Journal of Biological Chemistry · 1999 · 80 citations · 20 references
Life SciencesBiochemistryMolecular BiologyProtein-protein InteractionsSignal TransductionStructural BiologyCalcium-binding ProteinsProtein Complex AssemblyProtein DimerizationProtein DomainsProtein Interaction DomainsProtein StructureCalcium Binding DomainsCalcium Signal TransductionDimerization SpecificityEf-hand MotifsEf-hand ProteinsProtein Dimerization DomainsS100 ProteinsStructural Basis Of Dimerization
Calcium-binding S100 proteins are thought to play a central role in calcium-mediated signal transduction pathways. They consist of two helix-loop-helix, calcium-binding EF-hand domains. A characteristic feature is their tendency to form homo- and/or heterodimeric complexes. This report presents for the first time a functional “in vivo” approach to the analysis of S100 protein dimerization. Using the two-hybrid system we analyzed the dimerization of MRP8 (S100A8) and MRP14 (S100A9), two S100 proteins expressed in myeloid cells. It is reported that the MRP8-MRP14 heteromer is the clearly preferred complex in both man and mouse. The ability to homodimerize, however, appears to be restricted to the murine MRPs. Interaction analysis of chimeric murine/human MRP14 proteins indicates, that the C-terminal EF-hand domain plays a prominent role in MRP8-MRP14 interaction and determines the specificity of dimerization. Site-directed mutagenesis of four evolutionary conserved hydrophobic amino acids, which have been recently supposed to be essential for S100 protein dimerization, suggests that at least one of these, namely the most N-terminal located residue, is not critical for dimerization. Calcium-binding S100 proteins are thought to play a central role in calcium-mediated signal transduction pathways. They consist of two helix-loop-helix, calcium-binding EF-hand domains. A characteristic feature is their tendency to form homo- and/or heterodimeric complexes. This report presents for the first time a functional “in vivo” approach to the analysis of S100 protein dimerization. Using the two-hybrid system we analyzed the dimerization of MRP8 (S100A8) and MRP14 (S100A9), two S100 proteins expressed in myeloid cells. It is reported that the MRP8-MRP14 heteromer is the clearly preferred complex in both man and mouse. The ability to homodimerize, however, appears to be restricted to the murine MRPs. Interaction analysis of chimeric murine/human MRP14 proteins indicates, that the C-terminal EF-hand domain plays a prominent role in MRP8-MRP14 interaction and determines the specificity of dimerization. Site-directed mutagenesis of four evolutionary conserved hydrophobic amino acids, which have been recently supposed to be essential for S100 protein dimerization, suggests that at least one of these, namely the most N-terminal located residue, is not critical for dimerization. polymerase chain reaction galactosidase. S100 proteins comprise a family of low molecular weight calcium-binding proteins, which are specifically expressed in a variety of tissues and cell types. It is thought that they are primarily involved in Ca2+-mediated signal transduction. Changes in intracellular calcium levels alter the structure and function of these proteins. However, the specific functions of individual members of the S100 protein family in cell metabolism are not yet clear. They have been frequently described in association with cell growth and differentiation, with cell cycle regulation and with several human diseases as cancer and skin disorders (see Refs. 1Niki I. Yokokura H. Sudo T. Kato M. Hidaka H.J. J. Biochem . (Tokyo). 1996; 120: 685-698Crossref PubMed Scopus (121) Google Scholar, 2Schafer B.W. Heizmann C.W. Trends Biochem. Sci. 1996; 21: 134-140Abstract Full Text PDF PubMed Scopus (1035) Google Scholar, 3Zimmer D.B. Cornwall E.H. Landar A. Song W. Brain Res. Bull. 1995; 37: 417-429Crossref PubMed Scopus (816) Google Scholar, 4Hessian P.A. Edgeworth J. Hogg N. J. Leukoc. Biol. 1993; 53: 197-204Crossref PubMed Scopus (350) Google Scholar, for review). Recent investigations have given much emphasis on the structure of S100 proteins (5Kilby P.M. Van Eldik L.J. Roberts G.C. Structure. 