Characterization and expression of murine PRELP

Matrix Biology 20 Ž2001. 555᎐564

Characterization and expression of murine PRELP 夽 Judy Grover a,b, Peter J. Roughley U,a,b a

Genetics Unit, Shriners Hospital for Children, 1529 Cedar A¨ enue, Montreal, Quebec, Canada H3G 1A6 b Department of Surgery, McGill Uni¨ ersity, Montreal, Quebec, Canada Received 27 March 2001; received in revised form 16 August 2001; accepted 16 August 2001

Abstract The cDNA sequence of the murine prolinerarginine-rich end leucine-rich repeat protein ŽPRELP. gene was cloned by PCR-based techniques. The gene encodes a protein of 378 amino acids, which is four amino acid residues shorter than its human counterpart. This difference resides mainly in the amino terminal region of the mature protein, which is five amino acids shorter in the mouse than the human and has a lower arginine content. The remainder of the protein, including the structure of the leucine-rich repeats, the potential sites for N-linked glycosylation, and the disulfide-bonded domains are well conserved between species. In common with humans, the murine gene possesses three exons, with the translation initiation codon residing in exon 2 and the termination codon in exon 3. Exons 1 and 2 are separated by an intron of approximately 6.7 kbp, whereas exons 2 and 3 are separated by an intron of approximately 1.7 kbp. Western blot analysis of mouse cartilage extracts indicates that PRELP exists as a glycoprotein of approximately 55 kDa, as in human cartilage. Immunohistochemical and in situ hybridization analysis reveal that PRELP is expressed in cartilage throughout both fetal development and post-natal life, in contrast to the human where expression in cartilage is not apparent prior to birth. Northern blot analysis indicates that PRELP mRNA is also expressed in the developing embryo prior to skeletogenesis. The promoter region of the mouse PRELP gene possesses no TATA box in its proximal region, in common with humans, and shows differences in the conservation of elements known to be involved in regulating expression of the human PRELP gene. 䊚 2001 Elsevier Science B.V.rInternational Society of Matrix Biology. All rights reserved. Keywords: PRELP; Gene; Expression; Cartilage; Mouse

1. Introduction Prolinerarginine-rich end leucine-rich repeat protein ŽPRELP. belongs to the family of leucine-rich

Abbre¨ iations: DIG, digoxygenin; PAGE polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; PRELP, prolinerarginine-rich end leucine-rich repeat protein; RT, reverse transcription; SDS, sodium dodecyl sulfate; SLRP, small leucine-rich proteoglycan; SSC, saline sodium citrate 夽 The DNA sequence data reported in this manuscript have been deposited in the GenBank database under the accession numbers AF 261886, AF 261887 and AF 261888. U Corresponding author. Genetics Unit, Shriners Hospital for Children, 1529 Cedar Avenue, Montreal, Quebec, Canada H3G 1A6. Tel.: q1-514-842-5964; fax: q1-514-842-5581. E-mail address: [email protected] ŽP.J. Roughley..

repeat proteins ŽKobe and Deisenhofer, 1994., which are characterized by a series of adjacent leucine-rich motifs ŽLXXLXLXXNXL. flanked by disulfidebonded domains. The presence of 10 central leucinerich repeats places PRELP in the same sub-family as the small leucine-rich proteoglycans ŽSLRPs. ᎏ decorin, biglycan, fibromodulin, lumican and keratocan ŽIozzo and Murdoch, 1996; Hocking et al., 1998.. The major structural difference between sub-family members resides at the amino terminal region of their core proteins, where decorin and biglycan are substituted with chondroitin sulfate or dermatan sulfate chains, but PRELP, fibromodulin, lumican and keratocan are not. It is in this region that PRELP is rich in proline and arginine residues ŽBengtsson et al., 1995.. While fibromodulin, lumican and keratocan may be substituted with N-linked keratan sulfate in

