MOLECULAR EVENTS LEADING TO PICORNAVIRUS GENOME REPLICATION

Transcription

1 jf. Cell Sci. Suppl. 7, (1987) Printed in Great Britain The Company of Biologists Limited MOLECULAR EVENTS LEADING TO PICORNAVIRUS GENOME REPLICATION ECKARD WIMMER1, RICHARD J. KUHN1, STEVEN PINCUS1, CHEN-FU YANG1, HARUKA TOYODA1, MARTIN J. II. N ICKLIN 1 and NAOKAZU TAKEDA 2 1D epartm ent o f Microbiology, School of Medicine, State University o f N ew York at Stony Brook, Stony Brook, N ew York 11794, U.S.A. 2C entral V im s Diagnostic Laboratory, N ational Institute of Health, Gakuen 4-7-1, M usashim urayam a, Tokyo , Japan INTRODUCTION The Picornaviridae are a family of numerous human pathogens (enteroviruses and rhinovirus, altogether over 180 serotypes) that cause a bewildering array of disease syndromes. Well-known animal viruses such as encephalomyocarditis virus (EMC), a cardiovirus, and the economically important foot-and-mouth disease virus (FM DV), an aphthovirus, also belong to the Picornaviridae. The chemical and three-dimensional structures of several of these viruses have been solved (Kitamura et al. 1981; Forss et al. 1984; Palmenberg et al. 1984; Stanway et al. 1984; Potratz ei al. 1984; Rossmann et al. 1985; Hoglt e t a l. 1985). Moreover, the fine structure of the genetic map of some picornaviruses has been elucidated. Thus, the Picornaviridae may be the best characterized virus family to date (for references, see Koch & Koch, 1985; Rueckert, 1985; Kuhn & Wimmer, 1986; Nicklin^ al. 1986; Toyodaei al ; Wimmer ei al. 1986; Semler et al ; Palmenberg, 1987; Nomoto & Wimmer, 1987). Poliovirus, an enterovirus, has always been a prototype of the picornaviruses, and many key discoveries emerged from investigations with this human pathogen. The leading role of poliovirus in picornavirus research is undoubtedly due to the devastating and often lethal disease it inflicts upon its victims (even though only a small percent of infected individuals develop clinical syndromes). Following the successful cultivation of poliovirus in non-neural tissue by Enders, Weller & Robbins in 1949, every animal virologist seemed to have become involved in some aspect of poliovirus research. An enormous wealth of information on poliovirus has thus been gathered (Koch & Koch, 1985). The most important result of poliovirus research, of course, was the development of two efficacious vaccines (see references in Horstmann et al. 1984) and this has led to the control of poliomyelitis in developed and also in many developing countries. But it should be emphasized that poliomyelitis remains unchecked in many parts of the world. Genetic studies of the neutralization antigenic sites and of the attenuated phenotype have yielded fascinating results and an unexpected complexity (reviewed by Nomoto & Wimmer, 1987).

2 252 E. Wimmer and others We will review here mainly the molecular events leading to poliovirus genome replication. In nearly every aspect the events are similar for all picornaviruses, with the possible exception of the human hepatitis A virus (enterovirus 72). Virion structure The poliovirion is small (26 nm) and naked (non-enveloped). It consists of 60 copies of each of the capsid proteins VP1, VP2, VP3 and VP4, and one copy of a single-stranded RNA of plus-strand polarity (VP4 and VP2 are produced during the maturation cleavage of their precursor, VPO; 1-2 VPO molecules remain uncleaved in each virion). The RNA molecule has an unusual structure: (1) it is covalentlylinked to a peptide at the S' end (Lee et al. 1977; Nomoto et al. 1977a; Wimmer, 1982) and polyadenylated at the 3' end (Yogo & Wimmer, 1972) and (2) it contains an untranslated 5'-terminal region of 742 nucleotides, a single open reading frame of 6627 nucleotides followed by 72 untranslated 3'-terminal nucleotides (Fig. 1). The 7441 heteropolymeric bases of poliovirus type 1 (Mahoney) were sequenced first by Kitamura et al. (1981) who also reported the precise map position of all poliovirus proteins known at the time (see below). It is noteworthy that the genomic sequence, the first of any animal RNA virus, was obtained without the aid of molecular cloning (Kitamura & Wimmer, 1980). When permission was given by government agencies for molecular cloning of the poliovirus genome, total sequences of the RNA of poliovirus type 1 (Mahoney) and its attenuated Sabin 1 derivative were also obtained by Racaniello & Baltimore (1981a) and Nomoto et al. (1982), respectively. Since then, the virion RNA of all three poliovirus serotypes and of many other picornaviruses have been sequenced (see review by Palmenberg, 1987). The bizarre structure of the 5'-end of the poliovirus genome (VPg-pUU...), in which a 22 amino acid-long peptide (VPg) is covalently bound to the RNA via a 0 4-(5'-uridylyl)- tyrosine linkage (Lee et al. 1977; Nomoto et al. 1977a; Rothberg et al. 1978; Ambros & Baltimore, 1978; and references therein), has now been recognized in all picornavirus genomes. Similar 5 '-terminal peptides are present on many plant viral RNAs (see the article by van Kammen, this volume). The crystal structure of poliovirus (Hogle et al. 1985) has recently been obtained and was found to be nearly identical to that of human rhinovirus 14 (Rossmann et al. 1985). The virion is an icosahedron in which the capsid polypeptides VP1, VP2 and VP3 are to some extent intertwined (Rossmann et al. 1985; Hogle et al. 1985). The X-ray data confirmed earlier biochemical studies (Rueckert, 1985), in that the large capsid proteins are exposed to the outside whereas VP4 (7000 daltons) is entirely internal. Significantly, the large capsid polypeptides, VP1, VP2 and VP3, although different in amino acid sequence, form eight-stranded beta barrels that resemble each other. Their tertiary folds and quaternary organization form the building units of the icosahedron and are remarkably similar to structural elements found in viruses of other species, e.g. in plant viruses (for references see Rossmann et al. 1985). The elucidation of the crystal structure also allowed the precise mapping of the neutralization antigenic sites to different regions of the virion surface (Rossmann et a l. 1985; Hogle et al. 1985; see a review by Wimmer et al. 1986). Moreover,

3 Molecular events leading to picomavirus genome replication 253 Rossmann et al. (1985) have suggested that a distinct canyon formed in each structural unit of the virion may serve as the receptor attachment site. The genetic map The long open reading frame of the genome RNA, identified by nucleotide sequence analysis (Kitamura et al. 1981; Racaniello & Baltimore, 1981a; Nomoto et al. 1982) directs the synthesis of a polyprotein ( Mr) that was first predicted to exist by Jacobson & Baltimore (1968). The polyprotein is the only translation product known to be synthesized by poliovirus (Fig. 1). Indeed, a computer-aided evaluation of the coding capacity of the viral RNA does not reveal any other capsid -P I - * R - non-capsid P2 P3 Initiation of translation VPg-pU 5' *V-Poly (A) 60 ± n \ Termination of translation / VPO (37) n r - - VP4 (7) 1 VP2 (30) VP3(26)i VPI (34)i J J\ Í2C (38) VPg (2) \ 3Cpro (20) 3Dpo (52) \ 13C' (36-4) \3D' (35-6)\ Fig. 1. Gene organization of poliovirus and map of proteolytic processing. Virion RN A, terminated at the S' end with the genome-linked protein VPg and at the 3' end with poly (A ), is shown as a solid line, the translated region being more pronounced than the non-coding regions. T h e numbers above the virion RN A refer to the first nucleotide of the codon specifying the N-term inal amino acid for the viral specific proteins. T he coding region has been divided into three regions (PI, P2, P3), corresponding to rapid cleavages of the polyprotein. T h e newly adopted nomenclature of polypeptides is according to Rueckert & W immer (1984). Numbers in parentheses are calculated molecular weights. Open circles indicate that the terminal amino acids have been determined by sequence analysis. Closed circles indicate that the N -term ini are known to be blocked. Closed triangles: G ln-g ly pairs that are cleaved during proteolytic processing of a polypeptide by the virus-coded proteinase 3C. Open triangles: T yr-g ly pairs cleaved by viral proteinase 2A. Open diamond: Asn-Ser pair cleaved only during morphogenesis. Polypeptides 3C' and 3 D ' are products of an alternative cleavage, the biological significance of which is unknown (modified after Kitamura et al. 1981). T h e genome structure and processing map of poliovirus is probably applicable to all entero- and rhinoviruses, with the exception of hepatitis A virus (human enterovirus 72) that has a truncated VP4. Cardio- and aphthoviruses have a segment of poly(c) inserted into their 5' non-coding region. Moreover, their polyprotein begins with a leader polypeptide ( L ) that precedes the P I region. Finally, aphthoviruses have most of their polypeptide 2A within the polyprotein deleted. For further details, see text.

4 254 E. Wimmer and others overlapping reading frames within the long open reading frame that could code for extra polypeptides. Moreover, there exists no evidence suggesting that any of the very small reading frames in the long 5' untranslated region are expressed to yield small peptides in the infected cell (Dorner et al. 1982). As will be shown below, the mrna coding for the poly protein is identical in sequence to virion RNA except that the 5'-terminal VPg has been removed (VPg-pUU» pu U...). Thus, poliovirus mrna (and all picornavirus mrna) is uncapped, a unique feature among mammalian mrnas. Subgenomic poliovirus mrna species are not produced in the infected cell (see Fig. 2). The polyprotein is cleaved proteolytically at numerous sites to yield the final protein products (Fig. 1). Tedious micro-sequence analyses of all poliovirusspecified proteins (Kitamura et al. 1980a; Semlereia/. 1981a,b, 1982; Larsen et al. 1982; Dorner et al. 1982; YLmixnetal. 1982; Adler et al. 1983; Pallansch e ta l. 1984) led to the identification of the cleavage sites (see below) and confirmed the genetic map that was previously established by biochemical methods (reviewed by Rueckert et a l. 1978). As can be seen in Fig. 1, the capsid proteins map at the NH2-terminus of the polyprotein (the PI region), whereas the replication proteins (P2 and P3 regions) follow downstream. This organization of genetic elements contrasts with that of many other plus-strand viruses whose genomes often code for the capsid proteins in the 3'-terminal half of the RNA (see articles by Kaariainen and van Kammen; this volume). The pathway of gene expression, shown in Fig. 1, predicts that all viral proteins are synthesized in equimolar amounts. Experimental results seemed to contradict this prediction. It was proven to be correct, however, through careful quantitative analyses of all gene products, particularly after those polypeptides were taken into account that are the result of alternate cleavages of a single precursor (as seen for polypeptide 3CD) (Rueckert et al. 1978). The strategy of gene expression via polyprotein processing may be considered wasteful. On closer examination, however, it becomes apparent that the compacting of genetic information confers considerable advantages to the RNA virus. For example, regulatory sequences defining early v. late gene products can be eliminated and indeed do not exist in poliovirus RNA. As a consequence potential targets have been removed for random, and usually lethal mutation(s), that may occur as the result of error-prone RNA replication (Toyoda et al. 1986a; also see reviews by Holland et al. 1982; Reanny, 1984). In any event, the apparent over-production of the viral non-structural proteins cannot be unfavourable simply because poliovirus is one of the fastest and most productively growing animal viruses: the growth cycle is completed in 6-7 h, during which time a single virion has multiplied fold per cell. In vitro manipulations o f the poliovirus genome Two experimental achievements have greatly enhanced the opportunities for studying poliovirus molecular biology and genetics. The first achievement was the molecular cloning of the poliovirus genome (van der Werf et al. 1981; Racaniello &

