Type II insulin-like growth factor receptor is present in rat serum*

Transcription

1 Proc. Nati. Acad. Sci. USA Vol. 84, pp , November 1987 Medical Sciences Type II insulin-like growth factor receptor is present in rat serum* (insulin-like growth factor carrier proteins/affinity crosslinking/immunoblotting/triton X-114) WIELAND KIESStt, LAWRENCE A. GREENSTEINt, ROBERT M. WHITE, LILLY LEEt, MATTHEW M. RECHLER, AND S. PETER NISSLEYtII tmetabolism Branch, National Cancer Institute, National Institutes of Health, NMolecular, Cellular and Nutritional Endocrinology Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD 20892; and Cancer Center, Howard University, Washington, DC Communicated by David R. Davies, July 16, 1987 (received for review May 12, 1987) ABSTRACT We previously identified in fetal rat serum a component capable of specifically binding radiolabeled insulinlike growth factor type II (IGF-II) that is considerably larger than both the fetal (40 kda) and the adult (150 kda) carrier proteins. We now present immunologic and affinity crosslinking data to show that this binding species is the type II IGF receptor. Rat serum was gel-filtered on a Sephadex G-200 column (0.05 M NH4HCO3, ph 8), and 125I-labeled IGF-ll (125I-IGF-ll) binding was measured in individual column fractions. 125I-IGF-II binding activity was found in the void volume of the column in addition to the carrier protein regions. Competitive binding studies using 125I-IGF-II and binding activity from the Sephadex G-200 void volume showed the characteristics of the type II receptor: IGF-H was more potent than IGF-I, and insulin did not compete. Moreover, a specific anti-type II IGF receptor antibody (no. 3637) completely blocked 125I-IGF-II binding. 125I-IGF-I did not bind to the void volume pool, demonstrating the absence of the type I IGF receptor in rat serum. Affinity crosslinking of 125I-IGF-II to the Sephadex G-200 void volume material demonstrated a specific band at 210 kda without reduction and at 240 kda after reduction of disulfide bonds. The serum type II IGF receptor sie was confirmed by immunoblotting the void volume material with antiserum 3637, which revealed a band slightly smaller (=10 kda) than the type II IGF receptor from rat placental membranes. Immunoquantitation by immunoblotting using pure type II IGF receptor from rat placental membranes as standard showed a developmental pattern. In fetal rat serum (19-days gestation) and in sera from 3-, 10-, and 20-day-old rats, the concentrations of receptor protein were similar (1-5 gug/ml). The level of the type II IGF receptor in serum declined dramatically between age 20 and 40 days, but the receptor was still measurable at age 12 mo. We conclude that the type H IGF receptor is present in rat serum and is developmentally regulated. The insulin-like growth factors (IGF-I and IGF-II), which show extensive amino acid sequence homology with insulin (1), are mitogens for cells in culture, frequently acting in concert with other growth factors (2). IGF-I mediates the anabolic action of pituitary growth hormone in vivo (3). The in vivo role for IGF-II is less clear, but the observation that IGF-II concentration is high in fetal rat serum, with a decline postnatally, has led to the proposal that IGF-II may be important for fetal growth in the rat (4, 5). IGF-I and IGF-II bind to high-affinity receptors present on many cells and tissues (6). There are two types of IGF receptors. The type I receptor binds IGF-I with higher affinity than IGF-II and binds insulin weakly. The type II IGF The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C solely to indicate this fact. receptor prefers IGF-II over IGF-I and does not interact with insulin at all. Further structural and biochemical differences between the two types of receptors include the following: (i) the type I IGF receptor is a heterotetramer, made of two a-binding subunits (130 kda) and two p subunits (95 kda) (7-9); the type II IGF receptor consists of a single binding unit (250 kda) (7, 10); (ii) the type I IGF receptor belongs to a superfamily of tyrosine kinases that includes the insulin receptor, whereas the type II IGF receptor does not have intrinsic