0945-053Xr01r$ - see front matter 䊚 2001 Elsevier Science B.V.rInternational Society of Matrix Biology. All rights reserved. PII: S 0 9 4 5 - 0 5 3 X Ž 0 1 . 0 0 1 6 5 - 2

556

J. Gro¨ er, P.J. Roughley r Matrix Biology 20 (2001) 555᎐564

their central regions, this does not appear to be the case with PRELP, which exits as a glycoprotein. Indeed, it was first identified as a 58-kDa protein present in bovine cartilage ŽHeinegard ˚ et al., 1986.. At least in the human, one of the major features of PRELP is its relatively restricted localization to cartilagenous tissues, where its expression is abundant in the adult but not at or prior to birth ŽMelching and Roughley, 1990; Grover et al., 1996.. At present the in vivo function of PRELP has not been elucidated, and hence it is unclear why it should exhibit a restricted tissue distribution, or why it should not be required in cartilage prior to birth. However, it has been demonstrated that the amino terminal region of PRELP is able to interact with heparan sulfate ŽBengtsson et al., 2000., suggesting a possible link between the extracellular matrix and the plasma membrane of the chondrocytes, via PRELPrheparan sulfate᎐proteoglycan interactions. Surprisingly, this region of PRELP shows the lowest level of amino acid conservation between the human, bovine and rat ŽBengtsson et al., 2000.. In the present work the structure and expression pattern of the mouse PRELP gene were determined to further study species variation in these parameters and to evaluate the suitability of transgenic or knockout mice as models for the study of PRELP function in humans.

2. Materials and methods 2.1. Source of tissue Mouse tissues were obtained from C57B mice aged e7.5, e11.5, e14.5, e17.5, 10 days, 1 month, 3 and 6 months. Human cartilage was collected from the distal femur at the time of autopsy and within 20 h of death. The specimens were from individuals aged 17 weeks of gestation Žfetus., 17, 27, 64 and 70 years. 2.2. Polymerase chain reaction (PCR) amplification Genomic DNA was isolated from mouse tails using the QIAamp Tissue Kit ŽQIAgen. and standard procedures were used for PCR amplification ŽSaiki, 1990.. First-strand cDNA was synthesized from adult mouse lung total RNA ŽClontech. and RT-PCR amplification was performed using standard procedures ŽKawasaki, 1990.. Three sets of primers Ž5⬘-CTTCTAGCTGGCTCTCTGCT and 5⬘-CGGGAGCTCAGTGATGAAGTT, 5⬘-CTGCGAAAGGTCCCTGTCAT and 5⬘-GGCCTAGATGACCACGGACT, and 5⬘-TCAACGGAACCCAGATTTGC and 5⬘TAGACACACTGTGTTTCTC. from the human PRELP sequence ŽGenBank Accession Nos. U41292,

U41343 and U41344. were used to amplify overlapping fragments spanning the entire PRELP coding sequence as well as part of the 5⬘- and 3⬘-untranslated sequences. All amplified fragments described above were cloned into the pCRII vector using the TA cloning kit ŽInvitrogen., and subjected to sequence analysis using the dideoxy chain termination method ŽSanger et al., 1977.. Intronrexon junctions were assigned by comparing cDNA and genomic sequences, and intron sizes were determined by comparison of cDNA and genomic DNA PCR products sized on 1% agarose gels. 2.3. Northern blotting analysis Human chondrocytes were isolated as previously described ŽGrover and Roughley, 1995., and total RNA was extracted by the acid guanidinium thiocyanaterphenolrchloroform method ŽChomczynski and Sacchi, 1987.. Similar amounts of mouse polyAq RNA preparations blotted onto nylon membranes were obtained from Clontech. A 426-bp human PRELP probe Žbp 926᎐1351, GenBank Accession Nos. U41343 and U41344. was labeled using the Multiprime DNA labeling system ŽAmersham., and hybridization was carried out using Rapid-hyb buffer ŽAmersham. for 2 h at 65 ⬚C. The blot was then washed with 2 = SSC for 10 min with several buffer changes at room temperature, followed by 2 = 15-min washes in 0.1= SSC at 65 ⬚C and subsequent exposure to X-ray film. Blots were also hybridized using a human actin probe to verify mRNA loading though it is known that the actin mRNA is not expressed equally in all tissues and cannot be used to normalize mRNA loading. 2.4. SDS r PAGE and immunoblotting Femoral heads were dissected from the hip joints of mice aged 3 and 6 months. This tissue and adult human femoral condylar cartilage were extracted with 4 M guanidinium chloride ŽRoughley and White, 1980.. Proteins in the tissue extracts were analyzed by SDSrPAGE, then electroblotted onto nitrocellulose and immunolocalized as described previously ŽRoughley et al., 1993.. An antiserum recognizing both mouse and human PRELP was raised in a rabbit to a combination of three synthetic peptides ŽH-CGGKNQLEEVPSALPRGG-OH, H-CGGSNKIETIPNGYFKGG-OH and H-CGGFRLLQSVVIGGOH., each conjugated to ovalbumin as described previously ŽRoughley et al., 1993.. The central peptide sequences of all three peptides are fully conserved between mouse and human PRELP.

