Cis and trans-acting factors that regulate hepatitis A virus translation

Picornaviruses have been shown to translate their genomes by binding ribosomes to internal sequences withing the 5'-noncoding region (5'NCR) of the viral RNA genome. The internal ribosome binding sites (RES) has been mapped to the 5'NCR of all picornaviruses studies so far. The experiments presented in this dissertation have been designed to identify the cis- and trans- acting factors that effect hepatitis A virus (HAV) translation. Studies provided in chapter II describe the minimal sequence required for cap-independent translation of HAV monocistronic RNA in vitro. The region between nucleotides 347 and 734 is required for translation in this system. By utilizing bicistronic RNAs containing the HAV 5'NCR, Chapter II further defines those sequences by determining that HAV also translated by internal initiation and contains an IRES element that exists between nucleotides 45 and 734. Comparative analysis using capped and uncapped bi- and monocistronic RNAs showed that RNA synthesized with 5'-terminal methylated cap group allows for more efficient translation that RNA containing only the HAV IRES. Studies in Chapter IV describe the existence of a trans-acting activity fournd in liver cells that stimulates HAV translation in vitro. This activity appears to be specific to HAV RNA and could not be found in other cell of tissue culture origin or kidney, heart and brain extracts derived from the mouse. In combination, the studies presented in this dissertation may help elucidate the possible causes for the slow-growth and pathogenicity exhibited by HAV.

CIS AND TRANS-ACTING FACTORS THAT REGULATE HEPATITIS A VIRUS TRANSLATION by Michael John Glass A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Cellular, Viral and Molecular Biology U ni versi ty of U tab December 1992 Copyright © Michael John Glass 1992 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a dissertation submitted by Michael John Glass This dissertation has been read by each member of th owing supervisory commi nee and by majority vote has been foun sat acto . David Stillman David Low THE UNIVERSITY OF UTAH GRADUATE SCHOOL FIN AL READING APPROV AL To the Graduate Council of the University of Utah: ~~~~------------------- in its final hie style are consistent bles, and charts are in . ory committee and is L Donald F. 'll mers Chair. Supervisory Committee Approved for the Major Department Approved for the Graduate Council B. Gale Dick Dean of The Graduate School ABSTRACT Picornaviruses have been shown to translate their genomes by binding ribosomes to internal sequences within the 5'-noncoding region (5'NCR) of the viral RNA genome. The internal ribosome binding site (IRES) has been mapped to the 5'NCR of all picornaviruses studied so far. The experiments presented in this dissertation have been designed to identify the cis- and trans­acting factors that effect hepatitis A virus (HA V) translation. Studies provided in chapter II describe the minimal sequence required for cap-independent translation of HA V monocistronic RNA in vitro. The region between nucleotides 347 and 734 is required for translation in this system. By utilizing bicistronic RNAs containing the HAV 5'NCR, Chapter III further defines those sequences by determining that HA V also translates by internal initiation and contains an IRES element that exists between nucleotides 45 and 734. Comparative analysis using capped and uncapped bi- and monocistronic RNAs showed that RNA synthesized with the 5'-terminal methylated cap group allows for more efficient translation than RNA containing only the HAVIRES. Studies in Chapter IV describe the existence of a trans-acting activity found in liver cells that stimulates HA V translation in vitro. This activity appears to be specific for HA V RNA and could not be found in other cells of tissue culture origin or kidney, heart and brain extracts derived from the mouse. In combination, the studies presented in this dissertation may help elucidate the possible causes for the slow-growth and pathogenicity exhibited by HA V. v This work is dedicated in memory of my parents, Caroline 'Penny' Glass and Michael Arthur William Glass, the driving force behind my career. TABLE OF CONTENTS ABSTRACT... ....... .... .... .................................... ...................................................... i v LIST OF FIG'URES............................................................................................... ix ACKNOWl.EI)GEMENTS ..................................................................................... xi CHAPTER I . INTRODUCTION......... ....... ... .... ... ....... .... ............ ......... ..... ......... ..... .......... 1 Picomaviruses.... ............................................................................. 1 Life-Cycle/R.eplication.................................................................. 4 The 5' Noncoding Region .............................................................. 4 Hepatitis A Virus........ ........ ...... ..... ..... ...... ........ ................ ........ ...... 2 0 Rationale............................................................................................. 2 5 References.......................................................................................... 2 7 II. A CIS-ACTING ELEMENT WITHIN THE HEPATITIS A VIRUS 5t NONCODING REGION REQUIRED FOR IN VITRO TRANSLATION ..... ............ ............ ............. ............... .......... ..................... 3 8 Abstract......... .............. ............. .............. ............................................ 3 8 Introduction.. .......... ..... ..... ............... ...... ... ..... .......... .......... ..... ..... ..... 3 9 Materials and Methods ................................................................. 42 Results ................................................................................................. 47 Discussion.......................... .............. ................................................... 5 7 References .......................................................................................... 65 III. IDENTIFICATION OF THE HEPATITIS A VIRUS INTERNAL RIBOSOME ENTRY SITE: IN VIVO AND IN VITRO ANALYSIS OF BICISTRONIC RNAs CONTAINING THE HA V 5t NONCODING REGION............................................................................. 7 1 Abstract...................................................................................... ........ 7 1 Introduction.. ................ .......... .............. .... ................................ ........ 7 2 Materials and Methods ................................................................. 74 Results................................................................................................. 8 1 Disc:ussion................. ....... .... .... ...... ....... ............ ...... ............ ...... ...... .... 1 0 2 References............................................. ............ ................................. 1 0 7 IV. IDENTIFICATION OF A TRANS-ACTING FACTOR THAT STIMULATES HEPATITIS A VIRUS TRANSLATION IN VITRO....................................................................................................... 1 1 0 Abstract. ..... ....... ...... ...... ....... .... ............ ... ....... ............. ........ ... ....... ..... 1 1 0 Intr<Xluction.. ..... ....... ..... ........ .......... ..... .... ..... ....... ... ....... ... ..... ..... ..... 1 1 0 Results........ .......................................................... ............................... 1 1 3 Disc:ussion...... ........................................ ...... ......................... .............. 1 2 0 References................................................................ .......................... 1 2 3 V. STUDIES OF MUTATIONS CREATED IN THE FULL-LENGTH INFECTIOUS cDNAs OF POLIOVIRUS AND HEPATITIS A VIRUS.................................................................................................... 127 Intr<Xluction.. ...... ..... .... ..... ....... ........ ..... ............... ......... ..... ..... .... ...... 1 2 7 Materials and Meth<Xls ................................................................. 1 30 Results... ...................................... ........................................................ 1 3 4 Disc:ussion......................... .................................................................. 1 4 1 References.......................................................................................... 1 4 3 VI. SUMMARy................................................................................................ 146 Disc:ussion... ............ ...... ...... ...... ....... ..... ...... ...... ..... ...... ........ .... ........... 1 4 6 Future Directions ............................................................................. 14 7 viii LIST OF FIGURES Fi&ure 1.1. Picornavirus genome organization and protein cleavage ........ 3 1.2. Eukaryotic translation initiation ......................................................... 1 1 1.3. Diagram of the model for picornaviral internal rioosome entry ........................................................................................... 2 1 2.1. Construction of the transcription vectors used for in vitro translation.... ........................ ....... .... ............ ........... ......... ............ 4 3 2.2. RNA saturation of the in vitro rabbit reticulocyte lysate translation system ......................................................................... o ••• 0 0" 0.. 4 9 2.3. Translation efficiency of the 5'NCR mutanats in vitro .............. 50 2.4. Immunoprecipitation of the translation products made from RNA of the constructs with deleted sequences..................................................................................... 5 4 2.5. Translation in vitro of RNA transcripts synthesized with and without a 5f-terminal methylated cap structure ..... 5 5 2.6. Model of the HAV IRES region and how the deletion mutations map to this region and effect translation ................. 5 9 3.1. Schematic representation of the bicistronic construction and activities of both cistrons........................... .............................. .... 7 7 3.2. The effect on in vitro translation of capping bicistronic RNAs ............................................................................................................... 84 3.3. The effect on in vitro translation of capping monocistronic RNAs................................................................................. 8 7 3.4 Activity and protein expression from transcripts of the bicistronic mutants translated in an in vitro rabbit reticulocyte lysate system................................................................ .... 8 9 3.5. The effect of capping the bicistronic RN As on CAT and luciferase activity in vivo.................................................. .................... 9 6 3.6. Activity of the mutant bicistronic constructs transfected into BS-C-1 cells ................................................................ 9 8 4.1. Diagramatic representation of the truncated HA V construct showing the 51NCR and coding region used ............... 114 4.2. Autoradiogram of the translation products from serial dilutions of RNA with and without the addition of liver extract translated in vitro..... .......... ... ....... ...... ....... ... ........... ...... 1 1 6 4.3. Autoradiogram of translation products from both 1489 RNA (HAV) and Luciferase RNA in the mixed liver-reticulocyte lysate system ......................................................... 11 9 4.4. Autoradiogram of translation products from 1489 RNA supplemented with tissue culture cell S 10 extracts .................. 1 21 5. 1. Diagram of the chimeric constructs and diagram of the sequence of the mutations pHAV/7(741), pHAV/7(735) and pHA V n34( +4-) ................................................................................... 1 32 5.2. Autoradiogram of the chimeric RNAs translated in vitro.. ..... 1 37 5.3. Autoradiogram of the translation products from mutant RNAs ............................................................................................... 140 x ACKNOWLEDGEMENTS I would like to thank Don Summers for his patience and guidance in matters not only in science but in life in general. His attitude and outlook on life made my graduate experience over the last few years a bearable and even pleasurable one. Also, I'd like to thank the other members of my committee for their understanding and suggestions throughout my graduate career. Lastly, I'd like to give a special thanks to all my friends who kept me sane and helped me realize the 'big picture' when things got difficult, especially my best friend, 10Hannah, who made this whole ordeal possible. CHAPTER I INTRODUCTION Picornaviruses Picornaviridae are one of the largest and most important families of agricultural and human pathogens known. This family is currently divided into five genera: Enteroviruses, which include poliovirus; cardioviruses; rhinoviruses, the single most important cause of the common cold; aphthoviruses, including foot and mouth disease virus (FMDV), a major pathogen of livestock, and hepatitis A virus (HA V), currently assigned to its own genus. Picornaviruses are nonenveloped icosahedral particles ranging in size from 24 to 30 nanometers in diameter as measured by electron microscopy. The virions are generally resistant to organic solvents and vary in acid stability, with HA V the entero- and cardioviruses being stable at pH 3.0 or lower, whereas the other genera being labile at pH less than 6. The virion consists of four nonidentical polypeptide chains (VPl, VP2, VP3, VP4) arranged into identical protein subunits called protomers. Sixty protomers make up the virion; therefore the virus capsid contains 240 virus encoded structural proteins. To date several poliovirus virion three-dimensional structures to a resolution of 2.8 angstroms have been determined by X-ray crystallography (Filman, et aI., 1989; Hogle, et aI., 1985). 2 Picornaviral genomes are single-strand RNA molecules of positive polarity, and thus the naked viral RNA is infectious when introduced into a susceptible cell. The genomic RNA is polyadenylated at the 3' terminus and covalently linked at the 5' terminus to a small viral protein called VPg. The genomic length varies from approximately 7,200 nucleotides in rhinoviruses to approximately 8,500 nucleotides in FMDV. The RNA of cardioviruses and aphthoviruses contains a polycytidylic acid [poly(C)] tract not found in rhino- or enteroviruses. These poly(C) regions, whose function is unclear, are located in the 5' noncoding region and are heterogeneous in length between different virus isolates (80-250 nt for cardioviruses; 100-170 nt for aphthoviruses). Sequencing data have revealed a single long open reading frame that is translated into a single polyprotein. During and posttranslation this precursor protein is cleaved by two to three virus-specific proteases to ultimately generate 11 or 12 end-products (see Figure 1.1). The PI region of poliovirus codes for the four capsid proteins, while functions such as the tyrosine­glycine specific protease (2A) and replication functions reside in the P2 region. The P3 proteins include the 5' terminal VPg (3B), a protease (3C) responsible for many of the polyprotein cleavages and the RNA-dependent RNA polymerase (3D) (Koch and Koch, 1985). 3 O.73kb virion RNA (7.5 kb) 7.4 kb 0 I poly (A) VPg AUG UGA Translation P1 P2 P3 ... ..... ...... .... tt Cleavage t VP4 VP2 VP3 VP1 VPg • -- .. - 1A 1B lC 10 2A 28 2C 3A3B 3C 3D ... .............1. ..... ....... PR..O. TEA..S.E.. .. .... CAPSID PROTEASE TRANSCRIPTION? RNA POLYMERASE FIGURE 1.1. Picornavirus genome organization and protein cleavage. This figure is based on poliovirus nomenclature, cleavage and gene function. 