Expression of the measles virus nucleoprotein gene in Escherichia coli and assembly of nucleocapsid-like structures

Gene, 160 (1995) 173-178 © 1995 Elsevier Science B.V. All rights reserved. 0378-1119/95/$09.50

173

GENE 08945

Short Communications

Expi'ession of the measles virus nucleoprotein gene in Escherichia coli and assembly of nucleocapsid-like structures (Measles; recombinant protein; paramyxovirus; expression system)

Alan Warnes, Anthony R. Fooks, A. Barry Dowsett, Gavin W.G. Wilkinson * and John R. Stephenson Centre for Applied Microbiology and Research, Porton Down, Salisbury SP4 0JG, UK Received by G. Chinnadurai: 29 August 1994; Revised/Accepted: 9 November 1994; Received at publishers: 27 March 1995

SUMMARY

To investigate the use of fusion systems to aid the purification of recombinant proteins for structure/function studies and potential uses as diagnostic reagents, the measles virus (MV) gene encoding the nucleoprotein was cloned and expressed in Escherichia coli in three forms: as a full-length intact protein and as two fusion proteins. Expression of the intact N gene under the control of the tac promoter in the pTrc99c plasmid produced a protein of the correct size (60 kDa) which represented approx. 4% of the total cellular protein, and was recognised by known measles positive human sera. 'Herringbone' structures characteristic of paramyxovirus nucleocapsids (NuC) were identified in fractured cells examined by electron microscopy. The production of NuC-like structures in a prokaryotic cell indicates folding of the nucleoprotein can occur in the absence of MV genomic RNA, other MV-encoded gene products and eukaryotic cell proteins or RNA, to produce structures which are morphologically and antigenically similar to those seen in virusinfected cells. Conversely, synthesis of N protein as a fusion protein with either E. coli 13-galactosidase or the E. coli maltose-binding protein resulted in the production of fused proteins which could not be assembled into NuC-like structures or readily used as diagnostic reagents. However, the ability of MV N protein to form NuC-like structures in E. coli will facilitate structure/function and mutational analysis of the NuC protein.

INTRODUCTION

Although successful vaccination programmes have had a major impact in controlling MV-associated disCorrespondence to: Dr. J.R. Stephenson, Research Division, Centre for Applied Microbiology and Research, Porton Down, Salisbury, Wiltshire SP4 0JG, UK. Tel. (44-1980) 612-100; Fax (44-1980) 611-096. * Current address: Department of Medicine, University of Wales College of Medicine, Heath Park, Cardiff CF4 4XW, UK. Abbreviations: A, absorbance (1 cm); 13Gal, 13-galactosidase; E., Escherichia; ELISA, enzyme-linkedimmunosorbent assay; EM, electron microscopy; MBP, maltose-binding protein; MV, measles virus; N, gene encoding measles nucleoprotein; NuC, nucleocapsid(s); PAGE, polyacrylamide-gel electrophoresis; PCR, polymerase chain reaction; re-, recombinant; SDS, sodium dodecyl sulfate; SSPE, subacute sclerosing panencephalitis; [], denotes plasmid-carrier state. SSDI 0378-1119(95)00227-8

ease, measles remains a major public health problem in many countries. Measles is particularly important in developing countries where vaccination programmes have been less successful. The WHO estimates MV to be responsible for over a million deaths annually (WHO, 1993), particularly in malnourished children. Even in developed countries, the occurrence of sporadic outbreaks of measles has elicited concern over the longevity of protection induced in vaccinees (Matson et al., 1993). A need has been identified to improve serological tests for MV in order to monitor the immunological status in large populations for seroconversion, primarily for vaccine surveillance, and also increasingly for routine diagnosis of acute disease. Since the MV N protein is a major immunogen, we set out to express this protein in several prokaryotic vector systems to facilitate structure/

