Intrinsically disordered regions ofnucleophosminB23 regulate its RNA bindingactivity through their inter- andintra-molecular association

著者 Hisaoka Miharu Nagata Kyosuke OkuwakiMitsuru

journal orpublication title

Nucleic acids research

volume 42number 2page range 1180-1195year 2014-01権利 (C) The Author(s) 2013 Published by Oxford

University Press This is an Open Accessarticle distributed under the terms of theCreative Commons Attribution License(httpcreativecommonsorglicensesby30)which permits unrestricted reusedistribution and reproduction in any mediumprovided the original work is properly cited

URL httphdlhandlenet2241121193doi 101093nargkt897

Intrinsically disordered regions of nucleophosminB23 regulate its RNA binding activity throughtheir inter- and intra-molecular associationMiharu Hisaoka Kyosuke Nagata and Mitsuru Okuwaki

Faculty of Medicine University of Tsukuba 1ndash1ndash1 Tennoudai Tsukuba 305ndash8575 Japan

Received June 10 2013 Revised August 30 2013 Accepted September 14 2013

ABSTRACT

Nucleophosmin (NPM1B23) is a nucleolar proteinimplicated in growth-associated functions inwhich the RNA binding activity of B23 plays essen-tial roles in ribosome biogenesis The C-terminalglobular domain (CTD) of B23 has been believed tobe the RNA binding domain because the splicingvariant B232 lacking the CTD binds considerablyless efficiently to RNA However the recognitionof target RNAs by B23 remains poorly understoodHerein we report a novel mechanism by whichB23 recognizes specific RNA targets We observedthat the nucleolar retention of B233 lacking thebasic region of B231 was lower than that of B231because of its low RNA binding activity Circulardichroism measurements indicated that thebasic region and adjacent acidic regions of B23are intrinsically disordered regions (IDRs)Biochemical analyses revealed that the basic IDRalone strongly binds to RNA with low specificityThe excessive RNA binding activity of the basicIDR was restrained by intra-molecular interactionwith the acidic IDR of B23 Chemical cross-linkingexperiments and fluorescent labeling of bipartitetetracysteine-tagged proteins suggested that theinter- and intra-molecular interactions betweenthe two IDRs contribute to the regulation of theRNA binding activity of CTD to control the cellularlocalization and functions of B23

INTRODUCTION

The biological activities of proteins are often exertedby structurally ordered domains therefore structure-function analyses have been central in understandingseveral biological processes However recent studieshave documented that the structurally disordered flexible

domain in the native state [intrinsically disordered regions(IDRs)] play crucial roles in the regulation of proteinndashprotein and proteinndashnucleic acid interactions (1) TheIDRs often adopt a structured conformation on bindingto their target molecules For example the N-terminaltransactivation domain in the transcription factor p53 isintrinsically disordered and folded on binding to the Taz2domain of the co-activator p300 (2) In addition severalDNA binding proteins containing zinc fingers helixndashloopndashhelix motifs and homeodomains have N- or C-terminalextended IDRs that assist efficient recognition andbinding to target sequences (3) Furthermore variouspost-translational modifications occur in the IDRs sug-gesting that the IDRs play an integral role in regulationof the protein functions (4)

The nucleolar phosphoprotein NPM1B23 is involvedin the regulation of pre-ribosome RNA (rRNA) transcrip-tion processing and pre-ribosome transport (5ndash7) Inaddition B23 plays important roles in the regulation ofcentrosome duplication (8) genomic stability (9) andresponse to cellular stresses (10) Because of its widerange of functions the deregulation of B23 functions isclosely associated with tumorigenesis (10) Mutations ofthe C-terminal domain (CTD) of B23 are frequentlyobserved in normal karyotype adult acute myeloidleukemia (11) Moreover B23 overexpression is oftenobserved in solid tumors (12ndash17) Therefore it is crucialto understand the mechanism of B23 action

We previously identified B23 as a factor involved inadenovirus chromatin remodeling (18) B23 was foundto be involved not only in adenovirus chromatinremodeling but also in the regulation of the cellularrRNA gene chromatin to stimulate transcription (5) B23mainly localizes in the nucleolus but shuttles between thenucleolus nucleoplasm and cytoplasm Three splicingvariants of B23 were reported to be expressed in humancells (19) (see Figure 1A) B231 is the longest variant andthe most studied protein among the B23 variants B232has 259 amino acids and lacks the C-terminal 35 aminoacids (dark gray bar with black dots) the third variant

To whom correspondence should be addressed Tel +81 29 853 7950 Fax +81 29 853 3942 Email mokuwakimdtsukubaacjp

1180ndash1195 Nucleic Acids Research 2014 Vol 42 No 2 Published online 7 October 2013doi101093nargkt897

The Author(s) 2013 Published by Oxford University PressThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (httpcreativecommonsorglicensesby30) whichpermits unrestricted reuse distribution and reproduction in any medium provided the original work is properly cited

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

Intrinsically disordered regions of nucleophosminB23 regulate its RNA binding activity throughtheir inter- and intra-molecular associationMiharu Hisaoka Kyosuke Nagata and Mitsuru Okuwaki

Faculty of Medicine University of Tsukuba 1ndash1ndash1 Tennoudai Tsukuba 305ndash8575 Japan

Received June 10 2013 Revised August 30 2013 Accepted September 14 2013

ABSTRACT

Nucleophosmin (NPM1B23) is a nucleolar proteinimplicated in growth-associated functions inwhich the RNA binding activity of B23 plays essen-tial roles in ribosome biogenesis The C-terminalglobular domain (CTD) of B23 has been believed tobe the RNA binding domain because the splicingvariant B232 lacking the CTD binds considerablyless efficiently to RNA However the recognitionof target RNAs by B23 remains poorly understoodHerein we report a novel mechanism by whichB23 recognizes specific RNA targets We observedthat the nucleolar retention of B233 lacking thebasic region of B231 was lower than that of B231because of its low RNA binding activity Circulardichroism measurements indicated that thebasic region and adjacent acidic regions of B23are intrinsically disordered regions (IDRs)Biochemical analyses revealed that the basic IDRalone strongly binds to RNA with low specificityThe excessive RNA binding activity of the basicIDR was restrained by intra-molecular interactionwith the acidic IDR of B23 Chemical cross-linkingexperiments and fluorescent labeling of bipartitetetracysteine-tagged proteins suggested that theinter- and intra-molecular interactions betweenthe two IDRs contribute to the regulation of theRNA binding activity of CTD to control the cellularlocalization and functions of B23

INTRODUCTION

The biological activities of proteins are often exertedby structurally ordered domains therefore structure-function analyses have been central in understandingseveral biological processes However recent studieshave documented that the structurally disordered flexible

domain in the native state [intrinsically disordered regions(IDRs)] play crucial roles in the regulation of proteinndashprotein and proteinndashnucleic acid interactions (1) TheIDRs often adopt a structured conformation on bindingto their target molecules For example the N-terminaltransactivation domain in the transcription factor p53 isintrinsically disordered and folded on binding to the Taz2domain of the co-activator p300 (2) In addition severalDNA binding proteins containing zinc fingers helixndashloopndashhelix motifs and homeodomains have N- or C-terminalextended IDRs that assist efficient recognition andbinding to target sequences (3) Furthermore variouspost-translational modifications occur in the IDRs sug-gesting that the IDRs play an integral role in regulationof the protein functions (4)

The nucleolar phosphoprotein NPM1B23 is involvedin the regulation of pre-ribosome RNA (rRNA) transcrip-tion processing and pre-ribosome transport (5ndash7) Inaddition B23 plays important roles in the regulation ofcentrosome duplication (8) genomic stability (9) andresponse to cellular stresses (10) Because of its widerange of functions the deregulation of B23 functions isclosely associated with tumorigenesis (10) Mutations ofthe C-terminal domain (CTD) of B23 are frequentlyobserved in normal karyotype adult acute myeloidleukemia (11) Moreover B23 overexpression is oftenobserved in solid tumors (12ndash17) Therefore it is crucialto understand the mechanism of B23 action

We previously identified B23 as a factor involved inadenovirus chromatin remodeling (18) B23 was foundto be involved not only in adenovirus chromatinremodeling but also in the regulation of the cellularrRNA gene chromatin to stimulate transcription (5) B23mainly localizes in the nucleolus but shuttles between thenucleolus nucleoplasm and cytoplasm Three splicingvariants of B23 were reported to be expressed in humancells (19) (see Figure 1A) B231 is the longest variant andthe most studied protein among the B23 variants B232has 259 amino acids and lacks the C-terminal 35 aminoacids (dark gray bar with black dots) the third variant

To whom correspondence should be addressed Tel +81 29 853 7950 Fax +81 29 853 3942 Email mokuwakimdtsukubaacjp

1180ndash1195 Nucleic Acids Research 2014 Vol 42 No 2 Published online 7 October 2013doi101093nargkt897

The Author(s) 2013 Published by Oxford University PressThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (httpcreativecommonsorglicensesby30) whichpermits unrestricted reuse distribution and reproduction in any medium provided the original work is properly cited