4 Life-Cycle/Replication The picornavirus life-cycle is best understood in the enterovirus, poliovirus. Poliovirus infection begins with binding to a specific receptor (Mendelson, et al.,1989) on the the host cell followed by uncoating and release of the viral RNA within the cell cytoplasm. The viral RNA then serves as a message to translate the large viral precursor protein that is subsequently processed into the smaller functional proteins. The parental (+) RNA strand also serves as template for the synthesis of (-) RNA that, in turn will direct the synthesis of more (+) RNA. The newly synthesized (+) RNA is used in the infected cell in three different ways: first, as a template for more (-) RNA synthesis early in infection. Secondly, as a message from which more proteins are synthesized (this mRNA species does not have the VPg covalently linked to the 5'­end), and lastly, to act as virion RNA for encapsidation into new virions late in infection (this RNA species does have VPg covalently linked to the 5' -end). Late in infection progeny virus IS released from the cell by lysis of the host cell (Koch and Koch, 1985 ). The 5' NoncodinK ReKion Structure Sequence analysis of various picornaviruses revealed a long 5' noncoding region (5'NCR) that comprised approximately 10% of the viral genome. While remarkably constant between strains of the same genera, the 5' noncoding region showed diversity in length among the different genera, varying in length 5 from 500 to 1000 nucleotides. Also, the primary sequence is conserved between strains of the same genera although little conservation is found between genera. The three serotypes of poliovirus exhibited a greater than 90 % sequence homolgy in the 5'NCR (Toyoda, et al., 1984). Further analysis showed that the 5'NCRs of other enteroviruses as well as rhinoviruses share many nearly identical conserved sequence elements (Pilipenko, et al., 1989a; Rivera, et al., 1988). A significant similarity also exists between the cardioviruses' and the aphthoviruses' primary structure (Pilipenko, et al., 1989b). However, there is no noticeable similarity with these two groups and that of the entero- and rhinoviruses. Interestingly, hepatitis A virus exhibits no discernible sequence homology in the 5'NCR with any of the other members of the family (Cohen, et al., 1987a). This observation has led to the classification of HAVas a separate genus (Palmenberg, 1989). Once the primary structure of the picornaviral 5'NCRs was determined, secondary structure models were proposed. Initially a hairpin structure was identified at the 5' end of poliovirus from nucleotides (nt) 10 to 33 (Larson, et aI., 1981). This feature was also seen in all the genomes of the entero- and rhinoviruses studied (Chang, et al., 1989; Rivera, et aI., 1988). The physiological importance of this structure was determined by Racaniello and Meriam who showed that deletion of one nucleotide at the base of the hairpin (position 10) led to a viable but temperature sensitive virus (Racaniello and Meriam, 1986). Isolated revertants of this mutant all had a compensatory single 6 base mutation at position 34 which restored the original hairpin stem length. By using chemical and RNase analysis another report described the existence of a cloverleaf structure for the first 100 nucleotides of poliovirus (Andino, et al., 1990b). Mutational studies established that this structure was essential for viability of the virus and possibly played a role in positive strand RNA replication by binding both cellular and viral proteins required for replication. When the entire 5f NCRs of all the picomaviruses were modeled using computer analysis, no general consensus in structure was found. However, predictions of highly conserved structure could be derived for the enteroviruses and rhinoviruses (Pilipenko, et al., 1989a; Rivera, et al., 1988) and for the cardio­and aphthoviruses (Pilipenko, et al., 1989b). These models have also been supported by chemical and RNAase probing (Skinner, et al., 1989). Recently, the predicted secondary structure for HA V has been reported (Brown, et al., 1991). This model showed some sinlilarities with structures predicted for cardioviruses (Pilipenko, et al., 1989b), but was otherwise unrelated to other picorna viruses. Involvement in Replication When considering the role of the 5fNCR in replication of the virus, it should be noted that polymerase recognition and binding is presumed to be to the negative or complementary strand of the 5'NCR. The polymerase may initially bind to the 3'­end of the positive strand to synthesize the negative strand and 7 then bind to the complement of the 5'NCR on this newly synthesized strand. Several reports suggest that the 5' NCR does in fact affect the replication process. While cloning the poliovirus genome, it was noticed that removal of the first 100 nucleotides resulted in a loss of viability, presumably due to defects in replication (Racaniello and Baltimore, 1981). Further studies indicated that the terminal hairpin structure may be involved in viral replication. As mentioned previously, Racaniello and Meriam (1986) isolated a temperature sensitive mutant that had deleted 1 nucleotide at the base of the potential hairpin structure at nt 10- 33. This mutant was deficient in both positive and negative strand synthesis at the nonpermissive temperature. Revertants of this mutation contained a second-site transition that restored the original hairpin structure length. Further downstream, another element was identified that may play a role in poliovirus replication. A 4 base deletion around nt 70 resulted in a loss of viability (Kuge and Nomoto, 1987). This deletion might be predicted to cause the destruction of a hairpin structure in this area, whereas, an insertion in this region which led to an enlargement of the loop exhibited a temperature sensitive (ts) phenotype (Trono, et al., 1988a). This ts mutant showed a severe deficiency in viral RNA synthesis. Revertants of this mutation mapped to the 3C viral protease region suggesting an interaction of this protein or a precursor form with the region around nt 70 that affected viral RNA synthesis (Andino, et al., 1990a). Such an interaction was supported by the report (Andino, et al., 1990b) which identified 8 the existence of a cloverleaf structure for the first 100 nucleotides of the polio 51NCR by RNase, chemical and mutational analysis. This structure was necessary for binding not only the viral 3C protease, but also the viral 3D polymerase and a cellular protein of unknown function. This evidence strongly suggests a role of the 51NCR and particularly the first 100 nucleotides in poliovirus RNA replication. Due to the conservation of these 51 terminal structures among the entero- and rhinovoruses, it can be assumed that similar mechanisms exist for RNA replication among these groups (Rivera, et al., 1988). A ttenuation/Virulence Involvement of the poliovirus 51NCR in determining the level of neurovirulence was first suggested by the sequence differences in this region between the attenuated (Sabin) strains and their neurovirulent counterparts (Nomoto, et al., 1982; Stanway, et al., 1984). Direct evidence was reported by two groups that showed that reversion to neurovirulence by the Sabin type 3 strain correlated with a back mutation at nucleotide 472 (Cann, et al., 1984; Evans, et al., 1985). Neurovirulent strains and revertants of the attenuated strains contained a C residue at position 472, whereas the Sabin strains had a U residue. Two engineered type 3 strains, containing only the C and U difference at nucleotide 472, also exhibit different growth properties when grown in neuroblastoma cell lines (Agol, et al., 1989; LaMonica and Racaniello, 1989). The attenuated type 3 strain, containing 9 only the single change of a U at nucleotide 472, was shown to produce 10-fold less virus than the "neurovirulent" strain (containing the C at nt 472) when either virus was used to infect neuroblastoma cells. Minor and Dunn (1988) have shown that the gastrointestinal tract of vaccinated infants also selects for the U47 2 to C reversion in the Sabin type 3 vaccine leading to neurovirulence. Further experiments with type 1 and type 2 using intertypic switching of segments between attenuated strains and their neurovirulent counterparts has identified a region in the 5'NCR between nt 472-481 important for the attenuated phenotype (Kawamura, et aI., 1989; Moss, et aI., 1989). Several lines of evidence have suggested that the poly(C) tract observed in the cardio- and aphthovirus 5' NCRs may play a role in pathogenicity of these viruses. The first observation was that selection of an attenuated form of the virulent FMDV strain by serial passage in tissue culture cells was accompanied by a decrease in the poly(C) tract length by approximately seventy nucleotides (Harris and Brown, 1977). Using EMCV, it was shown that removal of the majority of the poly(C) tract led to an attenuated phenotype but had no effect on growth in tissue culture cells (Duke, et aI., 1990; Duke and Palmenberg, 1989). The virus lacking the poly(C) tract was found to protect mice from challenge with the wild-type EMCV. Although there has been much speculation, the exact molecular mechanism for virulence involving the 5' NCR of the picornaviruses is not understood at this time. Some suggestions indicate that the attenuated phenotype may be linked to a deficiency in translation of these viral genomes. However, even with the amount of data accumulated it is still not possible to predict the effect of specific mutations in the 5'NCR on the virulence of picornaviral strains. Translation 10 To have a better understanding of the complex method of translation initiation of the picornaviruses it is best to review the mechanism of eukaryotic translation initiation. The initiation pathway is shown in Figure 1.2. The mechanism for eukaryotic translation initiation has been mostly determined from biochemical studies using rabbit reticulocyte lysates. This approach has led to the identification and characterization of initiation factors involved in translation (for review see Hershey, 1991) Binding of the ribosomes to the mRNA is a very complex step and appears to be rate-limiting. This process involves the ribosomes themselves, mRNA (cis-acting sequences within the RNA) and the aminoacylated initiator tRNA. In concert with specific initiation factors, these components are converted into an active complex that is capable of translation initiation. Specifically, the smaller of the two ribosomal subunits (40S) is complexed with the initiation factors eIF-3 (eukaryotic initiation factor-3) and eIF-4C to prevent reassociation with the 60S ribosomal subunit. This 43S ribosomal complex then combines with a ternary complex of the aminoacylated initiator tRNA, eIF-2 and GTP to form the 43S preinitiation complex. The mRNAs to be Figure 1.2. Eukaryotic translation initiation. This figure was modified from the model presented by Sonenberg (1990b). 1 1 e1F-3 (3) eIF-4C (4C) \ 3/4C o ~).... ------... 0 40S Ribosomal Subunit 43S Ribosomal Comptex eIF-2/GTP-Met-tRNA Cf*) erF-4A e1F-48 3/4C/2* o 43 S Preinitiation Complex elf-4F r mRNA m7GpppG-AUG-A(n) ~ LComplex ATP ADP+Pi 3/4c/2* m7Gpp~A(n) 48S Preinitiation Complex etF-S - __ Met-tRNA f m7GPP~A(n) ~f m7GPpp@A(n) 80S Initiation Complex eIF-Z/GDP+Pi,eIF4C,eIF-S 1 2 1 3 translated also must form a complex with initiation factors in order to become competent for initiation. mRN A species in eukaryotic cells generally contain a m7Gppp(Aa) cap structure at the 5'-end which is recognized by the initiation factor eIF-4E (cap­recognizing factor). This factor, eIF-4E, can be found in a complex with two other factors, eIF-4A and p220 (together this complex is called eIF-4F). It is thought that this interaction between eIF-4F and the mRN A requires the hydrolysis of ATP to form the mRNA complex. The following step in translation initiation is the interaction of the 43S preinitiation complex with the mRN A eIF-4F complex. This association also requires the initiation factors eIF-4A and elF4B along with ATP hydrolysis. It has been shown that eIF-4F, eIF-4A along with eIF-4B contain helicase activity and it was proposed that their function is to unwind regions of RNA secondary structure to allow ribosome binding and/or scanning. Following binding of the ribosomal subunits, this complex, along with an undefined set of initiation factors, is thought to scan the RNA in an ATP-dependent manner until an initiating codon (AUG) is found in a favorable context (Kozak, 1989). Kozak has shown that leader length, RNA secondary structure, number of AUGs preceding the initiator codon and sequence surrounding the initiating AUG all effect the scanning and initiation by the ribosomal complex. A consensus sequence surrounding the initiating AUG has been determined that is optimal for translation initiation (GCCNCCAUGG, N=purine). Next, the larger 60S ribosomal subunit is added, requiring GTP and elF- 5 and the subsequent release of an eIF-2-GDP complex, eIF-3, elF- 14 4C, and eIF-5. The resulting structure, the 80S ribosomal initiation complex, is competent for translation from the selected AUG and elongation. Several studies have attempted to determine the initiation factor requirements for binding picornavirus mRN As. Due to the lack of a 51-terminal cap structure on picornaviral RNAs, it could be expected that the requirements differ from other eukaryotic messages. It has been shown that the initiation factor eIF-4F is not needed for picornaviral translation and in fact, the p220 subunit is cleaved during enterovirus (Buckley and Ehrenfeld, 1987; Etchison, et al., 1982), rhinovirus (Etchison and Fout, 1985) and aphthovirus infections (Devaney, et al., 1988). This cleavage renders the translational machinery incapable of translating capped messages, but capable of recognizing and translating picornaviral RN As. Several approaches have been undertaken to identify novel factors that specifically interact with picornaviral RNAs. Initially, it was noted that poliovirus RNA was translated from aberrant initiation sites within the viral coding region when translated in a rabbit reticulocyte lysate system (Dorner, et al., 1984). Brown and Ehrenfeld (1979) had earlier shown that addition of a ribosomal salt wash from HeLa cells could correct the aberrant pattern seen with translation in rabbit reticulocytes, presumably by directing the ribosomes to the correct AUG. Other laboratories confirmed that addition of translation initiation factors from extracts of HeLa and Krebs-2 cell lines would correct the aberrant pattern, thus the activity was termed the "initiation 1 5 correcting factor" (ICF) (Domer, et al., 1984; Phillips and Emmert, 1986; Svitkin, et al., 1985; Svitkin, et al., 1988). Although attempts have been made to identify this factor(s), the exact nature of the ICF is not known. It was shown to copurify with eIF-2 through several steps of purification, but pure extracts of eIF-2 possess little correcting activity (Svitkin, et aI., 1988). Gel retardation assays have been used to identify specific interactions between cellular factors and poliovirus RNA. Meerovitch et al. (1989) first identified a specific interaction between a short fragment of the 5'NCR RNA (nt 559-624) and a cellular protein of 52 kDa (p52) using this method. When purified, this protein could stimulate polio translation in vitro, but was not identified to be any of the known initiation factors (Meerovitch, et al., 1989; Sonenberg and Meerovitch, 1990a). Rabbit reticulocyte lysates were shown to