Portal de Periódicos da CAPES

Structural Characterization of TSC-36/Flik

2004; Elsevier BV; Volume: 279; Issue: 12 Linguagem: Inglês

10.1074/jbc.m309318200

1083-351X

Harald O. Hambrock, Brigitte Kaufmann, Stefan Müller, Franz‐Georg Hanisch, Kiyoshi Nose, Mats Paulsson, Patrik Maurer, Ursula Hartmann,

Cancer Research and Treatments

Recombinant forms of the glycoprotein TSC-36/Flik were expressed in human cells and used to compare their structural and functional properties with those described for other members of the BM-40/SPARC/osteonectin protein family. TSC-36 was found to occur in two charge isoforms that differ in the extent of sialylation of otherwise identical N-linked, complex type oligosaccharides. Conformational analysis with both circular dichroism and intrinsic fluorescence spectroscopy showed a lack of significant structural changes upon calcium addition or depletion. This finding is in contrast to results obtained for several other BM-40 family members and indicates that the extracellular calcium-binding domain in TSC-36 is non-functional. The lack of conservation of important functional features common to several other members of the BM-40 family indicates that TSC-36, despite its sequence homology to BM-40, has evolved clearly distinct properties. Recombinant forms of the glycoprotein TSC-36/Flik were expressed in human cells and used to compare their structural and functional properties with those described for other members of the BM-40/SPARC/osteonectin protein family. TSC-36 was found to occur in two charge isoforms that differ in the extent of sialylation of otherwise identical N-linked, complex type oligosaccharides. Conformational analysis with both circular dichroism and intrinsic fluorescence spectroscopy showed a lack of significant structural changes upon calcium addition or depletion. This finding is in contrast to results obtained for several other BM-40 family members and indicates that the extracellular calcium-binding domain in TSC-36 is non-functional. The lack of conservation of important functional features common to several other members of the BM-40 family indicates that TSC-36, despite its sequence homology to BM-40, has evolved clearly distinct properties. TSC-36 is an extracellular glycoprotein belonging to the BM-40/SPARC/osteonectin family of proteins containing both extracellular calcium-binding (EC) 1The abbreviations used are: EC, extracellular calcium-binding; FS, follistatin-like; fuc, fucosidase; hex, β-N-acetylhexosaminidase; HPLC, high-performance liquid chromatography; MALDI-MS, matrix-assisted laser desorption ionization mass spectrometry; MALDI-TOF, matrix-assisted laser desorption ionisation-time of flight; neur, neuraminidase; RFU, relative fluorescence unit(s); TGF-β, transforming growth factor β; VWC, von Willebrand factor type C-like; CD, circular dichroism; BisTris, 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol. and follistatin (FS)-like domains (1Shibanuma M. Mashimo J. Mita A. Kuroki T. Nose K. Eur. J. Biochem. 1993; 217: 13-19Crossref PubMed Scopus (229) Google Scholar). The other members of this family are the glycoproteins BM-40/SPARC/osteonectin (2Termine J.D. Kleinman H.K. Whitson S.W. Conn K.M. McGarvey M.L Martin G.R. Cell. 1981; 26: 99-105Abstract Full Text PDF PubMed Scopus (863) Google Scholar), SC1/hevin (3Johnston I.G. Paladino T. Gurd J.W. Brown I.R. Neuron. 1990; 4: 165-176Abstract Full Text PDF PubMed Scopus (124) Google Scholar), QR1 (4Guermah M. Crisanti P. Laugier D. Dezelee P Bidou L. Pessac B. Calothy G. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 4503-4507Crossref PubMed Scopus (63) Google Scholar), and SMOC-1 (5Vannahme C. Smyth N. Miosge N. Gösling S. Frie C. Paulsson M. Maurer P. Hartmann U. J. Biol. Chem. 2002; 277: 37977-37986Abstract Full Text Full Text PDF PubMed Scopus (84) Google Scholar) and -2 (6Vannahme C. Gösling S. Paulsson M. Maurer P. Hartmann U. Biochem. J. 2003; 373: 805-814Crossref PubMed Scopus (93) Google Scholar) as well as the proteoglycans testican-1 (7Alliel P.M. Perin J.P. Jolles P. Bonnet F.J. Eur. J. Biochem. 1993; 214: 347-350Crossref PubMed Scopus (107) Google Scholar), -2 (8Vannahme C. Schübel S. Herud M. Gösling S. Hülsmann H. Paulsson M. Hartmann U. Maurer P. J. Neurochem. 