 103 M (n ¼ 2.1 ± 0.05). In the presence of 1 mM Cu2+, binding constants are Ka1 ¼ (2.0 ± 0.3) · 105 M and Ka2 ¼ (4.6 ± 0.6) · 103 M (n ¼ 2.2 ± 0.2), thus Cu2+ binding does not significantly change the affinity of calhepatin for Ca2+ (values are all mean ± SD; n ¼ 3). Tissue expression of calhepatin Fluorescence titration The intrinsic emission spectrum of calhepatin and those of the protein with increasing amounts of Ca2+ are shown in Fig. 6A. Figure 6B displays the corrected maximum of fluorescence intensity for each Ca2+ concentration and allows detection of one binding site with Ka(app) ¼ (3.6 ± 0.5) · 105 M (mean ± SD, n ¼ 3). Intrinsic fluorescence determinations were also applied to study the binding of Zn2+, Mg2+ and Cu2+. Neither Zn2+ nor Mg2+ changes calhepatin fluorescence, suggesting that they have no binding sites in the protein. In addition, they have no effect on Ca2+ binding (not shown). Calhepatin fluorescence intensity decreased, and kmax changed with Cu2+ additions (Fig. 7A). The analysis of the corrected maximum of fluorescence (Fig. 7B) provides evidence of the presence of a single site with Ka(app) ¼ (1.5 ± 0.2) · 107 M (mean ± SD, n ¼ 3). To investigate the pattern of calhepatin expression in lungfish tissues, cytosolic fractions from liver, skeletal muscle, intestine, lung, brain, adipose tissue, heart and skin were submitted to electrophoresis and immunoblotting. Rabbit antibodies to calhepatin only detected the protein in liver and at a much lower level in intestine (Fig. 9). Furthermore, the antibodies did not cross-react with cytosolic proteins from rat liver or intestinal tissues (Fig. 9), suggesting that calhepatin-like proteins are not expressed in rat and that the antibodies do not cross-react with calbindin D9k. Consistently with the immunoblotting results, when cytosolic fractions from lungfish and rat tissues were submitted to immunoprecipitation experiments with the calhepatin antibodies followed by SDS/PAGE analysis, only lungfish liver and intestine showed the calhepatin band (data not shown). 3438 S. M. Di Pietro and J. A. Santomé (Eur. J. Biochem. 269)
FEBS 2002 To confirm that the lungfish intestinal protein recognized by the antibodies is calhepatin, it was submitted to in-gel tryptic digestion by the procedure of Rosenfeld et al. [14]. The peptide mixture was fractionated by RP-HPLC, and the two peptides sequenced (SGTLSVDELY and IIEK) were found to be identical with those corresponding to calhepatin fragments 19–28 and 46–49, respectively. DISCUSSION Fig. 4. Evolutionary tree of the S100 protein family. Unrooted evolutionary tree based on the multiple sequence alignment of 59 S100 protein primary structures constructed with CLUSTAL W (http:/ www.ebi.ac.uk/clustalw/) by the Neighbour Joining method and using the Kimura correction of distances. hu, human (Homo sapiens); ca, catfish (Ictalurus punctatus); ra, rat (Rattus norvegicus); bo, bovine (Bos taurus); mo, mouse (Mus musculus); rb, rabitt (Oryctolagus cuniculus); ch, chiken (Gallus gallus); ho, horse (Equus caballus); lf, lungfish (Lepidosiren paradoxa); pi, pig (Sus scrofa). The Ca2+ signal is transduced by a variety of CaBPs. Whereas a number of Ca2+-dependent responses are mediated by calmodulin, a ubiquitous CaBP universally present in cells, the S100 proteins are cell-type-specific mediators of the Ca2+ signal [3]. Kligman & Hilt [3] described the structural features that determine whether a protein is a member of the S100 family. They have two EF-hands per monomer. One of them, located in the C-terminal region, comprises 12 amino-acid residues and is similar to those found in calmodulin. The other differs from the calmodulin-related protein EF-hand as it contains 14 residues. The N-terminal and C-terminal regions contain conserved hydrophobic amino-acid domains. The CaBP reported here, calhepatin, shares all these structural characteristics. Members of the S100 family are acidic CaBPs comprising between 78 (calbindin D9k) and 119 (MRP126) residues, whereas calhepatin consists of 75 residues, this being the smallest S100 protein reported. As far as we know, this is the first time that an S100 has been described in liver. Calhepatin probably has specific functions in this organ taking into account that most S100 proteins are often expressed in a tissue-specific manner [3–5,24] and calhepatin is expressed almost exclusively in liver. In the evolutionary tree of the S100 family (Fig. 4), calhepatin appears as a new member, calbindin D9k being the most closely related to it. The two S100 proteins share the characteristic of having a low number of residues, although divergence between their genes seems to have occurred long ago. The lack of cross-reaction between the calhepatin antibodies and rat intestinal calbindin D9k agrees with their gene divergence. Interestingly, calbindin D9k is the Fig. 5. Structural modelling of calhepatin monomer. (A) Three-dimensional structure of lungfish calhepatin predicted using the SwissModel automated modelling package based on the crystal and/or NMR structures of other members of the family. The molecule is depicted in strand representation, and a helices are numbered from I to IV. (B) Residues Leu7, Arg8 and Phe11 from a-helix I, and Trp60, Phe63, Ala66 and Phe67 from a helix IV (located at equivalent positions to those crucial for S100 A4 dimerization and monomer stability [23]) are shown in dark and light grey representation, respectively. The figure was generated using the RASMOL program. 
