HAL Id: hal-00899027https://hal.archives-ouvertes.fr/hal-00899027

Submitted on 1 Jan 1989

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Growth hormones. II. Structure–function relationshipsN. Chêne, J. Martal, P. de la Llosa, J. Charrier

To cite this version:N. Chêne, J. Martal, P. de la Llosa, J. Charrier. Growth hormones. II. Structure–function relation-ships. Reproduction Nutrition Development, EDP Sciences, 1989, 29 (1), pp.1-25. �hal-00899027�

Review

Growth hormones. II. Structure—function relationships

N. Chêne J. Martal P. de la Llosa J. Charrier

Unité Endocrinologie de l’Embryon, Station de Physiologie Animale, INRA, 78350 Jouy-en-Josas,France;2 Laboratoire des Hormones polypeptidiques, CNRS, 91118 Gif sur-Yvette, France;3 Station de Physiologie Animale, INRA-ENSA, Place Viala, 34060 Montpellier Cedex, France

(received 14-6-1988, accepted 27-9-1988)

Summary ― Localization of the somatotropic activity of growth hormones from several species andfrom different organs was attempted using different approaches. Sequences were compared in orderto detect one or several regions with a common homology. The technique of peptide recombinantsas well as chemical changes affecting some amino acids was applied to these hormones; the biolo-gical function in vivo of growth or binding to somatotropic receptors was then estimated. The fewdata available on biosynthetic molecules and secondary structures of natural growth hormones arereported. This study indicates the somatotropic function of particular sites.

growth hormones ― somatotropic activity - localization

Résumé ― Hormones de croissance. Il. Relations structure-fonction. L’étude des relationsstructure-fonction d’hormones à activité de croissance, d’origines hypophysaires (GH) ou placen-taires (CS ou PL) et d’espèces variées, nous a conduit à formuler quelques hypothèses. Uneséquence particulière et restreinte est-elle suffisante pour révéler une activité hormonale ? Existe-t-il un ou plusieurs sites biologiquement actifs impliqués dans la liaison hormone―récepteur ? Quelleest l’incidence de la conformation moléculaire sur la fonction somatotrope ? Pour tenter derépondre à ces questions, nous avons proposé les approches suivantes :- l’étude comparée des séquences d’hormones de croissance et de variants ainsi que des don-nées, encore trop restreintes, sur la structure tridimensionnelle de quelques hormones nous a per-mis de mettre en évidence, d’une part, des régions d’homologie commune et de dégager, d’autrepart, des déterminants structuraux caractéristiques de domaines à vocation somatotrope;- l’activité biologique d’hormones de croissance soumises à l’action d’enzymes protéolytiques oudont certains résidus d’acides aminés ont été altérés par des modifications chimiques, a été esti-mée soit in vivo, soit par liaison à des récepteurs somatotropes;- la méthode originale utilisant les recombinants peptidiques entre hormones pourvues ou nond’activité de croissance a été appréciée pour la localisation de sites actifs, neutres, voire mêmeinhibiteurs dans l’activité de croissance;- l’emploi de peptides synthétiques, d’anticorps monoclonaux et de molécules biosynthétiques acomplété cette étude.

hormones de croissance - activité somatotrope ― localisation

Introduction

Study of structure―function relationshipsmust be carried out on physically andchemically homogeneous as well as onbiologically active products. Increasingsophistication in purification methods,such as HPLC (high performance liquidchromatography), allows the highly puri-fied polypeptide hormones required forsuch studies to be obtained. Elucidationof the primary structure of proteins hasbecome easier with the availability of themicro-sequencing devices which requireonly very small amounts of product, onthe nanomole scale. During the last fewyears, prediction of secondary and tertiarystructures has improved, and allowsassessment of the spatial conformation ofpolypeptides. The criteria used to definebiological activities are based on the wellknown multiple action of GHs (see&dquo;Growth Hormones. I. Polymorphism&dquo;;Charrier and Martal, 1988, Tables II and

III).

Growth-promoting activities were

demonstrated in vivo in hypophysec-tomized rats by the tibia test (Greenspanet aL, 1949) or the weight gain test (Wil-helmi, 1973), and in vitro through hor-mone affinity to hepatocyte cellular mem-branes (Tsushima and Friesen, 1973).Insulin activity was demonstrated by thebinding of hormones to adipocyte cellularmembranes (Posner et al., 1974; Faginet al., 1980) or lymphocytes (Lesniak e tal., 1973). Such binding is used in theso-called radioreceptor assay (RRA) to

demonstrate and quantify biological activi-ties and is easier than the conventionaland time consuming in vivo methods.

Somatotropic effect also refers to somato-medin or insulin-like growth factor (IGFS)stimulation. Human placental lactogenand growth hormone increase IGF-I onhuman fetal pancreas in tissue culture

(Swenne et al., 1987).

The use of monoclonal antibodies spe-cific for an epitope leads to a better know-ledge of immunological determinants.

Analysis of immunological activity siteswill not be studied here. However, inhibi-tion of biological activities may be ob-tained using monoclonal antibodies. Syn-thesis of polypeptide hormones similar tonatural hormones is now possible as aresult of advances in biotechnologicalmethods.

The purpose of the present paper is tostudy structure―function relationships ofgrowth hormones of pituitary (GHs) andplacental origin (placental lactogen [PL] orchorionic somatomammotropin [CS] andhuman placental growth hormone [hPGH])from several species. Some of these hor-mones possess another biological acti-

vity : for example, hGH also has a lacto-genic activity. Some placental hormoneshave either only a lactogenic activity, suchas human (hPL), rat (rPL) and mouse(mPL) placental lactogen (Friesen, 1965;Robertson and Friesen, 1975; Tala-

mantes, 1975), or both growth and lacto-genic activities, such as bovine (Forsyth,1973; Bolander and Fellows, 1976; Kellyet al., 1976), caprine (Becka et aL, 1977;Currie et al., 1977), ovine (Handwerger e tal., 1974; Chan et al., 1976; Martal and

Djiane, 1976; Martal, 1978), simian

(Shome and Friesen, 1971) and guinea-pig (Talamantes, 1975, Kelly ei al., 1976)chorionic somatomammotropins (CS).

The study of hormone structure―func―tion relationships gives rise to many ques-tions. The question may be raised as towhether there exists one or several biol-

ogically active sites, as already shown insome enzyme proteins. The following pre-liminary hypotheses have been put for-ward on structure―function relationships.

Are there particular regions of a hor-mone liable for specific biological activityand what is the function of the hormonal

molecule ? In order to answer these ques-tions, we made a comparative analysis ofalready known GH primary structures.

Some recently obtained information on

molecular conformation of some GHs mayalso constitute an original approach to theunderstanding of structure―function

relationships. Proteolytic enzymes are uti-lized in the so-called peptide recombinantmethod, which allows localization of activesites of growth hormones in the same

way as utilisation of polypeptide ana-

logues. Do the chemical modifications

altering some amino acid residues affectthe biological activity of these hormones ?Does mutagenesis modify biological hor-monal activity ?

Analysis of invariants

Growth hormones

Pituitary growth hormones are globularholoproteins consisting of about 191amino acids with a predominantly a-helixstructure (= 50%). They contain a singletryptophan residue at position 86 and 4cysteine residues which form a small loopin the COOH-terminal part (Cys 182―Cys189) and a large one in the NH2-terminalpart (Cys 53―Cys 165) (Table I). In all thealready identified growth hormones, theNH2-terminal end mainly consists of

phenylalanine or sometimes alanine resi-dues.

To our knowledge, growth hormone

sequences have already been determinedin several species: horse (Zakin et al.,1973; Daurat-Larroque et al., 1977), man(Bewley et al., 1972), sheep (Li et al.,1972), pig (Seeburg et al., 1983), beef(Wallis, 1973; Graf and Li, 1974a), rat

(Wallis and Davis, 1976; Barta et al.,1981), chicken (Souza et al., 1984),mouse (Linzer and Talamantes, 1985),

monkey (Li et al., 1986), sei whale (Pan-kov et al., 1982) and fin whale (Tsuboka-wa and Kawauchi, 1985). The method ofcomparison of 8 GH sequences consistedof lining up amino acids while complyingwith the largest position identity betweenone another according to Dayhoff (1976):human growth hormone (hGH), monkeygrowth hormone (MGH), ovine growthhormone (oGH), bovine growth hormone(bGH), equine growth hormone (eGH),chicken growth hormone (cGH), rat

growth hormone (rGH) and mouse growthhormone (mGH). In this text, amino acid

numbering of GH was based on primarysequence of hGH (Table I).

