Characterization of the human secreted phosphoprotein 24 gene (SPP2) and comparison of the protein sequence in nine species

Matrix Biology 22 (2004) 641–651

0945-053X/04/$30.00� 2003 Elsevier B.V.yInternational Society of Matrix Biology. All rights reserved.doi:10.1016/j.matbio.2003.12.001

Characterization of the human secreted phosphoprotein 24 gene(SPP2)and comparison of the protein sequence in nine species�

Clare S. Bennett, Hamid R. Khorram Khorshid, J. Alexandra Kitchen , David Arteta ,1 2

Raymond Dalgleish*

Department of Genetics, University of Leicester, University Road, Leicester LE1 7RH, UK

Received 1 July 2003; received in revised form 11 November 2003; accepted 3 December 2003

Abstract

Secreted phosphoprotein 24(spp24) is a member of the cystatin superfamily, which was first identified in cattle as a minorcomponent of cortical bone and subsequently has been identified as a component of the fetuin-mineral complex. We havelocalized the humanSPP2 gene, which encodes spp24 to chromosome 2q37.1, determined its structure and mapped the start oftranscription in liver. There is no CAAT orTATA box in the promoter region but potential transcription factor(TF)-binding siteshave been identified. The gene comprises eight exons spread over a region of approximately 27 kb with the cystatin-like regionof spp24 encoded by four exons, rather than the three-exon structure typical of the genes encoding the archetypal cystatins. Arare single amino acid polymorphism(p.S38F) has been identified within the mature protein and its significance has been assessedby comparing the sequence of human spp24 with that of eight other species.� 2003 Elsevier B.V.yInternational Society of Matrix Biology. All rights reserved.

Keywords: Secreted phosphoprotein 24;SPP2; Cystatin superfamily; Fetuin-mineral complex

1. Introduction

Secreted phosphoprotein 24(spp24) was originallydescribed as a 24-kDa bone matrix protein from bovinecortical bone(Hu et al., 1995). The N-terminal 107residues of the protein are related in sequence to thecystatin family of thiol protease inhibitors(Turk andBode, 1991; Bobek and Levine, 1992; Henskens et al.,1996; Brown and Dziegielewska, 1997). This is fol-lowed by a region of phosphorylated serine residuesseparating the cystatin-like region from an arginine-richC-terminal domain that shows no homology to any

� The human liver cDNA library described in this paper has beendeposited with the UK MRC Geneservice, Babraham, UK from whomaliquots may be obtained(http:yywww.hgmp.mrc.ac.ukygeneserviceyindex.shtml).*Corresponding author. Tel:q44-0-116-252-3425; fax:q44-0-

116-252-3378.E-mail address: [email protected](R. Dalgleish).

Present address: Department of Pathology, University Hospitals1

of Leicester NHS Trust, Leicester, UK.Present address: PROGENIKA-MEDPLANT GENETICS, S.L.,2

Edificio 801, Parque Tecnologico de Zamudio, 48160, Derio, Spain.´

known protein. Several functions have been suggestedfor spp24 including that it might act as a proteaseinhibitor in the manner of the cystatins or that the C-terminal domain is possibly cleaved off to become abiologically active peptide, the cystatin-like region act-ing as a precursor carrier. This latter suggestion issupported by the close relatedness of the spp24 cystatin-like domain to the third cystatin domain of kininogenand to the cystatin domain of the neutrophil antibioticprecursors(Hu et al., 1995). A third possibility is thatspp24 has a function that is either similar or comple-mentary to fetuin-A which, like spp24, accumulates inthe extracellular matrix of bone and is synthesized inbone and liver(Ohnishi et al., 1993). Fetuin-A has twocystatin domains, a phosphorylated region and a C-terminal domain that may be a precursor to a biologi-cally active peptide. Many functions, some of whichremain controversial, have been suggested for the humanfetuin-A, a -HS glycoprotein. However, it is generally2

accepted that it inhibits calcium salt precipitation inserum and modulates apatite formation in bone(Schinkeet al., 1996) and plays a central role in regulating

642 C.S. Bennett et al. / Matrix Biology 22 (2004) 641–651

Fig. 1. Nucleotide and deduced amino acid sequences of human spp24cDNA. The putative start codon, stop codon, polyadenylation signaland signal peptide are underlined. Numbers to the left correspond tothe single-letter amino acid sequence. Numbers to the right correspondto the nucleotide sequence.

Fig. 2. RNA dot blot analysis of the expression of theSPP2 gene. AcDNA probe for the gene was hybridized to an array of human RNAsfrom various tissues(Clontech MTE). Strongest hybridization was tofetal liver (D11) and adult liver(A9) with weaker hybridization tofetal kidney (C11). A low level of hybridization toE. coli DNA(D12) is probably due to slight contamination of the probe withE.coli DNA.

osteogenesis(Binkert et al., 1999). Recent evidenceindicates that fetuin-A is a systemically-acting inhibitorof ectopic calcification(Schafer et al., 2003) functioning¨by way of the formation of a fetuin-mineral complex(FMC) (Price and Lim, 2003) of which spp24 is acomponent(Price et al., 2003).The humanSPP2 gene, encoding spp24, has previ-

ously been mapped to the tip of the long arm ofchromosome 2(2q37-qter) by in situ hybridization(Swallow et al., 1997). We report here the structure ofthe humanSPP2 gene, an analysis of its promoter,discovery of an amino acid variant and a comparison ofthe protein sequence with that of eight other species.