1999; 73: 12-20Crossref PubMed Scopus (56) Google Scholar), and -3 (9Hartmann U. Maurer P. Matrix Biol. 2001; 20: 23-35Crossref PubMed Scopus (108) Google Scholar) (Fig. 1). For several of the members the calcium-binding domain, comprising two EF-hands, has been shown to be functional and for BM-40 and SC-1 an affinity for collagens was shown, leading to an anchorage of these proteins in extracellular matrix structures (10Sasaki T. Hohenester E. Göhring W. Timpl R. EMBO J. 1998; 17: 1625-1634Crossref PubMed Scopus (124) Google Scholar, 11Hambrock H.O. Nitsche D.P. Hansen U. Bruckner P. Maurer P. Hartmann U. J. Biol. Chem. 2003; 278: 11351-11358Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar). For BM-40 the binding was shown to be calcium-dependent, and a collagen-binding surface could be mapped on the EC domain. The many proposed functions of this protein family include inhibition of cell adhesion, influencing tumor invasion and metastasis, growth factor binding, and protease inhibition (12Bradshaw A.D. Sage E H. J. Clin. Invest. 2001; 107: 1049-1054Crossref PubMed Scopus (521) Google Scholar, 13Nakada M. Yamada A. Takino T. Miyamori H. Takahashi T. Yamashita J. Sato H. Cancer Res. 2001; 61: 8896-8902PubMed Google Scholar). TSC-36 is one of the family members with the least similarity to BM-40. The N-terminal region consists of only 12 amino acid residues, and the FS and EC domains are followed by a C-terminal domain with homology to the von Willebrand factor type C-like (VWC) domain. TSC-36 (TGF-β-stimulated clone 36) was first cloned as a TGF-β-induced protein from a mouse osteoblast cell line (1Shibanuma M. Mashimo J. Mita A. Kuroki T. Nose K. Eur. J. Biochem. 1993; 217: 13-19Crossref PubMed Scopus (229) Google Scholar). It was shown that the expression of TSC-36 in fibroblasts could be suppressed through transfection with the oncogenes v-ras and v-myc, whereas v-src, v-raf, and v-abl did not have any influence (14Mashimo J. Maniwa R. Sugino H. Nose K. Cancer Lett. 1997; 113: 213-219Crossref PubMed Scopus (42) Google Scholar). In aggressive, strongly proliferating lung cancer cells TSC-36 mRNA could not be demonstrated, and overexpression of TSC-36 in such cells had an antiproliferative effect (15Sumimoto K. Kurisaki A. Yamakawa N. Tsuchida K. Shimizu E. Sone S. Sugino H. Cancer Lett. 2000; 155: 37-46Crossref PubMed Scopus (64) Google Scholar). By use of a subtractive cDNA library of non-transformed versus v-fos-transformed fibroblasts (FBR-v-fos), TSC-36 was shown to be down-regulated by v-fos and transfection of FBR-v-fos cells with TSC-36 inhibited their migration into Matrigel in an in vitro assay for tumor invasion (16Johnston I.M.P. Spence H.J. Winnie J.N. McGarry L. Vass J.K. Meagher L. Stapleton G. Ozanne B.W. Oncogene. 2000; 19: 5348-5358Crossref PubMed Scopus (62) Google Scholar). The orthologous proteins from rat, human, chick, Xenopus, and macaque have been cloned and sequenced (17Zwijsen A. Blockx H. van Arnheim W. Willems J. Fransen L. Devos K. Raymackers J. Van de Voorde A. Slegers H. Eur. J. Biochem. 1994; 225: 937-946Crossref PubMed Scopus (60) Google Scholar, 18Patel K. Connolly D.J. Amthor H. Nose K. Cooke J. Dev. Biol. 1996; 178: 327-342Crossref PubMed Scopus (53) Google Scholar, 19Okabayashi K. Shoji H. Onuma Y. Nakamura T. Nose K. Sugino H. Asahima M. Biochem. Biophys. Res. Commun. 1999; 254: 42-48Crossref PubMed Scopus (27) Google Scholar, 20Tochitani S. Liang F. Watakabe A. Hashikawa T. Yamamori T. Eur. J. Neurosci. 2001; 13: 297-307Crossref PubMed Scopus (51) Google Scholar). The chick TSC-36 orthologue, also called Flik for follistatin-like, appears to play a role in mesodermal dorsalization, neural induction (18Patel K. Connolly D.J. Amthor H. Nose K. Cooke J. Dev. Biol. 1996; 178: 327-342Crossref PubMed Scopus (53) Google Scholar), and in subsequent maintenance of midline sonic hedgehog signaling thus influencing axial patterning and forebrain development (21Towers P. Patel K. Withington S. Isaac A. Cooke J. Dev. Biol. 1999; 214: 298-317Crossref PubMed Scopus (34) Google Scholar) in the chick embryo. It further influences somite compartmentalization and myogenesis (22Amthor H. Connolly D. Patel K. Brand-Saberi B. Wilkinson D.G. Cooke J. Christ B. Dev. Biol. 1996; 178: 343-362Crossref PubMed Scopus (83) Google Scholar). We have recombinantly expressed and characterized mouse TSC-36 with the purpose of comparing its properties to those of other members of the BM-40 protein family. In doing so we identified two charge isoforms that could be shown to differ in their sialylation pattern. Our work provides a molecular basis for further studies of TSC-36 function. Construction of Expression Vectors for Full-length and Truncated forms of TSC-36—The plasmid pGEM/36-A1 (1Shibanuma M. Mashimo J. Mita A. Kuroki T. Nose K. Eur. J. Biochem. 