J. Gro¨ er, P.J. Roughley r Matrix Biology 20 (2001) 555᎐564

2.5. Immunohistochemistry and in situ hybridization Tissue samples were fixed in 4% fresh paraformaldehyde in PBS, pH 7.2, at 4 ⬚C overnight. Specimens were embedded in paraffin ŽParaplast Xtra, Fisher. and sectioned at 6᎐8 ␮m. For immunohistochemistry, tissue sections were pretreated for 1 h with chondroitinase ABC Ž0.25 unitsrml; Sigma. at 37 ⬚C, then immunostaining was performed using the Vectastain ABC Elite kit ŽVector Laboratories., following the manufacturer’s instructions. The primary antibody was either the anti-peptide antiserum described in the previous section or preimmune serum. The preimmune serum showed no significant reactivity in all tissues that were analyzed. For in situ hybridizations, sense and antisense riboprobes encompassing bp 1018᎐1230 of the murine PRELP cDNA

557

sequence ŽFig. 1. were DIG-labeled using a ŽSP6rT7. DIG RNA labeling kit ŽBoehringer Mannheim.. Probes were hybridized at 42 ⬚C, washed in 0.1= SSC at 60⬚C and detected using a DIG nucleic acid detection kit ŽBoehringer Mannheim., following the manufacturer’s instructions. 2.6. Promoter analysis The mouse PRELP promoter was PCR-amplified using nucleotide sequence derived from an unidentified BAC clone ŽGenBank Accession No. AC026760. that was found to include PRELP cDNA sequence. Amplification used mouse genomic DNA and the primers 5⬘-CTAATGGTGACTGAGTAAGTGT from the GenBank sequence and 5⬘-TGGACTCTGGGGCACTCTT from exon 1 of the cDNA sequence

Fig. 1. Sequence of murine PRELP cDNA and protein. The sequence of 1499 bases of the mouse PRELP cDNA is depicted commencing with the major transcription start site. The translation initiation and termination codons are indicated Žunderline., together with the 378 residues of the deduced amino acid sequence, beneath the cDNA sequence.

558

J. Gro¨ er, P.J. Roughley r Matrix Biology 20 (2001) 555᎐564

ŽFig. 1.. The resulting product of 1020 bp was TAcloned and sequenced as above. The location of known transcription factor DNA-binding motifs was analyzed by matrix search software ŽSignal Scan 4.05. at the Institute for Transcriptional Informatics web-site Žwww.ifti.org.. The transcription start site of the mouse gene was determined by primer extension analysis ŽGrover et al., 1996., using 10 ␮g total RNA prepared from the cartilage of 3-day-old mice and the exon 1 primer described above. The product of this reaction was analyzed on a 6% polyacrylamide sequencing gel together with the sequence obtained from the genomic clone with the same primer.