2. Results

2.1. Messenger RNA sequence, pattern of expressionand the start of transcription

We identified a number of human cDNAs similar tothe published bovine mRNA sequence(accession num-ber U03872) by FASTA searches of the EMBL DNAsequence database. We aligned these cDNAs, and otherscomprising human UniGene cluster Hs.12230, into aconsensus cDNA sequence that was initially incompleteat the 59 end relative to the bovine sequence. To isolatefull-length sequences, we prepared a cDNA library of2.5=10 clones from normal human liver and a screen6

of 10 000 cDNAs by hybridization with a part-lengthcDNA probe, derived from I.M.A.G.E. ConsortiumClone ID 204242(Lennon et al., 1996), yielded threepositive clones whose identity was confirmed by DNAsequencing. The longest of the three contains an addi-tional 253 bp relative to the initial consensus sequence.We have deposited the sequence of the 1019-bp full-

length mRNA(Fig. 1) in the EMBL database with theaccession number AJ308009.Translation of the mRNA is predicted to initiate at

the 59-most ATG as assessed using the ATGpr analysistool (Salamov et al., 1998) resulting in a predictedsignal peptide of 29 amino acids. The mature peptide is182 amino acids in length and comprises the same threedomains originally identified in the bovine protein.There is no evidence that either the 59-UTR (113 bp)or the 39-UTR (298 bp) of the mRNA contains anyfunctional elements that might influence gene expressionas judged by analysis with the UTR scan analysis tool(Pesole and Liuni, 1999).We assessed the pattern of expression of spp24 in

humans by hybridization of a full-length spp24 cDNAprobe to a human multiple tissue expression(MTE)array (Clontech-BD Biosciences) (Fig. 2) which con-tains 76 adult and fetal tissue-specific poly A RNAs.q

The strongest hybridization was to adult liver(A9) andfetal liver (D11), with weaker hybridization to fetalkidney (C11). There was also hybridization toEscheri-chia coli DNA (D12) probably reflecting a small amountof contaminatingE. coli DNA in the probe preparation.Very weak hybridization(visible in longer exposures)was also seen with the human genomic DNA sampleson the array(G12 and H12). The array does not containmRNA from bone so comparisons with bovine expres-sion are not possible.We mapped the start of transcription using primer

extension with RNA isolated from human adult liver toa single strong start site that is depicted as base 1 ofthe sequence in Fig. 1(Fig. 3). The sequence aroundthe transcription start site in liver corresponds well withthe consensus transcriptional initiator(Inr) sequence(Lo and Smale, 1996), the only mismatch being atpositionq4 which is consistent with there being low

643C.S. Bennett et al. / Matrix Biology 22 (2004) 641–651

Fig. 3. Identification of the transcription start point in human liver byprimer extension. Messenger RNA from human liver was reverse tran-scribed using a radiolabeled primer(indicated by the horizontalarrow). The products were run on a denaturing polyacrylamide gelalongside the DNA sequencing products obtained from the corre-sponding region of genomic DNA using the same primer. The majortranscription start point is indicated in the DNA sequence by the ver-tical arrow. The translation start codon is boxed.

Fig. 4. Map of humanSPP2 depicting the eight exons of the gene.Exons 1–4 encode the cystatin-like region and exons 5–7 the non-cystatin-like region. Exon 8 encodes only untranslated sequence. Thelocations of three di-nucleotide repeats and one tri-nucleotide repeatare indicated.

levels of mRNA detected in the mRNA array hybrid-ization.

2.2. Chromosomal location and gene organization

Using the Ensembl Genome Browser(Hubbard et al.,2002) we have located the gene encoding human spp24(designatedSPP2) to chromosome 2q37.1, in the inter-val 233.64–233.67 Mb which is consistent with ourinitial assignment to chromosome 2q37™qter by fluo-rescence in situ hybridization(Swallow et al., 1997).The immediately neighboring identifiable genes areTRPM8 (transient receptor potential cation channel,subfamily M, member 8), which lies 60 kb proximal,and ARL7 (ADP-ribosylation factor-like 7), which lies410 kb distal. TheSPP2 gene comprises eight exonsand seven introns(Fig. 4), with all the intron boundariesconforming to the GTyAG consensus. The ATG startcodon is located in the first exon and the TAA stopcodon in the penultimate exon, the final exon codingexclusively for the 39-UTR. This location of the stopcodon in the penultimate exon is a feature of only 7%

of genes, of which 70% encode secreted or cell-surfaceproteins (Nagy and Maquat, 1998), and is consistentwith the presence of a putative signal peptide in spp24.The cystatin-like domain of spp24 is encoded by thefirst four exons of the gene, with exon 1 encoding solelythe signal peptide; the non-cystatin-like region is encod-ed by exons 5–7. The four-exon organization of thecystatin-like region contrasts with that of the classicalcystatins, which are encoded by three exons(Betts etal., 2001), suggesting that the gene encoding spp24 is anovel member of the cystatin superfamily.We carried out a comprehensive analysis of the region

of DNA encodingSPP2 using the NIX DNA sequenceanalysis environment(Williams et al., 1998). Each ofthe exons identified by our comparison of the cDNAsequence with genomic DNA was confirmed by NIXwith no evidence for any additional alternatively splicedexons. Four simple tandem repeats(STRs) and a numberof dispersed repetitive elements were identified in andaroundSPP2 and we have shown that each of the fourSTRs(Fig. 4) are polymorphic in the population(Ben-nett and Dalgleish, unpublished). We have deposited afully annotated genomic DNA sequence in the EMBLdatabase with the accession number AJ272265.

2.3. Sequence variation of SPP2

We designed PCR primers to allow amplification ofthe seven protein-coding exons ofSPP2. The amplifiedproducts from 75 unrelated normal northern EuropeanCaucasian individuals were analyzed by conformationsensitive gel electrophoresis(CSGE) and we detected asingle variant in the coding region which we verifiedby direct sequencing of the amplified DNA(Fig. 5).The variant is a C to T transition at base 198 of thecDNA sequence presented in Fig. 1 which results in thesubstitution of serine by phenylalanine at amino acid 38of the precursor protein; amino acid 9 of the matureprotein. The systematic designations(Antonarakis andthe Nomenclature Working Group, 1998; den Dunnenand Antonarakis, 2000) of the sequence alterations at


Fig. 5. DNA sequence chromatogram depicting the human p.S38Fvariant. The DNA sequence encoding amino acids 37–39 is depictedfrom an individual heterozygous for the variant which is indicated by‘N’ at amino acid 38(amino acid 9 of the mature protein).

the cDNA and protein level are c.113C)T and p.S38F,respectively. The alteration to the DNA sequence in thevariant allele results in loss of a recognition site for therestriction enzymeFokI (GGATG) (data not shown).Using this as a basis for detecting variant alleles, wescreened 285 human DNA samples of Caucasian originand identified three individuals heterozygous for thevariant and no homozygotes. This indicates an allelefrequency of approximately 0.0053 and an expectationof approximately one person in 36 000 being homozy-gous for the allele if the sampled population is inHardy–Weinberg equilibrium.