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

Figure 1 Characterization of B23 variants (A) Schematic representation of B23 splicing variants Black oligomerization domain stripes acidicregions black with white dots basic region dark gray with black dots CTD (B) Localization of EF-tagged B23 variants HeLa cells weretransfected with vectors for expression of EF-B231 EF-B232 and EF-B233 Localization of the protein was observed by fluorescent microscopeBar at the bottom indicates 10 mm (C) FRAP analysis of B23 variants The mobility of EF-B23 proteins expressed in HeLa cells as in Figure 1B wereexamined by FRAP assay The fluorescence at the bleached nucleoli relative to that before bleaching (10) was calculated and plotted as a function oftime The data were represented as mean valuesplusmnSD from 15 15 and 13 experiments for B231 (black) B232 (blue) and B233 (red) respectivelyThe t12 of fluorescence recovery was estimated by curve fitting and graphically represented (right panel) P-values were calculated by t-tests andindicated with Plt 001 (D) Immunoprecipitation of EF-tagged B23 variants 293T cells expressing EF EF-B231 EF-B232 and EF-B233 (lanes1ndash4 respectively) were subjected to immunoprecipitation with anti-Flag tag antibody in the buffer containing 150mM (lanes 5ndash8) or 300mM (lanes9ndash12) of NaCl Precipitated proteins were separated by SDSndashPAGE and detected with western blotting using anti-Flag tag and -B23 antibodies(bottom panels) RNAs co-precipitated with EF-tagged proteins were purified and separated on 6 denaturing PAGE and visualized by Gel Redstaining (top panel) Lane M indicates RNA molecular markers (E) RNA binding activity of B23 variants Recombinant GST-tagged B23 proteinswere analyzed by SDSndashPAGE followed by CBB staining (top panel) Purified proteins (5 mg) were incubated in the absence or presence of totalRNA (10 mg) prepared from 293T cells and loaded on a 15ndash40 sucrose gradient Proteins in fractions collected from the top were analyzed bySDSndashPAGE and visualized by CBB staining For the experiments incubated with RNA the amount of the B23 proteins in each fraction relative tothat of input was calculated and shown in graph (bottom panel)

Nucleic Acids Research 2014 Vol 42 No 2 1181

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

which we have named B233 has 265 amino acids andlacks 29 amino acids in the basic amino acid-rich region(shown as black bar with white dots in Figure 1A) B23forms a pentamer and decamer through the N-terminaloligomerization domain (black bar in Figure 1A) (20)A recent study reported that the CTD of B231 formsglobular structure consisting of three a-helices and itsdestabilization abolishes its nucleolar localization (21)Because B232 does not bind to RNA CTD is believedto be crucial for its RNA binding activity (22) It was alsoshown that the CTD of B23 binds to G-quadruplex DNAand this activity is reinforced by at least 17 adjacentresidues located in the basic region (2324) Several predic-tion programs suggest that the acidic and basic regions inthe central part of B23 are IDRs The two acidic regions[acidic IDR (aIDR) stripe bar in Figure 1A] are import-ant for the histone chaperone activity of B23 however thestructure of this region and relationship to its RNAbinding activity have not been analyzed Recently it wasreported that a synthetic peptide derived from the basicregion of B23 which shows a typical random-coilspectrum by circular dichroism (CD) stabilizes its CTD(25) We previously demonstrated that the cdc2cyclin Bkinase phosphorylates the basic IDR (bIDR) of B231and this phosphorylation significantly decreases its RNAbinding activity (26) However the mechanism by whichcdc2cyclin B-mediated phosphorylation decreases theRNA binding activity of B231 is currently unknownIn this study we examined the bIDR involved in the

regulation of the RNA binding activity of B231 Weobserved that B233 lacking a part of the bIDR showedsignificantly lower RNA binding activity than B231Further analyses strongly suggested that the associationbetween acidic and bIDRs of B231 allowed its recognitionand binding to target RNAs

MATERIALS AND METHODS

Cell culture and transfection

HeLa and 293T cells were maintained in Dulbeccorsquosmodified Eaglersquos medium (Nacalai Tesque) supplementedwith 10 fetal bovine serum Transient transfectionof plasmid DNA was performed using GeneJuice(Novagen) according to the manufacturerrsquos instructionsAntibodies Anti-B23 serum was generated in rabbit

using the His-B231-CR protein as an antigen Anti-Flagtag antibody (M2 SIGMA-ALDRICH) and anti-His tagantibody (HIS-1 SIGMA-ALDRICH) were commerciallyavailablePlasmids pEGFPC1-Flag-B231 pGEX2T-B231 and

pET14b-B231 pET14b-B231-C1 (used as B232)pET14b-B231-C2 (used as B231-C188) pET14b-B231-C3 and pEGFPC1-Flag-B232 pET14b-B232and pGEX2T-B232 and pET14b-B231-T4sD weredescribed previously (1826ndash28) To construct pEGFP-Flag-B233 the cDNA fragment was generated using atwo-step PCR based on the sequence NM_199185 Inthe first PCR two DNA fragments for B231 (1ndash194)and B231 (233ndash294) were amplified using T7 promoterprimer and 50-AGTGAAGAAAGGACAAGAATCCTT

CAAG-30 and T3 promoter primer and 50-ATTCTTGTCCTTTCTTCACTGGCGCTTT-30 with pBS-Flag-B231 as a template The amplified DNA fragments wereused in the second PCR as templates with T7 and T3promoter primers Amplified DNA fragment wasdigested by NdeI and BamHI and cloned into the samesites of pET14b The B233 cDNA was cloned into pBS-Flag by NdeI and HindIII sites of pBS-Flag vector TheFlag-B233 cDNA was cut out from pBS-Flag-B233 bydigestion with BamHI and subcloned into the same sitesof pGEX2T (GE Healthcare) or pEGFPC1 (ClonetechLaboratories) To construct the expression vectors ofGST- or His-tag fusion protein if not otherwise specifycDNAs prepared as following were digested by BamHIor NdeI and BamHI and cloned into the same sitesof pGEX2T or pET14b respectively To constructpGEX2T-B231-CR pGEX2T-B231-CRA pGEX2T-B231-CRD and pGEX2T-B231-CR1 the cDNA frag-ments were amplified by PCR using 50-AAGGATCCCATATGAAAGCGCCAGTGAAGAAA-30 and T7 termin-ator primer with appropriate templates The B231-CR15cDNA was amplified using 50-AAGGATCCCATATGAGTTCTGTAGAAGACATTAA-30 and GEX-30 primerand pGEX2T-B231-CR as a template To constructpGEX2T-B231-A2 the cDNA fragment was generatedusing a two-step PCR Two DNA fragments wereamplified using the T7 promoter primer and 50-CTGGCGCTTTAGCAGCAAGTTTTACTTT-30 and the T3promoter primer and 50-ACTTGCTGCTAAAGCGCCAGTGAAGAAA-30 with pBS-Flag-B231 as a templateThe second PCR and cloning into pGEX2T of thisfragment was performed as described earlier in the textTo construct pGEX2T-A1A2 the cDNA was cut outfrom pBS-Flag-B23-A (5) by digestion with BamHIand subcloned into the same site of pGEX2T TheB231-N and B232-N cDNA fragments wereamplified using 50-AAGAATTCAACATATGGAGGAAGATGCAGAGTCAGA-30 and T7 terminator primerand pET14b-B231 and pET14b-B232 as templates Toconstruct pET14b-B231-Nc N-terminal CysndashCys- (CC-)tag was introduced by PCR using 50-AAGGATCCATATGTTGAACTGCTGCGAGGAAGATGCAGAGTCA-30

and T7 terminator primer and pET14b-B231-N as atemplate To construct pET14b-B231-Cc and pET14b-B231-NCcc CysndashCys- (CC-) tag was introduced byPCR using 50-GCAGCAAGGTCCTTTTGGTGTTTTAGGAG-30 and 50-GTAGAAGACATTAAAGCAAAAATG-30 and pET14b-B231-N and pET14b-B231-Nc astemplates To construct pET14b-B231-Ccc tetracysteine(CCPGCC) tag was introduced by PCR using 50-GCAGCATCCTGGGCAGCAGTTCAAAGGTCCTTTTGGTGTTTTAGGAGT-30 and 50-GTAGAAGACATTAAAGCAAAAATG-30 and pET14b-B231-N as a templateAmplified DNA fragments containing pET14b vectorsequence were phosphorylated with T4 polynucleotidekinase (TOYOBO) for following ligation reactionpET14b-B232-Nc pET14b-B232-Cc pET14b-B232-NCcc and pET14b-B232-Ccc were constructed with thesame procedure as pET14b-B231-N using T7 promoterprimer and 50-GCCTAAGGATCCTTAACCACCTTTTTCTATACTTGC-30 and appropriate vectors as

1182 Nucleic Acids Research 2014 Vol 42 No 2

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

templates The fragments of pET14b-B231-C160 -C194-C206 -C223 and -C241 and pET14b-B231-A-C160-C194 -C206 -C223 and -C241 were amplified using T7promoter primer and 50-ccggatccttaagcagcaagttttacttttttctg-30 50-ccggatccttatttcttcactggcgctttttc-30 50-ccggatccttacttttgtgcatttttggctgg-30 50-ccggatccttattttgatcttggtgttgatga-30 and 50-ccggatccttaaggtccttttggtgttttagg-30 andpET14b-B231 and pET14b-B231-A1A2 as templatesrespectively