Some characteristics can be deducedfrom this comparison. There exists a largesequence similarity between human andmonkey GHs (Li et al., 1986), since only 4amino acid residues differ at positions105, 107, 133 and 173 due to the muta-tion of only one codon nucleotide.

hGH exhibits a 25% difference in theresidue composition as compared to bGH(Bewley and Li, 1970; Wallis, 1975). In

contrast, ovine and bovine growth hor-mones are quite similar since only 1% oftheir residues are different, indicating aclose relationship between these 2 spe-cies (Wallis, 1975). A 77% similarity wasobserved by Souza et al. (1984) betweenchicken and bovine GH (Wallis, 1973).Analysis of rGH and mGH sequencesshows a 95% and 92% identity, respecti-vely, with that of ruminants. The rat andmouse GH sequences only differ from oneanother in 2 amino acids at positions 11and 186. Only avian GH possesses anisoleucine residue as COOH-terminalamino acid, while phenylalanine is com-

mon to all the considered mammalianGHs. It has been suggested that PRLsand GHs are derived from a common ini-tial peptide consisting of 25―50 amino

acids, which by successive duplicationswould have generated these hormones

(Niall et al., 1971). Repetition of a modelwith a very close amino acid compositionis found in hGH at positions 15―32,93―110, 128―147, 163―180 (Table 11and Table 111). These observations regar-ding the existence of 4 homologous inter-nal regions suggest that they might beinvolved in hormonal biological activities.

These regions do not always corre-

spond to particularly well conserved seg-ments of GHs and PRLs. Thus, examina-tion of 7 mammalian prolactin sequencesand 9 growth hormones (of which 8 arefrom mammals) and analysis of either

homologous or closely related invariantamino acids led to identification by Nicollet al. (1986) of several high identity clus-ters separating poorly homologous re-

gions. These clusters are located in frag-ments 8―24, 51-68, 69―86, 114―125and 158―167. Only 2 of them partlycorrespond to 2 regions identified by Niall,including residues 15―32, 163―180

(Table II).Recently, Kawauchi and Yasuda

(1987) have compared amino acid

sequences of fish, chicken and mamma-lian GHs and PRLs, according to 204

alignment positions. They have delinea-ted 4 highly conserved domains for GHsin alignment positions : A (5―37), B

(49―89), C (104―126) and D (138―181) (Tables I I and I I I).

Three of these domains, A, B and C,partially overlap those of Nicoll, and onlyA and D imperfectly correspond to the

patterns identified by Niall (15―32 and163―180). It would be interesting to

determine whether some of these highlyconserved domains possess a specificsomatotropic-like bioactivity.

Table II gives a consensus sequenceof the GH family (Con GH). If we take intoaccount functional homologies accordingto Dayhoff (1976) between some aminoacids such as Arg-Lys, Arg-Glu, Leu-Ile-

Val, Ser-Thr, and amino acids with

frequent or very frequent mutation suchas Asp-Asn, Met-Val, Asn-Ser, His-Asn,Val-Ile, Gln-Arg, His-Gin, Ile-Phe, Met-Arg,we obtain a molecular archetype with a70% common primary structure (Martal,1980).

The 20K variant of hGH

This variant was first noted by Lewis e tal. (1978). These authors emphasized thedeletion of amino acids 32!6 (Lewis e taL, 1980). The features of this variantwere reviewed by Charrier and Martal

(1988). It exerts in vivo somatogeniceffects analogous to the 22K form of hGH,as shown by the rat tibia test (Spencer e tal., 1981) and the weight gain test (Lewiset al., 1981; Kostyo et al., 1985). It alsostimulates somatomedin production(Spencer et al., 1981 Lactogenic effectsof the 20K hGH are similar to those ob-served in the presence of the 22K com-

pound in the pigeon crop-sac assay(Lewis et al., 1978). On the other hand,the very marked decrease in the insulin-like activity was observed by Frigeri et al.(1979) using a purified preparation of 20Kvariant hGH, and by Kostyo et aL (1985)using a biosynthetic methionyl hGH-20K(Met-hGH 20K). Moreover, the latter havedemonstrated the preservation of thechronic diabetogenic effect of this 20Kvariant of hGH as probably due to theinsulin residual activity. It should be men-tioned that the 20 and 22K forms of hGH

represent an interesting model for stu-

dying structure―function relationships,since a deletion of 15 amino acidsinduces differences in biological activity.

Sigel et al. (1981) and Wohnlich andMoore (1982) have compared the biologi-cal activity by means of hormone-receptorbinding. In a radioreceptor assay, wherethe 22K hGH is used as tracer, displace-ment by the 20K relative to the 22K

in the presence of liver membranes from

pregnant doe rabbits, and between22―53% in mammary gland membranesfrom pregnant doe rabbits. In contrast,Closset et al. (1983) and Smal (1986)reported a competitive capacity of the 20KhGH in the presence of the 22K hGHused as tracer ranging from 50―100%

compared to the shift obtained with the22K hGH in both receptor systems used(liver and mammary gland of pregnantdoe rabbits). These results clearly showthat the 20K hGH exhibits comparablebehaviour to the 22K hGH towards these

receptors. Purity and homogeneity of the20 and 22K hGH preparations confirmthese results. It may be concluded thatnormal hGH and its 20K variant havesimilar binding activities towards lactoge-nic or somatotropic receptors as well asidentical biological activities in vivo.

In order to obtain further knowledge onthe binding affinity of the 20 and 22K com-pounds for insulin-like receptors, Smal

(1986) used hGH specific receptors isola-ted from cultured human lymphocytes(line IM-9) and rat adipocytes. Lympho-cyte receptors similarly recognize the 22Kform and the 20K variant with, neverthe-less, a slightly lower affinity for the latter.However, the 20K functions as an agonistwith a very low affinity (= 3%) for rat adi-pocyte receptors as compared to the 22K.Thus, although the 20K hGH is not totallydeprived in vivo of insulin activity (Frigeriet al., 1979; Kostyo et al., 1985), it pos-sesses a very small intrinsic insulin acti-

vity in vitro. Thus, peptide 32―46 is mostlikely involved in the insulin activity; and itshould be noted that the basic amino acidat position 41 is particularly conservativein each GH (Table I).

Placental growth hormoneshPGH. A placental human growth hor-mone has recently been noted by Hennen

et al. (1985) using monoclonal antibodies.This hormone exhibits an immunologicalcross-reaction with the 20K variant, butfew data are as yet available to assess itsstructure-function relationships.Placental hormones. Some placental hor-mones only have a lactogenic activity,such as human placental lactogen hor-mone (hPL), also called hCS (Florini e taL, 1966; Tsushima and Friesen, 1973),rat (Robertson and Friesen, 1975), mouse(Talamantes, 1975; Colosi et aL, 1982)and hamster (Kelly et al., 1976; Southardet al., 1986) placental lactogen hormones.Others have a lactogenic function asso-ciated with a significant somatotropicfunction, such as ovine (Chan et al., 1976;Martal and Djiane, 1976; Martal, 1978),bovine (Fellows et aL, 1975; Bolander andFellows, 1976), caprine (Becka et al.,1977; Currie et aL, 1977), simian (Shomeand Friesen, 1971 and guinea-pig (Kellyet al., 1976) chorionic somatomammo-

tropin hormones.

Unfortunately, only the , sequence of

placental hormones deprived of growthactivity has been elucidated, such as hPL(Sherwood et al., 1971; Li et al., 1973)which exhibits an 85% identity and a 95%similarity with hGH, as well as rPL (Duck-worth et al., 1986) and mPL (Jackson e taL, 1986). The sequence of the latter twohormones was estimated from that of their

complementary DNA. Analysis of these

sequences showed a 50% homology ofplacental hormones with prolactins and30% with growth hormones from corres-ponding species.

Although it has been established that

placental hormones belong to the super-family that includes PRLs and GHs, verylittle information is available on the

sequence of amino acids constitutinggrowth-promoting placental hormones, sothat it is not possible to make structuralcomparisons.

Prolactin-related proteins. Recently, Lin-zer and Nathans (1984) have identified inmouse placental tissue a prolactin-relatedglycoprotein exhibiting a growth-promot-ing activity. It was termed proliferin be-cause of its stimulating effect on prolifer-ation and differentiation of mouse culturedcells (Lee and Nathans, 1987).

Analysis of its primary structureshowed a 31% and a 46% identity ascompared to that of mouse and bovine

prolactins, respectively. Moreover, itcontained 2 tryptophan residues and 6

cysteine residues, as in most mammalianprolactins.

Molecular structure

It is interesting to determine whether

secondary structures of growth hormonesshare a number of common features andto what extent they are related to somato-tropic function.

Disulfide bonds

Mills and Wilhelmi (1965) reported thatafter partial reduction of bGH (one disul-fide bond out of two) using sodium sulfite,this hormone retained its biological poten-cy in vivo.

Bewley et al. (1968) demonstrated thatthe existence of 2 hGH disulfide bonds isnot essential for the expression of its bio-logical activity. A marked somatotropicactivity in vivo has been observed by Li

et al. (1977) with a recombination of2 hGH fragments containing the NH2-ter-minal part (1―134) non-covalently boundto the COOH-terminal part recovered afterreduction of disulfide bonds, and blockingcysteine residues by carbamidometh-

ylation.