2.4. Protein sequence comparison with other species

To investigate the significance of the amino acidsubstitution with respect to its potential to alter thefunction of the protein, we compared the sequences ofspp24 from as many species as possible. The cattle andrat protein sequences were obtained from the Swissprotprotein database(accession number Q27967) and theNCBI RefSeq database(accession number XP_217472),respectively. We confirmed the amino acid sequence forbovine spp24 by sequencing cDNA obtained by PCRamplification of reverse-transcribed bovine liver mRNA.We were also able to PCR amplify the correspondingregion of ovine spp24 cDNA using the bovine primersat the same annealing temperature, providing us withthe protein sequence for the entire open reading frame.We deduced the mouse protein sequence from a full-

length cDNA sequence compiled by alignment of theESTs for spp24 comprising the mouse UniGene clusterMm.28247 using the original sequence traces whereavailable. For pig, salmon and chicken, we identifiedspp24 cDNAs by BLAST searching of the GenbankDNA sequence database. Because the single pig cDNAsequence for spp24 was incomplete, we obtained theoriginal cDNA clone(127266 MARC 1PIG; accessionnumber BE015092), re-sequenced it and made correc-tions. The five salmon spp24 cDNAs that were identified(accession numbers BG934982, BG935174, BG935259,

BG935680, and BI468172) were aligned and a consen-sus cDNA produced using these data along with addi-tional unpublished sequencing traces provided by theoriginators of the cDNAs and re-sequencing data fortwo of the cDNAs(LRR3-c07 and LRR5-a06(Daveyet al., 2001)). A consensus cDNA sequence was com-piled for trout spp24 from the alignment of 69 ESTs,which were identified by searching dbEST using theconsensus salmon spp24 cDNA sequence as the query.The single matching chicken cDNA(accession numberU20160), which had previously been identified as agrowth hormone responsive gene(GHRG-1) (Agarwalet al., 1995), predicted a protein approximately one thirdshorter than spp24. The published account of GHRG-1also included the partial sequence of the promoter region(accession number U20161). Our re-appraisal of thesepublished cDNA and promoter sequences reveals thatthe former is in fact a chimera of cDNA and genomicsequence and the latter covers the promoter, exon 1, andthe start of intron 1 of the gene. Further cDNA sequenceswere identified by BLAST searching of the Universityof Delaware Chicken EST database(http:yywww.chickest.udel.edu) and, from the available data, aconsensus sequence was derived.We have deposited the cDNA sequences for mouse,

pig, salmon, trout, chicken, and sheep in the EMBLDNA database with the accession numbers AJ315513,AJ308100, AJ428527, AJ580808, AJ428876 andAJ544160, respectively. A revised sequence for thechicken promoter, exon 1, intron 1 and the start of exon2 has also been deposited in the EMBL DNA databasewith the accession number BN000081.The amino acid sequences for spp24 from human,

mouse, rat, cattle, pig, sheep, chicken, salmon and troutwere aligned(Fig. 6) using GCG pileup and displayedin GCG Seqlab where the alignments were optimizedmanually. The alignment reveals that the amino acidcorresponding to human amino acid 38(amino acid 9of the mature protein) is not absolutely conserved acrossspecies. Human, mouse, rat, and pig have serine at thisposition whereas cattle and sheep have alanine. Themore evolutionarily distant salmon, trout and chickenhave alanine, alanine and proline, respectively, at thisposition. Even though amino acid 9 of the mature proteinis not absolutely conserved, the nine species exhibit ahigh level of amino acid conservation in the region,highest among the mammals. The consequences, if any,of the substitution of serine by phenylalanine at aminoacid position 9 are likely to be structural as the bulkierside chain of the latter would probably not be as easilyaccommodated as that of the former into the three-dimensional structure of the protein. We also consideredthe potential of the substitution to affect the possiblephosphorylation of the protein using the NetPhos proteinphosphorylation prediction computer program(Blom etal., 1999). Analysis of the wild-type and mutant protein


Fig. 6. Alignment of the amino acid sequence of spp24 from mouse, rat, cattle, sheep, pig, human, chicken, salmon and trout with the thirdcystatin-like domain of bovine kininogen. The alignment as generated using GCG pileup and was optimized manually. Identical and similar aminoacids are shown on black and gray backgrounds, respectively. The disulfide bridges are indicated by brackets.

sequences predicts that the serine at amino acid 9 is nota substrate for phosphorylation, but also that the substi-tution has no effect on the strongly-predicted phospho-rylation of the serine at the following residue which isconserved in all species except chicken. Beyond this, itis not possible at present to speculate further on thesignificance of the variant.The alignment of the amino acid sequences for spp24

from nine species(Fig. 6) allows a preliminary assess-

ment of the conservation of amino acid sequence acrossthe entire protein. To help in that assessment, the thirdcystatin-like domain of bovine kininogen(Swissprotaccession number P01045) was also included as thishad previously been identified as a cystatin-like domainto which spp24 has similarity(Hu et al., 1995). The sixmammals for which we have data each produce proteinsof approximately the same length with considerableconservation of sequence. In contrast, the chicken, salm-


on and trout sequences exhibit greater divergence, beingboth less conserved and shorter in the C-terminal non-cystatin region. Among the mammals, there is markedconservation of positively- and negatively-charged ami-no acids within this region, especially at the extreme C-terminal end of the protein sequence. The region alsocontains several serines, tyrosines and threonines, someof which are predicted by NetPhos analysis to bephosphorylated, though the predicted residues differfrom species to species(data not shown). In spite ofthe divergence observed in chicken, salmon and trout,all nine species have glutamic acid at the extreme C-terminus which may be of functional significance.The N-terminal 41 amino-acid region which is encod-

