Protein Expression and Purification 20, 265–273 (2000)doi:10.1006/prep.2000.1290, available online at http://www.idealibrary.com on

High-Level Expression of Biologically Active HumanProlactin from Recombinant Baculovirus in Insect Cells

Tapas Das,* Paul W. Johns,* Vincent Goffin,† Paul Kelly,† Bruce Kelder,‡ John Kopchick,‡Kristi Buxton,* and Pradip Mukerji*,1

*Ross Products Division, Abbott Laboratories, Columbus, Ohio 43215; †INSERM, Endocrinologie Moleculaire,Paris Cedex 15, France; and ‡Edison Biotechnology Institute, Ohio University, Athens, Ohio 45701

Received April 17, 2000, and in revised form June 7, 2000

We examined the feasibility of high-level productionof recombinant human prolactin, a multifunctionalprotein hormone, in insect cells using a baculovirusexpression system. The human prolactin cDNA withand without the secretory signal sequence was clonedinto pFastBac1 baculovirus vector under the controlof polyhedrin promoter. Prolactin was produced uponinfection of either Sf9 or High-Five cells with the re-combinant baculovirus containing the human prolac-tin cDNA. The production of recombinant prolactinvaried from 20 to 40 mg/L of monolayer culture, de-pending on the cell types. The prolactin polypeptidewith its own secretory signal was secreted into themedium. N-terminal amino acid sequence analysis ofthe recombinant polypeptide purified from the culturemedium indicated that the protein was processed sim-ilar to human pituitary prolactin. Carbohydrate anal-ysis of the purified protein indicated that a fraction ofthe recombinant prolactin made in insect cells ap-peared to be glycosylated. Also, both secreted and non-secreted forms of the recombinant prolactin in insectcells were biologically equivalent to the native humanprolactin (pituitary derived) in the Nb2 lymphoma cellproliferation assay. © 2000 Academic Press

Prolactin (PRL)2 is one of the most versatile hor-mones of the pituitary gland in terms of biologicalactions. It is involved in more physiological processesthan all other pituitary hormones combined. Amongthese are the regulation of mammary gland develop-ment, initiation and maintenance of lactation, immune

1 To whom correspondence should be addressed. Fax: (614) 624-3570. E-mail: pradip.mukerji@rossnutrition.com.

2 Abbreviations used: PRL, prolactin; hPRL, human prolactin; moi,multiplicity of infection; LHRE, lactogenic hormone responsive ele-ment; RLU, relative light units; hsPRL, prolactin cDNA with secre-

tory signal.

1046-5928/00 $35.00Copyright © 2000 by Academic PressAll rights of reproduction in any form reserved.

modulation, osmoregulation, and behavioral modifica-tion (1, 2). Human prolactin is synthesized as a precur-sor polypeptide (preprolactin) of 227 amino acids inlength. The first 28 residues constituting the signalpeptide sequence are cleaved during processing in thesecretory pathway and generate 199 amino acid longmature prolactin with a molecular mass of 23 kDa (3,4). In most mammalian species, the mature hormoneconsists of 197–199 amino acids, with variable se-quence homology among the species. PRL was origi-nally identified in pituitary extract; however, researchin recent years has shown the presence of PRL andPRL-like molecules in the placenta, brain, and gut andin the tissues of immune system as well (4).

PRL has also been detected in milk of the cow, goat,pig, sheep, rat, and primate (5). Although PRL plays anessential role in the initiation and maintenance of lac-tation in the mother, the results of several animalstudies lend support to a role of milk-borne PRL in thedevelopment of neuroendocrine, reproductive, and im-mune function in the neonate (5, 6). In breast-fed hu-man infants, prolactin activity has been detected inserum when no endogenous hPRL production takesplace. The possibility that PRL may play an importantrole in the development of human neonate is suggestedfrom reports that serum PRL levels are positively cor-related to the stage of development (gestational age atbirth) and that serum prolactin levels below 33 ng/mlare associated with poor clinical outcome of the infant(5). The transport of milk prolactin into the infant’scirculation is believed to occur through receptors thatare abundantly expressed by the mucosal cells alongthe gut. Gut receptors have high affinity for prolactinand increase in number with age in the various speciestested (7). How pituitary PRL is able to serve as a keylactogenic agent in the mother, whereas milk-bornePRL seemingly influences the development of unre-

lated organ systems in the neonate, is poorly under-

265

c

dpg

266 DAS ET AL.

stood, but may in part be due to modification of prolac-tin structure and function as it is transferred into milk.In suckling neonatal animals, it has been demon-strated that milk-borne prolactin is responsible for theproliferation and differentiation of T-cells and B-cells(5, 6). The role of prolactin as an immunomodulatorhas been demonstrated both in vitro and in vivo (5, 6).A survey of the literature and data indicate that pro-lactin may be of potential clinical use after bone mar-row transplantation to promote hematopoietic and im-mune recovery (8). Prolactin promotes proliferation ofmurine B-cell hybridomas and has been shown to actas an antagonist of TGF-b, an immunosuppressive cy-tokine (9). These findings provide an additional ratio-nale for using recombinant PRL clinically in immuno-suppressed patients in certain disease settings such asAIDS and cancer, where overexpression of TGF-b hasbeen implicated in disease development and progres-sion (9).