3. Results 3.1. Mouse PRELP structure PCR-based techniques were used to establish the cDNA sequence of the murine PRELP gene ŽFig. 1.. The sequence contains an open reading frame of 1137 bp, commencing with an ATG methionine codon and ending with a TAG termination codon, and encodes a protein of 378 amino acids. The cDNA sequence also describes 164 bp of 5⬘-untranslated sequence and 198 bp of partial 3⬘-untranslated sequence. The mouse PRELP possesses a consensus site for signal peptide cleavage ŽVon Heijne, 1985, 1986. following amino acid residue 21, which would give rise to a mature protein of 357 residues. Upon analysis by SDSrPAGE and immunoblotting, the murine PRELP from adult cartilage behaved as a protein with a molecular size of 55 kDa ŽFig. 2., which is identical to that for the human PRELP. The size of this component is compatible with substitution by N-linked oligosaccharides, and its discrete nature suggests that they are not modified further by polylac-

Fig. 2. Western blotting of murine and human PRELP. Protein extracts from adult human Žlane 1. and mouse Žlane 2. cartilage were analyzed by SDSrPAGE and immunoblotting, using an antiserum raised against amino acid sequences conserved between the two species. Mouse cartilage was obtained from animals aged 3 months. The migration position of molecular weight standards is indicated.

tosamine elongation and sulfation to form keratan sulfate. 3.2. Mouse PRELP gene Comparison of cDNA and genomic DNA sequences revealed that the murine PRELP gene consists of three exons ŽFig. 3.. Exon 1 encodes only 5⬘-untranslated sequence and is separated from exon 2 by an intron of approximately 6.7 kbp. Exon 2 encodes the final 16 bp of the 5⬘-untranslated region, and the first 961 bp of the coding sequence. Exon 3 encodes the final 176 bp of the coding sequence and the 3⬘-untranslated region. Exons 2 and 3 are separated by an intron of approximately 1.7 kbp. This is a phase 1 intron splitting a lysine codon ŽAAA. between its first and second residues. Both introns 1 and 2

Fig. 3. Structure of the murine PRELP gene. The mouse PRELP gene is depicted as a sequence of exons Žopen boxes. and connecting introns Žlines.. The position of the translation initiation codon ŽATG. and termination codon ŽTAG. are indicated, as are the nucleotide sequences surrounding exonrintron junctions. Exon nucleotide sequence is shown in upper case while intron sequence is shown in lower case. Nucleotide numbering refers to the cDNA sequence depicted in Fig. 1. The size of exon 3 is calculated from the size of the largest mRNA isoform.

J. Gro¨ er, P.J. Roughley r Matrix Biology 20 (2001) 555᎐564

559

possess a gt sequence at their 5⬘-end and ag sequence at their 3⬘-end. These sequences represent the traditional donor and acceptor sites, respectively, involved in message splicing ŽBreathnach and Chambon, 1981.. Primer extension analysis revealed that the mouse PRELP gene may have several transcription start sites ŽFig. 4.. The major transcription start site results in exon 1 having a size of 148 bp, and the 5⬘-untranslated region of the mouse cDNA having 164 bp. Such variance in the transcription start site is not unexpected in a gene that possesses a TATA-less promoter ŽFig. 5.. The mouse PRELP promoter does possess a series of Sp1-binding sites residing upstream of the transcription start sites, which have been associated with the initiation of transcription in other genes ŽWeis and Reinberg, 1997., including the human PRELP gene ŽGrover and Roughley, 1998.. In addition, the mouse PRELP promoter possesses several Ets-binding sites, which have been associated with the regulation of transcription in the human PRELP promoter ŽGrover and Roughley, 1998. and other genes ŽMacleod et al., 1992.. 3.3. Mouse PRELP expression In the mouse embryo, PRELP mRNA expression is detected by day 7.5, which is prior to skeletogenesis, but then declines ŽFig. 6a.. Following the onset of skeletogenesis PRELP mRNA is detected in the developing fetus from day 14.5 onwards. At embryonic and fetal ages the mouse PRELP mRNA shows size heterogeneity, with two components of approximately 4.5 and 3.8 kb of similar abundance. This contrasts to adult tissues where one major mRNA size of 3.8 kb is detected, together with a minor component of approximately 1.7 kb ŽFig. 6b.. The largest mRNA size would result in exon 3 of the mouse PRELP gene having a size of approximately 3.3 kbp, assuming that no splice sites are present within the distal 3⬘-untranslated region ŽFig. 3.. This being the case, the mouse PRELP gene would span approximately 12.9 kbp of genomic DNA. Amongst the adult mouse tissues examined, PRELP mRNA expression was highest in the lung, presumably due to the presence of cartilagenous elements. Lower levels of PRELP mRNA expression were also observed in cardiac and skeletal muscle. This pattern of expression in adult tissues was similar to that previously observed in the human ŽGrover et al., 1996.. To further study the tissue expression of the mouse PRELP gene, sections from fetal and mature mice were examined by in situ hybridization. In the fetus ŽFig. 7a., cartilage was by far the most reactive tissue, with all cartilage of the skeletal system showing the presence of PRELP mRNA. In the adult, PRELP expression was again prominent in cartilage, and ap-