FEBS 2002 Hepatic S100 calcium-binding protein (Eur. J. Biochem. 269) 3439 Fig. 6. Ca2+ fluorescence titration. (A) Fluorescence spectra of 2 lM calhepatin in 25 mM Tris/HCl, pH 7.4, with 0–810 lM Ca2+. (B) Ca2+ titration curve showing corrected maximum fluorescence at each Ca2+ concentration. Curve fitting was performed as indicated in Materials and methods. only family member identified so far that does not form dimers, acting as a Ca2+ modulator, rather than as a Ca2+ sensor [25]. Analysis of apo-calhepatin and holo-calhepatin by Superose 12 gel filtration showed that the protein is in a monomer–dimer equilibrium and that the dissociation constant is in the micromolar range for the apoprotein and in the submicromolar range for the holoprotein, as reported for other S100 family members [7]. Tarabykina et al. [23] studied crucial residues for dimerization in Sl00A4 and found that three residues in helix I and four in helix IV are critical for S100A4 dimerization. Figure 3 shows these residues and those present in equivalent positions in the other S100 proteins. Most of the critical residues are present in calhepatin and calbindin D9k. However, the latter has a shorter helix IV, a characteristic that could explain its inability to dimerize. Kligman & Hilt [3] proposed that the interaction of a particular S100 protein with an effector protein occurs after Ca2+ binding induces a conformational change, exposing hydrophobic domains which then interact with corresponding hydrophobic domains in the effector. According to this currently accepted mechanism [2,26], calhepatin should undergo the Ca2+-dependent conformational changes responsible for the transmission of Fig. 7. Cu2+ fluorescence titration. (A) Fluorescence spectra of 2 lM calhepatin in 25 mM Tris/HCl, pH 7.4, with 0–100 lM Cu2+. (B) Cu2+ titration curve showing corrected maximum fluorescence at each Cu2+ concentration. Curve fitting was performed as indicated in Materials and methods. Fig. 8. 45Ca2+-Binding isotherms. 45Ca2+ binding to 15 lM calhepatin in 25 mM Tris/HCl, pH 7.4, was determined by the method of Mani & Kay [17] following the procedure of Dell’Angelica et al. [10]. Curve fitting was performed as indicated in Materials and Methods. information to effector proteins. Our fluorescence experiments indicate a Ca2+-induced change in the environment of at least the tryptophan residue located in one of the 3440 S. M. Di Pietro and J. A. Santomé (Eur. J. Biochem. 269)
FEBS 2002 from a single ur-domain by two cycles of gene duplication and fusion [8], calhepatin may be related to that ancient domain. ACKNOWLEDGEMENTS Fig. 9. Western-blot analysis. The 105 000 g supernatant of lungfish liver (lane 1), skeletal muscle (lane 2), intestine (lane 3), lung (lane 4), brain (lane 5), adipose tissue (lane 6), heart (lane 7) and skin (lane 8) homogenates and those of rat liver (lane 9) and intestine (lane 10) were subjected to SDS/PAGE and transferred to nitrocellulose membranes. Immunodetection was carried out with polyclonal rabbit anti-calhepatin IgG. four positions critical for dimerization in helix IV (Figs 3 and 5). The binding of Ca2+ to calhepatin increases its intrinsic fluorescence intensity and kmax. This result, the decrease in the protein hydrodynamic volume, and the fact that Ca2+-loaded calhepatin is retained on a phenylSuperose column and can be eluted with EDTA suggests that calhepatin undergoes a conformational change on Ca2+ binding that exposes hydrophobic regions. This probably involves residues shown in spacefill representation in Fig. 5B. Unfortunately, although preliminary immunoprecipitation experiments do precipitate calhepatin, they fail to coprecipitate calhepatin effector protein partners. The metal-binding properties of calhepatin were studied by a direct 45Ca2+-binding assay and