Recent reports indicate that PRL exhibits microhetero-geneity (4, 5, 10), which may account for these variedbiological activities. Posttranslational modifications ofprolactin include glycosylation, phosphorylation, oli-gomerization, and proteolytic cleavage. The consistentpresence of these modified forms of prolactin in humanmilk and serum suggests that they might have physio-logical relevance. Although the physiological role of gly-cosylated and phosphorylated variants is unknown, theN-terminal fragments of rat and human prolactin havebeen shown to have antiangiogenic activities as demon-strated both in vitro and in vivo (11). The lack of purifiedforms from either human pituitary or human milk hashindered studies to delineate the functional differencesamong different prolactin variants. To overcome thisproblem, we attempted to produce recombinant humanprolactin in large quantities in eukaryotic cells. A majoradvantage of recombinant hPRL is its lack of contamina-tion with other related hormones (e.g., growth hormone)that might interfere in functional assays. To date, humanprolactin has been produced in bacteria (12, 13) as well asin mammalian cells (14). Although the posttranslationalmodifications have been observed with PRL produced inmouse C127 cells (14), the limitation of the stable trans-fection system is that the amount of protein produced isnot high enough to obtain purified forms in large quan-tities (14). Also, the commercial use of recombinant pro-tein purified from mammalian cells has regulatory safetyconcerns since most of the mammalian cell lines aretransformed with pathogenic or infectious viruses. Wechose the baculovirus expression system since this sys-tem produces foreign proteins in large quantity and, inmost cases, the recombinant proteins are processed andmodified in a manner similar to that of their authenticcounterparts (15, 16). Moreover, baculoviruses have arestricted host range, limited to specific invertebrate spe-

cies. These viruses are safer to work with since they are c

noninfectious to vertebrates. Most of the susceptible in-sect cell lines are not transformed with pathogenic orinfectious viruses and can be cultured under minimalcontainment conditions. Here, we report the successfulproduction of large amounts of biologically active humanprolactin in insect cells. Further biochemical analysis ofthe purified protein indicates that the recombinant hu-man prolactin produced in these cells is processed andmodified in a manner similar to that of its native coun-terpart.

MATERIALS AND METHODS

Subcloning of Human Prolactin cDNA in BaculovirusTransfer Vector pFastBac1

The human prolactin (hPRL) cDNA in a pBR322vector was generously provided by Dr. John Kopchik,Ohio University, Athens, OH. To amplify a hPRLcDNA lacking a secretory signal sequence, the forwardprimer RO 468 (59-ATATCGCGGATCCATGTTGC-CCATCTGTCCCGGC-39) was used. This primer alsoontained a BamHI site and a start codon 59 of the

hybridizing sequence. The reverse primer RO 469 (59-ATATCCGCTCGAGTTAGCAGTTGTTGTTGTGG-39)had an XhoI site after the stop codon. To amplify thehPRL cDNA encoding a signal sequence, the forwardprimer RO 510 (59-ATATCGCGGATCCATGAACAT-CAAAGGATCGCC-39), which contained a BamHI siteupstream of the ATG, was used. RO 469 was used asthe reverse primer. DNA fragments were amplified byusing PCR. The polymerase chain reaction in a finalvolume of 100 ml was carried out for 25 cycles in tem-perature conditions of 30 s at 94°C, 30 s at 45°C, and30 s at 72°C. The PCR-amplified products were gelpurified, digested with BamHI and XhoI, and ligatedonto pFastBac1 baculovirus transfer vector (Life Tech-nologies, Gaithersburg, MD) restricted with BamHIand XhoI. This vector contains a mini-Tn7 element, anE. coli origin of replication, and an ampicillin marker.The mini-Tn7 contains an expression cassette consist-ing of gentamycin resistance marker, polyhedrin pro-moter, a multicloning site, and an SV40 PolyA signalinserted between the left and right arms of Tn7. Initialtransformation and screening were done in XL1 blue(Invitrogen, Carlsbad, CA), and the positive cloneswere confirmed by dideoxy sequencing and used fortransformation of E. coli DH10Bac.

Cell Culture and Virus

The Bac-To-Bac baculovirus (Life Technologies) is aderivative of Autographa californica nuclear polyhe-

rosis virus (AcMNPV). The recombinant virus wasropagated in a monolayer culture of Spodoptera fru-iperda cells, Sf9. Serum free medium adapted Sf9

ells (Life Technologies) were grown in a monolayer

taT7orvai

P

Ftrt4pdph(D

wcdpC

267HUMAN PROLACTIN PRODUCTION IN INSECT CELLS

culture at 27°C in serum-free medium, Sf-900 II SFM(Life Technologies) in the presence of a final concentra-tion of 0.53 penicillin–streptomycin–glutamine (PSG(1003) stock purchased from Life Technologies). High-Five cells adapted in serum-free medium were pur-chased from Invitrogen and grown in a monolayer cul-ture at 27°C in High-Five serum-free mediumaccording to the manufacturer’s protocol (Invitrogen).