Fig. 4. Analysis of transcription start site of the murine PRELP gene. Products obtained by primer extension of mouse chondrocyte RNA Žlane 1. and sequence analysis of the non-coding strand of genomic DNA Žlanes GATC., using the same primer in exon 1, were analyzed by polyacrylamide gel electrophoresis. The position of the major transcription start site is indicated Žarrow..

peared much more pronounced in growth plates rather than at the articular surface of joints ŽFig. 7b.. The sites of mRNA expression were also the sites at which PRELP protein could be detected by immunohistochemistry, with immune staining in the fetus being prominent in all skeletal cartilage from skeletogenesis onwards ŽFig. 8a,b.. In the newborn, PRELP was again present throughout skeletal cartilage, being equally abundant in the unresorbed cartilage of the epiphyses and the growth plates of the long bones ŽFig. 8c.. This is in contrast to the mature mouse, where PRELP expression appeared greater in the inactive growth plate than in the articular cartilage at the joint surface ŽFig. 8d.. A different age-related pattern of PRELP expression is observed in the human, where fetal cartilage of the epiphyses or growth plates exhibits little immunostaining for PRELP ŽFig. 8e., whereas adult articular cartilage shows strong

560

J. Gro¨ er, P.J. Roughley r Matrix Biology 20 (2001) 555᎐564

Fig. 5. Nucleotide sequence of the murine and human PRELP promoters. The sequences of 853 and 861 bases, respectively, preceding the major transcription start site of the mouse and human PRELP genes are depicted. The position of Spl-binding sites Ždashed underline. and Ets-binding sites Žsolid underline. are indicated.

PRELP immunoreactivity particularly in the vicinity of the cells ŽFig. 8f.. Thus, while both human and mouse have similar PRELP genes, their temporal and site regulation of PRELP expression is distinct.

4. Discussion

Fig. 6. Northern blotting of murine PRELP mRNA. Mouse polyAq RNA preparations were analyzed by Northern blotting using radiolabeled probes derived from the human PRELP Župper panels. and actin Žlower panels. cDNA sequences. Ža. RNA from mouse embryos aged: 1, 7.5 days; 2, 11.5 days; 3, 14.5 days; 4, 17.5 days. ŽB. RNA from adult mouse tissues: 1, heart; 2, brain; 3, spleen; 4, lung; 5, liver; 6, skeletal muscle; 7, kidney; 8, testes. The migration of RNA standards is indicated.