1993; 217: 13-19Crossref PubMed Scopus (229) Google Scholar) was used to generate a full-length mouse TSC-36 cDNA, which encompasses the entire open reading frame but lacks the signal peptide, by PCR using the forward primer 1 (5′-GCCCCGCTAGCCCACGGCGAGGAGGAACC) and reverse primer 2 (5′-CAATGACTGCGGCCGCTTAGATCTCTTTGGTGTTCAC). The primers introduced new restriction sites (NheI and NotI) and a stop codon. The NheI/NotI-restricted PCR products were purified and inserted between the same restriction sites of pCEP-Pu-BM40 (23Kohfeldt E. Maurer P. Vannahme C. Timpl R. FEBS Lett. 1997; 414: 557-561Crossref PubMed Scopus (203) Google Scholar). Sequencing revealed a 9-bp deletion (nucleotides 573-581) coding for the amino acids FLK. The deletion was corrected by replacing a EcoRI/NotI fragment with a corresponding 2-kb fragment from the pRcCMV/36 c1-2 containing the missing 9 bp. The corrected vector was subjected to a second PCR using the above-mentioned primers. This vector also served as a template to generate cDNA fragments representing the pair of EC-VWC domains and the EC domain using the following primers: forward primer 3 (5′-GCCCCGCTAGCCGTTGTCTGCTATCAAGCTAAC) and reverse primer 2 (see above) were used for the pair of EC-VWC domains and forward primer 3 (see above) and reverse primer 4 (5′-CAATGACTGCGGCCGCTTACTCAGGAGGGTTGAAGGAT) for the EC domain. The NheI/NotI-restricted PCR products were purified and, in the case of full-length TSC-36, inserted between the same restriction sites of pCEP-Pu-BM40 (23Kohfeldt E. Maurer P. Vannahme C. Timpl R. FEBS Lett. 1997; 414: 557-561Crossref PubMed Scopus (203) Google Scholar), to obtain the final expression vector pCEP-Pu-TSC-36. The cDNA fragments of EC-VWC and EC were cloned in the modified expression vector pCEP-Pu with an N-terminal tag consisting of six histidines, a myc epitope, and an enterokinase recognition site for cleavage of the tag (24Wuttke M. Müller S. Nitsche D.P. Paulsson M. Hanisch F.-G. Maurer P. J. Biol. Chem. 2001; 276: 36839-36848Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar), to obtain expression vectors pCEP-Pu-TSC-36-EC-VWC and pCEP-Pu-TSC-36-EC, respectively. The correct insertion and sequence of the inserts was verified by cycle sequencing of both strands using an ABI Prism 377 Automated Sequencer (PE Biosystems). Cell Culture and Transfection—The human embryonic kidney cell line EBNA-293 (Invitrogen), which constitutively expresses the EBNA-1 protein from Epstein-Barr virus, was used for transfection with pCEP-Pu-TSC-36, pCEP-Pu-TSC-36-EC-VWC, and pCEP-Pu-TSC-36-EC and selected with puromycin as described previously (23Kohfeldt E. Maurer P. Vannahme C. Timpl R. FEBS Lett. 1997; 414: 557-561Crossref PubMed Scopus (203) Google Scholar). Conditioned serum-free media were collected 48 h after replacement of the serum-containing medium and stored at -80 °C. Purification of Full-length TSC-36 and Separation of TSC-36 A and B—Conditioned serum-free medium (4 liters) was dialyzed against 50 mm Tris-HCl, pH 8.6, and passed over a DEAE-Sepharose FF (Amersham Biosciences) column equilibrated in the same buffer at 4 °C. The column was eluted with a linear NaCl gradient, and the recombinant protein eluted at 0.1-0.2 m NaCl. The TSC-36-containing fractions were dialyzed against 0.18 m ammonium sulfate in 40 mm sodium phosphate, pH 7.0, and applied to a C4-Butyl-Sepharose FF (Amersham Biosciences) column that bound several contaminants. TSC-36 was found in the flowthrough and was dialyzed against 50 mm Tris-HCl, pH 8.6, concentrated on a DEAE-Sepharose FF column, and eluted with 0.3 m NaCl. Final purification was achieved by gel filtration on Sephadex G-75 in 5 mm Tris-HCl, pH 7.4. The correct N-terminal sequence was confirmed by Edman degradation performed on a protein sequencer (model 473A, PE Biosystems) following the manufacturer's protocol. Separation of the two charge isoforms of TSC-36 was achieved by chromatography on a ResourceQ (Amersham Biosciences) column eluted with a non-linear NaCl gradient in 50 mm Tris-HCl, pH 9.5. Isoform A was eluted at 0.15 m and isoform B at 0.20 m NaCl. Purification of the EC-VWC Domain Pair—The serum-free medium was concentrated by ultrafiltration, dialyzed against 0.3 m NaCl, 50 mm sodium phosphate, pH 7.0, and applied to a column with cobalt Talon Metal Affinity Matrix (Clontech). The EC-VWC protein eluted at a concentration of 50-130 mm imidazole and was dialyzed against 50 mm BisTris, pH 6.0, and applied to a ResourceQ column eluted with a non-linear NaCl gradient in the same buffer. Isoform A was eluted at 0.15 m and isoform B at 0.25 mm NaCl. Purification of the EC Domain—The serum-free medium was dialyzed against 50 mm Tris-HCl, pH 7.4, applied to a column with Q-Sepharose FF (Amersham Biosciences) for