Purification of recombinant proteins

For expression and purification of recombinant proteinsBL21 (DE3) was transformed with pGEX2T or pET14bplasmids containing B231 mutant cDNAs The trans-formed Escherichia coli were grown at 37C until OD600

reached 04 Expression of the recombinant proteins wasinduced by addition of isopropyl b-D-thiogalacto-pyranoside at 30C for 3 h or at 18C for 16 h For puri-fication of GST-tagged proteins sonicated bacterialcell lysates were incubated with Glutathione sepharose(GE Healthcare) in buffer A [50mM TrisndashHCl (pH 79)01 Triton X-100 and 01mM PMSF] containing150mM NaCl The resin was washed with buffer A con-taining 500mM NaCl treated with micrococcal nuclease(MNase) (Nuclease S7 Roche Applied Science) andextensively washed with buffer A containing 500mMNaCl GST-tagged proteins retained on the resin wereeluted in a buffer containing [20mM reduced glutathione100mM TrisndashHCl (pH 79) and 120mM NaCl] His-tagged proteins were purified with HIS-Select Nickelaffinity gel according to the method suggested by themanufacturer (SIGMA-ALDRICH) Purified proteinswere dialyzed against buffer H [20mM HepesndashNaOH(pH 79) 50mM NaCl 05mM EDTA 1mM DTT01mM PMSF and 10 Glycerol] His-tagged proteinsused in proteinase digestion assays and FlAsH assayswere further purified by separation on SDSndashPAGEfollowed by extraction in a buffer containing [01SDS 50mM TrisndashHCl (pH 79) 01mM EDTA150mM NaCl and 5mM DTT] and precipitated withacetone Precipitated proteins were denatured in a buffercontaining [6M guanidinendashHCl 50mM TrisndashHCl (pH79) and 5mM DTT] refolded by dialysis against abuffer containing [50mM HepesndashNaOH (pH 80)125mM MgCl2 100mM KCl 02mM EDTA 01NP40 and 1mM DTT] and dialyzed against buffer HThe protein concentrations were estimated by comparisonwith proteins in a standard molecular mass marker (Bio-Rad) on SDSndashPAGE stained with Coomassie BrilliantBlue (CBB) Identity of a part of recombinant proteinswas confirmed by matrix-assisted laser-desorption time-of-flight mass spectrometry (MALDI-TOF-MS) and thepurity of the proteins was also analyzed by SDSndashPAGE(Supplementary Figure S1) The phosphorylation ofrecombinant proteins was performed as described in(26) using mitotic cell extracts as an enzyme sourcein a buffer containing [20mM TrisndashHCl (pH 74)10mM MgCl2 50mM NaCl 3mM ATP 1mM b-D-glycerophosphate 1mM NaVO4 and 1mM NaF] at37C for 1 h After phosphorylation reaction CaCl2(final 5mM) and MNase (Nuclease S7 Roche Applied

Science) were added and proteins were incubated foradditional 30min followed by purification using GST-tag as described earlier in the text Phosphorylationstatus was examined by phospho affinity gel electrophor-esis using phos-tag (Wako Pure Chemical Industries)(29) in the presence of Mn2+

Immunofluorescence and fluorescence recovery afterphotobleaching analysis

To observe the localization of B23 splicing variants HeLacells were transiently transfected with EF-tagged B23vectors Eighteen hours after transfection cells wereanalyzed by fluorescence microscope (Olympus IX71N-22FLPH-TS) For fluorescence recovery after photo-bleaching (FRAP) analysis HeLa cells were transientlytransfected with EF-B231 EF-B232 and EF-B233 andgrown on 35-mm glass base dishes (IWAKI) The dish wasset on inverted microscope (LSM EXCITER Carl ZeissMicroimaging Inc) in an air chamber containing 5CO2 at 37C and analyzed as previously described (30)The data were represented as mean valuesplusmnSD from 1515 and 13 experiments for B231 B232 and B233respectively

Immunoprecipitation

293T cells were transiently transfected with EF EF-B231EF-B232 and EF-B233 with Gene Juice (Novagene)Sixteen hours after transfection the cells were collectedand sonicated in buffer A containing 150mM NaCl Thecell lysates were incubated with anti-Flag M2 affinity gels(SIGMA-ALDRICH) in buffer A containing 150mM or300mM NaCl and 35 units RNase inhibitor (NacalaiTesque) The resins were washed with the same bufferand treated with 10 units of DNase I (Invitrogen) andwashed extensively with the buffer same The proteinsbound with the resin were eluted with buffer A containing150mM NaCl and Flag peptide (SIGMA-ALDRICH)separated by SDSndashPAGE and analyzed by westernblotting RNAs co-precipitated with EF-tagged proteinswere treated with proteinase K purified with phenol-chloroform extraction and ethanol precipitation andseparated by 6 denatured PAGE RNAs were visualizedby Gel Red staining (Biotium)

Sucrose density gradient centrifugation assays

GST-tagged B23 proteins (5mg) preincubated in theabsence or presence of total RNA purified from 293Tcells (10mg) or total RNA alone (10 mg) (100ml) wereloaded on 15ndash45 sucrose density gradient in the buffercontaining [20mM HepesndashNaOH (pH 79) 50mM NaCland 05mM EDTA] (21ml) The samples werecentrifuged at 54 000 rpm for 2 h at 4C in S55S rotor(Hitachi Koki SC100GXII) and fractions (150ml) werecollected from the top Proteins in each fraction wereseparated by SDSndashPAGE and visualized by CBBstaining RNAs in each fraction were purified separatedby 1 denatured agarose gel and visualized by Gel Redstaining The distribution patterns of proteins and RNAswere analyzed using CS analyzer (ATTO)

Nucleic Acids Research 2014 Vol 42 No 2 1183

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

CD spectral analysis

CD spectral analyses were performed using a JASCOspectrometer J-710 instrument (JASCO Inc) at roomtemperature in the wavelength interval 190ndash260 nmData were collected at 02 nm resolution 20 nmminscan speed 10 nm bandwidth and 2 s response using01 cm path-length quartz cuvette Data for each proteincollected from four scans were averaged The signal wasconverted to mean molar residue ellipticity in units ofmdeg cm2 dmol1 The concentration of His-tagged B23proteins was 5 mM in 50mM sodium phosphate buffer(pH 70)

Filter binding assays

Filter binding assays were performed as described (26)RNAs used in this assay was purified from 293T cellsusing Sepasol RNA I Super G (Nacalai tesque) andtreated with DNase I Experiments performed withdoublet were repeated more than twice The data wererepresented as mean valuesplusmnSD

Mobility shift assays

GST-B231-A was incubated with GST-tagged B231-CR1 B231-CR15 and B231-CR in 10 ml of buffer Hfor 30min at room temperature Incubated samples wereseparated on 6 native-PAGE in 05 TBE andvisualized by CBB staining

Trypsin digestion assays

His-B231 and His-B231 -A1A2 (2 mg) were incubated in10 ml of 50mM phosphate buffer (pH 76) containingTrypsin for 2min at 37C The proteolysis was stoppedwith the addition of 35 ml of 4SDS sample buffer andincubation for 5min at 95C and proteins were separatedby SDSndashPAGE followed by western blotting usinganti-His tag antibody

FlAsH labeling

Recombinant proteins containing Cys-tag (11 22 4466 88 and 132mM respectively) were mixed with10 mM FlAsH-EDT2 (Toronto Research Chemicals) and05mM EDT (Nacalai Tesque) in a buffer containing[10mM HepesndashNaOH (pH 80) 100mM NaCl 1mMEDTA 25mM TrisndashHCl (pH 79)] in 384-well plate(Corning) The fluorescence emission at 535plusmn5nm wasdetected by Varioskan (Thermo Fisher Scientific)

Cross-link assays

His-B231 -N and His-B232 -N (15ndash45 pmoles)were incubated in the buffer [20mM HepesndashNaOH(pH 79) 50mM NaCl 05mM EDTA 10 Glycerol1mM DTT 05mM PMSF] containing 01glutaraldehyde (10ml) for 10min at room temperatureThe reaction was stopped by the addition of 1 ml of 1MTrisndashHCl (pH 79) and 35ml of 4 SDSndashPAGE samplebuffer and incubation at 95C for 5min Cross-linkedproteins were separated by SDSndashPAGE and visualizedby CBB staining

RESULTS

Localization and RNA binding activity of B23 splicingvariants

Three splicing variants of NPM1B23 B231 B232 andB233 are expressed in HeLa cells (Figure 1A) The thirdvariant which we named B233 has not beencharacterized thus far We first characterized the cellularlocalization and biochemical activity of this variantprotein compared it with that of B231 and B232EGFP-Flag (EF)-tagged B231 B232 and B233 weretransiently expressed in HeLa cells and their localizationwas observed (Figure 1B) Consistent with previous obser-vations (26) EF-B231 was primarily detected in thenucleolus whereas EF-B232 was detected in both thenucleoplasm and nucleolus EF-B233 was primarilylocalized in the nucleolus and it was also clearlydetected in the nucleoplasm We subsequently examinedthe nucleolar retention of B23 splicing variants usingFRAP assays (Figure 1C) EGFP signal in a small nucle-olus was targeted for bleaching with a 488-nm laser lineand the intensity of EGFP was measured every 05 s Therecovery curves indicated that the mobility of B232 wasconsiderably faster than that of B231 and the half time(t12) of EF-B231 and -B232 recovery was 548plusmn180 sand 233plusmn096 s respectively This finding was consistentwith the previous observation that CTD plays a crucialrole in the nucleolar retention of B23 (2126) Despiteintact CTD the mobility of B233 was similar to that ofB232 (t12=211plusmn102 s)

