scieee Open visual document viewer

Molecular basis of FIR-mediated c-myc transcriptional control

Cukier, Cyprian D.; Hollingworth, David; Martin, Stephen; Kelly, Geoff; Díaz Moreno, Irene; Ramos, Andrés

Abstract

The far upstream element (FUSE) regulatory system promotes a peak in the concentration of c-Myc during cell cycle. First, the FBP transcriptional activator binds to the FUSE DNA element upstream of the c-myc promoter. Then, FBP recruits its specific repressor (FIR), which acts as an on/off transcriptional switch. Here we describe the molecular basis of FIR recruitment, showing that the tandem RNA recognition motifs of FIR provide a platform for independent FUSE DNA and FBP protein binding and explaining the structural basis of the reversibility of the FBP-FIR interaction. We also show that the physical coupling between FBP and FIR is modulated by a flexible linker positioned sequentially to the recruiting element. Our data explain how the FUSE system precisely regulates c-myc transcription and suggest that a small change in FBP-FIR affinity leads to a substantial effect on c-Myc concentration.

Full text

Molecula basis o FIR-media ed c-myc ansc ip ional con ol Cyp ian D. Cukie 1, Da id Hollingwo h1, S ephen R. Ma in2, Geo Kelly3, I ene Díaz- Mo eno1,4, and And es Ramos1 1Molecula S uc u e Di ision, MRC Na ional Ins i u e o Medical Resea ch, The Ridgeway, Mill Hill, London NW7 1AA, UK 2Physical Biochemis y Di ision, MRC Na ional Ins i u e o Medical Resea ch, The Ridgeway, Mill Hill, London NW7 1AA, UK 3MRC Biomedical NMR Cen e, The Ridgeway, Mill Hill, London NW7 1AA, UK 4Ins i u o de Bioquimica Vege al y Fo osin esis, US-CSIC, A da. Ame igo Vespucio 49, 41092, Se illa, Spain Abs ac The Fa UpS eam Elemen (FUSE) egula o y sys em p omo es a peak in he concen a ion o c- Myc du ing cell cycle. Fi s , he FBP ansc ip ional ac i a o binds o he FUSE DNA elemen ups eam o he c-myc p omo e . Then, FBP ec ui s i s speci ic ep esso (FIR) which ac s as an on/ o ansc ip ional swi ch. He e we desc ibe he molecula basis o FIR ec ui men showing ha he andem RNA ecogni ions mo i s o FIR p o ide a pla o m o independen FUSE DNA and FBP p o ein binding and explaining he s uc u al basis o he e e sibili y o he FBP-FIR in e ac ion. We also show ha he physical coupling be ween FBP and FIR is modula ed by a lexible linke posi ioned sequen ially o he ec ui ing elemen . Ou da a explain how he FUSE sys em egula es p ecisely c-myc ansc ip ion and sugges ha a small change in FBP–FIR a ini y leads o a subs an ial e ec on c-Myc concen a ion. The c-myc p o o-oncogene egula es he p oli e a ion, g ow h and di e en ia ion o soma ic cells, in eg a ing a numbe o in acellula and ex acellula p og ams o gene con ol. c-Myc a ge s he PolII-dependen p omo e o 10–15% o genes, including ansc ip ion ac o s, mRNA me abolism p o eins, DNA epai ac o s, elome ases and cy okines 1,2. Fu he , i modula es he ansc ip ion o a numbe o non-coding RNAs 3. c-myc mis- egula ion has been co ela ed wi h a b oad ange o cance pa hologies 4. The Fa Ups eam Sequence Elemen (FUSE) ups eam o he c-myc p omo e media es a as ansc ip ion- esponsi e mechanism esponsible o an upsu ge o c-Myc le els du ing he cell cycle. FUSE-based c-Myc up egula ion is media ed by he unwinding and opening o an AT- Use s may iew, p in , copy, download and ex and da a- mine he con en in such documen s, o he pu poses o academic esea ch, subjec always o he ull Condi ions o use: h p://www.na u e.com/au ho s/edi o ial_policies/license.h ml# e ms Co espondence should be add essed o A.R. ([email p o ec ed]). Au ho con ibu ions Cloning was pe o med by D.H. and C.C. Exp ession and pu i ica ion o he FBP and FIR cons uc s we e pe o med by C.C. and D.H. NMR spec a we e eco ded by C.C. and G.K. and analyzed by C.C. The s uc u es in his pape we e calcula ed by C.C. All NMR i a ions we e pe o med by C.C. SIA analysis was pe o med by I.D.M. CD da a we e eco ded by A.R. and analyzed by S.M. BLI da a we e eco ded and analyzed by C.C. and S.M. The pape was w i en by C.C. and A.R. All au ho s we e in ol ed in planning he expe imen s. Accession Codes: P o ein Da a Bank: Coo dina es o he s uc u e o FIR RRM1-RRM2 alone and in complex wi h FBP Nbox ha e been deposi ed wi h accession codes 2kx and 2kxh espec i ely. Supplemen a y In o ma ion accompanies he pape on www.na u e.com/na u e. UKPMC Funde s G oup Au ho Manusc ip Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. Published in inal edi ed o m as: Na S uc Mol Biol. 2010 Sep embe ; 17(9): 1058–1064. doi:10.1038/nsmb.1883. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip ich s e ch ( he FUSE sequence) loca ed abou 1.7 kb ups eam o he c-myc p omo e as a consequence o he nega i e supe coiling ha accumula es du ing p e-exis ing lowe -le el ansc ip ional ac i i y 5,6 (Fig. 1a). The FUSE DNA non-coding s and (hence o h e e ed o as ssFUSE o ssFUSE DNA) ec ui s he FUSE-Binding P o ein (FBP) ha , in u n, in e ac s wi h T ansc ip ional Fac o IIH (TFIIH) – a complex mul i-componen machine in ol ed in ansc ip ional ini ia ion and p omo e escape – and inc eases he a e o p oduc i e c-myc ansc ip ion 7,8. This inc ease in ansc ip ion leads o u he nega i e supe coiling and o u he opening o he FUSE. The FBP–FUSE complex ec ui s he FBP-In e ac ing Rep esso (FIR) p o ein ha in e ac s wi h bo h FBP and ssFUSE DNA. Upon binding o FBP and ssFUSE, FIR in e ac s wi h TFIIH 7 educing FBP-media ed ansc ip ion and he e o e nega i e supe coiling. FIR-media ed ansc ip ional ep ession leads o e-winding o he FUSE, ejec ion o FBP and, e en ually, ejec ion o FIR i sel , b inging he sys em back o a basal ansc ip ion s a e (Fig. 1a). The FUSE egula o y sys em can be ep esen ed by a simple h ee-componen mechanism o con ol c-myc ansc ip ion and is a p omising ool in colo ec al cance he apy 9,10. Cen al o FUSE egula ion is he FIR–FBP in e ac ion ha media es FIR ec ui men and ansc ip ional shu -o . I has been shown ha ssFUSE DNA binds o he andem RNA Recogni ion Mo i s (RRMs) o FIR (RRM1-RRM2) and o he ou K-Homology (KH) domains o FBP (Fig. 1b) 11-13 bu also ha a di ec FBP–FIR in e ac ion akes place ia a 26-aa elemen in he amino- e minus o FBP (called he Nbox, Fig. 1b) and he FIR RRM1-RRM2 andem domains 11. The impo ance o his in e ac ion has been es ablished by Chung and co-wo ke s 11 who analyzed he ac i i y o FIR in he FBP–DNA and FBP3–DNA ansc ip ional ac i a ion sys ems. FBP3 (Fig. 1b), a p o ein o he FBP amily, has been shown o be exp essed a di e en imes han FBP du ing he cell cycle, and no o be a ec ed by FIR-media ed ep ession 11,14. Chung and co-wo ke s ha e shown ha FBP Nbox media es an in e ac ion be ween FBP and FIR while FBP3 Nbox lacks his capabili y. They ha e also shown ha when he low e iciency o FIR ec ui men o he DNA–FBP3 sys em is by-passed by using FIR o a high a ini y DNA binding domain, FIR can ep ess FBP3. This indica es ha Nbox-media ed ec ui men o FIR o he DNA is a c ucial s ep o con ol o FIR ep essional ac i i y 11. Despi e he a ailabili y o a b oad ange o unc ional in o ma ion, ou molecula unde s anding o FUSE media ed egula ion – and he e o e ou capabili y o in e e e wi h i – is se e ely limi ed by he lack o s uc u al da a on he FBP–FIR in e ac ion and by he pauci y o quan i a i e da a on he p o ein–p o ein and p o ein–DNA in e ac ions ha build FUSE egula ion. He e we ocus on he molecula mechanism o FIR ec ui men and on he FIR– FBP in e ac ion ha de ines he leng h and in ensi y o c-myc up egula ion du ing cell cycle. We use human (Homo sapiens) FIR and FBP p o eins and FUSE DNA o explain why FBP is ec ui ed be o e FIR in he FUSE unc ional cycle and how he necessa y combina ion o speci ici y and e e sibili y is achie ed by he Nbox–FIR in e ac ion. Fu he , we quan i y di ec ly he e ec o he Nbox in a model FBP–FIR–FUSE sys em and he decoupling e ec o an un olded linke loca ed be ween he Nbox and he DNA-binding egion o FBP. We conclude ha he swi ch o he FUSE sys em om an ac i a ing o a ep essing s a e is based on a ela i ely modes inc ease in he a ini y o FIR o he FBP–FUSE DNA complex. Resul s S uc u e, dynamics and DNA binding p ope ies o FIR RRM1-RRM2 To unde s and he molecula basis o FIR ec ui men we need o cha ac e ize he s uc u e o he FIR RRM1-RRM2 andem domains and hei in e ac ions wi h FBP and wi h FUSE DNA. We ha e sol ed he solu ion s uc u e o FIR RRM1-RRM2 alone and in complex wi h he FBP Nbox and showed ha in FIR RRM1-RRM2 he helical ace o RRM1 packs on o he β- Cukie e al. Page 2 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip shee ace o RRM2, c ea ing a s able in e ace (1,100 Å2 o bu ied su ace) (Figu e 2a, Supplemen a y Fig. 1 and Table 1). This o ien a ion is simila o wha is obse ed in he s uc u e o he DNA-bound FIR RRM1-RRM2 15 (Supplemen a y Fig. 1b). NMR elaxa ion da a con i m ha binding o FIR o he FBP and FUSE binding pa ne s does no lead o con o ma ional ea angemen s in he p o ein (Supplemen a y Fig. 2). The a angemen o RRM1 and RRM2 desc ibed abo e indica es ha only he β-shee nucleic acid binding su ace o RRM1 is a ailable o DNA binding. Ou NMR Chemical Shi Pe u ba ion (CSP) da a con i m ha DNA binding in ol es canonical nucleic acid binding RiboNucleoP o ein (RNP2 and RNP1) mo i s plus esidues in he α2/β4, β1/α1 and β2/β3 loops o RRM1, including he esidues making con ac wi h he single nucleo ide isible in he published s uc u e o he DNA-bound FIR RRM1-RRM2 15 (Fig. 2b). Based on he known RRM–nucleic acid s uc u es, his su ace is likely o accommoda e a 4–5 nucleo ides long DNA sequence 16 and he e o e we ha e explo ed he nucleobase p e e ence o FIR o ou successi e nucleo ides posi ions using Sca old Independen Analysis (SIA) 17. We show ha FIR RRM1-RRM2 has a mode a e sequence p e e ence o T o T/G in all o he ou posi ions analyzed by SIA (Fig. 2c, Supplemen a y Table 1). We use ollow-up NMR binding assays wi h DNA 5-me s o di e en base composi