1996; 4: 1041-1052Abstract Full Text Full Text PDF PubMed Scopus (119) Google Scholar, 6Potts B.C. Smith J. Akke M. Macke T.J. Okazaki K. Hidaka H. Case D.A. Chazin W.J. Nat. Struct. Biol. 1995; 2: 790-796Crossref PubMed Scopus (168) Google Scholar). The primary sequence, around 90–100 amino acids long, consists of two calcium-binding EF-hand domains bridged by a so called hinge region. Particularly the sequences of the N-terminal 14-amino acid variant EF-hand and the 12-amino acid C-terminal invariant EF-hand are highly conserved within the family and among family members of different species (3Zimmer D.B. Cornwall E.H. Landar A. Song W. Brain Res. Bull. 1995; 37: 417-429Crossref PubMed Scopus (816) Google Scholar, 4Hessian P.A. Edgeworth J. Hogg N. J. Leukoc. Biol. 1993; 53: 197-204Crossref PubMed Scopus (350) Google Scholar). The three-dimensional structure of calcyclin (S100A6), a typical S100 protein, has been determined by NMR spectroscopy (5Kilby P.M. Van Eldik L.J. Roberts G.C. Structure. 1996; 4: 1041-1052Abstract Full Text Full Text PDF PubMed Scopus (119) Google Scholar, 6Potts B.C. Smith J. Akke M. Macke T.J. Okazaki K. Hidaka H. Case D.A. Chazin W.J. Nat. Struct. Biol. 1995; 2: 790-796Crossref PubMed Scopus (168) Google Scholar). These data revealed, that a single monomer is composed of two calcium-binding helix-loop-helix motifs and forms a globular monomer containing four helices (helix I, II, III, and IV). Two monomers are oriented in an antiparallel fashion forming a noncovalent antiparallel homodimer. Dimerization, which is a characteristic feature for the S100 proteins, appears to be mediated by several evolutionary conserved hydrophobic residues located in helix I and IV, which contribute to the dimer interface. This interface has been called a homodimeric fold and appears to be unique among calcium-binding proteins (5Kilby P.M. Van Eldik L.J. Roberts G.C. Structure. 1996; 4: 1041-1052Abstract Full Text Full Text PDF PubMed Scopus (119) Google Scholar, 6Potts B.C. Smith J. Akke M. Macke T.J. Okazaki K. Hidaka H. Case D.A. Chazin W.J. Nat. Struct. Biol. 1995; 2: 790-796Crossref PubMed Scopus (168) Google Scholar). The dimeric complex is widely regarded as the biologically active form of most S100 proteins. The S100 proteins MRP8 and MRP14 (calgranulin A and B; S100A8 and S100A9) are expressed in myeloid cells and in epithelial cells like keratinocytes upon inflammatory activation (see Refs. 4Hessian P.A. Edgeworth J. Hogg N. J. Leukoc. Biol. 1993; 53: 197-204Crossref PubMed Scopus (350) Google Scholar, 7Sorg C. Behring Inst. Mitt. 1992; 91: 126-137PubMed Google Scholar, and 8Lagasse E. Clerc R.G. Mol. Cell. Biol. 1988; 8: 2402-2410Crossref PubMed Scopus (196) Google Scholar, for review). There is overwhelming evidence that they are involved in inflammatory processes (4Hessian P.A. Edgeworth J. Hogg N. J. Leukoc. Biol. 1993; 53: 197-204Crossref PubMed Scopus (350) Google Scholar, 7Sorg C. Behring Inst. Mitt. 1992; 91: 126-137PubMed Google Scholar, 9Odink K. Cerletti N. Brüggen J. Clerc R.G. Tarcsay L. Zwadlo G. Gerhards G. Schlegel R. Sorg C. Nature. 1997; 330: 80-82Crossref Scopus (559) Google Scholar). On the amino acid level the sequences show 60% homology among different species (10Lagasse E. Weissman I.L. Blood. 1992; 79: 1907-1915Crossref PubMed Google Scholar). Based upon the conserved amino acid sequence, the structural similarities, their tissue-specific expression pattern, and their ability to form dimeric complexes (11Teigelkamp S. Bhardwaj R.S. Roth J. Meinardus-Hager G. Karas M. Sorg C. J. Biol. Chem. 1991; 266: 13462-13467Abstract Full Text PDF PubMed Google Scholar) they are regarded as typical members of the S100 protein family. MRP8 and MRP14 represent an ideal model system to study the dimerization of S100 proteins. MRP8 and MRP14 homologues are known from a number of species, which facilitate the identification of important residues by sequence comparison and “domain swap” experiments. Additionally, a naturally occurring “deletion derivative” is known from human MRP14, the so-called MRP14* isoform, which is 4 amino acids shorter using a second ATG encoding methionine at position 5 as alternative translational start site (12van den Bos C. Roth J. Koch H.G. Hartmann M. Sorg C. J. Immunol. 