tyrosine kinase activity (11-13); (iii) recent evidence shows that the primary structure of the type I IGF receptor as deduced by cloning and sequencing of cdna is highly homologous to the insulin receptor (14). The primary structure of the type II IGF receptor has not been elucidated. There are also two classes of serum IGF carrier proteins (15), one from adult serum that is growth hormone dependent and immunoglobulin sied (150-kDa carrier protein) (16), and the other class predominates in fetal serum and is slightly smaller than albumin (40-kDa carrier protein) (17). We previously identified in fetal rat serum a component capable of specifically binding 1251I-labeled IGF-II (125I-IGF-II) that is considerably larger than the 150- and 40-kDa carrier proteins (17). We now present immunologic and affinity crosslinking data to show that this binding species is the type II IGF receptor and provide evidence that circulating type II IGF receptor is developmentally regulated, being high in fetal and neonatal sera and then rapidly declining after day 20. MATERIALS AND METHODS Peptides and Radioligands. Rat IGF-II (multiplication stimulating activity, MSA-III-2) was purified to homogeneity from BRL 3A-conditioned medium according to method B of Greenstein et al. (18). [Thr5]IGF-I and insulin were purchased from AmGen (Thousand Oaks, CA) and Sigma. IGF-I and IGF-II were iodinated by a modified chloramine-t procedure (19) to specific activities of Ci/g (1 Ci = 37 GBq) (Meloy Laboratories, Springfield, VA). Rat Sera. Sprague-Dawley rats were purchased from Zivic-Miller (Pittsburgh). Blood was collected by decapitation under CO2 anesthesia from 19-day gestation fetuses and postnatal rats aged 3, 10, 20, and 40 days, and 12 mo. Blood was allowed to clot at room temperature and serum was collected after centrifugation and stored at -20'C. Abbreviations: IGF, insulin-like growth factor; 1251-IGF, 125I-labeled IGF; IL-2, interleukin 2. *A preliminary report of this work was presented at the 69th Annual Meeting of the Endocrine Society, Indianapolis, June 10-12, ton leave from the Children's Hospital, University of Munich, Federal Republic of Germany, supported by Stiftung Volkswagenwerk. 'To whom reprint requests should be addressed at: Building 10, Room 4N115, National Institutes of Health, Bethesda, MD

2 Medical Sciences: Kiess et al. Gel Filtration of Rat Sera on Sephadex G-200. Rat serum (1 ml) was diluted with 1 ml of 0.05 M NH4HCO3 (ph 8.0) and applied to a 2.5 x 60-cm Sephadex G-200 column (Pharmacia) that had been equilibrated with 0.05 M NH4HCO3 (ph 8.0) at 4TC. The protein elution profile was determined by measuring A280nm in individual fractions (2.5 ml). Competitive Binding Experiments. Binding of 1251-IGF-I or 1251-IGF-II was measured in a competitive binding assay as described (20). Aliquots (100 1p) of column fractions were incubated with 125I-IGF-I or 1251-IGF-II and unlabeled peptides or IgGs in a final vol of 0.4 ml for 16 hr at 4TC in binding buffer [phosphate-buffered saline (PBS), ph 7.4, containing 2 mg of bovine serum albumin per ml] (20). Albumin-treated charcoal was used to separate bound from free iodopeptide. Preparation of Anti-Type II Receptor IgG. A blocking anti-type II IGF receptor antibody was raised in a rabbit (no. 3637) by immuniation with purified type II IGF receptor from rat chondrosarcoma cells, and IgG from this antiserum and control serum was purified on a protein A affinity column (21). Chemical Crosslinking of '25I-IGF-II to Receptor-Like Material from Rat Sera. Aliquots (100,ul) from pooled void (YO) fractions from the Sephadex G-200 column were incubated with 1251-IGF-II and unlabeled peptides or IgGs for 16 hr at 4TC in crosslinking buffer (50 mm phosphate, ph 7.4) at a final vol of 0.4 ml. At the end of the incubation, 10,l of 50 mm disuccinimidyl suberate solution in dimethyl sulfoxide was added, and the incubation was continued for 15 min at room temperature. Aliquots (100 Au) were transferred to microfuge tubes and boiled for 3 min in 2% NaDodSO4 with or without 100 mm dithiothreitol. NaDodSO4/PAGE (0.1% NaDodSO4/6% acrylamide) was performed using the discontinuous system of Laemmli (22). The gels were