Localization and chromatin association of B23 areclosely related to its RNA binding activity (2628)Therefore we examined the RNA binding activity ofB23 splicing variants in cells (Figure 1D) EF-taggedB23 proteins were transiently expressed in 293T cellsand immunoprecipitation assays were performed Underthe assay conditions EF-tagged B23 proteins were equallyprecipitated (Figure 1D bottom panels) RNAs co-precipitated with EF-tagged proteins were extracted andanalyzed by denaturing polyacrylamide gel and agarosegel electrophoresis (Supplementary Figure S2A)EF-B231 co-precipitated with 28S 58S and 5S rRNAsas previously reported (28) but not with 18S rRNAwhereas B232 did so inefficiently (lanes 6 and 7) rRNAassociation of EF-B233 was significantly lower than thatof EF-B231 in the presence of 150mM NaCl and washardly detectable in the presence of 300mM NaCl (lanes10 and 12) The 28S 58S and 5S rRNA binding activityof EF-B233 was approximately half compared withthat of EF-B231 in the presence of 150mM NaCl(Supplementary Figure S2B) To investigate the directRNA binding activity we performed sucrose densitygradient centrifugation assays using GST-tagged recom-binant proteins and total RNAs purified from 293T cells(Figure 1E) All three proteins were recovered from thelow-density fractions in the absence of RNA (Figure 1Elanes 2ndash4) In the presence of RNA GST-B231 was pri-marily recovered from the 28S rRNA peak fractions(Figure 1E lanes 6ndash10 and Figure 3C) In contrast frac-tions of GST-B232 did not clearly change regardlessof the presence of RNA Although a minor population

1184 Nucleic Acids Research 2014 Vol 42 No 2

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

of GST-B233 was recovered from the 28S rRNA peakfractions the fractionation pattern of GST-B233 wasnot significantly changed by the presence of RNABecause the fractionation pattern of GST alone was notaffected by the presence of RNA and non-tagged andGST-tagged B231 showed similar fractionation patternin the presence of RNA (Supplementary Figure S3A andB) GST-tagging did not affect the RNA binding activityof B23 proteins These results indicate that not only CTDbut also the region lacking in B233 play important rolesin the RNA binding activity of full-length B231 in cellsand in vitro

Central regions of B23 are intrinsically disordered

Recent studies showed that the IDRs regulate the func-tions of structured regions such as proteinndashprotein orproteinndashnucleic acid binding activity (1) The acidic andbasic regions of B23 have low mean hydrophobicity and

high net charge properties known to be associated withthe disordered status Therefore we hypothesized that thecentral acidic and basic regions are IDRs and contributeto regulate the RNA binding activity of the structuredRNA binding domain CTD Computational predictionof the disordered status by POODLE-I (httpmbscbrcjppoodlepoodle-ihtml) (31) showed that the centralregion of B23 is IDR (Figure 2A) To confirm this predic-tion we performed CD spectrum analyses with His-taggedB231-C3 harboring the N-terminal oligomerizationdomain B231-N and B231-CR1 harboring bIDR(amino acids 1ndash119 120ndash294 and 189ndash259 respectively)(Figure 2B) Consistent with the reported structural datademonstrating that the oligomerization domain of B231consists of b-sheets (32) the spectrum recorded for His-B231-C3 showed a negative peak at 215 nm (greenline) His-B231-CR1 showed a typical random-coilpattern peaking at 195 nm (blue line) as previouslyreported (25) The spectrum recorded for His-B231 -N

Figure 2 Structural analysis of aIDR and bIDR (A) Prediction of disorder tendency of B231 B231 structure was schematically represented at thetop The patterns in the scheme are as defined in the legend of Figure 1A Disorder tendency of B231 was predicted using POODLE-I (BndashD) CDspectral analyses of B23 mutants His-tagged B23 mutant proteins (5 mM) were analyzed His-B231-C (green) His-B231-N (purple) andHis-B231-CR1 (blue) (B) His-B231-N (purple) His-B231-CR (blue) and His-B231-CR15 (light purple) (C) His-B231-N (purple) andHis-B232-N (pink) (D)

Nucleic Acids Research 2014 Vol 42 No 2 1185

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

showed negative peaks at 222 and 200 nm (purple line)Because negative peaks at 222 and 208 nm and thatat 200 nm are indicative of a-helix and random-coilstructure respectively we inferred that His-B231 -Ncontained both a-helix and random-coil structuresConsistent with previous nuclear magnetic resonancestudies (24) the spectrum of B231-CR15 (amino acids242ndash294) peaked at 222 and 208 nm (Figure 2C lightpurple line) When the sequences adjacent to CTD wereelongated (His-B231-CR and His-B231 -N) the peak at208 nm shifted to 200 nm In addition the spectrum ofHis-B232 -N (acidic and basic regions with a shortportion of CTD) showed a negative peak at 200 nm(Figure 2D red purple line) These results indicated thatthe central region of B23 containing acidic and basicregions was disordered in solution To exclude thepossible contribution of His-tag to the B23 protein struc-ture we prepared non-tagged B231 B23-CR1 B231 -Nand B232 -N and compared their spectra with those ofHis-tagged proteins (Supplementary Figure S4) In allcases the overall pattern of the spectra was similaralthough the peak at 200ndash210 nm of B231 B23-CR1B231-N was slightly shifted to the left when His-tagwas added Thus we concluded that His-tag did notsignificantly affect B23 protein structure

bIDR and CTD of B231 associate with RNA

To address the function of bIDR of B23 we ivestigatedthe RNA binding activity of this region Recent studiesindicated that in addition to CTD of B231 a part ofbIDR was required for its stable G-quadruplex DNAbinding activity (24) These results raised the possibilitythat bIDR is also involved in the RNA binding activityof B23 To test this hypothesis GST-tagged B231-CR1and B231-CR15 were purified (Figure 3A and B) andexamined for RNA binding activity using sucrosedensity gradient centrifugation assays In the absence ofRNA GST-B231 and mutant proteins were recovered inlow-density fractions (Figure 3C lanes 1ndash4) When totalRNA was subjected to a sucrose density gradient centri-fugation assay 28S and 18S rRNAs were recovered infractions 6ndash10 and 4ndash9 respectively (Figure 3C righttop panel) GST-B231 preincubated with total RNAswas recovered in the fractions yielding 28S rRNA(Figure 3C right panel) Consistent with this result thefractionation pattern of 28S rRNA but not of 18S rRNAshifted slightly to the bottom (Figure 3D top graphs)suggesting that B231 preferentially associated with 28SrRNA When GST-B231-CR1 was preincubated withRNA its distribution pattern shifted to high-density frac-tions (Figure 3C left panel lanes 5ndash10) In contrast GST-B231-CR15 was not clearly cofractionated with rRNAsConsistent with this result GST pull-down assays showedthat GST-B231-CR1 efficiently precipitated both 18S and28S rRNA but GST-B231-CR15 and GST did so ineffi-ciently (data not shown) These results indicated thatbIDR had potential RNA binding activity and thatCTD alone could not efficiently associate with rRNAsFractionation patterns of both 28S and 18S rRNAsshowed shifts to higher-density fractions when incubated

with the GST-B231-CR1 protein (Figure 3D) Inaddition 28S rRNA preincubated with CR1 shifted tohigher-density fractions than that preincubated withwild-type B231 On the basis of these results we predictedthat bIDR of B231 strongly associates with RNA withlow specificity

B231 was recovered primarily in the 28S rRNA peakfractions but we did not know whether B231 directlybinds 28S rRNA To investigate the same we examinedthe RNAs contained in the 28S peak fractions TotalRNAs were fractionated by sucrose density gradient cen-trifugation assay and the RNAs were separated byagarose gel and polyacrylamide gel electrophoresis(Supplementary Figure S5A) 18S and 28S rRNAs wererecovered in fractions 4ndash9 and 6ndash10 respectively Inaddition 58S rRNA and a small but distinct level of 5SrRNA clearly cofractionated with 28S rRNA To testwhether B231 directly binds to 28S but not to 18SrRNA both 28S and 18S rRNAs were purified 18S and28S rRNAs alone were primarily found in fractions 4ndash6and 7ndash9 respectively We next examined the 28S and 18SrRNA binding activity of B231 by sucrose densitygradient centrifugation assay (Supplementary Figure S5Band C) Surprisingly we found that GST-B231 wascofractionated with both purified 18S and 28S rRNAsHowever when total RNAs were mixed with purified18S rRNA and examined for B231 binding GST-B231recovery in the 18S rRNA peak fractions (fractions 4ndash7)decreased and those in the 28 S rRNA peak fractions(fractions 7ndash10) increased These results suggest thatB231 has some preference for RNA structure orsequence found in RNAs in the 28 S peak fractions suchas 28 S and 58 S rRNAs

Phosphorylation of bIDR contributes to the inactivation ofB231 RNA binding activity

Previously we have demonstrated that the phosphoryl-ation of B23 by cdc2cyclin B kinase reduces its RNAbinding activity (26) The phosphorylation sites threonine199 219 234 and 237 are located in bIDR Consistentwith previous results (28) the phosphomimetic mutant offour threonine residues (T4sD) (Figure 4A and B) showedsignificantly lower RNA binding activity than wild-typeB231 (Figure 4B) Therefore we hypothesized thatmitotic phosphorylation regulated the RNA bindingactivity of bIDR To test this hypothesis B231-CR wasphosphorylated by mitotic extracts prepared fromHeLa cells and the RNA binding activity of phosphory-lated protein was assessed (Figure 4C and D)Phosphorylation status of the protein was examined byMn2+-phos-tag SDSndashPAGE (compare lanes 1 and 2 inFigure 4C) Although the CR protein was not completelyphosphorylated the RNA binding activity ofphosphorylated GST-B231-CR was lower than that ofuntreated GST-B231-CR (Figure 4D) To gain clearerinsight into this point phosphomimetic (CRD) andunphosphorylatable (CRA) mutants of GST-taggedB231-CR in which threonine (T) residues were replacedwith aspartic acid (D) and alanine (A) residues respect-ively were expressed and examined for the RNA binding