Secondary structure

A method for predicting the secondarystructures of proteins has been devisedby Chou and Fasman (1978). Jibson and

Li (1979) used it for predicting the secon-dary structure of different human antipi-tuitary hormones such as hGH. Theyreported an overall a-helix content for thishormone. Bewley and Li (1986), throughstudies on circular dichroism spectra,concluded that human and monkeygrowth hormones exhibited a similar a-helix content of 55 ± 5%. The 3-dimensio-nal structure of methionyl porcine somato-tropin (MPS) has been determined byAbdel-Meguid et al. (1987) by use of theX-ray diffraction technique. It consists

mainly of 4 antiparallel a-helices whichaccount for 54% of the amino acid resi-dues. The amino acid content within a-helices represents the predominantlyinvariant amino acid residues of other

growth hormones. The hGH secondarystructure has been comprehensively stu-died by Nicoll et al. (1986) and the diffe-rent a-helix, (3-turn, B-sheet and randomcoil forms have been quantified and posi-tioned along exons II, III, IV, V. Recently,recombinant human growth hormone

(r-hGH) has been crystallized by Joneset al. (1987).

Evaluation of the secondary structureis also compatible with the hydropathyprofile obtained by the method of Kyteand Doolittle (1982). A hydropathy indextakes into account hydrophilic and hydro-phobic features of amino acid side-chains.These authors studied the tridimensionalstructures of some proteins by X-ray crys-tallography. There was a good correlationbetween the inner regions shown by theexperiments and the hydrophobic areason the one hand; and between the outerregions determined by the experimentsand the hydrophilic areas on the other.Nicoll ef al. (1986) examined the nature ofthe secondary structure and hydropathyprofiles, studying either the hormone-spe-cific regions or the regions with a highdegree of amino acid homology. Theyobserved a good correlation between

hydropathy profiles and secondary struc-tures of regions with a high degree of

homology. Nevertheless, although hor-

mone-specific regions possess identical

hydropathy profiles, they have noticeablydifferent secondary structures. Thus, hGHand hPL, with nearly 95% sequencehomology, exhibit very similar hydropathyprofiles. The only differences concern 2regions (59-87) and (100―117). Thefirst one belongs to a region with commonhomology between GH and PRL and thesecond one is involved in a hormone-spe-cific region (Fig. 3). Only one amino aciddiffers in the peptide 59―67 of hCS andhGH molecules located at position 64.

hCS Methionine is replaced by a basicamino acid arginine in hGH. It should be

pointed out that all the primary structuresof GH studied (Table I) present a basicamino acid (arginine or lysine) at position64; this fact suggests that this amino acidhas a significant role in growth-promotingactivity. The respective hCS (1) and hGH(2) sequences are the following from posi-tion 100 to 117 :

(1) Asn-Leu-Val-Tyr-Asp-Thr-Ser-Asp-Ser-Aps-Asp-Tyr-His-Leu-Leu-Lys-Asp-Leu

(2) Ser-Leu-Val-Tyr-Gly-Ala-Ser-Asn-Ser-Asp-Val-Tyr-Asp-Leu-Leu-Lys-Asp-Leu

Dissimilar amino acids at positions104, 107, 110 and 112 correspond to no-ticeably different charges. All growth hor-mones (Table I) contain glycine at position104, valine at position 110, glutamic oraspartic acid at position 112 like hCS.

Thus, amino acids 104, 110 and 112 ofhCS might explain the absence of growthactivity in this molecule.

Structure of biologically active frag-ments of bGH has been defined by Chenand Sonenberg (1977); it was determined

by methods of fluorescence spectroscopyand measurements of circular dichroism

spectra. Predictive analysis of helix, B-sheet and f3-turn regions in bGH, accord-ing to these authors, is shown in Fig. 1.

The small biologically active peptideAll (96-133) of bGH shifts from an a-helixto a random coil structure between pH 5and 10.

Segments Al (1―95, 134―191 ) whichhave a low biological activity mainly ex-hibit a very rigid (3-sheet structure. On theother hand, Chou and Fasman’s predic-tion method determined a 45% a-helixstructure in fragment All and a 34% a-helix and a 22% f3-sheet structure in frag-ment Al.

Studies on the secondary structure

provided different results according to themethods used (CD spectra or Chou andFasman prediction). The latter seems tobe more reliable for studying native globu-lar proteins at neutral pH. The method ofChen and Sonenberg is better adapted tothe study of peptides in solution at differ-ent pH and ionic strengths.

Thus, there are discrepancies regard-ing the structural determinants of the bio-logical activity of these proteins. However,the spatial proximity of lysine residues 70,115 and 134 more particularly involved insomatotropic-like activity, as will be seenafterwards (Fig. 3), must be emphasizedwhen analysing the representation of

secondary structure of bovine growth hor-mone (Fig. 1). ).

Enzymatic hydrolysis

Numerous experiments have been carriedout with GH submitted to proteolyticdigestion in order to define biologicallyactive sites.

The action of plasmin, thrombin, subti-lisin and trypsin on the hGH molecule hasbeen widely studied. hGH partial hydroly-

sis by trypsin does not affect its growthactivity (Li and Samuelson, 1965).

After treatment of hGH with plasmin,Reagan et al. (1975) and Li and Bewley(1976) isolated 2 peptide fragments(1―134) and (141―191) linked only by adisulfide bridge between Cys 53―Cys165. This new polypeptide retains a

growth-promoting effect in vivo (weightgain and tibia tests in hypophysectomizedrats) and a diabetogenic effect in the pan-createctomized rat similar to that of thenative hormone. Moreover, only the NH2-terminal fragment (1―134) obtained afterreduction and S-carbamidomethylationexhibits a low somatotropic activity in n

vivo (Reagan et al., 1978).The restrictive cleavage of hGH by

thrombin at Arg―Thr bonds (134―135)was first observed by Li and Graf (1974),Graf et al. (1976) and confirmed by Millset al. (1980). It gives rise to 2 peptides,1―134 and 135―191, which after reduc-tion and S-carbamidomethylation producea new recombinant. The latter displays asignificant activity in the rat tibia test.

Similar experiments have been per-formed by Reagan et al. (1981) on hGHafter treatment with thrombin. The non-

covalent recombinant obtained between

peptides 1―134 and 135―191 developsa noticeable growth activity estimated bythe weight gain of hypophysectomizedrats and equivalent to 35% of the nativehGH activity. Its binding to liver plasmamembranes of female rats only represents20% compared to that of native hGH.

These lower activities may result eitherfrom structural modifications or from a dif-ferent half-life as compared to that of na-tive hGH in experiments in vivo. Attemptsto recombine peptide 42―134 (obtainedby digestion of peptide 1―134 by plas-min) either to peptide 135―191 or to pep-tide 141―191 were unable to induce a

positive response in the rat weight gaintest (Mills et al., 1978).

More recently, Graf et aL (1982) usingthe limited action of trypsin on hGH didnot observe any modification in the biolo-

gical activity in spite of the excision of asegment of 11 amino acids at positions135―145, during radioreceptor assays(on rabbit liver membranes) and tibia testin hypophysectomized rats.

The proteolytic action of subtilisin, simi-lar to that of plasmin, gives rise to hGHmolecules characterized by the deletion ofamino acids 140―146 (Lewis et al.,1977). Another modified form of hGH,deprived of residues 135―146, termed

a3, has been reported by other authors(Singh et al., 1974). These 2 forms ofhGH exhibit a growth activity (evidencedby the tibia test) which is even higher thanthat of native hGH.

Analysis of the results regarding pro-teolytic effects on hGH indicates conser-vation of biological activity in most cases.It seems to be well established that resi-dues 135―146 of hGH are not necessaryto this activity.

In order to determine the distribution ofone or more active sites of growth hor-

mones, some authors have used therecombinant method between differenthormone molecules. Of particular interestfor the study of structure―function rela-

tionships are the experiments of Bursteinet al. (1978) and Russell et al. (1981). Ahybrid is formed by the peptides obtainedafter plasmin action on hGH and hPL : anNH2-terminal peptide of hGH (1―134)linked to a COOH-terminal peptide of hPL(141―191 ) by a disulfide bridge betweenCys 53 and Cys 165. This hybrid hasbeen tested for its binding capacity to

mammary gland or liver receptors. Thisrecombinant, whose NH2-terminal struc-

ture corresponds to that of hGH, pos-sesses the hGH activities, i. e. a bindingcapacity to somatotropic and lactogenicreceptors and a secondary and tertiarystructure similar to that of hGH as shown

by circular dichroism spectra. On theother hand, the recombinant obtained bylinking NH2-terminal peptide of hPL

(1―134) to the COOH-terminal peptide ofhGH (141―191) does not display the

somatotropic activity characteristic of thehGH molecule. An NH2-terminal fragmentof hGH (1―134) recombined to an hPLfragment (141―191 ) (Li, 1978) shows50% growth activity in vivo (rat tibia test)as compared to an hGH recombinant

composed of 1―134 and 141―191 partsobtained under the same conditions, i. e.

non-covalently linked (after reduction andS-carbamidomethylation of cysteine resi-

dues) (Li et al., 1977) (Fig. 2). Thisdecreased activity seems to be attrib-

utable to the existence of 2 differentamino acids between hPL and hGH andto be located in the last third of the mole-cule. Histidine 153 and methionine 179 ofhPL replace aspartic acid and isoleucinefrom hGH, respectively. Isoleucine (179)and methionine (179) are both hydropho-bic amino acids, whereas basic histidine(153) and aspartic acid (153) present adifferent charge. This difference might

explain the marked inhibition of the hybridhormone hGH―hPL growth activity.