ions o con i m his sequence p e e ence and o de ine a ange o a ini ies o FIR in e ac ion wi h op imal and non-op imal DNA sequences (Fig. 2c). Knowing FIR's sequence p e e ence and he ange o a ini ies o di e en DNA sequences allows a mo e accu a e unde s anding o FIR–FUSE in e ac ions. FIR–FUSE and FBP–FUSE in e ac ions In o de o de ine accu a e models o FBP and FIR ec ui men and ejec ion i is necessa y o quan i y he s eng h o he FIR–FUSE in e ac ion and o compa e i wi h he FBP–FUSE in e ac ion. A p elimina y model o FBP ec ui men pu o wa d by Liu and co-wo ke s p edic s ha ssFUSE has a highe a ini y o FIR han o FBP and p oposes ha he ini ial ec ui men o FBP – a he han FIR – o he ssFUSE is due o he highe concen a ion o FBP in he cell 8 while a ecen pape published du ing e ision o ou wo k indica es ha he a ini y o FIR o he FUSE is he same o lowe by a ac o o 5 han he one o FBP, depending on he FUSE DNA used 18. Conside ing ha FIR and FBP can bind o o e lapping sequences in he FUSE 12 a simila a ini y o he wo p o eins o he DNA would indica es ha he choice o ec ui ing FBP a he han FIR is dependen on he a ia ions o hei concen a ion in he cell. To cla i y i he ea lie binding o FBP o he FUSE is dependen on he concen a ion o FBP and FIR, we ha e eco ded BioLaye In e e ome y (BLI) expe imen s using an immobilized 40-me ssFUSE (ssFUSE40, Fig. 1a) and inc easing concen a ions o FIR (Fig. 2d). This DNA has been designed o ecapi ula e he a eas o in e ac ions o FIR and FBP as de ined by Benjamin e al. 12 . We show ha FIR binds o he FUSE wi h a Kd alue o ~7 μM, as de e mined om he a io o he dissocia ion and associa ion a e cons an s, and independen ly om analysis o he dependence o he ampli ude o he BLI signal on FIR concen a ion. Ins ead, BLI expe imen s epo ha he complexes be ween ssFUSE40 DNA and wo FBP KH domains cons uc s wi h and wi hou he Nbox (Fig. 1b) ha e dissocia ion cons an s in he ~1.5–2 nM ange (Fig. 2d, Supplemen a y Fig. 3, Supplemen a y Table 2). So, FBP–ssFUSE binding is h ee o de s o magni ude s onge han FIR–ssFUSE binding and he e o e he ini ial ec ui men o FBP in he cell is d i en by i s highe a ini y o he ssFUSE. The associa ion o FBP wi h he DNA is long las ing (ko ~4 × 10−4 s−1, Supplemen a y Fig. 3), consis en wi h he p olonged TFIIH-media ed enhancemen o p oduc i e c-myc ansc ip ion obse ed upon elimina ing FIR down egula ion o TFIIH 8. FIR dime iza ion on he FUSE has been p oposed o lead o FBP ejec ion 15 o o guide ea angemen o he FUSE–FBP–FIR complex 18. FIR was p oposed o o m a dime while bound o ssDNA based on he X- ay s uc u e o ha complex (whe e he p o ein c ys allizes Cukie e al. Page 3 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip as a dime in he asymme ic uni , Fig. 3a) and on he 2:1 s oichiome y ob ained om Size Exclusion Ch oma og aphy coupled wi h Ligh Sca e ing, Re ac i e Index and Ul aViole abso bance (SEC-LS/RI/UV) da a. We used NMR o in es iga e his possibili y. Ou assays show ha wo molecules o FIR bind o a single ssFUSE-de i ed 29-me used in unc ional s udies 11 (ssFUSE29, Fig. 1a) and ha he a ini y o he wo equi alen o quasi-equi alen p o ein binding si es is in he low mic omola ange (Fig. 3b), consis en wi h he BLI da a and wi h he as exchange egime o he shi ing esonances. This is consis en wi h he 2:1 s oichiome y ha has been epo ed o a simila FIR–FUSE DNA complex 15. Howe e , ou da a also show ha FIR does no dime ize upon binding he FUSE (ssFUSE29). We analyzed he chemical shi s o all o he amide p o ons wi hin 10 Å o he p o ein–p o ein in e - molecula in e ace, as de i ed om he X- ay s uc u e (Fig. 3a). In RRM2, which is impo an o dime iza ion and is a away om he ac ual DNA-binding su ace, none o he well dispe sed p o ons equencies o he monome ic ee RRM1-RRM2 is a ec ed upon ssFUSE29 binding, sugges ing ha he dime is no o med. In RRM1 he p oposed dime in e ace is in close p oximi y o he DNA binding su ace and some chemical shi changes om DNA binding a e likely o a ec esidues in he pu a i e dime in e ace. To decouple dime iza ion om DNA binding, we compa ed he CSP eco ded upon RRM1–ssFUSE29 binding o he ones eco ded upon RRM1 in e ac ion wi h he DNA pen ame s desc ibed in he p e ious pa ag aph – ha bind wi h a 1:1 FIR:DNA s oichiome y. Ve y simila chemical shi changes a e ob ained o he wo DNA molecules indica ing ha RRM1 is no in ol ed in dime iza ion ei he (Fig. 3c, Supplemen a y Resul s). Fu he , NMR elaxa ion da a show ha he o a ional co ela ion ime o he DNA-bound p o ein is lowe han he one expec ed i he wo p o eins would make con ac o ming a compac 50 kDa complex (Supplemen a y Fig. 2a, Supplemen a y Resul s). The binding o mo e han one FIR molecule on he FUSE is consis en wi h he elaxed T/TG sequence p e e ence o FIR, as desc ibed abo e bu he chemical shi changes we obse e upon FIR–DNA in e ac ions show unambiguously ha dime iza ion does no ake place. I is unclea whe he he binding o mo e han one FIR on he ee FUSE DNA