1186 Nucleic Acids Research 2014 Vol 42 No 2

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

activity using filter binding assay (Figure 4E and F)Labeled RNA was retained on the membrane whenincubated with GST-B231-CR and GST-B231-CRAmutant proteins However the activity of GST-B231-CRD was significantly lower than that of GST-B231-CR and GST-B231-CRA (30 of GST-B231-CR)The GST pull-down assay also indicated that the 28 Sand 18 S rRNA binding activities of GST-B231-CRD

were 35 and 25 of those of GST-B231-CR(Supplementary Figure S6) CD spectral analyses of thethree mutant proteins showed peaks at 222 and 208 nmthe typical pattern of a-helix (Figure 4G and H) suggest-ing that the phosphomimetic mutation at the phosphoryl-ation sites does not affect the C-terminal helical structureMoreover in the presence of a low concentration of RNAthe CD spectrum of bIDR (CR1 see Figure 3B) was not

Figure 3 The RNA binding activity of bIDR and CTD of B231 (A) Schematic representation of B231 and mutants The patterns in the schemeare as defined in Figure 1A (B) Recombinant GST-tagged B231 and its mutant proteins were analyzed by SDSndashPAGE followed by CBB stainingLane M indicates molecular markers (C) Sucrose density gradient centrifugation assays The RNA binding activities of GST-tagged B231 andmutant proteins were examined by Sucrose gradient centrifugation as in Figure 1E Proteins in fractions were analyzed by SDSndashPAGE and visualizedby CBB staining (left panels) RNA alone (right top panel) and RNAs preincubated with B23 proteins were purified from each fraction separated on1 denatured agarose gels and visualized by Gel Red staining (right panels) (D) Distribution pattern of 28S and 18S rRNAs The amounts of 28S(left panels) and 18S rRNAs (right panels) in each fraction in Figure 3C right panels relative to total 28S and 18S rRNAs were calculated and shownin graphs Dashed lines in each graph indicate the sucrose gradient results of RNAs in the absence of proteins (Figure 3C top right panel)

Nucleic Acids Research 2014 Vol 42 No 2 1187

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

Figure 4 The effect of mitotic phosphorylation on the RNA binding activity of bIDR (A) Schematic representation of B231 andunphosphorylatable or phospho-mimetic CR mutants B231-T4sD is a mutant in which threonine residues shown by arrowheads are replacedwith aspartic acid (D) and B231-CRA and B231-CRD are B231-CR mutants in which the same residues are replaced with alanine (A) and asparticacid (D) respectively (B) RNA binding activity of B231 and phospho-mimetic mutant GST-B231 and GST-B231-T4sD (50 100 200 and 400 ngfor lanes 3ndash6 respectively) were mixed with (lanes 3ndash6) or without (lane 2) 32P-labeled total RNA (10 ng) RNA alone was also examined (lane 1)The intensity of each spot was analyzed by FLA7000 and the RNA binding activity obtained with the same amount of B23 proteins were calculatedrelative to that of B231 (10) The relative RNA binding activity at each protein amount was averaged MeansplusmnSD obtained from twice ofduplicate independent experiments are shown (bottom panel) (C) Phosphorylation of GST-B23-CR GST-B231-CR was incubated with mitoticextracts in the absence (lane 1) or presence (lane 2) of 3mM ATP The incubated proteins were analyzed by phos-tag SDSndashPAGE and visualized byCBB staining (D) The RNA binding activity of GST unphosphorylated GST-B231-CR and phosphorylated GST-B231-CR (pGST-B231-CR) wereexamined for filter binding assays as in (B) The bound RNA for each amounts of GST (dashed line) GST-B231-CR (black line) and GST-B231-CRD (gray line) were plotted and shown in the right panel (E) Recombinant GST-tagged B23 proteins were analyzed by SDSndashPAGE followedby CBB staining Lane M indicates molecular markers (F) RNA binding activity of B231-CR mutants Filter binding assays were carried outas described in (B) using GST-B231-CR -CRA and -CRD The amount of RNA retained on the membrane with GST-B231-CR is set as 10Statistical P-values were calculated by t-tests and indicated with Plt 001 (G) Recombinant His-tagged B231-CR -CRA and -CRD were analyzedby SDSndashPAGE and visualized by CBB staining Lane M indicates molecular markers (H) CD spectral analysis of B231-CR B231-CRA andB231-CRD CD spectra of B231-CR (blue gray) B231-CRA (light purple) and B231-CRD (purple) (5 mM) were analyzed

1188 Nucleic Acids Research 2014 Vol 42 No 2

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

significantly changed (Supplementary Figure S7) Thisfinding suggests that a specific structure of bIDR is notinduced on binding to RNA

bIDR of B23 showed strong RNA binding activitywhereas B232 containing identical amino acid sequencesin this region did not efficiently associate with RNA sug-gesting that the RNA binding activity of bIDR ismodulated by full-length B23 protein Because B23 hasan aIDR it is likely that negative charge of the aIDRneutralizes the positive charge of bIDR To examine thispossibility we studied the effect of aIDR deletion on B23

RNA binding activity using B231 -A2 B231 -A1A2and B231-AC1 (Figure 5A) Sucrose density gradientcentrifugation assays clearly showed that the aIDRmutants associated with RNA (fractions 5ndash10) The frac-tionation pattern of 28S and 18S rRNAs indicated thatthe aIDR mutants associated with both rRNAs Thefraction shifts in fractions of both rRNAs incubatedwith GST-B231-A1A2 were more prominent thanthose of RNAs incubated with the B231-CR1 mutant(see Figure 3C and D) possibly due to the oligomer for-mation of GST-B231-A1A2 In addition both 28S and

Figure 5 aIDR is required for the proper RNA binding of B23 (A) Schematic representation of B231 and deletion mutants (B) Recombinant GST-tagged B23 mutant proteins were analyzed by SDSndashPAGE and visualized by CBB staining Lane M indicates molecular markers (C) Sucrose densitygradient centrifugation assays Total RNAs (10 mg) preincubated in the absence (top panel) or presence of GST-tagged B23 proteins (5 mg) wereloaded on a 15ndash40 sucrose gradient Proteins in fractions collected from the top were analyzed by SDSndashPAGE and visualized by CBB staining (leftpanels) RNAs recovered in each fraction were analyzed as in Figure 3C (right panels) Right top panel is data for RNA alone (D) Distributionpatterns of 28S and 18S rRNAs The intensity of 28 and 18S RNAs in C right panels were scanned and analyzed by CS analyzer (ATTO) andgraphically represented Dashed lines are the sucrose gradient results of RNA alone

Nucleic Acids Research 2014 Vol 42 No 2 1189

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

18S rRNAs incubated with GST-B231-A1A2 wererecovered in higher-density fractions than those incubatedwith GST-B231-A2 (Figure 5C and D)Furthermore GST-B231-AC1 associated with

RNAs as did GST-B231-A1A2 indicating that CTDwas not required for RNA binding when aIDR wasdeleted From these results we concluded that theRNA binding activity of bIDR is negatively regulatedby aIDR and that aIDR confers RNA binding specificityon B231We subsequently tested whether the two regions

associated with each other GST-tagged B23-A (aminoacids 120ndash188) was mixed with increasing amounts ofB231-CR1 B231-CR15 B231-CR B231-CRA orB231-CRD (Figure 6A and B) and the mixtures wereloaded on native PAGE (Figure 6C) GST-B231-Amigrated toward the cathode faster than other mutantsbecause of its negative charge (the theoretical pI ofaIDR is 36) whereas GST-B231-CR1 -CR -CRA and-CRD did not enter the gel (pI 102 100 100 and 97respectively) When GST-B231-A was incubated withincreasing amounts of GST-B231-CR1 -CR or -CRAfree form of GST-B231-A decreased and low-mobilitybands appeared (lanes 8ndash11 16ndash19 and 20ndash23) Incontrast addition of GST-B231-CR15 did not changethe mobility of GST-B231-A (lanes 12ndash15)Furthermore the binding activity of GST-B231-A withGST-B231-CRD was lower than that of GST-B231-CR1 -CR and -CRA (lanes 24ndash27) In addition thebinding activity of GST-B231-CR phosphorylated bymitotic extracts to GST-B231-A was weaker than thatof the unphosphorylated protein (Supplementary FigureS8) These results indicate that aIDR regulates the non-specific RNA binding activity of bIDR through directassociation Moreover phosphorylation in bIDR dimin-ished the RNA binding activity and the interaction withaIDR To assess the association between aIDR and bIDRin full-length B231 we tested the sensitivity of His-taggedB231 and B231-A1A2 toward trypsin (Figure 6D)Digested peptides were detected by anti-His-tag antibodyto map digested sites A series of C-terminal deletionmutants (His-B231-C160 -C188 -C194 -C206 -C224and -C241 and a corresponding set of mutants His-B231-A1A2-C160 -C194 -C206 -C224 and -C241)was used as size markers On treatment with trypsinfull-length B231 proteins decreased whereas fourdigested bands were detected by anti-His tag antibodyThe digested bands d and drsquo were lower than those ofHis-B231-C160 (amino acids 1ndash160) and His-B231-A1A2-C160 suggesting that the bands ddrsquo correspondto the product digested at amino acids 140ndash150 The sizesof bands aarsquo bbrsquo and ccrsquo were similar to those of theC223 C194 and C160 mutants respectively Theintensities of the bands a b and c relative to full-lengthproteins were lower than those of arsquo brsquo and crsquo relative tofull-length His-B231-A1A2 These results stronglysuggest that bIDR and the region 133ndash160 of B231 aresusceptible to trypsin degradation when aIDR is deletedThese results support the idea that aIDR of B231 regu-lates its bIDR by direct binding in its native full-lengthform