ed by exon 2 in human is highly conserved between themammals, more so than rest of the protein sequence.This suggests that this region may be under greaterselective pressure than the disulfide-bonded region thatimmediately follows(encoded by exons 3 and 4), whichperhaps plays a structural role that is not as tightlyconstrained. There is absolute conservation of the twopairs of cysteine residues characterizing the protein ascystatin-like and the sequence immediately before thefirst of the four cysteines is highly conserved. Theamino acid motif ETTC is absolutely conserved in allsix mammals as well as in the aligned kininogensequence and also in the corresponding region of human,mouse and bovine fetuin-A cystatin-like domain 1(datanot shown). In chicken spp24 the corresponding motifis ETEC and in salmon and trout it is ETDC which isa motif conserved in the human and mouse fetuin-Bcystatin-like domain 2(data not shown). NetPhos pre-dicts that both threonines in the ETTC motif will bephosphorylated in human, cattle, sheep and pig spp24,as will the single threonine in chicken and both threo-nines in the identical motif in bovine kininogen. How-ever, neither of the pair of threonines in mouse and rat,nor the single threonine in salmon and trout, are pre-dicted to be phosphorylated. The presence of theseacidic amino acids and threonines with the potential tobe phosphorylated suggests that a net negative chargein the vicinity of the first cysteine may be functionallyimportant.We scanned the amino acid sequence of the individual

mature spp24 proteins from all nine species using NPS@wNetwork Protein Sequence Analysis(Combet et al.,2000)x to identify protein motifs defined in the PROS-ITE database of protein families and domains(Falquetet al., 2002) and identified several phosphorylationsignals previously identified using NetPhos. However,no consistent matches with known protein motifs werefound in the nine protein sequences.

2.5. Analysis of the SPP2 promoter

None of the promoter-identification components ofNIX (GRAILypolIIprom, TSSWyPromoter, GEN-

SCANyProm and FgenesyProm) were able to identify aconventional promoter containing either a CAAT orTATA box in the vicinity of the start of transcriptionthat we had mapped by primer extension. This too wasthe case with independent analyses using Promoter-Inspector (http:yywww.genomatix.de) and FirstEF(http:yyrulai.cshl.orgytoolsyFirstEFy) and the GRAILyCpG component in the NIX analysis did not predict anyCpG islands.A cross-species genome comparison strategy(Jegga

et al., 2002) was adopted to identify potential TF-binding sites common to the human and mouse genepromoters using the TraFaC program(TranscriptionFactor Comparison; http:yytrafac.chmcc.org). Such astrategy is less likely to lead to spurious identificationswhich are common where single genes are consideredin isolation. The BLASTZ comparison performed byTraFaC of the humanSPP2 gene sequence with thecorresponding region of the mouseSPP2 gene(acces-sion number BN000084) established that there is exten-sive similarity between the two promoters for a distanceof approximately 300 bp upstream of the human tran-scription start point but further 59 there is little observ-able conservation of sequence(data not shown). Withinthis 300-bp region of similarity, TraFaC identified threetypes of cis-element in common: nuclear factor ofactivated T-cells(NFAT) (Kel et al., 1999), cut-likehomeodomain protein(CLOX) (Harada et al., 1995),and CCAAT enhancer binding protein(CyEBP)(Grange et al., 1991) (Fig. 7a). NFAT and CLOXelements were also identified in the chicken gene pro-moter using MatInspector(http:yywww.genomatix.ge),as was a TATA boxjust upstream of the transcriptionstart site. Regions of sequence similarity in the vicinityof the identified TF-binding sites in the promoters ofthe three genes are shown in Fig. 7b. A segment of 67bp in the chicken promoter(y131 to y65) has 46identities (69%) with the human promoter(y81 toy13) which, in turn, shares 80% identity with thecorresponding region of the promoter of the mouseSPP2 gene(y93 toy25).

3. Discussion

In this study we have characterized the human geneSPP2 which encodes secreted phosphoprotein 24 andhave shown that the gene structure identifies it as anovel member of the cystatin superfamily. In addition,we have derived cDNAs for mouse, pig, sheep, salmon,trout and chicken allowing the identification of con-served regions of the protein which may be functionallyimportant. The discovery of a potentially negatively-charged region adjacent to the first cysteine, along withthe previously identified region of phosphorylated ser-ines, suggests that the role of spp24 in the fetuin-mineralcomplex might be similar to that of fetuin-A, acting


Fig. 7. Alignment of the promoter regions of the human, chicken and mouse genes.(a) Schematic representation of the promoters which arealigned relative to the 39 ends of exon 1 of each gene. The open boxes represent TF-binding sites. The 59 UTRs of the exons are depicted in grayand the amino acid-coding regions in black. The bent arrows indicate the mapped human and chicken transcription start points in liver.(b) Alignment of the regions of close sequence similarity in the proximal region of the three promoters. The coordinates are relative to the mappedtranscription start points in liver, except for mouse(for which the transcription start point has not been formally determined) where it is relativeto the start of the longest cDNA that we have identified at present.

through dense arrays of acidic residues(Heiss et al.,2003). It is also possible that the highly-charged C-terminal non-cystatin domain also plays a role in thisprocess.Analysis of the promoters of the human, mouse and

chicken genes reveals the presence of TF-binding sitesfor NFAT and CLOX in all three and CyEBP sites inhuman and mouse. NFAT was originally identified inthe context of the regulation of early T-cell activationgenes(Shaw et al., 1988). Hence, the presence of theNFAT sites in the promoters is consistent with spp24sequences having been found in cDNA libraries con-structed from chicken T-cell-enriched activated spleno-cytes (Tirunagaru et al., 2000) (http:yywww.chickest.udel.edu) and with the detection of human spp24 mRNAin peripheral white blood cells by RT-PCR(KhorramKhorshid and Dalgleish, unpublished). NFAT proteinsare now also recognized as transducers of Ca signals2q

