J. gen, Virol. (I977), 34, I-8 .Printed in Great Britain REVIEW ARTICLE Vesicular Stomatitis Virus : M o d e o f Transcription By A. K. B A N E R J E E , G. A B R A H A M AND R. J. C O L O N N O Department of Cell Biology, Roche Institute of Molecular Biology, Nutley, New Jersey o7I Io, U.S.A. SUMMARY Recent studies on the mechanism by which the virion-associated RNA polymerase of vesicular stomatitis virus transcribes RNA have revealed several new biological features of general interest. The mode of synthesis of the 5'-terminal cap structure of the mRNAs, the sequential transcription of the genes and the presence of a transcribed 'leader' RNA segment are properties which are either not shown by other viruses, or have not yet been described. These features are probably inter-related with the primary transcription process, which itself may be a useful mode/for future studies on mRNA biosynthesis in eukaryotic systems. It has become apparent in recent years that the animal viruses that contain single-stranded RNA as their genetic material employ different strategies for their expression during productive infection (Baltimore, I97i ). For example, the negative stranded viruses which include the rhabdoviruses, the paramyxoviruses and the myxoviruses, are different from other RNA-containing virus groups in that their genomes do not appear to be used directly as messenger RNAs. Instead, the RNA isolated from the polysomes of infected cells is complementary to the genome RNA. The transcription and replication mechanisms occurring in these virus groups have become the centre of study in several laboratories. Since vesicular stomatitis virus (VSV), which is the prototype of the rhabdovirus group, multiplies rapidly in a valiety of mammalian tissue culture cells, its biology has been studied in some detail. In particular, recent studies on its mode of transcription have revealed several unique features which may have broader implications for transcription in general. This article summarizes recent advances made in understanding the mechanism of VSV transcription and its possible relationship to genome replication. Transcription and translation of VSV mRNAs VSV contains a linear single-stranded genome RNA of approximate mol. wt. 4 x io 6 which is packaged within a characteristic bullet-shaped particle containing five structural proteins (Wagner, I975). The structural proteins consist of the glycoprotein (G, tool. wt. 7oooo) located as surface projections on the outer membrane of the virion; the matrix polypeptide (M, mol. wt. 29ooo) which is non-glycosylated, and the constituent protein of the virus membrane; the nucleocapsid protein (N, mol. wt. 5o ooo) which is tightly associated with the single-stranded genome RNA to form the ribonucleoprotein core with two other proteins present in low levels (L, mol. wt. ~7o ooo and phosphoprotein NS, mol. wt. 40 ooo) (Wagner et aL I972). Purified VSV genome RNA is non-infectious, and this observation led to the discovery of 2 A.K. BANERJEE, G. ABRAHAM AND R.J. COLONNO an RNA-dependent RNA polymerase in the purified virions (Baltimore, Huang & Stampfer, I97o). This RNA polymerase, upon tleatment with a non-ionic detergent, is activated and, in the presence of the four ribonucleoside triphosphates, synthesizes RNA in vitro which is complementary to the genome RNA. The RNA synthesizing activity is carried out by the ribonucleoprotein core particles containing only the N, NS, and L proteins using the coreassociated template genome RNA (Bishop & Roy, I972; Szilagyi & Uryvayev, I973). Moreover, Emerson & Wagner (I972) have shown that the RNA-N protein complex serves as the template for the transcriptase, and both the L and the NS proteins are required for the RNA synthesis (Emerson & Yu, ~975; [mblum & Wagner, I975). in addition, the core particles contain an activity that polyadenylates the newly-synthesized mRNAs in vitro by a tratlscription-dependent mechanism (Villarreal & Holland, 1973; Banerjee & Rhodes, I973). Recent studies using purified VSV have demonstrated that four distinct classes of RNA species are syntheaized in vitro having the sedimentation coefficients of 31S, x7S, ~4'5S, and IzS (Meyer & Banerjee, I975). By polyacrylamide gel electrophoresis and