J. gen. Virol. (1988), 69, 2359-2367. Printed in Great Britain

Key words: BNYVV/translation initiation~transcription

2359

In vitro Translation of Natural and Synthetic Beet Necrotic Yellow Vein Virus RNA-I

By I. J U P I N , L. Q U I L L E T , V. Z I E G L E R - G R A F F , H. G U I L L E Y , K. R I C H A R D S * AND G. J O N A R D

Institut de Biol6gie Mol&ulaire des Plantes, 12 rue du G~n&al Zimmer, 67000 Strasbourg, France

(Accepted 10 June 1988)

SUMMARY

The in vitro translation of beet necrotic yellow vein virus RNA-1 was investigated in rabbit reticulocyte lysate and in wheatgerm extract using as messenger either R N A extracted from virions or a synthetic RNA-1 produced by in vitro transcription of full- length cDNA. In wheatgerm extract, RNA-1 directed the synthesis of approximately equivalent amounts of two long polypeptides of approximate Mr 220000 (220K) and 240K. The size of the translation products of Y-truncated RNA-1 transcripts suggested that synthesis of the 240K translation product was initiated at an A U G near the 5' terminus, probably AUG(154), the first initiation codon in the RNA-1 sequence. The 220K polypeptide is initiated internally, at AUG(496). In rabbit reticulocyte lysate only the internal initiation codon, AUG(496), was used for initiation on full-length RNA-1 although AUG(154) was accessible o'n short T-truncated transcripts.

INTRODUCTION Particles of beet necrotic yellow vein virus (BNYVV), the agent of rhizomania disease of

sugarbeet (Tamada, 1975), contain four plus-strand R N A molecules, all of which have been cloned and sequenced (Bouzoubaa et al., 1985, 1986, 1987). BNYVV RNA-1 [6746 nucleotides (nt)] contains a single long open reading frame of 2109 codons which can encode a polypeptide of Mr 237389 (Bouzoubaa et al., 1987). In earlier experiments, translation of RNA-1 in rabbit reticulocyte lysate was found to give rise to abundant polypeptides with apparent Mr values of 50000 (50K) and 170K as well as lesser amounts of a 220K species (Ziegler et al., 1985). The relative amounts of the 50K, 170K and 220K polypeptides produced and the time course of their appearance suggested that the two smaller species resulted from premature termination of translation. In this paper we describe translation experiments using both rabbit reticulocyte lysate and wheatgerm extract primed either with natural BNYVV R N A or with a synthetic RNA-1 transcript prepared by in vitro transcription of cloned cDNA. The results show that the translation products made depend upon the translation system used; translation was either initiated exclusively at an internal initiation codon, AUG(496), in rabbit reticulocyte lysate, or at both this codon and at an upstream codon [probably AUG(154), the first A U G in the RNA-1 sequence] in wheatgerm extract.

METHODS Recombinant plasmids. RNA from BNYVV isolate F 13 (Ziegler et al., 1985) was used in all experiments. The

recombinant cDNA plasmids referred to below are named according to the following convention: pBXY(a-b), where X indicates from which of the four BNYVV RNAs the cDNA is derived (X is always I in this study), Y is an arbitrary plasmid identification number and a and b give the coordinates of the cDNA insert with respect to the established RNA-1 sequence (Bouzoubaa et al., 1987). Double-stranded cDNA corresponding to the 5'-terminal region of BNYVV RNA-1 was obtained by specific primer-directed cDNA synthesis using a synthetic oligodeoxynucleotide complementary to nt 1527 to 1545 of RNA-I to prime first-strand synthesis and an oligodeoxynucleotide identical to nt 1 to 23 of the RNA-1 sequence to prime second-strand synthesis. The double-

0000-8288 © 1988 SGM

2360

He

I. J 'UPIN AND OTHERS

.e SO. S 435 1420 1954 3741

pB11(1-1545) pBl 16(1816-4617)PUC

B S S X H II | I | I |

4730 pB 164(1150-2212)PUC pB 18(3028-4874)PUC

X I pB 15(3816-6746)PUC

I I I I t I

2 4 6 Kilobases

H I

Fig. 1. StructureofrecombinantcDNAplasmidsusedtoconstruetthe full-lengthRNA-I transcription vector pB 11(1-6746). Numbering refers to the coordinates of the published BNYVV RNA-1 sequence (Bouzoubaa et al., 1987). The filled circle represents the T7 polymerase promoter, eDNA insert sequences are represented by heavy lines and vector sequences (which are only shown if they contain relevant restriction sites) are represented by thin lines. Restriction sites are indicated as: He, HincII; B, BstEII; S, SphI; X, XmaIII; H, HindlII.

