CORE brought to you by brought Helsingin yliopiston digitaalinen arkisto provided by Helsingin yliopiston 3/2006 University of Helsinki KATRI MÄKELÄINEN KATRI Institute of Biotechnology and Institute of Biotechnology Department of Applied Biology Department of Viikki Graduate School in Biosciences School Graduate Viikki Forestry, and Forestry, of Agriculture and Faculty Department of Applied Chemistry and Microbiology Applied Chemistry Department of Lost in : Translation Mechanisms Translation Lost in in Translation: Production of Cocksfoot Mottle Virus Virus Mottle of Cocksfoot Production Dissertationes bioscientiarum molecularium Universitatis Helsingiensis in Viikki Helsingiensis in Dissertationes bioscientiarum molecularium Universitatis

KATRI MÄKELÄINEN Lost in Translation: Translation Mechanisms in Production of Cocksfoot Mottle Virus Proteins 3/2006 KATRI MÄKELÄINEN Lost in Translation: Translation Mechanisms in Production of Cocksfoot Mottle Virus Proteins 3/2006

Helsinki 2006 ISSN 1795-7079 ISBN 952-10-3016-X

Recent Publications in this Series: in this Publications Recent Edelman 6/2005 Sanna Characteristics Commensal and Pathogenic Lactobacilli: Mucosa-Adherent Karhinen 7/2005 Leena Saccharomyces Reticulum of the Yeast in the Endoplasmic Proteins and Sorting Of Secretory Glycosylation cerevisiae Sen 8/2005 Saurabh alpha2-Adrenergic Receptor Subtypes Functional Studies on 9/2005 Tiina E. Raevaara in Colon Cancer Patients of Minor MLH1 Germline Alterations Found Functional Significance 10/2005 Katja Pihlainen Spectrometry in Analysing Drug Seizures and Atmospheric Pressure Ionisation Mass Liquid Chromatography Pietri Puustinen 11/2005 VPg and CP Movement Related Proteins A of Potato Virus Posttranslational Modifications 12/2005 Irmgard Suominen Tree Related to Paper and Spruce Paenibacillus and Bacillus 13/2005 Heidi Hyytiäinen Ssp. Carotovora Carotovora in the Plant Pathogen Erwinia Virulence Networks Controlling Regulatory 14/2005 Sanna Janhunen Mesolimbic Dopaminergic Pathways to Nicotinic Receptor Agonists Different Responses of the Nigrostriatal and 15/2005 Denis Kainov Packaging Motors of Cystoviruses 16/2005 Ivan Pavlov (HB-GAM) in Activity-Dependent Neuronal Plasticity in Heparin-Binding Growth-Associated Molecule Hippocampus 17/2005 Laura Seppä and Mammalian Cells Regulation of Heat Shock Response in Yeast Jaakola 18/2005 Veli-Pekka Membrane Proteins Functional and Structural Studies on Heptahelical 19/2005 Anssi Rantakari carotovora System in Erwinia Secretion Three Characterisation of the Type Airaksinen 20/2005 Sari Physical Stability of Solid Pharmaceutical Formulations Role of Excipients in Moisture Sorption and 21/2005 Tiina Hilden Integrin is Regulated by Phosphorylation of the LFA-1 Avidity Affinity and Ari Pekka Mähönen 22/2005 Arabidopsis thaliana Root Morphogenesis in the Cytokinins Regulate Vascular 23/2005 J. Matias Palva in the Developing and Adult Brain Interactions Among Neuronal Oscillations 25/2005 Michael Stefanidakis and ß2 Integrins: Role of the Complexes in Leukocyte Cell-Surface Association between Progelatinases Migration 26/2005 Heli Kansanaho Pharmacies Implementation of the Principles of Patient Counselling into Practice in Finnish Community 1/2006 Julia Perttilä Virus Semliki Forest and Activities and Subcellular Localization of Hepatitis E Virus Enzymatic Expression, Replicase Proteins Wennberg 2/2006 Tero Computer-Assisted Separation and Primary Screening of Bioactive Compounds View metadata, citation and similar papers at core.ac.uk similar papers and metadata, citation View Lost in Translation: Translation Mechanisms in Production of Cocksfoot Mottle Virus Proteins

Katri Mäkeläinen

Institute of Biotechnology and Department of Applied Biology Department of Applied Chemistry and Microbiology Faculty of Agriculture and Forestry, and Viikki Graduate School in Biosciences University of Helsinki

ACADEMIC DISSERTATION

To be presented, with the permission of the Faculty of Agriculture and Forestry of the University of Helsinki, for public criticism in the auditorium 1041 at Viikki Biocenter (Viikinkaari 5) on March 31th, 2006, at 12 noon. Supervisor: Docent Kristiina Mäkinen Department of Applied Chemistry and Microbiology University of Helsinki Finland

Reviewers: Docent Maija Vihinen-Ranta Department of Biological and Environmental Science University of Jyväskylä Finland

And

Professor Carl-Henrik von Bonsdorff Department of Food and Environmental Hygiene University of Helsinki Finland

Opponent: Professor W. Allen Miller Plant Pathology Department Iowa State University USA

3/2006 ISBN 952-10-3016-X (paperback) ISBN 952-10-3017-8 (PDF, online) ISSN 1795-7079 (paperback) ISSN 1795-8229 (PDF, online) Gummerus Kirjapaino Oy Vaajakoski 2006

To my family ORIGINAL PUBLICATIONS

I Mäkeläinen K., Aspegren K., Wahlström E., Teeri T.H., and Mäkinen K. Comparison of the translational properties of Cocksfoot mottle virus 5’leader sequence with known translational enhancers from other plant viruses. (Manuscript)

II Mäkeläinen K., and Mäkinen K. Testing of internal initiation via dicistronic constructs is complicated by production of extraneous transcripts. (Submitted)

III Mäkinen K., Mäkeläinen K., Arshava N., Tamm T., Merits A., Truve E., Zavriev S., and Saarma M. 2000. Characterization of VPg and the polyprotein processing of Cocksfoot mottle virus (genus Sobemovirus). J. Gen. Virol. 81: 2783-2789.

IV Lucchesi J., Mäkeläinen K., Merits A., Tamm T., and Mäkinen K. 2000. Regulation of –1 ribosomal frameshifting directed by Cocksfoot mottle sobemovirus genome. Eur. J. Biochem. 267: 3523-3529.

V Mäkeläinen K., and Mäkinen K. 2005. Factors affecting translation at the programmed –1 ribosomal frameshifting site of Cocksfoot mottle virus RNA in vivo. Nucleic Acids Res. 33: 2239-2247.

Also unpublished data will be presented.

CONTENTS

LIST OF ORIGINAL PUBLICATIONS

ABBREVIATIONS

ABSTRACT ...... 1

1. INTRODUCTION ...... 2 1.1 TRANSLATION INITIATION ...... 3 1.1.1 Regulation of translation initiation during virus infection ...... 6 1.1.2 Plant viral translational enhancers ...... 7 1.1.3 Internal entry sites (IRESs) ...... 9 1.1.3.1 Mechanism of IRES-mediated translation initiation ...... 10 1.1.3.2 IRES trans-acting factors (ITAFs) ...... 11 1.1.4 Leaky scanning ...... 12 1.1.5 Reinitiation ...... 12 1.1.6 SgRNAs ...... 13

1.2 TRANSLATION ELONGATION ...... 13 1.2.1 Regulation of elongation ...... 15 1.2.2 Programmed -1 ribosomal frameshifting (-1 PRF) ...... 16 1.2.2.1 The -1 PRF signals ...... 17 1.2.2.2 Mechanism of -1 PRF ...... 19

1.3 TERMINATION OF TRANSLATION ...... 21 1.3.1 Programmed termination codon readthrough ...... 22

1.4 POSTTRANSLATIONAL REGULATION OF .. 22 1.4.1 Polyprotein processing ...... 23 1.4.2 Viral proteases ...... 24

2. AIMS OF THE STUDY ...... 26

3. MATERIALS AND METHODS ...... 27 4. RESULTS AND DISCUSSION ...... 31 4.1 Translation initiation from CfMV RNA ...... 31 4.1.1 Comparison of production from CfMVε with known plant viral translational enhancers (I) ...... 31 4.1.2 Identifi cation of regions important for gene expression from CfMVε in barley suspension cells (I) ...... 32 4.1.3 Transient expression from in vitro transcribed mRNAs (I) ...... 34 4.1.4 Functioning of viral leader sequences in S. cerevisiae (II) ...... 34 4.1.5 Translational properties of CfMVε in vitro (I) ...... 35 4.1.6 Contribution of CfMV 3’UTR on translation initiation from CfMVε (I, unpublished) ...... 37

4.2 Does CfMVε promote internal initiation of translation? ...... 38 4.2.1 Studies on internal initiation in WGE (I) ...... 38 4.2.2 Internal initiation in barley suspension cells (unpublished) ...... 42 4.2.3 Internal initiation in yeast (II) ...... 43 4.2.3.1 Identifi cation of regions important for gene expression from internally positioned CfMVε in yeast (unpublished) ...... 48 4.2.3.2 Determination of 3’ cistron translation from dicistronic mRNAs in yeast spheroplasts (II) ...... 49 4.2.4 Evaluation of the dicistronic approach in IRES studies ...... 49

4.3 Proteolytic processing of CfMV polyprotein (III) ...... 50 4.3.1 N-terminal sequencing of CfMV VPg...... 50 4.3.2 Polyprotein processing in infected plants ...... 50 4.3.3 Putative processing sites of CfMV polyprotein ...... 52

4.4 Synthesis of CfMV polyprotein ...... 54 4.4.1 CfMV RNA programmed -1 ribosomal frameshifting in WGE (IV, V) 54 4.4.2 The -1 PRF in vivo (V) ...... 55 4.4.3 Regulation of -1 PRF by CfMV proteins (V) ...... 57

5. CONCLUDING REMARKS ...... 59

6. ACKNOWLEDGEMENTS ...... 61

7. REFERENCES ...... 70

REPRINTS OF ORIGINAL PUBLICATIONS ABBREVIATIONS aa-tRNA aminoacyl tRNA AMV Alfalfa mosaic virus A-site aa-tRNA binding site ATP adenosine triphosphate ATPase adenosine triphosphatase bp BYDV Barley yellow dwarf virus CaMV Caulifl ower mosaic virus cDNA complementary DNA CfMV Cocksfoot mottle virus CfMVε Cocksfoot mottle virus 5’UTR CP coat protein CrPV Cricket paralysis virus CrTMV Crucifer-infecting tobamovirus ds double-stranded 4E-BP eIF4E-binding protein eEF eukaryotic elongation factor eIF eukaryotic initiation factor ER endoplasmic reticulum eRF eukaryotic release factor E-site exit site GDP guanosine diphosphate GFP green fl uorescent protein gRNA genomic RNA GUS β-glucuronidase GTP guanosine triphosphate HC-Pro helper component proteinase HCRSV Hibiscus chlorotic ringspot virus HCV Hepatitis C virus HIV Human immunodefi ciency virus HP hairpin Hsp101 heat-shock protein 101 icDNA infectious complementary DNA ICS intercistronic spacer/sequence IRES internal ribosome entry site ITAF IRES trans-acting factor kDa kilodalton La lupus antigen lacZ β-galactosidase encoding gene lacZ β-galactosidase protein luc fi refl y luciferase encoding gene LUC fi refl y luciferase protein m7G 7-methylguanosine

Met-tRNAi initiator methionine-tRNA mRNA messenger RNA MuLV Moloney murine leukemia virus NIaPro nuclear inclusion protein a proteinase NMD nonsense-mediated RNA decay pathway Nt nucleotide ODC ornithine decarboxylase ORF open reading frame Paip PABP-interacting protein PABP poly(A) binding protein PCBP poly(rC) binding protein PGK1 phosphoglycerate kinase 1 PKR dsRNA-activated protein kinase PLRV Potato leafroll virus Poly(A) polyadenylic acid PRF programmed ribosomal frameshifting Pro protease P-site peptidyl-tRNA binding site PTB polypyrimidine tract binding protein PV Poliovirus PVA Potato virus A PVX Potato virus X RdRp RNA-dependent RNA polymerase RLU relative light unit RRL rabbit reticulocyte lysate rRNA ribosomal RNA RT-PCR reverse transcription polymerase chain reaction Ruc Renilla luciferase encoding gene RUC Renilla luciferase protein SeMV Sesbania mosaic virus SBMV Southern bean mosaic virus sgRNA subgenomic RNA ssRNA single-stranded RNA Tat trans-acting transcriptional activator TAV translational transactivator/viroplasmin TBP TATA-binding protein 3’TE 3’ translation element TEV Tobacco etch virus TLS tRNA-like structure TMV Tobacco mosaic virus TMVΩ Tobacco mosaic virus 5’UTR tRNA transfer RNA TS test sequence T-site arriving aa-tRNA binding site TYMV Turnip yellow mosaic virus UAS upstream activator sequence uidA β-glucuronidase encoding gene uORF upstream open reading frame UNR upstream of NRAS UTR untranslated region VDS vector-derived intercistronic sequence VPg viral protein, genome-linked WGE wheat germ extract wt wild type XRN1 5’-3’exoribonuclease 1

Amino acids may be described by one or three letter code

Abstract

ABSTRACT

The present study focuses on the via programmed -1 translational strategies of Cocksfoot mottle (-1 PRF). Two signals in the mRNA at virus (CfMV, genus Sobemovirus), which the beginning of the overlap program infects monocotyledonous plants. CfMV approximately every fi fth ribosome to slip RNA lacks the 5’cap and the 3’poly(A) one nucleotide backwards and continue tail that ensure effi cient translation of translation in the new -1 frame. This leads cellular messenger (mRNAs). to the production of C-terminally extended Instead, CfMV RNA is covalently linked polyprotein, which encodes the viral RNA- to a viral protein VPg (viral protein, dependent RNA polymerase (RdRp). The genome-linked). This indicates that the -1 PRF event in CfMV was very effi cient, viral untranslated regions (UTRs) must even though it was programmed by a functionally compensate for the lack of simple stem-loop structure instead of a the cap and poly(A) tail. We examined pseudoknot, which is usually required for the effi cacy of translation initiation in high -1 PRF frequencies. Interestingly, CfMV by comparing it to well-studied regions surrounding the -1 PRF signals viral translational enhancers. Although improved the -1 PRF frequencies. Viral insertion of the CfMV 5’UTR (CfMVε) protein P27 inhibited the -1 PRF event in into plant expression vectors improved vivo, putatively by binding to the -1 PRF gene expression in barley more than the site. This suggested that P27 could regulate other translational enhancers examined, the occurrence of -1 PRF. studies at the RNA level showed that Initiation of viral replication requires CfMVε alone or in combination with the that viral proteins are released from the CfMV 3'UTR did not provide the RNAs polyprotein. This is catalyzed by viral translational advantage. Mutation analysis serine protease, which is also encoded revealed that translation initiation from from the polyprotein. N-terminal amino CfMVε involved scanning. Interestingly, acid sequencing of CfMV VPg revealed CfMVε also promoted translation that the junction of the protease and VPg initiation from an intercistronic position of was cleaved between glutamate (E) and dicistronic mRNAs in vitro. Furthermore, asparagine (N) residues. This suggested internal initiation occurred with similar that the processing sites used in CfMV effi cacy in translation lysates that had differ from the glutamate and serine (S) reduced concentrations of eukaryotic or threonine (T) sites utilized in other initiation factor (eIF) 4E, suggesting that sobemoviruses. However, further analysis initiation was independent of the eIF4E. In revealed that the E/S and E/T sites may contrast, reduced translation in the eIF4G- be used to cleave out some of the CfMV depleted lysates indicated that translation proteins. from internally positioned CfMVε was eIF4G-dependent. After successful translation initiation, leaky scanning brings the to the second open reading frame (ORF). The CfMV polyprotein is produced from this and the following overlapping ORF

1 Introduction

1. INTRODUCTION

The majority of viruses have positive- strategies that make use of the advantages sense (+) single-stranded RNA (ssRNA) and limitations inherent within the cap- genomes. The (+)ssRNA viruses replicate dependent translation of cellular mRNAs in the cytoplasm of their host cells. Thus, and enable effi cient translation of viral they cannot utilize the host machinery mRNAs (reviewed by Gallie, 1996, Gale et for regulation of their gene expression at al., 2000). The small-sized genomes of the the transcriptional level and translational (+)ssRNA viruses limit their capacity to control is the most important tool for gene encode proteins by conventional strategies, expression regulation. Cellular enzymes but by using various translational strategies cannot copy RNA templates and viruses simultaneously, viruses can increase their encode their own RNA-dependent RNA coding capacity and multiply the steps that polymerases (RdRps) for the job. These can be regulated. The (+)ssRNA viruses enzymes lack proofreading capacity may also use additional gene expression and viral genomes have to be small strategies to avoid polycistronic mRNAs. to maintain low levels of detrimental Viral genomes may be divided into mutations. Because of the size limitations, segments (multipartite genomes), in which the viruses do not have the capacity the individual RNAs are monocistronic. to encode translational machineries of Alternatively, viral RdRps may synthesize their own. Consequently, the (+)ssRNA (+)-sense subgenomic RNAs (sgRNAs) viruses are totally dependent on the host that lack the 5’terminal open reading machinery. Eukaryotic mRNAs are usually frames (ORFs) of the polycistronic viral monocistronic, capped at their 5’ends, and mRNAs (reviewed by Bustamante and polyadenylated at their 3’termini. These Hull, 1998). structures both stabilize the mRNAs and In this thesis the main focus was synergetically affect translation effi cacy on gene expression of Cocksfoot mottle (Gallie et al., 1991, reviewed by Gallie, virus (CfMV), which is a member of the 1998). In contrast, viral genomes are often Sobemovirus genus (Mäkinen et al., 1995a). polycistronic and many viral mRNAs Sobemoviruses have small (+)ssRNA lack the 5’cap, the poly(A) tail, or both. genomes of approximately 4000 to 4500 Thus, one may believe that viral mRNAs nucleotides (nts) that are packed into small are poorly translated. However, viruses icosahedral particles (reviewed by Tamm have evolved multiple translational and Truve, 2000a). According to their

Fig. 1. Genome organization and translation strategy of CfMV.

2 Introduction putative RdRp sequences, sobemoviruses apart from translation. As a result of all this, are closely related to poleroviruses. at least 11 eIFs are involved in eukaryotic Sobemoviruses infect monocotyledonous protein synthesis initiation, whereas for and dicotyledonous plants, but usually three initiation factors (IF1, 2, the host range is narrow. The viruses are and 3) are suffi cient (reviewed by Kapp and transmitted by insect vectors, seeds, or by Lorsch, 2004). For example, eIFs 4A, 4B, mechanical inoculation. The type member 4E, 4G, and 4F do not have counterparts is Southern bean mosaic virus (SBMV). in prokaryotes, due to the differences CfMV infects some grass species, such as in ribosome recruitment to the mRNAs. cocksfoot (Dactylis glomerata L.), but also Although translation initiation is similar cereals including wheat, oat, and barley. in all , there are differences The CfMV RNA encodes four ORFs, of in the eIFs of mammalian, plant, and which ORFs 2A, 2B, and 3 overlap (Fig. 1) yeast cells (Browning, 2004). Most eIFs (Mäkinen et al., 1995a). The 5’end of the are composed of several subunits and RNA lacks the 5’cap and the 3’terminus individual eIFs form multiple interactions the poly(A) tail, suggesting that translation with other eIFs, ribosomal subunits, initiation occurs via a cap-independent and mRNA. The interactions stabilize mechanism. The translation initiation and induce conformational changes in codon of ORF1 is in poor context, which the eIFs and result in the colocalization allows the second ORF encoding the of the factors and coordination of their polyprotein to be produced by ribosomes binding and release (reviewed by Kapp that pass the fi rst AUG codon (Mäkinen and Lorsch, 2004). Whether the binding et al., 1995a). In most sobemoviruses of eIFs to mRNA occurs individually or the polyprotein is encoded from a large as preformed multicomponent complexes continuous ORF (reviewed by Tamm and remains unclear. The exact order and Truve, 2000a). In CfMV the polyprotein mechanism of most binding and release is arranged into two overlapping parts events are also still largely unknown. (ORF2A and ORF2B). The synthesis The majority of cellular mRNAs have of ORF2B encoding the viral RdRp is a 5’terminal 7-methylguanosine (m7G) regulated via -1 programmed ribosomal cap structure. This structure recruits eIFs frameshifting (-1 PRF) (Mäkinen et al., 4E, 4A, and 4G, which together form 1995b). The last ORF3 encodes the coat an eIF4F complex (Grifo et al., 1983). protein (CP), which is produced from Plants have two eIF4E isoforms: eIFiso4F sgRNA (Mäkinen et al., 1995a). (eIFs 4A, iso4G, and iso4E) promotes translation preferably from unstructured mRNAs, whereas eIF4F (eIFs 4A, 4G, 1.1 TRANSLATION INITIATION and 4E) also mediates translation from Initiation is usually the rate-limiting step structured, uncapped, and polycistronic of translation and is facilitated by soluble mRNAs (Gallie and Browning, 2001). The cytoplasmic eukaryotic initiation factors 5’cap structure is specifi cally recognized (eIFs) that prepare mRNA for the 40S by eIF4E (or eIFiso4E) via a network of subunit binding, assist in AUG selection, interactions, the most important being base and promote 60S subunit binding (Fig. stacking of the positively charged m7G 2). Eukaryotic genomes are complex, and between two electron-rich tryptophans transcription is spatially and temporally set in eIF4E (Marcotrigiano et al., 1997).

3 Introduction

The eIF4G plays an important role in 1985). The eIF4A function is stimulated by translation initiation by functioning as a eIF4B and eIF4F (Ray et al., 1985, Pestova scaffolding protein that binds mRNA and and Kolupaeva, 2002). The eIF4G-PABP interacts with the poly(A) binding protein interaction enhances translation putatively (PABP) and with several eIFs (Lamphear by bringing the termini of mRNA together, et al., 1995). Simultaneous interaction which enables the recycling of ribosomes of eIF4G with eIF4E and eIF4A escorts back to the 5’termini (Fig. 3) (Gallie, eIF4A to the 5’end of mRNA. The 1991, Gallie and Tanguay, 1994, Preiss and adenosine triphosphatase (ATPase)/ Hentze, 1998, reviewed by Gallie, 2002b). helicase activity of eIF4A then renders This interaction could thus prevent the the cap-proximal region accessible to the synthesis of N-teminally truncated proteins incoming ribosomal subunit (Ray et al., (Preiss and Hentze, 1998). The eIF4G-

Fig. 2. Schematic representation of the recruitment of eIFs to mRNA during translation initiation.

4 Introduction

PABP interaction may also function as a the 43S preinitiation complex (Benne quality control mechanism for dissecting et al., 1976, Pestova et al., 2001). In a mRNAs, which are truncated at their manner similar to that of eIF4G, eIF3 also 3’ends, as judged by their inability to plays a scaffolding role during translation bind PABP. Finally, the eIF4G-mediated initiation. It binds to the 40S subunits and binding of PABP to eIF4E stabilizes the together with eIF1 and 1A enhances the cap interaction and protects it against binding of the ternary complex close to decapping and degradation (Gallie, 1991, the ribosomal peptidyl-tRNA binding site reviewed by Gallie, 1998). (P-site, see Fig. 5B) (Benne and Hershey, The initiator methionine transfer 1978, Pestova et al., 1998, Majumdar et

RNA (Met-tRNAi) is selected from a pool al., 2003). Furthermore, simultaneous of tRNAs by eIF2 bound to guanosine interaction of eIF3 with the cap-bound triphosphate (GTP). Discrimination eIF4G brings the 43S complex and the against the tRNAs used for elongation 5’end of mRNA together (Lamphear et al., occurs on the basis of certain specifi c 1995). bases, base pairings, and modifi cations in The initiation codon is located in a the Met-tRNAi (Åström et al., 1993). The scanning process, during which the 43S

Met-tRNAi and eIF2-GTP form a stable complex migrates along the mRNA from ternary complex, which together with the the very 5’end towards the 3’end (Kozak, 40S subunit and eIFs 1, 1A, and 3 form 1989). The scanning-competent form is

Fig. 3. The closed-loop model of translation. In plants the two termini are brought together via PABP-eIF4G and PABP-eIF4B interactions. In mammalian cells, circularization is mediated by PABP-interacting protein (Paip), which interacts with PABP, eIF4A and eIF4B. Adapted from Gallie, 2002b.

5 Introduction achieved via binding of eIFs 1 and 1A certain eIFs. In these cases, the ribosomes to the 40S subunit (Pestova et al., 1998). are directly bound to sequence elements The forthcoming secondary structures that are located near the initiation codons are unwound by the helicase activity (reviewed by Martinez-Salas et al., 2001). present in the eIF4F complex (Ray et These possibilities are discussed in the al., 1985, Kozak, 1989, Pestova and following sections. Kolupaeva, 2002). Scanning continues until the initiation codon is recognized at 1.1.1 Regulation of translation initiation the ribosomal P-site. Correct base-pairing during virus infection between the Met-tRNAi anticodon and Many of the eIFs are present in low AUG results in a 48S complex formation amounts and thus their function is tightly and induces eIF5-catalyzed hydrolysis of regulated via specifi c responses to external eIF2-GTP (Pestova et al., 2000, Unbehaun and internal stimuli (reviewed by Dever, et al., 2004). Thus, eIF5 controls the 1999). Most of the regulation occurs fi delity of initiation by controlling the GTP via phosphorylation, which changes the hydrolysis. The accuracy of initiation codon affi nities of the eIFs towards other eIFs or selection is also regulated by eIF1 and 1A, mRNA. The eIFs 2, 2B, 3, 4A, 4B, 4G, 4E, which recognize and destabilize aberrant 5, and PABP are at least phosphoproteins preinitiation complexes and discriminate (Gallie et al., 1997, Kapp and Lorsch, between poor and good initiation codon 2004). contexts (Pestova et al., 1998, Pestova and Mammalian cells respond to viral Kolupaeva, 2002, Unbehaun et al., 2004). infections in a manner similar to other The AUG selection fi xes the reading stressful conditions by shutting off their frame and eIF2-guanosine diphosphate protein synthesis (reviewed by Gale et al., (GDP) is released (Pestova et al., 2000). 2000). By this means cells try to prevent The remaining mRNA-bound complex is virus multiplication. The down-modulation stabilized by eIF3 (Unbehaun et al., 2004). of protein synthesis occurs largely via The 60S subunit joining triggers the the regulation of eIF2 phosphorylation. hydrolysis of eIF5B-GTP and the release Viral double-stranded (ds) replication of eIF5B-GDP (Pestova et al., 2000) and is intermediates activate dsRNA-activated accompanied by the dissociation of eIFs 1, kinases (PKR), which phosphorylate 1A, and 3 (Unbehaun et al., 2004). Finally, the eIF2α subunit (Crum et al., 1988, an empty ribosomal aminoacyl-tRNA (aa- Bilgin et al., 2003, reviewed by Gale et tRNA) binding site (A-site) is ready to al., 2000). Phosphorylation stabilizes the accept the fi rst incoming aa-tRNA. interaction between eIF2 and eIF2B and A complete set of eIFs is not always blocks the recycling of eIF2-GDP to eIF2- required for translation initiation. For GTP, which results in cessation of protein example, the 5’leaders with unstructured synthesis (Pavitt et al., 1998). However, leaders can be translated in the absence of many animal viruses encode RNAs or eIFs having helicase activity (Sonenberg proteins that prevent PKR from activating et al., 1982, Jobling and Gherke, 1987, and functioning (reviewed by Gale et al., Browning et al., 1988, Pestova and 2000). Although not as much is known on Kolupaeva, 2002). Interestingly, some viral plant viruses, Tobacco mosaic virus (TMV, mRNAs with highly structured 5’leaders genus Tobamovirus) infection induces may also be translated in the absence of PKR-like activity in plants (Crum et al.,

6 Introduction

1988). Furthermore, successful tobamo- may disrupt transcription, capping, and and potyvirus infections are dependent on export of host mRNAs or affect translation cellular protein P58IPK that inhibits PKR initiation directly via modulation of eIFs activity (Bilgin et al., 2003). (reviewed by Gale et al., 2000). For Cap recognition is also tightly instance, picornavirus infections render regulated. Phosphorylation increases 4E-BP to the hypophosphorylated state eIF4E’s affi nity for the cap and eIF4G (Bu (Gingras et al., 1996). Some picornaviruses et al., 1993). In mammalian cells, eIF4G impair cap recognition by modulating interacts with a kinase that regulates the eIF4F function via proteolytic cleavage of eIF4E activity (Pyronnet et al., 1999). eIF4G and PABP (Sonenberg et al., 1982, However, plant eIF4G and eIFiso4G lack Lamphear et al., 1995, Kerekatte et al., the motif for the kinase binding (Browning, 1999). 2004). In animal cells, phosphorylation of eIF4 is suppressed by binding of an eIF4E- 1.1.2 Plant viral translational binding protein (4E-BP) (Whalen et al., enhancers 1996), whereas eIF4E-4E-BP association Several characteristics of the 5’untranslated is prevented by phosphorylation of 4E- regions (5’UTRs) affect their capacity BP (Lin et al., 1994, Gingras et al., 1996). to mediate effi cient translation (Fig. 4). The 4E-BP phosphorylation correlates Usually the 5’UTRs of cellular mRNAs with translational stimulation, whereas as well as many plant viral leaders dephosphorylation occurs during cellular are relatively AU-rich and lack strong stress such as viral infection or extreme secondary structures (Sonenberg et al., temperatures (Dever, 1999). No 4E-BP 1982, Gallie, 1996, Kozak, 2003). This homologs have been found from plants, is benefi cial, since simple 5’UTRs may suggesting that plants may regulate be translated with an incomplete set of protein synthesis by different mechanisms eIFs, because there is no need to unwind (Browning, 2004). Since translation of secondary structures (Sonenberg et al., cellular mRNAs is highly cap-dependent, 1982, Gallie and Browning, 2001, Pestova inhibition of cap-dependent translation and Kolupaeva, 2002). Viral 5’UTRs may is a good target for viruses that use cap- also be long and highly structured, due independent translation strategies. Viruses to their involvement in the regulation of

Fig. 4. Factors affecting the effi cacy of translation.