(Graef et al., 2001) and it may be significant in thiscontext that spp24 expression is increased 3.5-fold inhypocalcemic rat kidney relative to controls(MatthewBeckman, personal communication). The cut-like hom-eodomain(CLOX) proteins are known to be involvedin the control of genes at different stages of developmentand the presence of CLOX sites is consistent with themouse spp24 gene being expressed in 10-day neonatecerebellum but not in adult(Miki et al., 2001) (http:yyread.gsc.riken.go.jpy) and also with its expression inimmature mouse bone at 12 days but not in adult boneat 16 weeks(Khorram Khorshid and Dalgleish, unpub-lished). CCAAT enhancer binding protein(CyEBP)(Grange et al., 1991) is a liver-enriched transcriptionfactor and the presence of CyEBP sites in the humanand mouse promoters is explicable. The absence of sucha site in the chicken promoter suggests that the control

of expression of the spp24 gene may be modulated bydifferent transcription factors in birds. In this respect, itwould be interesting to analyze other avian spp24 genepromoters.The chicken spp24 gene was originally designated

GHRG-1 (Agarwal et al., 1995) and was first isolated,using differential display, as a cDNA present at higherlevels in normal chickens relative to growth hormone(GH) receptor-deficient dwarf chickens. The gene’sresponse to GH has subsequently been confirmed invivo (Radecki et al., 1997) and it is expressed widelyin the developing embryo in response to GH(Harveyet al., 2001). In the initial characterization of the gene,the authors claimed to have identified a nucleotidesequence similar to the rat Spi 2.1 GH response element(GHRE) (Sliva et al., 1994). However, the match ispoor and our revised sequence of the chicken promoterplaces the putative GHRE within the first exon, ratherthan upstream of the transcription start point. Morerecently, a high affinity GHRE(haGHRE) (Bergad etal., 1999) based on the rat liver Spi 2.1 GH responseelement has been defined. A search of the revisedchicken promoter sequence for this element yields nomatches, even when several mismatched bases wereallowed, suggesting that the true GHRE in the chickengene lies outside of the region that has been sequencedso far.Apart from the demonstration of the widespread

expression of chicken spp24 during chicken embryonicdevelopment(Harvey et al., 2001) and the up-regulationof the gene during the differentiation of rat pancreaticAR42J cells into insulin-secreting cells(Mashima et al.,1999), information about the function of spp24 is limitedat present. A better appreciation of the function will bepossible once the proteins with which spp24 interacts


have been identified using techniques such as the yeasttwo-hybrid system. This would afford the opportunityto test the effects of specific alterations in the sequenceof spp24 on such interactions and might provide infor-mation on whether or not the serine to phenylalaninesubstitution identified in this study is disease-related.Other possible approaches include the construction of amouse in which theSPP2 gene is disrupted to see ifthere is a demonstrable phenotype, and the screening ofchemically-mutagenized mice or embryonic stem cells(Brown and Hardisty, 2003) for the presence of delete-rious mutations inSPP2 might be informative.

4. Experimental procedures

4.1. Preparation of total RNA and mRNA

Human liver was obtained from the UK Human TissueBank (De Montfort University, Leicester, UK). Cattleand sheep liver were obtained via Biomedical Services,University of Leicester. Tissues were homogenized inTRIzol Reagent(Invitrogen, Paisley, UK) and total RNAwas extracted according to the manufacturer’s instruc-tions. Contaminating glycogen was solubilized in 4 MLiCl and centrifugation was used to recover the insolubleRNA which was washed in 70%(vyv) ethanol anddissolved in DEPC-treated water. Messenger RNA wasprepared from total RNA using the MessageMakermRNA Isolation System(Life Technologies, Paisley,UK) or the Oligotex mRNA Purification System(Qia-gen, Crawley, UK) each according to the manufacturer’sinstructions.

4.2. Preparation and screening of the cDNA library

A human cDNA library was constructed from livermRNA using the SuperScript Plasmid System for cDNASynthesis and Plasmid Cloning(Invitrogen, Paisley, UK)according to the manufacturer’s instructions. The cDNAswere ligated into the plasmid vector pSPORT 1 andelectroporated into ElectroMaxE. coli DH10B cellsusing a Cell-Porator Electroporation System(Invitrogen,Paisley, UK) according to the manufacturer’s instruc-tions. This yielded approximately 2.5=10 transfor-6

mants of which approximately 10 000 were grown on atotal of four 82 mm diameter Hybond-N(AmershamPharmacia Biotech, Little Chalfont, UK) nylon filtersplaced on Luria agar plates with ampicillin selection.Duplicate filters were made and, on one set, colonieswere lysed in situ and the DNA fixed to the membraneswith UV light according to standard protocols. Probepreparation and hybridization were carried out as pre-viously described(Dalgleish, 1987).

4.3. Isolation of genomic DNA clones

A human male genomic DNA PAC-vector library(RPCI-1) (Osoegawa et al., 1999), containing approxi-mately 120 000 clones, was obtained from the UK MRCGeneserivce(Babraham, UK) in the form of high-density gridded colony filters which were screened byhybridization using a human spp24 cDNA(IMAGEclone 204242), also obtained from the UK MRC Gene-service, as a probe. Segments from the inserts of PACclones isolated from the library were sub-cloned, usingstandard procedures, into plasmid vectors.

4.4. Hybridization to the human RNA array

The human multiple tissue expression(MTE) array(Clontech, BD Biosciences, Oxford, UK) was hybrid-ized with the same probe as used for the screening ofthe genomic DNA library using the ExpressHyb hybrid-ization solution supplied with the array.

4.5. Mapping the start of transcription by primerextension

The procedure was carried out essentially as describedby Sambrook et al.(1989) using 10mg of total RNAand a gene-specific primer(GAGAGTGTCTCTCT-ATGTG) end labeled withwg- PxATP. The products of33

reverse transcription were electrophoresed on a 6%denaturing polyacrylamide gel alongside conventionalsequencing reactions of the sub-cloned correspondingregion of genomic DNA using the same primer. Theresults were visualized by autoradiography.