concentration, eluted with a linear NaCl gradient in the same buffer and was dialyzed against 50 mm sodium phosphate, pH 8.0. The protein was applied to a column with cobalt Talon Metal Affinity Matrix (Clontech) in this buffer and eluted at a concentration of 20-50 mm imidazole. On non-reducing SDS-PAGE the preparation showed two diffuse bands, out of which the upper was sensitive to reduction, indicating dimerization due to incorrect disulfide bonding. The monomeric EC domain was separated from this contaminant by gel filtration on a column of Sephadex G50 (Amersham Biosciences) equilibrated in 50 mm ammonium bicarbonate, pH 8.0. SDS-PAGE—SDS-PAGE was performed according to the protocol of Laemmli (25Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207233) Google Scholar), and proteins were detected by staining with Coomassie Brilliant Blue R250. Circular Dichroism Spectroscopy—CD spectra were recorded in a Jasco model 715 CD spectropolarimeter at 25 °C in thermostatted quartz cells of optical path length 1 mm. The molar ellipticity [θ] (expressed in degrees·cm2·dmol-1) was calculated on the basis of a mean residue molecular mass of 110 Da. The Ca2+ dependence of the CD spectrum was measured by addition of 2 mm CaCl2. Reversibility of the conformational change was tested by subsequent addition of 4 mm EDTA. A baseline with buffer (5 mm Tris-HCl, pH 7.4) was recorded separately and subtracted from each spectrum. Protein concentrations were determined using extinction coefficients of ϵ280 = 34,200 m-1 cm-1 for TSC-36, 32,320 m-1 cm-1 for the EC-VWC domain pair, and 23,590 m-1 cm-1 for the EC domain calculated according to Gill and von Hippel (26Gill S.C. von Hippel P.H. Anal. Biochem. 1989; 182: 319-326Crossref PubMed Scopus (5060) Google Scholar). Fluorescence Emission Spectroscopy—Intrinsic fluorescence was measured with a PerkinElmer Life Sciences LS50B spectrofluorometer in 10-mm path length rectangular cells at 25 °C with excitation at 280 nm. Emission spectra for the full-length TSC-36, the EC-VWC domain pair, and the EC domain (each 1 μm) were recorded in 5 mm Tris-HCl, pH 7.4. Spectra were also recorded with 2 mm CaCl2 in the same buffer and after addition of 4 mm EDTA. Mass Spectrometry—For determination of protein mass and N-glycan content purified proteins (20 pmol) were reduced with 0.5% mercaptoethanol and denatured by heating to 100 °C for 5 min. After denaturation, the proteins were incubated with 0.3 unit of N-glycosidase F (Roche Molecular Biochemicals) per milligram in 50 mm Tris-HCl, pH 7.4, overnight at 37 °C. Digested and non-digested samples were analyzed by MALDI-TOF mass spectrometry. For analysis of potential truncations and N-glycosylation sites, proteins were further digested with trypsin (sequencing grade, Promega, 1:50 enzyme:substrate in 50 mm ammonium bicarbonate, pH 8.0) for 2-4 h at 37 °C and, after reduction with dithiothreitol and alkylation with iodoacetamide, subjected to MALDI-TOF mass spectrometry. For MALDI-TOF mass spectrometry analysis, the samples were dissolved in 5 μl of 0.1% aqueous trifluoroacetic acid. MALDI-MS was carried out in linear mode on a Bruker Reflex IV equipped with a video system, a nitrogen UV laser (λmax = 337 nm), and a HiMass detector. 1 μl of the sample solution was placed on the target, and 1 μl of a freshly prepared saturated solution of sinapinic acid in acetonitrile/H2O (2:1) with 0.1% trifluoroacetic acid was added. The spot was then recrystallized by addition of another 1 μl of acetonitrile/H2O (2:1), which resulted in a fine crystalline matrix. For recording of the spectra, an acceleration voltage of 20 kV was used and the detector voltage was adjusted to 1.9 kV. Approximately 500 single laser shots were summed into an accumulated spectrum. Calibration was carried out using the single and doubly protonated ion signal of bovine serum albumin for external calibration. Structural Analysis of N-Glycans—The N-linked oligosaccharides were released by treatment with N-glycosidase F, labeled with 2-aminobenzamide, digested with neuraminidase and other exoglycosidases, and analyzed by anion exchange or normal phase HPLC as previously described (24Wuttke M. Müller S. Nitsche D.P. Paulsson M. Hanisch F.