5 Molecular events leading to picornavirus genome replication 255 REPLICATION COMPLEX Fig. 2. Schem atic presentation of the life cycle of poliovirus. T h e virion enters the cell by an unknown mechanism (possibly v ia endosomes) and releases its genome (solid line) that will engage in protein synthesis. Virus-specific m RN A is identical to virion RN A in nucleotide sequence but is unlinked from the 5'-term inal peptide VPg (black dot). T h e RN A replication com plex is m em brane bound; RN A structures found associated with this com plex are the R I, R F and ssr N A. T h e multi-stranded R I is usually of the type containing nascent plus-strands; R I with nascent minus strands have not been observed. M inus strand synthesis may therefore proceed via R F molecules (not shown in this schem e). W hether R I molecules have a mainly single-stranded (as shown here) or a double-stranded back-bone structure, is unknown. M inus strands carry a VPg-linked poly(u ) at their 5 ' ends; the poly(u ) (black bar) serves as template for the 3'-term inal poly(a ) of plus-strand R N A. U nlinking of VPg from plus strands destined to becom e m R N A is catalysed by a cellular enzyme (after Kitam ura et al ). Baltimore, 1981a) and the successful construction of infectious cdna clones, by Racaniello & Baltimore (19816) that was later confirmed by Semler et al. (1984) and Omata et a l. (1984). The initial low specific infectivity of the cdna clones in mammalian cells (1 p.f.u. jug-1 DNA) was increased to greater than 1000 p.f.u. ^ig 1 of DNA when replication signals of SV40 were cloned into the plasmids containing the polio cdna and the constructs were transfected into T-Ag expressing cos cells

6 256 E. Wimmer and others (Semler et al. 1984). If the gene for T-Ag and its promoters are directly cloned into these plasmids together with the SY40 origin of DNA replication, the expression of the viral cdna is much more efficient in all permissive cells tested (1000 p.f.u. jwg-1 DNA; Kohara et al. 1986). The combination of the cdna clones with the appropriate regulatory signals of SV40 undoubtedly leads to a high copy number of the plasmids in the cell nuclei. Increased specific infectivity of the cdna clones, however, does not solely depend upon the copy number of the plasmid in the transfected cell (Kean et al. 1986; Kuhn et al. 1987c ), but also upon the nature of the SV40 promoters used and the orientation in which the SV40 promoters are inserted into the vectors (Kuhn et al. 1987c). The mechanism by which the poliovirus infection is initiated in cells transfected with cdna clones is entirely mysterious. Long, nonvirus-specific RNA sequences are present on either side of the virus-specific RNA transcripts and must be ignored to initiate a replicative cycle or processed away by specific RNases or random RNA degradation. The second important achievement for in vitro manipulations of the poliovirus genome was the construction of a transcription system using phage T7 RNA polymerase. For this purpose the T7 RNA polymerase promoter was engineered directly in front of the cdna sequence coding for the viral RNA. The resulting transcripts, that were found to carry only two extra G residues at the 5' end of the RNA (pppgpgpupu..., where the G residues are non-viral), have a specific infectivity of lo3 p.f.u. ^g-1 RNA, that is, within the range of virion RNA (106p.f.u. ;tig_1 RNA) (van der Werf et al. 1986). By this route virtually unlimited amounts of highly infectious poliovirus genetic material, and derivatives thereof, can be produced in a simple test tube experiment. Four strategies have been followed to modify poliovirus genetic information in vitro. The first involved the expression of segments of poliovirus cdna clones in suitable cells (usually Escherichia coli) to study the function(s) of the expressed viral proteins. This approach was successfully applied in investigations of the poliovirusencoded proteinases (Hanecak et al. 1984; Toyoda et al. 1986a ; Ivanoff et al. 1986) and in the identification of a neutralization antigenic site of poliovirus (van der Werf et al. 1984). Most recently, such gene segments have been transcribed with phage T7 or SP6 RNA polymerases, and the transcripts have been used as mrna in in vitro protein synthesis (Nicklin et al. 1987; Ypma-Wong & Semler, 1987). Second, allele replacements between wild type and mutant strains of poliovirus have been performed to assess the effect of multiple or single known mutations on viral replication and pathogenicity (Kohara et al. 1985; Omata et al. 1985, 1986; Nomoto et al. 1986; Semler et al. 1986a; La Monica et al. 1986; Westrop et al. 1986; Pincus & Wimmer, 1986). For example, by this route it was confirmed that the mutations in 2C are sufficient to confer guanidine resistance of poliovirus replication (Pincus & Wimmer, 1986; see below). The third strategy evolves the generation of mutations over the entire infectious cdna clone by linker insertions or deletions (usually at restriction sites). The mutated clones are then screened for new phenotypes that are subsequently correlated with specific genes or genetic elements of the viral RNA

7 Molecular events leading to picornavirus genome replication 257 (Bernstein et al. 1985, 1986; Sarnow et al. 1986). In the fourth and final strategy, a specific segment of the viral genome can be selected and earmarked for saturation mutagenesis in vitro. This strategy has been followed with a segment coding for VPg: suitable restriction sites were genetically engineered into the genome located just outside the VPg coding sequence. Multiple mutations in VPg can be generated by chemical synthesis of derivatives of the entire gene segment (Kuhn et al. 1987a,b). The replicative cycle: an overview The replicative cycle of poliovirus is shown schematically in Fig. 2 (Kitamura et al ). The first step is the attachment of the virion to an as yet unidentified cellular receptor that presumably mediates the particle s uptake via endosomes (Madshus et al. 1984). The molecular cloning of the host cell receptor for poliovirus, encoded only in the genome of primates, is in progress (Mendelsohn et al. 1986). The receptor proteins for other picornaviruses, such as Coxsackie B3 and human rhinovirus 14, are also currently being characterized (Crowell et al. 1985; Tomassini & Colonno, 1986) but, so far, a comparison between the receptor molecules has not been possible since sufficient information is not available as yet. The processes of uncoating and penetration of the RNA into the cell are as yet obscure. The incoming virion RNA is thought to be immediately processed such that its VPg is cleaved from the 5' end (Ambros et al. 1978; Dorner et al. 1981). Unlinking of the VPg is catalysed by a host-cellular enzyme (Ambros et al. 1978). The pu-terminated viral RNA then functions as mrna in the synthesis of the polyprotein. The steps that follow include proteolytic processing, RNA replication and morphogenesis, of which we will discuss protein synthesis and replication in some detail. It should be noted that purified VPg-containing virion RNA from which the VPg has been partially degraded with proteases, as well as RNA totally devoid of covalently linked amino acids (mrna or the above-mentioned T7 transcripts) are infectious. Thus, there is no requirement for virion-associated enzymes for infectivity and, indeed, no proteins other than the capsid proteins and VPg are usually found associated with infectious particles. RNA replication occurs in a membranous complex, but how template RNA and replication proteins are sequestered in this complex is unknown. Several authors reported that poliovirus protein synthesis is also a membrane-associated process (Koch & Koch, 1985). Single-stranded RNA (ssrna) and double-stranded RNA (the replicative form, RF) are released from the replication complex. The ssrna may then participate in RNA replication, be encapsidated during morphogenesis, or function as mrna whereupon it is destined to lose its 5'-terminal VPg. Poliovirus mrna isolated from polyribosomes of infected cells has never been found to be VPg-linked; virion RNA, on the other hand, is always VPg-linked. The significance of this observation is not known, but it could be interpreted to mean that (1) VPg prevents the proper binding of ribosomes to the mrna, and must therefore be removed prior to translation, and/or (2) VPg is part of an encapsidation signal. If the latter were indeed the case,

8 258 E. Wimmer and others capsid proteins would ignore viral mrna (which has the same nucleotide sequence as virion RNA) and thus would not interfere with translation (Nomoto et al ). The entire replicative cycle of poliovirus is cytoplasmic. It occurs efficiently in enucleated cells, even of non-primate origin (Detjen et al. 1978, and references therein). Proteolytic processing The polyprotein, the only known product of translation of poliovirus mrna, is a fascinating entity. It harbours a wealth of very different functions already known to us, yet the biological significance of many of its features remains obscure (e.g. polypeptide 2B and even polypeptide 2C). During the last few years it has become evident that the polyprotein is a very active molecule that can cleave itself into different domains without the help of cellular components (Toyoda et al. 1986a). It does this so efficiently that, under normal circumstances, the polyprotein cannot be observed at all; before its synthesis is completed, the PI region has already been severed from the growing polypeptide strand by an intramolecular cleavage event. It is not known, and even unlikely, that the completed, intact polyprotein would fold properly to induce its own cleavage (see discussions by Toyoda et al. 1986b ; Nicklin et al. 1986). Polyprotein synthesized at supraoptimal temperature (43 C), for example, is not processed when the temperature is shifted down to 37 C (Baltimore, 1971). Certain portions of the polyprotein must therefore fold properly so that it can serve as its own substrate. In addition, its individual components, once released from the polypeptide chain, must achieve pjroper conformations such that they can perform a variety of different structural and enzymatic functions. It is therefore not surprising that the polyprotein is quite sensitive to perturbation for, by example, site-directed mutagenesis, initiated by naive, if enthusiastic, investigators. Indeed, in vitro manipulation of cdna clones with the aim of deciphering functions has yielded a surprisingly sparse harvest; most often, the mutation introduced proved lethal to the virus for whatever reason. Nevertheless, a number of publications are beginning to emerge in which the in vitro manipulated poliovirus genome has generated novel biochemical phenotypes that may shed light on some fundamental questions of the molecular biology of poliovirus (see section In vitro manipulations of the poliovirus genome ). When the newly synthesized polypeptides in poliovirus-infected HeLa cells were analysed by SD S PAGE, a large number of virus-specific proteins were found that taken together exceeded the coding capacity of the viral genome (Maizel & Summers, 1968). This paradox was solved when it was observed that many of the intracellular polypeptides were precursors to other virus-specific polypeptides. The precursorproduct relationship among the poliovirus proteins was proven by a variety of experiments, such as pulse-chase experiments (Summers & Maizel, 1968; Holland & Kiehn, 1968), the use of amino acid analogues (Jacobson & Baltimore, 1968) and tryptic peptide analyses (Rueckert et al. 1978; and references therein). The genetic order of the peptides was quite accurately mapped by studies with pactamycin (an inhibitor of initiation of protein synthesis) and by pulse-chase and salt-shock

