Eur. J. Biochem. 207, 541 -547 (1992)

0FEBS f 992

SlOOP, a novel Ca2+-binding protein from human placenta cDNA cloning, recombinant protein expression and Ca2 binding properties +

Tilmann BECKER, Volker GERKE, Eckhard KUBE and Klaus WEBER Max Planck Institute for Biophysical Chemistry, Department of Biochemistry, Gottingen, Federal Republic of Germany (Received March 2/April 30, 1992)

-

EJB 92 0286

A novel member of the SlOO protein family, present in human placenta, has been characterized by protein sequencing, cDNA cloning, and analysis of Ca2+-binding properties. Since the placenta protein of 95 amino acid residues shares about 50% sequence identity with the brain SlOO proteins a and p, we proposed the name SlOOP. The cDNA was expressed in Escherichia coli and recombinant SlOOP was purified in high yield. SlOOP is a homodimer and has two functional E F hands/polypeptide chain. The low-affinity site (Kd = 800 pM), which, in analogy to Sloop, seems to involve the Nterminal EF hand, can be followed by the Ca2+-dependent decrease in tyrosine fluorescence. The high-affinity site, provided by the C-terminal E F hand, influences the reactivity of the sole cysteine which is located in the C-terminal extension (Cys85). Binding to the high-affinity site (Kd = 1.6 pM) can be monitored by fluorescence spectroscopy of SlOOP labelled at Cys85 with 6-proprionyl-2dimethylaminonaphthalene (Prodan). The Prodan fluorescence shows a Ca2+-dependentred shift of the maximum emission wavelength from 485 nm to 502 nm, which is accompanied by an approximately twofold loss in integrated fluorescence intensity. This indicates that Cys85 becomes more exposed to the solvent in Ca2+-bound SIOOP, making this region of the molecule, the so-called Cterminal extension, an ideal candidate for a putative Ca2+-dependent interaction with a cellular target. In p l l , a different member of the SlOO family, the C-terminal extension which contains a corresponding cysteine (Cys82 in p l I), is involved in a Ca2+-independentcomplex formation with the protein ligand annexin 11. The combined results support the hypothesis that SlOO proteins interact in " general with their targets after a Ca2 -dependent conformational change which involves hydrophobic residues of the C-terminal extension +

v

The SIOO multigene family covers a group of small dimeric molecules around 100 residues in length. SlOOa and Sloop are the principal members of this protein family which were originally identified in bovine brain (Moore, 1965). They share around 55% sequence identity, but show some significant differences, e. g. Sloop contains two cysteine residues while only one cysteine is present in S100a; Sloop also has more histidine residues and consequently can bind Zn2 with high affinity (for review see Hilt and Kligman, 1991). SlOOcl and Sloop, as well as all other members of the SlOO family identified in recent years, belong to the superfamily of Ca2+-binding proteins built from consecutive E F hands where the loop between two cc-helices provides the Ca2+-binding site. However, unlike calmodulin, which seems present in all eukaryotic cells, the various SlOO proteins show a cell-type-specific ex+

Correspondence to K. Weber, Max Planck Institute for Biophysical Chemistry, Department of Biochemistry, P.O. Box 2841, W-3400 Gottingen, Federal Republic of Germany Fux: + 49 551 201 578. Abbreviations. PCR, polymerase chain reaction; Prodan, 6-proprion yl-2-dimethylaminonaphthalene. Note. The novel nucleotide sequence data published here have been deposited with the EMBL sequence data bank and are available under the accession number X65614. The novel amino acid sequence data have also been deposited with the EMBL sequence data bank.

pression pattern, which is not currently understood. SlOO polypeptides consist of two E F hands and display an additional C-terminal extension. Due to their relative abundance, the SlOO proteins are thought to be involved in cytoplasmic Ca2+ control (reviewed by Kligman and Hilt, 1988; Hilt and Kligman, 1991; Persechini et al., 1989). Some members, such as 2A9 (calcyclin) and 18A2, seem to be involved in cell cycle progression since they are expressed in a cell-cycledependent manner (Hirschhorn et al., 1984; Jackson-Grusby et al., 1987), while p l l acts as a protein ligand of annexin I1 and modulates the Ca2+-dependent properties of this Ca2' 1 phospholipid-binding protein (Gerke and Weber, 1984, 1985b; Powell and Glenney, 1987). Finally, a disulfidebonded SlOOj? homodimer seems to act extracellularly as a neurite extension factor (Kligman and Marshak, 1985; Winningham-Major et al., 1989). Except for calcyclin, p l l and the classical SlOO proteins SlOOa and SlOOfi, most other members of the family are defined only by DNA sequences. Since pl 1 is unusual in that it has lost Cazf-binding due to amino acid changes in both EF hands (Gerke and Weber, 1985a ; Glenney, 1986), most studies on Ca2+-binding center on SlOOa and Sloop prepared from brain tissue (Baudier and Gtrard, 1983; Mani et al., 1983; Baudier et al., 1986; Baudier and Cole, 1989; for review see Hilt and Kligman, 1991). Only lately has

