0021-972X/90/7101-0015$02.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1990 by The Endocrine Society
Vol. 71, No. 1 Printed in U.S.A.
Identification of Placental Human Growth Hormone as the Growth Hormone-V Gene Expression Product* FRANCIS FRANKENNE, MARIE-LOUISE SCIPPO, JOZEF VAN BEEUMEN, AHMED IGOUT, GEORGE HENNEN Experimental and Clinical Endocrinology, CHU B-23, B-400 Liege; and Laboratorium uoor Microbiologie (J. V.B.), Ledeganckstraat 35, B-900 Gent, Belgium
ABSTRACT. A GH variant of placental origin, placental GH, has recently been shown to replace pituitary GH in maternal serum during pregnancy. Besides, the GH variant (GH-V) gene has been demonstrated to be expressed in the placenta. The similarities between their known properties strongly suggest that the placental GH and the GH-V protein are the same molecular
species. Here we provide final evidence that this is indeed the case by sequence analysis of both the 22K and 25K forms. Furthermore, the 25K form is shown to be glycosylated, while the 22K form is not. Both size variants of placental GH are, thus, likely to reflect the partial glycosylation of a unique peptidic chain. (J Clin Endocrinol Metab 7 1 : 15-18, 1990)
D
URING the second half of pregnancy, pituitary GH is progressively replaced in the maternal serum by a GH analog originating from placenta, distinct from placental lactogen (hPL), and which has been tentatively named placental GH (1-4). It can be distinguished from pituitary GH by its specific pattern of reactivity with two high affinity monoclonal antibodies (MAb) insensitive to physiological concentrations of hPL: the 5B5MAb, recognizing both pituitary and placental GH and the K24-MAb, which only reacts with the pituitary hormone main variant (22K hGH). Its concentration steadily rises to term, then drops sharply to an undetectable level within 1 h after delivery (4). So far, placental GH has only been found in the maternal compartment, where pituitary GH steadily disappears to become undetectable near term (4). Placental GH is, therefore, likely to actively participate in the control, by the placenta, of maternal metabolism in late pregnancy. Placental GH has been partially purified from term placenta through ion exchange and immunoaffinity chromatography (3, 4). Due to some structural unstability of the anti-GH MAb used in the latter step (our unpublished results), the recovery yields were very low, and the described material was contaminated by MAb, which hindered full structural analysis by classical means. A
partial characterization of placental GH has, nevertheless, been achieved. It displays an apparent size dimorphism, being composed of two forms of 22K and 25K, respectively (3, 4). Preliminary data have shown that the latter form is probably glycosylated (5). Both forms share an N-terminal epitope with pituitary GH, but lack an internal epitope (4). Placental GH is also more basic than the pituitary hormone variants with which it competes with an apparent high potency for binding to liver GH membrane receptors (4). The five genes of the hGH-human chorionic somatomammotropin (hCS) family are clustered on the long arm of the chromosome 17 in the following 5' to 3' order: hGH-N, hCS-L, hCS-A, hGH-V, and hCS-B (abbreviated for hGH-normal, hCS-like, and hGH-variant) (6, 7). hGH-N is expressed in the pituitary, yielding a major 22K, a minor 20K as well as posttranslational variants (8), while both CS-A and CS-B encode hCS (hPL), which is massively produced by the placenta (9). By mRNA hybridization with specific oligonucelotidic probes and cDNA cloning from a placental library, we have shown that the GH-V gene, previously considered as silent, is actually expressed in the placenta (10,11). This has been confirmed in a recent comprehensive study (12) showing that the GH-N gene is expressed only in the pituitary, whereas the four other genes are expressed in the placenta. A minor alternative GH-V pre-mRNA splicing, leading to a longer mRNA, has also been described (13). The GH-V protein (11, 14) differs from the main 22 K pituitary variant by 13 amino acid differences evenly distributed along the peptidic chain.
Received June 15,1989. Address requests for reprints to: Dr. Francis Frankenne, Experimental and Clinical Endocrinology, CHU B-23, B-4000 Liege, Belgium. * This work was supported by the Science Policy Programming Department of the Prime Minister's Office and grants from the Belgian FNRS and the Region Wallonne.
15
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:04 For personal use only. No other uses without permission. . All rights reserved.
FRANKENNE ET AL.
16
JCE & M • 1990 Vol71«Nol
carbohydrate-staining technique was also applied on Western blotted material to assess the glycosylation of the 25K form. Materials and Methods Partial purification of placental GH
FIG. 1. SDS-PAGE of a placental GH-enriched fraction, followed by electrotransfer and staining either for protein with amido black (A) or for placental GH immunoactivity using the 5B4-MAb (B).