174 function analysis and to produce antigen as a basis for an immunoassay. MV N protein forms nucleocapsids (NuC) which enshroud the viral RNA, and when viewed by EM exhibit a helical structure which has a characteristic 'herringbone' appearance (Nakai et al., 1969; Finch and Gibbs, 1970). However, unlike other negative-strand RNA viruses where capsid assembly has been shown to be dependent upon the presence of the viral genome (Blumberg et al., 1983; Baker and Moyer, 1988), the MV N protein folds into its mature form in the absence of viral RNA sequences (Gombart et al., 1993); and typical nucleocapsid-like structures are produced when the MV N protein is expressed in mammalian cells using a vaccinia virus vector (Spehner et al., 1991) and in insect cells using a baculovirus vector (Fooks et al., 1993). MV N protein has also been produced as a fusion with staphylococcal protein A in E. coli, to map B-cell antigenic determinants (Buckland et al., 1989); however, such a protein would be unsuitable for use as a diagnostic reagent as protein A binds to most human immunoglobulins. The MV N fusion protein, with [3Gal, was used to analyse the CD4 ÷ T-cell response in Lewis rats (Reich et al., 1992), by excision from PA gels before injection, although no analysis was performed on the final product. We have, therefore, investigated the potential for using fused systems in the expression of MV N protein for viral diagnosis and to elucidate whether eukaryotic proteins or RNA are necessary for folding and assembly of MV NuC structures.

(Boehringer-Mannheim) to permit inframe synthesis of N as a [~Gal fusion protein (pMV71). Expression of the fused gene product was detected by SDS-PAGE of E. coli cell extracts and immunoblot analysis. Two major protein species of 172 and 140 kDa were identified on gels stained with Coomassie brilliant blue, which were not present in control samples (Fig. 1); the larger band corresponded to the predicted size of the intact fused protein. SDS-PAGE of centrifuged extracts

1

2 i~!~i~i~ ~ i~ ~ ~ ~

BN-

k Da -

205

-

116

-

80

-

50

EXPERIMENTALAND DISCUSSION (a) Production of M V N as a fusion protein A complete and sequenced clone containing the MV (human 2 strain) N gene, was kindly provided by Prof. B.K. Rima and was used throughout this investigation. The pEX vector system (Stanley and Luzio, 1984) is designed to produce high-level expression from open reading frames as in frame fusion proteins with [3Gal. Expression is driven by the strong )~ phage PR promoter which in the E. coli host strain AH1 (Castellazi et al., 1972) is subject to transcriptional repression by an integrated copy of the )~clts857 gene. High level expression of the cloned gene can be induced by inactivating the temperature sensitive repressor at 42°C. Most of the N gene coding region was excised from pMV58 (Fooks et al., 1993) on a convenient BamHI-PstI fragment (BamHI cleaves 67 bp downstream from the start codon and PstI cleaves 4 bp downstream from the structural gene) and inserted between the B a m H I and PstI sites in pEX2

Fig. 1. Production of the [3Gal fusion protein. Plasmids pMV71 and pEX2 were transformedinto E. coli AH1 bacteria and grown at 30°C to an A54o of 0.9, and expressioninduced by incubationat 42°C. Samples of the cultures were taken after 1 h, treated as described by the pMAL protocol (as recommendedby the manufacturers)and subjectedto 0.1% SDS 10% PAGE and stained with Coomassie blue (Laemmli, 1970). Lanes: 1, E. coli[pMV71 ]; 2, E. coli[pEX2]; the migrationof molecular mass standards is indicated on the right margin (in kDa). BN is the I]Gal fusion protein.

175 of sonicated cells indicated that most of the fusion protein was produced in a form that was insoluble, even in 8 M urea (data not shown). Corresponding bands were seen on immunoblots probed with known human measles sera, which confirmed the antigenicity of the final products. EM analysis of expressing cells demonstrated the appearance of inclusion bodies which correlated with the N protein being insoluble, but NuC-like structures could not be identified (data not shown). While large amounts of the 13Gal fusion protein could readily be produced using this expression system, most of the N protein was insoluble, presumably due to its association with inclusion bodies. The relative insolubil•ity of the N protein produced from these vectors hinders its potential to be exploited as a source of antigen. As the M B P fusion system has been used successfully to provide high level expression of a range of soluble fusion protein products (di Guan et al., 1988), we used this as our second system of choice. The B a m H I - P s t I fragment, used in the construction of pMV71, was this time inserted into the plasmidpMALc to generate pMV73 and transformed into the protease-deficient host E. coli BL21 (DE3) (Studier et al., 1990). Expression from the tac promoter was induced with 0.5 m M I P T G for 1 h at 37°C and detected by PAGE as described earlier. No inducible bands could be observed directly on PA gels when compared to a negative control (results not shown). In immunoblotting experiments with SSPE sera, a number of antigenic proteins were identified at and below the predicted 96-kDa size, suggesting that the fused protein was susceptible to proteolysis, even though a protease-deficient host was used. Although apparently degraded, the fusion protein could be purified by affinity chromatography on maltose columns producing low levels of product in a soluble form. However, EM analysis did not reveal the presence of NuC-like structures or inclusion bodies in the host organism (results not shown).

