BARE retrotransposons are translated and replicated via distinct RNA pools

BARE Re o ansposons A e T ansla ed and Replica ed ia Dis inc RNA Pools Wei Chang1, Ma ko Jääskeläinen1, Song-ping Li2¤, Alan H. Schulman1,3* 1 Ins i u e o Bio echnology, Viikki Biocen e , Uni e si y o Helsinki, Helsinki, Finland, 2 Genome-Scale Biology P og am, Uni e si y o Helsinki, Biomedicum, Helsinki, Finland, 3 Bio echnology and Food Resea ch, MTT Ag i ood Resea ch Finland, Jokioinen, Finland Abs ac The eplica ion o Long Te minal Repea (LTR) e o ansposons, which can cons i u e o e 80% o highe plan genomes, esembles ha o e o i uses. A majo ques ion o e o ansposons and e o i uses is how he wo con lic ing oles o hei ansc ip s, in ansla ion and e e se ansc ip ion, a e balanced. He e, we show ha he BARE e o ansposon, despi e i s o ganiza ion in o jus one open eading ame, p oduces h ee dis inc classes o ansc ip s. One is capped, polyadenyla ed, and ansla ed, bu canno be copied in o cDNA. The second is no capped o polyadenyla ed, bu is des ined o packaging and ul ima e e e se ansc ip ion. The hi d class is capped, polyadenyla ed, and spliced o a o p oduc ion o a subgenomic RNA encoding only Gag, he p o ein o ming i us- like pa icles. Mo eo e , he BARE2 sub amily, which canno syn hesize Gag and is pa asi ic on BARE1, does no p oduce he spliced sub-genomic RNA o ansla ion bu does make he eplica ion compe en ansc ip s, which a e packaged in o BARE1 pa icles. To ou knowledge, his is i s demons a ion o dis inc RNA pools o ansla ion and ansc ip ion o any e o ansposon. Ci a ion: Chang W, Jääskeläinen M, Li S-p, Schulman AH (2013) BARE Re o ansposons A e T ansla ed and Replica ed ia Dis inc RNA Pools. PLoS ONE 8(8): e72270. doi:10.1371/jou nal.pone.0072270 Edi o : Chen Liang, Lady Da is Ins i u e o Medical Resea ch, Canada Recei ed May 28, 2013; Accep ed July 14, 2013; Published Augus 6, 2013 Copy igh : © 2013 Chang e al. This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed. Funding: This wo k was suppo ed by he Academy o Finland (h p://www.aka. i/ i/A/), P ojec 123074. Funding o open access cha ge: Uni e si y o Helsinki. The unde s had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o he manusc ip . Compe ing in e es s: The au ho s ha e decla ed ha no compe ing in e es s exis . * E-mail: [email p o ec ed] ¤ Cu en add ess: Depa men o Pa hology, Uni e si y o Tu ku and Tu ku Uni e si y Hospi al, Tu ku, Finland In oduc ion Long e minal epea (LTR) e o ansposons o Class I ansposable elemen s a e ubiqui ous in he euka yo es and can comp ise o e 80% o he la ge genomes o plan s [1,2]. They p opaga e simila ly o he in acellula phase o e o i uses: by a “copy and pas e” cycle o ansc ip ion o genomic copies, ansla ion, packaging o ansc ip s in o i us- like pa icles (VLPs) composed o Gag, e e se ansc ip ion, and a ge ing o he cDNA copy o he nucleus o in eg a ion in o he genome [3,4]. The li ecycle depends upon p o eins encoded by he e o ansposon i sel . The s uc u al Gag is o en in a sepa a e open eading ame (ORF) om Pol, which encodes he enzymes e e se ansc ip ase (RT), RNAse H (RH), aspa ic p o einase (PR), and in eg ase (IN). The s oichiome y be ween gag and pol p oduc s is c i ical o eplica ion because he assembly o he VLP equi es excess Gag ela i e o he enzymes [5,6]. A common s a egy among e o i uses and e o ansposons o p oduce mo e Gag is -1 o +1 ansla ional ameshi ing be ween gag and pol [6–8]. Howe e , he Copia supe amily o e o ansposons [3] and also mos plan e o ansposons ha e only a single ORF [9]. While al e na i e splicing in copia o D osophila dele es i ually all o pol, 2950 n , gene a ing an RNA dedica ed o Gag ansla ion [10], his has no been seen o o he membe s o he supe amily. In some cases pos - ansla ional p o ein deg ada ion se es o achie e a mola excess o Gag [11]. Ano he majo conund um is ha e e se ansc ip ion o e o ansposon and e o i us RNA, which des oys he RNA empla e, con lic s wi h i s u he ansla ion [12]. Al e na i ely, ins ead o a single pool o RNA, sepa a e popula ions may se e each unc ion wi