7 Introduction mRNA stability and localization as well as viral leaders function as translational virus replication and assembly. However, enhancers and effi cient translation of viral highly structured 5’leaders impede mRNAs may be achieved via production scanning of the ribosomal subunits (Kozak, of large amounts of viral mRNAs, which 1989). Therefore, initiation at these outcompete the cellular mRNAs for mRNAs requires the participation of all translation machinery. eIFs (Kozak 1989, Pestova and Kolupaeva, The 5’UTR of TMV (TMVΩ; genus 2002), unless scanning is replaced by the Tobamovirus) is one of the best-studied use of alternative initiation mechanisms examples of translational enhancers. It such as internal ribosomal entry. enhances translation in prokaryotes and Certain plant viruses, such as eukaryotes, albeit the mechanism probably alfamo-, tobamo-, and potexviruses, have differs (Gallie et al., 1987, Sleat et al., capped RNAs. In contrast, in some plant 1988, Tanguay and Gallie, 1996, Wells et viruses a viral genome-linked protein al., 1998, Gallie, 2002a). The (+)ssRNA (VPg) replaces the 5’cap. This group of of TMV is capped but the enhancement viruses includes polero-, sobemo-, and function of the 68-nt-long 5’UTR is picornavirus-like viruses. There are also cap-independent (Sleat et al., 1988). In viruses, such as luteo- and tombusviruses, plants, translation initiation is stimulated which do not have specifi c 5’terminal via binding of a host heat-shock protein, structures. In addition to the poly(A) Hsp101, to the poly(CAA) sequence tails found for instance in the 3’end of of TMVΩ (Tanguay and Gallie, 1996). picornavirus-like viruses, the genomes of Hsp101 further interacts with eIF4G and the (+)ssRNA viruses may terminate at eIF3, which stimulates the recruitment of tRNA-like structures (TLSs) as in tobamo- 40S subunits to the RNA (Wells et al., 1998, and tymoviruses or at heteropolymeric Gallie, 2002a). The TMV 3’UTR contains sequences not forming TLSs as in sobemo- a TLS and an upstream pseudoknot region. and luteoviruses (reviewed by Dreher et al., The TMV 3’UTR boosts the translational 1999). Although many (+)ssRNA viruses enhancement conferred by the Ω element lack either the 5’cap, the poly(A) tail, or (Gallie, 2002a). Hsp101 also interacts both, viral RNAs are effi ciently translated, with the 3’UTR (Tanguay and Gallie, indicating that the viral replacements can 1996). Thus, further stimulation of functionally complement the 5’cap and the translation via the 3’UTR probably results 3’poly(A) tail. In fact, many viral UTRs from the circularization of TMV RNA enhance translation initiation by attracting and recycling of translational components ribosomal subunits, eIFs, or some other from the 3’termini to the 5’end in a manner host proteins aiding in translation initiation similar to that proposed for the capped and (reviewed by Gallie, 1996). Translational polyadenylated cellular mRNAs (reviewed enhancers have been identifi ed from both by Gallie, 1998). termini of several genera of the (+)ssRNA The 5’UTRs from potyviruses also plant viruses. The heterogeneity in enhance gene expression (Nicolaisen et these UTRs indicates that translation al., 1992, Levis and Astier-Manifacier, enhancement is accomplished in several 1993, Gallie et al., 1995). The potyviruses ways. Depending on the enhancer, the belong to the Picornavirus supergroup. functional range can be narrow or broad As in other group members, potyviral (Gallie et al., 1987). However, not all (+)ssRNA genomes are polyadenylated

8 Introduction and linked to VPg at their 5’ends. The binding of AMV CP to the viral 3’UTR mechanism of translational enhancement stimulates translation (Neeleman et al., conferred by the Tobacco etch virus 2001). The AMV CP interacts with eIF4G (TEV) 5’UTR (143 nt) is the best studied. and eIFiso4G and thus appears to mimic The 5’UTR folds into two pseudoknots, of the action of PABP in circularization of which the 5’proximal pseudoknot is crucial mRNA (Krab et al., 2005). for cap-independent translation initiation In many viruses the translation (Zeenko and Gallie, 2005). Capping does regulatory element is located in the 3’UTR not improve translation from the TEV instead of 5’UTR. This arrangement 5’leader, showing that the 5’UTR function prevents translation initiation on mRNAs overlaps with that of the 5’cap (Gallie having truncated 3’UTRs. The cap- et al., 1995). However, enhancement is independent translation element (3’TE) of strongly stimulated by the polyadenylated Barley yellow dwarf virus (BYDV; genus 3’terminus. The TEV 5’UTR effi ciently Luteovirus) functionally mimics the cap recruits eIF4G to the mRNA, which gives structure. The 3’TE effi ciently competes the virus a competitive advantage (Gallie, for the eIF4F complex, which is putatively 2001). PABP boosts TEV RNA translation delivered to the uncapped 5’leader via the further, putatively via the contact formed kissing loop structure formed between the between the 5’and 3’termini as a result of complementary bases of the stem-loops eIF4G-PABP interaction (Gallie, 2001). present in the UTRs (Wang et al., 1997, The potyviral 5’leaders also function at Guo et al., 2001). The 5’leader is then the internal positions of polycistronic scanned by the ribosomal subunits in a mRNAs, providing further evidence for the conventional manner (Wang et al., 1997, occurrence of cap-independent translation Guo et al., 2001). The TLS in the 3’UTR initiation from these viral 5’UTRs (Levis of Turnip yellow mosaic virus (TYMV; and Astier-Manifacier 1993, Niepel and genus Tymovirus) stimulates translation Gallie, 1999a, Gallie 2001, Akbergenov via a totally different but still unknown et al., 2004). Interestingly, potyviral VPgs mechanism that however appears to be interact with eIF4E, eIFiso4E, and PABP, cap-dependent and to involve specifi c suggesting that VPg may participitate in recognition of the aminoacylated TLS by a translation initiation (Wittmann et al., eukaryotic elongation factor 1A (eEF1A) 1997, Léonard et al., 2004). (Matsuda and Dreher, 2004). The sgRNA (RNA4) encoding the CP of Alfalfa mosaic virus (AMV; genus 1.1.3 Internal ribosome entry sites Alfalfamovirus) is effi ciently translated (IRESs) under conditions in which cap-dependent Prokaryotic ribosomes are bound to translation initiation is impaired mRNAs via base-pairing of Shine- (Sonenberg et al., 1982). This suggested Dalgarno sequences with 16S ribosomal that translation from this simple AU-rich RNA (rRNA). Sequences complementary leader could occur with an incomplete set to the rRNA are also frequently encountered of eIFs. Consistent with this hypothesis, in eukaroytic and viral mRNA 5’UTRs the AMV RNA4 leader decreases the (Smirnyagina et al., 1991, Nicolaisen et al., amounts of eIFs 3, 4A, 4E, and 4G needed 1992, Mauro and Edelman 1997, Wang et for translation initiation (Browning et al., al., 1997, Niepel and Gallie, 1999, Koh et 1988). Recently, it was found that the al., 2003, Akbergenov et al., 2004), which

9 Introduction suggests that direct contact of the mRNA leafroll virus (PLRV, genus Polerovirus) and the translational apparatus could (Jaag et al., 2003). Although the 5’leader be used to induce initiation (Mauro and of TEV is more complex, i.e. contains Edelman, 1997, Akbergenov et al., 2004). two pseudoknot domains, only the fi rst Some evidence supports this hypothesis, pseudoknot and a region complementary e.g. the 40S subunits have high affi nity for to 18S rRNA in it, are important for IRES mRNAs, which are complementary to the activity (Zeenko and Gallie, 2005). In a exposed regions of 18S rRNA (Akbergenov manner similar to that for cap-dependent et al., 2004). Furthermore, mutations translation initiation, IRES-mediated in these complementary regions reduce translation initiation may be stimulated translation effi ciencies of the particular via interaction of the IRES with the mRNAs (Zeenko and Gallie, 2005). 3’terminus (Svitkin et al., 2001, Koh et Among the DNA viruses, such as al., 2003, reviewed by Martinez-Salas et Caulifl ower mosaic virus (CaMV, genus al., 2001). Caulimovirus), stable secondary structures may be bypassed by `jumping´ of the 1.1.3.1 Mechanism of IRES-mediated migrating ribosomes over the complex translation initiation regions (Ryabova and Hohn, 2000). During IRES-mediated translation However, (+)ssRNA viruses utilize internal initiation, ribosomes attach directly to binding of the 40S subunits to regions called the initiation codon or a short distance internal ribosome entry sites (IRESs). upstream of it. Since cap recognition or Internal initiation was fi rst described in the scanning is not required, certain eIFs picornaviruses, but IRESs were later found may be unnecessary for initiation. This from several other virus groups as well as improves the competitiveness of the from some cellular mRNAs (reviewed by viral mRNAs against cellular mRNA Martinez-Salas et al., 2001). Among the translation. In general, IRESs do not share animal viruses, IRES-mediated translation conserved structural requirements, which initiation is characteristic of viruses suggests that the mechanisms of ribosome containing complex and long 5’leaders. recruitment vary. The requirements for Thus, using internal initiation in addition IRES-mediated initiation have been studied to avoiding the need to scan the 5’UTRs, mainly in reconstituted in vitro assays viruses circumvent the inhibitory effect of (Pestova et al., 1996, 2001, Martinez-Salas the upstream ORFs (uORFs) often present et al., 2001), as illustrated in the following in long leaders. However, the IRESs examples. characterized from plant viruses are rather Picornaviral translation initiation simple (Ivanov et al., 1997, Koh et al., is dependent on all eIFs, except the cap- 2001, Jaag et al., 2003). The 148 nt IRES recognizing factor eIF4E (Pestova et al., element of Crucifer-infecting tobamovirus 1996). Thus, translation of viral mRNAs (CrTMV; genus Tobamovirus) contains is not affected by regulation of eIF4E or two putative stem-loops but the element 4E-BP activities (Pestova et al., 1996). important for IRES activity resides in the Furthermore, cap independency allows unstructured GA-rich region (Ivanov et picornaviruses to disrupt translation al., 1997, Dorokhov et al., 2002). A similar of cellular mRNAs by impairing cap purine-rich region is important for IRES- recognition using proteolytic cleavage of mediated translation initiation in Potato eIF4G. Cleavage separates the C-terminal

10 Introduction binding domains of eIF3 and eIF4A from 1.1.3.2 IRES trans-acting factors the N-terminal binding domains of eIF4E (ITAFs) and PABP (Lamphear et al., 1995). The Several IRESs interact with proteins that C-terminal fragment binds to the IRES, do not have previously identifi ed roles recruits other eIFs, and enables translation in translation initiation. For instance, initiation from the picornaviral RNAs in picornaviral IRESs interact with the absence of eIF4E (Lamphear et al., polypyrimidine tract binding protein 1995, Pestova et al., 1996). (PTB), proliferation-associated factor

The binding of Hepatitis C virus ITAF45, lupus antigen (La), poly(rC) (HCV, genus Hepacivirus) IRES to the binding protein (PCBP), and upstream of 40S subunits occurs in the absence of NRAS (UNR) (reviewed by Martinez- eIFs. The binding causes conformational Salas et al., 2001). In a survey of HCV changes in the 40S subunit that place the IRES-interacting proteins, approximately initiation codon directly at the ribosomal 90 proteins were shown to bind to this P-site (Spahn et al., 2001). Subsequent 48S IRES, of which ~20 were specifi c for complex formation occurs after correct the HCV sequence (Lu et al., 2004). In codon-anticodon recognition between the addition to translation-related proteins,

Met-tRNAi-eIF2-GTP ternary complex these proteins included RNA binding and AUG (Pestova et al., 1998). Finally, proteins, cytoskeletal proteins, as well as eIF3 is recruited to assist in the 60S subunit proteins involved in signal transduction, joining (Pestova et al., 1998). apoptosis, cell differentiation, and cell Cricket paralysis virus (CrPV, genus cycle regulation. The most popular Cricket paralysis -like viruses) promotes model for ITAF function is that they act internal initiation in the absence of any as chaperones that direct and stabilize eIFs, Met-tRNAi, or GTP hydrolysis folding of the IRESs, thus enabling the (Wilson et al., 2000). Translation initiates subsequent recruitment of ribosomal by incorporation of the Ala-tRNA into the subunits. Consequently, variation in the ribosomal A-site while the P-site remains ITAF content could explain the cell- empty. Thus, translation appears to initiate type-specifi c restriction of certain viruses directly from the elongation phase (Wilson (reviewed by Martinez-Salas et al., 2001). et al., 2000). Structural data show that the However, concurrent agreement of the CrPV IRES forms specifi c intermolecular role played by ITAFs is lacking, since contacts with both ribosomal subunits and many ITAFs are nuclear proteins, whereas introduces conformational changes in them viral RNAs are located in the cytoplasm (Spahn et al., 2004). The conformation of (Kozak, 2001, 2003). In fact, many of the the IRES also changes during the initiation ITAFs have previously characterized roles process. The ability of CrPV IRES to in regulation of RNA stability, splicing, modify the conformation of the ribosome or transcription. Thus, rather than being while its conformation is also changed is a involved in translation directly, ITAFs characteristic similar to protein translation could be involved in generating templates factors. Therefore, CrPV IRES appears for the translational apparatus (Kozak, to function as an RNA-based translation 2003). In fact, several sequences initially initiation factor. identifi ed as IRESs were later shown to contain cryptic promoters or splicing sites (Han and Zhang, 2002, Hecht et al., 2002,

11 Introduction

Van Eden et al., 2004, Verge et al., 2004). are too closely located at the 5’end of Thus, the involvement of the ITAFs in the mRNAs (Kozak, 1989, Pestova and IRES-mediated translation initiation is Kolupaeva, 2002). Actually, very few confusing. (+)ssRNA viral mRNAs have the fi rst AUG in optimal context. Leaky scanning 1.1.4 Leaky scanning may also be used to produce two isoforms Since the 5’UTRs are scanned from the of a single protein (reviewed by Kozak, very beginning of the 5’ends, usually the 2002). The effi cacy of initiation on the fi rst initiation codon is used for translation downstream AUG again is dependent on initiation (Kozak, 1989). However, the the sequence context. AUG-bypassing sequence surrounding the AUG determines results in the down-regulated expression the effi cacy of translation initiation. The of the fi rst cistron. On the other hand, eIF1 participates in the discrimination translation initiation from a good-context between favourable and poor context AUG reduces the protein yield produced initiation codons (Pestova and Kolupaeva, from the downstream AUGs. Thus, leaky 2002). The optimal sequence contexts for scanning can be used to regulate protein AUG recognition vary among eukaryotic production. However, only two or rarely species and cell types, but the purine at three sequential proteins are usually the –3 position (with the A of AUG as produced via leaky scanning (Kozak, 0) is highly conserved (Kozak, 1989, 2002). In CfMV, the initiation codon Lukaszewich et al., 2000). In mammalian context of ORF1 contains a pyrimidine cells the –3 purine residue stimulates at the –3 position and lacks the G residue effi cient translation initiation even if at position +4 (Mäkinen et al., 1995a). the rest of the consensus sequence is The initiation codon context of ORF2A is imperfect (Kozak, 1989), whereas in plant in better context, because it contains the cells the –1 and –2 position residues are purine at the –3 position (Mäkinen et al., also important (Lukaszewich et al., 2000). 1995a). Therefore, leaky scanning down- In some rare cases, translation may initiate regulates ORF1 expression and enables from a nonconventional ACG or CUG the trailing viral polyprotein to be encoded codons, but then the rest of the codon from the same mRNA. context must be optimal (Kozak, 1989, 2002). Initiation at non-AUG codons is 1.1.5 Reinitiation often observed among mRNAs whose The uORFs usually limit translation of the 5’UTRs have high GC contents and strong downstream ORFs by making it dependent secondary structures. Slow scanning may on leaky scanning or reinitiation (reviewed increase the time the mismatched codons by Kozak, 2002). In contrast to leaky can base-pair with Met-tRNAi (Kozak, scanning, both the uORF and the following 2002). ORF become translated in reinitiation Since translation initiation is usually during the same round of protein restricted to the fi rst AUG, downstream synthesis. The factor requirements for ORFs are usually silent. However, leaky reinitiation are not yet fully known but re- scanning through nonoptimal AUGs recruitment of the eIF2-GTP-Met-tRNAi provides viruses one mechanism for ternary complex appears to be a necessity. translating their polycistronic mRNAs. It is hypothesized that translation of short AUGs may also be bypassed, if they uORFs would not lead to dissociation of

12 Introduction all eIFs and that the remaining eIFs could truncations made to the 5’terminal part then enable scanning and reinitiation during their synthesis by viral RdRPs (Kozak, 1987). This is supported by the (reviewed by Miller and Koev, 2000, fact that reinitiation effi ciency usually Kozak, 2002). SgRNAs are subjected to the decreases as the uORFs become larger. same translational control as any mRNAs. However, recent data indicate that the The sgRNAs may encode overlapping duration of the elongation phase rather ORFs and viruses may encode several than the length of the translation product sgRNAs. A large percentage of (+)ssRNA determines the feasibility of reinitiation viruses synthesize sgRNAs (Miller and (Kozak, 2001). On the other hand, the Koev, 2000). Usually the sgRNAs encode effi ciency of reinitiation improves as the structural and movement proteins, e.g. the distance to the uORF increases, probably CfMV CP is putatively encoded from a because there is more time to bind Met- 1.2-kb sgRNA that is produced from the tRNAi (Kozak, 1987). 3’proximal part of the genome (Mäkinen Only about 10% of the eukaryotic et al., 1995a). BYDV produces three mRNAs have uORFs, whereas among sgRNAs: sgRNA1 encodes the structural viral and prokaryotic mRNAs they are proteins, whereas sgRNA2 and 3 do not more common (Ryabova and Hohn, encode any proteins. Both the gRNA and 2000). Although ineffi cient reinitiation the sgRNA2 contain the 3’TE element, can be benefi cial in expression of toxic thus, the sgRNA2 also sequesters eIF4F proteins (reviewed by Kozak, 2002), and eIF4iso4F via the 3’TE. This inhibits some viruses have evolved to overcome translation of the gRNA and renders it the ineffi ciency by encoding proteins that available for replication and encapsidation stimulate reinitiation. CaMV encodes (Shen and Miller, 2004). Thus, sgRNAs the translational transactivator protein may also be utilized as riboregulators to (TAV), which interacts with eIF3 and the regulate translation. 60S subunit (Park et al., 2001). These interactions keep the eIF3 associated 1.2 TRANSLATION ELONGATION with the elongating ribosomes. After The elongation step is highly conserved termination eIF3 can re-recruit the ternary across the three kingdoms of life (reviewed complex, resume scanning, and reinitiate by Kapp and Lorsch, 2004). During at the downstream AUG. elongation the initiator Met-tRNAi at the P-site of the ribosome is elongated to a 1.1.6 SgRNAs polypeptide via subsequent addition of aa- Several (+)ssRNA viruses have tRNAs (Fig. 5). The incoming aa-tRNAs 3’proximal genes that are not expressed are bound to the A-site. The A- and the from the genomic RNA (gRNA), since P-site tRNAs interact with both subunits leaky scanning or reinitiation at the of the ribosomes (Moazed and Noller, 3’proximal part of the genome would be 1989a). The anticodon binding sites are too ineffi cient processes to produce the located in the small subunit, whereas the proteins in adequate amounts. These silent aa-tRNA acceptor ends (tRNA 3’ends) 3’cistrons can be made accessible to the are located close to each other within the translational apparatus by synthesizing large ribosomal subunit. The used acylated sgRNAs in which the translation initiation tRNAs bound at the exit site (E-site) codon is brought closer to the 5’end via interact only with the large subunit via the

13 Introduction acceptor end. The binding and movement initial binding is rapid and reversible and of the tRNAs between the A- and P-sites does not involve codon recognition. The in the small subunit are uncoupled from aa-tRNA is initially placed in an A/T those occurring between the A, P, and hybrid state; the anticodon loop is bound E-sites of the large ribosomal subunit to the A-site in the small subunit and the (Moazed and Noller, 1989b, reviewed 3’end is bound to the large subunit T- by Wilson and Noller, 1998). Therefore, site via eEF1A. Interaction of the 3’end elongation generates hybrid states, in CCA sequence of aa-tRNA with eEF1A which the aminoacyl acceptor ends and prevents premature peptide bond synthesis the anticodon loops are differentially by impeding the access of the 3’terminus positioned between the small and large of the aa-tRNA to the peptidyl transferase subunit A, P, or E-sites. center of the large subunit (Moazed and The aa-tRNAs are transported as Noller, 1989a). Cognate tRNA binding a complex with GTP and eEF1A. The induces conformational changes in the

Fig. 5. A) The cloverleaf structure of tRNA. The activated amino acids are attached to the 3’- OH end of tRNA. B) Hybrid state model for the elongation cycle. 1) Initiator or peptidyl-tRNA is placed in the P-site (P/P-state). 2) The anticodon of the new incoming aa-tRNA interacts with the 40S subunit, whereas the acceptor end bound to eEF1A interacts with the 60S subunit at the T-site (A/T-state). 3) GTP hydrolysis releases eEF1A-GDP and the acceptor end also enters the A-site (A/A-state). 4) Following peptide bond formation, both tRNAs are in hybrid states. 5) Movement of the anticodon ends relative to the 40S subunit and as a result the peptidyl-tRNA is moved into the P/P-state and the deacylated tRNA into the E-state. Adapted from Wilson and and Noller, 1998.

