N-Linked Glycosylation and Sialylation of the Acid-labile Subunit
1999; Elsevier BV; Volume: 274; Issue: 9 Linguagem: Inglês
10.1074/jbc.274.9.5292
ISSN1083-351X
AutoresJackie Janosi, Sue M. Firth, J. J. Bond, Robert C. Baxter, Patric J. D. Delhanty,
Tópico(s)Enzyme Structure and Function
ResumoOver 75% of the circulating insulin-like growth factors (IGF-I and -II) are bound in 140-kDa ternary complexes with IGF-binding protein-3 (IGFBP-3) and the 84–86-kDa acid-labile subunit (ALS), a glycoprotein containing 20 kDa of carbohydrate. The ternary complexes regulate IGF availability to the tissues. Since interactions of glycoproteins can be influenced by their glycan moieties, this study aimed to determine the role of ALS glycosylation in ternary complex formation. Complete deglycosylation abolished the ability of ALS to associate with IGFBP-3. To examine this further, seven recombinant ALS mutants each lacking one of the seven glycan attachment sites were expressed in CHO cells. All the mutants bound IGFBP-3, demonstrating that this interaction is not dependent on any single glycan chain. Enzymatic desialylation of ALS caused a shift in isoelectric point from 4.5 toward 7, demonstrating a substantial contribution of anionic charge by sialic acid. Ionic interactions are known to be involved in the association between ALS and IGFBP-3. Desialylation reduced the affinity of ALS for IGFBP-3·IGF complexes by 50–80%. Since serum protein glycosylation is often modified in disease states, the dependence of IGF ternary complex formation on the glycosylation state of ALS suggests a novel mechanism for regulation of IGF bioavailability. Over 75% of the circulating insulin-like growth factors (IGF-I and -II) are bound in 140-kDa ternary complexes with IGF-binding protein-3 (IGFBP-3) and the 84–86-kDa acid-labile subunit (ALS), a glycoprotein containing 20 kDa of carbohydrate. The ternary complexes regulate IGF availability to the tissues. Since interactions of glycoproteins can be influenced by their glycan moieties, this study aimed to determine the role of ALS glycosylation in ternary complex formation. Complete deglycosylation abolished the ability of ALS to associate with IGFBP-3. To examine this further, seven recombinant ALS mutants each lacking one of the seven glycan attachment sites were expressed in CHO cells. All the mutants bound IGFBP-3, demonstrating that this interaction is not dependent on any single glycan chain. Enzymatic desialylation of ALS caused a shift in isoelectric point from 4.5 toward 7, demonstrating a substantial contribution of anionic charge by sialic acid. Ionic interactions are known to be involved in the association between ALS and IGFBP-3. Desialylation reduced the affinity of ALS for IGFBP-3·IGF complexes by 50–80%. Since serum protein glycosylation is often modified in disease states, the dependence of IGF ternary complex formation on the glycosylation state of ALS suggests a novel mechanism for regulation of IGF bioavailability. insulin-like growth factor acid-labile subunit insulin-like growth factor-binding protein endo-β-N-acetylglucosaminidase peptide-N 4-(acetyl-β-glucosaminyl)-asparagine amidase N-acetylneuraminidase III endo-α-N-acetylgalactosaminidase 2,3-dehydro-2-deoxy-N-acetylneuraminic acid Chinese hamster ovary polyacrylamide gel electrophoresis bovine serum albumin α-modified Eagle's medium 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid Insulin-like growth factors (IGF)1 I and II are peptide hormones that regulate the differentiation and proliferation of a large number of cell types and also have a role in glucose homeostasis (1Jones J.I. Clemmons D.R. Endocr. Rev. 1995; 16: 3-34Crossref PubMed Google Scholar). At least 75% of the total circulating IGFs are carried in 