542 some interest been focused on a homologous protein which can be purified from chicken gizzard in relatively good yield (Mani and Kay, 1990). In the course of some experiments on proteins of human placenta, we encountered a mixture of polypeptides containing a component of apparent molecular mass 11 kDa. When this component was subjected to N-terminal sequencing, a sequence potentially related to SlOO proteins was recognized. Here we report the cDNA cloning of this novel member of the SlOO protein family and the successful expression of the recombinant protein in Escherichia coli. The purified recombinant SlOOP was analyzed for its Ca2+-dependent properties by fluorescence spectroscopy. MATERIALS AND METHODS

SlOOP cDNA cloning Direct N-terminal sequencing of SlOOP from human placenta yielded, with the exception of position 27, an unambiguous sequence for the first 44 residues. Based on this protein sequence, two degenerate oligonucleotide primers of opposite orientation were synthesized (Hen 1 (forward primer, corresponding to amino acids 1 8); 5'-ATGAC[ATGC]GA[AG][CT]T[ATGC]GA[AG]AC[ATGC]GC[ATGC]ATGG; Hen 2 (reverseprimer,corresponding to amino acids 37 -44); 5'-A[AG][AG]AA [ATGC]CC [ATGC]G G [ATGC]A [AG][TC]TC[TC]TT[TC]TCC) and employed in a polymerase chain reaction (PCR) using human placenta cDNA as a template. The reaction was carried out in a buffer containing 50 mM KCl, 100 mM Tris/HCl pH 8.5,4.5 mM MgC12 and 2.5 mM each of dATP, dGTP, dCTP, and dTTP. 100 pg cDNA template and 1 pg of each oligonucleotide (Hen 1 and Hen 2) were added, the reaction volume was adjusted to 100 p1 and the mixture was incubated at 95°C for 8 min. Subsequently, 2 units Taq polymerase (Cetus-Perkin Elmer) was added and 35 cycles of denaturation (95"C, 1 min), annealing (37"C, 2 min), and polymerization (72"C, 2 min) were performed. The amplification yielded a 130-bp product which was gel-purified and cloned in to EcoRV-linearized pBluescript for sequence analysis. Dideoxy sequencing (Sanger et al., 1977) was carried out using the universal T3 and T7 primers (Stratagene) and a T7 sequencing kit (Pharmacia). To obtain a full-length cDNA clone, the PCR product was labelled by nick-translation (Sambrook et al., 1989) and employed to screen a human placenta cDNA library in 1 Zap I1 (Stratagene). Hybridization was carried out in 50% formamide, 50 mM sodium phosphate pH 7.5, 1 mM sodium pyrophosphate, 0.75 M sodium chloride, 0.075 M trisodium citrate, 5 mM EDTA, 5 x Denhardt's solution and 100 pg/ml sonicated salmon sperm DNA at 40°C. The final wash was at 55°C in 0.15 M sodium chloride, 0.015 M trisodium citrate, 0.1 YOSDS. Positive clones were grown on XLI-Blue cells and infected with the R408 helper phage (Stratagene) to excise the pBluescript plasmid from the 1Zap I1 vector. Plasmids were purified on Qiagen columns (Diagen) and cDNA inserts were characterized by restriction analysis and dideoxy sequencing. ~

an EcoRI site directly adjacent to the ATG start codon and TB2 (reverse primer; 5'-CCGAAGCTTGCATTTATTAATCAGAGGTAC) introduced a unique HindIII site 3' to the polyadenylation signal. PCR was carried out with TB1 and TB2 (1 pg each) on pBluescript DNA containing the Sloop cDNA insert (50 ng template DNA). Reaction buffer and conditions for denaturation, annealing, and polymerization were as described for the initial PCR with oligonucleotides Hen 1 and Hen 2 on the human placenta cDNA (see above). The amplification yielded a product of 422 bp which covered the entire-protein-coding region plus 116 bp of 3' nontranslated sequence. To generate the appropriate ends, the PCR product was digested with EcoRI and HindIII. gelpurified, and cloned into the EcoRI/HindIII-linearized expression vector pkk 223-3 (Pharmacia). Transformation of E. coli XL1 -Blue with the pkk-S100P construct led to high-level expression of S1OOP. For purification, large-scale cultures of E. coli cells (800 ml) bearing the pkk SlOOP plasmid were grown overnight at 37 "C in Luria broth medium containing 75 pg/ml ampicillin and 1 mM isopropyl 0-D-thiogalactopyranoside. Cells were harvested by low-speed centrifugation and resuspended in 20 ml40 mM imidazole/HCl pH 7.5,O.l M NaCl and 30 mM MgC12. To prevent proteoIysis, the buffer was made 2 mM in phenylmethylsulfonyl fluoride (Sigma) and 2.5 pM in [ ~ - 3 trans-carboxyoxiran-2-carbonyl]-leucyl-agmatin) (Peptide Institute, Osaka, Japan). The cell suspension was sonicated on ice (six 1-min periods at 30 W) using a Bransson sonifier. The supernatant obtained after high-speed centrifugation (1 h at 35000 rpm in a Beckman centrifuge) was dialyzed against 20 mM imidazole/HCl pH 7.5, 20 mM NaC1, 1 niM EGTA and 1 mM dithiothreitol, then applied to a 10-ml column of DEAE-cellulose (Whatman) DE-52 equilibrated in the same buffer. Using a shallow salt gradient, SlOOP was eluted at around 100 mM NaCl. Relevant fractions were pooled, dialyzed against 40 mM Tris/HCl pH 7.2, 80 mM NaCI, 2 mM CaClz and 1 mM dithiothreitol and applied to an 8ml phenyl-Sepharose column (Pharmacia) equilibrated in the same buffer. The column was washed with several column volumes of buffer 2 mM in CaCl, and then eluted with buffer containing 8 mM EGTA. Approximately 12 mg SlOOP was isolated from the original 800-ml culture. Fluorescence spectroscopy

