Cloning and Expression of the Growth Hormone-Dependent Insulin-Like Growth Factor-Binding Protein

Cloning and Expression of the Growth Hormone-Dependent Insulin-Like Growth Factor-Binding Protein William I. Wood, George Cachianes, William J. Henzel, Genine A. Winslow, Steven A. Spencer, Renate Hellmiss, Janet L. Martin, and Robert C. Baxter Departments of Developmental Biology and Molecular Biology Genentech, Inc. South San Francisco, California 94080 Department of Endocrinology (J.L.M., R.C.B.) Royal Prince Alfred Hospital Camperdown, New South Wales 2050, Australia N-terminal as well as internal amino acid sequence data were obtained from the GH dependent, insulinlike growth factor (IGF) binding protein, BP-53, purified from human plasma. Based on these sequence data, full-length cDNA clones of BP-53 have been isolated, and the complete deduced sequence of BP-53 determined. This sequence contains a 27 amino acid putative signal sequence followed by a mature protein of 264 amino acids containing 18 cysteine residues clustered near the N- and C-terminus. The deduced protein sequence of BP-53 has 33% amino acid identity including conservation of all 18 cysteine residues with the recently cloned BP28, a smaller human IGF-binding protein identified in amniotic fluid and also secreted by the cell line HEP G2. Expression of the cloned BP-53 cDNA in mammalian tissue culture cells results in secretion of the protein into the culture medium. This expressed protein is identical to plasma-derived BP53 in its immunoreactivity, high affinity binding of IGF-I and IGF-II, and mobility on sodium dodecyl sulfate gel electrophoresis. (Molecular Endocrinology 2:1176-1185,1988) INTRODUCTION The insulin-like growth factors (IGF) or somatomedins, are generally found in the circulation and in cell culture medium associated with one or more binding proteins (1-3). Two distinct human IGF-binding proteins have been extensively characterized. One predominates in adult serum, where it carries most of the circulating IGFs (4-6). The other is found in highest concentrations 0888-8809/88/1176-1185$02.00/0 Molecular Endocrinology Copyright © 1988 by The Endocrine Society in amniotic fluid (7-10), but has also been identified in the circulation, apparently not fully occupied by IGF (5, 11-14), in medium conditioned by HEP-G2 human hepatoma cells (15), and in human placental tissue (16). The amniotic fluid protein has been designated BP-28 based on its apparent molecular mass [28 kilodalton (kDa)] on nonreduced sodium dodecyl sulfate (SDS) gel electrophoresis (8). It appears slightly larger (32-40 kDa) on gel permeation chromatography (7, 9) or reduced SDS gel electrophoresis (7,16). Purified binding protein, BP-28, binds IGF-I and IGF-II with similar affinities (K a = 3-7 nM"1) (8). Recently, cDNA clones of BP28 have been isolated which encode a cysteine-rich protein of 234 amino acids (17,18). IGF binding proteins of similar size to BP-28 have also been isolated from culture medium conditioned by the rat liver-derived cell line, BRL-3A (19-21), and from the bovine kidneyderived cell line, MDBK (22). Limited amino acid sequence data from these proteins suggest that while they have some sequence similarity to BP-28, they may in fact constitute a distinct class of IGF-binding proteins. The major IGF-binding protein in adult serum is found in greatly decreased concentrations in GH deficiency and in increased concentrations in acromegaly (4, 2 3 25); thus this protein has been referred to as the GHdependent IGF-binding protein. It circulates as a 125150 kDa glycoprotein complex which dissociates at low pH, releasing free IGFs and an acid-stable IGF-binding subunit (24, 26, 27). This binding subunit, of apparent molecular mass 53 kDa (nonreduced) and 43 kDa (reduced), has been designated BP-53 (8, 28). The circulating complex also contains an acid-labile subunit which combines with BP-53 only when it is occupied by IGF-I or IGF-II (27, 29), both of which bind with high affinity (Ka = 20-30 nivr1) (28). Other forms of the GHdependent binding protein have also been identified in affinity-labeling studies (30); however, it is not known at present how these proteins relate to BP-53. BP-53, 1176 Cloning and Expression of the IGF Binding Protein, BP-53 purified from human plasma together with a minor 47 kDa form (28), has a distinctly different amino-terminus from that of BP-28 (31). The rat homolog of BP-53 has also been purified, and shows considerable N-terminal sequence similarity to the human protein (31). We present here the isolation of cDNA clones encoding the full sequence of BP-53. The deduced amino acid sequence of BP-53 is that of a cysteine-rich protein of 264 amino acids. When this clone is expressed in mammalian tissue culture cells, active BP-53, indistinguishable from the plasma protein by its mobility on SDS gel electrophoresis, IGF binding activity, and immunoreactivity, is secreted into the culture medium. Results and Discussion Amino Acid Sequencing of BP-53 BP-53 was purified from human plasma as described previously (28) and contained two bands of 53 and 47 kDa when electrophoresed on a nonreduced SDS gel. Both these bands contained IGF binding activity (31), and no attempt was made to separate them prior to sequencing or cleavage of the protein. Amino acid sequence data obtained from the purified BP-53 are shown in Table 1. The 60 amino acid N-terminal sequence matched the 15 amino acid sequence reported previously (31) except at position 5 where an alanine had been found previously. The current sequence gave approximately equal yields of alanine and glycine suggesting that the currently isolated protein is a mixture containing both residues at this position. After the Nterminal sequence of 60 residues was complete, the protein remaining on the sequencer filter was cleaved with cyanogen bromide (CNBr) (32). A clear internal sequence was then found (Table 1, CNBr). Trypsin, and lysine-C proteolytic cleavage of the protein followed by isolation of the peptides by reverse phase chromatography, gave a unique sequence for an additional nine internal peptides and three mixture sequences. 1177 Table 1. Amino Acid Sequence of BP-53 Peptides Approximate Initial Yield (pmol) Position in the cDNA Sequence N:GASSGGLGPVVRCEP CDARALAQCAPPPAV C A E L V R E P G(C)G(C C) L(L)X A L(C)E G Q P(P)X X X (T) 185 1-60 CNBr: E X T L N H L K F L N V L S PRGVHIPNXDXKGFY Trypsin Peptides T15-3: A Y L L P A P P A P G[N]A 66 191-219 32 98-117 60 118-132 33 64 35 138-144 145-149 159-178 64 129 18 14 84 17 189-198 199-206 207-216 180-187 217-220 233-251 16 252-263 89 199-214 45 1-18 13 179-196 28 241-251 N-terminal S E S E E D(D) T2-6: S A G S V E S P S V S S T H R T7: F H P L H S K T3-18: I I I I K T3-14: Y K V D Y E S Q S T D T Q[N]F S S E (E) K T2-12: E M E D T L N H L K T4-19: F L N V L S P R T2-10-8a: G V H I P N(N)D(W)K T2-10-8b:ETE YGP(Y)R T2-2: G F Y K T2-20a: G F(C)W X V D K Y G Q P LPGYTTK T2-20b: G K E D V H X Y S M Q S Lysine C-Peptides KC-20a: F L N V L S P R G V (...truncated)