9 Molecular events leading to picomavirus genome replication 259 experiments (reviewed by Koch & Koch, 1985). Finally, the precise map positions of viral polypeptides, as shown in Fig. 1, was established by sequence analyses of both the genome RNA and of the virus-specific proteins for which it codes (see above). This yielded the amino acid pairs that appear to function as signals in proteolytic processing and led to the identification of the enzymes responsible for cleavage. Nicklin ei a/. (1986) and Toyoda et al. (19866) have extensively reviewed this topic; here only the highlights of proteolytic processing will be discussed. Most cleavages of the polyprotein (Fig. 1) occur between Q-G pairs (closed triangles), two cleavages occur between Y-G pairs (open triangles) and one cleavage occurs between a N-S pair (open diamond). Earlier studies had produced conflicting results as to the nature of the proteinase(s) involved in processing of the poliovirus polyprotein, and as late as 1985 it was thought that host cellular activities participated in these events. It is now certain that the Q-G cleavages are carried out by polypeptide 3Cpro (Hanecakei al. 1982, 1984) and the Y-G cleavages by polypeptide 2Apro (Toyoda et al. 1986a). The N-S cleavage which occurs only in the last step of morphogenesis (when procapsid and vrna unite) may be autocatalytic as was proposed recently by Rossmann et al. (1985) on the basis of the crystal structures of human rhinovirus 14 and poliovirus (see also Arnold et al. 1987). It should be noted that elegant studies on proteolytic processing of the EMC polyprotein had earlier suggested that a virus-encoded protein functions as proteinase (Pelham, 1978); this protein was subsequently identified to be 3Cpro by Palmenberg et al. (1979) and Gorbalenya et al. (1979). The cleavage specificity of 3Cpro of EMC has recently been found to be similar to that of polio 3Cpro (Parks et al. 1986). Other specific considerations regarding the processing events are as follows (Nicklin et al. 1986; Toyoda et al ): (1) The polyprotein is not observed to occur in infected cells because the Y-G cleavage between PI and P2 by 2Apro occurs in statu nascendi, immediately after 2A has been synthesized. This event severs the capsid proteins (PI) from the nonstructural proteins. (2) 2Apro and 3Apro can act in cis as well as in trans (Nicklin et al. 1987, and references therein). Proteinase 3C is a sulphydryl proteinase (Pelham, 1978; Gorbalenya & Svitkin, 1983). The nature of its active site has been predicted by comparisons of the amino acid sequences of several picornaviral 3Cpro (Argos et al. 1984), and this prediction received support recently by site-directed mutagenesis of a 3Cpro gene segment of poliovirus expressed in E. coli (Ivanoff et al. 1986). From comparisons of amino acid sequences of the picornavirus polypeptides and through studies with specific inhibitors, we suggest that 2Apro is also likely to be a sulphydryl proteinase (Toyoda et al. 1986; Nicklin et al. 1987; B. Rosenwirth, personal communication). (3) All picornaviruses studied produce 3Cpro, but an active 2Apro appears to be a product of the polyprotein of entero- and rhinoviruses only. Cardioviruses and aphthoviruses have a leader polypeptide ( L ) preceding the PI region. In FMDV the L polypeptide is a proteinase that appears to be capable of self

10 260 E. Wimmer and others cleavage from PI (Strebel & Beck, 1986). A polypeptide corresponding to 2A in FM DV is absent altogether. In EMC, the L polypeptide has no known proteolytic activity, and the function of 2A is as yet uncertain (Parks et al. 1986). It has thus become clear that the mechanism of polyprotein processing of entero- and rhinoviruses v. cardio- and aphthoviruses is quite different in regard to the separation of PI from P2. (4) The proteinases cleave with high specificity: poliovirus 3Cpro does not cleave precursor proteins of encephalomyocarditis virus and vice versa (Nicklin & Palmenberg, unpublished results). This specificity of 3Cpro is remarkable in view of the fact that the cleavage signal recognized by 3Cpro of both poliovirus and EMC is mainly Q-G. But 3Cpro of poliovirus does not even cleave all 13 Q-G present in its own polyprotein. Instead, it recognizes only 8-9 Q-G sites by an unknown mechanism. Surprisingly, those Q-G pairs that are cleaved have surrounding amino acid sequences that differ from each other, although an additional determinant of recognition may be the amino acid in position -4 relative to the Q-G cleavage site; in poliovirus this amino acid is most often an alanine residue (Nicklin et al. 1986; Toyoda et al ). In view of these observations, it is unlikely that the selection of the proper Q-G sites is based solely upon accessibility. Instead, it is possible that 3Cpro requires structurally flexible contexts surrounding the active Q-G sites as has been proposed recently by Arnold et al. (1987). The high specificity of 3Cpro may be an explanation for the observation that no cellular proteins have been identified that are cleaved by 3Cpro; the same may be true for 2Apro (Korant, 1980; Lloyd et al. 1985, 1986; Lee et al. 1985). Also, no oligopeptides (n = 10) have been found that can serve as substrate for 3Cpro although some have been synthesized (with Q-G in the centre) that correspond precisely to poliovirus sequences known to be cleaved in the context of the polyprotein (B. Rosenwirth, personal communication; Nicklin & Wimmer, unpublished results). Poliovirus is the only virus whose 3Cpro cleavage sites are exclusively Q-G pairs. In other picornaviruses the 3Cpro cleavage sites can vary considerably but are generally Glx-Gly, where the glycine can be replaced by S, T, A, V or M residues (Q-G, -S, -T, -A, -V, -M; and E-G, -S) (Nicklin et al. 1986; Palmenberg, 1987). It should be mentioned that poliovirus rapidly and effectively turns off host cell protein synthesis (Ehrenfeld, 1982), and this event is accompanied by the cleavage of a large polypeptide (termed p220 corresponding to its molecular weight of 220x103Mr) from the cell s cap binding complex. Neither 2Apr0 nor 3Cpr cleave p220 directly, but based on genetic and biochemical experiments, polypeptide 2A appears to activate the degradation of p220 (Bernstein et al. 1985; Krausslich et al. 1987). Quite fittingly, infection with EMC (a virus whose 2A is not homologous in function to polio s 2A) does not induce the cleavage of this p220 cellular polypeptide (Mosenkis et al. 1985). (5) The Y-G specific cleavage in polypeptide 3CD that yields 3C' and 3D' (Fig. 1) may be fortuitous and of no biological significance other than lowering the yield of 3Cpro and 3Dpo1. Site-directed mutation of the tyrosine residue of this Y-G site in 3CD to a phenylalanine residue (F-G) did not abolish the infectivity of the

11 Molecular events leading to picomavirus genome replication 261 altered genome and, surprisingly, 3C' and 3D' were still produced (Lee & Wimmer, 1987). A change of the cleavage site in 3CD from TY-G to AY-G, however, completely abolished the production of 3C' and 3D' without detectable effects on virus growth (Lee & Wimmer, 1987). (6) Polypeptide 3AB, a membrane-associated protein (Semler et al. 1982; Takegami et al ), is relatively stable in pulse-chase experiments of infected cells although it contains an active Q-G site. It has been speculated that 3AB is processed only to 3A and VPg after a tyrosine residue in the 3B portion of 3AB has been uridylylated (Takegami et al. 1983a; Takeda et al. 1986; Kuhn & Wimmer, 1986). This hypothesis will be discussed in detail below. RNA replication Poliovirus RNA replication proceeds via a general pathway observed for all lytic plus-strand RNA viruses: the virion RNA is transcribed into complementary RNA strands that, in turn, function as template for the synthesis of an excess of virion RNA strands. For poliovirus, the intermediates in this reaction have been described many years ago: the replicative intermediate (R I), a partially single-stranded, partially double-stranded structure, and the replicative form (R F), a doublestranded structure whose poly(a) is longer than the poly(u) of minus strands and thus protrudes from the heteroduplex (for recent reviews, see Kuhn & Wimmer, 1986; Rueckert, 1985; Koch & Koch, 1985; Wimmer, 1982). Moreover, the virusencoded RNA polymerase involved in RNA replication has been identified (polypeptide 3Dpo1; see Fig. 1) and its properties have been characterized in great detail (Flanegan & Baltimore, 1977, 1979). In spite of this, no in vitro system has been developed that can faithfully replicate exogenously added poliovirus RNA. The biochemical analyses of virus-specific RNA structures found in infected cells strongly suggested that poliovirus RNA synthesis may proceed in steps unique among replication schemes of RNA viruses. First, the 3'-terminal poly(a) of plus strands is genetically encoded and not the product of end addition (Yogo & Wimmer, 1975; Dorsch-Hasler et al. 1975). Second, the 5'-terminal nucleotide of plus and minus strands was found to be a pyrimidine (U) and not a purine, the preferred base for enzymes that initiate RNA synthesis de novo (note that the 5' end of minus strands is poly(u)). Third, newly synthesized RNAs were all VPg-linked, even the nascent strands of R I, and no trace of pppn-termini could be found (reviewed by Wimmer, 1982). These observations were interpreted to mean that VPg is involved in the initiation of RNA synthesis (Wimmer, 1982). This notion was strongly supported by the finding that the poliovirus-specific RNA polymerase 3Dpo1 is a primer-dependent enzyme: it was considered possible that uridylylated VPg could serve as a primer for 3Dpo1 similar to the deoxycytidylylated terminal protein of adenovirus that serves as primer for the adenovirus-specific DNA polymerase (reviewed by Wimmer, 1982). 3Dpo1 is most likely the only RNA polymerase poliovirus produces. It is template and primer-dependent, properties that resemble all template-dependent DNA polymerases. Thus, the enzyme can conveniently be assayed using polyadenylylated

