pseudovirus particles in Escherichia coli - Europe PMC

Proc. Natl. Acad. Sci. USA Vol. 91, pp. 9067-9071, September 1994 Microbiology

Expression of tobacco mosaic virus coat protein and assembly of pseudovirus particles in Escherichia coli (plant virus/origin-of-assembly sequence/protein acetation/RNA pa

DUK-JU HWANG*, IAN M. ROBERTSt,

AND

glng/T7 RNA polymerase)

T. MICHAEL A. WILSONtt

*AgBiotech Center, Cook College, Rutgers University, P.O. Box 231, New Brunswick, NJ 08903; and tDepartment of Virology, Scottish Crop Research Institute, Invergowrie, Dundee DD2 5DA, United Kingdom

Communicated by Lawrence Bogorad, April 7, 1994

In addition to native TMV RNA, or 3' coterminal nested genome fragments, TMV CP was shown to package chimeric ssRNAs efficiently (13, 14) in a length- and sequenceindependent manner, provided that a contiguous viral sequence of only 432 nt (genome coordinates 5112-5543) was present. This can occur in vitro (13-15) or in vivo in transgenic tobacco plants (16). Only the 3' proximal stemloop, loop 1 in ref. 5, of the extended OAS (genome coordinates 5444-5518) was found necessary and sufficient to initiate nucleocapsid assembly in vitro (17). Significantly, only ssRNA, not ssDNA (18), can be packaged by TMV CP. The exceptional stability of TMV or TMV-like particles to proteases and RNases can be exploited to recover, store, and protect otherwise labile mRNA molecules to deliver them into plant or animal cells (19) for subsequent cotranslational disassembly (20). To study the molecular mechanism of plant virus assembly in vivo requires an easily and rapidly manipulated model system. Whole plant infection is a complex, slow, and asynchronous process. Nonviable virus mutants can be created (and lost) for reasons other than inefficient particle assembly in vivo. A synthetic TMV U1 CP gene with a lacUVS promoter and codons "optimized" for Escherichia coli was used to express native CP or a derivative carrying an additional C-terminal spacer (Gly-Gly) and an 8-amino acid epitope from poliovirus 3 VP1 (21). CPs were purified from sonicated bacterial cell supernatants by dialysis at pH 5.0 to induce helical aggregation (12). The resulting particles were shorter than with plant-derived TMV CP and had a striated or stacked cylindrical appearance. More recently, Shire et al. (22) expressed TMV U1 CP in E. coli from a trp promoter. CP was purified from cleared cell lysates by chromatography and dialysis (precipitation) at pH 3.5. The N-terminal N-formyl-Met had been removed, but the penultimate N-terminal Ser was not acetylated, unlike plant-encoded TMV CP. This feature, or the added positive charge, disrupted formation of helical CP aggregates and the 20S assembly nucleation species (7-10) and created assembly-incompetent CP when incubated with TMV RNA in vitro (22, 23). In our study, two point mutants of the U1 CP sequence were created to change the mature N terminus to Ala or Pro, which exist without acetylation in the NM and U2 strains of TMV, respectively (24). Recently, several groups have reported high-level expression ofCPs from a variety of plant RNA viruses in both E. coli and yeast cells (25-27, 42). Although virus-like (RNA-free) protein aggregates have been claimed, in no case has the efficient assembly of viral or pseudoviral ribonucleocapsids been engineered in vivo. We report plant virus or pseudovirus

The bidirectional self-assembly of tobacco ABSTRACT mosaic virus (TMV, common or Ul strain) has been studied extensively in vitro. Foreign single-stranded RNA molecules containing the TMV origin-of-assembly sequence (OAS, 75432 nt in length) are also packaged by TMV coat protein (CP) in vitro to form helical pseudovirus particles. To study virus assembly in vivo requires an easily manipulated model system, independent of replication in plants. The TMV assembly machinery also provides a convenient means to protect and recover chimeric gene transcripts of almost any length or sequence for a variety of applications. Native TMV CP expressed in and purified from Escherichia coli formed nonhelical, stacked aggregates after dialysis into pH 5 buffer and was inactive for in vitro assembly with TMV RNA. U1 CP derivatives in which the second amino acid was changed from Ser to Ala or Pro, nonacetylated N termini found in two natural strains of the virus, failed to remediate these anomalous properties. However, in vivo coexpression of CP and singlestranded RNAs (up to =2 kb) containing the TMV OAS gave high yields of helical pseudovirus particles of the predicted length (up to 7.4 ± 1.4 jpg/mg of total bacterial protein). If the OAS-containing RNA was first recruited into bacterial polyribosomes, elongation of pseudovirus assembly was blocked. In vivo, E. coli expression of a full-length cDNA clone of the TMV genome (6.4 kb) resulted in high, immunodetectable levels of CP and assembly of sufficient intact genomic RNA to initiate systemic infection of susceptible tobacco plants.

