UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A structure-function analysis of bovine prolactin Huyer, Marianne 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1995-059812.pdf [ 5.37MB ]
JSON: 831-1.0088174.json
JSON-LD: 831-1.0088174-ld.json
RDF/XML (Pretty): 831-1.0088174-rdf.xml
RDF/JSON: 831-1.0088174-rdf.json
Turtle: 831-1.0088174-turtle.txt
N-Triples: 831-1.0088174-rdf-ntriples.txt
Original Record: 831-1.0088174-source.json
Full Text

Full Text

A STRUCTURE - FUNCTION ANALYSIS OF BOVINE PROLACTINbyMarianne HuyerB.Sc. The University of Alberta, 1985M.Sc. The University of Alberta, 1989A THESIS SUBMITFED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinThe Faculty of Graduate StudiesDepartment of Biochemistry and Molecular BiologyWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAMay 1995© Marianne Huyer, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)______________________________Department of_______________The University of British ColumbiaVancouver, CanadaDate b.20. Sep1. 5DE-6 (2188)ABSTRACTThe 23 kDa pituitary protein hormone bovine prolactin (bPRL) belongs to a family ofhormones including growth hormones, prolactins, and placental lactogens, and is involved inregulating a variety of physiological processes including lactation and immune response. Thebiological actions of bPRL are mediated by cell surface receptors and it is anticipated, asreported for human growth hormone (hGH) and based on sequence homology among membersof this hormone family, that the hormone-receptor interaction involves receptor dimerization by asingle hormone molecule. The structure of bPRL is not known, but is believed to resemble thatof hGH in consisting of four a-helices arranged in a left twisted helical bundle. Variants ofrecombinant methionyl bovine prolactin (met-bPRL) containing single amino acid changes weregenerated by site-directed mutagenesis in order to carry out a structure-function analysis of theinteraction of bPRL with the Nb2 PRL receptor.The changes involved the replacement, in most cases with alanine, of residues in the loopregion joining putative helices 1 and 2, in putative helix 3, and in putative helix 4; regionshomologous with functionally important portions of closely related proteins including humanprolactin (hPRL) and hGH. The variant proteins were produced as inclusion bodies in E. coli,extracted with N-lauryl sarcosine (sarcosyl), and renatured by air oxidation at pH 10. Thecontribution of the residues to the biological activity of bPRL was assessed using the Nb2lymphoma cell bioassay. In this assay, in which met-bPRL was as active as pituitary-derivedbPRL, growth factor activity is measured, and the rate of cell proliferation is proportional to theconcentration of lactogen present. The findings are discussed in the light of a putative three-dimensional structure of bPRL, modelled using the structural coordinates of hGH.None of the exchanges of residues of the loop region resulted in drastic reductions in themitogenic activity. The variants H59A, L63A, Q7 lv, Q73V, and Q74V exhibited anapproximate two- to three-fold reduction in bioactivity compared to unmodified met-bPRL,suggesting that these residues may be involved in the mitogenic activity of bPRL. Of the 18putative helix 4 residues examined only substitutions of R177 and K181 led to marked decreases11in mitogenic activity, indicating that these residues are very important to the bioactivity of bPRLReplacements of L171, R176, and D183 resulted in variants with bioactivities 2.5- to 3-fold lessthan that of unmodified met-bPRL and only 9% activity was exhibited by the mutant D178E.However, according to the putative model of bPRL the side chains of 1159, L63, L171, and D178point into the molecule; thus, these residues may play a structural role in maintaining the shapeof the binding site rather than directly contacting the receptor. The variant G129R had nomitogenic activity, and E128A had reduced bioactivity, suggesting that bPRL contains a secondreceptor binding site analogous to that described for hGH.Comparison of these results with those reported for related hormones has confirmed thatthe bPRL residues involved in mediating the mitogenic activity of the hormone, and presumablyalso in interacting with the Nb2 PRL receptor, are located in regions similar to those identifiedfor other hormones. However, many differences exist in that the specific residues involved ineliciting the biological actions are not at equivalent positions; these differences likely contributeto the specificity of hormone-receptor interactions. The results of this study thus not onlyprovide insight into the interaction between bPRL and the Nb2 PRL receptor, but also, whenconsidered in light of the putative three-dimensional structure of bPRL and with respect torelated hormones, help to give a greater understanding of the general mechanism of receptorbinding by members of the growth hormone/prolactin/placental lactogen family of hormones.inTABLE OF CONTENTSAbstract iiTai,Ie of (on4gn+c wList of Tables viiList of Figures viiiAcknowledgements xList of Abbreviations xiDedication xiiINTRODUCTION 11.1 Biological Actions of Prolactin 11.2 Hormone Structure 41.3 Receptors 71.3.1 Extracellular Domains 91.3.2 Intracellular Domains 101.4 Signal Transduction 111.5 Hormone-Receptor Interactions 131.6 Mutational Analysis of Hormone Binding to Receptors 151.6.1 Human Growth Hormone hGH Fragments, Variants, and Chimeric Proteins Site-Specific Mutants of hGH hGH Interactions with PRL Receptors 261.6.2 Placental Lactogen 291.6.3 Prolactin 321.6.3.1 Prolactin Variants 321.6.3.2 Site-Specific Mutants of PRL 331. Site-Specific Mutants of bPRL 361.7 Objective of Thesis Project 39ivMATERIALS AND METHODS 412.1 Materials 412.2 Strains and Growth Conditions 412.2.1 Bacterial Strains 412.2.2 Yeast Strains 412.2.3 Media for Growth of Bacterial and Yeast Strains 412.3 Plasmids 432.4 Transformation of Cells 442.5 Isolation of Plasmid DNA 452.6 Isolation of Single-Stranded DNA 462.7 Site-Directed Mutagenesis 472.8 DNA Sequencing 472.9 Preparation and Analysis of RNA 492.9.1 Isolation of Yeast RNA 492.9.2 Isolation of E. coli RNA 492.9.3 Primer Extension Analysis of Yeast RNA 502.9.4 Competitive PCR Analysis of E. coli RNA 512.10 Isolation of Met-bPRL 522.10.1 Extraction of Met-bPRL and Met-bPRL Variants 522.10.2 Renaturation of Extracted Met-bPRL and Met-bPRL Variants 532.10.3 Quantification of Met-bPRL and Met-bPRL Variants 532.10.4 In Vitro Bioassay of Met-bPRL and Met-bPRL Variants 552.11 Analysis of Production of Met-bPRL at Different Temperatures 552.12 E. coli Secretion System 552.12.1 Plasmids 552.12.2 Secretion System 562.12.3 Fractionation Procedures for Periplasmic Proteins 57V2.13 Yeast Secretion System 582.14 Molecular Modelling of bPRL 59RESULTS 613.1 Production of Recombinant bPRL 613.1.1 Yeast Expression System 623.1.2 E. coli Secretion System 633.1.3 Intracellular E. coil System for Production of Met-bPRL andMet-bPRL Variants 663.1.3.1 E. coil Strains for Production of Met-bPRL 713.2 Extraction of Met-bPRL and Met-bPRL Variants from Inclusion Bodies 763.3 Analysis of Met-bPRL mRNA Levels by Competitive PCR 793.4 Renaturation of Met-bPRL and Met-bPRL Variants 813.5 Structure-Function Analysis ofMet-bPRL 833.5.1 Putative Lactogen-Specific Residues Identified Via ComparativeStudies 843.5.2 Substitutions of Residues in the Loop Region Between PutativeHelices 1 and2 863.5.3 Substitution of Met-bPRL Residues in Putative Helix 4 893.5.4 Substitution of Met-bPRL Residues in the Putative Site 2 BindingSite 92DISCUSSION 944.1 Production of Bovine Prolactin 944.2 Structure-Function Analysis of Met-bPRL 994.3 Residues Responsible for the Bioactivity of Met-bPRL in the Nb2 Assay 1094.4 Receptor Dimerization 1144.5 Species Specificity of the Hormone Binding Site 1184.6 Conclusions and Future Directions 119REFERENCES 124viLIST OF TABLESTable 1 Binding of Double Mutants of hGH to the hPRL and hGH BindingProteins (hPRLbp and hGHbp) 27Table 2 List of E. coli Strains 42Table 3 Oligonucleotides Used for Site-Directed Mutagenesis 48Table 4 Mitogenic Activities of Met-bPRL Variants Produced by Amino AcidSubstitutions at Positions R21, R177, and K187 86Table 5 Comparison of the Effect of Single Amino Acid Substitutions in theLoop Region Joining Putative Helices 1 and 2 on the Bioactivity ofbPRL, on the Bioactivity and Binding of hPRL to the Nb2 PRLReceptor, and on Binding of hGH to the hPRL and hGH Receptors 102Table 6 Comparison of the Effect of Single Amino Acid Substitutions inPutative Helix 4 on the Bioactivity of bPRL and on the Binding ofhGH to the hPRL and hGH Receptors 106viiLIST OF FIGURESFigure 1 Representation of the structure of porcine growth hormone (pGH). 5Figure 2 Schematic representation of members of the cytokine/hematopoietinreceptor superfamily. 8Figure 3 Diagram of hGH showing its two disulfide bridges and the locationsof the helical regions 16Figure 4 Structural model of hGH indicating the location of sites 1 and 2for binding to the hGHbp. 23Figure 5 Structural model of hGH indicating the binding site for the hPRLbp. 29Figure 6 Nucleotide sequence of the mRNA coding for bPRL and its aminoacid sequence. 37Figure 7 Map of pESP4. 43Figure 8 Primer extension analysis of bPRL mRNA in the S. cerevisiaeexpression system. 63Figure 9 Coomassie-blue stained denaturing polyacrylamide gels of bPRL invarious fractions from the E. coli secretion system. 64Figure 10 Western blot analysis of bPRL in fractions from the E. coli secretionsystem. 65Figure 11 Western blot analysis of bPRL in fractions from the E. coli secretionsystem. 68Figure 12 Coomassie blue-stained denaturing polyacrylamide gels of met-bPRLand several met-bPRL variants. 69Figure 13 Western blot analysis of met-bPRL and several met-bPRL variants. 70Figure 14 Coomassie blue-stained denaturing polyacrylamide gels of pelletedmaterial folowing lysis of various E. coli strains transformed withpESP4. 73Figure 15 Western blot analysis of met-bPRL production in E. coli TOPPstrains 1,2, and 3. 74Figure 16 Western blot analysis of renatured met-bPRL extracts from variousE. coli strains transformed with pESP4. 75Figure 17 Western blot analysis of met-bPRL and various met-bPRL mutants. 77Figure 18 Coomassie blue-stained denaturing polyacrylaniide gels of met-bPRLand various met-bPRL mutants. 78Figure 19 Quantitation of RNA by competitive PCR. 80viuFigure 20 Western blot analysis of non-renatured met-bPRL incubated in Nb2cell growth medium. 82Figure 21 Diagram of bPRL showing its three disulfide bridges and the locationsof the putative helical regions. 83Figure 22 The complete amino acid sequences of the lactogenic hormones bPRL,hPRL, mPL-ll, and hGH and of the related but non-lactogenic bGH. 85Figure 23 Bioactivities of met-bPRL variants obtained by single alanine substitutions of met-bPRL residues within the loop region joining putativehelices 1 and 2. 88Figure 24 Helical wheel diagram for putative helix 4 of bPRL. 90Figure 25 Western blot analysis of sarcosyl extracts of met-bPRL and variousmet-bPRL mutants. 91Figure 26 Western blot analysis of sarcosyl extracts of various met-bPRLmutants. 91Figure 27 Bioactivities of met-bPRL variants with single amino acid substitutions in putative helix 4 residues. 93Figure 28 Ribbon representation of the putative structure of bPRL indicatingthe binding site for activation of the Nb2 PRLreceptor. 108Figure 29 Ribbon representation of the structure of hGH indicating residueswithin the site 1 and site 2 binding regions for the hPRLbp. 111Figure 30 Ribbon representation of the putative structure of bPRL indicatingresidues within the binding site for the Nb2 PRL receptor. 112Figure 31 Backbone structure of the hGH-(hGHbp)2complex. 115Figure 32 Ribbon representation of the putative structure of bPRL indicatingresidues within the putative site 1 and site 2 binding sites. 117ixACKNOWLEDGEMENTSOf the many people who provided me with support and encouragement during the courseof my Ph.D. studies I especially want to thank my friends and colleagues in the Smith lab:Jeanette Beatty, for our very many conversations on all sorts of topics and for the wonderfulfriendship we have developed; Yip Ho, for his support and sympathy both when thoseexperiments just didn’t work out and also when they did; Heather Merilees, for her technicalassistance, friendship, and for occasionally indulging us all with Blue Chip Cookies; and GuyGuillemette, Chris Overall, and Lindsay Eltis for their constant encouragement and unflaggingfriendships. The route through my Ph.D. studies would have been much more difficult withoutthem all.Thanks are also owed to my supervisor, Michael Smith, for providing me with a supportivelab environment for my studies, for guidance, and for critically reading this thesis. A thank youalso to the rest of my supervisory committee: Roger Brownsey, Peter Gout, and Michel Roberge.Peter, by doing all the bioassays on all my mutants and by constantly giving me goodsuggestions, ideas, and references helped me a great deal, and I wish to thank him for that.Thanks also to Dennis Luck, for his valuable contributions to all aspects of this project. Inaddition, an enormous thank you is owed to Lawrence Mcintosh and Logan Donaldson for theirhelp in preparing a molecular model of bPRL. Without Logan’s many hours on the computer andhis perseverance in the face of all the problems we ran into, my thesis would have been muchless complete.There are also many people outside of UBC whom I wish to acknowledge for their helpand support. A special thank you goes to my brother, Greg, who painstakingly proof-read mythesis and learnt, in the process, more about prolactin than he ever wanted to know. My father,Rinus Huyer, and grandfather, Isaac Braunstein, have both provided me with an enormousamount of support and encouragement for which I want to thank them very much. Their beliefand pride in me is very gratifying. And especially, a big, big, big thank you to Mark Ring, foreverything!xLIST OF ABBREVIATIONSA260 absorbance at 260 nmamp ampicilimbGH bovine growth hormonebp base pairbPRL bovine prolactindH2O distilled waterDli’ dithiothreitolEDTA ethylenediaminetetraacetic acidhGH human growth hormonehGHbp human growth hormone binding proteinhPL human placental lactogenhPRL human prolactinhPRLbp human prolactin binding proteinIPTG isopropyl-13-D-thiogalactopyranosidekDa kilodaltonmPL-II mouse placental lactogen II0D600 optical density at 600 nmoPL ovine placental lactogenoPRL ovine prolactinpGH porcine growth hormonePCR polymerase chain reactionPRL-R prolactin receptorsarcosyl N-lauryl sarcosineSDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresisU unitvol volumexiDEDICATIONTo my mother, Sandra Anne Huyer (1935 - 1992)With all my love.xl’INTRODUCTIONMolecular recognition is a key component of many biochemical processes, includingDNA-protein interactions, ligand-receptor binding, and enzyme-substrate recognition. All ofthese systems require very selective recognition mechanisms to maintain the proper interactions.This selectivity is often exquisite, permitting the discrimination of molecules or targets that arestructurally very similar. The bases of this specificity are often poorly understood, and a greatdeal of scientific effort has been and continues to be aimed towards understanding the problem.A classic example of molecular recognition is ligand-receptor binding, such as peptidehormones binding to their receptors. With the relative ease of molecular biology techniques, ithas been possible to carry out structural and functional studies to delineate the contact surfacesbetween the proteins, to reveal the individual amino acids involved in the interactions, and todetermine the relative importance of each interaction. This type of approach has been appliedhere to carry out a structure-function analysis of bovine prolactin as it interacts with the prolactinreceptors on Nb2 lymphoma cells which depend on prolactin for growth.1.1 Biological Actions of ProlactinThe pituitary gland is the source of a number of important hormones including growthhormone and prolactin. Prolactin, which is now known to exist in all classes of vertebrates, wasfirst identified as a pituitary factor with the capability to induce milk secretion in rabbits (S trickerand Grueter, 1928). Soon afterwards it was discovered that proliferation of the mammary glandand lactation could be induced in spayed virgin rabbits by injecting the rabbits with extracts ofsheep pituitaries (Corner, 1930). In addition, extracts of bovine pituitaries were able to inducethe growth of pigeon crop sacs (Riddle and Braucher, 1931). A search for the pituitary factorcausing these effects led to the identification of a previously unknown hormone produced by theanterior pituitary; this hormone was named prolactin (Riddle et al., 1932).Prolactin, like growth hormone, is synthesized in acidophilic cells of the pituitary(prolactin in lactotroph cells, growth hormone in somatotroph cells) as a precursor, pre-prolactin,1with a signal peptide of approximately 25 amino acids attached to the N-terminus. As found forpre-growth hormones, the signal peptide directs pre-prolactin to the endoplasmic reticulum of theacidophilic cells where rapid processing to the mature hormone occurs (Wallis et at., 1985).Secretion of the anterior pituitary hormones is controlled by factors from the hypothalamus, suchas dopaniine; however, in contrast to growth hormone, control of prolactin secretion in mammalsis largely inhibitory. That is, removal of the hypothalamus results in marked increases inprolactin secretion whereas the release of other pituitary hormones, including growth hormone,decreases. In birds and reptiles, however, prolactin is under positive regulatory control of thehypothalamus. Many different factors including other hormones and neural signals are involvedin the physiological control of the secretion of both growth hormone and prolactin. Thus,regulation of secretion of these hormones is very complex and, indeed, is not yet fullyunderstood (Wallis et a!., 1985).Prolactin is a member of a large family of hormones that also includes growth hormone andplacental lactogen. Sequence analysis has confirmed that these hormones are evolutionarilyrelated, and it is thought that a single ancestral gene diverged to give rise to the separate prolactinand growth hormone lineages while genes for the placental lactogens may have arisen from boththese lineages (Miller and Eberhardt, 1983; Nicoll et at., 1986). These globular polypeptidehormones are found in species ranging from teleost fish to primates and as a group regulate alarge number of biological processes. Growth hormones have been shown to have somatogenic,anabolic, lipolytic, insulin-like, and diabetogenic effects (Chawla et a!., 1983), while placentallactogens, although their role has been less well defmed, seem to be involved in regulatinglactation, fetal growth, and fetal metabolism (Brinck-Johnsen and Benirschke, 1982; Ogren andTalamantes, 1988). More than 85 biological functions have been described for prolactin (Nicolland Bern, 1972) including such diverse processes as osmoregulation, amphibian metamorphosis,development of the incubation patch of birds and seahorses, and hyperplasia of the crop sacmucosa in pigeons.In mammals, some of the most important physiological effects of prolactin relate tolactation, to growth and development of the mammary gland, and to immune stimulatory2activities (Clarke and Bern, 1980). Prolactin acts in concert with insulin and glucocorticoids tostimulate milk protein gene expression and has been shown to influence the rate of genetranscription and the stability of the milk protein mRNAs (Guyette et a!., 1979). Growthhormones from primates also have lactogenic activity (Li, 1973), indicating some overlapbetween the functions of growth hormone and prolactin in these species. In fact, the lactogenicactivity of growth hormone from humans is such that, until human prolactin was purified in 1972(Hwang et al., 1972), many researchers doubted that humans produced both of these hormones.Prolactin and growth hormone are also involved in the regulation of the immune system (Gala,1991) and cells involved with immunity, such as lymphocytes, express the genes for thesehormones and for their receptors. Other evidence suggests that growth hormone and prolactinhave roles as hematopoietic growth and differentiation factors and thus they can be considered ashaving autocrine or paracrine roles in addition to their well-established endocrine actions (Gala,1991; Hooghe et at., 1993).As described above, prolactin exerts a strong effect on a number of physiological processesand has the ability to promote cell growth and/or differentiation in various tissues. Themitogenic effects of prolactin have been utilized in the development of a sensitive bioassay forprolactin and other lactogenic hormones. Proliferation of cultured cells derived from atransplantable lymphoma, designated Nb2 lymphoma, that arose in the lymph nodes of anestrogenized male rat (Nb strain) requires the presence of lactogens, and the rate of cellproliferation is proportional to the concentration of lactogen in the range of 5 - 500 pg/mL (Goutet al., 1980). Thus, the Nb2 cell line provides a sensitive and highly specific in vitro bioassay forlactogenic hormones, including prolactin (Tanaka et a!., 1980). Prior to the development of theNb2 cell assay, bioassays commonly used to determine levels of prolactin included the pigeoncrop sac assay (Nicoll, 1967) and in vitro assays employing mammary gland cultures (Forsythand Myres, 1971; Kleinberg and Frantz, 1971; Loewenstein eta!., 1971; Turkington, 1971).Since these assays are more time-consuming to perform and do not exhibit the same sensitivity tolevels of prolactin as the Nb2 cell assay, the latter assay has become the method of choice fordetermining concentrations and specific activity of biologically active lactogens.31.2 Hormone StructureIn addition to the many biological features which prolactin, growth hormone, and placentallactogen have in common, there are many shared sequences at the amino acid level (Nicoll et at.,1986; Wallis, 1978). These similarities have led to the grouping of these proteins into a singlehormone family. In turn, this family of hormones and the receptors with which they interact arepart of a larger cytokine-hormone superfamily, based on homology of the receptors (Bazan,1989, 1990a,b; Kelly et at., 1991). This family includes the interleukins, granulocytemacrophage colony-stimulating factors (GM-CSF and G-CSF) and erythropoietin (EPO). Inspite of the fact that the degree of sequence similarity between the proteins in the largersuperfamily is small, the sequence features which are shared suggests that their structures have arelated three-dimensional fold of four-helix bundles.Although there is a great deal of interest in determining the structural features of theprolactins, growth hormones, and placental lactogens which give rise to their various biologicaleffects, to date only the three-dimensional structures of porcine growth hormone (pGH) (AbdelMeguid et a!., 1987) and human growth hormone (hGH) (de Vos et a!., 1992; Ultsch et al., 1994)have been solved by X-ray crystallography. Both proteins are 191 amino acids in length andhave molecular weights of 22 kDa. The structure of pGH was solved to 2.8 A resolution andrevealed a molecule containing four a-helices arranged in a left twisted helical bundle. Althoughthe tightly packed helix bundle has a similar structure to that of other proteins such ascytochrome b-562 (Lederer eta!., 1981), cytochrome c’ (Weber et a!., 1980), andmyohemerytluin (Hendrickson and Ward, 1977), the connectivity of the helices is unusual.Helices 1 and 2 of pGH are parallel to each other and antiparallel to helices 3 and 4. In order toachieve this arrangement, long loop regions link the two sets of parallel helices while a shortsegment connects helix 2 to helix 3 (Figure 1). Porcine growth hormone was the first moleculefor which this type of connectivity was demonstrated, and because of the sequence homologybetween the growth hormones, placental lactogens, and prolactins, it is believed that this is thegeneral three-dimensional structure for this family of hormones. Recently, the elucidation ofstructures for other members of the larger cytokine superfamily including IL-2, hGH, IL-4, and4GM-CSF, has confirmed the generality of this type of fold (Bazan and McKay, 1992; de Vos eta!., 1992; Diederichs et at., 1991; Powers et at., 1992; Smith et at., 1992; Walter et al., 1992).Other structural features of pGH include a bend in helix 2 at P89, and disulfide loops whichconnect C53 in the first crossover connection to C164 in helix 4 and C181 in helix 4 to C189.Figure 1. Representation of the structure of porcine growth hormone (pGH). The four ahelices are represented by cylindrical rods and non-helical regions are shown as thin tubes. Helix1 is blue, helix 2 is pink, helix 3 is green, and helix 4 is orange. The N-terminus is located in theupper left-hand corner and the C-terminus in the lower left-hand corner. The disulfide shownlinks the centre of the loop joining helices 1 and 2 with the centre of helix 4. A hidden disulfidejoins two cysteine residues in the region between helix 4 and the C-terminal. The arrows pointtoward the position where introns have been found within the coding sequence. The figure istaken from Abdel-Meguid et a!., 1987.5The structure of hGH bound to the extracellular domain of its receptor (de Vos et at., 1992)is almost identical to that of free pGH. Differences exist in the structures of the connectingloops: hGH has two small helical segments (K38 - N47 and R64 - K70) in the ioop joininghelices 1 and 2 and a small helix from R94 - S100 in the ioop between helices 2 and 3. The firsttwo helical segments were not observed for pGH. These two minihelices of hGH are involved incontacts with the receptor and it was originally speculated that they may representconformational changes in the hormone due to receptor binding (de Vos et at., 1992). However,crystal structures of two hGH variants not complexed to receptors also reveal the existence ofthese minihelices and suggest that these are characteristic features of hGH (Ultsch et at., 1994).In contrast to the small helical section in the loop joining helix 2 to helix 3 in hGH, an omega-loop conformation was described for this section of pGH (Abdel-Meguid et at., 1987). Thereason for this apparent structural difference is not clear, since the sequences of pGH and hGHare almost identical in this region and this segment is apparently not involved in receptorbinding. However, since the structure of hGH was refined to an R factor of 0.204 while the Rfactor for the pGH structure was 0.33, the difference in the conformation of the loop joininghelices 2 and 3 may simply reflect the more preliminary state of refinement of the pGH molecule(Wells and de Vos, 1993) (R factor is the residual disagreement between the experimentallyobserved diffraction amplitudes and those calculated for a model crystal: values range from 0.0for exact agreement to around 0.59 for total disagreement).Refined structures of other members of the cytokine superfamily have recently beenreported and these show significant resemblances to each other (Wiodawer et at., 1993).Superimpositions of the available structures indicate that the greatest degree of structuralhomology is in the four-helix bundle, with helix 4 being the most highly conserved structuralelement. Each cytokine has a hydrophobic core within the interior of the four-helix bundle andalthough there is very little sequence conservation among the residues forming the core, there is agreat deal of structural similarity. In addition, there is significant sequence and structuralsimilarity between the extracellular domains of cytokine receptors which are involved in bindinghematopoietins (Bazan, 1990b). Based on the similarities between the receptors as well as6between the ligands, it has been suggested that the overall mode of binding between the ligandsand the receptors might be very similar within the cytokine superfamily (Wiodawer et a!., 1993).1.3 ReceptorsThe initial step in the action of growth hormones, prolactins, and placental lactogens is thebinding of the hormone to cell surface receptors. Once bound, intracellular signals are generatedwhich initiate the cellular response. As noted above, these proteins are members of a singlehormone family which is itself part of a large cytokine hormone family. The superfamily wasrecognized on the basis of receptor homologies, not on the basis of hormone and cytokinesequences; therefore, it is not surprising that the receptors for this group of related proteins forma family. In addition to the receptors for growth hormone and prolactin, this family includes thereceptors for granulocyte colony-stimulating factor, erythropoietin, granulocyte-macrophagecolony-stimulating factor, the interleukins IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, the 13- and ‘y-chainsof the IL-2 receptor, as well as a number of other proteins (Bazan, 1989, l990b; D’Andrea eta!.,1989; de Vos et a!., 1992; Fukunaga et a!., 1990; Gearing et a!., 1989; Goodwin eta!., 1990;Idzerda et a!., 1990; Itoh eta!., 1990; Kelly eta!., 1991; Patthy, 1990; Savino eta!., 1993). Thereceptors in this family are single subunit proteins composed of an extracellular domain whichinteracts with the activating ligand, a short transmembrane section, and a cytoplasmic domain(Figure 2). Most of the diversity between these related receptors is seen in the cytoplasmicportion of the proteins. Since all of the members of this receptor family are found on, andmediate effects on, cells of the hematopoietic lineage, it has been suggested that this group ofproteins be referred to as the hematopoietin receptor superfamily (Idzerda et a!., 1990).7IL-6RG-CSFR00Figure2.Schematicrepresentationofmembersofthecytokine/hematopoietmreceptorsuperfamily.Abbreviations:GHbp, growthhomionebindingprotein;GH-R, growthhormonereceptor;PRL-R, prolactinreceptor;EPOR, erythropoietinreceptor; JL-2R, interleukin-2receptor;IL-3R,interleukin-3receptor;IL-4R, interleukin-4receptor;IL-5R, interleukin-5receptor;GM-CSFR, granulocyte-macrophagecolony-stimulatingfactorreceptor;IL-6R, interleukin-6receptor;IL-7R, interleukin-7receptor;IL-9R, interleukin-9receptor;G-CSFR, granulocytecolony-stimulatingfactorreceptor.TheNb2formofthePRLreceptorisduetoamutationinthePRLreceptorgenewhichresultsinthelossof594bpfromthecytoplasmicdomainofthelongformofthePRL-R.Thethinblacklinesindicatetheconservedcysteines;thethickblacklinesindicatetheWSXWSmotif;thetransmembraneregionisstippled;andtheproline-richmotif(PRM)isindicatedbyahatchedbox.FigureisadaptedfromKellyetal.,1994.IL-3RPRL-RGFftGH-REPORIL2RashortNb2longIL.-5RGM-CSFR4.—Conserved4—•cysteines4—wsxwsmotifExtracellulardomainPRMIntracellulardomain1.3.1 Extracellular DomainsAlthough the sequence homology between the extracellular, hormone/growth factorbinding region of these receptors is low (14 - 25%), certain characteristics are shared and theoverall structures are likely similar. The shared features include four cysteine residues near theN-terminus, a tryptophan-serine-X-tryptophan-serine sequence (the WSXWS box, where X isany amino acid) near the C-terminus of the extracellular domain, and common patterns ofhydrophobic and hydrophiic residues (Figure 2). Based on predictive structural analysisbetween the binding domains of the cytokine receptors and interferon receptors, the overallstructure of the extracellular domain of this family of receptors was predicted to consist of twodistinct immunoglobulin-lilce domains each containing seven B-strands (Bazan, l990b).Currently, the only three-dimensional structures available for members of this receptorfamily are that of the extracellular domain of the hGH receptor (hGHbp) complexed to hGH (deVos et aL, 1992) and the extracellular domain of the human prolactin receptor (hPRLbp)complexed to hGH (Somers et at., 1994). Although the sequence identity between the hGHbpand the hPRLbp is only 28%, the structures are very similar (Somers et at., 1994). Theextracellular domain of each binding protein does, as predicted, contain two distinctimmunoglobulin-like domains. The four cysteines which are conserved in the superfamily areburied in the core of the N-terminal B-sandwich of the hGHbp and the hPRLbp. The WSXWSbox present near the C-terminus of the extracellular domain of all other members of thesuperfamily is not present in the hGHbp and is instead represented by the sequence tyrosineglycine-glutamate-phenylalanine-serine (YGEFS) beginning at residue 222. In spite of the factthat the WSXWS box is highly conserved among these related receptors, the hGHbp structureindicates that this portion of the receptor is located away from the binding interfaces with hGHand does not appear to be involved in ligand binding or in the interactions between receptormolecules which are required for signal transduction. The WSXWS box in the hPRLbp,represented by the sequence WSAWS, is also quite removed from the ligand binding interface(Somers et a!., 1994).9Although structures for other receptors in the hematopoietic superfamily are not yetavailable, the presence of conserved motifs within all members of the receptor family as well asthe many structural similarities between the two reported extracellular domains suggests that thestructure of the hGHbp can be used as a model for studying these related proteins as they interactwith their ligands. Sequence alignments of receptors in the superfamily show similar patterns ofhydrophobic and hydrogen-bonding residues, suggesting strong similarities in the folding andarrangement of the B-strands in the C-terminal region of the extracellular domain of thereceptors. Thus, residues which are conserved may be anticipated to occupy equivalent positionsin the different proteins, and the ligand-binding determinants of other receptors may correspondto those of the hGHbp. If this is indeed the case, then the residues of the WSXWS box, whichhad been predicted to form the “floor” of the binding crevice (Bazan, 1990) but which in thehGHbp and the hPRLbp are on a surface located away from the binding interfaces (de Vos et at.,1992; Somers eta!., 1994) therefore may not be involved in ligand binding by any of thereceptors. However, mutations in this region have been shown to disrupt ligand binding for theIL-2 (Miyazaki et al., 1991), erythropoietin (Chiba eta!., 1992), and prolactin (Rozakis-Adcockand Kelly, 1992) receptors, suggesting that, at least for these proteins, this region plays animportant role in the formation of the ligand binding site. It is also possible that this region maybe required to bind a putative accessory protein (Wells and de Vos, 1993). Further study isneeded to determine the role(s) of the WSXWS box in this receptor superfamily.1.3.2 Intracellular DomainsAlthough the receptors in the hematopoietic superfamily exhibit homology in theirextracellular ligand binding domains, the intracellular, cytoplasmic domains show considerablevariation (Figure 2). These intracellular portions of the receptors vary greatly in length andsequence and the only region of homology, located near the membrane-spanning region, is aproline-rich motif (PRM) with the consensus sequence aliphatic-aromatic-proline-X-aliphaticproline-X-proline (Al-Ar-P-X-Al-P-X-P) (O’Neal and Yu-Lee, 1993). It has been speculated thatthe PRM may be involved in binding to Src homology 3 (SH3) domains of downstream signalingmolecules (Horseman and Yu-Lee, 1994).10In the case of the prolactin receptor (PRL-R), a number of different receptor forms havebeen identified. These include a short form of 291 amino acids containing a cytoplasmic domainof 57 amino acids, a long form of 592 residues with a 358 amino acid cytoplasmic domain, andan intermediate length receptor (393 amino acids) found in rat Nb2 lymphoma cells (Kelly et a!.,1991, 1993). The different mRNAs which give rise to the short and long receptor forms arepresumed to arise from alternative splicing of a single gene, while the Nb2 form of the receptorappears to be due to a mutation resulting in the loss of a significant portion of the cytoplasmicdomain of the long form of the PRL-R (Kelly et at., 1989, 1991, 1993). These receptor isoformshave been shown to have tissue-specific expression and are also functionally different in terms ofthe biological effects which they mediate (Buck eta!., 1992; Kelly eta!., 1991; Lesueur eta!.,1991). In addition, different receptor forms exhibit differences in hormone binding affinity:compared to the long form of the receptor, the intermediate length Nb2 PRL receptor showed a3.3-fold increase in affinity for ovine PRL (oPRL). These results suggest that the intracellulardomain of the receptor may influence hormone binding affinity (Ali et a!., 1991).Polymorphism of receptor structure has been noted for other members of the hematopoietinreceptor superfamily including hGH, for which a soluble binding protein consisting of theextracellular domain of the receptor only has been identified (Baumann et a!., 1986; Herington eta!., 1986; Leung et a!., 1987). Variation in receptor structures may therefore be a general featureof these related proteins (Kelly eta!., 1991; Lesueur eta!., 1991; Smith et a!., 1988).1.4 Signal TransductionElucidation of the signal transduction mechanisms involved in the actions of the receptorsin the superfamily has been difficult because these proteins do not possess intrinsic enzymaticactivity found in a variety of protein ligand receptors. A number of possible signal transductionpathways have been suggested for prolactin and growth hormone (reviewed in Kelly et aL, 1991,1993) and there is evidence that, although the growth hormone receptor does not show anyhomologies to known tyrosine kinases, cytoplasmic tyrosine kinase activity is associated withthis receptor (Carter-Su eta!., 1989). Many of these studies were carried out using Nb2 cells11which have become a paradigm for studying mitogenic signaling pathways of prolactin. Furtherinvestigations indicated that tyrosine kinase activation may be an early event associated withligand binding to the prolactin receptor (Rui et at., 1992) and indeed, inhibition of tyrosinekinase activity was found to block the anabolic and mitogenic actions and thus, presumably, thesignal transduction mechanism of lactogens (Carey and Liberti, 1993). Signaling from receptorto nucleus by prolactin appears to occur via cytoplasmic proteins associated with the prolactinreceptor. Tyrosine phosphorylation of these associated proteins occurs following ligand bindingto the receptor and this initiates a