-G. Maurer P. J. Biol. Chem. 2001; 276: 36839-36848Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). Antiserum Production—Recombinant TSC-36 was used to immunize rabbits. TSC-36 was covalently coupled to CNBr-Sepharose (Amersham Biosciences), and specific antibodies were prepared by repeated passage of the antiserum over the affinity matrix and elution with 3 m KSCN in 50 mm Tris-HCl, pH 7.4. Eluted antibodies were immediately dialyzed against 0.15 m NaCl, 50 mm Tris-HCl, pH 7.4. Immunoblot Analysis of Cell Culture Media—Cultures of mouse embryonic fibroblasts, bovine articular chondrocytes, and of several established cell lines were grown to confluency. Serum-free conditioned media, harvested 48 h after replacement of the serum-containing medium, were submitted to SDS-PAGE on 12% gels under reducing conditions. Proteins were transferred electrophoretically onto nitrocellulose and incubated with affinity-purified antibodies to TSC-36, followed by a swine anti-rabbit IgG coupled to peroxidase (DAKO). ECL chemiluminescence (Amersham Biosciences) was used to detect peroxidase activity. Recombinant Production of Full-length and Truncated TSC-36 Proteins—We expressed TSC-36 in human embryonic kidney 293 cells to ensure proper folding and post-translational modification. In addition to the full-length protein we produced two truncated forms corresponding to the EC domain and the domain pair EC and VWC, respectively. For simplicity of purification, these truncated proteins were provided with an N-terminal His-Myc tag (24Wuttke M. Müller S. Nitsche D.P. Paulsson M. Hanisch F.-G. Maurer P. J. Biol. Chem. 2001; 276: 36839-36848Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). As a last step in the chromatographic purification of full-length TSC-36 from cell culture medium, the protein was applied to a ResourceQ ion exchanger in 50 mm Tris-HCl, pH 9.5. Under those conditions the material eluted in two distinct peaks at 0.15 and 0.2 m NaCl, respectively, indicating a heterogeneity (Fig. 2A). Rechromatography of pools made from the first and second peak resulted in their elution in identical positions as in the initial run (Fig. 2, B and C), showing that they represent stable variants of TSC-36. Similarly, the protein EC-VWC eluted in two peaks when analyzed by chromatography on ResourceQ (not shown), after initial purification by affinity chromatography on a cobalt matrix, showing that the difference between the two isoforms, designated A and B, must reside in the domain pair EC-VWC. Because of low yields, the recombinant EC domain was not analyzed for heterogeneity in this manner. SDS-PAGE of the purified proteins (Fig. 3), performed under both reducing and non-reducing conditions, showed a major diffuse band for each sample, with TSC-36 A migrating with an apparent molecular mass of 45-51 kDa, TSC-36 B with 48-55 kDa, EC-VWC A with 42-48 kDa, and EC-VWC B with 45-50 kDa. A lower molecular mass component seen in the sample of EC-VWC B was by immunoblotting shown to represent a degradation fragment. The isolated EC domain appeared even more polydisperse than the other proteins and was seen as a smear in the region of 30 kDa. In each case the migration was considerably slower than expected from the molecular mass calculated from amino acid sequence, which indicated the presence of post-translational modifications and/or an abnormal electrophoretic behavior.Fig. 3SDS-polyacrylamide gel electrophoresis and domain structure of TSC-36 derived proteins. A, SDS-PAGE of the purified recombinant proteins was performed on a 12% gel without (lanes 1-5) or with prior reduction with β-mercaptoethanol (lanes 6-10). The samples were full-length TSC-36 isoform A (lanes 1 and 6) and B (lanes 2 and 7), the EC-VWC domain pair isoform A (lanes 3 and 8) and B (lanes 4 and 9), and the EC domain (lanes 5 and 10). B, domain structures of the recombinant proteins.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Conformation of TSC-36—For each of the recombinant proteins, circular dichroism spectra were recorded in the far UV region. For full-length TSC-36 and the EC-VWC domain pair, the measurements were made on pools containing both isoforms. No significant differences could be discerned between the spectra recorded in the presence of 2 mm calcium chloride and after addition of excess EDTA (Fig. 4). The spectrum for the full-length TSC-36 was typical for a protein with high α-helical content, with a minimum at 208 nm and a shoulder at 222 nm. The EC domain and the EC-VWC domain pair showed similar spectra, but with higher contents of β-sheet, indicating that β-structure predominates in the VWC domain (Table I).Table ISecondary structure of recombinant TSC-36 proteins calculated from circular dichroism spectraSecondary structureProteinsTSC-36 flEC-VWCEC%α-Helix29 (32)11 (11)22 (19)β-Sheet14 (10)38 (37)24 (24)β-Turns22 (28)19 (19)17 (22)Unordered structure35 (33)33 (33)33 (33)Total100 (103)101 (100)96 (98) Open table in a new tab The apparent lack of conformational changes upon