Generation of Recombinant Baculovirus

Generation of recombinant baculovirus expressinghuman prolactin in Sf9 cells was carried out by usingthe Bac-To-Bac baculovirus expression kit purchasedfrom Life Technologies. DNA of the recombinant pFas-Bac1 transfer vector containing cDNA corresponding toeither human preprolactin or the mature prolactin se-quence was introduced into E. coli DH10Bac for thetransposition of the prolactin cDNAs into baculovirusgenomic DNA (bacmid) according to the manufactur-er’s protocol. Colonies containing recombinant bacmidswere identified by disruption of the lacZa gene, and thewhite colonies were picked for recombinant bacmidDNA isolation. DNA was isolated using a Qiagen plas-mid isolation kit (Qiagen Inc., Valencia, CA) specific forDNA over 135 kb long. The recombinant bacmid DNAswere analyzed on a 0.6% agarose gel followed by PCRanalysis using pUC/M13 forward and reverse primersto confirm the presence as well as correct insert size oftransposed prolactin cDNA within the bacmid. Therecombinant bacmid DNA was then used to transfectSf9 cells which were grown in serum-free medium. At adensity of 1 3 106 cells per 35-mm well, the cells wereransfected with the bacmid DNA using Cellfectin re-gent according to the manufacturer’s protocol (Lifeechnologies). The recombinant virus was harvested at2 h posttransfection. Plaque assays were performedn the supernatants to determine the titer of recoveredecombinant virion particles, and each recombinantiral stock was made for expression studies. Plaquessay and propagation of virus were carried out accord-ng to the manual provided with the kit.

roduction of Recombinant Protein

To analyze the production of prolactin, Sf9 or High-ive cells were grown in 75-mm tissue culture flasksill 70–80% confluency. The cells were infected withecombinant virus at different multiplicities of infec-ion (moi) and the infection was carried out at 27°C for8 h. To examine the intracellular production, cyto-lasmic extract was prepared from the infected cells asescribed previously (17). Cells were washed withhosphate-buffered saline and suspended in 0.5 ml ofomogenization buffer containing 10 mM Tris–HClpH 8.0) and 25 mM NaCl. Cells were disrupted with a

ounce homogenizer and the lysate was centrifuged at b

10,000g at 4°C for 5 min. The supernatant was ana-lyzed for the presence of prolactin. To examine theextracellular production, the culture medium was col-lected 48 h postinfection followed by centrifugation at10,000g at 4°C for 5 min and the supernatant wascollected and analyzed for the presence of secretedprolactin.

Western Blot Analysis

For Western blot analysis, approximately 2 mg ofcytoplasmic extract or 5 ml of culture medium super-natant was electrophoresed on a 14% precast poly-acrylamide gel (Novex, San Diego, CA), and the re-solved protein was transferred onto an Immobilon-Pmembrane (Millipore, Bedford, MA) according to themanufacturer’s protocol. After blocking with BSA, themembrane was probed with anti-human prolactinmonoclonal antibody (Biomeda Corp., Foster City, CA)followed by peroxidase-linked goat anti-mouse IgG.Prolactin was visualized by ECL (Amersham Pharma-cia Biotech Inc., Piscataway, NJ) according to the man-ufacturer’s protocol.

Immunoassay

The amount of prolactin present in the cell extract orin the culture medium was determined by IMx immu-noassay according to the manufacturer’s protocol (Ab-bott Laboratories, Chicago, IL). The IMx system is afully automated immunoassay analyzer designed torun assays using enzyme immunoassay and fluores-cence polarization immunoassay technologies. The IMxprolactin assay is based on microparticle enzyme im-munoassay technology, which makes use of an “anti-body–antigen–antibody” complex and the generation ofa fluorescent reaction product (4-methylumbelliferyl)to determine part per billion (ppb) levels of prolactin ina sample. The mock-infected cell extract and the cul-ture medium were used as controls.

Prolactin Bioassay

The Nb2 rat lymphoma cell line (18), which is depen-dent on prolactin for growth, was used to assay bioac-tivity of the recombinant human prolactin according tothe method described previously (12). Increasingamounts of recombinant prolactin (as diluted cell ex-tract or culture medium) were added to Nb2 cellsplated in multiwell Falcon plates at 2 3 105 cells per

ell. Each sample was assayed in quadruplicate andell cultures with mock extract or mock culture me-ium were used as controls. After 72-h incubation withrolactin, cell numbers were determined with aoulter counter (12).Since rat Nb2 cell proliferation assay depends on the

inding of human prolactin with rat receptors, a new

268 DAS ET AL.

bioassay of human prolactin has been developed in ahomologous system (19). The new bioassay involves theuse of human embryonic kidney cells 293 stably ex-pressing the human PRLR cDNA and the lactogenichormone responsive element (LHRE) luciferase re-porter (20). LHRE is the DNA binding element of thesignal transducer and transactivator Stat5 (21), one ofthe signaling proteins activated by dimerized PRLR(22). Thus, the expression of luciferase in these doublytransfected cells (HL) depends exclusively on the hu-man prolactin present in the culture medium. Lucif-erase assays were performed as described previously(19). Increasing amounts of recombinant prolactinwere added to HL cells plated in multiwell plates at5 3 104 cells per well. Each sample was assayed induplicate. After 18–24 h of stimulation, cells were har-vested and lysed in 50 ml of lysis buffer. Luciferaseactivity of each experiment was counted in 10–20 ml ofcell lysates for 10 s. The difference between duplicatesnever exceeded 15% of relative light unit (RLU) values.