Tobacco mosaic virus (TMV) particles are extremely stable and retain infectivity for decades. Over 2100 copies of a 17.6-kDa coat protein (CP) fully protect the 6.4-kb singlestranded RNA (ssRNA) genome against degradation by RNases. TMV was the model system of choice for early studies on the spontaneous "self-assembly" of multimeric, biological structures in vitro (1-3). TMV assembly is initiated by a specific interaction between a prefabricated (20 S) protein aggregate (the "disk" or protohelix) and an RNA sequence centered either 0.4 kb or 0.9 kb from the 3' end of the genomic RNA (4). An origin-of-assembly sequence (OAS) was characterized (5) and proposed to exist as three stem-loop structures, before the complete 6395-nt genome of the common (or Ul) strain of TMV had been sequenced (6). The structures of the prefabricated, oligomeric forms of TMV CP used for efficient assembly initiation and/or bidirectional elongation, and the precise assembly pathway (i.e., which CP form is used in which RNA direction, and whether simultaneously or sequentially) remain controversial (7-9), and the data continue to evade a consensus model (10). Two simple protocols exist to isolate CP from virions for subsequent encapsidation of ssRNA in vitro (11, 12).

Abbreviations: TMV, tobacco mosaic virus; OAS, origin-ofassembly sequence; CP, coat protein; IPTG, isopropyl a-Dthiogalactopyranoside; CAT, chloramphenicol acetyltransferase; GUS, P-glucuronidase; ss, single-stranded; mAb, monoclonal antibody; EM, electron microscopy; ISEM, immunosorbent EM; RT, reverse transcriptase; DAS, double-antibody sandwich. tTo whom reprint requests should be addressed.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 9067

Microbiology: Hwang et al.

.,,P,%nj8

O(VI

Proc. Natl. Acad. Sci. USA 91 (1994)

particle "self-assembly" in vivo, in a prokaryotic system that is economical, homogeneous, and easily manipulated and offers a rapid means by which to study plant virus assembly events in a cellular milieu. MATERIALS AND METHODS Virus Propagation, TMV RNA and CP Preparation, and in Vitro Self-Assembly Reactions. TMV Ul strain was grown in Nicotiana tabacum cv. Xanthi and purified from leaves (28). Viral RNA was extracted with phenol/chloroform, precipitated with ethanol, dissolved and stored in sterile water at -1960C. Viral CP was prepared (12) and stored at 40C in sodium acetate (pH 5.0) (I = 0.1) at 10-20 mg/ml. In vitro ribonucleocapsid assembly was performed as described (12) using the A310 turbidometric criteria of Zimmern (5). Plasmid Constructs. DNA manipulations were standard (29). Construction details for pET3a-derived plasmids pET301, pET302, pETAla301, and pETPro301 and for pLysE-derived plasmids pLys1O2 [E. coli promoter-CAT-OAS (CAT, chloramphenicol acetyltransferase)] and pLyslO32A [T7 promoterGUS-OAS (GUS, glucuronidase)] are contained in ref. 30 or available upon request. Final constructs were expressed in E. coli BL21(DE3), the T7 RNA polymerase-inducible system of Studier et al. (31). The salient features of each plasmid are shown in Fig. 1. Analytical-Scale Detection and Large-Scale Purffication of TMV CP and Derivatives from E. coli. E. coli BL21(DE3) pLysE cells (30) transformed with pET301, pET302, pETAla301, or pETPro301 were grown to midlogarithmic phase (A600 = 0.6) at 37°C and T7 transcription was induced with 0.4 or 2.0 mM isopropyl a-D-thiogalactopyranoside (IPTG) for 2-18 hr at 250C. At 2 hr, cells from 1-ml samples were recovered, washed once in TE (10 mM Tris HCl, pH 7.5/1 mM EDTA), resuspended, and heated (100°C, 5 min) in 80 ,ul of TE plus 80 ,u of Laemmli sample buffer (32) and 15 Al was analyzed by SDS/PAGE in a 15% (wt/vol) gel. Proteins were stained with Coomassie blue or immunoblotted (33) with rabbit polyclonal anti-TMV CP serum to confirm their identity. After 18 hr, cells from 500 ml of culture (M9ZB medium) were harvested, stored at -800C, thawed to 40C, and resuspended in lysis buffer (0.1 M Tris-HCl, pH 7.2/50 mM EDTA/0.2 M NaCl/0.1% 2-mercaptoethanol) for sonication (30 min). Debris was removed by centrifugation (8500 x g) and supernatant CP was precipitated by adding (NH4)2SO4 to 80%. Supernatant and pellet protein compositions were mon-

CP cassettes TE77 .MV CP

pET 3a

ATGTCT:... ATGTCT.... ATGGCT ...