an additional single protein species (60 kDa) could be detected which was the same size as intact MV N protein (Fig. 2a). This protein species represented about 4% of the total soluble proteins in the cell when stained gels were analysed by scanning densitometry. Analysis by immunoblotting and probing with SSPE sera identified a major antigenic 60-kDa protein (Fig. 2b), corresponding to the band seen on the Coomassie-stained gel. Immunoprobing also demonstrated that production reached a maximum by 2 h post induction and that there was no evidence of significant proteolysis. Separation of soluble and insoluble material by low-speed centrifugation demonstrated that the protein was in the soluble fraction (data not shown). A comparison of the N protein produced in the different expression systems and from MV-infected cells (positive control) can be seen in Fig. 3, which highlights the levels of proteolysis occurring in the pMAL system.

A

2

1

B

1

2

3

4

k Da

k Da - 106 - 80

-- 8 0

N-N -

-- 5 0 -

49.5

-

32.5

--33

-

~

27.5

i¸

(b) Synthesis of full-length M V N

As the fusion expression systems failed to produce a product of the desired quality, the pTrc99c expression plasmid was used (Amann et al., 1988), which places expression of the desired gene under the control of the strong tac promoter. Prior to cloning the N gene into the pTrc99c vector (Pharmacia) between the N c o l and P s t I sites, a N c o I site was inserted at the point of translational initiation to generate the plasmidpMV75: the gene sequence was confirmed by direct sequencing,. The plasmid was transformed into a protease deficient host (E. coli BL21(DE3)) where expression was induced with 0.5 m M I P T G for 1 h at 37°C. When an extract from cells containing pMV75 was compared with an extract of cells containing pTrc99c on a Coomassie-stained PA gel,

Fig. 2. Expressionof intact MV N gene.(A) As a first step to expressing the full-length N gene, the 5' end of the gene was tailored by PCR so as to insert an NcoI site at the translational start site, thus ensuring that upstream MV non-coding sequences could not interfere with prokaryotic transcription or translation (Fooks et al., 1993). PlasmidspMV75 and pTrc99c were transformed into E. coli BL21(DE3) and subsequently grown and induced. After induction, samples of the cultures were taken after 1 h and processedas described in the legend to Fig. 1 and subjectedto SDS-PAGE(see Fig. 1). Lanes: 1, E. coli[pMV75]; 2, E. coli[pTrc99c]. (B) Immunoblot analysis of expression from pMV75. After fractionation on SDS-PAGE (see A), samples were electrophoreticallytransferred to nitrocellulose membranes, and processed as describedby Warnes et al. (1994), except the time of induced expression was varied. Lanes: 1, E. coli[pTrc]; 2, E. coli[pMV75] after 1 h induction;3, E. coli[pMV75] after 2 h induction; 4, E. colilpMV75].

176 1

2 3 4 5 6

7 K Da

-106

N

-

80

-

49.5

-

32.5

-

27.5

-

18.5

u

Fig. 3. Comparison of all three expression systems using immunoblotting. Samples ofpMV71, pMV73 and pMV75, including negative controls of vectors without the MV N protein and the positive control of MV-infected Vero cells were analysed as described by Fooks et al. (1993). Immunoblotting was performed as described in the legend to Fig. 2B. Lanes: 1, E. coli AH1 [pEX2]; 2, E. coli AHI[pMV71]; 3, E. coli BL21(DE3)[pMAL]; 4, E. coli BL21(DE3)[pMV73]; 5, E. coil BL21 (DE3)[pTrc99c]; 6, E. coli BL21(DE3) [pMV75]; 7, MV-infected Vero cells. II

EM analysis of E. coli containing the intact N protein encoded by pMV75 and fractured by sonication, demonstrated the presence of typical MV NuC 'herringbone' structures (Fig. 4). These NuC-like structures, although similar to those seen in MV-infected cells, were much

Fig. 4. EM of E. coli containing the MV N protein, synthesised as a full-length intact form. Samples of E. coli were taken 1 h after induction and centrifuged at 3000 x g for 5 rain, the pellet being resuspended in lysis buffer (10raM NazHPO4, pH7.0/30mM NaC1/0.25% (v/v) Tween-20/10 mM EDTA/10 mM EGTA) and kept at - 2 0 ° C for 18 h. After thawing the sample was sonicated at medium amplitude 14 for 10 s (MSE soniprep), while on ice. The ruptured organisms were deposited onto formvar-carbon filmed, 400 mesh copper specimen grids and negatively stained with 1% (w/v) Na.silicotungstate (pH 6.8). The grids were air dried and examined in a Philips EM 400T EM operated at 80 kV. EM analysis showing the presence of typical herringbone structures which are indicative of MV NuC forms. The bars represent 100 nm.