stranded eDNA was dC-tailed and cloned into PstI-cut and dG-tailed pUC9 (Pharmacia) to give pBll(1- 1545)PUC, where PUC indicates that the vector is pUC9. The eDNA insert was excised from this plasmid with PstI and ligated into the PstI site of the transcription vector pBS(-) (Stratagene) to give pBl1(1-1545). The orientation of the insert is such that plus-strand RNA synthesis is under the control of the bacteriophage T7 promoter, pB11(1-638) was constructed in a similar fashion except that a synthetic oligodeoxynucleotide complementary to nt 622 to 642 was used to prime first-strand cDNA synthesis.

pB 11(1-1545) was extended at the insert 3" terminus by stepwise addition of restriction fragments derived from other RNA-1 eDNA clones (Fig. 1). The clone pB15(3816-6746)PUC was obtained during our earlier sequence work and was originally referred to as pBF5 (Bouzoubaa et al., 1987). The recombinant plasmids pB164(1150- 2212)PUC, pB 116(1816-4617)PUC and pB 18(3028-4874)PUC were obtained using a synthetic oligodeoxynucleo- tide complementary to nt 4854 to 4874 to prime first-strand synthesis and an oligodeoxynueleotide corresponding to nt 1 to 23 to prime second-strand synthesis. As in the first step, the BstEII-SphI fragment (nt 1420 to 1954) was excised from pB164(1150-2212)PUC and fused between the insert BstEII and the polylinker SphI sites of pB11(1- 1545) to produce pB 11 ( 1-1954). Then, an SphI fragment (nt 1954 to 3741 ) from pB 116( 18 l 6-4617)PUC was cloned in the proper orientation into the SphI site of pBl1(1-1954) to give pbl1(1-3741), pBl1(1-3741) was partially digested with SphI and fullqength linear DNA was isolated from a low melting point agarose gel. The linear DNA was then cut with HindlII, which cuts in the polylinker flanking the 3" end of the insert, and the long SphI-HindlII fragment containing the insert sequence (nt 1 to 3741) plus the vector was isolated and fused with the SphI-HindIII fragment from pB18(3028-4874)PUC giving pBl1(1-4874) (Fig. 1). Finally, pBl1(1-4874) and pB15(3816- 6746)PUC were each digested with XmalII (nt 4730) and with HindlII, which cuts in the polylinker on the 3' side of the insert of both clones, and the XmalII-HindlII fragment containing the 3"-terminal sequence of pB15(3816- 6746)PUC was substituted for the short XmalII-HindlII fragment of pB 11 (1-4874) to yield the recombinant DNA plasmid pB11(1-6746) which contains a complete DNA copy of RNA-1.

In vitro transcription. Transcription reactions (25 ~tl final volume) contained 40 mM-Tris-HCl pH 7.5, 6 mM- MgC12, 2 mM-spermidine, 10 mM-dithiothreitol (DTT), 80 ktg/ml bovine serum albumin (Bethesda Research Laboratories), 400 IXM each of ATP, UTP and CTP, 25 p.M-GTP, 500 ixM-mTGpppG (Pharmacia), 40 units of RNasin (Stratagene), 2 ~tg of linearized plasmid DNA and 10 units of bacteriophage T7 RNA polymerase (Stratagene). Incubation was at 37 °C. At intervals of 30 min during the reaction, 1 ~tl of 625 ttM-GTP was added to the mixture. After 120 rain, the mixture was extracted with an equal volume of phenol/chloroform and then with diethyl ether. An equal volume of 4 M-ammonium acetate and 10 Ixg of calf liver tRNA (Boehringer) were added to the collected aqueous phase and nucleic acid was precipitated overnight at - 2 0 °C from 67~ ethanol. After centrifugation, the pellet was washed twice with 70~ ethanol, dried and dissolved in 3 to 5 Ixl of sterile water. The length of the transcript was evaluated by formaldehyde-agarose get electrophoresis and Northern hybridization using 32p-labeUed antisense transcripts as probe (Lemaire et al., 1988). Sucrose gradient centrifugation of transcripts was carried out for 16 h at 30000 r.p.m, at 5 °C in a Beckman SW41 rotor. The gradient was 5~o to 20~ sucrose in 10 mM Tris-HCl pH 7.5, 1 mM-EDTA, 0.1 ~ SDS. Fractions of 0.5 ml were collected and concentrated by ethanol precipitation.