ha we obse e has any physiological ele ance, as in he cell FBP binds be o e FIR educing he ssDNA a ailable o FIR binding and one copy only o FIR can be ec ui ed by he FBP Nbox. FIR–FBP in e ac ion We hen examined he in e ac ion be ween FIR RRM1-RRM2 and he FBP Nbox ha is essen ial o FIR ec ui men o FUSE DNA 11. Ou NMR da a e eal ha he FBP Nbox assumes a helical con o ma ion and docks on a hyd ophobic su ace de ined by α1, α2 and he β1/α1 and α2/β4 loops (Fig. 4 and Supplemen a y Fig. 4 and 5) o he RRM2 domain o FIR (Kd ~15 μM, Fig. 4a and Supplemen a y Table 2). The binding o a helical elemen in an o ien a ion quasi-pa allel o RRM1 α1 has ne e been obse ed be o e in p o ein–RRM in e ac ions and he FBP–FIR s uc u e p esen ed he e de ines a no el p o ein in e ac ion su ace on he mul i- unc ional RRM pla o m (Fig. 5) 16. The pep ide–p o ein con ac s a e mos ly hyd ophobic (Fig. 4d and Supplemen a y Fig. 4b) and an o e iew o he in e ace shows ha wo alanine esidues om he Nbox (Ala34 and Ala38) occupy he cen al pa o he pep ide ecogni ion su ace and a e lanked by bulkie hyd ophobic esidues (e.g. Phe31, Leu35, Ile41) (Fig. 4c and Supplemen a y Fig. 4b). This a angemen esul s in a e y limi ed in e -digi a ion be ween side chains o FBP and FIR esidues and in low equency mo ions a he FBP–FIR in e ace (Supplemen a y Fig. 2 and Supplemen a y Resul s). The posi ion o hese wo alanine esidues explains he esul s o a ecen mu a ional s udy 11 in which a e y s ong binding impai men was obse ed o A34V and A38V mu an s. The mu a ions place a la ge hyd ophobic chain in he cen e o he in e -molecula in e ace, esul ing in s e ic hind ance (Supplemen a y Fig. 4b). In he same s udy, Chung and co-wo ke s showed ha mu a ion o Leu35 o Phe31 o alanine has a subs an ial bu smalle e ec , which can be explained by he loss o he speci ic in e ac ions media ed by hei wo pa ially packed side chains (Fig. 4c and Supplemen a y Fig. 4b). By con as , mu a ion o Ala42 (A42V) – ha we Cukie e al. Page 4 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip show o be e y loosely packed agains he RRM2 su ace (Supplemen a y Fig. 4b) – does no lead o subs an ial loss o binding. Ou s uc u e e eals ha he side chains a angemen on he FBP–FIR in e ac ion su ace is designed o p o ide a low a ini y bu speci ic ecogni ion. The d ama ic loss o binding ob ained by inse ing a bulky hyd ophobic side chain in he sol en -excluded pa o he in e ace indica es ha he helix, al hough in a dynamic in e ac ion, canno e-o ien eely on he RRM2 su ace. I seems likely ha he global speci ici y is p o ided by he sum o he hyd ophobic con ac s made by he di e en side chains o he hyd ophobic ace o he Nbox helix. Indeed when Chung and co-wo ke s mu a e simul aneously bo h Leu35 and Phe31 o alanine he loss o binding is s onge ha he one obse ed o he single mu an s. Relaxa ion da a on he bound RRM1-RRM2 and he b oadening o se e al pep ide esonances a he p o ein–pep ide in e ace (Supplemen a y Me hods, Supplemen a y Fig. 2) a e consis en wi h he s uc u al in o ma ion on he complex and wi h he ansien na u e o he FIR–FBP con ac s in he FUSE-media ed c-myc egula ion – in which he ejec ion o FBP p ecedes FIR ejec ion and is dependen on he dis up ion o he FBP–FIR in e ac ion (Fig. 1a). The FBP–FIR–FUSE sys em Wha is he ela ionship be ween p o ein binding and DNA binding on he FIR RRM1-RRM2? DNA and p o ein in e ac ion su aces a e on di e en domains and on opposi e sides o he RRM1-RRM2 p o ein (Fig. 6a). The chemical shi changes induced by he wo binding pa ne s a e addi i e (Fig. 6b), indica ing ha he wo binding e en s a e independen . Indeed, he a ini y o RRM1-RRM2 o he sho TGTGT DNA oligo does no change when he FBP Nbox pep ide is also bound and ice e sa (Supplemen a y Fig. 6, Supplemen a y Table 2). This con i ms ha no allos e ic coope a i i y is p esen be ween Nbox and ssFUSE DNA binding. The e o e he epo ed e ec o he Nbox on FIR ec ui men is due o he physical e he ing ha exis s be ween FBP and FIR in he la ge h ee-componen FBP–FIR–ssFUSE sys em. The lack o an allos e ic e ec in RRM1-RRM2 is consis en wi h he lack o con o ma ional changes upon binding he FBP Nbox pep ide and he ssFUSE DNA (Supplemen a y Fig. 1b). The e ec o physical e he ing by FBP on FIR DNA binding ac i i y was di ec ly measu ed using BLI and a la ge FBP–FIR–DNA h ee-componen sys em (Fig. 7a). 