130–150-kDa ternary complexes containing IGF-binding protein-3 (IGFBP-3) (2Rajaram S. Baylink D.J. Mohan S. Endocr. Rev. 1997; 18: 801-831Crossref PubMed Scopus (963) Google Scholar) and an 85-kDa glycoprotein, the acid-labile subunit (ALS) (3Baxter R.C. Martin J.L. Beniac V.A. J. Biol. Chem. 1989; 264: 11843-11848Abstract Full Text PDF PubMed Google Scholar). Recently, IGFBP-5 was also found to form a similar ternary complex with the IGFs and ALS in serum (4Twigg S.M. Baxter R.C. J. Biol. Chem. 1998; 273: 6074-6079Abstract Full Text Full Text PDF PubMed Scopus (161) Google Scholar). It is thought that the size of the ternary complex restricts the passage of IGFs to target cells, while free IGFs, or IGFs in binary complexes with IGFBPs, can cross the capillary endothelial barrier. Therefore, ALS regulates the hypoglycemic and mitogenic potential of the circulating IGFs via the formation of the ternary complexes. Furthermore, ALS plays a vital role in maintaining a circulating store of the IGFs, IGFBP-3, and possibly IGFBP-5, by significantly increasing their serum half-lives (5Guler H.P. Zapf J. Schmid C. Froesch E.R. Acta Endocrinol. 1989; 121: 753-758Crossref PubMed Google Scholar, 6Lewitt M.S. Saunders H. Baxter R.C. Endocrinology. 1993; 133: 1797-1802Crossref PubMed Scopus (64) Google Scholar).Despite the importance of the ternary complexes in regulating serum IGF bioactivity, nothing is known about the structural aspects of ALS that enable it to interact with IGFBPs. There are two features of ALS structure that may play a part. First, the protein backbone of ALS is made up of repeating blocks each containing 24 amino acids, of which 6 are typically leucine residues. This places ALS in the leucine-rich repeat superfamily of proteins (7Leong S.R. Baxter R.C. Camerato T. Dai J. Wood W.I. Mol. Endocrinol. 1992; 6: 870-876Crossref PubMed Scopus (66) Google Scholar), all of which are involved in protein-protein interactions (8Kobe B. Deisenhofer J. Curr. Opin. Struct. Biol. 1995; 5: 409-416Crossref PubMed Scopus (322) Google Scholar). Second, serum ALS is heavily and heterogeneously glycosylated with N-linked glycan chains (3Baxter R.C. Martin J.L. Beniac V.A. J. Biol. Chem. 1989; 264: 11843-11848Abstract Full Text PDF PubMed Google Scholar), and glycosylation is known to influence the interactions of many proteins.Electrophoretic studies have shown that human ALS circulates as two glycoforms. Serum-purified ALS displays a characteristic doublet on SDS-PAGE at 84–86 kDa, which is reduced to a single band of less than 70 kDa after enzymatic removal of the N-linked sugars (3Baxter R.C. Martin J.L. Beniac V.A. J. Biol. Chem. 1989; 264: 11843-11848Abstract Full Text PDF PubMed Google Scholar). There are seven consensus N-linked sugar attachment sites within the amino acid sequence of ALS that are conserved between primate and rodent (7Leong S.R. Baxter R.C. Camerato T. Dai J. Wood W.I. Mol. Endocrinol. 1992; 6: 870-876Crossref PubMed Scopus (66) Google Scholar, 9Dai J. Baxter R.C. Biochem. Biophys. Res. Commun. 1992; 188: 304-309Crossref PubMed Scopus (30) Google Scholar, 10Delhanty P. Baxter R.C. Biochem. Biophys. Res. Commun. 1996; 227: 897-902Crossref PubMed Scopus (14) Google Scholar, 11Boisclair Y.R. Seto D. Hsieh S. Hurst K.R. Ooi G.T. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 10028-10033Crossref PubMed Scopus (37) Google Scholar). One site occurs almost in the center of the sequence, and a cluster of three sites is found toward each terminus. Between six and seven bands are observed upon partial deglycosylation of ALS derived from serum, suggesting that multiple sites are used (12Liu F. Hintz R.L. Khare A. Diaugustine R.P. Powell D.R. Lee P.D. J. Clin. Endocrinol. Metab. 