Vlak J.M., de Gooijer C.D., Tramper J., Miltenburger H.G. (Eds.) Insect Cell Cultures: Fundamental and Applied Aspects

Подождите немного. Документ загружается.

272

Expression and characterization of E1

Our approach to express E1 with the baculovirus sys-

tem was based on our experiences with expression of

this protein with the PRV vector (Van Zijl et al

1991).

Two versions of the E1 gene of classical swine fever

strain Brescia (Moormann et al

1990), one encoding

E1 with a C-terminal membrane anchor (transmem-

brane region, TMR), and one encoding E1 without a

C-terminal TMR, were inserted into the PRV glycopro-

tein gX locus (Rea et al

1985), and expressed in swine

kidney cells. E1 without TMR was secreted, whereas

E1 with TMR was retained in the membranes of the

endoplasmic reticulum. To express both versions of

the Brescia E1 gene in Sf21 insect cells, the sequences

encoding E1 with and without TMR were fused to the

signal sequence of gX of PRV and inserted into the p10

locus of Autographa californica nuclear polyhedrosis

virus (AcNPV), using pAcAS3 (Vlak et al., 1990) as

transfer vector (Fig. 1). Polyhedrin-positive plaques

expressing -galactosidase were isolated and tested for

expression of E1 by immunostaining with an E1 spe-

cific monoclonal antibody (Wensvoort et al., 1989).

One plaque purified E1-TMR virus (BacE1[-]) and

one plaque purified E1+TMR virus (BacE1[+]) were

selected for further characterization by radio immuno-

precipitation (Fig. 2).

The E1 protein precipitated from the BacE1 [+] cell-

lysate was similar in size to wild-type E1 (compare

lanes 1 and 6; wild-type E1 is a doublet with a molec-

ular mass of 51 to 54 kD (Wensvoort et al., 1990)). As

expected the E1 protein precipitated from the BacE1 [-]

lysate (lane 4) was slightly smaller (49 to 52 kD) than

wild type E1. E1-TMR was secreted from insect cells

(lane 5) with a somewhat lower molecular weight (46

to 48 kD) than cell-associated E1-TMR. No E1+TMR

protein was secreted from insect cells (lane 7).

The N-glycans of the cell-associated E1+TMR

were completely sensitive to endo H, indicating that

similarly as in CSFV infected cells, the E1+TMR pro-

tein is anchored in the membranes of the endoplasmic

reticulum or cis-Golgi region of insect cells (Hulst et

1993). The N-linked glycans of the secreted E1

were partially resistant to endo H, indicating that a

part of the high mannose units were trimmed to a small-

er endo H resistant form (Kuroda et al

1990). This

explains the lower molecular weight of the secreted

E1 -TMR (lane 5) compared to cell associated E1 -TMR

(lane 4).

Similarly as in CSFV infected cells (Wensvoort et

al., 1990; Thiel et al

1991), E1 secreted from insect

cell was efficiently dimerized. Furthermore, E1+TMR

and both the secreted and cell-associated forms of E1 -

TMR reacted identically as native E1 in an ELISA

with four E1 specific monoclonal antibodies, which

each recognize a discontinuous antigenic domain on E1

(van Rijn et al., 1993). This indicated that E1 expressed

in insect cells is antigenically indistinguishable from

native E1.

Production level of E1

The level of expression of E1 in Sf21 cells and in

the medium, determined at different time points after

infection in an E1 specific ELISA (Wensvoort et al

273

1988), showed that the total amount of E1-TMR (cell-

associated plus secreted forms) synthesized in Sf21

cells is about 10 times higher than for E1+TMR

(Fig. 3.) Because no significant growth differences

between BacE1[-] and BacE1[+] virus were observed

and the level of E1 specific RNA present in Sf21 cells

infected with both viruses was almost identical (Fig. 4),

this lower production level of E1+TMR is probably

due to inhibition of protein synthesis as a result of the

accumulation of E1+TMR in the membranes of the

endoplasmic reticulum (see above).