Therefore, the somatotropic and struc-tural properties of hGH are related to theNH2-terminal part of the hormone, in par-ticular peptide 1―134.

Graf and Li (1974b) and Yamasaki andShimanaka (1975) have shown that a

fragment composed of 38 amino acidresidues (96―133) isolated from bGH oroGH following restricted trypsin hydrolysisexhibits significant growth activity in vivo.

Thrombin action on oGH is character-ized by a cleavage between residues 133and 134 (Arg-Ala). Two peptides (1―133and 134―191 ) are obtained. According tothe results of Graf et al. (1976), thesepeptides are totally inactive in the rat tibia

test, whereas peptide 96―133 of oGHobtained by trypsin retains its biologicalpotency in the same test. It seems thatthrombin action more drastically modifiesthe tertiary structure of oGH than trypsinand produces different metabolically ac-tive fragments.

Inactivity of peptide 1―133 of oGH

seems to contradict the previous experi-ments of Li and Graf (1974). Using hGHplasmin digests, these authors observedthat only the 1―134 NH2-terminal peptidemanifested a 10―20% biological activityin vivo as compared to native hGH. Thesedifferences in the activity of NH2-terminalfragments of oGH and hGH are probablydue to differences in the structure of the

peptides obtained from human and rumi-

nant (ovine and bovine) GHs (Bewley andLi, 1972; Li, 1972).

Li (1975) has attempted to localizemore accurately the biological activity sitein peptide 1―134 of hGH. Cyanogen bro-mide gives rise to peptide 15―125 devoidof secondary and tertiary structure, andwhich nevertheless retains a growth-pro-moting activity in the rat tibia test as wellas a lactogenic activity in the pigeon crop-sac assay. Thus, part 15―125 of hGHmolecule retains most of the biologicalactivities.

Liberti (1981) observed a small soma-tomedin-like activity due to 87―124 pepti-de of bGH, as shown by the stimulation ofsulfate and thymidine incorporation by thecostal cartilage of the hypophysectomizedrat, while native bGH is devoid of this ac-

tivity. Liberti and Durham (1983) haveobtained 1―133 peptide after thrombinaction on bGH. This peptide exhibits amarked somatomedin-like activity similarto that of peptide 96―133 (Liberti and Mil-ler, 1978). These experiments clearly sug-gest that residues 96―124 are essentialfor the expression of the somatomedin

activity. Nevertheless, Li and Bewley(1976), and Reagan et al. (1979) havedemonstrated that total expression of

growth-promoting activity requires the pre-sence of the COOH-terminal peptide.

Synthetic peptides

Very few experiments have been per-formed on hGH synthetic peptides. Chille-mi and Pecile (1971) have measured in nvivo the biological activity of 2 syntheticpeptides obtained by Merrifield’s tech-

nique (1969) (a monotetracontapeptide81―121 and a ditriacontapeptide 122―

153). Each peptide provokes the thicken-ing of the rat tibia cartilage. This test also

shows a remarkable synergy of actionbetween both polypeptides. Nevertheless,the potency of these synthetic peptidesremains very low, which emphasizes therole of the tertiary structures conservationof growth hormones.

Blake and Li (1973) have reported agrowth-promoting activity of a syntheticpeptide : N-a-acetyl-hGH (95―136) de-tected by the rat tibia test. Another locali-zation of somatotropic activity has beendemonstrated by Retegui et aL (1982),using monoclonal antibodies specific for 3hGH synthetic peptides: 19―128,73―128, 98―128. They investigated theinterference of these different anticlonalantibodies with the binding of hGH to its

receptors. The highest inhibition of hGHbinding to receptors occurs when usingthe antibody directed against the 98―128peptide.

Results of this work suggest that thehGH binding site to liver receptors is loca-ted in the NH2-terminal part of the hor-mone and in particular at positions98―128.

A growth-promoting activity test moresensitive although more partial than theconventional test on the rat tibia was used

by Morikawa et al. (1984). It measures theeffect of 2 synthetic peptides of hGH,1-43 and 32―46, on the conversion of3T3 preadipocytes into adipose cells. It

does not reveal any activity, whereas wenoticed the important insulin-like effect ofpeptide 32―46 of the 20K hGH-variant

binding to adipocytes. According to theseresults, the whole molecule does not

seem necessary for the expression of anygrowth-promoting activity.

Finally, localization of the sites respon-sible for somatotropic activity, in particularthat of hGH, has been defined as follows:a global localization in the NH2-terminalpart (1―134 peptide) on the one hand,

and a more restricted localization in pep-tides 15―125, 81―121 or 122―153 or95―136 on the other. Moreover, the bin-ding capacity of hGH to liver receptorsseems to be more particularly supportedby peptide 98-128.

The hypoglycemic action of growthhormone has been reported by Milmanand Russell (1950). Ng et al. (1974) firstattributed the insulin-like effects to theamino-terminal (1―15) region of the hGHmolecule. In particular, the synthetichuman growth hormone fragment (hGH4--15) stimulates in vivo and in vitro 2-

deoxyglucose uptake in rat adipocytes(Ng and Harcourt, 1986). These authorsexplain the hypoglycemic action of hGHby an interaction of the fragment 4―15 5

with the plasma membranes of the targetcells (adipocytes and hepatocytes) indu-cing the release of a cellular mediatorwhich would enhance insulin binding andhexose transport (Ng et al., 1985). A long-er peptide, hGH 1―43, was found to

occur naturally in pituitary extracts (Singhet al., 1983) in significant amounts (atleast 12 gg/gland) (Frigeri et al., 1988).These authors partially reinforce the

hypothesis of Ng et al. in showing thatthis fragment hGH 1―43 is able to en-hance the in vitro sensitivity of adiposetissue of yellow obese mice to insulinaction. Insulin-stimulated glucose oxida-tion was enhanced over 100% withoutserum insulin concentrations beingincreased.

Let us recall that 20K hGH, whichlacks fragment 32―46, exhibits little or noearly insulin-like actions (Frigeri et al.,1979; Smal et al., 1986). One implicationof these observations is that most of theinsulin-like properties of the native hor-

mone could reside in the deletion peptide.Stevenson et al. (1987) tested this dele-tion peptide hGH 32―46 in the consciousdog and found that it increased glucose-induced insulin secretion, which in turn

enhanced tissue uptake of glucose. Rud-man and Stebbing (cited by Stevenson e tal., 1988) showed in the rat that the smal-ler peptide hGH 32―38 is even several-fold more active than the deletion peptideitself.

Thus, it appears that insulin-like prop-erties of GH are limited to the N-terminal

sequence of the molecule. However, thedata are not clear enough to attribute apure function to a cluster of few amino

acids, since the synthetic hGH 1―43 ismore potent than any of the shorter pep-tides studied in enhancing insulin-stimula-ted action.

Chemical modifications

Reduction of disulfide bonds

This reduction followed by alkylation ofSH groups of a growth-promoting hor-mone does not modify its biological ac-

tivity, as shown by Li and Bewley (1976)on a plasmin-treated hGH. The bindingcapacity to liver membranes of other

growth hormones has been estimatedafter submitting the S-S bridges to sodiumborohydride. In these experiments, the

bindings of hGH, oCS (ovine chorionic

somatomammotropin or ovine placentallactogen) and bGH were almost similar tothose of their respective native hormones(Chdne et aL, 1984; Martal et al., 1985).

A partial reduction of bGH in the pre-sence of dithiothreitol followed by iodo-acetamide alkylation of cysteine residues182 and 189 did not affect its growth ac-tivity (Graf et al., 1975).

Punctual chemical modifications

Punctual chemical modifications havebeen applied to some amino acids. An

example of this is the work of Cascone e t

al. (1980) on oxidation of methionine at

positions 4, 125, 150 and 180 of bGH.These authors did not report any reduc-tion in the growth activity of this modifiedhormone in the body weight gain test ofhypophysectomized rats.

In contrast, nitration of 6 tyrosine resi-dues out of the 8 present in hGH usingtetranitromethane leads to a reduction of= 60% in the in vivo growth activity com-pared to native hGH. Moreover, the samereaction in the presence of guanidine-HCI5M provokes the nitration of 8 tyrosineresidues and the total loss of somatotropicactivity of this modified hormone (Ma e tal., 1971). This inactivation may be attrib-uted to structural changes in the hormonecaused by nitration. In fact, the structureof the nitrated molecule is far less rigidthan that of the native hormone, as shownby its higher sensitivity to tryptic digestion.