kinase cascade which leads to phosphorylation and activationof transcription factors. Once modified, these proteins bind enhancer elements present in thepromoters of certain genes such as the interferon regulatory factor 1 (IRF- 1) gene and the Bcasein gene (David eta!., 1994; Gilmour and Reich, 1994; Schmitt-Ney et a!., 1991; Stevens andYu-Lee, 1994).Recent investigations have demonstrated that prolactin induces tyrosine phosphorylationand activation of the tyrosine kinase JAK2 (David et a!., 1994; Rui et a!., 1994a), suggesting thatthis enzyme is responsible for initiating the subsequent signal transduction events. A region ofthe prolactin receptor located proximal to the membrane-spanning region and comprising theproline-rich motif (PRM) has been identified as essential for interaction with JAK2 andsubsequent signal transduction (DaSilva et a!., 1994; Edery et at., 1994). However, although theassociation between JAK2 and the prolactin receptor appears to be very important for signaltransduction, JAK2 is not the only kinase associated with the prolactin receptor. Anotherimportant kinase is the RAF- 1 kinase: all forms of the PRL receptor were shown to be associatedwith the RAF-1 kinase in the PRL-dependent rat T-cell line Nb2, leading to phosphorylation andactivation of RAF-1 (Clevenger eta!., 1994). PRL stimulation of Nb2 cells was also found toactivate the protein tyrosine kinase p59fYfl, and revealed an association between the prolactinreceptor and this src family kinase (Clevenger and Medaglia, 1994). Thus, both these kinasesmay act as intermediaries in the signaling pathway of PRL-mediated T lymphocyte proliferation.JAK2 has also been shown to associate with the growth hormone receptor and to act as asignaling molecule for growth hormone (Argetsinger et a!., 1993; Carter-Su et at., 1994). As is12the case for prolactin, the proline-rich region located near the transmembrane-spanning region ofthe growth hormone receptor is required for association with JAK2 and for signal transduction(Billestrup et at., 1994; Frank et a!., 1994; Goujon et at., 1994; VanderKuur et at., 1994; Wangand Wood, 1995). There is also evidence that JAK2 associates with the receptor forerythropoietin and with other members of the receptor superfamily (Argetsinger et at., 1993;Silvennoinen et at., 1993; Witthuhn et at., 1993) and is required for interferon-’y-activated geneexpression (Wailing et at., 1993). Therefore, it seems likely that a common signaling cascademechanism involving JAK2 and/or other members of the JAK family exists for members of thecytokine/hematopoietin superfamily. However, while the signaling mechanisms may be similar,the marked differences in the physiological responses, as well as the variety of cell typesinvolved, suggest that there must be many parameters which regulate specific gene expression.Further study is required before a complete understanding of the mechanism of action of thesehormones and cytokines is obtained.1.5 Hormone-Receptor InteractionsA proposed mechanism of signal transduction for receptors with a single transmembranedomain is that of ligand-induced receptor oligomerization (reviewed in Ulirich and Schiessinger,1990). In brief, binding of the ligand induces receptor oligomerization which stabilizesinteractions between the cytoplasmic domains and activates the kinase functions which turn onthe signal transduction pathway. Investigations into the mechanism of hGH-receptor bindinghave revealed that receptor oligomerization is involved in signal transduction in this system.However, in contrast to the two ligand/two receptor stoichiometry which is commonly seen withgrowth factor receptors, the hGH-hGHbp complex consists of one hGH molecule bound to tworeceptor molecules (Ultsch et a!., 1991).Both physical and biochemical analyses have confirmed that hGH contains two receptorsites (sites 1 and 2) which each bind to the same region of the hGHbp (Cunningham eta!., 1991;de Vos et at., 1992). Thus, each hGH molecule is bivalent while the hGHbp, which usesessentially the same residues to bind to either site on hGH, is univalent. The hGH-(hGHbp)213complex forms sequentially: binding of a receptor molecule to site 2 only occurs followingbinding of a receptor to site 1. Thus, receptor binding at site 1 creates the binding site for thesecond receptor. Excess hGH can cause the hGH-(hGHbp)2complex to dissociate into one inwhich the hGHbp is bound exclusively at site 1 (Cunningham eta!., 1991). Binding of tworeceptors to one molecule of hGH leads to receptor dimerization, a prerequisite for signaltransduction: bivalent monoclonal antibodies (mAbs) to the hGHbp were as able as hGH toinduce cell proliferation while monovalent Fab fragments had no effect. In addition, a hGHvariant containing a mutation blocking receptor binding to site 2 was unable to activateproliferation (Fuh et a!., 1992).Receptor dimerization has also been demonstrated for the binding, in solution, of bGH tothe extracellular domain of the bGH receptor, using gel filtration (Staten et at., 1993). The samestudy also showed that bovine placental lactogen (bPL), which is able to elicit both GH- andPRL-like responses in vivo by binding to bGH receptors, formed a 1:1 complex with bGHbpunder the same experimental conditions. Although this result suggests the intriguing idea thatvery different mechanisms could be used to activate the same receptor, lack of evidence fordimerization of the extracellular portion of the receptor does not preclude the possibility ofdimerization of the full-length receptor (Wells, 1994). As is the case for glycophorin A(Lemmon et at., 1992), the transmembrane and/or intracellular portions of the growth hormonereceptor could contain some of the determinants for dimerization.Evidence has been obtained which suggests that dimerization is required to activate theprolactin receptors. Experiments with mAbs have shown that bivalent, but not monovalent antiPRL receptor antibodies possess PRL-like activity (Elberg et a!., 1990; Shiu et at., 1983).Furthermore, bivalent anti-PRLbp antibodies, with the concomitant dimerization of the PRL-R,were required for activation of JAK2 (Rui et a!., 1994b). Monovalent anti-PRL-R Fab fragmentswere unable to dimerize the receptor and activate JAK2, but activity could be partially restoredvia cross-linking with bivalent anti-Fab antibodies. Binding of hGH to Nb2 PRL receptors hasbeen shown to follow the two-site model of hGH-hGHbp interactions in that activity was seenwith low concentrations of hGH while high concentrations resulted in inhibition of activity (Fuh14et a!., 1993). In addition, although the X-ray structure of hGH bound to the extracellular domainof the hPRL-R (hPRLbp) revealed a 1:1 association between these proteins, comparisonsbetween this structure and the structure of the hGH-(hGHbp)2led to the suggestion that anintermediate form ready to bind a second hPRLbp and form an active complex was captured inthe crystal hGH-hPRLbp complex (Somers et a!., 1994). However, conflicting results have beenobtained with respect to the interaction of PRL with the PRL-R. One group using theextracellular domain of the rabbit PRL-R did not observe a 1:2 complex of hormone withreceptor (Gertler et aL, 1993), while another group using the extracellular domain of the rat liverPRL-R obtained data supporting the model of dimerization of the PRLbp by one PRL molecule(Hooper et al., 1993). A mutational analysis of hPRL binding to the hPRLbp revealed that whilealanine replacements of residues in the putative site 2 binding site of hPRL did not diminish themitogenic activity of the hPRL variants, replacements of some of these residues (A22, L25, andG129) with tryptophan or arginine resulted in variants with bioactivities in the Nb2 cell assaytwo to three orders of magnitude less than hPRL (Goffin et al., 1994). These results thussuggested the existence of a second binding site on hPRL for the Nb2 PRL receptor. Defmitiveproof of whether PRL forms a complex with two molecules of PRL-R will require adetermination of the crystal structure of the complex; however, the evidence so far collectedgenerally supports this two-site model.1.6 Mutational Analysis of Hormone Binding to ReceptorsA complete understanding of the interactions between a ligand and its receptor requirescomplementary information obtained from mutational mapping of the binding sites and from thethree-dimensional structure of the ligand-receptor complex. Although three-dimensionalstructures are available for only a few members of the hematopoietic superfamily, a number ofmutational analyses have been carried out and these have provided a wealth of informationregarding the interactions of these proteins with their receptors.151.6.1 Human Growth HormoneA consequence of the many studies which have focussed on examining the interactions ofhGH with its receptor is that this hormone is now one of the best characterized members of thehematopoietic superfamily. Human growth hormone is a 191 amino acid globular proteincontaining two disulfide bridges, four ct-helices in the core of the molecule, and an additionalthree short helical segments in the connecting loops between the heices (de Vos et aL, 1992;Figure 3). Naturally occurring growth hormone is a very heterogeneous polypeptide hormone(Baumann, 1991), and prior to the development of efficient bacterial expression systems for hGH(Chang eta!., 1987; Hsiung et a!., 1986) and the concomitant increased ease of performing site-directed mutagenesis studies, hormone-receptor interactions were investigated using natural andchemically-generated hormone variants and peptide fragments (Aubert et a!., 1986; Chawla eta!., 1983; Chêne eta!., 1989; Kostyo, 1974; Lewis, 1984). These studies were unable to providea clear understanding of which portions of the hGH molecule are responsible for biologicalactivity; however, they did indicate possible general areas of biological importance. Theapplication of molecular biology techniques such as site-directed mutagenesis to this problemhave since allowed a thorough analysis of the interactions of this hormone with its receptor.Helix 1 Helix 2 Helix 3 Helix 49 34 38 47 64 7072 92 94100 106 128 155 184N ((C53 165 182 189Figure 3. Diagram of hGH showing its two disulfide bridges and the locations of the helicalregions. Numbers indicate the position of the cysteine residues and the limits of the helicalregions. Labels indicate the four x-helices which form the core of the three-dimensionalhormone molecule. The short helical sections in the loops connecting the helices (residues 38 -47, 64-70, and 94- 100) are itidicated only as hatched boxes. hGH Fragments, Variants, and Chimeric ProteinsAnalysis of the growth promoting activity of recombined peptide fragments of bovine GH,namely residues 1 - 95 linked by a disulfide bond to 151 - 191, residues 96 - 133, and residues134 - 150, obtained from a tryptic digest of hormone isolated from pituitary extracts, indicatedthat residues 134 - 150 of the hormone were not necessary for biological activity (Hara et a!.,1978). The N-terminal 134 amino acid section of hGH, obtained from a plasmin digest of apreparation of pituitary hGH, was also sufficient to promote biological activity, although fullexpression of activity required the entire molecule (Reagan et a!., 1975). Recombinant proteinsobtained through the noncovalent interaction of the N-terminal 134 amino acid portion of hGHwith parts or the whole of the C-terminal portion (Li et at., 1976; Li and Blake, 1979; Reagan eta!., 1981) exhibited nearly full biological activity. These results suggested that the disulfide loopjoining C53 and C164 is not necessary for activity.A number of studies of the N-terminal two-thirds of the molecule suggested a morerestricted localization of growth-promoting activity to several different peptides, includingresidues 15 - 125 and 96 - 124 (reviewed in Chêne et at., 1989). A fragment consisting ofresidues 96 - 133, containing helix 3 of growth hormone (amino acids 109 - 126), was isolatedfrom tryptic digests of bovine growth hormone (bGH) and had growth promoting activity (Grfand Li, 1974; Gráf et a!., 1976; Yamasaki et at., 1970). Recently, the role of helix 3 has beeninvestigated further through the synthesis of several peptides corresponding to residues in thisregion. No stimulation of Nb2 cell growth, as normally observed with the presence of primateGH’s, was seen with a peptide with the same amino acid sequence as that from residues 110 - 127of hGH; however, replacement of El 17 with leucine enhanced the amphipathic nature of thehelix and stimulated mitogenesis (Komberg and Liberti, 1992). Although intriguing, theseresults contradict findings from mutagenesis studies (discussed in Section and have notbeen followed up. Another peptide encompassing residues 108 - 129 of hGH elicited amitogenic response in a system using preadipocytes wherein native hGH is anti-mitogenic, butits actions appeared to be mediated through binding to a site other than the growth hormonereceptor (Jeoung et aL, 1993). Unfortunately, while these results suggest that helix 3 may play17an important role in signal. transduction they do not provide a clear picture of the region’scontribution to the biological activity of hGH.In general, whilst suggestive, the use of peptide fragments to identify regions of growthhormone important for biological activity, or to determine the contribution of particular segmentsto activity, has not been very informative. Although these studies led to the idea that growthhormone may contain several “active sites,” the weak receptor binding typically displayed bypeptide fragments and the possibility that differences in activity between the fragments and thecomplete hormone could simply reflect differences in structure prevent a thorough dissection ofhormone structure as related to function.The use of natural variants of the hormones is one way to get around the potential problemsof peptide fragments. Among the natural variants of hGH examined is a 20,000 Da proteinlacking the 15 amino acids between residues 32 - 46 (Lewis et a!., 1980). The growth-stimulating abilities and lactogenic effects of this variant were found to equal those of full-lengthhGH; however, a marked decrease in the insulin-like activity suggested that residues 32 - 46 maybe involved in this activity (Chêne et at., 1989). Although the 20 kDa variant was found to be aseffective as wild-type hGH in causing dimerization of the hGHbp (Cunningham et at., 1991),differences in affinities between the two forms of hGH in binding to receptors have beenreported and these may give rise to the variations in biological effects (Beattie, 1993). In anycase, residues 32 - 46 of hGH appear to define a significant but less critical part of the bindingdomain of hGH (Sigel et al., 1981; Wohnlich and Moore, 1982).Other attempts to relate structural features to biological properties focussed on sequencecomparisons between related hormones (Nicoll eta!., 1986). These results implicated certainclusters of residues as being potential components of a hormone-specific binding domain;however, these conclusions also exhibited a number of inconsistencies, including identificationof putative prolactin-specific clusters that were not found for other lactogenic hormones such ashGH and human placental lactogen (hPL). One drawback to using sequence comparisons todelineate regions of functional importance is that the multiple amino acid variation between18hormone sequences means that the conthbution of specific residues to hormone function cannotbe determined.Chimeric hormones, created by combining fragments from related hormones, have allowedthe contribution of specific portions of the hormones to biological activity to be examined whileavoiding severe disruptions to the protein structure. Recombinant hormones containingfragments of hGH and hPL in which the N-terminal fragment (amino acids 1 - 134) of each wascombined with the C-terminal fragment (residues 141 - 191) of the other, prepared by thecovalent joining of peptide fragments generated by limited plasmin digestion and reduction,revealed that the biological activities of each hormone resided in the N-terminal fragment. Thus,the chimeric hormone consisting of the N-terminal portion of hGH fused to the C-terminalportion of hPL (hGH-hPL) exhibited activities similar to native hGH, while the hPL-hGHchimeric hormone had lactogenic activity only (Russell et at., 1981). The N-terminal sectiontherefore appeared to contain binding site(s) responsible for specificity while the C-terminal wasinvolved in stabilizing the active conformation of the hormone.The affinity towards lactogenic receptors of a chimeric hormone consisting of residues 1 -23 of the nonlactogenic bGH joined to residues 24 - 191 of hGH, prepared by recombinant DNAtechnology and produced in E. coli, was found to be several orders of magnitude lower than thatof hGH alone while the affinity towards somatogenic receptors was decreased to a much lesserdegree. This result indicated that the binding site of hGH with high affinity for lactogenicreceptors includes residues in the N-terminal 23 amino acid portion of the hormone (Binder etat., 1989). Replacement of the N-terminal 13 amino acids of hGH by the corresponding bGHsequence had only a slight effect on binding to lactogenic receptors while binding tosomatogenic receptors was reduced to a higher degree (Binder et at., 1990). In earlier studies thebiological activity of hGH molecules lacking the N-terminal 7 or 13 residues, created by site-directed mutagenesis, was examined and it was found that, compared to full-length hGH, themolecule lacking 13 amino acids had reduced affinity towards the lactogenic receptors of Nb2lymphoma cells while the loss of the N-terminal 7 residues had no effect (Ashkenazi et at., 1987;Gertler et at., 1986). After taking all these results into account, along with data from an attempt19to identify functionally important domains through hydropathy plots (Nicoll et a!., 1986) theinvestigators suggested that the high-affinity hGH binding site for lactogenic receptors may belocated between residues 8 and 18 (Binder et a!., 1989). Similar conclusions were reachedfollowing a study in which high affinity mAbs with a defmed epitope specificity profile againsthGH were used to inhibit‘25I-hGH binding to lactogen and somatogen receptors (Strasburger eta!., 1989). Monoclonal antibodies that affected binding to one type of receptor did notnecessarily affect binding to the other. These researchers suggested that the lactogen receptorbinding site is mainly confined to the N-terminal portion of hGH, particularly residues 8 - 13,while sections of the N-terminal and C-terminal are necessary for binding to the somatogenreceptor.Further investigations into the biological activities of the hGH mutant lacking the first 7amino acids revealed that, although this hormone retained full ability to bind to lactogenicreceptors, it exhibited a markedly reduced ability to bind to somatogenic receptors and was alsounable to promote diabetogenic or insulin-like activity. Therefore, the initial 7 residues at the N-terminal end of hGH appear to be required for expression of the growth promoting activity(assessed on the basis of weight gain in hypophysectomized rats), diabetogenic activity(determined with respect to glucose tolerance in ob/ob mice), and insulin-like activity (measuredas the ability to stimulate14C-glucose oxidation by epididymal adipose tissue ofhypophysectomized rats) (Towns et at., 1992). These effects are presumably mediated by theinvolvement of this portion of hGH in somatogenic receptor dimerization: removal of the first 8residues of hGH has been noted to interfere with binding at site 2 and therefore prevent hGHbpdimerization (Cunningham et at., 1991). However, since removal of the first 7 residues did notdecrease affinity for the Nb2 lactogenic receptors while removal of the first 13 residues did(Ashkenazi et a!., 1987; Gertler et a!., 1986), dimerization of the lactogenic receptors by hGHmay involve slightly different residues than those necessary for dimerization of the somatogenicreceptors.While the conclusions from the various studies discussed in this section do not distinguishthe specific residues involved in the interaction of hGH with somatogenic and lactogenic20receptors they do identify regions of the hormone that are critical for activity. Thus, earlyinvestigations, analyzing peptide fragments, localized biological activity to the N-terminalportion of hGH since the first 134 amino acids were sufficient to promote most of the function ofthe hormone. Correlation of various activities of hGH to discrete sequences within this regioncame from later studies analyzing chimeric hormones. Specifically, residues 32-46 were foundto be required for insulin-like activity, amino acids 8 - 18 were necessary for lactogenic activity,and the first 7 residues were important for somatogenic activity. In addition, although theevidence was not clear-cut, there was an indication that residues within helix 3 of hGH (aminoacids 109 - 126) may be involved in signal transduction. Site-Specific Mutants of hGHA high resolution functional analysis of hGH-receptor binding has followed thedevelopment of efficient bacterial expression systems for recombinant hGH (Chang et a!., 1987;Hsiung eta!., 1986), the determination of a structural model for pGH (Abdel-Meguid eta!.,1987), the cloning of the hGHbp (Leung et at., 1987), and efficient bacterial production of thisextracellular domain of the hGH receptor (Fuh et a!., 1990). The ability to easily produce theextracellular portion of the receptor led to the development of an assay wherein the contributionof specific hGH residues in binding to the hGHbp, and thus presumably to the activity of thehormone, was determined with respect to whether these side chains strongly modulated bindingto the receptor. In contrast, the activities of the variants and chimeras discussed in the previoussection were determined by a number of bioassays: loss of bioactivity was presumed to relate todisruption of binding interactions between the hormone and various receptors.Regions of the hormone involved in receptor binding were initially mapped using astrategy called homolog-scanning mutagenesis (Cunningham et a!., 1989). Using the model ofpGH (Abdel-Meguid et a!., 1987) as a guide, a set of chimeric hormones that collectivelymodified 85 out of 191 residues in hGH was created by substituting segments (7 to 30 residues)of hGH with corresponding regions from homologous hormones that do not bind to the hGHbp(namely hPL, hPRL, and pGH). Sequences chosen for substitution were within borders ofsecondary structure and were thus limited to specific putative helical or ioop regions. In21addition, alteration of residues on the hydrophobic faces of the helices was avoided sincemutations of buried residues in hydrophobic cores are generally destabilizing (Alber et al., 1987;Pakula et a!., 1986). It was found that the segment substitutions which affected binding ofcertain of the chimeric proteins to the hGHbp were from three discontinuous regions of thehormone: the N-terminal section of the first helix, a portion of the ioop connecting helices 1 and2, and the C-terminal section of the fourth helix. When these segments were mapped on thethree-dimensional model of the hormone it was apparent that although discontinous in terms ofprimary structure, they are continuous with respect to tertiary structure in that they form a singlepatch (Cunningham et al., 1989).The specific residues that modulate binding to the hGHbp were identified using a techniquecalled alanine-scanning mutagenesis (Cunningham and Wells, 1989). In this technique manydifferent mutants are prepared, each containing a single alanine substitution of a residue withinthe region of interest. Although analysis of these variants can identify which residues areinvolved in function, the results can also be misleading since it is possible for nearby side chainsin the three-dimensional structure of the protein to substitute for the residue replaced by alanine.Therefore, a lack of loss of activity following an alanine replacement does not guarantee that thatspecific residue is not involved in the function of the protein. In the case of hGH, eachindividual residue within the three segments already implicated in receptor binding was mutatedto alanine, the mutant proteins were expressed in a secreted form from E. coli, and the bindingconstant to the extracellular domain of the hGH-R was determined for each mutant bycompetitive displacement of125-labeled hGH. Within helix 1, alanine substitutions at residues6, 10, and 14, all located on the same face of the helix, had the most disruptive effect on binding.In addition, residues 54, 56, 58, 64, and 68 of the ioop region joining helices 1 and 2 and residues171, 172, 174, 175, 176, 178, 182, and 185 of the C-terminal portion of helix 4 were alsoidentified as being involved in binding to the receptor (Site 1 in Figure 4). In general, the sidechains of these various residues extend from the same side of hGH and together form thediscontinuous binding site first identified through homolog-scanning mutagenesis. The inability22of peptide fragments to mimic this discontinuous binding site explains why earlier studies usingpeptides were not able to precisely define this region.The data obtained from the mutagenesis studies support a number of earlier investigationsin which chemical modification was used to determine the contribution of specific residues toreceptor binding. Ethoxyformylation of H19 and/or H21 (both in helix 1) interfered with thereceptor binding ability of equine GH (Fukushima et a!., 1987). Covalent labeling of K70 ofhGH (located in the loop between helices 1 and 2) with fluorescein isothiocyanate resulted in ahormone variant with decreased mitogenic ability in the Nb2 cell bioassay and lowered affinitytowards somatogen receptors in bovine liver and towards lactogen receptors in Nb2 cells (Sakaleta!., 1991). In addition, acetylation of K167 or K171 (both located in helix 4) with3H-aceticanhydride disrupted binding of hGH to rat liver somatogenic receptors (Teh and Chapman,1988).Figure 4. Structural model of hGH indicating the location of sites 1 and 2 for binding to thehGHbp. Residues where alanine substitutions reduced binding to the extracellular domain of thereceptor are indicated by closed circles (for site 1) and squares (for site 2), with increased symbolsize correlating with increased reduction in binding affinity. The open circle indicates that thevariant E174A displayed an increase in binding affmity for the hGHbp. The site 1 residuesindicated were identified as those in which alanine mutations caused a fourfold or greater changein the dissociation constant for the soluble portion of the receptor. Site 2 residues were identifiedusing a fluorescence quenching dimerization assay wherein mutants were examined for theirability to disrupt receptor dimèrization without affecting binding to site 1. The figure is takenfrom Cunningham eta!., 1991.23As discussed above, the mutagenic analysis of the interaction between hGH and the hGHbpidentified three regions of the hormone that form the binding domain for the receptor: the N-terminus of helix 1; the ioop between helices 1 and 2; and the C-terminus of helix 4(Cunningham and Wells, 1989). The functionally important residues in these regions form asingle binding site, designated Site 1 in Figure 4. These results were obtained prior to thediscovery that a single hGH molecule binds two receptor molecules. Indeed, the second site wasnot identified in these early studies because the monoclonal antibody used to detect receptorbinding blocked binding of a second hGHbp to hGH and thus only allowed formation of amonomeric complex (Cunningham et at., 1991). Receptor dimerization was demonstrated usinga sensitive solution assay in which quenching of the fluorescence from a fluorescein-labelledhGHbp was measured following the addition of hGH: the addition of 0.5 M equivalents of hGHcaused maximal fluorescence quenching (Cunningham eta!., 1991). Using this assay a numberof hGH variants, including deletion mutants, chimeric proteins of hGH with hPL or hPRLsegment substitutions, and alanine-substitution mutants were examined for their ability to reducereceptor dimerization without affecting site 1 binding. Those residues for which substitutionswere the most disruptive were mapped to the N-terminus and to the centre of helix 3 (Site 2 inFigure 4) (Cunningham eta!., 1991). These results supported an earlier study wherein exonexchange between the hGH gene and the rat PRL or rat GH genes revealed that the regionencoded by exon 4, which includes helix 3, is required for somatogen receptor binding (Ray eta!., 1990) and were confirmed by the crystal structure of the hGH-(hGHbp)2complex whichindicated that residues in this region form binding determinants (de Vos et a!., 1992).The importance of helix 3 residues (amino acids 109 - 126 in hGH) to receptor binding andbiological activity has been confirmed by several other studies. Transgenic mice expressinghGH analogs containing the mutation G12OR exhibited a growth-suppressed phenotype, with thedegree of growth suppression being directly correlated with the serum levels of G12OR. On theother hand, an enhanced growth phenotype was noted with mice expressing a G12OA mutation(Chen et a!., 1994). Studies on bGH indicated the same involvement of helix 3 in biologicalactivity as was shown for hGH (Chen eta!., 1991a,b, 1995). Human GH molecules containing24mutations in helix 3 which disrupt site 2 binding retain a functional site 1 and are able to bind tothe hGHbp. However, the G12OR mutation, by sterically blocking site 2, prevents receptordimerization and thus this variant acts as a hGH antagonist (Fuh et at., 1992).As listed above, the various mutagenic analyses of the interaction of hGH with the hGHbphave identified a number of residues that contribute to the binding of the hormone to thereceptor. However, although alanine scanning mutagenesis identified approximately 30 sidechains as being involved in binding, an analysis of the energetic importance of these side chainsindicated that approximately 85% of the binding energy is accounted for by only 8 of theseresidues (Cunningham and Wells, 1993). Thus, the functional binding site of hGH, consisting ofthose residues which when substituted with alanine caused a greater than twofold reduction inbinding affinity, is much smaller than the structural binding site, which is composed of theresidues at the contact interface between the hormone and receptor. Some of the structuralcontacts identified by X-ray crystallography are functionally silent or even deleterious asmeasured by alanine scanning mutagenesis. These residues include H18, H21, R167, K168, andE174 in site 1 and N12, R16, R19 and N109 in site 2. It is not yet clear why contacts betweenthe side chains of these residues and the receptor may be functionally silent but it is possible thatthe energy required to desolvate these charged and hydrogen-bonding groups may not be offsetby the free energy gained from binding (Cunningham and Wells, 1993).From the crystal structure of the growth hormone-receptor complex it can been seen thatmost of the side chains comprising the functional binding site do contact the receptor (de Vos eta!., 1992). The remaining residues identified by the mutagenesis studies, including FlO, F54,158, and F176, are buried in hydrophobic clusters and substitutions of these may disrupt thebinding of contact residues via structural changes (Wells and de Vos, 1993). An analysis of thecontributions of residues in the structural binding site of the hGHbp to the binding affinitysimilarly revealed that only a few of these residues, those mapping to the centre of the region,form the functional binding site (Clackson and Wells, 1995). When examined in the context ofthe hGH-(hGHbp)2complex, the functional binding sites of hGH and the hGHbp join together toform a tightly packed hydrophobic core surrounded by hydrogen bonds and five intermolecular25salt bridges. Thus, only a few interactions seem to be important for tight binding. However, theresidues identified in the structural but not the functional binding site may play important rolessuch as increasing the rate of hormone-receptor association or contributing to the specificity ofbinding (Clackson and Wells, 1995). Indeed, as discussed in the following section, the locationof these functionally “null” contact residues is in the region important for binding hGH to thehPRLbp (Cunningham and Wells, 1991). hGH Interactions with PRL ReceptorsThere are many similarities between growth hormones, prolactins, and placental lactogens,but in general the functions and binding specificities of each are distinct. Among the fewexceptions to this are growth hormones from primates which, in addition to having growthhormone activities, exhibit lactogenic activity as a result of binding to lactogenic receptors.However, as discussed in section, there is much evidence to suggest that differentresidues of the hGH molecule are responsible for the somatogenic and lactogenic activities(Ashkenazi et at., 1987; Binder et at., 1989, 1990; Gertler et at., 1986; Strasburger et at., 1989;Towns et at., 1992). This was also suggested by the observation that chimeric hormones createdby switching exons 3 or 4 of the hGH gene with the corresponding exons of the rat PRL or ratGH genes retained some degree of lactogen receptor binding while losing the ability to bind tosomatogen receptors (Ray et at., 1990). The differences between the somatogen and lactogenreceptor binding determinants were confirmed by a mutagenic analysis of the interaction of hGHwith the hPRLbp which indicated that the binding determinants on hGH for the growth hormoneand prolactin receptors overlap but are not identical (Cunningham and Wells, 1991). Alaninescanning mutagenesis identified a number of hGH residues which, when mutated to alanine,caused a greater than fourfold reduction in binding affinity to the prolactin receptor. Theseresidues are found in the central section of helix 1 (H18, H21, and P25), in the loop regionbetween helices 1 and 2 (158, N63, and S62), and in the middle of helix 4 (R167, K168, K172,E174, F176, and R178). When mapped onto the structural model of hGH, the patch formed bythese residues is similar to that formed by residues comprising the binding site for the hGHbp,but is shifted towards the centre of helices 1 and 4 and does not include the N-terminus of helix 126or the extreme C-terminus of helix 4 (Cunningham and Wells, 1991) (Figure 5). Within thispatch are residues M64 and M179: oxidation of either one or both of these amino acids resultedin a marked reduction in affinity of hGH and hPL for lactogenic receptors (Teh et at., 1987).The differences between the binding determinants of hGH for the growth hormone andprolactin receptors are great enough that receptor-specific variants were created by mutatingresidues identified as being essential for binding to one receptor but not the other (Table 1).Thus, the double mutant resulting from a combination of the mutations K168A and E174A,which each preferentially disrupt binding to the hPRLbp, exhibited a 34,000-fold shift inpreference for binding to the hGHbp. Similarly, binding to the hPRLbp was enhanced nearly150-fold by combining the mutations R64A and D171A, each of which disrupt binding to thehGHbp. Binding to both receptors was deleteriously affected by combining mutations atresidues such as K172 and F176 which are important for binding of hGH to both the hPRLbp andthe hGHbp.Table 1. Binding of Double Mutants of hGH to the hPRLand hGH Binding Proteins (hPRLbp and hGHbp)aChange in receptorhPRLbp hGHbp preference:CMutant Kd(mut)/Kd(hGH)b Kd(mut)/Kd(hGH) hPRLbpIhGHbphGH 1 1 1K168A1E174A 9100 0.27 34000R64AJD171A 1.9 280 0.0068K172A/F176A 8400 560 15a Data are taken from Cunningham and Wells, 1991.b The dissociation constants for variants of hGH to the hGHbp, [Kd(mut)], weremeasured by competitive displacement of 115 I-labeled hGH from the hGHbp. Amonoclonal antibody to the hGHbp was used to precipitate the‘251-hGH-hGHbpcomplex. The ratio Kd(mut)/Kd(hGH) indicates the relative reduction in bindingaffmity of the hGH variants to that of wild-type hGH for the hGHbp.The change in receptor preference was calculated by dividing the ratio of therelative reduction in binding affinity for the hPRLbp by that for the hGHbp.27A major difference between hGH binding to the growth hormone and prolactin receptorswas revealed in an earlier study which indicated that Zn2 is required for binding of hGH to thehPRLbp (Cunningham and Wells, 1990a) (Figure 5). Zn2 is not required for fonnation of thehGH-(hGHbp)2complex or for binding of hPRL to the hPRLbp. The ligands identified inmediating binding of hGH to the hPRLbp in the presence ofZn2are H18 and E174 which, inconjunction with two ligands on the receptor (D217 and H218) form aZn2-binding site (Somerset at., 1994). Initially, it was surmised that H21 is also involved in binding Zn2 (Cunningham etat., 1990a); however, crystallographic analysis revealed that this residue is not a direct ligand tothe Zn2 but is instead required to correctly orient the side chain of E174 (Somers et a!., 1994).The possible importance of residues within the loop region (residues 54 - 74) for PRLreceptor-binding specificity was investigated by a comparison of the sequence of the lactogenicand somatogenic hGH with the somatogenic but non-lactogenic bGH: residues 57 and 60 differin these hormones. T57 and A60 of bGH were changed to the corresponding hGH residues(serine and threonine, respectively) in order to determine if these residues are important for thelactogenic activities of hGH (Chen et at., 1991c). In fact, the PRL receptor-binding properties ofthe variants were unchanged from that of wild-type bGH, suggesting that these residues do notdetermine the receptor-binding properties. More recently, another mutagenic analysis of residueswithin the 54- 74 loop of hGH was carried out in order to investigate the involvement of theseresidues in lactogenic and somatogenic function (Sakal et a!., 1993). The affmity of the P61Amutant for the hGHbp was 12.5% that of the wild-type hGH and the affinity of a double mutantP59A/P61A was 7%. In both instances the variant hormones formed 1:2 complexes with thereceptor. However, while these mutations affected binding to the hGHbp, the ability of thesevariants to stimulate proliferation of Nb2 cells was unchanged from that of unmodified hGH.These results suggest that different amino acids in the residue 54 - 74 loop region are responsiblefor