be deficient in p52 and it was postulated that this protein may account for the ICF activity described above. Another specific protein-RNA interaction has been described with poliovirus using gel retardation assays. N ajita and Sarnow reported that a 50 kDa protein from HeLa cells, distinct from p52, eIF-2, eIF-3, and eIF-4A, specifically bound to a 46 nucleotide sequence of RNA corresponding to nt 178-224 of the poliovirus 5'NCR (N ajita and Sarnow, 1990). This membrane­associated protein was implicated to play a role in either poliovirus RNA replication or translation, both processes of which are found to occur on cellular membranes. This method of gel mobility shift has its drawbacks however. In designing these experiments, shorter RNA molecules 1 6 were used to produce a greater mobility shift when bound to proteins. It has been suggested that length of the RNA may influence the binding of specific proteins due to the high degree of secondary structure associated with the 5INCR. When a shorter RNA molecule was used, it was believed that the secondary structure might have been lost and not recognized by all the proteins that normally bind to a given region. Gebhard and Ehrenfeld (1992) have recently attempted to alleviate this problem by cross-linking proteins from cellular extracts with longer RNA molecules (approximately 800 nt). This method has led to the identification of at least 4 cellular proteins from HeLa cells that specifically interact with the poliovirus 5INCR. Translation competition studies also showed that these regions to which proteins bind are important for translation of the poliovirus polyprotein. Similarly, several groups have identified by UV­crosslinking proteins that specifically bind to EMCV and FMDV RNA (Borovjagin, et aI., 1990; Jang and Wimmer, 1990b; Luz and Beck, 1991). A 58 kDa protein from Kreb-2 cells was shown to specifically bind within the sequence 315-485 of the EMCV RNA (Borovjagin, et aI., 1990). Also, Jang and Wimmer (1990b) demonstrated the ability of a 57 kDa protein (p57) to bind to the EMCV 5INCR. Luz and Beck (1991) also showed that p57 can bind to the FMDV 51 NCR and mapped the binding sites within this region. Due to the location of the binding sites it was assumed that these proteins play a role in translation of the viral mRNAs. In 1 7 fact, p57 was isolated from rabbit reticulocytes in association with ribosomes (Luz and Beck, 1990). Because the picornaviruses lack the terminal m7Gppp( A G) cap structure characteristic of most eukaryotic messages, another mechanism for ribosome binding was suggested to exist. In 1985, it was noted that mutations within the 5'NCR of poliovirus (at nt 472-480) affected the ability of the viral genome to be translated in a cell free system (rabbit reticulocyte and Krebs-2) (Svitkin, et aI., 1985). Further studies using in vitro systems attempted to identify the existence of sequences within the 5'NCR that affected the translation of the viral RNA (Bienkowska-Szewczyk and Ehrenfeld, 1988; Pelletier, et aI., 1988c; Pelletier, et aI., 1988a; Trono, et aI., 1988b). These studies identified a minimal region in the poliovirus 5'NCR (between nt 567 -627) required for in vitro translation that acted in a cap­independent manner. In vivo studies using monocistronic RNA molecules supported the results found in vitro (Pelletier, et aI., 1988c; Trono, et aI., 1988a). Studies of mutant polioviruses also determined that the 5'NCR played a role in translation of the viral polyprotein (Dildine and Semler, 1989; Kuge and Nomoto, 1987; Trono, et aI., 1988b). These viruses, which contained engineered mutations within the 5'NCR, exhibited a small-plaque phenotype indicative of a defect in viral replication. Upon further analysis, it was determined that the defect was probably caused by an inability to translate mRNA at levels of the wild-type. These studies further extended the 5' boundaries of the region required for translation from nt 220 to 627. 1 8 A direct demonstration of internal ribosome binding within the polio 5'NCR was provided by Pelletier and Sonenberg (1988b). By creating a bicistronic RNA molecule that contained the poliovirus 5' NCR inserted in between the thymidine kinase gene (TK) and the chloramphenicol transferase gene (CAT), they attempted to show internal ribosome binding. Cells transfected with these bicistronic cDNAs showed that the second cistron (CAT) could be expressed under conditions that would not allow translation of the first cistron (TK), i.e., in poliovirus infected cells where cap-dependent translation is shut-off and in cells grown in hypertonic media. In vitro studies using these constructs further suggested the existence of an internal ribosome entry site (IRES) by demonstrating that expression of the second cistron required the presence of the poliovirus 5' NCR. Mutations created within the 5' NCR of these constructs identified a region from nt 79-632 of the 5' NCR that contained the poliovirus IRES. Similar studies with other picornaviruses have identified regions in the 5' NCRs of these viruses that are required for internal initiation of translation (Alsaadi, et aI., 1989; Belsham and Brangwyn, 1990a; Brown, et aI., 1991; Glass and Summers, 1992; Jang, et aI., 1989; Jang, et al., 1988; Jang, et al., 1990a; Jang and Wimmer, 1990b; Shih, 1987). Like poliovirus, these IRES regions map to the 3'-half of the 5' NCR, although the cardioviruses and hepatitis A virus showed some interesting differences. The 3' border of poliovirus IRES appears to end approximately 100 nt from the initiator codon at nt 743, indicating that the last 100 nt of the poliovirus 5'NCR are not 19 important for viral translation. In fact, Kuge and N omoto have shown that removal of this sequence had no effect on the virus viability (Kuge and Nomoto, 1987). In contrast, the IRES for the cardiovirus, EMCV, has been shown to include sequences up to nt 810, very close to the initiator codon, that are necessary for polyprotein synthesis (Jang and Wimmer, 1990b). The case for hepatitis A virus is very similar. Deletion analysis of the 5' NCR of HAV has shown that sequences up to nt 734 are involved in translation of the virus (Glass and Summers, 1992), suggesting that HAVis more similar to the cardioviruses than the enteroviruses. Another feature common to all picornavirus 5' NCRs studied was a pyrimidine-rich tract of approximately 17 nt which has been implicated as an important factor for internal ribosome binding. In recent studies, it has been demonstrated that this tract is essential for poliovirus translation in vivo (Meerovitch, et aI., 1991). In cardioviruses this tract lies approximately 25 nt upstream of the initiation site (Jackson, et aI., 1990). This V-rich tract is also found to overlap the sequence required for poliovirus in vitro translation (Bienkowska-Szewczyk and Ehrenfeld, 1988). A pyrmidine-rich sequence has also been noted in HA V that is similar to other picornaviruses (Cohen, 1989b), and is located between nt 709-721, starting 26 nt before the initiating codon at nt 736 and in the region which was identified as being necessary for HA V translation in vitro (Glass and Summers, 1992) Studies of the unique translation initiation of the picornaviruses have led to a proposed model for internal ribosome 20 entry and initiation of protein synthesis. Figure 1.3 demonstrates a current model for picornaviral translation. In this scheme, the IRES is recognized by cellular factors (including p52) that aid in the binding of eIF-4A and eIF-4B. Following binding, eIF-4A and eIF-4B catalyze the unwinding of RNA secondary structure to facilitate binding of the 40S ribosomal subunit. In poliovirus, the ribosomal complex then scans the RNA until it encounters the correct initiating codon some 100 nt downstream. With the cardioviruses and HA V, scanning of the RNA is minimal because ribosome entry is just upstream of the initiating codon. Hepatitis A Virus Hepatitis A virus (HA V) presently IS characterized as the single member of its own genus of the picornavirus family (Palmenberg, 1989). It is a 27 nanometer nonenveloped spherical virus with icosahedral symmetry and demonstrates a high resistance to a variety of physical and chemical agents, being stable at pH 3.0 and resistant to the presence of organic sol vents such as ether and chloroform. Like other members of the picornaviral family, the genome of HAV is a linear single-stranded monocistronic RNA molecule of positive polarity with a molecular weight of approximately 2.8 x 106 daltons. The 5' end is covalently linked to a viral protein designated VPg and the RNA molecule contains a 3' poly(A) tract of approximately 40-80 nt. Sequence analysis of several HA V isolates has been determined, showing that the viral genome is approximately 7500 nt long with an open reading Internal ribosome entry site 2 1 poliovirus pup------------------ ---- EJ elF-4A ------ FA *1 +ATP ATP ADP+Pi V / -AUG 743 scanning 1) Recognition by cellular factor( s) e1F-48 @ 2)Unwinding of local secondary structure PUP-------I4-A-l.. @... -sr--AUG-------AUG743 43 S Preinitiation Complex 48 S Initiation Complex Figure 1.3. Diagram of the model for picornaviral internal ribosome entry. The model suggests that cellular factor(s) recognize a specific RNA structure followed by eIF-4A and eIF-4B binding and unwinding of local secondary structure to allow ribosome binding. This figure was adapted from Ago!, et aI., (1989), Jackson, et aI., (1990), and Sonenberg, (1990b) 22 frame of approximately 6700 nt (Cohen, et al., 1989a; Cohen, et al., 1987b; Cohen, et al., 1987a; Linemeyer, et al., 1985; Najarian, et al., 1985). Preceding the open reading frame is the 5' noncoding region of approximately 730 nt (different isolates vary in length). Contained within the 5'NCR of HAV are 10 AUG codons preceding the initiating codon. The need for these is unclear, but a study by Pelletier, et al. (1988c) using poliovirus, determined that all the upstream AUGs except the most proximal (AUG7) to the initiating AUG were dispensable for viral replication. Mutation of the seventh AUG to UUG led to a small plaque phenotype and a decrease in translation in vitro. Presumably, this AUG, located at nt 588, was needed for poliovirus translation and/or replication. With the successful adaptation of HA V by serial passage in tissue culture in 1979 (Provost and Hilleman, 1979), growth properties of the virus could be studied. Unlike other picornaviruses, HA V growth produced little effect on host cell metabolism (de Chastonay and Siegl, 1987; Gauss-Muller and Dienhardt, 1984b; Locarnini, et al., 1981). Although the virus was adapted to growth in a variety of cell cultures of both human and nonhuman primate origin, the life-cycle was protracted (de Chastonay and Siegl, 1987; Gauss-Muller and Dienhardt, 1984b; Locarnini, et al., 1981; Provost and Hilleman, 1979; Siegl, et al., 1984; Vallbracht, et al., 1984; Wheeler, et al., 1986). Infection of the virus generally leads to an established persistent infection in cell cultures after initial replication cycles with the continuous production of low levels of infectious virus. Maximum titers were reached after 2 to 3 weeks postinfection and virus yield was 23 approximately 10-fold lower than that seen in a typical poliovirus infection. Several reports have described variants of HA V that can grow to maximal titers within 3 to 6 days postinfection and some even exhibited a cytopathic effect in specific cells (Anderson, 1987; Cromeans, et al., 1987; Shen, et al., 1986; Venuti, et al., 1985). Selection by forced passage at shorter time intervals created the faster replicating strains, but unfortunately viral yield was no greater than the wild-type isolates. Several hypotheses have been proposed in an attempt to address the slow-growth properties of HA V in tissue culture cells (Anderson, et al., 1988; Cho and Ehrenfeld, 1991; Harmon, et al., 1989; Updike, et al., 1991). It has been proposed that perhaps the progeny RNA is rapidly packaged and therefore replication templates are sequestered and unavailable for further RNA amplification (Anderson, et al., 1988). Alternatively, ideas have been posed in which all cells are initially infected, but, for replication of the virus to commence, a limiting host cell factor must be available (Cho and Ehrenfeld, 1991; Harmon, et al., 1989). Thus a given infection is asynchronous in nature, explaining the observation that viral protein and RNA are detected at early times in only a small proportion of cells (Harmon, et al., 1989). It has also been postulated that HA V may be deficient in functions necessary for efficient viral translation (Ticehurst, et al., 1989). This suggestion is based on the observations that it is difficult to detect HA V specific proteins in infected cells (Locarnini, et al., 1981; Updike, et al., 1991), and in vitro translation of HAV virion RN A or HA V transcripts produces a complex protein pattern due 24 to internal initiation within the viral coding region (Gauss-Muller, et aI., 1984a; Jia, et aI., 1991 a). In 1987, it was reported that an infectious cDNA had been constructed from the HA V HM-17S strain (Cohen, et aI., 1987c). With the availability of this cDNA, molecular genetics could be performed in an attempt to identify and determine the roles of the various nonstructural proteins of HA V. Using the infectious cDNA, Jia et al. (1991b) identified a protease activity that correlated with the HA V 3C region. Using mutational analysis and an artificial HA V substrate the protease was found to cleave both the P2-P3 junction and the 3C-3D junction. A study of HA V translation attempted to identify the HA V nonstructural proteins and their specific cleavage pattern by translating HA V transcripts in vitro in a rabbit reticulocyte lysate system (Jia, et aI., 1991a). Unfortunately the protein pattern was extremely complex, showing a range of proteins from less than 20 kDa to greater than 220 kDa. Characterization of this pattern led the investigators to believe that the pattern observed by translation of the HA V transcripts in vitro was due to internal initiation within the viral coding sequences, predominantly in the P3 coding region. Unlike poliovirus, addition of cellular extracts did not significantly alter this translation pattern, but replacement of the HA V S'NCR with the EMCV S'NCR increased translation at the correct polyprotein initiation site and reduced the overall aberrant pattern observed, suggesting a role of the HA V S'NCR In the aberrant translation seen in vitro. 