1994; 79: 1883-1886Crossref PubMed Google Scholar).Although there are no studies that directly identify physical features of ALS involved in ternary complex formation, there is evidence that charge-charge interaction exists between ALS and the IGF·IGFBP-3 binary complex. Polyanions, polycations, and increasing ionic strength all decrease the affinity of ALS for IGFBP-3 (13Holman S.R. Baxter R.C. Growth Regul. 1996; 6: 42-47PubMed Google Scholar, 14Baxter R.C. Biochem. J. 1990; 271: 773-777Crossref PubMed Scopus (55) Google Scholar). Recently we have shown that the removal of basic residues in the carboxyl-terminal region of IGFBP-3 decreased its affinity for ALS by 90%, indicating the importance of positive charge in this region (15Firth S.M. Ganeshprasad U. Baxter R.C. J. Biol. Chem. 1998; 273: 2631-2638Abstract Full Text Full Text PDF PubMed Scopus (129) Google Scholar). From these observations, it is likely that negative charges on ALS may be involved in the interaction. At physiological pH, ALS binds tightly to weak anion exchange columns, indicating that it has a net negative charge (3Baxter R.C. Martin J.L. Beniac V.A. J. Biol. Chem. 1989; 264: 11843-11848Abstract Full Text PDF PubMed Google Scholar).Since carbohydrates are a potential source of negative charge in glycoproteins, as well as being involved in protein interactions, we have investigated whether N-linked sugars on ALS play a role in the formation of ternary complexes between IGFs, IGFBP-3, and ALS.EXPERIMENTAL PROCEDURESReagentsPreparations of natural human ALS, human IGFBP-3, and rabbit antiserum against IGFBP-3 were similar to those used in previous studies (3Baxter R.C. Martin J.L. Beniac V.A. J. Biol. Chem. 1989; 264: 11843-11848Abstract Full Text PDF PubMed Google Scholar, 16Baxter R.C. Bayne M.L. Cascieri M.A. J. Biol. Chem. 1992; 267: 60-65Abstract Full Text PDF PubMed Google Scholar). ALS was radioiodinated and purified by ion-exchange chromatography as described previously (17Baxter R.C. J. Clin. Endocrinol. Metab. 1990; 70: 1347-1353Crossref PubMed Scopus (247) Google Scholar). IGF-I (Genentech, San Francisco, CA) was iodinated and cross-linked to IGFBP-3 as in previous studies (18Baxter R.C. J. Clin. Endocrinol. Metab. 1988; 67: 265-272Crossref PubMed Scopus (202) Google Scholar). Restriction enzymes and materials for site-directed mutagenesis were from Promega Corp. (Madison, WI). Bovine serum albumin (BSA; radioimmunoassay grade, fraction V), γ-globulin, hexadimethrine bromide (Polybrene), dexamethasone, hypoxanthine, xanthine, thymidine, and mycophenolic acid were purchased from Sigma. Aminopterin was obtained from Life Technologies Inc.. Nucleoside-free α-modified Eagle's medium (α-MEM) and fetal calf serum were from Cytosystems (New South Wales, Australia). Centricon 30 microconcentrators were obtained from Amicon (Beverly, MA). n-Octyl glucoside,O-glycosidase (endo-α-N-acetylgalactosaminidase), and peptide-N-glycosidase F (PNGase F; peptide-N 4-(acetyl-β-glucosaminyl)-asparagine amidase) were obtained from Boehringer Mannheim.N-Acetylneuraminidase III (NANase III) was from Glyko (Novato, CA). Endo-β-N-acetylglucosaminidase (Endo F containing <0.1% PNGase F; catalog no. 324706), 2,3-dehydro-2-deoxy-N-acetylneuraminic acid (DANA), and the sialic acid-specific lectin from Tritichomonas mobilensiswere purchased from Calbiochem (La Jolla, CA).Enzymatic Deglycosylation and Desialylation of ALSCharacterization of ALS GlycosylationPNGase F (5 units) and NANase III (25 milliunits) were used to remove N-linked sugars and sialic acids, respectively. Reactions contained [125I]ALS (4 × 105 cpm, ∼40 ng), 50 mm sodium phosphate buffer (pH 6.5), and 0.1% (w/v)n-octyl glucoside, and were incubated at 37 °C for 16 h. Identical reactions were set up without enzyme as controls. To