interacting with somatogenic and lactogenic receptors or that this region is not involved inlactogenic receptor activation.The X-ray structure of a hGH molecule complexed to an hPRLbp (Somers et at., 1994)showed that although essentially the same hormone and receptor residues form the hormone-28receptor interfaces for both the hGH-hGHbp and hGH-hPRLbp complexes, the orientations ofthe N- and C-terminal domains of the receptors are not identical. Since each hormone moleculebinds two receptor molecules, this difference in domain orientation may be important in order toprevent the formation of mixed complexes wherein the hormone is bound to both a hGH-R and ahPRL-R.Figure 5. Structural model of hGH indicating the binding site for the hPRLbp. Residues wherealanine substitutions caused a reduction in binding affinity are indicated by closed, graduatedcircles, with increased symbol size correlating with increased reduction in binding affinity. Thedotted lines joining residues H18 and E174 to the Zn2 molecule indicate the role that theseamino acids play in forming part of the binding determinant for Zn2. Zii2 is required forbinding hGH to the hPRLbp but not for binding hPRL to the hPRLbp. Figure is taken fromCunningham and Wells, 1991.1.6.2 Placental LactogenVery few studies have been carried out to investigate which residues of placental lactogenare involved in receptor binding. Indeed, very little is known in general about this member of thehormone family. Placental lactogen, in common with other lactogens, binds to the PRL-R but aspecific PL-R also has been identified (Freemark and Corner, 1989). Detailed analyses of PLreceptor binding have not yet been carried out. However, the data that have been obtained froma few structure-function analyses of placental lactogen have provided insight into the receptor-29mediated activities of this hormone and have reinforced the concept that the receptor-bindingdeterminants of members of this hormone family are very similar.A comparison of the sequences of six related mouse proteins, three lactogenic (mPRL,mPL-I, and mPL-II), and three non-lactogenic [mGH, proliferin (mPLF), and proliferin-relatedprotein (mPRP)], identified five residues that are conserved among the lactogenic hormones butare different in the nonlactogenic hormones (Davis and Linzer, 1989b). The locations of theseresidues (R14 near the N-terminus of helix 1 and R165, R169, K179, and N189 in the C-terminalhalf of helix 4) are in regions which in the related protein hGH are important for hGH-hGHbpinteractions (Cunningham and Wells, 1989) and it therefore seemed likely that these mPL-I1residues might be involved in binding to the prolactin receptor. A mutagenic analysis of theseresidues revealed that mPL-ll variants with mutations at R14 (in putative helix 1), R169, or K179(in putative helix 4) were unable to stimulate the growth of Nb2 cells or bind to the hepatic PRLreceptor (Davis and Linzer, 1989b). Thus, these residues appear to be critically important forhormone interaction with the lactogenic receptor. R165 and N189 were not, in themselves,absolutely required for receptor-binding; however, a combination of mutations at either of thesepositions with mutations at R169 or K179 caused a further decrease in the ability of theseproteins to stimulate Nb2 cell growth. The involvement of helices I and 4 in binding of mPL-IIto the mPRL receptor was further indicated by the inability of chimeric proteins, in which helix 1or 4 of mPL-ll were replaced with the corresponding regions of the non-lactogenic mPLF, tobind to the receptor.Another mutational analysis of mPL-II focussed on determining what, if any, role theconserved cysteine residues play in receptor binding and hormone function (Davis and Linzer,1989a). Placental lactogen, like growth hormone, contains two disulfide bonds: one connectingthe loop region between helices 1 and 2 with helix 4 (C51 and C166 in mPL-II); and one at theC-terminus (C183 and C191 in mPL-fl). These cysteines are conserved throughout the growthhormone/prolactin/placental lactogen family of hormones. Disruption of either disulfide bondthrough the mutation C51S or C183S did not lead to a change in the affinity for the PRLreceptor. However, while the mitogenic activity of the C183S mutant was unaffected, the C51S30mutant was unable to stimulate the growth of Nb2 cells. Therefore, disruption of the C51 - C183disulfide bond and/or mutation of residue C5 1 resulted in a variant hormone able to bind the PRLreceptor but not leading to receptor-mediated mitogenesis. In order to cause this effect the C51Svariant may interfere in some manner with receptor dimerization, but the exact nature of theinteraction of placental lactogen, or of this variant, with the prolactin receptor is as yet unknown.Bovine placental lactogen differs from that of other species including human and mouse inthat this hormone has an additional 13 N-terminal amino acids. A third disulfide bond iscontained in this region, a feature commonly found in prolactins from mammalian species. Theimportance of the N-terminal region, including the N-terminal “nose” as well as helix 1(postulated to begin at residue 19), to receptor-binding and subsequent signal transduction wasinvestigated through the creation and analysis of a number of N-terminus-truncated bPL variants(Gertler et a!., 1992). Removal of the first 15 residues caused an approximately 50% decrease instimulation of Nb2 cells but resulted in an increase in somatogenic receptor-mediated 3T3-F442A preadipocyte bioactivity. Further truncations by removal of up to 20 residues essentiallyabolished lactogenic activity and also decreased somatogenic activities, albeit not to the sameextent as the lactogenic actions. It was thus suggested that the N-terminal “nose” of bPL may berequired for binding to the PRL receptors but not for the interaction between bPL andsomatogenic or unique bPL receptors (Gertler et at., 1992). However, while this may be the casefor bPL, placental lactogens of other species lack this N-terminal 13 amino acid sequence but arenonetheless able to bind to the prolactin receptor and cause receptor-mediated mitogenesis(Davis and Linzer, l989b; Deb eta!., 1989; Lowman et at., 1991).While the affinities of hPL and hGH for the hPRLbp are almost identical (in the presenceof 50 p.M ZnC12)hPL binds very weakly to the hGHbp, even though hPL shares 85% sequenceidentity with hGH (as opposed to 23% with hPRL) (Lowman et at., 1991). This difference inbinding affinities is not unique: mammalian growth hormone receptors have high affinity formammalian growth hormones and sheep or goat placental lactogens but low affinity formammalian prolactins and primate placental lactogens; and lactogenic receptors bind lactogenichormones but have low affinity for somatogenic hormones (Freemark and Corner, 1989; Lesniak31et al., 1977). The similarity in binding affinities of hGH and hPL for the hPRL receptor appearsto be a consequence of the similarity in critical amino acid sequence in these hormones for thehPRLbp (Lowman et al., 1991): the only difference in the side chains involved in binding to thehPRLbp is residue 25 (phenylalanine in hGH, isoleucine in hPL). The affinity of hPL for thehGHbp is 2,300-fold less than that of hGH for the hGHbp; however, by altering residues withinthe receptor-binding site the specificity of hPL for the hPRL and hGH receptors could beselectively changed (Lowman et al., 1991). Thus, substitution of four residues from hGH intohPL (V411, D56E, M64R, and M1791) raised the affinity to the hGHbp from 2,300-fold weakerthan hGH to only 11-fold weaker. The binding affinity of this variant for the hPRLbp was onlyslightly reduced. However, substitution of E174, involved in coordinating Zn2 in both hPL andhGH, by alanine decreased the affinity for the hPRLbp by 1,400-fold. The combination of thismutation with those listed above increased the affinity of hPL for the hGHbp to only 6-fold lessthan hGH. One additional mutation, M64K, shown previously to enhance hGH binding to thehGHbp (Cunningham et al., 1990b), increased hPL affinity to 1.6-fold less than hGH.1.6.3 ProlactinAs described above, the members of the growth hormone/prolactin/placental lactogensuperfamily of hormones have many similarities with respect to receptor interactions, althoughsubtle but important differences have been noted. Investigations have been carried out to definethe structure-function relationships of prolactin, and these have confirmed the resemblance togrowth hormone and placental lactogen. However, much work remains to be done before acomplete understanding of how prolactin elicits all of its various physiological effects isobtained. Prolactin VariantsProlactin is a heterogeneous hormone in that it undergoes a number of different types ofpost-translational modifications including glycosylation, phosphorylation, deamidation,aggregation, and proteolysis. Prolactin also plays a role in regulating many different biologicalprocesses; however, whether the different PRL variants mediate different biological functions is32unknown (see Cole et at., 1991 and Smith and Norman, 1990 for reviews of variants andactivities).One of the prolactin variants which has been studied in terms of its ability to give rise toPRL-like effects is a 16 kDa N-terminal fragment, generated by proteolytic cleavage followed byreduction of the disulfide bonds. Obtained initially from rat and then from human, this fragmentwas reported to act as a mitogen for Nb2 cells, albeit at a lower activity than full-length rPRL(Clapp et at., 1988). However, since this fragment was obtained by in vitro cleavage of full-length rPRL it is possible that some of the observed bioactivity may have been a result ofcontamination with native 23 kDa rPRL. In support of this, independent studies showed noactivity of 16 kDa fragments of recombinant hPRL in the Nb2 cell assay (Dr. Goffin, InsermUnite 344, Endocrinologie Moleculaire, Paris, France, personal communication). Aninvestigation into the binding characteristics of the 16 kDa rPRL fragment provided evidence forspecific 16 kDa rPRL-binding sites, particularly in brain and kidney membranes (Clapp et at.,1989). Further research showed that the rat and human 16 kDa fragments, but not full-lengthrPRL or hPRL inhibited in vitro angiogenesis, that is, the regulation of new blood vesselformation (Clapp eta!., 1993; Ferrara eta!., 1991). It was suggested that this effect is mediatedthrough a receptor specific for the cleaved form of PRL (Ferrara et at., 1991). The 16 kDafragment is missing helix 4, which in GH and PL contains many of the receptor bindingdeterminants. Therefore, the lack of this region in the 16 kDa PRL suggests that the manner inwhich this PRL variant binds to its own receptor may be distinct. Site-Specific Mutants of PRLSequence comparisons of growth hormones, placental lactogens, and prolactins revealedseveral clusters of residues whose sequences were conserved in the lactogenic hormones but notin the non-lactogenic proteins of this group. The locations of these regions are the N-terminus ofputative helix 1, the C-terminal part of the ioop connecting putative helices 1 and 2, the middleof putative helix 2, and the N-terminal section of the loop connecting putative helices 3 and 4.The existence of these “lactogen-specific” residues suggested that lactogenic activities could bespecifically associated with these portions of the hormones (Nicoll et at., 1986). However, as33mentioned previously (see section this sequence comparison was unable to define thefunctional contribution of specific residues because the sequences of the hormones examined donot vary systematically. The role of individual amino acids is best determined through chemicalmodification and/or site-directed mutagenesis.The contribution of arginine residues to binding of PRL to lactogenic receptors wasexamined by treating oPRL with 1,2-cyclohexanedione which selectively modifies arginines:modification of the hormone resulted in decreased binding to receptors (Cymes et at., 1994).The reduction in affinity appeared to be due to modification of two arginine residues whichreacted with the reagent faster than the other two arginines. The identity of the two morereactive arginines could not be precisely determined: they could be any of residues 21, 125, 176,and 177. Chemical treatment was also used to determine the importance of the three disulfidebonds of PRL to biological activity. Selective reduction and alkylation of the disulfide bonds ofoPRL revealed that neither the N- nor C-terminal loop is required for mammotrophic activity.However, cleavage of all three disulfide bridges abolished biological activity, suggesting that thedisulfide bridge joining the loop region between helices 1 and 4 with helix 4 was crucial forhormone activity (Doneen et at., 1979). The importance of this disulfide bond was confirmedwhen site-directed mutagenesis was used to replace the cysteine residues with serines in eitherbPRL (C58 and C174, Luck et at., 1992) or hPRL (C58, Goffin et at, 1992). The resultingmutant protein was unable to stimulate growth of Nb2 cells. Replacing the cysteines which formthe N- and C-terminal loops with serines had no effect on the mitogenic activity of the hormone.The function of the C58 - C174 bond may be to maintain the proper spatial arrangement of thereceptor binding determinants: disruption of the bridge could allow portions of the hormonewhich are normally held together to separate, and this could affect the ability of the hormone tobind to the receptor and induce mitogenesis.The importance of the loop region connecting putative helices 1 and 2 to the bioactivity ofhPRL was investigated by alanine-scanning mutagenesis (Goffin et at., 1992). Ten alaninemutants were generated in order to determine whether any of these residues is important foreither binding of the hormone to the Nb2 lactogenic receptor or stimulation of Nb2 cell growth.34Binding affinties of the hPRL mutants was determined by measuring the ability of each mutant tocompete with 125 I-labeled unmodified hPRL for binding to Nb2 cell homogenates. Mutation ofresidues P66 and K69 had the most disruptive effect on the receptor binding and mitogenicability of the hormone. The concentration at which half-maximal displacement of the 1251.labeled hPRL occurred (the IC50)was 3.3-fold higher for P66A and 10-fold higher for K69Athan for unmodified hPRL. Similarly, the bioactivity of P66A, as measured by the ability of themutants to stimulate propagation of the Nb2 cells, was 25% that of wild-type and the bioactivityof K69A was lower than that of wild-type hPRL by two orders of magnitude. Mutation of theother residues within the ioop region did not severely affect hormone activity; in fact, S61A andQ74A were more active than wild-type hPRL (122% and 177%, respectively) and exhibited IC50values slightly lower than that of unmodified hPRL. Thus, as for hGH in its interaction withboth the hGHbp and the hPRLbp, the loop region connecting putative helices 1 and 2 of hPRLappears to be involved in receptor binding. However, the specific residues involved in binding tothe receptor differ between these related hormones: amino acids 158, S62, and N63 of hGH(corresponding to L63, E67, and D68 of hPRL) are important for binding to the hPRLbp(Cunningham and Wells, 1991) while P66 and K69 of hPRL are required for binding (Goffin eta!., 1992). Interestingly, while R64 of hGH, which is equivalent in position to K69 of hPRL, isnot required for binding of hGH to the hPRLbp, it is involved in the interaction between hGHand the hGHbp (Cunningham and Wells, 1989).The contribution of putative helix 4 to receptor binding and the bioactivity of hPRL wasinvestigated using a novel approach. Instead of mutating specific hPRL residues, the activity ofa hPRL molecule containing an additional nine residues at the C-terminus was examined (Goffinet a!., 1993). The conformation of the mutant protein was assessed by circular dichroism (CD)and Fourier-transform infrared spectroscopy: the lack of any major differences in the resultsobtained from the mutant and from wild-type hPRL suggested that the additional nine residuesdid not affect the global folding of the hormone. However, the presence of the extra amino acidsseverely affected the ability of hPRL to bind to the Nb2 lactogenic receptor as indicated by arequirement for a 25- to 30-fold greater concentration of the mutant hPRL than of unmodified35hPRL for 50% displacement of125-labeled unmodified hPRL (the 1C50). In addition, the abilityof the mutant to stimulate the growth of Nb2 cells was less than 3% that of wild-type hPRL.Although the authors did not precisely determine the three-dimensional location for this tail withrespect to the native hormone, modeling studies suggested that the additional nine amino acids atthe C-terminus lay within a small pocket defined by the N-terminal portion of putative helix 1,the C-tenuinal portion of putative helix 4, and the C-terminal half of the large loop betweenputative helices 1 and 2. This location is well away from K69, previously shown to be criticalfor binding of the hormone to the receptor, and the deleterious effect of the tail on hormoneactivity cannot therefore be atthbuted to disruption of this important interaction. The resultssuggest that the region of hPRL which comprises the small pocket described above might formpart of the binding site of hPRL. Site-Specific Mutants of bPRLA structure-function analysis of bPRL was made possible by the development of a systemfor the efficient production of met-bPRL, that is, the mature 199 amino acid hormone with amethionine residue at the N-terminus, in transformed E. coli cells (Luck et al., 1986). Thesequence encoding the signal peptide was removed from the 982-bp bPRL eDNA of clonepBPRL 72 (Sasavage eta!., 1982) (Figure 6) as was most of the sequence 5’ to the ATGinitiation codon. The resulting eDNA, consisting of the entire coding sequence of mature bPRLplus six nucleotides 5’ to the initiation codon, was cloned as a 686 bp insert into the EcoRI site ofa modified pEMBL 8(+) vector [designated A’3 pEMBL 8(+)]. This vector contained theconsensus Shine-Delgarno sequence AGGA overlapping with a unique EcoRI recognition site(Luck et a!., 1986). Very little met-bPRL was produced following transformation of E. coli withthis plasmid. In an effort to determine the reason for the lack of met-bPRL production the bPRLcoding sequence was fused to the N-terminus of the wild-type E. coli lacZ sequence. E. coli cellscontaining this plasmid synthesized large amounts of a lacZ-bPRL fusion protein, suggesting thatthe low level of production of met-bPRL from the A13 pEMBL 8(÷) was most likely aconsequence of the sequence 5’ to the initiation codon of lacZ or of the sequence immediatelydownstream. Therefore, the sequence around the initiation codon for met-bPRL was replaced by36—30Met AspUGCUUGGCUGAGGAGCCAUAGGACGAGAGCUUCCUGGUGAAGUGUGUUUCUUGAAAUCAUCACCACC AUG GAC—20 —10Ser Lys Giy Ser Ser Gin Lys Giy Ser Arg Leu Leu Leu Leu Leu Val Vai Ser AsnAGC AAA GGU UCG UCG CAG AAA GGG UCC CGC CUG CUC CUG CUG CUG GUG GUG UCA AAU4-i 10Leu Leu Leu Cys Gin Giy Val Vai Ser Thr Pro Vai Cys Pro Asn Giy Pro Gly AsnCUA CUC tJUG UGC CAG GGU GUG GUC UCC ACC CCC GUC UGU CCC AAU GGG CCU GGC AAC20Cys Gin Val Ser Leu Arg Asp Leli Phe Asp Arg Ala Val Met Val Ser His rr lieUGC CAG GUA UCC CUU CGA GAC CUG tJUU GAC CGG GCA GUC AUG GUG UCC CAC UAC AUC30 40His Asp Leu Ser Ser Glu Met Phe Asn Giu Phe Asp Lys Arg Tyr Ala Gin Gly LysCAU GAC CUC UCC UCG GAA AUG UUC AAC GAA tJUU GAU AAA CGG UAU GCC CAG GGC AAA50 60Giy Phe lie Thr Met Aia Leu Asn Ser Cys His Thr Ser Ser Leu Pro Thr Pro GiuGGG UUC AUU ACC AUG GCC CUC AAC AGC UCC CAU ACC UCC UCC CUU CCU ACC CCU GAA70 80Asp Lys Giu Gin Aia Gin Gin Thr His His Giu Vai Leu Met Ser Leu lie Leu GiyGAU AAA GAA CAA GCC CAA CAG ACC CAC CAU GAA GtJC CUU AUG AGC UUG AtJU CUU GGG90 100Leu Leu Arg Ser Trp Asn Asp Pro Leu Tyr His Leu Vai Thr Glu Vai Arg Giy MetUUG CUG CGC UCC UGG AAU GAC CCU CUG UAU CAC CUA GUC ACC GAG GUA CGG GGU AUGii0 120Lys Giy Aia Pro Asp Ala lie Leu Ser Arg Aia lie Giu lie Glu Giu Giu Asn LysAAA GGA GCC CCA GAU GCU AUC CUA UCG AGG GCC AUA GAG AUU GAG GAA GAA AAC AAA130 140Arg Leu Leu Giu Giy Met Glu Met lie Phe Giy Gin Vai lie Pro Giy Ala Lys GiuCGA CUU CUG GAA GGC AUG GAG AUG AUA UUU GGC CAG GUA AUU CCU GGA GCC AAA GAGiSO i60Thr Glu Pro Tyr Pro Vai Trp Ser Gly Leu Pro Ser Leu Gin Thr Lys Asp Giu AspACU GAG CCC UAC CCU GUG UGG UCA GGA CUC CCG UCC CUG CAA ACU AAG GAU GAA GAU170 i80Ala Arg Tyr Ser Ala Phe Tyr Asn Leu Leu His Cys Leu Arg Arg Asp Ser Ser LysGCA CGU UAU UCU GCU UUU UAU AAC CUG CUC CAC UGC CUG CGC AGG GAU UCA AGC AAG190lie Asp Thr Tyr Leu Lys Leu Leu Asn Cys Arg lie lie Tyr Asn Asn Asn Cys OCAUU GAC ACU UAC CUU AAG CUC CUG AAU UGC AGA AUC AUC UAC AAC AAC AAC AGC UAAGCCCACAUUCCAUCCUAUCCAUUUCUGAGAUGGUUCUUAAUGAUCCAUUCCCUGGCAAACUUCUCUGAGCUtJUAUAGCUUUGUAAUGCAUGCUUGGCUCUAAUGGGUtJUCAUCUUAAAUAAAAACAGACUCUGUAGCGAUGUCAAAAUCUFigure 6. Nucleotide sequence of the mRNA coding for bPRL and its amino acid sequence.The niRNA sequence was deduced from pBPRL72 (Sasavage et at., 1982) with changes in thethird positions of codons 66,76, and 137 as described (Luck et at., 1986). The coding portion ofbovine preprolactin (690 bp) starts from the methionine AUG codon labeled -30 and terminateswith the ochre UAA codon, labeled OC. The beginning of the processed bPRL protein is thethreonine at position +1. The polyA sequence, not indicated, is at the extreme 3’ end of thesequence.37one containing nucleotides from the consensus sequence around the initiation codon of E. coligenes. Transformation of E. coli cells with the resulting plasmid, pESP4, resulted in productionof met-bPRL to approximately 5% of total cell protein.Although the E. coli expression system described above resulted in efficient production ofmet-bPRL, the protein was produced in the form of insoluble inclusion bodies and this led todifficulties in isolating soluble, active protein. Therefore, as part of the work for this thesis, anattempt was made to avoid the problems associated with inclusion bodies by developingalternative expression systems for secreted bPRL, in yeast and in E. coti. Unfortunately, thesedid not result in the production of soluble active bPRL. However, several changes were made tothe E. coli intracellular expression system that improved the production of met-bPRL.Initial studies by Luck et a!. (1989) to determine which bPRL residues interact with thelactogenic receptor, using the above E. coli expression system and site-directed mutagenesis,focussed on amino acids which are conserved in lactogens but are not found in the related butnon-lactogenic bGH (Luck et al., 1989). Seven bPRL residues were identified by thiscomparison: four of these residues are within regions of the amino acid sequence previouslyidentified as lactogen-specific by sequence comparisons (Nicoll et at., 1986). The seven residueswere mutated to the corresponding bGH residues and the ability of the recombinant proteins tostimulate the lactogen-dependent growth of the Nb2 cells was determined. None of the singlesubstitution mutations had a significant effect on the bioactivity of bPRL. However, a doublemutant, S62T/T65A, had a bioactivity of 45% that of wild-type bPRL and the deletion of Y28(located in the putative helix 1) abolished the bioactivity. The importance of the putative helicalregions to the interaction of bPRL with the lactogenic receptor was investigated further in a studywherein single amino acids were specifically deleted from various regions of bPRL (Luck et at.,1990). Whereas removal of residues from non-helical regions had little effect on the bioactivity,removal of a residue from the centre part of a helical region led to essentially complete loss ofbioactivity. It was suggested that this loss of bioactivity reflected an inability of the proteins toassume a native conformation (Luck eta!., 1990).381.7 Objective of Thesis ProjectThe members of the cytokine-hormone superfamily are involved in regulating manyimportant biological processes, including growth, lactation, and the immune response. As such,there is a great deal of scientific, medical, and commercial interest in these proteins. Forexample, an understanding of the structure-function relationships of each protein with its ownreceptor or other receptors may potentially lead to improved methods of treatment of disorderssuch as acromegaly and hyperprolactinemia and to the design and production of varianthormones with altered binding specificity and improved pharmaceutical properties. Manystudies to date have focussed on hGH, and because of the high homology among members of thegrowth hormone/prolactin/placental lactogen family, many of the results with hGH can beextrapolated to the other hormones. By extension, then, similar in-depth studies of otherhormones will not only shed light on the particular hormones but also give more insights into theother members of the hormone family.The homology between members of the growth hormone/prolactiri/placental lactogenfamily extends beyond sequence and structure to the mechanisms by which each interacts withreceptor proteins. The interaction of growth hormone, particularly hGH, with lactogenic andsomatogenic receptors has been well studied, and the determination of the crystal structures ofhGH complexed to the hGllbp and hPRLbp has provided a great deal of insight into how thisprotein elicits its effects. However, while the general mechanism of hormone-receptor bindingappears to be conserved throughout this family, there are many differences with regard to theresidues which form the functional binding sites. Ideally, an analysis of the crystal structures ofthe hormones and their receptors is required in order to obtain a full understanding of theinteractions between the proteins; however, this information is not yet available for mostmembers of the hematopoietic hormone family. In the absence of such structural information, agreat deal of knowledge as to which residues are involved in eliciting biological responses fromthe receptors can be obtained from molecular biology techniques such as site-directedmutagenesis. In this thesis, site-directed mutagenesis coupled with in vivo analysis using Nb239cell cultures to measure mitogenic activity are used to investigate the structure-functionrelationships of bovine prolactin.The studies which have been carried out on the interactions of bPRL with lactogenicreceptors have confirmed that the mechanism of hormone-receptor interaction appears to besimilar to that of growth hormone, but a complete definition of the prolactin residues whichmediate its mitogenic and bioactivity is lacking. In this study of bPRL the residues involved inthe mitogenic activity of this hormone in the Nb2 cell assay have been investigated in regions ofthe hormone that are homologous with functionally important regions of related proteins. Theinformation obtained from the studies on growth hormone-receptor binding has identified the N-terminus of helix 1, the loop region between helices 1 and 2, a portion of helix 3, and the C-terminus of helix 4 as being involved in receptor binding. The focus of the current investigationwas on bPRL residues within the loop joining putative helices 1 and 2 and within putative helices3 and 4. Amino acids in these areas were selectively mutated and the ability of the variantproteins to stimulate the growth of Nb2 cell cultures was determined.The results obtained in this study highlight both the similarities and the differencesbetween the members of this hormone family. Although, as expected, the general areas of bPRLwhich appear to interact with the lactogenic receptor were found to be homologous to thoseinvolved in receptor binding in related proteins, there were differences with regard to specificresidues. While this result was expected for more distantly related proteins such as bPRL andhGH it was unanticipated for the closely related hormones bPRL and hPRL. Clearly, althoughthe mechanisms for binding and activation of the receptors in this hormone family are verysimilar, the exact details are different in many important ways. Thus, detailed structure-functionanalyses appear to be required for each member of the cytokine hematopoietin family in order toelaborate the precise details of hormone-receptor interactions.40MATERIALS AND METHODS2,1 MaterialsRestriction enzymes and DNA-modification enzymes were purchased from BoehringerMannheim, BRL, New England Biolabs, Pharmacia, Promega, or USB. DNA sequencing kits(Sequenase Version 2.0) were obtained from USB and deoxynucleotides from Pharmacia.Rabbit anti-ovine PRL was obtained from USB and the horseradish peroxidase and alkalinephosphatase complexes of goat anti-rabbit IgG were purchased from Bio-Rad. The ECL kitwas obtained from Amersham. Polyvinylidene difluoride (PVDF) microporous membrane(Immobilon-P) was obtained from Mihipore. Film used for detection of chemiluminescence waspurchased from Amersham (Hyperfilm-EC) and Du Pont (Reflection) while film used forautoradiography was obtained from Kodak (X-Omat) and Agfa (Curix RP-1). AuthenticbPRL (USDA-bPRL-B-1) was a generous gift from Dr. D.J. Bolt of the USDA Animal HormoneProgram. Fischer’s medium was purchased from Gibco and non-lactogenic horse (gelding)serum was from the National Biological Laboratory Ltd. All other reagents were of analyticalgrade and/or for biochemical use.2.2 Strains and Growth Conditions2.2.1 Bacterial StrainsThe Escherichia coil strains used in this study are listed in Table Yeast StrainsThe Saccharomyces cerevisiae strains used for production of bPRL were SR1O69B (MAThhis4-912 leu2-3 ura3-52) and XV700-24C (MATa leu2 ura3 pep4). Total RNA from strain S25(MATa his4-912 ura3-52 leu2-3 leu 2-112) was used as a control for primer extension analysis.2.2.3 Media for Growth of Bacterial and Yeast StrainsThe liquid media used for growth of bacterial strains included YT (0.8% tryptone, 0.5%yeast extract, 0.5% NaC1) supplemented with ampicillin (amp) at 100 g/mL for selection ofampicillin resistant transformants; superbroth (3.2% tryptone, 2% yeast extract, 0.5% NaC1,410.5% (v/v) 1 M NaOH) supplemented with 200 pg amp/mL; terrific broth (1.2% tryptone, 2.4%yeast extract, 0.4% (v/v) glycerol, 17 mMKH2PO4,72 mMK2HPO4)supplemented with 200 pgamp/mL; and M9+B 1 (50 mMNa2HPO4,25 mMK112P04,8.5 mM NaCI, 20 mMNH4C1, 10mM glucose, 0.1 mM CaCI2, 1 mMMgSO4,0.0001% thiamine) supplemented with 100 .tgamp/mL. Transformants were selected by plating on YT + amp plates (made with 1.5% agar andcontaining 100 ig amp/mL).Yeast strains were grown in complete YPD medium (1% yeast extract, 2% peptone, 2%dextrose) and in SD minimal media (0.67% yeast nitrogen base [without amino acids], 2%dextrose) supplemented with amino acids as required.Table 2. List of E. coli StrainsGenotypeAlac-pro ara thi F[traD36 proAB !aclq 1acZbM15]K12 ara &ac-pro thi F[proAB lacIq lacZtM15]K12 ara zlac-pro thi mutL::TnlO F[proAB lacIq 1acZAM15JsupE thi b(lac-proAB) F[traD36 proAB laclq 1acZzM15]supE endA sbcB 15 hsdR4 rpsL (hi A(lac-pro AB)F[traD36 proAB laclq 1acZAM15]Le392F supE44 supF58 hsdR5l4 galK2 ga1122 metBi Borck et aL, 1976trpR55 lacYl F’[traD36 proAB laciq lacZAMl5JMV1 193 A(lac-pro AB) rpsL (hi endA spcB 15 hsdR4 E(srl-recA)306::Tn1O(tef)F[traD36proAB laciq 1acZAM15]supE thi A(lac-pro AB) hsd5 F[traD36 proAB laclq 1acZzM15]siC LJacX74 gal ISII::0P308 rpsLHFrKL16 P0/45 [lysA(61-62)] Zbd-279::TnlOlysA thu relAl spoTl supE44 dutl unglsupE hsdA5 (hi A(lac-proAB) F[traD36 proAB laciq 1acZzM15]rif [F proAB lacIl lacZtiMl5 TnlOQd)]rif [F proAB lacN !acZAMI 5 TnlOQef)] kan’StrainCSH5OHB215 1HB2154JM1O1JM1O5ReferenceKunkel. 1985Carter et a!., 1985Carter et a?., 1985Messing, 1983Yanisch-Perron eta?., 1985NM522RV308RZ1032Zoller and Smith, 1987TG1TOPP1, 2TOPP3Gough and Murray, 1983Schoner et a!., 1985Kunkel, 1985Gibson, 1984StratageneStratagene422.3 PlasmidsThe E. coil expression plasmid pESP4 (Figure 7) used for production of met-bPRL is a 4.7kbp A13 pEMBL 8(-t-) vector containing a modified bPRL eDNA under the inducible control ofthe E. coil lac promoter (Luck et a!., 1986). As discussed in the Introduction, the eDNA wasmodified by removal of both a 5’ nontranslated sequence and the signal sequence of preprolactin.In addition, the sequence around the ATG initiation codon was replaced with one thaf conformedas closely as possible to the consensus sequence in this region of E. coil genes (Luck eta!.,1986). As well as the gene for bPRL, the plasmid pESP4 contains the E. coil on (for replicationin E. coil), a marker for ampicilhin resistance (for selection in E. coil), and the fi intergenicregion (for in vivo production of ssDNA by the filamentous bacteriophage R408). All bPRLvariants used in this study were derived from pESP4 by site-directed mutagenesis.Other plasmids used in attempts to obtain secreted bPRL from either E. coil or S. cerevisiaeare described below in sections 2.12 and 2.13.NaeIFigure 7. Map of pESP4. The 4700 bp vector containing the bPRL coding sequence undercontrol of the iac promoter is derived from pEMBL 8(+) as discussed in the Introduction.NcI432.4 Transformation of CellsFor high efficiency yeast transformation, spheroplasts of S. cerevisiae were transformed asdescribed (Ausubel et a!., 1994).Bacterial strains were transformed by the calcium chloride procedure or by electroporation.AllCaCl2-competent cells, except TOPP2 cells, were prepared and transformed essentially aspreviously described (Ausubel et at., 1994). Cells from a 50 mL culture of mid-log phase E. colicells were harvested by centrifugation at 2,000 x g for 8 mm, gently resuspended in 25 mL ice-cold 50 mM CaC12,and incubated on ice for 30 mm. Cells were again pelleted by centrifugationand gently resuspended in 3 mL ice-cold CaC12. For storage of competent cells 140 1iL aliquotsof theCaCI2-competent cells were mixed with 60 L ice-cold 50% glycerol and frozen at -70°C.For transformation, 10 - 50 ng plasmid DNA was added to 200 iL of theCaCl2-competent cells,the cells were incubated on ice for 30 mm and then heated at 42°C for 2 mm. 6 - 8 1i.L of thesuspension was plated directly on YT plates containing 100 .ig amp/mL (YT + amp plates) forselection of transformants.CaCl2-competent TOPP2 cells, prepared as described above, were transformed asrecommended by Stratagene (protocol supplied with competent cells). B-mercaptoethanol wasadded to 100 tL of competent TOPP2 cells to a final concentration of 25 mM. Cells wereincubated on ice for 10 mm prior to the addition of 10 - 100 ng plasmid DNA. Following anadditional 30 mm incubation on ice cells were heated at 42°C for 45 sec, incubated on ice for 2mm, and then diluted with 0.9 mL SOC medium (2% tryptone, 0.5% yeast extract, 10 mM NaC1,2.5 mM KC1, 10 mM MgC12,10 mMMgSO4,20 mM glucose) and incubated at 37°C for 1 hwith shaking. Transformants were selected by plating on YT + amp plates and incubatingovernight at 37°C.Electroporation was performed as recommended by Bio-Rad (protocol outlined in the PulseController Instruction Manual). E. coli cells at early to mid-log (OD600 0.5 - 0.8) were harvestedby centrifugation at 4,000 x g for 15 mm, resuspended in 1 vol ice-cold dH2O, centrifuged,resuspended in 0.5 vol ice-cold dH2O, centrifuged, and then resuspended in 0.02 vol ice-cold10% glycerol and centrifuged. The pellet was resuspended to 0.002 - 0.003 vol in 10% glycerol.44For transformation, 1 -2 pL of plasniid DNA in TE buffer (10 mM Tris-HC1, 1 mM EDTA, pH8.0) was added to the electro-competent cells, the mixture incubated on ice for 1 mm and thenpulsed at 2.5 kV, 25 1iF, and 200 2 using 0.2 cm cuvettes in a Gene Pulser® TransfectionApparatus (Bio-Rad). A 1 mL volume of SOC medium was added to the mixture and aliquotswere plated on YT + amp plates for selection of transformants.2.5 Isolation of Plasmid DNAPlasmid DNA was isolated from bacteria by a modified version of the alkaline lysisprocedure (Ausubel eta!., 1994; Peiham, 1985). For small scale plasmid production 1.5 mLcultures were grown overnight in YT + amp, the cells harvested by centrifugation in amicrocentrifuge at maximum speed (12,700 x g), and the cell pellets resuspended in 100 L TB(10 mM Tris-HC1, 1 mM EDTA, pH 8.0). After incubation at room temperature for 5 miii, 200iiL of 1% SDS, 0.2 M NaOH were added and the solution was incubated on ice for 5 mm. Then150 JIL of 5 M potassium acetate, pH 4.8 was added and the mixture incubated on ice for anadditional 5 miii. The DNA in the supernatant was precipitated by the addition of 900 LLL of 95%ethanol, collected by centrifugation for 2 mm in a microcentrifuge, then dissolved in 100 jtL TBplus 100 jiL 5 M LiC1 and incubated on ice for 5 mm. Following centrifugation for 5 miii in amicrocentrifuge the supernatant was decanted and the DNA precipitated with 400 L of 95%ethanol. The DNA was collected by centrifugation in a microcentrifuge for 2 miii, washed with 1mL of 70% ethanol, centrifuged, and dried. The pellets were dissolved in 50 iL TB.Larger amounts of plasmid DNA (from 50 mL cultures) were prepared by a slightlymodified version of another procedure (Sambrook et a!., 1989). Cells were harvested bycentrifugation (12,000 x g for 10 miii), the pellets resuspended in 20 mL ice-cold 50 mM TrisHC1, 5 mM EDTA, 50 mM NaC1, pH 8.0 and centrifuged (12,000 x g for 10 miii). The cells thenwere resuspended in 3 mL 50 mM glucose, 25 mM Tris-HC1, 10 mM EDTA, pH 8.0 containing5 mg lysozyme/mL and incubated at room temperature for 10 mm. To this was added 6 mL of1% SDS, 0.2 M NaOH and the suspension was incubated on ice for 5 mm. Then 4.5 mL of 5 Mpotassium acetate, pH 4.8 was added and the mixture incubated for an additional 5 miii on ice,45The suspension was centrifuged (17,500 x g for 30 mm) and the DNA precipitated by adding 8mL isopropanol to the supernatant and incubating for 30 mm at room temperature. The DNAwas pelleted by centrifugation (8,000 x g for 30 mm) in a warm rotor and resuspended in 1,200tL TE plus 1200 1iL 5 M LiC1. The suspensions were incubated on ice for 5 mm and thencentrifuged in a microcentrifuge for 5 mm. The DNA was precipitated by the addition of 2 vol95% ethanol and collected by centrifugation for 2 mm in a microcentrifuge. The DNA pellet wasdissolved in 162 1iL TE, cleaned by the addition of 18 tL RNase A (20 ig/mL, 30 mm at 37°C)followed by 20 jiL proteinase K (50 p.g/mL, 30 mm at 37°C), extracted once with phenolICHCl3(1:1), and precipitated by the addition of 0.5 vol 7.5 M ammonium acetate and 2 vol 95%ethanol. The DNA was collected by centrifugation for 5 mm in a microcentrifuge, washed with 1mL of 70% ethanol, and dissolved in TE buffer.2.6 Isolation of Single-Stranded DNASingle-stranded DNA (ssDNA) templates from plasmids containing the fi on wereprepared for DNA sequencing (from strain HB2151 containing the F factor genetic element (F’)which allows the cells to be infected by filamentous phages such as R408) and site-directedmutagenesis (from the dur ung F’ strain RZ1032 for uracil-containing ssDNA) as described(Dente et al., 1983). Briefly, YT + amp medium was inoculated with cells from a single colonyof freshly transformed E. coli and the culture was incubated at 37°C with shaking forapproximately 1 h (ODcyy0.2) prior to the addition of R408 helper phage at a multiplicity ofinfection of 20:1. Following infection, the R408 phage packages the fi on-containing plasmidDNA and produces ssDNA. The superinfected cells were incubated at 37°C with shaking for anadditional 5 - 6 h. Phage particles were precipitated from the supernatant with 0.25 vol 20%PEG 8,000, 2.5 M NaC1 at room temperature for 15 mm and collected by centrifugation in amicrocentrifuge (12,700 x g for 5 mm). The phage pellets were resuspended in TE, extractedwith phenol/CHC13(1:1), and the ssDNA precipitated by the addition of 0.1 vol 3 M sodiumacetate and 2 vol 95% ethanol. The ssDNA was collected by centrifugation for 5 mm in amicrocentrifuge, washed with 1 mL 70% ethanol, dried, and dissolved in TE buffer.462.7 Site-Directed MutagenesisSite-directed mutagenesis (Zoller and Smith, 1982) was performed essentially as described(Kunkel et a!., 1987). Phosphorylated oligonucleotides synthesized using an AppliedBiosystems 380B or 380A DNA synthesizer were purified either on C18 Sep Pak cartridges(Millipore) (Atkinson and Smith, 1984) or by extraction with 30%NH4O and butanol(Sawadogo and Van Dyke, 1991). The mutations created by oligonucleotide-directedmutagenesis and the oligonucleotides used are listed in Table 3.The procedure used for site-directed mutagenesis is as follows. One pmol of theoligonucleotide was mixed with 0.1 - 0.5 pmol of the uracil-containing ssDNA template (UssDNA, prepared from the dur ung E. coil strain RZ1032) in 10 1tL of core buffer (500 mMNaC1, 100 mM Tris-HC1, 100 mMMgCI2,10 mM DTT, pH 8.0). Annealing was promoted byheating for 5 mm at 55°C followed by a 10 mm incubation at room temperature. Theextension/ligation step was catalyzed by 1 U DNA polymerase I (Kienow fragment) and 1 U T4DNA ligase in ligase buffer (660 mM Tris-HC1, 50 mM MgC12,50 mM DTr, pH 7.5) withadded ATP (to 0.5 mM) and dNTPS (to 0.5 mM). The final reaction volume was 20 j.iL. Thereaction was incubated for 2 h at room temperature. CaC12-competent E. coil HB2151 cellswere transformed with 5 iL aliquots of the mutagenesis mixture while 0.5 1L aliquots were usedto transform electro-competent HB2151 cells.2.8 DNA SequencingDNA sequence analysis was performed on ssDNA templates (isolated from the E. coiistrain HB2151 as described above) using the Sanger dideoxy chain termination method (Sangeret a!., 1977) with modified Ti DNA polymerase (Sequenase®) and [cx-35S]thio-dATP accordingto the procedure recommended and supplied by the manufacturer (USB). dITP nucleotides wereused to avoid compression regions within the bPRL gene. The sequence was visualized byautoradiographic detection on Curix RP-1 film (Agfa).47Table 3. Ollgonucleotides Used for Site-Directed MutagenesisOligo Mutagenic Primera Mutation(s)MH4 CCATITGGACAAACTGGTGTFGCAGATATITfGGCTGCA Attach invertase signalAAACCAGCCAAAAGGAAAAGGAAAGC1TGCAAAAGC sequence to 5 end ofATAATAGAXITCCTG113TG bPRL geneMH6 CAGGTCCAGGACAAACTGGTGTGGCCGCGCTA Attach bPRL gene in-frameCGGTAGCGAAACCA to ompA signal sequenceMH13 CCATGACTGCCGCGTCAAACAG R21AMH14 GCTGAATCCGCGCGCAGGCAG R177AMH15 ATFCAGGAGCGCAAGGTAAGTG K187AMH26 AGGGAGGAGGTAGCGCAGCTGTGAG H59AMH28 AAGGGAGGAGGTGGAGCAGCTGTGAG H59SMH25 TAGGAAGGGAGGA/CGGT/cATGGCAGCTGTrb S61A and T6OAMH37 CAGGGGTAGGAGCGGAGGAGGTATG L63AMH46 TCAGOGGTAGCAAGGGAGGAG P64AMH38 GTrCTrrATCTT/GCAGG/cGGTAGGAAGGci P66A and E67AMH39 GGGCTGTCTGCATCTCAGGGG K69AMH49 CTGTTGGGCTGCTCTATC Q71AMH52 CGTGGGCTACTCT1TATC Q71VMH48 GTGGGTCTGTGCGGCITGTC Q73AMH48 GTGGGTCTGTACGGCTGTC Q73VMH47 ATGGTGGGTCGCTfGOOCTG Q74AMH5O ATGGTGGGTCACTGGGCTfG Q74VMH4 1 TCTCCATGCCTT/GCCAGAAGTC/GG/cmGrITICT R125A and E128AMH42 CITCCAGAAGTGC’ITTGTnICTrCC R12SAMH36 CATCTCCATGCGTTCCAGAAGTC G129RMH29 GTGGAGCAGGGCATAAAAAGCAGA N17OAMH43 CAGTGGAGCGCGTATAAAAAGC L171AMH24 CTGCGCAGGCAGGCGAGCAGGTrATA H173AMH23 GAATCCCTGCGCGCGCAGTGGAGCAG L175AMH22 CTGAATCCCTGGCCAGOCAGTGGAG R176AMH27 AATCTGCGAGGACCTGCGCAX3CA D178SMH32 GTCAATCTGCTrGA/CA/.IIXDCcTGCGCAGOC S 179A and D178EMH21 AGTGTCAATCTrGC/GT/CTGAAT/GCCCTGCGCAGGCA S 180A and D178AMH2O TAAGTGTCAATCGCGCTGAATCCCTG K181AMH33 CAGGAGCTrAAGGT/GA/CAGTGT/GCAATCTrGcTrGA D183A and Y185AMH44 TAAGGTAAGCGTCAATCTG T184AMH45 GCAATCAGGGCCYAAGGTAAG L188AMH3O TCTGCAATCGCGAGCTAAGGT L189AMH3 1 GTfGTAGATGGCTCTGCAATCAG 1193Aa The sequences of the mutagenic primers are presented from the 5’ to the 3’ end. Each primer was phosphorylatedat the 5’ end during synthesis.b Mixed oligonucleotides are shown with the two possible bases at each mixed position divided by a “1”.482.9 Preparation and Anaiysis of RNA2.9.1 Isolation of Yeast RNAWhere appropriate, solutions were treated with diethylpyrocarbonate (DEPC) as follows.To inhibit RNase activity, 0.2 mL of DEPC was added per 100 mL of the solution to be treatedand the mixture shaken vigorously to get the DEPC into solution, stirred for approximately 1 h atroom temperature, and then autoclaved. In order to collect yeast cells for extraction of RNA, 50mL of S. cerevisiae culture grown in SD selective medium to an 0D600 of 0.5 - 1 were filteredthrough a 0.45 urn nitrocellulose filter. The harvested cells were then washed with 500 p.L ofbreakage buffer (0.2 M Tris-HC1, 0.5 M NaC1, 10 mM EDTA, pH 7.5) and resuspended in 200pL of the same buffer. Acid-washed glass beads and 200 tL of 4:4:1:1 phenol/CHC13/10%SDS/3 M potassium acetate, pH 5.5 were added and the suspension was vortexed vigorously for5 mm. The organic phase was removed and discarded and the aqueous phase was extractedtwice more as described above. The aqueous phase was then extracted once with CHC13. TheRNA in the aqueous phase was precipitated with 0.4 vol 5 M ammonium acetate, pH 7.4 and 2.5vol 95% ethanol, washed with 70% ethanol, dried, and dissolved in 40 JIL dH2Oby heating at85°C for 2 mm. The amount of RNA was estimated by theA260 (using the formula C (jig/mL) =A260/0.025) and adjusted to 10 tg/mL for use in primer extension analysis.2.9.2 Isolation of E. coli RNATotal E. coli RNA was isolated according to published methods (Ausubel et al., 1994).Where appropriate, solutions were treated with DEPC as described above and autoclaved in orderto inhibit RNase activity. 100 mL cultures of TOPP2 transformed with vectors expressingvarious bPRL mutants were grown to log phase in terrific broth containing 200 ig amp/niL and 3mM IPTG. Cell growth was stopped by the addition of 5 mL of stop buffer (200 mM Tris-HCI,20 mM EDTA, 20 mM sodium azide, pH 8.0) and the cells were harvested by centrifugation at5,500 x g for 5 mm. The pelleted cells were resuspended in 2 mL lysing solution (8% w/vsucrose, 5% v/v Triton X-100, 50 mlvi EDTA, 50 mlvi Tris-HC1, pH 7.0) plus 100 i.tL 200 mlvivanadyl ribonucleoside complex (VRC), extracted with 1 mL phenol, and then extracted with 1mL CHC13. Nucleic acids were precipitated from the aqueous layer with 0.1 vol 3 M sodium49acetate and 2 vol ice-cold 100% ethanol, collected by centrifugation at 10,000 x g for 10 mm,and resuspended in 2 mL 10 mM VRC. The nucleic acid in the VRC solution was extractedtwice more with 1:1 phenolICHCl3and precipitated with sodium acetate and ethanol as describedabove. The nucleic acid pellet was then resuspended in 2 mL DEPC-treated dH2O, 1 g CsC1 wasadded, and 2.25 mL of the solution was layered onto a 0.75 mL CsCl cushion (5.7 M CsCl in 100mM EDTA, pH 7.0) in a 13 x 51 mm TLA-l00.3 polycarbonate tube. The RNA was pelleted bycentrifugation for 1 h at 80,000 rpm (280,000 x g) at 20°C in a TLA-l00.3 rotor. The RNA wasresuspended in 0.36 mL DEPC-treated dH2O, 36 jiL 3 M sodium acetate and 2.5 vol ice-cold100% were added, and the RNA was precipitated by a 20 mm incubation at -70°C followed bycentrifugation in a microcentrifuge for 5 mm at 4°C. The pellet was washed with 1 mL of ice-cold 70% ethanol, centrifuged in a microcentrifuge for 5 mm at 4°C, dried, dissolved in 200 p.LDEPC-treated dH2Oand quantified by theA260. The concentration was adjusted to 4 .tg/mL inDEPC-treated dH2Oand the RNA solution stored at -70°C.2.9.3 Primer Extension Analysis of Yeast RNAOne pmol of oligonucleotide primer was incubated in 20 ji.L PNK buffer (100 mM TrisHC1, 10 mM MgCl2,10 mM DTT, pH 8.0) containing 9 U T4 polynucleotide kinase and 3 pmol.32pATP (10 pCL4tL) for 30 mm at 37°C. The reaction was stopped by incubation at 65°C for 5mm. The labelled primer was annealed to the RNA by incubating 1 jiL of the kinase-labelledprimer with 5 pL of yeast RNA (10 j.tg/iL) and 1 iL buffer A (500 mM Tris-HC1, 1 M KC1, 100mMMgCl2,pH 8.5) at 85°C for 2 mm. The reaction was then incubated on ice for 10 mm. Theextension reaction was carried out by adding 1 i1 AMV reverse transcriptase (diluted to 1 U4Lin 50 mM Tris-HC1, 100 mM KC1, 10 mM DTT, 10 mM MgC12,pH 8.5) and 3 L of 500 pMdNTPs, 10 mM DTT and incubating at 42°C for 1 h. The reaction was stopped by the addition of10 tL formamide dye mix (80% formamide, 10 mM EDTA pH 8.0, 1 mg xylene cyanol FF/mL,1 mg bromophenol blue/mL) and 1 jiL 0.2 M NaOH. The final mixture was boiled for 5 mm andanalyzed by denaturing polyacrylamide gel electrophoresis. The gel was soaked in a solution of10% glacial acetic acid, 12% methanol, dried, and exposed to film for visualization of the bands.502.9.4 Competitive PCR Analysis of E. coli RNADetection and quantitation of mRNA species from selected recombinant bPRL variants wasperformed by the sequential use of reverse transcriptase and the polymerase chain reaction (PCR)(Foley et a!., 1993; Gilliland eta!., 1990; Perrin and Gilliland, 1990). Briefly, RNA isolatedfrom E. coli cells expressing recombinant bPRL mutants was reverse transcribed with AMVreverse transcriptase. The resulting cDNA was then amplified by PCR in a reaction mixturecontaining a known amount of DNA from a bPRL variant containing the signal sequence for theyeast invertase gene (pMH4).For the preparation of the cDNA, 100 ng RNA (from a 4 g RNA/mL solution prepared asdescribed above) was mixed with 20 pmol of bPRL primer, heated at 65°C for 10 mm, and thenreverse transcribed in a reaction mixture containing 20 U RNasin® ribonuclease A inhibitor(Promega), 25 U AMV reverse transcriptase, and 10 nmol each dNTP in 20 1iL total volume ofreverse transcriptase buffer (50 mM Tris-.HC1, 8 mM MgC12,30 mM KC1, 1 mM DTT, pH 8.5)for 1 hour at 42°C. The reaction mixture was added to 200 iL of a PCR master mixturecontaining 3 mM MgC12,44 pmol each primer, 200 LtM dNTPs, and 3.75 U Taq DNApolymerase in PCR buffer (20 mM Tris-HCI, 50 mM KCI, pH 8.4). The master mixconcentrations were adjusted to account for input primer and dNTPs from the reversetranscription reaction. 20 j.tL aliquots of the PCR mixture were added to 10 tubes eachcontaining 5 j.iL of a dilution series of pMH4 plasmid DNA in PCR buffer. The 25 .tL reactionvolumes were overlayed with oil, heated at 94°C for 2 mm, and 25 cycles of PCR wereperformed (94°C for 25 sec. 52°C for 30 sec, and 72°C for 1 mm) followed by aS mm extensionat 72°C. An aliquot of each reaction mixture was subjected to electrophoresis in a 1.9%MetaPhor agarose gel (FMC BioProducts) prepared in TBE buffer (50 mM Tris, 50 mM boricacid, 1 mM EDTA). For all PCR reactions the controls included (i) reverse transcription reactionmixture from E. coli cells lacking the bPRL expression plasmid; (ii) no added cDNA; (iii) pMH4and pESP4 plasmid DNA; (iv) cDNA not mixed with competitor pMH4; and (v) reversetranscription reaction mixture from which AMV reverse transcriptase was omitted.512.10 Isolation ofMet-bPRLThe total concentration of protein in extracts from various steps in the isolation procedurefor bPRL was determined using the Lowry protein assay (Lowry eta!., 1951).2.10.1 Extraction of Met-bPRL and Met-bPRL VariantsTransformed E. coil cells (generally strains HB2151 or TOPP2) from which met-bPRL ormet-bPRL variants were to be isolated were grown, lysed, and extracted essentially as previouslyreported (Luck et at., 1989) with minor modifications detailed below. Overnight cultures oftransformed cells grown in 1.5 mL terrific broth + 200 ig amp/mL were diluted 1:100 into 100mL of terrific broth containing 200 ig amp/mL and 3 mM IPTG. Cultures were incubated at37°C with shaking (275 rpm) for 16 h and the cells were harvested by centrifugation (8,000 x gfor 10 mm). The cells were resuspended in 20 mL cold 0.1 M sodium phosphate buffer, pH 7.4,centrifuged (8,000 x g for 10 mm), and resuspended in 20 mL 0.1 M sodium phosphate buffer,pH 7.4, containing 40 mM Dl]’ and 1 mM phenylmethylsulfonyifluoride (PMSF). The cells inthe suspension were lysed by a single passage through a French pressure cell at maximumpressure (20,000 psi). The insoluble material was collected by centrifugation at 20,000 x g for30 mm and resuspended using a Dounce homogenizer in 8 mL cold 0.1 M sodium phosphatebuffer, pH 7.4, containing 5 mM Dl]’. The suspension was divided into 1 mL aliquots whichcould be stored indefinitely at -70°C. To remove some contaminating E. coil proteins 32 j.iL of10% sodium desoxycholate were added to 800 itL of the suspension and the mixture wasincubated at 37°C for 1 h. The suspension was centrifuged in a microcentrifuge at maximumspeed for 10 mm. The pellets were washed by the addition of 1 mL of dH2Ofollowed byvigorous vortexing and the insoluble material was collected by centrifugation in amicrocentrifuge for 10 mm. Using a 22 gauge needle attached to a 1 mL syringe each pellet wasresuspended in 500 L 0.1 M sodium phosphate buffer, pH 7.4, containing 0.1% N-laurylsarcosine (sarcosyl). Care was taken to avoid excessive aeration of the solutions. Thesuspensions were incubated overnight at 4°C in order to extract the met-bPRL. The insolublematerial was then removed by 10 mm centrifugation in a microcentrifuge and the supernatants,containing the solubiized met-bPRL, stored at 4°C.522.10.2 Renaturation of Extracted Met-bPRL and Met-bPRL VariantsRenaturation of met-bPRL and its variants was carried out essentially as previouslydescribed (Luck and Huyer et at., 1992). The extracts of met-bPRL and its variants in 0.1%sarcosyl (475 1iL volumes, prepared as described above) were diluted with 950 iL 0.1 M sodiumborate, pH 10, containing 0.1% sarcosyl. In order to promote renaturation by air oxidation thesamples were incubated at room temperature for 4 h in Eppendorf microcentrifuge tubes with thelids open. The samples were then neutralized by addition of 42 jiL 6 N HC1. The renatured metbPRL extracts were stored at 4°C prior to quantification by Western blot analysis. Samplescould be stored in this state for approximately 2 weeks without loss of bioactivity. Followingquantitation, samples for bioassay were diluted to 1 ig hormone/mL in Fische?s mediumcontaining 10% gelding serum and 15 mM HEPES (N-[2-hydroxyethyl]piperazine-N’-[2-ethane-sulfonic acid]) buffer and stored at -20°C. Once diluted into the Fische?s medium the met-bPRLsamples could be stored for long periods (e.g., 1 year) at -20°C; they retained bioactivity throughrepeated freeze-thaw cycles.2.10.3 Quantification of Met-bPRL and Met-bPRL VariantsThe concentration of renatured met-bPRL and its variants in the 0.1% sarcosyl extracts ofE. coil cells (obtained as described above) was determined using Western blot analysisessentially as previously described (Luck and Huyer et a!., 1992). Aliquots of the extracts and ofstandard pituitary bPRL (USDA-bPRL-B- 1) (diluted in 0.1 M sodium phosphate buffer, pH 7.4,containing 0.1% sarcosyl) were subjected to discontinuous SDS-PAGE (Laemmli, 1970) andanalyzed by Western blotting. Samples were diluted in 2X sample buffer (0.2 M Tris-HC1, pH6.8, 5% SDS, 30% glycerol) without B-mercaptoethanol (to provide non-reducing conditions)and loaded onto a discontinuous polyacrylamide gel (4% stacking gel over a 13% separating gel).Electrophoresis was carried out at room temperature using currents of 14 mA through thestacking gel and 28 mA through the separating gel. Following electrophoresis the separatedproteins were transferred to a polyvinylidene difluoride (PVDF) microporous membrane at 100mA for 2- 2.5 h at 4°C using a Tris/glycine/methanol transfer buffer (20 mM Tris, 150 mMglycine, 20% methanol). Blots were blocked in blotto (5% Carnation® instant skim milk powder53in TBS-Tween [20 mM Tris, 137 mM NaCI, 0.1% Tween-20, pH 7.6]) for 1 h to overnight. ThebPRL bands were visualized using enhanced chemiluminescence. The Western blot protocolfollowed was that recommended and provided for use with the Amersham ECLThI system. Theprimary antibody used was polyclonal rabbit anti-ovine PRL (USB) (diluted 1:10,000 in TBSTween), and the secondary antibody was the horseradish peroxidase conjugate of goat anti-rabbitIgG (Bio-Rad) (diluted 1:5,000 - 1:10,000 in TBS-Tween). Following exposure of the treatedmembrane to the detection solution the bands were visualized on Hyperfilm-EC (Amersham) oron Reflection autoradiography film (Du Pont). The densities of the bands of the non-reducedpituitary bPRL standards and of the corresponding bands of the met-bPRL variants weremeasured using a computerized laser scanning densitometer (Model 300A, MolecularDynamics). Alternatively, the light emitting from the bPRL bands was detected using a highperformance luminescence imaging system (Luminograph LB 980, EG&G Berthold). Theamount of met-bPRL or met-bPRL variant in each extract was estimated by reference to astandard curve constructed from the densities of the bands, or the amount of light emitted, of thevarious concentrations of the pituitary bPRL standard.Western blot analysis was also performed using a colourimetric system with alkalinephosphatase-linked secondary antibody (Harlow and Lane, 1988). Following electrophoresis theseparated bands were transferred to nitrocellulose using the conditions described above. Thenitrocellulose blot was blocked in incubation buffer (10 mM Tris-HC1, 150 mM NaC1, pH 7.5)containing 5% skim milk powder, incubated with polyclonal rabbit anti-oPRL (diluted 1:2,000with incubation buffer with 2.5% skim milk powder), and then with the alkaline phosphataseconjugate of goat anti-rabbit IgG (Bio-Rad) (diluted 1:1,500 with incubation buffer with 2.5%skim milk powder). The blot was washed with incubation buffer containing 0.25% sarcosyl, Forvisualization of the bands the blot was immersed in 30 mL 0.1 M NaHCO3,1 mM MgC12,pH9.8, 9 mg nitro blue tetrazolium (NET, made as a 50 mg/mL solution in 70%dimethylformamide) and 4.5 mg bromochloroindolyl phosphate (BCIP, made as a 25 mg/mLsolution in dimethylformamide) was added, and the blot was incubated at room temperature untilthe bands were visible (approximately 10 mm).542.10.4 In Vitro Bioassay ofMet-bPRL and Met-bPRL VariantsExtracts of met-bPRL and its variants, diluted to 1 j.tg/mL in Fischer’s medium containing10% gelding serum and 15 mM HEPES buffer, were assayed for mitogenic activity using theNb2 lymphoma cell proliferation assay (Tanaka et al., 1980). The ability of the met-bPRLvariants to stimulate the growth of lactogen-dependent Nb2 lymphoma cell cultures wascompared with that of the pituitary bPRL standard preparation. The met-bPRL preparations wereassayed in a series of dilutions to cover the growth response of the cells in a 48 h period tolactogens in the useful working range of 0.01 - 0.25 ng/mL. Cells were counted with a Coulterelectronic cell counter. Mitogenic activities of the met-bPRL variants are expressed aspercentages of the bioactivity of the pituitary bPRL standard.2.11 Analysis of Production ofMet-bPRL at Different TemperaturesE. coli strains HB2151, Le392F’, and TOPP2 transformed with pESP4 were incubated withshaking inS mL of YT, superbroth, and terrific broth medium containing 100 jig amp/mL (forYT) or 200 jig amp/nth (for superbroth and terrific broth) and 5 mM IPTG at room temperature(23°C), 30°C, and 37°C for 15 - 17.5 h. Cells were harvested by centrifugation (12,000 x g for10 mm), the cell pellets were resuspended in 20 mM Tris-HC1 buffer, pH 8.0 containing 20 mMEDTA and 2 mg lysozyme/mL, and the cells lysed by four successive rounds of freezing inliquid nitrogen followed by thawing in a room temperature waterbath. After lysis the DNA wassheared by sonication using the microtip probe for two 10 sec bursts, cooling on ice betweeneach burst. The suspensions were centrifuged in a microcentrifuge for 10 mm and the solubleand insoluble fractions loaded on polyacrylamide gels for analysis by coomassie blue-stainingand Western blotting. V2.12 E. coli Secretion System2.12.1 PlasmidsThe vectors pompPRL and pompPRLF are derived from the E. coli secretion vector pINIllOmpA3 (Ghrayeb et at, 1984). They were constructed as part of the system for secretion of55bPRL from E. coil. All DNA manipulations were carried out according to published protocols(Ausubel et a!., 1994).pompPRLF: The phage fi on was subcloned into pINIII-OmpA3 in order that single-stranded vector DNA would be obtained following infection of E. coil with the filamentoushelper phage R408. The bPRL gene was subcloned as an EcoRI-BamHI fragment into pINIllOmpA3,creating the vector pINPR. The phage fl intergenic region (f1 on) was subcloned fromthe yeast/E. coil shuttle vector pVT100U (Vernet et a!., 1987) into pINPR as follows. Theplasmid pVT100U was digested with EcoRI, the ends of the linearized vector filled in by DNApolymerase I (Kienow fragment) and the 1.2 kbp fragment containing the fi on isolatedfollowing digestion with NdeI. The fi on-containing fragment was subcloned into pINPR thathad been digested with AvaI, blunt-ended with DNA polymerase I (Kienow fragment) and thendigested with NdeI. The vector pompPRLF was created by looping out the DNA sequencebetween the 3’ end of the ompA signal sequence and the 5’ end of the bPRL gene.pompPRL: The fi on was removed from pompPRLF as an EcoRl-Bgill fragment, theends filled in with DNA polymerase I (Kienow fragment) and the blunt ends ligated togetherwith T4 DNA ligase.The vectors pJL3 and pSW1 were obtained from Drs. Oudega and Luirink (VrijeUniversiteit, Amsterdam, The Netherlands). pJL3 codes for chloramphenicol resistance andcontains the bacteriocin release protein (BRP) gene under the control of the lpp-iac promoteroperator system. pSW1 codes for tetracycline resistance and contains the BRP gene under thecontrol of the mitomycin C inducible pCloDFl3 promoter.2.12.2 Secretion SystemThe BRP expression plasmids pJL3 and pSW1 were co-transformed with the vectorpompPRLF into E. coil JM1O1. The double transformants of pJL3 and pompPRLF were selectedon YT plates containing chlorarnphenicol (20 ig/mL) and ampicillin (100 ig/mL) and the doubletransformants of pSW1 and pompPRLF were selected on YT plates containing tetracycline (12i.tg/mL) and ampicillin (100 ig/mL). The isolates were analyzed for bPRL release followinginduction of BRP and bPRL as follows. An overnight culture of cells containing pJL3 and56pompPRLF was diluted 1:50 with YT medium containing 100 jig amplmL and 20 jigchloramphenicollmL. After 1.5 h incubation with shaking at 37°C JPTG was added to 40 jiM,MgSO4 was added to 10 jiM and the incubation was continued for 3 h. An overnight culture ofcells containing pSW1 and pompPRLF was diluted 1:50 with YT medium containing 100 jigamp/mL and 12 jig tetracycline/mL. After 1.5 h incubation with shaking at 37°C IPTG wasadded to 100 jiM, mitomycin C was added to 20 ng/mL and the incubation at 37°C was continuedfor 3 h. IPTG induces the expression of the bPRL gene in pompPRLF and the BRP gene inpJL3, and mitomycin C induces the expression of the BRP gene in pSW1. Samples analyzed bySDS-PAGE and Western blotting were diluted in sample buffer containing 13-mercaptoethanoland boiled for 5 mm prior to being loaded on a polyacrylamide gel.2.12.3 Fractionation Procedures for Periplasmic ProteinsTo determine whether bPRL expressed from the vectors pompPRL and pompPRLF wasexported to the periplasm, E. coil strains JM1O1 and RV308 were transformed with each of thesevectors and the resulting cells fractionated in order to obtain the periplasmic contents. The cellfractionation procedures were essentially as previously described (Koshland and Botstein, 1980).Strains JM1O1 and RV308 were transformed with the plasmids pompPRL and pompPRLF.Duplicate cultures in 20 mL YT + amp were incubated with shaking at 37°C for 1.5 h prior toinduction of one of the duplicates with 2 mM IPTG. The RV308 cells were incubated for anadditional 4 h (final0D600=3.8) and the .TM 101 cells incubated for an additional 7 h (final0D600=2.8). Aliquots (1.5 mL) of the cultures were centrifuged in a microcentrifuge for 5 mm,the pellets resuspended in sample buffer and saved as samples of total cell protein. Additional1.5 mL aliquots of the cultures were centrifuged in a microcentrifuge for 5 mm and the pelletsplaced on ice prior to the following fractionation procedures.Lysis by Osmotic Shock: Cells were resuspended in 150 jiL of cold 20% sucrose, 10 mMTris-HC1, pH 7.5 and incubated on ice for 30 mm. Then 0.5 jIL of 0.5 M EDTA, pH 8.0 wasadded, the suspension incubated on ice for an additional 10 mm. A 50 jiL aliquot was removed,diluted with 2X sample buffer, and saved as the periplasm control. The remainder of thesuspension was centrifuged in a microcentrifuge at 4°C for 5 mm. The pellet was rapidly and57vigorously resuspended in 100 L of ice-cold dH2O, incubated on ice for 10 mm, and centrifugedin a microcentrifuge at 4°C for 5 miii. The supematant was diluted with 2X sample buffer andsaved as the periplasmic contents fraction. The pellet was resuspended in 100 iL of dH2Oplus100 ii.L 2X sample buffer and saved as the periplasmic membrane fraction.Formation ofSpheroplasts: Cells were resuspended in 75 iL of ice-cold 0.1 M Tris-HC1,0.5 mM EDTA, 0.5 M sucrose (pH 8.0) and incubated on ice for 30 mm. Then 7.5 j.tL of 2 mglysozyme/mL and 75 jiL cold dH2Owas added and the suspension incubated on ice for anadditional 25 mm. A third of each sample was removed, diluted with 2X sample buffer, andsaved as the spheroplast control. 3 j.tL of 1M MgC12 was added to the remainder and thesuspension centrifuged for 5 mm in a microcentrifuge. The supematant was diluted with 2Xsample buffer and saved as the periplasmic fraction while the pellet resuspended in 100 j.tL dH2Oplus 100 L 2X sample buffer and saved as the membrane fraction.All samples were analyzed by SDS-PAGE and Western blotting.2.13 Yeast Secretion SystemThe yeast bPRL expression vector pVTUP was constructed by inserting the gene for bPRL,with the signal sequence from the yeast invertase gene fused to the 5’ end, into the yeastlE. coilshuttle vector pVT1O1-U (Vernet et at., 1987). The vector contains the E. coli on (forreplication in E. coil), a marker for ampicillin resistance (for selection in E. coil), the fi intergenicregion (for production of ssDNA by filamentous phage), the URA3 gene (for selection in yeast),the yeast 2 on (for replication in S. cerevisiae), and the bPRL gene under the control of theconstitutive ADHJ promoter. The bPRL gene was isolated as a EcoRI-BamHI fragment in whichthe EcoRI ends were blunted with DNA polymerase I (Kienow fragment) and subcloned intopVT100-U digested with Pvull and BamHI.pVTUP was transformed into the S. cerevisiae strains SR1O69B and XV700-24C.Saturated cultures of pVTUP in SR1O69B and XV700-24C were diluted 1:10 into 25 mL of YPDmedium and incubated with shaking at 30°C for 24 h. Cells were harvested by centhfugation at5,000 x g for 10 mm. PMSF was added to the supernatañts to a final concentration of 0.5 mM58and the supernatants were then concentrated using a Centriprep®10 concentrator (Amicon) to afinal volume of approximately 1 mL. The cell pellets were resuspended in 10 mL dH2O,centrifuged as above, resuspended in 4 mL 100 mM EDTA (pH 8), 0.5% 13-mercaptoethanol, andincubated at 37°C for 15 mm. The suspensions were centrifuged (5,000 x g for 10 mm) and thecell pellets resuspended in 100 .tL 50 mM Tris-HC1 (pH 7.5), 10 mM MgC12, 1 M sorbitol, 2 mMDTT. Zymolyase was added to 7.5 U/mL and the suspensions incubated at 30°C with gentleshaking for 1.75 h. The spheroplasts were centrifuged in a microcentrifuge for 5 mm,resuspended in 75 iL 18% Ficoll, 0.5 mM PMSF and centrifuged in a microcentrifuge for 5 mm.The pellets were twice resuspended in 100 tL dH2Oand collected by centrifugation in amicrocentrifuge and then resuspended in 100 jiL dH2Oand stored on ice.The various cell extracts were diluted with 2X sample buffer containing B-mercaptoethanol,boiled for 5 mm and loaded on a polyacrylamide gel for analysis by SDS-PAGE and Westernblotting.2.14 Molecular Modelling of bPRLThe bovine prolactin primary sequence was aligned to that of human growth hormone.Coordinates of homologous regions were transferred from hGH to bPRL using Homologyversion 2.35 (Biosym Technologies). Loop regions with no steric overlap and with goodgeometry were spliced into the bPRL chain once all residues in regions of regular secondarystructure had been modelled. In three cases, helices spanning the bPRL residues 95 - 102, 111 -138, and 162 - 193 had to be manufactured as perfect (0 = -57°) a-helices in the Biopolymermodules and individually docked into the growing bPRL structure. This intervention wasnecessary due to differences in the number of residues present within certain regions of the hGHand bPRL sequences. Fourteen N-terminal residues did not have homologous counterparts inhGH and were not predicted to form a regular secondary structure; hence, these residues (Ti -S 14) were deleted from the bPRL model. Once all coordinates had been assigned, amino andcarboxy termini for the polypeptide were created. Steric overlap was reduced using 100 steps ofconjugate gradient minimization in the Discover module of the Biosym software suite.59Further refmement of the bPRL model was achieved using X-PLOR version 3.1 (Brunger,1992). Two disulfide pairs (residues 58, 174 and 191, 199) were patched. The model was thenregularized using three cycles of unrestrained conjugate gradient minimization and 0.5 psec ofmolecular dynamics at 300°K followed by a fmal minimization. The target energy functionincluded improper angle, Van der Waals, and bond-length components. Subsequent analysis ofbackbone (phi, psi, omega) torsion angles and side-chain rotomers using the PROCHECK suiteof programs (Laskowski et at., 1993) indicated that the model had favourable geometry. Inaddition, no solvent accessible gaps in the structure were present when a molecular surface wasconstructed in the program GRASP (Nicholls et a!., 1993). Figures of the bPRL molecularmodel were produced using the program SETOR (Evans, 1993).60RESULTS3.1 Production of Recombinant bPRLAlthough the expression system originally developed for production of met-bPRL in E. colistrain HB2151 (Luck eta!., 1986, 1989) resulted in efficient production of the hormone,accumulation of the met-bPRL in insoluble inclusion bodies led to difficulties in obtainingbioactive protein. Therefore, in an attempt to avoid the problems associated with obtainingactive met-bPRL from inclusion bodies several avenues were explored to obtain the hormone in asoluble form, including reducing the growth temperature of the transformed cells and developingalternative expression systems, in Saccharomyces cerevisiae and E. coli, for production of bPRL.It has been shown that while certain human proteins (e.g., interferon-a2, interferon-’y andthe interferon-inducedmurine protein Mx) are normally found in aggregates when produced in E.coil grown at 37°C, they were produced in soluble form when growth temperatures in the rangeof 23 - 30°C were used (Schein and Noteborn, 1988). Aggregate formation was not dependentupon the proteins being synthesized in the E. coli cells: interferon-a2 could be aggregated invitro by incubating the protein at 37°C with E. coli cell lysate obtained from growth at either30°C or 37°C. This result suggested that the production of inclusion bodies may be linked toprotein turnover in the E. coil cells, a process which is affected by temperature.In order to test whether growth temperature affects the solubiity of bacterially-producedbPRL, E. coli strains HB2151, Le392F’, and TOPP2 transformed with the bPRL expressionplasmid pESP4 were incubated at room temperature (23°C), 30°C, and 37°C. The soluble andinsoluble fractions of the cells obtained by freeze-thaw lysis were analyzed by SDS-PAGE andWestern blotting. A small amount of soluble bPRL was produced in cells grown in superbrothand terrific broth at 23°C and 30°C however, the amount of hormone present was barelydetectable on a coomassie blue-stained gel and comprised much less than 1% of the total cellprotein. The vast majority of the bPRL produced at all three growth temperatures was in theform of insoluble aggregates (data not shown).613.1.1 Yeast Secretion SystemAn expresssion vector for the production of bPRL in yeast, pVTIJP, was created by fusingthe 5’ end of the bPRL coding sequence to the signal sequence from the yeast invertase gene andsubcloning the product into a yeast/E. coli expression vector, pVT1O1-U (Vernet et at., 1987).Expression of the bPRL gene was under the control of the constitutive ADH1 promoter and thepresence of an ampidilhin resistance gene and the ura3 gene allowed selection of the plasmid inboth E. coli and yeast, respectively. pVTUP was transformed into S. cerevisiae strains SR1O69Band XV700-24C. XV700-24C, canying the pep4 marker, is a protease-deficient strain defectivein processing precursors of vacuolar enzymes (Hemniings et at., 1981). Following 24 h growthin YPD medium, cells were harvested by centrifugation and the culture supernatantsconcentrated. The contents of the cells were isolated by lysing spheroplasts formed by treatmentwith Zymolyase. Samples consisting of the concentrated culture supernatants and the cellcontents were loaded on a denaturing polyacrylamide gel and analyzed by Western blotting usinga colourimetric detection system, polyclonal rabbit anti-oPRL, and an alkaline phosphataseconjugate of goat anti-rabbit IgG. No bPRL was detected in the concentrated supernatants or inthe fractions from the spheroplasting procedure (data not shown).A primer extension experiment was performed in order to determine if the lack of bPRLproduction in S. cerevisiae was due to an absence of bPRL mRNA. Total RNA was isolatedfrom strains SR1O69B and XV700-24C transformed with plasmid pVTUP which contained thebPRL cDNA. As a negative control, RNA was isolated from strain SR1O69B transformed withthe parent plasmid pVT1O1-U. All RNA preparations were examined using the primers MH7,which hybridized specifically to the bPRL cDNA; CYH26 for the yeast ura3 gene; and P20 forthe yeast leu2 gene. Controls included total RNA isolated from strains SR1O69B and XV700-24C not transformed with any plasmid as well as total RNA isolated from ura and ura strains ofS. cerevisiae. Results from the analysis of the above RNA samples indicated that bPRL mRNAwas present in the strain transformed with the bPRL expression vector pVTI.JP (see Figure 8 forresults from strain SR1O69B). The absence of bPRL in the S. cerevisiase strains containing thebPRL expression vector cannot be due to a lack of mRNA and may be due to inefficient62translation of the bPRL mRNA or to degradation of the hormone. These possibilities were notinvestigated and no further effort was made to obtain secreted bPRL from S. cerevisiae.Figure 8. Primer extension analysis of bPRL mRNA in the S. cerevisiae expression system.Total RNA samples analyzed were: 1, from strain SR1O69B not transformed with any plasmid;2, from strain SR1O69B transformed with the parent vector pVT1O1-U; and 3, from strainSR1O69B transformed with bPRL expression vector pVTUP. Primers used were a, MH7, whichhybridizes to the bPRL gene; b, CYH26, which hybridizes to the ura3 gene; and c, P20, whichhybridizes to the leu2 gene. Lane 4 contains total yeast RNA from strain S25 carrying apromoter-less ura3 gene (ura) strain and lane 5 contains total yeast RNA from strain S25carrying the ura3 gene fused to the cyci promoter (uraj. Lanes 4 and 5 were probed withprimer CYH26.3.1.2 E. coli Secretion SystemSecretion of bPRL into the E. coli culture medium was attempted using a system developedfor the secretion of hGH (Hsuing et a!., 1989). The expression vector pompPRLF, containing anampicilin marker and the bPRL gene fused to the ompA signal sequence under the control of thelpp-lac promoter-operator system, was constructed and co-transformed into E. coli HB2151 withthe bacteriocin release protein (BRP) expression vectors pJL3 or pSW1. Dual antibioticresistance (ampicillin plus chioramphenicol or tetracycline) was used to select for transfonnantscarrying both the bPRL secretion vector and one of the BRP expression vectors. pJL3 codes forchloramphenicol resistance and contains the BRP gene under control of the lpp-lac promoter-63a bc bcab’Loperator. pSW1 codes for tetracycline resistance and contains the BRP gene under control of thepC1oDF13 promoter. Bovine PRL was detected in cells of HB2151 transformed either withpompPRLF alone or with pompPRLF plus pJL3 or pSW1 (Figures 9 and 10, section 1).However, no hormone was detected by Western blot analysis of concentrated samples of theculture medium (data not shown). In addition, bPRL was not present in the soluble fractionobtained following lysis of the cell cultures with a French pressure cell (Figures 9 ‘and 10, section2). Therefore, contrary to expectations, the bPRL joined to the ompA signal sequence was notpresent in a soluble form and did not appear to be exported to the periplasm. The bPRL wasinstead found associated with the insoluble material (Figures 9 and 10, section 3), suggesting thatthe hormone accumulated in inclusion bodies. The presence of two closely spaced bands at theposition corresponding to bPRL that were reactive with the anti-PRL antibodies (Figure 10) mayindicate cleavage of the signal sequence from part of the population of bPRL.