177 shorter, with a size distribution between 23 to 178 nm, with a mean length of 62 nm, in contrast with a length of 1000 nm reported for MV (Waters et al., 1972). This may be due to physical constraints within E. coli restricting filament elongation or breakage of filaments during the sonication process required to rupture the host organism. The presence of NuC-like structures in a prokaryote indicates that the protein can self-assemble independently of any eukaryotic function. However, the assembly of these structures could be dependent on an interaction with RNA or mRNA from the prokaryotic host, although this would be non-specific. Other workers have mentioned this possibility (Fooks et al., 1993; Meric et al., 1994), although this hypothesis has yet to be confirmed. The re-protein structures produced in E. coli were analysed using CsC1 density-gradient centrifugation to determine the buoyant density, as described by Waters et al. (1972). Analysis of fractions from the gradient identified a peak of protein concentration (as estimated by A28o) at a buoyant density of 1.28 g/ml (using an Abbe refractometer), which coincided with the highest concentration of N protein when fractions were analysed by immunoprobing (results not shown). Thus the buoyant density of prokaryotic re-NuC-like structures was similar to that of NuC produced by insect cells (Fooks et al., 1993), although slightly lower than the value of 1.31 g/ml obtained for NuC formed in MV-infected cells (Waters et al., 1972). The lower buoyant density of re-NuC could be due to the absence of RNA, or a reduction in the RNA content. The assembly into NuC-like structures is strong evidence that the N re-protein is synthesised in a native form, is folded appropriately and can perform at least one biological function, i.e., oligomerisation. In order to form polymeric structures, MV N protein must contain at least two sites involved in protein:protein interactions. Furthermore, it will also be possible now to test the biological activity of the purified prokaryotic N protein using an in vitro transcription assay (Blumberg et al., 1983). Genetic modifications can readily be performed on plasmids to rapidly generate insertion, deletion and point mutants. Expression in E. coli potentially provides a powerful technology with which to map biologically important domains within this protein. Additionally, as the N gene is conserved, (Rozenblatt et al., 1985; Cattaneo et al., 1989) and the N protein produces similar structures with other Paramyxoviridae (including mumps virus, respiratory syncytial virus, parainfluenza virus and Sendai virus; Miyahara et al., 1992), the potential exists to test the capacity of elements from the genes encoding the nucleoprotein of other paramyxoviruses to substitute for elements in the MV genome.

(c) Use of full-length N protein derived from E. coil in an ELISA A primary aim in expressing the N gene was to generate an economic, reliable source of antigen for use in an immunoassay. Antigens produced from either fusion system were not suitable for use in a diagnostic assay as they were either insoluble or degraded. Intact, full-length protein, i.e., produced from pMV75, a plasmid derived from the pTrc99c plasmid, was clarified by centrifugation and used in an ELISA for the detection of measles antibodies in a cohort of sera and compared directly with MV N protein produced in insect cells using a baculovirus vector, mammalian cells using a defective adenovirus and extracts from MV-infected cells (Warnes et al., 1994). The results indicated that although the E. coli derived antigen could be used as a diagnostic reagent to detect positive sera there were problems with high backgrounds using the negative control antigen which highlighted the problems with cross reactions between E. coli protein and human sera. There were also high backgrounds with some of the negative control sera. In comparison, the eukaryotic-derived antigen suffered little background effects and also no false positives were generated. This would indicate that there is a potential use of MV N protein produced in E. coli as a diagnostic reagent; however, the protein would have to be purified from the contaminating bacterial antigens and background effects assessed with the appropriate controls. The advantages of using a prokaryotic host to produce MV N protein would be considerable due to the ease of scale-up, and the low costs involved in growing bacteria.

