Portal de Periódicos da CAPES

Co-translational, Intraribosomal Cleavage of Polypeptides by the Foot-and-mouth Disease Virus 2A Peptide

2003; Elsevier BV; Volume: 278; Issue: 13 Linguagem: Inglês

10.1074/jbc.m211644200

1083-351X

Pablo de Felipe, Lorraine E. Hughes, Martin D. Ryan, Jeremy D. Brown,

RNA and protein synthesis mechanisms

During co-translational protein import into the endoplasmic reticulum ribosomes are docked onto the translocon. This prevents inappropriate exposure of nascent chains to the cytosol and, conversely, cytosolic factors from gaining access to the nascent chain. We exploited this property of co-translational translocation to examine the mechanism of polypeptide cleavage by the 2A peptide of the foot-and-mouth disease virus. We find that the scission reaction is unaffected by placing 2A into a co-translationally targeted protein. Moreover, the portion of the polypeptide C-terminal to the cleavage site remains in the cytosol unless it contains its own signal sequence. The pattern of cleavage is consistent with the proposal that the 2A-mediated cleavage reaction occurs within the ribosome itself. In addition, our data indicate that the ribosome-translocon complex detects the break in the nascent chain and prevents any downstream protein lacking a signal sequence from gaining access to the endoplasmic reticulum. During co-translational protein import into the endoplasmic reticulum ribosomes are docked onto the translocon. This prevents inappropriate exposure of nascent chains to the cytosol and, conversely, cytosolic factors from gaining access to the nascent chain. We exploited this property of co-translational translocation to examine the mechanism of polypeptide cleavage by the 2A peptide of the foot-and-mouth disease virus. We find that the scission reaction is unaffected by placing 2A into a co-translationally targeted protein. Moreover, the portion of the polypeptide C-terminal to the cleavage site remains in the cytosol unless it contains its own signal sequence. The pattern of cleavage is consistent with the proposal that the 2A-mediated cleavage reaction occurs within the ribosome itself. In addition, our data indicate that the ribosome-translocon complex detects the break in the nascent chain and prevents any downstream protein lacking a signal sequence from gaining access to the endoplasmic reticulum. foot-and-mouth disease virus endoplasmic reticulum signal recognition particle CPY, carboxypeptidase Y Positive-strand RNA viruses typically encode polyproteins that are cleaved by viral or host-encoded proteinases (proteolytic processing) to produce mature, individual proteins (reviewed in Refs. 1Ryan M.D. Monaghan S. Flint M. J. Gen. Virol. 1998; 79: 947-959Google Scholar and 2Seipelt J. Guarne A. Bergmann E. James M. Sommergruber W. Fita I. Skern T. Virus Res. 1999; 62: 159-168Google Scholar). Alternatively proteins may be generated by translational effects such as ribosomal frameshifting or read-through of “leaky” stop codons. Such programmed alterations of translation are not virus-specific but widespread (although rare) mechanisms of gene expression (reviewed in Refs. 3Baranov P.V. Gurvich O.L. Fayet O. Prere M.F. Miller W.A. Gesteland R.F. Atkins J.F. Giddings M.C. Nucleic Acids Res. 2001; 29: 264-267Google Scholar and 4Farabaugh P.J. Annu. Rev. Genet. 1996; 30: 507-528Google Scholar). In foot-and-mouth disease virus (FMDV)1 and some other picornaviruses the oligopeptide (∼20 amino acid) 2A region of the polyprotein mediates cleavage at its own C terminus to release it from the 2B region. 2A is also active when placed between reporter proteins and, therefore, cleavage requires no viral (proteinase) sequences outside this short peptide (5Ryan M.D. Drew J. EMBO J. 1994; 13: 928-933Google Scholar). Similarly, no host proteinases are known that cleave the 2A/2B site.Scission of 2A-containing polyproteins requires the correct protein rather than mRNA sequence (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar). However, synthetic peptides containing 2A and “2A-like” sequences from other viruses do not autoproteolyse (7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar). Furthermore, on translation in vitro the portion of a polyprotein N-terminal to 2A typically accumulates in excess over the C-terminal portion. This imbalance is not because of protein degradation or nonspecific transcription/translation termination (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar, 8Donnelly M.L. Gani D. Flint M. Monaghan S. Ryan M.D. J. Gen. Virol. 