Biosynthesis of rat insulin-like growth factor II. II. Localization of mature rat insulin-like growth factor II (7484 daltons) to the amino terminus of the approximately 20-kilodalton biosynthetic precursor by radiosequence analysis

In previous studies, we have identified possible biosynthetic precursors of rat insulin-like growth factor II (rIGF-II) using specific immunoprecipitation, approximately 22-kDa prepro-rIGF-II and 20-kDa pro-rIGF-II. We now provide chemical evidence that amino acid sequences corresponding to mature 7484-dalton rIGF-II are present at the NH2 terminus of the putative approximately 20-kDa pro-rIGF-II. BRL-3A cultures have been labeled individually with several radioactive amino acid precursors, the cells have been lysed, and the lysates have been immunoprecipitated with antiserum to rIGF-II. Following electrophoresis of the immunoprecipitated proteins, labeled approximately 20-kDa pro-rIGF-II was eluted from the gels and subjected to automated radiosequence analysis. Discrete peaks of radioactivity were observed in 12 of the first 30 cycles of Edman degradation. The deduced partial amino acid sequence was identical at each position with that of mature 7484-dalton rIGF-II. These results directly demonstrate that mature rIGF-II sequences are present in the approximately 20-kDa protein, as required if the approximately 20-kDa protein were pro-rIGF-II. In addition, they localize the 7484-dalton rIGF-II to the NH2 terminus of the precursor molecule. A second NH2-terminal sequence differing only in the absence of the NH2-terminal residue, alanine, also was present in an approximately equal amount. Similar NH2-terminal heterogeneity has been reported for 7484-dalton rIGF-II and most likely reflects ambiguity in the cleavage sites for the signal peptidase.     

Nucleotide sequence and expression of a cDNA clone encoding a fetal rat binding protein for insulin-like growth factors

The insulin-like growth factors (IGFs), IGF-I and IGF-II, occur in plasma and tissue fluids complexed to specific binding proteins. Although the role of the binding proteins is not completely defined, they are capable of modulating the biological activity of the IGFs. In order to better understand the function of these proteins, we have isolated a clone from the BRL-3A rat liver cell line that encodes a protein corresponding to the IGF binding protein in fetal rat serum. The cDNA clone encodes a precursor protein of 304 amino acids (32,886 daltons), comprised of a 34-residue hydrophobic prepeptide and a 270-residue mature protein (29,564 daltons). The deduced amino acid sequence agrees with the sequence of 173 amino acid residues determined by Edman degradation. The mature protein contains 18 cysteines and no N-glycosylation sites. It contains an Arg-Gly-Asp (RGD) sequence near the carboxyl terminus. A similar sequence is present on many extracellular matrix proteins and contributes to their recognition by cellular adhesion receptors. The cloned cDNA has been transcribed in vitro and the resulting RNA expressed in Xenopus oocytes. Injected oocytes secrete a 33-kDa protein that is immunoprecipitated by polyclonal antibodies to the BRL-3A binding protein and binds IGF-I and IGF-II with the same affinity and specificity as does purified BRL-3A binding protein. The binding protein cDNA probe hybridizes to an approximately 2-kilobase mRNA in BRL-3A cells and in multiple fetal rat tissues including liver, kidney, intestine, and lung. Levels of this mRNA are greatly reduced in the corresponding adult tissues. The rat IGF binding protein is closely related to the partial amino acid sequences reported for a bovine IGF binding protein and more distantly related to a human IGF binding protein that recently has been cloned. No significant homologies were identified to other proteins. Thus, the rat IGF binding protein that we have cloned appears to be a distinct member of a family of related IGF binding proteins. We postulate that the structurally distinct IGF binding proteins may have different biological functions.     

Purification of the insulin-like growth factor II (IGF-II) receptor from an IGF-II-producing cell line, and generation of an antibody which both immunoprecipitates and blocks the type 2 IGF receptor

18,54-SF cells, which secrete rat insulin-like growth factor II (rIGF-II), have abundant type 2 IGF receptors. We have purified the type 2 receptor from these cells by solubilization of crude membranes in Triton X-100, followed by chromatography on agarose-immobilized rIGF-II. A partially purified receptor preparation, obtained by chromatography of solubilized membranes over wheat germ agglutinin, was used to immunize a rabbit. The antibody generated both immunoprecipitates the type 2 receptor, and specifically inhibits IGF-II binding to a variety of rat tissues, including 18,54-SF cells, BRL-3A cells and placenta. The presence of abundant type 2 receptors on an rIGF-II-secreting cell line is consistent with an autocrine role for IGF-II in select cells.     

