National Academy of Sciences is collaborating with JSTOR to digitize preserve and extend access to Proceedings of the National Academy of Sciences of the United States of America http www jstor org A Translation Attenuating Intraleader Open Reading Frame is Selected on Coronavirus mRNAs During Persistent Infection Author s Martin A Hofmann Savithra D Senanayake and David A Brian Source Proceedings of the National Academy of Sciences of the United States of America Vol 90 No 24 Dec 15 1993 pp 11733 11737 Published by National Academy of Sciences Stable URL http www jstor org stable 2363537 Accessed 01 11 2015 03 35 UTC Your use of the JSTOR archive indicates your acceptance of the Terms ORF open reading frame PCR polymerase chain reaction N nucleocapsid protein M mul tispanning membrane protein S spike protein I second protein encoded by the bicistronic N mRNA Present address Institut fur Viruskrankheiten und Immunoprophy laxe CH 3147 Mittelhausern Switzerland iTo whom reprint requests should be addressed 11733 This content downloaded from 128 6 218 72 on Sun 01 Nov 2015 03 35 23 UTC All use subject to JSTOR Terms and Conditions 11734 Microbiology Hofmann et al Proc Natl Acad Sci USA 90 1993 anneal to a site within the coding region of the mRNA but near the 5 untranslated region and used in a primer exten sion reaction with reverse transcriptase to copy the 5 end of the mRNA Extension amplification and cloning reactions were carried out separately for each mRNA species Primer 1 for N mRNA bound to a 26 nt region beginning 34 nt downstream from the N initiation codon for M mRNA to a 20 nt region beginning 19 nt downstream from the M initiation codon and for S mRNA to a 21 nt region beginning 23 nt downstream from the S initiation codon Extended primers were purified by electrophoresis in a 6 polyacrylamide sequencing gel containing 50 urea and were eluted Eluted extension products were ligated head to tail with T4 RNA ligase and amplified by PCR For this a second primer primer 2 complementary in sequence to primer 1 and a third primer primer 3 complementary to nt 30 48 of the leader 13 were used in the thermocycling reaction Ampli fied products therefore represented the ligated head to tail junctions of molecules that contained the first 29 nt of the leader assuming a leader of consensus length e g one having a type I terminus The PCR products were ligated into Sma I linearized nondephosphorylated pGEM 3Zf vec tor Promega DNA and used to transform Escherichia coli JM109 cells An oligodeoxynucleotide homologous to nt 7 33 of the leader 13 was used to screen by colony hybridization for clones containing a leader specific insert DNA from the identified colonies was asymmetrically amplified by PCR and sequenced by the dideoxynucleotide chain termination method using the forward and reverse primers for the pGEM 3Zf vector 14 15 Testing the Effect of the Intraleader ORF on Translation of Downstream ORFs cDNA of BCV N mRNA was cloned into the pGEM 3Zf vector to give pLN I 13 T7 polymerase generated transcripts of Mlu I linearized pLN I began with 5 GAUUGUG type I terminus and ended with a 3 poly A of 21 nt pLN II generated from pLN I by an oligonucle otide directed G A mutation at the fifth nucleotide yielded transcripts with a 5 terminal sequence of GAUUAUG a type II terminus For quantitative analysis of translation efficien cies transcripts of pLN I and pLN II were separately pre pared and quantitated before translation Transcripts were digested with RNase free DNase Promega chromato graphed on a Bio Spin 6 column Bio Rad and quantitated by spectrophotometry A portion was resolved by formaldehyde agarose gel electrophoresis and quantitated by blotting and RNA hybridization to 32P end labeled primer 1 of known specific activity 1 9 x 107 cpm g in order to normalize the amount of full length transcript to be used in translation Blots were quantitatively analyzed with the Ambis radioanalytic imager 7 Proteins were synthesized in vitro