investigate O-glycosylation, [125I]ALS (∼40 ng) was treated at 37 °C with 5 units of PNGase F with the addition of 12.5 milliunits of NANase III after 8 h, then 2 milliunits of O-glycosidase 3 h later. The mixture was subsequently incubated for an additional 12–16 h at 37 °C.Enzymatic Deglycosylation for IGFBP-3 Binding StudiesA series of increasingly deglycosylated [125I]ALS preparations was generated by incubation with 2.3, 11.7, 29.3, and 58.6 milliunits of Endo F/ng of [125I]ALS in 50 mmsodium phosphate buffer, pH 6.5, containing 0.0005% (w/v)n-octyl glucoside, for 16 h at 37 °C. For complete deglycosylation, [125I]ALS was treated with 640 milliunits of Endo F/ng of [125I]ALS in 0.0075% (w/v)n-octyl glucoside.Desialylated [125I]ALS was prepared using 0.035 milliunits of NANase III/ng of [125I]ALS in 50 mm sodium phosphate buffer (pH 6.5) at 37 °C for 16 h. In experiments where DANA was used, it was added at 1 nmol/milliunit NANase III to specifically inhibit the sialidase activity of NANase III.Site-directed Mutagenesis of ALSThe ALS cDNA was generated by reverse transcriptase-polymerase chain reaction from normal human liver total RNA as described previously (10Delhanty P. Baxter R.C. Biochem. Biophys. Res. Commun. 1996; 227: 897-902Crossref PubMed Scopus (14) Google Scholar). This cDNA was inserted into the BamHI/XbaI site of pSELECT (Promega Corp.) for mutagenesis according to the manufacturer's recommended protocol. Mutagenic deoxyoligonucleotides were used to generate seven different ALS cDNAs, each containing Asn to Ala codon substitutions at one of the seven consensusN-glycan linkage sites (Asn37 → Ala, 5′-CTGCAGCTCCAGGgcCCTCACGCGCCTG-3′; Asn58 → Ala, 5′-GCTGGACGGCAACgcCCTCTCGTCCGTC-3′; Asn69 → Ala, 5′-GGCAGCCTTCCAGgcCCTCTCCAGCCTG-3′; Asn341 → Ala, 5′-CGTGGCGGTCATGgcCCTCTCTGGGAAC-3′; Asn448 → Ala, 5′-CCTCAGCCTCAGGgcCAACTCACTGCGG-3′; Asn527 → Ala, 5′-CTTCGCCCTGCAGgcCCCCAGTGCTGTG-3′; Asn553 → Ala, 5′-CCCGCGTACACCTACAACgcCATCACCTG-3′; substituted nucleotides are shown in lowercase). Oligonucleotides were synthesized using an Oligo 1000 DNA Synthesizer (Beckman, Palo Alto, CA). The mutations were confirmed by sequencing (T7 sequencing kit; Amersham Pharmacia Biotech), and then the cDNAs were excised from pSELECT and inserted into the NheI/SalI site of pMSG (Amersham Pharmacia Biotech), an expression vector that contains the constitutively active and glucocorticoid-inducible murine mammary tumor virus promoter.Cell Culture and TransfectionsChinese hamster ovary (CHO) cells were transfected with either wild-type or mutant ALS expression constructs, or pMSG alone using the Polybrene/Me2SO technique (19Chaney W.G. Howard D.R. Pollard J.W. Sallustio S. Stanley P. Somat. Cell Mol. Genet. 1986; 12: 237-244Crossref PubMed Scopus (99) Google Scholar). pMSG contains the guanine phosphoribosyltransferase gene, which confers resistance to mycophenolic acid. Stable transfectants were selected for 3 weeks in α-MEM supplemented with 10% fetal calf serum, 25 μg/ml mycophenolic acid, 2 μg/ml aminopterin, 250 μg/ml xanthine, 15 μg/ml hypoxanthine, and 10 μg/ml thymidine. Some ALS transfectants were cultured from single foci to produce clonal lines. Flasks of stably transfected cells were grown to confluence in α-MEM supplemented with 10% fetal calf serum, then the medium was changed to α-MEM supplemented with 10 μm dexamethasone. After 3–4 days, the conditioned medium was collected and cell debris was removed by centrifugation. Supernatants were then concentrated and equilibrated into 50 mm sodium phosphate buffer (pH 6.5) using Centricon 30 microconcentrators. Conditioned media from CHO cells transfected with pMSG alone were concentrated to the same extent as media containing the lowest concentration of mutant ALS, and used as