25 Another study of the HAV 5f-end determined that the complete 5f -terminal sequence was required for viability when the RNA was transfected into BS-C-1 cells (Harmon, et al., 1991). By creating mutations that lacked two and three nucleotides at the 5f-terminus of HAV RNA, it was found that these mutant RNAs were not viable when transfected into BS-C-1 cells. Conversely, poliovirus RNAs containing the same mutations were infectious, suggesting that the HA V 5' -terminus may form a predicted hairpin structure thought to be important for viral replication as has been previously postulated (Rivera, et aI., 1988). Further studies of the 5'NCR of HA V determined which of the two possible initiating AUGs (at nt 736 and 741) could act as the start site for polyprotein translation. Tesar et al. (1992), using mutational analysis to alter each of the two in-frame AUGs at nt 735 and 741, both of which could be used as potential initiation sites, identified the preferential start site as the second AUG at nt 741. Interestingly, when either AUG was modified to GCG the resulting RNA was infectious when transfected into BS-C- 1 cells. Rationale Due to the slow, asynchronous growth cycle of HA V, molecular biological studies of HA V replication have been difficult. HA V growth in tissue culture cells does not result in inhibition of host cell macromolecular synthesis and therefore in vivo identification of the nonstructural proteins is a difficult task. This is compounded by the observation that translation of HA V in vitro results in aberrant initiation within the coding region and precise identification of the nonstructural proteins is difficult (Gauss­Muller, et al., 1984a; Jia, et al., 1991a). In an attempt to 26 understand the control mechanisms for translation of HA V, the data in this dissertation describe the identification of cis and trans-acting factors that play a role in regulating HA V translation. Identification of these features of HA V may help in understanding the observed slow-growth of HA V in tissue culture cells. Chapter II describes studies to identify the region in the 51NCR required for translation of a HA V monocistronic construct in vitro. These studies have identified a part of the HA V 51NCR required for translation and show that this region acts in a cap-independent manner to promote translation of HA V capsid sequences. In order to further define this element, Chapter III describes the use of bicistronic RNAs which contain the HA V 51NCR inserted in between the chloramphenicol acetyl transferase gene (CAT) and the firefly luciferase gene to show that this region functions as an IRES to bind ribosomes internally within the HA V 51NCR to direct synthesis of the downstream luciferase gene. These studies have defined the IRES region both in vitro and in vivo to include sequences from nt 45 to 734. In an attempt to identify trans-acting factors that may affect translation of the hepatotrophic HA V, Chapter IV describes the use of liver extracts to supplement an in vitro rabbit reticulocyte lysate translation reaction programmed with HA V RNA. Stimulation of HAV translation by the addition of liver SID extracts was demonstrated and specificity was determined by completing analogous experiments using heterologous RNAs and 27 cell extracts from mouse kidney, heart, brain and HeLa, L-cell and BS-C-1 tissue culture cell lines. The studies presented in this dissertation define and characterize some of the requirements for translation of HA V and may shed light on the reasons for the slow-growth phenotype exhibited by this member of the picornaviral family. References Agol, V. I., Drozdov, S. G., Ivannikova, T. A., Kolesnikova, M. S., Korolev, M. B., and Tolskaya, E. A. (1989). Restricted growth of attenuated poliovirus strains in cultured cells of neuroblastoma. 1. Viro!., 63, 4034-4038. Alsaadi, S., Hassard, S., and Stanway, G. (1989). Sequences in the 5' non-coding region of human rhinovirus 14 RNA that affect in vitro translation. 1. Gen. Viro!., 70, 2799-2804. Anderson, D. A. (1987). Cytopathology, plaque assay, and heat inactivation of hepatitis A virus strain HM-175. 1. Med. Viro!., 22, 35-44. Anderson, D. A., Ross, B. C., and Locarnini, S. A. (1988). Restricted replication of hepatitis A virus in cell culture: Encapsidation of viral RNA depletes the pool of RNA available for replication. 1. Viro!., 62, 4201-4206. Andino, R., Rieckhof, G. E., and Baltimore, D. (1990b). A functional ribonucleoprotein complex forms around the 5' end of poliovirus RNA. Cell, 63, 369-380. Andino, R., Rieckhof, G. E., Trono, D., and Baltimore, D. (1990a). Substitutions in the peotease (3Cpro) gene of poliovirus can suppress a mutation in the 5' noncoding region. J. Viro!., 64, 607- 612. Belsham, G. J., and Brangwyn, 1. K. (1990a). A region of the 5' noncoding region of foot- and- mouth disease virus RNA directs efficient internal initiation of protein synthesis within cells: involvement with the role of L protease in translation control. J. Virol., 64, 5389-5395. 28 Bienkowska-Szewczyk, K., and Ehrenfeld, E. (1988). An internal 5'­noncoding region required for translation of poliovirus RNA in vitro. J. Virol., 62, 3068-3072. Borovjagin, A. V., Evstafieva, A. G., Yu, T., and Shatsky, I. N. (1990). A factor that specifically binds to the 5'-untranslated region of encephalomyocarditis virus RNA. FEBS, 261, 237-240. Brown, B., and Ehrenfeld, E. (1979). Translation of poliovirus RNA in vitro: changes in cleavage pattern and initiation sites by ribosomal salt wash. Virology, 97, 396-405. Brown, E. A., Day, S. P., Jansen, R. W., and Lemon, S. (1991). The 5' nontranslated region of hepatitis virus RNA: secondary structure and elements required for translation in vitro. J. Virol., 65, 5828- 5838. Buckley, B., and Ehrenfeld, E. (1987). The cap binding protein complex in uninfected and infected poliovirus-infected HeLa cells. J. BioI. Chern., 262, 13599-13606. Cann, A. J., Stanway, G., Hughes, P. 1., Minor, P. D., Evans, D. M. A., Schild, G. C., and Almond, J. W. (1984). Reversion to neurovirulence of the live-attenuated Sabin type 3 oral poliovorus vaccine. Nucleic Acids Res., 12, 7787-7792. Chang, K. H., Auvinen, P., Hyypia, T., and Stanway, G. (1989). The nucleotide sequence of coxsackievirus A9; implications for receptor binding and enterovirus classification. J. Gen. Virol., 70, 3269-3280. Cho, M., and Ehrenfeld, E. (1991). Rapid completion of the replication cycle of hepatitis A virus subsequent to reversal of guanidine inhibition. Virology, 180, 770-780. Cohen, J. I. (1989b). Hepatitis A virus: Insights from molecular biology. Hepatology, 9, 889-895. 29 Cohen, J. I., Rosenblum, B., Feinstone, S. M., Ticehurst, J. R., and Purcell, R. H. (1989a). Attenuation and cell culture adaptation of hepatitis A virus (HAV): a genetic analysis with HA V cDNA. J. Virol., 63, 5364-5370. Cohen, J. I., Rosenblurm, B., Ticehurst, J. R., Daemer, R. J., Feinstone, S. M., and Purcell, R. H. (1987b). Complete nucleotide sequence of an attenuated hepatitis A virus: comparison with wild-type virus. Proc. Natl. Acad. Sci. USA, 84, 2497-2501. Cohen, J. I., Ticehurst, J. R., Feinstone, S. M., Rosenblum, B., and Purcell, R. H. (1987c). Hepatitis A virus cDNA and its RNA transcripts are infectious in cell culture. J. Virol., 61, 3035-3039. Cohen, J. I., Ticehurst, J. R., Purcell, R. H., Buckler-White, A., and Baroudy, B. M. (1987a). Complete nucleotide sequence of wild­type hepatitis A virus: comparison with different strains of hepatitis A virus and other picornaviruses. J. Virol., 61, 50-59. Cromeans, T., Sobsey, M. D., and Fields, H. A. (1987). Development of a plaque assay for a cytopathic, rapidly replicating isolate of hepatitis A virus. J. Med. Virol., 22, 45-56. de Chastonay, J., and Siegl, G. (1987). Replicative events in hepatitis A virus-infected MRC-5 cells. Virology, 157, 268-275. Devaney, M. A., Vakharia, V. N., Lloyd, R. E., Ehrenfeld, E., and Grubman, M. J. (1988). Leader protein of foot-and-mouth disease virus is required for cleavage of the p220 component of the cap­binding protein complex. J. Virol., 62, 4407-4409. Dildine, S. L., and Semler, B. L. (1989). The deletion of 41 proximal nucleotides reverts a poliovirus mutant containing a termperature-sensitive lesion in the 5' noncoding region of genomic RNA. J. Virol., 63, 847-862. Domer, A. J., Semler, B. L., Jackson, R. J., Hanecak, R., Duprey, E., and Wimmer, E. (1984). In vitro translation of poliovirus RNA: utilization of internal initiation sites in reticulocyte lysate. J. Virol., 50, 507-514. 30 Duke, G. M., Osorio, J. E., and Palmenberg, A. C. (1990). Attenuation of Mengo virus through genetic engineering of the 5' noncoding poly(C) tract. Nature (London), 343, 474-476. Duke, G. M., and Palmenberg, A. C. (1989). Cloning and synthesis of infectious cardiovirus RNAs containing short, discrete poly(C) tracts. J. ViroI., 63, 1822-1826. Etchison, D., and Fout, S. (1985). Human rhinovirus 14 infection of HeLa cells results in the proteolytic cleavage of the p220 cap­binding complex subunit and inactivates globin mRNA translation in vitro. J. ViroI., 54, 634-638. Etchison, D., Milburn, S. C., Edery, I., Sonenberg, N., and Hershey, J. W. B. (1982). Inhibition of HeLa cell protein synthesis following poliovirus infection correlates with the proteolysis of a 220,000- dalton polypeptide associated with eucaryotic initiation factor 3 and a cap binding protein complex. J. BioI. Chern., 257, 14806- 14810. Evans, D. M. A., Dunn, G., Minor, P. D., Schild, G. C., Cann, A. J., Stanway, G., Almond, J. W., Currey, K., and Maizel, J. V. J. (1985). Increased neurovirulence associated with a single nucleotide change in a noncoding region of the Sbin type 3 poliovaccine genome. Nature (London), 314, 548-550. Filman, D. J., Syed, R., Chow, M., Macadam, A. J., Minor, P. D., and Hogle, J. M. (1989). Structural factors that control conformational transitions and serotype specificity in type 3 poliovirus. EMBO J., 8, 1567-1579. Gauss-Muller, V., von der Helm, K., and Deinhardt, F. (1984a). Translation in vitro of hepatitis A virus RNA. Virology, 137, 182- 184. Gauss-Muller, V., and Dienhardt, F. (1984b). Effect of hepatitis A virus infection on cell metabolism in vitro. Proc. Soc. Exp. BioI. Med.,175, 10-15 Gebhard, 1. R., and Ehrenfeld, E. (1992). Specific interactions of HeLa cell proteins with the proposed translation domains of the poliovirus 5' noncoding region. J. ViroI., 66, 3101-3109. 3 1 Glass, M. J., and Summers, D. F. (1992). A cis-acting element within the hepatitis A virus 5'· noncoding region required for in vitro translation. Vir. Res., in press. Harmon, S. A., Richards, O. C., Summers, D. F., and Ehrenfeld, E. (1991). The 5'-terminal nucleotides of hepatitis A virus, but not poliovirus RNA, are required for infectivity. J. Virol., 65, 2757- 2760. Harmon, S. A., Summers, D. F., and Ehrenfeld, E. (1989). Detection of hepatitis A virus RNA and capsid antigen in individual cells. Virus Res., 12, 361-370. Harris, T. J. R., and Brown, F. (1977). Biochemical analysis of a virulent and an avirulent strain of foot-and-mouth disease virus. J. Gen. Virol., 34, 87-105. Hershey, J. W. B. (1991). Translational control in mammalian cells. Ann. Rev. Biochem., 60, 717-756. Hogle, J. M., Chow, M., and Filman, D. J. (1985). Three-dimensional structure of poliovirus at 2.9A resolution. Science, 229, 1358- 1365. Jackson, R. J., Howell, M. T., and Kaminski, A. (1990). The novel mechanism of initiation of picornavirus RNA translation. TIBS, 15, 477-483. Jang, S. K., Davies, M. V., Kaufman, R. J., Wimmer, E., and. (1989). Initiation of protein synthesis by internal entry of ribosomes into the 5' nontranslated region of encephalomyocarditis virus RNA in vivo. J. Virol., 63, 1651-1660. Jang, S. K., Krausslich, H. G., Nicklin, M. J. H., Duke, G. M., Palmemberg, A. C., and Wimmer, E. (1988). A segment of the 5' nontranslated region of encephalomyocarditis virus RNA directs internal entry of ribosomes during in vitro translation. J. Virol., 62, 2636-2643. Jang, S. K., Pestova, T., Hellen, C. U. T., Witherell, G. W., and Wimmer, E. (1990a). Cap-independent translation of picornavirus RNAs: structure and function of the internal ribosome entry site. Enzyme, 44, 292-309. 32 J ang, S. K., and Wimmer, E. (1990b). Cap-independent translation of encephalomyocarditis virus RNA: structural elements of the internal ribosomal entry site and involvement of a 57-kD RNA­binding protein. Genes Dev., 4, 1560-1572. Jia, X. Y., Ehrenfeld, E., and Summers, D. F. (1991b). Proteolytic activity of hepatitis A virus 3C protein. J. Virol., 65, 2595-2600. Jia, X., Scheper, G., Brown, D., Updike, W., Harmon, S., Richards, 0., Summers, D., and Ehrenfeld, E. (1991a). Translation of hepatitis A virus in vitro: Aberrant internal initiations influenced by 5' noncoding region. Virology, 182, 712-722. Kawamura, N., Kohara, M., Abe, S., Komatsu, T., Tago, K., Arita, M., and Nomoto, A. (1989). Determinants in the 5' noncoding region of poliovirus Sabin 1 RNA that influence the attenuation phenotype. J. Virol., 63, 1302-1309. Koch, F., and Koch, G. (1985). The molecular biolo~y of poliovirus. Wien-New York: Springer-Verlag. Kozak, M. (1989). The scanning model for translation: an update. J. Cell Biol., 108, 229-241. Kuge, S., and Nomoto, A. (1987). Construction of viable deletion and insertion mutants of the sabin strain of type 1 poliovirus: Function of the 5' noncoding sequences in viral replication. J. Virol., 61, 1478-1487. LaMonica, N., and Racaniello, V. R. (1989). Differences in replication of attenuated and neurovirulent poliovirus in human neuroblastoma cell line SH-SY5Y. J. Virol., 63, 2357-2360. Larson, G. R., Semler, B. L., and Wimmer, E. (1981). Stable hairpin structure within the 5'-terminal 85 nucleotides of poliovirus RNA. J. Virol., 37, 328-335. Linemeyer, D. L., Menke, J. G., Martin-Gallardo, A., Hughes, J. V., Young, A., and Mitra, S. W. (1985). Molecular cloning and partial sequencing of hepatitis A viral eDNA. J. Virol., 54, 247-255. Locarnini, S. A., Coulepis, A. G., Westaway, E. G., and Gust, I. D. (1981). Restricted replication of human hepatitis A virus in cell culture: intracellular bichemical studies. J. Virol., 37, 33 Luz, N., and Beck, E. (1990). A cellular 57 kDa protein binds to two regions of the internal translation initiation site of foot-and-mouth disease virus. FEBS Lett, 269, 311-314. Luz, N., and Beck, E. (1991). Interaction of a cellular 57 -kilodalton protein with the internal translation initiation site of foot-and - mouth disease virus. J. Virol., 65, 6486-6494. Meerovitch, K., Nicholson, R., and Sonenberg, N. (1991). In vivo mutational analysis of cis-acting RNA translational elements within the poliovirus type 2 5' untranslated region. J. Virol., 65, 5895-5901. Meerovitch, K., Pelletier, J., and Sonenberg, N. (1989). A cellular protein that binds to the 5' -noncoding region of poliovirus RNA: implications for internal translation initiation. Genes & Dev., 3, 1026-1034. Mendelson, C. L., Wimmer, E., and Racaniello, V. R. (1989). Cellular receptor for poliovirus: Molecular cloning, nucleotide sequence, and expression of a new member of the immunoglobulin supetfamily. Cell, 56, 855-865. Minor, P. D., and Dunn, G. (1988). The effect of sequences in the 5' non-coding region on the replication of poliovirus in the human gut. J. Gen. Viro!., 69, 1091-1096. Moss, E. G., O'Neill, R. E., and Racaniello, V. R. (1989). Mapping of attenuating sequences of an avirulent poliovirus type 2 strain. J. Virol., 63, 1884-1890. Najarian, R., Caput, D., Gee, W., Potter, S. J., Renard, A., Merryweather, J., Van Nest, G., and Dina, D. (1985). Primary structure and gene orginization of human hepatitis A virus. Proc. Natl. Acad. Sci. USA, 82, 2627-2631. Najita, L., and Sarnow, P. (1990). Oxidation-reduction sensitive interaction of a cellular 50 kDa protein with a RNA hairpin in the 34 5' noncoding region of the poliovirus genome. Proc. Natl. Acad. Sci. U.S.A., 87, 5846-5850. Nomoto, A., Ornata, T., Toyoda, H., Kuge, S., Horie, H., Kataoka, Y., Genba, Y., Nakaano, Y., and Imura, N. (1982). Complete nucleotide sequence of the attenuated poliovirus Sabin 1 strain genome. Proc. Natl. Acad. Sci. U.S.A., 79, 1258-1262. Palmenberg, A. C. (1989). Sequence alignments of picornaviral capsid proteins. In B. L. Semler and E. Ehrenfeld (Eds.) Molecular aspects of picornavirus infection and detection . American Society for Microbiology, Washington, D. C. Pelletier, J., Flynn, M. E., Kaplan, G., Racaniello, V., and Sonenberg, N. (1988c). Mutational analysis of upstream AUG codons of poliovirus RNA. J. Virol., 62, 4486-4492. Pelletier, J., Kaplan, G., Racaniello, V. R., and Sonenberg, N. (1988a). Cap-independent translation of poliovirus mRN A is conferred by sequence elements within the 5' noncoding region. Molec. Cell. BioI., 8, 1103-1112. Pelletier, J., and Sonenberg, N. (1988b). Internal initiation of translation of eukaryotic mRN A directed by a sequence derived from poliovirus RNA. Nature, 334, 320-325. Phillips, B. A., and Emmert, A. (1986). Modulation of the expression of poliovirus proteins in reticulocyte lysates. Virology, 148, 255-267. Pilipenko, E. M., Blinov, V. M., Chernov, B. K., Dmitrieva, T. M., and Agol, V. I. (1989b). Conservation of the secondary structure element of the 5'-untranslated region of cardio- and aphthovirus RNAs. Nucl. Acids Res., 17, 5701-5711. Pilipenko, E. V., Blinov, V. M., Romanova, L. I., Sinyakov, A. N., Maslova, S. V., and Agol, V. I. (1989a). Conserved structural domains in the 5' -untranslated region of picornaviral genomes: an analysis of the segment controlling translation and neurovirulence. Virology, 168, 201-209. 