in rabbit retic ulocyte lysate in the presence of 35S methionine and analyzed by SDS PAGE as described 13 Quantitative protein mea surements were made from duplicate analyses on two inde pendent preparations of RNA transcripts A Type I Type 11 Type III Type IV s e IG5A G5 G G GATC GAT GATC B 10 20 30 40 5 GAUUAUGAGCGAUUUGCGUGCGUGCAUCCCGCUUCACUGA met ser asp Ieu arg ala cys ile pro Ieu his peptide 50 60 70 UC UCUUGU UAGAUCUUUUUAUMAUCUMAACUUUMAGGAUG FIG 1 A Four basic types of 5 termini found on N mRNA species Asymmetrically amplified single stranded DNAs from cloned 5 termini shown as head to tail ligated products were sequenced by the chain termination method B ORF in the leader with a type II terminus on the N mRNA The peptide potentially translated from the intraleader open reading frame is shown below RESULTS An Intraleader ORF Is Selected During Persistent Infection in Cell Culture In initial efforts to establish the leader sequence on BCV subgenomic mRNAs we were unable to resolve an apparent heterogeneous 5 terminal sequence by chemical sequencing of end labeled extended primers un published data To resolve the heterogeneity we developed a method to sequence extended primers that had been first ligated head to tail with RNA ligase amplified by PCR and cloned refs 12 and 14 Fig 1 The analysis was applied to the N M and S mRNA species isolated from virions and from cells in culture during an acute infection the first 18 hr postinfection and throughout a period of persistent infection through 432 days The results are summarized in Tables 1 and 2 and Fig 2 Three major points emerge from these data i The 5 terminal sequences of mRNA were heterogeneous and hy pervariable for over a 120 day period of infection The heterogeneous termini furthermore could be classified into two broad categories In the first 285 out of a total of 352 clones were termini that made up one of the four basic types I 5 GAUUGUG II 5 GAUUAUG III 5 GAAUAUG or IV 5 GAUAUG Fig 1 and Table 1 Among these were termini with a truncated first base but still identifiable as one Table 1 Distribution of 5 termini found on the N M and S mRNAs at various times postinfection No of clones Time N mRNA M mRNA S mRNA post infection I II III IV Aber I II III IV Aber I II III IV Aber Virion 7 3 2 4 6 1 7 3 1 5 8 2 18 hr 11 2 7 4 2 1 8 6 11 10 4 days 9 11 1 5 5 1 1 1 1 8 7 2 4 120days 13 4 5 5 2 1 5 1 1 2 5 7 7 296 days 7 14 5 1 1 5 7 14 14 1 1 1 432 days 21 1 12 6 16 The 5 termini were identified as types I GAUUGUG II GAUUAUG III GAAUAUG and IV GAUAUG or as aberrant Aber as described in Table 2 Numbers in parentheses indicate clones with a truncated first base Packaged mRNAs were isolated from purified virus harvested at 18 hr postinfection This content downloaded from 128 6 218 72 on Sun 01 Nov 2015 03 35 23 UTC All use subject to JSTOR Terms and Conditions Microbiology Hofmann et al Proc Natl Acad Sci USA 90 1993 11735 Table 2 Aberrant 5 terminal sequences on N M and S mRNAs at various times postinfection Time post in fection N mRNA M mRNA S mRNA 1 5 10 15 1 5 10 15 1 5 10 15 Virion guguuGAUUU gGAUUGUGAGCGAUUU aaGAUAUGAGCGAUUU uUGAGAGAUUU aGAUUGUGAGCGAUUU ugGAUUGUGAGCGAUUU gccGgccCuGAUUU uGAUUGUGAGCGAUUU ggAggGaauGcGAUUU uaacccUGcauauGAUUU X3 ggacuuuagcGAUGUuuauGAUUU 18 hrt uaauGgcAUUU GAccAUGcugGuUUU gGAUAUGAGCGAUUU x2 GgccGuGAUUU guAccAcGAaaauUUU aGAUAUGAGCGAUUU guUccGguAUUU gGAUUGUGAGCGAUUU aaGAUAUGAGCGAUUU aagcAGCGAUUU aGAUUGUGAGCGAUUU x2 uGAUUGUGAGCGAUUU gGgccGgGAUUU aGAUUAUGAGCGAUUU ucAUUGUGAGCGAUUU X2 GAUAcuAGCGAUUU gaccauGcUgGUGAGCGAUUU uggcaGAUUGUGAGCGAUUU guucaaaGUcGguAauGAUUU gaccaugcuGgUGaGgGCuuUUU uggugGAAUGUGAGCGAUUU 4 dayst AUcGcGAGCGAUUU GAccAUGcugGuUUU aGAUAUGAGCGAUUU GAUcGcGAGCGAUUU ugaGAUAUGAGCGAUUU GAUUGUaAGCGAUUU ugagGAUAUGAGCGAUUU auuauuAUGAGCGACAU ggagcaGAcUAUaAagGuUUU 120 days gcAUAUGAGCGAUUU gGAAUAUGAGCGAUUU uCcuUaaa guAUAUGAGCGAUUU AcuuAgaca x2 gauAUAUGAGCGAUUU X3 cGAUAUGAGCGAUUU aGAUAUGAGCGAUUU gauaUAUGAGCGAUUU ccaAaUAUGAGCGAUUU 296 days GuUAUGAGCGAUUU