150 nM FIR RRM1- RRM2 does no show de ec able binding o he ee DNA o o he DNA bound by he ou KH domain FBP cons uc , while he addi ion o he Nbox ec ui ing elemen o he FBP cons uc s imula es he in e ac ion (Fig. 7a). The Kd o FIR in he h ee-molecule complex, as es ima ed om he associa ion and dissocia ion a e cons an s, is lowe by a ac o o 10–50 han he Kd o FIR o he ee DNA (7 μM). This inc ease in a ini y is modes , much lowe han he one expec ed i he Nbox was s uc u ally coupled o he i s o he DNA binding modules (KH1) o FBP. The pa ial decoupling be ween DNA binding by FBP and FIR is explained by he p esence o a ~50-amino acid uns uc u ed linke ha sepa a es he Nbox om KH1 (Fig. 1b). The decoupling e ec es ima ed o he FUSE sys em (by a ac o o ~5,000) is consis en wi h he one calcula ed based on a model commonly used o wo in e ac ing egions sepa a ed by a lexible linke 19,20. Such a decoupling e ec has been o en epo ed o RNA ecogni ion by mul i-domain p o eins 21. In o de o alida e he model abo e we p o ide di ec expe imen al e idence ha FIR RRM1- RRM2 and FBP Nbox do no in e ac wi h he KH egion o FBP and ha he Nbox-FIR RRM1- RRM2 uni is s uc u ally decoupled om he DNA binding egion o FBP. We ha e i a ed an unlabeled sample o he FBP cons uc comp ising bo h he Nbox and he KH domains in o a 15N labeled sample o FIR RRM1-RRM2 and eco ded inge p in NMR 15N-1H co ela ion NMR spec a. We ha e compa ed hese spec a wi h he ones eco ded du ing an equi alen i a ion pe o med using an unlabeled sample o he Nbox-only cons uc . The same pa e n Cukie e al. Page 5 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip o chemical shi changes is obse ed in bo h i a ions (Supplemen a y Fig. 7a). This indica es ha he Nbox–FIR in e ac ion does no change whe he he KH domains a e p esen o no in he FBP cons uc and ha nei he he KH domains no he linke in e ac wi h RRM1-RRM2. Fu he , he mode a e inc ease in he line-wid h o he RRM1-RRM2 esonances (Supplemen a y Fig. 7b) p o ides independen e idence ha he lexible linke be ween he Nbox and he KH domains s uc u ally decouples FIR RRM1-RRM2 om he FBP KH domain egion – i ha was no he case a e y se e e b oadening would be obse ed upon binding o he ~45 kDa FBP cons uc . Nex , we ha e added FUSE DNA o he FIR RRM1-RRM2–FBP Nbox KH1-KH4 complex desc ibed abo e and eco ded again a inge p in NMR 15N-1H co ela ion NMR spec um. A d ama ic b oadening is obse ed o FIR RRM1-RRM2 esonances upon addi ion o DNA. The b oadening is mo e se e e o esonances in he mo e igid pa s o he molecules (e.g. in he β-shee ) and is absen o esonances o he lexible C- e minus (Supplemen a y Fig. 7c). This indica es ha bo h FBP and FIR bind a he same ime o he FUSE DNA. Simul aneous binding es ablishes a second DNA-media ed physical link be ween FBP and FIR – s ongly impai ing he capabili y o he wo molecules o umble independen ly. FBP3 in e ac ions wi h FUSE and FIR In o de o in es iga e whe he a 10–50 old di e ence in a ini y can jus i y ec ui men o FIR in a cellula en i onmen , we compa ed p o ein–p o ein and p o ein–DNA in e ac ions in he FBP and FBP3 sys ems. Using BLI we show ha a FBP3 Nbox-KH1-KH4 cons uc binds o he FUSE elemen (ssFUSE40) wi h an a ini y compa able o he one o FBP (Kd = 1.1 nM) (da a no shown, Supplemen a y Table 2). Howe e , when we added FIR o he FBP3 Nbox- KH1-KH4–FUSE sys em no Nbox-dependen binding o FIR o he DNA was obse ed (Fig. 7b). This is consis en wi h he published da a ha indica e ha FBP3 Nbox does no ec ui FIR in he cell, and con i ms ha he Nbox-dependen ec ui men o FIR we obse e by BLI mi o s he unc ionally ele an ec ui men o FIR. Nex , we es ablished he di e ence in a ini y o FBP and FBP3 Nboxes o FIR RRM1-RRM2 ha leads o he loss o unc ional ec ui men . We measu ed he a ini y o he FBP3 Nbox–FIR RRM1-RRM2 in e ac ion: he Kd o he complex is 280 ± 36 μM and is highe han he one o he FBP Nbox–FIR RRM1- RRM2 complex by a ac o o ~20 (Fig. 7c). Tha is, a d op by a ac o o ~20 in a ini y co esponds o a loss o unc ion in he FBP(s) Nbox–FIR in e ac ion. Discussion The wo k p esen ed he e explains he molecula basis o FIR-media ed c-myc egula ion. We show ha he ini ial ec ui men o FBP by he pa ially open FUSE is jus i ied by i s high a ini y o a speci ic si e on he DNA. FIR is hen ec ui ed by concu en FBP and DNA binding, and sepa a e FIR–FBP and FIR–DNA in e ac ions a e es ablished on FIR RRM1- RRM2. I has been shown ha a single RRM domain (RRM2 o Ra e 1) can bind simul aneously o p o ein and RNA a ge s 22,23. The use o wo RRM domains in FIR is he e o e no an absolu e equi emen imposed by he small size o he RRM domain bu a cha ac e is ic o his speci ic sys em. In FIR, he physical sepa a ion o he DNA and FBP binding su aces (Fig. 6a) minimizes in e e ence be ween binding e en s in he h ee-subuni complex. This sepa a ion is unc ionally impo an , because he FBP–FIR in e ac ion mus be dis up ed upon FBP ejec ion while FIR emains bound o he DNA. The use o a cen al pa ch o alanines su ounded by la ge hyd ophobic esidues o c ea e a ansien and ye speci ic in e ac ion is, as a as we a e awa e o , desc ibed he e o he i s ime. I will be o in e es o explo e whe he such a pa e n de ines o he ansien p o ein–p o ein in e ac ions. Assembly o he FIR-con aining complex (and he e o e ansc ip ional ep ession) is con oled by he ela i ely low a ini y, high speci ici y FBP–FIR in e ac ion 11, which is based on a Cukie e al. Page 6 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip no el, non-canonical mode o RRM–p o ein ecogni ion. The sequence o FBP and FIR