1997; 78: 13-21Google Scholar). Modeling of 2A and 2A-like sequences indicates that the majority of each peptide can form an amphipathic helix whereas the amino acids immediately preceding the “cleavage” site (-NPG⇓P-) form a tight turn (7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar). A co-translational model for the cleavage reaction has been proposed (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar, 7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar) in which the conformation of 2A places strain on the peptidyltransferase center of the ribosome, re-positioning the peptidyl(2A)-tRNA ester linkage. This steric effect prohibits nucleophilic attack by the incoming (prolyl)-tRNA amide nitrogen that normally creates the new peptide bond. Instead, the N-terminal product is released from the ribosome by hydrolysis of the peptidyl(2A)-tRNA ester bond. A proportion of ribosomes then cease translation, while the remainder continue, effectively “initiated” by the prolyl-tRNA, to produce the downstream product as a discrete (“cleaved”) entity.Although published data support the proposal that 2A acts within the ribosome no direct evidence has been provided for this. Intraribosomal cleavage would not require cytosolic factors such as proteinases, and demonstration of this requires translation of 2A-containing proteins in a situation where cytosolic factors have no access to the nascent chain. Such “screening” occurs during co-translational translocation into the endoplasmic reticulum (ER). Here the nascent chain is shielded from cytosolic factors by first the ribosome, then the translocation apparatus, before being partitioned into the ER lumen. Establishment of co-translational translocation requires the signal recognition particle (SRP; reviewed in Refs. 9Keenan R.J. Freymann D.M. Stroud R.M. Walter P. Annu. Rev. Biochem. 2001; 70: 755-775Google Scholar and 10Bui N. Strub K. Biol. Chem. 1999; 380: 135-145Google Scholar). SRP binds cis-acting hydrophobic signal sequences at the N terminus of nascent ER-targeted proteins as they emerge from the ribosome and concomitantly slows translation by the ribosome, a phenomenon termed elongation arrest (11Walter P. Blobel G. J. Cell Biol. 1981; 91: 557-561Google Scholar, 12Wolin S.L. Walter P. J. Cell Biol. 1989; 109: 2617-2622Google Scholar, 13Mason N. Ciufo L.F. Brown J.D. EMBO J. 2000; 19: 4164-4174Google Scholar). This ensures that ribosome-nascent chain complexes targeted to the translocon by SRP arrive with a short length of cytosolically exposed nascent chain. Once the ribosome-translocon junction is established the ribosomal nascent chain exit site lies directly over the translocon (14Beckmann R. Bubeck D. Grassucci R. Penczek P. Verschoor A. Blobel G. Frank J. Science. 1997; 278: 2123-2126Google Scholar, 15Beckmann R. Spahn C.M. Eswar N. Helmers J. Penczek P.A. Sali A. Frank J. Blobel G. Cell. 2001; 107: 361-372Google Scholar, 16Menetret J. Neuhof A. Morgan D.G. Plath K. Radermacher M. Rapoport T.A. Akey C.W. Mol. Cell. 2000; 6: 1219-1232Google Scholar). This junction is sufficiently tight to protect the nascent chain from externally added proteases (17Matlack K.E. Walter P. J. Biol. Chem. 1995; 270: 6170-6180Google Scholar, 18Connolly T. Collins P. Gilmore R. J. Cell Biol. 1989; 108: 299-307Google Scholar) and even excludes small ions (19Crowley K.S. Liao S. Worrell V.E. Reinhart G.D. Johnson A.E. Cell. 1994; 78: 461-471Google Scholar,20Crowley K.S. Reinhart G.D. Johnson A.E. Cell. 1993; 73: 1101-1115Google Scholar). Thus comparing the translation products of 2A-containing proteins with and without SRP-dependent signal sequences will reveal if 2A functions without the influence of cytosolic factors.The translocon has been proposed to have a signal sequence recognition function independent from that of targeting factors such as SRP (21Jungnickel B. Rapoport T.A. Cell. 1995; 82: 261-270Google Scholar,22Mothes W. Jungnickel B. Brunner J. Rapoport T.A. J. Cell Biol. 1998; 142: 355-364Google Scholar). Co-translational translocation of a 2A-containing polypeptide will result in amino acids C-terminal to 2A being presented to the translocon immediately after they have emerged from the ribosome. Therefore if 2A is active in a co-translationally targeted protein we reasoned that we could use this to ask two interdependent questions. First, where does cleavage take place? Second, can the translocon discern the presence or absence of a signal sequence on a nascent chain presented to it directly by the ribosome and in vivo? If cleavage does not take place inside the ribosome the portion of the protein downstream of 2A will arrive in the lumen of the ER. However, if the nascent chain cleaves within the ribosome a gap will occur in the polypeptide. The translocon may then “detect” this discontinuity in the nascent chain as it does the normal termination of translation, closing, and excluding the downstream protein from the ER. In contrast addition of a signal sequence to the N terminus of protein downstream of the 2A site and expression in vivo should result in reopening of the translocon and translocation of this protein into the ER.We chose the yeast Saccharomyces cerevisiae for these studies. We demonstrate that the FMDV 2A sequence is functional in yeast, and that targeting through the SRP-dependent co-translational translocation pathway does not impair the cleavage reaction. In addition the released downstream product remains in the cytosol unless it contains its own signal sequence. Thus nascent chain cleavage takes place before the C-terminal portion of the protein initiates translocation and therefore within the ribosome. These data provide significant support for the proposal that 2A modifies the activity of the ribosome to promote scission of the nascent chain (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar,7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar) and are consistent with the notion that the translocon itself examines nascent chains for the presence of signal sequences.DISCUSSIONThe 2A sequence of FMDV directs production of separate proteins from one polyprotein open reading frame in a variety of higher eukaryotic and plant systems. As such it has great potential as a biotechnological tool, reducing the need for multiple vectors (e.g. Refs. 39Halpin C. Cooke S.E. Barakate A. El Amrani A. Ryan M.D. Plant J. 1999; 17: 453-459Google Scholar, 40Thomas C.L. Maule A.J. J. Gen. Virol. 