X2 uUUAUGAGCGAUUU auuAUUAUGAGCGAUUU x3 432 days Lowercase sequences are those that do not conform to one of the four basic terminus types Italicized sequences except for the underlined bases are identical to stretches of sequence found within the genome Packaged mRNAs were isolated from purified virus harvested at 18 hr postinfection tThe 5 sequence uucuauuagcacugauacuguagcuguuacuaaugguuuaggacuuGAUUU was found on one clone of S mRNA at 18 hr and the sequence uucuauuagcacugauacuguagcuguuacuaaugguuuagguacuuaAUUU was found on an S mRNA clone at 4 days of the four basic types The single base truncation could have resulted from a premature termination of the reverse tran scriptase during synthesis of first strand cDNA in the cloning procedure or alternatively could represent the actual Type I Type II Type III Type IV GAUUGUG GAUUAUG GAAUAUG GAUAUG Virion i50 100 4 dpi IIIIIof 7777 77 I 1O O total 120 dpi 50 100 296 dpi I50 100 432 dpi Z50 N M S N M S N M S N M S mRNA species FIG 2 Occurrence of the four basic types of 5 termini at various times postinfection hr hpi days dpi Results for the N M and S mRNA species are depicted separately Aberrant termini listed in Table 1 were included in the total number when percentages were calculated mRNA terminus as suggested by the large number of such termini at 432 days postinfection Table 1 Regardless of origin for this study clones truncated by one base were scored as one of the four major types in Table 1 In the second category were clones the remaining 67 out of 352 clones containing aberrant sequences some replacing up to 16 nt of the 5 terminal sequence and some extending 5 ward from base 1 for a distance of 1 37 nt Table 2 Of the 39 5 ward extensions 30 represented extensions of 1 5 nt on one of the four basic terminus types described in Table 1 The origin of the extended sequences is not apparent but they appear not to have resulted from an enzymatic copy back mechanism Some of the aberrant sequences included stretches of 6 nt identified by italics in Table 2 that could have been derived from within of one of the eight open reading frames in the sequenced 3 terminal 8955 nt of the BCV genome The mechanism by which these sequences were incorporated is likewise not understood Two pieces of evidence indicate that the sequence hetero geneity was not a result of the method used to make the analyses First when the leader sequence with the type I 5 terminus 5 GAUUGUG was cloned and the resulting T7 RNA polymerase generated capped transcripts from a single clone were further cloned and sequenced by the protocol described above only the type I terminus was recovered in the cloned products 6 of 6 clones 12 When transcripts of this construct were transfected into uninfected cells reisolated and used as templates for cloning again only the original sequence 18 of 18 clones was recovered 12 Second because cloned termini obtained after a long period of persis tent infection prepared both at 296 and at 432 days had a clearly different makeup predominantly 5 GAUUAULG type This content downloaded from 128 6 218 72 on Sun 01 Nov 2015 03 35 23 UTC All use subject to JSTOR Terms and Conditions 11736 Microbiology Hofmann et al Proc Natl Acad Sci USA 90 1993 II than those obtained during acute infection predominantly 5 GAUUGUG type I Table 1 and Fig 2 more than a mere random generation of sequences occurring during our analyt ical procedure would have had to play a role ii The 5 termini on the separate N M and S messages coevolved coordinately to yield a predominant terminus type at various times postinfection Table 1 and Fig 2 In the virion and early in infection continuing through 4 days postinfection the predominant terminus was type I 5 GAUUGUG and we infer that this may have been the progenitor on the basis of its predominance Terminus types II 5 GAUUAUG and IV 5 GAUAUG but not type III 5 GAAUAUG were found in low numbers during this time By 120 days postinfection the predominant termini were