TABLE 1. Sequence analysis data Step
18
21
Arg 0.51 (His 0.02)
Tyr 1.3 (His 0.25) Tyr 0.39 (His 0.04)
22K PGH 25K PGH
25 Tyr (Phe Tyr (Phe
0.64 0.12) 0.58 0.26)
See Materials and Methods. Recorded increases in amounts (picomoles) of the phenylhydantoin amino acids for the residues attributed to the positions 18, 21, and 25. The increases in phenylhydantoin amino acids corresponding to the GH-N sequence at the same positions are also indicated (in parentheses).
As determined from expression experiments with SV40-transfected eukariotic cells (15,16), the GH-V protein displays a basic isoelectric point and a high potency for binding to GH receptors in addition to an Asn-glycosidebinding site. Also, the first 18-amino acid portion is similar to pituitary GH (14). The similar properties of the GH-V protein and placental GH strongly suggest their identity. Sequence data on placental GH would definitely assess this identity. Since homogeneous placental GH was not available, we attempted to perform its sequencing directly on material bound to Western blot fragments. In addition, a specific 10
1
The enriched placental GH was obtained from placental extract essentially as described previously (4), but with the final immunoaffinity step modified as follows. To 3 mL ReactiGel GF-2000 (Pierce, Rockford, IL) 5 mg purified anti-GH 5b4 (4, 17) in 3 mL 0.1 M NaHCO3, pH 9, were added. The suspension was gently agitated at 4 C for 48 h. The protein solution was then replaced by 3 mL 1 M ethanolamine, pH 9, for 4 h at 20 C. Finally, the resin was sequentially washed using 1 M NaCl, H2O, and 0.2 M phosphate-0.14 M NaCl buffer, pH 7.4 (PBS). To that affinity matrix were added 100 mL PBS
containing 3 g of a crude placental GH fraction. After mild stirring for 16 h at 4 C, the resin was centrifuged, washed with PBS, then poured into a column (0.9 x 5 cm), washed again with 100 mL PBS, and eluted using 3 mL 3 M KSCN in PBS. The desorbed material was finally concentrated, desalted, and lyophilized. Electrophoresis and blotting The partially purified placental GH preparation was submitted to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, 15% acrylamide), then electroblotted using a semidry transfer apparatus (Bio-Lyon, Lyon, France), with the following set of buffers: 0.3 M Tris-20% methanol (anodic 1); 25 mM Tris-20% methanol (anodic 3); 40 mM 6-amino-iVhexanoic acid-20% methanol (cathodic); either A) a polyvinylidene difluoride filter (PVDF; Immobilon, Millipore, Milford, MA) (18) or B) a nitrocellulose (NC) filter was used. After applying 24 V for 2 h, the PVDF filter (A) was stained with 0.1% amido black in 2% acetic acid-20% methanol and then destained in 7% acetic acid-20% methanol, and a NC blot (B) stained for 5B4-MAb-reacting material as described previously (4). On a second NC blot, a specific detection of glycoconjugates was obtained by mild periodate oxidation, followed by staining of the resulting aldehydes using digoxigenin-succinyl-E-amidocaproic acid hydrazide and antidigoxin alkaline phosphatase15
20
25
J 22K PGH :
Pro-Thr- He -Pro-Leu-Ser-Arg-Leu-Phe-Asp-Asn-Ala-Met-Leu-Arg-Ala
25K PGH: Phe-Pro-Thr- De -Pro-Leu
Leu-Phe-Asp-Asn-Ala-Met-Leu-Arg-Ala-Arg
30
1
Leu-Tyr-Gln-Leu-Ala-Tyr-Asp-Thr-Tyr-Gln-Glu-Phe Leu-Tyr
Tyr-Asp-Thr
Gin
GH-V protein: Phe-Pro-Thr-He-Pro-Leu-Ser-Arg-Leu-Phe-Asp-Asn-Ala-Met-Leu-Arg-Ala Arg Arg-Leu Tyr Gln-Leu-Ala Tyr Asp-Thr-Tyr-GIn GH-N protein: Phe-Pro-Thr-lie-Pro-Leu-Ser-Arg-Leu-Phe-Asp-Asn-Ala-Met-Leu-Arg-Ala His Arg-Leu His Gin-Leu-Ala Phe •Asp-Thr-Tyr-GIn CS protein:
VaJ-GIn
Thr Val Pro-Leu-Ser-Arg-Leu-Phe-Asp His •Ala-Met-Leu Gin-Ala- His Arg Ala- His Gln-Leu-Ala He Asp-Thr-Tyr-GIn
FIG. 2. N-Terminal amino acid sequence of the 22K and 25K forms of placental GH compared to the corresponding regions of the GH-N (23), GH-V (14), and CS (10) proteins.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:04 For personal use only. No other uses without permission. . All rights reserved.