competition annealing with VSV genome RNA, these RNA species are found to be identical in size and base sequences to the mRNA species isolated from VSV-infected cells (Meyer et aL I975 ). The RNA species synthesized in vitro are functional mRNAs since they cart be translated efficiently in in vitro protein-synthesizing systems (Both, Meyer & Banerjee, I975a). Such studies indicate that the ~7S and the ~4"5S RNA species code for the G and the N polypeptides respectively, whereas the IzS RNA species contains two mRNA species (Rose & Knipe, I975; D. P. Rhodes & A. K. Banerjee, unpublished observations) and codes for both the NS and M polypeptides. A similar analysis using the corresponding mRNA species isolated from infected cells yielded identical results (Both, Meyer & Banerjee, I975b; Knipe, Rose & Lodish, 1975). Although the in vitro 3JS RNA has not yet been translated in vitro, its reel. wt. is compatible with its serving as the template for the L protein. The corresponding mRNA species isolated from infected cells codes for a protein similar in size to the L protein in a cell-free protein-synthesizing system (Morrison et al. I974). From the above results, it can be concluded that almost all of the genetic potential of the VSV genome RNA is expressed in vitro and in vi~o and presumbly the virion-associated RNA polymerase is responsible for initiating infection and the eventual replication of the virus. Mechanism o f m R N A transcription in vitro The intriguing problem that is faced in understanding the mechanism of VSV mRNA transcription is that the complete complement of the genome RNA (i.e. the 4zS + strand) has not yet been detected in in vitro synthesizing systems although it has been found in infected ceils (Morrison et al. 1975). Evidently, it is a required intermediate in the replication of the genome RNA. Since only the mRNAs coding for the virus structural proteins are transcribed in vitro, there must be either an alternate pathway for the synthesis of the 42S + strand by the virion-associated polymerase or a replicase activity that is different from the RNA polymerase(s). So far, no other virus-specific proteins apart from the five structural proteins have been detected in infected cells although a few minor protein bands are always seen in the gels of highly purified VSV (Abraham & B anerjee, 1976 a). It thus seems possible that the virion polymerase under certain conditions may serve dual functions as both the transcriptase and the replicase. Such an inter-relationship between transcription and replication in VSV has also been suggested by the use of a ts mutant of VSV belonging to the complementation Group I class which does not synthesize RNA at the non-permissive temperature (Pearhnan & Huang, T973). In order to study the relationship between transcription and replication in the VSV system, Review: VSV: mode of transcription 3 it was imperative to understand the precise mechanism by which the VSV mRNAs are transcribed in vitro using the genorne RNA as template. Two mechanisms have been proposed (Rhodes & Banerjee, I976). First, the virion polymerase may initiate RNA transcription independently at five sites on the single-stranded genome RNA. Presumably, the polymerase molecules must be located at these sites. Secondly, the polymerase may initiate transcription at the 3'-end of the genome and copy the entire length of the genome. By this method the completed monocristronie mRNAs are formed in sequential order and are released by a possible processing mechanism (Fig. ~). Alternatively, the polymerase may initiate at the 3'-end of the genome RNA and synthesize and complete each mRNA in turn before re-initiating transcription of the next. An indication of which of the above mechanisms was operative in the VSV system, should have been provided by a knowledge of the 5'- and 3'-terminal sequences of both the mRNA species and the genome RNA. If the mRNAs were initiated molecules one would expect the 5'-terminal nucleosides to be polyphosphorylated. The situation was complicated by the observation that all of the VSV mRNA species contained a unique 5'-terminal blocked structure having the form G(5')ppp(5')Ap... (Abraham, Rhodes & Banerjee, 1975 a). In the presence of the methyl