-d[1 2abc a b C abc—-— —— — — —— — —bPRL a —1 ui_iiZ!iFigure 9. Coomassie blue-stained denaturing polyacrylaniide gels of bPRL in various fractionsfrom the E. coil secretion system. E. coil strain HB2151 was transformed with the bPRLexpression vector, pompPRLF, alone or with pompPRLF plus one of the BRP expressionvectors. Cultures were grown to stationary phase in YT + amp +1- 12 g tetracyclinelmL (forpSW1) +1-20 tg chloramphenicollmL (for pJL3). Samples analyzed were (1) total cell protein;(2) soluble portion following cell lysis by a French pressure cell; and (3) pelleted materialfollowing cell lysis. The expression vectors were (a) pompPRLF; (b) pJL3 and pompPRLF; and(c) pSW1 and pompPRLF. The bPRL standard is 1000 ng of pituitary bPRL. The upper band inthe lane containing the standard bPRL corresponds to BSA, added as a stabilizer for the bPRL.a641 2 3abc abc abcFigure 10. Western blot analysis of bPRL in fractions from the E. coli secretion system. E. colistrain HB215 1 was transformed with the bPRL expression vector, pompPRLF, alone or withpompPRLF plus one of the BRP expression vectors. Cultures were grown to stationary phase inYT + amp +1- 12 ig tetracycline/mL (for pSWl) +1- 20 ig chloramphenicollmL (for pJL3).Samples were loaded on a denaturing polyacrylamide gel, and analyzed by a colourimetricWestern blot system using polyclonal rabbit anti-oPRL and an alkaline phosphatase conjugate ofgoat anti-rabbit IgG. Samples analyzed were (1) total cell protein; (2) soluble portion followingcell lysis by a French pressure cell; and (3) pelleted material following cell lysis. The expressionvectors were (a) pompPRLF alone; (b) pJL3 and pompPRLF; and (c) pSW1 and pompPRLF.The bPRL standard is 750 ng of pituitary bPRL.The lack of bPRL secretion from the E. coli system was investigated further. Onepossibility for the lack of soluble bPRL was that high levels of bPRL production may haveoverwhelmed the ability of the cells to secrete the hormone to the periplasm and thus led to theformation of inclusion bodies. Since YT medium contains components which can induce the lacpromoter, growth of the cells in minimal medium (M9+B1) provided a means of more strictlycontrolling the level of bPRL expression. However, SDS-PAGE and Western blot analysis ofthe concentrated culture supernatants and the soluble and insoluble fractions of cultures of E. coliJM1O1 containing pompPRLF alone or in combination with pJL3 or pSW1 and grown inminimal medium with the appropriate antibiotics and inducers did not reveal increased amountsof soluble bPRL (data not shown). Another possible reason for the accumulation of insolublerather than soluble bPRL was the host strain used for expression of the plasmids. The strain used65for secretion of hGH was E. coli RV308 (Hsiung et al., 1989); therefore, the bPRL producedfrom the secretion vectors was compared in strains RV308 and JM1O1 in order to determine theeffect of the host strain on bPRL secretion. An additional difference between the hGH and bPRLsecretion systems is that the hGH secretion vector did not contain the fi on: this was inserted intothe plasmid for the bPRL secretion system. Although it seemed unlikely that the presence of thefi intergenic region could affect secretion of bPRL, this possibility was investigated by removingthis region thereby creating the vector pompPRL. Duplicate cultures of JM1O1 and RV308transfomied with pompPRL and pompPRLF were grown in YT + amp for 1.5 h prior to theinduction of one culture of each with 2 mM IPTG. Cells in the exponential phase of growth wereharvested and fractionated in order to specifically release the periplasmic components (Koshlandand Botstein, 1980). Samples from stages in the fractionation procedures were examined for thepresence for bPRL by denaturing SDS-PAGE and Western blotting. No differences wereobserved between strains JM1O1 and RV308 nor between vectors pompPRL and pompPRLF.The results of the fractionation of RV308 are shown in Figure 11. Two different methods wereused to release the periplasmic components: in each case the bPRL remained associated with theinsoluble fraction.3.1.3 Intracellular E. coli System for Production ofMet-bPRL and Met-bPRL VariantsThe met-bPRL and met-bPRL variants synthesized in E. coli cells are produced asinsoluble inclusion bodies and the method developed for the isolation of met-bPRL involvesdisruption of the cells, centrifugation to pellet the insoluble met-bPRL, and treatment of thepellet with detergents in order to remove contaminating E. coli proteins followed bysolubilization and renaturation of the met-bPRL. Cells from stationary phase cultures, with atotal protein concentration in the range of 0.8 - 1.2 mg/mL (Figures 12A and l3A) weredisrupted in the presence of 25-50 mM DTT by a single passage at maximum pressure througha French pressure cell. The insoluble material, of which the met-bPRL is a major constituent,was collected by centrifugation and resuspended in buffer containing 5 mM DT1’ to a proteinconcentration of approximately 3 - 3.5 mg/mL (Figures 12B and 13B). Treatment of thissolution with 0.4% sodium desoxycholate solubiized some E. coli proteins (Figures 12C and6613C). A portion of the met-bPRL was solubilized by extraction with 0.1% N-lauryl sarcosine(sarcosyl) and renatured at pH 10 (Figures 12D and 13D). The protein concentration of the finalsolutions, consisting primarily of met-bPRLs, was generally in the range of 0.1 - 0.14 mg/mL.The yield of met-bPRL from the E. coil cells was 1 mg from 100 mL of cell culture or less.However, the amount of protein obtained was more than sufficient for the in vitro bioassaywherein the ability of the met-bPRL extracts to stimulate the growth of Nb2 cell cultures isdetermined (Tanaka et a!., 1980). In addition to providing more than enough recombinantprotein for the bioassay, other positive features of the extraction procedure are that it istechnically very simple and lends itself well to the processing of many different samples at onetime. For bioassay, the hormone preparations were diluted in Fischer’s medium containing 10%non-mitogenic horse (gelding) serum and assayed at a number of different concentrations to givea growth response within the useful working range of the assay, i.e., 0.01 - 0.25 ng standardpituitary bPRL/mL.671 2 3 4 5abab abab abab ab ab abab++-- ++ -- ++- -6 7abab ababOFigure 11. Western blot analysis of bPRL in fractions from the E. coli secretion system. E. colistrain RV308 was transformed with the bPRL expression vectors pompPRL or pompPRLF andcultures were grown to stationary phase in YT + amp. Samples obtained following fractionationof the cells to release the periplasmic components were loaded on a denaturing polyacrylamidegel and analysed by a colourimetric Western blot system using polyclonal rabbit anti-oPRL andalkaline phosphatase-linked goat anti-rabbit IgG. Samples analyzed were (1) total cell protein;(2) periplasm control; (3) periplasmic contents; (4) periplasmic membrane components; (5)spheroplast control; (6) periplasmic contents from spheroplasts; and (7) membrane fraction fromspheroplasts. (a) vector pompPRL; (b) vector pompPRLF. (+) indicates cells induced with 2mM IPTG; (-) indicates no added IPTG.68A -J0.0 12345678— — — — -Figure 12. Coomassie blue-stained denaturing polyacrylamide gels of met-bPRL and severalmet-bPRL variants. The met-bPRLs were synthesized as insoluble inclusion bodies in E. colistrain TOPP2 and samples at stages of the bPRL extraction procedure were analyzed underreducing conditions by SDS-PAGE. Samples were (A) total cell protein; (B) pelleted materialfollowing lysis of cell cultures by a French pressure cell; (C) insoluble material remainingfollowing extraction of the pelleted material with 0.4% sodium desoxycholate and 0.1% sarcosyl;and (D) renatured sarcosyl extract of the sodium desoxycholate-extracted pellet. Lane 1,unmodified met-bPRL; lane 2, R176A; lane 3, D178E; lane 4, S179A; lane 5, S18OA; lane 6,K181A; lane 7, D183A; lane 8, L189A. Variants R176A, D178E, S179A, K181A and L189Awere produced efficiently while pESP4 and Si 80A were synthesized in lesser amounts. Verylittle of the variant D183A was produced in the TOPP2 cells. The bPRL standard in each case is500 ng of pituitary bPRL.C-Ja-12345678 .0974kDa68kDa43kDaI25.7kDa18.4 kDal43kDa• — ——— — -———__.0 1234 587897.4 kDa68 kDa43 kDa25.7 kDa18.4 kDa14.3 kDaB -j12345678“69V1234567897.4 kDa68kDa43kDa25.7kDa18.4 kDa14.3 kDaD12345678 1234567897.4kDa68 kDaE14.3 kDaFigure 13. Western blot analysis of met-bPRL and several met-bPRL variants. The met-bPRLssynthesized in E. coli TOPP2 were extracted with 0.1% sarcosyl from 0.4% sodiumdesoxycholate-treated insoluble cell material obtained after cell lysis. The sarcosyl-soluble metbPRL samples were renatured at pH 10. Gels A, B, and C were run under reducing conditionsand gel D was run under nonreducing conditions. All samples were analyzed by Westernblotting using a chemiluminescent detection system, polyclonal rabbit anti-oPRL, and ahorseradish peroxidase conjugate of goat anti-rabbit IgG. Samples were (A) total cell protein;(B) pelleted material following cell lysis; (C) insoluble material remaining following extractionwith sodium desoxycholate and sarcosyl; and (D) renatured sarcosyl-soluble met-bPRLs. Lane1, unmodified met-bPRL; lane 2, R176A; lane 3, D178E; lane 4, S179A; laneS, S18OA; lane 6,K181A; lane 7, D183A; and lane 8, L189A. The bPRL standard in each case is 50 ng ofpituitary bPRL. In gel D the faint slow-moving bands in all samples including the standardscorrespond to high molecular weight aggregates of bPRL, reduced met-bPRL is present in theband slightly above the major band corresponding to oxidized met-bPRL, and the very faint fast-moving band in the met-bPRL samples is most likely a degradation product.-J070Although bPRL is a monomeric protein (23 kDa molecular weight) high molecular weightaggregates of met-bPRL were often obtained following sarcosyl extraction of transformed E. colicultures. These aggregates, which react with anti-PRL antibodies, are visible on non-reducinggels as a doublet of bands of approximately 45 - 50 kDa molecular weight along with a numberof other less prominent bands of higher molecular weights (Figure 13). The aggregates arepresumed to consist of met-bPRL molecules linked covalently through disulfide bonds to itself orto thiol-containing E. coli proteins (Luck and Huyer et a!., 1992). Lysis of the E. coli cells in theabsence of DTT resulted in the presence of very high amounts of the high molecular weightaggregates in the sarcosyl-soluble extracts. Inclusion of DT1 in the buffers used when lysing thecell cultures and isolating the inclusion bodies reduces but does not eliminate the formation ofthe high molecular weight aggregates (Luck and Huyer et at., 1992). For reasons as yet notunderstood, the amount of high molecular weight material in the met-bPRL extracts variedbetween preparations and occasionally a substantial amount of the aggregates was present inspite of the inclusion of DTf. Such preparations were not suitable for quantitation of thehormone and its bioactivity since in addition to being immunoreactive they are probablybioactive. Thus, in this study, the met-bPRLs were quantified for bioassay based on the amountof monomeric oxidized protein present in extracts containing only small amounts of the highmolecular weight contaminants. In other experiments, high bioactivity values (in the range of200% for umnodifled met-bPRL) based on the amount of monomeric met-bPRL were obtainedfrom extracts containing large amounts of the high molecular weight aggregates (data notshown). It is not known whether certain of the aggregates stimulated Nb2 cell mitogenesisdirectly or whether the met-bPRL present in the high molecular weight material dissociatedduring the course of the bioassay to produce bioactive monomeric bPRL3.1.3.1 E. coli Strains for Production ofMet-bPRLAs a means of optimizing met-bPRL production in E. coli, a number of different E. colistrains were transformed with the bPRL expression vector pESP4. The level of met-bPRLsynthesized in each strain was analyzed by SDS-PAGE and Western blotting as follows. 50 mLand 100 mL YT + amp +5 mM IPTG cultures of the E. coli strains transformed with pESP471were grown overnight with shaking at 37°C in 250 mL Erlenmeyer flasks. The insolublematerial obtained following lysis of the stationary phase cells was analyzed by denaturing SDSPAGE. The level of met-bPRL produced varied greatly between the different E. coli strains(Figure 14). Very little, if any, met-bPRL was synthesized in strains 1M105, RV308, CSH5O orMV1193. The lack of met-bPRL production in strain RV308 is interesting since this strain wasused for high-level expression in E. coli of the closely related hormone bGH (Schoner et at.,1985). The highest level of met-bPRL expression was obtained in strain Le392F’ and the amountof hormone synthesized was slightly higher than in strain HB215 1 (data not shown).The E. coli strains TOPP1, TOPP2, and TOPP3, marketed by Stratagene for use inexpressing hard-to-express proteins, were also transformed with pESP4 and the level of metbPRL production examined. Western blot analysis of the renatured sarcosyl-soluble met-bPRLextracts from each of these strains revealed that, amongst the TOPP strains, bPRL productionwas highest in TOPP2 (Figure 15). Although the amount of met-bPRL synthesized in theTOPP2 cells was less than in Le392F’, the amount of high molecular weight aggregates presentin the sarcosyl-soluble met-bPRL samples was much less in the TOPP2 extracts than in extractsfrom the other E. coli strains (Figures 15 and 16). For this reason TOPP2 was selected as thehost E. coli strain for production of met-bPRL.72-to to‘ 0) 0c4J — C,)w >I - -J a:IIab a •b ab a b— -— £. — .— —— ——— w -. —18.4 kDa14.3 kDa97.4 kDa68kDa43 kDa25.7 kDaFigure 14. Coomassie blue-stained denaturing polyacrylamide gels of pelleted materialfollowing lysis of various E. colt strains transformed with pESP4. E. coli strains HB2154,3M105, Le392F, RV308, CSH5O, MV1193, NM522, and TG1 were transformed with pESP4and grown in YT + amp +5 mM IPTG to stationary phase. The insoluble fraction of each cellculture obtained after cell lysis was analyzed by SDS-PAGE under reducing conditions. Theextracts loaded in lane (a) were from 50 mL cultures in 250 niL Erlenmeyer flasks; and in lane(b) from 100 niL cultures in 250 niL Erlenmeyer flasks. The bPRL standard is 500 ng pituitarybPRL4-0-Ja-.097.4 kDa68 kDa43 kDa25.7 kDabPRL—*bPRL -+0)00)If)C,)-j Cl)>C.)I—,.0 babCsjc’j100z I—II I,a b a b18.4 kDa14.3 kDa734-TOPP1 TOPP2 TOPP3SB TB SB TB SB TB4 1-ugh Mw aggregates-- bPRL (reduced)— bPRL (oxidized)Figure 15 Western blot analysis of met-bPRL production in E. coli TOPP strains 1,2, and 3.The cell cultures were grown to stationary phase in superbroth (SB) and terrific broth (TB), andthe insoluble fraction of each treated with 0.4% sodium desoxycholate and 0.1% sarcosyl. Thesarcosyl-soluble met-bPRL samples were renatured at pH 10 and loaded on a nondenaturingpolyacrylamide gel for Western blot using a chemiluminescent detection system. The bPRLstandard is 50 ng of pituitary bPRL. The very faint slowest-moving anti-oPRL-reactive bandsreveal that the high molecular weight met-bPRL aggregates were present in very small amounts.The almost equal intensity for each met-bPRL sample of the band corresponding to the reducedform of bPRL (slower moving) with that of the band corresponding to the oxidized form ofbPRL (faster moving) indicates that the renaturation was approximately 60% complete.74U) U)—C bPRLstd.Z - -i a:11 r9 ) IIU)abab abab c,High MW aggregatesbPRL bPRL (recliced)I bPRL (oxdzed)CI’0)14)ZCi)>C)High MW aggregatesbPRL (rechiced)bPRL - dP bPRL (oxidzed)Figure 16. Western blot analysis of renatured met-bPRL from various E. coil strainstransformed with pESP4. The E. coil strains HB2154, JM1O5, Le392F, RV308, CSH5O,MV1 193, NM522, and TG1 were transformed with pESP4 and grown to stationary phase in YT+ amp +5 mM IPTG. Following cell lysis, met-bPRL was isolated by 0.2% sarcosyl-extractionof 1% sodium desoxycholate-treated insoluble material. The sarcosyl-soluble met-bPRLs wererenatured at pH 10 and loaded on a nondenaturing polyacrylamide gel for Western blot analysisusing a chemiluminescent detection system. Met-bPRL samples in lane (a) are from 50 mLcultures; and in lane (b) from 100 mL cultures. Approximately 20% of each met-bPRL samplewas present in the reduced form, as shown by the intensity of the band slightly above thatcorresponding to the oxidized form of bPRL (indicated with an arrow). The met-bPRL fromHB2154 (lane b), Le392F (lanes a and b), MV1 193 (lane b), NM522 (lanes a and b) and TG1(lanes a and b) contained significant amounts of the high molecular weight aggregates of bPRL,as indicated by the presence of the slowest-moving anti-oPRL-reactive bands. The bPRLstandards are the indicated amounts of pituitary bPRL.Cs1c’JU)z‘Ib a b ab.— _hbPRL std.0(I’753.2 Extraction ofMet-bPRL and Met-bPRL Variants from Inclusion BodiesWhile the amount of monomeric met-bPRL extracted from E. coil cultures could beincreased by the use of higher concentrations of sarcosyl, this also increased the amount ofcontaminating E. coil proteins (Luck eta!., 1989) and of the high molecular weight met-bPRLaggregates. In order to examine the extent to which sarcosyl solubilizes the high molecularweight aggregates, the insoluble material from cultures of E. coil TOPP2 cells expressingunmodified met-bPRL and a number of met-bPRL variants was extracted with 0.4% sodiumdesoxycholate followed by either 0.1% or 0.2% sarcosyl. The final extracts were loaded on anondenaturing polyacrylamide gel and analyzed by Western blotting (Figure 17). The figureclearly shows that the amount of high molecular weight aggregates present in the extracts wasmuch lower in the extracts obtained with 0.1% sarcosyl. However, although the proportion ofmonomeric met-bPRL is higher in the 0.1% sarcosyl extracts the total amount of met-bPRL wasmuch less. It may be noted that the volume of the 0.1% sarcosyl extracts loaded on the gel wasdouble that of the 0.2% extracts.Differences in the relative amount of hormone extracted by incubation with 0.1% sarcosylhave been observed between met-bPRL variants. in general, a high level of met-bPRL or metbPRL variant expression in the E. coil cells correlated with a high concentration of hormone inthe sarcosyl extract. However, this was not always the case. For example, the amount of thevariant Q71V synthesized in E. coil TOPP2 was quite high (Figure 18A, lane 5) but the amountextracted with 0.1% sarcosyl was relatively low (Figure 1 8B, lane 5). Certain mutations cantherefore affect characteristics of the protein other than those relating specifically to function,which in the case of bPRL is the ability to interact with the receptor. Another effect on thesolubility of met-bPRL was noted with disulfide bridge variants and with C-terminal truncatedmutants. Variants lacking either the C-terminal disuffide bridge or C-terminal residues werealmost completely solubilized by treatment with 0.4% sodium desoxycholate (Luck and Huyer etai., 1992; Dr. Luck, Oberlin College, Oberlin, OH, personal communication).76bPRL std.ACC1 2 3 4567 8 9• HighMWagegatesbPRL (reduced)— bPRL (oxidized)BbPRLstd.0)0)C C C C0 Lt) 0 L()1234 5678 ‘- Lf)N- . pa — ugh MWIIIbPRL (reduced)bPRL (oxidized)Figure 17. Western blot analysis of met-bPRL and various met-bPRL mutants. Cultures of E.coil TOPP2 were grown to stationary phase in terrific broth + amp +3 mM IFTG. The insolublematerial obtained following cell lysis was treated with 0.4% sodium desoxycholate and the metbPRLs were extracted from the sodium desoxycholate-insoluble material with (A) 0.1% sarcosyland (B) 0.2% sarcosyl, renatured at pHlO, and loaded on a nondenaturing polyacrylaniide gel forWestern blot analysis using a chemiluminescent detection system. The volume of the 0.1%sarcosyl extracts loaded was twice that of the 0.2% sarcosyl extracts. Lane (1) unmodified metbPRL; (2) T6OA; (3) N17OA; (4) H173A; (5) L175A; (6) D178A; (7) Y185A; (8) 1193A; and (9)G129R. The amount of hormone extracted varied between samples: T6OA and H173A werepresent in the highest amount while very little D178A was obtained. Renaturation of the metbPRLs was more than 90% complete as indicated by the very faint band of the reduced form ofbPRL visible just above the major band of oxidized bPRL. The slowest-moving anti-oPRLreactive bands correspond to the high molecular weight aggregates of bPRL. The bPRLstandards are pituitary bPRL in the amounts indicated.77Au)-J12345678 .0- 974kDa68kDa— — ——-— — 43 kDa— — — -. — ,—“ !! 257kDa— —_ —..— I_18.4 kDal43kDaVBu)-Ja.01 23456797.4 kDa68kDa—— 43 kDa— — -25.7kDa18.4 kDa14.3 kDaFigure 18. Coomassie blue-stained denaturing polyacrylamide gels of met-bPRL and variousmet-bPRL mutants. Cultures of E. coil TOPP2 were grown to stationary phase in terrific broth +amp +5 mM IPTG. The insoluble material obtained folowing cell lysis was treated with 0.4%sodium desoxycholate and the met-bPRLs were extracted with 0.1% sarcosyl. Samples analyzedby SDS-PAGE were (A) pelleted material following cell lysis; and (B) sarcosyl-soluble metbPRLs, not renatured at pH 10. Lane (1) unmodified met-bPRL; (2) T6OA; (3) L63A; (4) P64A;(5) Q71V; (6) Q73V; (7) Q74V; and (8) E128A. The amount of the met-bPRL variantssynthesized (gel A) varied considerably: Q7 1V was produced at a very high level; T6OA, P64A,and E128A were produced at a slightly lower level; unmodified met-bPRL, L63A, and Q73Vwere synthesized at even lower levels; and no Q74V could be detected. However, the amount ofthe met-bPRL samples extracted by 0.1% sarcosyl (gel B) showed less variation: approximatelyequal amounts of all met-bPRL samples except T6OA (slightly higher) and Q74V (notdetectable) were present in the extracts. The bPRL standard is 500 ng of pituitary bPRL.783.3 Analysis ofMet-bPRL mRNA Levels by Competitive PCRMutations may have unexpected secondary effects on gene transcription or on the stabilityof the protein during synthesis. These effects may be responsible for the poor expression in E.coli of the bPRL variants R125A and D178A. In order to investigate this possibility, reversetranscription was coupled with competitive PCR to quantitate the amount of RNA present in E.coli cells expressing different met-bPRL variants using methods previously described for othersystems (Foley et a!., 1993; Gilliland et al., 1990; Penin and Gilliland, 1990).Competitive PCR was carried out using a single set of primers on the cDNA of interest anda known amount of a competitor DNA. The competitor chosen was the bPRL eDNA containingthe signal sequence for the yeast invertase gene: the PCR products from the amplification of thevariant bPRL cDNA and the competitor bPRL DNA differed in length by approximately 50 bpand were separable by agarose gel electrophoresis. Comparison of the amount of mRNA presentin TOPP2 cells transformed with vectors for poorly expressed mutants (R125A and D178A) withthat from cells transformed with more highly expressed variants (unmodified met-bPRL andD178E) did not reveal a significant difference in the relative amounts of mRNA (see Figure 19for comparison between unmodified met-bPRL and R125A). Therefore, the differences in levelsof protein expression did not seem to be related to differences in transcription efficiency.Poor levels of expression of certain variants, could result from inefficient translation orcould possibly be a result of decreased protein stability. Thus, substitution of certain residues,particularly those with charged and/or bulky side chains, with a small amino acid such as alaninecould destabilize the protein and thus ultimately lead to low levels of expression. Alternatively,problems in mRNA translation could lead to low levels of production of some variants. Codonusage can affect the level of gene expression; therefore, use of E. coli preferred codons oftenimproves expression of heterologous proteins (Gouy and Gautier, 1982; Grosjean and Fiers,1982; Robinson et at., 1984). A single amino acid can be represented by any one of severalcodons but the frequency with which each synonymous codon is used is not identical. Morefrequently used codons may correspond to more frequent tRNAs and thus the high translationefficiency seen with the genes for highly expressed proteins may be a consequence of bias79towards the most common codons. However, codon bias is not absolute and it is thereforedifficult to predict which codon at any one position will lead to efficient expression. Thus, whilemutations in the bPRL cDNA were designed to incorporate the more frequently used codons,certain choices may have deleteriously affected gene translation.AA B C 1 2 3 45 7 8° in— — — — — —BFigure 19. Quantitation of RNA by competitive PCR. eDNA prepared using AMV reversetranscriptase from total RNA from E. coli TOPP2 expressing (A) R125A and (B) unmodifiedmet-bPRL was amplified in the presence of a dilution series of pMH4 DNA (bPRL eDNA fusedto the signal sequence from the yeast invertase gene). The flanking primers were: MH53, 5’-GTGAGCGGATAACAA1TFCACACAGGA-3’; and MH19, 5’-TCTCAGAAATGGATAG-GATCCAATGTG-3’. Following 25 rounds of PCR amplification samples loaded on a 1.9%MetaPhor agarose gel in TBE buffer were visualized by UV light following staining withthidium bromide (0.5 jig/mL in dH2O). Lanes 1 - 10 are from amplification reactionscontaining pMH4 plasmid DNA ranging in the relative amounts: 100, 30, 20, 8,4, 1.2,0.8,0.32,0.16, and 0.016 (reaction 1 contained approximately 5 amol pMH4). Controls are amplificationreactions carried out in the absence of competitor DNA on (A) pMHl DNA; (B) pMH4 DNA;(C) TOPP2 cDNA; and (D) dH2O.A B D 1 2 45 6 7 8 9803.4 Renaturation ofMet-bPRL and Met-bPRL VariantsIn previous studies it was found that met-bPRL and met-bPRL variants extracted frominsoluble inclusion bodies by consecutive extractions with 1% sodium desoxycholate and 0.2%sarcosyl was soluble, but exhibited only about 40% of the bioactivity of the native hormone. Aslight increase in the bioactivity of sarcosyl extracts of met-bPRL was noted to occurspontaneously during storage of the samples at 4°C although fully bioactive protein was notobtained (Luck et at., 1989). Western blot analysis of met-bPRL extracts electrophoresed on anondenaturing acrylamide gel revealed bands corresponding to the reduced and oxidized formsof bPRL, as can be seen in Figure 20 (Luck and Huyer et at., 1992). The low level of bioactivitywas therefore felt to be due to problems in disulfide bond formation, as bPRL contains threedisulfide bridges: incorrect and incomplete disulfide bond formation can cause proteins to fail toattain the correct conformation and therefore form inclusion bodies (Pigiet and Schuster, 1986;Schoemaker et at., 1985). Exposure of the samples to air may have promoted correct disulfidebond formation between some of the initally reduced cysteine residues but would not havecaused incorrectly formed disulflde bonds to break. In this study, renaturation of met-bPRL wasfound to be also promoted by the growth medium used for the Nb2 cell bioassay. Thus, extractsof met-bPRL incubated with Fische?s medium containing 8-mercaptoethanol in aC02-enrichedatmosphere over a 48 h period exhibited an increase in the amount of oxidized met-bPRL and acorresponding decrease in the quantity of reduced protein (Figure 20).The refolding of proteins obtained from inclusion bodies is generally accomplished byremoving the denaturing agent, usually by dialysis or dilution, and is influenced by a number ofdifferent variables (Marston, 1987). Factors such as the rate at which the denaturing agent isremoved, the concentration and purity of the protein preparation, and the pH affect the yield ofactive protein. Methods which have been used successfully to obtain active PRL include dialysisof inclusion body material solubilized by 8 M urea (Gilbert et a!., 1991; Paris et at., 1990) andincubation at high pH (Dr. Byatt, Monsanto Corp., St. Louis, MS, personal communication).The method initially used to convert the sarcosyl-soluble met-bPRL to a fully active form was airincubation with a 50-fold excess of thioredoxin used as a 10:1 redox mixture of the oxidized and81bPRL Standard Controls Incubation in growth mediumrFigure 20. Western blot analysis of non-renatured met-bPRL incubated in Nb2 cell growthmedium. Met-bPRL was extracted using 0.4% sodium desoxycholate and 0.2% sarcosyl from E.coil TOPP2 transformed with pESP4. The sarcosyl-solublized met-bPRL was diluted 1:50 withFischer’s medium containing 10% gelding serum and 0.1 mM f3-mercaptoethanol. Theapproximate final concentration of hormone was 0.5 tg/mL. Samples were sparged with 5%C02/air, incubated at 37°C for the indicated amounts of time, loaded on a nondenaturingpolyacrylamide gel, and analyzed by Western blotting using a chemiluminescent system,polyclonal rabbit anti-oPRL, and a horseradish peroxidase conjugate of goat anti-rabbit IgG. ThepH of all samples remained at pH 7.4 during incubation. Control samples were (a) met-bPRL in0.1 M sodium phosphate buffer, pH 7.4; (b) met-bPRL in Fischer’s medium containing geldingserum but not sparged with 5% C02; and (c) met-bPRL in Fischer’s medium containing geldingserum and p-mercaptoethanol and sparged with 5% CO2. The slow-moving bPRL-reactive bandsindicates that significant amounts of high molecular weight aggregates of bPRL were present inall samples. The bPRL standards are the indicated amounts of pituitary bPRLreduced forms (Luck et a!., 1989). However, thioredoxin is not an ideal renaturing agent becauseit is expensive and its use leads to the introduction of a large amount of foreign protein into thesamples. Therefore, an alternative method of renaturing met-bPRL and met-bPRL variants byincubating the sarcosyl-solublized met-bPRL extracts at pH 10 in the presence of air wasdeveloped (Luck and Huyer et al., 1992). This procedure generally resulted in the efficient andrapid refolding of the hormone. However, occasionally incubation at pH 10 did not lead tocomplete renaturation of the met-bPRL. In this study, attempts to determine the source of the0) 0) 0) 0) CC C C CU•) 0 0 0 0— c) co co —.C .Cc’J.?a b c— ‘,ill ..,,,bPRL (reduced) —bPRL (oxidized) —) —_— — ——..*—. S .82variability in the degree of renaturation focussed on the length of time of sample incubation atpH 10, the volume and concentration of each sample, the amount of air introduced to thesamples, and the effect of different concentrations of sodium desoxycholate and sarcosyl (datanot shown). No reason was found for the variability.35 Structure-Function Analysis ofMet-bPRLBovine PRL is a 23 kDa globular protein containing three disulfide bridges and fourputative cz-helices (Figure 21) and although no three-dimensional structure has been established,it is generally anticipated that the structure of bPRL is similar to that of growth hormone (seeIntroduction). In an effort to determine the relationship between amino acid residues and themitogenic activity of met-bPRL site-directed mutagenesis was used to examine the effect ofchanges of selected amino acids on the mitogenic activity of the hormone in the Nb2 cell assay.It has previously been shown that the mitogenic activity of met-bPRL is equivalent to that ofnative bPRL (Luck et a!., 1989). The amino acid changes involved the substitution, primarily byalanine, of bPRL residues which either corresponded directly to those involved in receptorinteractions in related hormones or which fell within the specific regions of the hormone that,based on structural homologies with hGH, are postulated to play a role in contacting thelactogenic receptor.Helix 1 Helix 2 Helix 3 Helix 4__15 42 77 102 112 137 163 193s-s s-s s-sN± tIC4 11 58 174 191 199Figure 21. Diagram of bPRL showing its three disulfide bridges and the locations of theputative helical regions. Numbers indicate the position of the cysteine residues and the putativelimits of the helical regions.833.5.1 Putative Lactogen-Specifc Residues Identified Via Comparative StudiesAn initial study of the relationship between the structure of met-bPRL and its mitogenicactivity in the Nb2 cell assay focussed on seven residues conserved amongst a group of 10pituitary lactogens (bovine, equine, human, mouse, ovine, porcine, rat, and whale PRLs, andhuman and monkey GHs) that are not found in the related, but non-lactogenic bGH (Luck et a!.,1989 and Figure 22). Variants of met-bPRL which had been made by substituting in met-bPRLcorresponding bGH amino acids had bioactivities of the same order as that of unmodified metbPRL. However, although this strategy of substituting a conserved amino acid from a lactogenichormone with the corresponding residue from a non-lactogenic growth hormone failed todistinguish any lactogen-specific amino acids in bPRL, a similar approach was used successfullyto identify three residues which play a role in the mitogenic activity of mPL-ll (Davis andLinzer, 1989b). In the case of mPL-II, five residues common to a group of lactogenic hormonesthat differed at the corresponding positions in a group of nonlactogenic hormones were altered tothe residues in the nonlactogenic hormone proliferin.A comparison of the sequences of bPRL and mPL-ll indicated that residues R21, R177,and K187 of bPRL are equivalent to the mPL-ll residues identified as being critically importantfor binding to the PRL receptor and for PRL-induced mitogenesis of Nb2 cells (namely R14,R169, and K179) (Figure 22). The importance of the equivalent three bPRL residues to themitogenic activity of the hormone was investigated (Luck and Huyer et a!., 1991). Each of thesethree residues was substituted with alanine, asparagine, leucine, and lysine or arginine. Alanine,a small, nonpolar residue, was chosen for substitution because it can be perceived to be the mostinnocuous replacement (Cunningham andWells, 1989) while asparagine and lysine or argininewere chosen because these residues are in the equivalent positions to R21, R177, and K187 ofbPRL in the non-lactogenic bGH. Leucine was chosen because the RI4L rnPL-ll variant showeda large decrease in mitogenic activity (Davis and Linzer, 1989b). The mitogenic activities of thevariants are shown in Table 4. Changes of any of these three residues to alanine, leucine, orasparagine resulted in decreases in mitogenic activity while an effect from the more conservativeinterchange of arginine and lysine was seen only at position 177. The importance of R177 to the84mitogenic activity of bPRL was exanined further through a number of additional substitutions.Replacement of this residue with aspartate, glutamate, tyrosine, glycine, serine, glutaniine,histidine, and phenylalanine resulted in variants all with very low bioactivity ( 5%) (Luck andHuyer et aL, 1991).10 20 30 40 50 60bPRL - -AQGKGITMALNS-cHTSSLPIhPRL - -THGRGEITKAINS-HTSSLAImPL- II LPNYRLPTESYQVIVVS-HNAjDASKAMEMKF- - -GRTAW-TYGLMLSPHIAAILIhGH FPTIPSRjN RAliRLijQLLAFDTYQEEAXIPKEQKYSELQNPQTSLFSEIP1bGH PEGQRYS - IQNTQVAFFSETI PA* * * *70 80 90 100 110 120 130bPRLhPRL PEDKEOAQQNNQKDFLSLLIVSmPL- IIhGH £SNREETQQKSNLELLRISLLLIQWLE.EVQFLRS -VFANSLVYGASDNVYDL- - -LDJ.EEQIQTLMCbGH £NEQQKSDLELLRISLLLIQWLGQFSR-VFTNSLVFGTSD-RVYEK-- -LDEILALMR* *140 150 160 170 180 190bPRL 2-VI.EGAETEP-hPRL -VH.ETiENEI-mPL- II NRVYPGAVASDYTFWSAWSDLQSDES- - -TKNSALRTLWRCVRR1YPHKVNYLKVjKCRDVHNNNhCH -RLEDGSPRTGQbGH*Figure 22. The complete amino acid sequences of the lactogenic hormones bPRL, hPRL, mPLII, and hGH and of the related but non-lactogenic bGH. The sequences are aligned to maximizethe homology between the hormones (Luck et a!., 1989). The numbers refer to the amino acidlocations in bPRL. The underlined residues are conserved in eight mammalian pituitary PRLs(bovine, equine, human, mouse, ovine, porcine, rat, and whale). The seven residues found in 10pituitary hormones (bovine, equine, human, mouse, ovine, porcine, rat, and whale PRLs plushuman and monkey GHs) but not in bGH are marked by asterisks (*).85Table 4. Mitogenic Activities of Met-bPRL Variants Producedby Amino Acid Substitutions at Positions R21, R177, and K187Amino Acid % Activity RelativeSubstitution to Pituitary bPRLUnmodified met-bPRL 101.5 ± 5.0R21A 20.9 ± 0.2R21L 29.2 ± 0.2R21N 50.0 ± 4.0R21K 88.6 ± 5.6R177A 1.1±0.05R177L 0.9±0.1R177N 2.4 ± 0.03R177K 12.8 ± 1.6K187A 46.3 ± 7.3K187L 41.8 ± 8.8K187N 81.0±2.0K187R 99.1 ± 4.9a The data obtained for each variant are the means (±SEM) of theresults of bioassays done with three different mutant met-bPRLpreparations.3.5.2 Substitutions of Residues in the Loop Region Between Putative Helices I and 2Investigations into the interaction of hGH with the hGHbp and the hPRLbp have revealedthat residues within the second half of the loop joining helices 1 and 2 are involved in binding ofthe hormone to these receptors (Cunningham and Wells, 1989, 1991). In addition, this regionwas shown to be important to the mitogenic activity of hPRL (Goffm et a!., 1992). Althoughthese amino acids (comprising residues 58 -74 in bPRL) do not, in any of the related hormonesthat have been examined (Figure 22), appear to form part of a specific structural element such asa helix or p-sheet, the sequence of this segment is highly conserved within the hormone familysuggesting that for all of these proteins this region is structurally and/or functionally important(Goffin et al., 1992).The importance of this putative loop region to the mitogenic activity of bPRL wasinvestigated through alanine-scanning mutagenesis. Met-bPRL variants, each containing a single86mutation, were produced as described in Materials and Methods. Residues T60, S61, L63, P64,P66, E67, and K69 were each replaced by alanine. Residues Q71, Q73, and Q74 were eachreplaced by alanine and valine and H59 was replaced by alanine and serine. The contributions ofthe amino acids S62, T65, D68, and E70 to the mitogenic activity of bPRL were examined in anearlier study: substitutions at any of these positions had only a minor effect on activity (Luck eta!., 1989). Position 72 was not examined because it already contains an alanine. All of thevariants were produced as insoluble inclusion bodies in E. coli TOPP2 cells except the Q71,Q73, or Q74 variants which were produced in HB2151 cells; they were solubiized with 0.1%sarcosyl and renatured by air oxidation at pH 10. Western blot analysis of the variants subjectedto polyacrylaniide gel electrophoresis under non-reducing conditions revealed that greater than80% of each preparation had the same electrophoretic mobility as pituitary bPRL, thus indicatingthat each had renatured with the C58 - C174 disulflde bond intact (Luck and Huyer et a!., 1992).Most of the met-bPRL variants were expressed in E. coli strain TOPP2 cells at a levelcomparable to that of unmodified met-bPRL The low levels of expression observed when Q71,Q73, and Q74 were substituted led to difficulties in extracting sufficient quantities of thehormone for the Nb2 cell bioassay. In fact, the variants with alanine substitutions at thesepositions were produced at such low levels in the TOPP2 cells that their mitogenic activity couldnot be determined. Expression of the Q71, Q73,and Q74 variants, containing either alanine orvaline substitutions, was much higher in E. coli strain HB2151 (data not shown) and bioactivityvalues were determined for hormone extracted from these cells.The results of the bioassays on these variants are presented in Figure 23. Variants T6OA(72%), S61A (7 1%), P66A (70%) and E67A (90%) were nearly as bioactive as unmodified metbPRL while the mitogenic activity of L63A (5 1%) and P64A (61%) was slightly reduced. H59A(38%), H59S (23%), Q71V (35%), Q73V (38%), and Q74V (36%) each showed an approximatethree-fold reduction of Nb2 cell activity whereas the K69A variant (151%) exhibited abioactivity 1.5-fold greater than that of unmodified met-bPRL. The bioactivities of the variantscarrying alanine substitutions at Q71, Q73, and Q74 (48%, 45%, and 44%, respectively) weredetermined using a single extract of each mutant.87160140120100>..4-I>4-IC-,06040200Loop Region VariantsFigure 23. Bioactivities of met-bPRL variants obtained by single alanine substitutions of metbPRL residues within the loop region joining putative helices 1 and 2. The bioactivities weredetermined using the Nb2 cell assay and are expressed as a percentage relative to pituitary bPRL(average ±SEM). Three to five different extracts of each mutant (except one extract of Q74V)were assayed and each extract was assayed over a range of concentrations. No error bar is shownfor Q74V because only one extract of this variant was assayed for mitogenic activity. Forsimplicity, the mitogenic activities of H59S (23%), Q7 lÀ (48%), Q73A (45%), and Q74A (44%)are not included in the Figure.