In vitro translation o f B N Y V V RNA-1 2361

Cell-free translation. In vitro translation in rabbit reticulocyte lysate and analysis of the 35S-labelled translation products by SDS-PAGE was as described by Ziegler et al. (1985) except that immediately following translation the reaction mixture was incubated at 50 °C for 15 min with 3 mM-iodoacetamide to prevent possible reoxidation of translation product sulphhydryl groups (Lane, 1978). Translation in wheatgerm extract was as described by Marcu & Dudock (1974) with minor modifications. Wheatgerm (General Mills, Vallejo, Ca., U.S.A.) was separated from the bulk of the heavy debris by flotation on a mixture of carbon tetrachloride and cyclohexane (4.5 : 1 v/v) . Dried, intact embryos were gleaned, by visual inspection, from damaged embryos and residual debris. The rest of the procedure was carried out at 4 °C. Selected wheatgerm (1.5 g) was mixed with an equal volume of sterile glass powder and pulverized for 1 min in a cold mortar. A portion (2 ml) of extraction buffer (20 mM-HEPES pH 7°6, 1 raM-magnesium acetate, 100 mM-potassium acetate, 2 mM-calcium acetate and 1 mM-DTT) was added and the slurry was immediately centrifuged at 23000 g for 12 min in a Beckman JA20 rotor. The supernatant fraction was layered on a 15 ml Sephadex G-25M column equilibrated with 20 mM-HEPES pH 7.6, 160 mM- potassium acetate, 1"66 mM-magnesium acetate and 6 mM-DTT and 0.5 ml fractions were eluted with the same

t% buffer. Fractions with A260 greater than 0.09 and A260/Azso ratios of about 1.5 were pooled and centrifuged at 23000 g for 20 min. Aliquots of the supernatant fraction were kept under liquid nitrogen.

Translation mixture (10 Ixl final volume) contained 4 ~tl wheatgerm extract, 17.5 mM-HEPES pH 7.6, 300 IXM- spermidine, 1 mM-ATP, 200 ktM-GTP, 2.8 raM-magnesium acetate, 40 rnM-potassium acetate, 10 n~- phospho(enol)pyruvate, 25 ~tg/ml calf liver tRNA, 100 Ixg/ml Escherichia coli rRNA (a gift from L. Pinck), 100 ~tra of each essential amino acid except methionine, 25 ixM-methionine, 600 IxCi/ml [35S]methionine (Amersham; 1500 Ci/mmol) and 50 to 100 Ixg/ml BNYVV RNA or transcript. Incubation was at 25 °C for 150 min followed by iodoacetamide treatment and SDS-PAGE as described above. Partial digestion of translation products with Staphylococcus aureus V8 protease was as described by Ziegler et al. (1985). Polypeptide M r markers (Bethesda Research Laboratories) were myosin heavy chain (223K) (Molina et al., 1987), phosphorylase b (97.4K), bovine serum albumin (68K), ovalbumin (43K), ct-chymotrypsinogen (25.7K), fl-lactoglobulin (18.4K), lysozyme (14.3K), cytochrome c (12.3K), bovine trypsin (6.2K) and insulin (3K).

RESULTS

Translation o f B N Y V V R N A in wheatgerm extract

In our earl ier investigation of cell-free translat ion of BNYVV R N A (Ziegler et al., 1985) the longest t ranslat ion product observed for RNA-1 in reticulocyte lysate was about Mr 220K. This polypept ide was previously described as 200K (Ziegler et al., 1985) because it comigrated with the heavy chain of myosin, a marker assigned an Mr of 200K in the manufacturer ' s (Bethesda Research Laboratories) product description. However, nucleotide sequence analysis has recently established an Mr for chicken myosin heavy chain of 222 559 (Molina et al., 1987) and we have accordingly adjusted upward the est imated Mr of the RNA-1 translat ion product. Similarly, the second longest RNA-1 reticulocyte lysate translat ion product, previously assigned an Mr of 150K has had its Mr value revised upward to 170K. Notwi ths tanding these Mr revisions, it is evident that size est imates for such long polypeptides by P A G E are subject to considerable uncertainty and, consequently, it cannot be said whether the largest observed reticulocyte lysate RNA-1 translat ion product in fact corresponds in size to the maximum RNA-1 coding capaci ty of Mr 237 389.