Protein Purification

The secreted recombinant human prolactin in theculture medium from either mouse L-cells (positivecontrol) or Sf9 cells was first dialyzed against buffercontaining 10 mM Tris–HCl (pH 8.0) and 25 mM NaCl.After extensive dialysis (2–3 changes) at 4°C, the me-dium was concentrated two- to threefold by an Amiconfiltration device using a 10-kDa cutoff membrane. Theconcentrated, dialyzed material was then applied to aMono Q FPLC column (Amersham Pharmacia BiotechInc.). The column was equilibrated with a buffer con-taining 10 mM Tris–HCl (pH 8.0) and 25 mM NaCl.After loading the protein, the column was washed withthe same buffer and eluted with a gradient of 25–300mM NaCl in Tris–HCl (pH 8.0). One-milliliter fractionswere collected and analyzed on a 14% SDS–polyacryl-amide gel stained with Coomassie blue.

N-Terminal Sequencing and Carbohydrate Analysis

Purified prolactin was electrophoresed on a 14%SDS–polyacrylamide gel at constant power. The pro-tein bands were transferred to the polyvinylidene di-fluoride (PVDF) membrane as described elsewhere(23), excised after staining with Coomassie blue R-250,and subjected to N-terminal amino acid sequencing ina Procise sequencing system, Applied Biosystem Model494.

For carbohydrate analysis, the purified protein wastransferred to the PVDF membrane as describedabove. The protein bands were excised and subjected tomonosaccharide analysis of glycoconjugates by anionexchange chromatography with pulsed amperometricdetection. This analysis was performed on a Dionex

BioLC according to the methods described earlier (24).

Reagents

All molecular biology reagents were purchased fromBoehringer Mannheim (Indianapolis, IN). Reagents forprotein purification and analysis were purchased fromVWR Scientific and Bio-Rad (Richmond, CA). Purifiedpituitary human prolactin was purchased from Biode-sign (Kennebunk, ME). The culture medium of mouseL-cells (stably expressing human prolactin cDNA withsecretory signal sequence) containing recombinant hu-man prolactin was received from the laboratory of Dr.John Kopchick, Ohio University, Athens, OH.

RESULTS

Production of Human Prolactin in Insect Cells

The human prolactin cDNA was cloned into the bac-

FIG. 1. Outline of the procedure used for the expression of humanprolactin gene in insect cells. The prolactin cDNA fragments with(hsPRL) and without (hPRL) secretory signal sequences were clonedinto baculovirus transfer vector, pFastBac1, under the control ofpolyhedrin promoter, pPolh. Ap and Gm indicate ampicillin andgentimycin resistance genes, respectively.

ulovirus vector pFastBac1 (Fig. 1) as detailed under

capdbap

269HUMAN PROLACTIN PRODUCTION IN INSECT CELLS

Materials and Methods. The cDNA either correspond-ing to 227 amino acid long preprolactin (for extracellu-lar expression) or corresponding to 199 amino acid longmature prolactin (for intracellular expression) wasused for subcloning into baculovirus transfer vector,pFastBac1. The recombinant plasmid was then intro-duced into E. coli DH10Bac for the transposition ofprolactin cDNAs into baculovirus genomic DNA (bac-mid). To generate recombinant baculovirus, the recom-binant bacmid DNA was used to transfect Sf9 insectcells and the recombinant virus was harvested 72 hposttransfection.

To analyze the production of prolactin, the Sf9 cellmonolayer was infected with recombinant virus at dif-ferent moi. At 48 h postinfection, the medium wascollected and saved, and the cytoplasmic extract wasprepared. The cell extract and the culture fluid wereelectrophoresed in a 14% polyacrylamide gel, and theprotein bands were electroblotted onto Immobilon-Pmembrane. The blot was probed with monoclonal anti-human prolactin antibody and developed with ECLreagent. The results shown in Fig. 2A indicate that theexpression of prolactin in the case of Sf9(hPRL) (pro-lactin cDNA without secretory signal sequence) is ex-clusively intracellular and almost no prolactin polypep-tide was detected in the culture medium as expected(data not shown). In the case of Sf9(hsPRL) (prolactin

FIG. 2. Production of human prolactin in Sf9 cells. (A) Cytoplasmicextract was prepared from Sf9 cells infected with either wild-type[Sf9(vec)] or recombinant virus [Sf9(hPRL)]. Approximately 2 mg ofytoplasmic extract was electrophoresed on a 14% SDS–polyacryl-mide gel and subjected to Western blot analysis with anti-humanrolactin monoclonal antibody. (B) Five microliters of culture me-ium obtained from Sf9 cells infected with either wild-type or recom-inant virus [Sf9(hsPRL)] was electrophoresed on a similar gel asbove and subjected to Western blot analysis. Nat indicates nativerolactin purified from human pituitary.

cDNA with secretory signal), a majority of the prolactin

was recovered in the culture medium, indicating thatpreprolactin was processed in insect cells and the ma-ture prolactin was secreted into the medium (Fig. 2B).The secreted prolactin from insect cells was comigratedwith native prolactin in the polyacrylamide gel (Fig.2B).