h no all RNA being seques e ed in o capsids o e e se ansc ip ion. Among he e o i uses, Mu ine Lukemia Vi us (MLV) may use dis inc RNA pools [13], whe eas HIV-1 and -2 do no [14]. The ques ion has no been in es iga ed o e o ansposons. Al hough e o ansposons comp ise much o mos plan genomes, he de ails o hei li ecycle ha e been in es iga ed o only a ew. The BARE e o ansposon o supe amily Copia accoun s o o e 10% o he ba ley genome [2,15]. BARE1 has no ameshi be ween gag and pol [16,17]. A a ian called BARE2 canno exp ess Gag [16,18]. BARE1 and BARE2 p oduce mul iple classes o RNA ansc ip s om wo TATA PLOS ONE | www.plosone.o g 1 Augus 2013 | Volume 8 | Issue 8 | e72270 boxes, o which only 15 o 25% a e polyadenyla ed [19]. Mo eo e , hose which a e polyadenyla ed lack he R domain needed o e e se ansc ip ion. These obse a ions aise he ques ions add essed he e: which BARE RNAs se e o ansla ion, which ones a e packaged, and does BARE use an al e na i e o ameshi ing o Gag p oduc ion. We we e able o demons a e no only di e en RNA pools o ansla ion and e e se ansc ip ion bu also a no el splicing pa e n o Gag syn hesis. Ma e ials and Me hods Plan Ma e ials and RNA Isola ion Ba ley (Ho deum ulga e L.) plan ma e ials and callus cul u es used o RNA isola ion, as well as he me hods used o isola e he RNA, a e desc ibed in he Ma e ials S1. RNA End S uc u e The p esence o a 5’ 7-me hylguanosine cap on BARE ansc ip s was assayed by he p ocedu e called RLM- media ed apid ampli ica ion o 5’ cDNA ends (5’ RLM-RACE) using a ki (Fi s Choice®RLM-RACE, Ambion AM1700) wi h small changes and a cus om adap e . To examine 3’ polyadenyla ion o poly ibosome-associa ed RNA, 3’ RLM- RACE was ca ied ou . De ails o bo h me hods a e p esen ed in he Ma e ials S1. Analysis o BARE1 and BARE2 Exp ession Le els To e alua e he ela i e exp ession le els o BARE1 and BARE2, RT-PCR was ca ied ou using p ime AP4 (Table S1) o p ime cDNA syn hesis and hen p ime s LS1 (Table S1) and AP4 o he ampli ica ion eac ion. The p ime pai binds bo h o BARE1 and BARE2, and ampli ies hem equally well [16]. The wo e o ansposon amilies we e ampli ied om genomic DNA using he same p ime pai . Poly ibosome Isola ion and RT-PCR Ba ley callus cul u es cells we e collec ed, ozen unde liquid N2, and hen pul e ized wi h mo a and pes le unde liquid N2. Poly ibosomes we e isola ed la gely as p e iously desc ibed [20]. The p ocedu e is desc ibed in de ail in he Ma e ials S1. Splicing assays Splicing analysis was made wi h 1µg o RNA ea ed wice wi h DNase, e e se ansc ibed in o cDNA as using he BARE- speci ic p ime 81567. The BARE ansc ip s we e ampli ied ollowing cDNA p oduc ion using wo p ime s close o he spliced egion, F1593 and F1594 (Table S1), and a PCR p og am consis ing o 94oC o 5 min, 40 cycles o 94oC o 30 sec, 56oC o 1 min, and 72oC o 1 min, wi h a inal ex ension a 72oC o 5 min. To in es iga e splicing in polyadenyla ed RNA, p ime 81567 was used o ini ia e cDNA syn hesis, ollowed by F1594 and AP4 o PCR ampli ica ion. Fo de ec ion o capped RNA splicing, an RNA linke was i s liga ed o dephospho yla ed and decapped RNA and hen cDNA syn hesized as abo e. Two PCR eac ions we e p epa ed by 5’ RACE using he linke and 81567 as he p ime pai . One eac ion was ampli ied, he o he se ed as a con ol. The second RACE eac ion used 1µl o ei he he i s PCR p oduc o he con ol as he empla e and p ime s F1594 and 81567. Con ols p oduced no signal om he second ampli ica ion. The decapped RNA ga e he same size p oduc as did o al RNA RT-PCR using same p ime pai . Resul s BARE1 bu no BARE2 RNA Is Spliced Ampli ica ion o BARE om o al RNA p oduced wo p oduc s, one consis en wi h he size o genomic BARE copies and he o he somewha smalle (Figu e 1). Ampli ica ions om genomic DNA (Figu e 1B) p oduced no smalle p oduc . A smalle p oduc was ampli ied om he RNAs o all issues es ed, which we e callus and emb yo (Figu e 1C) as well as lea and oo (da a summa ized in Table 1). A o al o 60 clones we e sequenced om he mo e abundan , longe p oduc ; BARE1 and BARE2 we e equally p esen among he