removal of calcium indicated that the EC domain in TSC-36 may not be functional in calcium binding, but could also be due to that a change is too small or too local to be detected by circular dichroism spectroscopy. We therefore also measured intrinsic fluorescence at calcium saturation or depletion. The amino acid sequence of TSC-36 contains eight tyrosines and four tryptophans that are distributed between domains so that all three recombinant proteins are able to give signals in fluorescence spectroscopy. Only minor, non-reversible changes in fluorescence intensity were observed, again indicating a lack of calcium-dependent conformational changes (Fig. 4). Mass Spectrometry Analysis of TSC-36 and Its Post-translational Modifications—The mass of each of the recombinant TSC-36-derived proteins was determined by MALDI-TOF mass spectroscopy and compared with the theoretical mass calculated from the amino acid sequence (Table II). For each protein the measured mass was at least 20% higher than the calculated one, showing the presence of extensive post-translational modifications. The contribution of N-glycosidically linked glycans was determined by treatment of proteins with N-glycosidase F and renewed MALDI-TOF analysis of the deglycosylated products. The results showed that the bulk of the mass discrepancy was due to N-linked oligosaccharides and that both for the full-length TSC-36 and the EC-VWC domain pair the mass in N-linked glycans was higher in the more anionic B isoforms. Prediction of N-glycosylation sites using the NetNGlyc software indicated substitution of asparagines in positions 142, 173, 178, and 221, which all reside in the EC domain. Similar amounts of N-linked oligosaccharides were found in the full-length TSC-36 and the EC-VWC domain pair in good agreement with this prediction. However, the isolated EC fragment carried a lower glycosylation than the larger proteins, possibly due to the fact that Asn221 is only three amino acid residues from the C terminus of this fragment (Fig. 5). This N-glycosylation site may be less efficiently used when close to the artificial C terminus than in a larger protein containing also a VWC domain. In each case the mass after N-glycosidase F digestion was slightly higher than the theoretical one, possibly indicating the presence of additional, unidentified forms of post-translational modifications. Prediction of O-glycosylation with the NetOGlyc software indicated a possible substitution of the serines in position 101 and 105.Table IIMolecular masses of the recombinant TSC-36 proteins and their glycan portionsTSC-36 fl isoform ATSC-36 fl isoform BEC-VWC isoform AEC-VWC isoform BECMolecular masscalculated, Da33,00833,00827,64327,64318,979Molecular mass, Da42,310aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.41,880aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.35,387aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.37,985aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.25,023aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.Molecular mass after N-glycosidase F, Da33,828aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.33,467aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.28,148aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.28,168aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.19,688aValues were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans.Molecular mass of N-glycans, Da8,4828,4137,2399,8175,335a Values were experimentally determined by MALDI-TOF mass spectrometry and used to calculate the mass of N-glycans. Open table in a new tab To determine the potential usage of the four N-glycosylation and two O-glycosylation sites both isoforms of full-length TSC-36 were digested with trypsin, and the peptides were analyzed by MALDI-TOF mass spectrometry. The identified peptides covered 86% of the amino acid residues in TSC-36 A and 80% of those in TSC-36 B (Fig. 5). The seven C-terminal residues could not be recovered, probably because of the presence of multiple trypsin cleavage sites in this region of the sequence, likely to result in very small fragments. Peptides spanning the regions of the protein corresponding to asparagines 142, 173, and 178 were missing in both isoforms, indicating that these N-glycosylation sites are used. On the other hand, peptides were found containing Asn221 and Ser101 and Ser105. This result does, however, not exclude that these sites are used in some of the protein molecules. The peptide Gly272-Lys281 was recovered only from the A isoform, but the two threonine residues at positions 275 and 277 in this peptide gave low scores