'

o t A :clt:

ATGCCG: ...t.-.

I[-

A

3c

OAS RNA cassettes §OAS

CAT

-

Y:s:;' Pv

pLyVs

T7

T7

PI

_

terminator ,

_

2

,

Il

Scj

prit !:,(I~1

,(CAA)21CACCATGG tV'. :-,!

FIG. 1. Plasmids used to express TMV CP (Ul or N-terminal derivatives), CAT-OAS RNA (1.6 kb), or GUS-OAS RNA (2.8 kb). Significant restriction sites and genetic elements are indicated. TMV CP constructs (pET301, pET302, pETAla301, and pETPro301) were in pET3a. CAT-OAS and GUS-OAS constructs used pLysE (31). OAS (dotted) is the 432-bp OAS of TMV. Bold arrowheads denote promoter sequences. Hatched areas in pLyslO2 depict a duplication of the 3' end of the CAT gene generated by adding the OAS. Not to scale.

itored by SDS/PAGE. Precipitated proteins were resuspended in 15% (NH4)2SO4 and dialyzed overnight at 40C against sodium acetate (pH 5.0) (I = 0.1; ref. 12) to form helical aggregates of TMV CP (12). Any precipitate remaining was removed (1000 x g, 10 min, 40C) and polymerized TMV CP in the supernatant was recovered by ultracentrifugation (130,000 x g, 90 min, 40C). Dialysis and differential centrifugation were repeated several times until only one protein band (17.5 kDa) was seen by SDS/PAGE. CPs were stored in sodium acetate (pH 5.0) for electron microscopy (EM) or prepared for in vitro assembly reactions (5, 12) with virusderived TMV RNA. Isolation of TMV-Like Particles from E. coli. E. coli BL21 (DE3) were cotransformed with plasmids expressing either TMV CP or a reporter-OAS RNA (Fig. 1). Cultures (100 ml) were grown at 370C for 4 hr, induced with 0.4 or 2 mM IPTG, and then grown for 6 or 18 hr at 200C. Cells were harvested by centrifugation (6000 x g, 4TC, 15 min), resuspended in 1 ml of TE, and incubated with lysozyme (1 mg/ml) at 200C for 5-10 min. Bacteria were lysed by osmotic shock [1 vol 40%o (wt/vol) sucrose, then 4 vol of TE buffer]. DNase I was added to 5 Mg/ml. Standard Triton X-100 lysis procedures (29) caused nonspecific disassembly of exogenous TMV (control reconstruction) and in vivo assembled virus-like particles (data not shown). Cell debris was removed (10,000 x g, 40C, 20 min) and the supernatant was centrifuged (130,000 x g, 4 hr, 4°C) in a sucrose gradient [10-40% (wt/vol) in TE]. The pellet was resuspended in 1 ml of TE for EM. Reverse Transcriptase (RT)-PCR Assay for Encapsidated Sequences. Pseudovirus particles were purified from E. coli by polyethylene glycol precipitation and sucrose gradient centrifugation and then used for RT-PCR analysis. RT reactions used a primer complementary to the TMV OAS (5'-CCGGTTCGAGATCGA-3'). PCR (30 or 45 cycles: 24 sec at 94°C/30 sec at 54°C/36 sec at 720C) used the following primer pairs (0.2 pg each per 100-/p reaction with 5 units of Taq polymerase): the common 3' primer above with either 5'-GTTGTTCGTCACGG-3' (5' end TMV OAS) or 5'-ATGGTACGTCCTGTC-3' (5' end GUS gene). PCR detection of the CAT sequence used the primer pair: 5'-CAGGAGCTAAGGAGG-3' (5' end CAT gene) and 5'-CGCCCCGCCC-3' (3' end CAT gene). Amounts of Total Bacterial Protein, TMV CP, and Helical TMV-Like Particles. Total bacterial protein was measured by dye-binding assay (34). Total TMV CP expressed was judged by comparison to known amounts of TMV-derived CP coelectrophoresed in Coomassie blue-stained gels or by doubleantibody sandwich (DAS) ELISA using alkaline phosphatase-conjugated polyclonal IgG against TMV in phosphate-buffered saline (PBS) containing 0.05% (vol/vol) Tween 20 and 1% (wt/vol) bovine serum albumin (BSA). Yields of helical TMV-like particles were measured by DASELISA with a mouse monoclonal antibody (mAb) specific for an epitope (neotope) in the TMV CP helix (ref. 35; TMV253P). ELISA followed (36) except plates were coated with 10 Mg of rabbit polyclonal anti-TMV serum per ml and the p-nitrophenyl phosphate reagent was at 4 mg/ml. EM. Immunosorbent electron microscopy (ISEM; ref. 37) used carbon filmed grids coated with 10 /4 of rabbit polyclonal anti-TMV serum (1:1000) for 30-60 min at 37°C. Washed, drained grids were placed on drops of extracted, dialyzed TMV CP or cleared bacterial cell lysates for 1-4 hr at 4°C. Negative staining was with 2% ammonium molybdate (pH 5.0) or 1% uranyl acetate. Immunolabeling of TMV CP-only aggregates or pseudovirus particles trapped by ISEM used mAb TMV-253P diluted 1:10 in sodium acetate, pH 5.0/1% BSA for 15, 30, or 60 min at 20C. TMV CP-only samples were then allowed to react for 30 min with 10-nm gold-conjugated goat anti-mouse IgG, diluted 1:50 in 1% BSA. Negative staining was with 2% ammonium molybdate (pH 5.0). Immunospecific