14 Introduction small ribosomal subunit, which facilitate tRNA (reviewed by Joseph, 2003). During close contact and subsequent base-pairing translocation the interactions between the of the fi rst two bases of the codon and tRNA-mRNA complex and the ribosome anticodon (Ogle et al., 2001). During this are broken and renewed at a new position interaction the ribosome senses whether the three bases downstream towards the mRNA base-pairs have Watson-Crick geometry 3’end. This is probably achieved most and discrimination against near-cognate accurately by performing the translocation tRNAs occurs. In contrast, the binding site in a stepwise manner (Moazed and of the third wobble position base remains Noller, 1989b). The fi rst hybrid state relatively open and thus the geometry of the develops spontaneously after the peptide base-pairing is not as closely monitored. bond formation. The deacetylated tRNA This explains why certain noncanonical anticodon remains at the P-site, but the bases may occupy the wobble position acceptor stem is moved to the E-site (Ogle et al., 2001). Irreversible tRNA (P/E hybrid state). Simultaneously the selection occurs when a conformational peptidyl-tRNA´s anticodon is retained change activates the eEF1A-catalyzed at the A-site, whereas the acceptor stem hydrolysis of GTP. As a result, eEF1A- carrying the growing peptide is moved to GDP is released and the freed 3’end of the the P-site (A/P hybrid state) (Moazed and aa-tRNA can enter the large subunit A- Noller, 1989b). The next step is catalyzed site (A/A hybrid state). The acceptor ends by the hydrolysis of eEF2-GTP (Rodnina of the tRNAs become closely arranged et al., 1997). Movement of the anticodon with each other and with the peptidyl ends together with the mRNA relative transferase center of the large ribosomal to the small subunit places the peptidyl- subunit, which catalyzes the peptide bond tRNA in the P/P-state and the deacylated formation (Nissen et al., 2000). Peptide tRNA at the E-site (Moazed and Noller, bond formation is entirely catalyzed 1989b). The translocation machinery acts by rRNA. In this reaction a conserved directly on the tRNAs and the movement adenosine residue at the catalytic site of the mRNA is driven by its association attracts a proton from the α-amino group with the tRNAs. The ribosomes undergo of the amino acid moiety of the A-site aa- signifi cant conformational changes tRNA. This is facilitated by the active site during translocation (Joseph, 2003). environment, which increases the pKa of The movement of the ribosome by three the functional adenosine residue and thus nucleotides in the 3’direction also places its capacity to attract protons (Nissen et the next codon at the A-site. Finally, the al., 2000). Next, the nucleophilic α-amino deacetylated tRNA leaves the ribosome group attacks the electrophilic carbonyl via the E-site, and the eEF1A-GDP is carbon of the ester bond linking the peptide recycled back to eEF1A-GTP by eEF1B. moiety to the peptidyl-tRNA bound to the The aa-tRNA binding, transpeptidation, P-site. The resulting intermediate goes and translocation are repeated as many through rearrangements that fi nally yield times as there are sense codons in the a discharged tRNA bound to the P-site and mRNA. a one-amino acid-longer peptidyl-tRNA bound to the A-site. 1.2.1 Regulation of elongation Next, the translocation step vacates Elongation is the fastest step in translation the A-site for binding the following aa- and is not generally considered to play a

15 Introduction major role in the regulation of translation. or protein-membrane complexes, and However, elongation consumes a large cotranslational transmembrane transport. amount of energy, since at least four high- Translational pauses may be induced via energy bonds are used per each amino codons encoding rare aa-tRNAs as well acid addition (Browne and Proud, 2002). as via strong secondary RNA structures Therefore, it is advantageous for the cells (Tu et al., 1992, Somogyi et al., 1993). to be able to reduce the rate of protein Interaction of the nascent polypeptide synthesis during increased energy demand with itself or with another protein can or decreased energy supply so that the also inhibit elongation. For example, energy can be used for more important translation of proteins that are transported processes. The benefi t from inhibition at to the endoplasmic reticulum (ER) ceases the elongation phase is that the mRNAs after the leader peptide is synthesized remain polysome-associated (Browne and recognized by the signal recognition and Proud, 2002). This enables rapid particle (Wolin and Walter, 1988). resumption of translation after the shortage Translation commences after the complex in energy is overcome. is targeted to the ER. All eEFs are phosphoproteins and targets for various kinases and phos- 1.2.2 Programmed -1 ribosomal phatases. However, how phosphorylation frameshifting (-1 PRF) modulates eEF activities and thus protein PRF provides viruses an additional synthesis is not yet well understood mechanism for increasing the diversity (reviewed by Browne and Proud, 2002). of proteins produced from their small Phosphorylation of eEFs 1A and B by genomes. However, natural frameshift certain kinases appears to enhance their errors occur rarely, thus, certain cis-acting activity, whereas the reverse occurs signals have been built into the mRNAs, for the phosphorylated eEF2. Human which increase the frameshift frequency immunodefi ciency virus 1 (HIV-1; genus signifi cantly. Depending on the signals, Lentivirus) encodes a trans-acting some of the elongating ribosomes slip transcriptional activator (Tat) protein that either backwards (-1) or forwards (+1), interacts with eEF1B. This interaction after which translation continues in the reduces the translation of cellular mRNAs new overlapping ORF (Fig. 6). The but not viral mRNAs (Xiao et al., 1998). nonshifted ribosomes continue translation Several characteristics of the mRNAs in the original ORF (0-frame) (reviewed affect the rate of elongation (Fig. 4). As a by Farabaugh, 1996). Both products are result, ribosomes move along the mRNA at synthesized until the termination codons non-uniform rates and become distributed are encountered. As an outcome, two unevenly along the mRNA (Wolin and proteins are manufactured that share Walter, 1988). Occasionally the elongating identical N-termini but differ at their C- ribosomes stall and the trailing ribosomes terminal parts, beginning at the frameshift stack behind the leading ribosome (Wolin site. and Walter, 1988, Tu et al., 1992, Somogyi Usually, transframe products et al., 1993). Coordinated pauses provide represent the minority. Therefore, PRF is a correct time frame for many recoding often used for production of viral RdRps, events, cotranslational folding of proteins, which are needed in small amounts. The cotranslational assembly of protein +1 frameshift is less common among

16 Introduction

(+)ssRNA viruses, but widespread among 1998, Barry and Miller 2002). In general, bacterial, yeast, and mammalian genes viral propagation occurs in a narrow -1 PRF (reviewed by Farabough, 1996). The -1 effi ciency window. Therefore, compounds PRF was fi rst identifi ed among retroviruses that infl uence -1 PRF frequencies may be (Jacks et al., 1985, 1987, 1988a), but since of use in antiviral therapy (reviewed by then -1 PRF sites have been found in Dinman et al., 1998). dsRNA and (+)ssRNA viruses, prokaryotic genes, prokaryotic and eukaryotic mobile 1.2.2.1 The -1 PRF signals elements, bacteriophage genomes, and in The fi rst cis-acting signal directing -1 PRF some cellular mRNAs (Hammell et al., is the slippery site at which the actual shift 1999, Baranov et al., 2003). Retro- and in frames occurs. To prevent premature totiviruses use -1 PRF for regulating the termination during slippage, the slippery ratio between their structural proteins site is composed of nucleotides that allow (Gag) produced from the 0-frame and the nonwobble bases of the slipped tRNA the enzymatic proteins (Pol) produced as anticodons to rebase-pair with the new a transframe Gag-Pol fusion via -1 PRF -1 frame codons. These requirements (Jacks et al., 1988a, 1988b, Dinman et al., lead to the fact that usually the signal is 1991). Among the (+)ss RNA viruses, -1 a heptanucleotide with the sequence X- PRF determines the ratio between viral XXY-YYZ in the 0-frame and thus XXX- RdRps and other nonstructural proteins YYY-Z in the new -1 frame. In addition such as viral proteases. In some viruses to enabling the slippage to occur, the such as CfMV and PLRV, the -1 PRF also heptamer sequence affects the frequency of affects the amount of VPg synthesized -1 PRF. X can be any nucleotide (Dinman (Mäkinen et al., 1995b, Prüfer et al., 1999). et al., 1991), but it acts as a multiplicative Since the frequency of -1 PRF determines factor of the -1 PRF frequency. The bases the amount of RdRp produced, it is an U and G program the highest and C the important determinant of viral viability lowest -1 PRF effi ciencies in eukaryotes (Dinman and Wickner 1992, Hung et al., (Bekaert et al., 2003). Usually A or U are

Fig. 6. Two cis-acting signals, a slippery heptamer and downstream secondary structure, program -1 PRF. The elongating ribosomes are stalled over the heptamer, since the downstream secondary structure cannot enter the ribosomal mRNA tunnel. During unwinding of the structure, a fraction of the ribosomes slips one nt backwards and continues translation from the overlapping -1 reading frame. Alternatively, the ribosomes may continue translation in the original reading frame.

17 Introduction found at the Y position, probably because more effi cient -1 PRF (Somogyi et al., the higher amount of energy would be 1993, Marczinke et al., 2000, Kontos et needed to unpair tRNAs paired to the CCZ al., 2001), possibly because they may or GGZ codons (Dinman al., 1991). Z resist the unwinding better than stem- also varies, the infl uence of Z on -1 PRF loops (Dinman, 1995, Plant and Dinman, is dependent on the identity of X and Y 2005). The pseudoknots are formed when (Dinman et al., 1991, Marczinke et al., the loop of the fi rst stem base-pairs with 2000). Usually C at this position gives the downstream sequences, which thus the highest, whereas G gives the lowest forms the second stem (Fig. 7). Thus, -1 PRF in eukaryotes (Marczinke et al., unwinding at the basal part of the fi rst stem 2000). The tRNA modifi cations affect of the pseudoknot causes the downstream translation effi cacy and fi delity. Therefore, structure to induce supercoiling in the it was also speculated that alterations in remainder of the fi rst stem, because it the modifi cation status of tRNAs could cannot rotate freely due to the anchoring modulate their shiftiness by affecting of the loop regions with the downstream the bulkiness of the anticodon loop and sequences (Dinman, 1995). This results in the stability of the anticodon-codon inhibition of the unwinding. The unwinding interaction (Urbonavicius et al., 2001). of the pseudoknots is hardened further by However, changes in the modifi cation nonstandard base-pairings, tertiary and state of tRNAs encoding the heptamers quadruple interactions, triple helixes, as affect -1 PRF effi ciency by at most 2- well as the presence of coordinated cations fold (Brierley et al., 1997, Marczinke that stabilize these structures (Su et al., et al., 2000, Urbonavicius et al., 2003), 1999). In stem-loops the loop regions and the effect is highly dependent on the are free to rotate during the unwinding neighboring A- and P-site codons and on process and only the basal base pairs the type of modifi cation (Carlson et al., resist unwinding, thus, the unwinding of 2001, Urbonavicius et al., 2003). stem-loops is easier. Mutational studies The second signal programming -1 have shown that certain sequences, which PRF is a downstream secondary structure are not important for secondary structure that may be a hairpin (HP) (Mäkinen et formation, are necessary for -1 PRF. These al., 1995b, Dulude et al., 2002), but more regions may be involved in modifying the often a complex pseudoknot is involved contacts with the ribosomes or the putative (Kim et al., 1999, Su et al., 1999, Paul trans-acting proteins (Shen et al., 1995, et al., 2001) (Fig. 7). The secondary Kim et al., 1999). structure acts to pause the ribosomes over For successful -1 PRF to occur, the the heptamer so that the XXY and YYZ translational apparatus must pause correctly codons are placed at the P- and A-sites of over the heptamer. Therefore, correct the ribosome (Tu et al., 1992, Somogyi et spacing between the cis-acting signals is al., 1993, Lopinski et al., 2000, Kontos critical (Tu et al., 1992, Somogyi et al., et al., 2001). The shift in frames occurs 1993, Lopinski et al., 2000). The length of during the time the secondary structure is the mRNA protected by the ribosome and unwound. However, pausing as such does the length of the spacer indicates that the not ensure effi cient -1 PRF (Somogyi et intact pseudoknot cannot enter the mRNA al., 1993, Lopinski et al., 2000, Kontos tunnel of the ribosome (Plant et al., 2003). et al., 2001). Usually pseudoknots drive Thus, the spacer becomes stretched before

18 Introduction the pseudoknot is unwound. Therefore, cis-acting signals. In BYDV, -1 PRF is the spacer may account for the -1 PRF regulated via interactions that occur across effi ciency via its capacity to be stretched several kilobases between the cis-acting (Kim et al., 2001, Bekaert et al., 2003). secondary structure and the 3’UTR (Paul Alternatively, the effect of the spacer et al., 2001, Barry and Miller 2002). nucleotides may rely on their interactions No viral proteins have been shown with the translational machinery, on to affect -1 PRF. Due to the ease of the stabilities of the anticodon-codon genetic manipulation of Saccharomyces interactions in the spacer sequence, or cerevisiae, it has been the main target on the availability of the corresponding in studies focused on the search for host tRNAs (Bekaert et al., 2003). factors involved in the regulation of -1 Sequences up- and downstream of PRF (reviewed by Dinman et al., 1998). the slippery heptamer also affect -1 PRF To date, no candidates having direct (Honda et al., 1996, Kim et al., 2001, Barry effect on the -1 PRF effi ciency have been and Miller, 2002, Dinman et al., 2002). identifi ed. For the upstream sequences this seems plausible because they are in contact with 1.2.2.2 Mechanism of -1 PRF the elements that compose the ribosomal The fi rst clues for the mechanism of -1 PRF mRNA channel (Kim et al., 2001). In fact, came from the sequencing of transframe the two bases upstream of the heptamer proteins produced from retroviral RNAs affect the -1 PRF frequencies, which (Jacks et al., 1988a, 1988b). According to suggests that the identity of the E-site these data, a simultaneous slippage model codon is important (Kollmus et al., 1994, was described which proposed that the Bekaert and Rousset, 2005). The E-site- P-site peptidyl-tRNA and the A-site aa- tRNA interaction may affect -1 PRF by tRNA occupying the slippery heptamer infl uencing the stability of the P-site- shift frames simultaneously (Jacks et anticodon interaction directly or indirectly. al., 1988b). Later the model included In some cases, the essential parts of the refi nements based on genetic, biochemical, frameshift signals are located far from the molecular, and pharmacological studies

Fig. 7. Schematic presentation of stem-loop and pseudoknot structures. In addition to the loop regions, stem structures may contain bulges that contain single unpaired bases. In pseudoknots, loop regions are base-paired to fl anking regions via standard and non standard base-pairings. Sometimes the upper stem is bent with respect to the fi rst stem and the stems may also be rotated with respect to each other (indicated by arrows).

19 Introduction

(reviewed by Harger et al., 2002). These affect -1 PRF (Paul et al., 2001, Barry and studies have revealed that the effi cacy of Miller, 2002). High initiation rates increase -1 PRF is dictated by the kinetics of the the ribosomal loads on the mRNAs and elongation steps, which affects the time thus also the amount of ribosomes, which during which the ribosomes occupy the become stacked behind the ribosome -1 PRF signals. Changes in the tRNA unwinding the secondary structure. As anticodons, rRNAs, or ribosomal proteins a result, more ribosomes could pass the may also affect the -1 PRF frequencies via slippery site without pausing and the their effects on the stabilities of the tRNA- frequency of -1 PRF would be lowered ribosome interactions. Furthermore, (Barry and Miller, 2002). defects in the translational apparatus for Current data indicate that only a detecting and correcting errors account for small portion of the mRNA, rather than the occurrence of -1 PRF. the whole ribosome, is involved in the The impacts of the mutations and slippage (Plant et al., 2003). During aa- antibiotics that affect the kinetics of each tRNA accommodation, the transition from specifi c substep in elongation on the the A/T hybrid state to the A-site moves occurrence of -1 PRF have been especially the anticodon loop 9 Å in the 5’direction. important for polishing the mechanism Normally this movement would be of -1 PRF. These studies revealed that accompanied by the mRNA movement, but eEF1A mutations that slow down the aa- the cis-acting secondary structure resists tRNA selection increase the likelihood the movement because it cannot enter of -1 PRF, probably by lengthening the the mRNA tunnel of the ribosome. Thus, time the ribosomes spend on the -1 PRF tension is generated in the spacer region signals (Harger et al., 2002). Antibiotics (Plant et al., 2003). This local tension can and mutations in the ribosomal proteins be relieved via uncoupling of the A- and resulting in reduced peptidyl transfer rates P-site anticodon-codon interactions and and increased residence times at -1 PRF subsequent re-pairing in the -1 frame. The sites also promote -1 PRF (Brunelle et al., energy released by the eIF1A-catalyzed 1999, Meskauskas et al., 2003). However, hydrolysis of GTP is theoretically enough since inhibition of translocation has no to cover the energetic costs of the slippage effect on -1 PRF, the postpeptidyl transfer (Plant et al., 2003). Tension may also ribosomes must be incapable of slipping be relieved by aborted translation or by (Brunelle et al., 1999, reviewed by unwinding of the structure followed by the Harger et al., 2002). Thus, the conclusion movement of the slippery heptamer distal from these studies is that the slippage region of mRNA by one base forward, must occur prior to the peptidyl transfer after which translation can continue in the reaction. Furthermore, since peptide bond original 0-frame (Lopinski et al., 2000, formation occurs spontaneously shortly Plant et al., 2003). after the A-site becomes occupied, the More careful analysis of the amino slippage most probably occurs before or acid sequencing data revealed that a less immediately after the transition of the aa- frequent mechanism of -1 PRF occurs in tRNA from the A/T hybrid state to the A/ parallel with the simultaneous slippage A state (Brunelle et al., 1999, Plant et al., mechanism (Jacks et al., 1988a, Yelverton 2003)(see Fig. 5B). et al., 1994, Horsfi eld et al., 1995). In this High translation initiation rates also case, the A-site amino acid is determined

20 Introduction by the -1 frame, which suggests that the tRNAs. Instead, termination codons are P-site peptidyl-tRNA slips before the aa- recognized by eukaryotic release factor 1 tRNA incorporation. The occurrence of (eRF1), which has the omnipotent capacity single peptidyl-tRNA slippage is induced to decode all three termination codons in situations where the decoding of the 0- (Frolova et al., 1994, Drugeon et al., 1997, frame A-site occurs slowly (Yelverton et reviewed by Kapp and Lorsch, 2004). al., 1994). Slow decoding occurs when The eRF1 acts at the A-site and mimics the A-site decodes a termination codon or aa-tRNAs structurally (Song et al., 2000). when the decoding amino acid is in short A second type of release factor, eRF3, supply. Slow recognition appears to result also takes part in translation termination. in an additional ribosomal pause, during It plays a postulated role in aiding eRF1 which the peptidyl-tRNA slips in the -1 in termination codon recognition and direction. subsequent release of eRF1 (Salas-Marco The rules concerning -1 PRF are and Bedwell, 2004). Binding of eRF1 conserved not only among the lower facilitates the entry of a water molecule and higher eukaryotes (Stahl et al., into the active site of the ribosome and 1995), but also between the eukaryotes induces hydrolysis of the ester bond and prokaryotes (Yelverton et al., 1994, linking the polypeptide chain and the P-site Brierley et al., 1997, Brunelle et al., tRNA (Song et al., 2000). The reaction is 1999, Napthine et al., 2003). The tRNAs the same as during elongation, except that encoding the heptamers determine which the nucleophilic group is a water molecule -1 PRF mechanism is most frequently instead of the amino group of aa-tRNA. used and the same mechanism is utilized Thus, both reactions are catalyzed by the in all organisms (Napthine et al., 2003). peptidyl transferase center of the ribosomal However, the -1 PRF frequencies at large subunit (Song et al., 2000). After the same heptamers vary signifi cantly hydrolysis, the peptidyl-tRNA is released. in prokaryotic and eukaryotic hosts, As the fi nal step, all ligands including indicating that the translation apparatuses the eRFs are released from the ribosome are not exactly identical (Brierley et al., and translation is terminated (Kapp and 1997, Napthine et al., 2003). For example, Lorsch, 2004). the identity of the last nt of the heptamer The effi cacy of termination is has nearly opposite effects on the -1 PRF regulated by the termination codon and frequencies in and in its sequence context. The most important eukaryotic cells (Brierley et al., 1997). determinant of termination effi cacy is Furthermore, E. coli ribosomes usually the nucleotide following the stop codon; promote equal -1 PRF effi ciencies in the termination is favored if it is a purine presence of stem-loops and pseudoknots (C or U) (McCaughan et al., 1995). The (Brierley et al., 1997). suggestion that stop signals are actually composed of at least four nucleotides is 1.3 TERMINATION OF supported by the fi nding that eRF1 also TRANSLATION interacts with bases downstream from Each polypeptide is elongated until a the stop codons (reviewed by Bertram termination codon UGA, UAA, or UAG et al., 2001). However, the capacity of is placed at the A-site of the ribosome. the tetranucleotides to promote peptide These codons do not possess cognate aa- release varies signifi cantly (McCaughan

21 Introduction et al., 1995). Highly expressed genes CPs (Skuzeski et al., 1991, Wills et al., appear to contain the most effi cient 1991, Li et al., 1993, Brown et al., 1996). termination signals recognized with high TMV RdRp is produced via translational affi nity. In contrast, genes expressed at readthrough of UAG, but the downstream low rates accommodate a diversity of nucleotides CARYYA (where R = A/G, Y stop signals that are decoded with lower = C/T) are also detrimental to successful effi ciency due to their lower affi nity for readthrough (Skuzeski et al., 1991). In this eRF1 (McCaughan et al., 1995). However, context, the relatively frequent recoding the incidence of certain recoding events, effi cacy at UGAC can be at least partly such as stop codon readthrough or explained by its ineffi cient recognition selenocysteine incorporation, increases by the eRF1 (McCaughan et al., 1995). In when a weak stop codon is coupled to (+)ssRNA alphaviruses the signal is the certain cis-acting signals in the mRNAs. simplest possible, UGAC (Li et al., 1993), whereas in Moloney murine leukemia 1.3.1 Programmed termination codon virus (MuLV; genus Gammaretrovirus) readthrough signals directing leaky termination consist In programmed termination codon of a stop codon, a spacer sequence, and a readthrough, the stop codons are downstream pseudoknot structure (Wills programmed to be read as sense codons. et al., 1991). The leakiness of termination Instead of eRF1, the stop codons are can also be regulated via long-term recognized by tRNAs that have been interactions. In BYDV two regions reassigned via mutations to decode stop following the stop codon are involved: a codons as sense codons. Alternatively, C-rich region closely after the termination wild-type (wt) tRNAs may wobble base- codon and a distal element nearly 700 pair at the third position. These tRNAs nt downstream. These elements form compete against eRF1 for the same binding putative kissing-loop interactions (Brown site. Therefore, the concentration of the et al., 1996). tRNAs and their affi nity for the termination codon and the A-site environment relative 1.4 POSTTRANSLATIONAL to that of eRF1 determines the readthrough REGULATION OF GENE effi cacy (Drugeon et al., 1997). In viral EXPRESSION readthroughs, suppressor tRNAs usually Translation products may undergo different outcompete the release factor complexes types of posttranslational modifi cations with 1-10% effi ciency. after or during their synthesis. These Programmed termination codon modifi cations affect the stabilities and readthrough is used as a gene expression activities of the corresponding proteins. strategy among many plant, animal, and Well-known examples of the biological bacterial viruses (reviewed by Bertram et consequences of protein modifi cations al., 2001). However, `leaky´ termination include phosphorylation for signal has been observed only at the UAG or transduction, ubiquitination for proteolysis, UGA stop codons. Readthrough ends up attachment of fatty acids for membrane in the production of protein fusions with anchoring, and glycosylation for protein extended C-termini. Among the (+)ssRNA half-life regulation, targeting, and cell–cell viruses, this mechanism is usually used interactions. However, in the following to produce RdRps or extended forms of sections only proteolytic processing of

22 Introduction viral polyproteins is reviewed, since other silencing, aphid-mediated transmission, posttranslational modifi cations were not and viral cell-to-cell movement the focus of this thesis. (reviewed by Rajamäki et al., 2004). The intermediate processing products may 1.4.1 Polyprotein processing also have important functions that differ Many (+)ssRNA viruses encode large from those of the fi nal end products, polyproteins from which the functional resulting in further increases in the coding proteins are released via proteolytic capacity of viral genomes. For instance, processing. For this purpose, viruses encode in alphaviruses, the partially processed proteases that catalyze the hydrolysis of nonstructural polyproteins P123 and specifi c peptide bonds located between two P4 catalyze the minus-strand synthesis, specifi c amino acids. Proteolytic processing whereas completely cleaved proteins may commence during translation and are needed for the plus-strand synthesis proceeds via several intermediate products (Shirako and Strauss, 1994). Thus, it is to the fi nal end products that are present important that viruses can regulate the in equimolar amounts (Merits et al., occurrence of intermediate products. This 2002). Polyprotein processing can be the can be achieved by regulating the timing principal mode of gene expression or it of cleavages and modifying the effi cacy can be used in conjunction with other of the processing, but the stabilities of gene expression strategies (Gorbalenya et the intermediates also play a role (Merits al., 1988, reviewed by Spall et al., 1997). et al., 2002). Potyviral NIaPro recognizes For instance, sobemoviral gene expression a seven-amino acid stretch and cleavage employs polyprotein processing in concert occurs between Gln (Q) and Ser (S), Gly with leaky scanning, -1 PRF, and sgRNA (G), or Ala (A), which are the last two production (Tamm and Truve, 2000a). In amino acids of the recognized sequence contrast, members of the picornavirus-like (Shukla et al, 1994). The recognition superfamily rely solely on the production sequence is composed of residues that of a single large polyprotein for their are highly conserved and required for gene expression (reviewed by Spall et effi cient cleavage, whereas few residues al., 1997). In Potato virus A (PVA; genus vary and adjust the cleavage effi ciency. Potyvirus), a plant virus member of the Thus, some of the cleavages are rapid, group, the polyprotein is processed into as whereas other sites are processed more many as 10 mature proteins by three viral slowly (Merits et al., 2002). However, the proteases (reviewed by Riechmann et al., effi cacy of processing is also affected by 1992). Two of these, helper component other interactions between the enzyme proteinase (HC-Pro) and P1, cleave only and substrate, the most important being their respective C-termini (Verchot et al., the accessibility of the potential cleavage 1991). The third protease, nuclear inclusion site by the protease. Sometimes proteases protein a proteinase (NIaPro), resembles may be regulated by cofactors, such the picornaviral 3C-like proteases and as membranes, RNA, metal cations, or processes the remainder of the polyprotein polypeptide cofactors. The protease of (Gorbalenya et al., 1989). Sesbania mosaic virus (SeMV: genus Viral proteases may be multifunctional Sobemovirus) is active if viral VPg is proteins. For instance, potyviral HC-Pro uncleaved from the Pro-VPg precursor. also functions in suppression of gene (Satheshkumar et al., 2005). Similarly,

23 Introduction in Poliovirus (PV: genus Enterovirus) enzyme. The amino acids of the catalytic effi cient processing by 3C protease occurs triad His (H), Asp (D), Ser (S), or Cys (C) only before the viral RdRp (3D) is released are precisely spaced HX8DX30-31GXSG. from the intermediate (Jore et al., 1988). These proteases are found for instance Some viruses exploit the spectra of in sobemoviruses and the Picornavirus host proteases, which are mostly utilized supergroup (Gorbalenya et al., 1988). In in the maturation of viral envelope proteins serinelike proteases of the Picornavirus (Dougherty and Semler, 1993). Enveloped supergroup, the Ser residue is replaced with viruses may target their proteins to the host Cys, whereas sobemoviral proteases more secretion route by utilizing signal peptides, resemble the cellular serine proteases in which are then removed by cellular target having the Ser residue at the active site. signal peptidases. Along the secretion On cleavage, the Ser or Cys residue acts as route, viral proteins are further processed a nucleophile and donates an electron for and modifi ed by the cellular enzymes the carbonyl carbon of the peptide bond to present in different compartments of the be cleaved (reviewed by Dougherty and secretion route (ER and Golgi). Semler, 1993). Nucleophilic attack forms a covalent acyl-Ser complex between the 1.4.2 Viral proteases enzyme and the polyprotein substrate. The Viral proteases are structurally and active site His residue donates a proton functionally related to cellular serine, for the departing amino group, facilitating cysteine, aspartic, or metalloproteinases its release. Hydroxylation of the acyl- (Gorbalenya et al., 1988). Proteases Ser complex releases the carboxylic acid usually contain two globular domains, product and regenerates the active site. cellular proteases are composed of two Cysteine proteinases have a catalytic separately encoded protein domains, dyad that is composed of closely arranged whereas viral proteases are dimers of two interacting Cys and His residues. These identical proteins (Dougherty and Semler, proteases are also known as papainlike or 1993). The amino acids involved in the thiol proteinases. One viral representative catalysis are located in the crevice formed is the HC-Pro of potyviruses (reviewed by between these domains. The spacing and Shukla et al., 1994). The sulfhydryl group arrangement of the catalytic amino acids is of the Cys acts as a nucleophile that attacks highly conserved and used to classify the the carbonyl carbon of the target peptide proteases. The specifi city of the proteases bond. An enzyme forms a temporary is mainly determined by the three- covalent acyl-enzyme complex through dimensional structure of the substrate- the carbonyl carbon of the substrate and the binding pocket, which is located near the sulfhydryl group of the active Cys residue active site (reviewed by Dougherty and (reviewed by Dougherty and Semler, Semler, 1993). However, the substrate- 1993). The departing amide is protonated binding pockets are highly heterologous by the active site His, the active site is and may consist of a single binding site for regenerated, and the cleavage product is one amino acid or involve several amino released by hydrolysis. acids. Aspartic and metalloproteases Serine and serinelike proteases are operate on acid-base catalysis and do not named according to the presence of a serine form covalent intermediates (reviewed by or cysteine residue at the active site of the Dougherty and Semler, 1993). Aspartic

24 Introduction proteases have the Asp-Thr-(Ser)-Gly Zn2+, is present at the catalytic site. The signature sequence in their active sites. catalytic site also contains Glu and His The catalytic dyad is composed of two Asp residues, where the Glu residue probably residues originating from the individual donates the electron to the carbonyl carbon members of the dimer. Aspartic proteases during peptide bond cleavage. Only a have not been found in (+)ssRNA viruses single characterized (+)ssRNA virus, HCV, but they do exist in retroviruses and plant is known to encode metalloproteinases pararetroviruses (Spall et al., 1997). In (Spall et al., 1997). metalloproteinases a divalent cation, often

25 Aims of the Study

2. AIMS OF THE STUDY

1) To determine the mechanism of translation initiation of CfMV by comparing it with well-studied viral 5’UTRs.

2) To examine the regulation of -1 PRF in CfMV, which probably affects the viral viability by determining the amount of VPg and RdRp produced.

3) To examine the processing of CfMV polyprotein.

26 Materials and Methods

3. MATERIALS AND METHODS

A summary of the methods used is given. Detailed methods can be found in the original publications and references therein. Standard protocols were used for DNA manipulations.