PLACENTAL hGH AS THE GH-V GENE EXPRESSION PRODUCT
17
for 26 of the first 32 residues for the 22K form and for 21 of the first 29 residues for the 25K form (Fig. 2). On specific staining for glycoconjugates of Western blotted placental GH (Fig. 3), the 25K form appears as a colored band opposite the 22K form, which is not stained.
Discussion A
A
A
B
FlG. 3. SDS-PAGE of a placental GH-enriched fraction (lane B), followed by electrotransfer and specific staining for glycoconjugates (see Materials and Methods). The darkly stained bands correspond to glycoproteins, whereas nonglycosylated proteins appear as nonstained bands, due to the lesser background staining. Biosynthetic Met-GH was used as negative control (lane A).
labeled antibody (Glycan detection kit, Boeringher, Mannheim, West Germany). Biosynthetic Met-GH (Genentech, San Francisco, CA) was used as the negative control. Sequencing Sequence data were obtained using an Applied Biosystems 477A sequencer loaded with PVDF filter fragments (19) containing either the 22K or the 25K form of placental GH after SDS-PAGE and Western blotting.
Results Partially purified material was first prepared, using improved techniques for both immobilizing the MAb and placental GH binding. That preparation was submitted to SDS-PAGE, followed by an electrotransfer to a PVDF filter that was then stained for protein (Fig. 1A). A parallel Western blot on an analytical sample revealed the position of the placental GH bands through their specific immunoactivities (Fig. IB). Filter fragments containing 22K and 25K placental GH were submitted to automated amino acid sequence analysis. The positions for which no net increase in the amount of phenylthiohydantoin amino acid was obtained compared to the preceding cycle were left unassigned. In Table 1 are plotted the raw data obtained for the critical 18, 21, and 25 positions, which correspond to different amino acids on GH-V and GH-N proteins. Based on the amount of Ile-4, the amounts of 22K and 25K bound to the blot were calculated to be 15 ± 3 and 10 ± 2 pmol, respectively. Taking the presumed losses due to electrophoresis and electroblotting into account (19), this would mean that about 30 and 20 pmol, respectively, were loaded on the gel. In spite of the very low amounts of material bound to the PVDF filters, a clear-cut assignment could be made
Although in each of the sequence runs, some (but different, except for Arg19) positions could not be identified, the data for positions 18, 21, and 25 are sufficient to show that both the 22K and 25K forms have an Nterminal sequence allowing their definitive identification as GH-V gene expression products and not as GH-N gene products. Preliminary evidence of glycosylation of the 25K form was previously given by Concanavalin-A binding experiments (5). The present results, obtained using a totally independent technique, confirm that the 25K form of placental GH is actually a glycoprotein, while the 22K form is not. Besides the GH-V mRNA coding for the 22K GH-V protein, a minor alternative pre-mRNA splicing has been described, which results in a larger GH-V2 mRNA encompassing intron 4 (13). The resulting protein would be of a larger size than 22K GH (230 instead of 191 amino acids), and due to a shift in the reading frame, the amino acid sequence of a large C-terminal portion (residues 127-230) would be totally different from the GH-N or GH-V 22K product. As a consequence, it would be devoid of the JV-glycosylation site present on the 22K GH-V protein. It has recently been reported that the GH-V2 protein was not secreted by a GH-V gene-transfected fibroblastic cell line, whereas the GH-V protein was secreted as three size variants, one nonglycosylated 22K and two glycosylated of 24K and 26K (20). The 25K placental GH displays physiochemical and immunochemical characteristics very similar to the 22K form (4) and has been shown here to be glycosylated. It is, therefore, much more likely to result from translation of the GH-V mRNA and not the alternatively spliced GH-V2 mRNA. Four genes belonging to the GH family are known to be expressed in the placenta: GH-V, CS-A, CS-B, and CS-L (10, 12). So far, the CS-L protein has not been in either the placenta or the circulation, and to our knowledge, nothing is known of its potential GH-like activity. The CS-A and CS-B genes yield mass amounts of hCS (hPL), exhibiting a very low GH-like activity (21). The third gene, GH-V, encodes placental GH, which is likely to be a potent analog of GH (4). These activities probably account for the disappearance of maternal circulating pituitary GH, resulting from inhibition of somatotroph
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:04 For personal use only. No other uses without permission. . All rights reserved.
FRANKENNE ET AL.
18
secretion by increasing levels of serum insulin-like growth factor-I (22). This is consistent with the placenta taking control of the maternal metabolism through expression of the hCS gene, the hGH-V gene, or both.