donor, S-adenosylmethionine, the blocked structure was methylated as 7mG(5')ppp(5') A~... (Abraham, Rhodes & Banerjee, I975b ) indicating that purified VSV particles contain enzyme activities that will catalyse both the blocking and methylation of the 5'-termini. The latter type of structure has also been found in a variety of virus and eukaryotic mRNAs (Shatkin, r976). However, the origin of the phosphate groups forming the pyrophosphate linkage in the 'capped' structure of VSV mRNAs differed in that the s-phosphate of ATP and both the c~- and fl-phosphates of GTP were incorporated into the blocked structure. In the corresponding structures of vaccinia and reovirus mRNAs, only the o-phosphate of the blocking GTP is incorporated, indicating that the biosynthesis of these virus mRNAs is different from that found for VSV. Since radioactivity derived from /3, y-'~P-ATP was not il~corporated into the 5'-termini of the VSV mRNA species, it was difficult to ascertain whether these mRNA molecules were separately initiated. Moreover, all of the VSV mRNA species were found to contain a commonI 5'-terminal sequence G(5')ppp (5')ApApCpApGp... (Rhodes & Banerjee, I976). When the 3'-terminal trinucleotide of the genome RNA, ...PypGpUo~ (Banerjee & Rhodes, ~976) was compared with the 5'-terminal sequence of the VSV mRNAs, it became apparent that they were not complementary, suggesting that the mRNAs were transcribed fi'om regions remote from the 3'-end of the genome RNA. Furthermore, to synthesize the complementary 42S + strand, a necessary intermediate in virus replication, the transcription must be initiated at the 3'-terminus. A possible interpretation of these observations was that both types of RNA initiated similarly at the same site and that mRNA production involved subsequent processing while the 4zS + strand was not cleaved. Sequential transcription of the genes The effect of ultraviolet (n.v.) irradiation on purified VSV particles was used to resolve whether individual mRNAs result from five independent initiations or from a single initiation event with sequential read-through of the genes. Ultraviolet light produces pyrimidine dimers in nucleic acids (Wacker, I963), and when purified VSV is subjected to such irradiation, it is expected that the transcription of the genome RNA will be prevented beyond the damaged site (Michalke & Bremer, ~969). IfVSV mRNAs are synthesized by independent initiation events the sensitivity of each gene to u.v. irradiation should be proportional to the size of the gene, i.e. the target-sizes of the genes should be proportional to the tool. wt. 4 A.K. BANERJEE, G. ABRAHAM AND R. J. COLONNO of the corresponding mRNA species. On the other hand, if the mRNAs are synthesized sequentially from a single initiation site, the transcription of a particular gene will depend on the prior transcription of its 3'-proximal genes. Using this approach, Ball & White U976) studied the expression of the genes of VSV in a cell-free system derived from mouse L cells which executed coupled transcription and translation of the virus genes. Detergent-activated VSV was allowed to synthesize mRNAs in vitro, which were immediately translated, and the synthesis of each individual gene product was then measured after polyacrylamide gel electrophoresis, in reponse to varying doses of irradiation, the kinetics of sy,thesis of each virus polypeptide in the coupled system displayed a characteristic single-hit inhibition. The apparent target-sizes for the expression of the different virus genes were calculated relative to the entire genome as measured by the loss of infectivity. It was observed that the target size of the N protein gene corresponded to the mol. wt. of its mRNA. On the other hand, the target sizes of the other genes appeared to be cumulative, indicating that transcription was initiated at a single point on the VSV genome RNA and proceeded in the order 3'-N-NS-M-G-(L)-5'. Although the L protein gene was not expressed in this sytem, it was concluded from a consideration of its size and from genetic data that it must be located close to the 5'-region of the template RNA. Independent studies by Abraham & Banerjee (I976 b) showed a similar polar effect of u.v. irradiation on the transcription of