controls.Electrophoretic AnalysesSDS-PAGE AnalysisRadioiodinated ALS (5,000 cpm, ∼0.5 ng) in Laemmli buffer was loaded, without heat treatment, onto 7.5% Ready gels (Bio-Rad) and electrophoretically separated under nonreducing conditions. The gels were then dried and exposed to Hyperfilm MP (Amersham, Bucks, UK) overnight at −80 °C.Isoelectric Focusing of ALSPreparations of [125I]ALS (10,000 cpm, ∼1 ng), untreated or treated with NANase III, were added to sample buffer containing 4% (w/v) CHAPS (BDH Ltd, Poole, UK), 100 mm Tris-HCl, pH 7.2, and 2% (v/v) ampholytes (Pharmalyte 3–10, Amersham Pharmacia Biotech). Samples were loaded onto the alkaline end of an immobilized pH gradient strip (pH 3–10, Amersham Pharmacia Biotech) that had been rehydrated overnight in 0.5% (w/v) CHAPS, 10 mm dithiothreitol, 6m urea, and 2% (v/v) Pharmalyte 3-10. Isoelectric focusing was performed in a Multiphor II electrophoresis unit (Amersham Pharmacia Biotech). A broad pI isoelectric focusing calibration kit (pH 3–10; Amersham Pharmacia Biotech) was run in parallel with the ALS samples. The theoretical pI for the human ALS protein backbone was calculated using the program ISOELECTRIC from the Genetics Computer Group, Inc. (Madison, WI).Binding AssaysGel Filtration Studies with Mutated ALSSize fractionation chromatography was used to determine whether the seven site-directed mutant recombinant ALS species lacking individual consensusN-glycan linkage sites were able to form ternary complexes (18Baxter R.C. J. Clin. Endocrinol. Metab. 1988; 67: 265-272Crossref PubMed Scopus (202) Google Scholar). Briefly, cross-linked [125I]IGF-I·IGFBP-3 (1 × 105 cpm) was incubated for 30 min at 25 °C with conditioned medium containing 150 ng of mutant ALS equilibrated in 50 mm sodium phosphate buffer (pH 6.5) containing 1% (w/v) BSA, in a total volume of 200 μl. The mixture was then injected into a Superose-12 column (Amersham Pharmacia Biotech), and 0.5-ml fractions of eluate were collected at a flow rate of 1 ml/min. The degree of conversion from binary to ternary complex was evaluated by the shift of radioactivity from fractions corresponding to 50 kDa to those corresponding to 140 kDa (18Baxter R.C. J. Clin. Endocrinol. Metab. 1988; 67: 265-272Crossref PubMed Scopus (202) Google Scholar).Lectin Solution Binding AssayA lectin solution binding assay was used to indicate the presence or absence of sialic acids on [125I]ALS after enzymatic desialylation with NANase III and deglycosylation with PNGase F as described above. Identical reactions were set up without enzymes as controls. The assay was modified from that of Abidi et al. (20Abidi F.E. Bishayee S. Bachhawat B.K. Bhadra R. Anal. Biochem. 1987; 166: 257-266Crossref PubMed Scopus (5) Google Scholar). Briefly, approximately 35,000 cpm [125I]ALS (∼3.5 ng), either treated with enzyme or untreated control, was incubated for 1 h with 2 μg of the sialic acid-specific lectin from T. mobilensis in 50 mm sodium phosphate buffer with 0.01% (w/v) BSA, pH 6.5 at 22 °C (final volume 100 μl). [125I]ALS complexed to the lectin was then precipitated using γ-globulin (35 μl) and 6% polyethylene glycol (1 ml) for 10 min at 4 °C. Both BSA and γ-globulin were acid-hydrolyzed and dialyzed against water to remove contaminating sialic acids (21Dorai D.T. Bachhawat B.K. Bishayee S. Anal. Biochem. 