2000; 81: 1851-1855Google Scholar, 41Varnavski A.N. Khromykh A.A. Virology. 1999; 255: 366-375Google Scholar, 42De Felipe P. Izquierdo M. Hum. Gene Ther. 2000; 11: 1921-1931Google Scholar). Previously the activity of FMDV 2A had not been examined in yeast. We found that it is active in vivo in yeast (Fig. 2). Thus, whatever features of higher eukaryotic cells allow 2A to promote scission of the polypeptide chain at the 2A C terminus are conserved to this “simpler” organism. This opens up the possibility of examining, in a genetically tractable system, how 2A functions and what trans-acting factors are required for or influence the cleavage reaction. A similar approach has been successfully applied to analysis of ribosomal frameshifting (43Harger J. Meskauskas A. Dinman J. Trends Biochem. Sci. 2002; 27: 448-454Google Scholar), indicating that such examination of alterations to normal ribosome function in yeast is feasible.The proportion of 2A-containing proteins that were cleaved in yeast was similar to that seen previously in in vitro translation reactions, and the 2A-dependent reaction was unaffected by sequestering the nascent chain from cytosolic factors. As discussed above although SRP-dependent targeting functions similarly in yeast to higher eukaryotes (13Mason N. Ciufo L.F. Brown J.D. EMBO J. 2000; 19: 4164-4174Google Scholar, 44Hann B.C. Walter P. Cell. 1991; 67: 131-144Google Scholar), translation and translocation are not efficiently coupled at short nascent chain lengths (37Johnsson N. Varshavsky A. EMBO J. 1994; 13: 2686-2698Google Scholar). Both SRP-dependent substrates used in our analysis are of sufficient length to expect tight coupling between targeting and translocation for the great majority of individual protein chains. Thus if cytosolic factors were required for 2A-dependent cleavage, a significant reduction of the proportion of these proteins that were cleaved would have been expected. Instead the efficiency of cleavage was maintained when the protein containing it was co-translationally translocated and we conclude that extraribosomal cytosolic factors such as proteinases are not necessary for the 2A reaction.Determination of the fate of the separated C-terminal (GFP) portions of 2A-containing polyproteins yielded information both on (i) where cleavage took place and (ii) the gating function of the translocon. Even when polyproteins were co-translationally translocated into the ER, GFP remained cytosolic if it lacked a signal sequence. This exclusion was, as far as we could determine from fractionation experiments, quantitative and our view of how this occurs is shown in Fig. 6. For GFP to be excluded from the ER the cleavage reaction must have taken place before the N terminus of GFP was translocated and therefore our data demonstrate that the proposal that the break in the nascent chain is generated within the ribosome is correct (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar, 7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar). We suggest that the N terminus of GFP is treated as a new translocation substrate and is “surveyed” by the translocon for the presence of a signal sequence. Without a signal sequence the translocon remains closed and the ribosome-translocon complex dissociates. When a signal sequence is added to the N terminus of GFP (substrates DNαF-2A-Kar2-GFP and DNαF-2A-CPY-GFP) the translocon recognizes this and reopens. In addition to indicating that cytosolic proteases are not responsible for the cleavage reaction the cytoplasmic location of GFP released from ER-targeted proteins refutes other models of 2A activity such as one 2A sequence acting in trans on another nascent 2A-containing protein to cleave it, or 2A requiring release from the ribosome and folding before it becomes active in cis. Each of these possibilities would result in cleavage within the ER lumen, and thus the GFP portion of the polyproteins would have been found in the ER.The model above and in Fig. 6 is consistent with the recent finding that, upon completion of co-translational translocation in vitro, ribosomes remain attached to the translocon, competent to initiate translation of new mRNA species (45Potter M.D. Nicchitta C.V. J. Biol. Chem. 2000; 275: 33828-33835Google Scholar, 46Potter M.D. Nicchitta C.V. J. Biol. Chem. 2002; 277: 23314-23320Google Scholar). If these ribosomes encode proteins with signal sequences they are translocated directly into the ER independent of targeting by SRP, whereas if they do not contain a signal sequence the translating ribosomes detach and release the protein into the cytosol. This has been argued to represent a physiologically important mechanism by which the cell ensures correct partitioning of cytosolic and ER-targeted proteins. Assuming that the in vitro experiments (45Potter M.D. Nicchitta C.V. J. Biol. Chem. 2000; 275: 33828-33835Google Scholar, 