The foregoing considerations prompted us to reinvestigate the translat ion of BNYVV RNA-1. For this purpose we have used synthetic RNA-1 transcripts and an addi t ional cell-free translat ion system, wheatgerm extract. The concentrat ions of K +, Mg 2+ and spermidine added to the extract were adjusted to give about 65 % of the max imum extent of incorporat ion of [35S]- methionine and the most efficient synthesis of long polypeptides. When translated under these conditions, R N A from BNYVV isolate F13 directed ~ynthesis of equal amounts of two very long polypeptides (Fig. 2, lane 1), one comigrat ing with the 220K product synthesized in reticulocyte lysate (Fig. 2, lane 2) and the other migrat ing slightly more slowly, which we shall refer to as the 240K polypeptide. The prominent reticulocyte lysate t ranslat ion products P75 and P25 seen in Fig. 2 lane 2 are encoded by BNYVV R N A - 2 and RNA-3, respectively (Ziegler et al., 1985). Nei ther species is synthesized in abundant amounts in wheatgerm extract.

As noted above the maximum coding capaci ty of RNA-1 is Mr 237389 and thus the open reading frames for the 220K and 240K translat ion products in the wheatgerm system must overlap extensively. The products of V8 protease digestion of the two polypeptides were indist inguishable (data not shown), further indicat ing that the two coding regions overlap in

2362

l 2

240K x,__ 220K /

I. JUPIN AND OTHERS

1 2

1 2 3 4

,40K~_ 20K /

Fig. 2 Fig. 3 Fig. 4

Fig. 2. Cell-free translation products of BNYVV (F13) RNA in wheatgerm extract (lane 1) and rabbit reticulocyte lysate (lane 2). Autoradiogram of an 8~ polyacrylamide-SDS gel containing [3sS]- methionine-labelled translation products. Mr values and the prominent reticulocyte lysate products, P75 and P25, are indicated.

Fig. 3. Autoradiogram of a Northern blot of in vitro transcript tl 1(1-6746) (lane 1) and BNYVV RNA isolate (F3) (lane 2).. Transcript was synthesized from HindlII-linearized pBl1(1-6746) and electrophoresed in a 1~ formaldehyde-agarose gel. After transfer to nitrocellulose, RNA was vizualizcd by hybridization using 32p-labelled antisense RNA as probes (Lemaire et al., 1988). The arrow indicates the incomplete RNA-1 transcript referred to in the text. The numbers to the right indicate the positions of BNYVV RNA-1, -2, -3 and -4.

Fig. 4. Autoradiograms of cell-free translation pr~iucts of BNYVV RNA (lanes 1 and 3) or full length RNA-1 transcript (lanes 2 and 4) in rabbit reticulocyte lysate (lanes l and 2) and wheatgerm extract (lanes 3 and 4) fractionated in an 8~ polyacrylamide-SDS gel. Arrows indicate transcript-specific translation products referred to in the text. Mr values and the prominent reticulocyte lysate product, p75, are indicated.

frame. BNYVV RNA-1 also directed synthesis in wheatgerm extract of a 15K polypept ide (see below; Fig. 6) but polypeptides corresponding to the 170K and 50K species made in reticulocyte lysate were not produced.

Translation o f a B N Y V V R N A - 1 transcript

In vitro t ranscript ion of pB 11 (1-6746) l inearized with Hindl I I produced a major R N A species, which we shall refer to as t l 1(1-6746), with an electrophoretic mobil i ty very similar to that of authentic RNA-1 (Fig. 3) and a minor approx. 3400 nt RNA-l -spec i f ic t ranscr ipt (see arrow). The exact map coordinates of the minor species on the RNA-1 sequence have not yet been established, but it hybridized specifically with a 32p-labelled antisense R N A probe corresponding to the 3 '- terminal por t ion of the RNA-1 sequence. Thus it appears that the minor species shown in Fig. 3 arose by transcript ion ini t iat ion in the interior of the RNA-1 c D N A sequence rather than premature terminat ion of t ranscript ion ini t iated at the consensus T7 promoter.