fluorescence titration. The binding of 2 Ca2+/monomer is consistent with the presence of two EF-hand motifs. The affinity constants determined agree with the fact that S100 protein affinity for Ca2+ is low, the affinity of the C-terminal EF-hand being greater than that of the N-terminal EF-hand [3]. Such characteristics suggest that S100 proteins may be activated only in subcellular compartments where the Ca2+ concentration reaches a relatively high level [3]. Additional modes of affinity control may involve other factors such as other cations. The affinity of S100 proteins for Ca2+ can be modulated by Zn2+ binding in some subfamilies such as S100B [27], S100A5 [28], S100A6 [20] and S100A12 [10]. Our results indicate that Cu2+, unlike Zn2+ and Mg2+, binds to calhepatin. Copper binding does not change calhepatin affinity for Ca2+, but it is not unlikely that in some cases calhepatin biological activity could be regulated by Cu2+ instead of Ca2+ [5]. Preliminary cross-linking experiments show that both Ca2+ and Cu2+ increased the dimer/monomer ratio, suggesting that, like Ca2+, Cu2+ also enhance calhepatin dimerization. Surprisingly, according to the BLASTP program [22], the higher scores of similarity when the calhepatin amino-acid sequence is compared with all database proteins correspond to one or more segments of CaBPs of higher molecular mass than S100 proteins. This contains several EF-hands such as Ca2+-dependent protein kinases from Arabidopsis thaliana, Zea mays, Glycine max, Dunaliella tertiolecta, Picea mariana, Solanum tuberosum, Plasmodium falciparum and other species. In addition, calmodulin-like proteins from A. thaliana, Z. mays, Mus musculus, Homo sapiens skin and Suberites domuncula have two fragments with similar characteristics. As some CaBPs appear to have evolved We thank Dr Ulf Hellman and the Ludwig Institute for Cancer Research (Uppsala, Sweden) for MALDI-TOF MS analysis. We acknowledge S. B. Linskens and E. V. Dacci for amino-acid analysis and sequence determination by Edman degradation. We also thank R. Davis for language supervision. This work was supported by Consejo Nacional de Investigaciones Cientı́ficas y Técnicas de la República Argentina Grant 4115 and Universidad de Buenos Aires Grant TB75. REFERENCES 1. Heizmann, C.W. & Hunziker, W. (1991) Intracellular calciumbinding proteins: more sites than insights. Trends Biochem. Sci. 16, 98–103. 2. Donato, R. (1999) Functional roles of S100 proteins, calciumbinding proteins of the EF-hand type. Biochim. Biophys. Acta 1450, 191–231. 3. Kligman, D. & Hilt, D.C. (1988) The S100 protein family. Trends Biochem. Sci. 13, 437–443. 4. Schäfer, B.W. & Heizmann, C.W. (1996) The S100 family of EF-hand calcium-binding proteins: functions and pathology. Trends Biochem. Sci. 21, 134–140. 5. Heizmann, C.W. & Cox, J.A. (1998) New perspectives on S100 proteins: a multi-functional Ca2+-, Zn2+- and Cu2+-binding protein family. Biometals 11, 383–397. 6. Potts, B.C.M., Smith, J., Akke, M., Macke, T.J., Okazaki, K., Hidaka, H., Case, D.A. & Chazin, W.J. (1995) The structure of calcyclin reveals a novel homodimeric fold for S100 Ca2+-binding proteins. Nat. Struct. Biol. 2, 790–796. 7. Drohat, A.C., Nenortas, E., Beckett, D. & Weber, D.J. (1997) Oligomerization state of S100B at nanomolar concentration determined by large-zone analytical gel filtration chromatography. Protein Sci. 6, 1577–1582. 8. Kawasaki, H., Nakayama, S. & Kretsinger, R.H. (1998) Classification and evolution of EF-hand proteins. Biometals 11, 277–295. 9. Dell’Angelica, E.C., Schleicher, C.H., Simpson, R.J. & Santomé, J.A. (1996) Complex assembly of calgranulins A and B, two S100-like calcium-binding proteins from pig granulocytes. Int. J. Biochem. Cell Biol. 28, 53–62. 