1981; 115: 130-137Crossref PubMed Scopus (10) Google Scholar). The tubes were then spun at 3500 rpm in a swing bucket centrifuge for 10 min at 22 °C. The radioactive pellet in each tube was measured as a percentage of the total radioactivity added and these data were used to generate histograms. Nonspecific binding was determined to be the radioactivity measured when no lectin was added to the tube during the 1-h incubation. Control and desialylated forms of [125I]ALS gave nonspecific binding ranging from 10% to 14% of total, and the PNGase F deglycosylated forms of [125I]ALS gave approximately 28% of total.Solution Binding Assay and Scatchard AnalysisSolution binding assays were carried out as described previously (16Baxter R.C. Bayne M.L. Cascieri M.A. J. Biol. Chem. 1992; 267: 60-65Abstract Full Text PDF PubMed Google Scholar). Briefly, 10,000 cpm [125I]ALS, either treated with enzyme or untreated control, was incubated for 2 h with 10 ng of IGF-I or -II and a range from 0 to 10 ng of IGFBP-3 in 50 mm sodium phosphate buffer, pH 6.5, at 22 °C (final volume 0.3 ml). ALS complexed to IGFBP-3 was then precipitated using IGFBP-3 antiserum. The radioactivity in each tube was measured as a percentage of the total radioactivity added, and these data were used to generate binding curves. Endo F (58.3 milliunits/ng ALS) was added to a control untreated [125I]ALS preparation during the 2-h incubation to ensure that the presence of the enzyme did not adversely affect complex formation. Nonspecific binding was calculated as the percentage of radioactivity present after precipitation when there was no IGFBP-3 in the reaction mixture. For the Endo F deglycosylation experiments, nonspecific binding ranged from 3% to 18% of total for the partially deglycosylated forms and was approximately 30% of total for the fully deglycosylated forms. For the the NANase III binding curves, nonspecific binding was between 3% and 8% of the total. Scatchard analysis was carried out as described previously (16Baxter R.C. Bayne M.L. Cascieri M.A. J. Biol. Chem. 1992; 267: 60-65Abstract Full Text PDF PubMed Google Scholar), except that IGF-I, IGF-II, and IGFBP-3 were held constant at 1 ng/0.3 ml. ALS was added over the range of 0–200 ng/0.3 ml.Statistical AnalysesBinding curve data were analyzed by repeated measures analysis of variance, followed by Fisher's protected least significant difference test, using Statview 4.02 (Abacus Concepts Inc., Berkeley, CA). The value was considered significant if thep value was less than 0.05.DISCUSSIONThe interaction of ALS with IGF·IGFBP-3 complexes in the serum is believed to regulate both the function and the stability of the bound IGFs and IGFBP-3. Previous reports indicate that the affinity of IGFBP-3 for its ligands may be affected by post-translational modifications such as limited proteolysis (25Baxter R.C. Skriver L. Biochem. Biophys. Res. Commun. 1993; 196: 1267-1273Crossref PubMed Scopus (29) Google Scholar, 26Bang P. Brismar K. Rosenfeld R.G. J. Clin. Endocrinol. Metab. 1994; 78: 1119-1127Crossref PubMed Scopus (92) Google Scholar, 27Baxter R.C. Suikkari A.M. Martin J.L. Biochem. J. 1993; 294: 847-852Crossref PubMed Scopus (19) Google Scholar). In this study, we found that the carbohydrate chains on ALS play an influential role in determining its interaction with IGFBP-3. Our data therefore suggest that modifications to ALS as well as IGFBP-3 may be important in fine-tuning the bioavailability of the IGFs.In addition to the seven putative N-linked carbohydrate attachment sites (NXS/T) in the ALS sequence, there are two putative O-linked glycan sites. The first, at Ser60, conforms to a mucin-type O-glycosylation site as predicted by NetOglyc 2.0 (28Hansen J.E. Lund O. Tolstrup N. Gooley A.A. Williams K.L. S. B. Glycoconj. J. 1998; 15: 115-130Crossref PubMed Scopus (448) Google Scholar) and is part of theN-linked carbohydrate attachment site at Asn58. The second, FT494PQP, corresponds to theXTPXP sequence recently described as the minimal requirement for O-linked carbohydrates (29Yoshida A. Suzuki M. Ikenaga H. Takeuchi M. J. Biol. Chem. 1997; 272: 16884-16888Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar) and lies adjacent to