Fig. 4 shows a comparison of [35S]methionine-labelled translat ion products obta ined using t l 1(1-6746) as messenger in rabbi t reticulocyte lysate (Fig. 4, lane 2) and wheatgerm extract

In vitro translation of BNYVV RNA-1 2363

Table 1. Calculated Mr of possible translation products of truncated RNA-1 transcripts for different translation initiation sites

Transcript

t 11 (1-638)

t11(1-1545)

Translation Translation product initiation s i t e (calculated Mr)

AUG(154) 18710 AUG(181) 17676 AUG(232) 15916

AUG(154) 52868 AUG(439) 42002 AUG(496) 39972 AUG(559) 36646

(Fig. 4, lane 4) with the translation products directed by authentic BNYVV RNA (Fig. 4, lanes 1 and 3). Non-fractionated virion RNA was used in preference to purified RNA-1 in this experiment because of the difficulty of eluting the long RNA-1 molecule from polyacrylamide- agarose gels in reasonable yield without degradation. Comparison of the gel profiles reveals that the translation products directed by the synthetic transcript and authentic RNA-1 in the two systems are very similar. The 220K, 170K and 50K polypeptides previously identified as major RNA-1 translation products in reticulocyte lysate (Ziegler et al., 1985) were also produced by the synthetic transcript (Fig. 4, lanes 1 and 2). Furthermore, many of the minor bands produced by translation of BNYVV virion RNA are also obtained with the RNA-1 transcript (Fig. 4). This means that these minor products were derived from RNA- I, probably by premature termination of translation, and that the synthetic transcript faithfully reproduces this behaviour. Note that one comigrated with the 75K band obtained by translation of virion RNA, indicating that the latter band is probably in fact a mixture containing the abundant RNA-2 translation product mentioned above and lesser amounts of an RNA-1 partial translation product.

Translation of tl 1 (1-6746) in wheatgerm extract (Fig. 4, lane 4) produced the same 220K and 240K species observed with natural BNYVV RNA (Fig. 4, lane 3). There was, however, one translation product of approx. Mr 120K in the wheatgerm extract and 115K in reticulocyte lysate which was synthesized with t11(1-6746) but that did not have a counterpart when natural BNYVV RNA was the template (see arrows in Fig. 4, lanes 2 and 4). In order to determine the size of the RNA encoding this polypeptide, the RNA-1 transcription products were first fractionated by sucrose gradient centrifugation before in vitro translation of the gradient fractions in reticulocyte lysate. The messenger activity for the 115K polypeptide did not comigrate with the full-length transcript in the gradient but rather with the minor 3400 nt species referred to above (data not shown). We conclude that the supplementary polypeptide arose by translation of an incomplete RNA-1 transcript, probably the aforesaid 3400 nt species, rather than incomplete translation of full-length tl 1(1-6746).

Translation of truncated RNA-1 transcripts

The 240K wheatgerm RNA-1 translation product and the 220K reticulocyte lysate product presumably have either their N termini or their C termini in common. The 220K polypeptide could thus arise by premature termination of translation, post-translation cleavage or initiation at a different site. In order to test for different initiation sites we constructed transcription vectors pB 11 (1-638) and pB 11 (1-1545) containing the 5'-terminal 638 nt or 1545 nt of the RNA-1 cDNA insert (Fig. 5). When the run-off transcript of tl 1(1-638) was translated, a polypeptide of approx. 18K was synthesized in both translation systems (Fig. 6, lanes 3 and 4). A polypeptide of about this Mr would be produced after translation initiation at either AUG(154), the first initiation codon in the RNA-1 sequence, or at the next in-phase initiation codon, AUG(181) (Table 1). From the measured Mr of the tl 1(1-638) translation product one cannot distinguish, between these two possibilities, but circumstantial evidence strongly favours AUG(154) as the functional initiation site because its sequence context, CAACAUGGC, resembles the consensus initiation sequences AACAAUGGC and CACCAUG reported for plants and animals

2364

pB11(1-638)-~ ~L • J k

AUG 154181

I. JUPIN AND OTHERS

H

II I i i 496

18.7K

H c ~Hc] I H /

pBl1(1-1545) "__ ', ~- k I I I k I I I I I I U I

AUG 154181 439 496

40.0K

52.9K

H e

pBl 1 (436-1545)--~ ~ . I I l I I I II II I

AUG 439496

,~ 40.0K

42.0K r

Fig. 5. Structure of recombinant cDNA plasmids containing 5'-terminal portions of the RNA-1 sequence. The filled circle represents the T7 promoter and the small vertical arrow the site of transcription initiation. Thin lines represent vector sequences and the heavy lines and hollow rectangles represent 5' non-coding and coding portions of the RNA-1 sequence, respectively. Short vertical bars indicate the position of in-phase AUG codons with the position of the A residue in certain of the triplets written underneath. Horizontal arrows indicate possible translation products of run-off transcripts and their Mr values. Restriction sites are marked as in Fig. 1.