< < < < < < < < > > >O 0 w ç CD r’ C) ‘tIt) CD CD CD CD CD CD CD -= f- U) .J O LU S d 0’883.5.3 Substitution ofMet-bPRL Residues in Putative Helix 4The importance of residues within the fourth a-helix of hGH for binding to both thesomatogenic and lactogenic receptors has been established through a series of alanine-scanningmutagenesis studies (Cunningham and Wells, 1989, 1991) and confirmed by analysis of thecrystal structures for the hGH-(hGHbp)2complex (de Vos et at., 1992) and the hGH-hPRLbpcomplex (Somers et a!., 1994). Although the binding site of PRL to the lactogenic receptor isassumed to be very similar to that described for hGH, few studies have been carried out todetermine the involvement of helix 4 residues in PRLs in their interaction with.the receptor.In this study the contribution of residues in the C-terminal two-thirds of helix 4 to themitogenic activity of bPRL was investigated by alanine-scanning mutagenesis. Met-bPRLvariants, each containing a single substitution, were prepared and the mitogenic activity of eachvariant measured using the Nb2 cell bioassay system. Residues N170, L171, H173, L175, R176,S179, S180, D183, T184, Y185, L188, L189, R192, and 1193 were all replaced with alanine,D178 was replaced with glutamate and serine, and K181 was replaced by alanine, leucine, andarginine. Residues C174 and C191 were investigated in a previous study which showed that themitogenic activity of the C191S variant was equivalent to that of unmodified met-bPRL whileessentially no activity could be detected with the C174S variant (Luck and Huyer et at., 1992).However, the effect of this latter mutation is believed to be due to disruption of the C58 - C174disulfide bond which is critical for maintaining the three-dimensional structure of bPRL. WhileR177 is very important, K187 does not play a significant role in the mitogenic activity (section3.5.1 and Luck and Huyer eta!., 1991). Residues L172, 1182, L186, and N190 were notexamined because they are on the hydrophobic side of the putative helix and are therefore in theinterior of the hormone molecule and unlikely to interact with the receptor (Figure 24).All of the variants were produced as insoluble inclusion bodies in TOPP2 cells, solubiizedwith 0.1% sarcosyl, and renatured by air oxidation at pH 10. All the mutant hormones wereexpressed at levels similar to or greater than unmodified met-bPRL, except for those containingsubstitutions at positions D178 and Dl83 which were produced in much lower amounts.However, while the level of expression of D178A was so low that sufficient protein could not be89extracted for analysis in the Nb2 cell bioassay, enough of the more conservative mutants D178Eand D178S was produced to permit a determination of the contribution of this residue to themitogenic activity. Typical Western blots of some of the bPRL variants, made from gels rununder non-reducing conditions, are shown in Figures 25 and 26. Approximately 80% of eachmet-bPRL preparation had the same electrophoretic mobility as pituitary bPRL, indicatingalmost complete renaturation of the hormones.Figure 24. Helical wheel diagram for putative helix 4 of bPRL. The helix is stronglyamphipathic, as shown by the distribution of charged (filled symbols), polar but uncharged(shaded symbols), and hydrophobic (open symbols) residues.====E190Figure 25. Western blot analysis of sarcosyl extracts of met-bPRL and various met-bPRLmutants. The met-bPRLs were extracted with 0.1% sarcosyl from 0.4% sodium desoxycholatetreated insoluble material from transformed E. coil TOPP2 cells. The sarcosyl-soluble met-bPRLsamples were loaded on a nondenaturing polyacrylamide gel and analyzed by Western blottingusing a chemiluminescent detection system. Lane (1) unmodified met-bPRL; (2) T6OA; (3)E128A; (4) N17OA; and (5) L175A. The major band corresponds to oxidized hormone, the faintband slightly above the major band corresponds to the reduced form of the hormone and the veryfaint slow-moving bands are high molecular weight immunoreactive aggregates. The bPRLstandards are pituitary bPRL in the amounts indicated.bPRL Standard1234 81-Igh MW aggregates e a IbPRL (reduced) —bPRL (oxidized) — —Figure 26. Western blot analysis of sarcosyl extracts of various met-bPRL mutants. The metbPRLs were extracted with 0.1% sarcosyl from 0.4% sodium desoxycholate-treated insolublematerial from transformed E. coil TOPP2 cells. Samples were loaded on a nondenaturingpolyacrylamide gel and analyzed by Western blotting using a chemiluminescent detectionsystem. Lane (1) S179A; (2) D178S; (3) D178E; (4) R176A; and (5) S18OA. The major bandrepresents the oxidized, renatured form of the hormone, the faint band above the major bandcorresponds to the reduced form of the hormone and the very faint slow-moving bandscorrespond to high molecular weight immunoreactive aggregates. The bPRL standards arepituitary bPRL in the amounts indicated.bPRL Standard0 0 0 U)C) Il) CD0 012 3 45-IIugh MW aggregates‘ .IbPRL (reduced)bPRL (oxidized)91The mitogenic activities of each putative helix 4 met-bPRL variant are shown in Figure 27.Variants T184A (92%), L189A (84%), and R192A (94%) were essentially as bioactive asunmodified met-bPRL In addition, little effect resulted from the alanine replacements of H173(58%), L175 (58%), S179 (64%), S180 (53%), Y185 (66%), L188 (66%) and 1193 (62%). Anapproximate three-fold decrease in mitogenic activity was observed with the recombinantproteins L171A (42%), R176A (32%), and D183A (36%). The greatest effect on bioactivity wasobtained with the variants containing substitutions at positions 178 (serine and glutamatereplacements both gave mitogenic activities of 9%) and 181 (the alanine and leucine variants had5% and 4% activity, respectively). From these data it therefore appears that D178 and K181, inaddition to the previously identified R177 (section 3.5.1) play very important roles in themitogenic activity of met-bPRL. Of the other 11 residues examined only replacements of L17 1,R176, and D183 seem to affect the interaction of bPRL with the Nb2 receptor, albeit to a muchlesser extent than replacements of D178, R177, or K181.3.5.4 Substitution ofMet-bPRL Residues in the Putative Site 2 Binding SiteHuman GH contains two binding sites for the hGHbp: the hormone first binds a receptormolecule at site 1 and then a second hGHbp is bound at site 2 (Cunningham eta!., 1991; de Voset a!., 1992). The residues comprising site 2 are located at the N-terminus of the hGH moleculeand in the middle of the third helix. Although defmitive proof of receptor dimerization by othermembers of the growth hormone/prolactin/placental lactogen family has not been reported it isanticipated that this mechanism of receptor binding is common to this family of hormones.The present.study of bPRL, using the Nb2 cell assay, should be able to detect residuescritical for site 1 and site 2 function. Therefore, some residues equivalent to those which in hGHaffect site 2 binding were examined. Amino acid replacements were performed at R125, E128,and 0129 of bPRL. Unfortunately, while the E128A variant was synthesized in TOPP2 cells at alevel equal to or exceeding that of unmodified met-bPRL (Figure 18) the amount of R125Aproduced was so low that the contribution of this residue to the mitogenic activity of bPRL couldnot be determined. The bioactivity of E128A was 44% that of unmodified met-bPRL whilesubstitution of G129 with arginine resulted in a variant with no activity in the Nb2 cell assay.92140120100>..804bC-)040200Helix 4 VariantsFigure 27. Bioactivities of met-bPRL variants with single amino acid substitutions in putativehelix 4 residues. The bioactivities were determined using the Nb2 cell assay and are expressedas a percentage relative to the bioactivity of pituitary bPRL (average ±SEM). Three to sixdifferent extracts of each mutant were used and each extract was assayed over a range ofconcentrations.< << < <w < < < < <<< << <0 ,- ‘.D co 0 - c - u, a o C\IF— F— i.. 1-. r-.. r-. co co co co o— _ — e — r — — e- — , -z = _i C1, LI) S — >- JJ —93DISCUSSION4.1 Production of Bovine ProlactinMany different factors are involved in the successful intracellular production of soluble andactive proteins. In the bacterial and eukaryotic cell a complex process involving the protein,chaperones, co-chaperones, and foldases determines whether the protein will accumulate in thecytoplasm as the native structure or in aggregates (reviewed by Gilbert, 1994). However, manyfolding mechanisms exist and different proteins behave differently in each system: a set ofconditions allowing the production of one protein in native form will not necessarily result inproduction of another protein in its natural form. Further, since each protein requires its owncombination of chaperones and foldases, achieving this in heterologous systems is obviouslydifficult.In contrast, the production of proteins secreted in the periplasm of E. coli is fairly wellestablished as a means of generating recombinant proteins (reviewed in Blight et a!., 1994) andoffers many advantages over cytoplasmic production. The greatest advantage is that ofproducing a correctly folded and soluble protein. Efficient production and secretion of hGH intothe E. colt periplasm has been achieved by fusing the hGH gene to a signal sequence from an E.coli gene, in particular that from alkaline phosphatase (phoA) (Gray et at., 1985), that from heatstable-enterotoxin II (STU) (Chang et at., 1987; Chang et a!., 1989), or that from the outermembrane protein A (ompA) (Becker and Hsiung, 1986; Hsiung et al., 1986). The signalsequence aids in the translocation of proteins across the cytoplasmic membrane and cleavage bya signal peptidase then releases the mature protein into the periplasmic space (Inouye et a!.,1984). The hGH produced in these systems was processed properly, had formed the correctdisuffide bonds, and appeared to have a secondary structure identical to that of native hGH.Secretion to the periplasmic space of E. colt is thus generally more convenient thancytoplasmic production, particularly when the latter results in the formation of insolubleinclusion bodies. However, a system wherein the foreign protein is released to the culturemedium offers even more advantages. In addition to obtaining a correctly folded and soluble94protein the medium provides a larger volume for accumulation of the foreign protein andcontains few proteases which could degrade the protein. As well, the absence of many otherproteins in the medium simplifies the purification of the protein of interest. Two systems forsecretion of hGH to the culture medium of E. coli have been reported (Hsiung et a!., 1989; Katoet a!., 1987). In one system the hGH expression vector carried the weakly activated ku gene, thepenidilinase promoter, and the hGH gene fused to a signal sequence from an alkalophilicBacillus sp. The hGH was secreted to the periplasm and then released to the culture medium bythe action of the kit gene which permeabilized the outer membrane (Kato et a!., 1987). In theother system secretion of hGH to the culture medium of E. coli was obtained by refining asystem designed to export hGH to the periplasm (Hsiung et at., 1989). The gene for thebacteriocin release protein (BRP) was cloned into a vector and placed under the inducible controlof the lpp-lac promoter-operator system. BRP activates the detergent-resistant phospholipase Aof the outer membrane and thus causes the membranes to be permeabiized. Low levels ofexpression of the BRP, achieved by controlling the level of the inducer JPTG, resulted in thespecific release of proteins from the periplasniic space to the culture medium. Thus,transformation of E. coli with a vector containing hGH fused to the ompA signal sequence alongwith a BRP expression vector resulted in efficient secretion of hGH to the medium (Hsiung etat., 1989).As described above, there has been some success in producing soluble hGH from E. coli;however, in spite of the homologies between GH and PRL, PRL is much more commonlyproduced as insoluble inclusion bodies. Attempts in this study to obtain soluble, secreted bPRLfrom S. cerevisiae and from E. coli were unsuccessful: no hormone could be detected in the yeastcontaining the bPRL expression vector in which the bPRL gene was fused to the yeast invertasesignal sequence, and all bPRL produced in the E. coli cells was found associated with theinsoluble cell fraction. The reason for the lack of bPRL production in S. cerevisiae is unknown.The E. coli secretion system, derived from the dual vector system for hGH secretion describedabove, in which a vector containing the bPRL gene fused to the ompA signal sequence was cotransformed into E. coti with a vector for expression of the bacteriocin release protein, did not95produce soluble bPRL. This result is puzzling since this system was used successfully forsecretion of the related hormone hGH.Only a few examples of expression of biologically active PRL have been reported. Solublechicken PRL (chPRL) was produced in E. coil; however, although the level of expression ofchPRL was approximately 1.5% of total cell protein, 95% of the chPRL accumulated in inclusionbodies and only 2 - 3% could be isolated as the soluble and biologically active form (Hanks eta!., 1989). Efficient production of biologically active hPRL was obtained in E. coli using a high-expression vector containing a chimeric gene encoding a fusion of protein A and hPRL (Hiraokaet aL, 1991). The amount of hormone produced was 0.06 - 0.2% of total E. co/i protein;however, approximately 60% of the hPRL molecules were inactive, possibly as a result ofincorrect disulfide bond formation. In addition, the hPRL obtained after digestion withcollagenase to remove the protein A component had an additional four amino acids at the N-terminus, a tag resistant to collagenase action. Periplasmic secretion of a biologically activevariant of hPRL has recently been obtained (Morganti, L., Huyer, M., Gout, P.W. and Bartolini,P. 1995. Production and Characterization of Biologically ActiveAla-Ser-(His)6-l e Glu-Gly-Arg-Human Prolactin, Secreted in the Periplasmic Space of Escherichia co/i. Manuscriptsubmitted). Secretion of hPRL to the periplasm was obtained using a leader sequence from thecellulase gene of Celiulomonasfimi. The hPRL produced contained a 12 amino acid peptide tagat the N-terminus but had biological activity of the same order as pituitary hPRL Other reportsof bacterial production of PRL describe the accumulation of the hormone in inclusion bodies(Cunningham eta!., 1990; Gilbert et at., 1991; Luck eta!., 1986, 1989; Paris eta!., 1990).Production of soluble, biologically active, authentic PRL has not yet been reported. Although inanimals PRL is produced in soluble form and is naturally secreted, further work needs to becarried out to determine the conditions necessary to obtain soluble bacterially-produced PRL.As in the case for bPRL, the high-level expression of eukaryotic proteins in E. coli oftenresults in the formation of insoluble inclusion bodies which can accumulate to over 20% of thetotal E. coli cell protein (reviewed in Mitraki and King, 1989). Inclusion bodies are amorphousnon-native protein aggregates clearly separated from the rest of the cytoplasm, but not96surrounded by a membrane (Schoemaker et a!., 1985; Schoner et a!., 1985; Williams et a!., 1982)and they remain particulate after the cells are broken open. Inclusion body formation is believedto be due to the specific association of partially folded intermediates rather than to theprecipitation of native proteins synthesized at levels above their solubility. Thus, production ofthe aggregates is affected by environmental conditions such as temperature and by the presenceor absence of factors such as chaperones during the maturation process (Mitraki and King, 1989).Although the aggregates are not composed of correctly folded protein, there are positivefeatures to the production of protein in inclusion bodies. These advantages include the highlevels of protein expression generally obtained and the relative ease of isolating fairly pureprotein since the E. coil cytoplasmic proteins can be easily washed away. However, there arealso a number of disadvantages associated with inclusion bodies, the most notable amongst thesebeing the difficulty in converting the insoluble, aggregated heterologous protein to a soluble andactive form. A number of agents can be used to solubilize inclusion bodies including denaturantssuch as urea and guanidinium chloride, detergents, organic solvents, and extremes of pH(Marston, 1987). These agents disrupt the ionic and/or hydrophobic interactions between thepolypeptide chains and result in denaturation of the protein. Thus, a renaturation step isnecessary in order to obtain active protein.The procedure for extracting met-bPRL from the insoluble fraction of E. coil cellstransformed with the bPRL expression vector involves mechanical lysis of the cells using aFrench pressure cell, centrifugation to harvest the insoluble met-bPRL, treatment with sodiumdesoxycholate to solubiize some contaminating E. coil proteins, and extraction with sarcosyl tosolubiize the met-bPRL. Fully active met-bPRL is not obtained from this procedure (Luck eta!., 1989); therefore, the sarcosyl-soluble met-bPRL was renatured by air oxidation at pH 10(Luck and Huyer et a!., 1992). The renaturation was assessed by Western blot analysis of thesamples run under nondenaturing conditions: the oxidized form of bPRL, containing the C58 -C174 disulfide necessary for biological activity, is more compact than the reduced form; thus,these two forms can be distinguished on a nondenaturing polyacrylamide geL Although themajority of the met-bPRL obtained by this procedure is in the monomeric and oxidized form, a97portion of the met-bPRL is usually present as high molecular weight aggregates, presumablyconsisting of the hormone met-bPRL covalently linked via disulfide bonds to itself or to thiolcontaining E. coli proteins (Luck and Huyer et al., 1992).In previous studies, E. coil HB2151 was used as the host strain for the met-bPRLexpression vector pESP4 and transformed cells were grown in YT medium (Luck et al., 1989,1990, 1991, 1992). In this study, in an attempt to optimize production of met-bPRL, a number ofdifferent E. coil strains were transformed with pESP4 and cultures were grown in terrific broth, aricher medium than YT. Production of met-bPRL varied greatly among different E. coli strains,ranging from very little or no production in strains JM1O5 and CSH5O to high levels ofproduction in strains Le392F’, NM522, and TOPP2 (Figures 14, 15, and 16). Although the totalamount of hormone synthesized in the TOPP2 strain was not maximal, this strain was chosen forroutine production because the amount of high molecular weight met-bPRL aggregates presentwas lower in TOPP2 cell extracts than in those from any of the other E. coil strains tested(Figures 15 and 16). The use of 0.1% sarcosyl, rather than the 0.2% sarcosyl used in earlierstudies, to extract the met-bPRL from the E. coil extracts was also found to decrease theproportion of the high molecular weight aggregates in the met-bPRL samples and 0.1% sarcosylwas therefore used for routine extraction.E. coil expression systems are often chosen for the production of recombinant proteinssince, in addition to the ease of manipulating E. coil, the vast amount of knowledge that has beenobtained about this organism has led to the engineering of a number of different vectors andstrains which can be used to achieve efficient and controlled synthesis of almost any geneproduct (Das, 1990). However, there are disadvantages to expression of heterologous proteins inE. coli. Some eukaryotic proteins produced in E. coil are toxic to the cell and are degraded onsynthesis; others could only be produced as fusion proteins of which none could be processedappropriately. In addition, a common consequence of high levels of expression is theaggregation of the protein in insoluble inclusion bodies. Although the latter problem is notlimited to E. coil, eukaryotic expression systems can be used to circumvent some àf the problemsencountered with the bacterial systems (Bradley, 1990).98Eukaryotic expression systems used successfully for the production of eukaryotic proteinsinclude expression in yeast, in mammalian cells, and in insect cells using baculovirus vectors(Ausubel et a!., 1994; Bradley, 1990). Advantages of these expression systems are that therecombinant proteins are generally properly processed and usually end up in the correct cellularcompartment. Disadvantages of these systems include greater complexitiy, cost, and longerprocessing times. However, in spite of these drawbacks the use of eukaryotic expression systemsfor the production of eukaryotic proteins is becoming more common and more popular, and asthe knowledge about, and familiarity with these systems grows, the common problems becomeeasier to handle. At the time that the expression system for bPRL was developed (Luck et a!.,1986) E. coli systems were by far the preferred routes for production of recombinant proteins.Indeed, bacterial systems, as well as yeast systems, are still often preferred when the intention isto create and analyze many different mutants. However, given the problems that have beenencountered in the production of active met-bPRL from E. coli and the advances that have beenmade in the development of eukaryotic expression systems, investigation of the latter systems forthe synthesis of bPRL would likely be worthwhile.4.2 Structure-Function Analysis of Met-bPRLTherelationship between the primary structure and biological activity of bPRL wasinvestigated by determining the effects of specific amino acid changes on the mitogenic activityof the hormone. All amino acid changes were produced by site-directed mutagenesis, and, forthe most part, involved the substitution by alanine of residues corresponding to those whicheither corresponded directly to those involved in receptor interactions in related hormones orwhich fell within the specific regions of the hormone that, based on structural homologies withhGH, are postulated to play a role in contacting the receptor.The mitogenic activity of bPRL was significantly decreased by replacement of the arginineat position 177 with any one of 12 different amino acids, suggesting that this residue is veryimportant for the biological activity of the hormone. The reduced mitogenic activities resultingfrom replacement of residue R21 suggests that this residue, whilst not essential, may also be99involved in the activity of bPRL. However, since substitution of this arginine with thefunctionally similar residue lysine had little effect on the mitogenic activity of bPRL (activity ofR21K was 87% that of unmodified met-bPRL), the activity of the hormone is not as dependenton an arginine at position 21 as it is on one at position 177. In contrast, the residue K187 doesnot appear to play an important role in the mitogenic activity of the hormone: no substitutions atthis position severely affected the bioactivity of bPRL.A study of the lactogenic hormone mPL-II had revealed that three lactogen-specificresidues (R14, R169, and K179) play significant roles in the mitogenic activity of this hormonein the Nb2 cell bioassay system and in the binding of mPL-II to the Nb2 PRL receptor (Davisand Linzer, 1989b). However, an examination of the roles of the corresponding bPRL residues(R2 1, R177, and Ki 87) in the mitogenic activity of this hormone revealed differences frommPL-ll in its response to mutational changes at these positions (Luck and Huyer et al., 1991).Whereas all of the mPL-II variants (R14L, R167N, and K179N) had mitogenic activities in theNb2 cell assay that were 20- to 30-fold lower relative to wild-type mPL-ll; of the correspondingbPRL variants (R21L, R177N, and K187N) only one, R177N, had very low mitogenic activity(2.4% relative to unmodified met-bPRL). The substitution at position 21 resulted in a muchlesser decrease in mitogenic activity (29% for R2 1L), and the mitogenic activity of variantK187N (87%) was close to that of unmodified met-bPRL (Table 4). Other substitutions at thesepositions resulted in somewhat similar decreases in mitogenic activity, indicating that of thethree residues found to have important roles in the mitogenic activity of mPL-II, only the bPRLresidue corresponding to R167 of mPL-ll (namely R177) plays an equally important role in themitogenic activity of met-bPRL. Interestingly, the hGH residues equivalent to R177 and K187of bPRL (K168 and R178, respectively) have both been shown to be involved in binding hGH tothe hPRLbp (Cunningham and Wells, 1991) while of these two only R178 is required for bindingto the hGHbp (Cunningham and Wells, 1989). These series of studies, on mPL-ll, hGH, andbPRL, demonstrates that one cannot generalize about features of a family of proteins with thesame biological activity from a study of one member.100None of the substitutions used in the ioop region connecting putative helices 1 and 2 of thebPRL molecule led to a severe reduction in mitogenic activity (Figure 23), suggesting that theseresidues are not critically important for the interaction of the hormone with its receptor. Of thepositions examined, replacement of residues H59 (with alanine and serine) and Q71, Q73, andQ74 (with valine) did result in a greater than 60% decrease in bioactivity; therefore, these aminoacids may have a subsidiary role in the mitogenic activity of met-bPRL. Interestingly, thesubstitution of alanine for lysine at position 69 resulted in a more active met-bPRL molecule,suggesting that removal of the lysine side chain increases the ability of the hormone to interactwith the receptor. Therefore, in bPRL this lysine residue could act to modulate receptor binding.The sequences of the second half of the loop region joining putative helices 1 and 2 showstrong homology across the growth hormone/prolactin/placental lactogen family (Figure 22; alsosee Goffin et al., 1992 for a more complete list of sequences in this region) and it has beensuggested that this region may be important for the activity of all of the hormones in this family(Goffin et al., 1992). Residues in this region have been shown to be involved in the binding ofhGH to the hGH and hPRL receptors (Cunningham and Wells, 1989, 1991) and in the binding ofhPRL to the Nb2 PRL receptor as well as its mitogenic activity in the Nb2 cell assay (Goffin etal., 1992). The homology between bPRL and hPRL is particularly strong in this region: the onlydifference is position 64 (proline in bPRL, alanine in hPRL). However, in spite of the almostidentical sequences in this region there are significant differences in the contribution ofequivalent residues to the mitogenic activities on Nb2 cells of hPRL and bPRL (Table 5).The role of individual amino acids in the 58 -74 region of hPRL to the activity of thishormone was investigated in an alanine-mutagenesis study (Goffin et al., 1992) and while thebioactivities of the hPRL variants T6OA, L63A, E67A were very similar to the correspondingmet-bPRL variants, differences in activity were observed at the other positions (Table 5). Mostnotably, the hPRL variants S61A and Q74A showed increased levels of mitogenic activity(122% and 171%, respectively) while the activities of the met-bPRL variants (S6 1A and Q74V)were lower than that of unmodified met-bPRL (71% and 36%, respectively). A very strikingdifference between the two hormones was seen with the K69A variant: in hPRL this substitution101essentially abolished Nb2 cell bioactivity whereas in bPRL the mitogenic activity increased 1.5-fold. No bioactivity values were reported for hPRL variants carrying substitutions at positionsQ71 orQ73.TableS. Comparison of the Effect of Single Amino Acid Substitutionsin the Loop Region Joining Putative Helices 1 and 2 on the Bioactivity ofbPRL, on the Bioactivity and Binding of hPRL to the Nb2 PRL Receptor,and on Binding of hGH to the hPRL and hGH ReceptorsbPRL ppjb hGHCResiduea % Bio- % Bio- Binding Binding Ability!b, hPRL hGH activityd activity” Abiitye hPRLbp hGHbpH59 F54 38 58 2.4 1.4 4.4T60 S55 72 59 1.03 1.5 1.2S61 E56 71 122 0.79 0.8 4.1S62 S57 77g 1 1.4L63 158 51 62 1.47 18 1764k P59 61 1.3 1.9T65 T60 98P66 P61 70 25 3.3 8’E67 S62 90 90 1.06 11 2.8D68 N63 87g 4.3 3.3K69 R64 151 1 12.5 1.8 21E70 E65 87g 0.9 2.5 0.59Q71 E66 35 1.1 2.1Q73 Q68 38 1.2 5.2Q74 Q69 36 171 0.76 0.7 0.91a Position of corresponding residues in bPRL, hPRL, and hGH. Variantsat these positions are all single alanine substitutions except residues 62, 68,and 70 of bPRL which were mutated to threonine, glycine, and asparagine,respectively.b Data taken from Goffin et a!., 1992.C Data for all variants except P61A taken from Cunningham and Wells, 1991.d Bioactivity refers to mitogenic activity in the Nb2 cell bioassay and isexpressed as a percentage relative to unmodified bPRL or hPRL.Binding ability of hPRL to the Nb2 receptor is expressed as the ratioIC50 (mutant)/1C50(hPRL). 1C50 is the concentration at which each proteincaused half-maximal displacement of1I-labe1ed hPRL from the Nb2receptor in Nb2 cell homogenates. Higher values indicate decreased abilityto bind to the Nb2 receptor.f Binding ability of hGH mutants for the extracellular part of the hPRLbp isexpressed as Kd (mutant)/Kd(hGH). The dissociation constant Kd wasdetermined for each mutant by competitive displacement of125-labeledhGH from the hPRLbp in the presence of 50 mM ZnCl2. Higher numbersindicate decreased ability to bind to the hPRLbp.g Data for these variants have been taken from Luck et at., 1989.h Residue 64 is a proline in bPRL and an alanine in hPRL.The binding affmity of this variant is taken from Sakal et at., 1993.102h Residue 64 is a proline in bPRL and an alanine in hPRL.The binding affinity of this variant is taken from Sakal et at., 1993.The homology between the sequences of bPRL and hGH in the second half of the loopregion is not as strong as that between bPRL and hPRL. Although most of the differences areconservative (such as serine/threonine, isoleucine/leucine, arginine/lysine, and glutamate!glutamine), a number of the differences between the sequences of hGH and bPRL are marked,i.e., at positions 59, 61, 67, and 72 of bPRL (residues 54, 56, 62, and 67 of hGH). Interestingly,when compared across the hormone family the most divergent residues in this portion of the loopregion are found in the first three of these positions. The binding affinity for the hPRL receptorof hGH variants containing alanine substitutions in the loop region has been examined(Cunningham and Wells, 1991; Sakal et at., 1993) (Table 5). Zn2 was required for binding ofall hGH variants to the hPRLbp, a situation unique amongst the hormone-receptor interactionsdiscussed here. Human GH variants 158A, P61A, S62A, and N63A (corresponding to L63, P66,E67, and D68 of bPRL) each exhibited decreased binding to the lactogenic receptor. In addition,the hGH variant 158A had decreased binding affinity for the hGHbp. The other loop regionamino acids which in hGH play a role in contacting the somatogenic receptor are F54, E56, R64,and Q68 (corresponding to H59, S61, K69, and Q73 of bPRL) (Cunningham and Wells, 1989).Although in the present study the binding affinity of the met-bPRL variants to the Nb2receptor was not determined, it was observed that none of the variants were able to inhibit theaction of unmodified met-bPRL on the Nb2 cells (data not shown) indicating a possiblecorrelation between low mitogenic activity and reduced receptor binding ability. Indeed, thehPRL variants least able to stimulate the growth of the Nb2 cells, namely P66A and K69A, werealso the most affected in their ability to bind to the Nb2 PRL receptor while the variants S61Aand Q74A showed an increase in both mitogenic activity and binding affinity (Table 5 andGoffin et at., 1992).The differences between the contributions of equivalent residues within this highlyconserved region of these related hormones again illustrates that although the same generalregion of each hormone may be involved in binding to the receptor molecules, the functionalresidues can be different in each case. Certain residues may be universally important:103abolished mitogenic activity in bPRL (Luck and Huyer, et a!., 1992), hPRL (Goffin et a!, 1992),and mPL-ll (Davis and Lmzer, 1989a). Substitution of this cysteine residue would result indisruption of the disulfide; thus, the importance of this amino acid to biological activity is mostlikely due to its role in stabilizing the conformation of the loop joining helices 1 and 2 withrespect to the three-dimensional structure of the hormone. In addition, disruption of the disulfidebond likely would prevent the various residues, which together form the binding site, fromattaining the conformation required for hormone-receptor binding. However, in contrast to thegeneral importance of the cysteine residue to biological activity, the contribution of the otheruniversally conserved residue, P66 (numbering according to bPRL), varies between hormones.For both hPRL and hGH this residue is important for activity: substitution of this amino acidwith alanine decreased the mitogenic activity of hPRL by 4-fold and the binding to the Nb2 PRLreceptor by more than 3-fold (Goffin et a!., 1992) and decreased hGH binding to the hGHbp by8-fold (Sakal et a!., 1993). However, the P66A variant of bPRL did not show a significantdecrease in mitogenic activity (Figure 23). It is also very interesting that in spite of the almosttotal conservation of sequence in the ioop region between hPRL and bPRL, the role of residuesK69 and Q74 in the mitogenic activity of each hormone is very different: in hPRL K69A hasalmost no activity while in bPRL this variant has increased activity; and in hPRL Q74A hasincreased activity while in bPRL the mitogenic activity is reduced 3-fold. It is clear, therefore,that differences exist not only at the residue level between binding of PRLs and hGH to thelactogenic receptor, which might be anticipated because of the role of Zn2 in the activity of thelatter, but also in the binding of different, but closely related, PRLs.In general, the contributions of each of the ioop region residues to the biological activitiesand/or receptor-binding ability of bPRL, hPRL, and hGH are quite different (Table 5). None ofthe residues are crucial to activity in all the hormones although replacement of H59 (P54 ofhGH) does have a slightly deleterious effect on the mitogenic activity and/or binding ability ofeach of these hormones. Similarly, while there are ioop region residues which are not involvedin the activity of the individual hormones, there is no one residue in this region which is notinvolved in activity of any of these hormones except possibly T60 (S55 of hGH) which appears104to have a universally minor role. Some residues are required only for binding of a hormone toone of the receptors studied [e.g., E56 of hGH (for binding to the hGHbp) and S62 of hGH (forbinding to the hPRLbp)] while others play roles in more than one set of activities [e.g., 158 ofhGH (for binding to the hGH and hPRL receptors)] and P66 of hPRL/P61 of hGH (for binding tothe Nb2 PRL receptor and hGHbp, respectively)]. These results ifiustrate that while the generalregion for receptor binding may be the same for members of the hormone family the identities ofthe specific residues involved are often quite different. These differences likely contribute to thespecificity of the interactions between these related hormones and their receptors.Of the many putative helix 4 residues examined, only replacements of R177 (discussedabove), D178, and K181 (Figure 27) severely affected the mitogenic activity of bPRL.Substitution of the aspartate at position 178 with either serine or glutamate (similar to aspartatein terms of both size and charge) gave variants with bioactivities of 9%, suggesting that thisresidue is very important for the mitogenic activity of bPRL. The bioactivities of the variantsKi 8 1A and Ki8 1L were both approximately 5% that of unmodified met-bPRL, and thebioactivity of K181R was 63%. Thus, while K181 appears to play a very important role in themitogenic activity of bPRL, the requirement for a lysine at that position is not absolute sincesubstitution with the functionally similar residue arginine gave rise to a variant with close to two-thirds the activity of the unmodified hormone. It is possible that the interaction between bPRLand the Nb2 PRL receptor is enhanced by a charge-charge interaction involving the lysine atposition 181. Alternatively, the positive charge on the lysine could be required to stabilize thebPRL structure via a hydrogen bond: the corresponding residue in hGH, K172, is important forthe interactions between hGH and the hGH and hPRL receptors, but only the aliphatic portion ofthe side chain contacts the receptor molecules, the charged portion is involved in forming ahydrogen bond to D169 in the core of the hormone (Cunningham and Wells, 1993; de Vos eta!.,1992). In this regard, it may be noted that D178 of bPRL corresponds to D169 of hGH;therefore, it