sequences. The wo la ge p oduc s seen in callus RNA (Figu e 1C) di e only by ampli ica ion om a seconda y PCR p iming si e. All he sequences om he sho p oduc , howe e , we e om BARE1 and con ained a dele ion a he beginning o p domain, comp ising a segmen o 104 n lanked by GT and AG espec i ely a he le and igh bo de s (Figu e 2A). The p esence o he sho o m o BARE1 in he RNA bu no he genome sugges ed ha i is a spliced ansc ip . The BARE1 genomic sequences con ain conse ed CAG/GTAT and CAG/GA mo i s espec i ely ma ching he 5’ and 3’ junc ions CAG/---/GA ha lank he segmen missing in he mino cDNA sequence (Figu e 2A). These a e a e y good ma ch o he consensus sequence AG/GT o he dono si e in a genome-wide su ey o he model species B achypodium dis achyon [21] and o consensus dono and accep o splice si es, espec i ely C(A) AG/GTA and CAG/G, in A abidopsis and ice [22,23]. The BARE1 junc ions a e also well iden i ied by he Ne gene2 splice si e p edic o (h p://www.cbs.d u.dk/ se ices/Ne Gene2/) wi hin he A abidopsis genome and by SpliceP edic o (h p://deepc2.psi.ias a e.edu/cgi-bin/sp.cgi) agains maize and human genomes, suppo ing he in e p e a ion ha he sho e BARE ansc ip is a splicing p oduc . No ably, he p edic ed splicing signals a e no ound in he BARE2 genomic o RNA sequences (Figu e 2A). The a eas immedia ely 5’ o he dono si e and 3’ o he accep o si e o BARE1 a e also di e gen in BARE2, al hough he sequence o he acul a i e in on is qui e simila in bo h. While he sho o m comp ised a mino ac ion o o al RNA comp ising bo h BARE1 and BARE2, abou 12.5% using he p ime s (Figu e 1A) ha ampli ied bo h, i ep esen ed ully hal o he BARE1- speci ic p oduc ampli ied om polyadenyla ed RNA (Figu e 1C). The p edic ed and sequenced splice junc ion is 2 n beyond he end o he gag coding domain [16,17], he eby c ea ing a s op codon h ee amino acids beyond he end o Gag (Figu e 2B,C,D) ollowed by many mo e wi hin pol. Consequen ly, he spliced RNA can exp ess only Gag (Figu e 2D). The p edic ed molecula weigh o he Gag om he BARE Re o ansposon T ansc ip ion o Replica ion PLOS ONE | www.plosone.o g 2 Augus 2013 | Volume 8 | Issue 8 | e72270 Figu e 1. Splicing o BARE ansc ip s. A. Aga ose gel elec opho esis o he RT-PCR ampli ica ion p oduc om callus RNA using p ime s LS1 and AP4. The uppe band (a ow, 1.3 kb) con ains p oduc s om bo h BARE1 and BARE2 and is he same size as he ampli ica ion p oduc om genomic DNA using he same p ime pai (B); he lowe , ain band (A, a ow, 1.2 kb) is he spliced BARE1 o m and i is no seen in he genomic DNA ampli ica ion. B. RT-PCR (+) om o al RNA (p ime s LS1 and AP4); he lanes display eac ions con aining e e se ansc ip ase (+), nega i e con ol lacking e e se ansc ip ase (-), o genomic DNA ins ead o RNA (G) as he posi i e con ol. C. De ec ion o splicing in emb yo, E, and callus, C, o al and poly(A) RNA (labelled p (A)) using a BARE1-speci ic p ime pai (81567, F1594). A ows indica e he unspliced and spliced o ms. Size ma ke s (m), 100 bp ladde , ma ked band is 1 kb (A, B) o 0.5 kb (C). doi: 10.1371/jou nal.pone.0072270.g001 spliced RNA is 32.1 kDa, he same as p edic ed om he elec opho e ic mobili y o Gag om VLPs [18]. In ons in plan s a e gene ally 15% mo e U- ich han he lanking exons, while exons a e 15% mo e GC- ich han he co esponding in on [24–26]. In BARE1, he 52 n lanking he slicing signals a e 13.5% U s. 37.5% U o he 104 n in on, making he in on 24% and 2.8- old mo e U- ich han he exon segmen s. In BARE2, which does no splice, he in on egion is only 16.7% and 1.9- old mo e U- ich han he su ounding egion. Mo eo e , hese lanking exon egions in BARE1 a e GC- ich, being 53.8% GC, 16.3% mo e GC han he in on, whe eas he co esponding BARE2 segmen is 47.2% GC, only a 7.6% di e ence wi h he su ounding egions. Bo h hese measu es and he splice si e compa ison show ha he BARE1 in on con o ms o expec a ions o plan in on unc ionali y and sugges ha he e has been selec i e p essu e on BARE1 o splicing compa ed wi h BARE2. T ansc ip s om BARE TATA1 a e uncapped, bu hose om TATA2 a e capped Because he