On the basis of this ELISA, it was estimated that

75% of the E1-TMR protein was secreted from insect

cells. The concentration of E1 secreted in the medium,

determined by ELISA using a pure E1 preparation as

standard, was approximately 30 to cells (50

to when Sf21 cells were grown as monolay-

ers in serum-free medium ( SF900; Gibco BRL).

In a pilot experiment, Spodoptera exiqua larvae

were infected with BacE1[-] and BacE1[+] virus. As

expected, only the E1 -TMR protein was secreted in the

hemolymph fluid of the larvae. The production level of

E1-TMR in the hemolymph fluid was approximately

300 to 400 Analysis of the E1 -TMR protein by

SDS-PAGE and Western blotting showed that 50% of

the E1 was C-terminally cleaved probably by a host

protease to a smaller protein with a molecular weight

of 29 to 31 kD (Fig. 5). Because the neutralizing epi-

topes on E1 are positioned in the N-terminal half of

the protein (Van Rijn et al., 1993), E1 produced in

larvae could still be suitable as a subunit vaccine. To

improve production of E1, further experiments with

larvae producing more hemolymph fluid, are needed.

Furthermore, vaccination experiments in pigs with E1

secreted in the hemolymph fluid should be performed

to establish whether E1 produced in larvae is as effi-

cacious in inducing protection against CSFV as E1

produced in insect cells.

274

Vaccination of pigs with E1 produced in insect cells

For the first vaccination trial with E1 produced with the

baculovirus system, E1-TMR purified from the medi-

um of insect cells by immunoaffinity chromatography,

was used (Hulst et al., 1993). Because no data were

available regarding the dose of E1 required for the

induction of a protective immune response in pigs by a

dead subunit vaccine, groups of pigs were vaccinated

with high doses, i.e., 20 and of E1. The vaccine

was applied in a double water-oil emulsion (Herbert,

1965; Barterling & Vreeswijk, 1991). After vaccina-

tion all pigs developed high neutralizing antibody liters

against classical swine fever virus (Table 1). In fact, no

significant differences in liters were observed between

pigs vaccinated with 20 or of E1 42 days after

inoculation. After intranasal challenge with 100

(50% lethal dose) of classical swine fever virus strain

Brescia, all pigs vaccinated with E1 were complete-

ly protected and did not show any signs of disease or

viraemia. In contrast, the non-vaccinated pigs devel-

oped fever from day 4 and 5 on, became recumbent at

day 6, and were killed when moribund at day 8 post

challenge.

The results of this limited vaccination experiment

with E1 are highly promising in several aspects. Firstly,

after vaccination with the conventional Chinese (C)-

strain vaccine, pigs with neutralizing antibody titers of

30 or higher are protected from classical swine fever,

and do not transmit the virus to seronegative contact

pigs, after challenge with virulent CSFV (Terpstra et

1988). Pigs inoculated once with of E1

produced in insect cells, develop neutralizing antibody

titers of 3000 and higher, indicating that a much lower

dose of E1 might be sufficient to induce protection

against CSF. Preliminary results with lower doses of

E1, directly prepared from serum-free SF900 insect

cell medium, confirm that the dose needed to protect

pigs, and to prevent transmission of classical swine

fever virus, is significantly lower than E1 (the

50% protective dose is about E1).

Secondly, the rise of antibody titers against E1 pro-

ceeded much faster in pigs vaccinated with E1 pro-

duced in insect cells (dead E1) than in pigs infected

275

with low-virulent field strains of classical swine fever

virus or with the PRV-E1 recombinant viruses (Van Zijl

et al

1991; G. Wensvoort, personal communication).

All these observations indicate that E1 expressed by the

baculovirus-insect cell system may be a very effective

dead subunit vaccine.

Thirdly, pigs infected with field virus can be dis-

criminated from pigs vaccinated with E1 on the basis of

presence (field virus infection) or absence (vaccination

with E1) of antibodies directed against other immuno-

genic structural or non-structural viral proteins. Such a

discriminating serological test has to be developed and

should accompany the E1 subunit vaccine in a vaccina-

tion campaign for a controlled eradication of classical

swine fever during outbreaks of the disease.