dissocia ion in he cell mus be, a leas pa ially, con oled by o he p o ein–p o ein in e ac ions. The binding a ini y we measu e o he FIR–FUSE in e ac ion (Kd ~7 μM) is al oge he no su icien o main ain he obse ed FIR–DNA binding a e FBP has dissocia ed 8. The FUSE mechanism is based on a single ac i a o (FBP) ec ui ing i s own (single) ep esso (FIR). This allows he es ablishmen o a unc ional ela ionship be ween he leng h o he ac i a ion s ep and he weak coupling be ween FBP and FIR binding o ssFUSE. A s ong coope a i i y o FBP–FIR binding on he DNA would speed FIR ec ui men beyond wha is desi able and would educe he peak o c-myc exp ession (Fig. 7d). Ins ead, he long lexible linke and he weak Nbox–FIR in e ac ion c ea e a weak coupling be ween he wo p o eins ha can be egula ed mo e sub ly. In e es ingly, he analysis o he in e ac ion be ween FBP3 Nbox and FIR shows ha a d op by a ac o o 20 in he s eng h o he in e ac ion impai s unc ional FIR ec ui men in he cell, con i ming a egula o y model whe e a change in a ini y o one– wo o de s o magni ude egula es an impo an ansc ip ional swi ch. Recen wo k has asse ed he po en ial o a ge ing c-myc as an an icance s a egy 24 sugges ing ha c- myc ep ession could be explo ed as a he apeu ic s a egy. Ou wo k e eals how he modula ion o FBP–FIR coupling allows an ini ial pe iod o unhinde ed FBP-p omo ed ansc ip ion. This opens he exci ing possibili y ha compounds ha s abilize he Nbox– RRM2 in e ac ion could speed FIR ec ui men and educe p oduc i e ansc ip ion o c-myc (Fig. 7d). Me hods P o ein p epa a ion FIR RRM1-RRM2 (amino-acids 103–297, NP_055096) we e cloned in o pETM-30 ec o (EMBL-Heidelbe g, P o ein Exp ession Facili y), in oducing TEV p o ease-clea able HisTag-GST usion N- e minal o he inse . The HisTag-GST usion p o ein was pu i ied om he soluble ac ion by nickel-a ini y ch oma og aphy (Qiagen) ollowed by gel il a ion. The inal p o ein was concen a ed o 0.4–0.6 mM and s o ed in 10 mM T is-HCl pH 8.0 (o 7.4), 50 mM NaCl, 2 mM T is(2-ca boxye hyl)phosphine (TCEP), 0.05% (w/ ) NaN3 in he p esence o p o ease inhibi o s (Roche) a −80 °C. Unlabeled FBP Nbox pep ide (amino-acids 27–52, NP_003893) wi h and wi hou an N- e minal Y esidue and unlabeled FBP3 Nbox (amino-acids 15–40, NP_003925) we e chemically syn hesized (Pep ide Syn hesis Facili y, Uni e si y o B is ol and in-house, espec i ely). The addi ion o Y esidue does no a ec FBP Nbox–FIR RRM1-RRM2 in e ac ion as assessed by NMR i a ions (da a no shown). Labeled (15N o 15N13C) FBP Nbox pep ide (wi h N- e minal Y) was cloned and exp essed as desc ibed o FIR RRM1- RRM2 and hen pu i ied om he soluble ac ion by nickel-a ini y ch oma og aphy (Qiagen) ollowed by HPLC pu i ica ion. The pep ide was s o ed in 10 mM T is-HCl pH 5.0 (o 8.0), 0.05% (w/ ) NaN3 in he p esence o p o ease inhibi o s (Roche) a −80 °C. FBP Nbox-KH1-KH4 (amino-acids 27–455, NP_003893), FBP KH1-KH4 (amino-acids 85– 455, NP_003893) and FBP3 Nbox-KH1-KH4 (amino-acids 13–431, NP_003925) we e cloned, exp essed and pu i ied by nickel-a ini y ch oma og aphy as desc ibed o FIR RRM1-RRM2. The p o ein was u he pu i ied on anion exchange and hepa in a ini y columns. The p o eins we e s o ed in 10 mM T is-HCl pH 7.4, 50 mM NaCl, 2 mM TCEP, 0.05% (w/ ) NaN3 in he p esence o p o ease inhibi o s (Roche) a −80 °C. Fo de ails o he p o ein p epa a ions see he online Supplemen a y Me hods sec ion. Cukie e al. Page 7 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip ssDNA oligonucleo ides All ssDNA oligonucleo ides and SIA oligonucleo ides pools we e chemically syn hesized (Sigma and In eg a ed DNA Technologies). ssFUSE40 co esponds o nucleo ides 769–808 in he complemen a y s and o X00364 en y. NMR spec oscopy All NMR expe imen s we e eco ded a 37 °C o 45 °C on Va ian Ino a and B uke A ance spec ome e s equipped wi h c yop obes and ope a ing a 600, 700 and 800 MHz 1H equencies. The spec a we e p ocessed wi h he NMRPipe package 27 and analyzed wi h Spa ky 28. Fu he de ails can be ound in Supplemen a y Me hods. S uc u e calcula ions Dis ance and angle es ain s we e used o pe o m s uc u e calcula ions wi h ARIA 1.2 29. Expe imen al dis ance es ain s we e achie ed by in eg a ing Spa ky assigned NOE peaks wi h XEASY 30. Dihed al es ain s (φ and ψ) we e ob ained om he chemical-shi -based TALOS da abase 31. Fo he ee and bound p o eins, H-bond es ain s we e added in subsequen calcula ions i an exchange-p o ec ed 1HN was H-bonded in a leas 50% o he p elimina y s uc u es gene a ed in a gi en un. Fo he pep ide, he H-bond es ain s o a s anda d α-helix we e added in he helical egion (Ala30–Lys44). In i e a ions 0–7 o ARIA calcula ions i y andomized con o me s we e subjec o simula ed annealing wi h a s anda d CNS p o ocol 32. The en lowes -global-ene gy s uc u es we e used o assignmen a he nex i e a ion. In i e a ion 8 he numbe o gene a ed s uc u es was inc eased o 200. Finally, he 20 lowes -ene gy ob ained s uc u es we e wa e e ined 