Inter- and intra-molecular interaction between IDRsin B23

bIDR of B231 was essential for efficient RNA bindingactivity although the RNA binding activity of the sameamino acid sequences (bIDR) in B232 was suppressedTherefore we hypothesized that the interaction mode ofacidic and bIDRs in B231 was different from that inB232 In addition it was unclear whether the interactionbetween the two IDRs was inter- or intra-molecularTo address these points we tested the interactionbetween the two IDRs of B23 proteins using FlAsHlabeling (Figure 7AndashD) FlAsH a biarsenical dye fluor-esces on binding to the thiol moieties of cysteines (Cys)in the tetracysteine (CysndashCysndashXaandashXaandashCysndashCys) (33) Iftetracysteine is separated into two CysndashCys pairs FlAsHfluoresces only when two CysndashCys pairs are in close prox-imity (6ndash8 A) (3334) For this assay we constructed aseries of B231-N and B232-N mutants containingCysndashCys tags (Figure 7A) One CysndashCys tag wasintroduced at the N-terminal of aIDR (B23-Nc andB232-Nc) and the other was introduced at positions242 and 243 (both serines) located at the border betweenbIDR and CTD (B231-Cc and B232-Cc) The lsquo-NCccrsquoproteins have two CysndashCys tags at both sites B231-Cccand B232-Ccc in which serines at 242 and 243 werereplaced with the known FlAsH binding tetracysteinesequence (CysndashCysndashProndashGlyndashCysndashCys) (35) were usedas positive controls (Figure 7A and B) When increasingamounts of B232-Ccc were mixed with FlAsH fluores-cence was detected in a dose-dependent manner(Figure 7C) Fluorescence was also detected when B231-Ccc was examined However the fluorescence intensity ofB231-Ccc was only 10 of that of B232-Ccc(Plt 00005) This result suggests that CTD of B231 re-stricted the access of FlAsH molecule to the tetracysteinesequence located between bIDR and CTD Whenincreasing amounts of B231-Nc B231-Cc B232-Ncand B232-Cc harboring a single CysndashCys pair weremixed with FlAsH very weak fluorescence was detected(Figure 7D) A mixture of B231-Nc and B231-Cc or ofB232-Nc and B232-Cc also showed weak fluorescence(Figure 7D) In contrast in case of B231-NCcc andB232-NCcc harboring two CysndashCys pairs fluorescentsignals were dependent on protein concentration Theseresults suggest that both B231-N and B232-N bentto form a hairpin-like structure and the two CysndashCyspairs in one molecule were located close to each otherAlthough we could not detect inter-molecular interactionusing FlAsH labeling it was possible that it could not bedetected because of the distance between the two CysndashCyspairs To examine this possibility chemical cross-linkassays were performed (Figure 7E) His-tagged B231-N and B232-N were cross-linked with 01glutaraldehyde and separated using SDSndashPAGEfollowed by CBB staining Both His-B231-N and His-B232-N were found to be monomers at low proteinconcentrations because of the absence of the N-terminaloligomerization domain (Figure 7E lanes 1 and 2) Bandscorresponding to dimers appeared when His-B231-Nprotein concentration was increased whereas the dimer

1190 Nucleic Acids Research 2014 Vol 42 No 2

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

Figure 6 Association between two IDRs (A) Schematic representation of B23 and deletion mutants (B) Recombinant GST-tagged B23mutant proteins were analyzed by SDSndashPAGE and visualized by CBB staining Lane M indicates molecular markers (C) Mobility shift assaysGST-B231-A (05 mg) mixed with increasing amounts of GST-B231-CR1 (lanes 8ndash11) GST-B231-CR15 (lanes 12ndash15) GST-B231-CR (lane 16ndash19)GST-B231-CRA (lanes 20ndash23) and GST-B231-CRD (lanes 24ndash27) (05 10 20 and 40 mg) were incubated for 30min separated by 6 nativePAGE in 05 TBE and visualized by CBB staining (D) Trypsin digestion of B231 and B231 -A1A2 His-B231 (lanes 7ndash11) and His-B231 -A1A2 (lanes 12ndash16) were incubated with trypsin (0 2 8 20 and 40 ng) for 2min at 37C and digested proteins were separated onSDSndashPAGE and detected with western blotting using anti-His-tag antibody As references for the digestion sites a series of C-terminal deletionmutants (lanes 1ndash6 and 17ndash21) were also separated by the same SDSndashPAGE C160 C188 C194 C206 C223 and C241 indicate B231 orB231 -A1A2 proteins containing amino acids 1ndash160 ndash188 ndash194 ndash206 ndash223 and ndash241 respectively Amino acid numbers are from the full-length protein The digested peptides of His-B231 and His-B231-A1A2 are indicated by andashd and arsquondashdrsquo respectively B231 amino acids141ndash294 are shown at the bottom Putative digestion sites (aarsquo bbrsquo and ccrsquo) are mapped by black arrowheads Dashed underline black underlineand double underline indicate aIDR bIDR and CTD respectively

Nucleic Acids Research 2014 Vol 42 No 2 1191

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

Figure 7 Intra- and inter-molecular interaction between two IDRs of B231 and B232 (A) Schematic representation of B23 mutants B231B232-N and their CysndashCys-tagged mutants are schematically represented The patterns in the scheme of B231-N are as defined in Figure 1AN-terminal CysndashCys tag was represented as CC- (CysndashCys-) C-terminal Cys-Cys-tag was introduced by replacement of serines [S] 242 and 243with cysteines [C] and represented as -CC- Tetracysteine-tag (CysndashCysndashProndashGlyndashCysndashCys CCPGCC) was introduced into the same site with -CC-(B) Purified proteins The Cys-Cys-tagged mutant proteins (200 ng) as indicated at the top of each lane were separated by SDSndashPAGE and visualizedwith CBB staining (C) FlAsH labeling of B231-Ccc and B232-Ccc His-tagged B231-Ccc (diamond) or B232-Ccc (square) (11 22 44 66 88and 132 mM) was incubated with 10 mM of FlAsH-EDT2 for 30min at room temperature and the fluorescence intensity was recorded withVarioskan (D) FlAsH labeling of B23 mutants His-tagged B23-Nc (light blue) B23-Cc (pink) B23-NCcc (purple) and mixture of B23-Nc andB23-Cc (light purple line) [12 24 49 73 97 and 144mM (B231) or 11 22 44 66 88 and 132mM (B232)] were examined for FlAsH labelingassay as in (B) (E) Cross-link assay His-tagged B231-N and B232-N (15 30 and 45 mM) (lanes 1 and 2 3 and 4 and 5 and 6) were incubatedin the presence of 01 Glutaraldehyde for 10min separated on SDSndashPAGE and visualized by CBB staining Positions of size markers areindicated at the left of the panel (F) A model for B231-N and B232-N structures The globular structure of CTD is illustrated by circleand stripes and black with white dots bars indicate aIDR and bIDR respectively

1192 Nucleic Acids Research 2014 Vol 42 No 2

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

formation of His-B232-N was inefficient (Figure 7Elanes 5 and 6) This result suggests that inter-molecularinteraction between aIDR and bIDR in B231 occursmore frequently compared with that in B232 Monomerforms of these proteins were likely to form intra-molecularinteractions as shown by FlAsH labeling assays(Figure 7D) Therefore we conclude that aIDR andbIDR of both B231-N and B232-N interact intra-molecularly and that B231-N also has the potentialto form inter-molecular interactions (Figure 7F)

DISCUSSION

In this study we investigated the molecular mechanism bywhich B231 associates with RNA CD spectral analysisrevealed that the central acidic and basic regions of B23were intrinsically disordered (Figure 2) Our resultssuggest that B231 recognizes RNA through two frag-ments bIDR and CTD and that aIDR modulates thenon-specific RNA binding activity of bIDR allowingB23 CTD to recognize and interact with target RNAsIn the absence of CTD the non-specific RNA bindingactivity of bIDR was inhibited by aIDR indicating thatB232 inefficiently associates with RNA This model isbased on three findings (i) bIDR binds directly toRNAs (Figures 3 and 4) (ii) the RNA binding activityof bIDR of B23 is regulated by its aIDR (Figures 5 and6) and (iii) the association mode of the acidic and bIDRsis different between B231 and B232 (Figure 7) The regu-latory functions of IDRs could be biologically importantbecause bIDR deletion (B233) increased and aIDRdeletion (B231-A1A2) decreased cellular mobility ofB23 (Figure 1C and Supplementary Figure S9)