Nitration of some tyrosine residueswas performed on equine and bovine GHsby Daurat-Larroque et aL (1977). The

reactivity of the different tyrosines towardstetranitromethane is variable. The mostsensitive are tyrosine residues 35 and 176for bGH and 35 for eGH, followed by tyro-sine 42 for both hormones. Residues 111 1for bGH and 111, 143 and 176 for eGHreact partially, while tyrosines 143 and160 for bGH and 28 and 160 for eGHremain unchanged. In these conditions,nitration also affects the tryptophan resi-due of both hormones. The growth activityestimated by the weight gain test repre-sents 100% for nitrated-bGH and 84% fornitrated-eGH compared to native hor-

mones, respectively. On the other hand,nitration of bGH and eGH, under denatur-ing conditions (in the presence of 8M

urea) totally inhibits the biological activityof both hormones. A change in the struc-ture of these hormones may account forthis loss of activity, but it has not been

proved. These authors concluded that

tyrosine residues 35, 42, 111 and 176 and

tryptophan 86 for bGH and eGH and tyro-sine residue 143 of eGH do not affect the

biological activity of these hormones in

non-denaturing conditions.

Acetylation of bGH by N-acetylimida-zole of the following tyrosine groups givenby order of decreasing reactivity : Tyr 42,160, 35, 111, 143 and 176, is also accom-panied by acetylation of some lysine resi-dues (Blumgrund de Satz and Santom6,1981 ). In these conditions, bGH exhibitsin vivo a small growth activity and is notable to compete with iodinated bovine

growth hormone in the binding to liver

membranes, although modified bGHretains its a-helix structure. Growth activi-

ty of bGH in vivo is maintained when ami-dination of lysine residues is performed inparallel with acetylation of tyrosine resi-dues. These results are in agreement withthose of Daurat-Larroque et al. (1977).They indicate that first tyrosine residuesare not involved in growth activity. lodina-tion of 3 tyrosine residues of bGH doesnot interfere with the structure of this hor-

mone, which exhibits a similar circulardichroism spectrum (Mattera et al., 1982)and does not modify its activity for soma-totropic sites (Mattera and Dellacha,1982). Biscoglio de Jimenez Bonino et al.(1979) have observed a 50% reduction in nvivo in the bGH growth activity after trini-trophenylation of lysines 181, 145, 70,113, 172 and 168 (residue 158 is ex-

cluded in this reaction), without change inthe a-helix structures suggesting that

lysines are implicated in the expression ofthis activity through the positive chargescarried by lysine.

Different alkylation reactions (methyla-tion and ethylation) as well as guanidina-tion and acetiminidation have been per-formed on lysine residues of 3 growth-promoting hormones such as bGH, hGHand oCS (Chene et aL, 1984; Martal e t

al., 1985). They showed that whatever thedegree and kind of chemical modification,lysine residues participate in the hor-

mone-hepatic receptor interaction. Amongthem may be mentioned the followingbasic amino acid residues (lysine or argi-nine) at positions 16, 41, 64, 70, 77 and134 as well as lysine 115, highly con-served in prolactins, and which are liableto be involved in one or several sites of

recognition of somatotropic receptors(Chene, 1987).

In conclusion, among the numerouschemical modifications performed up tonow, only those affecting lysine havebeen found to have a negative effect onthe in vivo biological activity and on thebinding capacity to the liver receptors ofGHs.

Nicoll et al. (1986) have identifiedseveral G clusters specifically associatedwith GH activity (binding affinity and invivo potency), and PG clusters connectedwith both GH and PRL activities, by ana-lysis of predicted hydropathy profilescombined with the secondary structure ofconsensus GH, hGH and pPRL (porcineprolactin). Among clusters located in the1-134 fragment molecule there are clus-ters G (G1-G2-G4-G5-G6) at the followingamino acid positions: 17―20, 33―50,70―74, 106―113 and 130―158 (Table111). Moreover, clusters PG1 (30-31),PG2 (67―69), PG3 (114-122) characte-rize common binding determinants for GHand PRL.

Among the basic amino acids studiedby Chene et al. (1984) and Martal et al.(1985) and liable to participate in the

growth hormone-receptor relationship,most are included or closely related to thebinding determinant clusters identified byNicoll et al. (1986) (Fig. 3). Consequently,arginine at positions 64 and 77 of hGHseems to be excluded from a binding siteto somatotropic receptors. Moreover, Lys

41 is not essential for hGH binding to

somatotropic receptors, since the 20Kvariant of hGH, lacking 32 to 46 aminoacids, binds both lactogenic and somato-tropic receptors (Fig. 3; Table 111). Clustersidentified by Nicoll et al. (1986) and basicamino acids selected by Martal et al.

(1985) and involved in a somatotropic-likeactivity do not correspond precisely withthe invariant sequences proposed by Niallet al. (1971), ), Nicoll et al. (1986) andKawauchi and Yasuda (1987). The ex-

pected overlap between invariant and bio-logically active areas of GH could not beclearly verified (cf. Table III).

Genomic modifications

hGH genes

Two genes coding for hGH have beenidentified (Fiddes et aL, 1979; Seeburg,1982). hGH-N gene encodes for the

sequence of the major form of hGH, the22K, as well as the subsequent forms thatare derived from it (see &dquo;Growth Hor-mone. I. Polymorphism&dquo;; Charrier and

Martal, 1988) including the 20K variantfrom an alternative splicing of premessen-ger RNA (Wallis, 1980; De Noto et al.,1981), and a new natural 17.5K variant(Lecomte et al., 1987). The 20K variantrepresents the product of N gene in whichthe second intron has been extended atthe expense of part of the exon encodingfor the 22K hGH sequence.

It has been previously demonstratedthat deletion of amino acids 32―46 led tothe almost total loss (80%) of insulin-likeactivity of the 20K as compared to the22K (Frigeri et al., 1979; Kostyo et al.,1985). The second hGH-V gene seems toencode for a polypeptide of 191 aminoacids experimentally transcribed in vitro,in which 13 amino acids differ from the22K of hGH (Pavlakis et al., 1981). It hasbeen verified that V gene encodes for pla-cental hGH (Hennen et al., 1985; Fran-kenne et al., 1987) and the &dquo;invisible&dquo; hGH

(Bistritzer et al., 1988).

Methionyl-GHA methionyl-hGH synthetized by E. colihas been purified by Olson et al. (1981 ).They demonstrated that natural hGH andthe Met-hGH derivative obtained by gen-etic engineering are similar in both weightgain and tibia tests in hypophysectomizedrats. The same growth activity has beendescribed by Kostyo et al. (1985) for the20,000-dalton Met-hGH.

A Met-bGH exhibits a growth activity of1.4 U/mg detectable by the increased

weight gain of hypopituitary-dwarf-mouseversus 1 U/mg for a natural pituitary bGH(Hart et al., 1984).

Accordingly, growth activity is not

affected when growth hormones are bio-

synthesized, i.e., obtained with or withoutsupplementary methionine at the NH2-ter-minal (Moore et al., 1988).

Conclusion

Structure-function relationships of growthhormones, especially hGH and bGH, weremainly studied with reference to growthactivities in vivo and binding to somatotro-pic receptors. The data obtained willcontribute to a better fundamental know-

ledge of biologically active sites of this GHfamily. However, some results are stillinconsistent.

Thus, although the gene reduplicationof the common ancestral peptide seemsto be highly probable between the differ-ent GH and PRL families, determination ofpeptide sequences corresponding to inter-nal homologies between these moleculesis still uncertain and variable.

Estimation of the secondary structuresof hGH, bGH, oGH and mGH, whichrequires the use of different methods, isstill restricted. Accordingly, it is not

possible to specify a relationship betweensecondary and tertiary structures and

somatotropic function. Clearly the pres-ence of two disulfide bonds is not implica-ted in the expression of its biological acti-vity. On the other hand, it has been wellestablished that the first two-thirds of thehGH molecule (peptide 1―134) displaythe specific biological activity, althoughthe full expression of the latter requiresthe presence of the last third (peptide141-191 So, the whole molecule wouldmaintain the correct structure of one ormore restricted binding sites. The frag-ment 135―145 does not contribute to the

somatotropic activity. The involvement inthe insulin-like activity of peptide 32―46,missing in the 20K hGH, is most probableas is also peptide 4―15 of hGH.

The somatomedin-like activity is attrib-uted more particularly to fragment96―124 of bGH. The same part of thehGH molecule (98―128) intervenes in

growth activity in vivo, as well as in highlyspecific binding to liver membranes. Amarked decrease in the growth activity ofhGH (1―134) and hPL (141―191)recombinant might be mediated by theexistence in hPL of histidine replaced byaspartic acid in hGH at position 153, sug-gesting that the latter residue could be

important for the growth hormone ac-

tivities of some placental lactogens suchas ruminant CS. In the same manner, hPLexhibits a methionine residue at position64 instead of Arg or Lys in every moleculeof GH.

In addition, hPL displays an asparticacid residue at position 104 versus gly-cine residue in every GH, an aspartic acidresidue at position 110 versus valine resi-due in every GH and a histidine residue at

position 112 versus an aspartic or gluta-mic acid residue in every GH studied.

The integrity of some amino acids, par-ticularly lysines, is essential for the full

expression of the biological activity. It maybe asserted that lysine or arginine at posi-tions 16, 70, 115 and 134 is involved inthe growth hormone-receptor binding.

In the future, directed mutagenesisstudies should lead to a more accuratedefinition of the structure―function rela-

tionships of growth-promoting hormones.

Acknowledgments

We are grateful to Mrs A. Bouroche for thetranslation into English of the manuscript.