12 262 E. Wimmer and others RNA plus oligo(u) as primer (Flanegan & Baltimore, 1977). Another virus-encoded polypeptide thought to be involved in RNA replication is 2C (see Fig. 1) by virtue of the fact that mutations to guanidine resistance that lead to an altered phenotype of poliovirus RNA synthesis map in 2C (Pincus & Wimmer, 1986; Pincus ei al ; see also reviews by Caliguiri & Tamm, 1973, and by Pincus et al. 1987a). In contrast to 3D po1, 2C has not been purified in any active form as yet owing to the fact that no specific function of this protein, and hence no assay for its function, is known. It should be stressed that in vivo all poliovirus-specific RNA synthesis appears to proceed in a membranous environment (Caliguiri & Tamm, 1970a,b; Takegami et al. 1983a,6, and references therein). Indeed, all virus-encoded polypeptides can be found associated with a membrane-bound replication complex, and, with the exception of 3Cpro, none of the polypeptides can be removed with high salt (2 m NaCl) or 4 M urea. As we shall see, the models for RNA synthesis and initiation either take involvement of membranes into account or ignore them for the time being. The prototype in vitro replication system of an RNA virus was developed many years before by Spiegelman and his colleagues (Spiegelman & Hayashi, 1963). Working with the bacteriophage Q(3, they demonstrated that incubation of the phage RNA along with a purified viral protein (termed replicase ) and appropriate nucleoside triphosphates resulted in the de novo synthesis of authentic viral RNA. It was subsequently discovered, however, that the purified phage replicase was a complex of a phage-encoded RNA polymerase along with three host-encoded proteins that normally functioned in protein translation (Kamen, 1975). This novel concept in viral RNA replication profoundly influenced all subsequent investigations in that most studies were aimed at finding not only viral polymerases but host factors that may or may not be involved in cellular RNA metabolism. An in vitro replication system, as its name implies, is strictly defined as a system which has the ability to synthesize replicas of the input template RNA. That is, plus strands are copied into minus strands which in turn serve as templates for the synthesis of new plus strands. To date, there is no in vitro replication system of picornavirus RNA which meets this criterion. To understand this, it is necessary to examine the strategies that have been used to study poliovirus replication. The strategies can be clearly divided into two general approaches: the first utilizes purified polypeptides and an exogenous source of RNA, while the second utilizes a crude membrane mixture from infected cells and is dependent on endogenous RNA. R econstituted systems fo r poliovirus RNA synthesis Following the purification and characterization of 3Dpo1 by Flanegan and his colleagues (Flanegan & Baltimore, 1977, 1979; Van Dykeei al. 1982; and references therein), numerous experiments have been performed to reconstitute an RNA replication system. Originally, template RNA (poliovirion RNA) and 3Dp()1 needed the addition of oligo(u) for transcription to initiate. Dasgupta et al. (1980) then implicated a host factor ( H F ) ( molecular weight) in poliovirus RNA replication that rendered RNA synthesis independent of added primer such as oligo(u). In the presence of HF the product RNA is supposedly full-length, that is,

13 Molecular events leading to picornavirus genome replication 263 the product RNA represents complete, non-covalently linked complementary strands (Morrow et al. 1985). Indeed, if negative-stranded RNA (prepared from cdna clones with the aid of SP6 polymerase) is used as template for 3Dpo1 and HF, infectious poliovirus RNA (containing non-viral sequences at 5' and 3' termini) can be synthesized (Kaplan et al. 1985). It was originally thought that RNA transcribed with 3Dpo1 and HF was linked to a protein containing VPg sequences and that the reaction was thus indicative of protein-primed initiation (Baron & Baltimore, 1982a; Morrow et al. 1983, 1985). This conclusion, however, has been disputed recently by several investigators (Andrews & Baltimore, 1986a; Young et al. 1986) who found that the majority if not all of the VPg-linked product RNA originates from the covalent linking of poliovirion template RNA to the transcripts. Currently, there are opposing views as to what host factor is and what role it plays in poliovirus replication. None of the reconstituted systems have so far produced authentic poliovirion RNA, that is, polyadenylylated RNA linked to VPg. The major obstacle is the inability of these soluble systems to initiate RNA synthesis with concomitant VPglinkage. The product RNA is usually very heterogeneous in chain length and, dependent upon what HF preparation is used, the product RNA may be covalently linked ( end-linked ) to the template RNA. This latter observation has led to the formulation of the hairpin model of poliovirus RNA replication which will be discussed below. The hairpin m odel Flanegan and his colleagues reported that in the in vitro transcription of poliovirion RNA, using purified 3Dpo1 and HF, molecules were produced in which template and product RNA is covalently linked at one end (Young et al. 1985), an observation corroborated by Hey et al. (1986). In contrast, when oligo(u) was added to this transcription mixture, none of the products were end-linked since initiation occurred on the oligonucleotide primer. It was concluded that, in the absence of oligo(u), HF serves to modify the 3' end of the template RNA topographically such that its 3' terminal nucleoside can function as a primer (Fig. 3). Flanegan and his colleagues (personal communication) argue that the formation of end-linked products is suppressed if 3Dpo1 is used that was purified by affinity chromatography on poly(u)-sepharose and possibly containes trace amounts of oligo(u). Such poly(u)-sepharose purified 3Dpo1 has been used by Dasgupta and his colleagues (Morrow et al. 1985) and, indeed, these investigators do not find end-linked RF molecules in their reaction products. The hairpin model of initiation received support by Andrews et al. (1985) and Andrews & Baltimore (19866) who reported that HF is a terminal uridylyl transferase (T U T ). Accordingly, these authors suggested that the 3' end of the template is uridylylated. The product then supposedly forms a snapback structure whose 3' end serves as primer for 3Dpo1.

14 E. Wimmer and others 264 The addition of U residues to 3' termini of RNA was found to occur most efficiently on poly (A) tails. This fits well with the existence of the 3 '-terminal poly(a) of plus strands, but makes it difficult to explain initiation of plus strands at the 3' end of minus strands (see below). The self-priming hairpin model has recently suffered from new data that strongly suggests the involvement of endonuclease(s) in the in vitro reactions described above leading to random degradation of template with the fortuitous formation of snapback structures (Lubinski et a l. 1986; Hey et al. 1987). Briefly, it was suggested that the protein preparations contain endonuclease activities, not previously recognized, that nick the RNA and produce truncated template molecules with hairpins at their 3' ends. These in turn can then function to prime 3Dpo1. Hey et al. (1987) analysed end-linked RNA molecules synthesized by transcription of plus-stranded virion RNA and found that none of the products contain homopolymeric regions (poly(a)poly(u)) as would have been predicted if snap-back structures exist shown in Fig. 3. The data by Hey et al. (1987) do explain why reactions that lead to end-linked TEM PLATE RNA AS PRIM ER 5' Poliovirion RNA ^ (A )75 VPg* + Pol Host Factor Initiation and Elongation ZÜ -w n m m x jc + V Pg# #\JUlUJUUUUJJUUL ju oju u u u u u lu ü ^D Cleavage and Attachment M M II Fig. 3. Self-prim ing by tem plate RN A. A host factor (either a kinase or a terminal uridylate transferase) produces a hairpin at the 3 ' end of the template RN A that serves as a prim er for 3 D po1. A fter some elongation has taken place, VPg (or its precursor) will cleave the hairpin thereby attaching itself to the 5 ' end of the nascent RN A strand. If nicking does not occur, an end-linked R F m olecular is form ed; as shown here, the hom opolymeric segments of the R F would be covalently bound. We have named those structures hom o-linked R F. Hairpin-mediated initiation at the 3 ' end of m inus-stranded tem plate would yield hetero-linked R F molecules. (T h is drawing was kindly provided by D r J. B. Flanegan.)

15 Molecular events leading to picomavirus genome replication 265 structures give products that are extremely heterogeneous in length and end-linked poliovirus RF (twice the length of genome RNA), if it exists at all, is only a minute fraction of the total. In addition to the nicking mechanism, Lubinski et al. (1986) have added yet another mechanism that may function in the formation of end-linked products. Their hypothesis states that snapback structures occur in certain full-length product RNA (a phenomenon seen in reverse transcription of RNA), but not in template RNA. Taken altogether, the transcription experiments with purified or semi-purified polypeptides in a completely soluble phase have yielded rather conflicting results that do not give us much insight into how initiation of poliovirus RNA synthesis occurs in vivo. Moreover, even if end-linked molecules are intermediates in poliovirus replication, the mechanism by which these structures are cleaved to yield minus and plus strands containing VPg at their 5' termini remains unknown. Viral RNA synthesis on endogeneous m em brane complexes The separation of disrupted poliovirus-infected HeLa cells by differential centrifugation into nuclei, mitochondria, membranous material and cytoplasmic supernatant, has revealed that all of the detectable viral RNA synthesis occurs on membranous structures. This fact has been known for many years (Girard & Baltimore, 1967; Caliguiri & Tamm, 1971a,b, and references therein), but the precise nature of the complex producing RNA remains a mystery. Only very recently has it been possible to analyse initiation and elongation in a membranous replication complex (also termed crude replication complex, CRC). The properties of CRC, reviewed recently by Kuhn & Wimmer (1986) and Semler et al. (19866) may be summarized as follows: (1) The CRC synthesizes RI, RF and ssrna. If the infected cell extract is separated by centrifugation in a step-wise sucrose gradient, most of the RNA synthesis is found to band with the smooth membrane (Caliguiri & Tamm, 1971a,b). Upon disruption of the membranes with 0-5 % NP-40, however, only RF is produced (Girard, 1969; McDonnel & Levintow, 1979; Etchison & Ehrenfeld, 1981; Takeda et al. 1986). It is unknown why the release of ssrna is abolished through the addition of non-ionic detergent. Most likely, the detergent interferes with proper initiation of RNA synthesis. (2) The CRC is capable of synthesizing authentic VPg-linked viral RNA. Apparently, CRC can initiate as well as elongate by the same mechanisms functioning in the infected cell (Takeda et al. 1986). (3) The CRC can synthesize VPg-pU and VPg-pUpU if incubated only with (a- 32P)U TP (Takegami et al. 1983a,6; Takeda et al. 1986). This has been observed not only with poliovirus-specific CRC but also with EMC-specific CRC (Vartapetian et a l. 1984). Synthesis of VPg-pU is template dependent (Takeda et al. 1987). (4) The CRC contains all known viral proteins, and these are tightly complexed together (e.g. none of the virus-specific proteins, with the possible exception of 3Cpro, can be removed by washing with 2m NaCl, etc.) (Takegami et al. 1983a,b\ Takeda et al. 1986; Tershak, 1984).