References 1. Hennen G, Frankenne F, Pirens G, Gomez F, Closset J, El Khayat N. New chorionic GH-like antigen revealed by monoclonal antibody radioimmunoassays. Lancet. 1985;16:399. 2. Hennen G, Frankenne F, Pirens G, et al. A chorionic GH-like antigen: increasing levels during second half of pregnancy with pituitary GH suppression as revealed by monoclonal antibody radioimmunoassays. Int J Fertil 1985;30:27-33. 3. Hennen G, Frankenne F, Closset J, et al. Monoclonal antibody to growth hormone: the discovery of a new variant, human placental growth hormone. In: Forti G, Serio M, Lipsett MB, eds. Monoclonal antibodies: basic principles, experimental and clinical applications in endocrinology. New York: Raven Press; 1985;3O:29-4O. 4. Frankenne F, Closset J, Gomez F, Scippo ML, Smal J, Hennen G. The physiology of growth hormones in pregnant women and partial characterization of the placental GH variant. J Clin Endocrinol Metab. 1988;66:1171-80. 5. Frankenne F, Scippo ML. Circulating forms and partial glycosylation of human placental growth hormone hPGH. Proc of the 70th Annual Meet of The Endocrine Soc. 1988;1129. 6. Owerbach D, Rutter WJ, Martial JA, Baxter JD, Shows TB. Genes for growth hormone, chorionic somatomammotropin, and growth hormones-like gene on chromosome 17 in humans. Science. 1980;209:289-92. 7. Hirt H, Kimelman J, Birnbaum MJ, et al. The human growth hormone gene locus: structure, evolution and allelic variations. DNA. 1987;6:59-70. 8. Barsh GS, Seeburgh PH, Gelinas RE. The human growth hormone gene family: structure and evolution of the chromosomal locus. Nucleic Acids Res. 1983;83:3939-59. 9. Barrera-Saldana HA, Seeburg PH, Saunders GF. Two structurally different genes produce the same secreted human placental lactogen hormone. J Biol Chem. 1983;258:3787-93. 10. Frankenne F, Rentier-Delrue Fr, Scippo ML, Martial H, and
11. 12. 13.
14. 15.
16.
17. 18. 19. 20. 21. 22.
23.
JCE & M • 1990 Vol 71 • N o l
Hennen G. Expression of the growth hormone variant gene in human placenta. J Clin Endocrinol Metab. 1987;64:635-637. Igout A, Scippo ML, Frankenne F, Hennen G. Cloning and nucleotide sequence of placental HG-V and c-DNA. Arch Int Physiol Biochem. 1988;96:63. Chen EY, Liao YC, Smith SH, Barrera-Saldana HA, Gelinas RE, Seeburg PH. The human growth hormone locus: nucleotide sequence, biology, and evolution. Genomics. 1989;4:479-97. Cooke NE, Ray J, Emery JG, Liebhaber SA. Two distinct species of human growth hormone-variant mRNA in the human placenta predict the expression of novel growth hormone proteins. J Biol Chem. 1988;263:9001-6. Seeburg PH. The human growth hormone gene family: nucleotide sequences show recent divergence and predict a new polypeptide hormone. DNA. 1982;82:239-249. Plavlakis GN, Hizuka N, Gorden PH, Seeburg PH, Hamber DH. Expression of two human growth hormone genes in monkey cells infected by simian virus 40 recombinants. Proc Natl Acad Sci USA. 1981;78:7398-402. Hizuka N, Hendricks CM, Pavlakis GN, Hamer DH, Gorden P. Properties of human growth hormone polypeptides: purified from pituitary extracts and synthesized in monkey kidney cells and bacteria. J Clin Endocrinol Metab. 1982;55:545-50. Gomez F, Pirens G, Schaus Ch, Closset J and Hennen G. A highly sensitive radioimmunoassay for human growth hormone using a monoclonal antibody. J Immunol. 1984;4:145-57. Matsudaira P. Sequence from picomole quantities of proteins electroblotted only polyvinylidene difluoride membranes. J Biol Chem. 1987;262:10035-8. Yuen SW, Chut AH, Wilson G, Yuan PM. Microanalysis of SDSPAGE electroblotted proteins. Appl System Users Bull. 1988;36:125. Ray J, Jones BK, Liebhaber SA, Cooke NE. Glycosylated human growth hormone variant. Endocrinology. 1989; 125:566-8. Carr D, Friesen HG. Growth hormone and insulin binding to human liver. J Clin Endocrinol Metab. 1976;42:484-93. Wilson DM, Benett A, Adamson GD, et al. Somatomedins in pregnancy: a cross-sectional study on insulin-like growth factors I and II and somatomedin peptide content in normal human pregnancies. J Clin Endocrinol Metab. 1982;55:858-61. Martial JA, Hallewell RA, Baxter JD, Goodman HM. Human growth hormone: complementary DNA cloning and expression in bacteria. Science. 1979;205:602-6.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:04 For personal use only. No other uses without permission. . All rights reserved.