wo RNA splice a ian s o BARE1 possess di e en ansla ion capaci ies, we in es iga ed he ansc ip s o ea u es associa ed wi h ansla ion. The RNA des ined o ansla ion in euka yo ic cells commonly ecei es a 7- me hylguanosine cap as pa o he ma u a ion p ocess [27], al hough many plan i uses as well as HIV exploi cap- independen ansla ion ins ead [28,29]. We ea lie showed [19] ha BARE p oduces en classes o ansc ip s om wo TATA boxes (Figu e 3A), i e each om TATA1 and TATA2. In o de o in es iga e which migh be ansla ed, we i s looked a hose which ha e 5’ caps. Capped and uncapped RNAs we e dis inguished by enzyma ic ea men be o e RNA-ligase-media ed (RLM) PCR, espec i ely by py ophospha ase decapping and phospha ase 5’ dephospho yla ion. Py ophospha ase-media ed cap emo al gene a es a 5’ phospha e in i s place, which will allow liga ion o an RNA adap e and PCR. Uncapped 5’ ends can be liga ed di ec ly wi hou py ophospha ase ea men . The expe imen s we e ca ied ou wi h p ime s posi ioned (Figu e 3A) so ha RNA p oduc s o bo h TATA1 and TATA2 could be de ec ed, bu only hose om in ac BARE elemen s con aining in e nal domains and no om ead- h ough Table 1. Summa y o BARE RNA species de ec ed. TE S uc u e Occu ence Sp Cap pA R C L E BARE1 + + + - + + + BARE1 - + + - + + + BARE1 - - - + + + + BARE2 - - - - + + + BARE2 - + + - + + + Abb e ia ions: TE, e o ansposon amily; Sp, ansc ip spliced; Cap, cap(Gppp) p esen ; pA, polyadenyla ion; R, R-domain p esen ; C, callus; L, lea ; E, emb yo Each able ow co esponds o one RNA ype ha ing he ea u es ma ked as p esen (+) o absen (-) BARE Re o ansposon T ansc ip ion o Replica ion PLOS ONE | www.plosone.o g 3 Augus 2013 | Volume 8 | Issue 8 | e72270 ansc ip s o solo LTRs. Capped RNAs ansc ibed om BARE we e ound in all issues examined (Figu es 2, 4B). Con ol eac ions lacking py ophospha ase ga e no PCR p oduc (Figu e 3E). Sequenced PCR p oduc s showed ha capped ansc ip s de i e only om TATA2; he longe p oduc s om he emb yo (Figu e 3B) we e sequenced and a e non-speci ic. The capped ansc ip s s a a n 1686-1689 (Z17327), co esponding wi h he published RACE-PCR da a [19], which could no dis inguish capped om uncapped RNA. No ably, he 5’ ends o he capped RNAs a e a posi ions shown ea lie o be oo a downs eam o pe mi o ma ion o an R domain needed o eplica ion by e e se ansc ip ion [19]. In a complemen a y expe imen , he uncapped ansc ip s, which o igina e only om TATA1, we e cloned and sequenced om o al RNA o emb yo and callus (Figu e 3C). The s a si es o hese RNAs o callus a e a n 1351, 1379, and 1382 (numbe ing om Z17327). The wo majo bands om emb yo issue ep esen BARE1 and BARE2 RNA and espec i ely s a a n 1350 (numbe ing om Z17327) and 1682 (numbe ing om AJ279072). All s a si es co esponded o hose ound in he ea lie RACE-PCR da a o TATA1 in BARE1 and BARE2 [19]. Bo h Spliced and Unspliced Capped TATA2 RNAs A e Polyadenyla ed As desc ibed abo e, TATA2, bu no TATA1, ansc ip s om bo h BARE1 and BARE2 a e capped; hose om BARE1 a e spliced abou hal he ime and BARE2 ansc ip s a e no spliced. Whe eas ansla ed cellula RNAs a e gene ally bo h capped and polyadenyla ed, plan i al RNAs and mos posi i e-s and RNA i uses a e ansla ed no only wi hou caps bu also wi hou poly(A) ails [30]. To cla i y he si ua ion o BARE, polyadenyla ed RNAs isola ed om lea and callus we e subjec ed o RLM-PCR diagnos ic o he p esence o caps and hen he PCR p oduc s sequenced. Caps we e Figu e 2. Splicing o BARE1. A. Alignmen o a se o BARE genomic DNA sequences and cDNA clones showing he o ms wi h he consensus splice junc ions (SD and SA) wi hin he gag domain. Nucleo ides shaded blue ma ch he consensus slice junc ions, hose in ed do no . Genomic DNA sequences a e labelled as “G”, cDNA as “T”. Genomic sequences in he alignmen a e: G1, AJ279072; G2, Z17327; G3, AY66155; G4, AY485643; G5, BQ900685. B. Schema ic diag am o BARE e o ansposons showing he LTRs, encoded p o eins o he open eading ame, and he cDNA p iming si es (PBS, PPT), oge he wi h he posi ion o he diagnos ic p ime s as a ows below. The in e ed