16 266 E. Wimmer and other's (5) Pulse-chase analyses using CRC have supported a model in which VPg-pU can function as primer in the elongation reaction (Takeda et al. 1986). If VPg is directly involved in the priming of RNA synthesis, one would have expected to find some of it in infected cells. In fact, no free VPg was found in cells until Crawford & Baltimore (1983) observed that the detection of unbound VPg and uridylylated derivatives thereof is complicated by their physical properties. Takegami et al. (1983a,6), on the other hand, argued that a precursor of VPg, most likely 3AB (Semler et al. 1982; Baron & Baltimore, 19826; Takegami et al. 1983a), may participate in the initiation reaction. Polypeptide 3AB (see Fig. 1) contains a Q-G cleavage signal. Since proteins larger than VPg are never observed to be linked to the nascent RNA strands, it was assumed that the precursor is rapidly cleaved by the proteinase 3Cpro to yield VPg-RNA. 3AB is a membrane-bound polypeptide (Semler et al. 1982) and thus well suited to serve as donor for VPg in the membraneassociated machinery of poliovirus RNA replication. The model depicting the events leading to VPg-primed RNA synthesis is shown in Fig. 4 (Takegami et al. 1983a,6; Vartapetian et al. 1984). 3AB is uridylylated and subsequently cleaved by 3Cpr0; at the same time, the uridylyl-vpg functions as primer for 3Dpo1. Whether or not any host factor is involved in this process remains to be seen. Also, the role of polypeptide 2C in the events of initiation, if any, is unclear. One might speculate that 2C functions as helicase for the release of the nascent strands. Apart from studies of virus-specific proteins such as 3AB in CRC, this model is based upon experiments of RNA synthesis in the same system. As mentioned above, VPg-pU and VPg-pUpU can be synthesized in vitro (Takegami et al. 1983b), and preformed VPg-pU can be chased into VPg-pUpU (Takeda et al. 1986; Toyodaei al. 1987). Moreover, kinetic evidence suggests that the preformed VPg-pUpU can be chased into authentic elongation products as assayed by immunoprecipitation of the RNase T1 generated 5'-terminal oligonucleotide, VPg-pUUAAAACAGp. Indeed, CRC can produce VPg-linked, full-length RNA (Takeda et al. 1986). All of these reactions are highly sensitive to the addition of non-ionic detergents such as NP40, an observation suggesting that an intact membranous environment must be maintained to uridylylate VPg and to release single-stranded RNA from CRC (Takegami et al ; Takeda et al. 1986). The possibility cannot be excluded, however, that the uridylylation itself can proceed in the absence of membrane but that the enzyme catalysing this reaction is sensitive to detergent. A major question remains to be solved: Is the activity that catalyses the uridylylation of the Tyr residue of VPg virusencoded or of cellular origin? Using mutants mapping in the 3' end of the viral RNA, Toyoda et al. (1987) have obtained evidence that implicates 3Dpo1 in the formation of VPg-pU and VPg-pUpU. Such a conclusion is not unreasonable since the adenovirus-specific DNA polymerase has also been shown to link dcmp on to the adenovirus terminal protein. A large panel of mutants in VPg are currently being constructed (using recombinant DNA techniques) by Kuhn, Tada, Dunn & Wimmer (see Kuhn et al. 1987a,6) with the objective of finding a phenotype that would shed light onto the events shown in Fig. 4 (see section Tn vitro manipulation of the poliovirus genome ).

17 Molecular events leading to picornavirus genome replication 267 F ig. 4. M odel of V Pg-pU pu primed initiation of plus-strand RN A synthesis. Polypep tide 3A B, the m em brane-bound poliovirus protein whose CO O H term inus (wavy line) is V Pg, is thought to be the precursor to the genome-linked protein. Polypeptide 3C pro is the virus-encoded proteinase (responsible for the cleavage between 3A and 3 B ). 3C P' may cleave only after VPg has been uridylylated. Polypeptide 3 D po1 is the primer-dependent R N A polymerase and 2C an auxiliary viral protein, mutations of which (to guanidine resistance) lead to an altered phenotype of RN A synthesis in viv o. T h e possibility of the involvement of a host factor is indicated although there is no evidence to support such factor in the m em brane bound replication com plex. T h is model can account for the initiation of both plus and minus strand RN A. (M odified from Takegam i et al. 1983/?.) It has become evident already that the NH 2 -terminal sequence of VPg is highly sensitive to alteration. For example, if the VPg sequence H2N -G A Y T G L... (where the Y is the amino acid linked to the RNA) is changed to H2N -G A YYG L... or H2 N -G A T Y G L..., the cdna clones are rendered non-infectious (Kuhn el al. 1987a). A change at amino acid No. 6, on the other hand, from L >M (H2N -G A Y TG M...) was found to yield infectious virus. This result is also of practical value, since VPg can now be labelled with i;,s-methionine which makes its detection easier. So far no ts phenotype mapping in VPg has been found. The problem o f tem plate recognition an d o f initiation at different termini Whatever the mechanism, the poliovirus RNA replication machinery must be capable of initiating RNA synthesis at two very different 3' termini: the homopoly meric segment (poly(a)) of plus strands and the heteropolymeric terminus (-CAGUUUUAA qh ) of minus strands. The only feature common to these ends are

18 268 E. Wimmer and others two terminal adenosine residues. Clearly, VPg-pUpU, a structure that can be synthesized in a replication complex in vitro (Takegami et al ; Takeda et al. 1986) and that has been found to occur in infected cells (Crawford & Baltimore, 1983) could recognize and bind to such a 3 '-terminal dinucleotide prior to initiation. Similarly, the termini of both plus- or minus-strands could be oligouridylylated by T U T to form snapback structures. If the latter were the case, one would predict that the termini of minus strands isolated from infected cells would carry extra U residues that would remain there after the proposed cutting by VPg. When minus strands of R F were analysed, however, the 3' termini were found to be CAGUUUU(A)n, where n 2 in most molecules (Larsen et al. 1980) (to form blunt ends). A small fraction of the minus strand also carried some extra A residues (n = 3, 4, 5, 6) (Richards & Ehrenfeld, 1980). No additional U residues were found. Unfortunately, a similar study on minus strands of RI has not been performed. Two A residues at the 3' terminus of an RNA can hardly be the sole determinant of template selection since nearly all cellular mrnas are polyadenylylated. Thus, the poliovirus polypeptides must recognize sequences other than the ends of the RNA. Of course, numerous primary or secondary structures distant to the termini could serve as recognition signals. No such signals have been discovered as yet. Experiments to assay for these signals are difficult because (1) an in vitro replication system is not available and (2) deletions in the RNA of D I particles are restricted to the PI capsid region (see below). These observations, which contrast with other RNA virus replication systems (e.g. vesicular stomatitis viruses or togaviruses), do not allow an easy search for the minimum genome sequence still capable of replicating through complementation with wild type virus. The in vitro construction of D I genomes, based upon the highly efficient phage T 7 RNA polymerase transcription system (van der Werf et al. 1986) has recently been achieved, but no viable deletion mutants have been found as yet (Bradley, Wimmer, Girard & van der Werf, unpublished results). Elongation The RNA polymerase 3Dpo1 is probably the only activity produced by poliovirus capable of elongating the nascent RNA strand. It is unknown as yet whether the newly synthesized RNA is released from the template by strand displacement (in RI molecules with a double-stranded backbone) or prevented from hybridizing to the template RNA by polypeptides in the replication complex (in single-stranded RI molecules). If strand replacement is the mechanism, one would predict that an RNA specific helicase would be involved in replication also. When genomes of D I particles were sequenced, it was found by Kuge et al. (1986) that the deletions in the capsid region always occur in the PI region and in frame of the polyprotein. These authors suggested that poliovirus genomes may need their own replication proteins (originating from the P2 and P3 regions) for RNA synthesis, that is, that some P2 or P3 proteins may function only in cis. This speculation finds support in a genetic analysis of Bernstein et al. (1986) who found

19 Molecular events leading to picomavirus genome replication 269 that specific mutants in the P2 and P3 region generated in vitro by manipulations of cdna clones cannot be complemented. More data, however, must be generated to define the specificity of viral proteins for heterologous or homologous template RNA. For example, if 3Dpo1 can indeed only act in cis (that is only on the RNA from which it was translated) it is difficult to envisage the events that lead to RI structures with multiple growing strands. It-is possible that 3Dpo1 acts in cis only to produce minus strands but can function in trans to produce plus strands. CONCLUSION The study of poliovirus RNA replication, now in its 24th year and pursued by many investigators, has proven to be exceptionally challenging. As has been pointed out before (Kuhn & Wimmer, 1986), the search for the ultimate in vitro replication system followed the classical strategy by Spiegelman and his contemporaries and required the purification of cellular and viral polypeptides and their subsequent reconstitution to a functional complex. So far, this strategy has achieved little. The authors of this review are convinced that an essential component of the in vitro replication complex is a hydrophobic environment provided by cellular membranes where polypeptides can be anchored and processed to participate in RNA replication. Indeed, only a membranous replication complex has yielded authentic virion RNA so far (Takeda et al. 1986). Clearly, the membrane may be replaceable by an artificial hydrophobic environment, but attempts to achieve this have so far not been successful. The in vitro replication of picornavirus may thus involve a complex set of components: membrane, proteinase, precursor for VPg, polymerase 3D. The participation of 2C is likely, that of one or more host cellular polypeptides may be envisioned. However, neither T U T nor the proteinase kinase are good candidates to participate in RNA synthesis. The extraordinary stability of the membranous replication complex to treatment with high salt, proteinase or nuclease makes the CRC an interesting entity. Is the CRC a vesicle with pores large enough to let nucleoside triphosphates in, but to keep degradative enzymes out? Does the template strand exist in the form of a ribonucleoprotein, the structure of which nobody has observed as yet? Does the strict dependence of RNA replication upon viral protein synthesis, which is known to occur in vivo, imply that newly synthesized viral polypeptides have to be constantly supplied to the CRC in order to maintain its function? Or can only specific viral polypeptides, possibly short-lived precursors, function in RNA replication? We thank our colleagues, particularly Aniko Paul, Michael Murray, Chong-Kyo Lee, Sung Key Jang and Hiroomi Tada for many stimulating discussions. We are grateful to Lynn Zawacki for the preparation of the manuscript. Work described here was supported by Public Health Service grants AI and CA from the National Institutes of Health. M.N. is a fellow of the Damon Runyon-Walter W inchell Cancer Fund.