Extracellular release as the major degradative pathway of the insulin-like growth factor II/mannose 6-phosphate receptor.

The presence of a soluble, truncated form of the IGF-II/Man-6-P receptor in serum has suggested that cleavage from the cell surface may be an initial step in the degradation of this protein (MacDonald, R. G., Tepper, M. A., Clairmont, K. B., Perregaux, S. B., and Czech, M. P. (1989) J. Biol. Chem. 264, 3256-3261). In order to test this hypothesis, we pulse-labeled cultured BRL-3A rat liver cells with [35S]methionine and [35S]cysteine and measured the fate of labeled receptor at various times after incubation with unlabeled amino acids. It was found that the appearance of labeled IGF-II/Man-6-P receptor in the medium accounts quantitatively for the loss of labeled receptor from the BRL-3A cells. In similar experiments with Chinese hamster ovary cells, L6 rat myoblasts, and chick embryo fibroblasts, labeled receptor from the cell membranes decreases with a time course corresponding to the appearance of soluble receptor in the medium. The release of labeled receptor into the medium can be blocked by the addition of the protease inhibitors aprotinin, chymostatin, or phenylmethylsulfonyl fluoride, but not antipain, leupeptin, and benzamidine. The results are consistent with the hypothesis that the degradation and loss of cellular IGF-II/Man-6-P receptors occurs by a nonlysosomal mechanism involving their proteolysis and removal into the extracellular fluid.     

Evidence for the phosphorylation of the type II insulin-like growth factor receptor in cultured cells

The ATP pools of monolayer cultures of rat embryo fibroblasts and rat liver cells (BRL-3A2) were labeled with [32P]H3PO4. The type II insulin-like growth factor (IGF) receptor was purified by affinity chromatography on wheat germ lectin-Sepharose and IGF-II-Sepharose columns. A phosphorylated species having the expected size of the type II receptor (Mr = 220,000 without reduction, Mr = 260,000 with reduction) was identified by sodium dodecyl sulfate polyacrylamide gel electrophoresis and autoradiography. IGF-II stimulated phosphorylation of the type II receptor in BRL-3A2 rat liver cells. Lability of the receptor phosphate bonds to alkaline pH suggests that the bulk of phosphorylation was occurring on serine residues.     

Production of monoclonal antibodies to the rat insulin-like growth factor II (IGF-II) receptor

BALB/c mice were immunized with rIGF-II receptors purified from 18-54, SF cells by chromatography of solubilized receptors over agarose-immobilized rIGF-II. Two fusions of splenic lymphocytes with FO mouse myeloma cells yielded 27 stable hybrids which were positive by ELISA. Cloning of seven of these hybrids yielded 30 positive clones by ELISA. At least seven of these clones (minimum of one from each parent hybrid) were capable of specifically immunoprecipitating the rIGF-II receptor.     

Serum and growth factors rapidly elicit phosphorylation of the Ca2+/calmodulin-dependent protein kinase II in intact quiescent rat 3Y1 cells

A 50-kDa protein was recognized in rat embryo fibroblast 3Y1 cells with an affinity-purified antibody against rat brain Ca2+/calmodulin-dependent protein kinase II (CaM kinase II). When the cytosolic extract from quiescent 3Y1 cells was immunoprecipitated with the antibody, the 50-kDa protein in the immunoprecipitate became phosphorylated in a Ca2+- and calmodulin-dependent manner following exposure to [gamma-32P]ATP. Moreover, the reaction proceeded through an intramolecular mechanism. These results suggest that the 50-kDa protein is a subunit of CaM kinase II in rat 3Y1 cells. The addition of 10% fetal calf serum to quiescent 3Y1 cells caused a rapid increase in the phosphorylation of the 50-kDa protein, which was immunoprecipitated with the affinity-purified anti-CaM kinase II antibody. The phosphorylation of CaM kinase II was detected as early as 20 s after the addition of serum, reached the maximal level at 2 min, and decreased to the basal level within 60 min. Platelet-derived growth factor and epidermal growth factor also elicited the phosphorylation of the 50-kDa protein in quiescent 3Y1 cells, while neither insulin nor 12-O-tetradecanoylphorbol-13-acetate did. Calcium ionophores, A23187 and ionomycin, also caused the phosphorylation of the protein in 3Y1 cells. Moreover, phosphopeptide mappings of the phosphorylated 50-kDa subunit generated in response to serum, EGF, and A23187 yielded patterns similar to that generated from the immunoprecipitated 50-kDa subunit phosphorylated in vitro. Phosphoamino acid analysis of the phosphorylated subunit demonstrated that serine residue was the major amino acid labeled under any condition. These results suggest that CaM kinase II undergoes phosphorylation in response to various stimuli that can increase the free Ca2+ concentration in the cytoplasm of quiescent fibroblast cells and therefore probably mediates at least some of the biological actions of growth factors.     