in NetOGlyc and are unlikely to be substituted. Sialic Acid Content and Structure of TSC-36 N-Glycans—N-Glycans were enzymatically released from the A and B isoforms of TSC-36 and labeled with the fluorescent dye 2-aminobenzamide for detection. Sialo-structures were separated by anion exchange HPLC (Fig. 6), which revealed glycans with up to four sialic acid residues. The N-glycans from the TSC-36 B isoform had a significantly higher sialic acid content that those from the A isoform (Table III). The sialic acid residues could be efficiently removed from the isolated N-glycans by neuraminidase digestion (Fig. 6). When the neuraminidase was applied to the full-length TSC-36 A and B isoforms, these eluted at the exactly same position upon ion exchange fast protein liquid chromatography on a ResourceQ column, indicating that differences in sialic acid content is the sole cause of the two charge isoforms (not shown).Table IIISialic acid composition of the N-glycan structures of TSC-36 isoforms A and BTSC-36 ATSC-36 B%Neutral N-glycansN020.011.0Monosialo N-glycansS141.331.9Disialo N-glycansS217.920.7Trisialo N-glycansS316.024.5Tetrasialo N-glycansS44.811.9 Open table in a new tab The neutral glycans obtained after treatment with neuraminidase were separated on a normal phase HPLC column, and the asialo- structures were identified by sequential digestion with different exoglycosidases (Fig. 7). The same glycan structures were obtained for both isoforms of TSC-36. These were three different completely galactosylated complex type oligosaccharides, all carrying a core fucose (Table IV). Both isoforms show the same composition of 18% biantennary (A2G2F), 23% triantennary (A3G3F), and 59% tetraantennary (A4G4F) complex type oligosaccharides. Therefore, the difference between the two isoforms is not due to different N-glycan structures but to variations in sialic acid content.Table IVAsialo N-glycan structures present in TSC-36 Open table in a new tab Analysis of TSC-36 Expressed by Primary Cells and by Established Cell Lines—An antiserum raised in rabbit against the full-length TSC-36 was affinity-purified by binding to the same antigen coupled to CNBr-Sepharose. The affinity-purified antibody did not show cross-reactivity with any known member of the BM-40 family (BM-40; testican-1, -2 and -3; SC1; and SMOC-1 and -2) when tested native in slot blots or in Western blot after SDS-PAGE (not shown). This antibody was used to determine the expression and N-linked glycosylation of TSC-36 in supernatants from primary chondrocytes and fibroblasts and from established cell lines of osteoblastic (MG-63) and fibroblastic (Wi-26) origins and to compare endogenous TSC-36 from these cells with the recombinant protein (Fig. 8). TSC-36 from these sources gave diffuse bands with mobilities similar but not identical to that of recombinant TSC-36. All samples were susceptible to digestion with N-glycosidase F, showing the presence of N-linked oligosaccharides, but the digested proteins still showed heterogeneity, running as one or two bands with differing mobility. Because the cell lines MG-63 and Wi-26 are both of human origin and the TSC-36 molecules produced showed clearly different mobilities after N-glycosidase F digestion, a heterogeneity must exist beyond that introduced by N-linked glycosylation. A broader screen of established cell lines was performed to determine cellular sources of TSC-36 (Fig. 9). TSC-36 was detected in the medium of all osteosarcoma and chondrosarcoma cell lines tested. Cells of the fibroblast lineage gave variable results with HT1080 and Wi-26 being positive and L132 negative. The astrocyte cell line S6E and the glioma cell line C6 were strongly positive. The mammary carcinoma cell lines MDA MB-231 and ZR 75-1 gave moderately strong signals, whereas T47D and MCF-7 were negative. A large number of other epithelial or carcinoma cell lines were also negative. Among the positive cell lines were those of mouse, rat, and human origin, showing that the antibody detects TSC-36 from these three species. The TSC-36 forms produced by the different cell lines all migrated in the range of 45-55 kDa seen for recombinant murine TSC-36 expressed in 293 cells, with minor differences indicating a cell-type-specific variability in post-translational modification. A central aspect of our study was to determine if the homology in sequence and domain structure between TSC-36 and BM-40 is reflected in a similarity between the two proteins in structural and functional properties. A comparison of the two sequences showed the presence of two putative EF-hands in the TSC-36 sequence (27Maurer P. Hohenadl C. Hohenester E. Göhring W. Timpl R. Engel J. J. Mol. Biol. 1995; 253: 347-357Crossref PubMed Scopus (114) Google Scholar). The first EF-hand in TSC-36 shows a deletion of one amino acid residue compared with the canonical EF-hand sequence, a feature that does not exclude a calcium affinity as demonstrated by the crystal structure of scallop myosin essential light chain (28Xie X. Harrison D.H. Schlichtling I. Sweet R.M. Kalabolis V.N. Szent-Gyorgyi A.G. Cohen C. Nature. 