the putative N-linked glycosylation site, Asn448. Treatment of ALS with O-glycosidase, which hydrolyzes Galβ(1–3)GalNAc, a common core structure of manyO-linked glycans (30Umemoto J. Bhavanandan V.P. Davidson E.A. J. Biol. Chem. 1977; 252: 8609-8614Abstract Full Text PDF PubMed Google Scholar), had no effect on the molecular mass of ALS as judged by SDS-PAGE analysis. While it is possible that fucose residues may have interfered with the enzyme, or that a shift in size was below the level of detection, the putative O-linked sites do not appear to be occupied by sugar side chains with the common core structure described above. Furthermore, a lectin binding assay that was specific for sialic acids, a common residue onO-linked carbohydrates, also failed to provide evidence forO-linked sugars on ALS.Therefore, we focused on the N-linked carbohydrates of ALS and their role in ternary complex formation. Enzymatic removal of theN-linked sugars from ALS decreased its ability to form the complex with IGFBP-3 in a manner that was related to the level of deglycosylation. Intermediate levels of ALS deglycosylation, probably involving the loss of more than a single sugar chain, measurably disrupted ternary complex formation, whereas complete removal of theN-linked glycans abolished complex formation. However, when site-directed mutagenesis was used to mutate each of theN-linked attachment sites individually, no single glycan appeared to have a major impact on ALS binding activity. Therefore, the carbohydrates on ALS clearly influence the affinity that ALS has for the IGF·IGFBP-3 complex, although this influence relies on a number of N-linked sugars rather than any single chain. However, it is not clear whether deglycosylated ALS is less able to form the ternary complex because of conformational changes in ALS induced by removal of the glycans or disruption of interactions between the ALS carbohydrates and the other two proteins.Sialic acids are common anionic residues that can be attached to bothN- and O-linked sugars. They can contribute significant charge to glycoproteins as many sialic acids can be attached to one highly branched N-linked sugar. For example, the acute-phase protein α1-acid glycoprotein has a pI of 2–3 mostly due to the large number of sialic acids attached to its highly branched complex N-linked sugars (31Van Dijk W. Havenaar E.C. Brinkman-van der Linden E.C.M. Glycoconj. J. 1995; 12: 227-233Crossref PubMed Scopus (133) Google Scholar). The calculated pI of ALS, based on its amino acid sequence, is 6.56; however, we observed six discrete isoforms of pI 4.6–5.3 by isoelectric focusing. After NANase III treatment, the pI of ALS increased to between 5.5 and 7, close to the predicted value for the amino acid backbone. The contribution of negative charge by the sialic acid may explain, in part, the high affinity with which ALS binds to weak anion exchange columns used in ALS purification (3Baxter R.C. Martin J.L. Beniac V.A. J. Biol. Chem. 1989; 264: 11843-11848Abstract Full Text PDF PubMed Google Scholar). Given that the sialic acid contributes significantly to the negative charge on ALS and that the interactions within the ternary complex are dependent on charge, it could be predicted that the sialic acid on ALS would affect the formation of the complex. Indeed, ALS treated with NANase III to remove sialic acid displayed a 50–80% decrease in affinity for the IGF-I and IGF-II binary complexes compared with that of untreated ALS. However, it is noteworthy that desialylation only lowered the affinity of ALS for the IGF·IGFBP-3 complex unlike deglycosylation, which abolished complex formation. This suggests that the effects of ALS glycosylation on IGFBP-3 binding are not entirely due to the negative charges imparted