the VSV genome. In these studies, the inhibition of synthesis of the individual VSV mRNA species (rather than their gene products) was examined in response to the u.v. irradiation. Target size calculations showed that only the tool. wt. of the mRNA coding for Nprotein was proportional to its target size. The remaining genes had target sizes larger than expected, ranging up to the tool. wt. of the entire genome RNA of VSV. This analysis also allowed the mapping of the genes of VSV in the order 3'N-(NS-M)-G-L-5'. Since the mRNAs coding for the NS and M proteins could not be separated due to their very similar sizes, the exact orientation of their genes in the map was not possible. Subsequently, the order of these two genes was determined by examination of their in vitro translation products after appropriate u,v. irradiation (D. P. Rhodes & A. K. Banerjee, unpublished observation) and found to be the same as that observed by Ball & White 0976). The ' leader" R N A The above results strongly indicated a compulsory order for the synthesis ofVSV mRNAs, with the N-gene, located close to the 3'-terminus of the VSV genome RNA, being synthesized first and the L-gene, located close to the 5'-terminus, synthesized last. In addition, previous studies had shown that the 3'-terminal sequence of the genome RN A ...pPypGpL~oa was not complemerttary to the 5'-terminal sequence G ( 5 ' ) p p p ( 5 ' ) A p A p C p A p G p . . . found in VSV mRNA species. Thus, if the RNA polymerase initiated transcription exactly at the 3'-terminus of VSV genome RNA, an RNA product should be present which contained a 5'-terminal sequence complementary to the 3'-terminal sequence of the genome RNA. Indeed, a small RNA product has recently been isolated from the R N A products of a VSVprimed in vitro transcription reaction which contained an unblocked, polyphosphorylated and unmethylated 5'-structure with the sequence (p)ppApCpGp (Colonno & Banerjee, r976). This sequence is complementary to the 3'-terminal trinucleotide of the genome R N A . The small RNA product (approx. 7o bases long) is rich in adenosine (48 ~) but was not polyadenylated since it did not bind to oligo(dT)-cellulose. Further experiments have shown that this RNA product maps at the 3'-terminal region of the genome RNA and is the first product to be synthesized during VSV RNA transcription in vitro (R. J. Colonno & A. K. Banerjee, Review: VSV: mode of transcription 5' (p}ppApCpGp ] G I M I Ns I N ] PypGpU 3' / 5"3' ~ / .... GpApCpApApNpNpNp ........ pGpCpApp(p) 5 iI CLEAVAGE LEADER RNA I ./ .... GpApCpApAp.N~p-ppG ~5' 5~ Pi .... GpApCpApAp-ppG CAP NpNpNp .... 3" NpNpNp .... POLY (A) Fig. 1. Modelfor the biosynthesisof VSV mRNAs. unpublished results). These results indicated that this small RNA molecule probably represented an initiated lead-in RNA segment which is removed during formation of VSV mRNAs by a possible processing mechanism. A model for VSV transcription and replication As yet no specific function has been assigned to this 'leader' RNA. The following model for VSV mRNA biosynthesis has recently been proposed (Fig. I; Colormo, Abraham & Banerjee, t976). A similar mechanism has also been proposed by Ball, White & Collirts (1976). First, the virion-associated RNA polymerase initiates transcription at the 3'-terminus of the template genome RNA to give a product with an unblocked, 5'-terminal sequence (p)ppAp CpGp..., the y-phosphate of which is presumably removed by the virioo-associated phosphohydrorase (Wagner, ~975). Secondly, the remainder of the genome RNA continues to be transcribed as the RNA polymerase moves toward the 5'-end of the genome. Thirdly, a 'processing' enzyme recognizes specific sequences (...GpApCpApApNpNpNp...)on the transcribed product RNA, cleaving it into cistron-sized mRNAs which retain a free 5'monophosphate, and the 'leader' RNA molecule. Fourth, the capping of the 5'-structures of the mRNAs with GDP takes place with concomitant polyadenylation of the 3'-ends of the cleaved RNA. Finally, irt the presence of SAM, the guanosine and the adenosine bases in the blocked structure are methylated as rinG and A respectively. A modification of this modelmay suggest a mechanism for the synthesis of the 4zS + strand. If the putative