Table 1. Methods used in the present study. Method Reference Amino acid sequencing III Antisera production III Bacterial transformation I-V Barley suspension cell culturing I, II Electroporation I,II Immunoprecipitation III Iodination III In vitro transcription I, II, V In vitro translation I, II, IV, V Northern blotting I, II Particle bombardment I, II Plasmid construction I-V Protein overexpression; III Protein overexpression; yeast V Protoplast isolation I Reporter gene expression analysis I, II, V RT-PCR V SDS-PAGE analysis I-V Site directed mutagenesis by PCR I, II, IV, V Total RNA isolation I, II, V Virus inoculation III Virus purifi cation III Western blotting III, V Yeast cell culturing II, V Yeast spheroplasting II Yeast transformation II, V

Table 2. Viruses or viral sequences studied. Virus or sequence element Reference Alfalfa mosaic virus (AMV, genus Alfamovirus) I Cocksfoot mottle virus (CfMV, genus Sobemovirus) I-V Crucifer-infecting tobamovirus (CrTMV, genus Tobamovirus) I, II Human immunodefi ciency virus (HIV, genus Lentivirus) V Potato virus X (PVX, genus Potexvirus) I Tobacco mosaic virus (TMV, genus Tobamovirus) I, II

27 Materials and Methods

Table 3. Reporter genes and proteins. Protein and gene Origin Activity and measurement β-galactosidase, LacZ E. coli Hydrolyzes fl uorogenic substrate (o- (lacZ) nitrophenyl-beta-D-galactopyranoside) to yield colored product (o- nitrophenyl). β-glucuronidase, GUS E. coli Chemiluminescent or fl uoregenic (uidA) substrate decomposed by the enzyme to yield light emission or coloured product. Enhanced green Jellyfi sh Fluoresces on irradiation with UV. fl uorescent protein, GFP (Aequorea (GFP) victoria) Luciferase LUC, (luc) Firefl y Oxidiation of luciferin results in (Photinus light production and in an inactive pyralis) oxyluciferase. Renilla luciferase, RUC Sea pansy Catalyzes coelenterate-luciferin (Ruc) (Renilla (coelentrazine) oxidation to produce reniformis) light.

28 Materials and Methods

Table 4. The most relevant plasmids used in the study.

Plasmid name Description, vector backbone, and Sequence studied Reference expression system at the 5’UTR or intercistronic position (ICS)

Monocistronic gene constructs 5’UTR pANU5-series 35S-ICS-uidA, pRT101 (Töpfer et AMV 5’UTR, I al., 1993), plant expression CfMVε, PVXαβ, TMVΩ, reference pANU6-series 35S-ICS-luc, pRT101 (Töpfer et al., AMV 5’UTR, I 1993), plant expression CfMVε, CrTMV IRES, PVXαβ, TMVΩ, reference pMKJM GAL1-T7-ICS-luc, pYES2 CfMVε, TMVΩ, II (Invitrogen), yeast expression, in CrTMV IRES, vitro transcription and translation reference pMKJMΔGAL ICS-luc, pYES2 (Invitrogen), GAL CfMVε, TMVΩ, II and T7 promoters deleted, yeast CrTMV IRES, expression reference Dicistronic gene constructs ICS pHKJM GAL1-T7-HP-GFP-ICS-luc, CfMVε, TMVΩ, II pYES2 (Invitrogen), yeast CrTMV IRES, expression and in vitro transcription reference and translation pHKJMΔGAL HP-GFP-ICS-luc; pHKJM in CfMVε, TMVΩ, II which GAL and T7 promoters CrTMV IRES, deleted, pYES2 (Invitrogen), yeast reference expression pHKJMB GAL1-T7-HP-lacZ-ICS-luc, pYES2 CfMVε, TMVΩ, II (Invitrogen), yeast expression and in CrTMV IRES, vitro transcription and translation reference pKJM GAL1-T7-lacZ-ICS-luc, pYES2 CfMVε, TMVΩ, II (Invitrogen), yeast expression CrTMV IRES, reference pKM T7-HP-GFP-ICS-uidA (Ivanov CfMVε, CrTMV I et al., 1996), pBluescript SK+ IRES, reference (Stratagene), in vitro transcription and translation pPKM 35S-HP-GFP-ICS-luc, pRT101 CfMVε, CrTMV I (Töpfer et al., 1993), plant IRES, reference expression

pKAH T7-ICS-luc-(A)35, pSP73 (Promega), CfMVε, TMVΩ, I in vitro transcription and translation reference

29 Materials and Methods

-1 PRF -1 PRF casettes

pAC-A/B/C/AB/ SV40-lacZ-ICS-luc, pAC74 dual reporter A (CfMV V BA/AC/CA plasmid (Stahl et al., 1995), bacterial and 1602-1720), B yeast expression. The -1 PRF induces (1386-2137), C translation to continue to the -1 frame luc (1551-1990), AB gene. (1602-2137), BA (1386-1720), AC (1602-1900), CA (1551-1720) pAC-Am/Bm/Cm/ SV40-lacZ-ICS-luc inframe controls, in CfMV Am, Bm, V ABm/BAm/ which one nt insertion upstream the -1 PRF Cm, ABm, BAm, ACm/CAm site enables the luc gene to be translated ACm, CAm without the -1 PRF event. pAC1789, SV40-lacZ-HIV-1-luc test (1789) and HIV-1 -1 PRF V pAC1790 inframe control (1790) constructs that region contain a 53 bp sequence from the HIV-1 -1 PRF region, (Stahl et al., 1995), bacterial and yeast expression pACRF-A/B/C SV40-Ruc-ICS-luc, pAC74 (Stahl et A/Am, B/Bm, V al., 1995), in which the lacZ gene was C/Cm replaced with Ruc gene, bacterial and yeast expression pACRF-Am/Bm/ SV40-Ruc-ICS-luc inframe controls, one nt A/Am, B/Bm, V Cm insertion upstream the -1 PRF site enables C/Cm the luc gene to be translated without the -1 PRF event. pYES2/NT-A/Am, GAL1-T7-lacZ-ICS-luc, pYES2/NT A/Am, B/Bm, V B/Bm, C/Cm (Invitrogen), N-terminal 6-His tag, yeast C/Cm expression

pAB-21 T7-ORF2A2B, pGEM-5Zf(-) (Promega), in CfMV polyprotein IV vitro transcription and translation encoding region

pFSC1 T7-uidA-CfMV(1621-2521)-uidA, CfMV 1621-2521 IV pGEM3Z(-) (Promega), in vitro transcription and translation Protein expression pQE-VPg T5-VPg, pQE30 (Qiagen), N-terminal 6- CfMV 1386-1724 III His tag, bacterial expression pYES-P27 GAL1-P27, pYES2/NT (Invitrogen), N- CfMV 1385-2137 V terminal 6-His tag, yeast expression pYES-Rep GAL1-RdRp, pYES2/NT (Invitrogen), N- CfMV 1669-3255 V terminal 6-His tag, yeast expression

30 Results and Discussion

4. RESULTS AND DISCUSSION

4.1 Translation initiation from CfMV translation initiation in CfMV differs from RNA cap-mediated translation initiation. Due to the simplicity of viral genomes, viruses serve as irreplaceable tools for 4.1.1 Comparison of protein production the study of cellular processes such as from CfMVε with known plant viral translation (reviewed by Gale et al., 2000). translational enhancers (I) Studies of translation can be performed in To initiate studies on the translational vivo or in vitro, both approaches having properties of the 5’UTR of CfMV (CfMVε), benefi ts and drawbacks. A complex set we compared it with known enhancer of factors affects the translation process sequences TMVΩ, CrTMV IRES, the in vivo and therefore the results may be 5’UTR of AMV RNA4, and Potato virus diffi cult to interpret (Kozak, 2002). In X αβ (PVXαβ, genus Potexvirus) (Fig. 1A the in vitro assays the reaction conditions in I). All these leaders promote effi cient are more easily controlled, but since the translation in wheat germ extract (WGE) endogenous mRNAs are removed any and in tobacco cells (Gallie et al., 1987, added transcript is usually translated Jobling and Gherke, 1987, Browning et readily. Therefore, mRNAs that would al., 1988, Gallie et al., 1989, Smirnyagina not promote signifi cant expression under et al., 1991, Pooggin et al., 1992, Ivanov in vivo conditions, may be translated in et al., 1997, Dorokhov et al., 2002, Gallie, vitro. However, in vitro conditions may be 2002a), whereas with the exception of the brought closer to the in vivo conditions by CrTMV sequence (Dorokhov et al., 2002) supplementing reactions with competitor they do not appear to function in yeast cells RNAs or by depleting the lysates from eIFs (Van den Heuvel and Raue, 1992, Everett by incubating them with cap analogues or and Gallie, 1992). We placed the viral poly(A) sequences (Gallie and Tanguay, sequences into plant expression vectors 1994, Gallie, 2001, 2002a). upstream from a reporter gene (luc or Several 5’UTRs from plant viruses can uidA) and compared the activities obtained stimulate translation initiation (reviewed with a reference construct, in which the by Gallie, 1996). The enhancement results 5’leader was composed of a multicloning from the reduced requirement for certain site (Fig. 1A in I). The effect of the host on eIFs (Browning et al., 1988, Pestova and the functioning of viral leader sequences Kolupaeva, 2002) or from the capacity was studied by performing the analysis in of viral sequences to effi ciently recruit tobacco protoplasts (Nicotiana tabaccum essential eIFs (Gallie, 2001, 2002a, Krab et L.) (I), in barley suspension cells (Hordeum al., 2005). Since CfMV multiplies to high vulgare L. cv. Pokko) (I), and in yeast (S. titers in its host plants (Truve et al., 1997), cerevisiae) (II). we examined whether this was due to the The competitive advantage of the strong translation activity of viral mRNA. AMV RNA4 leader arises from the fact Sequencing of CfMV RNA showed that that the 5’UTR does not contain secondary it lacks the poly(A) tail (Mäkinen et structures. As a result the 43S preinitiation al., 1995a). The 5’terminus of RNA is complexes can reach the initiation codon covalently linked to a viral protein, VPg in the absence of adenosine triphosphate (Fig. 2 in III). These facts suggest that (ATP) and factors associated with ATP

31 Results and Discussion hydrolysis, i.e. eIFs 4A, 4B, and 4F (Pestova known to contain high levels of Hsp101 and Kolupaeva, 2002). Although TMVΩ (data not shown). The CrTMV sequence is also unstructured (Sleat et al., 1988), it also functioned poorly in the barley cells appears to use a different strategy. TMVΩ at the 5’leader position (Fig. 2B in I), does not recruit eIFs directly (Tanguay although it promotes effi cient translation and Gallie, 1994), instead it binds a host at the intercistronic position in tobacco and protein Hsp101 that presumably assists in WGE (Dorokhov et al., 2002) (Fig. 2B in the recruitment of eIF3 and eIF4F in I). Together our results suggested that (Wells et al., 1998, Gallie, 2002a). In our there may be substantial variations in the experiments both sequences promoted requirements for effi cient gene expression effi cient gene expression in the tobacco from viral 5’UTRs in barley and tobacco. cells, in close correlation with existing data. In some cell types longer 5’leaders CfMVε also increased protein expression, promote higher translation yields (Niepel but the 5’leaders from viruses infecting and Gallie, 1999a, 1999b, Gallie et al., dicotyledonous plants were stimulating 2001). The sequences studied varied even higher protein yields (Fig. 2A in I). greatly in length, with the CrTMV and In contrast, CfMVε was the only element PVX leaders being six times longer than the that positively affected gene expression reference 5’UTR. However, we observed in the barley suspension cells (Fig. 2B no correlation between the 5’leader in I). PVX αβ can promote translation in length and the downstream reporter gene barley protoplasts (Zelenina et al., 1992). expression levels. The expression from In our experiments it mediated the highest CfMVε-containing plasmids occurred expression among the rest of the 5’UTRs approximately at 12-times higher levels (Fig. 2B in I). Surprisingly, we observed in barley cells when TMVΩ and CfMVε no translational enhancement from TMVΩ as equal-length sequences were compared. in barley suspension cells, although it can This suggests that CfMVε could be used enhance translation in suspension cultures to enhance heterologous gene expression of rice and maize, both monocots (Gallie in cereals. et al., 1989). However, ΤΜVΩ cannot stimulate translation in all plant cell types, 4.1.2 Identifi cation of regions important since no translational enhancement was for gene expression from CfMVε in obtained in Orychophragmus violaceus barley suspension cells (I) cells (Family Brassicaceae) (Akbergenov CfMVε was mutagenized to identify et al., 2004). The Hsp101 content varies, regions that were crucial to enhanced gene depending on the developmental stage expression in barley suspension cells. and type of the cells (Young et al., 2001). Mutagenesis was performed for the 35S- Although Hsp101 is well expressed CfMVε-luc plasmid during PCR. Deletion in barley seeds (Tangyay and Gallie, of almost half of the 3’terminal part of 1994), the expression levels in the barley CfMVε had no effect on the extent of LUC suspension cells used may have been too yield (DelII, Fig. 3B in I). This showed low to provide competitive advantage for that the deleted region was not needed the TMVΩ−containing mRNAs. When for improved expression from CfMVε- we tested the Hsp101 expression with containing constructs and further proved Western blotting, a very weak signal was that the enhanced expression relative to observed compared withthat in WGE the reference leader did not result from the

32 Results and Discussion difference in leader lengths. Introduction of translation yields than the wt sequence or a small uORF to CfMVε (Destab II, Fig. 3B the other 5’UTRs studied, it could be the in I) abolished translation initiation from best choice for stimulation of heterologous the downstream luc AUG, indicating that protein production in barley. translation initiation involved scanning. Computer analysis identifi ed several This is in agreement with studies on relatively short (7-14 nt) GC-rich translation initiation from SBMV (Hacker segments from 18S rRNA that can base- and Sivakumaran, 1997). pair with complementary sequences of We suggest that CfMVε folds into a mRNAs (Matveeva and Shabalina, 1993). stem-loop structure that begins at fi rst nt Therefore, it was suggested that stretches of the RNA (Fig. 3A in I). Destabilization in mRNAs that were complementary with of the putative structure by mutating the the 18S rRNA could function in attaching 5’stem sequence resulted in an almost mRNAs to the small ribosomal subunits. 2-fold improvement in LUC expression This compared favorably with results in relative to the wt sequence. In contrast, which 5’UTRs having short complement- restabilization of the structure by ary regions with 18S rRNA stimulated complementary mutations in the 3’stem translation in yeast, plant, and mammalian sequence reduced translation to half of the cells (Zhou et al., 2003, Akbergenov et expression measured from the wt CfMVε al., 2004, Chappell et al., 2004). Several (Destab I and Restab, Fig. 3B in I). Similar translational enhancers from plant viruses results were obtained from the in vitro also show complementarity with 18S translations, suggesting that the inhibition rRNA (Akbergenov et al., 2004). CfMVε occurred at the translational level (data not also has several 6-23-nt-long stretches shown). Thus, the 5’proximal position of with 70-100% complementarity with plant the secondary structure appeared to inhibit 18S rRNA (data not shown). In plants, either the initial binding of the 43S complex the 5’UTRs complementary to the central or subsequent scanning, or both. In carrot region of plant 18S rRNA (nt 1105-1124 suspension cells, a 5’proximal stem-loop according to the rice 18S rRNA sequence) of seven bases reduced luc translation to show the highest affi nity for the ribosomal 10% (Niepel and Gallie, 1999b). The stem- 40S subunits and also stimulate the loop formed in the restabilization mutant strongest protein expression (Akbergenov had a slightly higher predicted free energy et al., 2004). The C-rich region immediately (–17.4 kcal/mol) than the wt structure (−12 downstream from the CfMVε stem-loop kcal/mol). Thus, reduced translation from structure can potentially base-pair with this mutant may have resulted from the this central 18S rRNA region. However, inability of the preinitiation complexes to deletion of the corresponding region from bind and to initiate scanning. Alternatively, CfMVε did not affect expression from the sequence of the 3’part of the stem may this element (DelI, Fig. 3B in I). The have contributed to the stimulated gene sequence forming the 3’part of the stem expression from CfMVε. However, since and the upstream loop region contained the entire 5’proximal sequence of CfMVε potential sites for 18S rRNA binding (data was altered in the restabilization mutant, not shown). To determine whether these solid conclusions cannot be made before potential base-pairings play a role in gene less drastic mutations are studied. Since the expression from CfMV RNA, it should fi rst destabilization mutation resulted in higher be shown whether the corresponding sites

33 Results and Discussion are accessible in the plant 40S subunits. multiplication. Under normal growth A more detailed mutation analysis at the conditions, Hsp101 expression is usually single-nt level could also be performed. low in nondeveloping tissues, such as adult leaves (Young et al., 2001). This could 4.1.3 Transient expression from in vitro indicate that the cap-mediated binding of transcribed mRNAs (I) 43S preinitiation complex is utilized in the To determine whether the observed presence of low concentration of Hsp101. enhancement in gene expression occurred Under these circumstances, TMV would at the transcriptional or translational level, benefi t from having a simple 5’UTR, since in vitro transcribed mRNAs were delivered scanning of unstructured leaders does to tobacco protoplasts or barley suspension not require eIF4F and eIF4A (Pestova cells. In tobacco protoplasts, LUC yields and Kolupaeva, 2002). However, certain from capped mRNAs containing CfMVε stressful conditions disrupt cap-mediated or TMVΩ were 1.6 ± 0.6- and 2.4 ± 1.1- translation initiation via modifi cation of fold higher than the expression driven eIF activities (reviewed by Dever, 1999). from the reference mRNA, respectively. In TMV translation may possibly overcome barley suspension cells, the same mRNAs some stress responses by recruiting eIFs drove LUC expression only ~1.2 ± 0.4- via Hsp101. In tobacco cells, heat shock- fold over the reference (Fig. 4 in I). Thus, mediated induction of Hsp101 expression since neither CfMVε nor TMVΩ promoted results in as much as 10-fold stimulated relative LUC expression from the in vitro translation from TMVΩ (Gallie, 2002a). It transcribed mRNAs to levels similar would be interesting to examine whether to those in plant expression plasmids, TMVΩ-mediated expression would it appeared that the enhanced LUC improve after heat treatment of the barley expression from plant expression vectors suspension cells. Hsp101 expression did not result solely from the stimulated appeared to be induced in the barley translation activity of these mRNAs. The suspension cells used after 30-45-min heat TMVΩ function overlaps with that of treatments at +37 °C and +45 °C (data not the 5’cap (Gallie, 2002a). As a result, the shown). extent of translation enhancement from polyadenylated mRNAs containing TMVΩ 4.1.4 Functioning of viral leader is higher from uncapped than from capped sequences in S. cerevisiae (II) mRNAs in comparison to a corresponding To determine whether the plant host was reference mRNA lacking the viral needed for enhanced gene expression 5’UTR (Sleat et al., 1988, Gallie, 2002a). from the plant viral leaders studied, we However, capping of the mRNAs did not also analyzed gene expression from mask the detection of the TMVΩ effect in CrTMV, TMVΩ, and CfMVε in yeast barley suspension cells, since translational (II). The studied sequences were inserted stimulation was likewise not observed in into yeast expression plasmids between the uncapped mRNAs (data not shown). an inducible galactokinase 1 (GAL1) The gRNA of TMV contains both the 5’cap promoter and the luc gene (Fig. 1B in and the translational enhancer element. II). All viral sequences tested inhibited So why does it need both? TMV may downstream reporter expression, whereas possibly use two parallel mechanisms for high levels of expression were measured translation initiation to ensure its effi cient from the reference construct containing

34 Results and Discussion a polylinker 5’UTR (Table 2 in II). The in yeast (Coward et al., 1992, Evstafi eva yeast expression plasmids had identical et al., 1993). CfMVε possibly contains extensions of 189 nt upstream from the a single stem-loop structure, whereas studied sequences. Therefore, the initial CrTMV IRES may have two of them binding of the cap-binding complex was (Ivanov et al., 1997). These structures may assumed to occur with similar affi nity have reduced translation from the CrTMV for all mRNAs. However, Northern blot and CfMV sequences by inhibiting analysis indicated that several transcripts scanning. In fact, a 5’UTR stem-loop were produced from all plasmids (Fig. 3B,C structure with a free energy as low as – 4.5 in II). One explanation for these additional kcal/mol reduces translation to 5% in yeast RNAs is that they may have originated (Niepel and Gallie, 1999b). The predicted from opportunistic transcription initiation free energies of CrTMV IRES and CfMVε in the plasmid or in the studied sequences. are clearly higher (Fig. 3A in I). Since For instance, the GAL promoter is known translation initiation in yeast occurs almost to contain several minor transcription exclusively via cap-mediated recruitment initiation sites (Johnston and Davis, of initiation complexes (Preiss and Hentze, 1984). The shorter mRNAs may also be 1998), the structures present in CfMV and RNAs that have lost the major part of their CrTMV 5’UTRs may have caused the poly(A) tails via deadenylation, which is low translatability of the corresponding the fi rst step in mRNA degradation. In mRNAs. Although TMVΩ is unstructured contrast, the longer forms would represent (Sleat et al., 1988), it inhibits translation in the most recently synthesized mRNAs yeast (Everett and Gallie, 1992, Van den with intact poly(A) tails (Caponigro Heuvel and Raue, 1992). In our case, LUC and Parker, 1996). Whatever the origin expression was inhibited by ~90% (Table 2 of these shorter transcripts, they were in II). One reason for the poor functioning present in comparable amounts in all of TMVΩ in yeast appears to be the transformants including the reference. inability of yeast homologue Hsp104 to Thus, the differences in the expression complement the corresponding protein of levels probably did not result from the plant origin (Wells et al., 1998). variable mRNA amounts but most likely from differences in the translation effi cacy 4.1.5 Translational properties of CfMVε of the mRNAs. in vitro (I) The 5’UTR length does not usually The translational properties of CfMVε, affect translation initiation effi cacy in TMVΩ, and CrTMV IRES were also yeast, but long runs involving Gs and Us compared in WGE, which allowed us to are deleterious (reviewed by Romanos et examine more easily the role of eIFs in al., 1992). The yeast scanning complex is translation initiation from these 5’UTRs. also more sensitive to secondary structures The translations were programmed with in present in the 5’UTRs than are complexes vitro transcribed luc mRNAs having a viral from higher eukaryotes (Kozak, 1986, or polylinker 5’UTR and a 35-nt poly(A) Vega Laso et al., 1993). Furthermore, tail. Commercial translation mixes have secondary structures are equally inhibitory abundant translational capacity, whereas at all positions of the 5’UTRs (Vega Laso strong competition for eIFs and the et al., 1993). These facts may explain translational apparatus prevails in living why many viral 5’UTRs function poorly cells. Therefore, elements improving

35 Results and Discussion the competitiveness of mRNAs do not containing a 189-bp extension upstream necessarily show up under the conditions from the studied sequences. Tobamoviral recommended by the suppliers of lysates. sequences promoted approximately 10- For instance, translation initiation in times higher LUC yields than the reference vitro is not enhanced by the cap and the 5’UTR (data not shown). CfMVε improved poly(A) tail to the same extent as in vivo, expression ~7-fold. where a strong synergy between these Next, the functional half-lives of the terminal elements is observed (Gallie, mRNAs were determined to compare the 1991, Gallie and Tanguay, 1994). This periods during which the mRNAs studied compared favourably with our results, remained translationally active. This was in which we measured only a 2-3-fold done by measuring the duration of LUC higher expression from capped mRNAs synthesis in the translation mix. Cessation relative to uncapped mRNAs (data not of LUC accumulation was taken as an shown). In the barley suspension cells indication of complete degradation of the corresponding difference between the mRNAs programming translation. The capped and uncapped mRNAs was ~42- functional half-life was then designated fold (data not shown). as the amount of time required for 50% of Translation lysates can be made to the mRNAs to become degraded (Gallie resemble more the in vivo conditions by and Tanguay, 1994). This analysis did not using depleted lysates or high mRNA show differences in the degradation rates concentrations (Gallie and Browning, of reference mRNA or mRNAs containing 2001, Gallie, 2002a). For instance, the viral 5’UTRs (data not shown). Thus, the translational enhancement conferred by capacities of the viral leaders to promote TMVΩ becomes detectable in WGE only translation in WGE did not result from under these types of conditions (Sleat et their stabilizing effect on the mRNAs. al., 1988, Gallie, 2002a). Therefore, we TMVΩ showed increasing stimulation fi rst determined the mRNA concentration of translation with rising mRNA in which translation in WGE became concentrations, whereas less improvement saturated and thus competition-dependent. was measured from CfMVε. This may This point was chosen based on reduced have resulted from differences in the LUC expression, which indicated that the complexity of the 5’UTRs. The predicted high mRNA amounts sequestered eIFs, stem-loop of CfMVε initiates from the fi rst thus reducing translation initiation (Fig. nucleotide of CfMV RNA and is formed 5A in I). Below the saturation point (60 of 10 bps (Fig. 5A in I). It was previously ng/μl), no strong translational advantage shown that under in vitro conditions from TMVΩ or CfMVε was observed secondary structures are more inhibitory at (Fig. 5B in I). However, at higher mRNA higher mRNA concentrations, putatively concentrations both viral sequences due to titration of the eIF4A helicase showed improved translation in relation to activity (Gallie and Browning, 2001). the reference mRNA, indicating that viral A 7-bp 5’proximal G-C-rich stem-loop leaders succeeded better under conditions with free energy of –4.5 kcal/mol reduced in which the eIFs became limiting. translation in WGE to 60% with RNA Enhanced translation from viral sequences concentrations as low as 10 ng/μl (Gallie was also observed in coupled transcription and Browning, 2001). Another study and translation reactions with mRNAs showed that scanning through even weak

36 Results and Discussion

A-U-rich secondary structures required 4.1.6. Contribution of CfMV 3’UTR on that all eIFs participated in the process translation initiation from CfMVε (I, (Pestova and Kolupaeva, 2002). This unpublished) suggests that translation initiation from In several cases viral 3’UTRs substitute CfMVε also requires the entire set of eIFs, the poly(A) tail functionally and cooperate including eIFs 4A and 4F that are putatively with the viral 5’UTRs in promoting required to unwind the 5’proximal stem- effi cient translation initiation. Hsp101 loop of CfMVε. Translation from TMVΩ also binds to the TMV 3’UTR (Tanguay is improved via effi cient recruitment of and Gallie, 1996) and thus, may bridge the eIFs 3 and 4F to the element via a protein interaction of the two termini of TMV. This bridge formed of Hsp101 (Wells et al., correlates favorably with the translation 1998, Gallie, 2002a). Since Hsp101 is an effi ciency of mRNAs containing both abundant protein in WGE (Gallie, 2002a), TMV UTRs, which is higher than that effi cient recruitment of eIFs to TMVΩ of mRNAs containing TMVΩ and the can presumably continue even under poly(A) tail (Gallie, 2002a). In BYDV, highly competitive conditions. As already effi cient translation is achieved when discussed, some translational advantage the 5’UTR and 3’TE base-pair (Guo et can also arise from the simplicity of al., 2001). The 3’TE binds eIFs and this TMVΩ. Reconstitution assays performed interaction may serve to deliver the eIFs to in rabbit reticulocyte lysate (RRL) have the 5’UTR (Guo et al., 2001). shown that eIFs 4A, 4B, and 4E are not The combined effect of CfMV UTRs required for 48S complex formation in on translation effi ciency was compared