Vol. 71, No. 1 Printed in U.S.A.
Identification of Placental Human Growth Hormone as the Growth Hormone-V Gene Expression Product* FRANCIS FRANKENNE, MARIE-LOUISE SCIPPO, JOZEF VAN BEEUMEN, AHMED IGOUT, GEORGE HENNEN Experimental and Clinical Endocrinology, CHU B-23, B-400 Liege; and Laboratorium uoor Microbiologie (J. V.B.), Ledeganckstraat 35, B-900 Gent, Belgium
ABSTRACT. A GH variant of placental origin, placental GH, has recently been shown to replace pituitary GH in maternal serum during pregnancy. Besides, the GH variant (GH-V) gene has been demonstrated to be expressed in the placenta. The similarities between their known properties strongly suggest that the placental GH and the GH-V protein are the same molecular
species. Here we provide final evidence that this is indeed the case by sequence analysis of both the 22K and 25K forms. Furthermore, the 25K form is shown to be glycosylated, while the 22K form is not. Both size variants of placental GH are, thus, likely to reflect the partial glycosylation of a unique peptidic chain. (J Clin Endocrinol Metab 7 1 : 15-18, 1990)
D
URING the second half of pregnancy, pituitary GH is progressively replaced in the maternal serum by a GH analog originating from placenta, distinct from placental lactogen (hPL), and which has been tentatively named placental GH (1-4). It can be distinguished from pituitary GH by its specific pattern of reactivity with two high affinity monoclonal antibodies (MAb) insensitive to physiological concentrations of hPL: the 5B5MAb, recognizing both pituitary and placental GH and the K24-MAb, which only reacts with the pituitary hormone main variant (22K hGH). Its concentration steadily rises to term, then drops sharply to an undetectable level within 1 h after delivery (4). So far, placental GH has only been found in the maternal compartment, where pituitary GH steadily disappears to become undetectable near term (4). Placental GH is, therefore, likely to actively participate in the control, by the placenta, of maternal metabolism in late pregnancy. Placental GH has been partially purified from term placenta through ion exchange and immunoaffinity chromatography (3, 4). Due to some structural unstability of the anti-GH MAb used in the latter step (our unpublished results), the recovery yields were very low, and the described material was contaminated by MAb, which hindered full structural analysis by classical means. A
partial characterization of placental GH has, nevertheless, been achieved. It displays an apparent size dimorphism, being composed of two forms of 22K and 25K, respectively (3, 4). Preliminary data have shown that the latter form is probably glycosylated (5). Both forms share an N-terminal epitope with pituitary GH, but lack an internal epitope (4). Placental GH is also more basic than the pituitary hormone variants with which it competes with an apparent high potency for binding to liver GH membrane receptors (4). The five genes of the hGH-human chorionic somatomammotropin (hCS) family are clustered on the long arm of the chromosome 17 in the following 5' to 3' order: hGH-N, hCS-L, hCS-A, hGH-V, and hCS-B (abbreviated for hGH-normal, hCS-like, and hGH-variant) (6, 7). hGH-N is expressed in the pituitary, yielding a major 22K, a minor 20K as well as posttranslational variants (8), while both CS-A and CS-B encode hCS (hPL), which is massively produced by the placenta (9). By mRNA hybridization with specific oligonucelotidic probes and cDNA cloning from a placental library, we have shown that the GH-V gene, previously considered as silent, is actually expressed in the placenta (10,11). This has been confirmed in a recent comprehensive study (12) showing that the GH-N gene is expressed only in the pituitary, whereas the four other genes are expressed in the placenta. A minor alternative GH-V pre-mRNA splicing, leading to a longer mRNA, has also been described (13). The GH-V protein (11, 14) differs from the main 22 K pituitary variant by 13 amino acid differences evenly distributed along the peptidic chain.
Received June 15,1989. Address requests for reprints to: Dr. Francis Frankenne, Experimental and Clinical Endocrinology, CHU B-23, B-4000 Liege, Belgium. * This work was supported by the Science Policy Programming Department of the Prime Minister's Office and grants from the Belgian FNRS and the Region Wallonne.
15
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:04 For personal use only. No other uses without permission. . All rights reserved.
FRANKENNE ET AL.
16
JCE & M • 1990 Vol71«Nol
carbohydrate-staining technique was also applied on Western blotted material to assess the glycosylation of the 25K form. Materials and Methods Partial purification of placental GH
FIG. 1. SDS-PAGE of a placental GH-enriched fraction, followed by electrotransfer and staining either for protein with amido black (A) or for placental GH immunoactivity using the 5B4-MAb (B).