Characterization of insulin-like growth factor binding proteins of cultured rat astroglial and neuronal cells

Insulin-like growth factor binding proteins produced by cultured rat neurons, astrocytes, and rat cell lines BRL-3A and B104 were compared to binding proteins found in rat serum, using affinity labeling, deglycosylation, and Western ligand blotting studies. Each source elaborated an unique pattern of heterogeneous binding proteins. Some of the binding proteins from different sources behaved similarly in each experimental system suggesting that subsets of these binding proteins may be structurally related. In particular, our data suggest that cultured astrocytes and neurons make the major binding protein produced by BRL-3A cells.     

Stimulation of the phosphorylation of cytoskeletal 350-kDa and 300-kDa proteins by insulin-like growth factor-I, platelet-derived growth factor and phorbol ester in rat 3Y1 cells

Insulin-like growth factor-I (IGF-I) stimulated the phosphorylation of cytoskeletal 350-kDa and 300-kDa proteins which were immunoprecipitated with antibodies against brain high molecular weight microtubule-associated proteins in quiescent rat 3Y1 cells. The data on the effective concentrations of IGF-I and 125I-labeled IGF-I binding indicated that type I IGF receptors mediate this IGF-I effect. Platelet-derived growth factor (PDGF) as well as phorbol ester (TPA) also stimulated the phosphorylation of these proteins. These proteins, whether immunoprecipitated from cells stimulated by insulin, IGF-I, TPA, PDGF, or epidermal growth factor, produced very similar phosphopeptide mapping patterns irrespective of the stimulant. The results suggest the possibility that these growth factors and phorbol esters may activate a common protein kinase which is responsible for the phosphorylation of the 350-kDa and 300-kDa proteins in cells.     

Production and characterization of mammary-derived growth factor 1 in mammary epithelial cell lines.

A mammary-derived growth factor, MDGF1, which stimulates collagen synthesis and proliferation in mammary epithelial cells was previously detected and purified from human milk and primary human breast tumors. MDGF1 binds to putative cell-surface receptors of 120-140 kDa and stimulates proliferation of normal and malignant human mammary epithelial cells. Partial protein sequence (N-terminal 18 amino acid sequence) shows that MDGF1 has no homology to any other known growth-promoting peptides. Polyclonal antiserum raised against this synthetic peptide recognizes native milk-derived MDGF1. We hypothesize that MDGF1 might be an autocrine or paracrine factor produced by and acting on normal and malignant human breast epithelial cells possessing MDGF1 receptors. As a first step in testing this possibility, we examined whether human breast epithelial cells in culture produce the growth factor. A protein with the size of MDGF1 was immunologically detected in the concentrated conditioned medium prepared from human breast cancer cell line MDA-MB 231, the mammary-derived but nontumorigenic HBL-100 line, and the normal reduction mammoplasty-derived, nonimmortalized 184 cell strain. A competitive radioreceptor assay (RRA) was used to estimate the level of MDGF1 in the conditioned medium. MDGF1 was present in the nanogram range per 1 million cells. A 62-kDa protein was detected in the above cell lysates by Western immunoblotting or by immunoprecipitation of metabolically labeled cell-conditioned media. The polyclonal antisera directed against the 18 amino acid peptide sequence from milk-derived MDGF1 could adsorb MDGF1 biological activity from conditioned medium. In vitro translation of cell mRNA yielded a protein of 55 kDa which was immunoprecipitated by anti-MDGF1 antibody.(ABSTRACT TRUNCATED AT 250 WORDS)     

The E-domain peptide of rat pro-insulin-like growth factor II (proIGF-II): properties of the peptide in serum and production by rat cell lines