Application of CSFV proteins expressed with a

baculovirus vector in diagnostic tests

The pestiviruses CSFV, BVDV and BDV are struc-

turally, antigenically, and genetically, closely related.

The genomic organization of these viruses are similar.

Sera induced in pigs infected with CSFV and in pigs

and ruminants infected with BVDV, cross-react in neu-

tralization assays (Wensvoort et al

1989). Therefore,

a serological test which specifically detects antibod-

ies in field sera directed against CSFV specific epi-

topes, is essential to differentiate between pigs infected

with CSFV and pigs infected with BVDV or BDV. For

CSFV, such a serological test (Wensvoort et al., 1988)

is available commercially for several years already. In

this diagnostic ELISA, which detects antibodies in field

sera, which are specifically directed against conserved

(present on all CSFV strains) epitopes on E1 of CSFV

(Wensvoort, 1989), native E1 was used as antigen.

This antigen, however, has now been replaced by E1

synthesized in insect cells. Replacement of native El

appeared to have several advantages. Instead of inten-

sive large scale isolation of native E1 from porcine

kidney cells infected with CSFV, serum-free medium

of Sf21 cells infected with BacE1[-], grown in small

culture flasks (75 ), can be used directly in this

assay. Because of the constant quality of the E1 pro-

duced in insect cells, 75% less inconclusive results with

field sera were scored in this ELISA. Furthermore, E1

synthesized in insect cells improved the sensitivity of

the test (R. Bloemraad, personal communication).

Animals infected with pestiviruses also raise anti-

bodies against envelope glycoprotein E2 (Kwang et al.,

1992). Glycoprotein E2 of CSFV, recently identified as

a ribonuclease, can also be expressed in large amounts

in insect cells (Hulst et al

1994). The ribonuclease

specific activity of E2 produced in insect cells is com-

parable to that of native E2 (Schneider et al

1993),

indicating that the conformation of E2 synthesized in

insect cells is indistinguishable from native E2. There-

fore, E2 synthesized in insect cells could also be suit-

able as antigen in an serological test for the detection

of field infections. Preliminary results in our laborato-

ry show that such a test can indeed be developed on

the basis of E2. Nevertheless, before a reliable, sensi-

tive, and CSFV specific serological test based on E2

produced in insect cells can be employed commercial-

276

ly, monoclonal antibodies, specifically directed against

conserved epitopes (present on all CSFV strains) on

E2 of CSFV, have to be produced. For the production

of these monoclonal antibodies, properly synthesized,

and processed viral proteins synthesized in insect cells

should be very suitable.

Because animals infected with pestiviruses also

raise antibodies to the non-structural viral protein, p80

(Donis & Dubovi, 1987), a diagnostic test for the detec-

tion of field infections could also be developed on the

basis of this protein. Using p80 expressed in insect

cells as antigen in an ELISA, Petric et al. (1992) and

Vanderheijden et al. (1993), showed that anti-BVDV

antibodies could be detected in field sera from cattle

infected with BVDV. Because the degree of homolo-

gy between the amino acid sequences encoding p80

of BVDV, BDV and CSFV is high (more than 86%;

Moormann et al

1990), sera from pigs infected with

BVDV and BDV cross-react with CSFV p80. There-

fore, a serological test on the basis of p80 will not

be suitable to differentiate between pigs infected with

CSFV, BVDV and BDV.

Concluding remarks

The baculovirus-insect cell expression system has pro-

vided us with properly synthesized envelope proteins

E1 and E2 of CSFV. Because the production level in

insect cells of these proteins is high, we are able to

apply these expression products for various purposes.