The mouse PRELP shows 90% sequence identity in structure to its human counterpart, with the major differences residing in the 17 amino acid residues at the extreme amino terminus of the mature core protein ŽFig. 9.. This region shares only 45% identity with the equivalent region in the human, which also has five additional amino acids. Furthermore, several of the arginine residues that are abundant in this region in human PRELP are replaced by lysine in the mouse PRELP. The bovine and rat PRELP cDNA sequences have recently been published ŽBengtsson et al., 2000., with the bovine showing similar features to the human, and the rat showing the amino terminal trunca-

J. Gro¨ er, P.J. Roughley r Matrix Biology 20 (2001) 555᎐564

Fig. 7. In situ hybridization analysis of murine PRELP mRNA expression. Paraformaldehyde-fixed tissue sections from day 17.5 fetal Žpanel a. or 3-month post-natal Žpanel b. mouse were analyzed by in situ hybridization. The locations of the upper limb ŽUL., mandible ŽM. and ribs ŽR. are indicated in the fetal section, and the locations of the articular cartilage ŽAC. and growth plate ŽGP. are indicated in the post-natal section. Each panel is presented at the same magnification, and the bar on panel a indicates a distance of 100 ␮m.

tion and arginine replacement by lysine characteristic of the mouse ŽFig. 9.. The remainder of the protein is highly conserved between the four species, including the position of the six cysteine residues, the location of the 11 leucine-rich repeats, and the location of the four potential sites for substitution by N-linked oligosaccharides. The amino terminal region is unique to PRELP among the various leucine-rich repeat proteinsrproteoglycans in its sub-family, and this has led to speculation that this region of PRELP may be involved in a specific functional role ŽBengtsson et al., 1995.. At present, the in vivo function of PRELP remains unresolved, but if a common function exists, it is perhaps surprising that the amino terminal region of PRELP does not show greater species conservation. In vitro, it has been shown that the amino terminal region of PRELP is necessary for interaction with heparin and heparan sulfate, leading to speculation that matrix-associated PRELP may interact with chondrocyte-membrane-associated heparan sulfateproteoglycans to provide a link between the matrix

561

and its constituent cells ŽBengtsson et al., 2000.. However, while human, bovine and rat PRELP all showed the ability to interact, their rates of interaction with various heparinrheparan sulfate structures were different, suggesting that species differences in PRELP function could exist due to the substitution of lysine for arginine or truncation within the amino terminal region. This being the case, one might speculate that mouserrat and humanrbovine PRELP could have different functional properties in vivo, with respect to the fine structure of the heparan sulfate chains with which they interact. Whether this merely reflects species differences in heparan sulfate structure, or is of more functional significance, remains to be established. With respect to the PRELP gene, there is considerable homology between the mouse and human ŽGrover et al., 1996.. Both have three exons with the coding region being divided between exons 2 and 3, and in both cases there are 16 bp of -5⬘-untranslated sequence separating the translation initiation codon from the beginning of exon 2. Furthermore, in both genes the splice junction between exons 2 and 3 splits an AAA lysine codon following its first base, at an equivalent position in the protein sequence. Exon 2 is, however, 12 bp shorter in the mouse Ž977 bp. than the human Ž989 bp. due to the amino acid variation in the amino terminal regions of their proteins. In both cases the beginning of exon 3 includes the final 176 bp of coding sequence. The introns of the PRELP gene do show variation in size between the mouse and human, with intron 2 being smaller in the mouse. Intron 1 sizes are approximately 6.7 kbp in both species, whereas intron 2 sizes are approximately 1.7 kbp vs. 2.6 kbp in the mouse and human, respectively. This results in the mouse gene spanning approximately 0.9 kbp less genomic sequence than its human counterpart. The mouse mRNA shows heterogeneity in size with isoforms of 4.5, 3.8 and 1.7 kb being observed. The two larger isoforms are most abundant in the fetus, whereas that of intermediate size is most abundant in the adult. In contrast, the human PRELP mRNA shows three isoforms of 6.7, 4.6 and 1.7 kb, with that of 1.7 kb being most abundant ŽGrover et al., 1996., and the bovine PRELP mRNA has been reported to exist as a single component of 3.8 kb ŽBengtsson et al., 1995.. It is likely that the different mouse PRELP mRNA isoforms reflect the existence of three polyadenylation signals within exon 3 of the gene, but at present it is unclear whether variation in the use of these polyadenylation signals is of any functional significance. The greatest difference observed in this work between the mouse and human PRELP is related to their expression during fetal development. In the hu-