Fluorescence emission spectra were recorded on an SLM model 8000 spectrofluorometer (Urbana, IL) between 300 400 nm, with excitation at 280 nm for the tyrosine fluorescence, or between 400 - 630 nm, with excitation at 380 nm for the 6-proprionyl-2-dimethylaminonaphthalene(Prodan) fluorescence. Spectroscopy was carried out at 20°C in 25 mM imidazole/HCl pH 7.5, 100 mM NaCl, 1 mM dithiothreitol in a total volume of 0.4 ml. The protein concentration was adjusted to 3.1 pM (tyrosine fluorescence) or 1.2 pM (Prodan fluorescence). Ca2+titrations of the tyrosine and Prodan fluorescence signals were performed by recording individual spectra after adjusting the Ca2 concentration in the cuvette by addition of CaCl, from appropriate stock solutions. +

Expression and purification of recombinant SlOOP

Two synthetic oligonucleotides were constructed and used in a PCR to generate unique restriction sites for straightforward cloning of the protein-coding region of SlOOP into procaryotic expression vectors. TB1 (forward primer; 5'CGTGAATTCATGACGGAACTAGAGACAGCC) installed

Chemical crosslinking

All crosslinking reagents were from Pierce (Rockford, IL, USA). To a 25 pM solution of SlOOP in 20 mM potassium phosphate pH 7.2, ethyl-3-(3-dimethylaminopropyl)-carbodiimide and N-hydroxysulfosuccinimidewere added to a final

543 I I

hl

~ 121 '4

181 54

GGTGGGTCTGAATCTAGCACC~~LMACGGAACTAGAGACAGCCATGGGCATGATCATAGAC M

T

E

L

E

T

A

M

G

M

l

i

GTCTTTTCCCGATATTCGGGCAGCGAGGGCAGCACGCACGCAGACCCTGACCAAGGGGGAGCTC ~ V F S R Y S G S E G S T ~ T L T K

GTGGATAAATTGCTCAAGGACCTGGACGCCAATGGAGATGCCCAGGTGGACTTCAGTGAG V D K L L K D L O A N G O A ~ V D F S TTCATCGTGTTCGTGGCTGCAATCACGTCTGCCTGTCACAAGTACTTTGAGAAGGCAGGA F I V F V A A I T S A C H K Y F E K A G

301

CTCAAATGATGCCCTGGAGATGTCACAGATTCCTGCAGAGCCATGGTCCCAGGCTTCCCA

9

G

E

L

ARPLEEALDVI

AAGGTGCTGATGGAGAAGGAGCTACCAGGCTTCCTGCAGAGTGGAAAAGACAAGGATGCC K V L M E K E L P G F L ~ S G K D K D A

74

241

52

D

4

L

K

T IGSKLQOA QNFCKKENKNEK E

53

95

+

361

AAAGTGTTTGTTGGCAATTATTCCCCTAGGCTGAGCCTGCTCATGTACCTCTGATTAATA

421

AATGCTTATGAAAAAAAAA -

~

Fig. 1. cDNA and protein sequence for SlOOP, a CaZ+-bindingprotein from human placenta. Except for position 27, the first 44 amino acid residues (underlined) were directly obtained by protein sequencing. The cDNA was isolated from a human placenta cDNA library made in IZapII. The start codon is boxed and the stop codon is indicated by an asterisk. The canonical polyadenylation signal in the 3' untranslated region is underlined.