processing is inhibited in the infected cell, the RNA polymerase may transcribe the genome RNA ertd-to-end. Siuce the 42S + strand has not yet been detected in the in vitro system, the processing of the mRNAs is the dominant event in this situation. Presumably, a modification of the transcriptase occurs dining the infection cycle and the rep[icase thus formed can synthesize the 4zS + strand by the mechanism suggested above. Alternatively, the processing event may be prevented by a factor which accumulates in the infected cell. TM 6 A.K. BANERJEE, G. A B R A H A M A N D R. J. C O L O N N O The problems Although the above model for VSV transcription and replication seems to be straightforward, there are various results and observations which have to he reconciled before coming to a definite conclusion. Hefti & Bishop (I976) have recently shown that in addition to the 5'-terminal sequences pppApCpGp (e 5 to 35 ~) and G(5')ppp(5')ApApPyp (3o to 40 700),there are initiated sequences present in the VSV RNA products synthesized in vitro, such as pppApApPyp... (5 to IO ~), pppGpCp... (5 to I5 ~o), pppGpAp...(5 to I5 ~) and pppGpCp...(io to I5 70). The presence of a small amount of the unblocked sequence pppApApPyp.., suggests that it could be the precursor for the capped sequences. Thus, in this situation, each mRNA species could be independently initiated, and the capping reaction may occur on the 5'-terminus of the initiated RNA molecules. The origin and the role of the othel initiated RNA products have not been determine4. Since the RNA species synthesized by the VSV strain used in these studies are small and do not correspond to the sizes of any of the mRNA species synthesized in vivo, the implications of these results are unclear. However, these observations indicate that there may exist certain strain and growth differences between VSV Indiana isolates and that these differences may be reflected in the inability of the virus transcriptase to initiate or modify RNA correctly. It should also be pointed out that although the u.v. experiments described above indicate a compulsory order for the synthesis of VSV mRNAs they do not preclude the possibility of a ' stop-start' mechanism. By this process, the RNA polymerase may initiate transcription at the 3'-terminus to synthesize the 'leader RNA' and the same enzyme would re-initiate transcription of the N-protein gene, terminate, and initiate transcription of the adjacent gene and so on. The capping of the mRNAs would be concomitant with initiation. This model can explain the presence of some uncapped mRNA 5'-terminal sequences (Hefti & Bishop, ~976; Rose, I975) if the efficiency of the capping reaction is not one hundred percent. The production of the 42S + strand would then depend on the total inhibition of the initiation events except the initiation of the leader RNA. If the capping reaction with pppG indeed takes place on pppA-ended RNA molecules, then this reaction is expected to be inhibited by PP~ orP~, a fact that has been demonstrated in the vaccinia (Martin & Moss, 1975) and the reovirus sytems (Furuichi et al. 1976). But recent experiments show neither of these compounds inhibit the capping of VSV mRNAs suggesting that the capping reaction is more complicated in the VSV system than simple condensation of pppG and pppAp... (G. Abraham & A. K. Banerjee, unpublished observations). If the mRNA formation is mediated by processing (Fig. I), then the capping reaction can occur with pppG and pA-ended RNA molecules. This possibility is consistent with the experimental observations concerning the origin of the phosphate groups in the VSV mRNA cap structure and provides an explanation for the different mechanisms of cap formation found for the vaccinia and reovirus systems. Although the processing of VSV mRNAs has not been demonstrated directly, the evidence is compatible with such a mechanism. A similar type of processing of an mRNA precursor with the production of leader RNA has been demonstrated in the T 7 (Dunn & Studier, 1973; Rosenberg, Kramer & Steiz, 1974) and T3 (U. Maitra, personal comm unication) phage systems. If VSV mRNA processing from a transcribed precursor does occur, it implies the presence of a nucleolytic activity which must be associated