iangle indica es he 8 bp dele ion o he s a codon in BARE2 ha elimina es syn hesis o Gag; he ollowing ATG o he Pol domain is indica ed. C. Diag ams o he unspliced (1) and spliced (2) o ms o he BARE1 ansc ip s as well as he ansla ed p oduc o he spliced o m (3). D. Concep ual ansla ion o he BARE1 ORF (Accession Z17327) co e ing he Gag and pa o he PR egion o he unspliced (Gag) and spliced (Gag_S) ansc ip o ms. Amino acids al e ed by he splice-induced ameshi a e shown in ed, he s op codon as *. doi: 10.1371/jou nal.pone.0072270.g002 BARE Re o ansposon T ansc ip ion o Replica ion PLOS ONE | www.plosone.o g 4 Augus 2013 | Volume 8 | Issue 8 | e72270 Figu e 3. Capping o BARE ansc ip s. A. Schema ic ep esen a ion o he BARE LTR and pa o he ORF. The black box be ween gag and p ep esen s he speci ic dele ion in BARE2 (dele ion no o scale). The 5’ LTR is shown as a hick box, he egion be ween he LTR and he s a codon o gag as a hin box. The posi ion o TATA1 (T1) and TATA2 (T2) a e ma ked wi h ben a ows, hei posi ion and hose o he beginning and end o he LTR and he beginning o gag numbe ed acco ding o acc. Z17327. P ime s used o PCR and making cDNA a e indica ed by a ows below. Wa y lines indica e he capped TATA2 and uncapped TATA1 ansc ip s iden i ied by RLM-RACE PCR. B. RLM-RACE PCR analysis o he ansc ip ion s a si es o BARE om o al RNA in di e en issues (C, callus; E, emb yo; R, oo ) ollowing phospha ase and py ophospha ase ea men o selec o capped ansc ip s. (+) and (-) indica e he p esence o absence (con ol) o e e se ansc ip ase in he assay. The p oduc size (a ow, 244 bp) co esponds o ampli ica ion om he 5’ adap e p ime and E1625. C. De ec ion o uncapped BARE ansc ip s in emb yos and callus o al RNA by RLM-RACE PCR wi hou phospha ase and py ophospha ase ea men . The la ge band (a ow) co esponds o BARE1, he smalle (a ow) o BARE2. D. De ec ion o capped BARE poly(a) RNA (a ow) in callus (Cp(A)) and emb yo (Ep(A)) by RLM-RACE PCR; RNA ea ed as in (B). The uppe band in emb yo is due o a seconda y p iming si e; he ampli ica ion gene a es a p oduc o he same size o BARE1 and BARE2. 100 bp ladde s (m) a e shown. E. Con ol eac ions ea ed wi h phospha ase bu no subsequen py ophospha ase. doi: 10.1371/jou nal.pone.0072270.g003 BARE Re o ansposon T ansc ip ion o Replica ion PLOS ONE | www.plosone.o g 5 Augus 2013 | Volume 8 | Issue 8 | e72270 Figu e 4. Spliced BARE RNA is associa ed wi h poly ibosomes. A. Diag am o he 3’ LTR, indica ing o e e ence he posi ion o he wo TATA boxes TATA1 (T1) and TATA2 (T2); only hose in he 5’ LTR se e o ansc ibe he e o ansposon. The posi ions o o wa d p ime s RLM1 and RLM2 o 3’ RLM-RACE a e shown. The wa y lines show, espec i ely, he app oxima e e mina ion posi ions o he poly ibosome-associa ed poly(A) RNA. B. Supe na an (Sup) and poly ibosome pelle (Pel) ac ions om o al callus RNA (To ) ul acen i uged on a 10-45% suc ose g adien . RNA bands a e labelled. C. Elec opho esis o RLM-RACE eac ions om poly ibosome-associa ed callus and lea RNA ampli ying bo h BARE1 and BARE2. A 100 bp size ladde (m) is shown, 500 bp ma ked. D. Ampli ica ion o BARE2 and BARE1 om poly ibosome-associa ed callus RNA (p ime s 1965 and 1966 o BARE2, p ime s Gag5 and AP4 o BARE1, Figu e 2). Nega i e con ols (-) o he p esence o genomic DNA con amina ion lack e e se ansc ip ase; posi i e con ols, G, con ain genomic DNA. Ba s poin o 1000 bp and 500 bp size ma ke s, m. E. RT-PCR assay (p ime s F1594 and F1593, Figu e 2, which a e no speci ic o BARE1) om he ac ions in (B). Fo size compa ison, PCR om genomic DNA wi h he same p ime s is shown on he igh . Unspliced and spliced ansc ip s a e indica ed as U and S, espec i ely. The RT-minus con ols ga e no ampli ica ion. doi: 10.1371/jou nal.pone.0072270.g004 BARE Re o ansposon T ansc ip ion o Replica ion PLOS ONE | www.plosone.o g 6 Augus 2013 | Volume 8 | Issue 8 | e72270 p esen in bo h BARE1 and BARE2 p oduc s om he polyadenyla ed RNA ac ion; hese ansc ip s s a only a e TATA2 (Figu e 3D). In con ol expe imen s, he 5’ adap e s we e di ec ly liga ed o polyadenyla ed RNA