concentration of 2 mM. After 30 min at room temperature, the reaction was quenched with 10 mM 2-mercaptoethanol (Vandekerckhove et al., 1989). A 25 pM solution of SlOOP in 0.1 M potassium phosphate pH 7.5, 100 mM NaC1, was treated with 1 mg/ml of ethylene glycobis(succinimidy1 succinate) for 30 min. The reaction was terminated with 10 mM glycine. Finally, a 25 pM solution of SlOOP in 0.1 M potassium phosphate pH 7.1 was treated with 1 mg/ml of disuccinimidyl tartrate for 30 min. The reaction was stopped with 10 mM glycine. Reaction products were analyzed by SDS/ PAGE. Miscellaneous procedures Protein sequencing was on an automated gas-phase sequenator (Knauer model 810) equipped with an on-line phenylthiohydantoin analyzer. For Prodan labeling, SlOOP in 0.1 M potassium phosphate pH 7.1, 100 mM NaCl, 5 pM C a 2 + ,1 mM dithiothreitol, was treated with a twofold molar excess of Acrylodan (Molecular Probes, Eugene, OR), the thiol-reactive derivative of Prodan (Prendergast et al., 1983), over the total amount of SH groups. After 30 min at 25°C an excess of 2-mercaptoethanol was added and the Prodanlabelled SlOOP was isolated by gel filtration through a PDlO column. SDSjPAGE was performed with the Tricine buffer system (Schagger and von Jagow, 1987). Electrophoretic blotting from SDSjPAGE on polyvinyldifluoride membranes (Millipore, Bedford, MD, USA) was done as described (Bauw et al., 1987). Protein concentrations of SlOOP were determined by quantitative amino acid analysis and refer to the monomer concentration. Concentrations of free Ca2+ were obtained from Ca2+/EGTA-buffered media using the procedure described by Miller and Smith (1984).

RESULTS Protein sequencing and cDNA cloning In a fraction of proteins from human placenta, a polypeptide of apparent molecular mass 11 kDa was partially characterized due to the absence of an N-terminal blocking group. SDSjPAGE followed by blotting onto a polyvinyldifluoride membrane provided the possibility to obtain the 15 N-ter-

Fig. 2. Alignment of amino acid sequences for the SlOO protein family. The sequence at the top (shaded throughout) is that of human SlOOP taken from Fig. 1. Other sequences from top to bottom: bovine SlOOa (Isobe and Okuyama, 1981), bovine SlOOJ (Isobe and Okuyama, 1978), bovine SlOOL (lung) (Glenney et al., 1989), murine 18A2 (Jackson-Grusby et al., 1987),porcine pl1 (Gerke and Weber, 1985a), porcine SlOOC (cardiac) (Ohta et al., 1991), human MRP8 (Odink et al., 1987), human calcyclin or 2A9 (Calabretta et al., 1986), human MRPl4 (Odink et al., 1987) and rabbit gizzarin (Watanabe et al., 1991). For alternative names which have been given to the individual members of the protein family, see Hilt and Kligman (1991). The sequence of the vitamin-D-dependent Ca2+-binding protein from bovine intestine is not incuded as this protein lacks the C-terminal extension typical for SlOO proteins (Szebenyi et al., 1981). Dashes indicate gaps necessary for optimal alignment. The Caz+-binding loops of the two consecutive EF hands are indicated by lines above the top sequence. Note that these loops are distorted in p l l , which is known to lack Ca2+ binding. Residues shared by SlOOP and one of the other 10 proteins are shaded in the latter sequences. Note the particularly close relationship of SlOOP and SlOOu (48 identical residues).

minal residues. Since these showed a strong similarity to bovine brain SlOOa, the N-terminal sequencing was repeated for 44 residues. Except for position 27 an unambiguous sequence was obtained. Since this partial sequence (Fig. 1) and the molecular mass argued for a novel member of the SlOO protein family, we decided to isolate a full-length cDNA clone from a human placenta cDNA library to establish the complete protein sequence. To obtain an appropriate hybridization probe, we employed sense and anti-sense oligonucleotides whose sequences were derived from the known protein sequences in polymerase chain reactions using human placenta cDNA as a template. This approach led to the amplification of a 130-bp fragment whose sequence corresponded to the SlOOP cDNA encoding amino acid residues 1- 44 (not shown). When this PCR product was used to screen a human placenta cDNA library constructed in /1 Zap 11, 30 positive clones were identified in a total of 400000 plaques. The cDNA inserts of several positive clones were characterized by dideoxy sequencing. The cDNA sequence obtained is given in Fig. 1. It is 439 bp in length, a value which is in good agreement with the size of the S1OOP transcript determined in Northern blot analysis (data not shown). The open reading frame of the SlOOP cDNA predicts a unique protein of 95 residues. The initiator methionine of the deduced sequence is kept as residue 1 of the placenta protein (Fig. 1). The polypeptide has a molecular mass of 10.4 kDa

544 two of these proteins have been called SlOOL and C, for lung and cardiac muscle respectively (Glenney et al., 1989; Ohta et al., 1991), we feel that the even greater similarity of the placenta protein and SlOOcl and Sloop is best characterized by the provisional name SlOOP.