TABLE 1. Sequence analysis data Step
18
21
Arg 0.51 (His 0.02)
Tyr 1.3 (His 0.25) Tyr 0.39 (His 0.04)
22K PGH 25K PGH
25 Tyr (Phe Tyr (Phe
0.64 0.12) 0.58 0.26)
See Materials and Methods. Recorded increases in amounts (picomoles) of the phenylhydantoin amino acids for the residues attributed to the positions 18, 21, and 25. The increases in phenylhydantoin amino acids corresponding to the GH-N sequence at the same positions are also indicated (in parentheses).
As determined from expression experiments with SV40-transfected eukariotic cells (15,16), the GH-V protein displays a basic isoelectric point and a high potency for binding to GH receptors in addition to an Asn-glycosidebinding site. Also, the first 18-amino acid portion is similar to pituitary GH (14). The similar properties of the GH-V protein and placental GH strongly suggest their identity. Sequence data on placental GH would definitely assess this identity. Since homogeneous placental GH was not available, we attempted to perform its sequencing directly on material bound to Western blot fragments. In addition, a specific 10
1
The enriched placental GH was obtained from placental extract essentially as described previously (4), but with the final immunoaffinity step modified as follows. To 3 mL ReactiGel GF-2000 (Pierce, Rockford, IL) 5 mg purified anti-GH 5b4 (4, 17) in 3 mL 0.1 M NaHCO3, pH 9, were added. The suspension was gently agitated at 4 C for 48 h. The protein solution was then replaced by 3 mL 1 M ethanolamine, pH 9, for 4 h at 20 C. Finally, the resin was sequentially washed using 1 M NaCl, H2O, and 0.2 M phosphate-0.14 M NaCl buffer, pH 7.4 (PBS). To that affinity matrix were added 100 mL PBS
containing 3 g of a crude placental GH fraction. After mild stirring for 16 h at 4 C, the resin was centrifuged, washed with PBS, then poured into a column (0.9 x 5 cm), washed again with 100 mL PBS, and eluted using 3 mL 3 M KSCN in PBS. The desorbed material was finally concentrated, desalted, and lyophilized. Electrophoresis and blotting The partially purified placental GH preparation was submitted to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, 15% acrylamide), then electroblotted using a semidry transfer apparatus (Bio-Lyon, Lyon, France), with the following set of buffers: 0.3 M Tris-20% methanol (anodic 1); 25 mM Tris-20% methanol (anodic 3); 40 mM 6-amino-iVhexanoic acid-20% methanol (cathodic); either A) a polyvinylidene difluoride filter (PVDF; Immobilon, Millipore, Milford, MA) (18) or B) a nitrocellulose (NC) filter was used. After applying 24 V for 2 h, the PVDF filter (A) was stained with 0.1% amido black in 2% acetic acid-20% methanol and then destained in 7% acetic acid-20% methanol, and a NC blot (B) stained for 5B4-MAb-reacting material as described previously (4). On a second NC blot, a specific detection of glycoconjugates was obtained by mild periodate oxidation, followed by staining of the resulting aldehydes using digoxigenin-succinyl-E-amidocaproic acid hydrazide and antidigoxin alkaline phosphatase15
20
25
J 22K PGH :
Pro-Thr- He -Pro-Leu-Ser-Arg-Leu-Phe-Asp-Asn-Ala-Met-Leu-Arg-Ala
25K PGH: Phe-Pro-Thr- De -Pro-Leu
Leu-Phe-Asp-Asn-Ala-Met-Leu-Arg-Ala-Arg
30
1
Leu-Tyr-Gln-Leu-Ala-Tyr-Asp-Thr-Tyr-Gln-Glu-Phe Leu-Tyr
Tyr-Asp-Thr
Gin
GH-V protein: Phe-Pro-Thr-He-Pro-Leu-Ser-Arg-Leu-Phe-Asp-Asn-Ala-Met-Leu-Arg-Ala Arg Arg-Leu Tyr Gln-Leu-Ala Tyr Asp-Thr-Tyr-GIn GH-N protein: Phe-Pro-Thr-lie-Pro-Leu-Ser-Arg-Leu-Phe-Asp-Asn-Ala-Met-Leu-Arg-Ala His Arg-Leu His Gin-Leu-Ala Phe •Asp-Thr-Tyr-GIn CS protein:
VaJ-GIn
Thr Val Pro-Leu-Ser-Arg-Leu-Phe-Asp His •Ala-Met-Leu Gin-Ala- His Arg Ala- His Gln-Leu-Ala He Asp-Thr-Tyr-GIn
FIG. 2. N-Terminal amino acid sequence of the 22K and 25K forms of placental GH compared to the corresponding regions of the GH-N (23), GH-V (14), and CS (10) proteins.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:04 For personal use only. No other uses without permission. . All rights reserved.