1996; 156: 1247-1254PubMed Google Scholar). Similarly, different MRP14 isoforms were observed by two-dimensional gel electrophoresis in murine cells. 1J. Roth, unpublished data. With respect to the biological relevance of these isoforms it might be interesting to know whether the dimerization properties of the two proteins are different. Last, these S100 proteins are members of a subgroup of the S100 family, which are all coexpressed in the myeloid cell lineage as well as in activated and/or fetal keratinocytes. Beside MRP8 and MRP14 two other proteins belong to this group, namely human S100A12 (homologous to bovine CAAF1) (13Wicki R. Marenholz I. Mischke D. Schafer B.W. Heizmann C.W. Cell Calcium. 1996; 20: 459-464Crossref PubMed Scopus (66) Google Scholar, 14Yamamura T. Hitomi J. Nagasaki K. Suzuki M. Takahashi E. Saito S. Tsukada T. Yamaguchi K. Biochem. Biophys. Res. Commun. 1996; 221: 356-360Crossref PubMed Scopus (24) Google Scholar) and psoriasin (S100A7, homologous to bovine CAAF2) (15Hitomi J. Maruyama K. Kikuchi Y. Nagasaki K. Yamaguchi K. Biochem. Biophys. Res. Commun. 1996; 228: 757-763Crossref PubMed Scopus (17) Google Scholar, 16Madsen P. Rasmussen H.H. Leffers H. Honore B. Dejgaard K. Olsen E. Kiil J. Walbum E. Andersen A.H. Basse B. Lauridsen J.B. Ratz G.P. Celis A. Vandekerckhove J. Celis J.E. J. Invest. Dermatol. 1991; 97: 701-712Abstract Full Text PDF PubMed Google Scholar). Although these four S100 proteins are expressed by the same cells and are structurally highly homologous, they do not randomly interact with each other. Despite the potential biological relevance of the interaction specificity nothing is known about how this specificity is mediated in S100 proteins. Using the yeast two-hybrid system (17Fields S. Song O. Nature. 1989; 340: 245-246Crossref PubMed Scopus (4863) Google Scholar) this report presents evidence that the C-terminal domain of S100 proteins is primarily involved in protein-protein dimerization. The results are discussed with respect to recently published structural data (6Potts B.C. Smith J. Akke M. Macke T.J. Okazaki K. Hidaka H. Case D.A. Chazin W.J. Nat. Struct. Biol. 1995; 2: 790-796Crossref PubMed Scopus (168) Google Scholar) and may have implications for S100 proteins in general. Polymerase chain reaction (PCR)2-amplified cDNA fragments encoding the entire murine and human MRP's were fused in-frame to either the DNA-binding domain or activation domain of yeast plasmids pAS2-1 or pACT2 (CLONTECH Laboratories Inc., Palo Alto, CA), respectively. Vector pAS2-1 encodes the C-terminal GAL4 DNA-binding domain (amino acids 1 to 147) and the TRP1 selection marker. pACT2 encodes the N-terminal acidic GAL4 activation domain (amino acids 768 to 881) and the LEU2 selection marker. To facilitate directional cloning of the PCR products into the yeast expression vectors the 5′-primers contained either a NcoI orNdeI restriction site and the 3-primer contained aBamHI restriction site. This resulted in the in-frame fusion of each MRP to the 3′ end of either the GAL4 (1–147) DNA-binding domain (pAS2–1), or the GAL4 (768–881) activation domain (pACT2). To construct a vector encoding a chimeric S100A12/mMRP14 a PCR fragment of the N-terminal domain of human S100A12 was generated, which contained an EcoRI site at the 3′ end. The murine MRP14 fragment was restricted with EcoRI (human S100A12 amino acid lysine, position 32, murine MRP14 position 39, amino acid phenylalanine) before ligation (Fig. 1). The chimeric murine/human MRP14 encoding constructs hm14-pAS2-1 and mh14-pAS2-1 were also cloned by enzymatic restriction of the PCR-amplified fragments of human MRP14 and murine MRP14 with EcoRI and subsequent ligation of the fragments. The desired chimeric fragments were isolated via PCR with the corresponding primers and cloned into pAS2-1 (Fig.1). Deletion mutants (Fig. 2) were cloned into the pAS2-1 yeast expression vectors as described above. The deletions were generated by PCR using the murine MRP14 cDNA as template and the corresponding primer