4.6. Conformation sensitive gel electrophoresis (CSGE)

Human genomic DNA was isolated from peripheralblood using standard procedures. Using 20 ng of tem-plate DNA, PCR amplifications were carried out asdescribed by Jeffreys et al.(1990) using primersdesigned to amplify either one or two exons and theirimmediate flanking DNA sequences. The forward andreverse primer sequences, the length of the PCR productproduced, the optimum Mg concentration and primer2q

annealing temperature were: Exons 1q2: CAGAAA-TATTGACCCCAGGA, GACAGCATTGGAAGGAG-GAG, 511 bp, 4.5 mM, 61 8C; Exon 2:CTGCTCTGGATCATGCAGAG, GACAGCATTGGAAGGAGGAG, 245 bp, 4.5 mM, 638C; Exon 3:GCTTTCATGGTGGACAATTC, CATTTCTGGGATGGGTCTC, 268 bp, 3.5 mM, 608C; Exon 4:CAATGGAGGCTATCCCTTTCC, CCTAAGAGGTGGGGTCTGG, 217 bp, 4.0 mM, 598C; Exon 5:TTTCATGTGCTGACACATCC, AAATGACTCAC-TAACAAAGAGTTGC, 173 bp, 4.0 mM, 598C; Exon6: AACATTCTGGAACAGTGAGAGG, TGATCA-


GAAAAGGGTCTGGTG, 153 bp, 4.0 mM, 598C; Exon7: AGAGCCTATGCTTCCCTTTTC, CAGCAGTTT-TAAGGCGTTCAC, 199 bp, 4.0 mM, 598C. Thereactions were cycled for 30 cycles each comprisingdenaturation at 968C for 30 s, primer annealing at theoptimum temperature for 30 s, and extension at 728Cfor 30 s. CSGE was performed as described by Korkko¨ ¨et al. (1998) with modifications. Piperazine diacrylam-ide (PDA) was used as the cross linker in 15 and 20%polyacrylamide gels and samples were loaded in 7%(wyv) sucrose, 0.04%(wyv) bromophenol blue, 0.04%(wyv) xylene cyanol FF.

4.7. DNA sequencing

DNA sequencing reactions were carried out usingABI PRISM BigDye 2.0 sequencing reagents(AppliedBiosystems, Warrington, UK) according to the manufac-turer’s recommendations. Electrophoresis of thesequencing reactions was carried out by the Protein andNucleic Acid Chemistry Laboratory(PNACL), Univer-sity of Leicester using an ABI PRISM 377 DNASequencer.

4.8. RT-PCR amplification of bovine and ovine mRNA

Messenger RNA was copied to cDNA using an oligo-dT primer and SuperScript II reverse transcriptase(Invi-trogen, Paisley, UK) according to the manufacturer’srecommended conditions. PCR amplification of thecDNA was carried out using forward(TCTGA-ACGGAAATTGTTCTTCC) and reverse (TGGA-ACTTCTATTCCTCCCAGTG) primers designed usingthe published bovine cDNA sequence(Agarwal et al.,1995). The reactions were cycled for 30 cycles eachcomprising 968C 30 s, 598C 30 s, and 728C 40 susing the same reaction buffer as described for thegenomic DNA amplifications.

4.9. Computer analysis of DNA and protein sequences

All DNA and protein sequences, and their alignmentswere edited using the SeqLab editor of the GCGWisconsin Package, version 10.3(Accelrys, San Diego,CA, USA) and GCG pileup was used to create sequencealignments. Searching of DNA sequences for specificsequence motifs was carried out using GCG findpatterns.Analyses of genomic DNA sequence organization werecarried out using the NIX DNA sequence analysisenvironment(Williams et al., 1998) at the UK HumanGenome Mapping Project Resource Centre, Hinxton,UK (http:yywww.hgmp.mrc.ac.uk). Comparison of thehuman and mouse promoter sequences and their analysisfor TF-binding sites was performed using TraFac(Jeggaet al., 2002) (http:yytrafac.cchmc.org) and MatInspector(Quandt et al., 1995) (http:yywww.genomatix.de). All

PCR primers were designed using Primer3(Rozen andSkaletsky, 2000) (http:yywww-genome.wi.mit.eduycgi-binyprimeryprimer3_www.cgi). The phosphorylationpotential of protein sequences was analyzed usingNetPhos(Blom et al., 1999) (http:yywww.cbs.dtu.dkyservicesyNetPhosy). Searches of protein sequences forpatterns in the PROSITE database were performedusing NPS@ (Combet et al., 2000) (http:yynpsa-pbil.ibcp.fry). Initiation codons were identified in m-RNAs using ATGpr (Salamov et al., 1998) (http:yywww.hri.co.jpyatgpry) and the 59 and 39 UTRs weresearched for functional elements using UTRscan(Pesoleand Liuni, 1999) (http:yybighost.area.ba.cnr.ityBIGyUTRScany).

Acknowledgments

We are grateful to the UK MRC Geneservice forproviding the I.M.A.G.E. cDNA clone and the filters toscreen the human genomic DNA PAC library. We aregrateful to Tim Smith (US Meat Animal ResearchCenter, Clay Center, NE, USA) for providing the pigcDNA clone and sequencing traces, to Richard Powell(Department of Microbiology, NUI, Galway, Eire) forsalmon cDNAs and sequencing traces, to Susan Douglas(Institute for Marine Biosciences, Halifax, Nova Scotia,Canada) for salmon cDNA sequencing traces, and toJoan Burnside(Delaware Biotechnology Institute, New-ark, DE, USA) for the chicken cDNA sequencing traces.We thank Matthew Beckman for sharing results prior topublication, Larry Mertz for advice on the constructionof the cDNA library, Anil Jegga for help and advicewith the TraFaC promoter analysis, Tim Francis forcarrying out some of the DNA sequencing, and RichardTrembath for providing human genomic DNA samples.CSB was supported by a studentship from the MedicalResearch Council, UK and HRKK was supported by ascholarship from the Ministry of Health and MedicalEducation, Iran.