33. The quali y o each gene a ed amily was e alua ed wi h PROCHECK_NMR 34. Ramachand an plo s a is ics o FIR RRM1-RRM2: Mos a o ed egions 83.9%, Addi ional allowed egions 14.3%, Gene ously allowed egions 0.9%, Disallowed egions 0.9%. Ramachand an plo s a is ics o FIR RRM1-RRM2–FBP Nbox complex: Mos a o ed egions 82.2%, Addi ional allowed egions 15.3%, Gene ously allowed egions 1.0%, Disallowed egions 1.5%. The s uc u es we e displayed and analyzed wi h MOLMOL 26, PYMOL (www.pymol.o g) and Insigh II (Accel ys). Ci cula Dich oism (CD) Fa -UV CD spec um o 50 μM FBP Nbox in 10 mM T is-HCl pH 7.4, 20 mM NaCl was eco ded a 5 °C on a Jasco J-715 spec opola ime e (Jasco) equipped wi h a PTC-348 Pel ie empe a u e-con ol sys em. The CD signal is epo ed as he mean esidue CD ex inc ion coe icien (ΔεMRW). Binding assays: NMR NMR i a ions we e pe o med by i a ing unlabeled in e ac ing pa ne in o solu ion o 15N- labeled sample ( o de ails see Supplemen a y Me hods). 15N so as HMQC spec a 35 we e eco ded a each i a ion poin a 37 °C o 45 °C on Va ian Ino a and B uke A ance spec ome e s equipped wi h c yop obes and ope a ing a 600, 700 and 800 MHz 1H equencies. To ob ain Kd alues he a e age chemical shi pe u ba ions (Δδa g = ((ΔδN/ 10)2 + ΔδH2)1/2) o 7–10 peaks we e plo ed as a unc ion o ligand:p o ein a io, ollowed by wo-pa ame e nonlinea leas -squa es i o he da a o each peak in he p og am O igin (O iginLab), using a one-si e binding model which co ec s o dilu ion e ec 36. Kd alues a e epo ed as a e age ± 2 s anda d de ia ions. Cukie e al. Page 8 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip Binding assays: SIA SIA analysis was pe o med as desc ibed in Beu h e al. 17 wi h 16 pools o 5-me DNA oligonucleo ides (nANNN, nGNNN, nCNNN e c.). In b ie , solu ions o 25 μM 15N-labeled FIR RRM1-RRM2 in 10 mM T is-HCl pH 7.4, 50 mM NaCl, 2 mM TCEP we e i a ed wi h indi idual DNA pools ( a ios 1:0, 1:1 and 1:4). 15N so as HMQC spec a 35 we e eco ded a each i a ion poin a 37 °C on a Va ian Ino a spec ome e equipped wi h c yop obe and ope a ing a 800 MHz 1H equency. A e age chemical shi pe u ba ions (Δδa g = ((ΔδN/ 10)2 + ΔδH2)1/2) o 15 peaks we e analyzed o ob ain SIA sco es. The esul s we e isualized wi h WebLogo (h p://weblogo.be keley.edu/logo.cgi). Binding assays: BLI All BLI expe imen s we e pe o med in 10 mM T is-HCl pH 7.4, 150 mM NaCl, 2 mM TCEP, 0.5 mg ml−1 BSA on an Oc e Red ins umen (Fo eBio, Inc., Menlo Pa k, CA) ope a ing a 25 °C. S ep a idin coa ed biosenso s wi h immobilized bio inyla ed ssFUSE40 (Sigma) we e exposed o di e en concen a ions o FIR RRM1-RRM2, FBP Nbox-KH1-KH4, FBP KH1- KH4 o FBP3 Nbox-KH1-KH4 o combina ion o hem as desc ibed in he Supplemen a y Resul s sec ion. Supplemen a y Ma e ial Re e o Web e sion on PubMed Cen al o supplemen a y ma e ial. Acknowledgmen s We would like o hank D s A. O eggioni and T. F enkiel o help in eco ding NMR expe imen s, D s C. deChia a and G. Nicas o o ad ice on he ARIA p o ocols used in s uc u e calcula ions, D . A. M. Candel o help wi h spec oscopic da a and S. K alo ico a o gene al suppo in he lab. We would also like o hank D Pe e Rosen hal o use ul discussions. All NMR spec a we e eco ded a he MRC Biomedical NMR Cen e. We would like o hank D S. Kindle (Uni e si y o Hambu g) o he gi o a plasmid wi h he FIR RRM1-RRM2 gene. This wo k has been unded by he MRC G an -in-aid U117574558. Re e ences 1. Le ens D. Disen angling he MYC web. P oc Na l Acad Sci U S A 2002;99:5757–5759. [PubMed: 11983876] 2. Wie s a I, Al es J. The c-myc p omo e : s ill Mys e Y and challenge. Ad Cance Res 2008;99:113– 333. [PubMed: 18037408] 3. Kenne h NS, Whi e RJ. Regula ion by c-Myc o ncRNA exp ession. Cu Opin Gene De 2009;19:38– 43. [PubMed: 19179065] 4. Meye N, Penn LZ. Re lec ing on 25 yea s wi h MYC. Na Re Cance 2008;8:976–990. [PubMed: 19029958] 5. Kouzine F, Liu J, San o d S, Chung HJ, Le ens D. The dynamic esponse o ups eam DNA o ansc ip ion-gene a ed o sional s ess. Na S uc Mol Biol 2004;11:1092–1100. [PubMed: 15502847] 6. Kouzine F, San o d S, Elisha-Feil Z, Le ens D. The unc ional esponse o ups eam DNA o dynamic supe coiling in i o. Na S uc Mol Biol 2008;15:146–154. [PubMed: 18193062] 7. Liu J, e al. De ec i e in e play o ac i a o s and ep esso s wi h TFIIH in xe ode ma pigmen osum. Cell 2001;104:353–363. [PubMed: 11239393] 8. Liu J, e al. The FUSE/FBP/FIR/TFIIH sys em is a molecula machine p og amming a pulse o c- myc exp ession. EMBO J 2006;25:2119–2130. [PubMed: 16628215] 9. Ma sushi a K, e al. An essen ial ole o al e na i e splicing o c-myc supp esso FUSE-binding p o ein- in e ac ing ep esso in ca cinogenesis. Cance Res 2006;66:1409–1417. [PubMed: 16452196] 10. Ma sushi a K, e al. c-myc supp esso FBP-in e ac ing ep esso o cance diagnosis and he apy. F on Biosci 2009;14:3401–3408. [PubMed: 19273283] Cukie e al. Page 9 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip Figu e 5. FIR–FBP in e ac ion ep esen s a no el ecogni ion mode