types III 5 GAAUAUG and IV 5 GAUAUG on all three mRNAs and by 296 days and remaining through 432 days postinfection the predominant terminus was type II 5 GAUUAUG on all three mRNAs iii Throughout establishment of the persistent infection leader sequences with a 33 nt internal ORF were selected Fig 1 That is by 120 days postinfection the vast majority of 5 termini were types II III and IV or the aberrant variants all with a methionine codon initiating a 33 nt ORF For types II and III the methionine codon begins at nt 5 and for type IV it begins at nt 4 At 296 days postinfection and beyond all termini possessed the methionine codon Accord ing to the ribosomal scanning model for initiation of protein synthesis on eukaryotic mRNAs 16 the sequence surround ing the methionine codon fits a consensus that is favorable for initiation of translation since there is a purine in the 3 position A in the case of types II and III and G in the case of type IV and an A in the 4 position Fig 1 The potential peptide Met Ser Asp Leu Arg Ala Cys Ile Pro Leu His en coded by the ORF is neutral and hydrophilic and its syn thesis would be terminated by a UGA codon The Intraleader ORF Attenuates Translation of Downstream ORFs in Vitro To test whether the intraleader ORF is translated into a product and whether it attenuates translation of downstream ORFs equimolar amounts of the two forms of N mRNA differing only by the absence or presence of the intraleader ORF i e a type I or a type II terminus were translated in vitro and the products were quantitated To date we have found no radiolabeled polypeptide of the size predicted by the intraleader ORF 1 1 kDa when either 35S methionine or 35S cysteine was included in the transla tion mixture and when products were analyzed in a denatur ing gel of 20 polyacrylamide 17 We observed however a 47 5 3 mean SD reduction in the amount of N and a 17 4 5 reduction in the amount of I produced when the intraleader ORF was present Fig 3 DISCUSSION By exploiting an experimental approach that has enabled us to sequence the 5 ends of individual BCV mRNA species we have learned that the 5 terminal sequence of the leader is heterogeneous and hypervariable throughout early infection and during establishment of a persistent infection and that it becomes essentially homogeneous and stable by 432 days of persistent infection Although the mechanism by which the heterogeneity was generated remains unknown we have experimentally ruled out the possibility that it was artifactu ally derived from the procedures of amplification and cloning used in our analysis since transcripts of cloned constructs demonstrated no such heterogeneity when carried through the procedure 12 We think it is also unlikely that the heterogeneity resulted from a cap snatching mechanism sim ilar to that described for the myxoviruses 18 19 bunyavi ruses 20 21 or arenaviruses 22 23 in which the 5 terminal sequences are obtained from cellular mRNAs during the transcription process since this mechanism would have A21 L vector T7 a Mlul 2 NrA21 vector ORF I Without With ORF ORF N a 1 2 FIG 3 Effect of the intraleader ORF on the translation of N and I proteins in vitro from the bicistronic N mRNA 13 Equal amounts of each full length transcript 1 ug made in vitro by T7 RNA polymerase were translated in rabbit reticulocyte lysate in the presence of 35S methionine and the radioactive products were analyzed by gel electrophoresis and quantitated by an Ambis radio analytic imager Lanes 1 pLN I 2 pLN II yielded a heterogeneous pattern throughout long term infec tion and not the observed periodic predominance of one kind of terminus over another What is the origin of the 5 terminal hypervariability It has been clearly established that the common 5 leader encoded only at the 5 end of the coronavirus genome can be the only source of the 5 leader on subgenomic mRNAs when the genome alone is the infectious unit and when mRNAs are transcribed from the antigenome 3 5 Once subgenomic