References

Agarwal, S.K., Cogburn, L.A., Burnside, J., 1995. Comparison ofgene expression in normal and growth hormone receptor-deficientdwarf chickens reveals a novel growth hormone regulated gene.Biochem. Biophys. Res. Commun. 206, 153–160.

Antonarakis, S.E. and the Nomenclature Working Group, 1998.Recommendations for a nomenclature system for human genemutations. Hum. Mutat. 11, 1–3.

Bergad, P.L., Towle, H.C., Berry, S.A., 1999. Definition of a highaffinity growth hormone DNA response element. Mol. Cell. Endo-crinol. 150, 151–159.

Betts, M.J., Guigo, R., Agarwal, P., Russell, R.B., 2001. Exon´structure conservation despite low sequence similarity: a relic ofdramatic events in evolution? EMBO J. 20, 5354–5360.

Binkert, C., Demetriou, M., Sukhu, B., Szweras, M., Tenenbaum,H.C., Dennis, J.W., 1999. Regulation of osteogenesis by fetuin. J.Biol. Chem. 274, 28 514–28 520.


Blom, N., Gammeltoft, S., Brunak, S., 1999. Sequence and structure-based prediction of eukaryotic protein phosphorylation sites. J.Mol. Biol. 294, 1351–1362.

Bobek, L.A., Levine, M.J., 1992. Cystatins—inhibitors of cysteineproteinases. Crit. Rev. Oral Biol. Med. 3, 307–332.

Brown, S.D.M., Hardisty, R.E., 2003. Mutagenesis strategies foridentifying novel loci associated with disease phenotypes. Semin.Cell Dev. Biol. 14, 19–24.

Brown, W.M., Dziegielewska, K.M., 1997. Friends and relations ofthe cystatin superfamily—new members and their evolution. Pro-tein Sci. 6, 5–12.

Combet, C., Blanchet, C., Geourjon, C., Deleage, G., 2000. NPS@:´network protein sequence analysis. Trends Biochem. Sci. 25,147–150.

Dalgleish, R., 1987. Southern blotting. In: Boulnois, G.J.(Ed.), GeneCloning and Analysis: a Laboratory Guide. Blackwell ScientificPublications, Oxford, pp. 45–60.

Davey, G.C., Caplice, N.C., Martin, S.A., Powell, R., 2001. A surveyof genes in the Atlantic salmon(Salmo salar) as identified byexpressed sequence tags. Gene 263, 121–130.

den Dunnen, J.T., Antonarakis, S.E., 2000. Mutation nomenclatureextensions and suggestions to describe complex mutations: adiscussion. Hum. Mutat. 15, 7–12.

Falquet, L., Pagni, M., Bucher, P., Hulo, N., Sigrist, C.J.A., Hofmann,K., et al., 2002. The PROSITE database, its status in 2002. Nucl.Acids Res. 30, 235–238.

Graef, I.A., Chen, F., Crabtree, G.R., 2001. NFAT signaling invertebrate development. Curr. Opin. Genet. Dev. 11, 505–512.

Grange, T., Roux, J., Rigaud, G., Pictet, R., 1991. Cell-type specificactivity of two glucocorticoid responsive units of rat tyrosineaminotransferase gene is associated with multiple binding sites forCyEBP and a novel liver-specific nuclear factor. Nucl. Acids Res.19, 131–139.

Harada, R., Berube, G., Tamplin, O.J., Denis-Larose, C., Nepveu, A.,´ ´1995. DNA-binding specificity of the Cut repeats from humanCut-like protein. Mol. Cell. Biol. 15, 129–140.

Harvey, S., Lavelin, I., Pines, M., 2001. Growth hormone(GH)action in early embryogenesis: expression of a GH-response genein sites of GH production and action. Anat. Embryol. 204, 503–510.

Heiss, A., DuChesne, A., Denecke, B., Grotzinger, J., Yamamoto, K.,¨Renne, T., et al., 2003. Structural basis of calcification inhibition´by a -HS glycoproteinyfetuin-A. Formation of colloidal calcipro-2

tein particles. J. Biol. Chem. 278, 13 333–13 341.Henskens, Y.M.C., Veerman, E.C.I., Nieuw Amerongen, A.V., 1996.Cystatins in health and disease. Biol. Chem. Hoppe–Seyler 377,71–86.

Hu, B., Coulson, L., Moyer, B., Price, P.A., 1995. Isolation andmolecular cloning of a novel bone phosphoprotein related insequence to the cystatin family of thiol protease inhibitors. J. Biol.Chem. 270, 431–436.

Hubbard, T., Barker, D., Birney, E., Cameron, G., Chen, Y., Clark,Y., et al., 2002. The Ensembl genome database project. Nucl. AcidsRes. 30, 38–41.

Jeffreys, A.J., Neumann, R., Wilson, V., 1990. Repeat unit sequencevariation in minisatellites: a novel source of DNA polymorphismfor studying variation and mutation by single molecule analysis.Cell 60, 473–485.

Jegga, A.G., Sherwood, S.P., Carman, J.W., Pinski, A.T., Phillips,J.L., Pestian, J.P., et al., 2002. Detection and visualization ofcompositionally similarcis-regulatory element clusters in ortholo-gous and coordinately controlled genes. Genome Res. 12,1408–1417.

Kel, A., Kel-Margoulis, O., Babenko, V., Wingender, E., 1999.Recognition of NFATpyAP-1 composite elements within genesinduced upon the activation of immune cells. J. Mol. Biol. 288,353–376.

Korkko, J., Annunen, S., Pihlajamaa, T., Prockop, D.J., Ala-Kokko,¨ ¨L., 1998. Conformation sensitive gel electrophoresis for simple andaccurate detection of mutations: comparison with denaturing gra-dient gel electrophoresis and nucleotide sequencing. Proc. Natl.Acad. Sci. USA 95, 1681–1685.