20 270 E. Wimmer and others REFERENCES A d le r, C. J., E lz in g a, M. & W im m er, E. (1983). T h e genome-linked protein of picornavirus. V I II. Complete amino acid sequence of poliovirus VPg and carboxy-terminal analysis of its precursor. J. gen. Virol. 64, A m bros, V. & B a ltim o r e, D. (1978). Protein is linked to the 5' end of poliovirus R N A by a phosphodiester linkage to tyrosine. J. biol. Chem. 60, A m bros, V., P e tte r s s o n, R. F. & B a ltim o r e, D. (1978). An enzymatic activity inuninfected cells that cleaves the linkage between poliovirion RN A and the 5'-term inal protein. Cell 15, A n d re w s, N. C. & B a ltim o r e, D. (1986a). Lack of evidence for VPg priming of poliovirus RNA synthesis in the host factor-dependent in vitro replicase reaction. J. Virol. 58, A n d re w s, N. C. & B a ltim o r e, D. (19866). Purification of aterm inal uridylyl-transferase that acts as host factor in the in vitro poliovirus replicase reaction. Proc. natn. Acad. Sci. U.S.A. 83, A n d re w s, N. C., L e v in, D. & B a ltim o r e, D. (1985). Poliovirus replicase stimulation by terminal uridylyl transferase. J. biol. Chem. 260, A rg o s, P., K a m e r, G., N ic k lin, M. J. H. & W im m er, E. (1984). Similarity in gene organization and homology between proteins of animal picornaviruses and a plant comovirus suggest common ancestry of these virus families. N ucl. Acids Res. 12, A r n o ld, E., L u o, M., F r ie n d, G., R o ssm a n n, M. G., P a lm e n b e rg, A. C., P a r k s, G. D., N ic k lin, M. J. H. & W im m er, E. (1987). Implication of the picornavirus capsid structure for polyprotein processing. Proc. natn. Acad. Sci. U.S.A. 84, B a l t im o r e, D. (1971). Polio is not dead. In Perspectives in Virology, vol. 1 (ed. M. Pollard), pp New York: Academic Press. B a r o n, M. H. & B a ltim o r e, D. ( a ). Anti-VPg antibody inhibition of poliovirus replicase reaction and production of covalent complexes of VPg-related proteins and RN A. Cell 30, B a r o n, M. H. & B a ltim o r e, D. (19826). Antibodies against the chemically synthesized genomelinked protein of poliovirus react with native, virus-specific proteins. Cell 28, B e r n s t e in, H. D., S a r n o w, P. & B a l tim o r e, D. (1986). Genetic complementation among poliovirus mutants derived from an infectious cd N A clone. J'. Virol, (in press). B e r n s t e in, H. D., S o n e n b e r g, N. & B a l tim o r e, D. (1985). Poliovirus mutant that does not selectively inhibit host cell protein synthesis. Mol. Cell. Biol. 5, CALIGUIRI, L. A. & T a m m, I. (1970a). T h e role of cytoplasm ic m em branes in poliovirus biosynthesis. Virology 42, C a lig u ir i, L. A. & Tam m, I. (19706). Characterization of poliovirus-specific structures associated with cytoplasmic membranes. Virology 42, C a l ig u ir i, L. A. & T a m m, I. (1973). Guanidine and 2-(/?-hydroxybenzyI)-benzimidazole (HBB): selective inhibitors of picornavirus multiplication. In Selective Inhibitors of Viral Functions (ed. W. A. Carter), pp Cleveland, Ohio: CR C Press. C r a w fo r d, N. M. & B a ltim o r e, D. (1983). Genome-linked protein VPg of poliovirus is present as free VPg and VPg-pU pu in poliovirus-infected cells. Proc. natn. Acad. Sci. U.S.A. 80, C r o w e l l, R. L., R e a g a n, K. J., S c h u lt z, M., M a p o le s, J. E., G r u n, J. B. & L a n d a u, B. J. (1985). Cellular receptors as determinants of viral tropism. In Genetically A ltered Viruses and the E nvironm ent (ed. B. Fields, M. A. Martin & D. Kamely), pp Cold Spring Harbor Banbury Report 22. D a s g u p ta, A., Z a b e l, P. & B a ltim o r e, D. (1980). Dependence of the activity of poliovirus replicase on a host cell protein. Cell 19, D e t je n, B. M., L u c a s, J. & W im m er, E. (1978). Poliovirus single-stranded RN A and doublestranded RN A : differential infectivity in enucleate cells. J. Virol. 27, D o r n e r, A. J., D o r n e r, L. F., L a r s e n, G. R., W im m e r, E. & A n d e r s o n, C. W. (1982). Identification of the initiation site of poliovirus polyprotein synthesis. J 1. Virol. 42, D o r n e r, A. J., R o th b e r g, P. G. & W im m er, E. (1981). T he fate of VPg duringm vitro translation of poliovirus RN A. F E B S L e tt. 132,

21 Molecular events leading to picornavirus genome replication 271 D o r sc h -HAs l e r, K., Y o g o, Y. & W im m e r, E. (1975). Replication of picornaviruses. I. Evidence from in vitro RN A synthesis that poly(a) of the poliovirus genome is genetically coded. J. Virol. 16, E h r e n f e ld, E. (1982). Poliovirus-induced inhibition of host-cell protein synthesis. Cell 28, E m ini, E. A., E lz in g a, M. & W im m er, E. (1982). Carboxy-terminal analysis of poliovirus proteins: termination of poliovirus translation and location of unique poliovirus cleavage sites. J. Virol. 42, E n d e r s, J. F., W e l l e r, T. H. & R o b b in s, F. C. (1949). Cultivation of the Lansing strain of poliomyelitis virus in cultures of various human embryonic tissues. Science 109, E tc h is o n, D. & E h r e n f e ld, E. (1981). Comparison of replication complexes synthesizing poliovirus RN A. Virology 111, F l a n e g a n, J. B. & B a ltim o r e, D. (1977). Poliovirus-specific primer-dependent RNA polymerase able to copy poly(a). Proc. natn. Acad. Sei. U.S.A. 74, F la n e g a n, J. B. & B a ltim o r e, D. (1979). Poliovirus polyuridylic acid polymerase and RNA replicase have the same viral polypeptide.^. Virol. 29, F o r s s, S., S t r e b e l, K., B e c k, E. & S c h a l l e r, FI. (1984). Nucleotide sequence and genome organization of foot-and-mouth disease virus. Nucl. Acids Res. 12, G ir a r d, M. (1969). In vitro synthesis of poliovirus ribonucleic acid: role of the replicative intermediate. J. Virol. 3, 376. G ir a r d, M. & B a ltim o r e, D. (1967). T h e poliovirus replication com plex: site for synthesis of poliovirus RN A. J. molec. Biol. 24, G o r b a le n y a, A. E. & S v itk in, Y. V. (1983). Encephalomyocarditis virus protease: purification and role of the SH groups in processing of the precursor of structural proteins. Biochemistry (U.S.S.R.) 48, ' G o r b a le n y a, A., S v itk in, Y., K a z a c h k o v, Y. & A g o l, V. (1979). Encephalomyocarditis virusspecific polypeptide p22 is involved in the processing of the viral precursor polypeptides. F E B S Lett. 108, 1-5. H a n e c a k, R., S e m le r, B. L., A n d e r s o n, C. W. & W im m er, E. (1982). Proteolytic processing of poliovirus polypeptides: antibodies to polypeptide P3-7C inhibit cleavage at glutamine-glycine pairs. Proc. natn. Acad. Sei. U.S A. 79, H a n e c a k, R., S e m le r, B. L., A rig a, H., A n d e rs o n, C. W. & W immer, E. (1984). Expression of a cloned gene segment of poliovirus in E. coli: evidence for autocatalytic production of the viral proteinase. Cell 37, H ey, T. D., R ic h a r d s, O. C. & E h r e n f e ld, E. (1986). Synthesis of plus- and minus-strand RNA from poliovirion RNA template in vitro. J. Virol. 58, H ey, T. D., R ic h a r d s, O. C. & E h r e n f e ld, E. (1987). Host factor induced template modification during synthesis of poliovirus R N A ot vitro. J. Virol. 61, H o g l e, J. M., C h o w, M. & FiLMAN, D. J. (1985). Three-dimensional structure of poliovirus at 2-9 Ä resolution. Science 229, H o l la n d, J. J. & K ie h n, E. D. (1968). Specific cleavage of viral proteins as steps in the synthesis and maturation of enteroviruses. Proc. natn. Acad. Sei. U.S.A. 60, H o l la n d, J., S p in d le r, K., H o ro d y sk, F., G r a b a u, E., N ic h o l, S. & v a n d e P o l, S. (1982). Rapid evolution of RNA genomes. Science 215, H o r stm a n n, D. M., Q u in n, T. C. & R o b b in s, F. C. (1984). International Symposium on Poliomyelitis Control. R eview s o f Infectious Diseases 6, Supp. 2. I v a n o f f, L. A., T o w a t a r i, T., R a y, J., K o r a n t, B. D. & P e tte w a y, S. R. J r (1986). Expression and site-specific mutagenesis of the poliovirus 3C protease in Escherichia coli. Proc. natn. Acad. Sei. U.S.A. 83, Ja c o b s o n, M. F. & B a ltim o r e, D. (1968). Polypeptide cleavages in the formation of poliovirus proteins. Proc. natn. Acad. Sei. U.S.A. 61, K a m e n, R. I. ( ). Structure and Function o f the OB R N A Replicase in R N A Phages. In Cold Spring Harbor Monograph Series (ed. N. D. Zinder), pp K a p la n, G., L u b in sk i, J., D a s g u p ta, A. & R a c a n ie llo, V. R. (1985). In vitro synthesis of infectious poliovirion R N A. Proc. natn. Acad. Sei. U.S.A. 82,