562

J. Gro¨ er, P.J. Roughley r Matrix Biology 20 (2001) 555᎐564

Fig. 8. Immunohistochemical analysis of murine and human PRELP expression. Paraformaldehyde-fixed tissue sections from day 14.5 fetal Žpanel a., day 17.5 fetal Žpanel b., day 10 post-natal Žpanel c. and 3-month post-natal Žpanel d. mouse, and 17-week fetal Žpanel e. and 17-year adult Žpanel f. human, were analyzed by immunohistochemistry. The locations of the upper limb ŽUL., mandible ŽM., ribs ŽR., heart ŽH., lung ŽLg., liver ŽL., intestines ŽI. and vertebral column ŽVC. are indicated in the fetal mouse sections; the locations of the epiphyseal cartilage ŽE., articular cartilage ŽAC. and growth plate ŽGP. are indicated in the post-natal mouse and fetal human sections; and the locations of the superficial ŽS. and deep ŽD. zones of the articular cartilage are indicated in the adult human sections. The bar on each panel indicates a distance of 100 ␮m.

man there is little evidence for expression of either protein or mRNA during this period ŽGrover et al., 1996., whereas in the mouse both mRNA and protein are expressed ŽFigs. 7 and 8.. Little work has been carried out on the fetal expression of PRELP in other species, though PRELP mRNA has been detected in

fetal bovine cartilage ŽBengtsson et al., 1995.. The origin of the different fetal expression patterns of human and mouse PRELP are likely to reside in their promoter regions via the control of transcription, though comparison of the mouse and human PRELP promoters provided little information on the reason

J. Gro¨ er, P.J. Roughley r Matrix Biology 20 (2001) 555᎐564

563

Fig. 9. Amino acid sequence of mammalian PRELPs. The amino acid sequence of mouse ŽM., human ŽH., rat ŽR. and bovine ŽB. PRELP are depicted. The site at which cleavage of the signal peptide is predicted to occur is indicated Žarrow., as are the locations of cysteine residues Žblack boxes., leucine-rich repeats Žopen boxes. and consensus sequences for the attachment of N-linked oligosaccharides Žshaded boxes..

for this difference. Both promoters lack a TATA box prior to their transcription start sites, and in the case of the human a proximal Sp1-binding site has been shown to be essential for basal transcription and a distal Ets-binding site has been shown to be necessary for repression of transcription ŽGrover and Roughley, 1998.. While this study contains no functional assessment of the mouse PRELP promoter, similar potential regulatory elements are present, though in the case of the Ets-binding sites they are not present at equivalent positions. At present, there is no information on whether any of the Ets-binding sites are functional in the mouse, and it is possible that species differences in gene expression may not be a consequence of variation in regulatory elements, but rather a result of differences in the expression of the corresponding transcription factors. The existence of species differences in the fetal expression of PRELP raises a number of questions. Firstly, is PRELP needed to ensure normal fetal cartilage development in the mouse, but not the human? Secondly, does the variation in amino terminal structure alter the functional properties of murine and human PRELP and does this influence their need

for fetal expression? Thirdly, are the post-natal requirements for PRELP the same in different species? Obviously, the present work cannot answer these issues, but they should be addressed in future work related to the study of PRELP function using transgenic or knock-out mice.