dried and autoradiographed with Kodak X-Omat film (Eastman Kodak) and DuPont enhancing screens (DuPont Instruments, Sorvall Division) at -70 C. Immunoblotting of Type II IGF Receptor. Sera, Sephadex G-200 VO pools, washed rat placental membrane preparations, and pure type II IGF receptor were subjected to NaDodSO4/PAGE under nonreducing or reducing conditions. The electrophoresed proteins were transferred onto nitrocellulose paper (Schleicher & Schuell) overnight under constant cooling in Tris/glycine/methanol buffer (15 mm glycine/20 mm Tris/20% methanol) at 0.5 A and 70 V. The nitrocellulose sheets were treated with anti-type II IGF receptor antiserum, developed with 1251-labeled protein A, and autoradiographed as described (21). For immunoquantitation of receptor-like material in sera and Sephadex G-200 VO material, type II IGF receptor from rat placental membranes, purified as described (23), was used as a standard. Protein content of the purified receptor was measured by a modified Lowry method (24). Phase Separation of Rat Placental Membrane Type I Receptor in Triton X-114. Triton X-114 (Sigma) was precondensed three times to obtain a more homogenous preparation (25). Phase separation was performed as described (25, 26). A microsomal membrane preparation from rat placenta was solubilied in 1% Triton X-114 in 0.02 M Tris-HCl, ph 7.4/0.15 NaCl (4 mg of membrane protein per ml of detergent) at 4 C for 1 hr and centrifuged at 100,000 x g for 90 min. A 0.3-ml aliquot of the supernatant was layered on 0.15 ml of 6% (wt/vol) sucrose/0.06% Triton X- 114/Tris/NaCl, warmed to 35 C for 3 min, and centrifuged at 340 X g for 3 min at 22 C. The detergent-poor phase (upper phase) was removed and washed once with Triton X-114. The original detergent-rich phase (lower phase) was made to the same vol as the upper phase, and type II IGF receptor was measured in aliquots of the upper and lower phases by binding of 125I-IGF-II and immunoblotting as described above. Proc. Natl. Acad. Sci. USA 84 (1987) 7721 RESULTS Evidence That the Type II IGF Receptor Is Present in Rat Serum. When fetal and postnatal rat sera were gel filtered on a Sephadex G-200 column in 0.05 M NH4HCO3 (ph 8.0) and binding of 1251-IGF-II was measured in aliquots from individual column fractions, binding activity was found in the void volume, in addition to the 150- and 40-kDa IGF carrier protein regions (Fig. 1 Left). The growth hormone-dependent 150-kDa carrier protein appeared postnatally between day 10 and 20 and predominated over the 40-kDa carrier protein in the adult rat as reported (16, 17). To characterie the 1251_IGF-II binding activity in the void volume of the Sephadex G-200 column, the relative ability of unlabeled IGF-I, IGF-II, and insulin to compete for binding of 1251-IGF-II was measured (Fig. 1 Right). In addition, we used a polyclonal rabbit antiserum (no. 3637) that specifically blocks the binding of 1251-IGF-II to the type II IGF receptor (21). IGF-II was more potent than IGF-I, and insulin did not compete for binding of 125I-IGF-II. In addition, the anti-type II receptor IgG completely blocked 1251-IGF-II binding to VO material from fetal and neonatal rat sera. The anti-type II IGF receptor IgG only partially inhibited 125I-IGF-II binding to the V0 pool from the sera of 40-day-old and 12-mo-old rats, consistent with contamination of those V0 pools with 150-kDa carrier protein. When binding of 1251-IGF-I to individual fractions from the Sephadex G-200 gel filtrations of 10- and 40-day-old rats was measured, only minimal binding activity was detected (Fig. 1). In addition, no specific binding of 1251-IGF-I to pooled Sephadex G-200 V. fractions from sera of 3- and 40-day-old rats was measured (data not shown), demonstrating the absence of the type I IGF receptor in rat serum. Independent support for identification of the '25I-IGF-II binding activity in the Sephadex G-200 VO as the type II IGF receptor came from affinity crosslinking experiments using disuccinimidyl suberate. Typically, affinity crosslinking of 125[_IGF-II to the membrane-associated type II receptor results in formation of a 250-kDa complex if the analysis on NaDodSO4/PAGE is carried out in the