respectively (Lutcke et al., 1987) somewhat more than does that of A U G ( I 81) ( U C C A A U G G A ) and because eukaryotic translation usually initiates at the first contextually favourable codon encountered from the 5' terminus (Kozak, 1986). Transcript t11(1-638) also directed the synthesis of a polypeptide that comigrated with the previously mentioned 15K species synthesized when natural BNYVV RNA-1 was translated in wheatgerm extract (Fig. 6, lanes 1 and 3). As shown in Fig. 6 (lane 2), a similar species is also encoded by tl 1(1-1545). The 15K species from natural and synthetic R N A produced identical patterns of partial degradation products after V8 protease digestion (data not shown). These observations thus demonstrate that the 15K species is derived from the Y-terminal portion of RNA-1, presumably as a result of premature translation termination in the wheatgerm system.

Translation of t 11 (1-1545) in wheatgerm extract produced equal amounts of two polypeptides of approx. Mr 51K and 40K (Fig. 6, lane 2). The same two species were produced in reticulocyte lysate, but the 40K polypeptide was considerably more abundant (Fig. 6, lane 5). The 51K polypeptide had about the Mr expected for translation initiation at AUG(154), which we have argued above to be the site of initiation for synthesis of the 18K polypeptide for t 11 (1-638). The 40K polypeptide, on the other hand, almost certainly resulted from an internal initiation event. Initiation at either AUG(439) or AUG(496) would give a polypeptide of about the observed Mr (Table 1). To discriminate between these two possible initiation sites, pBl1(1-1545) was modified by deletion of the sequence between the HinclI site in the polylinker preceding the c D N A insert and the HinclI site at nt435 (Fig. 5). In the resulting construction, pB 11 (436-1545), the first 435 nt of the RNA-1 insert have been eliminated and AUG(439) positioned as the first

In vi t ro translation of BNYVV RNA-1 2365

C P - -

15K- -

1 2 3

- - 5 1 K

- - 4 0 K - -

- - 1 8 K _ _

4

Fig. 6. Autoradiogram of a gel containing [3SS]methionine-labelled translation products of BNYVV RNA (lane 1), tl 1(1-1545) (lanes 2 and 5) and tl 1(1-638) (lanes 3 and 4) obtained in wheatgerm extract (lanes 1, 2 and 3) or rabbit reticulocyte lysate (lanes 4 and 5). Electrophoresis was in a 15~ polyacrylamide-SDS gel. The arrows indicate the top of the resolving portion of the gel. The positions of viral coat protein (CP) and the 15K polypeptide referred to in the text are indicated on the left. Other My values are also indicated.

1 2 3 4 5

C P - -

~ 5 1 K

- - 4 1 K

40I<

15K- -

Fig. 7. Autoradiograrn of a gel containing [35S]methionine-labelled translation products of BNYVV RNA (lane 1), tl 1(1-1545) (lanes 2 and 5) and tl 1(436-1545) (lanes 3 and 4) obtained in wheatgerm extract (lanes 1, 2 and 3) or rabbit reticulocyte lysate (lanes 4 and 5). Electrophoresis was in a 15~ polyacrylamide-SDS gel. Annotations as in Fig. 6.

potent ia l in i t ia t ion codon in the sequence. The t ranscr ip t t l 1 (436-1545) directed the synthesis of two polypeptides, one of which comigra ted wi th the 40K polypept ide while the other, more a b u n d a n t species migra ted slightly more slowly (Fig. 7, lanes 3 and 4). The difference in mobi l i ty

2366

(a)

- " ~ 15K

I . J U P I N A N D OTHERS

224K

~- 237K AUG 154496

-[ J--AAA

(b) ~ 224K

50K

170K

Fig. 8. Proposed map of major RNA-1 cell-free translation products in wheatgerm extract (a) and rabbit reticulocyte lysate (b).

corresponds to no more than a 1K to 2K difference in Mr. The most probable interpretation of these observations is that the abundant 41K polypeptide resulted from initiation at the 5'- proximal initiation codon AUG(439), as the surrounding sequence, CAACAUGGG, offers a good initiation context, and the 40K species resulted from initiation at AUG(496). Consequently, it follows that AUG(496) is the probable site tbr internal translation initiation on tl 1(1-1545) and by inference on full-length BNYVV RNA-1 (Fig. 8).