46Potter M.D. Nicchitta C.V. J. Biol. Chem. 2002; 277: 23314-23320Google Scholar) reflect in vivo events, then translocation of Kar2-GFP and CPY-GFP released from DNαF-2A-Kar2-GFP and DNαF-2A-CPY-GFP likely represents direct recognition of signal sequences in vivo by the translocon. Thus 2A may provide a means by which we can, for the first time, examine the signal sequence recognition and gating activities of the translocon in vivo independent of targeting.Analysis of 2A and 2A-like sequences has revealed that the FMDV 2A reaction is far from unique. It is a strategy for creating separate proteins from one polypeptide chain used by a variety of picornaviruses and insect viruses, and active 2A-like sequences are also found in repeated sequences within the genomes of Trypanosoma species (8Donnelly M.L. Gani D. Flint M. Monaghan S. Ryan M.D. J. Gen. Virol. 1997; 78: 13-21Google Scholar, 26Donnelly M.L. Hughes L.E. Luke G. Mendoza H. ten Dam E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1027-1041Google Scholar, 47Hahn H. Palmenberg A.C. J. Virol. 1996; 70: 6870-6875Google Scholar). Neither is 2A the only example of a nascent peptide chain that modifies the activity of the ribosome while within the ribosome exit tunnel (reviewed in Refs. 48Tenson T. Ehrenberg M. Cell. 2002; 108: 591-594Google Scholar and 49Morris D.R. Geballe A.P. Mol. Cell. Biol. 2000; 20: 8635-8642Google Scholar). Other examples are sequences within the bacteriophage T4 gene 60 that promote “hopping” of the ribosome along the mRNA (50Herr A.J. Gesteland R.F. Atkins J.F. EMBO J. 2000; 19: 2671-2680Google Scholar), the TnaC leader peptide of theEscherichia coli tryptophanase operon that, in the presence of tryptophan, causes ribosomes to stall allowing expression of the downstream open reading frames (51Gong F. Yanofsky C. Science. 2002; 297: 1864-1867Google Scholar) and a peptide within SecM that causes ribosomes to stall unless the protein is engaged with the protein export machinery. Mutations that suppress the stalling activity of SecM have been identified in both RNA and protein components of the ribosome, these mapping to the nascent chain exit tunnel (52Nakatogawa H. Ito K. Cell. 2002; 108: 629-636Google Scholar). Thus interactions between the nascent chain and the ribosome can have significant effects on translation and we expect that our further analysis may reveal similar interactions important for the activity of the 2A peptide. Positive-strand RNA viruses typically encode polyproteins that are cleaved by viral or host-encoded proteinases (proteolytic processing) to produce mature, individual proteins (reviewed in Refs. 1Ryan M.D. Monaghan S. Flint M. J. Gen. Virol. 1998; 79: 947-959Google Scholar and 2Seipelt J. Guarne A. Bergmann E. James M. Sommergruber W. Fita I. Skern T. Virus Res. 1999; 62: 159-168Google Scholar). Alternatively proteins may be generated by translational effects such as ribosomal frameshifting or read-through of “leaky” stop codons. Such programmed alterations of translation are not virus-specific but widespread (although rare) mechanisms of gene expression (reviewed in Refs. 3Baranov P.V. Gurvich O.L. Fayet O. Prere M.F. Miller W.A. Gesteland R.F. Atkins J.F. Giddings M.C. Nucleic Acids Res. 2001; 29: 264-267Google Scholar and 4Farabaugh P.J. Annu. Rev. Genet. 1996; 30: 507-528Google Scholar). In foot-and-mouth disease virus (FMDV)1 and some other picornaviruses the oligopeptide (∼20 amino acid) 2A region of the polyprotein mediates cleavage at its own C terminus to release it from the 2B region. 2A is also active when placed between reporter proteins and, therefore, cleavage requires no viral (proteinase) sequences outside this short peptide (5Ryan M.D. Drew J. EMBO J. 1994; 13: 928-933Google Scholar). Similarly, no host proteinases are known that cleave the 2A/2B site. Scission of 2A-containing polyproteins requires the correct protein rather than mRNA sequence (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar). However, synthetic peptides containing 2A and “2A-like” sequences from other viruses do not autoproteolyse (7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar). Furthermore, on translation in vitro the portion of a polyprotein N-terminal to 2A typically accumulates in excess over the C-terminal portion. This imbalance is not because of protein degradation or nonspecific transcription/translation termination (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar, 8Donnelly M.L. Gani D. Flint M. Monaghan S. Ryan M.D. J. Gen. Virol. 