pairs. Amino acid substitutions located at the very ends of the MRP14 molecule were generated by using mutated primers for the PCR reaction. Construction of the internal MRP14 substitution mutants (Fig. 2) were achieved with the QuickChangeTM Site-directed Mutagenesis Kit (Stratagene, Heidelberg). Briefly, two oligonucleotide primers, which encode the mutation and are complementary to each other, extend the entire plasmid during temperature cycling by means of Pfu DNA polymerase. Parental DNA was digested with DpnI endonuclease (target sequence: 5′-Gm6ATC-3′), while the unmethylated PCR-amplified DNA could not be restricted by DpnI. The PCR-generated plasmid DNA was then transformed into Escherichia coli and the inserts of several clones analyzed. The integrity of all constructs has been verified by sequence analysis. To exclude nonspecific interactions, each vector was also solely transformed into yeast. Additionally, yeast containing human or murine MRP8 and MRP14 encoding vectors and the corresponding two-hybrid vector encoding human S100A12 were transformed as negative control. Transformation was performed by the high-efficiency lithium acetate method modified by Gietz et al. (20Gietz D. St. Jean A. Woods R.A. Schiestl R.H. Nucleic Acids Res. 1992; 20: 1425Crossref PubMed Scopus (2895) Google Scholar) using yeast strain Y190 (MATa, ura3–52, his3–20, ade2–101, lys2–801, trp1–901, leu2–3, 112, gal4D, gal80D, cyhr2, LYS2::Gal1UAS-His3TATA-His3, URA3::GAL1UAS-GAL1TATA-lacZ). Generally, vectors were sequentially transformed into yeast. A human leukocyte MATCHMAKER cDNA library (CLONTECH) was transformed into yeast Y190 carrying human MRP14 as bait. Screening procedure was performed according to the manufacturer's instructions. Blue colonies were isolated, inserts amplified via PCR, and sequenced according to standard methods. β-Galactosidase activity was assayed on plates and in liquid following the manufacturer's instructions (CLONTECH). Colony filter were performed by of from each into were at for to upon filter which was in and 1 with and in To the activity a single yeast was into 5 of and and at were to of liquid and the was to The yeast cells were for of activity following the Briefly, of the was in and the cells were was to the protein at and the was The activity is expressed in according to in Scholar). were performed following the manufacturer's instructions also described by et al. R. Smith K. in J. Inc., Scholar). least colonies were and standard were which were as negative were not to that each desired fusion protein was expressed in analysis with the GAL4 were The GAL4 activation domain and GAL4 DNA-binding domain which specifically to the activation domain (amino acid or the DNA-binding domain (amino acid of the GAL4 protein, were The of yeast protein and of the the manufacturer's instructions (CLONTECH). an was The yeast two-hybrid system (17Fields S. Song O. Nature. 1989; 340: 245-246Crossref PubMed Scopus (4863) Google Scholar) was to the interaction MRP8 and MRP14 and human MRP8 and MRP14 and murine MRP8 and MRP14 were cloned in-frame with the corresponding domains of the two-hybrid Additionally, the of the human S100A12 cDNA was cloned in-frame with the domain as a negative control. could not activity with or MRP or in with with protein from yeast clones containing the constructs described could the integrity of the construct and the expression of the fusion proteins not of all of MRP8 and MRP14 encoding plasmids that the of is preferred This was for both the human and murine MRP proteins. human and murine to with respect to The MRP8 and MRP14 form homodimeric this interaction was the heterodimeric complex could not of the human (Fig. To the that human MRP14 may interact with so MRP14 or other members of the S100 family, which in a leukocyte library was to for human MRP14 proteins. colonies were analyzed in and all of the human MRP8 The amino acid sequence of the MRP8 and MRP14 proteins are highly homologous to each other among and whether and human are to form could human MRP14 murine MRP14 homodimeric complex the structural is to protein-protein a human and murine To which of the MRP14 