Nevertheless, in the literature their are many exam-

ples that the production levels of viral proteins in the

baculovirus insect cell system can be poor. In our lab-

oratory, the production levels of the structural proteins

of porcine reproductive and respiratory syndrome virus

(Lelystad virus) in insect cells were poor compared to

CSFV E1 and E2 (J. Meulenberg, personal communi-

cation). Instability of the messenger RNAs of the viral

proteins in the nuclei of insect-cells or instability of the

proteins themselves, as a result of a different process-

ing or routing in insect cells compared to mammalian

cells, may be responsible for this low level of pro-

duction. It therefore should be recognized that RNA

viruses like CSFV and Lelystad virus replicate in the

cytoplasm of the mammalian cell. For the production of

proteins of these kind of viruses, eukaryotic expression

systems based on vectors replicating in the cytoplasm,

like the semliki forest virus vector system (Liljeström

& Garroff, 1991), or the vaccinia virus vector system

(Mackett et al

1982), might be more suitable.

Acknowledgments

We thank Just Vlak and John Martens for providing the

Spodoptera exiqua larvae and for their help to infect

the larvae with the baculovirus-E1 recombinants. The

data of the diagnostic ELISA using E1 were kindly

provided by Rinus Bloemraad.

Addendum in proof

In this paper we still used the old pestivirus protein

nomenclature. A new nomenclature of these proteins

will, however, be proposed to the International Com-

mittee on Taxonomy of Viruses by the Flaviviridae

Study Group. In recent papers this new nomenclature

has already been used by several research groups. The

nomenclature of pestivirus proteins used in this paper

compared to the new one is as follows: E1 is renamed

E2, E2 is renamed , E3 is renamed El, and p80 is

renamed NS3.

References

Barteling S & Vreeswijk J (1991) Developments in foot and mouth

disease vaccines. Vaccine 9: 75–88.

Donis R & Dubovi E (1987) Molecular specificity of the antibody

responses in cattle naturally and experimentally infected with

cytopathic and noncytopathic bovine viral diarrhea virus bio-

types. American Journal of Veterinary Research 48: 1549–1554

Francki R, Fauquet C, Knudson D & Brown F (1991) Flaviviridae

Arch. Virol. Suppl. 2: 223–233.

Herbert W (1965) Multiple emulsions: A new form of mineral-oil

antigen adjuvant. Lancet 11: 771

Hulst M, Westra D, Wensvoort G & Moormann R (1993) Glyco-

protein E1 of hog cholera virus expressed in insect cells protects

swine from hog cholera. J. Virol. 67: 5435–5442.

Hulst M, Himes G, Newbigin E & Moormann R (1994) Glycoprotein

E2 of classical swine fever virus: expression in insect cells and

identification as a ribonuclease. Virology 200: 558–565.

Kuroda K, Geyer H, Geyer R, Doerfler W & Klenk H (1990) The

oligosaccharides of influenza virus hemagglutinin expressed in

insect cells by a baculovirus vector. Virology 174: 418–429.

Kwang J, Littledike E, Donis R & Dubovi E (1992) Recombinant

polypeptide from the gp48 region of the bovine viral diarrhoea

virus (BVDV) detects serum antibodies in vaccinated and infected

cattle. Vet. Microbiol. 32: 281–292.

Liljeström P & Garoff H (1991) A new generation of animal cell

expression vectors based on the Semliki forest virus replicon.

Biotechnology 9: 1356–1362.

Mackett M, Smith G & Moss B (1982) Vaccinia virus expression

vector. PNAS 79: 7415–7419.

Moennig V & Plagemann (1992) The pestiviruses. Adv. Virus Res.

41:53–98

Moormann R, Warmerdam P, van der Meer B, Schaaper W,

Wensvoort G & Hulst M (1990) Molecular cloning and nucleotide

277

sequence of hog cholera virus strain Brescia and mapping of the

genomic region encoding envelope protein El. Virology 177:

184–198.

Petric M, Yolken R, Dubovi E, Wiskerchen M & Collett M (1992)

Baculovirus expression of pestivirus non-structural proteins. J.

Gen. Virol. 73: 1867–1871.

Rea T, Timmins G, Long G & Post L (1985) Mapping and sequence

of the gene for the pseudorabies virus glycoprotein which accu-

mulates in the medium of infected cells. J. Virol. 54: 21–29.

Schneider R, Unger G, Stark R, Schneider-Scherzer E & Thiel H-J

(1993) Identification of a structural glycoprotein of a RNA virus

as a ribonuclease. Science 261: 1169–1171.