1997; 78: 13-21Google Scholar). Modeling of 2A and 2A-like sequences indicates that the majority of each peptide can form an amphipathic helix whereas the amino acids immediately preceding the “cleavage” site (-NPG⇓P-) form a tight turn (7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar). A co-translational model for the cleavage reaction has been proposed (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar, 7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar) in which the conformation of 2A places strain on the peptidyltransferase center of the ribosome, re-positioning the peptidyl(2A)-tRNA ester linkage. This steric effect prohibits nucleophilic attack by the incoming (prolyl)-tRNA amide nitrogen that normally creates the new peptide bond. Instead, the N-terminal product is released from the ribosome by hydrolysis of the peptidyl(2A)-tRNA ester bond. A proportion of ribosomes then cease translation, while the remainder continue, effectively “initiated” by the prolyl-tRNA, to produce the downstream product as a discrete (“cleaved”) entity. Although published data support the proposal that 2A acts within the ribosome no direct evidence has been provided for this. Intraribosomal cleavage would not require cytosolic factors such as proteinases, and demonstration of this requires translation of 2A-containing proteins in a situation where cytosolic factors have no access to the nascent chain. Such “screening” occurs during co-translational translocation into the endoplasmic reticulum (ER). Here the nascent chain is shielded from cytosolic factors by first the ribosome, then the translocation apparatus, before being partitioned into the ER lumen. Establishment of co-translational translocation requires the signal recognition particle (SRP; reviewed in Refs. 9Keenan R.J. Freymann D.M. Stroud R.M. Walter P. Annu. Rev. Biochem. 2001; 70: 755-775Google Scholar and 10Bui N. Strub K. Biol. Chem. 1999; 380: 135-145Google Scholar). SRP binds cis-acting hydrophobic signal sequences at the N terminus of nascent ER-targeted proteins as they emerge from the ribosome and concomitantly slows translation by the ribosome, a phenomenon termed elongation arrest (11Walter P. Blobel G. J. Cell Biol. 1981; 91: 557-561Google Scholar, 12Wolin S.L. Walter P. J. Cell Biol. 1989; 109: 2617-2622Google Scholar, 13Mason N. Ciufo L.F. Brown J.D. EMBO J. 2000; 19: 4164-4174Google Scholar). This ensures that ribosome-nascent chain complexes targeted to the translocon by SRP arrive with a short length of cytosolically exposed nascent chain. Once the ribosome-translocon junction is established the ribosomal nascent chain exit site lies directly over the translocon (14Beckmann R. Bubeck D. Grassucci R. Penczek P. Verschoor A. Blobel G. Frank J. Science. 1997; 278: 2123-2126Google Scholar, 15Beckmann R. Spahn C.M. Eswar N. Helmers J. Penczek P.A. Sali A. Frank J. Blobel G. Cell. 2001; 107: 361-372Google Scholar, 16Menetret J. Neuhof A. Morgan D.G. Plath K. Radermacher M. Rapoport T.A. Akey C.W. Mol. Cell. 2000; 6: 1219-1232Google Scholar). This junction is sufficiently tight to protect the nascent chain from externally added proteases (17Matlack K.E. Walter P. J. Biol. Chem. 1995; 270: 6170-6180Google Scholar, 18Connolly T. Collins P. Gilmore R. J. Cell Biol. 1989; 108: 299-307Google Scholar) and even excludes small ions (19Crowley K.S. Liao S. Worrell V.E. Reinhart G.D. Johnson A.E. Cell. 1994; 78: 461-471Google Scholar,20Crowley K.S. Reinhart G.D. Johnson A.E. Cell. 1993; 73: 1101-1115Google Scholar). Thus comparing the translation products of 2A-containing proteins with and without SRP-dependent signal sequences will reveal if 2A functions without the influence of cytosolic factors. The translocon has been proposed to have a signal sequence recognition function independent from that of targeting factors such as SRP (21Jungnickel B. Rapoport T.A. Cell. 1995; 82: 261-270Google Scholar,22Mothes W. Jungnickel B. Brunner J. Rapoport T.A. J. Cell Biol. 1998; 142: 355-364Google Scholar). Co-translational translocation of a 2A-containing polypeptide will result in amino acids C-terminal to 2A being presented to the translocon immediately after they have emerged from the ribosome. Therefore if 2A is active in a co-translationally targeted protein we reasoned that we could use this to ask two interdependent questions. First, where does cleavage take place? Second, can the translocon discern the presence or absence of a signal sequence on a nascent chain presented to it directly by the ribosome and in vivo? If cleavage does not take place inside the ribosome the portion of the protein downstream of 2A will arrive in the lumen of the ER. However, if the nascent chain cleaves within the ribosome a gap will occur in the polypeptide. The translocon may then “detect” this discontinuity in the nascent chain as it does the normal termination of translation, closing, and excluding the downstream protein from the ER. In contrast addition of a signal sequence to the N terminus of protein downstream of the 2A site and expression in vivo should result in reopening of the translocon and translocation of this protein into the ER. We chose the yeast Saccharomyces cerevisiae for these studies. We demonstrate that the FMDV 2A sequence is functional in yeast, and that targeting through the SRP-dependent co-translational translocation pathway does not impair the cleavage reaction. In addition the released downstream product remains in the cytosol unless it contains its own signal sequence. Thus nascent chain cleavage takes place before the C-terminal portion of the protein initiates translocation and therefore within the ribosome. These data provide significant support for the proposal that 2A modifies the activity of the ribosome to promote scission of the nascent chain (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar,7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar) and are consistent with the notion that the translocon itself examines nascent chains for the presence of signal sequences. DISCUSSIONThe 2A sequence of FMDV directs production of separate proteins from one polyprotein open reading frame in a variety of higher eukaryotic and plant systems. As such it has great potential as a biotechnological tool, reducing the need for multiple vectors (e.g. Refs. 39Halpin C. Cooke S.E. Barakate A. El Amrani A. Ryan M.D. Plant J. 1999; 17: 453-459Google Scholar, 40Thomas C.L. Maule A.J. J. Gen. Virol. 