is possible that the role of D178 in the activity of bPRL may be a structural one• rather than one of directly contacting the receptor.105Table 6. Comparison of the Effect of Single Amino AcidSubstitutions in Putative Helix 4 on the Bioactivity of bPRLand on the Binding of hGH to the hPRL and hGH ReceptorshGHaResidueb bPRL Binding AbilitycbPRL hGH % Bioactivity” hPRLbp hGHbpN170 G161e 122L171 L162e 42H173 Y164 58 2.1 3.6L175 F166e 58R176 R167 32 770 0.75R177 K168 1 18 1.1D178g D169f 9S179 M170f 64S180 D171 53 1.1 7.1K181 K172 5 220 14D183 E174 36 356 0.22T184 T175h 92 2.3 3.5Y185 F176 66 25 16K187 R178 46 7 58L188 1179 66 1.8 2.7L189 V180 84 0.6 1R192 R183 94 2.6 2.11193 S184 62 0.8 0.911194 V185 1.6 4.5a Data are taken from Cunningham and Wells, 1991.b Identity and position of corresponding residues in bPRL andhGH. All were replaced with alanine unless otherwise indicated.C Binding ability of hGH to the extracellular portions of thehPRLbp and hGHbp is expressed as Kd (mutant)/Kd(hGH). Thedissociation constant Kd was determined for each mutant bycompetitive displacement of1I-labe1ed hGH from the hPRLbp(in the presence of 50 IIM ZnC12)or the hGHbp.d % bioactivity is the mitogenic activity in the Nb2 cell bioassayand is expressed relative to wild-type bPRL (100%).e No replacements of G161, L162, and F166 of hGH were reported.f Not enough of D169A and M17OA of hGH were produced inE. coil for analysis of receptor-binding abilities.g D178 of bPRL was replaced with glutamate (D178E).h T175 of hGH was replaced with serine (T175S).106Alanine-scanning mutagenesis of hGH revealed that helix 4 plays a very important role inthe interactions between this hormone and the human growth hormone and prolactin receptors(Cunningham and Wells, 1989, 1991). However, inasmuch as detailed studies of thecontribution of this region to the biological activities of related hormones has not yet beencarried out, the general importance of this region of helix 4 in hormone-receptor interactions hasnot been established. In hGH helix 4 amino acids are involved in binding to both the hGHbp andthe hPRLbp, and although the binding sites are not identical they do overlap. A comparisonbetween the contribution of helix 4 residues to the activities of hGH and bPRL is given in Table6. As was noted for the residues in the ioop region joining helices 1 and 2, the residues in helix 4have a variety of roles in binding to and activating the different receptor molecules.Thus, in hGH, residues K172, P176, and R178 interact with both the hGH and hPRLreceptors; D171 and Vl85 are only involved in binding to the hGHbp; and R167, K168, andE174 are only required for binding to the hPRLbp (Cunningham andWells, 1991). The verylarge reduction in the binding ability of hGH to the hPRLbp for the variant E174A is a result ofdisruption of theZn2-binding site. The corresponding residues in bPRL and hGH which haveimportant roles in the activities of both these hormone are R176, R177 and K181 of bPRL(R167, K168 and K172 of hGH): the hGH variants R167A, K168A and K172A exhibitedreduced affinity for the hPRLbp but not for the hGHbp. Marked differences between thehormones are seen at positions Tl84, Y185, K187, and R192 of bPRL: replacements of thecorresponding hGH residues (T’175, Fl76, R178, and R183) disrupted binding to both the hGHand hPRL receptors while replacements of the bPRL had little or no effect on the ability of thehormone to activate the Nb2 PRL receptor. Thus, while there appear to be some similaritiesbetween the binding sites of hGH for the hPRLbp and of bPRL for the Nb2 PRL receptor thereare also some significant differences.107Figure 28. Ribbon representation of the putative structure of bPRL indicating the binding sitefor activation of the Nb2 PRL receptor. The model does not contain the first 14 N-terminalamino acids or the C4 - Cli disulfide. The truncated N-terminus is located in the upper right-hand corner and the C-terminus, joined in a disulfide bond (coloured greenish-yellow) to acysteine near the C-terminus (C191 - C199) is positioned just below the N-terminus. The sidechains indicated are those which in the present study were identified as being involved in themitogenic activity of bPRL Substitutions of these amino acids, primarily with alanine, resultedin a reduction in mitogenic activity. The disulfide bondjoining the loop region between helices1 and 2 to the cysteine within helix 4 is also shown (C58 - C174) (coloured greenish-yellow) andit located in the middle bottom of the representation of the hormone structure.1084.3 Residues Responsible for the Bioactivity ofMet-bPRL in the Nb2 AssaySystematic replacement of residues within the ioop region joining putative helices 1 and 2and within the C-terminal two thirds of helix 4 has identified a number of residues which play arole in the mitogenic activity of bPRL. These include H59, Q71, Q73, Q74, L171, R176, R177,D178, K181, and D183. When mapped onto a putative structural model of bPRL, obtained bymolecular modeling of bPRL from the structure of hGH, these residues form a patch covering themiddle of helix 4 and the latter third of the ioop region (Figure 28). The general location of theseamino acids is very similar to that identified as being involved in the interaction between hGHand the hPRLbp (Figure 29 and Cunningham and Wells, 1991). Indeed, as discussed above, thesimilarity between the binding sites extends beyond general location to certain specific residues:most notably, the side chains of the equivalent residue pairs R177 and K181 of bPRL and K168and K172 of hGH are similarly positioned for interacting with the receptors.The putative molecular model of bPRL can be used to shed light on the contribution ofparticular amino acids to the mitogenic activity of the hormone. Replacements of any of residuesH59, L171, and D178 resulted in a reduction in mitogenic activity; however, the side chains ofL171 and D178 are on the interior-facing sides of helices (Figure 30). The internal location ofthe side chains of L171 and D178 was predicted by the helical wheel diagram (Figure 24).Therefore, the role of these residues is most likely a structural one rather than one of receptorbinding. Replacements of these residues may distort the protein structure and indirectly causereductions in mitogenic activity. Interestingly, L171 and D178 correspond to the hGH residuesL162 and D169 which are found in hydrophobic clusters and therefore likely have structuralroles: L162 is part of a group of hydrophobic residues from the region between helices 2 and 3,and from helices 1 and 4; while D169 is part of the core of the four-helix bundle (de Vos eta!.,1992). Other amino acids within the binding site which may also support the general structureare L63, P64, and P66 in the loop region joining helices 1 and 2. L63 corresponds to 158 ofhGH, and although alanine scanning mutagenesis identified this hGH residue as being critical forinteraction with both the hGHbp and the hPRLbp (Cunningham and Wells, 1991), the crystalstructure of the hormone-receptor complex revealed that 158 is part of a hydrophobic cluster and109does not directly contact the receptor molecule (de Vos et a!., 1992). L63 is, like 158, ahydrophobic residue and in the putative structural model of bPRL its side chain points into thehormone molecule. The side chains of P64 and P66 are similarly positioned in that they alsoface into the interior of the protein (Figure 30). However, since replacing L63 or either of theproline residues with alanine did not cause a major reduction in mitogenicactivity (5 1%, 61%,and 70%, respectively), individually the contributions of these amino acids to the maintenance ofthe structure of the binding site does not appear to be very great.The functional roles of K69, Q7 1, Q73, and Q74 can also be inferred from the putativemodel of bPRL. The side chains of all of these amino acids appear to be very solvent-accessibleand are therefore well positioned for interacting with a receptor molecule and appear less likelyto have structural roles (Figure 28). The mitogenic activity of variants in which any of thesethree glutamine residues was replaced by valine or alanine was approximately 3-fold less thanthat of unmodified met-bPRL (Figure 23 and Table 5). It is possible that the space vacated bythe replaced glutamate could be filled by one of the two glutamate residues remaining and thatthis movement of side chains might mitigate the effect of replacement of any one of theglutamates in this region. Therefore, while the observed reduction in bioactivity from theindividual replaäement of these residues is not particularly large it does raise the possibility thatthese amino acids may play a role in interacting with the Nb2 PRL receptor. On the other hand,replacement of the lysine at position 69 with an alanine resulted in an increase in mitogenicactivity, to 1.5-fold greater than that of unmodified met-bPRL (Figure 23 and Table 5). Thislysine residue may, by virtue of its large side chain, physically hinder slightly the interactionbetween bPRL and the Nb2 PRL receptor: contact between these proteins may be improved bythe replacement of K69 with an amino acid that has a less bulky side chain. Alternatively, it ispossible that there may be a repulsion effect between bPRL and the Nb2 PRL receptor as a resultof the positive charge on the lysine residue. In this case, replacement of the lysine with anaspartate or glutamate may result in a variant with even greater activity.110Figure 29. Ribbon representation of the structure of hGH indicating residues within the site 1and site 2 binding regions for the hPRLbp. The N-terminus is located in the upper right-handcorner. A portion of the C-terminus could not be modelled; therefore the structure shown lacksthe terminal 4 amino acids and the Cl82 - C189 disulfide. The truncated C-terminus is locatedjust below the N-terminus in the middle of the right-hand side of the figure. The residuesindicated are those sites where alanine substitutions resulted in significant reductions in site 1binding affmity to the hPRLbp (H18, H2l, F25, 158, S62, N63, R167, K168, Kl72, E174, T175,and R178). The role of H18 and E174 is coordinating the Zn2molecule required for bindinghGH to the hPRLbp. Alanine substitutions of D116 and E119 decreased receptor binding at site2. Also indicated are residues within the site 1 binding region for the hPRLbp which areinvolvedin binding hGH to the hGHbp (E56, 158, R64, K172, F176, R178). The complete set ofalanine mutants that reduced binding to the hGHbp, comprising residues outside the binding sitefor the hPRLbp, is not shown. The disulfide joining the loop region between heices 1 and 2 tohelix 4 (C53 - C165) is indicated as a yellow tube.6.:.illFigure 30. Ribbon representation of the putative structure of bPRL indicating residues withinthe binding site for the Nb2 PRL receptor. The truncated N-terminus lacking the C4 - Clidisulfide is in the upper right-hand corner and the C-terminus, part of the C191 - C199 disulfide(coloured yellow) is located just below the N-terminus in the middle of the right-hand side of thefigure. The C58 - C174 disulflde is indicated as a shadowed greenish-yellow tube in the bottommiddle of the figure. Residues indicated include those which are involved in activating the Nb2PRL receptor (K69, Q71, Q73, Q74, R176, R177, K181, and D183), those whose replacementdecreased the mitogenic activity of bPRL but which most likely play a structural role (H59,L171, and D178) and those whose replacement had little effect on activity (R21, L63, P64, P66,S180, and K187). Residues R176, D178, and D183 are not labeled.112The hGH residues which correspond in general location to K69, Q71, Q73, and Q74 ofbPRL are N63, R64, E65, and E66 (Figures 28 and 29). Replacements of these residues alteredbinding to the hPRLbp and/or to the hGHbp (Cunningham and Wells, 1991; Table 5). The roleof R64, E65, and E66 is likely related to the charged nature of their side chains: arginine ispositively charged and glutamate is negatively charged. The E65A hGH variant exhibitedenhanced binding affmity for the hGHbp and a nearly two-fold increase in on-rate of thehormone to the receptor (Cunningham and Wells, 1993) but decreased affinity for the hPRLbp.In contrast, the E66A hGH variant exhibited decreased affinity for the hGHbp with no change inthe affinity for the hPRLbp (Cunningham andWells, 1991). These electrostatic interactions aswell as that involving the positive charge on the arginine at position 64 are believed to beimportant for modulating the association of hGH with the receptor molecules (Cunningham andWells, 1993). These results suggest that the positive charge on K69 in bPRL, rather than the sizeof its side chain, is likely important in the interaction between the hormone and the Nb2 PRLreceptor.The putative position of residues R21 and K187 in bPRL, of which the correspondingresidues in mPL-ll (namely R14 and K179) play important roles in mitogenic activity, are on theedge of the patch of amino acids in helix 4 and the loop region which together seem to form thebinding determinant for the Nb2 PRL receptor (Figure 30). The side chains of both these aminoacids appear to be located on the exterior of the bPRL molecule and are therefore in a favourableposition for interacting with the receptor. Non-conservative substitutions of both R21 and K187resulted in minor reductions in mitogenic activity, with the effect of the R21 substitutions beingslightly greater (Table 4). Consideration of the putative location of these residues along with theapparently minor roles that they play in the mitogenic activity of bPRL suggests that they arelocated on the edge of the binding site for the Nb2 PRL receptor. Thus, although there is a greatdeal of sequence homology between bPRL and mPL-ll within the portion of putative helix 4,which in bPRL contains the binding determinants for the Nb2 PRL receptor (Figure 22), thebinding sites for these related hormones to the same receptor molecule are obviously slightlydifferent.113The requirement for Zn2 in the formation of the hGH-hPRLbp complex provides a furthercomplexity to an analysis of lactogen-PRLbp interaction. Zn2 is required for binding of hGH tothe hPRLbp but not for the interaction of either hGH with the hGHbp or hPRL with the hPRLbp(Cunningham eta!., 1990a). The hGH ligands which coordinate the Zn2 are H18 and E174. Inaddition, although H21 does not bind the metal ion it is required to correctly orient the side chainof E174 (Somers et aL, 1994). Throughout the family of hormones all three of these residues arehighly conserved, particularly the two histidines (Figure 22). However, although in bPRL theresidue equivalent to E174 of hGH (namely D183) does appear to play a role in the mitogenicactivity, neither of the histidine residues is important and Zn2 does not appear to be required forbinding bPRL to the Nb2 PRL receptor (Luck et at., 1990; Dr. Luck, Oberlin College, Oberlin,OH, personal communication). Indeed, the only members of the hormone family for whichZn2-mediated binding to a receptor been reported are hGH and hPL: both require Zn2 to bindthe hPRLbp (Lowman eta!., 1991). Therefore, the role played by D183 of bPRL in activatingthe Nb2 PRL receptor is quite different from the role of the equivalent hGH residue, E174.4.4 Receptor DimerizationHuman GH contains two binding sites for the hGHbp: the hormone first binds a receptormolecule at site 1 and then a second hGHbp is bound at site 2 (Figure 31) (Cunningham et at.,1991; de Vos et a!., 1992). The residues comprising site 1 form a patch covering sections ofhelices 1 and 4 and part of the ioop region joining helices 1 and 2, while site 2 residues arelocated at the N-terminus of the hGH molecule and in the middle of the third helix. Receptoroligomerization is a common mechanism for signal transduction (Ulirich and Schlessinger, 1990)but the interaction between hGH and the hGHbp is unique in exhibiting a stoichiometry of onehormone to two receptors. There is a great deal of homology between members of the growthhormone family and it is likely that this extends to a common mechanism of receptor binding.Indeed, there is evidence indicating that binding of PRL to the PRLbp also occurs through a two-site mechanism (Elberg eta!., 1990; Gertler eta!., 1993; Goffin eta!., 1994; Shiu et a!., 1983)although definitive proof of whether this indeed occurs has not yet been obtained.114Figure 31. Backbone structure of the hGH-(hGHbp)2complex. In this orientation the C-terminiof the receptor molecules are at the bottom of the figure. The hGH molecule is shown withyellow cylinders representing the helices joined by red tubes. The loops of the first hGHbp arecoloured green while those of the second receptor are blue. For both receptor molecules the B-strands are shown in red. The figure is taken from de Vos et aL, 1992.Evidence that bPRL contains a binding site analogous to site 2 of hGH was obtained froman examination of the glycine at position 129. The G129R variant of bPRL was found to haveabsolutely no mitogenic activity in the Nb2 lymphoma cell system. This result strongly suggeststhat the introduction of a charged, bulky side chain at position 129 blocks the formation of aproductive interaction between bPRL and the Nb2 PRL receptor. The corresponding residue inhGH is G120, and substitution of this amino acid with arginine, with its charged, bulky sidechain, resulted in a mutant which was able to bind the hGHbp at site 1 but which blocked bindingat site 2 (Fuh et a!., 1992). This variant was a potent antagonist of hGH-induced cellproliferation: binding of the mutant prevented receptor dimerization and the concomitantmechanisms of signal transduction. Similarly, replacement of G129 of hPRL (equivalent toG120 of hGH) reduced the binding affinity and the Nb2 cell bioactivity by two to three orders of115magnitude (Goffin et a!., 1994) In addition, transgenic mice expressing bGH or hGH analogswith mutations in the third a-helix, including residue Gl20, exhibited a dwarf phenotype (Chenet a!., 1991a, 1994, 1995). Based on the homology which exists between the binding of hGH tothe hGH and hPRL receptors, it is likely that the effect of the G129R mutation in bPRL resultsfrom interference with site 2 binding. A model of bPRL indicating the putative locations of site1 and site 2 is shown in Figure 32.The site 2 residues, as identified in hGH and bGH, are predicted to form a cleft in the thirda-helix with the glycine (position 120 in hGH and 119 in bGH) located at the centre and flankedby larger residues such as aspartate and leucine (Dl 16 and L124 of hGH, and Dl 15 and Ll23 ofbGH) which may interact with the receptor molecules (Chen et at., 1995). That a small residue issatisfactory for function was confirmed by the creation of transgenic mice expressing a G12OAvariant of hGH: these mice did not exhibit a growth-suppressed phenotype. Therefore, theactivity of the hormone likely depends on the presence of a small amino acid at that locus andnot on the glycine per se (Chen et a!., 1994). The severe effects arising from either replacementof the glycine with arginine (Chen et a!., 1994, 1995; Fuh et a!., 1992; Goffm et at., 1994; thisstudy) or its deletion (Chen et at., 1995) further suggests that the primary structure of the third a-helix is critical for hormone activity.116Figure 32. Ribbon representation of the putative structure of bPRL indicating residues withinthe putative site 1 and site 2 binding sites. The truncated N-terminus lacking the C4 - Clidisulfide is located in the upper right-hand corner and the C-terminus, which forms part of theC191 - C199 disulfide (indicated in greenish-yellow), is positioned just below the N-terminus.The C58 - C174 disulfide is shown as a shadowed, greenish-yellow tube. The residues indicatedare those which when substituted with other amino acids, primarily alanine, resulted in a variantwith reduced mitogenic activity. These residues may therefore form the binding determinants forsite 1 and site 2 binding to the receptor. Residues K69, Q71, Q73, and Q74 are on the loopregion joining helices 1 and 2, residues R125, E128, and G129 are on helix 3, and residues R176,R177, D178, K181, and D183 are on helix 4.1174.5 Species Specificity of the Hormone Binding SiteWhile not a conmion characteristic of members of the hormone family, the ability to bindwith high affinity to different types of receptors is not limited to hGH. Two species of placentallactogen from the mouse (PL-I and PL-ll) are capable, along with mPRL, of binding to themouse prolactin receptor (Davis and Linzer, 1989b), and ovine placental lactogen (oPL) is one ofthe few hormones other than hGH that binds with high affinity to the hGHbp (Carr and Friesen,1976). The overall sequence identity between oPL and hGH is small (26%) and although thedegree of homology is greater amongst the residues which in hGH contact the hGHbp (Colosi eta!., 1989) there are enough differences to suggest that oPL and hGH have overlapping butdifferent sets of important binding residues. One dissimilarity is the residue occupying theposition equivalent to D171 of hGH: this amino acid is a serine in oPL. This substitution isparticularly interesting since it relates to recent evidence suggesting that the identity of the aminoacid at this position is a major determinant of whether a growth hormone from one species willbind to a receptor from another species, a phenomenon called species specificity (Souza et at.,1995). While the growth hormone receptors from many species bind hGH as well as their owngrowth hormones, the hGHbp will bind only growth hormones from primates. It was suggestedthat the basis for this specificity is the arginine at position 43 in the hGHbp. R43 contacts bothD17l and T175 of hGH and while the threonine is conserved, the aspartate is not. In all non-primate growth hormones, in many prolactins, and in PL-ll from rat and mouse a histidineresidue is found at the position corresponding to 171 of hGH while a few hormones includingoPL, oPRL, and bPRL display a serine at this locus. Neither bPRL nor oPRL bind to the hGHbp(Carr and Friesen, 1976); therefore, the major determinant of species specificity is more likely tobe the incompatibility of a histidine at position 171 with R43 of the receptor than an absoluterequirement for aspartate or serine at that locus. With respect to bPRL, the putative modelsuggests that the serine at this position (S 180) is located within the patch of amino acids that hasbeen identified as being important for the mitogenic activity of the hormone. In addition, its sidechain appears to be well positioned for interacting with the receptor (Figure 30). But sincereplacement of this serine with alanine (variant Si 80A) did not cause a large reduction in118replacement of this serine with alanine (variant Si 80A) did not cause a large reduction inmitogenic activity (Figure 27 and Table 6), the serine per se cannot be essential for activity.However, in light of the above evidence that this locus may play a role in species specificity, itwould be very interesting to examine the effect that replacement of S180 with a bulky residuesuch as histidine might have on the mitogenic activity of bPRL.4.6 Conclusions and Future DirectionsRegulation of the many biological processes in which prolactin, along with the othermembers of the hematopoietic family of hormones, is involved is accomplished through theinteraction of the hormone with receptor molecules. Receptor binding initiates the signaltransduction mechanisms which ultimately give rise to the various biological effects. Therefore,elucidation of the specific contacts involved in the molecular interaction between each hormoneand its receptor is an essential component of the body of knowledge required for each hormonein order to fully exploit their commercial and medical potentials, particularly as related to thetreatment of various diseases.The current structure-function investigation into the interaction of bovine prolactin withthe Nb2 PRL receptor has confirmed the generality of the binding site and highlighted thedifferences that exist at the residue level with respect to other related hormones. As predictedfrom the sequence homology between members of the family of growth hormones, prolactins,and placental lactogens, the binding site of bPRL for the Nb2 PRL receptor appears to be verysimilar to that reported for binding of hGH to both the hGH and hPRL binding proteins. Thedata show that bPRL residues within the C-terminal half of the loop region joining putativehelices 1 and 2 as well as within the central portion of putative helix 4 are involved in promotingthe mitogenic activity of the hormone. Amino acids R177 and K18 1 were demonstrated to becritically important for activity of bPRL while replacements of other residues, including Q71,Q73, Q74, R176, and D183, indicated that they may also play a role in contacting the receptor, ormay, in the cases of L171 and D178, play a structural role in maintaining the global structure ofthe binding site.119However, while the general location of these residues is similar to that of the residueswhich in related hormones are involved in contacting receptor molecules, the identity amonghormones does not extend to the level of specific amino acids. For example, residues R21 andK187 were found not to be very important for the mitogenic activity of bPRL, while thecorresponding mPL-ll residues were shown to be very important for the mitogenic activity ofthat hormone (Davis and Linzer, 1989b). Differences between the residues involved in receptorbinding exist between bPRL and all the other hormones for which structure-function analyseshave been performed and are particularly striking between bPRL and hPRL. While the sequencehomology between these hormones within the binding epitope region is very high, there aresignificant differences in the contribution that some of the corresponding residues in bPRL andhPRL make to the mitogenic activity. For example, P61A of hPRL exhibited a four-foldreduction in mitogenic activity (Goffin et at., 1992) while the activity of P61A of bPRL was onlyslightly reduced. In addition, the K69A variant of hPRL had very low bioactivity and wasseverely affected in its ability to bind to the Nb2 PRL receptor (Goffin et at., 1992) but the K69Avariant of bPRL exhibited increased mitogenic activity. The differences between the interactionof bPRL with the Nb2 PRL receptor and the interactions of hGH with the hGH and hPRLreceptors are also quite marked. One major difference between the interactions of these twohormones with a lactogenic receptor is that Zn2 is required for the binding of hGH to thehPRLbp (Cunningham and Wells, 1991) but not for bPRL binding to the Nb2 PRL receptor.Thus, although the corresponding residues D183 of bPRL and E174 of hGH are both importantfor the interaction of these hormones with a lactogenic receptor, the roles of these residues arenot identical since E174 of hGH is involved in coordinating the Zn2 ligand. Another markeddifference between the two hormones is seen with the contribution of the corresponding residuesE67 of bPRL and S62 of hGH: the mitogenic activity of the E67A met-bPRL variant was almostidentical to that of unmodified met-bPRL while the ability of the S62A variant of hGH to bind tothe hPRLbp was severely reduced (Cunningham and Wells, 1991).The comparison of sequences and analysis of residues conserved between the relatedgrowth hormones, prolactins, and placental lactogens has not, in general, been very informative120for identifying residues which form the receptor binding sites. As noted repeatedly above,residues which are highly conserved throughout the family of hormones do not play equallyimportant roles in interacting with the receptor molecules. However, as shown from acomparison of the hormones hPL, pGH, and hPRL with hGH, consideration of sequencedivergence as a basis for predicting which residues form binding sites can be misleading. Inthese examples the sequence variations within the binding site are the same as that over the entirehormone molecule (Cunningham and Wells, 1989). Thus, a number of sequence differencesthroughout the disparate regions that form the binding sites give rise to the functional differencesbetween the hormones.The many differences between the residues involved in the interactions among these relatedhormones and their receptors in spite of the similarities in the general locations of the bindingsites likely contribute to hormone-receptor binding specificity. Although the receptors for thisfamily of hormones also belong to a family of proteins (Idzerda et a!., 1990) they are quitedifferent in terms of amino acid sequence. Therefore, the portions of these receptors with whichthe hormones interact contain a variety of residues and it is impossible that a single set ofhormone residues could bind to each receptor. In addition, although the members of the growthhormone/prolactin/placental lactogen family of hormones are evolutionarily related (Miller andEberhardt, 1983) these hormones have continued to evolve separately from each other and thedifferent roles that individual residues have in activating receptors may be a consequence of thisevolutionary divergence.In this study evidence was also obtained which suggests that bPRL, like hGH, contains twoseparate receptor binding sites and that signal transduction therefore follows receptordimerization. Replacement of the glycine at position 120 with arginine totally abolished themitogenic activity of bPRL: this effect may be the result of the arginine residue hindering contactbetween the hormone and the receptor at site 2 either by virtue of its size or of its charge. Themechanism of the interaction between hGH and the hGHbp or the hPRLbp is one of sequentialdimerization wherein the affinity of site 1 for the receptor is higher than that of site 2 (Fuh et aL,1992, 1993). However, a study of the interaction between hPRL and the Nb2 lactogenic receptor121suggested that this mechanism may not be universal amongst the family of hormones: thebinding affinity of both hPRL sites for the Nb2 PRL receptor appeared to be similar (Goffm eta!., 1994). Further study is needed to determine which mechanism is followed with respect tobPRL binding to the Nb2 PRL receptor.Significant progress has been made towards obtaining an understanding of the structure-function relationship between bPRL and the Nb2 PRL receptor but much work remains to bedone. The binding site of hGH for the hGHbp and the hPRLbp includes residues within helix 1as well as within the loop region and helix 4 (Cunningham andWells, 1991), and amino acidswithin the N-terminal region of rat PRL have been shown to be involved in the niitogenicactivity of this hormone (Maruyama et a!., 1994). Therefore, the contribution of putative helix 1residues to the mitogenic activity of bPRL should be determined, both by alanine-scanningmutagenesis and by replacement of residues believed to be within the binding site by amino acidswhich might have a more disruptive effect on receptor binding by virtue of their size or charge.In this regard it is of interest that alanine-scanning mutagenesis of putative site 2 residues ofhPRL did not identify side chains involved in activating the Nb2 PRL receptor, but replacementof certain of these residues with tryptophan or arginine did provide evidence for a second bindingsite (Goffin eta!., 1994). The data obtained regarding the effect of the G129R mutation providesa tantalizing glimpse into the mechanism of receptor binding; i.e., that bPRL contains tworeceptor binding sites: additional mutations should be created in this region in order to furtherexplore this possibility. It may even be possible to generate variants which, by acting asantagonists to PRL, could be useful in the treatment of disorders such as hyperprolactinemia.While a structure-function analysis provides a great deal of information with respect to thecontribution of specific residues to the functional activity of a protein, a crystal structure isnecessary to determine whether residues which appear to be functionally important actually playa structural role in maintaining the binding site rather than directly interacting with the receptor.In addition, the validity of the model of receptor activation could be tested by the creation ofspecific mutants that, as suggested by molecular modeling based on the coordinates of the wildtype structure, would be expected to severely interfere with the hormone-receptor interaction.122However, a relatively large amount of pure protein is required in order to perform a structuralanalysis of prolactin. Unfortunately, the system used in this study for the isolation of active metbPRL is not amenable to large-scale production of protein. There are many problems associatedwith efficient production of recombinant proteins, not the least of which is producing a correctlyfolded and soluble protein. Although there are a wide variety of E. coli expression systemsavailable for high-level expression of eukaryotic proteins in E. coli, these often lead toaccumulation of the foreign protein in insoluble inclusion bodies. As it is possible that specificmutations generated in the course of a structure-function analysis could affect the ability of therecombinant protein to refold correctly, and thereby interfere with an evaluation of thecontribution of the residue in question to the biological activity of the protein, it is very importantto assess the conformation of the mutants. While production of recombinant proteins in a solubleform does not guarantee that mutations will not cause any structural modification of the protein,it does eliminate the problems associated with solubiization and refolding of recombinantproteins from inclusion bodies. Thus, an investigation into alternative methods of producing thehormone, particularly those involving eukaryotic or insect cell cultures, may prove fruitful interms of generating larger amounts of soluble, active bPRL than could be obtained withtransformed cultures of E. coli.The present work provides a significant contribution to the sum of knowledge about theinteraction between bPRL and the Nb2 PRL receptor, and, as a consequence of the homologybetween prolactin and other hormones, lends supports to the suggestion that there is a commonmechanism of receptor binding and activation within the hematopoietic hormone family. Furtherstudy into the interaction between bPRL and the Nb2 PRL receptor will allow identification ofall the bPRL residues that play a role in activating this receptor and will also, by virtue of thedifferences that exist between this hormone-receptor pair and others that have been examined,lead to greater understanding of the general mechanism. of receptor binding by members of thegrowth hormone/prolactin/placental lactogen hormone family.123REFERENCESAbdel-Meguid, S.S., Shieh, H.-S., Smith, W.W., Dayringer, H.E., Violand, B.N. and Bentle,L.A. 1987. Three-dimensional structure of a genetically engineered variant of porcine growthhormone. Proc. Nati. Acad. Sci. USA 84:6434-6437.Alber, T., Dao-pin, S., Nye, J.A., Muchmore, D.C. and Matthews, B.W. 1987. Temperaturesensitve mutations of bacteriophage T4 lysozyme occur at sites with low mobility and lowsolvent accessibility in the folded protein. Biochemistry 26:3754-3758.Ali, S., Pellegrini, I. and Kelly, P.A. 1991. A prolactin-dependent immune cell line (Nb2)expresses a mutant form of prolactin receptor. J. Biol. Chem. 266:20110-20117.Argetsinger, L.S. Campbell, G.S., Yang, X., Witthuhn, B.A., Silvennoinen, 0., Ihie, J.N. andCarter-Su, C. 1993. Identification of JAK2 as a growth hormone receptor-associated tyrosinekinase. Cell 74:237-244.Ashkenazi, A., Vogel, T., Barash, I., Hadary, D., Levanon, A., Gorecki, M. and Gertler, A. 1987.Comparative study of in vitro and in vivo modulation of lactogenic and somatotropic receptorsby native human growth hormone and its modified analogue prepared by recombinantdeoxyyribonucleic acid technology. Endocrinology 121:414-419.Atkinson, T. and Smith, M. 1984. Solid-phase synthesis of oligodeoxyribonucleotides by thephosphitetriester method. In: Gait, M.J. (ed.) Oligonucleotide Synthesis. A Practical Approach.IRL Press, Oxford, pp. 35-8 1.Aubert, M.L., Bewley, T.A., Grumbach, M.M., Kaplan, S.L. and Li, C.H. 1986. Structure-function studies on human growth hormone. hit. J. Peptide Protein Res. 28:45-57.Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A. and Struhl, K.(eds.) 1994. Current Protocols in Molecular Biology. John Wiley & Sons, Inc. N.Y.Baumann, G. 1991. Growth hormone heterogeneity: genes, isohonnones, variants, and bindingproteins. Endocr. Rev. 12:424-449.Baumann, G., Stolar, M.W., Amburn, K., Barsano, C.P. and DeVries, B.C. 1986. A specificgrowth hormone-binding protein in human plasma: initial characterization. 3. Clin. Endocrinol..Metab. 62:134-141.Bazan, J.F. 1989. A novel family of growth factor receptors: a common binding domain in thegrowth hormone, prolactin, erythropoietin and IL-6 receptors, and the p75IL-2 receptor B-chain.Biochem. Biophys. Res. Commun. 164:788-795.Bazan, J.F. 1990a. Haemopoietic receptors and helical cytokines. ImmunoL Today 11:350-354.Bazan, J.F. 1990b. Structural design and molecular evolution of a cytokine receptor superfamily.Proc. Nail. Acad. Sci. USA. 87:6934-6938.Bazan, J.F. and McKay, D.B. 1992. Unraveling the structure of IL-2. Science 257:410-413.Beanie, 3. 1993. Structural and functional aspects of the interaction between growth hormone andits receptor. Biochim. Biophys. Acta 1203: 1-10.Becker, G.W. and Hsiung, H.M. 1986. Expression, secretion and folding of human growthhormone in Escherichia coli: purification and characterization. FEBS Letts. 204:145- 150.124Billestrup, N., Allevato, G., Norstedt, G., Møldrup, A. and Nielsen J.H. 1994. Identification ofintracellular domains in the growth hormone receptor involved in signal transduction. Proc. Soc.Expt BioL Med. 206:205-209.Binder, L., Gertler, A., Elberg, G., Guy, R. and Vogel, T. 1990. Site-directed mutations of humangrowth hormone that selectively modify its lactogenic activity and binding properties. Mol.Endocrinol. 4:1060-1068.Binder, L., Vogel, T., Hadary, D., Elberg, G. and Gertler, A. 1989. Chimeric bovine-humangrowth hormone prepared by recombinant DNA technology: binding properties and biologicalactivity. Mol Endocrinol. 3:923-930.Blight, M.A., Chervaux, C. and Holland, I.B. 1994. Protein secretion pathways in Escherichiacoli. Curr. Opin. Biotechnol. 5:468-474.Bradley, M.K. 1990. Overexpression of proteins in eukaryotes. Methods Enzymol. 182:112-132.Brinck-Johnsen, T. and Benirschke, K. 1982. The placenta: maternal, fetal, and placentalendocrine interaction. In: Endocrine Pathology: General and Surgical. Bloodworth Jr., J.M.B.