Terpstra C & Wensvoort G (1988) The protective value of vaccine

induced neutralizing antibody titres in swine fever. Vet. Microbi-

ol. 16: 123–128.

Thiel H-J, Stark R, Weiland E, Rümenapf T & Meyers G (1991) Hog

cholera virus: Molecular composition of virions from a pestivirus.

J. Virol. 65: 4705–4712.

Vanderheijden N, De Moerloze L, Vandenbergh D, Chappuis G,

Renard A & Lecomte C (1993) Expression of the bovine viral

diarrhoea virus Osloss p80 protein: its use as antigen for cattle

serum antibody detection. J. Gen. Virol. 74: 1427–1431.

Van Rijn P, van Gennip R, de Meyer E & Moormann R (1993)

Epitope mapping of envelope glycoprotein El of hog cholera

virus strain Brescia J. Gen. Virol. 74: 2053–2060

Van Zijl M, Wensvoort G, de Kluyver E, Hulst M, van der Gulden H,

Gielkens A, Berns A & Moormann R (1991) Live attenuated

pseudorabies virus expressing envelope glycoprotein El of hog

cholera virus protects swine against both Pseudorabies and Hog

Cholera. J. Virol. 65: 2761–2765.

Vlak J, Schouten A, Usmany M, Belsham G, Klinge-Roode M,

Maule A, van Lent J & Zuidema D (1990) Expression of

cauliflower mosaic virus gene I using a baculovirus vector based

upon the p10 gene and a novel selection method. Virology 179:

312–320.

Wensvoort G, Terpstra C & de Kluyver E (1989) Characterization of

porcine and some ruminant pestiviruses by cross-neutralization.

Vet. Microbiol. 20: 291–306.

Wensvoort G, Bloemraad R & Terpstra C (1988) An enzyme immuno

assay employing monoclonal antibodies and detecting specifical-

ly antibodies to classical swine fever virus. Vet. Microbiol. 17:

129–140.

Wensvoort G (1989) Topographical and functional mapping of epi-

topes on hog cholera virus with monoclonal antibodies. J. Gen.

Virol. 70: 2865–2876.

Wensvoort G, Boonstra J & Bodzinga B (1990) Immuno-affinity

purification of envelope protein El of hog cholera virus. J. Gen.

Virol. 71: 531–540.

Address for correspondence: Rob J.M. Moormann, Department of

Mammalian Virology, Institute for Animal Science and Health (ID-

DLO), P.O. Box 65, NL–8200 AB Lelystad, the Netherlands.

This page intentionally left blank.

Cytotechnology

20:

279–288, 1996.

279

Production of multidomain complement glycoproteins in insect cells

Péter Závodzky & Sándor Cseh

Institute of Enzymology, Biological Research Center, Hungarian Academy of Sciences, Budapest, Pf. 7., H-1518,

Hungary

Key words: baculovirus, complement activation, genetic engineering, mosaic protein, serine-protease, zymogen

Abbreviations: AcMNPV – Autographa californica nuclear polyhedrosis virus; C1 – first component of comple-

ment; Clq, Clr, Cls – subcomponents of Cl; CCP – complement control protein; EGF – epidermal growth factor;

High5 – Trichoplusia ni cell line; SCR – short concensus repeat; Sf9 – Spodoptera frugiperda 9 cell line.

Introduction

The scarcity of sources makes the isolation of most

human proteins, peptides and proteinaceous materi-

als impossible or impractical. On the other hand there

exists a great demand for complex mammalian gly-

coproteins for research, diagnosis or therapy. Glyco-

proteins of various complexity have been expressed

in various laboratories (see proceddings of the Work-

shop on baculovirus and recombinant protein produc-

tion processes, edited by Vlak et al

1992) using the

baculovirus expression system. Since insect cells do

almost all of the posttranslational modifications asso-

ciated with mammalian cells, the expressed proteins

are expected to be near-authentic concerning their bio-

logical activity and antigenicity. To produce complex,

glycosylated and processed proteins we have the option

of choosing from a number of different expression sys-

tems, each of which has a particular inherent set of

advantages and disadvantages. Bacterial systems pro-

vide important advantages, like the ease of use, low

cost and high level of expression, but impose a num-

ber of limitations for synthesis of eukaryotic proteins

(folding, proteolytic processing, glycosylation, secre-

tion, subunit assembly, etc.). For this reason eukaryotic

cells are preferred for expression of mammalian genes.