REFERENCES Amann, E, Ochs, B. and Abel, K.-J.: Tightly regulated tac promoter vectors for the expression of unfused and fused proteins in .Escherichia coll. Gene 69 (1988) 301-315. Baker, S.C. and Moyer, S.A.: Encapsidation of Sendai virus genome RNAs by purified NP protein during in vitro replication. J. Virol. 62 (1988) 834-838. Blumberg, B.M., Giorgi, C. and Kolakofsky, D.: N protein of vesicular stomatitis virus selectively encapsidates leader RNA in vitro. Cell 32 (1983) 559-567. Buckland, R., Giraudon, P. and Wild, F.: Expression of measles virus nucleoprotein in Escherichia coli: use of deletion mutants to locate the antigenic sites. J. Gen. Virol. 70 (1989) 435 441. Castellazi, M., Brachet, P. and Eisen, H.: Isolation and characterization of deletions in bacteriophage ~. residing as prophage in E. coli K-12. Mol. Gen. Genet. 117 (1972) 211 218. Cattaneo, R., Schmid, A., Spielhofer, P., Kaelin, K., Baczko, K., ter Meulen, V., Pardowitz, J., Flanagan, S., Rima, B.K., Udem, S.A. and Billeter, M.A.: Mutated and hypermutated genes of persistent measles virus which cause lethal human brain diseases. Virology 173 (1989) 415-425. di Guan, C., Li, P., Riggs, P.D. and Inouye, H.: Vectors that facilitate

178 the expression and purification of foreign peptides in Eschericia coli by fusion to a maltose-binding protein. Gene 67 (1988) 21-30. Finch, J.T. and Gibbs, A.J.: Observations on the structure of the nucleocapsids of some paramyxoviruses. J. Gen. Virol. 6 (1970) 141-150. Fooks, A.R., Stephenson, J.R., Warnes, A., Dowsett, B.A., Rima, B.K. and Wilkinson, G.W.G.: Measles virus nucleocapsid protein expressed in insect cells assembles into nucleocapsid-like structures. J. Gen. Virol. 74 (1993) 1439-1444. Gombart, A.F., Hirano, A. and Wong. T.C.: Conformational maturation of measles virus nucleocapsid protein. J. Virol. 67 (1993) 4133-4141. Laemmli, U.K.: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227 (1970) 680-685. Matson, D.O., Byington, C., Canfield, M., Albrecht, P. and Feigin, R.D.: Investigation of a measles outbreak in a fully vaccinated school population including serum studies before and after revaccination. Pediatr. Infect. Dis. J. 12 (1993) 292-299. Meric, C., Spehner, D. and Mazarin, V.: Respiratory syncytial virus nucleocapsid protein (N) expressed in insect cells forms nucleocapsid-like structures. Virus Res. 31 (1994) 187 201. Miyahara, K., Kitada, S., Yoshimoto, M., Matsumura, H., Kawano, M., Komada, H., Tsurudome, M., Kusagawa, S., Nishio, M. and Ito, Y.: Molecular evolution of human paramyxoviruses. Arch. Virol. 124 (1992) 255-268. Nakai, T., Shand, F.L. and Howatson, A.F.: Development of measles virus in vitro. Virology 38 (1969) 50-67.

Reich, A., Erlwein, O., Nieswiesk, S., ter Meulen, V. and Liebert, U.G.: CD4 ÷ T cells control measles virus infection of the central nervous system. Immunol. 76 (1992) 185-191. Rozenblatt, S., Eizenberg, O., Ben-Levy, R., Lavie, V. and Bellini, W.J.: Sequence homology within the morbilliviruses. J. Virol. 53 (1985) 684-690. Spehner, D., Kirn, A. and Drillien, R.: Assembly of nucleocapsid-like structures in animal cells infected with vaccinia virus recombinant encoding the measles virus nucleoprotein. J. Virol. 65 (1991) 6296 6300. Stanley, K.K. and Luzio, J.P.: Construction of a new family of high efficiency bacterial expression vectors: identification of cDNA clones coding for human liver proteins. EMBO J. 3 (1984) 1429-1434. Studier, F.W., Rosenberg, A.H., Dunn, J.J. and Dubendorff, J.W.: Use of T7 RNA polymerase to direct expression of cloned genes. Methods Enzymol. 185 (1990) 60-89. Waters, D.J., Hersh, R.T. and Bushell, R.H.: Isolation and characterisation of measles nucleocapsid from infected cells. Virology 48 (1972) 278 281. Warnes, A., Fooks, A.R. and Stephenson, J.R.: Production of measles nucleoprotein in different expression systems and its use as a diagnostic reagent. J. Virol. Methods 49 (1994) 257 268. WHO, Expanded Programme on Immunisations Global Advisory Group Part 1. Wkly. Epidemiol. Rec. 68 (1993) 1-5.