and a calculated isoelectric point of 4.58. Fig. 2 gives the sequence alignment of SlOOP with other members of the SlOO protein family. It shows that the new placenta protein is particularly closely related to S1OOcl. It shares 48 identically placed residues with SlOOcr and 45 identities with SlOOg, while the latter two proteins show 53 identical residues. The similarity is somewhat lower in SIOOL, 18A2 and p l l (around 40 identical residues) and even further reduced in calgizzarin, SIOOC, MRP8, MRP14 and 2A9 (31 -36 identical residues). Since

Expression and purification of recombinant Sloop

24.014.3-

1

2

3

4

5

6

7

8

Fig. 3. Purification of recombinant SlOOP monitored by SDS/PAGE. Total extracts of E. coli with and without SlOOP are shown in lanes 3 and 2, respectively. Lane 4 shows the SlOOP preparation obtained after a high-speed cent-iifugation and chromatography on DEAEcellulose. A subsequent step using phenyl-Sepharose provides essentially pure S1 OOP (lane 5). A molecular mass standard is given in lane 1. Molecular masses of two standards are indicated at the side in kDa. Lanes 6-8 show the results of chemical crosslinking of SlOOP with ethyl-3-(3-dimethylamino-propyl)-carbodiimide plus N-hydroxysulfosuccinimide (lane 6), ethylene glycobis(succinimidy1 succinate) (lane 7) and disuccinimidyl tartrate (lane 8) followed by SDS/PAGE. Note the formation of a pronounced dimer band in lanes 7 and 8 and the nearly complete conversion to a dimer in lane 6.

a

To obtain quantities of SlOOP sufficient for a detailed biochemical analysis, we subcloned the protein region of the SlOOP cDNA into the bacterial expression plasmid pkk 2233. This was achieved by employing PCR to introduce an unique EcoRI site directly 5' to the ATG start codon and a HindIII site adjacent to the polyadenylation signal, and subsequent cloning of this amplified EcoRI - HindIII fragment into the appropriately linearized pkk 223-3 (see Materials and Methods). Due to this cloning protocol, the ATG start codon of SlOOP was placed six nucleotides downstream of the ribosome binding site of the expression plasmid, i. e. at a position optimal for efficient translation of the cloned gene (Amman et al., 1983). Transformation of E. colistrain JMlOl with the pkk SlOOP plasmid led to the synthesis of recombinant SIOOP. SDS/ PAGE of total cell extracts showed a strong band with an apparent molecular mass of 11 kDa, which was absent in the strain bearing the parental pkk 223-3 plasmid (compare lanes 2 and 3 in Fig. 3). After cell rupture and high-speed centrifugation, the recombinant protein was found in the supernatant. In line with its acidic PI, SlOOP was retained by DEAEcellulose and eluted from the anion-exchange column at around 0.1 M NaC1, when a shallow salt gradient was applied. Relevant fractions were pooled, dialyzed against a Ca2+-containing buffer and applied to a phenyl-Sepharose column. The column was washed with buffer plus 2 mM Ca2+ to remove contaminants and then step-eluted with a buffer containing 8 mM EGTA. Fig. 3, lane 5, shows that this procedure pro-

'

(a) 2 0 mM EGTA (bl 0 5 pM Ca (cl 1 0 p M C a (d) 2 0 VM Ca ( e ) 3 0 p M Ca ( f l 100 p M Ca

I

\\\\

(11

2 0 rnM Ca

0 400

450

500

550

600

Wavelength (nm)

Fig.4. Fluorescence emission spectra of Prodan-S100P at different Ca2+ concentrations. SlOOP labelled with Prodan (1.2 pM protein in 25 mM imidazole/HCI pH 7.5, 100 mM NaCl and 1 mM dithiothreitol), was excited a t 380 nm at the indicated concentrations of free C a 2 + .Spectra were recorded at 20°C. Note the red shift in the emission maximum with increasing C a Z +concentrations. This process is already completed at 10 pM C a 2 + .A plot of the shift versus the negative logarithm of the free C a 2 + concentration is given in Fig. 5a.

545

6.0

5.5

5.0

4.5

4.0

3.5

3.0

-Lag(Eal/Ml

Fig. 5. Caz+-binding sites of SlOOP monitored by fluorescence spectroscopy. (a) The shift in the maximum emission wavelength of Prodanlabelled SlOOP (Fig. 4) as a function of the negative logarithm of the free C a z + concentration. This titration curve reflects the high-affinity site (Kd around 1.6 pM). (b) The change in tyrosine fluorescence (excitation 280 nm, emission 306 nm) for SlOOP at 20°C (3.1 pM protein in 25 mM imidazole/HCl pH 7.5, 100 mM NaCI, 1 mM dithiothreitol) in arbitrary units as a function of the negative logarithm of the free Ca2+ concentration. This procedure measures the low-affinity site (Kd around 800 pM).