with the core particles. Already, there are considerably more enzymatic activities associated with such particles than there are virus proteins (Wagner, i975) and so it remains to be seen if some of these activities are due to cellular components which co-purify with the virus particles. The Review: VSV: mode of transcription 7 fact that the polyadenylatin8, methylating, capping (and perhaps processing) activities in VSV are all transcriptiort dependent (Abraham & Banerjee, I976a) suggests that RNA synthesis is effected by a large transcriptive complex associated with the RNA template. At present it is not understood if any similarities exist between the biosymhesis of VSV mRNAs and the corresponding mechanisms operative in the formation of vaccinia virus mRNAs. Indeed, it is unclear if either of these mechanisms resembles that occurring in the derivation of eukaryotic mRNAs from the heterogeneous nuclear RNAs found in such cells. Further studies with the VSV system should help to clarify these questions and to establish the precise relationship between virus transcription and replication. REFERENCES ABRAHAM, G. & BANERJEE,A. K. (I976a). The nature of the R N A products synthesized in vitro by subviral components of vesicular stomatitis virus. Virology 7x, 23o-241. ABRAHAM, G. & BANERJEE,A. K. (I976b). Sequential transcription of the genes of vesicular stomatitis virus. Proceedings of the National Academy of Sciences of the United States of America 73, 15o4-15o8. ABRAHAM,G., RHODES,D. P. & BANERJEE,A. K. (I975 a). Novel initiation of R N A synthesis in vitro by vesicular stomatitis virus. Nature, London ~-55, 37-4o. ABRAHAM,G., RHODES,D. P. & BANERJEE,A. K. (1975b). The 5'-terminal structure of the methylated m R N A synthesized in vitro by vesicular stomatitis virus. Celt 5, 51 58. BALI., L. A. & WHITE, C. N. (1976). Order of transcription of genes of vesicular stomatitis virus. Proceedings of the National Academy of Sciences of the United States of America 73, 442-446. ~AL~, L.A., WrflTE, C.N. & COta3NS, P.L. (I976). A transcriptional map of vesicu]ar stomatitis virus. In Animal Virology, ICN-UCLA Symposia on Molecular and Cellular Biology, vol. 4 (in the press). Edited by D. Baltimore, A. S. Huang & C. F. Fox. New York : Academic Press. BALXI~,IORE,D. (I97I) Expression of animal virus genomes. Bacteriological Reviews 35, 235-241. BALTIMORE,D., HUANG,A. S. & STAMPFER,M. (I970). Ribonucleic acid synthesis of vesicular stomatitis virus. 1I. An R N A polymerase in the virion. Proceedings of the National Academy of Sciences of the United States of America 66, 572-576. BAI~EF,IEE, A. X. & RHODES, D. 1". (I973). In yitro synthesis of R N A that contains po]yadenylate by virionassociated R N A polymerases of vesicular stomatitis virus. Proceedings of the National Academy of Sciences of the United States of America 7o, 3566-357o. BANERJEE, A. K. &RHOOES, D.e. 0976). 3'-terminal sequence of vesicular stomatitis virus genome RNA. Biochemical and Biophysical Research Communications 68, 1387-i 394. BISHOe, D. H. L. & ROe, 1". (1972). Dissociation of vesicular stomatitis virus and relation of the virion proteins to the viral transcriptase. Journal of Firology xo, 234-243. BOTH, ~. W., MOVER,S. A. & r~ANERJEE,A. K. (I975a). Translation and identification of the m R N A species synthesized in vitro by the virion-associated R N A polymerase of vesicular stomatitis virus. Proceedings of the National Academy of Sciences c f the United States of America 72, 274-278Boa-ri, o. w., MOWR, S. A. & BANERJEE, A.K. 0975b). Translation and identification of the viral rnRNA species isolated from subcellular fractions of vesicular stomatitis virus-infected cells. Journal of Virology "r~, loi2_ioi9 ' COLONNO, R. J., ABRAHAM,G. & BANERJEE, A. K, (1976). Blocked and unblocked 5'-termini in vesicular stomatitis virus product R N A in vitro: their possible role in m R N A biosynthesis. In Progress in Nucleic AcM Research and Molecular Biology, vol. I9 (in the press). Edited by E. Volkin & W, E. Cohn. New York: Academic Press. coI.or~No, R. J. & BA~E~tJ~E,A. K. 0976). A uniqtlc R N A species involved in initiation of vesicular stomatitis virus R N A transcription in vitro. Cell 8, 197-2o4. DUNN, J. J. & STtJDIER, r. W. (1973). T7 early RNAs are generated by site-specific cleavages. Proceedings of the National Academy of Sciences of the United States of America 7o, 1559-1563. EMERSON, S. O. & WAGNER, R. R. (I972). Dissociation and reconstitution of the transcriptase and template activities of vesicular stomatitis B and T virions. Journal of Virology io, 297-3o9. EraERSON, S. U. & YU, V. ~1. 