To quantify the production levels, either Sf9 or High-Five cells were infected with the recombinant virus at5 moi for 48 h postinfection. No significant increase inexpression level was achieved in Sf9 cells at 48 hpostinfection by increasing the moi from 5 to 10 (Fig.2). To measure the intracellular production levels, thecells were lysed and cytoplasmic extract was made andsubsequently used in immunoassay. To measure theamount of secreted prolactin, the culture medium wascollected and used directly in immunoassay. The intra-cellular production appeared to be twofold higher thanthe extracellular production (Table 1). Also, High-Fivecells are a better host than Sf9 in terms of prolactinproduction. In these cells, the intracellular productionlevels obtained were around 35–40 mg/L of culturewhile the amount of secreted prolactin varied from 18to 20 mg/L of culture medium (Table 1).

Bioactivity of Recombinant Prolactin Expressed inInsect Cells

To determine whether the human prolactin ex-pressed in insect cells was biologically active, an invitro cell proliferation assay was performed accordingto Tanaka et al. (25). However, in this assay, humanprolactin interacts with the heterologous rat receptorsand stimulates the Nb2 cell proliferation. To test thebioactivity of human prolactin in a homologous system,we have used a bioassay developed by generating hu-man embryonic kidney fibroblast 293 cells stably ex-pressing the human prolactin receptor cDNA and aprolactin-responsive (LHRE) luciferase reporter gene(19). The luciferase induction is dependent on theamount of prolactin added to the culture medium. In-creasing amounts of culture medium containing se-creted prolactin or cytoplasmic extract (in the case of

TABLE 1

Quantification of Recombinant hPRL Producedin Insect Cells

Recombinantprolactin

Production levels

In Sf9 cells(mg/L of culture)

In High-Five cells(mg/L of culture)

Intracellular(hPRL) 18–20 35–40

Extracellular(hsPRL) 8–10 18–20

270 DAS ET AL.

intracellular expression) were added to either Nb2 or293 (LHRE) cells in tissue culture plates. For Nb2 cellproliferation assay, the cell numbers were counted 72 hafter addition of prolactin, while for LHRE inductionassay, luciferase activity was measured in the cell ly-sate prepared from 293 (LHRE) cells after 18–24 h ofstimulation with recombinant hormone. The stimula-tion of Nb2 cell proliferation or induction of LHRE byrecombinant human prolactin produced in insect cellsis illustrated in Fig. 3. A high level of similarity (astrict parallelism) between the dose–response curves ofinsect cell derived prolactin and native or mammaliancell produced prolactin indicates that the recombinanthuman prolactin produced by insect cells is as active as

FIG. 3. Bioactivity analysis of recombinant prolactin. The concen-tration of prolactin in the cytoplasmic extract [in the case ofSf9(hPRL)] or in the culture medium [in the case of Sf9(hsPRL) andL-cell(hsPRL)] was determined by RIA with the use of the AmershamRIA kit. (A) Increasing amounts of recombinant prolactin as dilutedcell extract or culture medium were added to Nb2 lymphoma cells.Each point represents the mean of four determinations. (B) HL cellsstably expressing human prolactin receptor cDNA and LHRE-lucif-erase reporter gene were stimulated for 18–24 h by increasing con-centrations of recombinant prolactin as diluted cell extract or culturemedium. Data are expressed as relative units obtained from theluminometer reading. Each point represents the mean of duplicatemeasurements of luciferase activity.

its authentic counterpart.

Purification and Characterization of RecombinantProlactin from Insect Cells

The medium from a 100-ml culture of Sf9 cells in-fected with the recombinant baculovirus (containingprolactin cDNA with secretory signal sequence) at amoi of 5 plaque-forming units (pfu) per cell was col-lected 48 h postinfection. The secreted prolactinpresent in the culture medium was applied on a MonoQ FPLC column as described under Materials andMethods. The prolactin was eluted from the column bya linear 25–300 mM NaCl gradient and the fractionswere subjected to 14% SDS–PAGE analysis. The re-solved protein gel was stained with Coomassie blue(Fig. 4A). The prolactin protein from the Mono Q col-umn eluted in a distinct peak at approximately 150mM NaCl. The purified prolactin (pooled peak frac-tions) derived from insect cells comigrated with nativeprolactin in the SDS–polyacrylamide gel and appearedto be more than 90% pure (Fig. 4A).