35 Provost, P. J., and Hilleman, M. R. (1979). Propogation of human hepatitis A virus in cell culture in vitro. Proc. Soc. Exp. BioI. Med., 160, 213-221. Racaniello, V. R., and Baltimore, D. (1981). Molecular cloning of poliovirus DNA and determination of the complete nucleotide sequence of the viral genome. Proc. N ad. Acad. Sci. USA, 78, 4887- 4891. Racaniello, V., and Meriam, C. (1986). Poliovirus temperature sensitive mutant containing a single nucleotide deletion in the 5'noncoding region of the viral RNA. Virology, 155, 498-507. Rivera, V. M., Welsh, D., and Maizel, J., J.V. (1988). Comparative sequence analysis of the 5' noncoding region of the enteroviruses and rhinoviruses. Virology, 165, 42-50. Shen, W., Lian-cal, J., Hui-xun, S., Xi-tan, Z., Olng-hal, M., and Xue­guang, L. (1986). Strain of hepatitis A virus causing cytopathic effects isolated in A549 cell line. Chin. Med. J., 99, 387-392. Shih, D. s. (1987). Effects of cDNA hybridization on translation of encephalomyocarditis virus RNA. J. ViroI., 61, 2033-2037. Siegl, G., de Chastonay, J., and Kronauer, G. (1984). Propogation and assay of hepatitis A virus in vitro. J. Viro!. Methods, 9, 53-67. Skinner, M. A., Racaniell, V. R., Dunn, G., Cooper, J., Minor, P. D., and Almond, J. W. (1989). New model for the secondary structure of the 5' noncoding RNA of poliovirus is supported by biochemical and genetic data that also show that RNA secondary structure is important in neurovirulence. J. Molec. BioI., 207, 379-392. Sonenberg, N. (1990b). Poliovirus translation. Curro Top. Microbiol. Immunol., 161, 23-47. Sonenberg, N., and Meerovitch, K. (1990a). Internal initiation of translation on poliovirus RNA. Rinshoken Int. Conf., 5th, p. 22. Stanway, G., Hughes, P. J., Mountford, R. C., Reeve, P., Minor, P. D., Schild, G. C., and Almond, J. W. (1984). Comparison of the complete nucleoyide sequence of the genomes of the neurovirulent poliovirus P3ILeon/37 and its attenuated Sabin vaccine derivative P3lLeonl2alb. Proc. Natl. Acad. Sci. U.S.A., 81, 1539- 1543. 36 Svitkin, Y. V., Maslova, S. V., and Agol, V. I. (1985). The genomes of attenuated and virulent poliovirus strains differ in their In vitro translation efficiencies. Virology, 147, 243-252. Svitkin, Y. V., Pestova, T. V., Maslova, S. V., and Agol, V. I. (1988). Point mutations modify the response of poliovirus RNA to a translation initiation factor: a comparison of neurovirulent and attenuated strains. Virology, 166, 394-404. Tesar, M., Harmon, S. A., Summers, D. F., and Ehrenfeld, E. (1992). Hepatitis A virus polyprotein synthesis initiates from two alternative AUG codons. Virology, 186, 609-618. Ticehurst, J., Cohen, J. I., Feinstone, S. M., Purcell, R. H., Jansen, R. W., and Lemon, S. M. (1989). Replication of hepatitis A virus: new ideas from studies with cloned cDNA.In B. L. Semler and E. Ehrenfeld (Eds.) Molecular aspects of picornavirus infection and detection. American Society for Microbiology, Washington, D. C. Toyoda, H., Kohara, M., Kataoka, Y., Suganuma, T., Ornata, T., Imura, N., and Nomoto, A. (1984). Complete nucleotide sequences of all three poliovirus serotype genomes. Implications for genetic relationship, gene function and antigenic determinants. J. Mol. Biol., 174, 561-585. Trono, D., Andino, R., and Baltimore, D. (1988b). An RNA sequence of hundreds of nucleotides at the 5' end of poliovirus RNA is involved in allowing viral protein synthesis. J. Virol., 62, 2291- 2299. Trono, D., Pelletier, J., Sonenberg, N., and Baltimore, D. (1988a). Translation in mammalian cells of a gene linked to the poliovirus 5' noncoding region. Science, 241, 445-448. Updike, W. S., Tesar, M., and Ehrenfeld, E. (1991). Detection of hepatitis A virus proteins in infected BS-C-l cells. Virology, 185, 411-418. 37 Vallbracht, A., Hofmann, L., Wurster, K. G., and Flehmig, B. (1984). Persistent infection of human fibroblasts by hepatitis A virus. J. Gen. Virol., 65, Venuti, A., DiRusso, C., del Grosso, N., Patti, A. M., Ruggeri, F., DeStasio, P. R., Martiniello, M. G., Pagnotti, P., Degener, A. M., Midulla, M., Pana, A., and Perez-Bercoff, R. (1985). Isolation and molecular cloning of a fast-growing strain of human hepatitis A virus from its double-stranded replicative form. J. Virol., 561, 579-588. Wheeler, C. M., Fields, H. A., Schable, C. A., Meinke, W. J., and Maynard, J. E. (1986). Adsorption, purifiaction, and growth characteristics of hepatitis A virus strain HAS-15 propagated in fetal rhesus monkey kidney cells. J. Clin. Microbiol., 23, 434-440. CHAPTER II A CIS-ACTING ELEMENT WITHIN THE HEPATITIS A VIRUS 5'·NONCODING REGION REQUIRED FOR IN VITRO TRANSLATION Abstract Every picornavirus studied thus far has a sequence within the 5' noncoding region that is required for internal ribosome binding and translation of the polyprotein. In an attempt to identify this region in hepatitis A virus we constructed a truncated hepatitis A virus (HA V) cDNA clone that contains the entire 736 bp 5' noncoding region (5' NCR) and 754 base pairs of the viral capsid coding region (PI) under control of the SP6 promoter. In vitro transcription and translation of this transcript in a rabbit reticulocyte lysate yielded a protein product of about 29 kDa as analyzed by autoradiography following sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-P AGE). A series of mutations of this construct have defined a minimal sequence between bases 523 and 734 in the 5' NCR that is required for efficient in vitro translation. These deleted constructs (A 523-734 and A 632-734 showed a reduced ability to translate in the rabbit reticulocyte lysate system in comparison with the full-length 5'NCR construct, pHI489. The translation of these deleted 39 constructs was artificially restored by the addition of a 5'- terminal methylated cap structure, m7GpppG, to the RNA. This increase in translational efficiency could be competed away with cap analog (m7GDP) thus indicating that this region is required for cap-independent internal ribosome binding for HA V translation. Introduction Members of the picornavirus family have a single­stranded RNA genome of positive polarity which varies between 7200 to 8500 nucleotides (nt) in length. These genomes, which lack the characteristic 7 -methyl guanosine cap of eukaryotic mRNAs, contain one large open reading frame that is preceded by an unusually long 5' noncoding region (5' NCR) comprising approximately 10 % of the genome. This region is thought to play an essential role in virus replication because of its conservation within the picornavirus family (Pestova, et al., 1991; Provost and Hilleman, 1979) and observations that alterations within this region correlate with changes in attenuation (of poliovirus), growth properties, and translational efficiency (Evans, et aI., 1985; Kuge and Nomoto, 1987; LaMonica, et aI., 1986; Ornata, et aI., 1986; Racaniello and Meriam, 1986; Svitkin, et aI., 1985). By computer and chemical analysis, this 5'-NCR contains a high degree of secondary structure (Pilipenko, et aI., 1989; Rivera, et aI., 1988; Skinner, et al., 1989). In addition, this region includes many unused AUG codons that are located in the 51NCR prior to the initiating codon. When these AUG codons were individually mutated using the Lansing strain of poliovirus (Type 2), there was no effect on poliovirus growth or in vitro translation efficiency except when the seventh AUG was mutated (the first AUG 5' to the initiating AUG at base 745); this mutant showed a small plaque phenotype and a decreased translational efficiency (Pelletier, et aI., 1988d). 40 The absence of the 5' -terminal cap structure on picornaviral genomes suggests a different method of ribosome entry and translational initiation to what is currently understood for eucaryotic messages. In the current model, the 40S ribosomal subunit (carrying Meti-tRNA and various initiation factors) initially binds the 5'-capped structure and then scans the RNA until it reaches the first AUG in a favorable context (ANNAU GG) at which it then initiates translation (Kozak, 1989). Scanning by the 40s subunit can be affected by mRNA secondary structure, length of the 5'-NCR and the number of AUGs within the region. The above features~ all evident in picornaviral genomes, appear inconsistent with efficient translation using the scanning model. In fact, it has been demonstrated with members of the enteroviruses and cardioviruses that ribosome binding occurs by a cap­independent internal entry mechanism (lang, et al., 1989; Jang, et al., 1988; Pelletier and Sonenberg, 1988a). Using studies that utilize deletion analysis in both in vitro and in vivo systems, this region (designated IRES, internal ribosome entry site) has been mapped to the 3' half of the 5' NCR in poliovirus, EMCV, FMDV and rhinovirus (Alsaadi, et al., 1989; Bienkowska-Szewczyk and Ehrenfeld, 1979; lang, et al., 1989; Pelletier, et al., 1988a). 4 1 Hepatitis A virus (HA V), currently considered a member of the enterovirus genera of the Picornaviridae, has the same genome organization as other members of the family as predicted by cDNA sequence analysis (Cohen, et aI., 1987; Najarian, et aI., 1985). Although the genome organization is similar, the growth characteristics are drastically different, as analyzed by one-step growth kinetics in a variety of cultured cell lines (DeChastonay and Siegl, 1987; Locarnini, et aI., 1981; Provost and Hilleman, 1979; Wheeler, et aI., 1986). With a few exceptions (Anderson, 1987; Cromeans, et aI., 1989; Venuti, et aI., 1985), HAV grows slowly in tissue culture, requiring approximately 1 to 2 weeks to reach maximum titers. The infection is nonlytic and frequently establishes a persistent state in which the cells shed low amounts of virus. Although several hypotheses have been suggested to explain the unusual growth properties of this virus (Anderson, et aI., 1988; Cho and Ehrenfeld, 1991; Harmon, et aI., 1989), the mechanisms are not understood. It has been proposed that perhaps the progeny RNA is rapidly packaged and therefore replication templates are sequestered and unavailable for further RNA amplification (Anderson, et aI., 1988). On the other hand, ideas have been posed in which all cells are initially infected, but, for replication of the virus to commence, a limiting host cell factor must be available (Cho and Ehrenfeld, 1991; Harmon, et aI., 1989). Thus a given infection is asynchronous in nature, explaining the observation that one can detect only low levels of HA V RNA and protein in the early stages of an infection (Harmon, et aI., 1989). It has also been postulated that HA V may be deficient in functions ECCLES HEALTH SCIENCES liBRARY 42 necessary for efficient viral translation (Ticehurst, et aI., 1989). This suggestion is based on the observations that it is difficult to detect HA V specific proteins in infected cells (Locarnini, et aI., 1986; Updike, et al., 1991), and correct in vitro translation of HA V virion RNA or HA V transcripts is inefficient (Gauss-Muller, et al., 1986; Jia, et aI., 1991). These hypotheses cannot be completely supported without the further understanding of the mechanisms of HAV replication and which sequences are important in regulation of this growth. In this paper we have defined regions within the viral 5' NCR that are important for translation of HAV RNA in vitro. Using the infectious cDNA of HA V (Cohen, et aI., 1987) we have constructed various mutations within the 5' NCR and, after synthesizing RNA transcripts, translated these in a cell-free rabbit reticulocyte lysate system. We conclude that, like other picornaviruses, HAV contains a sequence within the 5'-NCR that is required for in vitro translation and that this sequence acts to bind ribosomes in a manner other than cap-dependent translation. Materials and Methods Plasmid DNA pHAV/7 contains the HAV (HM-175) cDNA inserted into the pGEMI plasmid (Promega) so that (+) strand RNA synthesis is under control of the SP6 promoter (Cohen, et aI., 1987). This plasmid was used to construct pH1489 (Figure 2.1) by digestion with Sca I (nucleotide 1489) and Sma I in the multiple 43 Figure 2.1. Construction of the transcription vectors used for in vitro translation. As shown in figure (A), pHA V 17 was digested with Sma I and Sca I and the purified 4.3kb fragment was ligated to yield pH1489. This plasmid was then digested with HindIII (7 nucleotides 3' to the SP6 promoter) and each of the appropriate enzymes as indicated. After purification in an agarose gel the fragments were either filled-in with the Klenow fragment or digested with T4 DNA polmerase and ligated to yield the constructs shown in Figure 2.1B. (B) Linear maps depicting the deletions constructed in part 2.1A. The DNA was transcribed as in Materials and Methods yielding RNA of the relative sizes indicated. pH1489 RNA, which contains the entire HAV 5INCR, would, upon translation, yield a protein product of approximately 29,000 daltons (containing VP4,VP2 and 6 amino acids of VP3). Serial truncations were made that deleted nt 1-347 ( A 1-347), 1- 437 (A 1-437), 1-528 ( A 1-528), 1-632 ( A 1-632), 1-734 ( A 1- 734), 632-734 ( A 632-734), 523-632 ( A 523-632), 523-734 (A 523-734) and 347-632 ( A 347-632) of the 51 NCR while leaving the coding region intact. The dashed line indicates sequences in the HA V 51NCR and the solid line corresponds to HA V coding sequences. Sea I (1489) pHAvn IO.5Kb Cut with Sma I t (MCS) and Sea I (1489), purify 4.3 Kb fragment, ligate. HlndlllOOO HPllOJS ECoHI 044 . Nsil 052 BamHI063 Mil II 0.73 EcoRI 1.49 I Cut with Hind 111 and appropriate enzyme, fill·ln with the Klenow fragment or E. coli DNA POL J or T4 DNA POL, ligate. Nested set of deletions pH1489 61·347 61·437 6 1·523 6 1-632 61·734 6632-734 6523·632 6523-734 6347-632 AUG·12 VP4/VP2/VP3ot (29kDa) ----------- - 1 736 1489 ------- 347 1489 ----- 437 1489 ---- 523 1489 . -- 632 1489 734 1489 -1 ------- ~3; V734 1489 ~ - - - - - - ;23Y.3; 1489 i - - - - - - 5'; V 734 1489 --1 --34~32 1489 +:0- +:0- 45 cloning site (MCS), yielding blunt ends which were then ligated to make a truncated HA V construct of approximately 4.3 kb. From this parental plasmid a series of deletions were made by digesting with Hind III (7 bases after the SP6 promoter start site thus deleting nucleotide 1 of the HA V 5'NCR), followed by complete or partial digestion with convenient restriction enzymes. The ends were made blunt by filling in or digesting single stranded overhangs with E. coli DNA polymerase I large fragment (Klenow) or T4 DNA polymerase, respectively. Ligation produced the products pHpal (L\ 1-347), p EcoNI (L\ 1-437), pNsil (L\ 1-523), pBamHI (L\ 1-632), p AflIl1 (L\ 1-734), each of which was deleted from base 1 to the indicated site (Figure 2.1). Constructs with internal deletions within the 5'NCR were also made in the same manner to yield the products pHpal/BamHI (L\ 347-632), pNsil/BamHI (L\ 523-632), pBamHI/Aflll1 (L\ 632-734), and pN sill AflIll (L\ 523-734). These constructs, upon ligation, were transformed into the competent bacteria, DH5a, and recombinant colonies were screened by restriction mapping for the appropriate deletions. In Vitro Transcription All HA V constructs were linearized using EcoRI restriction