References

Abdel-Meguid S.S., Shieh H.S., Smith W.W.,Dayringer H.E., Violand B.N. & Bentle L.A.

(1987) Three-dimensional structure of a geneti-cally engineered variant of porcine growth-hor-mone. Proc. NaN. Acad. Sci. USA 84, 6434-6437

Barta A., Richards R.L, Baxter J.D. & Shine J.

(1981) Primary structure and evolution of rat

growth hormone gene. Proc. Natl. Acad. Sci.USA 78, 4867-4871Becka S., Bilek J., Slaba J., Skarda J. & Miku-las 1. (1977) Some properties of the goat pla-cental lactogen. Experimentia 33, 771-772Bewley T.A., Dixon J.S. & Li C.H. (1968)Human pituitary growth hormone. XVI. Reduc-tion with dithiothreitol in the absence of urea.Biochim. Biophys. Acta, 154, 420-422Bewley T.A., Dixon J.S. & Li C.H. (1972)Sequence comparison of human pituitarygrowth hormone, human chorionic somato-

mammotropin and ovine pituitary growth andlactogenic hormone. Int. J. Peptide ProteinRes. 4, 281-287

Bewley T.A. & Li C.H. (1970) Primary structuresof human pituitary growth hormone and sheeplactogenic hormone compared. Science 168,1361-1362

Bewley T.A. & Li C.H. (1972) Molecular weightand circular dichroism studies of bovine andovine pituitary growth hormones. Biochemistry11, 927-931

Bewley T.A. & Li C.H. (1986) The conformationof monkey pituitary somatotropin. Arch. Bio-chem. Biophys. 248, 646-651

Biscoglio de Jimenez Bonino M.J., Cascone O.& Santome J.A. (1979) Trinitrophenylation ofbovine growth hormone. Int. J. Peptide ProteinRes. 14, 107-112 2

Bistritzer T., Chalew S.A., Lovchik J.C. &Kowarski A.A. (1988) Growth without growthhormone: the &dquo;invisible&dquo; GH syndrome. Lancet1, 321-323Blake J. & Li C.H. (1973) Synthesis and

growth-promoting activity of Na-acetyl-hGH(95―136). Int J. Peptide Protein Res. 5, 123-125

Blumgrund de Satz V. & Santome J.A. (1981)Acetylation of bovine growth hormone by N-acetylimidazole. Int J. Peptide Protein Res. 18,492-499

Bolander F.F. & Fellows R.E. (1976) Purificationand characterization of bovine placental lacto-gen. J. Biol. Chem. 251, 2703-2708

Burstein S., Grumbach M.M., Kaplan S.L. & LiC.H. (1978) Immunoreactivity and receptor

binding of mixed recombinants of human grow-th hormone and chorionic somatomammotro-

pin. Proc. Natl. Acad. Sci. USA 75, 5391-5394Cascone 0., Biscoglio de Jimenez Bonino M.J.& Santome J.A. (1980) Oxidation of methionineresidues in bovine growth hormone by chlora-mine T. Int J. Peptide Protein Res. 16, 299-305Chan J.S.D., Robertson H.A. & Friesen H.G.(1976) The purification and characterization ofovine placental lactogen. Endocrinology 98,65-76

Charrier J. & Martal J. (1988) Growth hormoneI. Polymorphism. Reprod. Nutr. Dev. 28, 857-886

Chen C.H. & Sonenberg M. (1977) Conforma-tion studies of biologically active fragments ofbovine growth hormone. Biochemistry 16,2110-2118 8

Chene N. (1987) Relations structure-fonctiond’hormones à activité lactogbne et de croissan-ce en vue d’applications A 1’6tude de la soma-tomammotrophine chorionique ovine (oCS).These de Doctorat d’Universite, OrsayChene N., Martal J. & De la Llosa P. (1984)Involvement of lysine residues in the binding ofovine chorionic somatomammotropin to lacto-genic and somatotropic receptors. FEBS Lett.166, 352-356Chillemi F. & Pecile A. (1971) Solid phase syn-thesis and bioassay by tibia test of monotetra-contapeptide 81―121 and of ditriacontapeptide122―153 of human growth hormone. Expe-rientia 27, 385-386Chou P.Y. & Fasman G.D. (1978) Empirical pre-dictions of protein conformation. Annu. Rev.Biochem. 47, 251-276Closset J., Smal J., Gomez F. & Hennen G.

(1983) Purification of the 22 000 and 20 000mol wt forms of human somatotropin and char-acterization of their binding to liver and mam-mary binding sites. Biochem. J. 214, 885-892Colosi P., Marr G., Lopez J., Haro H., Ogren L.& Talamantes F. (1982) Isolation, purificationand characterization of mouse placental lacto-gen. Proc. Natl. Acad. Sci. USA 79, 771-775Currie W.B., Kelly P.A., Friesen H.G. & Thorn-burn G.D. (1977) Caprine placental lactogen :levels of prolactin-like and growth hormone-likeactivities in the circulation of pregnant goatsdetermined by radioreceptor assays. J. Endo-cr. 73, 215-226

Daurat-Larroque S., Moya Portuguez M.E. &Satome J.A. (1977) Reaction of bovine and

equine growth hormones with tetranitrometh-ane. Int. J. Peptine Protein Res. 9, 119-128

Dayhoff M.O. (1976) Atlas of Protein Se-

quence and Structure. Nat. Biomed. Res.Found., Washington DC, Vol. 5 (suppl. 2), 120-121 .

De Noto F.M., Moore D.D. & Goodman M.M.

(1981) Human growth hormone DNA sequenceand mRNA structure: possible alternation spli-cing. Nucleic Acids Res. 9, 3719-3730Duckworth M.L., Kirk K.L. & Friesen H.G.

(1986) Isolation and identification of a cDNAclone of rat placental lactogen. 11. J. Biol.Chem. 261, 10871-10878

Fagin K.D., Lackey J.L., Reagan C.R. & DiGirolano M. (1980) Specific binding of growthhormone by rat adipocytes. Endocrinology 107,608-615 5

Fellows R.E., Bolander F.F., Hurley T.W. &

Handwerger S. (1975) Isolation and characteri-zation of bovine and ovine placental lactogen.In : Proc. 3rd Int Symp. Milan. Growth Hor-mone and Related Peptide (A. Pecile and E.E.Muller, eds.), Excerpta Medica Found., Amster-dam, pp. 315-326

Fiddes J.C., Seeburg P.H., De Noto F.M., Halle-well R.A., Baxter J.D. & Goodman H.M. (1979)Structure of genes for human growth hormoneand chorionic somatomammotropin. Proc.Natl. Acad. Sci. USA 76, 4294-4298

Florini J.R., Tonelli G., Breuer C.B., Coppola J.,Ringler I. & Bell P.H. (1966) Characterizationand biological effects of purified placental pro-tein (human). Endocrinology 79, 692-708

Forsyth LA. (1973) Secretion of a prolactin-likehormone by the placenta in Ruminants. In : LeCorps Jaune (R. Denamur and A. Netter, eds.),Masson, Paris, pp. 239-255Frankenne F., Rentier-Delrue F., Scippo M.L.,Martial J. & Hennen G. (1987) Expression ofthe growth hormone variant gene in human pla-centa. J. Clin. Endocrinol. Metab. 64, 635-637Friesen H. (1965) Purification of placental fac-tor with immunological and chemical similarityto human growth hormone. Endocrinology 76,369-381

Frigeri L.J., Peterson S.M. & Lewis V.J. (1979)The 20,000-dalton structural variant of humangrowth: lack of some early insulin-like effects.Biochem. Biophys. Res. Commun. 91, 778-782

Frigeri L.G., Teguh C., Ling N., Wolff G.L. &Lewis V.J. (1988) Increased sensitivity of adi-pose tissue to insulin after in vivo treatment of

yellow AvY/A obese mice with amino-terminalpeptides of human growth hormone. Endocri-nology 122, 2940-2945

Graf L., Barat E., Borvendeg J., Hermann L. &

Patthy A. (1976) Action of thrombin on ovine,bovine and human pituitary hormone. Eur. J.Biochem. 64, 333-340

Graf L. & Li C.H. (1974a) On the primary struc-ture of pituitary bovine growth hormone. Bio-chem. Biophys. Res. Commun. 56, 168-176Graf L. & Li C.H. (1974b) Isolation and proper-ties of two biologically active fragments fromlimited tryptic hydrolysis of bovine and ovinepituitary growth hormones. Biochemistry 13,5408-5415 5

Graf L., Li C.H. & Bewley T.A. (1975) Selectivereduction and alkylation of the COOH-terminaldisulfide bridge in bovine growth hormone. Int.J. Peptide Protein Res. 7, 467-473Graf L., Li C.H. & Jibson M.J. (1982) Humansomatotropin. Selection removal with trypsin ofresidues 135―145 from the hormone moleculewith no loss of biological activities. J. Biol.Chem. 257, 2365-2369

Greenspan F.S., Li C.H., Simpson M.E. &Evans H.M. (1949) Bioassay of hypophysealgrowth hormone: the tibia test. Endocrinology45, 455-463

Handwerger S., Maurer W., Barrett J., Hurley T& Fellows R.E. (1974) Evidence for homologybetween ovine and human placental lactogen.Endocr. Res. Commun. 1, 403-413 3

Hart I.C., Chadwick P.M.E., Boone TC., Lan-gley K.E., Rudman C. & Souza L.M. (1984) Acomparison of the growth promoting, lipolytic,diabetogenic and immunological properties ofpituitary and recombinant DNA-derived bovinegrowth hormone (somatotropin). Biochem. J.224, 93-100

Hennen G., Frankenne F., Closset J., GomezF., Pirens G. & El Khayat N. (1985) A humanplacental GH : increasing levels during secondhalf of pregnancy with pituitary GH suppressionas revealed by monoclonal antibody radio-

immunoassays. lnt. J. Fertil. 30 (2), 27-33Jackson L.L., Colosi P., Talamantes F. & LinzerD.I.H. (1986) Molecular cloning of mouse pla-cental lactogen CDNA. Proc. Natl. Acad. Sci.USA 83, 8496-8500

Jibson M.D. & Li C.H. (1979) Secondary struc-ture prediction of anterior pituitary hormones.Int J. Peptide Protein Res. 14, 113-122Jones N.D., De Honiesto J., Tackitt P.M. & Bec-ker G.W. (1987) Crystallization of authenticrecombinant human growth hormone.