CTD of B231 is known as the DNA and RNA bindingdomain Recent studies have shown that the small grooveformed by the first two a-helices in CTD recognizes thephosphate group of the G-quadruplex DNA (24) Basicamino acids located at the surface of CTD allow electro-static interaction with DNA The affinity of CTD aloneto G-quadruplex DNA was weak (KDgt 1 105M) butit increased on addition of flanking sequences (KD was2 106M) (23) Consistent with this the RNAbinding activity of CTD was low and full bIDR wasrequired for efficient RNA binding by CTD In thecomplex with G-quadruplex DNA bIDR does notdirectly associate with DNA (24) and the role of bIDRis inferred to stabilize the structure of CTD (25) In thisstudy we found that bIDR alone could bind to RNA(Figure 3) and B233 lacking most of bIDR showedsignificantly lower RNA binding activity than B231(Figure 1) These results indicate that bIDR is requiredfor the RNA binding activity of B231 This activity ofbIDR is likely to help CTD to identify and recognizetarget RNA sequences or structures by direct binding toRNA Considering that CTD preferentially binds tostructured RNA or DNA bIDR may stabilize thebinding by interacting with flanking RNADNAsequences The RNA recognition mode of B231 isextremely similar to that used by several transcriptionfactors and the linker histone H1 For example histoneH1 binds to the nucleosome at the DNA entry and exit

sites through its central globular domain (3637) Inhuman cells the absence of the basic C-terminal IDregion of histone H1 significantly decreases the fluorescentrecovery time after photobleaching (187plusmn57 s for thewild-type and 049plusmn011 s for the deletion mutant) (38)It is suggested that the C-terminal ID region confersthe stable nucleosome binding activity on histone H1 byinteracting with the linker DNA region between thenucleosomesHere we show that full-length B231 seems to bind pref-

erentially to 28S rRNA through its CTD However whichsequence or structure in the RNA is recognized by B23 iscurrently unknown It has been reported that B23 binds toG-quadruplex DNAs (39) Because G-rich RNAs alsoform G-quadruplex structures G-rich RNA regionsare possibly those recognized by B23 Both 18S and28S rRNAs have G-rich sequences predicted to formG-quadruplex structures (data not shown) and bothpurified rRNAs are associated with B23 (SupplementaryFigure S5B) Of interest 18S rRNA binding of B23decreased on total RNA addition These results suggestthat the RNA structure or sequence in 18S rRNAtargeted by B23 is disrupted or hindered by other RNAspecies It is also possible that B23 associates with 28SrRNA complexed with other RNA species with muchhigher affinity than that with18S rRNA or 28S rRNAalone It is well established that 28S rRNA is stronglybound to 58S rRNAs by hydrogen bonds (4041) Thusdouble-stranded region of the 28Sndash58S rRNAs may berecognized by B23B232 lacking CTD did not efficiently associate with

RNA (Figure 1) We assume that the RNA bindingactivity of bIDR of B232 was completely inhibited byaIDR possibly because of the hairpin-like structure forma-tion (see Figure 7F) Intra-molecular interaction betweenthese regions may be physically hindered by CTD in full-length B231 and this interaction becomes more dynamicand inter-molecular between adjacent molecules Thishypothesis was supported by the observation that theaccessibility of the FlAsH fluorophore to the site betweenbIDR and CTD was restricted in B231-Ccc (Figure 7C)Previously we have demonstrated that B231ndashB232 andB231ndashNPM3 hetero-oligomers show lower RNA bindingactivity than that by the B231 homo-oligomer (2630)B232 or NPM3 incorporation in the B231 oligomer maydecrease the chance to establish inter-molecular inter-actions between adjacent molecules of B231 possiblyinhibiting bIDR function by aIDR It should also be men-tioned that the stochastic interaction between the two IDRsincreases depending on protein concentration Because B23forms a pentamer or decamer in solution the local concen-tration of IDRs is increased and this may allow inter-molecular interactions in B231 homo-oligomers Giventhat two B23 pentamers form a decamer in a head-to-head fashion (2032) two possible interaction modes canbe considered One is that bIDR associates with aIDR ofa neighboring protein in a pentamer and the other is thatbIDR associates with aIDR in the other pentamer (seeSupplementary Figure S10)B23 is known to be subjected to various post-

translational modifications including phosphorylation

Nucleic Acids Research 2014 Vol 42 No 2 1193

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

acetylation and sumoylation (2642ndash47) Our resultsclearly explained why mitotic phosphorylation by cdc2cyclin B kinase at bIDR decreases the RNA bindingactivity It has been reported that lysines at positions257 and 267 of B231 located at the surface of CTD (24)are acetylated by p300 and that these acetylations inducethe translocation of B231 from the nucleolus to the nu-cleoplasm (46) Because the G-quadruplex DNA andRNA binding activity of B231 strongly affect its nucleolarlocalization (2839) it is likely that acetylation reduces theDNARNA binding activity by neutralizing the positivecharge required for its contact with target sequencesSer125 located in aIDR is known to be phosphorylatedby casein kinase 2 (43) Previous studies havedemonstrated that phosphorylation of Ser125 decreasesthe RNA binding activity (28) and nucleolar retentiontime of B231 (48) We hypothesize that the increasednegative charge in aIDR by phosphorylation augmentsits affinity to bIDR resulting in the decreased DNARNA binding activity of bIDR In summary post-translational modifications in aIDR bIDR and CTD allaffect the DNARNA binding activity of B231Regulation of B231 functions by post-translational modi-fication is mechanistically explained by the inter- or intra-molecular interactions between the two IDRs Moreoverbecause these regulatory regions are IDRs the accessiblesurface area of modification enzymes is believed to belarger than the structured domains thereby rapidlyregulating the functions of B231 Thus these modifica-tions contribute to fine-tuning the B231 functions tocontrol proper cell growth and proliferationB23 associates with various proteins in cells such as

p14ARF (49) histones (2750) and p53 (5152) throughits N-terminal domain aIDR and bIDR respectivelyBecause the RNA binding activity of B231 is regulatedby the association between acidic and bIDRs it ispresumed that when bIDR of B231 is occupied byRNA or aIDR is occupied by histones the associationof the two IDRs is disturbed These interactions withother molecules may reciprocally increase the affinity ofthese regions for target molecules Therefore it is likelythat B231 tethered by RNA molecules on chromatinaround the rRNA gene (28) functions as an activehistone chaperone

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online

ACKNOWLEDGEMENTS

The authors thank Drs Takashi Minowa and Xianglan Li(National Institute for Material Sciences Japan) for theirhelp to use CD spectroscopy

FUNDING

Grant-in-Aid for Scientific Research from the Ministry ofEducation Culture Sports Science and TechnologyJapan [grant numbers 22570132 and 21113005] (to MO

and KN) and Takeda Foundation NIMS Molecule ampMaterial Synthesis Platform in lsquolsquoNanotechnologyPlatform Projectrsquorsquo operated by the MEXT JapanFunding for open access charge Grant-in-Aid forScientific Research from the Ministry of EducationCulture Sports Science and Technology Japan[21113005 to MO]

Conflict of interest statement None declared

REFERENCES

1 TompaP (2002) Intrinsically unstructured proteins TrendsBiochem Sci 27 527ndash533

2 WellsM TidowH RutherfordTJ MarkwickP JensenMRMylonasE SvergunDI BlackledgeM and FershtAR (2008)Structure of tumor suppressor p53 and its intrinsically disorderedN-terminal transactivation domain Proc Natl Acad Sci USA105 5762ndash5767

3 GuoX BulykML and HarteminkAJ (2012) Intrinsic disorderwithin and flanking the DNA-binding domains of humantranscription factors Biocomputing 104ndash115

4 DunkerAK BrownCJ LawsonJD IakouchevaLM andObradovicZ (2002) Intrinsic disorder and protein functionBiochemistry 41 6573ndash6582

5 MuranoK OkuwakiM HisaokaM and NagataK (2008)Transcription regulation of the rRNA gene by a multifunctionalnucleolar protein B23nucleophosmin through its histonechaperone activity Mol Cell Biol 28 3114ndash3126

6 SavkurRS and OlsonMO (1998) Preferential cleavage inpre-ribosomal RNA byprotein B23 endoribonucleaseNucleic Acids Res 26 4508ndash4515

7 MaggiLB Jr KuchenruetherM DadeyDY SchwopeRMGrisendiS TownsendRR PandolfiPP and WeberJD (2008)Nucleophosmin serves as a rate-limiting nuclear export chaperonefor the Mammalian ribosome Mol Cell Biol 28 7050ndash7065

8 WangW BudhuA ForguesM and WangXW (2005)Temporal and spatial control of nucleophosmin by the Ran-Crm1complex in centrosome duplication Nat Cell Biol 7 823ndash830

9 GrisendiS BernardiR RossiM ChengK KhandkerLManovaK and PandolfiPP (2005) Role of nucleophosmin inembryonic development and tumorigenesis Nature 437 147ndash153

10 GrisendiS MecucciC FaliniB and PandolfiPP (2006)Nucleophosmin and cancer Nat Rev Cancer 6 493ndash505

11 FaliniB MecucciC TiacciE AlcalayM RosatiRPasqualucciL La StarzaR DiverioD ColomboE SantucciAet al (2005) Cytoplasmic nucleophosmin in acute myelogenousleukemia with a normal karyotype N Engl J Med 352254ndash266

12 TanakaM SasakiH KinoI SugimuraT and TeradaM(1992) Genes preferentially expressed in embryo stomach arepredominantly expressed in gastric cancer Cancer Res 523372ndash3377