(ed.) Wiffiams and Wilkins, Baltimore, pp. 293-308.Brunger, A.T. 1992. X-PLOR version 3.1. A system for X-ray crystallography and NMR. YaleUniversity, New Haven, CT.Borck, K., Beggs, J.D. Branimer, W.J., Hopkins, A.S. and Murray, N.E. 1976. The constructionin vitro of transducing derivatives of phage lambda. Mol. Gen. Genet. 146:199-207.Buck, K., Vanek, M., Groner, B. and Ball, R.K. 1992. Multiple forms of prolactin receptormessenger ribonucleic acid are specifically expressed and regulated in murine tissues and themammary cell line HC11. Endocrinology 130:1108-1114.Carey, G.B. and Liberti, J.P. 1993. Inhibition of tyrosine kinase activity abolishes anabolic andmitogenic effects of prolactin and growth hormone in Nb2 cells. Biochem. Mol. Biol. Internat.31:635-642.Carr, D. and Friesen, H.G. 1976. Growth hormone and insulin binding to human liver. J. ClinEndocrinol. Metab. 42:484-493.Carter, P., Bedouelle, H. and Winter, G. 1985. Improved oligonucleotide site-directedmutagenesis using M13 vectors. Nucleic Acids Res. 13:4431-4443.Carter-Su, C., Stubbart, J.R., Wang, X., Stred, S.E., Argetsinger, L.S. and Shafer, J.A. 1989.Phosphorylation of highly purified growth hormone receptors by a growth hormone receptorassociated tyrosine kinase. J. Biol. Chem. 264:18654-18661.Carter-Su, C., Argetsinger, L.S., Campbell, G.S., Wang, X., Ihle, 3. and Witthuhn, B. 1994. Theidentification of JAK2 tyrosine kinase as a signaling molecule for growth hormone. Proc. Soc.Expt. Biol. Med. 206:210-215.Chang, C.N., Rey, M., Bochner, B., Heyneker, H. and Gray, G. 1987. High-level secretion ofhuman growth hormone by Escherichia coil. Gene 55:189-196.Chang, J.Y.-H., Pal, R.-C., Bennett, W.F. and Bochner, B.R. 1989. Periplasmic secretion ofhuman growth hormone by Escherichia coli. Biochem. Soc. Trans. 17:335-337.125Chawla, R.K., Parks, J.S. and Rudman, D. 1983. Structural variants of human growth hormone:biochemical, genetic, and clinical aspects. Ann. Rev. Med. 34:519-547.Chen, W.Y, Chen, N.-Y., Yun, 3., Wagner, T.E. and Kopchick, J.J. 1994. In vitro and in vivostudies of antagonistic effects of human growth hormone analogs. J. Biol. Chem. 269:15892-15897.Chen, W.Y., Chen, N.-Y., Yun, J., Wight, D.C., Wang, X.Z., Wagner, T.E. and Kopchick, J.J.1995. Amino acid residues in the third a-helix of growth hormone involved in growth promotingactivity. Mo!. Endocrinol. 9:292-302.Chen, W.Y., Wight, D.C., Chen, N.Y.; Coleman, T.A., Wagner, T.E. and Kopchick, J.J. 1991a.Mutations in the third a-helix of bovine growth hormone dramatically affect its intracellulardistribution in vitro and growth enhancement in transgenic mice. J. Biol. Chem. 266:2252-2258.Chen, W.Y., Wight, D.C., Mehta, B.V., Wagner, T.E. and Kopchick, J.J. 199 lb. Glycine 119 ofbovine growth hormone is critical for growth-promoting activity. Mo!. Endocrinol. 5:1845-1852.Chen, X.Z., Yun, J.S., Li, Y.S., Wagner, T. and Kopchick, JJ. 1991c. Conversion of bovinegrowth hormone amino acid residues 57 and 60 to those of human growth hormone do not alterreceptor binding properties and growth patterns in transgenic mice. Proc. 73rd Ann. Meet.Endocr. Soc., p. 214, Endocrine Society, Bethesda (Abstr. 736).Chêne, N., Martal, J., de la Liosa, P. and Charrier, J. 1989. Growth hormones II: structure-function relationships. Reprod. Nutr. Develop. 29:1-25.Chiba, T., Amanuma, H. and Todokoro, K. 1992. Tryptophan residue of trp-ser-X-trp-ser motifin extracellular domains of erythropoietin receptor is essential for signal transduction. Biochm.Biophys. Res. Commun. 184:485-490.Clackson, T. and Wells, l.A. 1995. A hot spot of binding energy in a hormone-receptor interface.Science 267:383-386.aapp, C., Sears, P.S. and Nicoll, CS. 1989. Binding studies with intact rat prolactin and a 16Kfragment of the hormone. Endocrinology 125:1054- 1059.Clapp, C., Martial, J.A., Guzman, R.C., Rentier-Deirue, F. and Weiner, R.L 1993. The 16-kilodalton N-terminal fragment of human prolactin is a potent inhibitor of angiogenesis.Endocrinology 133:1292-1299.Clapp, C., Sears, P.S., Russell, D.H., Richards, J., Levay-Young, B.K. and Nicoll, C.S. 1988.Biological and immunological characterization of cleaved and 16K forms of rat prolactin.Endocrinology 122:2892-2898.Clarke, W.C. and Bern, H.A. 1980. Comparative endocrinology of prolactin. In: Li, C.H. (ed.)Hormonal Proteins and Peptides. Academic Press, New York vol. 8: 105-197.Clevenger, C.V. and Medaglia, M.V. 1994. The protein tyrosine kinase p59fYfl is associated withprolactin (PRL) receptor and is activated by PRL stimulation of T-lymphocytes. Mo!.Endocrinol. 8:674-681.Clevenger, C.V., Torigoe, T. and Reed, S.C. 1994. Prolactin induces rapid phosphorylation andactivation of prolactin receptor-associated RAF-1 kinase in aT-cell line. J. Biol. Chem.269:5559-5565.126Cole, E.S., Nichols, E.H., Lauziere, K., Edmunds, T. and McPherson, J.M. 1991.Characterization of the microheterogeneity of recombinant primate prolactin: implications forposttranslational modifications of the hormone in vivo. Endocrinology 129:2639-2646.Colosi, P., Thordarson, G., Helimiss, R., Singh, K., Forsyth, l.A., Gluckman, P. and Wood, W.I.1989. Cloning and expression of ovine placental lactogen. Mol. Endocrinol. 3:1462-1469.Cooke, N.E., Coit, D., Shine, J., Baxter, J.D. and Martial, J.A. 1981. Human prolactin: cDNAstructural analysis and evolutionary comparisons. J. Biol. Chem. 256:4007-4016.Corner, G.W. 1930. The hormonal control of lactation: I. Non-effect of the corpeus luteum; II.Positive action of extracts of the hypophysis. Am. 3. Physiol. 95:43-55.Cunningham, B.C., Bass, S., Fuh, G. and Wells, J.A. 1990a. Zinc mediation of the binding ofhuman growth hormone to the human prolactin receptor. Science 250: 1709-1712.Cunningham, B.C., Henner, D.J. and Wells, J.A. 1990b. Engineering human prolactin to bind tothe human growth hormone receptor. Science 247:1461-1465.Cunningham, B.C., Ihurani, P., Ng, P. and Wells, J.A. 1989. Receptor and antibody epitopes inhuman growth hormone identified by homolog-scanning mutagenesis. Science 243:1330-1336.Cunningham, B.C., Ultsch, M., de Vos, A.M., Mulkerrin, J.G., Clauser, K.R. and Wells, J.A.1991. Dimerization of the extracellular domain of the human growth hormone receptor by asingle hormone molecule. Science 254:821-825.Cunningham, B.C. and Wells, J.A. 1989. High-resolution epitope mapping of hGH-receptorinteractions by alanine-scanning mutagenesis. Science 244:1081-1085.Cunningham, B.C. and Wells, J.A. 1991. Rational design of receptor-specific variants of humangrowth hormone. Proc. Nail. Acad. Sci. USA 88:3407-3411.Cunningham, B.C. and Wells, J.A. 1993. Comparison of a structural and a functional epitope. J.MoL Biol. 234:554-563.Cymes, G.D., Atlasovich, F.M., Caridad, 3.3., Iglesias, M.M. andWolfenstein-Todel, C. 1994.Modification of arginine residues in ovine prolactin by 1,2-cyclohexanedione. mt. 3. PeptideProtein Res. 44:31-35.D’Andrea, A.D., Fasman, G.D. and Lodish, H.F. 1989. Erythropoietin receptor and interleukin-2receptor B-chain: a new receptor family. Cell 58:1023-1024.Das, A. 1990. Overproduction of proteins in Escherichia coli: vectors, hosts, and strategies.Methods Enzymol. 182:93-112.DaSilva, L., Howard, O.M.Z., Rui, H., Kirken, R.A. and Farrar, W.L. 1994. Growth signalingand JAK2 association mediated by membrane-proximal cytoplasmic regions of prolactinreceptors. 3. Biol. Chem. 269:18267-18270.David, M., Petricoin ifi, E.F., Igarashi, K.-I., Feldman, G.M., Finbloom, D.S. and Lamer, A.C.1994. Prolactin activates the interferon-regulated p91 transcription factor and the JAK2 kinase bytyrosine phosphorylation. Proc. Nat!. Acad. Sci. USA 91:7 174-7 178.Davis, J.A. and Linzer, D.I.H. 1989a. A mutant lactogenic hormone binds, but does not activate,the prolactin receptor. Mol. Endocrinol. 3:949-953.127Davis, J.A. and Linzer, D.I.H. 1989b. Mutational analysis of a lactogenic hormone reveals a rolefor lactogen-specific amino acid residues in receptor binding and mitogenic activity. Mol.Endocrinol. 3:1987-1995.Deb, S., Hashizume, K., Boone, K., Southard, J.N., Talamantes, F., Rawitch, A. and Soares, M.J.1989. Antipeptide antibodies reveal structural and functional characteristics of rat placentallactogen-il. Mo!. Cell. Endocrinol. 63:45-56.Dente, L., Cesareni, G. and Cortese, R. 1983. pEMBL: a new family of single stranded plasmids.Nuc!. Acids Res. 11: 1645-1655.de Vos, A.M., Ultsch, M. and Kossiakoff, A.A. 1992. Human growth hormone and extracellulardomain of its receptor: crystal structure of the complex. 255:306-3 12.Diederichs, K., Boone, T. and Karplus, P.A. 1991. Novel fold and putative receptor binding siteof granulocyte-macrophage colony-stimulating factor. Science 254:1779-1782.Doneen, B.A., Bewley, T.A. and Li, C.H. 1979. Studies on prolactin. Selective reduction of thedisulfide bonds of the ovine hormone. Biochemistry 18:4851-4860.Edery, M., Levi-Meyrueis, C., Paly, 3., Kelly, P.A. and Djiane, 3. 1994. A limited cytoplasmicregion of the prolactin receptor critical for signal transduction. Mol. Cell. Endocrinol. 102:39-44.Elberg, G., Kelly, P.A., Djiane, 3., Binder, L. and Gertler, A. 1990. Mitogenic and bindingproperties of monoclonal antibodies to the prolactin receptor in Nb2 rat lymphoma cells. 3. Biol.Chem. 265:14770-14776.Evans, S.V. 1993. SETOR - hardware-lighted 3-dimensional solid model representations ofmacromolecules. J. Mo!. Graphics 11: 134-138.Ferrara, N., Clapp, C. and Weiner, R. 1991. The 16K fragment of prolactin specifically inhibitsbasal or fibroblast growth factor stimulated growth of capillary endothelial cells. Endocrinology129:896-900.Foley, K.P., Leonard, M.W. and Engel, J.D. 1993. Quantitation of RNA using the polymerasechain reaction. Trends Genet. 9:380-385.Forsyth, l.A. and Myres, R.P. 1971. Human prolactin. Evidence obtained by the bioassay ofhuman plasma. J. Endocr. 51:157-168.Frank, S.J., Gilliland, G., Kraft, A.S. and Arnold, C.S. 1994. Interaction of the growth hormonereceptor cytoplasmic domain with the JAK2 tyrosine kinase. Endocrinology 135:2228-2239.Freemark, M. and Corner, M. 1989. Purification of a distinct placental lactogen receptor, a newmember of the growth hormone/prolactin receptor family. J. Clin. Invest. 83:883-889.Fuh, G., Colosi, P., Wood, W.L and Wells, J.A. 1993. Mechanism-based design of prolactinreceptor antagonists. 3. Biol. Chem. 268:5376-5381.Fuh, G., Cunningham, B.C., Fukunaga, R., Nagata, S., Goeddel, D.V. and Wells, J.A. 1992.Rational design of potent antagonists to the human growth hormone receptor. Science 256:1677-1680.Fuh, G., Mulkerrin, M.G., Bass, S., McFarland, N., Brochier, M., Bourell, H.H., Light, D.R. andWells, J.A. 1990. The human growth hormone receptor: secretion from Escherichia coli anddisulfide bonding pattern of the extracellular binding domain. 3. Biol. Chem. 265:3111-3115.128Fukunaga, R., Ishizaka-Ikeda, E., Seto, Y. and Nagata, S. 1990. Expression cloning of a receptorfor murine granulocyte colony-stimulating factor. Cell 61:341-350.Fukushima, 3.G., Cascone, 0., Santomé, LA. and Biscoglio de Jimenez Bonino, M.J. 1987.Ethoxyformylation of histidine residues in equine growth hormone. mt. J. Peptide Protein Res.30:365-370.Gala, R.R. 1991. Prolactin and growth hormone in the regulation of the immune system. Proc.Soc. Expt. Biol. Med. 198:513-527.Gearing, D.P., King, J.A., Gough, N.M. and Nicola, N.A. 1989. Expression cloning of a receptorfor human granulocyte-macrophage colony-stimulating factor. EMBO J. 8:3667-3676.Gertler, A., Hauser, S.D., Sakal, E., Vashdi, D., Staten, N., Freeman, J.J. and Krivi, G.G. 1992.Preparation, purification, and determination of the biological activities of 12 N terminus-truncated recombinant analogues of bovine placental lactogen. J. Biol. Chem. 267: 12655-12659.Gertler, A., Petridou, B., Kriwi, G.G. and Djiane, J. 1993. Interaction of lactogenic hormoneswith purified recombinant exctracellular domain of rabbit prolactin receptor expressed in insectcells. FEBS Letts. 319:277-28 1.Gertler, A., Shamay, A., Cohen, N., Ashkenazi, A., Friesen, H.G., Levanon, A., Gorecki, M.,Aviv, H., Hadary, D and Vogel, T. 1986. Inhibition of lactogenic activities of ovine prolactin andhuman growth hormone (hGH) by a novel form of a modified recombinant hGH. Endocrinology118:720-726.Ghrayeb. 3., Kimura, H., Takahara, M., Hsiung, H., Masui, Y. and Inouye, M. 1984. Secretioncloning vectors in Escherichia coli. EMBO 3. 3:2437-2442.Gibson, T.J. 1984. Studies on the Epstein-Barr virus genome. Ph.D. thesis. CambridgeUniversity, England.Gilbert, H.F. 1994. Protein chaperones and protein folding. Curr. Opin. Biotechnol. 5:534-539.Gilbert, M.S., Lowry, P.3., Castro, M.G. Woods, R.J. and Savva, D. 1991. Expression and partialpurification of human prolactin in Escherichia coli. mt. 3. Biochem. 23:107-114.Gilliland, G., Perrin, S., Blanchard, K. and Bunn, H.F. 1990. Analysis of cytokine mRNA andDNA: detection and quantitation by competitive polymerase chain reaction. Proc. Natl. Acad.Sci. USA 87:2725-2729.Gilmour, K.C. and Reich, N.C. 1994. Receptor to nucleus signaling by prolactin and interleuldn2 via activation of latent DNA-binding factors. Proc. Nail. Acad. Sd. USA 91:6850-6854.Goffin, V., Norman, M. and Martial, J.A. 1992. Alanine-scanning mutagenesis of humanprolactin: importance of the 58-74 region for bioactivity. Mol. Endocrinol. 6:1381-1392.Goffin, V., Struman, I., Goormaghtigh, E. and Martial, J.A. 1993. The addition of nine residuesat the C-terminus of human prolactin drastically alters its biological properties. Eur. 3. Biochem.214:483-490.Goffin, V., Struman, I., Mainfroid, V., Kinet, S. and Martial, LA. 1994. Evidence for a secondreceptor binding site on human prolactin. 3. Biol. Chem. 269:32598-32606.129Goodwin, R.G., Friend, D., Ziegler, S.F., Jerzy, R., Falk, B.A., Gimpel, S., Cosman, D., Dower,S.K., March, C.J., Namen, A.E. and Park, L.S. Cloning of the human and murine interleukin-7receptors: demonstration of a soluble form and homology to a new receptor superfamily. Cell60:941-951.Gough, J.A. and Murray, N.E. 1983. Sequence diversity among related genes for recognition ofspecific targets in DNA molecules. 3. Mol. Bid. 166:1-19.Goujon, L., Allevato, G., Simonin, G., Paquereau, L., Le Cam, A., Clark, 3., Nielsen, J.H.,Djiane, 3., Postel-Vinay, M.-C., Edery, M. and Kelly, P.A. 1994. Cytoplasmic sequences of thegrowth hormone receptor necessary for signal transduction. Proc. Nati. Acad. Sci. USA 91:957-961.Gout, P.W., Beer, C.T. and Noble, R.L. 1980. Prolactin-stimulated growth of cell culturesestablished from malignant Nb rat lymphomas. Cancer Res. 40:2433-2436.Gouy, M. and Gautier, C. 1982. Codon usage in bacteria: correlation with gene expressivity.Nucl. Acids Res. 10:7055-7074.Gráf, L., Barát, E., Borvendég, 3., Hermann, I. and Patthy, A. 1976. Action of thrombin on ovine,bovine and human pituitary growth hormones. Eur. J. Biochem. 64:333-340.Gráf, L. and Li, C.H. 1974. Isolation and properties of two biologically active fragments fromlimited tryptic hydrolysis of bovine and ovine pituitary growth hormones. Biochemistry 13:5408-5415.Gray, G.L., Baldridge, J.S., McKeown, K.S., Heyneker, H.L. and Chang, C.N. 1985. Periplasmicproduction of correctly processed human growth hormone in Escherichia coli: natural andbacterial signal sequences are interchangeable. Gene 39:247-254.Grosjean, H. and Fiers, W. 1982. Preferential codon usage in prokaryotic genes: the optimalcodon-anticodon interaction energy and the selective codon usage in efficiently expressed genes.Gene 18:199-209.Guyette, W.A., Matusik, R.J. and Rosen, J.M. 1979. Prolactin-mediated transcriptional and post-transcriptional control of casein gene expression. Cell 17: 1013-1023.Hanks, M.C. Talbot, R.T. and Sang, H.M. 1989. Expression of biologically active recombinant-derived chicken prolactin in Escherichia coli. J. Mol. Endocrinol. 3:15-21.Hara, K., Chen, C.-J.H. and Sonenberg, M. 1978. Recombination of the biologically activepeptides from a tryptic digest of bovine growth hormone. Biochemistry 17:550-556.Harlow, E. and Lane, E. (eds.) 1988. In: Antibodies: a laboratory manual. Cold Spring HarborLaboratory Press, NY.Hemmings, B.A., Zubenko, G.S., Hasilik, A. and Jones, E.W. 1981. Mutant defective inprocessing of an enzyme located in the lysozyme-like vacuole of Saccharomyces cerevisiae.Proc. Nati. Acad. Sci. USA 78:435-439.Hendrickson, W.A. andWard, K.B. 1977. Pseudosymmetry in the structure of myohemerythrin.3. BioL Chem. 252:3012-3018.Herington, A.C., Ymer, S. and Stevenson, J. 1986. Identification and characterization of specificbinding proteins for growth hormone in normal human sera. 3. Clin. Invest. 77:18 17-1823.130Hiraoka Y., Nomata, Y., Matsuo, K., Tsubota, N., Tanabe, K. and Fukasawa, T. 1991. Efficientproduction of biologically active human prolactin in Escherichia coli. Mo!. Cell. Endocrinol.81:147-154.Hooghe, R., Deihase, M., Vergani, P., Malur, A. and Hooghe-Peters, E.L. 1993. Growthhormone and prolactin are paracrine growth and differentiation factors in the haemopoieticsystem. Immunol. Today 14:212-214.Hooper, K.P., Padmanabhan, R. and Ebner, K.E. 1993. Expression of the extracellular domain ofthe rat liver prolactin receptor and its interaction with ovine prolactin. 3. Biol. Chem. 268:22347-22352.Horseman, N.D. and Yu-Lee, L.-Y. 1994. Transcriptional regulation by the helix bundle peptidehormones: growth hormone, prolactin, and hematopoietic cytokines. Endocr. Rev. 15:627-649.Hsiung, H.M., Canirell, A., Luirink, J., Oedega, B. Veros, A.J. and Becker, G.W. 1989. Use ofbacteriocin release protein in E. coli for excretion of human growth hormone into the culturemedium. Bio/Technology 7:267-27 1.Hsiung, H.M., Mayne N.G. and Becker, G.W. 1986. High-level expression, efficient secretionand folding of human growth hormone in Escherichia coli. Biofrechnology 4:991-995.Hwang, P., Guyda, H. and Friesen, H. 1972. Purification of human prolactin. 3. Biol. Chem.247:1955-1958.Idzerda, R.L., March, C.J., Mosley, B., Lyman, S.D., Vanden Bos, T., Gimpel, S.D., Din, W.S.,Grabstein, K.H., Widmer, M.B., Park, L.S., Cosman, D. and Beckmann, P. 1990. Humaninterleukin 4 receptor confers biological responsiveness and defmes a novel receptorsuperfamily. 3. Exp. Med. 171:861-873.Inouye, S. Vlasuk, G.P., Hsiung, H., and Inouye, M. 1984. Effects of mutations at glycineresidues in the hydrophobic region of the Escherichia coli prolipoprotein signal peptide on thesecretion across the membrane. J. BioL Chem. 259:3729-3733.Itoh, N., Yonehara, S., Schreurs, J., Gorman, D.M., Maruyama, K., Ishii, A., Yahara, I., Arai, K.and Miyajima, A. 1990. Cloning of an interleukin-3 receptor gene: a member of a distinctreceptor gene family. Science 247:324-327.Jeoung, D.-I., Allen, D.L., Guller, S., Yen, V. and Sonenberg, M. 1993. Mitogenic and receptoractivities of human growth hormone 108-129. J. Biol. Chem. 268:22520-22524.Kato, C., Kobayashi, T., Kudo, T., Furusato, T., Murakami, Y., Tanaka, T., Baba, H., Oishi, T.,Ohtsuka, E., Ikehara, M., Yanagida, T., Kato, H., Moriyama, S. and Horikoshi, K. 1987.Construction of an excretion vector and extracellular production of human growth hormone fromEscherichia coli. Gene 54:197-202.Kelly, P.A., All, S., Rozakis, M., Goujon, L., Nagano, M., Pellegrini, I., Gould, D., Djiane, 3.,Edery, M., Finidori, 3. and Postel-Vinay, M.C. 1993. The growth hormone/prolactin receptorfamily. Rec. Prog. Horm. Res. 48:123-164.Kelly, P.A., Boutin, J.-M., Jolicoeur, C., Okamura, H., Shirota, M., Edery, M., Dusanter-Fourt, I.and Djiane, J. 1989. Purification, cloning, and expression of the prolactin receptor. Biol.Reproduc. 40:27-32.Kelly, P.A., Djiane, 3., Postel-Vinay, M.C. and Edery, M. 1991. The prolactin/growth hormonereceptor family. Endocrine Rev. 12:235-25.1.131Kleinberg, D.L. and Frantz, A.G. 1971. Human prolactin: measurement in plasma by in vitrobioassay. 3. Clin. Invest. 50:1557-1568.Kornberg, L.J. and Liberti, J.P. 1992. Stimulation of Nb2 mitogenesis by an analogue of humangrowth hormone (110-127). Biochem. Internat. 28:873-879.Koshland, D. and Botstein, D. 1980. Secretion of beta-lactamase requires the carboxy end of theprotein. Cell 20:749-760.Kostyo, J.L. 1974. The search for the active core of pituitary growth hormone. Metabolism23:885-899.Kunkel, T.A. 1985. Rapid and efficient site-specific mutagenesis without phenotypic selection.Proc. Nati. Acad. Sci. USA 82:488-492.Kunkel, T.A., Roberts, J.D. and Zakour, R.A. 1987. Rapid and efficient site-specific mutagenesiswithout phenotypic selection. Methods Enzymol. 154:367-382.Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature 227:680-685.Laskowski, R.A., MacArthur, M.W., Moss, D. and Thornton, J.M. 1993. PROCHECK: aprogram to check the stereochemical quality of protein structures. 3. Appi. Cryst. 26:283-29 1.Lederer F., Glatigny, A., Bethge, P.H., Bellamy, H.D. and Mathews, R.S. 1981. Improvement ofthe 2.5 A resolution model of cytochrome b562 by redetermining the primary structure and usingmolecular graphics. 3. MoL Biol. 148:427-448.Lemmon, M.A., Flanagan, J.A., Hunt, J.F., Adair, B.D., Bormann, B.-J., Dempsey, C.E. andEngleman, D.M. 1992. Glycophorin A dimerization is driven by specific interactions betweentransmembrane a-helices. 3. Biol. Chem. 267:7683-7689.Lesniak, M.A., Gorden, P. and Roth, 3. 1977. Reactivity of non-primate growth hormones andprolactins with human growth hormone receptors on cultured human lymphocytes. J. Cliii.EndocrinoL Metab. 44:838-849.Lesueur, L., Edery, M., All, S., Paly, 3., Kelly, P.A. and Djiane, 3. 1991. Comparison of long andshort forms of the prolactin receptor on prolactin-induced milk protein gene transcription. Proc.Nati. Acad. Sci. USA 88:824-828.Leung, D.W., Spencer, S.A., Cachianes, G., Hammonds, R.G., Coffins, C., Henzel, W.J.,Barnard, R., Waters, M.J. and Wood, W.L 1987. Growth hormone receptor and serum bindingprotein: purification, cloning and expression. Nature 330:537-543.Lewis, U.J., Bonewald, L.F. and Lewis, L.J. 1980. The 20,000-dalton variant of human growthhormone: location of the amino acid deletion. Biochem. Biophys. Res. Commun. 92:511-516.Li, C.H. 1973. Human growth hormone: perspectives on its chemistry and biology. In: Raiti, S.(ed.) Advances in Human Growth Hormone Research. U.S. Government Printing Office,Washington, D.C., pp. 321-341.Li, C.H. and Blake, 3. 1979. Semisynthesis of human somatotropin analogs. Proc. Nati. Acad.Sci. USA 76:6124-6127.132Li, C.H., Hayashida, T., Doneen, BA. and Rao, A.J. 1976. Human somatotropin: biologicalcharacterization of the recombinant molecule. Proc. Nat!. Acad. Sci. USA 73:3463-3465.Loewenstein, J.E., Mariz, 1.K., Peake, G.T. and Daughaday, W.H. 1971. Prolactin bioassay byinduction of N-acetyllactosamine synthetase in mouse mammary gland explants. 3. Clin. Endocr.33:217-224.Lowman, H.B., Cunningham B.C. andWells, J.A. 1991. Mutational analysis and proteinengineering of receptor-binding determinants in human placental lactogen. 3. Biol. Chem.266:10982-10988.Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. 1951. Protein measurement with theFolin phenol reagent. J. Biol.Chem. 193:265-275.Luck, D.N., Gout, P.W., Beer, C.T. and Smith, M. 1989. Bioactive recombinant methionylbovine prolactin: structure-function studies using site-specific mutagenesis. Mol. Endocrinol.3:822-831.Luck, D.N., Gout, P.W., Kelsay, K., Atkinson, T., Beer, C.T. and Smith, M. 1990. Recombinantmethionyl bovine prolactin: loss of bioactivity after single amino acid deletions from putativehelical regions. Mol. Endocrinol. 4: 1011-1016.Luck, D.N., Gout, P.W., Sutherland, E.R. Fox, K., Huyer, M. and Smith, M. 1992. Analysis ofdisuiphide bridge function in recombinant bovine prolactin using site-specific mutagenesis andrenaturation under mild alkaline conditions: a crucial role for the central disuiphide bridge in themitogenic activity of the hormone. Protein Engng. 5:559-567.Luck, D.N., Huyer, M., Gout, P.W., Beer, C.T. and Smith, M. 1991. Single amino acidsubstitutions in recombinant bovine prolactin that markedly reduce its mitogenic activity in Nb2cell cultures. Mol. Endocrinol. 5:1880-1886.Luck, D.N., Ngsee, J.K., Rottman, F.M. and Smith, M. 1986. Synthesis of bovine prolactin inEscherichia coli. DNA 5:21-28.Marston, F.A.O. 1987. The purification of eukaryotic polypeptides expressed in Escherichia coli.In: Glover, D.M. (ed.) DNA Cloning: A Practical Approach. Vol. ifi. IRL Press, Oxford, pp. 59-88.Maruyama, 0., Kato, T., Wakabayashi, K. and Kato, Y. 1994. Amino acids in the amino terminalregion of the rat prolactin contribute to PRL-receptor binding and Nb2 cell proliferation activity.Biochem. Biophys. Res. Commun. 205:312-319.Messing, J. 1983. New M13 vectors for cloning. Methods Enzymol. 101:20-78.Miller, W.L. and Eberhardt, N.L. 1983. Strucure and evolution of the growth hormone genefamily. Endocr. Rev. 4:97-130.Mitraki, A. and King, J. 1989. Protein folding intermediates and inclusion body formation.Bio/Technology 7:690-697.Miyazaki, T., Maruyama, M., Yamada, G., Hatakeyama, M. and Taniguchi, T. 1991. Theintegrity of the conserved “WS motif common to IL-2 and other cytokine receptors is essentialfor ligand binding and signal transduction. EMBO J. 10:3191-3 197.Nicholls, A. Bharadwaj, R. and Honig, B. 1993. GRASP - graphical representation and analysisof surface properties. Biophys. J. 64:A166.133Nicoll, C.S. 1967. Bio-assay of prolactin. Analysis of the pigeon crop-sac response to localprolactin injection by an objective and quantitative method. Endocrinology 80:641-655.Nicoll, C.S. and Bern, HA. 1972. On the actionof prolactin among the vertebrates: is there acommon denominator? In Woisten-Holme, G.E.W.and Knight, 3. (eds.) Lactogenic Hormones.Churchill-Livingston, London, pp. 299-324.Nicoll, C.S., Mayer, G.L and Russell, S.M. 1986.Structural features of prolactins and growthhormones that can be related to their biological properties. Endocr. Rev. 7:169-203.Ogren, L. and Talamantes, F. 1988. Prolactins of pregnancy and their cellular source. Ira. Rev.Cytol. 112:1-65.O’Neal, K.D. and Yu-Lee, L.-Y. 1993. The proline-rich motif (PRM): a novel feature of thecytokine/hematopoietin receptor superfamily. Lympho. Cyto. Res. 12:309-312.Pakula, A.A., Young, V.B. and Sauer, R.T. 1986.Bacteriophage cro mutations: effects onactivity and intracellular degradation. Proc. Nati.Acad. Sci. USA 83:8829-8833.Paris, N., Rentier-Deirue, F., Defontaine, A., Goffin, V., Lebrun, 3.3., Mercier, L. and Martial,J.A. 1990. Bacterial production and purificationof recombinant human prolactin. Biotechnol.Appl. Biochem. 12:436-449.Patthy, L. 1990. Homology of a domain of the growth hormone/prolactin receptor family withtype ifi modules of fibronectin. Cell 61:13-14.Petham, H. 1985. Cleaning up plasmid minipreps with lithium chloride. Trends Genet. 1:6.Perrin, S. and Gililand, G. 1990. Site-specific mutagenesis using asymmetric polymerase chainreaction and a single mutant primer. Nucl. AcidsRes. 18:7433-7438.Pigiet, V.P. and Schuster, B.J. 1986. Thioredoxin-catalyzed refolding of disulfide-containingproteins. Proc. Nati. Acad. Sci. USA 83:7643-7647.Powers, R., Garrett, D.S., March, C.J., Frieden, E.A., Gronenbom, A.M. and Clore, G.M. 1992.Three-dimensional solution structure of human interleukin-4 by multidimensional heteronuclearmagnetic resonance spectroscopy. Science 256: 1673-1677.Ray, 3., Okamura, H., Kelly, P.A., Liebhaber, S.A.and Cooke, N.E. 1990. Alteration in thereceptor binding specificity of human growth hormone by genomic exon exchange. Mol.Endocrinoi 4:101-107.Reagan, C.R., Kostyo, J.L., Mills, 3.B., Gennick, S.E., Messina, J.L., Wagner, S.A. andWilhelmi, A.E. 1981. Recombination of fragments of human growth hormone: altered activityprofile of the recombinant molecule. Endocrinology109:1663-1671.Reagan, C.R., Mills, J.B., Kostyo, 3.L. and Wilhelmi, A.E. .1975. Isolation and biologicalcharacterization of fragments of human growth hormone produced by digestion with plasmin.Endocrinology %:625-636.Riddle, 0., Bates, R.W. and Dykshorn, S.W. 1932. A new anterior pituitary hormone. Proc. Soc.Expt. Biol. Med. 29:1211-1212.134Robinson, M., Liley, R., Little, S., Emtage, J.S. Yarranton, G., Stephens, P., Millican, A., Eaton,M. and Humphreys, G. 1984. Codon usage can affect efficiency of translation of genes inEscherichia coli. Nuci. Acids Res. 12:6663-6671.Rozakis-Adcock, M. and Kelly, P.A. 1992. Identification of ligand binding determinants of theprolactin receptor. 3. Biol. Chem. 267:7428-7433.Rui, H., Djeu, J.Y., Evans, G.A., Kelly, P.A. and Farrar, W.L. 1992. Prolactin receptortriggering: evidence for rapid tyrosine kinase activation. J. Biol. Chem. 267:24076-24081.Rui, H., Kirken, R.A. and Farrar, W.L: 1994i Activation of receptor-associated tyrosine kinaseJAK2 by prolactin. J. Biol. Chem. 269:5364-5368.Rui, H., Lebrun, J.J., Kirken, R.A., Kelly, P.A. and Farrar, W.L. 1994b. JAK2 activation and cellproliferation induced by antibody-mediated prolactin receptor dimerization. Endocrinology135:1299-1306.Russell, 3., Sherwood, L.M., Kowaiski, K. and Schneider, A.B. 1981. Recombinant hormonesfrom fragments of human growth hormone and human placental lactogen. 3. Biol. Chem.256:296-300.Sakal, E., Gertler, A. and Shechter, Y. 1991. Biological activity of a fluorescein human growthhormone derivative prepared by specific covalent labeling of lysine-70. Biochemistry 30:8899-8904.Sakal, E., Tchelet, A., Uchida, E., Shimokawa, S., Nishikawa, S., Hayakawa, T., Krivi, G.G. andGertler, A. 1993. Site-directed mutagenesis of hGH at the 54-74 loop selectively modifies itslactogenic receptor-mediated biological activity. MoL Cell. Endocrinol. 97:129-134.Sambrook, 3., Fritsch, E.F. and Maniatis, T. (eds.) 1989. Molecular Cloning: A LaboratoryManual. Cold Spring Harbor Laboratory Press, N.Y.Sanger, F., Nicklen, S. and. Coulson, A.R. 1977. DNA sequencing with chain-terminatinginhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467.Sasavage, N.L., Nilson, J.H., Horowitz, S. and Rottman, F.M. 1982. Nucleotide sequence ofbovine prolactin messenger RNA. J. BioL Chem. 257:678-68 1.Savino, R., Lahm, A., Giorgio, M., Cabibbo, A., Tramontano, A. and Ciliberto, G. 1993.Saturation mutagenesis of the human interleukin 6 receptor-binding site: implications for itsthree-dimensional structure. Proc. Nail. Acad. Sci. USA 90:4067-4071.Sawadogo, M. and Van Dyke, M.W. 1991. A rapid method for the purification of deprotectedoligodeoxynucleotides. Nuci. Acids Res. 19:674.Schein, C.H. and Noteborn, M.H. 1988. Formation of soluble recombinant proteins inEscherichia coil is favored by lower growth temperature. Bio/Technology 6:291-294.Schmitt-Ney, M., Doppler, W., Ball, R.K. and Groner, B. 1991. 13-casein gene promoter activityis regulated by the hormone-mediated relief of transcriptional repression and a mammary-gland-specific nuclear factor. Mol. Cell. Biol. 11:3745-3755.Schoemaker, J.M., Brasnett, A.H. and Marston, F.A.O. 1985. Examination of calf prochymosinaccumulation in Escherichia coil: disulphide linkages are a structural componet of prochymosincontaining inclusion bodies. EMBO 3. 4:775-780.135Schoner, R.G., Ellis, L.F. and Schoner, B.E. 1985. Isolation and purification of protein granulesfrom Escherichia coil cells overproducing bovine growth hormone. Bio/Technology 3:151-154.Shiu, R.P.C., Elsholtz, H.P., Tanaka, T., Friesen, H.G., Gout, P.W., Beer, C.T. and Noble, R.L.1983. Receptor-mediated mitogenic action of prolactin in a rat lymphoma cell line.Endocrinology 113:159-165.Sigel, M.B., Thorpe, N.A., Kobrin, M.S. Lewis, U.J. and Vanderlaan, W.P. 1981. Bindingcharacteristics of a biologically active variant of human growth hormone (20K) to growthhormone and lactogen receptors. Endocrinology 108:1600-1603.Silvennoinen, 0., Zitthuhm, B.A., Quelle, F.W., Cleveland, J.L., Yi, T. and Ihie, J.A. 1993.Structure of the murine JAK2 protein-tyrosine kinase and its role in interleukin 3 signaltransduction. Proc. Natl. Acad. Sci. USA 90:8429-8433.Smith, C.R. and Norman, M.R. 1990. Prolactin and growth hormone:molecular heterogeneityand measurement in serum. Ann. Clin. Biochem. 27:542-550.Smith, W.C., Linzer, D.I.H. and Talamantes, F. 1988. Detection of two growth hormone receptormRNAs and primary translation products in the mouse. Proc. Nati. Acad. Sci. USA 85:9576-9579.Smith, L.J., Redfield, C., Boyd, 3., Lawrence, G.M.P., Edwards, R.G., Smith, R.A.G. andDobson, C.M. 1992. Human interleukin 4: the solution structure of a four-helix bundle protein. 3.Mol. Biol. 224:899-904.Somers, W., Ultsch, M., de Vos, A.M. and Kossiakoff, A.A. 1994. The X-ray structure of agrowth hormone-prolactin receptor complex. Nature 372:478-48 1.Souza, S.C., Frick, G.P., Wang, X., Kopchick, 3.3., Lobo, R.B. and Goodman, H.M. 1995. Asingle arginine residue determines species specificity of the human growth hormone receptor.Proc. Nail. Acad. Sci. USA 92:959-963.Staten, N.R., Byatt, J.C. and Krivi, G.G. 1993. Ligand-specific dimerization of the extracellulardomain of the bovine growth hormone receptor. 3. Biol. Chem. 268:18467-18473.Stevens, A.M. and Yu-Lee, L.-Y. 1994. Multiple prolactin-responsive elements mediate Gi andS phase expression of the interferon regulatory factor-i gene. Mol. Endocrinol. 8:345-355.Strasburger, C.J., Binder, L., Elberg, G., Cohen-Chapnick, N., Kohen, F. and Gertler, A. 1989.Indication of different lactogen and somatogen binding sites in the human growth hormonemolecule as probed with monoclonal antibodies. Mol. Cell. Endocrinol. 67:55-62.Stricker, P. and Grueter, F. 1928. Action du lobe antérieur de l’hypophyse sur la montée laiteuse.Compt. Rend. Soc. de BioL 99: 1978-1980.Tanaka, T., Shiu, R.P.C., Gout, P.W., Beer, C.T., Noble, R.L. and Friesen, H.G. 1980. A newsensitive and specific bioassay for lactogenic hormones: measurement of prolactin and growthhormone in human sera. 3. Clin. Endocrinol. Metab. 51:1058-1063.Teh, L.-C. and Chapman, G.E. 1988. Determination of the effect of acetylation of specific lysineresidues in human growth hormone on its affinity for somatogenic receptors by an affinityselection technique. Biochem. Biophys. Res. Commun. 150:391-398.136Teh, L.-C., Murphy, L.J., Huq, N.L, Surus, A.S., Friesen, H.G., Lazarus, L. and Chapman, G.E.(1987) Methionine oxidation in human growth hormone and human chorionic.somatomammotropin: effects on receptor binding and biological activities. J. Biol. Chem.262:6472-6477.Towns, R., Kostyo, J.L., Vogel, T., Sakal, E., Tchelet, A., Maher, R. and Gertler, A. 1992.Evidence that the N-terminus of human growth hormone is involved in expression of its growthpromoting, diabetogenic, and insulin-like activities. Endocrinology 130:1225-1230.Turkington, R.W. 1971. Measurement of prolactin activity in human serum by the induction ofspecific milk proteins in mammary gland in vitro. 3. Chin. Endocr. 33:210-216.Ulirich, A. and Schiessinger, 3. 1990. Signal transduction by receptors with tyrosine kinaseactivity. Cell 61:203-212.Ultsch, M., de Vos, A.M. and Kossiakoff, A.A. 1991. Crystals of the complex between humangrowth hormone and the exracellular domain of its receptor. J. MoL BioL 222:865-868.Ultsch, M.H., Somers, W., Kossiakoff, A.A. and de Vos, A.M. 1994. The crystal structure ofaffinity-matured human growth hormone at 2 A resolution. J. Mol. Biol. 236:286-299.VanderKuur, J.A., Wang, X., Zhang, L, Campbell, G.S., Ahlevato, G. Billestrup, N., Norstedt, G.and Carter-Su, C. 1994. Domains of the growth hormone receptor required for association andactivation of JAK2 tyrosine kinase. 3. Biol. Chem. 269:21709-21717.Vernet, T., Dignard, D. and Thomas, D.Y. 1987. A family of yeast expression vectors containingthe phage fi intergenic region. Gene 52:225-233.Wallis, M. 1978. The chemistry of pituitary growth hormone, prolactin, and related hormones,and its relationship to biological activity. In: Weinstein, B. (ed.) Chemistry and Biochemistry ofAmino Acids, Peptides, and Proteins: A Survey of Recent Developments. Marcel Dekker, N.Y.5:213-320.Wallis, M., Howell, S.L. and Taylor, K.W. (eds.) 1985. The Biochemistry of the PolypeptideHormones. John Wiley & Sons, Inc., N.Y.Walter, M.R., Cook, W.J., Ealick, S.E., Nagabhushan, T.L., Trotta, P.P. and Bugg, C.E. 1992.Three-dimensional structure of recombinant human granulocyte-macrophage colony-stimulatingfactor. J. Mol. Biol. 224:1075-1085.Wang, Y.-D. and Wood, W.I. 1995. Amino acids of the human growth hormone receptor that arerequired for proliferation and JAK-STAT signaling. Mol. Endocrinol. 9:303-311.Watling, D., Guschin, D., Muller, M., Silvennoinen, 0., Witthuhn, B.A., Quelle, F.W., Rogers,N.C., Schindler, C., Stark, G.R., Ihie, J.N. and Keff, I.M. 1993. Complementation by the proteintyrosine kinase JAK2 of a mutant cell line defective in the interferon-7 signal transductionpathway. Nature 366:166-170.Weber, P.C., Bartsch, R.G., Cusanovich, M.A., Hamlin, R.C., Howard, A., Jordan, S.R., Kamen,M.D., Meyer, T.E., Weatherford, D.W., Xuong, N.H. and Salemme, F.R. 1980. Structure ofcytochrome C’: a dimeric, high-spin haem protein. Nature 286:302-304.Wells, J.A. 1994. Structural and functional basis for hormone binding and receptoroligomerization. Curr. Opin. Cell Biol. 6: 163-173.137Wells, J.A. and de Vos, A.M. 1993. Structureand function of human growth hormone:implications for the hematopoietins. Annu. Rev. Biophys. Biomol. Struct. 22:329-35 1.Williams, D.C., Van Frank, R.M., Muth, W.L.and Burnett, J.P. 1982. Cytoplasmic inclusionbodies in Escherichia coli producing biosynthetic human insulin proteins. Science 215:687-689.Witthuhn, B.A., Quelle, F.W., Silvennoinen, 0., Yi, T., Tang, B., Miura, 0. and Ihie, J.N. 1993.JAK2 associates with the erythropoietin receptor and is tyrosine phosphorylated and activatedfollowing stimulation with erytbropoietin. Cell74:227-236.Wiodawer A., Pavlovsky, A. and Gustchina, A. 1993. Hematopoietic cytokines: similarities anddifferences in the structures, with implicationsfor receptor binding. Protein Science 2:1373-1382.Wohnlich, L. and Moore, W.V. 1982. Bindingof a variant of human growth hormone to liverplasma membranes. Horm. Metabol. Res. 14: 138-141.Yamasaki, N., Kikutani, M. and Sonenberg, M.1970. Peptides of a biologically active trypticdigest of bovine growth hormone. Biochemistry9:1107-1114.Yanisch-Perron, C., Vieira, 3. and Messing, J. 1985. Improved M13 phage cloning vectors andhost strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33:103-119.Zoller, M.J. and Smith, M. 1982. oligonucleotiae-directed mutagenesis using M13-derivedvectors: an efficient and general procedure for the production of point mutations in any fragmentof DNA. Nuci. Acids Res. 10:6487-6500.Zoller, MJ. and Smith, M. 1987. Oligonucleotide-directed mutagenesis: a simple method usingtwo oligonucleotide primers and a single-stranded DNA template. Methods Enzymol. 154:329-350138


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items