endonuclease, phenol-chloroform extracted and analyzed for digestion by agarose gel electrophoresis. The linearized plasmids were then used to direct in vitro transcription using a commercially available kit (Promega) and SP6 RNA polymerase (Promega). The reaction included 1 - 2 flg DNA, 2.5 46 mM each of GTP, UTP, ATP, CTP, 10 mM orr, 35 U. RNasin (Promega), Ix Promega transcription buffer, and 16 U. SP6 RNA polymerase and was incubated at 37°C for 90 minutes. Capping reactions were carried out similarly, but with the addition of 300 JlM m7GpppG or 300 JlM GpppG and subsequent decrease in GTP concentration to 0.25 mM. The relative amount of transcript that was synthesized with the cap was assayed by radiolabeling the RNA with a-32P-UTP and analyzing capped and uncapped RNAs in a 6 % urea sequencing gel (Wu, et aI., 1977). After phenol­chloroform extraction, the integrity of the RNA was analyzed by methylmercury hydroxide agarose gel electrophoresis (Bailey and Davidson, 1976). In Vitro Translation Purified RNA transcripts were used to program an in vitro ribonuclease-treated rabbit reticulocyte lysate translation reaction (Promega) (Jackson and Hunt, 1983). The reactions were run containing 15 J,!Ci [35S]-methionine essentially as previously described (Jackson and Hunt, 1983). The reaction was incubated at 30°C for 60 minutes and analyzed by autoradiography following electrophoresis in a 10 % polyacryamide-SDS gel system (Maizel, 1971). Translations of capped messages were supplemented with 320J,!M S-adenosyl-homocysteine to inhibit methylation. Quantitation was carried out by integrating the area under the curves obtained by densitometry scanning the autoradiograph with a Bio Rad Model 620 video densitometer. All measurements were made within the linear range of the exposed autoradiograms so that band intensities could be directly correlated with radiolabeled protein amount. Immunopreci pi ta ti 0 n 47 One-tenth of each translation reaction was boiled 2 minutes in Laemmli sample buffer (Jackson and Hunt, 1983) and mixed with ice cold immunoprecipitation buffer (100 mM NaCI, 10 rrlM Tris pH 7.4, 1 % NP40 and 1 mM EDT A). To this was added 2 J.11 of polyclonal rabbit antisera prepared against a region of HAV VP2 (Tesar et aI., 1991) and mixed overnight at 4°C. The following day 25 J.11 of 50% (w/v) protein A-Sepharose (Pharmacia) solution in immunoprecipitation buffer was added and mixed for 40 minutes at 4°C. The mixture was then washed three times with 50 mM Tris pH 7.4, 0.5 M NaCI, 5% sucrose, 1 % NP40 and 5 mM EDT A. The beads were resuspended in Laemmli buffer, boiled, pelleted and the supernatant was analyzed on a 10% SDS-polyacrylamide gel. Results. Translation Efficiencies of Mutated HA V 5'-NCR Constructs A series of truncated constructs of the HAV cDNA was made starting at nucleotide 1 and deleting in a 3' direction up to nucleotide 734 (Figure 2.1). The open reading frame, beginning at nt 736 and ending at base 1489 (which includes the putative HAV viral capsid protein VP4 , VP2 and 6 amino acids of VP3), would, upon translation, yield a protein product of approximately 29,000 48 Daltons as predicted by nucleotide sequence analysis of the HAV genome (Cohen, et al., 1987). Initially, using transcripts made from the construct pH1489, saturating amounts of RNA were determined for the rabbit reticulocyte lysate system (Figure 2.2). RNA concentrations and integrity of the RNA were analyzed, either by absorbance at 260 nm and by comparison to a known concentration of poliovirus genomic RN A after electrophoresis on a formaldehyde or methylmercury hydroxide agarose gel. The RNA was then serially diluted, and 1 )ll of each dilution was translated using a nuclease-treated rabbit reticulocyte lysate mixture. One­tenth of the total volume of the lysate was then analyzed by SDS­PAGE, stained with Coomassie blue dye to determine if equivalent amounts of protein were loaded and, after drying, was subjected to autoradiography. As seen in Figure 2.2 (lanes 3 and 4), 100 )lg/ml of RNA were sufficient to saturate the system. It was found that RNA made from each of the other constructs saturated the system at approximately the same level (data not shown). All subsequent experiments were carried out at or slightly above these concentrations of RNA (between 100-200 )lg/ml). Transcripts of the mutated constructs were used to program an in vitro translation reaction. Figure 2.3 shows an autoradiogram of [35S]-methionine-radiolabeled proteins synthesized by the constructs containing the full length 5'NCR and by the various mutant constructs in such a set of reactions. As shown, the predominant polypeptide product was approximately 29,000 Daltons as would be predicted from the nucleotide MW 72 kDa - 52 kDa - 38 kDa - 29 kDa - 26 kDa - lane Q) Q') ro Ul Ul Ul Q) I: 0 z « « z z a: a: ,E ,E 0> 0> :J :J 0 0 If) 2 3 49 az«: a«z: az«: a«z: ,E ,E ,E 0> Ol Ol :J :J :J 0 0 0 0 0 0 0 0 N f't"') ~ 4 5 6 7 Figure 2.2. RNA saturation of the in vitro rabbit reticulocyte lysate translation system. RNA was transcribed from the linearized pH1489 plasmid and after quantitation by absorbance at 260 nm, was serially diluted. Each dilution of RNA was used to program an equivalent translation reaction the products of which were analyzed by 10 % SDS-PAGE and autoradiography. Lane 1 shows a no message control and lanes 2-7 show increasing amounts of RNA added to each translation. 50 I.J.J f«tf!)) "'1" N "'1" N ff) f't1 f't1 f't1 f't1 I.J.J '" f' f' f't1 N "'1" f' \.0 f' \.0 > L ex:> "'1" f't1 N f't1 f't1 I I I I L "'1" ,...., "'1" IJ") \.0 f' N f't1 f't1 f' co 0Z c.. I I I I I f\'t01 INJ" ) INJ" ) f"''t11" 35kDa .... 20kDa .... lane 2 345 6 7 8 9 10 11 12 Figure 2.3. Translation efficiency of 5'NeR mutants in vitro. After transcription of the linearized plasmids with deleted HA V sequences, the purified RNAs were used to program identical translation reactions. Lane 1, Brome Mosaic Virus RNA; lane 2, no message; lane 3, pH1489 RNA; lane 4, ~ 1-347 RNA; lane 5, ~ 1- 437 RNA; lane 6, ~ 1-528 RNA; lane 7, ~ 1-632 RNA; lane 8, ~ 1- 734 RNA; lane 9, ~ 632-734 RNA; lane 10, ~ 523-632; lane 11, ~ 523-734 RNA; lane 12, ~ 347-632 RNA. sequence (Cohen, et al., 1987). pH1489 (lane 3) and the other constructs produced a doublet band which has been shown to genome (Cohen, et al., 1987). Initially, using transcripts made from the construct pH1489, saturating amounts of RNA were genome (Cohen, et al., 1987). Initially, using transcripts made 51 from the construct pH1489, saturating amounts of RNA were (lane 4), deletion of the first 347 bases (pHpal) had little effect on the translational efficiency. However, when the HA V 5'NCR was deleted to 437 (pEcoNI), translational efficiency was lowered nearly twofold as determined by densitometry scanning. Deletions from 1-523 (pNsiI) began to decrease translational efficiency by about four-fold as determined by densitometric scanning. Deletions from 1-632 (pBamHI) and 1 to 734 (pAfllll) gave a somewhat surprising result in that there was an increased translation efficiency (lanes 7 and 8). Densitometric scanning showed up to a fivefold increase in translation of these two RNAs in comparison with the full length NCR pH1489. One explanation for this result can be based on the observation by Nicklin, et al. (1987) that translation in a rabbit reticulocyte lysate of an uncapped message which consisted of 27 nucleotides of leader sequences derived from nucleotides of the promoter region of the plasmid plus the poliovirus PI region showed greater translation than that of a message containing the full poliovirus 5' NCR and PI region. Pelletier, et al. (1988b) also have identified a translational inhibitory region within the poliovirus 5' NCR between nucleotides 70-381, deletion of which resulted in increased translation in vitro. It is possible that our deletions (Ll 1-632 and Ll 1-734) have 52 also removed a translational inhibitory region though we feel this is unlikely because the inhibitory region in poliovirus is much smaller (311 nts versus 632 nts) and does not encompass the IRES region (identified minimally in vitro between nt 567-627 [Bienkowska-Szewczyk and Ehrenfeld, 1988]). Also, deletions in the HA V 5' NCR that removed a similar region (A 1-347) did not show an increase in translation. Therefore, it is our belief that this increase is due to ribosome scanning from the 5'-end of the RNA rather than removal of a specific inhibitory element. Translation was decreased by nearly 20-fold when an internal deletion was created between 632 and 734 (lane 9). This was also observed with a larger deletion between nt 523 and 734 (lane 11), which showed a decrease in translational efficiency of 40-fold. Mutations which deleted sequences between 523 and 632 and 347 and 632 also decreased translational efficiency but to a much lesser extent (fourfold and fivefold, respectively; lanes 10 and 12). Dideoxy sequence analysis of pN sil AflIII (A 523-734) and pBamHII AflIII (A 632-734) verified that the constructs contained the correct sequence around AUG736 thus eliminating the idea that the decrease in translation was due to loss of the initiating codon at nt 736. Also the decreases in translation of these two constructs could not be explained by an increased RNA degradation, because these mutant RN As exhibited similar stabilities as that of pH1489 as assayed by methyl mercury hydroxide gel electrophoresis after 2 hours incubation in the In vitro system (data not shown). 53 Figure 2.4 shows an autoradiogram of an immunoprecipitation reaction using rabbit polyclonal anti-VP2 antisera (lanes 3 thru 8) to detect six of the translational products synthesized by the deleted mutants translated in vitro. As is seen in the last two lanes (lanes 7 and 8) this antisera precipitated HAV specific proteins that migrated at the same molecular weight as those seen in Figure 2.3. Furthermore, a doublet can be seen in the deletion of 1-623 (lane 7) which, as mentioned above, is characteristic of initiation at the two in frame AUG's at nt730 and 736. This anti-VP2 serum however was not able to detect the low levels of protein synthesized by some of the constructs (lanes 3- 6). However the fact that it did precipitate the proteins synthesized by pBamHI (A 1-623) and pAflIII (A 1-734) RNAs highly suggests that the in vitro translated products in Figure 2.3 are in fact HA V specific. The Effect of Capping on Translation of the Deleted Constructs To determine if the sequences that were identified as being important for translation in vitro (nt 523-734) were acting in a cap-independent manner, we synthesized RNAs with either a methylated (m7GpppG) or unmethylated (GpppG) form of the cap structure on the 5' end and translated these in a rabbit reticulocyte lysate system. The capping reactions were carried out as stated in Figure 2.5. To analyze the efficiency of the capping reaction under these conditions, the RNA was radiolabeled with 54 w (.!) « if) if) w 0" J"'-. J"'-. ,...., N v- I: ro v- ,...., N ,...., ,...., > v- ,...., v- If) I.D J"'-. I: 0 - I I I I I tn Z 0.. - 29kDa Lane 2 3 4 5 6 7 8 Figure 2.4. Immunoprecipitation of the translation products made from RNA of the constructs with deleted HA V sequences. RNAs were translated in a rabbit reticulocyte lysate system and 2Jll was immunoprecipitated using rabbit polyclonal antisera as stated in Materials and Methods. Products were analyzed by autoradiography after seperation by 10 % SDS-PAGE. Lane 1, Brome Mosaic Virus (not immunoprecipitated); lane 2, no message; lane 3, pH1489 RNA; lane 4, A 1-347 RNA; lane 5, A 1-437 RNA; lane 6, A 1-523 RNA; lane 7, A 1-632; lane 8, A 1-734 RNA.sequencing gel. This comparison showed a 90 % capping efficiency using transcripts synthesized with and without the 5'­terminal GpppG m7GpppG m7GDP Lane > .L co w t!) « C.f) C.f) w .L 0 Z 2 0'1 0'1 0'1 0'1 ex) (0 (0 (0 -"'l" "'l" "'l" "'l" ~ - - a. n a. a. ..-. ., ... ,,--.-, + + + + + + 3 4 5 6 55 "'l" "'l" "'l" "'l" M M M M r--.. r--.. r--.. r--.. I I I I N N N N M h h h \0 \0 \0 \0 29kDa + + + + + 7 8 9 10 Figure 2.5. Translation in vitro of RNA transcripts synthesized with and without a 51-terminal methylated cap structure. RNAs were synthesized in vitro with the addition of 300 J.1M m7GpppG or 300 J.1M GpppG plus GTP (250 J.1M). Lane 1, Brome Mosaic Virus RNA; lane 2, no message; lane 3, pH1489 RNA synthesized with an unmethylated cap (GpppG); lane 4, GpppG pH1489 RNA + O.lmM exogenous m7GDP; lane 5, pH1489 RNA synthesized with a methylated cap (m7GpppG); lane 6, m7GpppG pH1489 RNA + O.lmM exogenous m7GDP; lane 7, GpppG A 623-734 RNA; lane 8, GpppG A 623-734 RNA + O.lmM exogenous m7GDP; lane 9, m7GpppG A 623-734 RNA; lane 10, m7GpppG A 623-734 RNA + O.lmM exogenous m7GDP. a-32P-UTP and compared with uncapped message on a 6 % urea methylated.cap structure. These capped messages were then purified, isolated and used to program an in vitro rabbit reticulocyte lysate system supplemented with 320JlM S­adenosyl- homocysteine to inhibit SAM methylation of unmethylated (GpppG) capped RNAs. Figure 2.5 shows an autoradiograph of radiolabelled proteins synthesized in vitro 56 using RNAs made with and without a 5'-methylated cap group. Controls, seen in the third through sixth lanes, are products of translations programmed with messages that contained the full 5' NCR of HAV. These RNAs were either synthesized with the unmethylated (the third and fourth lanes) or in the next two lanes the methylated form of the cap structure. The third lane shows the translation products using RNA with an unmethylated form of the cap structure (GpppG). This translation could not be competed with a cap analog (m7GDP, fourth lane) as would be expected if the method of ribosome binding was cap dependent. Competition was not attempted with the unmethylated cap (GpppG) in these experiments. Methylation of the cap on this construct showed no increase in translation product (fifth lane), indicating that this RN A was already at its maximum translational efficiency. Once again, competition with the cap analog showed no effect. However, as seen in the next four lanes, capping of the RNA made from the construct pBamHI-AflIII, which deleted between nucleotides 632 and 734, had a marked effect. The seventh lane demonstrates the translation product of the transcript made with an unmethylated form of the cap structure (GpppG). As shown in Figure 2.3, this 57 construct was not as efficient in translation in vitro as the full 5' NCR construct, pH1489. Once again, this could not be and would not be expected to compete with a methylated cap analog (eighth lane). However, when this transcript was synthesized with a 5'­terminal m 7 GpppG cap structure and translated, the amount of product was restored to the level of the product seen when translating RNA with the full 5' NCR, pH1489. To prove that this stimulation was due to the presence of a 5'-terminal methylated cap group, we attempted to compete away the factors required for cap-dependent initiation by adding exogenous m 7 GDP. As seen in last lane, addition of m 7 GDP reduced the amount of translation product to levels comparable with the unmethylated (GpppG) RNA seen in the seventh and eighth lanes. This is the predicted result if the increase in translation was due to capping of the RNA containing a deletion from 632-734. These data would suggest that the decrease in translational efficiency seen in this deleted construct is due to a loss of the ability to translate in a cap­independent manner and that the deletion has removed a region required for ribosome binding at sequences within the 5' NCR of HA V. The construct that deleted 523-734 also showed similar effects as A 632-734 when RNA from this construct was used to program a similar set of experiments (data not shown). Discussion In this paper we have demonstrated that a cis-acting element exists within the HA V 5' NCR that is required for efficient, cap-independant, in vitro translation, and like similar sequences 58 in other picornaviruses this region IS located in the 31 -half of the 51NCR. A recent report (Jackson, et al., 1991), based on experimental observations with other picornaviruses, outlined basic requirements for initiation of translation by this family of viruses. According to this model (illustrated in Figure 2.6), all picornaviruses contain an internal ribosome entry site (IRES) within the 51-NCR dependent on primary sequence and secondary structure, but there are differences in location and the requirement for trans-acting factors present in host cells between different members of the family. The cardioviruses appear to possess an IRES in a 450 nt segment just preceding the authentic initiation site. These viruses do not need the supplementation of cellular factors to a rabbit reticulocyte lysate for efficient in vitro translation. Poliovirus translation is somewhat more complex, and studies have suggested that poliovirus RNA requires the addition of cell extracts, either HeLa S 10 or ribosomal salt wash, for efficient and accurate in vitro translation (Brown and Ehrenfeld, 1979; Phillips and Emmert, 1986). Deletion constructs of the 51 - NCR using the PI region as a reporter gene, when translated in vitro, suggested that the minimal region for poliovirus translation in a mixed HeLa-reticulocyte system was located between nt 567- 627 (Bienkowska-Szewczyk and Ehrenfeld, 1988). However, when thepoliovirus 51-NCR was inserted between two cistrons, the herpes simplex virus-l thymidine kinase gene and the bacterial chloramphenicol acetyltransferase gene, it was found that sequences 31 from nucleotide 70 of the poliovirus 51-NCR were 59 Figure 2.6. Model of the HA Y IRES region and how the deletion mutations map to this region and effect translation. A reproduction of Jackson, et al. (1991) model and regions assigned according to Brown, et al. (1991) in relation to the deletions reported in this article. Comparative translational efficiencies are shown compared to the construct containing the full HAY 5tNCR (pHI489). 