We previously identified a naturally occurring peptide fragment derived from the carboxyl terminal region of the E-domain of pro-insulin-like growth factor II (proIGF-II117-156) in medium conditioned by cultured BRL-3A rat liver cells. In the present study we utilized a radioimmunoassay (RIA) for this peptide to measure physiological concentrations of the peptide in media and serum. Serum levels of the E-domain peptide were very high in the 5-day neonatal rat and declined thereafter to reach low levels in adult rat serum. Chromatography of adult rat serum on Sephadex G-75 in 1 M acetic acid yielded a single broad peak of E-peptide immunoreactivity that coeluted with a synthetic E-peptide standard. However, chromatography of day 5 neonatal rat serum on Sephadex G-75 yielded two peaks of immunoreactivity. One of the peaks coeluted with a synthetic E-peptide standard, whereas the other peak eluted in a region where higher molecular weight proteins typically elute. Experiments aimed at determining whether adult rat serum contained a binding protein for the E-domain peptide revealed that: (1) serum contains little, if any, binding protein for the E-domain peptide, (2) serum contains a proteinase activity that degrades the E-domain peptide, and (3) the proteinase activity can be eliminated by acetic acid/ethanol extraction. Of several rat cell lines tested (BRL-3A, rat embryo fibroblasts (REF), hepatoma cell lines (H4, HTC), GH3 pituitary tumor cells, and normal rat kidney fibroblasts (NRK], only BRL-3A and REF cells secreted measurable E-domain peptide into the medium. In addition, it was found that some component(s) of serum could stimulate secretion of E-domain peptide from BRL-3A and REF cells. Chromatography of the immunoreactivity from BRL-3A and REF-conditioned media on Sephadex G-75 in 1 M acetic acid yielded a single peak that coeluted with a synthetic E-domain peptide standard. Since secretion of the E-domain peptide parallels the expression of IGF-II, the RIA for the proIGF-II E-domain peptide may be useful for studies of the biosynthesis and secretion of IGF-II under different physiological conditions. The RIA for the IGF-II E-domain peptide has two technical advantages over the RIA for IGF-II, namely, the lack of interference by IGF binding proteins and the relative ease with which large quantities of pure antigen can be synthesized.     

Photoaffinity labeling with a neuroactive steroid analogue. 6-azi-pregnanolone labels voltage-dependent anion channel-1 in rat brain

Neuroactive steroids modulate the function of gamma-aminobutyric acid, type A (GABA(A)) receptors in the central nervous system by an unknown mechanism. In this study we have used a novel neuroactive steroid analogue, 3 alpha,5 beta-6-azi-3-hydroxypregnan-20-one (6-AziP), as a photoaffinity labeling reagent to identify neuroactive steroid binding sites in rat brain. 6-AziP is an effective modulator of GABA(A) receptors as evidenced by its ability to inhibit binding of [(35)S]t-butylbicyclophosphorothionate to rat brain membranes and to potentiate GABA-elicited currents in Xenopus oocytes and human endothelial kidney 293 cells expressing GABA(A) receptor subunits (alpha(1)beta(2)gamma(2)). [(3)H]6-AziP produced time- and concentration-dependent photolabeling of protein bands of approximately 35 and 60 kDa in rat brain membranes. The 35-kDa band was half-maximally labeled at a [(3)H]6-AziP concentration of 1.9 microM, whereas the 60-kDa band was labeled at higher concentrations. The photolabeled 35-kDa protein was isolated from rat brain by two-dimensional PAGE and identified as voltage-dependent anion channel-1 (VDAC-1) by both matrix-assisted laser desorption ionization time-of-flight and ESI-tandem mass spectrometry. Monoclonal antibody directed against the N terminus of VDAC-1 immunoprecipitated labeled 35-kDa protein from a lysate of rat brain membranes, confirming that VDAC-1 is the species labeled by [(3)H]6-AziP. The beta(2) and beta(3) subunits of the GABA(A) receptor were co-immunoprecipitated by the VDAC-1 antibody suggesting a physical association between VDAC-1 and GABA(A) receptors in rat brain membranes. These data suggest that neuroactive steroid effects on the GABA(A) receptor may be mediated by binding to an accessory protein, VDAC-1.     