om emb yo and callus. The eac ions yielded no p oduc , indica ing ha no BARE RNA was simul aneously polyadenyla ed and uncapped. Polyadenyla ed BARE RNA Is Poly ibosome- Associa ed Gi en he mul iple BARE1 and BARE2 RNA species, we in es iga ed which pool is ansla ed by examining RNA in poly ibosomal ansla ion complexes. As desc ibed abo e, he BARE2 ansc ip s nei he exp ess Gag no a e spliced, aising he ques ion o whe he hey a e none heless p esen among he poly ibosomal RNAs. Poly ibosomes we e isola ed; he 28S and 18S ibosomal RNAs we e e ec i ely concen a ed in o he pelle (Figu e 4B). The supe na an e ained mainly he small RNAs (mos ly RNAs), indica ing he p esence o he poly ibosomes in he pelle . Capped RNA was de ec ed by 5’ RLM-RACE using a p ime ma ching bo h BARE1 and BARE2 (Figu e 4B). BARE1 and BARE2 RNAs we e dis inguished by sequencing and ound in poly ibosomes o lea , callus and emb yos, he h ee issues in es iga ed. Sequencing showed he s a si es a n 1689 (accession Z17327) o BARE1 and 25n downs eam o TATA2 o BARE2 (AJ279072) as be o e [19]. Fo compa ison, cloning and sequencing o he uncapped RNA p oduc s in o al RNA e ealed ha TATA1 ansc ip s s a a posi ion n 1351 in emb yos and 1412 in callus as desc ibed ea lie [19]. To look a he ela i e abundance o BARE1 and BARE2 ansc ip s in he poly ibosomes, speci ic p ime pai s we e used. The BARE2 poly ibosomal ansc ip s a e mo e abundan han hose o BARE1 (Figu e 4D), co esponding o ea lie esul s o he same ba ley cul i a (Bomi): BARE1 and TATA2 ansc ip s we e shown o ep esen espec i ely 32% and 6– 25% o he o al BARE pool [16,19]. Hence, he lack o a ansla able gag domain does no appea o in e e e wi h he BARE2 ansc ip s ei he being capped o o ming poly ibosomes and sugges s ha BARE2 pol is ansla ed, e en i gag canno be, by ei he ibosome scanning o in e nal en y. The p esence o he spliced BARE1 ansc ip ha codes only o Gag aises he expec a ion ha i he spliced o m con ibu es o p oduc ion o Gag i , oo, should be polyadenyla ed and associa ed wi h poly ibosomes. The o al RNA om callus was used as a empla e o RT-PCR using p ime s nea he splice junc ion; bo h spliced and unspliced o ms we e p esen (Figu e 3E). The pelle con aining he poly ibosomes also con ained he spliced o m, a a p opo ion o he o al BARE pool simila o ha expec ed. The da a show ha he spliced o m is no di e en ially excluded om poly ibosomes, al hough some o he unspliced o m emains in he supe na an . In o de o examine poly ibosomal BARE RNA o polyadenyla ion, o al RNA was i s isola ed om he poly ibosomes and wo ounds o 3’ RLM-RACE ca ied ou , i s o selec o he 3’ LTR segmen o he LTR (Figu e 4A) and hen o a poly(A) ail. The PCR p oduc s we e sequenced; he polyadenyla ed, poly ibosomal BARE ansc ip s sha ed hei 3’ e mini wi h hose in he polyadenyla ed BARE popula ion in o al RNA [19]. Non-Polyadenyla ed BARE RNA Is Packaged in o VLPs The TATA1 ansc ip s, as shown abo e, a e no capped, spliced, o polyadenyla ed. Because only hey con ain he R domain, i BARE is being eplica ed hen TATA1 ansc ip s should be packaged in o VLPs. To in es iga e his, we isola ed VLPs and examined he ends o RNAs associa ed wi h hem. The RNA was isola ed om he pooled and pu i ied VLP ac ions 9-11 desc ibed p e iously [18], and used as he empla e o 3’ RLM-RACE (Figu e 5). The sequenced p oduc s e eal ha only non-polyadenyla ed RNA is packaged in VLPs. Sequences o he packaged RNAs ha e he p e iously desc ibed end poin s ha a e expec ed o ansc ip s ini ia ed by TATA1 and no TATA2 [19]. Fu he mo e, he RNA sequences include TATA1 ansc ip s o BARE2, which is no able o p oduce he Gag o he VLPs in o which i s ansc ip s a e packaged. This clea ly demons a es he pa asi ism o BARE2 on BARE1. Discussion Re o ansposon and e o i us ansc ip s se e wo dis inc oles: as empla es o he p o eins needed o hei own eplica ion; as genomic RNA, which is i s packaged in o capsids comp ised o Gag, i s own ansla ion p oduc , and hen la e des oyed du ing i s e e se ansc ip ion in o cDNA. We ea lie showed ha e o ansposons BARE1 and i