Microbiology: Hwang et A labeling of CAT-OAS pseudovirus particles with TMVW253P was observed by negative staining only. Assembly of Full-Length TMV RNA in Vivo and Infectivity Assay. To show in vivo encapsidation of longer T7 transcripts (.6.4 kb), midlogarithmic cultures (200 ml) of E. coli [BL21 (DE3) pLysS; ref. 31] were induced with IPTG for 10 hr at 250C and the cells were harvested and lysed as above. Bacteria were either nontransformed (control) or transformed with pTMV212 (pUC118 with a full-length infectious clone of TMV Ul cDNA, preceded by a T7 promoter and two additional 5' G residues). Cleared lysates were sampled for expression of TMV CP by Western blotting and the remainders were centrifuged at high speed. The pellets were resuspended in 2 ml of TE buffer. Control reconstructions contained the resuspended pellets from lysates of nontransformed E. coli plus either natural TMV RNA (50 jtg/ml) or virus particles (10 pg/ml). To assay for in vivo assembly of pTMV212 transcripts and to check for survival of added TMV RNA or virions, lysates were stored for 4-5 weeks at 4°C and then manually inoculated onto lower leaves of young seedlings of N. tabacum cvs. Xanthi (genotype nn; systemic host), Petite Havana (SR1; systemic host), or Xanthi nc (genotype NN; hypersensitive host). After 10-20 days, newly emerged systemic leaves of the former were examined for mosaic symptoms. After 5-10 days, inoculated leaves of Xanthi nc plants were examined for necrotic lesions. Two plants of each type were used for each sample in each of three replicate experiments.

RESULTS AND DISCUSSION TMV CP Mutagenesis and Expression. The mature U1 CP N terminus was changed to Ala or Pro by PCR mutagenesis but no further mutations were made to convert the native U1 CP clone to NM or U2 sequences. The CP coding regions of pET301, pETAla301, pETPro301, and pET302 are shown in Fig. 1. SDS/PAGE ofcleared lysates of bacteria transformed with each TMV CP-expressing plasmid, induced for 2 hr with IPTG, is shown in Fig. 2. Western blotting with a rabbit polyclonal antiserum against TMV Ul CP revealed a single band coincident with each stained TMV CP band. By comparison with different amounts of coelectrophoresed TMV CP marker, CP expression was estimated to be 60 ug/ml of original bacterial cell culture for pET302 and 30 ug/ml for pET301 and pETAla301. pETPro301 was expressed at lower levels, possibly due to additional nucleotides cloned between the Shine-Dalgarno motif in pET3a and the AUG codon.