protein determines the specificity of protein-protein we chimeric MRP14 (Fig. 1). The domain domain with murine MRP8 and murine MRP14, not with human On the other the with human we could activity with human MRP14 with the murine (Fig. like a human MRP14 and like a murine MRP14 with respect to their dimerization the complex of the chimeric with MRP14 or was with of the MRP14 or A human MRP14 chimeric molecule not with human or murine analysis of MRP8-MRP14 encoding encoding encoding about of all which were for homo- and in the two-hybrid to the in in Scholar) according to et al. R. Smith K. in J. Inc., the in are Interaction of the murine and human were The of interaction on the of vector for MRP8 and MRP14, respectively. This results in for the interaction of and the observed in at least experiments. are were not negative the human S100A12 N-terminal N-terminal DNA-binding in a about of all which were for homo- and in the two-hybrid to the in in Scholar) according to et al. R. Smith K. in J. Inc., the in are Interaction of the murine and human were The of interaction on the of vector for MRP8 and MRP14, respectively. This results in for the interaction of and the observed in at least experiments. are were not negative the human S100A12 N-terminal N-terminal DNA-binding To the role of the and N-terminal a number of constructs were in the yeast of the N-terminal amino acids of murine MRP14 the of the interaction (Fig. complex was This to a potential murine MRP14* The corresponding was also for the naturally occurring human MRP14* isoform, which is 4 amino acids shorter using a second methionine in position 5 as translational start site. Deletion of these four N-terminal amino acids of the human MRP14 also not the interaction with A single amino acid at position from to to a of interaction A of C-terminal deletions was (Fig. It is known that the MRP14 molecule a C-terminal of about amino acids, which is among S100 proteins. A of these C-terminal amino acids in murine MRP14 not homo- or however, namely and which helix located in the C-terminal EF-hand the ability to interact with or (Fig. Recent the dimeric structure of S100 protein calcyclin via NMR spectroscopy (6Potts B.C. Smith J. Akke M. Macke T.J. Okazaki K. Hidaka H. Case D.A. Chazin W.J. Nat. Struct. Biol. 1995; 2: 790-796Crossref PubMed Scopus (168) Google Scholar). Based upon these data a number of evolutionary conserved amino acid residues in helix I and to be for the S100 protein dimerization. To the functional relevance of these residues several murine MRP14 mutants were by single amino acids (Fig. at position to at position to at position to and at position to amino acid resulted in a or of activity (Fig. was to with MRP8 is regarded to be an essential for the biological functions of S100 proteins. the two-hybrid system to the dimerization of the S100 proteins MRP8 and to the recently published which the structure of S100 proteins in (6Potts B.C. Smith J. Akke M. Macke T.J. Okazaki K. Hidaka H. Case D.A. Chazin W.J. Nat. Struct. Biol. 1995; 2: 790-796Crossref PubMed Scopus (168) Google Scholar, Chazin W.J. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar) the two-hybrid system a functional “in vivo” approach to the molecular analysis of S100 have that the two-hybrid system be to and/or protein-protein J. 1995; PubMed Scopus (66) Google Scholar). This report that it may also and about the S100 protein dimerization. the interaction the murine and human MRP8 and is that is preferred to in both This is in with which that the human form and (11Teigelkamp S. Bhardwaj R.S. Roth J. Meinardus-Hager G. Karas M. Sorg C. J. Biol. Chem. 1991; 266: 13462-13467Abstract Full Text PDF PubMed Google Scholar). Recent data that is a in the interface of human MRP8-MRP14 complex that be in the MRP8 and MRP14 Chazin W.J. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar). However, the of could be in one MRP species is of both proteins the is at least preferred Chazin W.J. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar). to it whether in human MRP at are to to the of the homodimeric complexes is with the This in human and murine is the primary structure of the of both species is highly It be interesting to which structural of the murine the that both and the of murine could recently be verified in by and experiments. was not in this report J. PubMed Scopus Google Scholar). that also in the heteromer is the naturally occurring form at least in which both MRPs. the human and murine dimerization that are not also functional this it is that the murine MRP8 a activity on myeloid cells M. P. G. S. J. Immunol. 1993; Google a which has not been observed with human It to the structural and a potential functional relevance of this human and murine MRPs. A comparison of the primary structure of S100 proteins that the MRP14 sequence a C-terminal which is unique among S100 proteins (Fig. It was that this and sequence is involved in protein and subsequent to might activity J. P. Hogg N. Nature. 1989; PubMed Scopus Google Scholar). To study the of this on dimerization a of C-terminal were The of the residues containing the not murine MRP14 homo- and deletions which helix MRP interaction the of this for MRP complex This that the unique is not essential for and may for the that this might be involved in interaction with other There is a second MRP14 form in human myeloid the so-called MRP14* four N-terminal amino It was that MRP14* may different biological functions (12van den Bos C. Roth J. Koch H.G. Hartmann M. Sorg C. J. Immunol. 1996; 156: 1247-1254PubMed Google Scholar). The reported that the naturally occurring human MRP14* is to with human However, complex is The analysis of the MRP14 that the at position of the MRP14 is involved in complex This is other that do not contribute to the S100 complex (11Teigelkamp S. Bhardwaj R.S. Roth J. Meinardus-Hager G. Karas M. Sorg C. J. Biol. Chem. 1991; 266: 13462-13467Abstract Full Text PDF PubMed Google Scholar, Chazin W.J. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar). Similarly, the N-terminal murine MRP14, the potential murine is also to form homo- and The murine MRP14 not encode a the may be of biological relevance to be that the N-terminal the C-terminal residues of MRP14 are critical for the of MRP dimerization. Despite their structural the murine and human are not to interact species of chimeric that the C-terminal domain determines the specificity of interaction with respect to species specificity as well as (Fig. the of protein-protein association is using the chimeric The chimeric human MRP14 molecule not with in that not N-terminal S100 domain the homologous that several conserved hydrophobic residues located in both the N-terminal EF-hand domain (helix and the C-terminal EF-hand domain (helix the S100 dimer (6Potts B.C. Smith J. Akke M. Macke T.J. Okazaki K. Hidaka H. Case D.A. Chazin W.J. Nat. Struct. Biol. 1995; 2: 790-796Crossref PubMed Scopus (168) Google Scholar). we mutated the corresponding amino acids by mutagenesis (Fig. The analysis of of these and to the of these with could a activity be in the one of these evolutionary conserved residues located in the C-terminal domain which were recently by NMR spectroscopy to be involved in S100 protein-protein association (6Potts B.C. Smith J. Akke M. Macke T.J. Okazaki K. Hidaka H. Case D.A. Chazin W.J. Nat. Struct. Biol. 1995; 2: 790-796Crossref PubMed Scopus (168) Google Scholar) is not essential for MRP8-MRP14 complex The data with the chimeric MRP14 and the single amino acid substitutions that the structural for dimerization to be for the N-terminal domain of S100 proteins. that the C-terminal most residues of helix IV, determines the specificity of interaction and is the critical domain for S100 protein dimerization. B. W. Schafer for the S100A12 cDNA T. C. and M. for critical the and
20
A novel genetic system to detect protein–protein interactions
Stanley Fields, Ok-Kyu Song · Nature · 1989 · 6.1K citations
Improved method for high efficiency transformation of intact yeast cells
Daniel Gietz, Andrew St. Jean, Robin A. Woods et al. · Nucleic Acids Research · 1992 · 3.3K citations · Full text
Two calcium-binding proteins in infiltrate macrophages of rheumatoid arthritis
Karel Odink, Nico Cerletti, Josef Brüggen et al. · Nature · 1987 · 699 citations