13 NozawaY Van BelzenN Van der MadeAC DinjensWNand BosmanFT (1996) Expression of nucleophosminB23 innormal and neoplastic colorectal mucosa J Pathol 178 48ndash52

14 ShieldsLB Gercel-TaylorC YasharCM WanTCKatsanisWA SpinnatoJA and TaylorDD (1997) Inductionof immune responses to ovarian tumor antigens by multiparityJ Soc Gynecol Investig 4 298ndash304

15 SubongEN ShueMJ EpsteinJI BriggmanJV ChanPKand PartinAW (1999) Monoclonal antibody to prostate cancernuclear matrix protein (PRO4-216) recognizes nucleophosminB23 Prostate 39 298ndash304

16 TsuiKH ChengAJ ChangPL PanTL and YungBY(2004) Association of nucleophosminB23 mRNA expression withclinical outcome in patients with bladder carcinoma Urology 64839ndash844

17 SkaarTC PrasadSC ShararehS LippmanME BrunnerNand ClarkeR (1998) Two-dimensional gel electrophoresisanalyses identify nucleophosmin as an estrogen regulated protein

1194 Nucleic Acids Research 2014 Vol 42 No 2

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from

associated with acquired estrogen-independence in human breastcancer cells J Steroid Biochem Mol Biol 67 391ndash402

18 OkuwakiM IwamatsuA TsujimotoM and NagataK (2001)Identification of nucleophosminB23 an acidic nucleolar proteinas a stimulatory factor for in vitro replication of adenovirusDNA complexed with viral basic core proteins J Mol Biol 31141ndash55

19 StrausbergRL FeingoldEA GrouseLH DergeJGKlausnerRD CollinsFS WagnerL ShenmenCMSchulerGD AltschulSF et al (2002) Generation and initialanalysis of more than 15000 full-length human and mouse cDNAsequences Proc Natl Acad Sci USA 99 16899ndash16903

20 NamboodiriVM AkeyIV Schmidt-ZachmannMS HeadJFand AkeyCW (2004) The structure and function of XenopusNO38-core a histone chaperone in the nucleolus Structure 122149ndash2160

21 GrummittCG TownsleyFM JohnsonCM WarrenAJ andBycroftM (2008) Structural consequences of nucleophosminmutations in acute myeloid leukemia J Biol Chem 28323326ndash23332

22 WangD BaumannA SzebeniA and OlsonMO (1994) Thenucleic acid binding activity of nucleolar protein B231 resides inits carboxyl-terminal end J Biol Chem 269 30994ndash30998

23 FedericiL ArcovitoA ScaglioneGL ScaloniF Lo SterzoCDi MatteoA FaliniB GiardinaB and BrunoriM (2010)Nucleophosmin C-terminal leukemia-associated domain interactswith G-rich quadruplex forming DNA J Biol Chem 28537138ndash37149

24 GalloA Lo SterzoC MoriM Di MatteoA BertiniIBanciL BrunoriM and FedericiL (2012) Structure ofnucleophosmin DNA-binding domain and analysis of its complexwith a G-quadruplex sequence from the c-MYC promoterJ Biol Chem 287 26539ndash26548

25 MarascoD RuggieroA VascottoC PolettoMScognamiglioPL TellG and VitaglianoL (2012) Role of mutualinteractions in the chemical and thermal stability of nucleophosminNPM1 domains Biochem Biophys Res Commun 430 523ndash528

26 OkuwakiM TsujimotoM and NagataK (2002) The RNAbinding activity of a ribosome biogenesis factor nucleophosminB23 is modulated by phosphorylation with a cell cycle-dependentkinase and by association with its subtype Mol Biol Cell 132016ndash2030

27 OkuwakiM MatsumotoK TsujimotoM and NagataK (2001)Function of nucleophosminB23 a nucleolar acidic protein as ahistone chaperone FEBS Lett 506 272ndash276

28 HisaokaM UeshimaS MuranoK NagataK and OkuwakiM(2010) Regulation of nucleolar chromatin by B23nucleophosminjointly depends upon its RNA binding activity and transcriptionfactor UBF Mol Cell Biol 30 4952ndash4964

29 KinoshitaE Kinoshita-KikutaE TakiyamaK and KoikeT(2006) Phosphate-binding tag a new tool to visualizephosphorylated proteins Mol Cell Proteomics 5 749ndash757

30 OkuwakiM SumiA HisaokaM Saotome-NakamuraAAkashiS NishimuraY and NagataK (2012) Function ofhomo- and hetero-oligomers of human nucleoplasminnucleophosmin family proteins NPM1 NPM2 and NPM3 duringsperm chromatin remodeling Nucleic Acids Res 40 4861ndash4878

31 HiroseS ShimizuK and NoguchiT (2010) POODLE-Idisordered region prediction by integrating POODLE series andstructural information predictors based on a workflow approachIn Silico Biol 10 185ndash191

32 LeeHH KimHS KangJY LeeBI HaJY YoonHJLimSO JungG and SuhSW (2007) Crystal structure ofhuman nucleophosmin-core reveals plasticity of the pentamer-pentamer interface Proteins 69 672ndash678

33 GriffinBA AdamsSR and TsienRY (1998) Specific covalentlabeling of recombinant protein molecules inside live cellsScience 281 269ndash272

34 LuedtkeNW DexterRJ FriedDB and SchepartzA (2007)Surveying polypeptide and protein domain conformation andassociation with FlAsH and ReAsH Nat Chem Biol 3779ndash784

35 AdamsSR CampbellRE GrossLA MartinBRWalkupGK YaoY LlopisJ and TsienRY (2002) Newbiarsenical ligands and tetracysteine motifs for protein labelingin vitro and in vivo synthesis and biological applications J AmChem Soc 124 6063ndash6076

36 AllanJ HartmanPG Crane-RobinsonC and AvilesFX(1980) The structure of histone H1 and its location in chromatinNature 288 675ndash679

37 SimpsonRT (1978) Structure of the chromatosome a chromatinparticle containing 160 base pairs of DNA and all the histonesBiochemistry 17 5524ndash5531

38 LeverMA ThrsquongJP SunX and HendzelMJ (2000) Rapidexchange of histone H11 on chromatin in living human cellsNature 408 873ndash876

39 ChiarellaS De ColaA ScaglioneGL CarlettiE GrazianoVBarcaroliD Lo SterzoC Di MatteoA Di IlioC FaliniBet al (2013) Nucleophosmin mutations alter its nucleolarlocalization by impairing G-quadruplex binding at ribosomalDNA Nucleic Acids Res 41 3228ndash3239

40 PeneJJ KnightE Jr and DarnellJE Jr (1968) Characterizationof a new low molecular weight RNA in HeLa cell ribosomesJ Mol Biol 33 609ndash623

41 PaceNR WalkerTA and SchroederE (1977) Structure of the58S RNA component of the 58S-28S ribosomal RNA junctioncomplex Biochemistry 16 5321ndash5328

42 PeterM NakagawaJ DoreeM LabbeJC and NiggEA(1990) Identification of major nucleolar proteins as candidatemitotic substrates of cdc2 kinase Cell 60 791ndash801

43 ChanPK AldrichM CookRG and BuschH (1986) Aminoacid sequence of protein B23 phosphorylation site J Biol Chem261 1868ndash1872

44 KrauseA and HoffmannI (2010) Polo-like kinase 2-dependentphosphorylation of NPMB23 on serine 4 triggers centrioleduplication PLoS One 5 e9849

45 KangYJ OlsonMO JonesC and BuschH (1975) Nucleolarphosphoproteins of normal rat liver and Novikoff hepatomaascites cells Cancer Res 35 1470ndash1475

46 ShandilyaJ SwaminathanV GadadSS ChoudhariRKodaganurGS and KunduTK (2009) Acetylated NPM1localizes in the nucleoplasm and regulates transcriptionalactivation of genes implicated in oral cancer manifestationMol Cell Biol 29 5115ndash5127

47 LiuX LiuZ JangSW MaZ ShinmuraK KangSDongS ChenJ FukasawaK and YeK (2007) Sumoylation ofnucleophosminB23 regulates its subcellular localization mediatingcell proliferation and survival Proc Natl Acad Sci USA 1049679ndash9684

48 NegiSS and OlsonMO (2006) Effects of interphase and mitoticphosphorylation on the mobility and location of nucleolar proteinB23 J Cell Sci 119 3676ndash3685

49 ItahanaK BhatKP JinA ItahanaY HawkeDKobayashiR and ZhangY (2003) Tumor suppressor ARFdegrades B23 a nucleolar protein involved in ribosome biogenesisand cell proliferation Mol Cell 12 1151ndash1164

50 GadadSS SenapatiP SyedSH RajanRE ShandilyaJSwaminathanV ChatterjeeS ColomboE DimitrovSPelicciPG et al (2011) The multifunctional proteinnucleophosmin (NPM1) is a human linker histone H1 chaperoneBiochemistry 50 2780ndash2789

51 LambertB and BuckleM (2006) Characterisation of theinterface between nucleophosmin (NPM) and p53 potential rolein p53 stabilisation FEBS Lett 580 345ndash350

52 ColomboE MarineJC DanoviD FaliniB and PelicciPG(2002) Nucleophosmin regulates the stability and transcriptionalactivity of p53 Nat Cell Biol 4 529ndash533

Nucleic Acids Research 2014 Vol 42 No 2 1195

at UN

IV O

F TSU

KU

BA

on April 2 2014

httpnaroxfordjournalsorgD

ownloaded from