(CAA)n 5’UTR (Pestova and Kolupaeva, with luc mRNAs having a vector-derived 2002), a sequence very similar to TMVΩ. 3’UTR (145 nt), CfMV 3’UTR (226 nt), Translation lysates do not contain or a vector derived 3’UTR combined into endogenous mRNAs. Thus, in vitro a poly(A) tail (145 nt + 35 nt of poly(A)). translations are performed in somewhat In WGE, uncapped mRNAs ending at artifi cial environments in the absence the CfMV 3’UTR or at the poly(A) tail of competition against cellular mRNAs promoted gene expression 2-fold relative (Gallie and Tanguay, 1994). To further test to the vector-derived 3’UTR. With capped the competitiveness of the viral leaders mRNAs the highest expression was studied, translations were programmed obtained from the polyadenylated mRNAs in the presence of total RNA from yeast (data not shown). In barley suspension (Fig. 5C in I). Under these conditions, cells, polyadenylated mRNAs gave the CfMVε programmed translation only best expression irrespective of whether 1.6-fold more than the reference mRNA. the mRNAs were capped or not (data not This correlated well with the results shown). The expressions from mRNAs obtained from the barley suspension cells. ending at the vector- or CfMV-derived Thus, CfMVε did not succeed under the 3’UTR were similar, but only about half competitive conditions very well, which the level of expression measured from could be explained by the requirement mRNAs ending at the poly(A) tail. of all eIFs for translation initiation from In general, it is probably a rule rather this 5’UTR. In contrast, TMVΩ competed than an exception that interaction of the successfully against the heterologous UTRs is required for effi cient translation mRNAs and promoted translation 11-fold (reviewed by Gallie, 1998). However, over the reference. the two termini of CfMV RNA did not

37 Results and Discussion appear to have a synergistic effect on mediating internal initiation of translation, translation as such. This may indicate which is presumably cap-independent. The that a sequence from the coding region is leader of the sgRNA encoding CrTMV needed or that a viral protein plays a role CP is a strong IRES in several cell types in connecting the UTRs. AMV RNAs are (Dorokhov et al., 2002) and was used as effi ciently translated when AMV CP binds a positive control. The key element in to the 3’UTR and recruits eIF4G and CrTMV IRES is a polypurine-rich region eIFiso4G to the viral RNA (Neeleman et (Dorokhov et al., 2002). Interestingly, al., 2001, Krab et al., 2005). Translation CfMVε also contains a GA-rich region. of VPg-containing SBMV RNA is less In fact, the only remarkable homology susceptible to inhibition by cap analogues at the 5’UTRs of sobemoviruses is the than are capped and uncapped SBMV GAAA sequence that is located in the loop RNAs (Hacker and Sivakumaran, 1997), of the putative 5’stem structure (Mäkinen which may suggest that this viral protein et al., 1995a, Ryabov et al., 1996). As participates in translation initiation of already mentioned, CfMVε also contains sobemoviral RNAs. In fact, certain plant several regions complementary with and animal virus VPgs are known to the 18S rRNA. Recently it was shown interact with eIFs, such as eIFs 3, 4E, that complementary interaction with the iso4E and PABP, suggesting that VPg may leader and 18S rRNA may enable cap- participate in translation initiation of viral independent binding of 43S preinitiation RNAs (Wittmann et al., 1997, Léonard et complexes into the intercistronic spacers al., 2004, Goodfellow et al., 2005). We (ICSs) of dicistronic mRNAs (Chappell et aim to conduct future tests to determine al., 2000, Akbergenov et al., 2004). whether CfMV proteins affect translation The in vitro translations were from mRNAs containing CfMV UTRs. In programmed with capped and the infected plants, CfMV CP is one of the polyadenylated dicistronic mRNAs, in most abundant proteins (Fig. 4 in III). This which the test sequences (TSs) were placed suggests that the sgRNA encoding CP may between GFP and luc genes (Fig. 1B in also contain elements that guarantee the I). Initially, we planned to use ΤΜVΩ high productivity of CP. as a negative control, because it was previously shown that it cannot promote 4.2 Does CfMVε promote internal translation from the internal position in initiation of translation? O. violaceus cells (Akbergenov et al., 2004). Supporting data come from the 4.2.1 Studies on internal initiation in RRL system, in which a stable secondary

WGE (I) structure placed upstream from a (CAA)n Increasing amounts of data indicate leader prevented the 48S complex that several 5’UTRs can also promote formation, indicating that the (CAA)n translation initiation from intercistronic sequence cannot mediate internal binding positions (Levis and Astier-Manifacier, of 40S subunits (Pestova and Kolupaeva, 1993, Niepel and Gallie, 1999a, Ivanov et 2002). In addition to viral sequences a al., 1997, Koh et al., 2003). Since CfMV reference control was included, in which RNA is covalently linked to a viral protein the reference multicloning site served as and not to the cap structure (III), we the ICS. Expression of the 3’proximal examined whether CfMVε was capable of LUC cistron was taken as an indication of

38 Results and Discussion internal translation initiation. To diminish the dicistronic mRNAs was determined the likelihood that the 3’cistron expression to verify that translation templates were arose from reinitiation or leaky scanning, not cleaved to monocistronic mRNAs translation of the 5’cistron was prevented encoding functional LUC protein. by a 5’proximal stable HP. However, no such degradation products In a manner similar to that of the results were detected from any of the mRNAs obtained with monocistronic mRNAs (I), (Fig. 6B in I). Furthermore, the physical low mRNA amounts showed no signifi cant stabilities of the individual mRNAs were variation at the level of 3’LUC production comparable. The functional half-lives of between viral and reference ICSs (Fig. the dicistronic mRNAs showed that the 6A in I). However, the IRES activity may reference mRNA remained translationally become detectable only under competitive active for the longest period of time: the in vitro conditions, similar to the results approximate t1/2 was 80 min (Fig. 8). The obtained with translational enhancers dicistronic mRNAs containing CfMVε (Gallie, 2001). Increased mRNA amounts and TMVΩ were degraded slightly revealed stimulated expression from more rapidly (t1/2 ~68 min), whereas the mRNAs containing viral TSs. Surprisingly, mRNA containing the CrTMV sequence TMVΩ promoted the highest 3’cistron had the shortest half-life (~58 min). The expression. The degradation pattern of expression data from the same experiment

6000 25000

5000 20000 ICS 4000 15000 TMVΩ CfMVε 3000 Ref 10000

RLU CrTMV CrTMV 2000 RLU TMV, CfMV, Ref 5000 1000

0 0 024 6 8 1 1 1 1 1 2 2 2 2 2 3 3 0 0 0 0 0 2 4 6 8 0 2 4 6 8 0 2 0 0 0 0 0 0 0 0 0 0 0 0 Time (min)

Fig. 8. Functional half-lives of dicistronic mRNAs were determined by incubating them within the WGE translation mix, which lacked the ribonuclease inhibitor. The duration of LUC expression was followed until the mRNAs became degraded and LUC accumulation ceased. This can be observed as a plateau in the curve. The functional half-life was then designated as the amount of time required for 50% of the mRNAs to become degraded. The fi nal concentration of RNA in the mix was 35 ng/μl. RLU, relative light unit.

39 Results and Discussion showed that the enhancement of LUC abundant Hsp101, which is putatively used expression relative to the reference mRNA to recruit eIF4G into TMVΩ (Wells et al., remained constant from the fi rst rounds of 1998, Gallie, 2002a). Thus, the internally translation until the end. Thus, the viral positioned TMVΩ may also be capable elements did not stabilize the dicistronic of recruiting eIF4G, which then further mRNAs and the stimulated expression recruits the eIFs needed for translation from mRNAs containing viral sequences initiation. Hsp101 binds to the CAA repeat did not result from the stability differences of TMVΩ (Tanguay and Gallie, 1994). or opportunistic expression from IRESs from Hibiscus chlorotic ringspot monocistronic mRNAs. In conclusion, virus (HCRSV: genus Carmovirus) and all viral sequences promoted internal TEV share the CA-richness with TMVΩ. initiation in WGE. The CA region is crucial for the HCRSV WGE contains two isoforms of the IRES function and as an unstructured cap-binding complex. The more abundant region it was proposed to serve as a landing complex eIFiso4F (eIFiso4E, eIFiso4G) pad for the ribosomes (Koh et al., 2003). promotes translation preferentially from Both IRESs function in WGE (Gallie, unstructured mRNAs, whereas eIF4F 2001, Koh et al., 2003). (eIF4E, eIF4G) also promotes translation To verify that the observed 3’cistron from mRNAs that contain multiple expression also occurred in other gene cistrons, structured leaders, or uncapped combinations, the 3’luc gene was replaced mRNAs (Gallie and Browning, 2001). with uidA. Alternatively, the 5’GFP was Interestingly, translation initiation from switched to a lacZ gene. Neither change TMVΩ is eIF4G- but not eIFiso4G- prevented the 3’cistron expression (Fig. dependent (Gallie, 2002a). WGE contains 9). However, we observed slightly lower

A) B)

7.0 5.84 6.0 4.75 5.0

4.0

3.0

2.0 1.56

Relative 3'Luc expression 1.0 1.0

0.0 ICS Reference CfMV 5'UTR TMV 5'UTR crTMV

Fig. 9. Translation of the 3’proximal gene of dicistronic construcs occurs in various 5’reporter contexts. A) Autoradiogram from coupled in vitro transcription and translation reaction in WGE preformed in the presence of 35S-methionine. Linearized dicistronic pSK: T7-HP-GFP-ICS-uidA plasmids were used as templates. The CfMVε (lane 1), CrTMV IRES (lane 2), or CfMV -1 PRF signal (lane 3) was inserted between the reporter genes. B) Relative expression of 3’LUC from coupled in vitro transcription and translation reactions programmed with linearized dicistronic pKJM plasmids (T7-lacZ-ICS-luc). LUC expression from the reference was set to 1.

40 Results and Discussion relative 3’LUC expression from constructs amount of eIF4A is also decreased (Gallie, containing the 5’lacZ gene instead of the 2002a). Incubation of WGE with poly(A) GFP gene, similar to earlier observations depletes the lysate mostly from PABP on the impact of the 5’gene on the and eIF4G (Gallie and Browning, 2001), effi cacy of downstream cistron expression but the concentrations of eIFs 4A and 4B (Ivanov et al., 1997, Chappell et al., 2000, are also reduced (Gallie and Tanguay, Hennecke et al., 2001). 1994). In both cases indirect interactions To test the competitiveness of internal may also reduce the amount of other eIFs initiation against cap-mediated translation, (Gallie, 2002a). Free cap analogue reduces dicistronic mRNAs were translated in the translation from the capped mRNAs and presence of total RNA from yeast (Fig. 7A stimulates translation from the uncapped in I). Under these conditions the relative mRNAs (Tanguay and Gallie, 1994). If 3’cistron expression from tobamoviral translation initiation from the internal sequences occurred at levels similar to position is cap-independent, the cap those observed in the absence of total analogue should not affect the 3’cistron RNA. The relative enhancement from translation. This compared favorably with TMVΩ and CrTMV was 5.1 and 12.9, our results, in which we measured higher respectively. When a similar experiment absolute LUC levels from translations was repeated in the presence of the supplemented with the cap than from monocistronic competitor Ruc mRNA, the translations lacking this additive (Table relative enhancement of 3’LUC expression 5). This indicated that titration of eIF4E from TMVΩ was reduced to 2.5, whereas did not affect the 3’LUC expression. that of CrTMV increased up to 22.3 (Fig. Stimulation of LUC expression was more 7A in I). The internal initiation effi ciency pronounced in the case of CfMVε and from CfMVε did not differ from that of TMVΩ, in which activities almost twice the reference. Coupled transcription and as high were obtained. In monocistronic translation reactions performed with equal mRNAs, the Hsp101-mediated translation amounts of monocistronic and dicistronic initiation from TMVΩ is eIF4E- templates indicated that the 3’cistron independent but eIF4G-dependent (Wells expression attained 25% of the level et al., 1998). For CrTMV and reference of monocistronic expression (data not mRNAs the extent of stimulation was shown). 30% and 48%, respectively. However, To increase the level of understanding the induced LUC expression did not alter of the eIFs required for translation the relative 3’cistron expression ratios initiation from the internally positioned signifi cantly in comparison to the situation viral sequences, dicistronic RNAs in which no addition was made (Fig. 7B were translated in WGE, which was in I). Supplementation of translation supplemented with cap analogue or mixes with poly(A) inhibits translation of poly(A) sequence to reduce the amount uncapped mRNAs more than the capped of eIFs interacting with them. Incubation mRNAs (Tanguay and Gallie, 1994). of WGE with the cap analogue decreases Poly(A) treatment renders the lysates the amount of components from the cap- cap-dependent so that the synergistic binding complex i.e. eIF4E, eIFiso4E, effect of the cap and poly(A) tail on eIF4G, and eIFiso4G (Gallie and Tanguay, translation effi ciency of mRNAs is also 1994, Browning and Gallie, 2001). The observed under in vitro conditions (Gallie

41 Results and Discussion and Browning, 2001). In our studies, CfMVε−uidA cassettes from constructs the poly(A) addition reduced translation used in the in vitro assays were transferred yields from all our dicistronic mRNAs to plant expression vectors between the (Table 5). This is expected, considering 35S promoter and transcription terminator. the scaffolding protein function of eIF4G. The resulting plasmids were then used to Depletion affected the reference mRNA analyze the transient expression of the most, in which expression attained 3’proximal GUS in barley suspension only 12% of the level measured in the cells. In these studies a 190-nt region nonsupplemented translations. With viral from the CfMV -1 PRF site was used sequences, translation was most reduced as a reference. Unfortunately, GUS from mRNAs containing CfMVε (~26%). expression was very low due to the low Tobamoviral sequences were less affected transfection effi ciency achieved via and translation was reduced to ~40%. particle bombardment combined with the Thus, the capacity of viral sequences to low activity of 35S promoter in barley. resist poly(A)-mediated depletion of eIFs Expression from CrTMV was barely 2- correlated with their capacity to increase fold above the reference (data not shown). the relative 3’cistron expression (Fig. 7B In comparison to the GUS expression in I). In general, translation initiation from measured from a monocistronic control ICSs appeared to be dependent on eIF4G. plasmid, expression from an internally positioned CrTMV sequence was less 4.2.2 Internal initiation in barley than 8%, while expression from CfMVε suspension cells (unpublished) was clearly lower. The value reported We next tested the capacity of for CrTMV-mediated IRES activity in CrTMV IRES and CfMVε to promote transgenic tobaccos is much higher: 30% internal initiation in barley suspension (Dorokhov et al., 2002). This suggested cells. Dicistronic HP-GFP-CrTMV/ that either the viral sequences studied did

Table 5. Effects of eIF depletion on 3’cistron expression. In vitro translations were programmed with capped and polyadenylated dicistronic HP-GFP-ICS-luc mRNAs and supplemented with 1 mM cap analogue or with poly(A) at fi nal concentration of 27 ng/μl. After a 90-min incubation, reactions were terminated on ice and the LUC activities were measured. Activities are presented as RLUs.

ICS No additive Cap Poly(A)

Reference 4428 ± 454 6555 ± 953 533 ± 155

CfMVε 9173 ± 1060 16594 ± 820 2432 ± 70

TMVΩ 38606 ± 3269 74283 ± 1549 13983 ± 1281 CrTMV 38649 ± 3100 50173 ± 1787 16579 ± 1765 IRES

42 Results and Discussion not function as effi cient IRESs in barley ORF. Since IRES studies have been or that a required host factor was lacking criticized for the uncertainty of whether from the barley cells. Due to the very low mechanisms other than internal initiation 3’GUS synthesis obtained with internally may give rise to the observed 3’cistron positioned CfMVε, further studies with expression (Kozak, 2003), we examined these constructs were not performed in the this possibility further. barley cells. Evaluation of the yeast expression data indicated that the actual expression 4.2.3 Internal initiation in yeast (II) levels of the 3’cistron were much lower In contrast to barley suspension cells, than the expression measured from the high 3’cistron expression was achieved monocistronic plasmids (Table 2 in in S. cerevisiae with constructs in which II). Even more, the plant viral 5’UTRs the CfMVε, TMVΩ, and CrTMV IRESs functioned poorly in the monocistronic were placed into the ICS of two reporters context compared with the reference leader. in yeast expression vectors (Fig. 1A in Thus, 3’cistron expression appeared to be II). All viral sequences promoted higher rather ineffi cient. One alternative cause 3’cistron expression relative to the for the low-level 3’cistron expression expression measured in the reference, was that it occurred via reinitiation or which contained the multicloning site as leaky scanning. However, both events the ICS (Fig. 2 in II). CrTMV programmed are dependent on translation of the fi rst the highest 3’cistron expression of the ORF. In two of the plasmid series used, studied sequences, closely paralleling pHKJM and pHKJMB, translation of the earlier reports on the effi cacy of CrTMV 5’proximal gene was blocked by 90% IRES in yeast (Dorokhov et al., 2002). by a stable HP structure. Thus, in these In general, few viral IRESs are known to cases leaky scanning or reinitiation would function in yeast cells and some are active have been unusually effi cient processes only if the cap-dependent translation if LUC was expressed via this means. initiation is compromised (Thompson Furthermore, the number of AUGs in the et al., 2001, Dorokhov et al., 2002, lacZ gene would have prevented leaky Rosenfeld and Racaniello, 2005). This scanning to the luc AUG (Kozak, 1989). may result from the fact that cap-mediated The fact that the 3’LUC was expressed translation initiation is a very effi cient at comparable levels from pKJM and process in yeast (Preiss and Hentze, 1998). pHKJM constructs, differing only in the A survey of 2000 yeast genes found no translatability of the 5’lacZ gene (Table 2 IRES activity from a single yeast 5’UTR in II), suggested that reinitiation or leaky (Thompson et al., 2001). Thus, IRES- scanning was an unlikely explanation mediated translation initiation appears for the observed expression from these to be a relatively rare event in yeast. In constructs. Viral sequences also stimulated our experiments the extent of 3’reporter 3’LUC expression signifi cantly from expression was dependent on the sequence pAGL constructs in which the GAL and translatability of the 5’cistron (Fig. 2 promoter was switched to an alcohol in II), even though one could assume that dehydrogenase 1 (ADH1) promoter and the binding of ribosomes to sequences the 5’GFP translation was not blocked by promoting internal initiation should the HP (Table 6). Since reinitiation does not occur independently of the 5’proximal usually occur after translation of full-length

43 Results and Discussion

ORFs (Kozak, 2001), it also appeared more appeared likely that the mRNAs produced likely that if the 3’LUC expression did not from the promoter-free plasmids lacked result from internal initiation, it would the inhibitory parts of the viral sequences. have originated from transcripts arising Stimulated LUC expression was also from splicing or transcription initiation observed in the dicistronic promoter- from cryptic promoters. containing constructs when transcription Splicing or cryptic transcription from the GAL promoter was repressed could generate transcripts, in which by growth on glucose (Table 4 in II). No the initiation codon of the 3’luc gene is detectable LacZ expression occurred under brought close to the 5’end of the mRNA. the same conditions. Thus, the mRNAs These monocistronic mRNAs could then programming LUC expression must have promote the LUC expression observed. lacked a functional lacZ gene. Since no The presence of shorter mRNAs encoding signifi cant expression occurred from functional LUC protein was supported the promoter-containing monocistronic by the expression data. The 3’cistron luc mRNAs during repression, binding was expressed when the GAL promoter of transcriptional regulators to the was deleted (Table 3 in II). Since high GAL promoter appeared to prevent the expression was also measured from occurrence of cryptic transcription. No promoter-free monocistronic constructs, cryptic promoter activity has been found the 5’reporter appeared to be unnecessary in the yeast expression plasmid backbone for the observed LUC expression. The used (Hecht et al., 2002). However, fact that transcripts synthesized from the introduction of a reporter gene between the promoter-free monocistronic constructs GAL promoter and the studied sequences containing viral sequences stimulated LUC allowed some cryptic transcription to expression, whereas those produced from occur, since LUC expression was observed. the GAL promoter inhibited translation In this case the increased distance to the (Table 2 in II), suggested that the 5’leaders GAL promoter may have enabled the differed in these mRNAs. Thus, it transcription initiation complex to form.

Table 6. The 3’LUC expression from dicistronic pAGD constructs, in which ADH1 promoter regulated transcription of GFP-ICS-luc mRNAs. The cells were grown in the presence of glucose or galactose and raffi nose. The LUC activities were normalized to μg of total protein concentration. Means (± SD) calculated from one experiment with three independent clones are shown. Galactose + ICS Glucose Raffi nose CrTMV IRES 28.8 ± 1.1 2.47 ± 0.63

TMVΩ 15.3 ± 3.2 1.66 ± 0.13

CfMVε 10.9 ± 1.0 0.86 ± 0.04

Reference 0.7 ± 0.2 0.09 ± 0.01

44 Results and Discussion

When the mRNAs produced during mRNAs produced from the corresponding induced transcription from the GAL monocistronic expression plasmids promoter were analyzed with Northern (pAL) (Fig. 10). However, larger mRNAs blot analysis, the dicistronic mRNAs were were also detected, which presumably readily detected (Fig. 3 in II). However, represented the dicistronic mRNAs. Thus, several additional RNAs were detected the Northern blot analysis indicated that with a probe recognizing the 3’luc gene shorter RNAs were produced from the but not with a probe recognizing the dicistronic plasmids, which could have 5’GFP gene (Fig. 3 in II, data not shown). programmed the LUC expression. Comparison of the sizes of these RNAs Surprisingly, although elevated with the mRNAs synthesized from the expression was measured from the monocistronic plasmids revealed that the promoter-free expression plasmids, no larger species most likely represented mRNAs were detected during growth on a monocistronic mRNA. Interestingly, glucose or galactose and raffi nose (Fig. the most abundant mRNAs produced 3C in II, Fig. 10). This suggested that from the dicistronic expression plasmids the RNAs detected from the promoter- containing the ADH promoter upstream containing dicistronic transformants from the GFP-ICS-luc casettes (pAGL) were not necessarily driving high LUC also had sizes very similar to those of the expression. Therefore, the short RNAs

Fig. 10. Transcript amounts detected in the Northern blot analysis did not correlate with the LUC expression data. Total RNA isolated from yeasts grown on different carbon sources was examined with antisense probe recognizing the 5’terminal part of the luc gene. Yeasts transformed with dicistronic pAGL (ADH-GFP-ICS-luc), pHKJM (GAL-HP-GFP-ICS-luc), and pHKJMΔGAL (HP-GFP-ICS-luc) plasmids and monocistronic pAL (ADH-ICS-luc) and pMKJM (GAL- ICS-luc) controls were studied. A total of 10 μg of total RNA was loaded into the gel. In vitro synthesized HP-GFP-ICS-luc and luc transcripts served as size markers.

45 Results and Discussion may have represented stable degradation the LUC expression observed originated products that lacked the 5’cap structure from cryptic promoters rather than from and were thus poorly translated. In yeast, spliced mRNAs. The CrTMV sequence mRNA degradation is usually initiated by also functions in inverted orientation shortening of the poly(A) tail, which is (Toth et al., 2001), which suggests that it followed by decapping and degradation can function as a transcriptional promoter by a 5’-to-3’ exonuclease (Caponigro (Kozak, 2001). and Park, 1996). Yeast mRNA encoding Usually yeast promoters consist of at phosphoglycerate kinase 1 (PGK1) is least three parts, which are an upstream destabilized if a stable HP in its 5’UTR activator sequence (UAS), a TATA box blocks translation (Muhlrad et al., 1995). (consensus TATAA), and the initiator Thus, the blockage of 5’cistron translation element (reviewed by Romanos et al., from dicistronic mRNAs may have targeted 1992). We next searched vector-derived them for degradation. Alternatively, intercistronic sequences (VDSs), studied degradation may have begun from the viral and reference sequences, and 5’end of the mRNAs independently of the 5’reporter genes for the putative binding deadenylation via a nonsense-mediated sites of yeast transcription factors (Zhu RNA decay (NMD) pathway. However, and Zhang, 1999). Several sites were NMD is triggered by premature translation found from all studied viral sequences but termination (reviewed by Weischenfeldt not from the reference sequence (Table 7). et al., 2005). In our case, this would For instance, both tobamoviral sequences take place at the termination codon of contained several putative binding sites 5’cistron. Absence of these shorter RNA for TATA-binding protein (TBP) 50- species in yeast mutants lacking the 5’- 130 nt upstream from the AUG. In yeast 3’exoribonuclease 1 (XRN1) would promoters, TATA elements are usually reveal whether the short RNAs were true located 40-120 bp upstream from the degradation products. However, it would transcription initiation site (Romanos et not explain why the 3’proximal part of the al., 1992). The largest number of putative mRNAs was left undegraded. binding sites for transcription factors High LacZ activities were measured was found from the lacZ gene (Tamle during induced transcription from the 7). Interestingly, the highest background GAL promoter from dicistronic constructs expression from the reference construct having a translatable lacZ gene (pKJM). was also observed from plasmids, in which Thus, splicing at the 5’lacZ gene appeared the 5’reporter was lacZ (Tables 2 and 4 highly unlikely. In addition, RT-PCR in II). This suggests that the combined analysis of transformants harboring effect of putative binding sites in viral dicistronic HP-GFP-ICS-luc-plasmids sequences and in the upstream sequences (pHKJM) revealed no PCR products may have accounted for the enhanced truncated at the GFP-ICS-luc junction (data LUC expression observed from viral not shown), suggesting that these mRNAs sequences. The variation in the expression were also not spliced. Finally, a computer- levels observed during growth on different based analysis of conserved yeast splicing sugars (Tables 2 and 4 in II) may thus have sites (Lopez and Séraphin, 2000) found no resulted from differences in the activities intron patterns from the upstream reporter of the various transcription factors binding sequences or ICSs. Thus, it appeared that to the viral sequences.

46 Results and Discussion

Table 7. Yeast transcription factors putatively interacting with the used 5’reporter genes, VDSs, or the studied viral and reference sequences. The numbers in the parentheses indicate the number of sites found. VDS = vector-derived intercistronic sequence. Sequence Transcription factor CfMVε GCN4, GCR1, MSN2 TMVΩ GRF10(4), TBP(3) CrTMV IRES ABF1, ADR1, GCN4, GCR10(8), MCM1, TBP(3) Reference -

VDSpKJM, pHKJMB GCN4, GRF10, REB1, SWI5, UME6

VDSpHKJM GCR1, LEU3, REB1, UME6(3) ABF1, ACE2(5), ADR1(4), GAL4, GCN4(21), GCR1(24), LacZ GRF10(3), HSF1, LEU3(4), MBP1(8), MCM1(2), MSN2(3), NBF, RAP1, REB1(4), SWI4, SWI5(15) GFP ABF1(5), ACE2, GCR1(2), LEU3, MAC1, MCM1, RAP1, SWI5(3) ABF1 (ARS binding factor): General transcriptional activator ACE2 (Activation of CUP1 Expression): Involved in regulation of histidine and adenine biosynthesis genes. ADR1 (Alcohol Dehydrogenase Regulator 1): Transcriptional activator of alcohol dehydrogenase 2 (ADH2). GAL4 (Galactose metabolism): Transcription factor in expression of galactose-induced genes. GCN4 (General Control Nondepressible): General control of nitrogen and purine metabolism. GCR1 (Glycolysis regulatory protein 1): Activator of glycolytic genes. GRF10 (General regulatory factor 10): Regulation of purine pathway genes. HSF1(Heat shock transcription factor 1): Regulation of transcription in response to heat shock. LEU3 (Leucine biosynthesis): Transcription regulator in branched chain amino acid biosynthesis pathways repressor and activator. MAC1 (Metal-binding activator): Repression of transcription of genes coding for copper transport proteins.