*.S>42 ;fet


Recovery and in Vitro Assembly Behavior of TMV CP Made in E. coli. Attempts to recover TMV CP from cleared lysates of IPTG-induced bacterial cells by low-pH dialysis proce-

dures alone (12, 21) resulted in large, insoluble precipitates,

particularly with the Ser -* Pro CP derivative. An extraction

protocol involving (NH4)2SO4 fractionation and low pH dialysis was optimized. SDS/PAGE showed that little CP was lost in the cell debris removed after sonication or lysozyme/ osmotic lysis procedures and no detectable proteolytic degradation of the 17.6-kDa CP, as occurs when bacteria contaminate virus-derived TMV CP preparations, resulting in irreversible polymerization (38). In all cases, TMV CP or derivatives were purified to homogeneity as judged by SDS/ PAGE and silver staining (not shown) and stored in pH 5 buffer. In contrast to virus-derived TMV CP (Fig. 3a), EM revealed that all E. coli-made CP, including both N-terminal derivatives, failed to form long helices but existed as heterodisperse populations of striated, stacked disk-like rods (Fig. 3 b-f). Similar structures were reported previously for TMV and Johnsongrass mosaic potyvirus CPs made in and purified from E. coli or yeast cells (21-23, 26, 27). The nonhelical nature of these CP-only assemblies was confirmed by their failure (Fig. 4a) to bind mAb TMV-253P specific for an epitope (neotope) present only in correctly assembled TMV CP helices (Fig. 4b) or in virions (35). None of the E. coli-made TMV CPs showed significant TMV RNA packaging activity in vitro in a turbidometric assembly assay (refs. 5 and 12; data not shown). In Vivo Assembly of TMV-Like Particles. Two plasmids provided ssRNA species to test for pseudovirus particle assembly in vivo (Fig. 1). pLyslO2 produces a 1.6-kb CATOAS transcript constitutively. pLyslO32A contained an IPTGinducible cassette expressing GUS-OAS RNA (2.8 kb). Coexpression of CAT-OAS RNA with native TMV CP (pET301 or pET302; not shown) or the N-terminal Ala derivative of TMV CP (pETAla301) resulted in helical ribonucleoprotein rodlets (Fig. SA b and c) of the predicted length (75 nm; Fig. SB). These particles reacted specifically with mAb TMV-253P, as in Fig. 4b, confirming their helical symmetry (data not shown). RT-PCR with RNA from puri-

6

d

a

XF'S 4

11.3-i'

.

FIG. 2. SDS/PAGE of total proteins expressed in E. coli BL21 (DE3) following a 2-hr induction with I1PIG. Cells were transformed with pETPro3Ol (lane 1), pET3O2 (lane 2), pETAla3O1 (lane 3), pE.T3O1 (lane 4), or pERTa control (lane 5). Prote-incus werestainedA with Coomassie blue. Positions of coelectrophoresed marker proteins are indicated. The arrowhead denotes virus-derived TMV CP

(17.6 kDa).

9069

Y

K

A A

,

--..

XW

FIG. 3. ISEM of TMV CPs alone. CP was prepared from U1 virus (a) or isolated from IPTG-induced cultures of E. coli transformed with pET301 (b), pET302 (c andf), pETAla301 (d), or pETPro3O1 (e). All samples were stored in pH 5.0 acetate buffer. Negative staining was with 1% uranyl acetate. [Bar = 100 nm (a-e) or 50 nm (f).]


9070

Proc. Natl. Acad. Sci. USA 91 (1994) A

a '7

I!il,

a

c

'.

4b

.

b

0

o B

.;k5,

'YZt

M

wr

i

.-

00I ?

Of'?

B . he

a.

.1e ...

..

FIG. 4. Immunospecific gold-labeling of TMV CP-only aggregates formed in pH 5.0 buffer using mAb TMV-253P, specific for helically assembled particles. (a) TMV CP expressed in E. coli from pET302. (b) TMV CP from virions prepared from infected tobacco (12). mAb TMV-253P was located with 10-nm colloidal goldconjugated goat anti-mouse IgG. (Bar = 100 nm.)