22 I l l E. Wimmer and others K e a n, K. M., W y ch o w sk i, C., K o p e c k a, H. & G ir a r d, M. J. (1986). Highly infectious plasmids carrying poliovirus cd N A are capable of replication in transfected simian cells. J. Virol. 59, K ita m u r a, N., A d le r, C. J., R o th b e r g, P. G., M a r tin k o, J., N a th e n s o n, S. G. & W im m er, E. (1980a). T h e genome-linked protein of picornaviruses. V II. Genetic mapping of poliovirus VPg by protein and R N A sequence studies. Cell 21, K ita m u r a, N., A d le r, C. & W im m er, E. (19806). Structure and expression of the picornavirus genome. 7th A nnals o f the N ew York Academ y o f Sciences, Genetic Variation o f Viruses 354, K ita m u r a, N., S e m le r, B. L., R o th b e r g, P. G., L a r s e n, G. R., A d le r, C. J., D o r n e r, A. J., E m ini, E. A., H a n e c a k, R., L e e, J. J., v a n d e r W e r f, S., A n d e r s o n, C. W. & W im m er, E. (1981). Primary structure, gene organization and polypeptide expression of poliovirus R N A. N ature, Land. 291, K it a m u r a, N. & W im m e r, E. (1980). Sequence of '-term inal nucleotides of poliovirus RN A as determined by a modification of the dideoxynucleotide method. Proc. natn. Acad. Sei. U.S.A. 77, K o ch, F. & K och, G. (1985). The M olecular Biology o f Poliovirus. W ien: Springer-Verlag. K o h a r a, M., Ab e, S., K u g e, S., S e m l e r, B. L., K o m a t su, T., A r it a, M., I t o h, H. & N om oto, A. (1986). An infectious cd N A clone of the poliovirus Sabin strain could be used as a stable repository and inoculum for the oral polio live vaccine. Virology (in press). K o h a r a, M., O m a ta, T., K a m ed a, A., S e m l e r, B. L., I t o h, H., W im m e r, E. & N om oto, A. (1985). In vitro phenotypic markers of a poliovirus recombinant constructed from infectious cd N A clones of the neurovirulent mahoney strain and the attenuated Sabin 1 strain. J 1. Virol. 53, K o r a n t, B. D., C h o w, N. L., L iv e ly, M. O. & P o w e rs, J. C. ( ). P roteolytic events in rep lication of anim al viruses. Ann. N.Y. Acad. Sei. 343, K r ä u s s l ic h, H. G., N ic k l in, M. J. T., T o y o d a, H., E t c h iso n, D. & W im m e r, E. (1987). Poliovirus proteinase 2A induces cleavage of the eukaryotic initiation factor 4 F polypeptide p220 on translation in vitro (submitted). K u g e, S., C h is a k a, O., K o h a r a, M., Im u ra, N., M a ts u b a r a, K. & N o m o to, A. (1984). Expression of poliovirus complementary DNA coding for viral antigenic determinants in Escherichia coli.j. Biochem. 96, K u g e, S., S a ito, I. & N om oto, A. (1986). Primary structure of poliovirus defective interfering particle genomes and possible generation mechanisms of the particles, j?. molec. Biol, (in press). K u h n, R. J., T a d a, H., Y pm a-w o n g, M. F., D u n n, J. J., S e m l e r, B. L. & W im m e r, E. (1987a). Construction of a mutagenesis cartridge for the genome-linked protein (VPg) of poliovirus and characterization of viable and non-viable mutants. Unpublished. K u h n, R. J., T a d a, H., Y pm a-w o n g, M. F., S e m l e r, B. L. & W im m e r, E. (19876). Construction and characterization of mutants in the genome-linked protein (VPg) of poliovirus. Unpublished. K u h n, R. J. & W im m e r, E. (1987). T h e replication of picornaviruses. In The M olecular Biology o f Positive S tra n d R N A Viruses (ed. D. J. Rowlands, B. W. J. Mahy & M. M ayo), pp London: Academic Press. K u h n, R. J., W im m er, E. & S e m le r, B. L. (1987c). Expression of the poliovirus genome from infectious cd N A is dependent upon arrangments of eukaryotic and prokaryotic sequences in recombinant plasmids. Virology 157, L a M o n ic a, N., M e r ia n, C. & R a c a n i e l lo, V. R. (1986). Mapping of sequences required for mouse neurovirulence of poliovirus type 2 Lansing. J. Virol. 57, L a r s e n, G. R., A n d e rs o n, C. W., D o r n e r, A. J., S e m le r, B. L. & W im m er, E. (1982). Cleavage sites within the poliovirus capsid protein precursors. J. Virol. 41, L a r s e n, G. R., D o r n e r, A. J., H a r r is, T. J. R. & W im m er, E. (1980). T h e structure of poliovirus replicative form. N ucl. Acids Res. 8, L e e, C.-K. & W im m e r, E. (1987). Studies on the cleavage specificity of poliovirus proteinase 2A and evidence that 3C ' and 3D ' are not required for virus proliferation. Unpublished. L e e, K. A. W., E d e r y, I., H a n e c a k, R., W immer, E. & S o n e n b e r g, N. C. (1985). Poliovirus protease 3C (P3-7C ) does not cleave P220 of the eukaryotic m RNA cap-binding protein com plex. J. Virol. 55,

23 Molecular events leading to picornavirus genome replication 273 L e e, Y. F., N o m o to, A., D e t je n, B. M. & W immer, E. (1977). T h e genome-linked protein of picornaviruses. I. A protein covalently linked to poliovirus genome RN A. Proc. natn. Acad. Sei. U.S.A. 74, L lo y d, R. E., E tc h is o n, D. & E h r e n f e ld, E. (1985). Poliovirus protease does not mediate cleavage of the D a component of the cap binding protein complex. Proc. natn. Acad. Sei. U.S.A. 82, L l o y d, R. E., T o y o d a, H., E t c h iso n, D., W im m er, E. & E h r e n f e l d, E. (1986). Cleavage of the cap binding protein complex polypeptide P220 is not effected by the second poliovirus proteinase 2A. Virology (in press). L u b in sk i, J. M., K a p la n, G., R a c a n i e l lo, V. R. & D a s g u p ta, A. (1986). Mechanism of in vitro synthesis of covalently linked dimeric R N A molecules by the poliovirus replicase. J. Virol. 58, M a d s h u s, I. H., O ls n e s, S. & S a n d v ig, K. (1984). Mechanism of entry into the cytosol of poliovirus type 1: requirement for low ph. J. Cell Biol. 98, M a iz e l, J. V. J r & SUMMERS, D. R. (1968). Evidence for differences in size and composition of the poliovirus-specific polypeptides in infected H ela cells. Virology 36, 48. M cd o n n e l, J. P. & L e v in t o w, L. (1970). Kinetics of appearance of the products of poliovirusinduced RN A polymerase. Virology 42, 999. M e n d e l s o h n, C., J o h n so n, B., L io n e t t i, K. A., N o b is, P., W im m er, E. & R a c a n ie ll o, V. R. (1986). Transform ation of a human poliovirus receptor gene into mouse cells. Proc. natn. Acad. Sei. U.S.A. (in press). M o r r o w, C. D., G ib b o n s, G. F. & D a s g u p ta, A. (1985). T he host protein required for in vitro replication of poliovirus is a protein kinase that phosphorylates eukaryotic initiation factor-2. Cell 40, M o r r o w, C. D., N avag, M., P e t e r s o n, C., H o ck o, J. & D a sg u pt a, A. (1984). A ntibody to poliovirus genom e-linked protein (V P g) precipitates in vitro synthesized R N A attached to VPgprecursor polypep tid e(s). Virus Res. 1, M o s e n k is, J., D a n ie l s-m cq u e e n, S., J a n o v ec, S., D u n c a n, R., H e r s h e y, J. W. B., G r if o, J. A., M e r r ic k, W. C. & T h a c h, R. E. (1985). Shutoff of host translation by encephalomyocarditis virus infection does not involve cleavage of the eucaryotic initiation factor 4 F polypeptide that accompanies poliovirus infection. J. Virol. 54, N ic k l in, M. J. H., K r a u s s l ic h, H. G., T o y o d a, H., D u n n, J. J. & W im m er, E. (1987). Poliovirus polypeptide precursors: expression in vitro and processing by exogenous 3C and 2A (submitted). N ic k lin, M. J. H., T o y o d a, H., M u r r a y, M. G. & W im m er, E. (1986). Proteolytic processing in the replication of polio and related viruses. Biotechnology 4, N o m o to, A., D e t je n, B., P o z z a t ti, R. & W immer, E. (1977a). T he location of the polio genome protein in viral RNAs and its implication for RNA synthesis. N ature, Lond. 268, N o m o to, A., K ita m u r a, N., G o lin i, F. & W immer, E. (19776). T h e 5'-term inal structures of poliovirion RN A and poliovirus m RNA differ only in the genome-linked protein VPg. Proc. natn. Acad. Sei. U.S.A. 74, N o m o to, A., K o h a r a, M., K u g e, S., K a w a m u r a, N., A r ita, M., K o m a t su, T., A b e, S., SEMLER, B., WlMMER, E. & I t o h, H. (1986). Study on virulence of poliovirus type 1 using in vitro modified viruses. In Positive S tra n d R N A Viruses (ed. M. A. Brinton & R. R. Rueckert), U C L A Symposia Molecular Cellular Biology (in press). N om o to, A., O m a ta, T., T o y o d a, H., K u g e, S., H o r ie, H., K a ta o ka, Y., G e n b a, Y., N a k a n o, Y. & I m u r a, N. (1982). Complete nucleotide sequence of the attenuated poliovirus Sabin 1 strain genome. Proc. natn. Acad. Sei. U.S.A. 79, N o m o to, A. & W im m e r, E. (1987). Genetic studies of the antigenicity and the attenuation phenotype of poliovirus. In M olecular Basis of Virus Disease (ed. W. C. Russell & J. W. Almond), pp Soc. gen. M icrobiol. Sym p., vol. 40. Cambridge University Press. O m a ta, T., K o h a r a, M., A b e, S., I to h, H., K o m a t su, T., A r it a, M., S e m l e r, B. L., W im m er, E., K u g e, S., K a m e d a, A. & N om oto, A. (1985). Construction of recombinant viruses between Mahoney and Sabin strains of type 1 poliovirus and their biological characteristics. In Vaccines 85: M olecular and Chemical Basis o f Resistance to Parasitic, Bacterial and Viral Diseases (ed. R. A. Lerner, R. M. Chanock & F. Brown), pp Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory.