BiolTechnology 5, 499-500

Kawauchi H. & Yasuda A. (1987) Evolutionaryaspects of growth hormones from nonmamma-lian species. J. Endocrinol. Invest. 10 (4), 4

Kelly P.A., Tsushima T., Shiu R.P.C. & FriesenH.G. (1976) Lactogenic and growth hormone-like activities in pregnancy determined by radio-receptor assays. Endocrinology 99, 765-774

Kostyo J.L., Cameron C.M., Olson K.C., JonesA.J.S. & Pai R.C. (1985) Biosynthetic 20-kilo-dalton methionyl-human growth hormone hasdiabetogenic and insulin-like activities. Proc.Natl. Acad. Sci. USA 82, 4250-4253

Kyte J. & Doolittle R.T (1982) A simple methodfor displaying the hydropathic character of aprotein. J. MoL Biol. 157, 105-132Lecomte C.M., Renard A. & Martial J.A. (1987)A new natural hGH variant -17,5 kd-producedby alternative splicing. An additional consensussequence which might play a role in branch-

point selection. Nucleic. Acids Res. 15, 6331-6348

Lee S.J. & Nathans D. (1987) Secretion of pro-liferin. Endocrinology, 120, 208-213 3Lesniak M.A., Roth J., Gorden P. & GavinJ.R.111 (1973) Human growth hormone radiore-ceptor assay using cultured human lympho-cytes. Nature (Lond.) 241, 20-22Lewis U.J., Bonewald L.F. & Lewis L.J. (1980)The 20,000-dalton variant of human growthhormone: location of the amino acid deletions.Biochem. Biophys. Res. Commun. 92, 511-516 6Lewis U.J., Dunn J.T., Bonewald L.F., SeaveyB.K. & Vanderlaan W.P. (1978) A naturallyoccurring structural variant of human growthhormone. J. Biol. Chem. 253, 2679-2687

Lewis U.J., Singh R.N.P. & Tutwiler G.F. (1981)Hyperglycemic activity of the 20,000-daltonvariant of human growth hormone. Endocr.Res. Commun. 8, 155-164Lewis U.J., Singh R.N.P., Vanderlaan N.P. &Tutwiler G.F. (1977) Enhancement of the hyper-glycemic activity of human growth hormone byenzymic modification. Endocrinology 101,1583-1603

Li C.H. (1972) Hormones of the adenohypophy-sis. Proc. Am. Phil. Soc. 116, 365-382

Li C.H. (1975) Human pituitary growth hor-mone : a biologically active hendekakaihekatonpeptide fragment corresponding to amino-acidresidues 15―125 in the hormone molecule.Proc. Natl. Acad. Sci. USA 72, 3878-3882

Li C.H. (1978) Noncovalent interaction of theNH2-terminal fragment of human somatotropin

with the COOH-terminal fragment of human

chorionomammotropin to generate growth-pro-moting activity. Proc. Natl. Acad. Sci. USA 75,1700-1702

Li C.H. & Bewley T.A. (1976) Human pituitarygrowth hormone : restoration of full biologicalactivity by noncovalent interaction of two frag-ments of the hormone. Proc. Natl. Acad. Sci.USA 73, 1476-1479Li C.H., Bewley T.A., Blake J. & Hasashida T.

(1977) Human somatotropin : restoration of fullbiological activity by non covalent interaction ofa natural and a synthetic fragment of the hor-mone. Proc. Natl. Acad. Sci. USA 74, 1016-1019 9

Li C.H., Chung D., Lahm H.W. & Stein S.

(1986) The primary structure of monkey pituit-ary growth hormone. Arch. Biochem. Biophys.245 (1), 287-291Li C.H., Dixon J.S. & Chung D. (1973) Amino-acid sequence of human chorionic somato-

mammotropin. Arch. Biochem. Biophys. 155,95-110 0

Li C.H., Dixon J.S., Gordon D. & Knorr J.

(1972) Amino acid sequence of sheep pituitarygrowth hormone. lnt. J. Peptide Protein 4, 151-153

Li C.H. & Graf L. (1974) Human pituitary growthhormone : isolation and properties of two bio-logically active fragments from plasmin digests.Proc. Natl. Acad. Sci. USA 71, 1197-1201Li C.H. & Samuelson G. (1965) Human pituitarygrowth hormone. IX. Rate of hydrolysis by tryp-sin, chymotrypsin, and pepsin; effect of trypsinon the biological activity. Mol. Pharmacol. (NY)1, 47-52Liberti J.P. (1981) Isolation and somatomedinactivity of bovine growth hormone fragment87―124. Biochim. Biophys. Acta 675, 239-247Liberti J.P. & Durham L.A. (1983) Bovine

growth hormone fragment (1―133) has in nvitro somatomedin-like activity. J. Endocrinol.96, 195-199Liberti J.P. & Miller M.S. (1978) Somatomedin-like effects of biologically active bovine growthhormone fragments. Endocrinology 103, 29-34Linzer D.LH. & Nathans D. (1984) Nucleotidesequence of a growth-related mRNA encodinga member of the prolactin-growth hormone

family. Proc. Natl. Acad. Sci. USA 81, 4255-4259

Linzer D.I.H. & Talamantes F. (1985) Nucleo-tide sequence of mouse prolactin and growthhormone mRNAs and expression of these

mRNAs during pregnancy. J. Biol. Chem. 260,9574-9579

Ma L., Brovetto-Cruz J. & Li C.H (1971) Hor-mone pituitary growth hormone. XXVII. Reac-tion with tetranitromethane. Biochem. Biophys.Acta 229, 444-450Martal J. (1978) Placental growth hormone insheep : purification, properties and variations.Ann. Biol. Anim. Biochim. Biophys. 18, 45-51Martal J. (1980) L’hormone lactogbne placen-taire ovine. Purification et propriétés physico-chimiques, immunologiques et biologiques.Etude de son r6le dans les mécanismes decontr6le de la gestation. These de Doctorat 6s-Sciences d’Etat en BiologieMartal J., Chene J. & De La Llosa P. (1985)Involvement of lysine residues in the binding ofhGH and bGH to somatotropic receptors.FEBS Lett. 180, 295-299

Martal J. & Djiane J. (1976) Mammotropic andgrowth promoting activities of a placental hor-mone in sheep. J. Steroid. Biochem. 8, 415-417

Mattera R. & Dellacha J.M. (1982) Biologicaland immunological characterization of iodina-ted bovine growth hormone. Int. J. PeptideProtein Res. 19, 181-186

Mattera R., Turyn D., Fernandez H.J. & Della-cha J.M. (1982) Structural characterization ofiodinated bovine growth hormone. lnt. J. Pep-tide Protein Res. 19, 172-180Merrifield R.B. (1969) Solid phase peptide syn-thesis. Adv. Enzymol. 32, 221-296Mills J.B., Kostyo J.L., Mosely M.H., ReaganC.R. & Wilhelmi A.E. (1978) Isolation of charac-terization of fragments of reduced and S-carba-midomethylated human growth hormone pro-duced by plasmin digestion. I. Chemistry.Endocrinology 102, 1366-1376Mills J.B., Kostyo J.L., Reagan C.R., WagnerS.A., Moseley M.H. & Wilhelmi A.E. (1980)Fragments of human growth hormone pro-duced by digestion with thrombin : chemistryand biological properties. Endocrinology 107,391-399

Mills J.B. & Wilhelmi A.E. (1965) Sulfitolysis ofbovine growth hormone. Endocrinology 76,522-525

Milman A.E. & Russell J.A. (1950) Someeffects of purified pituitary growth hormone oncarbohydrate metabolism in rat. Endocrinology47, 114-128Moore J.A., Rudman C.G., MacLachlan J.N.,Fuller G.B., Burnett B. & Frane J.W. (1988)