1994; 368: 306-312Crossref PubMed Scopus (266) Google Scholar). The second EF-hand conforms to the consensus sequence. In an attempt to test the predicted calcium binding, recombinant TSC-36 and the proteins representing the EC-VWC domain pair and the EC domain were analyzed by circular dichroism spectroscopy with and without addition of calcium and in the presence of surplus EDTA. For all three proteins calcium addition resulted in only marginal changes in the molar ellipticity and addition of EDTA resulted in a not significant reversal (Fig. 4, A-C). As circular dichroism spectroscopy did not give any indication of a calcium-induced conformational switch, the study was extended to measurement of fluorescence emission at calcium saturation and depletion (Fig. 4, D-F). Because of the favorable distribution of tryptophan and tyrosine residues in the TSC-36 sequence, this method was expected to give sensitive detection of conformational changes. Despite this, no calcium-induced rearrangements could be detected in either of the three recombinant proteins. Therefore, we conclude that if calcium binding to TSC-36 occurs, this does not have any major consequences for the conformation of the protein. For comparison, when the same experiments were performed with human and murine BM-40 calcium addition gave significant increases in the proportion of α-helix, as determined by circular dichroism spectroscopy, and a 95% decrease in the intensity of the emitted fluorescence (29Engel J. Taylor W. Paulsson M. Sage H. Hogan B. Biochemistry. 1987; 26: 6958-6965Crossref PubMed Scopus (162) Google Scholar, 30Maurer P. Mayer U. Bruch M. Jenö P. Mann K. Landwehr R. Engel J. Timpl R. Eur. J. Biochem. 1992; 205: 233-240Crossref PubMed Scopus (66) Google Scholar, 31Pottgiesser J. Maurer P. Mayer U. Nischt R. Mann K. Timpl R. Krieg T. Engel J. J. Mol. Biol. 1994; 238: 563-574Crossref PubMed Scopus (53) Google Scholar). In contrast, in the case of Caenorhabditis elegans BM-40 neither of the spectroscopic methods gave any indication of calcium binding. C. elegans BM-40 lacks 23 amino acid residues in the EC domain, which encompass the α-helix B. It also contains an additional cysteine pair, which is predicted to form an additional disulfide bond, locking the EC domain in a more rigid structure (32Maurer P. Göhring W. Sasaki T. Mann K. Timpl R. Nischt R. Cell Mol. Life Sci. 1997; 53: 478-484Crossref PubMed Scopus (29) Google Scholar). Interestingly, TSC-36 shows a similar deletion of 31 amino acid residues, which even affects two α-helices, B and C, in the EC domain. During the purification of full-length TSC-36 and the EC-VWC domain pair by anion exchange chromatography we found that the recombinant proteins occur in two differently charged isoforms. In earlier studies of TSC-36 secreted by C6 glioma cells, a related heterogeneity was noticed by isoelectric focusing where TSC-36 showed a spectrum of isoelectric points ranging from 5.0 to 6.0 (17Zwijsen A. Blockx H. van Arnheim W. Willems J. Fransen L. Devos K. Raymackers J. Van de Voorde A. Slegers H. Eur. J. Biochem. 1994; 225: 937-946Crossref PubMed Scopus (60) Google Scholar). We could conclusively show that this heterogeneity is due to differential sialylation (Table III) and that the two isoforms have an identical protein structure (Fig. 5) and the same oligosaccharide backbones in the same proportions (Table IV). Expression of TSC-36 was studied in primary chondrocytes and fibroblasts and in a variety of established cell lines of different origins (Figs. 8 and 9). All of these showed molecular masses for TSC-36 in the range of 45-55 kDa, when analyzed by SDS-PAGE, indicating that the high amount (more than 20%) of glycosylation seen for the recombinant proteins is likely to be representative for naturally occurring forms of TSC-36. The functional role of the high glycosylation and differential sialylation of TSC-36 remains unclear, but it is likely to make the protein strongly hydrophilic, and it may well affect its ligand interactions. We are grateful to Dr. Marcus Macht, University of Cologne, for his help and advice regarding mass spectrometry.