Stimulation by insulin-like growth factors is required for cellular transformation by type beta transforming growth factor

Medium conditioned by BRL-3A cells, a known source of insulin-like growth factor II (IGF-II), induced phenotypic transformation (anchorage-independent proliferation) of mouse BALB/c 3T3 fibroblasts but not rat NRK-49F fibroblasts, in the presence of 10% calf serum. A specific radioreceptor assay and a bioassay indicated that BRL-3A conditioned medium contained 0.5-1 ng/ml of type beta transforming growth factor (beta TGF). Purified IGF-II and beta TGF acting together reconstituted the transforming activity of BRL-3A conditioned medium on BALB/c 3T3 cells. Insulin was 5-10% as potent as IGF-II in supporting the transforming action of beta TGF on BALB/c 3T3 cells. NRK-49F cells were phenotypically transformed by beta TGF in the presence of EGF and 10% calf serum as the sole source of IGFs. However, transformation of NRK-49F cells under these conditions was inhibited by addition of purified IGF-binding protein. Addition of an excess of IGF-II prevented the inhibitory action of IGF-binding protein. The different sensitivity of the two cell lines to IGFs was correlated with lower levels of type I IGF receptor and higher levels of type II IGF receptor in NRK-49F cells as compared with BALB/c 3T3 cells. The results suggest that cellular stimulation by IGFs is a prerequisite for transformation of rodent fibroblasts by beta TGF. We propose that transformation of fibroblasts by beta TGF requires concomitant stimulation by the set of growth factors that support normal cell proliferation.     

A phosphatidylinositol-3 kinase binds to platelet-derived growth factor receptors through a specific receptor sequence containing phosphotyrosine.

Platelet-derived growth factor (PDGF) stimulates autophosphorylation of the PDGF receptor and association of the receptor with several cytoplasmic molecules, including phosphatidylinositol-3 kinase (PI3 kinase). In this study we examined the association of PI3 kinase with immunoprecipitated autophosphorylated PDGF receptor in vitro. The PI3 kinase from cell lysates bound to the wild-type receptor but not to a mutant receptor that had a deletion of the kinase insert region. A protein of an apparent size of 85 kDa bound to the receptor, consistent with previous observations that a protein of this size is associated with PI3 kinase activity. In addition, 110- and 74-kDa proteins bound to the phosphorylated receptor. Dephosphorylated receptors lost the ability to bind PI3 kinase activity as well as the 85-kDa protein. A 20-amino-acid peptide composed of a sequence in the kinase insert region that included one of the autophosphorylation sites of the receptor (tyrosine 719) as well as a nearby tyrosine (Y708) blocked the binding of PI3 kinase to the receptor, but only when the peptide was phosphorylated on tyrosine residues. A scrambled version of the peptide did not block PI3 kinase binding to the receptor even when it was phosphorylated on tyrosine. These tyrosine-phosphorylated peptides did not block binding of phospholipase C-gamma or GTPase-activating protein to the receptor. In separate experiments (receptor blots), soluble radiolabeled receptor bound specifically to an 85-kDa protein present in sodium dodecyl sulfate-polyacrylamide gel electrophoresis-fractionated 3T3 cell lysates that were transferred to nitrocellulose paper. The binding was blocked by the same tyrosine-phosphorylated peptides that prevented binding of PI3 kinase activity to immobilized receptors. These findings show that the PDGF receptor binds directly to an 85-kDa protein and to a PI3 kinase activity through specific sequences in the kinase insert region. The association of a 110-kDa protein with the receptor also involve these sequences, suggesting that this protein may be a subunit of the PI3 kinase. Phosphotyrosine is an essential structure required for the interactions of these proteins with the PDGF receptor.     