To confirm that during translation and membranetranslocation in insect cells the secretory signal at theN-terminal end of the recombinant protein was cleavedproperly, we carried out N-terminal sequence analysisof the recombinant prolactin purified from insect cellculture medium. To compare with the N-terminal se-

FIG. 4. N-Terminal sequence analysis of secreted human prolactinpurified from cell culture medium. (A) Coomassie blue stained pro-tein pattern of various fractions eluted from Mono Q column. Leftpanel represents the purification of secreted prolactin from Sf9 cellswhile the right panel shows the purification of secreted prolactinfrom mouse L-cells. M, molecular weight standards; Nat, nativeprolactin isolated from human pituitary; FT, flowthrough; Wash,column wash; Eluate, pooled peak fractions. (B) The purified proteinbands from Sf9(hsPRL) and L-cell(hsPRL) as shown above weretransferred onto a PVDF membrane and subjected to N-terminal

amino acid sequence analysis.

271HUMAN PROLACTIN PRODUCTION IN INSECT CELLS

quence of human prolactin produced in mammaliancells, we purified the secreted prolactin from the cul-ture medium of mouse L-cells (stably expressed humanprolactin) by Mono Q column chromatography. Thepurified protein samples were electrophoresed in a 14%SDS–polyacrylamide gel and transferred to the PVDFmembrane. The protein bands were excised and sub-jected to N-terminal sequencing. The results shown inFig. 4B indicate that in insect cells the human prolac-tin with its own secretory signal was processed in amanner similar to that which takes place in mamma-lian cells.

It can be noticed in Fig. 4A that the native or recom-binant prolactin band appeared as a doublet in the 14%SDS–polyacrylamide gel. This could be due to post-translationally modified forms present in the purifiedprolactin preparation. Since human prolactin producedin mammalian cells is known to exist as the glycosy-lated form (14), we were interested in determiningwhether human prolactin produced in insect cells isalso glycosylated. The carbohydrate composition of gly-cosylated prolactin (if any) present in the purified frac-tion from insect cells was analyzed. In parallel, thepurified prolactin fraction from mouse L-cells was alsoanalyzed for comparison. The monosaccharide analysisof the prolactin fractions purified from the Mono Qcolumn is presented in Table 2. The results indicatethat the recombinant human prolactin produced ininsect cells exists as a glycosylated form. The monosac-charide composition of the glycosylated prolactin de-rived from insect cells appears to be similar to thatderived from mouse L-cells, albeit to a lesser extent. Incontrast to earlier reports (14), we could not detect anyfucose moiety in our analysis, even in the mammaliancell expressed prolactin fraction.

DISCUSSION

Prolactin is one of the most versatile hormones of thepituitary in terms of its biological actions; how a singlemolecule exerts different responses in an organism is

TABLE 2

Carbohydrate Analysis of Recombinant Human Prolactin

Monosaccharide

mol/mol PRL

L-cell(hsPRL)Mono Q eluate

Sf9(hsPRL)Mono Q eluate

GlcNAc 0.2 0.13Galactose 0.6 0.14Mannose 0.15 0.07Fucose nda nd

a nd, not detected.

not known. Recent findings on structural polymor-

phism of prolactin lend credence to the hypothesis thatthe molecular heterogeneity of prolactin is one of themechanisms for creating diversity in the biological ac-tions of this hormone. The prolactin used in the struc-tural and biochemical analyses was primarily purifiedfrom different sources such as pituitary, placenta,milk, and other organs. Since the purification of pro-lactin from these sources is cumbersome, the lack ofpurified forms has hindered studies to delineate func-tional differences between these molecules. On theother hand, a major advantage of producing recombi-nant prolactin is its lack of contamination with otherpituitary hormones (e.g., GH and CS) that may inter-fere in immunoassays as well as functional assays.Moreover, the expression of different forms of prolactincan be regulated by genetic manipulation, thereby un-derscoring the utility of recombinant prolactin instudying the structure–function relationship of thisimportant hormone. So far, the recombinant expres-sion of human prolactin has been achieved in E. coli(12, 13) and in mammalian cells (14). Although theposttranslational modifications have been observedwith prolactin expressed in mouse C127 cells (14), thelimitation of the stable transfection system is that theamount of protein produced (5–10 mg/L of culture) isnot sufficient for purification in large quantities. Also,there are regulatory safety issues for the commercialuse of recombinant proteins purified from mammaliancells. Moreover, genetic manipulation is cumbersomeas well as time-consuming for stable expression sys-tems. To overcome these disadvantages, we chose thebaculovirus expression system (BEVS) since this ex-pression system produces foreign proteins in largequantity and, in most cases, the recombinant proteinsare processed and modified in a manner similar to thatof their authentic counterparts (15, 16).

Here, we have described a protocol for the synthesisof biologically active recombinant human prolactin inlarge amounts in insect cells using the baculovirusexpression system and subsequent purification andcharacterization. The level of recombinant prolactinproduced by insect cells was greater than that obtainedin mammalian cells and found to be as active as itsnative counterpart (Fig. 3). The quantitative analysispresented in Table 1 clearly indicates that the intra-cellular production is more efficient (;40 mg/L of cul-ture) than extracellular production (;20 mg/L of cul-ture medium). To rule out the possibility that in thecase of extracellular production 50% prolactin re-mained within the cells, we analyzed the cell pellet andfound that only 15–20% prolactin was retained withinthe cells (data not shown). Thus, the extracellular ex-pression level was 30–40% lower than the level whenthe same protein was produced intracellularly. How-ever, the major advantage of extracellular production