There are several host-vector systems for heterologous

gene expression in yeast or mammalian cells. The dis-

advantages of the mammalian expression systems are

the high cost and complexity of the cell culture main-

tenance and the difficulty of manipulating large viral

genomes, as well as the relatively low level of expres-

sion, that might be a high price for the correct folding

and full biological activity. Insect cell based expres-

sion systems are advantageous, since they are cost

effective, simple in handling while performing glyco-

sylation, fatty acid acylation, phosphorylation, disul-

fide formation, C-terminal amidation, hydroxylation

of aspartate residues, signal peptide cleavage, folding

and secretion. Therefore insect cells seem to be suit-

able for efficient expression of recombinant complex

mammalian glycoproteins in biologically active form.

Because of its complexity human complement

protein, Cl is an ideal object to test the potential

of baculovirus-insect cell expression for producing

sophisticated mammalian proteins.

Complement is a complex multicomponent system

having many similarities with blood coagulation and

fibrinolysis. The classical pathway of complement acti-

vation is initiated by the reaction of the first component,

Cl, with immune complexes. This recognition leads to

removal of the invading bacteria, virus or toxin.

The first component of complement is a supramole-

cular complex composed of five subcomponents: one

Clq, two Clr and two Cls. A specific inhibitor protein,

Cl-inhibitor, regulates the activation and function of

the C1 complex (reviewed by Arlaud et al

1987; Schu-

maker et al

1987). Cl is a complex multifunctional

molecule, and the complexity of its function and reg-

ulation is also reflected in the structure of its subunits:

Clq is the structural framework of the Cl complex

(Reid & Porter, 1976; Borsos et al

1980) and its inter-

esting flower boquet like structure is shown in Figure

1 and Figure 2. The dependent C1sC1rC1rC1s

280

tetramer is responsible for the enzymatic activity of

C1 (Tschopp et al., 1980). C1r and C1s are zymogen

serine-proteases in the inactive C1 and the tetramer is

folded around the collagenous stems of C1q (Figure 1).

When C1 binds – through the C1q heads – to immune

complexes, the cone formed by the spreading arms will

be distorted. This distortion might be the activitation

signal that induces a conformational change convert-

ing the C1r proenzyme into an enzyme. The activated

C1r in turn activates C1s by the cleavage of an Arg-Ile

bond. The following initiation of the complement cas-

cade is sustained through the proteolytic action of C1s

on C4 and C2.

The C1r and C1s zymogens are single chain gly-

coproteins (86–78 kD) containing 688 (C1r) and 673

(C1s) amino acid residues (Leytus et al

1986; Tosi

et al

1987). There is strong relationship between the

two molecules: sequence comparison revealed 40%

amino acid identity and conservation of all the cys-

teine residues (Tosi et al., 1987). Both C1r and C1s

have been cloned and sequenced and have mosaic like

structures typical of plasma serine-proteases (Leytus

et al

1986; Tosi et al

1987; Mackinnon et al

1987;

Patthy et al

1994). Both C1r and C1s can be divid-

ed into six structural motifs (domains or modules),

including two pairs of internal repeats (I/III and IV/V),

a single copy of motif II and the trypsin like serine-

protease module (Figure 3).

Modules I and III are homologous domains, they

may be involved in a function that is highly specif-

ic to C1r and C1s (for example: Clr-Cls interaction,

tetramer-C1q binding). Module II shows homology

with epidermal growth factor (EGF). In general such

domains are exposed to the extracellular environment

and participate in protein-protein interactions. This

domain harbours an unusual amino acid, hydroxy

asparagine in position 150 in C1r and 134 in C1s. The

level of hydroxylation in native human C1r is 100%

281