fled pseudovirus particles, and primer pairs specific for the TMV OAS region (0.43 kbp) or the CAT-coding region (0.78 kbp), confirmed the predicted composition of the nucleocapsids (Fig. SC, lanes 1, 2, 4, and 5). Expression of pETAla301 alone resulted in few, if any, helical particles and none of significant length (Fig. 5Aa). In vivo, the Pro derivative (pETPro301), alone or with CATOAS RNA, produced only striated, stacked CP assemblies of variable length (as seen in vitro; Fig. 3e). DAS ELISA results on the yield of CAT-OAS pseudovirus particles (with TMV-253P) and the total amount of TMV CP (with alkaline phosphatase-conjugated, polyclonal IgG against TMV) are summarized in Table 1. In contrast to the 2-fold increase in TMV CP expression estimated for pET302 in Fig. 2, the averaged ELISA data showed 6 times more TMV CP produced by the E. coli codon-optimized CP gene and, consequently, a 5-fold rise in CAT-OAS pseudovirus particles. In all cases, only about 20%o ofthe expressed CP assembled into CAT pseudovirus particles. In TMV-infected plants, the level of unassembled CP is low (=5-10% of total CP). In the absence of CAT-OAS transcripts, all DAS ELISA values with TMV-253P were low, confirming the absence of endogenous E. coli RNA in helical particles (Fig. 5Aa). Rboones (70 S) Block GUS-OAS RNA Ammbly in Vivo. GUS-OAS RNA (2.8 kb) was predicted to form 130-nm-long pseudovirus particles when coexpressed with TMV CP. In vitro, GUS-OAS transcripts assemble efficiently and completely with virus-derived TMV CP (15). However, in E. coli, ISEM (Fig. SA d-f) showed only partially assembled rodlets between 30 and 60 nm in length (i.e., 640-1280 nt encapsidated). After RNase treatment, the protected RNA was phenol-extracted and shown, by RT-PCR, to contain the OAS domain, but not the GUS portion of the chimeric RNA.

1-\ -_-.

FIG. 5. In vivo assembly of TMV-like CAT-OAS or GUS-OAS pseudovirus particles in E. coli. (A) ISEM of complexes containing TMV CP from IPTG-induced cultures of E. coli transformed with pETAla301 alone (a), cotransformed with pETAla301 and pLyslO2 (b and c), or cotransformed with pETAla301 and pLyslO32A (d-f). Negative staining used 1% uranyl acetate (a, b, and d-f) or 2% ammonium molybdate, pH 3.5 (c). Representative fields from separate micrographs are shown in d-f. (Bar = 100 am.) (B) Histogram showing the length (nm)-frequency (%) distribution of CAT-OAS pseudovirus particles isolated from E. coli cotransformed with pETAla301 and pLys1O2. (C) RT-PCR analysis of RNAs isolated from pseudovirus particles containing CAT-OAS (lanes 2 and 5) or GUS-OAS (lane 3) RNAs. Lanes 1 and 4 are control extracts from cells transformed with pETAla301 alone, using PCR primer pairs specific for the TMV OAS (lane 1) or CAT gene (lane 4). Lanes 2 and 3 used PCR primers targeted to the TMV OAS. Lane S is a reaction with PCR primers for the CAT gene. Double-stranded DNA marker positions are shown on the right, and the relevant PCR bands identified on the left.

Closer examination revealed that most, if not all, GUS-OAS pseudovirus particles had one or more ribosome-like comTable 1. Expression levels of TMV CP, a CP derivative and TMV-like CAT-OAS particles in E. coli Total TMV CP,* TMV-like particles,t TMV CP ug/mg of total ag/mg of total protein protein expression vector 0.7 ± 0.03 3.0 ± 0.5 pET301 3.6 ± 0.8 0.8 ± 0.02 pETAla301 19.0 ± 3.1 3.8 ± 0.7 pET302 7.4 ± 1.4* pET302 38.0-: 4.40 E. coli cells [BL21(DE3)] in logarithmic growth, constitutively expressing a CAT-OAS RNA (1.6 kb) from pLyslO2 (Fig. 1) and cotransformed with a pET3a-derived plasmid expressing TMV CP or a Ser-* Ala derivative (see text), were induced with 2 mM IPTG for 18 hr. Cleared lysates were prepared as described in the text. Values are expressed as mean ± 1 SD. *The total amount of TMV CP expressed was quantified by ELISA using rabbit polyclonal antiserum. tThe amount of CAT pseudovirus particles was quantified by ELISA using a mAb specific for helical TMV CP structures. tCells were induced with 0.4 mM IPTG for 18 hr.



must have been processed or fragmented and encapsidated correctly to survive and initiate plant infection. We thank Joel R. Haynes for pTMV CP used to construct pET302, Bill Dawson for pTMV212 and pTMV210 (used for pET301 and derivatives), and Marc H. V. van Regenmortel for mAb TMV-253P. We are grateful to Elizabeth Fyffe and Myra Purves for typing and to Ian Pitkethly for the graphics. This work was funded by the New Jersey Commission for Science and Technology and The Scottish Office Agriculture and Fisheries Department. The results form part of U.S. patent application no. 07/971, 101 (30).