DISCUSSION

The use of a synthetic BNYVV RNA-1 transcript rather than native virion RNA in translation studies such as those described above has several important advantages. First, the transcript is easy to prepare and is absolutely free of other viral components, a goal not easily achieved by purification of virion RNA. Thus the translation experiments with synthetic RNA-I confirm the origin of the various abundant polypeptides previously identified as RNA-1 translation products on the basis of experiments with gel-purified virion RNA (Ziegler et al., 1985). Furthermore, as the transcript is derived from a clone it has a unique sequence, in contrast to the possible mixture of sequence variants in virion RNA populations, Thus the long overlapping translation products observed with both natural RNA-1 and synthetic transcript cannot have been due to sequence heterogeneity in the RNA population but must represent a fundamental translation property of the RNA molecule. Finally, the various manipulations possible at the DNA level before transcription permit genome organization and expression to be studied in detail.

The results described in this paper identify two translation initiation sites on RNA-1 with comparable affinity for wheatgerm ribosomes. Translation from these two sites at AUG(154) and AUG(496) would yield polypeptides with calculated Mr values of 237 389 and 224 493 (Fig. 8). In reticulocyte lysate, on the other hand, it appears that initiation occurs exclusively at the internal site AUG(496), as only the smaller of the two long polypeptides can be detected when full-length RNA-I is used as messenger. When the T-truncated transcript t11(1-1545) was translated in reticulocyte lysate, both AUG(496) and AUG(154) initiate translation, although the internal site is considerably more active. Preliminary experiments with other T-truncated transcripts indicate that initiation at AUG(154) is essentially abolished on transcripts of half the full length or longer (data not shown). Thus it is possible that a longer RNA chain can fold back upon itself in such a way that AUG(154) is rendered inaccessible to reticulocyte ribosomes but not to wheatgerm ribosomes.

As noted above, the sequence around AUG(154) resembles the consensus sequence context for initiation. It matches the animal consensus at three of four positions, and the plant consensus at three of six positions including the A at position - 3 , which is thought to be of special importance (Kozak, 1986). Indeed no other of the 65 in-phase A U G codons in the RNA-1 sequence conforms better to the animal consensus context, and only one, AUG(1225) (AAAAAUGGU), conforms better to the plant consensus context. The codon context of AUG(496), UUAUAUGUU, on the other hand, deviates considerably from the consensus.

In vitro translation o f B N Y V V RNA-1 2367

Perhaps the folding of the R N A molecule or other as yet unrecognized characterist ics of the sequence compensates for the poor codon context. Cowpea mosaic virus (CPMV) M R N A resembles BNYVV RNA-1 in having two translat ion ini t iat ion sites. However, translat ion of CPMV M R N A initiates at both the 5 '-proximal A U G codon and at an internal A U G (at either posit ion 512 or 524) in both wheatgerm and rabbi t reticulocyte systems (Pelham, 1979; Franssen et al., 1982; Vos et al., 1984). Both the candidate internal ini t iat ion codons have sequence contexts similar to the consensus.

I t is likewise of interest to compare the in vitro t ranslat ion strategy employed by BNYVV with that of soil-borne wheat mosaic virus (SBWMV), a bipar t i te plus-strand R N A virus with a fungal vector, Polymyxa graminis, similar to Polymyxa betae and with a part icle morphology resembling that of BNYVV (Shirako & Brakke, 1984). R N A II of SBWMV, like BNYVV RNA-2 (Ziegler et al., 1985), can undergo translat ional suppression of the terminat ion codon of its 5"-terminally situated coat protein cistron to produce long readthrough polypeptides (Hsu & Brakke, 1985; Shirako & Ehara, 1986). In wheatgerm extract, S B W M V R N A I encodes two polypeptides of 220K and 150K (Shirako & Ehara, 1986). The N termini of these polypeptides have not yet been mapped but the possible existence of dual init iat ion sites on SBWMV R N A I would represent a second point of similari ty between the translat ional strategies employed by the two viruses. The two R N A s clearly differ, however, in their behaviour in reticulocyte lysate; as shown in this paper BNYVV RNA-1 initiates translat ion at the internal A U G in this system, while the results suggest that the more Y-proximal A U G is employed for SBWMV R N A I (Shirako & Ehara, 1986). It remains to be determined how the in vitro t ranslat ion propert ies of the long BNYVV and SBWMV R N A s are related to their behaviour in virus-infected plant tissue.