60 Intfrnal ribosome entry site poliovirus / IV TRANSI ATION II 11/ Hepatttls A VIrus 1.0 IV~ dt'l 1-347 1.0 V AUG 736 VI Ivk del 1-437 0.66 V AUG VI 736 ff vtUG 716 del 1-523 0.25 [f.AUG dt'l 1-632 4.0 VI 736 -AUG 736 dt'l 1-734 4.0 IV II III dt'1632-734 0.05 IV II III dt'1523-632 0.25 IV ,- -., 1'"- II III -:'- .. I I. - -AUG 736 dt'1523-734 0.025 I' 1"' ___ - 'E."'-- It III I •. f ! AUG H6 dt'1341-632 0.20 VI 6 1 necessary for the translation of the downstream cistron when a reticulocyte lysate was supplemented with HeLa cell extracts (Pelletier and Sonenberg, 1988c). These data suggested that the full IRES in polio lies between base 70 and 627, ending approximately 120 nt before the initiating AUG at base 743. Interestingly, deletion of nucleotides 564-726 of poliovirus type 1 yielded a viable virus when tested in vivo (Kuge and Nomoto,1987). This also suggested that the minimal region necessary for in vitro translation reflects only a portion of the sequence utilized for the virus growth. We are currently examining the effects of similar deletions in an in vivo system for HAV. Another feature in all picornavirus 5' NCRs studied was a pyrimidine-rich tract of approximately 17 nt which has been implicated as an important factor for internal ribosome binding. In recent studies, several groups have demonstrated that this tract is essential for translation in vitro and in vivo (Meerovitch, et al., 1991). In cardioviruses this tract lies approximately 25 nt upstream of the initiation site (Jackson, et al., 1991). This U-rich tract is also found to overlap the sequence required for poliovirus in vitro translation (Bienkowska-Szewczyk and Ehrenfeld, 1988). A pyrmidine-rich sequence has also been noted in HA V that is similar to other picornaviruses (Cohen, 1989), and is located between nt 709 and 721, starting 26 nt before the initiating codon at nt 736 and in the region which we have identified as being necessary for HA V translation in vitro. 62 Our results would support the current belief that the pyrimidine­rich sequence is a necessary part of the IRES. A recent report has also attempted to define the HA V IRES using transcripts from a series of 5' deletions of the HA V 5'­NCR translated in an in vitro system (Brown, et aI., 1991). The 5' boundry of the IRES was identified at approximately nt 355. However, our results showed no effect on translation until the RN A was deleted from nt 1-437, which decreased the efficiency 1.5-fold. This difference may be attributed to the methods of measuring translational efficiency, in that the former report used immunoprecipitation to detect products of translation which we found to be less sensitive. These authors also observed an increase in translational efficiency when the HAV 5'-NCR was deleted up to base 634, 740 and even 745 (deleting the whole 5'-NCR). This stimulation of translation has also been noted using poliovirus constructs deleting the entire 5'NCR (Ornata, et aI., 1986) and it is our belief that this stimulation is due to different initiation mechanism than internal ribosome entry; however, it cannot be ruled out that an inhibitory region was deleted as has been observed in poliovirus (Pelletier, et aI., 1988b). Our studies have extended these observations in that we have also examined the effects of internal deletions within the HAV 51-NCR on in vitro translation. Translation of these constructs in a rabbit reticulocyte was shown to be decreased 40-fold when internal deletions were used encompassing nucleotide 523-734. This would indicate that sequences between nt 523 and 734 encompass a major portion of the IRES of HA V. Studies have shown that HAV can initiate from the second of two in-frame AUGs at nt 730 and 736, and in fact, utilize this initiating codon at nt 736 in vitro and in vivo. These results rule out the possibility that these deleted constructs 63 cannot initiate translation due to loss of the AUG at nt 730. It should be noted that in wild type HAV, both AUGs (AUAAUGACCAUGU) are in suboptimal context according to Kozakfs rules (Kozak, 1989) which state the need of a concensus sequence, ANNAUGG (with requirement for a purine at position -3 and + 1), for efficient translation. The two mutations L\ 523-734 (pNsi/Afllll) and L\ 632-734 (pBamHI/Afllll), which translate poorly in rabbit reticulocyte lysates, are also in a non optimal context (GACAUGU and UCCAUGU, respectively). The purine at position -3 is absent in the mutation L\ 632-734, a factor that could influence translational ability. However, L\ 1-437 and L\ 1- 523 both have wild-type sequences surrounding the initiating AUG at nt736 but have decreased translational efficiency suggesting a loss of a region important for translation. Several reports studying translation of poliovirus RNA in vitro have used m7G-capped poliovirus RNAs to show the importance of the lack of a Sf-terminal cap structure and internal sequences located in the 3' half of the Sf-NCR for ribosome binding (Pelletier, et aI., 1988a; Trono, et aI., 1988), By synthesizing RNAs containing a methylated cap but lacking nt 523-734, we attempted to show that the region identified as being necessary for internal translation initiation of HA V acted in a cap­independent manner. Translational efficiency of these mutants (L\ 64 523-734 and A 632-734} could be restored to levels of pH1489 (RNA containing the full 5' NCR of HA V) when the RNA was synthesized with a m 7 G cap group on the 5' -end. This reaction could be competed with an exogenous cap analog (m 7 GDP) thus suggesting that this recovery of translation was due to the addition of a methylated cap group and ribosome entry by a different mechanism. These results thus imply that this region, when present, acts in ribosome binding by a cap-independent mechanism such as has been shown with the other picornaviruses studied. Jackson, et al. (1991) recently proposed a model for picornaviral translation, Figure 2.6 depicts our reproduction of this model and our experimental evidence that identifies the minimal sequences required for the HAV IRES. Our results suggest that HA V more closely resembles the cardiovirus, EMCV, in the location of the IRES (Jackson, et al., 1991). The twohundred 5'­nucleotides proximal to the initiating AUG at nt 736 encompass the minimal sequence required for translation of this virus in vitro. This region corresponds with domain V and VI according to the most recent published secondary structure map of HA V (Brown, et al., 1991). Interestingly, translation of our mutations (A 1-523, A 523-632 and A 523-734) also showed that approximately one-half of the domain IV large stem-loop structure is required but that deletion of the 5' -half has little effect on translation (A 1-437). Obviously, more mutations must be made in these regions to elucidate the exact structures/sequences required for translation of HA V. Currently, we are using bicistronic constructs containing the HA V 5'NCR to determine the regions required for internal ribosome binding in vitro and in vivo. References 65 Alsaadi, S., Hassard, S., and Stanway, G. (1989). Sequences in the 5' noncoding region of human rhinovirus 14 RNA that affect in vitro translation. J. Gen. Virol.,70, 2799-2804. Anderson, D.A. (1987). Cytopathology, plaque assay, and heat inactivation of Hepatitis A Virus strain HM175. J. Med. Virol.,22, 35-44. Anderson, D.A., Ross, B.C., and Locarnini, S.A. (1988). Restricted replication of Hepatitis A Virus in cell culture: Encapsidation of viral RNA depletes the pool of RNA available for replication. J. V irol., 62, 4201-4206. Bailey, J.M. and Davidson, N. (1976). Methyl mercury as a reversible denaturing agent for agarose gel electrophoresis. Anal. Biochem., 70, 75-85. Bienkowska-Szewczyk, K. and Ehrenfeld, E. (1988). An internal 5' noncoding region required for translation of poliovirus RNA in vitro. J. Virol., 62, 3068-3072. Brown, B.A. and Ehrenfeld, E. (1979). Translation of Poliovirus RNA in vitro: Changes in cleavage pattern and initiation sites by ribosomal salt wash. Virology, 97, 396-405. Brown, E.A., Day, S.P., Jansen, R.W., and Lemon, S.M. (1991). The 5' nontranslated region of hepatitis A virus RNA: secondary structure and elemins required for translation in vitro. J. Virol., 65, 5828-5838. Cho, M., and Ehrenfeld, E. (1991). Rapid completion of the replication cycle of hepatitis a virus subsequent to reversal of guanidine inhibition. Virology, 180, 770-780. Cohen, J.I. (1989). Hepatitis A Virus: Insights from molecular biology. Hepatology, 9, 889-895. 66 Cohen, J.I., Ticehurst, J.R., Feinstone, S.M., Rosenblum, B., and Purcell, R.H. (1987). Hepatitis A virus cDNA and its RNA transcripts are infections in cell culture. J. Virol., 61, 3035-3039. Cohen, J.I., Ticehurst, J.R., Purcell, R.H., Buckler-White, A., and Baroudy, B.M. (1987). Complete nucleotide sequence of wild-type Hepatitis A virus: Comparison with different strains of Hepatitis A virus and other picornaviruses. J. Virol., 61,50-59. Cromeans, T., Fields, H.A., and Sobsey, M.D. (1989). Replication kinetics and cytopathic effect of Hepatitis A Virus. J. Gen. Virol., 70, 2051-2062. De Chastonay, J., and Siegl, G. (1987). Replicative events in Hepatitis A Virus-infected MRC-5 cells. Virology, 157, 268-275. del Angel, R.M., Papavassiliou, A.G., Fernandez-Tomas, C., Silverstein, S., and Racaniello, V.R. (1989). Cell proteins bind to multiple sites within the 5' untranslated region of poliovirus RNA. PNAS (USA), 86, 8299-8303. Domer, A.I., Semler, B.L., Jackson, R.I., Hanecak, R., Duprey, E., and Wimmer, E. (1984). In vitro translation of poliovirus RNA: Utilization of internal initiation sites in reticulocyte lysate. J. Virol., 50, 507-514. Evans, D.M.A., Dunm, G., Minor, P.D., Schild, G.L., Cann, A.I., Stanway, G., Almond, J.W., Currey, K., and Maizel, J.V. (1985). Increased neurovirulence associated with a single nucleotide change in a noncoding region of the Sabin type 3 polio vaccine genome. Nature, 314, 548-550. Gauss-Muller, V., Lottspeich, E., and Deinhardt, F. (1986). Characterization of Hepatitis A Virus structural proteins. Virology, 155, 732-736. Harmon, S.A., Summers, D.F., and Ehrenfeld, E. (1989). Detection of Hepatitis A Virus RNA and capsid antigen in individual cells. Virus Research, 12, 361-370. 67 Jang, S.K., Davies, M.V., Kaufman, R.J., and Wimmer, E. (1989). Initiation of protein synthesis by internal entry of ribosomes into the 5' nontranslated region of Encephalomyocarditis virus RNA in vivo. J. Viro!., 63, 1651-1660. Jang, S.K., Krausslich, H.-G., Nicklin, MJ.H., Duke, G.M., Palmenberg, A.C., and Wimmer, E. (1988). A segment of the 5' nontranslated region of Encephalomyocarditis virus RNA directs internal entry of ribosomes during in vitro translation. J. Viro!., 62, 2636-2643. Jackson, R.J., Howell, M.T., and Kaminski, A. (1991). Initiation of translation of picornavirus RNAs: a novel mechanism with possible relevance to cellular mRNA translation. TIBS,.15, 477-483. Jackson, R.J., and Hunt, T. (1983). Preparation and use of nuclease­treated rabbit reticulocyte lysates for the translation of eukaryotic messenger RNA. Methods Enzymol., 96, 50-74. Jia, X-Y., Scheper, G., Brown, D., Updike, W., Harmon,S., Richards, 0., Summers, D., Ehrenfeld, E. (1991). Translation of Hepatitis A Virus in vitro: Aberrant internal initiations influenced by 5' non-coding region. Virology, 182, 712-722. Kozak, M. (1989). The scanning model for translation: An update. J. Cell. Biol., 108, 229-241. Kuge, S. and Nomoto, A. (1987). Construction of viable deletion and insertion mutants of the Sabin strain of type 1 poliovirus: Function of the 5' noncoding sequence in viral replication. J. Viro!., 61, 1478-1487. Kuhn, R., Luz, N., and Beck, E. (1990). Functional analysis of the internal translation initiation site of foot-and-mouth disease virus. J. Viro!., 64, 4625-4631. La Monica, N., Meriam, C., and Racaniello, V. (1986). Mapping of sequences required for mouse neurovirulence of poliovirus type 2 Lansing. J. Virol., 57, 515-525. Locarnini, S.A., Coulepis, A.G., Westaway, E.G., and Gust, I.D. (1981). Restricted replication of human Hepatitis A Virus in cell culture: Intracellular biochemical studies. J. Virol., 37, 216-225. Maizel, J.V. (1971). Polyacrylamide gel electrophoresis of viral proteins. Methods in Virology, 5, 180-244. Meerovitch, K., Nicholson, R., and Sonenberg, N. (1991). In vivo mutational analysis of cis-acting RNA translational elements within the poliovirus type 2 5' untranslated region. J. Virol., 65, 5895-5901. Meerovitch, K., Pelletier, J., and Sonenberg, N. (1989). A cellular protein that binds to the 5'-noncoding region of poliovirus RNA: Implications for internal translation initiation. Genes Dev., 3, 1026-1034. Najarian, R., Caput, D., Gee, W., Potter, S.J., Renard, A., 68 Merryweather, J., Van Nest, G., and Dina, D. (1985). Primary structure and gene organization of human Hepatitis A Virus. PNAS (USA), 82, 2627-2631. Najita, L., and Sarnow, P. (1990). Oxidation-reduction sensitive interaction of a cellular 50-KDa protein with an RNA hairpin in the 5' noncoding region of the poliovirus genome. PNAS (USA), 87, 5846-5850. Nicholson, R., Pelletier, J., Le, S.