presence of dithiothreitol (Fig. 2) (7, 10). Crosslinking of _251-IGF-II to the G-200 VO pool from sera of 10-day-old rats demonstrated a slightly smaller complex (240 kda) (Fig. 2). Formation of the 240-kDa radioligand-receptor complex was prevented by incubation with unlabeled IGF-II and anti-type II receptor IgG 3737, whereas insulin and control IgG did not block formation of the type II receptor complex. When NaDod- S04/PAGE was performed without disulfide bond reduction, the sie of the radioligand-receptor complex decreased to 210 kda (data not shown), analogous to the behavior of the membrane-associated type II IGF receptor (7, 10). Quantitation of the Serum Type II IGF Receptor: Developmental Regulation. Gel filtration of fetal and postnatal serum on Sephadex G-200 and measurement of 125I-IGF-II binding in the V0 fractions (Fig. 1) suggested that the amount of type II receptor was higher in fetal and neonatal sera than in sera from older animals. To measure the amount of type II receptor in these sera, we used two methods-scatchard analysis of the binding of IGF-II and immunoblotting with anti-type II IGF receptor antiserum and highly purified type II receptor as standard. Scatchard analysis of IGF-II binding to the Sephadex G-200 VO pool from fetal and 3-, 10-, and 20-day-old rat sera yielded a straight line consistent with a single class of binding sites. Assuming a binding stoichiometry of 1:1, we calculated the amount of receptor protein to be 4, 2, 7, and 2,ug/ml, respectively. In addition, immunoblotting was done on both untreated sera (data not shown) and aliquots of the Sephadex G-200 VO pools from rats at different ages (Fig. 3). Immunoquantitation revealed a developmental pattern; in fetal rat serum (19-days gestation)

3 7722 Medical Sciences: Kiess et al. Proc. Natl. Acad Sci. USA 84 (1987) 2- FETAL I vo y-g Albumin A B C D E AM.. -O f.i,:j s/ S 6 E, DAY I Vo I y-g jlabumin (D a co I- 0 MEMBRANES SERUM (l0d) To E ' m T FRACTION NUMBER 1 IgGs [pg/mil FIG I-IGF-II binding to column fractions after Sephadex G-200 gel filtration of sera from rats of different ages; competitive binding to VO pools. Rat serum (1 ml) from rats of the indicated ages was gel-filtered on a Sephadex G-200 column in 0.05 M NH4HCO3 (ph 8.0), and 125I-IGF-II binding to individual column fractions (0) was measured as described (Left). 125I-IGF-I binding to aliquots of column fractions (q) was also measured (sera from 10- and 40-day-old rats). The Vo and the elution positions of immunoglobulin (y-g) and albumin are indicated by arrows. VO fractions were pooled as FIG. 2. Affinity crosslinking of '251-IGF-Il to type II IGF receptors from rat placental membranes (lanes A-E) and to Sephadex G-200 VO pools from sera from 10-day-old rats (lanes 1-5). Affinity crosslinking of '251-IGF-II to membranes or G-200 VO pools was performed using disuccinimidyl suberate (0.1 mm) as the crosslinking agent as described. Incubations were done in the presence of unlabeled IGF-II (2.5,Ag/ml) (lanes B and 2), insulin (10 pg/ml) (lanes C and 3), IgG 3637 (160 jug/ml) (lanes D and 4), and control IgG (160 ug/ml) (lanes E and 5). NaDodSO4/PAGE was carried out under reducing conditions (100 mm dithiothreitol). Radiolabeled sie standards were as follows: ovalbumin, 46 kda; bovine serum albumin, 69 kda; phosphorylase b, 92.5 kda; myosin, 200 kda. The scale on the right is deduced from the molecular sie standards on the left. The radioligand-type II receptor complex is indicated by an arrow. and in sera from 3-, 10-, and 20-day-old rats, 1-5 4g of receptor protein per ml was measured. The level of type II IGF receptor declined dramatically between 20 and 40 days and was <60 ng/ml at 40 days and 12 mo. The Paradox of a Membrane Protein Being Present in Serum: Possible Explanations. We were surprised to discover that the type II IGF receptor, an intrinsic membrane protein, was present in solution in serum, and we sought possible explanations. First, the affinity crosslinking studies (Fig. 2) suggested that the serum type II IGF receptor (-240 kda) was slightly smaller than the rat placental membrane type II IGF receptor (=250 kda). This difference in sie was confirmed by immunoblotting