10. Dell’Angelica, E.C., Schleicher, C.H. & Santomé, J.A. (1994) Primary structure and binding properties of calgranulin C, a novel S100-like calcium-binding protein from pig granulocytes. J. Biol. Chem. 269, 28929–28936. 11. Hoffmann, M.A., Drury, S., Fu, C., Qu, W., Taguchi, A., Lu, Y., Avila, C., Kambham, N., Bierhaus, A., Nawroth, P., Neurath, M., Merser, J., Stern, D. & Schmidt, A.M. (1999) RAGE mediates a novel proinflammatory axis: a central cell surface receptor for S100/calgranulin polypeptides. Cell 97, 889–907. 12. Schägger, H. & von Jagow, G. (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal. Biochem. 166, 368–379. 13. Harlow, E. & Lane, D. (1988) Antibodies. A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 14. Rosenfeld, J., Capdevielle, J., Guillemot, J.C. & Ferrara, P. (1992) In-gel digestion of proteins for internal sequence analysis after one- or two-dimensional gel electrophoresis. Anal. Biochem. 203, 173–179. 15. Hellman, U., Wernstedt, C., Góñez, J. & Heldin, C.H. (1995) Improvement of an in-gel digestion procedure for the 
FEBS 2002 16. 17. 18. 19. 20. 21. 22. Hepatic S100 calcium-binding protein (Eur. J. Biochem. 269) 3441 micropreparation of internal protein fragments for amino acid sequencing. Anal. Biochem. 224, 451–455. Peitsch, M.C. (1996) ProMod and Swiss-Model: internet-based tools for automated comparative protein modelling. Biochem. Soc. Trans. 24, 274–279. Mani, R.S. & Kay, C.M. (1990) Isolation and characterization of a novel molecular weight 11000 Ca2+-binding protein from smooth muscle. Biochemistry 29, 1398–1404. Boyhan, A., Casimir, C.M., French, J.K., Teahan, C.G. & Segal, A.W. (1992) Molecular cloning and characterization of grancalcin, a novel EF-hand calcium-binding protein abundant in neutrophils and monocytes. J. Biol. Chem. 267, 2928–2933. Dianoux, A.C., Stasia, M.J., Garin, J. & Vignais, P.V. (1992) A 23-kilodalton protein, a substrate of protein kinase C, in bovine neutrophil cytosol is a member of the S100 family. Biochemistry 31, 5898–5905. Pedrocchi, M., Schäfer, B.W., Durussel, I., Cox, J.A. & Heizmann, C.W. (1994) Purification and characterization of the recombinant human calcium-binding S100 proteins CAPL and CACY. Biochemistry 33, 6732–6738. Moore, D.S. (1985) Amino acid and peptide net charges: a simple calculation procedure. Biochem. Ed. 13, 10–11. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. & Lipman, D.J. (1990) Basic local alignment tool. J. Mol. Biol. 215, 403–410. 23. Tarabykina, S., Scott, D.J., Herzyk, P., Hill, T.J., Tame, J.R.H., Kriajevska, M., Lafitte, D., Derrick, P.J., Dodson, G.G., Maitland, N.J., Lukanidin, E.M. & Bronstein, I.B. (2001) The dimerization interface of the Metastasis-associated protein S100A4 (Mts1). J. Biol. Chem. 276, 24212–24222. 24. Mandinova, A., Atar, D., Schäfer, B.W., Spiess, M., Aebi, U. & Heizmann, C.W. (1998) Distinct subcellular localization of calcium binding S100 proteins in human smooth muscle cells and their relocation in response to rises in intracellular calcium. J. Cell Sci. 111, 2043–2054. 25. Skelton, N.J., Kördel, J., Akke, M., Forsén, S. & Chazin, W. (1994) Signal transduction versus buffering activity in Ca2+binding proteins. Nat. Struct. Biol. 1, 239–245. 26. da Silva, A.C.R. & Reinach, F.C. (1991) Calcium binding induces conformational changes in muscle regulatory proteins. Trends Biochem. Sci. 16, 53–57. 27. Baudier, J., Glasser, N. & Gerard, D. (1986) Ions binding to S100 proteins I. Calcium- and zinc-binding properties of bovine brain S100aa, S100a (ab), and S100b (bb) protein: Zn2+ regulates Ca2+ binding on S100b protein. J. Biol. Chem. 261, 8192–8203. 28. Schäfer, B.W., Fritschy, J.M., Murmann, P., Troxler, H., Durussel, I., Heizmann, C.W. & Cox, J.A. (2000) Brain S100A5 is a novel calcium-, zinc-, and copper ion-binding protein of the EF-hand superfamily. J. Biol. Chem. 275, 30623–30630.