Acknowledgements We would like to thank the Shriners of North America for financial support, and the pathology departments of the Royal Victoria Hospital and the Montreal General Hospital for access to tissue. We would also like to thank Ms N. Nikolajew for typing the manuscript and Ms G. Bedard for preparation of ´ the figures. References Bengtsson, E., Aspberg, A., Heinegard, ˚ D., Sommarin, Y., Spillmann, D., 2000. The amino terminal part of PRELP binds to heparin and heparan sulfate. J. Biol. Chem. 275, 40695᎐40702. Bengtsson, E., Neame, P.J., Heinegard, ˚ D., Sommarin, Y., 1995. The primary structure of a basic leucine-rich repeat protein,

564

J. Gro¨ er, P.J. Roughley r Matrix Biology 20 (2001) 555᎐564

PRELP, found in connective tissues. J. Biol. Chem. 270, 25639᎐25644. Breathnach, R., Chambon, P., 1981. Organization and expression of eukaryotic split genes encoding for proteins. Annu. Rev. Biochem. 50, 349᎐383. Chomczynski, P., Sacchi, N., 1987. Single-step method of RNA isolation by acid guanidinium thiocyanate᎐phenol᎐chloroform extraction. Anal. Biochem. 162, 156᎐159. Grover, J., Roughley, P.J., 1995. Expression of cell surface proteoglycan mRNA by human articular chondrocytes. Biochem. J. 309, 963᎐968. Grover, J., Chen, X.-N., Korenberg, J.R., Recklies, A.D, Roughley, P.J., 1996. The gene organization, chromosomal location, and expression of a 55-kDa matrix protein ŽPRELP. of human articular cartilage. Genomics 38, 109᎐117. Grover, J., Roughley, P.J., 1998. Characterization of the human prolinerarginine-rich end leucine-rich repeat protein ŽPRELP. gene promoter and identification of a repressor element. Biochem. J. 336, 77᎐82. Heinegard, ˚ D., Larsson, T., Sommarin, Y., Franzen, ´ A., Paulson, M., Hedbom, E., 1986. Two novel matrix proteins isolated from articular cartilage show wide distribution among connective tissues. J. Biol. Chem. 261, 13866᎐13872. Hocking, A.M., Shinomura, T., McQuillan, D.J., 1998. Leucine-rich repeat glycoproteins of the extracellular matrix. Matrix Biol. 17, 1᎐19. Iozzo, R., Murdoch, A.D., 1996. Proteoglycans of the extracellular environment: clues from the gene and protein side offer novel perspectives in molecular diversity and function. FASEB J. 10, 598᎐614.

Kawasaki, E.S., 1990. Amplification of RNA. In: Innis, M.A. et al. ŽEd.., PCR Protocols. Academic Press, New York, pp. 21᎐27. Kobe, B., Deisenhofer, J., 1994. The leucine-rich repeat: a versatile binding motif. Trends Biochem. Sci. 19, 415᎐421. Macleod, K., Leprince, D., Stehelin, D., 1992. The ets gene family. Trends Biochem. Sci. 17, 251᎐256. Melching, L.I., Roughley, P.J., 1990. A matrix protein of M r 55,000 that accumulates in human articular cartilage with age. Biochim. Biophys. Acta 1036, 213᎐220. Roughley, P.J., White, R.J., 1980. Age-related changes in the structure of the proteoglycan subunits from human articular cartilage. J. Biol. Chem. 255, 217᎐224. Roughley, P.J., White, R.J., Magny, M.-C., Liu, J., Pearce, R.H., Mort, J.S., 1993. Non-proteoglycan forms of biglycan increase with age in human articular cartilage. Biochem. J. 295, 421᎐426. Saiki, R.K., 1990. Amplification of genomic DNA. In: Innis, M.A. et al. ŽEd.., PCR Protocols. Academic Press, New York, pp. 13᎐20. Sanger, F., Nicklen, S., Coulson, A.R., 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463᎐5467. Von Heijne, G., 1985. Signal sequences. The limits of variation. J. Mol. Biol. 184, 99᎐105. Von Heijne, G., 1986. A new method for predicting signal sequence cleavage sites. Nucl. Acids Res. 14, 4683᎐4690. Weis, L., Reinberg, D., 1997. Accurate positioning of RNA polymerase II on a natural TATA-less promoter is independent of TATA-binding protein-associated factors and initiator-binding proteins. Mol. Cell. Biol. 17, 2973᎐2984.