of serum-derived Sephadex G-200 VO material and rat placental membranes with antiserum 3637 (Fig. 3 Right). Under nonreducing conditions, the receptor protein from placental membranes was -220 kda, whereas the serum receptor was -210 kda. The sie difference was documented by applying the serum receptor and the membrane receptor to the same lane, which resulted in a double band. The same sie difference (-10 kda) was seen under reducing conditions (data not shown), but the bands were much fainter due to lower reactivity of antiserum 3637 with the reduced receptor. This result would be consistent with the serum type II receptor being truncated and representing only the extracellular domain. Second, we also asked whether the type II IGF receptor is circulating in serum as a large aggregate, in which case a hydrophobic core surrounded by hydrophilic domains could confer water solubility. Serum from 3-day-old rats was gel filtered on a Sepharose 6B column in Dulbecco's PBS (ph 7.4) without calcium and magnesium, and 1251-IGF-II binding to individual column fractions was measured in the presence or absence of IgG The type II IGF receptor did not elute indicated, and competitive binding experiments (Right) were performed on 100-plI aliquots of the pools using '25I-IGF-II and unlabeled IGF-I1 (o), IGF-I (e), anti-type II receptor IgG (e), control IgG (a), and insulin (A).

4 Medical Sciences: Kiess et al. A B C D E F 123-I K A B C 92.5 FIG. 3. Quantitation of the serum type II IGF receptor from rats of different ages by immunoblotting; measurement of the sie difference between the membrane and serum receptors. Immunoblotting of G-200 Vo pools (lanes A-F) was performed as described. (Left) Sephadex G-200 VO pool material (1 ml per lane) and pure type II IGF receptor from rat placental membranes as standard (lane 1, 250 ng; lane 2, 62.5 ng; lane 3, 31 ng of receptor protein) were subjected to NaDodSO4/PAGE without dithiothreitol, transferred onto nitrocellulose paper, and underwent reaction with antiserum 3637 and 1251-labeled protein A. Radioactive molecular sie standards are the same as in Fig. 2, and the scale on the right is deduced from the standards on the left. Sephadex G-200 V. material in lanes A-F are from fetal (19-days gestation), 3-, 10-, 20-, and 40-day-old, and 12-mo-old rats, respectively. (Right) Immunoblotting of rat placental membrane type 1I IGF receptor and serum type II IGF receptor. Washed microsomal membranes (lane A) and pooled Sephadex G-200 V. material (serum from 10-day-old rats) (lane B) were subjected to NaDodSO4/PAGE, proteins were transferred onto nitrocellulose paper, and the sheets underwent reaction with antitype II receptor antiserum 3737 and 251I-labeled protein A as described. Lane C contains both pooled Sephadex G-200 VO material and placental membranes. in the VO of the column but eluted between thyroglobulin (660 kda) and bovine serum albumin (67 kda) markers, showing that the type II IGF receptor is not circulating in rat serum as a large aggregate. Finally, we examined the relative preference of the intact membrane-derived type II receptor for a detergent-rich versus a detergent-poor environment by utiliing the technique of phase separation in Triton X-114 (25, 26). Typically, intrinsic membrane proteins are recovered from the detergent-rich phase (25). However, when rat placental membranes were solubilied and phase separation was carried out, 125I-IGF-II binding activity was predominantly measured in the detergent-poor phase; binding of 1251-IGF-II to the detergent-poor phase was 3- to 4-fold greater than binding to the detergent-rich phase (data not shown). Immunoblotting of detergent-poor and detergent-rich phases also revealed a 3- to 4-fold difference in the amount of receptor protein (data not shown). These results demonstrate the hydrophilic nature of the membrane-derived type II IGF receptor, consistent with the solubility of the type II receptor in serum. In spite of this behavior, which is not typical of an intrinsic membrane protein, the type II IGF receptor could not be freed from rat placental membranes by treatment with 0.1 M Na2CO3 (ph 11), 1 M NaCl, or 6 M guanidine hydrochloride, demonstrating that the type II IGF receptor is not an extrinsic membrane protein. DISCUSSION We previously described a component in fetal rat serum capable of specifically binding '25I-IGF-II that is considerably larger than the 150- and 40-kDa IGF carrier proteins (17). 