Equivalent potency and pharmacokinetics ofrecombinant human growth hormones with orwithout an N-terminal methionine. Endocrino-logy 122, 2920-2926Morikawa M., Green H. & Lewis U.J. (1984)Activity of human growth hormone and relatedpolypeptides on the adipose conversion of 3T3cells. Mol. Cell. Biol. 4, 228-231

Ng F.M., Blaskett E. & Larsen-Disney P. (1985)A cellular mediator for the hypoglycemic actionof a human growth hormone fragment (hGH4̂15). Biochem. Int. 11, 739-746Ng F.M., Bornstein J., Welker C., Zimmett P.Z.& Taft P. (1974) Insulin-potentiating action ofsynthetic peptide relating to the amino terminalsequence of human growth hormone. Dia-betes 23, 943-949

Ng F.M. & Harcourt J.A. (1986) Stimulation of2-deoxyglucose uptake in rat adipocytes by ahuman growth hormone fragment (hGH 4-15).Diabetologia 29, 882-887Niall H.D., Hogan M.L., Sauer R., RosenblumLY & Greenwood F.C. (1971) Sequences ofpituitary and placental lactogenic and growthhormones: evolution from a primordial peptideby gene reduplication. Proc. Natl. Acad. Sci.USA 68, 866-869Nicoll C.S., Mayer G.L. & Russell S.M. (1986)Structural features of prolactins and growthhormones that can be related to their biologicalproperties. Endocr. Rev. 7, 169-203Olson K.C., Fenno J., Lin N., Harkins R.N.,Snider C., Kohr W.H., Ross M.J., Fodge D.,Prender G. & Stebbing N. (1981) Purifiedhuman growth hormone from E. coli is biologi-cally active. Nature 293, 408-411 1Pankov Y.A., Bulatov A.A. & Osi Pova T.A.

(1982) Primary structure of seiwhale pituitarysomatotropin. Int. J. Peptide Res. 20, 396-399Pavlakis G.N., Hizuka N., Gorden P.L., See-burg P. & Hamer D.H. (1981) Expression of twohuman growth hormone genes in monkey cellsinfected by simian virus 40 recombinants.Proc. Natl. Acad. Sci. USA 78, 7398-7402Posner B.I., Kelly P.A., Shiu R.P.C. & FriesenH.G. (1974) Studies of insulin, growth hormoneand prolactin binding : tissue distribution, spe-cies variation and characterization. Endocrino-

logy 95, 521-531Reagan C.R., Kostyo J.L., Mills J.B., GennickS.E., Messian J.L., Wagner S.A. & WilhelmiA.E. (1981) Recombination of fragments ofhuman growth hormone : altered activity of therecombinant molecule. Endocrinology 109,1663-1671

Reagan C.R., Kostyo J.L., Mills J.B., Martha H.,Moseley M.H. & Wilhelmi A.E. (1978) Isolationand characterization of fragments of reducedand S-carbamidomethylated human growthhormone produced by plasmin digestion. 11.

Biological and immunological activities. Endo-crinology 102, 1377-1386Reagan C.R., Mills J.B., Kostyo J.L., WagnerS.A. & Wilhelmi A.E. (1979) Recombination offragments of growth hormone. Abstr. 61 stAnnual Meeting of the Endocrine Society, Ana-heim, CA, No. 607

Reagan C.R., Mills J.B., Kostyo J.L. & WilhelmiA.E. (1975) Isolation and biological characteri-zation of fragments of human growth hormoneproduced by digestion with plasmin. Endocri-nology 96, 625-633Retegui L.A., De Meyts P., Pena C. & MassonP.L. (1982) The same region of human growthhormone is involved in its binding to variousreceptors. Endocrinology 111, 668-676Robertson M.C. & Friesen M.G. (1975) Thepurification and characterization of rat placen-tal lactogen. Endocrinology 97, 621-629Russell J., Sherwood L.M., Kowalski K. &Schneider A.B. (1981) Recombinant hormonesfrom fragments of human growth hormone andhuman placental lactogen. J. Biol. Chem. 256,296-300

Seeburg P.H. (1982) The human growth hor-mone gene family : nucleotide sequences showrecent divergence and predict a new polypep-tide hormone. DNA 1, 239-249

Seeburg P.H., Sias S., Adelman I., De BoerH.A., Heyflick J., Jhurani P., Goeddel D.V. &

Heyneker H.L. (1983) Efficient bacterial expres-sion of bovine and procine growth hormones.DNA 2, 37-45Sherwood L.M., Handwerger S., McLaurin W.D.& Lanner M. (1971) Amino-acid sequence ofhuman placental lactogen. Nature (Lond.) 233,59-61

Shome B. & Friesen H.G. (1971) Purificationand characterization of monkey placental lacto-gen. Endocrinology 89, 631-641Sigel M.B., Thorpe N.A., Kobrin M.S., LewisV.J. & Vanderlaan W.P. (1981) Binding charac-teristics of a biologically active variant of humangrowth hormone (20K) to growth hormone andlactogen receptors. Endocrinology 108, 1600-1603

Singh R.N.P., Seavey B.K., Rice V.P., LindseyT.T. & Lewis U.J. (1974) Modified forms ofhuman growth hormone with increased biolo-gical activities. Endocrinology 94, 883-891

Singh R.N.P., Seavey B.K., Lewis L.J. & LewisU.J. (1983) Human growth hormone peptide1―43: isolation from pituitary glands. J. Pro-tein Chem. 2, 425-436

Smal J. (1986) Etude des proprietes biolo-

giques de I’hormone de croissance humaine etde son variant 20K. Those de l’Universit6

Catholique de Louvain, Faculte des SciencesAgronomiquesSmal J., Closset J., Hennen G. & De Meyts P.(1986) The receptor binding properties of the20K variant of human growth hormone explainits discrepant insulin-like and growth promotingactivities. Biochem. Biophys. Res. Commun.134, 159-165Southard J.N., Thordarson G. & Talamantes F.

(1986) Purification and partial characterizationof hamster placental lactogen. Endocrinology119, 508-514 4

Souza L.M., Boone T.C., Murdock D., LangleyK., Wypych J., Fenton D., Johnson S., Lai P.H.,Everett R., Rou-Yn H.S.U. & Bosselman R.

(1984) Application of recombinant DNA techno-logies to studies on chicken growth hormone.J. Exp. Zool. 232, 465-473

Spencer E.M., Lewis L.J. & Lewis V.J. (1981)Somatomedin generating activity of the 20,000-dalton variant of human growth hormone.

Endocrinology 109, 1301-1302Stevenson R.W., Stebbing N., Jones T., CarrK., Jones P.M., Hii C. & Cherrington A.D.

(1988) The synthetic human growth hormonefragment (32―38) increases glucose uptake inthe conscious dog. Acta Endocrinol. 117, 457-462

Stevenson R.W., Stebbing N., Rudman C.G.,Williams P.E. & Charrington A.D. (1987) Thesynthetic 32―46 fragment of human growthhormone increases insulin and glucogen levelsin the conscious dog. Metabolism 36, 400-404Swenne I., Hill D.J., Strain A.J. & Milner R.D.G.

(1987) Effects of human placental lactogen andgrowth hormone on the production of insulin

and somatomedin C/insulin-like growth factor Iby human fetal pancreas in tissue culture. J .Endocrinol. 113, 297-303Talamantes F. (1975) In vitro demonstration oflactogenic activity in the mammalian placenta.Am. Zool. 15, 279-284Tsubokawa M. & Kawauchi H. (1985) Completeamino acid sequence of fin whale growth hor-mone. Int. J. Peptide Res. 25, 297-304Tsushima T. & Friesen H.G. (1973) Radiorecep-tor assay for growth hormone. J. Clin. Endocri-nol. Metab. 37, 334-336

Wallis (M. (1973) The primary structure of bo-vine growth hormone. FE8S Lett 35, 11-14 4Wallis M. (1975) The molecular evolution of

pituitary hormones. Biol. Rev. 50, 35-98Wallis M. (1980) Growth hormone : deletions inthe protein and introns in the gene. Nature 284,512 2

Wallis M. & Davis R.V. (1976) Studies on thechemistry of bovine and rat growth hormones.In : Growth Hormone and Related Peptides.(Pecile A. and Muller E.E., eds.), ExcerptaMedica, Amsterdam, pp. 4-13 3

Wilhelmi A.E. (1973) Growth hormone. 3. Mea-surement. 3.1. Bioassay. In : Peptides Hor-mones (Berson S.A. and Yallow R.S., eds.),North-Holland Publishing Co., Amsterdam, pp.296-302

Wohnlich L. & Moore W.V. (1982) Binding ofvariant of human growth hormone to liver plas-ma membrane. Horm. Metab. Res. 14, 138-141

Yamasaki N. & Shimanaka J. (1975) Studies onthe common active site of growth hormone.Revision of the amino acid sequence of an ac-tive fragment of bovine growth hormone. J .Biol. Chem. 250, 2510-2514 4

Zakin M.M., Poskus E., Dellacha J.M., PaladiniA.C. & Santome J.A. (1973) The amino acidsequence of equine growth hormone. FEBSLett. 34, 353-355