0975). Both NS and L proteins are required for in vitro RNA synthesis by vesicular stomatitis virus. Journal of Virology 15, 1348-t356. FURU1CHI, Y., MUTHUI(RISHNAN, S., TOMASZ, J. & SHATKIN,A. J. (I976). Mechanism of formation of reovirus m R N A 5'-terminal blocked and methylated sequence mTGpppG~aC. Journal of Biological Chemistry (in the press). HEETI, E. & msnoP, D. H. L. (1976). The 5' sequence of VSV in vitro transcription product RNA (+SAM). Biochemical and Biophysical Research Communications 68, 393 4o0. 8 A . K . B A N E R J E E ~ G. A B R A H A M A N D R . J . C O L O N N O IMSLUM, R. L. & WAG~,'~.R, V,. R. (1975). Inhibition of viral transcriptase by immunoglobulin directed against the nucleocapsid NS protein of vesicular stomatitis virus. Journal of Virology xS, 1357-1366. KNWE, D., ROS~, g. J. & I_OD~St10 H. F. (1975). Translation of individual species of vesicular stomatitis viral mRNA. Journal of Virology I5, ~oo4-IOI 1. MARTIN, S. A. & MOSS,a. (1975). Modification of R N A by m R N A guanyltransferase and m R N A (guanine-7-) methyltransferase from vaecinia virions. Journal of Biological Chemistry 250, 933o-9335. MICHALKE, n ~ BRES~R, 11. (1969). R N A synthesis in Eseheriehia coli after irradiation with ultraviolet light. Journal of Molecular Biology 4x, 1-23. MORRL~ON~ T,, STAMPFER, M,, BALTIMORE, D. & LODISH, H. F. (I974). Translation of vesicular stomatitis messenger R N A by extracts from mammalian and plant cells. Journal of Virology z3, 6z-72. MORRI.C~N~ T. G., STAMPFER, M., LODISH, H. F. & BALTIMORE, D. 0975). In vitro translation of vesicular stomatifis virus messenger RNAs and the existence of a 4oS 'plus' strand. In Negative Strand Firuses, vol. I, Pp. 293-300. Edited by B. W. J. Mahy & D. Barry. New York : Academic Press. MOVER, S. A. & BANERJEE,A. K. (1975). Messenger RNA species synthesized in vitro by the virion-associated R N A polymerase of vesicular stomatitis virus. Celt 4, 37-43. MOYER, S. A,, GRAUBMAN, M. J., EHRENFELD, E. & BANERJEE, A. K. (1975). Studies on the in vivo and in vitro messenger RNA species of vesicular stomatitis virus. Virology 67, 463-473. PVARLMAN,S. M. & HUAI~G,A. S. (1973). R N A synthesis of vesicular stomatitis virus. V. Interactions between transcription and replication. Journal of Virology x2, I395-I4OO. P~ODZS, D. P. & BANERJEE,A. K. (1976)- 5"-terminal sequence of vesicular stomatitis virus mRNAs synthesized in vitro. Journal of Virology x7, 33-42. ROSE, J. K. (I975). Heterogenous 5'-terrninal structures occur on vesicular stomatitis virus mRNAs. Journal of Biological Chemistry 250, 8o98-8to4. ROSE, J. K. & K~IPE, D. (I975). Nucleotide sequence complexities, molecular weights, and poly (A) content of the vesicular stomatitis virus m R N A species. Journal of Virology ~5, 994-1oo3. ROSENBERG, M., KRAMER, R. A. & STEITZ, S.A. (1974)- T7 early messenger RNAs are the direct products of ribonuclease HI cleavage. Journal of Molecular Biology 89, 777-782. S~tATgat% A. s. (1976). Terminal caps in eukaryotic mRNAs. New Scientist (in the press). SZlLAGYI, S. ~. & tmYVAYEV,L. (I973). Isolation of an infectious ribonucleoprotein from vesicular stomatitis virus containing an active RNA transcriptasc. Journal of Virology xx, 279-286. WLLARREAL, L P. & hOLLAND, J.J. (I973). Synthesis of poly(A) in vitro by purified virions of vesicular stomatitis virus. Nature New Biology 246, 1 ~ 9 . WACKER, A. (~963). Molecular mechanisms of radiation effects. Progressive Nucleic Acid Research of Molecular Biology x, 369-399. WAGNER, g. R. (I975). Reproduction of rhabdoviruses. In Comprehensive Virotegy, vol. 4, PP- 1-8o. Edited by H. Fraenkel-Conrat & R. R. Wagner. New York: Plenum Press. WAGNER, R. R., PREVEC, L., BROWN, F., SUMMERS, D, F., SOKOL, F. & MACLEOD, R. (1972). Classification of rhabdovirus proteins: a proposal. Journal of Virology xo, 1228-123o. (Received 20 August 1976 )