300 Proc. Natl. Acad. Sci. USA 84 (1987) 7723 We have now utilied a polyclonal antiserum directed against the type II IGF receptor to demonstrate that the large serum binding species is the type II IGF receptor, heretofore found only on the cell surface or in membrane preparations. This antiserum (no. 3637) is specific for the type II receptor (21). It inhibits the binding of 125I-IGF-II to the type II IGF receptor but does not inhibit the binding of 125I-IGF-I to the type I IGF receptor nor does it block the binding of insulin to the insulin receptor. In addition, IgG 3637 does not block the binding of 125I-IGF-II to either the 150- or 40-kDa serum carrier proteins. Furthermore, immunoblotting experiments utiliing membrane preparations or solubilied whole cells demonstrate that IgG 3637 is only directed against the type II IGF receptor (21). Competitive binding experiments that utilied radiolabeled IGF-II and the large molecular weight material from fetal and neonatal serum showed a pattern typical of the type II IGF receptor: IGF-II was much more potent than IGF-I in competing for binding of 1251-IGF-II, high concentrations of insulin did not compete, and the anti-type II IGF receptor antibody fully blocked 125I-IGF-II binding. Affinity crosslinking experiments with 1251-IGF-II and the large molecular mass material from rat serum demonstrated that IgG 3637 completely prevented the formation of a complex that was only slightly smaller than the typical type II radioligandreceptor complex derived from plasma membranes. Immunoblotting with anti-type II receptor serum and the large molecular weight material from rat serum also identified a single component slightly smaller than the membrane-derived type II IGF receptor. Taken together, these results provide strong evidence that the type II IGF receptor is present in rat serum. The presence of other growth factor receptors in serum or in conditioned medium from cell cultures has been reported. Human epidermoid carcinoma A431 cells secrete a 105-kDa protein that is the extracellular domain of the epidermal growth factor receptor (27). A truncated form of the interleukin 2 (IL-2) receptor is secreted by activated T-cell lines and peripheral mononuclear cells (28). The IL-2 receptor is also present in human serum, and there is a 100-fold increase in levels of the receptor protein in sera from patients with T-cell leukemia (28). A specific growth hormone binding protein has been reported in rabbit and human sera (29, 30). In the rabbit, there is evidence for immunologic identity between the serum growth hormone binding protein and a membrane growth hormone receptor.** IM-9 lymphocytes have been shown to release into the culture medium growth hormone and insulin binding components, presumably intact receptors or receptor fragments (31, 32). In the case of the soluble forms of the epidermal growth factor and IL-2 receptors, there is evidence that these truncated forms represent the extracellular domains of the membrane receptors (27, 28). Our affinity crosslinking and immunoblotting data show that the serum type II IGF receptor is -10 kda smaller than the intact receptor from rat placental membranes, raising the possibility that the serum type II IGF receptor may also represent the extracellular domain. However, a similar sie difference has been reported between two membrane type II receptors from different tissues in the rat (adipocytes and liver) (8). If the serum type II IGF receptor is analogous to the truncated epidermal growth factor and IL-2 receptors and represents the extracellular domain, this means that the cytoplasmic domain of the type II receptor is relatively small. In fetal and neonatal rat sera, the concentration of the type II IGF receptor is at least several micrograms per milliliter **Waters, M. J. & Barnard, R., Program of the 68th Annual Meeting of the Endocrine Society, June 25-27, 1986, Anaheim, CA, abstr. 371.