(PPpp

20S

P

T;rticie morphoiog iI-mal RNA polarity

disk protoheiii

.

IC

FvO

FIG. 6. Schematic representation of the complexes shown in Fig. 5A d-f. The uncapped, free 5' end of the chimeric CAA leader-GUSOAS RNA recruits 30S ribosomal subunit(s) and assembles active translocating 70S ribosome(s). Subsequent transcription of the contiguous TMV OAS initiates encapsidation in the presence of prefabricated 20S TMV CP (pETAla301) aggregates. During RNA pack5' direction, the free 5' tail must pass through aging in the major 3' the lumen of the growing helix to form the "traveling loop" at the concave (5') face of the particle. Eventually this process is sterically blocked by the advancing ribosome. --

plexes associated with the convex end of the rodlet, where the 3' terminus of TMV RNA finally resides (39). During virus

assembly in the major 3'

--

5'

direction, the 5' end of

the RNA also protrudes from the lumen of the nucleocapsid at the convex end (8-10). Fig. 6 provides a model for the architecture of the complexes shown in Fig. SA d-f. Such complexes may have been more evident with GUSOAS than with CAT-OAS transcripts because (i) being longer, GUS-OAS RNA would have more time to recruit 30S (70S) ribosomes and begin translation before the 3' proximal OAS was transcribed; (ii) both plasmids (pLysl032A and pETAla301) require T7 RNA polymerase and are induced synchronously, thereby preventing prior accumulation of functional TMV CP and (by competition) reducing the total amount of TMV CP expressed (data not shown); and/or (iii) the 5' leader in pLysl032A [G(CAA)2,CACCaug] is highly effective in recruiting 30S ribosomal subunits despite the absence of a true Shine-Dalgarno sequence (40). Thus polysomes, not naked GUS-OAS RNA, may have initiated 3' 5' nucleocapsid assembly. Particle elongation was then un--

able to

dislodge

70S

ribosomes

translocating

5'

-*

3'

9071

during

GUS mRNA translation, resulting in a sterically blocked complex. This observation has important implications for the fate of each progeny plus-strand TMV RNA molecule in infected plant cells. It confirms that translation, virus assembly, and further activity as a template for RNA replication are sterically exclusive functions. Assembly of FuHl-Length Infectious TMV Particles inE. coli. E. coli BL21 (DE3) pLysS cells were transformed with pTMV212. IPTG induced TMV CP expression equivalent to that from pET302 (60 pg/ml culture), confirmning earlier results from in vitro translations of full-length TMV RNA in E. coli cell-free extracts (41). The identity of the induced protein was confirmed by Western blotting (not shown). ISEM of pTMV212-transformed bacterial extracts also revealed low numbers of TMV-like particles of variable length. After storage, this material initiated systemic TMV infections in two of three independent experiments on susceptible cultivars of N. tabacum but had insufficient titer to produce local lesions on hypersensitive tobacco plants. Control reconstruction experiments in which TMV, or TMV RNA extracted from virions, was added to lysates of E. coli BL21 (DE3) cells, processed as for pTMV212-transformed cell extracts, either did (TMV) or did not (TMV RNA) cause plant infection. In E. coli, pTMV212 transcripts will be uncapped and have two nonviral 5' G residues and an undefined 3' end. Nevertheless, some .6.4-kb RNA molecules from pTMV212

1. Fraenkel-Conrat, H. & Williams, R. C. (1955) Proc. Natl. Acad. Sci. USA 41, 690-698. 2. Caspar, D. L. D. (1963) Adv. Protein Chem. 18, 37-121. 3. Butler, P. J. G. & Klug, A. (1971) Nature New Biol. 229, 47-50. 4. Zimmern, D. & Wilson, T. M. A. (1976) FEBS Lett. 71, 294-298. 5. Zimmern, D. (1977) Cell 11, 463-482. 6. Goelet, P., Lomonossoff, G. P., Butler, P. J. G., Akam, M. E., Gait, M. J. & Karn, J. (1982) Proc. Natl. Acad. Sci. USA 79, 5818-5822. 7. Schuster, T. M., Scheele, R. B., Adams, M. L., Shire, S. J., Steckert, J. J. & Potschka, M. (1980) Biophys. J. 32, 313-329. 8. Okada, Y. (1986) Adv. Biophys. 22, 95-145. 9. Bloomer, A. C. & Butler, P. J. G. (1986) in The Plant Viruses, eds. van Regenmortel, M. H, V. & Fraenkel-Conrat, H. (Plenum, New York), Vol. 2, pp. 19-57. 10. Butler, P. J. G., Bloomer, A. C. & Finch, J. T. (1992) J. Mol. Biol. 224, 381-394. 11. Fraenkel-Conrat, H. (1957) Virology 4, 1-4. 12. Durham, A. C. H. (1972) J. Mol. Biol. 67, 289-305. 13. Sleat, D. E., Turner, P. C., Finch, J. T., Butler, P. J. G. & Wilson, T. M. A. (1986) Virology 155, 299-308. 14. Turner, D. R., McGuigan, C. J. & Butler, P. J. G. (1989) J. Mol. Biol.