MBP1 (MluI-box binding protein 1): G1/S-specifi c transcription. MCM1 (Minichromosome maintenance factor 1): Activator of a-specifi c genes. MSN2 (Multicopy suppressor of SNF1 mutation): Transcriptional activator for genes in multistress response. NBF (Nonamer binding factor): Transcriptional regulation of phospholipid biosynthesis genes. RAP1 (Repressor activator protein 1): Transcriptional regulation of most genes. REB1 (RNA polymerase I Enhancer Binding protein 1): General transcription factor.

SWI4 (Switching defi cient): G1/S-specifi c transcription. SWI5 (Switching defi cient): Transcription factor for control of cell cycle-specifi c transcription of homothallic switching endonuclease. TBP (TATA binding protein): Component of RNA polymerases I, II, and III; part of initiation factors TFIID and TFIIIB UME6 (Unscheduled Meiotic gene Expression): Negative transcriptional regulator involved in nitrogen repression and induction of meiosis.

47 Results and Discussion

4.2.3.1 Identifi cation of regions severely, suggesting that either the spacing important for gene expression from or the deleted sequence was crucial for internally positioned CfMVε in yeast the observed 3’cistron expression (DelII (unpublished) in Fig. 8B). CfMVε contained putative To determine which regions of CfMVε binding site for MSN2 and GCR1 in the were important for gene expression from C-rich region following the stem-loop the internal position, mutated CfMVε structure (Table 7). However, when this elements were introduced into the HP- region was deleted (PyrDel in Fig. 8B), no GFP-ICS-luc dicistronic plasmids and repression in LUC expression occurred. analyzed in yeast (Fig. 11). Deletion of the In CrTMV, the GA-rich module directs region spacing the 5’terminal secondary cross-kingdom IRES activity (Dorokhov structure of CfMVε and the AUG of et al., 2002). Deletion or mutation of the the luc gene reduced LUC expression GAAA motif from CfMVε reduced 3’LUC

Fig. 11. CfMVε was mutated to identify the regions, which were regulating the expression of 3’proximal luc gene from dicistronic constructs. A) Alignment of mutated CfMVε sequences. Alignment was performed with Multalin version 5.4.1. (Corpet, 1988). B) Relative 3’LUC expression from mutated CfMVε sequences. The measured activities were compared with wt CfMVε, which was given the value 1.

48 Results and Discussion expression, indicating that these regions (Table 2 in II). This indicated that the were also involved in the regulation of major expression from the dicistronic the 3’cistron expression from CfMVε plasmids arose from templates other than (GAAdel and GAAmut in Fig. 8B). the dicistronic mRNAs. Destabilization of the 5’stem-loop structure did not affect LUC expression (5’Destab 4.2.4 Evaluation of the dicistronic in Fig. 8B), suggesting that neither the approach in IRES studies stem structure nor the 5’stem sequence Currently, the dicistronic approach is the was crucial for LUC expression. However, main experimental setup used to identify nt changes in the 3’stem sequence reduced and study IRESs. However, the fact that the 3’LUC expression (3’Destab in Fig. 8B), 3’cistron expression actually arises from the suggesting that in addition to the GA- dicistronic mRNAs has often been poorly rich loop the downstream region was also studied (reviewed by Kozak, 2001, 2003). needed for expression from the internal Several follow-up studies have revealed position. that in many cases translation appears to originate from aberrant mRNAs derived 4.2.3.2 Determination of 3’ cistron from cryptic transcription or splicing translation from dicistronic mRNAs in rather than from the intercistronic position yeast spheroplasts (II) of dicistronic mRNAs (Kozak, 2001, Han To measure the level of the 3’cistron and Zhang, 2002, Hecht et al., 2002, Verge translation from the dicistronic mRNAs, et al., 2004, reviewed by Kozak, 2003, this capped and polyadenylated mRNAs study). IRES activities are also usually were electroporated to yeast spheroplasts presented as relative values, in which together with a transcript encoding RUC. expression is compared with a sequence RUC was readily expressed, which that should not promote internal initiation. indicated that electroporation had been Since the level of background expression successful (data not shown). However, varies greatly, depending on the ICS and no 3’LUC expression from the dicistronic cell type (Niepel and Gallie, 1999a, Gallie mRNAs was detected, even though the et al., 2000) as well as on the sensitivity assay is very sensitive. In contrast, LUC of the assay used to detect the 3’cistron expression from the monocistronic luc expression, evaluation of the importance mRNA was 170-fold over the background, of the expression level is impossible which was practically zero. Thus, we unless comparisons are made against conclude that expression from the monocistronic controls. One problem in dicistronic mRNAs was ineffi cient and the use of dicistronic expression plasmids less than 1% of that measured from the is also the fact that the mRNAs produced monocistronic mRNAs (1/170x100% may undergo some unwanted processing = 0.6%, if background would be given events in the nucleus, which are left value one). When the expression from the undetected due to the lack of suffi ciently dicistronic plasmids was compared with sensitive methods. The best way to that from the monocistronic reference circumvent these problems is to perform the plasmid, the 3’LUC expression from studies with in vitro-synthesized mRNAs. CrTMV was ~7% (7.2/106.7x100%), The synthesis of aberrant mRNAs from 3.7% from TMVΩ (3.9/106.7x100%), upstream reporter sequences may also be and 2.1% from CfMVε (2.2/106.7x100%) avoided by utilization of monocistronic

49 Results and Discussion expression constructs. However, in sequencing of CfMV VPg extracted from this approach cap-mediated translation viral RNA was performed. The 17-amino initiation should be prevented, which may acid sequence obtained corresponded to lead to problems similar to those observed amino acids 320-336 in the C-terminal with the dicistronic plasmids. part of ORF2A (Fig. 5 in III), verifying that CfMV proteins are organized in a 4.3 Proteolytic processing of CfMV Pro-VPg-Pol order similar to that in other polyprotein (III) sobemoviruses (van der Wilk et al., 1998) The CfMV polyprotein is encoded from and in related poleroviruses (van der Wilk the second ORF of CfMV RNA (Mäkinen et al., 1997). et al., 1995a). Studies of CfMVε suggested Based on sequence comparisons, that translation initiation from this 5’UTR Gorbalenya et al. (1988) suggested that involved scanning (I). However, since sobemoviral proteases cleave between the fi rst initiation codon is in suboptimal glutamate (E) and serine (S) or threonine context, some ribosomes most probably (T). Studies of SBMV and SeMV proved bypass the fi rst ORF and reach the that the theory is valid at least in these polyprotein ORF via a leaky scanning sobemoviruses (van der Wilk et al., 1998, mechanism. Sequence comparisons Satheshkumar et al., 2004). However, N- suggested that the proteolytic processing terminal sequence analysis revealed that of sobemoviral polyproteins involves the CfMV VPg cleavage occurred between proteases that are similar to picornaviral glutamate and asparagine (E319/320N) (III). 3C proteases and cellular serine proteases (Gorbalenya et al., 1988). In contrast 4.3.2 Polyprotein processing in infected to the picornaviral proteases, however, plants sobemoviral proteases contain a serine CfMV polyprotein is not processed in vitro, residue in the place of a cysteine residue at which indicates that some essential factors their active sites (Gorbalenya et al., 1988). are lacking from the system (Tamm et al., Therefore, sobemoviral proteases also 1999). Thus, we examined the processing closely resemble cellular serine proteases further in CfMV-infected plants to ensure and may represent the evolutionary link that all the putatively needed host factors between cellular and viral proteases were available. Barley plants were infected (Gorbalenya et al., 1988). In CfMV the with CfMV and plant samples collected conserved amino acids of the active site from infected and uninfected plants are located in the central part of the P2A analyzed with antisera raised against P2A, polyprotein (Mäkinen et al., 1995a). P2B, CP (Tamm et al., 1999), and VPg (III). The sizes of the detected bands were 4.3.1 N-terminal sequencing of CfMV then compared with the calculated sizes of VPg the hypothesized processing products that CfMV RNA has a viral protein VPg were predicted, based on the identifi ed covalently linked to its 5’end (Figs. 1 and N-terminal E/N processing site of CfMV 2 in III). Since VPg is packaged in the viral VPg, the size of the CfMV VPg, and the particles among the gRNA, it represents sizes of the proteases and RdRps in related a functional end-cleavage product. To viruses (Table 8). examine the cleavage sites used in CfMV P2A antisera simultaneously polyprotein processing, N-terminal recognized several products with estimated

50 Results and Discussion sizes of 12, 18, 19, 20, 23, 24, and 30 kDa in III). This suggested that this protein was from the same sample (Fig. 3A in III). The the C-terminal fragment of P2A, which predicted sizes of the serine and serine- included the CfMV VPg. The protein was like proteases range from 17 to 35 kDa named P27 according to its predicted mass. (Dougherty and Semler, 1993). Thus, one In PLRV, the C-terminal intermediate is of the 18-24-kDa bands could be the CfMV also readily detected (Prüfer et al., 1999). protease (Pro). The ~12-kDa protein could Unexpectedly, the mature 12-kDa VPg was have represented the VPg or the very N- hardly ever detected with the VPg antisera or C-terminal part of P2A (Fig. 5 in II, in the CfMV-infected plants. This may Table 5). Since the 12-kDa protein was not indicate that CfMV VPg does not exist detected by the VPg antisera, it most likely in its free form and that it is immediately represented either the N-or the C-terminal linked to the viral RNA. fragment of P2A. However, discrimination Infection was not synchronous and between these two was not possible with therefore it was possible that several the antisera used. The 24-kDa protein cleavage intermediates were detected recognized with the P2A antisera was also simultaneously. Some of the 18-24-kDa recognized with the VPg antisera (Fig. 4B bands may have also represented differently

Table 8. Sizes and detection of expected processing products of CfMV polyprotein. Sizes were calculated according to the amino acid sequence of the predicted products. Estimated size, Protein Recognizing antisera kilodalton (kDa) P2A2B * 103 P2A, P2B, VPg N131Pro-VPg-RdRp * 89 P2A, P2B, VPg N320VPg-RdRp * 69 P2A, P2B, VPg P2A 61 P2A, VPg P2B* 56 P2B T468RdRp* 53 P2B 42.8, 46.5, 47.8, P2A, VPg, N-term-Pro-VPgE397 / E432 / E445 / E467 * 50.4* (P2B)1 N131Pro-VPg-C-term 47 P2A, VPg N-term-ProE319 34 P2A N131Pro-VPgE397 / E432 / E445 / E467 * 29, 33, 34, 36.5* P2A, VPg, (P2B)1 N320VPg-C-term (P27) 27 P2A, VPg N131ProE319 20 P2A N-termE130 14 P2A N320VPgE397 / E432 / E445 / E467 * 8.7, 12.3, 13.7, 16.0* P2A, VPg, (P2B)1* T398 / S433 / S446C-term 18.0, 14.4, 13.0 P2A 1 Short transframe portion, unlikely detected with P2B antisera. * Transframe protein

51 Results and Discussion modifi ed isoforms or degradation products SeMV, the fi rst cleavage in the polyprotein of CfMV proteins. Amino acid sequencing occurs between the VPg and the RdRp of CfMV VPg failed to identify the second (Satheshkumar et al., 2004). Therefore, it and fi fth amino acids. The CfMV RNA is possible that in CfMV the polyprotein sequence (Mäkinen et al., 1995a) suggests processing also starts at the corresponding that these amino acids should be serine cleavage and that no other processing (S321) and tyrosine (Y324). These amino intermediates detectable with the P2B acids can be modifi ed from their hydroxyl antisera exist. groups; one possible modifi cation could Several of the predicted intermediates be uridylylation. In picornaviruses were never detected in the infected plant uridylylated VPg functions as a primer for samples (Table 8), which may indicate viral RdRp during RNA replication (Paul that the intermediates were short-lived. et al., 1998). Unfortunately, several of the predicted Sometimes a 58- and a 62-kDa intermediates and the viral CP were protein were detected with the P2A similarly sized. Thus, large amounts antisera (data not shown). These proteins of CfMV CP (~30 kDa) in the infected could have represented the full-length plants (Fig. 4A in III) may have masked P2A (61 kDa) precursor. However, in the detection of intermediates with similar these blots the antisera also cross-reacted masses. However, a CfMV protein of ~33 with some plant proteins from uninfected kDa was occasionally detected with the and infected plants. Thus, we cannot P2A antisera (data not shown). This protein absolutely rule out that some ‘specifi c’ could have represented the N-terminal Pro proteins detected in the infected plants or the Pro-VPg intermediate. represented host proteins whose expression was induced as a result of infection. The 4.3.3 Putative processing sites of CfMV full-length P2A2B polyprotein (103.4 polyprotein kDa) was never detected with the antisera Finally, we searched CfMV polyprotein used. Among many viruses, such as for similar E/N sites used to cleave the potyviruses, processing already initiates Pro-VPg junction. Another E130/131N site during polyprotein synthesis (Merits et was located upstream from the putative al., 2002). In PLRV and in SeMV the protease-encoding region (Fig. 5 in III). large polyprotein intermediates also Processing at this site would release a represent the minority (Prüfer et al., 1999, protease of ~20 kDa (Table 8). Good Satheshkumar et al., 2004). Therefore, candidate proteins of comparable sizes polyprotein processing of sobemoviruses were detected in the infected plant material may also initiate cotranslationally. CfMV (Fig. 3A in III). Amino acid comparison P2A2B transframe protein is synthesized between the N-terminal cleavage site of via -1 PRF (Mäkinen et al., 1995b) and VPg and this putative site also revealed it attains ~10-20% of the amount of P2A some consensus in the fl anking amino (IV, V). Therefore, detection of transframe acids (VE/NSRLQPLESS, conserved precursors with P2B antisera would be amino acids in bold), strongly supporting more diffi cult. In fact, only a single ~54- the hypothesis that this was used to 58-kDa protein, most likely representing release the N-terminus of CfMV protease. the viral RdRp, was detected in the Dissimilarities in the fl anking amino acids CfMV-infected plants (Fig. 3B in III). In could be used to fi ne-tune the timing of the

52 Results and Discussion processing events. In SeMV, the cleavage poleroviruses (van der Wilk et al., 1998, between protease and VPg is slow, probably Satheshkumar et al., 2005). Processing at because the VPg domain is needed to keep P2A E432/433S or E445/446S would yield a 12- the protease active (Satheshkumar et al., kDa protein, however, the fl anking amino 2004, 2005). These facts would suggest acids shared no similarity with the other that in CfMV the site E319/320N between the predicted E/N processing sites. protease and the VPg is also cleaved more The 54-58-kDa size of the RdRp slowly than the putative E130/131N site. detected in CfMV-infected plants indicated The C-terminus of VPg must be that the cleavage must occur in the vicinity released with further processing events. of the -1 PRF site. Once again, no suitable However, whether the C-terminus of E/N sites were found in that region. The CfMV VPg is encoded entirely from N-terminus of SBMV VPg is processed at the 0-frame or as part of the transframe RSQE326/327TLPPEL (van der Wilk et al., protein is not currently known. We can 1998). In SeMV the corresponding release hypothesize that the transframe VPg would occurs at RSNE325/326TLPPEL (Lokesh already be in close contact with the RdRp, et al, 2001). Interestingly, a similar site which most likely catalyzes the joining RAAE467/ 468TEFPEL is located at the between the VPg and RNA. In contrast, if beginning of the CfMV ORF2B transframe P27 served as a VPg donor, there would region. Processing at this site would yield be more VPg to be linked with the RNA. a replicase of 53 kDa and a transframe However, there are no E/N sites in the VPg of ~16 kDa. One of the 18-19-kDa 0- or in the -1 frames that would yield a proteins detected with the P2A antisera may 12-kDa VPg. This suggested that cleavage possibly have represented this transframe sites others than E/N must also be used to VPg (Fig. 3A in III). However, a VPg of process the CfMV polyprotein. In SBMV this size would probably be too large to and SeMV, the C-terminal processing of be linked to viral RNA without further VPg occurs at the E/T site about 80 amino processing. After its initial release, the acids downstream from the N-terminal SeMV RdRp undergoes further processing cleavage site (van der Wilk et al., 1998, at a suboptimal E/S site to yield an RdRp Satheshkumar et al., 2004). A similarly of ~52 kDa (Satheshkumar et al., 2004). located E397/T398 site can also be found In conclusion, the processing sites in CfMV P2A (Mäkinen et al., 1995a). used in CfMV differ clearly from those However, C-terminal cleavage at this site used in SeMV and SBMV. However, would produce a VPg of 8.7 kDa, whereas phylogenetic comparison of the N-terminal the virion-extracted VPg has a mass of 12 Pro-VPg region of sobemoviruses showed kDa. This indicates that if CfMV VPg was that CfMV is clearly distinguished from released from P27, it must have undergone SeMV and SBMV (Lokesh et al., 2001). signifi cant modifi cations. In contrast, a 12- In fact, the CfMV Pro-VPg domain shares kDa protein would be produced if the C- only 27% similarity with SeMV Pro- terminal cleavage site was close to residue VPg (Lokesh et al., 2001). Similarities 430. However, no suitable E/N or E/T sites in the genome expression strategies could be found around that region in the between poleroviruses and sobemoviruses 0- or the -1 frames. We next looked for E/S prompted us to adapt the model proposed sites, which are used in some processing for processing of PLRV polyprotein events of studied sobemoviruses and (Prüfer et al., 1999) to CfMV. Poleroviral

53 Results and Discussion and sobemoviral polyproteins share an the viral RdRp and thus the effi ciency of - N-terminal transmembrane domain, 1 PRF determines its amount. Two signals which may indicate that a membranous putatively directing the event, a slippery location is required for processing (Prüfer UUUAAAC heptamer (1634-1640) and et al., 1999, Satheshkumar et al., 2004). a downstream stem-loop structure (1648- This could explain why no processing is 1676), can be found in the N-terminal part observed under in vitro conditions. The of the ORF2A2B overlap (Mäkinen et al., membranous location is most probably also 1995b, Fig. 1 in IV). Chemical probing required for viral RNA replication. P27 of segment 1634-1690 from CfMV RNA contains motifs for RNA binding (Tamm has shown that a 12-bp stem with a 4- and Truve, 2000b), and this part could nt loop is formed 7 nt downstream from ensure that viral RNA is transported along the heptamer (Tamm, 2000c). Mutational the P2A to the same cellular location with analysis verifi ed that the slippery heptamer viral proteins. This would end up in the and the downstream region forming the colocalization of all required components secondary structure were essential for the in the same compartment of the cell. -1 PRF (Fig. 2 in IV). To be able to examine the timing Increasing evidence shows that both of the cleavage events, synchronous the nearby sequences as well as long- infections should be obtained. This would distance interactions may affect -1 PRF require that either viral RNA or infectious (Kollmus et al., 1994, Honda et al., 1996, complementary DNA (icDNA) would be Barry and Miller, 2002). In addition to the delivered to plant protoplasts to attain cis-acting signals, we examined whether suffi ciently high transfection effi ciency. some regions from the CfMV polyprotein This approach would also make metabolic region were required for effi cient -1 PRF. labeling of the translation products This was done with dual-reporter enzyme possible, which could ease the detection constructions, which can detect even small of transient processing intermediates. changes in -1 PRF frequencies (Stahl et The usage of icDNA would also allow al., 1995, Harger and Dinman, 2003). The the predicted processing sites to be beauty of this assay system resides in the fact mutagenized. Finally, a wider repertoire that the translation products can be easily of antisera would enable more precise quantifi ed by measuring their enzymatic identifi cation of the cleavage products. activities. Furthermore, experimental variation can be monitored as changes in 4.4 Synthesis of CfMV polyprotein the fi rst reporter activity. Thus, the fi rst reporter activity can be used to normalize 4.4.1 CfMV RNA programmed -1 the second reporter activity determining ribosomal frameshifting in WGE (IV, the effi cacy of the recoding event (Stahl V) et al., 1995). We selected three regions The strategy CfMV uses to produce its (A:1602-1720, B:1386-2137, and C:1551- RdRp differs from that of most other 1900) from the overlapping polyprotein sobemoviruses. Instead of polyprotein region of CfMV for the study (Fig. 1 in V). synthesis from a continuous ORF, The highest -1 PRF, ~36%, was measured CfMV polyprotein is produced from two from the longest CfMV sequence (B). In overlapping ORFs, 2A and 2B, via -1 PRF contrast, the medium-length region (C) (Mäkinen et al., 1995b). ORF2B encodes promoted -1 PRF with ~28% effi cacy,

54 Results and Discussion which was only slightly over the 25% -1 between CfMV and most other PRF frequency measured from the shortest sobemoviruses account for the fact that region (A) (Fig. 4B in V). This indicated less RdRp is produced in CfMV than that in CfMV the surrounding regions of in its relatives. This could indicate that the cis-acting signals may also infl uence CfMV RdRp is more operative and -1 PRF. One danger in the approach used therefore smaller amounts are needed. was that the varying C-terminal fusions in When CfMV replicase was changed to the transframe proteins could differentially be encoded from the 0-frame to resemble reduce the specifi c activity of the fi rst the genome organization found in the reporter used for normalization (Fig. 1A majority of sobemoviruses (reviewed in V); however, this was not the case here by Tamm and Truve, 2000a), the -1 PRF (Fig. 4A in V). Further shortening of the effi ciencies increased 4-5-fold (Fig. A region to 70 nt (CfMV RNA 1621- 2B in IV). However, the effect was not 1690) did not prevent -1 PRF (IV, Fig. specifi c for the CfMV sequences, since 2A; pJCL24). Thus, although high -1 PRF an equal increase was observed when a frequencies are usually obtained only with similar shift in frames was introduced to a pseudoknot structures (Dinman, 1995, reporter gene construct (pJCL24Δ), when Plant and Dinman, 2005), it appeared that long regions from the RdRp sequence in CfMV a simple stem-loop structure were deleted (pJCL16Δ), or when the was capable of promoting effi cient -1 gene rearrangements occurred after the PRF, because chemical probing of the entire P2A was synthesized (pJCL28, secondary structure revealed that the loop Table 1 in IV). Thus, the only thing does not interact with the surrounding these high -1 PRF effi ciency constructs regions (Tamm, 2000c). However, improved appeared to have in common was the -1 PRF from the B region suggested that shortening of the transframe product and some interactions leading to stimulated -1 the simultaneous increase in length of the PRF may have occurred between the cis- 0-frame product. Control experiments in acting signals and the ORF2A2B overlap. which the 0-frame product was gradually Alternatively, longer insertions between shortened (pJCL17Δ, pJCL17ΔMfeI, the reporter genes may have affected the pJCL17ΔSmaI) showed no change in the kinetics of the CfMV stem-loop folding -1 PRF frequencies, indicating that the 0- and induced more frequent translational frame length had no effect on the -1 PRF. pauses at the heptamer and thus enhanced Shortening of the transframe product -1 PRF. Purifi ed CfMV RNA appears to so that no gene rearrangements occurred, induce higher levels of -1 PRF than the alone improved -1 PRF 3-fold (pJCL22 polyprotein encoding region (Fig. 3A in in Fig. 2B in IV). Thus, only part of the IV). Therefore, we cannot exclude the increase in the -1 PRF frequency of possibility that regions outside the CfMV ORF rearrangement constructs could be polyprotein region may also affect -1 PRF. explained by relocation of the transframe For instance, interactions between the termination codon closer to the -1 PRF site. 3’UTR and the cis-acting stem-loop in Since a similar movement of termination BYDV regulate the switch from replication in the 0-frame did not affect -1 PRF to protein synthesis (Paul et al., 2001, much (pJCL22Δ in Fig. 2B in IV), the Barry and Miller, 2002). phenomenon appeared to be characteristic Differences in genome organization only for ribosomes shifting the frames.

55 Results and Discussion

Termination is a slower process than highly conserved nature of the elongation elongation and it induces pausing and phase from prokaryotes to higher stacking of ribosomes upstream from the eukaryotes, recoding events such as -1 PRF termination codons (Wolin and Walter, and the termination codon readthrough 1988). In WGE up to 10 ribosomes were used by viruses infecting mammalian and found stacked behind a termination codon plant cells can be recapitulated in yeast and each ribosome protected an ~27-29-nt (Stahl et al., 1995, Harger and Dinman, region from the mRNA (Wolin and Walter, 2003, Bekaert et al., 2005). Although 1988). For instance, in MuLV introduction identical mechanisms of -1 PRF are used of a stop codon 48 nt downstream from the in eukaryotes and prokaryotes, the effi cacy recoding site stimulated the readthrough of the process may differ (Garcia et al., 5-fold (Wills et al., 1991). In our studies 1993, Napthine et al., 2003). Thus, instead high -1 PRF frequencies were obtained of using plant cells for the in vivo studies, even if the termination codon was located we performed the studies in S. cerevisiae 462 nt downstream from the stem-loop and E. coli for convenience. The A, B, structure (IV). This would mean that and C regions from the CfMV polyprotein approximately 16 ribosomes should be were analyzed with dual-reporter vectors, queuing behind the termination codons in which the regions tested were inserted if the increase in -1 PRF was due to between the lacZ and luc genes (Stahl et stacking. We can hypothesize that the al., 1995, Fig. 1A in V). transframe ribosomes have a slower pace As expected, the CfMV -1 PRF than ribosomes translating the 0-frame, signals were functional both in yeast due to the translational pause at the -1 and in bacterial cells, although -1 PRF PRF signals. However, a major fraction of occurred at lower frequencies in the the ribosomes passes the -1 PRF signals prokaryotic cells (Fig. 2 in V). In general, without pause (Lopinski et al., 2001). the XXXAAAC heptamers function These ribosomes may possibly move more ineffi ciently in prokaryotes (Garcia et al., rapidly and thus could easily reach the 1993, Brierley et al., 1997, Napthine et al., trailing slow transframe ribosomes. This 2003), even though the prokaryotic Asp- could induce stacking, and a translational tRNA encoding the AAC triplet promotes pause at a nearby termination codon effi cient -1 PRF in eukaryotic ribosomes. would give a ribosome occupying the -1 This indicates that the differences in -1 PRF signal an opportunity to shift frames, PRF frequencies between eukaryotes even though the secondary structure was and prokaryotes must arise from the opened. translational apparatus itself (Napthine et al., 2003). Thus, further analyses were 4.4.2 The -1 PRF in vivo (V) performed in yeast. The -1 PRF is highly dependent on the Improved -1 PRF was measured kinetics of both translation initiation as for longer CfMV regions in vivo (Fig. 2 well as elongation (Barry and Miller, 2002, in V), comparing favorably with the in Harger et al., 2002). However, protein vitro assays performed with Ruc-luc dual- synthesis is signifi cantly slower under in reporter RNAs. Previously, it was shown vitro conditions than in vivo (Lopinski et that under in vitro conditions a fraction of al., 2000). Thus, we extended our studies the ribosomes terminates at the slippery to include in vivo conditions. Due to the heptamer during the ribosomal pause

56 Results and Discussion

(Lopinski et al., 2000, Plant et al., 2003). spacers of -1 PRF signals promoting high Our protein analysis showed that similar levels of -1 PRF (Bekaert et al., 2003). premature termination products were also produced in vivo (Fig. 3 in V). These 4.4.3. Regulation of -1 PRF by CfMV termination products appeared to be more proteins (V) abundant when the longer CfMV regions The correct ratio between RdRp and the 0- were programming -1 PRF. However, frame products is vital for viruses using -1 similar -1 PRF frequencies were obtained, PRF (Dinman and Wickner, 1992, Hung irrespective of whether the additional et al., 1998, Barry and Miller, 2002). CfMV sequences were placed upstream or However, the need for RdRp probably downstream from the -1 PRF site (Fig. 2C varies during infection. At the initial in V). Thus, the longer -1 PRF cassettes stages the full-length gRNA is needed for may have affected the folding of the stem- production of proteins with enzymatic loop and induced longer or more frequent properties, whereas at the later stages translational pauses. Higher frequencies of CP encoded from the sgRNA is needed ribosomal pausing would have presumably for particle formation. Thus, it could be induced termination and -1 PRF at the benefi cial for the viruses to regulate the heptamer. effi cacy of the -1 PRF to fi t the need for The current hypothesis suggests that the amount of RdRp. Thus far no viral or stem-loops are less resistant to unwinding cellular proteins are known to be directly than pseudoknots. This results in shorter involved in the regulation of -1 PRF. pausing at the -1 PRF signals (Dinman, However, the +1 PRF event encoding 1995, Plant et al., 2003, 2005). Keeping mammalian ornithine decarboxylase this in mind, the CfMV -1 PRF can be (ODC) antizyme is up-regulated by regarded as unusually high compared with increased cellular concentrations of reported effi ciencies of ~1-3% from other polyamines, which are the biosynthesis -1 PRF signals with stem-loop structures products of ODC (Matsufuju et al., 1995). (Prüfer et al., 1992, Stahl et al., 1995). One The antizyme binds to the enzyme and reason for the high -1 PRF frequency in directs it to degradation. Thus, ODC CfMV is the heptamer UUUAAAC, which concentration is reduced and polyamine directs effi cient -1 PRF in eukaryotic cells synthesis becomes down-regulated. We (Brierley et al., 1987, Napthine et al., tested whether CfMV proteins produced via 2003). The fi rst triplet of the heptamer -1 PRF could participate in the regulation plays a multiplicative role in determining of -1 PRF. We coexpressed CfMV RdRp or the -1 PRF effi ciency, and U at this position P27 (the C-terminus of P2A) together with promotes the highest -1 PRF frequencies the lacZ-luc dual-reporter vectors carrying (Bekaert et al., 2003). The Asp-tRNA the minimal CfMV -1 PRF region A or the encoding the AAC triplet also promotes corresponding inframe control, Am (Fig. 1 effi cient -1 PRF and also dictates that in V). The effect of the CfMV proteins on slippage occurs via the simultaneous dual- -1 PRF was then evaluated by monitoring tRNA slippage mechanism (Napthine the changes in reporter gene expression. et al., 2003). Finally, the CfMV spacer Control coexpressions were performed sequence between the cis-acting signals with empty expression plasmids or with also contains bases that are often found in constructs in which the translation initiation codon of P27 or RdRp was deleted.