5 7724 Medical Sciences: Kiess et al. (-10 nm). This is several times the concentration of the truncated IL-2 receptor in plasma of patients with adult T-cell leukemia and is approximately several hundredfold the concentration of the IL-2 receptor in normal plasma (28). The circulating IL-2 receptor is presumably derived from normal or leukemic T cells; the source of the serum type II receptor is unknown. With the exception of C6 rat glial cells, where a very low level of type II IGF receptor was measured in conditioned medium, we were unable to demonstrate release of type II IGF receptor from other cell strains or lines known to have high numbers of cell-surface type II IGF receptors (W.K. and L.L., unpublished results). Constituents of blood cells released during the clotting process do not account for the presence of the type II IGF receptor in rat serum, because the receptor was also present in EDTA or citrated plasma (data not shown). In addition, release of type II IGF receptors from tissue or tissue fluid during blood collection (decapitation) does not account for the serum type II IGF receptor either, because the type II IGF receptor was also present in blood collected by cardiac puncture (data not shown). How can an intrinsic membrane protein be soluble in serum? If the serum type II IGF receptor does represent the extracellular domain of the membrane receptor as discussed above, this could provide an explanation for solubility in serum. We also asked whether the type II IGF receptor was circulating as a large aggregate, in which case the hydrophobic nature of the membrane protein could be masked by burying lipophilic portions of the receptor inside the aggregate. However, when rat serum was gel filtered on a Sepharose 6B column and 15I-IGF-II binding to individual column fractions was measured, the receptor binding activity was found between thyroglobulin (660 kda) and albumin (69 kda) markers, indicating that the type II IGF receptor is not circulating as a large aggregate. In addition, we examined the overall hydrophilic versus hydrophobic nature of the intact membrane-derived type II IGF receptor by performing phase separation in Triton X-114 (25). Triton X-114 solubiliation of cellular type II IGF receptor from rat placental membranes and separation of the aqueous and detergent phases showed that the type II IGF receptor preferred the detergent-poor phase and, therefore, by this criterion is a hydrophilic protein. While phase separation in Triton X-114 was originally considered a test of intrinsic versus extrinsic membrane proteins (25), the distinction is apparently more complex. Thus, the acetylcholine receptor, which is clearly an integral membrane protein, also is found in the detergent-poor phase after phase separation in Triton X-114 (26). While all of the factors that determine the distribution of a membrane protein between the detergent-poor and detergentrich phase are not fully understood, the result with the type II IGF receptor provides empirical evidence that the intact membrane-derived type II receptor could be soluble in an aqueous environment such as serum. Thus, solubility in serum could be explained if the circulating type II IGF receptor represents the extracellular domain, or alternatively, the serum receptor could be intact and the sie difference could reflect differences among tissues, possibly due to variability of glycosylation. What is the function of the serum type II IGF receptor? The serum type II IGF receptor is not a significant carrier of circulating IGF-II in fetal rats; <3% of serum IGF-II is found associated with the type II receptor (17). This is, in part, explained by the lower concentration of serum type II IGF receptor (10 nm) compared with the 40-kDa carrier protein ( nm) (17). Nonetheless, if present in extracellular Proc. Natl. Acad. Sci. USA 84 (1987) fluid, it is possible that the type II IGF receptor could modulate the action of IGF-II. In this regard, it is of interest that the serum type II IGF receptor is developmentally regulated, being high in fetal and neonatal sera and then declining rapidly after day 20. This approximates the developmental pattern of IGF-II itself (4, 5). 