2000; 81: 1851-1855Google Scholar, 41Varnavski A.N. Khromykh A.A. Virology. 1999; 255: 366-375Google Scholar, 42De Felipe P. Izquierdo M. Hum. Gene Ther. 2000; 11: 1921-1931Google Scholar). Previously the activity of FMDV 2A had not been examined in yeast. We found that it is active in vivo in yeast (Fig. 2). Thus, whatever features of higher eukaryotic cells allow 2A to promote scission of the polypeptide chain at the 2A C terminus are conserved to this “simpler” organism. This opens up the possibility of examining, in a genetically tractable system, how 2A functions and what trans-acting factors are required for or influence the cleavage reaction. A similar approach has been successfully applied to analysis of ribosomal frameshifting (43Harger J. Meskauskas A. Dinman J. Trends Biochem. Sci. 2002; 27: 448-454Google Scholar), indicating that such examination of alterations to normal ribosome function in yeast is feasible.The proportion of 2A-containing proteins that were cleaved in yeast was similar to that seen previously in in vitro translation reactions, and the 2A-dependent reaction was unaffected by sequestering the nascent chain from cytosolic factors. As discussed above although SRP-dependent targeting functions similarly in yeast to higher eukaryotes (13Mason N. Ciufo L.F. Brown J.D. EMBO J. 2000; 19: 4164-4174Google Scholar, 44Hann B.C. Walter P. Cell. 1991; 67: 131-144Google Scholar), translation and translocation are not efficiently coupled at short nascent chain lengths (37Johnsson N. Varshavsky A. EMBO J. 1994; 13: 2686-2698Google Scholar). Both SRP-dependent substrates used in our analysis are of sufficient length to expect tight coupling between targeting and translocation for the great majority of individual protein chains. Thus if cytosolic factors were required for 2A-dependent cleavage, a significant reduction of the proportion of these proteins that were cleaved would have been expected. Instead the efficiency of cleavage was maintained when the protein containing it was co-translationally translocated and we conclude that extraribosomal cytosolic factors such as proteinases are not necessary for the 2A reaction.Determination of the fate of the separated C-terminal (GFP) portions of 2A-containing polyproteins yielded information both on (i) where cleavage took place and (ii) the gating function of the translocon. Even when polyproteins were co-translationally translocated into the ER, GFP remained cytosolic if it lacked a signal sequence. This exclusion was, as far as we could determine from fractionation experiments, quantitative and our view of how this occurs is shown in Fig. 6. For GFP to be excluded from the ER the cleavage reaction must have taken place before the N terminus of GFP was translocated and therefore our data demonstrate that the proposal that the break in the nascent chain is generated within the ribosome is correct (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar, 7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar). We suggest that the N terminus of GFP is treated as a new translocation substrate and is “surveyed” by the translocon for the presence of a signal sequence. Without a signal sequence the translocon remains closed and the ribosome-translocon complex dissociates. When a signal sequence is added to the N terminus of GFP (substrates DNαF-2A-Kar2-GFP and DNαF-2A-CPY-GFP) the translocon recognizes this and reopens. In addition to indicating that cytosolic proteases are not responsible for the cleavage reaction the cytoplasmic location of GFP released from ER-targeted proteins refutes other models of 2A activity such as one 2A sequence acting in trans on another nascent 2A-containing protein to cleave it, or 2A requiring release from the ribosome and folding before it becomes active in cis. Each of these possibilities would result in cleavage within the ER lumen, and thus the GFP portion of the polyproteins would have been found in the ER.The model above and in Fig. 6 is consistent with the recent finding that, upon completion of co-translational translocation in vitro, ribosomes remain attached to the translocon, competent to initiate translation of new mRNA species (45Potter M.D. Nicchitta C.V. J. Biol. Chem. 2000; 275: 33828-33835Google Scholar, 46Potter M.D. Nicchitta C.V. J. Biol. Chem. 2002; 277: 23314-23320Google Scholar). If these ribosomes encode proteins with signal sequences they are translocated directly into the ER independent of targeting by SRP, whereas if they do not contain a signal sequence the translating ribosomes detach and release the protein into the cytosol. This has been argued to represent a physiologically important mechanism by which the cell ensures correct partitioning of cytosolic and ER-targeted proteins. Assuming that the in vitro experiments (45Potter M.D. Nicchitta C.V. J. Biol. Chem. 2000; 275: 33828-33835Google Scholar, 46Potter M.D. Nicchitta