209, 407-422. 15. Jupin, I., Sleat, D. E., Watkins, P. A. C. & Wilson, T. M. A. (1989) Nucleic Acids Res. 17, 815. 16. Sleat, D. E., Gallie, D. R., Watts, J. W., Deom, C. M., Turner, P. C., Beachy, R. N. & Wilson, T. M. A. (1988) Nucleic Acids Res. 16, 3127-3140. 17. Turner, D. R., Joyce, L. E. & Butler, P. J. G. (1988) J. Mol. Biol. 203, 531-547. 18. Gallie, D. R., Plaskitt, K. A. & Wilson, T. M. A. (1987) Virology 158, 473-476. 19. Gallie, D. R., Sleat, D. E., Watts, J. W., Turner, P. C. & Wilson, T. M. A. (1987) Science 236, 1122-1124. 20. Wilson, T. M. A. (1984) Virology 137, 255-265. 21. Haynes, J. R., Cunningham, J., von Seefried, A., Lennick, M., Garvin, R. T. & Shen, S.-H. (1986) Bio/Technology 4, 637-641. 22. Shire, S. J., McKay, P., Leung, D. W., Cachianes, G. J., Jackson, E., Wood, W. I., Raghavendra, K., Khairallah, L. & Schuster, T. (1990) Biochemistry 29, 5119-5126. 23. Raghavendra, K., MacDonald, G., Schuster, T. M., Shire, S. J., Leung, D. W., Cachianes, G. J., Jackson, E., Wood, W. I. & McKay, P. (1989) Biophys. J. 55, 156 (abstr.). 24. Wittmann, H. G. (1965) Z. Naturforsch. B 20, 1213-1223. 25. Gal-On, A., Antignus, Y., Rosner, A. & Raccah, B. (1990) Gene 87, 273-277. 26. Jagadish, M. N., Ward, C. W., Gough, K. H., Tulloch, P. A., Whittaker, L. A. & Shukla, D. D. (1991) J. Gen. Virol. 72, 1543-1550. 27. Jagadish, M. N., Huang, D. & Ward, C. W. (1993) J. Gen. Virol. 74, 893-896.

28. Leberman, R. (1966) Virology 30, 341-347. 29. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning:A Laboratory Manual (Cold Spring Harbor Lab. Press, Plainview, NY), 2nd Ed. 30. Wilson, T. M. A. & Hwang-Lee, D. J. (1994) Int. Patent Appl. IPN WO 94/10329, pp. 30-36. 31. Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. W. (1990) Methods Enzymol. 185, 60-89. 32. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 33. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76, 4350-4354. 34. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. 35. Dore, I., Weiss, E., Altshuh, D. & van Regenmortel, M. H. V. (1988) Virology 162, 279-289. 36. van Regenmortel, M. H. V. & Burkard, J. (1980) Virology 106, 327-334. 37. Roberts, I. M. (1986) in Electron Microscopy of Proteins 5: Viral Topics, eds. Harris, J. R. and Horne, R. W. (Academic, Orlando), pp. 293-357. 38. Durham, A. C. H. (1974) FEBS Lett. 25, 147-150. 39. Wilson, T. M. A., Perham, R. N., Finch, J. T. & Butler, P. J. G. (1976) FEBS Lett. 64, 285-289. 40. Tzareva, N. V., Makhno, V. I. & Boni, I. V. (1994) FEBS Lett. 337, 189-194. 41. Glover, J. F. & Wilson, T. M. A. (1982) Eur. J. Biochem. 122,485-492. 42. Jagadish, M. N., Hamilton, R. C., Femandez, C. S., Schoofs, P., Davem, K. M., Kalnins, H., Ward, C. W. & Nisbet, J. T. (1993) Biol Technology 11, 1166-1170.