Estrogen regulation of tissue-specific expression of complement C3

The administration of estradiol to immature rats results in the increased synthesis and secretion of a 180-kDa protein, composed of 115- and 65-kDa subunits, by the uterine luminal epithelial cells. A monoclonal antibody against the 180-kDa protein was utilized to isolate the corresponding cDNA (LE-1) from a rat uterine luminal epithelial cell cDNA lambda gt11 expression library. This LE-1 cDNA was sequenced and shown to be homologous to complement component C3. The sequence was approximately 81 and 90% homologous to human and mouse C3, respectively. The LE-1 cDNA sequence was homologous with the 3' portion of the C3 mRNA containing the alpha subunit (115 kDa). Uterine mRNA isolated from immature rats treated with 1 microgram of estradiol for 24 h demonstrated a 25-fold increase in the concentration of a 6.0-kilobase mRNA by Northern hybridization with either LE-1 or authentic human C3 cDNA probes. To further examine the possibility that the estradiol-regulated secretory protein was C3, an aliquot of radiolabeled media protein from control and estradiol-stimulated rat uteri was incubated with goat anti-rat C3 antibody. The immunoprecipitated radiolabeled protein from estradiol-treated animals was increased significantly (p less than 0.01) compared to media from control animals. Analysis of the immunoprecipitated proteins on nonreducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed a single protein of 180 kDa from estradiol-stimulated uterine media, whereas no detectable proteins were immunoprecipitated from media obtained from control uteri. Also, when the immunoprecipitated protein was reduced (20 mM dithiothreitol) it dissociated into two subunits of 115 and 65 kDa. Immunohistochemical studies demonstrated the presence of C3 only in the epithelial cells of estrogen-stimulated rat uteri. In addition, the estradiol-stimulated mRNA was only detectable in uterine epithelial cell RNA. Analysis of liver RNA demonstrated a 6.0-kilobase mRNA, as in the uterus, when hybridized with LE-1. However, unlike the uterus, its concentration was not influenced by estrogen administration with up to three daily injections of 100 micrograms of diethylstilbestrol. Based on biophysical, DNA sequence, and antibody data we conclude that rat uterine epithelial cells produce C3 in response to estradiol whereas the expression in the liver was not modulated by estrogens.     

Insulin-like growth factor-I is internalized after binding to the type I insulin-like growth factor receptor

Receptor-mediated endocytosis may represent an important mechanism whereby peptide hormones exert their biological effects. The ability of recombinant insulin-like growth factor (IGF)-I to be internalized by cultured cells was evaluated in BRL-3A2 cells, a rat liver-derived cell line which lacks insulin receptors. Since recombinant IGF-I does not bind to the Type II IGF receptor, all specific binding of 125I-IGF-I in BRL-3A2 cells represents binding to the Type I receptor. Exposure of BRL-3A2 cells to IGF-I resulted in a rapid 50% downregulation of Type I IGF receptors. Only one-half of these binding sites were sensitive to treatment with trypsin, a phenomenon which indicates that the peptide and its receptor were internalized after the cells were exposed to IGF-I. In conclusion, these experiments demonstrate that IGF-I can be internalized by cultured cells via the Type I IGF receptor, and suggest that IGF hormone action may be exerted by receptor-mediated endocytosis.     

Purification and amino-terminal sequence of an insulin-like growth factor-binding protein secreted by rat liver BRL-3A cells

A protein preparation that specifically binds insulin-like growth factors (IGFs) I and II was purified from medium conditioned by rat liver BRL-3A cells using molecular sieve chromatography in 1 M acetic acid followed by affinity chromatography on IGF-II-agarose. The affinity-purified IGF-binding protein exhibits a single major band with apparent Mr = 36,300 under reducing conditions on sodium dodecyl sulfate-polyacrylamide gels. The IGF-binding protein is efficiently and specifically cross-linked to either 125I-IGF-I (human) or 125I-IGF-II (rat) using disuccinimidyl suberate. An IGF-binding protein of similar apparent molecular weight was also affinity purified from rat hepatoma H-35 cell conditioned medium and found to differ from the BRL-3A protein such that potent polyclonal antisera prepared in rabbits against the purified BRL-3A IGF-binding protein exhibited a much lower titer for the H-35 protein in an enzyme-linked immunosorbent assay and upon immunoblotting. In order to determine whether a single BRL-3A IGF-binding protein is present in the affinity-purified preparation, the protein was prepared for sequencing on a Sephacryl S-300 column in 6 M guanidine HCl after reduction and alkylation. The amino acid composition (expressed in percentages) of this IGF-binding protein was determined to be: Cys = 5.5, Lys = 4.8, His = 2.8, Arg = 7.8, Asx = 10.2, Thr = 5.1, Ser = 3.9, Glx = 15.7, Gly = 17.4, Ala = 7.3, Val = 4.6, Met = 1.4, Ile = 2.4, Leu = 8.3, Tyr = 1.0, Phe = 1.9. Sequencing of the NH2-terminal portion of this protein led to the identification of 31 amino acids in the following order: Phe-Arg-Cys-Pro-Pro-Cys-Thr-Pro-Glu-Arg-Leu-Ala-Ala-Cys-Gly-Pro-Pro-Pro- Asp-Ala-Pro-Cys-Ala-Glu-Leu-Val-Arg-Glu-Pro-Gly-Cys. We conclude that rat liver BRL-3A cells secrete a single major IGF-binding protein capable of binding both IGF-I and IGF-II.