C.V. J. Biol. Chem. 2002; 277: 23314-23320Google Scholar) reflect in vivo events, then translocation of Kar2-GFP and CPY-GFP released from DNαF-2A-Kar2-GFP and DNαF-2A-CPY-GFP likely represents direct recognition of signal sequences in vivo by the translocon. Thus 2A may provide a means by which we can, for the first time, examine the signal sequence recognition and gating activities of the translocon in vivo independent of targeting.Analysis of 2A and 2A-like sequences has revealed that the FMDV 2A reaction is far from unique. It is a strategy for creating separate proteins from one polypeptide chain used by a variety of picornaviruses and insect viruses, and active 2A-like sequences are also found in repeated sequences within the genomes of Trypanosoma species (8Donnelly M.L. Gani D. Flint M. Monaghan S. Ryan M.D. J. Gen. Virol. 1997; 78: 13-21Google Scholar, 26Donnelly M.L. Hughes L.E. Luke G. Mendoza H. ten Dam E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1027-1041Google Scholar, 47Hahn H. Palmenberg A.C. J. Virol. 1996; 70: 6870-6875Google Scholar). Neither is 2A the only example of a nascent peptide chain that modifies the activity of the ribosome while within the ribosome exit tunnel (reviewed in Refs. 48Tenson T. Ehrenberg M. Cell. 2002; 108: 591-594Google Scholar and 49Morris D.R. Geballe A.P. Mol. Cell. Biol. 2000; 20: 8635-8642Google Scholar). Other examples are sequences within the bacteriophage T4 gene 60 that promote “hopping” of the ribosome along the mRNA (50Herr A.J. Gesteland R.F. Atkins J.F. EMBO J. 2000; 19: 2671-2680Google Scholar), the TnaC leader peptide of theEscherichia coli tryptophanase operon that, in the presence of tryptophan, causes ribosomes to stall allowing expression of the downstream open reading frames (51Gong F. Yanofsky C. Science. 2002; 297: 1864-1867Google Scholar) and a peptide within SecM that causes ribosomes to stall unless the protein is engaged with the protein export machinery. Mutations that suppress the stalling activity of SecM have been identified in both RNA and protein components of the ribosome, these mapping to the nascent chain exit tunnel (52Nakatogawa H. Ito K. Cell. 2002; 108: 629-636Google Scholar). Thus interactions between the nascent chain and the ribosome can have significant effects on translation and we expect that our further analysis may reveal similar interactions important for the activity of the 2A peptide. The 2A sequence of FMDV directs production of separate proteins from one polyprotein open reading frame in a variety of higher eukaryotic and plant systems. As such it has great potential as a biotechnological tool, reducing the need for multiple vectors (e.g. Refs. 39Halpin C. Cooke S.E. Barakate A. El Amrani A. Ryan M.D. Plant J. 1999; 17: 453-459Google Scholar, 40Thomas C.L. Maule A.J. J. Gen. Virol. 2000; 81: 1851-1855Google Scholar, 41Varnavski A.N. Khromykh A.A. Virology. 1999; 255: 366-375Google Scholar, 42De Felipe P. Izquierdo M. Hum. Gene Ther. 2000; 11: 1921-1931Google Scholar). Previously the activity of FMDV 2A had not been examined in yeast. We found that it is active in vivo in yeast (Fig. 2). Thus, whatever features of higher eukaryotic cells allow 2A to promote scission of the polypeptide chain at the 2A C terminus are conserved to this “simpler” organism. This opens up the possibility of examining, in a genetically tractable system, how 2A functions and what trans-acting factors are required for or influence the cleavage reaction. A similar approach has been successfully applied to analysis of ribosomal frameshifting (43Harger J. Meskauskas A. Dinman J. Trends Biochem. Sci. 2002; 27: 448-454Google Scholar), indicating that such examination of alterations to normal ribosome function in yeast is feasible. The proportion of 2A-containing proteins that were cleaved in yeast was similar to that seen previously in in vitro translation reactions, and the 2A-dependent reaction was unaffected by sequestering the nascent chain from cytosolic factors. As discussed above although SRP-dependent targeting functions similarly in yeast to higher eukaryotes (13Mason N. Ciufo L.F. Brown J.D. EMBO J. 2000; 19: 4164-4174Google Scholar, 44Hann B.C. Walter P. Cell. 1991; 67: 131-144Google Scholar), translation and translocation are not efficiently coupled at short nascent chain lengths (37Johnsson N. Varshavsky A. EMBO J. 1994; 13: 2686-2698Google Scholar). Both SRP-dependent substrates used in our analysis are of sufficient length to expect tight coupling between targeting and translocation for the great majority of individual protein chains. Thus if cytosolic factors were required for 2A-dependent cleavage, a significant reduction of the proportion of these proteins that were cleaved would have been expected. Instead the efficiency of cleavage was maintained when the protein containing it was co-translationally translocated and we conclude that extraribosomal cytosolic factors such as proteinases are not necessary for the 2A reaction. Determination of the fate of the separated C-terminal (GFP) portions of 2A-containing polyproteins yielded information both on (i) where cleavage took place and (ii) the gating function of the translocon. Even when polyproteins were co-translationally translocated into the ER, GFP remained cytosolic if it