s pa asi ic ela i e BARE2, which canno syn hesize i s own capsid p o ein, p oduce wo se s o ansc ip s; one om each o he wo TATA boxes in he LTR [19,31]. We also showed ha only a mino i y o he BARE ansc ip s, 15 o 25%, we e polyadenyla ed, al hough nei he was ansc ip p ocessing u he examined no he eason o he incomple e polyadenyla ion ound. He e, we ha e unco e ed a eplica ion sys em whe eby BARE1 and BARE2 encode dis inc classes o RNAs o se e he wo dispa a e unc ions, one o ansla ion and he o he as he genomic RNA des ined o be e e se- ansc ibed in o cDNA (Figu e 6). The esul s a e eminiscen o he dis inc pools o ansla ion and e e se ansla ion pu po edly o med by he MLV e o i us [13], a he han he single pool o HIV [14]. The i s RNA pool, ansc ip s om TATA2 o bo h BARE1 and BARE2, is capped, polyadenyla ed, and poly ibosome- associa ed. These ea u es indica e ha TATA2 RNA se es o ansla ion o he p o ein p oduc s o BARE. Ea lie , we had shown ha epo e gene exp ession d i en by he BARE LTR is dependen on he p esence o TATA2 and no TATA1 [19,31], s ongly sugges ing ha all ansla ed BARE p o eins a e de i ed om TATA2. Capping and polyadenyla ion ha e no been well in es iga ed o LTR e o ansposons; Ty1 and copia o Supe amily Copia a e ansla ed om capped RNA [32,33] while Ide ix o supe amily Gypsy exploi s bo h cap- dependen and independen mechanisms [34]. Re o i uses, which a e likely de i ed om Gypsy e o ansposons [3], p oduce capped ansc ip s bu appea o exploi bo h cap- dependen and ‑independen ansla ion [35]. He e, he BARE Re o ansposon T ansc ip ion o Replica ion PLOS ONE | www.plosone.o g 7 Augus 2013 | Volume 8 | Issue 8 | e72270 p esence o na i e polyadenyla ed, capped TATA2 ansc ip s in he poly ibosome “ ansla ome” is he clea es indica ion o ac i e BARE1 ansla ion in he issues examined [36,37]. In e es ingly, he TATA2 ansc ip s o BARE2 a e also capped, polyadenyla ed and poly ibosome-associa ed, e en i he conse ed BARE2 dele ion abolishes he s a codon o he ORF and he eby ansla ion o Gag [16]. Howe e , he Figu e 5. BARE ansc ip s in VLPs. Elec opho esis o 3’ RLM-RACE PCR eac ions, pe o med on pu i ied VLP ac ions 9-11 [18]. The ampli ied p oduc s (a ows, 220 bp and 180 bp) ep esen wo o he ansc ip g oups seen ea lie in o al RNA, dis inc om hose in poly(A) [19]. Fo wa d nes ed p ime s a e RLM1 and RLM2; p ime E2146, which ma ches he liga ed linke sequence, was used as he e e se p ime . 100 bp ladde (m) is shown. doi: 10.1371/jou nal.pone.0072270.g005 subsequen AUG s a codon a he end o gag could make he es o BARE2 ansla able. T ansla ion o e o ansposons and e o i uses aises a majo challenge: balancing he s oichiome y o se e al gene p oduc s exp essed by a single p omo e . The s uc u al Gag is needed in g ea e abundance han he enzymes o pol; al e a ion o he a io in e e es wi h e o i al in ec i i y [38] and e o ansposon mobili y [39]. Mos e o i uses use -1 ameshi ing o yield wo eading ames, Gag and Gag-Pol [35] and mos Supe amily Gypsy elemen s use +1 ameshi ing [9,40]. In con as , sequence analysis sugges s ha he o e whelming majo i y o Copia e o ansposons such as BARE encode Gag and Pol as one ORF [9], al hough Ty1 uses a +1 ameshi [41]. The p oblem o p oducing enough Gag is especially acu e o BARE o se e al easons: BARE2 lacks i s own Gag [16], ye BARE2 comp ises abou 68% o he BARE ansc ip s [16]; he TATA2 ansc ip s, which a e he only ones ound in he poly ibosomes, comp ise on a e age 15% o he ansc ip s. Despi e he p esence o he la ge pool o uncapped TATA1 BARE ansc ip s, he BARE e o ansposons appea no o exploi cap-independen ansla ion on hese ansc ip s as does HIV [29] and many plan i uses [28]. Hence, BARE ansc ip s able o exp ess Gag amoun o only 4.8% o he o al. Fu he mo e, sequence analysis o BARE likewise had shown a single ORF o Gag and Pol [17]. Howe e , he e we show ha , o he BARE1 TATA2 p oduc s, abou hal a e spliced so as o exp ess only Gag, e en i he spliced o m ep esen s only 2.4% o he o al BARE ansc ip pool. Immunoblo ing p oduces much s onge signals o Gag han o IN om ba ley p o