vides essentially pure SIOOP. More than 12 mg protein was prepared from an 800-ml culture. Sequence analysis showed that the recombinant protein has the same N-terminal sequence as originally established for the protein from human placenta. Thus the initiator methionine is retained as residue 2 both in E. coli and in human placenta (Fig. 1). In line with the lack of a tryptophan and the presence of only two tyrosines (positions 28 and SS), the ultraviolet absorption spectrum of SlOOP (data not shown) revealed the fine structure of a protein rich in phenylalanine (five residues/polypeptide chain). To determine whether SlOOP forms a dimer in physiological buffers, the protein was treated with various crosslinking reagents and subsequently analyzed by SDS/PAGE. Treatment with ethylene glycobis(succinimidyl succinate), a 1.5-nm crosslinker, and disuccinimidyl tartrate, a 0.7-nm crosslinker, led to the formation of a covalent dimer in addition to remaining monomer (Fig. 3, lanes 7 and 8). Nearly complete conversion to the covalent dimer was obtained with the zero-length crosslinker ethyl-3-(3dimethylamino-propy1)-carbodiimide (Fig. 3, lane 6). Ca2+-binding sites of SlOOP

SlOOP contains a single cysteine residue (Cys85) which is located in the C-terminal extension and also conserved in other members of the family (Fig. 2). This cysteine is of particular interest since in pl1 chemical modification of Cys82, which in the sequence alignment corresponds in position to Cys85 of SIOOP, inhibits the interaction with the intracellular p l l ligand annexin I1 (Johnsson and Weber, 1990). In addition, the corresponding cysteines in SlOOa and /?,Cys85 and Cys84, show an increased chemical accessibility in the presence of Ca2+ (Baudier and Gerard, 1983; Calissano et al., 1976). Thus the C-terminal extension of these 5-100 proteins is exposed to the solvent either permanently (in p l l , which has lost the ability to bind Ca2+) or in the Ca2+-conformation (SIOOa, p). To determine whether Ca2+-binding to SlOOP induces similar conformational changes, we have labelled Cys85 with the fluorophor Prodan and studied the spectroscopic properties of this Prodan-S100P derivative. Fig. 4 shows that fluorescence emission from the Prodan group is indeed significantly altered upon Ca2+binding. Both a C a 2 + induced red shift in the emission maximum from 485 nm to 502 nm and a marked reduction in fluorescence intensity are

observed, indicating a less hydrophobic environment of the Prodan group in the presence of Ca2+.A C a 2 + titration of the red shift reveals a Kd value of 1.6 pM for the Ca2+-binding site whose occupation alters the environment of the Prodan group on Cys85 (Fig. 5a). The presence of this high-affinity Ca2+-bindingsite is also reflected by the increased chemical accessiblity of Cys85 in the presence of C a 2 + .The efficiency of Prodan labeling of Cys85 is already strongly enhanced at 0.5 pM Ca2+ (data not shown). In analogy with Sloop, the high-affinity Ca2+ site, which is monitored by Ca2+-induced changes in the environment of Cys85, should be located in the C-terminal EF-hand (Baudier and Cole, 1989). In Sloop a second Ca2+site of lower affinity was identified in the N-terminal EF-hand. Occupation of this site alters the spectroscopic properties of the sole tyrosine in Sloop, Tyrl7 (Baudier and Cole, 1989). Since one of the two tyrosine residues (Tyrl8) of SlOOP corresponds in position to Tyrl7 of SlOOfl (Fig. 2), we decided to study a potential influence of C a 2 +binding on the spectroscopic properties of tyrosine residues in SIOOP. Fluorescence spectroscopy with the excitation wavelength set at 280 nm specifically monitors the tyrosine emission since SlOOP lacks a tryptophan residue. Fig. 5 b shows that the fluorescence signal obtained upon excitation at 280 nm is indeed altered upon addition of C a 2 + .The intensity of the emission maximum situated at 306 nm is significantly reduced in Ca2+-boundSIOOP. A Ca2+titration of this intensity decrease provides a i(d of 800 pM, indicating the presence of a low-affinity C a 2 + binding probably located in the N-terminal half of the molecule (see Discussion).

DISCUSSION We have isolated a complete cDNA clone encoding a novel member of the family of SlOO proteins. This protein, which is present in human placenta, is particularly closely related 10 the classic SlOOa protein. Over its 95 residues it shares 48 residues with SlOOa (Fig. 2) and is therefore called SIOOP. Fig. 2 documents the rapid growth of the SlOO family of proteins in recent years. While restricted to three members in 1985 (SIOOa, Sloop and p l l ) , it now covers at least ten members, two of which (calgizzarin and SlOOC) were reported (Ohta et al., 1991; Watanabe et al., 1991) when we had completed the cloning of SlOOP.