1. Humbel, R. E. (1984) in Hormonal Proteins and Peptides, ed. Li, C. H. (Academic, New York), pp Van Wyk, J. J. (1984) in Hormonal Proteins and Peptides, ed. Li, C. H. (Academic, New York), pp Schoenle, E., Zapf, J., Humbel, R. E. & Froesch, E. R. (1982) Nature (London) 296, Moses, A. C., Nissley, S. P., Short, P. A., Rechler, M. M., White, R. M., Knight, A. B. & Higa, 0. Z. (1980) Proc. Natl. Acad. Sci. USA 77, Daughaday, W. H., Parker, K. A., Borowsky, S., Trivedi, B. & Kapadia, M. (1982) Endocrinology 110, Rechler, M. M. & Nissley, S. P. (1985) in Polypeptide Hormone Receptors, ed. Posner, B. I. (Dekker, New York), pp Kasuga, M., Van Obberghen, E., Nissley, S. P. & Rechler, M. M. (1981) J. Biol. Chem. 256, Massague, J. & Cech, M. P. (1982) J. Biol. Chem. 257, Jacobs, S., Kull, F. C. & Cuatrecasas, P. (1983) Proc. Natl. Acad. Sci. USA 80, Massague, J., Guillette, B. J. & Cech, M. P. (1981) J. Biol. Chem. 256, Sasaki, N., Rees-Jones, R. W., Zick, Y., Nissley, S. P. & Rechler, M. M. (1985) J. Biol. Chem. 260, Haskell, J. F., Nissley, S. P., Rechler, M. M., Sasaki, N., Greenstein, L. & Lee, L. (1985) Biochem. Biophys. Res. Commun. 130, Covera, S., Whitehead, R. E., Mottola, C. & Cech, M. P. (1986) J. Biol. Chem. 261, Ullrich, A., Gray, A., Tam, A. W., Yang-Feng, T., Tsubokawa, M., Collins, C., Henel, W., Le Bon, T., Kathuria, S., Chen, E., Jacobs, S., Francke, U., Ramachandra, J. & Fujita-Yamaguchi, Y. (1986) EMBO J. 5, Hint, R. L. (1984) Clin. Endocrinol. Metab. 13, Moses, A. C., Nissley, S. P., Cohen, K. L. & Rechler, M. M. (1976) Nature (London) 263, White, R. M., Nissley, S. P., Short, P. A., Rechler, M. M. & Fennoy, 1. (1982) J. Clin. Invest. 69, Greenstein, L. A., Nissley, S. P., Moses, A. C., Short, P. A., Yang, Y. W.-H., Lee, L. & Rechler, M. M. (1984) Methods for Preparation ofmedia, Supplements, and Substrata for Serum-Free Animal Cell Culture, eds. Sato, G. & Barnes, D. (Liss, New York), pp Rechler, M. M., Nissley, S. P., Podskalny, J. M., Moses, A. C. & Fryklund, L. (1977) J. Clin. Endocrinol. Metab. 44, Moses, A. C., Nissley, S. P., Passamani, J., White, R. M. & Rechler, M. M. (1979) Endocrinology 104, Kiess, W., Haskell, J. F., Lee, L., Greenstein, L. A., Miller, B. E., Aarons, A. L., Rechler, M. M. & Nissley, S. P. (1987) J. Biol. Chem., in press. 22. Laemmli, U. K. (1970) Nature (London) 227, August, G. P., Nissley, S. P., Kasuga, M., Lee, L., Greenstein, L. A. & Rechler, M. M. (1983) J. Biol. Chem. 258, Peterson, G. L. (1977) Anal. Biochem. 83, Bordier, C. (1981) J. Biol. Chem. 256, Maher, P. A. & Singer, S. J. (1985) Proc. Natl. Acad. Sci. USA 82, Weber, W., Gill, G. & Spiess, J. (1984) Science 224, Rubin, L. A., Kurman, C. C., Frit, M. E., Yarchoan, R. & Nelson, D. L. (1986) in Leukocytes and Host Defense, eds. Oppenheim, J. J. & Jacobs, D. M. (Liss, New York), pp Ymer, S. I. & Herington, S. C. (1985) Mol. Cell. Endocrinol. 41, Baumann, G., Stolar, M. W., Amburn, K., Barsano, C. P. & DeVries, B. C. (1986) J. Clin. Endocrinol. Metab. 62, McGuffin, W. L., Gavin, J. R., Lesniak, M. A., Gorden, P. & Roth, J. (1976) Endocrinology 98, Gavin, J. R., Buell, D. N. & Roth, J. (1972) Science 178,