57 Results and Discussion

Neither of the proteins affected the to unspecifi c but to specifi c binding of P27 expression of the fi rst reporter (lacZ) to a certain part of the downstream RNA. since there was no signifi cant difference We believe that this binding occurred at the in the β-galactosidase activities between CfMV -1 PRF region, which resulted in the protein coexpressions and control inhibition of downstream translation. Thus, expression with the empty expression during CfMV infection the accumulating plasmids (Table 2 in V) or with the AUG amounts of P27 in the cells could indicate deletion mutants (data not shown). In that no additional RdRp is needed, whereas contrast, both proteins appeared to affect subsequent processing of P27 could the downstream reporter expression to relieve the block. Alternatively, P27 could some extent when LUC activities were prevent -1 PRF, rendering the 3’proximal compared with those measured from the part of CfMV RNA free of ribosomes. coexpressions performed in the presence This would enable the 3’proximal end of the empty expression vector. However, to function as a template for sgRNA coexpressions with RepΔAUG showed synthesis in the absence of collisions that the RdRp expression did not affect with ribosomes moving in the opposite -1 PRF specifi cally. In contrast, similar direction. However, it is not yet known comparisons with P27 and P27ΔAUG whether CfMV sgRNA is synthesized expressions indicated a reproducible from the full-length (-) strand or from a reduction in the amount of LUC produced shorter template. In BYDV, translation in the presence of CfMV P27. The initiation and -1 PRF are regulated via repressing effect of P27 was stronger on direct interaction of specifi c elements in the inframe control than on the actual -1 the 3’UTR with complementary regions in PRF construct. This, however, resulted the 5’UTR and the stem-loop of the -1 PRF from the fact that since only ~15% of site (Barry and Miller, 2002). Initiation the elongating ribosomes translate the of the (-) strand synthesis from the 3’end transframe protein in the -1 PRF test of the (+) strands disrupts these base- constructs, the effect observed in LUC pairings and prevents translation initiation expression will also be relatively lower and -1 PRF. Further rounds of replication and more diffi cult to detect. would produce excess amounts of (+) The function of P27 is unknown, strands, which would then outcompete but it contains motifs for RNA binding. the RdRp molecules and be free to form In fact, P2A binds RNA in an unspecifi c the long-distance interactions needed for manner, putatively via this motif (Tamm translation. However, it is clear that the and Truve, 2000b). Since the synthesis of importance of P27 for -1 PRF and virus the fi rst reporter was not affected by P27 infection, and thus the specifi city of P27 expression, the effect of P27 was not due binding, should be studied further.

58 Concluding Remarks

5. CONCLUDING REMARKS

All the viral sequences studied, namely translation initiation in WGE. Depletion AMV 5’UTR, CfMVε, CrTMV IRES, of WGE from eIFs interacting with the PVXαβ, and TMVΩ, programmed higher cap analogue or the poly(A) sequence gene expression in tobacco compared with suggested that internal initiation was the 5’UTR composed of a polylinker. In eIF4E-independent but eIF4G-dependent. contrast, only CfMVε led to the production Poly(A) addition reduced the 3’proximal of higher protein yields in barley. cistron translation mostly from the Expression from constructs containing reference mRNA, whereas the viral CfMVε was ∼12-fold higher than that sequences recruited the eIFs interacting from TMVΩ-containing constructs, even with the poly(A) sequence more effi ciently. though TMVΩ functions as a translational However, in vivo studies revealed that enhancer in other monocots such as rice and the IRES activity of CfMVε and CrTMV maize. This suggests that the requirements was low, at least in barley and yeast. for effi cient gene expression may differ in Furthermore, our studies showed that barley. Transient expression studies with the combination of plant viral sequences luc mRNAs having the reference 5’UTR, with dicistronic reporter gene expression CfMVε, or TMVΩ as the 5’leader showed plasmids led to unpredictable behavior of comparable LUC accumulation from all the constructs and, thus, this approach was mRNAs. Therefore, the capacity of viral not applicable to yeast. sequences to enhance reporter expression After successful binding of the did not result merely from promoted preinitiation complexes to the 5’end of translation and the mechanism should be CfMV RNA, leaky scanning brings the studied further. Introduction of an uORF to preinitiation complex into the region the CfMVε abolished downstream reporter encoding CfMV polyprotein. The CfMV expression, suggesting that translation RdRp is synthesized via -1 PRF as initiation from CfMV RNA involves the C-terminal part of the transframe scanning. The fi nding that destabilization polyprotein. Thus, the occurrence of -1 of the putative 5’proximal structure PRF is extremely important for the viral improved downstream gene expression viability. Interestingly, CfMV protein further supported this observation. P27 repressed translation of proteins Competition assays performed in vitro encoded downstream from the -1 PRF showed that CfMVε did not compete as signals. Thus, P27 may play a role in successfully against eIFs as CrTMV IRES regulating the amount of RdRp produced. and TMVΩ. This may have resulted from Alternatively, P27 may function in the fact that the 5’proximal stem-loop regulating the ribosomal load at the structure in CfMVε renders translation 3’end of the RNA. Proteolytic processing initiation dependent on the complete set of releases the functional domains from the eIFs, whereas initiation from tobamoviral CfMV polyproteins. However, the E/N sequences may occur in the absence of site used to process the CfMV VPg differs some eIFs. In fact, translation initiation from the E/T or E/S sites recognized by from TMVΩ was reported to be eIF4E- the proteases of other sobemoviruses. independent (Gallie, 2002a). Interestingly, The CfMV polyprotein did not contain all these sequences also promoted internal a suffi cient number of E/N sites, which

59 Results and Discussion would have led to the complete processing relative to other viral sequences studied of the polyprotein. Thus, additional indicates that CfMVε could be utilized cleavage sites are clearly utilized. The in biotechnological applications to search for other putative cleavage sites increase heterologous protein expression suggested that the E/T or E/S sites could in cereals. Furthermore, CfMV -1 PRF also be used in CfMV for some processing signals could be utilized to synthesize a events. The usage of these sites could be fraction of proteins with certain C-terminal verifi ed by mutating the suspected sites in fusions. For instance, introduction of tags the icDNA of CfMV. would enable affi nity purifi cation and Although, CfMVε did not function immunodetection of a certain percentage at the translational level, the clear of the expressed protein. improvement of gene expression in barley

60 Acknowledgements

6. ACKNOWLEDGEMENTS

This study was carried out at the Institute of The previous and the current members Biotechnology and Department of Applied of the lab: Andres M., Kostya I., Pietri Biology, University of Helsinki, during P., Jimmy L., Deyin G., Minna R., Eva the years 1999-2005. Financial founding W., Anders H., Kimmo R. and Rasa G. provided by the Academy of Finland, the are thanked for creating splendid and University of Helsinki, and the National humoristic working environment. I am Technology Agency of Finland (TEKES) indepted to Andres, who initially thought is highly appreciated. me patiently the methods of molecular My warmest thanks go to my biology as well as the basics of cloning. supervisor Kristiina Mäkinen, who gave My special thanks go to my closest me the opportunity to carry out this colleagues in the group, Pietri and Kostya. work. During our years together she has Their skilfull assistance in the case of provided me constant guidance, support, technical problems during experiments and encouragement. I also appreciate is much appreciated. For Rasa, Anders, the positive attitude, amazing memory, and Kimmo, I wish good luck with your and endless fl ow of new ideas. I also projects, productive collaborations, and want to thank for the possibility to work nice karonkkas in the near future. Julia independently and freely. P. and Anne S. from the animal virus I would like to thank Prof. Mart laboratory are greatly appreciated of joined Saarma, the director of Institute of courses, lunches, and congress trips. I also Biotechnology, for providing excellent wish to acknowledge all the co-authors for facilities during the years 1999-2001. their contribution to our mutual papers. Equally, the help and support from the Warm thanks also go to my friends, personnel of SBL during years 2002- Jonna J., Jonna P., Petra, Soile, and Leena, 2005 is highly appreciated. Prof. Mirja who have actively organized events, which Salkinoja-Salonen is thanked for giving have forced me to switch from science me the opportunity to carry out my PhD to free time. Also Ninnu and Merlin are studies at the Department of Applied thanked for giving me an opportunity to Chemistry and Microbiology. I also wish relax and have some phycical exercise to thank the personnel of Viikki Graduate after work. School in Biosciences for providing many Finally, my very special thanks go to interesting courses. Mikko Frilander and my family. My parents Tuovi and Arvo Tero Ahola, the follow up group of my have provided me constant trust and PhD project, are warmly thanked for the support but also tought me to work hard. annual meetings and helpful comments. My sister Kirsi and his husband Marko are I also wish to thank reviewers docent thanked for their encouragement. Most of Maija Vihinen-Ranta and professor Carl- all, I want to thank Harri for the interest, Henrik von Bonsdorff for the time they endless inspiration and support. spent reading my text as well as for the comments, which led to the improvement Helsinki, March 2006 of the text.

61 References

7. REFERENCES Akbergenov R.Z., Zhanybekova S.S., Brierley I., Meredith M.R., Bloys A.J., Kryldakov R.V., Zhigailov A., Polimbetova and Hagervall T.G. 1997. Expression of a N.S., Hohn T., and Iskakov B.K. 2004. ARC- coronavirus ribosomal frameshift signal in 1, a sequence element complementary to Escherichia coli: infl uence of tRNA anticodon an internal 18S rRNA segment, enhances modifi cation on frameshifting. J. Mol. Biol. translation effi ciency in plants when present in 270: 360-373. the leader or intercistronic region of mRNAs. Brown C.M., Dinesh-Kumar S.P., and Miller Nucleic Acids Res. 32: 239-247. W.A. 1996. Local and distant sequences are Baranov P.V., Gurvich O.L., Hammer A.W., required for effi cient readthrough of the barley Gesteland R.F., and Atkins J.F. 2003. RECODE yellow dwarf virus PAV coat protein gene stop 2003. Nucleic Acids Res. 31: 87-89. codon. J. Virol. 70: 5884-5892. Barry J.K., and Miller W.A. 2002. A -1 Browne G.J., and Proud C.G. 2002. Regulation ribosomal frameshift element that requires of peptide-chain elongation in mammalian base pairing across four kilobases suggests cells. Eur. J. Biochem. 269: 5360-5368. a mechanism of regulating ribosome and Browning K.S. 2004. Plant translation replicase traffi c on a viral RNA. Proc. Natl. initiation factors: it is not easy to be green. Acad. Sci. USA 99: 11133-11138. Biochem. Soc. Trans. 32: 589-591. Bekaert M., Bidou L., Denise A., Duchateau- Browning K.S., Lax S.R., Humphreys J., Nguyen G., Forest J.P., Froidevaux C., Hatin Ravel J.M., Jobling S.A., and Gehrke L. 1988. I., Rousset J.P., and Termier M. 2003. Towards Evidence that the 5’-untranslated leader of a computational model for -1 eukaryotic mRNA affects the requirement for wheat frameshifting sites. Bioinformatics 19: 327- germ initiation factors 4A, 4F, and 4G. J. Biol. 335. Chem. 263: 9630-9634. Bekaert M., and Rousset J.P. 2005. An Brunelle M.N., Payant C., Lemay G., and extended signal involved in eukaryotic -1 Brakier-Gingras L. 1999. Expression of the frameshifting operates through modifi cation human immunodefi ciency virus frameshift of the E site tRNA. Mol. Cell. 17: 61-68. signal in a bacterial cell-free system: infl uence Benne R., and Hershey J.W.B. 1978. The of an interaction between the ribosome and mechanism of action of protein synthesis a stem-loop structure downstream from the initiation factors from rabbit reticulocytes. J. slippery site. Nucleic Acids Res. 27: 4783- Biol. Chem. 253: 3078-3087. 4791. Benne R., Wong C., Luedi M., and Hershey Bu X., Haas D.W., and Hagedorn C.H. 1993. J.W.B. 1976. Purifi cation and characterization Novel phosphorylation sites of eukaryotic of initiation factor IF-E2 from rabbit initiation factor-4F and evidence that reticulocytes. J. Biol. Chem. 23: 7675-7681. phosphorylation stabilizes interactions of the p25 and p220 subunits. J. Biol. Chem. 268: Bertram G., Innes S., Minella O., Richardson 4975-4978. J., and Stansfi eld I. 2001. Endless possibilities: translation termination and stop codon Bustamante, P.I., and Hull, R. 1998. Plant recognition. Microbiology 147: 255-269. virus gene expression strategies. Electronic Journal of Biotechnology 1(2). Bilgin D.D., Liu Y., Schiff M., and Dinesh- Kumar S.P. 2003. P58(IPK), a plant ortholog Caponigro G., and Parker R. 1996. of double-stranded RNA-dependent protein Mechanisms and control of mRNA turnover kinase PKR inhibitor, functions in viral in Saccharomyces cerevisiae. Microbiol. Rev. pathogenesis. Dev. Cell 4: 651-661. 60: 233-249.

62 References

Carlson B.A., Mushinski J.F., Henderson Dinman J.D. 1995. Ribosomal frameshifting D.W., Kwon S.Y., Crain P.F., Lee B.J., and in yeast viruses. Yeast. 11: 1115-1127. Hatfi eld D.L. 2001. 1-Methylguanosine in Dinman J.D., Ruiz-Echevarria M.J., and Peltz place of Y base at position 37 in phenylalanine S.W. 1998. Translating old drugs into new tRNA is responsible for its shiftiness in treatments: ribosomal frameshifting as a target retroviral ribosomal frameshifting. Virology. for antiviral agents. Trends Biotechnol. 16: 279: 130-135. 190-196. Chappell S.A., Edelman G.M., and Mauro Dorokhov Y.L., Skulachev M.V., Ivanov V.P. 2000. A 9-nt segment of a cellular mRNA P.A., Zvereva S.D. Tjulkina L.G., Merits A., can function as an internal ribosome entry site Gleba Y.Y., Hohn T., and Atabekov J.G. 2002. (IRES) and when present in linked multiple Polypurine (A)-rich sequences promote cross- copies greatly enhances IRES activity. Proc. kingdom conservation of internal ribosome Natl. Acad. Sci. U S A 97: 1536-1541. entry. Proc. Natl. Acad. Sci. USA 99: 5301- Chappell S.A., Edelman G.M., and Mauro 5306. V.P. 2004. Biochemical and functional analysis Dougherty W.G., and Semler B.L. 1993. of a 9-nt RNA sequence that affects translation Expression of virus-encoded proteinases: effi ciency in eukaryotic cells. Proc. Natl. Acad. functional and structural similarities with Sci. U S A 101: 9590-9594. cellular enzymes. Microbiol. Rev. 57: 781- Corpet F. 1988. Multiple sequence alignment 822. with hierarchical clustering. Nucleic Acids Drugeon G., Jean-Jean O., Frolova L., Le Res. 16: 10881-10890 Goff X., Philippe M., Kisselev L., and Haenni Coward P., and Dasgupta A. 1992. Yeast cells A. 1997. Eukaryotic release factor 1 (eERF1) are incapable of translating RNAs containing abolishes readthrough and competes with the poliovirus 5’untranslated region: evidence suppressor tRNAs at three termination codons for a translational inhibitor. J. Virol. 66: 286- in messenger RNA. Nucleic Acids Res. 1997: 295. 2254-2258. Crum C.J., Hu J., Hiddinga H.J., and Roth Dreher T.W. 1999. Functions of the 3’- D.A. 1988. Tobacco mosaic virus infection untranslated regions of positive stranded stimulates the phosphorylation of a plant regions of positive strand RNA viral genomes. protein associated with double-stranded RNA- Annu. Rev. Phytopathol. 37: 151-174 dependent protein kinase activity. J. Biol. Dulude D., Baril M., and Brakier-Gingras Chem. 263: 13440-13443. L. 2002. Characterization of the frameshift Dever T.E. 1999. Translation initiation: adept stimulatory signal controlling a programmed at adapting. Trends Biochem. Sci. 24: 398– -1 ribosomal frameshift in the human 403. immunodefi ciency virus type 1. Nucleic Acids Res. 30: 5094-5102. Dinman J.D., Icho T., and Wickner R.B. 1991. A -1 ribosomal frameshift in a double-stranded Everett J.G., and Gallie D.R. 1992. RNA RNA virus of yeast forms a gag-pol fusion delivery in Saccharomyces cerevisiae using protein. Proc. Natl. Acad. Sci. USA 88: 174- electroporation. Yeast 8: 1007-1014. 178. Evstafi eva AG, Beletsky AV, Borovjagin AV, Dinman J.D., and Wickner R.B. 1992. and Bogdanov AA. 1993. Internal ribosome Ribosomal frameshifting effi ciency and gag/ entry site of encephalomyocarditis virus RNA gag-pol ratio are critical for yeast M1 double- is unable to direct translation in Saccharomyces stranded RNA virus propagation. J. Virol. 66: cerevisiae. FEBS Lett. 335: 271-276. 3669-3676.

63 References

Farabaugh P.J. 1996. Programmed Gallie D.R. 1998. A tale of two termini: A translational frameshifting. Microbiol. Rev. functional interaction between the termini of an 60: 103-134. mRNA is a prerequisite for effi cient translation initiation. Gene 216: 1-11. Frolova L., Le Goff X., Rasmussen H.H., Cheperegin S., Drugeon G., Kress M., Arman Gallie D.R., and Browning K.S. 2001. eIF4G I., Haenni A.L., Celis J.E., and Philippe M. functionally differs from eIFiso4G in promoting 1994. A highly conserved eukaryotic protein internal initiation, cap-independent translation, family possessing properties of polypeptide and translation of structured mRNAs. J. Biol. chain release factor. Nature 372: 701-703. Chem. 276: 36951-36960. Gale M. Jr., Tan S.L., and Katze M.G. 2000. Gallie D.R. 2001. Cap-independent translation Translational control of viral gene expression conferred by the 5’ leader of tobacco etch virus in eukaryotes. Microbiol. Mol. Biol. Rev. 64: is eukaryotic initiation factor 4G dependent. J. 239-280. Virol. 75: 12141-12152. Gallie D.R., Sleat D.E., Watts J.W., Turner Gallie D.R. 2002a. The 5’-leader of tobacco P.C., and Wilson T.M. 1987. A comparison mosaic virus promotes translation through of eukaryotic viral 5’-leader sequences as enhanced recruitment of eIF4F. Nucleic Acids enhancers of mRNA expression in vivo. Res. 30: 3401-3411. Nucleic Acids Res.15: 8693-8711. Gallie D.R. 2002b. Protein-protein interactions Gallie D.R., Lucas W.J., and Walbot V. required during translation. Plant Mol. Biol. 1989. Visualizing mRNA expression in plant 50: 949-970. protoplasts: factors infl uencing effi cient Garcia A., Van Duin J., and Pleij C.W. 1993. mRNA upstake and translation. Plant Cell 1: Differential response to frameshift signals 301-311. in eukaryotic and prokaryotic translational Gallie D.R. 1991. The cap and poly(A) tail systems. Nucleic Acids Res. 21: 401-406. function synergistically to regulate mRNA Gingras A., Svitkin Y., Belsham G.J., Pause translational effi ciency. Genes Dev. 5: 2108- A., and Sonenberg N. 1996. Activation of the 2116. translational suppressor 4E-BP1 following Gallie D.R, and Tanguay R. 1994. Poly(A) infection with encephalomyocarditis virus binds to initiation factors and increase cap- and poliovirus. Proc. Natl. Acad. Sci. USA 93: dependent translation in vitro. J. Biol. Chem. 5578-5583. 269: 17166-17173. Guo L., Allen E.M., and Miller W.A. Gallie D.R, Tanguay R.L., and Leathers V. 2001. Base-pairing between untranslated 1995. The tobacco etch viral 5’ leader and regions facilitates translation of uncapped, poly(A) tail are functionally synergistic nonpolyadenylated viral RNA. Mol. Cell 7: regulators of translation. Gene 165: 233-238. 1103-1109. Gallie D.R. 1996. Translational control of Goodfellow I., Chaudhry Y., Gioldasi I., cellular and viral mRNAs. Plant Mol. Biol. 32: Gerondopoulos A., Natoni A., Labrie L., 145-158. Laliberte J., and Roberts L. 2005. Calicivirus translation initiation requires an interaction Gallie D.R., Le H., Caldwell C., Tanguay R.L., between VPg and eIF4E. EMBO Rep. 6: 968- Hoang N.X., and Browning K.S. 1997. The 972. phosphorylation state of translation initiation factors is regulated developmentally and Gorbalenya A.E., Koonin E.V., Blinov V.M., following heat shock in wheat. J. Biol. Chem. and Donchenko A.P. 1988. Sobemovirus 272: 1046-1053. genome appears to encode a serine protease related to cysteine proteases of picornaviruses. FEBS Lett. 236: 287-290. 64 References

Gorbalenya A.E., Donchenko A.P., Blinov Horsfi eld J.A., Wilson D.N., Mannering V.M. and Koonin E.V. 1989. Cysteine proteas- S.A., Adamski F.M., and Tate W.P. 1995. es of positive strand RNA viruses and chymo- Prokaryotic ribosomes recode the HIV-1 trypsin-like serine proteases. A distinct protein gag-pol-1 frameshift sequence by an E/P site superfamily with a common structural fold. post-translocation simultaneous slippage FEBS Lett. 243: 103-114. mechanism. Nucleic Acids Res. 23: 1487- 1494. Grifo J.A., Tahara S.M., Morgan M.A., Shatkin A.j., and Merrick W.C. 1983. New initiation Hung M., Patel P., Davis S., and Green S.R. factor activity required for globin mRNA 1998. Importance of ribosomal frameshifting translation. J. Biol. Chem. 258: 5804-5810. for human immunodefi ciency virus type 1 particle assembly and replication. J. Virol. 72: Hacker D.L., and Sivakumaran K. 1997. 4819-4824. Mapping and expression of Southern bean mosaic virus genomic and subgenomic RNAs. Ivanov P.A., Karpova O.V., Skulachev M.V., Virology 234: 317-327. Tomashevskaya M.V., Rodionova N.P., Dorokhov Y.L., and Atabekov J.G. 1997. A Hammell A.B. Taylor R.C., Peltz S.W., and Tobamovirus genome that contains an internal Dinman J.D. 1999. Identifi cation of putative ribosome entry site functional in vitro. Virology programmed -1 ribosomal frameshift signals 232: 32-43. in large DNA databases. Genome Res. 9: 417- 427. Jaag H.M., Kawchuk L. Rohde W., Fischer R., Emans N., and Prüfer D. 2003. An unusual Han B., and Zhang J. 2002. Regulation of gene internal ribosomal entry site of inverted expression by internal ribosome entry sites or symmetry directs expression of a potato leafroll cryptic promoters: the eIF4G story. Mol. Cell polerovirus replication-associated protein. Biol. 22: 7372-7384. Proc. Natl. Acad. Sci. USA 100: 8939-8944. Harger J.W., and Dinman J.D. 2003. An in Jacks T., and Varmus H.E. 1985. Expression of vivo dual-luciferase assay system for studying the Rous sarcoma virus pol gene by ribosomal translational recoding in yeast Saccharomyces frameshifting. Science 230: 1237-1242. cerevisiae. RNA 9: 1019-1024. Jacks T., Townsley K., Varmus H.E, and Harger J.W., Meskauskas A., and Dinman J.D. Majors J. 1987. Two effi cient ribosomal 2002. An “integrated model” of programmed frameshifting events are required for synthesis ribosomal frameshifting. Trends Biochem. Sci. of mouse mammary tumor virus gag-related 27: 448-454. polyproteins. Proc. Natl. Acad. Sci. U S A 84: Hecht K., Bailey J.E., and Minas W. 2002. 4298-4302. Polycistronic gene expression in yeast versus Jacks T., Power M.D., Masiarz F.R., Luciw cryptic promoter. FEMS Yeast Res. 2: 215- P.A., Barr P.J., and Varmus H.E. 1988a. 224. Characterization of ribosomal frameshifting Hennecke M, Kwissa M, Metzger K, Oumard in HIV-1 gag-pol expression. Nature 331: 280- A, Kroger A, Schirmbeck R, Reinmann J, and 283. Hauser H. 2001. Composition and arrangement Jacks T., Madhani H.D., Masiarz F.R., and of genes defi ne the strength of IRES-driven Varmus H.E. 1988b. Signals for ribosomal translation in bicistronic mRNAs. Nucleic frameshifting in the Rous sarcoma virus gag- Acids Res. 29: 327-3334. pol region. Cell 55: 447-458. Honda A., and Nishimura S. 1996. Suppression Jobling S.A., and Gehrke L. 1987. Enhanced of translation frameshift by upstream translation of chimaeric messenger RNAs termination codon. Biochem. Biophys. Res. containing a plant viral untranslated region. Commun. 221: 602-608. Nature 325: 622-625.