PLACENTAL hGH AS THE GH-V GENE EXPRESSION PRODUCT
17
for 26 of the first 32 residues for the 22K form and for 21 of the first 29 residues for the 25K form (Fig. 2). On specific staining for glycoconjugates of Western blotted placental GH (Fig. 3), the 25K form appears as a colored band opposite the 22K form, which is not stained.
Discussion A
A
A
B
FlG. 3. SDS-PAGE of a placental GH-enriched fraction (lane B), followed by electrotransfer and specific staining for glycoconjugates (see Materials and Methods). The darkly stained bands correspond to glycoproteins, whereas nonglycosylated proteins appear as nonstained bands, due to the lesser background staining. Biosynthetic Met-GH was used as negative control (lane A).
labeled antibody (Glycan detection kit, Boeringher, Mannheim, West Germany). Biosynthetic Met-GH (Genentech, San Francisco, CA) was used as the negative control. Sequencing Sequence data were obtained using an Applied Biosystems 477A sequencer loaded with PVDF filter fragments (19) containing either the 22K or the 25K form of placental GH after SDS-PAGE and Western blotting.
Results Partially purified material was first prepared, using improved techniques for both immobilizing the MAb and placental GH binding. That preparation was submitted to SDS-PAGE, followed by an electrotransfer to a PVDF filter that was then stained for protein (Fig. 1A). A parallel Western blot on an analytical sample revealed the position of the placental GH bands through their specific immunoactivities (Fig. IB). Filter fragments containing 22K and 25K placental GH were submitted to automated amino acid sequence analysis. The positions for which no net increase in the amount of phenylthiohydantoin amino acid was obtained compared to the preceding cycle were left unassigned. In Table 1 are plotted the raw data obtained for the critical 18, 21, and 25 positions, which correspond to different amino acids on GH-V and GH-N proteins. Based on the amount of Ile-4, the amounts of 22K and 25K bound to the blot were calculated to be 15 ± 3 and 10 ± 2 pmol, respectively. Taking the presumed losses due to electrophoresis and electroblotting into account (19), this would mean that about 30 and 20 pmol, respectively, were loaded on the gel. In spite of the very low amounts of material bound to the PVDF filters, a clear-cut assignment could be made
Although in each of the sequence runs, some (but different, except for Arg19) positions could not be identified, the data for positions 18, 21, and 25 are sufficient to show that both the 22K and 25K forms have an Nterminal sequence allowing their definitive identification as GH-V gene expression products and not as GH-N gene products. Preliminary evidence of glycosylation of the 25K form was previously given by Concanavalin-A binding experiments (5). The present results, obtained using a totally independent technique, confirm that the 25K form of placental GH is actually a glycoprotein, while the 22K form is not. Besides the GH-V mRNA coding for the 22K GH-V protein, a minor alternative pre-mRNA splicing has been described, which results in a larger GH-V2 mRNA encompassing intron 4 (13). The resulting protein would be of a larger size than 22K GH (230 instead of 191 amino acids), and due to a shift in the reading frame, the amino acid sequence of a large C-terminal portion (residues 127-230) would be totally different from the GH-N or GH-V 22K product. As a consequence, it would be devoid of the JV-glycosylation site present on the 22K GH-V protein. It has recently been reported that the GH-V2 protein was not secreted by a GH-V gene-transfected fibroblastic cell line, whereas the GH-V protein was secreted as three size variants, one nonglycosylated 22K and two glycosylated of 24K and 26K (20). The 25K placental GH displays physiochemical and immunochemical characteristics very similar to the 22K form (4) and has been shown here to be glycosylated. It is, therefore, much more likely to result from translation of the GH-V mRNA and not the alternatively spliced GH-V2 mRNA. Four genes belonging to the GH family are known to be expressed in the placenta: GH-V, CS-A, CS-B, and CS-L (10, 12). So far, the CS-L protein has not been in either the placenta or the circulation, and to our knowledge, nothing is known of its potential GH-like activity. The CS-A and CS-B genes yield mass amounts of hCS (hPL), exhibiting a very low GH-like activity (21). The third gene, GH-V, encodes placental GH, which is likely to be a potent analog of GH (4). These activities probably account for the disappearance of maternal circulating pituitary GH, resulting from inhibition of somatotroph
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:04 For personal use only. No other uses without permission. . All rights reserved.
FRANKENNE ET AL.
18
secretion by increasing levels of serum insulin-like growth factor-I (22). This is consistent with the placenta taking control of the maternal metabolism through expression of the hCS gene, the hGH-V gene, or both.