546 Several members of the SlOO protein family are essentially known only by cDNA cloning. Thus detailed molecular studies on physical-chemical properties and Ca2+ binding have been restricted to Sloop purified from brain (for references see Introduction), a smooth muscle protein, which is due to its high tyrosine content (Mani and Kay, 1990) is most likely not calgizzarin (Watanabe et al., 1991) and p l l (Gerke and Weber, 1985b; Glenney, 1986). Although the latter protein has lost Ca2+-bindingactivity due to mutations in both E F hand loops, it is currently the only member of the SlOO family of proteins for which a strongly binding protein ligand has been identified. p l l forms a tight complex (dissociation constant below 30 nM; Johnsson et al., 1988) with annexin 11, a Ca2 /phospholipid-binding protein of the annexin family of proteins (reviewed by Klee, 1988). To obtain an easily available SlOO protein, we have expressed the cDNA clone for SlOOP in E. coli using the expression plasmid pkk 223-3. The recombinant protein is efficiently synthesized in soluble form. After high-speed centrifugation, two column steps (DEAEcellulose and phenyl-Sepharose) provide the pure protein in very good yield (> 15 mg/l culture). Recombinant SlOOP behaves as a dimer in various crosslinking experiments and its N-terminal sequence retains the initiator methionine as does the natural protein from human placenta (Fig. 1). The sequences of the two EF-hand loops (Fig. 2) indicate that SlOOP should bind two C a 2 +ions/monomer. Previous work on Sloop has shown that the degenerate N-terminal E F hand provides a low-affinity Ca2+ site (& around 500 pM) while the C-terminal E F hand involves a high-affinity site (reviewed by Kligman and Hilt, 1988; Baudier and Cole, 1989; Hilt and Kligman, 1991). The view of two Ca2+sites of different affinities also holds for a related protein from chicken gizzard (Mani and Kay, 1990) whose relation to calgizzarin (Watanabe et al., 1991) remains unknown. I n Sloop the low-affinity Ca2+ site can be monitored by a decrease in fluorescence of Tyrl7 (Baudier and Cole, 1989), the sole tyrosine of this protein chain. A very similar decrease in tyrosine fluorescence at high Ca2 concentrations is seen in Sloop and monitors a C a 2 + site with a Kd around 800 pM (Fig. 5b). We tentatively attribute this effect to Tyrl8 of SIOOP, as it occupies the position corresponding to Tyrl7 of Sloop (Fig. 2). The second tyrosine of SlOOP (position 88) lies in the C-terminal extension past the second EF hand. Spectroscopic evidence for the high-affinity Ca2+ site of SlOOP is already reflected by the accessibility of the sole cysteine residue, which is strongly enhanced at 0.5 pM C a 2 + .Fluorescence spectra of SlOOP labelled with Prodan on Cys85 show a Ca”-dependent shift in the emission maximum from 485 nm to 502 nm. This red shift provides a titration curve indicating a Kd value of 1.6 pM for the strong Ca2+ site (Fig. 5a), which in analogy to Sloop should be located in the C-terminal E F hand (reviewed by Hilt and Kligman, 1991). In SlOOg this high-affinity Ca” site has a Kd value of 20 pM (Baudier and Gkrard, 1983). Cys85, the sole cysteine residue of SlOOP, is situated in the C-terminal extension and is relatively well conserved in other members of the SlOO family, which usually have an additional cysteine at variable positions (Fig. 2). p l l , another member of the protein family, lacks Ca2+ binding and forms a tight complex with its ligand annexin I1 (Gerke and Weber 1985a, b; Glenney, 1986). This interaction involves the N-terminal 12 - 14 residues of annexin II as an amphiphatic a-helix, which i s thought to make primarily hydrophobic contacts with the p l l molecule (Johnsson et al., 1988; Becker et al., 1990). The binding site on the p l l molecule seems to involve the C+

+

terminal extension since most hydrophobic residues in this region are well conserved in pl 1 from Xenopus to man (Kube et al., 1991) and chemical substitution of Cys82, which corresponds in position to Cys85 of SlOOP, abolishes the complex formation between p l 1 and annexin I1 (Johnsson and Weber, 1990). The combined results support the hypothesis that SlOO molecules interact with their cellular targets due to a Ca2+dependent conformational change of their C-terminal extensions (reviewed by Hilt and Kligman, 1991). While this concept is well supported in the Ca2+-independentcomplex between p l l and annexin 11, which seems to be in a permanent ‘on state’ in affecting its ligand, tightly binding ligands for other SlOO proteins, which are true Ca2+-binding proteins, still need to be identified. We thank Uwe Plessmann for protein sequence results, and Thomas Harder and Huub Dodemont for help in cDNA cloning and Northern blot analysis.