by sialic acid.The results from the N-linked glycan studies imply that a number of N-linked sugars are required to act in concert to enable ALS to interact with IGFBP-3 and the IGFs, rather than a specific carbohydrate chain being solely responsible. The placement of the glycans within the tertiary structure of ALS may shed light on this finding. In unpublished studies, 2J. B. M. Janosi, P. A. Ramsland, M. Mott, S. M. Firth, R. C. Baxter, and P. J. D. Delhanty, unpublished data. we have modeled the central leucine-rich repeat region of ALS on the only published crystal structure of a leucine-rich repeat protein, the porcine ribonuclease inhibitor (32Kobe B. Deisenhofer J. Nature. 1993; 366: 751-756Crossref PubMed Scopus (541) Google Scholar). If the modeled structure is a true representation of ALS, then six out of the seven potential N-linked sugar attachment sites lie very close to each other, suggesting a possible clustering of carbohydrate chains. Therefore, the loss of any singleN-linked glycan may be compensated by the potentially large number of other carbohydrates in the vicinity. This model also has implications with regard to the sialic acid moieties that we demonstrated to exist on ALS. A lectin solution binding assay suggested that all the sialic acids on ALS are attached to theN-linked sugars. These sialic acid moieties might therefore result in a region of negative charge where the N-linked sugars are concentrated.The binding affinity of ALS to the IGF·IGFBP-3 complex is relatively weak, 1–2 orders of magnitude less than the affinity of IGFBP-3 for either of the IGFs (13Holman S.R. Baxter R.C. Growth Regul. 1996; 6: 42-47PubMed Google Scholar), and in this sense ALS binding is the limiting step in ternary complex formation. This suggests that modulation of ALS affinity might directly influence complex formation, and thus the bioavailability of IGFs. Isoelectric focusing indicates that ALS purified from normal serum already exists as a number of differently sialylated isoforms. If the degree of ALS sialylation is subject to physiological regulation, as described for other proteins (31Van Dijk W. Havenaar E.C. Brinkman-van der Linden E.C.M. Glycoconj. J. 1995; 12: 227-233Crossref PubMed Scopus (133) Google Scholar), the potential exists for modulation of formation or stability of the ternary complexes. This might occur in addition to the previously described modulation of the circulating concentration of the ALS protein itself (33Delhanty P.J.D. Takano K. Hizuka N. Takahashi S.I. The Regulation and Actions of ALS: Molecular Mechanisms to Regulate the Activities of Insulin-like Growth Factors. Elsevier Science B. V., Amsterdam1998: 135-143Google Scholar) through cytokine suppression (34Delhanty P.J.D. Biochem. Biophys. Res. Commun. 1998; 243: 269-272Crossref PubMed Scopus (32) Google Scholar) or in patients who are critically ill (35Baxter R.C. Hawker F.H. To C. Stewart P.M. Holman S.R. Growth Regul. 1997; 7: 1-10Google Scholar) or have hepatic cirrhosis (36Donaghy A.J. Baxter R.C. Baillieres Clin. Endocrinol. Metab. 1996; 10: 421-446Abstract Full Text PDF PubMed Scopus (38) Google Scholar).In summary, we have found that the N-linked carbohydrates on ALS are a requirement for the formation of complexes with IGFBP-3 and IGFs. We have also shown that the removal of sialic acids from ALS significantly reduces the affinity of ALS for binding to IGFBP-3. Since the glycosylation of secreted proteins are often modified in certain physiological and pathological states, the modification of ALS glycosylation has the potential to be an important factor in the regulation of IGF access to the tissues. Insulin-like growth factors (IGF)1 I and II are peptide hormones that regulate the differentiation and p
Referência(s)