-Y., and Sonenberg, N. (1991). Structural and functional analysis of the ribosome landing pad of poliovirus type 2: In vivo translation studies. J. Virol., 65, 5886- 5894. Nicklin, M.1.H., Kriiusslich, H.G., Toyoda, H., Dunn, J.1., and Wimmer, E. (1987). Poliovirus polypeptide precursors: Expression in vitro and processing by exogenous 3C and 2A proteinases. PNAS (USA), 84, 4002-4006. Ornata, T., Kohara, M., Kuge, S., Komatsu, T., Abe, S., Semler, B.L., Kameda, A., Hoh, H., Arita, M., Wimmer, E., and Nomoto, A. (1986). Genetic analysis of the attenuation phenotype of poliovirus type 1. J. Virol., 58, 348-358. Pelletier, J., Kaplan, G., Racaniello, V.R., and Sonenberg, N. (1988a). Cap-independent translation of poliovirus mRN A is conferred by sequence elements within the 5' noncoding region. Mol. Cell. Biol., 8, 1103-1112. 69 Pelletier, J., Kaplan, G., Racaniello, V.R., and Sonenberg, N., (1988b). Translational efficiency of poliovirus mRN A: Mapping inhibitory cis-acting elements within the 5' noncoding region. J. Viro!., 62, 2219-2227. Pelletier, J. and Sonenberg, N. (1988c). Internal initiation of translation of eukaryotic mRNA directed by a sequence derived from poliovirus RNA. Nature, 334, 320-325. Pelletier, J., Flynn, M.E., Kaplan, G., Racaniello, V.R., and Sonenberg, N. (1988d). Mutational analysis of upstream AVG codons of poliovirus RNA. J. Virol., 62, 4486-4492 Pestova, T.V., Hellen, C.V.T., and Wimmer, E. (1991). Translation of poliovirus RNA: Role of an essential cis-acting oligopyrimidine element within the 5' nontranslated region and involvement of a cellular 57-kilodalton protein. J. Viro!., 65, 6194-6204. Phillips, B.A. and Emmert, A. (1986). Modulation of the expression of poliovirus proteins in reticulocyte lysates. Virology, 148, 255- 267. Pilipenko, E.V., Blinov, V.M., Romanova, L.I., Singakov, A.N., Maslova, S.V., and Agol, V.1. (1989). Conserved structural domains in the 5'-untranslated region of picornaviral genomes: An analysis of the segment controlling translation and neurovirulence. Virology, 168, 201-209 Provost, P.J. and Hilleman, M.R. (1979). Propogation of human Hepatitis A Virus in cell culture in vitro. Proceedings of the Society for Experimental Biology and Medicine, 160, 213-221. Racaniello, V.R. and Meriam, C. (1986). Poliovirus temperature­sensitive mutant containing a single nucleotide deletion in the 5'_ noncoding region of the viral RNA. Virology, 155, 498-507. Rivera, V.M., Welsh, J.D., and Maizel, J.V. (1988). Comparative sequence analysis of the 5' noncoding region of the enteroviruses and rhinoviruses. Virology, 165, 42-50. Skinner, M.A., Racaniello, V.R., Dunn, G., Cooper, J., Minor, P.O., and Almond, J.W. (1989). New model for the secondary structure of the 5' noncoding RNA of poliovirus is supported by biochemical and genetic data that also show that RNA secondary structure is important in neurovirulence. J. Mol. BioI., 207, 379-392. 70 Svitkin, Y.V., Maslova, S.V., and Agol, V.I. (1985). The genomes of attenuated and virulent poliovirus strains differ in their in vitro translation efficiencies. Virology, 147, 243-252. Tesar, M., Harmon, S.A., Summers, D.F., and Ehrenfeld, E. (1991). Hepatitis A virus polyprotein synthesis initiates from two alternative AUG codons. Virology, 186, 609-618. Ticehurst, J., Cohen, J.I., Feinstone, S.M., Purcell, R.H., Jansen, R.W., and Lemon, S.M. (1989). Replication of Hepatitis A Virus: New ideas from studies with cloned cDNA. p. 27-49. B.L.Semler and E. Ehrenfeld (Eds.), Molecular Aspects of Picornavirus Infection and Detection. American Society for Microbiology, Washington, D.C. Trono, D., Pelletier, J., Sonenberg, N., and Baltimore, D. (1988). Translation in mammalian cells of the gene linked to the poliovirus 5' noncoding region. Science, 241, 445-448. Updike, W.S., Tesar, M., and Ehrenfeld, E. (1991). Detection of hepatitis A virus proteins in infected BS-C-l cells. Virology, 185, 411-418. Venuti, A., Di Russo, C., Del Grosso, N., Patti, A.-M., Ruggeri, P.F., De Stasio, P.R., Mastiniello, M.G., Pagnotti, P., Degener, A.M., Midulla, M., Pana, A., and Perez-Bercoff, R. (1985). Isolation and molecular cloning of a fast-growing strain of human Hepatitis A Virus from its double-stranded replicative form. J. Virol., 56, 579-588. Wheeler, C.M., Fields, H.A., Schable, C.A., Meinke, W.J., and Maynard, J .E. (1986). Adsorption, purification and growth characteristics of Hepatitis A Virus strain HAS-15 propagated in fetal Rhesus monkey kidney cells. J. Clin. Micro., 23, 434-440. Wu, C.W., Wu, F.V.H., and Speckhard, D.C. (1977). Subunit location of the intrinsic divalent metal ions in RNA polymerase from E. coli. Biochemistry, 16, 5449-5454. CHAPTER III IDENTIFICATION OF THE HEPATITIS A VIRUS INTERNAL RIBOSOME ENTRY SITE: IN VIVO AND IN VITRO ANALYSIS OF BICISTRONIC RNAs CONTAINING THE HAV 5' NON CODING REGION. Abstract Hepatitis A virus (HA V), a RNA virus of positive polarity, contains a long 5' noncoding region (5tNCR) that lacks the characteristic m7GpppG cap group of most eukaryotic messages. By creating bicistronic constructs that contain the bacterial chloramphenicol acetyl-transferase gene followed by the HA V 5'NeR and the luciferase gene we have used in vitro and in vivo assays to demonstrate that ribosome entry for translation initiation occurs via binding to sequences within the HA V 5INCR. Using mutations created within this region we have identified that the HAV internal ribosome entry site (IRES) is located downstream of nucleotide 45 and including sequences up to nucleotide 734 of the HA V 5'NCR. Translation of a number of mutant constructs in vitro in a rabbit reticulocyte lysate and in vivo by transfection of the cDNAs into BS-C-I cells in the presence of the recombinant vaccinia virus, vTF7-3, gave similar results. However, a 4 nucleotide insertion at base 628 showed an increased activity over wild-type when transfected into BS-C-l cells that was not seen in vitro. This increase in activity correlated with an increase in luciferase gene product as assayed by immunoprecipitations of [35S]-methionine radio-labeled cells. Comparison of mono- and bicistronic RN As that were synthesized with or without a m7GpppG cap group showed a competition for ribosome binding when translated in a rabbit reticulocyte lysate system. The cap group on the RNA 5' terminus led to a preferred translation of the first cistron over the translation of the second gene directed by the HA V IRES. Introduction Hepatitis A virus (HA V), the causative agent of infectious hepatitis, is classified as a member of the Picornaviridae, a family of pathogens that includes such agents as poliovirus, rhinoviruses and foot and mouth disease virus (FMDV). The genome of HA V, like other members of this family, is a single­stranded polyadenylated RNA molecule of positive polarity. Sequence analysis has revealed one large open reading frame that is presumed to be translated into one polyprotein which undergoes proteolytic processing into smaller structural and nonstructural proteins. The coding region of HAVis preceded by a 5' noncoding region (5'NCR) comprising approximately 10 % of the genome (729 nucleotides in the pHA V/7 cDNA). Within the 5'NCR there are 10 AUG triplets before the initiating codon at nucleotide (nt) 730 (Cohen, et al., 1987). Unlike cellular mRNAs, the HAV 72 message lacks the characteristic 5'-terminal m7Gppp(G A) structure that is imperative for translation of most eukaryotic cellular mRNAs. The traditional model for eukaryotic translation initiation states that ribosomes recognize and bind the 5'-terminal cap structure of mRNAs and scan to the initiating AUG, which is generally the first AUG in favorable context (Kozak, 1989). The above mentioned features of HA V, and picornaviruses in general, seem incompatible with the traditional Kozak model for translation initiation (Kozak, 1989). In fact, it has been shown using several picornaviruses that translation initiation occurs by ribosomes binding to internal sequences or structures within the 5'NCR (Belsham and Brangwyn, 1990a; Jang, et aI., 1989; Jang, et al., 1988; Pelletier and Sonenberg, 1988b). Using constructs which contain the 5' NCR of poliovirus, encephalomyocarditis virus (EMCV) or FMDV inserted in between two cistrons, several groups have shown that translation of the second cistron (downstream of the 5'NCR) is dependent on sequences or structures within the 5'NCR, implying that ribosomes are binding internally, independent of the first cistron. These sequences, which span approximately 400 nucleotides of the 5'NCR, have been termed the IRES (internal ribosome entry site) and it has been postulated that a specific RNA secondary structure conformation is required for efficient utilization (Jackson, et aI., 1990). We and others have previously described a similar sequence within the 5'NCR of HA V (Brown, et aI., 1991; Glass and Summers, 1992). Using monocistronic constructs and in vitro 73 translation in rabbit reticulocyte lysate systems, the nucleotides required for translation initiation of HA V have been minimally defined to encompass sequences between nt 347 and 734 of the 5'NCR. By computer and chemical secondary structure analysis this region appears to contain a high degree of structure (Brown, et aI., 1991), as would be predicted by our current understanding of IRES function. These studies have not definitively shown, however, that ribosome entry acts by an internal binding mechanism nor where these sequences map on the HA V genome. In this report we demonstrate, by using bicistronic constructs in vitro and in vivo, that HA V utilizes an IRES in a manner similar to those previously identified in other picornaviruses. In addition, we have further defined the boundaries of the HA V IRES to include sequences between nt 45 and 734 of the HA V 5'NCR. However, although most mutations within this region decreased the ability to translate, some resulted in a greater inhibition than others. Interestingly, one mutant showed a stimulatory effect on translation when tested in the in vivo system, suggesting a complex interrelation of sequences or structures within this region of the HA V 5' NCR required for internal ribosome binding. Materials and Methods Plasmid Constructions The parental plasmid, pCHL, was constructed using sequences from three separate plasmids (Figure 3.1). The polymerase chain reaction (PCR) was used to construct pA5LIII-2, 74 a monocistronic plasmid containing the HA V 5'NCR and 7 amino acids of the HAV PI region (as determined by initiation at the first AUG at nt 730) fused in-frame to the firefly luciferase coding sequences. The oligonucleotide primers used in the above PCR reaction were: primer 1, 5'-GGA CTG AGC TCT AAA CGG TCC TGG CAA TTT GGA CTT TCC GCC CTT-3' (corresponding to nt 1953-1933 on the minus strand of the luciferase gene and introducing a Sacl site) and primer 2, 5'-AC ATG TCT AGA CAA GGT ATT TCC GAA GAC GCC AAA AAC ATA AAG-3' (corresponding to the plus strand of pHA V 17 sequences from nt 734-756 fused directly to nt 305- 325 of the luciferase gene). The template DNA was nt 305 to 1933 from the luciferase coding region. The product of this reaction created a fusion of the first seven amino acids of HAV in-frame with the luciferase gene. The blunt-end PCR product was ligated to a Blue-Script KS+ plasmid (Stratagene, Inc.) previously digested with Smal. White colonies were screened for the presence of the HAV -luciferase sequences in the correct orientation. Positive clones (pBL-Luclll) were digested with Sacl to yield a 1.7 kb HA V Iluciferase fragment which was purified using Geneclean (Biol0l) after electrophoresis in a 1 % agarose gel. This fragment was ligated to a Sacl digested, dephosphorylated vector pGEM2Ad-A5 (the pGEM2 [Promega] plasmid minus the AflIII site at nt 2656, containing a 743 bp Xbal/Hindll1 fragment corresponding to the 5'NCR of pHA V 17). After transformation of competent E. coli DH5a, positive colonies were screened for plasmids with inserted HA V -luciferase sequences (pAdA5- Luclll). These clones were digested with Afllll to remove a 0.6 75 Kb fragment containing repeated and MCS sequences, gel purified, and religated. The resulting plasmid, pA5LIII-2, contains the entire HA V 5'NCR plus 7 amino acids of PI fused directly to the luciferase gene, all under control of the T7 promoter. pCHL was constructed by digestion of pA5LIII-2 with HindIII, Klenow fill­in, and ligation to a blunt-end bacterial chloramphenicol acetyl transferase (CAT) gene 800 bp fragment which was derived from a BamHI digestion and subsequent Klenow fill-in of pPT -CAT (a plasmid containing the CAT gene inserted into a blue script KS+ vector). pCHL was used as the parental plasmid for all the mutant constructions. A series of deletions and insertions were derived from the parental plasmid, pCHL, by digestion with convenient restriction endonucleases, separation of these fragments by agarose gel electrophoresis and purification using Geneclean. The purified fragments were filled-in using Klenow or digested with T4 DNA polymerase to yield blunt ends. The resultant fragments were religated and transformed into E. coli DH5a.. Restriction endonuclease digestions were carried out with the following enzymes: partial digestions with Nco 1 (nt 45), N si 1 (nt 523), AflIII (nt 734) and complete digestions with Hpal (nt 347) and BamHl (nt 628). These digestions on their own and in various combinations yielded the constructs shown in Figure 3.1 B. 76 Figure 3.1. Schematic representation of the bicistronic construction and activities of both cistrons.(A) The construct, pCHL, was made by first fusing the firefly luciferase gene in frame with 21 nucleotides (nt) of the HA V PI sequence. The entire HA V 51NCR was then ligated upstream and finally, the bacterial chloramphenicol acetyl-transferase gene (CAT) was placed upstream of the HAV 51NCR. pCHL contains 14 nucleotides of noncoding sequences prior to the CAT gene, the entire 736 nt of the HAV 51NCR and 21 nt of HAV PI fused to the entire luciferase gene. (B) By using convenient restriction endonuclease sites, mutations were made throughout the HA V 51 NCR. The letters denote the plasmid nomenclature, whereas the numbers refer to the sequences where deletions or insertions were created. (C) Activities of CAT and luciferase after translation in vitro. Several RNA concentrations were used to program an in vitro translation reaction in a rabbit reticulocyte lysate system. After translation for 1 hour at 30°C, one-tenth of each reaction was assayed for both CAT and luciferase activities as described in Materials and Methods. 77 A Luciferase Primer 1 +2 Aflil. ~PCR Sacl BamHI BamHI CAT 8CObp Sac. Blunt-end ligation pHAVl7 Sacl 1.7Kb HindUI Klenow luciferase Blunt-end Ligation pCHl (5.8Kb) Nhel 1 Xbal Hindfll pGEM-2 1 Aflill Klenow fill-in Hindlll Ligation Sac. 78 C B HAY BICISTRONIC CONSTRUCTS 100 1 CAT HAY 5'NCR LUClfERASE ~ pCHL ." -Q) 80 co ~ 45 Ne ZT} >- 60 j 45+4 >-'S _al '> 523 :;:;::E 40 N ~< 523·4 -(TGCA)- 0 t- <iI- 20 1 / 628 0_ B 628+4 7\: o f I -GATC-i 0 20 40 SO 80 100 120 734 A ZT$ [RNA] ug/ml 734+4 45 347 NeH 12000 i V 45-347 347 523 10000 >~ 1 / 34H7N-5 23 V -E 8000 ~ 523 628 > 6000 NB V 523-628 U co 347 628 4000 0 H8 V ::;) 347-628 ..J 2000 347 734 HA O~ i I 347-734 V 0 20 40 S(} 80 100 120 [RNA] ug/ml 628 734 AB 62B-734 V -.) \0 In Vitro Transcription and Translation Transcriptions and translations were carried out as described previously (Glass and Summers, 1992). Plasmids were linearized with Nhel, extracted with phenol/chloroform, precipitated with ethanol, and 1-2 Jlg DNA used to program an in vitro transcription reaction driven by the T7 promoter on each construct. Capping reactions were supplemented with 300J,lM m7GpppG and GTP levels were decreased to 0.25 mM. The RNAs were phenol/chloroform extracted, precipitated with ethanol in the presence of 2.5 M Nf4 acetate and IJ,lg directly used in an in vitro translation reaction (Promega in vitro Translation kit), with or without the addition of exogenous [35S]-methionine (15J,lCi/ 10JlI reaction). Products were analyzed using CAT and luciferase assays and/or by electrophoresis on a 10 % polyacrylamide SDS gel. Immunopre