REFERENCES Amman, E., Brosius, J. & Ptashne, M. (1983) Gene 25, 167- 178. Baudier, J. & Cole, R. D. (1989) Biochem. J . 264, 79- 85. Baudier, J. &Gerard, D. (1983) Biochemistry 22, 3360-3369. Baudier, J., Glasser, N. & Gerard, D. (1986) J . B i d . Chem. 261, 8192- 8212. Bauw, G., Van Damme, J., Puype, M., Vandekerckhove, J., Gesser, B., Ratz, G. P., Lauridsen, J. B. & Celis, J. E. (1987) Proc. Nut1 Acad. Sci. U S A 86,7701 -7705. Becker, T., Weber, K. & Johnsson, N. (1990) EMBO J . 9, 42074213. Calabretta, B., Battini, R., Kaczmarek, L., de Riel, J. K. & Baserga, R. (1986) J . Biol. Chem. 261,12628-12632. Calissano, P., Mercanti, D. & Levi, A. (1976) Eur. J . Biochem. 71, 45-52. Gerke, V. & Weber, K. (1984) EM B O J. 3, 227-233. Gerke, V. & Weber, K. (1985a) EMBO J . 4, 2917-2920. Gerke, V. & Weber, K. (1985 b) J . Bid. Chem. 260,1688 - 1695. Glenney, J. R. Jr (1986) J . Biol. Chem. 261,7241-7252. Glenney, J. R. Jr, Kindy, M. S. & Zokas, L. (1989) J . Cell B id . 108, 569 - 578. Hilt, D. C. & Kligman, D. (1991) in Novel calcium-binding proteins (Heizmann, C . W., ed.) pp. 65-103, Springer Verlag, Berlin. Hirschhorn, R., Aller, P., Yuan, Z., Gibson, C. & Baserga, R. (1984) Proc. Nut1 Acad. Sci. U S A 81,6004 - 6008. Isobe, T. & Okujama, T. (1978) Eur. J . Biochem. 89,379-388. Isobe, T. & Okujama, T. (1981) Eur. J . Biochem. 116,79-86. Jackson-Grusby, L. L., Swiergiel, J. Q. & Linzer, D. 1. (1987) Nucleic Acids Res. 1.5, 6677 - 6690. Johnsson, N. & Weber, K. (1990) J . Biol. Chem. 26.5, 14464--4468. Johnsson, N., Marriott, G . & Weber, K. (1988) EMBO J . 7, 24352442. Klee, C. B. (1988) Biochemistry 27, 6645-6653. Kligman, D. & Marshak, D. (1985) Proc. Nut/ Acad. Sci. USA 82, 7136 - 71 39. Kligman, D. & Hilt, D. C. (1988) Trends Biochem. Sci. 3,437-443. Kube, E., Weber, K. & Gerke, V. (1991) Gene 102, 255-259. Mani, R. S., Shelling, F., Sykes, B. & Kay, C. M. (1983) Bioclzrmistry 22,1134-1740. Mani, R. S. & Kay, C. M. (1990) Biochemistry 29, 1398-1404. Miller, D. J. & Smith, G. L. (1984) Am. J . Physiol. 246, C16O-Cl66. Moore, B. (1965) Biochem. Biophys. Res. Commun. 19, 739-744. Odink, K., Cerletti, N., Brueggen, J., Clerc, R. G., Taresay, L., Zwadlo, G., Gerhards, C., Schlegel, R. & Sorg, C. (1987) Nature 330, 80 - 82. Ohta, H., Sasaki, T., Naka, M., Hiraoka, O., Miyamoto, C., Furuichi, Y . & Tanaka, T. (1991) FEBS Lett. 295,93-96. Persechini, A., Moncrief, N. D. & Kretsinger, R. H. (1989) Trends Neurosci. 12, 462 -467. Powell, M. A. & Glenney, J. R. Jr (1987) Biochem. J . 247,321 -328.

547 Prcndcrgast, F. G., Meycr, M., Carlson, G. L. Iida, S. & Potter, J. D. (1 983) J . Biol. Chem. 258, 7541 -7544. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular cloning, a laboratory manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl Acad. Sci. USA 73, 5463 - 5467. Schagger, H. & von Jagow, G. (1987) Anal. Biochenz. 166, 368-379.

Szebenyi, D. M. E., Obendorf, S. K. & Moffat, K. (1981) Nature 294, 327 - 332. Vandekerckhove, J., Kaiser, D. A. & Pollard, T. D. (1989) J . Cell Bid. 109, 619-626. Wdtanabe, M., Ando, Y., Todoroki, H., Minarni, H. & Hidaka, H. (1991) Biochem. Biophys. Rex Commun. 181, 644-649. Winningham-Major, F., Staecker, J. L., Barger, S. W., Coats, S. & Van Eldik, L. J. (1989) J . CeN Bid. 109, 3063 - 3071.

S100P, a novel Ca(2+)-binding protein from human placenta. cDNA cloning, recombinant protein expression and Ca2+ binding properties.

A novel member of the S100 protein family, present in human placenta, has been characterized by protein sequencing, cDNA cloning, and analysis of Ca(2...
804KB Sizes 0 Downloads 0 Views