ein ex ac s, consis en wi h he ac ions o a mechanism o inc ease he ela i e p opo ion o Gag [18]. These ac s oge he sugges ha splicing in BARE1 may se e o inc ease he con en o Gag compa ed o Pol. Fo copia, he spliced sub-genomic RNA also is p esen in abou equal amoun s wi h he ull-leng h ansc ip [10] ye he Gag p oduc is mo e abundan han Pol. The copia Gag RNA is ansla ed abou en- old mo e e icien ly han he genomic RNA [10]. The enhancemen in copia may be due o emo al o a 2.9 kb o Pol sequence; in BARE, only 104 n is spliced ou o c ea e a s op codon. The ole o his domain and he ansla ional e iciency o he spliced BARE1 RNA a e cu en ly unde in es iga ion. Splicing o e o ansposon ansc ip s in plan s is known in a ew o he cases. The en sub-genomic RNA in he e o i al- like clade o supe amily Gypsy, which includes Bagy2, esul s om splicing [42]. Two o he Gypsy elemen s, Og e and CRR, also splice ansc ip s; Og e splices be ween ORFs [20] whe eas CRR splices o emo e and c ea e wo ORFs. Fo e o i uses such as MMTV, exp ession o Gag ac ually equi es supp ession o splicing [43]. The BARE splicing epo ed he e is he only case demons a ed in supe amily Copia aside om ha o copia i sel and he only one whe eby a small, al e na i e in on wi hin pa o he gag ORF esul s in a nea ly ull-leng h sub-genomic RNA encoding jus a single p o ein. I emains o be seen i he s a egy employed by copia and BARE is gene al among ela ed e o ansposons. BARE Re o ansposon T ansc ip ion o Replica ion PLOS ONE | www.plosone.o g 8 Augus 2013 | Volume 8 | Issue 8 | e72270 In con as o ansla ion, o eplica i e compe ence e o ansposons mus a oid he packaging, e e se ansc ip ion, and in eg a ion o spliced RNA. Se e al lines o e idence suppo he iew ha BARE eplica es only TATA1 ansc ip s, which a e no spliced. Fi s , only TATA1 ansc ip s con ain he R domain, which is necessa y o s and swi ching du ing e e se ansc ip ion [19]. Second, DNA copies o he spliced RNA, p oduced by TATA2, canno be ampli ied om ba ley genomic DNA. Thi d, analyses o he BARE LTR demons a ed ha TATA2 was su icien o gi e ull epo e exp ession bu TATA1 alone ga e none [19,31]. The s onges a gumen ha he uncapped, un-polyadenyla ed TATA1 ansc ip s se e as he sole empla es o BARE1 cDNA syn hesis, in addi ion o he exclusi e occu ence o he R domain in hem, is hei p esence (and he con as ing absence o TATA2 ansc ip s) in VLPs. I is wi hin he VLPs ha Figu e 6. Schema ic model o BARE RNA exp ession, ansla ion, and eplica ion. A. BARE e o ansposon, d awn o scale, showing: 5’ LTR ( u quoise), including he posi ions o TATA1 (T1) and TATA2 (T2); un ansla ed leade (g ay box); gag (yellow), encoding he capsid p o ein Gag; dele ion in BARE2 (black in e ed iangle), which abla es gag s a codon; pol (g een), encoding aspa ic p o einase (PR), in eg ase (IN), and he e e se ansc ip ase – RNase H complex (RT-RH); he al e na i ely spliced in on (dashed box), which gene a es a ameshi ha knocks ou pol exp ession in BARE1; 3’ un ansla ed egion (g ay box); 3’ LTR ( u quoise), including e mina ion si e o ansc ip s om TATA2 (S2) and om TATA1 (S1). B. T ansc ip s om BARE1, including he al e na i ely spliced capped (Gppp) and polyadenyla ed (aaaa) RNA om TATA2 and he uncapped non-polyadenyla ed RNA om TATA1, he la e which ha e he e minal epea s (R, u quoise boxes) needed o eplica ion in o cDNA. Unexp essed ORFs a e shown as ha ched boxes labelled in g ay C. T ansc ip s om BARE2, including he TATA1 p oduc s and he capped and polyadenyla ed TATA2 p oduc s, which canno exp ess gag. D. Mapping o he o ma ion o he ansla ion p oduc s om he a ious RNAs, including: Gag om BARE1; he polyp o ein om BARE1, which is clea ed by PR in o unc ional uni s (GAG, yellow; PR, iole ; RT-RH ( ed and b own). A schema ic ep esen a ion o he assembly o he componen s in o he i us-like pa icle (VLP) is shown, in o which he TATA1 ansc ip s oge he wi h RT-RH and IN a e packaged. doi: 10.1371/jou nal.pone.0072270.g006 BARE Re o ansposon T ansc ip ion o Replica ion PLOS ONE | www.plosone.o g 9 Augus 2013 | Volume 8 | Issue 8 | e72270