is the ease of purification of the secreted protein from

ttimdttctmltgo(tohis

cmrpa

1

1

272 DAS ET AL.

the serum-free culture medium. It can be noticed inFig. 4A that a significant purification can be achievedby a single Mono Q column chromatography of theculture medium. Recently, an affinity purification ofhuman prolactin from pituitary extract has been de-veloped using a heparin column (26). Therefore, thissingle-step heparin affinity chromatography would beuseful to purify prolactin from insect cell extract. It isimportant to mention that our quantification of theproduction level of recombinant prolactin was based ona monolayer culture. The use of an insect cell suspen-sion culture in a bioreactor generally facilitates theproduction level of a foreign protein compared to thatin a monolayer culture (27). Therefore, using a suspen-sion culture in a bioreactor, the production level ofprolactin could be increased severalfold over the levelreported here using a monolayer culture.

The interesting observation is that the recombinanthuman prolactin, when expressed in insect cells withits own secretory signal, was processed and secretedinto the culture medium (Fig. 2). These results obvi-ously support the contention that prolactin produced ininsect cells undergoes processing and modificationssimilar to those found in mammalian cells or in humanpituitary. Moreover, the results of carbohydrate anal-ysis indicate that a portion of the recombinant humanprolactin produced in insect cells exists as a glycosy-lated form (Table 2). Based on the results presented inTable 2, the molar ratio of monosaccharides in a singlechain appears to be mannose1:GlcNAc2:galactose2 andhe predicted amount of glycosylated form present inhe Mono Q fraction of prolactin purified from thensect cells seems to be around 7%, assuming that a

onomer of mannose is present in a single carbohy-rate chain. However, in Fig. 4A, it can be noticed thathe slower migrating band in the doublet is more in-ense than the faster one. This slower migrating bandould be a mixture of various modified forms of prolac-in, e.g., glycosylated and phosphorylated, or the fasterigrating band could be a C-terminal truncated pro-

actin (since the N-terminal sequence is intact as de-ermined by sequence analysis). In the pituitary gland,lycosylated prolactin comprises a variable percentagef the total prolactin, ranging from 5 to 40% in humans4). The ratio of nonglycosylated to glycosylated prolac-in in human serum varies according to certain physi-logical states and potential disease states (14). Thus,igh-level production of bioactive human prolactin in

nsect cells will give us an opportunity to study itstructure–function relationship.The baculovirus expression system described here

an be used for large-scale production of bioactive hu-an prolactin variants, and the availability of these

eagents would provide a means to dissect the role ofrolactin variants in modulating the immune system

nd would serve to promote the study of the physiolog- 1

ical relevance of different variants, particularly prolac-tin fragments. In this regard, it is important to men-tion that the members of the human prolactin/growthhormone family; i.e., human prolactin, human growthhormone, and human placental lactogen, are angio-genic whereas their respective 16-kDa N-terminalfragments are antiangiogenic (11). The opposite ac-tions are regulated in part via activation or inhibitionof the mitogen-activated protein kinase signaling path-way. The concept that a single molecule encodes bothangiogenic and antiangiogenic peptides represents anefficient model for regulating the balance of positiveand negative factors controlling angiogenesis. This hy-pothesis has potential physiological importance for thecontrol of the vascular connection between the fetaland maternal circulations in the placenta, where hu-man prolactin, placental lactogen, and growth hor-mone variants are expressed. Thus, the 16-kDa variantrepresents a potential therapeutic agent for the treat-ment of cancer or other diseases whose etiology neces-sarily involves angiogenesis.

REFERENCES

1. Cooke, N. E. (1989) Prolactin: Normal synthesis, regulation, andactions in “Endocrinology” (DeGroot, L. J., Ed.), pp. 384–407,Saunders, Philadelphia.

2. Murphy, W. J., Rui, H., and Longo, D. L. (1995) Effects of growthhormone and prolactin immune development and function. LifeSci. 57, 1–14.

3. Cooke, N. E., Coit, D., Shine, J., Baxter, J. D., and Martial, J. A.(1981) Human prolactin: cDNA structural analysis and evolu-tionary comparisons. J. Biol. Chem. 256, 4007–4016.

4. Sinha, Y. N. (1995) Structural variants of prolactin: Occurrenceand physiological significance. Endocr. Rev. 16, 354–369.

5. Ellis, L. A., and Picciano, M. F. (1995) Bioactive and immunore-active prolactin variants in human milk. Endocrinology 136,2711–2720.

6. Ellis, L. A., Mastro, A. M., and Picciano, M. F. (1997) Do milk-borne cytokines and hormones influence neonatal immune cellfunction? J. Nutr. 127, 985S–988S.

7. Dusanter-Fourt, L. B., Gespach, C., and Djiane, J. (1992) Expres-sion of prolactin (PRL) receptor gene and PRL-binding sites inrabbit intestinal epithelial cells. Endocrinology 130, 2877–2882.

8. Woody, M. A., Welniak, L. A., Richards, S., Taub, D. D., Tian, Z.,Sun, R., Longo, D. L., and Murphy, W. J. (1999) Use of neuroen-docrine hormone to promote reconstitution after bone marrowtransplantation. Neuroimmunomodulation 6, 69–80.