References 1. Hennen G, Frankenne F, Pirens G, Gomez F, Closset J, El Khayat N. New chorionic GH-like antigen revealed by monoclonal antibody radioimmunoassays. Lancet. 1985;16:399. 2. Hennen G, Frankenne F, Pirens G, et al. A chorionic GH-like antigen: increasing levels during second half of pregnancy with pituitary GH suppression as revealed by monoclonal antibody radioimmunoassays. Int J Fertil 1985;30:27-33. 3. Hennen G, Frankenne F, Closset J, et al. Monoclonal antibody to growth hormone: the discovery of a new variant, human placental growth hormone. In: Forti G, Serio M, Lipsett MB, eds. Monoclonal antibodies: basic principles, experimental and clinical applications in endocrinology. New York: Raven Press; 1985;3O:29-4O. 4. Frankenne F, Closset J, Gomez F, Scippo ML, Smal J, Hennen G. The physiology of growth hormones in pregnant women and partial characterization of the placental GH variant. J Clin Endocrinol Metab. 1988;66:1171-80. 5. Frankenne F, Scippo ML. Circulating forms and partial glycosylation of human placental growth hormone hPGH. Proc of the 70th Annual Meet of The Endocrine Soc. 1988;1129. 6. Owerbach D, Rutter WJ, Martial JA, Baxter JD, Shows TB. Genes for growth hormone, chorionic somatomammotropin, and growth hormones-like gene on chromosome 17 in humans. Science. 1980;209:289-92. 7. Hirt H, Kimelman J, Birnbaum MJ, et al. The human growth hormone gene locus: structure, evolution and allelic variations. DNA. 1987;6:59-70. 8. Barsh GS, Seeburgh PH, Gelinas RE. The human growth hormone gene family: structure and evolution of the chromosomal locus. Nucleic Acids Res. 1983;83:3939-59. 9. Barrera-Saldana HA, Seeburg PH, Saunders GF. Two structurally different genes produce the same secreted human placental lactogen hormone. J Biol Chem. 1983;258:3787-93. 10. Frankenne F, Rentier-Delrue Fr, Scippo ML, Martial H, and
11. 12. 13.
14. 15.
16.
17. 18. 19. 20. 21. 22.
23.
JCE & M • 1990 Vol 71 • N o l
Hennen G. Expression of the growth hormone variant gene in human placenta. J Clin Endocrinol Metab. 1987;64:635-637. Igout A, Scippo ML, Frankenne F, Hennen G. Cloning and nucleotide sequence of placental HG-V and c-DNA. Arch Int Physiol Biochem. 1988;96:63. Chen EY, Liao YC, Smith SH, Barrera-Saldana HA, Gelinas RE, Seeburg PH. The human growth hormone locus: nucleotide sequence, biology, and evolution. Genomics. 1989;4:479-97. Cooke NE, Ray J, Emery JG, Liebhaber SA. Two distinct species of human growth hormone-variant mRNA in the human placenta predict the expression of novel growth hormone proteins. J Biol Chem. 1988;263:9001-6. Seeburg PH. The human growth hormone gene family: nucleotide sequences show recent divergence and predict a new polypeptide hormone. DNA. 1982;82:239-249. Plavlakis GN, Hizuka N, Gorden PH, Seeburg PH, Hamber DH. Expression of two human growth hormone genes in monkey cells infected by simian virus 40 recombinants. Proc Natl Acad Sci USA. 1981;78:7398-402. Hizuka N, Hendricks CM, Pavlakis GN, Hamer DH, Gorden P. Properties of human growth hormone polypeptides: purified from pituitary extracts and synthesized in monkey kidney cells and bacteria. J Clin Endocrinol Metab. 1982;55:545-50. Gomez F, Pirens G, Schaus Ch, Closset J and Hennen G. A highly sensitive radioimmunoassay for human growth hormone using a monoclonal antibody. J Immunol. 1984;4:145-57. Matsudaira P. Sequence from picomole quantities of proteins electroblotted only polyvinylidene difluoride membranes. J Biol Chem. 1987;262:10035-8. Yuen SW, Chut AH, Wilson G, Yuan PM. Microanalysis of SDSPAGE electroblotted proteins. Appl System Users Bull. 1988;36:125. Ray J, Jones BK, Liebhaber SA, Cooke NE. Glycosylated human growth hormone variant. Endocrinology. 1989; 125:566-8. Carr D, Friesen HG. Growth hormone and insulin binding to human liver. J Clin Endocrinol Metab. 1976;42:484-93. Wilson DM, Benett A, Adamson GD, et al. Somatomedins in pregnancy: a cross-sectional study on insulin-like growth factors I and II and somatomedin peptide content in normal human pregnancies. J Clin Endocrinol Metab. 1982;55:858-61. Martial JA, Hallewell RA, Baxter JD, Goodman HM. Human growth hormone: complementary DNA cloning and expression in bacteria. Science. 1979;205:602-6.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:04 For personal use only. No other uses without permission. . All rights reserved.