REFERENCES

BOUZOUBAA, S., GUILLEY, H., JONARD, G., RICHARDS, K. & PUTZ, C. (1985). Nucleotide sequence analysis of RNA-3 and RNA-4 of beet necrotic yellow vein virus, isolates F2 and G1. JournalofGeneral Virology 66, 1553-1564.

BOUZOUBAA, S., ZIEGLER, V., BECK, D., GUILLEY, H., RICHARDS, K. & JONARD, G. (1986). Nucleotide sequence of beet necrotic yellow vein virus RNA-2. Journal of General Virology 67, 1689-1700.

BOUZOUBAA, S., QUILLET, L., GUILLEY, n., JONARD, G. & RICHARDS, K. (1987). Nucleotide sequence of beet necrotic yellow vein virus RNA-1. Journal of General Virology 68, 615~526.

FRANSSEN, H., GOLDBACH, R., BROEKHUYSEN, M., MOERMAN, M. & VAN K.AMMEN, A. (1982). Expression of middle- component RNA of cowpea mosaic virus: in vitro generation of a precursor to both capsid proteins by a bottom-component RNA-encoded protease from infected cells. Journal of Virology 41, 8-17.

nsu, v. n. & BRAKKE, M. K. (1985). Cell-free translation of soil-borne wheat mosaic virus RNAs. Virology 143, 272-279.

KOZAK, M. (1986). Regulation of protein synthesis in virus-infected animal cells. Advances in Virus Research 31, 229-292.

LANE, L. C. (1978). A simple method for stabilizing protei~sulfhydryl groups during SDS-gel electrophoresis. Analytical Biochemistry 86, 655-664.

LEMAIRE, O., MERDINOGLU, D., VALENTIN, P., PUTZ, C., ZIEGLER-GRAFF, V., GUILLEY, H., JONARD, G. & RICHARDS, K. (1988). Effect of beet necrotic yellow vein virus RNA composition on transmission by Polymyxa betae. Virology 162, 232-235.

LUTCKE, H. A., CHOW, K. C., MICKEL, F. S., MOSS, K. A., KERN, H. F. & SCHEEK, G. A. (1987). Selection of A U G initiation codons differs in plants and animals. EMBO Journal 6, 43-48.

MAReU, K. & DUOOCK, B. (1974). Characterization of a highly efficient protein synthesizing system derived from commercial wheatgerm. Nucleic Acids Research 1, 1385-1397.

MOLINA, M., KROPP, K., GULICK, J. & ROBBINS, J. (1987). The sequence of an embryonic myosin heavy chain gene and isolation of its corresponding cDNA. Journal of Biological Chemistry 262, 6478-6488.

PELHAM, H. R. B. (1979). Synthesis and proteolytic processing of cowpea mosaic virus proteins in reticulocyte lysates. Virology 96, 463-477.

SHIRAKO, V. & BRAKKE, M. K. (1984). Two purified RNAs of soil~borne wheat mosaic virus are needed for infection. Journal of General Virology 65, 119 127.

SmRAKO, Y. & ErIARA, V. (1986). Comparison of the in vitro translation products of a wild-type and a deletion mutant of soil-borne wheat mosaic virus. Journal of General Virology 67, 1237-1245.

TAMADA, T. (1975). Beet necrotic yellow vein virus. CMI/AAB Descriptions of Plant Viruses, no. 144. vos, P., VERVER, J., VAN WE ZENBEEK, P., VAN KAMMEN, A. & GOLDBACH, R. (1984). Study of the genetic organization

of a plant viral RNA genome by in vitro expression of a full-length DNA copy. EMBO Journal3, 3049-3053. ZIEGLER, V., RICHARDS, K., GUILLEY, H., JONARD, G. & PUTZ, C. (1985). Cell-free translation of beet necrotic yellow

vein virus RNA: readthrough of the coat protein cistron. Journal of General Virology 66, 2079-2087.

(Received 5 February 1988)