9. Richards, S. M., Garman, R. D., Keyes, L., Kavanagh, B., andMcPherson, J. M. (1998) Prolactin is an antagonist of TGF-betaactivity and promotes proliferation of murine B cell hybridomas.Cell Immunol. 184, 85–91.

0. Ostrom, K. M. (1990) A review of the hormone prolactin duringlactation. Prog. Food Nutr. Sci. 14, 1–44.

1. Struman, I., Bentzien, F., Lee, H., Mainfroid, V., D’Angelo, G.,Goffin, V., Weiner, R. I., and Martial, J. A. (1999) Opposingactions of intact and N-terminal fragments of the human prolac-tin/growth hormone family members on angiogenesis: An effi-cient mechanism for the regulation of angiogenesis. Proc. Natl.Acad. Sci. USA 96, 1246–1251.

2. Paris, N., Rentier-Delrue, F., Defontaine, A., Goffin, V., Lebrun,

1

273HUMAN PROLACTIN PRODUCTION IN INSECT CELLS

J. J., Mercier, L., and Martial, J. A. (1990) Bacterial productionof recombinant human prolactin. Biotechnol. Appl. Biochem. 12,436–449.

3. Morganti, L., Huyer, M., Gout, P. W., and Bartolini, P. (1996)Production and characterization of biologically active Ala-Ser-(His)6-Ile-Glu-Gly-Arg-human prolactin (tag-hPRL) secreted inthe periplasmic space of Escherichia coli. Biotechnol. Appl. Bio-chem. 23, 67–75.

14. Price, A. E., Logvinenko, K. B., Higgins, E. A., Cole, E. S., andRichards, S. M. (1995) Studies on the microheterogeneity and invitro activity of glycosylated and nonglycosylated recombinanthuman prolactin separated using a novel purification process.Endocrinology 136, 4827–4833.

15. Luckow, V. A. (1995) “Baculovirus Expression Systems and Bio-pesticides” (Shuler, M. L., Wood, H. A., Granados, R. R., andHammer, D. A., Eds.), Wiley–Liss, New York.

16. Luckow, V. A. (1995) “Principles and Practice of Protein Engi-neering” (Cleland, J. L., and Craik, C. S., Eds.), Wiley, New York.

17. Mathur, M., Das, T., and Banerjee, A. K. (1996) Expression of Lprotein of vesicular stomatis virus indiana serotype from recom-binant baculovirus in insect cells: Requirement of a host factor(s)for its biological activity in vitro. J. Virol. 70, 2252–2259.

18. Gout, P. W., Beer, C. T., and Noble, R. L. (1980) Prolactin-stimulated growth of cell cultures established from malignantNb rat lymphomas. Cancer Res. 40, 2433–2436.

19. Kinet, S., Bernichtein, S., Kelly, P. A., Martial, J. A., and Goffin,V. (1999) Biological properties of human prolactin analogs de-pend not only on global hormone affinity, but also on the relativeaffinities of both receptors’ binding sites. J. Biol. Chem. 274,26033–26043.

20. Goffin, V., Kinet, S., Ferrag, F., Binart, N., Martial, J. A., and

Kelly, P. A. (1996) Antagonistic properties of human prolactinanalogs that show paradoxical agonistic activity in the Nb2bioassay. J. Biol. Chem. 271, 16573–16579.

21. Wakao, H., Gouilleux, F., and Groner, B. (1994) Mammary glandfactor (MGF) is a novel factor of the cytokines regulated tran-scription factor gene family and confers the prolactin response.EMBO J. 13, 2182–2191.

22. Bole-Feysot, C., Goffin, V., Edery, M., Binart, N., and Kelly, P. A.(1998) Prolactin and its receptor: Actions, signal transductionpathways and phenotypes observed in PRL receptors knockoutmice. Endocr. Rev. 19, 225–268.

23. Matsudaira, P. (1987) Sequence from picomole quantities of pro-teins electroblotted onto polyvinylidene difluoride membranes.J. Biol. Chem. 262, 10035–10038.

24. Hardy, M. R., Townsend, R. R., and Lee, T. C. (1988) Monosac-charide analysis of glycoconjugates by anion exchange chroma-tography with pulsed amperometric detection. Anal. Biochem.170, 54–62.

25. Tanaka, T., Shiu, R. P. C., Gout, P. W., Beer, C. T., Noble, R. L.,and Friesen, H. G. (1980) A new sensitive and specific bioassayfor lactogenic hormones: Measurment of prolactin and growthhormone in human serum. J. Clin. Endocrinol. Metab. 51, 1058–1063.

26. Khurana, S., Kuns, R., and Ben-Jonathan, N. (1999) Heparin-binding property of human prolactin: A novel aspect of prolactinbiology. Endocrinology 140, 1026–1029.

27. Canaan, S., Dupuis, L., Riviere, M., Faessel, K., Romette, J. L.,Verger, R., and Wicker-Planquart, C. (1998) Purification andinterfacial behavior of recombinant human gastric lipase pro-duced from insect cells in a bioreactor. Protein Expr. Purif. 14,

23–30.