24 274 E. Wimmer and others O m a ta, T., K o h a r a, M., K u g e, S., K o m a ts u, T., A be, S., S e m le r, B. L., K a m e d a, A., I to h, H., A r it a, M., W im m er, E. & N o m o to, A. (1986). Genetic analysis of the attenuation phenotype of poliovirus type \. J. Virol. 58, O m a ta, T., K o h a r a, M., S a k a i, Y., K a m ed a, A., I m u r a, N. & N om oto, A. (1984). Cloned infectious complementary D N A of the poliovirus Sabin 1 genome: biochemical and biological properties of the recovered virus. Gene 32, P a l l a n s c h, M., K e w, O. M., S e m le r, B. L., O m ilia n o w sk i, D. R., A n d e r s o n, C. W., W im m er, E. &. R u e c k e r t, R. R. ( ). P rotein processing map of poliovirus. J. Virol. 49, P a l m e n b e r g, A. C. (1987). Comparative organization and genome structure in picornaviruses. In Positive Stra n d R N A Viruses (ed. M. A. Brinton & R. R. Rueckert), U C L A Symposia Cellular Biology (in press). P a lm e n b e rg, A. C., K ir b y, E. M., J a n d a, M. R., D r a k e, N. L., D u k e, G. M., P o t r a t z, K. F. & C o l l e t t, M. S. (1984). T h e nucleotide and deduced amino acid sequence of the encephalomyocarditis viral polyprotein coding region. Nucl. Acids Res. 12, P a l m e n b e r g, A. C., P a l l a n sc h, M. A. & R u e c k e r t, R. R. (1979). Protease required for processing picornaviral coat protein resides in the viral replicase gene. J'. Virol. 32, P a r k s, G. D., D u k e, G. M. & P a lm e n b e rg, A. C. (1986). Encephalomyocarditis virus 3C protease: efficient cell-free expression from clones which link viral 5' noncoding sequences to the P3 region. J. Virol. 60, P e lh a m, H. R. B. (1978). Translation of encephalomyocarditis virus RNAm vitro yields an active proteolytic processing enzyme. Eur. J. Biochem. 85, P in c u s, S. E., D iam o n d, D. C., Em ini, E. A. & W im m er, E. (1986a). Guanidine-selected mutants of poliovirus: mapping of point mutations to polypeptide 2C. J. Virol. 57, P in c u s, S. E., K u h n, R. J., Y a n g, C.-F., T o y o d a, H., T a k e d a, N. & W im m er, E. (1987a). T h e poliovirus genome: a unique RN A in structure, gene organization and replication. In M olecular Biology o f RNA: N ew Perspectives (ed. M. Inouye & B. Dudock). New York: Academic Press (in press). P in c u s, S. E., R ö h l, FI. & W im m er, E. (19866). Guanidine dependent mutants of poliovirus: identification of three classes with different growth requirements. Virology 156, P in c u s, S. E. & W im m er, E. (1986). Production of guanidine resistant and dependent poliovirus from cloned cd N A : demonstration that mutations in polypeptide 2C are directly responsible for altered guanidine sensitivity. J. Virol. 60, P o t r a t z, K. F. & C o l l e t t, M. S. (1984). T h e nucleotide and deduced amino acid sequence of the encephalomyocarditis viral polyprotein coding region. Nucl. Acids Res. 12, R a c e n i e l lo, V. R. & B a ltim o r e, D. (1981a). Molecular cloning of poliovirus cd N A and determination of the complete nucleotide sequence of the viral genome. Proc. natn. Acad. Sei. U.S.A. 78, R a c a n i e l lo, V. R. & B a ltim o r e, D. (19816). Cloned poliovirus complementary DNA is infectious in mammalian cells. Science 214, R e a n n y, D. (1984). T h e molecular evolution of viruses. In The Microbe, part I (ed. B. W. J. Mahy & J. R. Pattison), pp Cambridge: Cambridge University Press. R ic h a r d s, O. C. & E h r e n f e ld, E. (1980). Heterogeneity of the 3' end of minus-strand R N A in the poliovirus replicative form. J. Virol. 36, R o s s m a n n, M. G., A r n o l d, E., E r ic k so n, J. W., F r a n k e n b e r g e r, E. A., G r if f it h, J. P., H ec h t, H.-J., J o h n so n, J., K a m e r, G., L u o, M., M o s s e r, A. G., R u e c k e r t, R. R., S h e r r y, B. & V r ie n d, G. (1985). Structure of a human common cold virus and functional relationship to other picornaviruses. N ature, Land. 317, R o th b e r g, P. G., H a r r is, T. J. R., N o m o to, A. & W immer, E. (1978). T he genome-linked protein of picornaviruses V. 0 4-(5'-uridylyl) tyrosine is the bond between the genome-linked protein and the RN A of poliovirus. Proc. natn. Acad. Sei. U.S.A. 75, R u e c k e r t, R. R. (1985). Picornaviruses and their replication. In Virology, chapter 32 (ed. B. N. Fields et a l.), pp New York: Raven Press. R u e c k e r t, R. R., M a t t h e w s, T. J., K e w, O. M., P a l l a n sc h, M., M cl e a n, C. & O m il ia n o w s k i, D. ( ). In The M olecular Biology o f Picornaviruses (ed. R. P erez-b ercoff), pp N A T O Advanced Study Institutes Series. New York: Plenum. R u e c k e r t, R. R. & W im m er, E. (1984). Systematic nomenclature of picornavirus proteins. J. Virol. 50,

25 Molecular events leading to picomavirus genome replication 275 S a r n o w, P., B e r n s t e in, H. D. & B a l tim o r e, D. (1986). A poliovirus temperature-sensitive RNA synthesis mutant located in a noncoding region of the genome. Proc. natn. Acad. Sei. U.S.A. 83, S e m l e r, B. L., A n d e r s o n, C. W., H a n e c a k, R., D o r n e r, L. F. & W im m e r, E. (1982). A membrane-associated precursor to poliovirus VPg identified by immunoprecipitation with antibodies directed against a synthetic heptapeptide. Cell 28, S e m l e r, B. L., A n d e r s o n, C. W., K ita m u r a, N., R o t h be r g, P. G., W is h a r t, W. L. & W im m e r, E. (1981a). Poliovirus replication proteins: RNA sequence encoding P3-lb and the sites of proteolytic processing. Proc. natn. Acad. Sei. U.S.A. 75, S e m le r, B. L., D o r n e r, A. J. & W im m er, E. ( ). Production of infectious poliovirus from cloned cd N A is dram atically increased by SV 4 0 transcription and rep lication signals. N ucl. Acids Res. 12, S e m l e r, B. L., H a n e c a k, R., A n d e r s o n, C. W. & W im m e r, E. (19816). Cleavage sites in the polypeptide precursors of poliovirus protein P2-X. Virology 114, S e m l e r, B. L., J o h n so n, V. H. & T r a c y, S. (1986a). A chimeric plasmid from cd N A clones of poliovirus and Coxsackievirus produces a recombinant virus that is temperature sensitive. Proc. natn. Acad. Sei. U.S.A. 83, S e m l e r, B. L., K u h n, R. J. & W im m er, E. (19866). Replication of the poliovirus genome. In iw A Genetics (ed. E. Domingo, P. Ahlquist & J. J. Holland). CRC Press (in press). SPIEGELMAN, S. & H a y a sh i, M. ( ). T h e present status of the transfer of genetic information and its control. Cold Spring Harbor Q uant. Biol. 28, 161. S t a n w a y, G., H u g h e s, P. Y., M o u n t fo r d, R. C., R e e v e, P., M in o r, P. D., S c h il d, G. G. & A l m o n d, J. W. (1984). Comparison of the complete nucleotide sequences of the genomes of the neurovirulent poliovirus P3/Leon/37 and its attenuated Sabin vaccine derivative P3/Leon 12a,b. Proc. natn. Acad. Sei. U.S.A. 81, S t r e b e l, K. & B e c k, E. (1986). A second protease of the foot-and-mouth disease virus. J. Virol. 58, S u m m e r s, D. F. & M a iz e l, J. V. (1968). Evidence for large precursor proteins in poliovirus synthesis. Proc. natn. Acad. Sei. U.S.A. 57, T a k e d a, N., K u h n, R. J., Y a n g, C.-F., T a k e g a m i, T. & W im m er, E. (1986). Initiation of poliovirus plus-strand RN A synthesis in a membrane complex of infected H ela cells, jf. Virol. 60, T a k e d a, N., Y a n g, C.-F., K u h n, R. J. & W im m er, E. (1987). Uridylylation of the genomelinked protein of poliovirus in vitro is dependent upon an endogenous RN A template. Virus Res. (in press). T a k e g a m i, T., K u h n, R. J., A n d e r s o n, C. W. & W im m e r, E. (1983a). Membrane-dependent uridylylation of the genome-linked protein VPg of poliovirus. Proc. natn. Acad. Sei. U.S.A. 80, T a k e g a m i, T., S e m l e r, B. L., A n d e r s o n, C. W. & W im m e r, E. (19836). Membrane fractions active in poliovirus R N A replication contain VPg precursor polypeptides. Virology 128, T e r s h a k, D. R. (1984). Association of poliovirus proteins with the endoplasmic reticulum. J. Virol. 52, T o m a s s in i, J. E. & C o lo n n o, R. J. (1986). Isolation of a receptor protein involved in attachment of human rhino viruses. J. Virol. 58, T o y o d a, H., N ic k l in, M. J. H., M u r r a y, M. G., A n d e r s o n, C. W., D u n n, J. J., S t u d ie r, F. W. & W im m e r, E. (1986a). A second virus-encoded proteinase involved in proteolytic processing of poliovirus polyprotein. Cell 45, T o y o d a, H., N ic k l in, M. J. H., M u r r a y, M. G. & W im m e r, E. (19866). Proteolytic processingof the poliovirus polyprotein by two virus-encoded proteinases. In Protein Engineering: Application in Science, M edicine and In d u stty (ed. M. Inouye & R. Sarm a), pp New York: Academic Press. T o y o d a, H., Y a n g, C.-F., T a k e d a, N., N om oto, A. & W im m er, E. (1987). Analysis of RNA synthesis of type 1 poliovirus using an in vitro molecular genetic approach. Unpublished. van DER W e r f, S. ( ). Clonage moleculaire du poliovirus type 1 et expression de ses proteines de capside chez Escherichia coli: identification d un epitope de neutralisation. These de Doctorat D E tat, Universität Paris 7.