Lennon, G., Auffray, C., Polymeropoulos, M., Soares, M.B., 1996.The I.M.A.G.E. consortium: an integrated molecular analysis ofgenomes and their expression. Genomics 33, 151–152.

Lo, K., Smale, S.T., 1996. Generality of a functional initiatorconsensus sequence. Gene 182, 13–22.

Mashima, H., Yamada, S., Tajima, T., Seno, M., Yamada, H., Takeda,J., et al., 1999. Genes expressed during the differentiation ofpancreatic AR42J cells into insulin-secreting cells. Diabetes 48,304–309.

Miki, R., Kadota, K., Bono, H., Mizuno, Y., Tomaru, Y., Carninci,P., et al., 2001. Delineating developmental and metabolic pathwaysin vivo by expression profiling using the RIKEN set of 18 816full-length enriched mouse cDNA arrays. Proc. Natl. Acad. Sci.USA 98, 2199–2204.

Nagy, E., Maquat, L.E., 1998. A rule for termination-codon positionwithin intron-containing genes: when nonsense affects RNA abun-dance. Trends Biochem. Sci. 23, 198–199.

Ohnishi, T., Nakamura, O., Ozawa, M., Arakaki, N., Muramatsu, T.,Daikuhara, Y., 1993. Molecular cloning and sequence analysis ofcDNA for a 59 kD sialoprotein of the rat: demonstration that it isa counterpart of humana -HS glycoprotein and bovine fetuin. J.2

Bone Miner. Res. 8, 367–377.Osoegawa, K., de Jong, P.J., Frengen, E., Ioannou, P.A., 1999.Construction of bacterial artificial chromosome(BACyPAC) librar-ies. In: Dracopoli, N.C., Haines, J.L., Krof, B.R., Morton, C.C.,Seidman, C.E., Seidman, J.G., Smith, D.R.(Eds.), Current Proto-cols in Human Genetics. John Wiley and Sons, New York, pp.5.15.11–15.15.33.

Pesole, G., Liuni, S., 1999. Internet resources for the functionalanalysis of 59 and 39 untranslated regions of eukaryotic mRNAs.Trends Genet. 15, 378.

Price, P.A., Lim, J.E., 2003. The inhibition of calcium phosphateprecipitation by fetuin is accompanied by the formation of afetuin–mineral complex. J. Biol. Chem. 278, 22 144–22 152.

Price, P.A., Nguyen, T.M.T., Williamson, M.K., 2003. Biochemicalcharacterization of the serum fetuin–mineral complex. J. Biol.Chem. 278, 22 153–22 160.

Quandt, K., Frech, K., Karas, H., Wingender, E., Werner, T., 1995.MatInd and MatInspector: new fast and versatile tools for detectionof consensus matches in nucleotide sequence data. Nucl. AcidsRes. 23, 4878–4884.

Radecki, S.V., McCann-Levorse, L., Agarwal, S.K., Burnside, J.,Proudman, J.A., Scanes, C.G., 1997. Chronic administration ofgrowth hormone(GH) to adult chickens exerts marked effects oncirculating concentrations of insulin-like growth factor-1(IGF-1),IGF binding proteins, hepatic GH regulated gene 1, and hepaticGH receptor mRNA. Endocrine 6, 117–124.

Rozen, S., Skaletsky, H.J., 2000. Primer3 on the WWW for generalusers and for biologist programmers. In: Krawetz, S., Misener, S.(Eds.), Bioinformatics Methods and Protocols: Methods in Molec-ular Biology. Humana Press, Totowa, NJ, pp. 365–386.

Salamov, A.A., Nishikawa, T., Swindells, M.B., 1998. Assessingprotein coding region integrity in cDNA sequencing projects.Bioinformatics 14, 384–390.

Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: aLaboratory Manual. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, NY.

Schafer, C., Heiss, A., Schwarz, A., Westenfeld, R., Ketteler, M.,¨Floege, J., et al., 2003. The serum proteina -Heremans–Schmid2

glycoproteinyfetuin-A is a systemically acting inhibitor of ectopiccalcification. J. Clin. Invest. 112, 357–366.


Schinke, T., Amendt, C., Trindl, A., Poschke, O., Muller-Esterl, W.,¨ ¨

Jahnen-Dechent, W., 1996. The serum proteina -HS glycoproteiny2

fetuin inhibits apatite formation in vitro and in mineralizing calvariacells. J. Biol. Chem. 271, 20 789–20 796.

Shaw, J.-P., Utz, P.J., Durand, D.B., Toole, J.J., Emmel, E.A., Crabtree,G.R., 1988. Identification of a putative regulator of early T cellactivation genes. Science 241, 202–205.

Sliva, D., Wood, T.J.J., Schindler, C., Lobie, P.E., Norstedt, G., 1994.Growth hormone specifically regulates serine protease inhibitorgene transcription via g-activated sequence-like DNA elements. J.Biol. Chem. 269, 26 208–26 214.

Swallow, J.E., Merrison, W.K., Gill, P.K., Harris, S., Dalgleish, R.,1997. Assignment of secreted phosphoprotein 24(SPP2) to humanchromosome band 2q37™qter by in situ hybridization. Cytogenet.Cell. Genet. 79, 142.

Tirunagaru, V.G., Sofer, L., Cui, J., Burnside, J., 2000. An expressedsequence tag database of T-cell enriched activated chicken spleno-cytes: sequence analysis of 5251 clones. Genomics 66, 144–151.

Turk, V., Bode, W., 1991. The cystatins: protein inhibitors of cysteineproteinases. FEBS Lett. 285, 213–219.

Williams, G.W., Woollard, P.M., Hingamp, P., 1998. NIX: a nucleotideidentification system at the HGMP-RC. URL http:yywww.hgmp.mrc.ac.ukyNIX y.

Documents

Characterization of the human secreted phosphoprotein 24 gene (SPP2) and comparison of the protein sequence in nine species