lacked a signal sequence. This exclusion was, as far as we could determine from fractionation experiments, quantitative and our view of how this occurs is shown in Fig. 6. For GFP to be excluded from the ER the cleavage reaction must have taken place before the N terminus of GFP was translocated and therefore our data demonstrate that the proposal that the break in the nascent chain is generated within the ribosome is correct (6Donnelly M.L. Luke G. Mehrotra A. Li X. Hughes L.E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1013-1025Google Scholar, 7Ryan M.D. Donnelly M.L. Lewis A. Mehotra A.P. Wilke J. Gani D. Bioorg. Chem. 1999; 27: 55-79Google Scholar). We suggest that the N terminus of GFP is treated as a new translocation substrate and is “surveyed” by the translocon for the presence of a signal sequence. Without a signal sequence the translocon remains closed and the ribosome-translocon complex dissociates. When a signal sequence is added to the N terminus of GFP (substrates DNαF-2A-Kar2-GFP and DNαF-2A-CPY-GFP) the translocon recognizes this and reopens. In addition to indicating that cytosolic proteases are not responsible for the cleavage reaction the cytoplasmic location of GFP released from ER-targeted proteins refutes other models of 2A activity such as one 2A sequence acting in trans on another nascent 2A-containing protein to cleave it, or 2A requiring release from the ribosome and folding before it becomes active in cis. Each of these possibilities would result in cleavage within the ER lumen, and thus the GFP portion of the polyproteins would have been found in the ER. The model above and in Fig. 6 is consistent with the recent finding that, upon completion of co-translational translocation in vitro, ribosomes remain attached to the translocon, competent to initiate translation of new mRNA species (45Potter M.D. Nicchitta C.V. J. Biol. Chem. 2000; 275: 33828-33835Google Scholar, 46Potter M.D. Nicchitta C.V. J. Biol. Chem. 2002; 277: 23314-23320Google Scholar). If these ribosomes encode proteins with signal sequences they are translocated directly into the ER independent of targeting by SRP, whereas if they do not contain a signal sequence the translating ribosomes detach and release the protein into the cytosol. This has been argued to represent a physiologically important mechanism by which the cell ensures correct partitioning of cytosolic and ER-targeted proteins. Assuming that the in vitro experiments (45Potter M.D. Nicchitta C.V. J. Biol. Chem. 2000; 275: 33828-33835Google Scholar, 46Potter M.D. Nicchitta C.V. J. Biol. Chem. 2002; 277: 23314-23320Google Scholar) reflect in vivo events, then translocation of Kar2-GFP and CPY-GFP released from DNαF-2A-Kar2-GFP and DNαF-2A-CPY-GFP likely represents direct recognition of signal sequences in vivo by the translocon. Thus 2A may provide a means by which we can, for the first time, examine the signal sequence recognition and gating activities of the translocon in vivo independent of targeting. Analysis of 2A and 2A-like sequences has revealed that the FMDV 2A reaction is far from unique. It is a strategy for creating separate proteins from one polypeptide chain used by a variety of picornaviruses and insect viruses, and active 2A-like sequences are also found in repeated sequences within the genomes of Trypanosoma species (8Donnelly M.L. Gani D. Flint M. Monaghan S. Ryan M.D. J. Gen. Virol. 1997; 78: 13-21Google Scholar, 26Donnelly M.L. Hughes L.E. Luke G. Mendoza H. ten Dam E. Gani D. Ryan M.D. J. Gen. Virol. 2001; 82: 1027-1041Google Scholar, 47Hahn H. Palmenberg A.C. J. Virol. 1996; 70: 6870-6875Google Scholar). Neither is 2A the only example of a nascent peptide chain that modifies the activity of the ribosome while within the ribosome exit tunnel (reviewed in Refs. 48Tenson T. Ehrenberg M. Cell. 2002; 108: 591-594Google Scholar and 49Morris D.R. Geballe A.P. Mol. Cell. Biol. 2000; 20: 8635-8642Google Scholar). Other examples are sequences within the bacteriophage T4 gene 60 that promote “hopping” of the ribosome along the mRNA (50Herr A.J. Gesteland R.F. Atkins J.F. EMBO J. 2000; 19: 2671-2680Google Scholar), the TnaC leader peptide of theEscherichia coli tryptophanase operon that, in the presence of tryptophan, causes ribosomes to stall allowing expression of the downstream open reading frames (51Gong F. Yanofsky C. Science. 2002; 297: 1864-1867Google Scholar) and a peptide within SecM that causes ribosomes to stall unless the protein is engaged with the protein export machinery. Mutations that suppress the stalling activity of SecM have been identified in both RNA and protein components of the ribosome, these mapping to the nascent chain exit tunnel (52Nakatogawa H. Ito K. Cell. 2002; 108: 629-636Google Scholar). Thus interactions between the nascent chain and the ribosome can have significant effects on translation and we expect that our further analysis may reveal similar interactions important for the activity of the 2A peptide. We thank members of the Brown laboratory for discussions and suggestions, and Janet Quinn, Simon Whitehall, and Susan Farrington for comments on the manuscript. We are grateful to Davis Ng, Peter Walter, Tom Stevens, Joachim Li, Caroline Shamu, Nils Johnsson, and Reid Gilmore for reagents that were used during this study.