65 References

Johnston M., and Davis R.W. 1984. Sequences Kontos H., Napthine S., and Brierley I. 2001. that regulate the divergent GAL1-GAL10 Ribosomal pausing at a frameshifter RNA promoter in Saccharomyces cerevisiae. Mol. pseudoknot is sensitive to reading phase Cell. Biol. 4: 1440-1448. but shows little correlation with frameshift effi ciency. Mol. Cell Biol. 21: 8657-8670. Jore J., de Geus B., Jackson R.J., Pouwels P.H., and Enger-Valk B.E. 1988. Poliovirus protein Kozak M. 1987. Effects of the intercistronic 3CD is the active protease for processing of length on the effi ciency of reinitiation by the precursor protein P1 in vitro. J. Gen. Virol. eukaryotic ribosomes. Mol. Cell Biol. 7: 3438- 69: 1627-1636. 3445. Joseph S. 2003. After the ribosome structure: Kozak M. 1989. The scanning model for how does translocation work? RNA 9: 160- translation: an update. J. Cell Biol. 108(2): 164. 229-41. Kapp L.D., and Lorsch J.R. 2004. The Kozak M. 2001. Constraints on reinitiation of molecular mechanics of . translation in mammals. Nucleic Acids Res. Annu. Rev. Biochem. 73: 657-704. 29: 5226-5232. Kerekatte V., Keiper B.D., Badorff C., Cai Kozak M. 2002. Pushing the limits of A., Knowlton K.U., and Rhoads R.E. 1999. the scanning mechanism for initiation of Cleavage of Poly(A)-binding protein by translation. Gene 299: 1-34. coxsackievirus 2A protease in vitro and in Kozak M. 2003. Alternative ways to think about vivo: another mechanism for host protein mRNA sequences and proteins that appear to synthesis shutoff? J. Virol. 73: 709-717. promote internal initiation of translation. Gene Kim Y.G., Maas S., and Rich A. 2001. 318: 1-23. Comparative mutational analysis of cis-acting Krab I.M., Caldwell C., Gallie D.R., and Bol RNA signals for translational frameshifting in J.F. 2005. Coat protein enhances translational HIV-1 and HTLV-2. Nucleic Acids Res. 29: effi ciency of Alfalfa mosaic virus RNAs 1125-1231. and interacts with the eIF4G component of Kim YG, Su L, Maas S, O’Neill A, and Rich initiation factor eIF4F. J. Gen. Virol. 86: 1841- A. 1999. Specifi c mutations in a viral RNA 1849. pseudoknot drastically change ribosomal Lamphear B.J., Kirchweger R., Skern T., frameshifting effi ciency. Proc. Natl. Acad. Sci. and Rhoads R.E. 1995. Mapping of functional USA 96: 14234-14239. domains in eukaryotic protein synthesis Koh D.C., Wong S.M., and Liu D.X. 2003. initiation factor 4G (eIF4G) with picornaviral Synergism of the 3’-untranslated region and proteases. Implications for cap-dependent and an internal ribosome entry site differentially cap-independent translational initiation. J. enhances the translation of a plant virus coat Biol. Chem. 270: 21975-21983. protein. J. Biol. Chem. 278: 20565-20573. Léonard S., Viel C., Beauchemin C., Kollmus H., Honigman A., Panet A., and Daigneault N., Fortin M.G., and Laliberté F. Hauser H. 1994. The sequences of and distance 2004. Interaction of VPg-Pro of Turnip mosaic between two cis-acting signals determine the virus with the translation initiation factor 4E effi ciency of ribosomal frameshifting in human and the poly(A) binding protein in planta. J. immunodefi ciency virus type 1 and human T- Gen. Virol. 85: 1055-1063. cell leukemia virus type II in vivo. J. Virol. 68: Levis C., and Astier-Manifacier S. 1993. The 5’ 6087-6091. untranslated region of PVY RNA, even located in an internal position, enables initiation of translation. Virus Genes 7: 367-79.

66 References

Li G., and Rice C.M. 1993. The signal for Martinez-Salas E., Ramos R., Lafuente E., translational readthrough of a UGA codon in and Lopes de Quinto S. 2001. Functional Sindbis virus RNA involves a single cytidine interactions in internal translation initiation residue immediately downstream of the directed by viral and cellular IRES elements. termination codon. J. Virol. 67: 5062-5067. J. Gen. Virol. 82: 973-984. Lin T., Kong X., Haystead T.A.J., Pause A., Matsuda D., and Dreher T.W. 2004. The Belsham G., Sonenberg N., and Lawrence J.C. tRNA-like structure of Turnip yellow mosaic 1994. PHAS-1 as a link between mitogen- virus RNA is a 3’-translational enhancer. activated protein kinase and translation Virology 321: 36-46. initiation. Science 266: 653-656. Matsufuji S., Matsufuji T., Miyazaki Y., Lokesh G.L., Gopinath K., Satheshkumar P.S., Murakami Y., Atkins J.F., Gestland R.F., and Savithri H.S. 2001. Complete nucleotide and Hayashi S-I. 1995. Autoregulatory sequence of Sesbania mosaic virus: a new frameshiting in decoding mammalian ornithine virus species of the genus Sobemovirus. Arch. decarboxylase antizyme. Cell 80: 51-60. Virol. 146: 209-223. Matveeva O.V., and Shabalina S.A. 1993. Lopez P.J. and Séraphin B. 2000. YIDB: The Intermolecular mRNA-rRNA hybridisation yeast intron database. Nucleic Acids Res. 28, and the distribution of potential regions in 85-86. murine 18S rRNA. Nucleic Acids Res. 21: 1007-1011. Lopinski J.D., Dinman J.D., and Bruenn J.A. 2000. Kinetics of ribosomal pausing during Mauro V.P., and Edelman G.M. 1997. rRNA- programmed -1 translational frameshifting. like sequences occur in diverse primary Mol. Cell. Biol. 20: 1095-1103. transcripts: implications for the control of gene expression. Proc. Natl. Acad. Sci. USA Lu H., Li W., Noble W.S., Payan D., and 94: 422-427. Anderson D.C. 2004. Riboproteomics of the Hepatitis C virus internal ribosomal entry site. McCaughan K.K., Brown C.M., Dalphin J. Proteome Res. 3: 949-957. M.E., Berry M.J., and Tate W.P. 1995. Translational termination effi ciency in Lukaszewicz M., Feuermann M., Jerouville mammals is infl uenced by the base following B., Stas A., and Boutry M. 2000. In vivo the stop codon. Proc. Natl. Acad. Sci. USA 92: evaluation of the context sequence of the 5431-5435. translation initiation codon in plants. Plant Sci. 54: 89-98. Merits A., Rajamäki M-L., Lindhom P., Runeberg-Roos P., Kekarainen T., Puustinen Majumdar R., Bandyopadhyay A., and Maitra P., Mäkeläinen K., Valkonen J.P.T., and Saarma U. 2003. Mammalian translation initiation M. 2002. Proteolytic processing of potyviral factor eIF1 functions with eIF1A and eIF3 in polyprotein processing intermediates in insect the formation of a stable 40 S preinitiation and plant cells. J. Gen. Virol. 83: 1211-1221. complex. J. Biol. Chem. 278: 6580-6587. Meskauskas A., Harger J.W., Jacobs Marcotrigiano J., Gingras A.C., Sonenberg K.L., and Dinman J.D. 2003. Decreased N., Burley S.K. 1997. Cocrystal structure of peptidyltransferase activity correlates the messenger RNA 5’cap-binding protein with increased programmed -1 ribosomal (eIF4E) bound to 7-methyl-GDP. Cell 89: 951- frameshifting and viral maintenance defects in 961. the yeast Saccharomyces cerevisiae. RNA 9: Marczinke B., Hagervall T., and Brierley 982-992. I. 2000. The Q-base of asparaginyl-tRNA Miller W.A, and Koev G. 2000. Synthesis of is dispensable for effi cient -1 ribosomal subgenomic RNAs by positive-strand RNA frameshifting in eukaryotes. J. Mol. Biol. 295: viruses. Virology 273: 1-8. 179-191. 67 References

Moazed D., and Noller H.F. 1989a. Interaction Nissen P., Hansen Jj., Ban N., Moore P.M., of tRNA with 23S rRNA in the ribosomal A, P, and Steitz T.A. 2000. The structural basis of and E sites. Cell 57: 585-597. ribosome activity in peptide bond synthesis. Science 289: 920-930 Moazed D., and Noller H.F. 1989b. Intermediate states in the movement of transfer Ogle J.M., Brodersen D.E., Clemons W.M. RNA in the ribosome. Nature 342: 142-8. Jr, Tarry M.J., Carter A.P., Ramakrishnan V. 2001. Recognition of cognate transfer RNA by Muhlrad D., Decker C.J., and Parker R. 1995. the 30S ribosomal subunit. Science 292: 897- Turnover mechanisms of the stable yeast PGK1 902. mRNA. Mol. Cell. Biol. 15: 2145-2156. Park H.S., Himmelbach A., Browning K.S., Mäkinen K., Tamm T., Naess V., Truve Hohn T., and Ryabova LA. 2001. A plant viral E., Puurand U. Munthe T., and Saarma M. “reinitiation” factor interacts with the host 1995a. Characterization of cocksfoot mottle translational machinery. Cell 106: 723-733. sobemovirus genomic RNA and sequence comparison with related viruses. J. Gen. Virol. Paul A.V., Van Boom J.H., Filippov D., 76: 2817-2825. and Wimmer E. 1998. Protein-primed RNA synthesis by purifi ed poliovirus RNA Mäkinen K., Naess V., Tamm T., Truve E, polymerase. Nature 393: 280-284. Aaspollu A., and Saarma M. 1995b. The putative replicase of the cocksfoot mottle sobemovirus Paul C.P., Barry J.K., Dinesh-Kumar S.P., is translated as a part of the polyprotein by -1 Brault V., Miller W.A. 2001. A sequence ribosomal frameshift. Virology 207: 566-571. required for -1 ribosomal frameshifting located four kilobases downstream of the frameshift Napthine S., Vidakovic M., Girnary R., Namy site. J. Mol. Biol. 310: 987-999. O., and Brierley I. 2003. Prokaryotic-style frameshifting in a plant translation system: Pavitt G.D., Ramaiah K.V.A., Kimball S.R., conservation of an unusual single-tRNA and Hinnebusch A.g. 1998. eIF2 independently slippage event. EMBO J. 22: 3941-3950. binds two distinct eIF2B subcomplexes that catalyse and regulate guanine-nucleotide Neeleman L., Olsthoorn R.C.L., Linthorst exchange. Genes Dev. 12: 514-526. H.J.M., and Bol J.F. 2001. Translation of a nonpolyadenylated viral RNA is enhanced by Pestova T.V., Hellen C.U., Shatsky I.N. binding of viral coat protein or polyadenylation 1996. Canonical eukaryotic initiation factors of the RNA. Proc. Natl. Acad. Sci. USA 98: determine initiation of translation by internal 14286-14291. ribosomal entry. Mol. Cell. Biol. 16: 6859- 6869. Nicolaisen M., Johansen E., Poulsen G.B., and Borkhardt B. 1992. The 5’ untranslated region Pestova T.V., Borukhov S.I, Hellen C.U. 1998. from pea seedborne mosaic potyvirus RNA as Eukaryotic ribosomes require initiation factors a translational enhancer in pea and tobacco 1 and 1A to locate initiation codons. Nature protoplasts. FEBS Lett. 303: 169-172. 394: 854-859. Niepel M., and Gallie D.R. 1999a. Identifi cation Pestova TV, Lomakin IB, Lee JH, Choi SK, and characterization of the functional elements Dever TE, Hellen CU. 2000. The joining of within the tobacco etch virus 5’ leader required ribosomal subunits in eukaryotes requires for cap-independent translation. J. Virol. 73: eIF5B. Nature 403: 332-335. 9080-9088. Pestova T.V., Kolupaeva V.G., Lomakin Niepel M., and Gallie D.R. 1999b. Secondary I.B., Pilipenko E.V., Shatsky I.N., Agol V.I., structure in the 5’leader or the 3’untranslated Hellen C.U. 2001. Molecular mechanisms region reduces protein yield but does not affect of translation initiation in eukaryotes. the functional interaction between the 5’cap Proc. Natl. Acad. Sci. USA 98: 7029-7036. and the poly(A) tail. FEBS Lett. 462: 79-84. 68 References

Pestova T.V., Kolupaeva V.G. 2002. The roles unwinding of messenger RNA structure by of individual eukaryotic translation initiation eukaryotic initiation factors. J. Biol. Chem. factors in ribosomal scanning and initiation 260: 7651-7658. codon selection. Genes Dev. 16: 2906-2922. Riechmann J.L., Lain S., and Garcia J.A. Plant E.P., Jacobs K.L., Harger J.W., 1992. Highlights and prospects of potyvirus Meskauskas A., Jacobs J.L., Baxter J.L., Petrov . J. Gen. Virol. 73: 1-16. A.N., Dinman J.D. 2003. The 9-A solution: Rodnina M.V., Savelsbergh A., Katunin V.I., how mRNA pseudoknots promote effi cient and Wintermeyer W. 1997. Hydrolysis of GTP programmed -1 ribosomal frameshifting. RNA by elongation factor G drives tRNA movement 9: 168-174. on the ribosome. Nature 385: 18-19. Plant E.P., and Dinman J.D. 2005. Torsional Romanos M.A., Scorer C.A., and Clare J.J. restraint: a new twist on frameshifting 1992. Foreign gene expression in yeast: a pseudoknots. Nucleic Acids Res. 33: 1825-33. review. Yeast 8: 423-488. Pooggin M.M., and Skryabin K.G.1992. Rosenfeld A.B., and Racaniello V.R. 2005. The 5’untranslated leader sequence of potato Hepatitis C virus internal ribosome entry virus X RNA enhances the expression of site-dependent translation in Saccharomyces heterologous gene in vivo. Mol. Gen. Genet. cerevisiae is independent of polypyrimidine 234: 329-331. tract-binding protein, poly(rC)-binding protein Preiss T., and Hentze M.W. 1998. Dual 2, and La protein. J. Virol. 79: 10126-10137. function of the messenger RNA cap structure Ryabov E.V., Krutov A.A., Novikov V.K., in poly(A)-tail-promoted translation in yeast. Zhelezkinova O.V., Morozov S.Y., and Zavriev Nature 393: 516-520. S.K. 1996. Nucleotide sequence of RNA from Prüfer D., Kawchuk L., Monecke M., Nowok Sobemovirus found in infected Cocksfoot S., Fischer R., Rohde W. 1999. Immunological shows a Luteovirus-like arrangement of analysis of potato leafroll luteovirus (PLRV) the putative replicase and protease genes. P1 expression identifi es a 25 kDa RNA- Phytopathology 86: 391-397. binding protein derived via P1 processing. Ryabova L.A., and Hohn T. 2000. Ribosome Nucleic Acids Res. 27: 421-425. shunting in the caulifl ower mosaic virus 35S Prüfer D., Tacke E., Schmitz J., Kull B., RNA leader is a special case of reinitiation Kaufmann A., Rohde W. 1992. Ribosomal of translation functioning in plant and animal frameshifting in plants: a novel signal directs systems. Genes Dev. 14: 817-829. the -1 frameshift in the synthesis of the putative Salas-Marco J., and Bedwell D.M. 2004. viral replicase of potato leafroll luteovirus. GTP hydrolysis by eRF3 facilitates stop EMBO J. 11: 1111-1117. codon decoding during eukaryotic translation Pyronnet S., Imataka H., Gingras A., Fukunaga termination. Mol. Cell. Biol. 2004. 24: 7769- R., Hunter T., and Sonenberg N. 1999. Human 7778. eukaryotic trasnlation initiation factor 4G Satheshkumar P.S., Gayathri P., Prasad K., (eIF4G) recruits Mnk1 to phosphorylate eIF4E. and Savithri H.S. 2004. “Natively Unfolded” EMBO J. 18: 270-279. VPg Is Essential for Sesbania Mosaic Virus Rajamäki ML., Mäki-Valkama T., Mäkinen Serine Protease Activity. J Biol. Chem. 280: K., and Valkonen J. 2004. Infection with 30291-30300. potyviruses. In: Talbot N.J. (ed) Plant-pathogen Satheshkumar P.S., Lokesh G.L., and Savithri interactions, Blackwell publishing, pp 68-91. H.S. 2004. Polyprotein processing: cis and Ray B.K., Lawson T.G., Kramer J.C., Cladaras trans proteolytic activities of Sesbania mosaic M.H., Grifo J.A. Abramson R.D., Merrick virus serine protease. Virology 318: 429-438. W.C., and Thach R.E. 1985. ATP-dependent 69 References

Shen S.L., and Tinoco I. Jr. 1995. The structure Sonenberg N., Guertin D., and Lee K.A.W. of an RNA pseudoknot that causes effi cient 1982. Capped mRNAs with reduced secondary frameshifting in mouse mammary tumor virus. structure can function in extracts from J. Mol. Biol. 247: 963-78. poliovirus-infected cells. Mol. Cell. Biol. 2: 1633–1638. Shen R., and Miller W.A. 2004. Subgenomic RNA as a riboregulator: negative regulation of Spahn C.M., Kieft J.S., Grassucci R.A., RNA replication by Barley yellow dwarf virus Penczek P.A., Zhou K., Doudna J.A., and subgenomic RNA 2. Virology 327: 196-205. Frank J. 2001. Hepatitis C virus IRES RNA- induced changes in the conformation of the 40s Shirako Y., and Strauss J.H. 1994. Regulation ribosomal subunit. Science 291: 1959-1962. of Sindbis virus RNA replication: Uncleaved P123 and nsp4 function in minus-strand RNA Spahn C.M., Jan E., Mulder A., Grassucci synthesis, whereas cleaved products from P123 R.A., Sarnow P., and Frank J. 2004. Cryo-EM are required for effi cient plus-strand synthesis. visualization of a viral internal ribosome entry J. Virol. 68: 1874-1885. site bound to human ribosomes: the IRES functions as an RNA-based translation factor. Shukla D.D., Ward C.W., and Brunt, A.A. Cell 118: 465-475. 1994. The potyviridae. CAB International, Wallingford, UK. Spall V.E., Shanks M. and Lomonossoff G.P. 1997. Polyprotein processing as a strategy for Skuzeski J.M., Nichols L.M., Gesteland R.F., gene expression in RNA viruses. Seminars in and Atkins J.F. 1991. The signal for a leaky virology 8: 15-23. UAG stop codon in several plant viruses includes the two downstream codons. J. Mol. Stahl G., Bidou L., Rousset J.P., and Cassan Biol. 218: 365-373. M. 1995. Versatile vectors to study recoding: conservation of rules between yeast and Sleat D.E., Hull R., Turner P.C., and Wilson mammalian cells. Nucleic Acids Res. 23: T.M. 1988. Studies on the mechanism of 1557-1560. translational enhancement by the 5’-leader sequence of tobacco mosaic virus RNA. Eur. Su L., Chen L., Egli M., Berger J.M., and Rich J. Biochem. 175: 75-86. A. 1999. Minor groove RNA triplex in the crystal structure of a ribosomal frameshifting Smirnyagina E.V., Morozov S.Y., Rodionova viral pseudoknot. Nat. Struct. Biol. 6: 285- N.P., Miroshnichenko N.A., Solovev A.G., 292. Fedorkin O.N., and Atabekov J.G. 1991. Translational effi ciency and competitive Svitkin Y.V., Imataka H., Khaleghpour K., ability of mRNAs with 5’-untranslated alpha Kahvejian A., Liebig H.D., Sonenberg N. beta-leader of potato virus X RNA. Biochimie 2001. Poly(A)-binding protein interaction with 73: 587-598. elF4G stimulates picornavirus IRES-dependent translation. RNA 7: 1743-1752. Somogyi P., Jenner A.J., Brierley I., and Inglis S.C. 1993. Ribosomal pausing during Tamm T., Mäkinen K., and Truve E. 1999. translation of an RNA pseudoknot. Mol. Cell. Identifi cation of genes encoding for the Biol. 13: 6931-40. cocksfoot mottle virus proteins. Arch. Virol. 144: 1557-1567. Song H., Mugnier P., Das A.K., Webb H.M., Evans D.R., Tuite M.F., Hemmings B.A., and Tamm T., and Truve E. 2000a. Sobemoviruses. Barford D. 2000. The crystal structure of human J. Virol. 74: 6231-6241. eukaryotic release factor eRF1-mechanism Tamm T., and Truve E. 2000b. RNA-binding of stop codon recognition and peptidyl-tRNA activities of cocksfoot mottle sobemovirus hydrolysis. Cell 100: 311-321. proteins. Virus Res. 66: 197-207.

70 References

Tamm T. 2000c. Cocksfoot mottle virs: Van den Heuvel M., and Raue H.A. 1992. the genome organization and translational Translational effi ciency of mRNA in yeast strategies. Ph.D. Thesis. cells is not increased by plant viral leader sequences. Appl. Microbiol. Biotechnol. 37: Tanguay R.L., and Gallie D.R. 1996. Isolation 84-87. and characterization of the 102-kilodalton RNA-binding protein that binds to the 5’ and Van Eden M.E., Byrd M.P., Sherrill K.W., and 3’ translational enhancers of tobacco mosaic Lloyd R.E. 2004. Demonstrating internal entry virus RNA. J. Biol. Chem. 271: 14316-14322. sites in eukaryotic mRNAs using stringent RNA test procedures. RNA 10: 720-730. Thompson S.R., Gulyas K.D., and Sarnow P. 2001. Internal initiation is Saccharomyces van der Wilk F., Verbeek M., Dullemans A.M., cerevisiae mediated by an initiator tRNA/eIF2- and van den Heuvel J.F. 1997. The genome- independent internal initiation site element. linked protein of potato leafroll virus is located Proc. Natl. Acad. Sci. USA 98: 12972-12977. downstream of the putative protease domain of the ORF1 product. Virology 234: 300-303. Toth R.L., Chapman S., Carr F., and Santa Cruz S. 2001. A novel strategy for the expression of van der Wilk F., Verbeek M., Dullemans A., foreign genes from palnt virus vectors. FEBS and van den Heuvel J. 1998. The genome- Lett. 489: 215-219. linked protein (VPg) of southern bean mosaic virus is encoded by the ORF2. Virus Genes 17: Truve E., Naess V., Blystad D.R., Järvekülg 21-24. L., Mäkinen K., Tamm T., and Munthe T. 1997. Detection of cocksfoot mottle virus particles Vega Laso M.R., Zhu D., Sagliocco F., and RNA in oat plants by immunological, Brown A.J., Tuite M.F., and McCarthy J.E. biotechnical and electronmicroscopical 1993. Inhibition of translation in the yeast techniques. Archiv für phytopathologie und Saccharomyces cerevisae as a function of the pfl anzenschutz 30: 473-485. stability and position of hairpin structures in the mRNa leader. J. Biol. Chem. 268: 6453- Tu C., Tzeng T.H., and Bruenn J.A. 1992. 6462. Ribosomal movement impeded at a pseudoknot required for frameshifting. Proc. Natl. Acad. Verge V., Vonlanthen M., Masson J.M., Trachsel Sci. USA 89: 8636-8640. H., and Altmann M. 2004. Localization of a promoter in the putative internal ribosome Unbehaun A., Borukhov S.I., Hellen C.U.T., entry site of the Saccharomyces cerevisiae and Pestova T. Release of initiation factors TIF4631 gene. RNA 10: 277-86. from 48S complexes during ribosomal subunit joining and the link between establishment of Wang S., Browning K.S., and Miller W.A. codon-anticodon base-pairing and hydrolysis 1997. A viral sequence in the 3’-untranslated of eIF2-boung GTP. Genes Dev. 18: 3078- region mimics a 5’ cap in facilitating translation 3093. of uncapped mRNA. EMBO J. 16: 4107- 4016. Urbonavicius J., Stahl G., Durand J.M., Ben Salem S.N., Qian Q., Farabaugh P.J., and Björk Weischenfeldt J., Lykke-Andersen J., and G.R. 2003. Transfer RNA modifi cations that Porse B. 2005. Messenger RNA surveillance: alter +1 frameshifting in general fail to affect neutralizing natural nonsense. Curr. Biol. 15: -1 frameshifting. RNA 2003 9: 760-8. 559-562. Urbonavicius J., Qian Q., Durand J.M., Wells D.R., Tanguay R.L., Le H., Gallie Hagervall T.G., and Björk G.R. 2001. D.R. 1998. HSP101 functions as a specifi c Improvement of reading frame maintenance translational regulatory protein whose activity is a common function for several tRNA is regulated by nutrient status. Genes Dev. 2: modifi cations. EMBO J. 20: 4863-4873. 3236-3251.

71 References

Whalen S.G., Gingras A., Amankwa L., Mader Young T.E., Ling J., Geisler-Lee J., Tanguay S., Branton P.E., Aebersold R., and Sonenberg R.L., Caldwell C., and Gallie D.R. 2001. N. 1996. Phosphorylation of eIF-4E on serine Developmental and thermal regulation of the 209 by protein kinase C is inhibited by the maize heat schock protein, Hsp101. Plant translational repressors, 4E-binding proteins. Physiol. 127: 777-791. J. Biol. Chem. 217: 11831-11837. Zeenko V., and Gallie D.R. 2005. Cap- Wills N.M., Gesteland R.F., and Atkins J.F. independent translation of tobacco Etch virus 1991. Evidence that a downstream pseudoknot is conferred by an RNA Pseudoknot in the 5’- is required for translational read-through of leader. J. Biol. Chem. 28: 26813-26824. the Moloney murine leukemia virus gag stop Zhou W., Edelman G.M., and Mauro V.P. 2003. codon. Proc. Natl. Acad. Sci. USA 88: 6991- Isolation and identifi cation of short nucleotide 6995. sequences that affect translation initiation in Wilson J.E., Pestova T.V., Hellen C.U., and Saccharomyces cerevisiae. Proc. Natl. Acad. Sarnow P. 2000. Initiation of protein synthesis Sci. USA 100: 4457-4462. from the A site of the ribosome. Cell 102: 511- Zhu J. and Zhang M.Q. 1999. SCPD: a 520. promoter database of the yeast Saccharomyces Wittmann S., Chatel H., Fortin M.G. and cerevisiae. Bioinformatics 15:607-611. Laliberte J.F. 1997. Interaction of the viral Zelenina D.A., Kulaeva O.I., Smirnyagina protein genome linked of turnip mosaic E.V., Solovyev A.G., Miroshnichenko N.A., potyvirus with the translational eukaryotic Fedorkin O.N., Rodionova N.P. Morozov initiation factor (iso) 4E of Arabidopsis S.Yu., and Atabekov J.G. 1992. Translation thaliana using the yeast two-hybrid system. enhancing properties of the 5’-leader of potato Virology 234: 84-92. virus X genomic RNA. FEBS Lett. 296: 267- Wolin S.L., and Walter P. 1988. Ribosome 270. pausing and stacking during translation of a Xiao H., Neuveut C., Benkirane M., and Jeang eukaryotic mRNA. EMBO J. 7: 3559-3569. K.T. 1998. Interaction of the second coding Wilson K.S., and Noller H.F. 1998. Molecular exon of Tat with human EF-1 delta delineates movement inside the translational engine. Cell a mechanism for HIV-1-mediated shut-off of 92: 337-349. host mRNA translation. Biochem. Biophys. Res. Commun. 244: 384-389 Yelverton E., Lindsley D., Yamauchi P., and Gallant J.A. 1994. The function of a Åström S.U., von Pawel-Rammingen U., and ribosomal frameshifting signal from human Byström A.S. 1993. The yeast initiator tRNAMet immunodefi ciency virus-1 in Escherichia coli. can act as an elongator tRNAMet in vivo. J. Mol. Mol. Microbiol. 11: 303-313. Biol. 233: 43-58.

72