mRNAs are present however it becomes less clear what molecules can serve as leader donors since anti mRNAs have been shown to possess a minus strand copy of the leader 9 a potential leader template and the leader fusion process may function in trans a process called leader switching 24 A subgenomic defective interfering RNA when present 7 could also theoretically serve as a leader donor Thus the presence of a wide variety of 5 termini on packaged virion mRNAs and on mRNAs extracted during the acute phase of the infection first 4 days lead us to conclude that despite our use of stock inoculum that had been prepared from sequen tially plaque purified virus a variety of preexisting termini from multiply packaged genomes or subgenomic molecules may have been introduced by the infecting virus Thus at least two possible mechanisms could explain the apparent hyper variability In the first predominant terminus types could have been periodically selected from a preexisting pool In the second base changes may have occurred as the result of the high error frequency 1o 3 to 10 4 inherent among RNA dependent RNA polymerases 25 giving rise to variant ter mini that were subsequently selected The type II terminus 5 GAUUAUG for example could have arisen from a type I terminus 5 GAUUGUG by a G A transition at base position 5 and subsequently given rise to a type III terminus 5 GAAUAUG by a U A transversion at position 3 A more rigorous analysis of the 5 end hypervariability must This content downloaded from 128 6 218 72 on Sun 01 Nov 2015 03 35 23 UTC All use subject to JSTOR Terms and Conditions Microbiology Hofmann et al Proc Natl Acad Sci USA 90 1993 11737 await the preparation of a virus inoculum that has been derived from a single infectious genomic RNA molecule We postulate that selection of leaders with an intraleader ORF during persistent infection is causally related to coro naviral persistence This idea is based on a large body of experimental evidence showing that an intraleader ORF on eukaryotic mRNAs attenuates translation of the mRNA ORF 16 and secondly on the notion that persistent infection by a virus is promoted when the cytolytic potential of a virus is diminished 1 Conceivably attenuating coronavirus mRNA translation in this manner would impair virus replication and bring it into equilibrium with cell growth and division This would be an especially potent mechanism if as found in this study the same intraleader ORF arises simultaneously on the genome and all subgenomic mRNAs Alternatively attenu ating expression of a specific cytopathogenic gene could inhibit the cytolytic virulent phenotype Persistent infections by cytoplasmically replicating RNA viruses have been correlated with such genetic changes as deletions and rearrangements resulting in subgenomic defec tive interfering RNAs reviewed in ref 26 point mutations in structural proteins affecting viral entry 27 or assembly 28 and point mutations in the 5 untranslated region affect ing translation rates through a non ORF mechanism 29 Here we identify a selectable upstream ORF that potentially functions as an attenuator of virulence leading to persistent infection Precedent for an intraleader ORF operating in a strain speciflc manner to attenuate virulence has been re ported for the barley stripe mosaic hordeivirus 30 but in this instance persistence was not a reported consequence of attenuation Inspection of known coronavirus leader se quences all of which were studied during acute lytic infection and show no intraleader ORF 4 9 31 35 verify the potential of an intraleader ORF as a selectable element Translation attenuating intraleader ORFs occurring as sta ble sequence elements in mRNAs of cells and viruses atten uate translation by one of two known mechanisms 16 either or both of which might function in BCV i The peptide product of the intraleader ORF as in the case of the yeast gene CPAJ 36 may be responsible for attenuating expres sion ii The intraleader ORF may d