in he RRM amily. Ribbon ep esen a ions o he s uc u es o p o ein–p o ein complexes be ween RRM domains (g ey) and hei binding pa ne s (blue and g een): Y14–Mago–PYM (1 k8), UPF3b–UPF2 (1uw4), p14–SF3b155 (2 9d), Ra e 1–Vinculin (3h2u), U2AF35–U2AF65 (1jm ), U2AF65–SF1 (1o0p), SPF45–SF3b155 (2peh), U2B″–U2A′ (1a9n) and FIR–FBP ( his s udy). In he PTB– Ra e 1 complex he Ra e 1 pep ide posi ion is epo ed (blue) on he s uc u e o PTB RRM2 in he complex wi h RNA (2adb), acco ding o he published model 22. U2AF35, U2AF65 and SPF45 RRM domains belong o he UHM sub amily, ha bind a conse ed T p esidue in he ligand pep ides and includes he hi d RRM o PUF60, an iso o m o FIR p o ein 25. The ca oon ep esen a ion o he di e en s uc u es we e gene a ed using he p og am MOLMOL 26. Cukie e al. Page 16 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip Figu e 6. FIR RRM1-RRM2 independen ly binds FBP Nbox and ssFUSE on wo physically sepa a ed si es loca ed on opposi e sides o he molecule. (a) Su ace ep esen a ion o FIR RRM1- RRM2. Residues showing subs an ial CSP upon addi ion o ssFUSE29 o FBP Nbox pep ide a e colo ed in blue and ed espec i ely. This s uc u al ep esen a ion was gene a ed using he p og am PYMOL (www.pymol.o g). (b) Supe imposed 15N-1H co ela ion spec a show ha FIR RRM1-RRM2 in e ac s independen ly wi h a ssFUSE29 and he FBP Nbox pep ide. The spec a o RRM1-RRM2+ssFUSE DNA, RRM1-RRM2+FBP Nbox and RRM1-RRM2 +ssFUSE DNA+FBP Nbox a e in blue, ed and g een espec i ely. A ep esen a i e egion con aining one RRM1 esonance pe u bed by DNA binding (Val114) and one RRM2 esonance pe u bed by pep ide binding (Asp222) is displayed. The chemical shi changes o RRM1-RRM2 peaks in he p o ein–DNA and p o ein–pep ide complexes a e addi i e in he h ee molecule complex, indica ing independen binding. Cukie e al. Page 17 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip Figu e 7. FBP Nbox ec ui s FIR o ssFUSE DNA. (a) BLI binding assays show ha he p esence o he Nbox in FBP cons uc inc eases he a ini y o FIR o he DNA. The BLI biosenso s we e de i a ized wi h ssFUSE40 and exposed o di e en combina ions o p o ein cons uc s (indica ed on he igh ), as epo ed in he igu e. RU changes in h ee pa allel expe imen s a e displayed. (b) FBP3 Nbox does no media e ec ui men o FIR o he DNA. BLI biosenso s we e de i a ized wi h ssFUSE40 and exposed o di e en combina ions o p o ein cons uc s (indica ed on he igh ), as epo ed in he igu e. The RU changes obse ed in h ee pa allel expe imen s a e displayed. (c) Iso he ms o FBP and FBP3 Nbox binding o FIR RRM1- RRM2 (NMR). Kd a e epo ed ± 2 × s.d. (d) Th ee di e en modes o FBP–FIR coupling and hei e ec on FIR binding and c-myc ansc ip ion. Le – In he absence o a coupling be ween FBP and FIR he a ini y o FIR o he ssFUSE (Kd ~7 μM) would be oo low o FIR o bind he ssFUSE. FBP ac i a ion o c-myc ansc ip ion (bo om) would he e o e con inue unpe u bed. Middle – Boxed, he physiological scena io. The weak binding o he Nbox o FIR RRM2 and he 50-amino acid linke be ween he Nbox and KH1 a e esponsible o he physiological weak coupling be ween FBP and FIR. This coupling is necessa y o c ea e he equi ed peak in c-myc exp ession (bo om) and egula e he cell cycle. Righ – A s onge coupling be ween FBP and FIR c ea ed by, o example, a small molecula weigh compound binding a he in e ace be ween he wo molecules ha would inc ease wo o de s o magni ude he appa en a ini y o FIR o he FBP–ssFUSE complex, would speed FIR ec ui men and lead o a sho e peak in c-myc exp ession (bo om). Cukie e al. Page 18 Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1. UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip UKPMC Funde s G oup Au ho Manusc ip Cukie e al. Page 19 Table 1 NMR and e inemen s a is ics o p o ein s uc u es FIR RRM1–RRM2 (103–297) FIR RRM1–RRM2 (103–297) – FBP Nbox (27–52) complex NMR dis ance and dihed al cons ain s Dis ance cons ain s 4,431 4,647 To al NOE 4,368 4,611 In a- esidue 1,588 1,632 In e - esidue 2,780 2,979 Sequen ial (|i – j| = 1) 906 878 Medium- ange (|i – j| < 4) 487 546 Long- ange (|i – j| > 5) 1,387 1,483 In e molecula – 72 Hyd ogen bonds 63 36 To al dihed al angle es ain s 236 280 φ118 140 ψ118 140 S uc u e s a is ics Viola ions (mean and s.d.) Dis ance cons ain s (> 0.3 Å) (Å) 0.32 ± 0.02 0.34 ± 0.03 Dihed al angle cons ain s (> 5°) – *6.37 ± 0.98 Max. dihed al angle iola ion (°) – *8.65 Max. dis ance cons ain iola ion (Å) 0.35 0.42 De ia ions om idealized geome y Bond leng hs (Å) 0.0036 ± 0.0001 0.0036 ± 0.0001 Bond angles (°) 0.480 ± 0.011 0.507 ± 0.008 Imp ope s (°) 1.46 ± 0.08 1.41 ± 0.06 A e age pai wise .m.s. de ia ion** (Å) Hea y 1.91 ± 0.23 2.03 ± 0.24 Backbone 1.28 ± 0.20 1.40 ± 0.21 *No iola ions > 5° **Pai wise .m.s. de ia ion was calcula ed among 20 e ined s uc u es ( esidues 105-288 o FIR RRM1-RRM2 and esidues 105-288 and 27-45 o FIR RRM1-RRM2–FBP Nbox complex) Na S uc Mol Biol. Au ho manusc ip ; a ailable in PMC 2011 Ma ch 1.