Interphase-specific association of intrinsic centromere protein … · 2001-05-03 · polymerase-treated pEG202 (Gyuris et al., 1993), which was also digested with BamHI. The correct

2029Journal of Cell Science 111, 2029-2041 (1998)Printed in Great Britain © The Company of Biologists Limited 1998JCS7256

Interphase-specific association of intrinsic centromere protein CENP-C with

HDaxx, a death domain-binding protein implicated in Fas-mediated cell death

Ann F. Pluta* ,**, William C. Earnshaw ‡ and Ilya G. Goldberg §

Department of Cell Biology and Anatomy, Johns Hopkins University School of Medicine, 725 N. Wolfe Street, Baltimore, MD21205, USA*Present address: Department of Molecular Biology and Biophysics, Medical Biotechnology Center, University of Maryland Biotechnology Institute, 725 West LombardStreet, Baltimore, MD 21201, USA**Author for correspondence (e-mail: [email protected])‡Present address: Institute of Cell and Molecular Biology, University of Edinburgh, Swann Building, Mayfield Road, Edinburgh EH9 3JR, UK§Present address: Department of Molecular and Cellular Biology, Howard Hughes Medical Institute, Harvard University, 7 Divinity Avenue, Cambridge, MA 02138, USA

Accepted 4 May; published on WWW 30 June 1998

CENP-C, one of the few known intrinsic proteins of thehuman centromere, is thought to play structural as well asregulatory roles crucial to proper chromosome segregationand mitotic progression. To further define the functions ofCENP-C throughout the cell cycle we have used the yeastinteraction trap to identify proteins with which it interacts.One specific CENP-C interactor, which we have namedHDaxx, was characterized in detail and found to behomologous to murine Daxx, a protein identified throughits ability to bind the death domain of Fas (CD95). Theinteraction between CENP-C and HDaxx is mediated bythe amino-terminal 315 amino acids of CENP-C and the

carboxyl-terminal 104 amino acids of HDaxx. This regionof Daxx is responsible for binding to death domains ofseveral apoptosis signalling proteins. The biologicalsignificance of the interaction between CENP-C andHDaxx was confirmed by immunofluorescencecolocalization of these two proteins at discrete spots in thenuclei of some interphase HeLa cells. We discuss thefunctional implications of the interphase-restrictedassociation of HDaxx with centromeres.

Key words: Interphase centromere, CENP-C, HDaxx

SUMMARY

f.lyrs the;ddetchintd

dn

1),d3;

s,ainsl.,se

INTRODUCTION

The centromere is the chromosomal region responsible forprecise and accurate segregation of eukaryotic genetic matduring mitosis and meiosis (reviewed by Pluta et al., 199Centromeres direct the segregation of mitotic chromosomthrough a differentiated multilayered structure called tkinetochore, which serves as the binding site for spindmicrotubules at prometaphase and for the mechanochemmotors that move chromosomes along those microtubuduring anaphase. It is becoming increasingly clear thkinetochores also act as cellular sensors that monitor prorequirements necessary for progression through the cell cyAs the last region of replicated chromosomes to remain paiprior to disjunction, centromeres both monitor the bipolattachment of chromosomes to the spindle and regulateseparation of the sister chromatids at the metaphase-anaptransition. Centromeres defective in any of these mitofunctions can result in aneuploidy, a condition that accoufor a high percentage of early embryonic lethality in human

A complete understanding of how the centromecoordinates these roles requires comprehensive knowledgits biochemical associations at all stages of the cell cycle,

theerial5).es

heleicallesatpercle.redar thehaseticntss. ree ofyet

the list of known human centromere proteins (CENPs) is brieMany CENPs associate transiently with centromeres onduring mitosis, and as the cell exits mitosis and enteinterphase these proteins are degraded or transferred fromkinetochore to the mitotic spindle (Brinkley et al., 1992Allshire, 1997). Some, such as cytoplasmic dynein ankinesin-related CENP-E, supply the microtubule-binding anmotor functions carried out by the kinetochore (Wordeman al., 1991; Wood et al., 1997; Schaar et al., 1997). Others, suas the 3F3/2 antigens and the spindle assembly checkpoproteins MAD2 and BUB1 are involved in tension sensing ancell cycle signaling (Campbell and Gorbsky, 1995; Li anBenezra, 1996; Taylor and McKeon, 1997). Less is knowabout the functions of others, including the INCENPchromosomal passengers (Earnshaw and Cooke, 199kinesin-related MCAK (Wordeman and Mitchison, 1995) anCENP-F/mitosin (Casiano et al., 1993; Rattner et al., 199Zhu et al., 1995).

Despite the absence of the mitotic-specific CENPcondensed centromeres persist as discrete chromatin domthroughout the cell cycle (Moroi et al., 1981; Matsumoto et a1989b; Cooke et al., 1990). Nevertheless the interphacentromere, unlike its mitotic counterpart, is functionally ill-

2030

lso

eceith

1--

d

s

-

in,

as

dd

s-e

n-

ona-nd

Cggrng

eit

as

A. F. Pluta and others

defined partly because so few of the proteins that comprisare known. CENPs-A, -B and -C, initially identified aautoantigens in patients with scleroderma spectrum dise(Moroi et al., 1980; Brenner et al., 1981), are the only humproteins known to remain associated with centromeric DNAall stages of the cell cycle (Cooke et al., 1990; Sullivan et 1994; Knehr et al., 1996). While the relationship between timmunoreactivity and the etiology of systemic sclerosis unclear, human autoimmune sera containing anti-centromantibodies (ACA) were used to clone and characterize thintrinsic CENPs. CENP-A, a novel centromere-specific cohistone related to histone H3, was recently immunolocalizto the inner kinetochore plate (Sullivan et al., 1994; Warburtet al., 1997). CENP-B is a sequence-specific α-satellite DNAbinding protein localized throughout the centromerheterochromatin located beneath the kinetochore (Cooke e1990; Earnshaw et al., 1987; Matsumoto et al., 1989a).functional significance for centromeres remains unclear siit is undetectable on human Y chromosomes (Earnshaw et1991; Matsumoto et al., 1993) and it was recently reported tthe mouse CENP-B gene is not essential for viability (Kapoet al., 1997). CENP-C is a structural component of the inkinetochore plate (Saitoh et al., 1992). All the intrinsic humCENPs display homology with yeast proteins involved centromere biology, indicating that their functions aimportant and likely to be conserved between species (Bro1995; Meluh and Koshland, 1995; Stoler et al., 1995; Leeal., 1997; Halverson et al., 1997). Howeveimmunofluorescence studies of human dicentric chromosomwhich are mitotically stable because they are functionamonocentric, have revealed that CENPs-A and -C are fouonly at active centromeres, while CENP-B is detectable at bthe active and inactive centromeres (Earnshaw et al., 19Page et al., 1995; Sullivan and Schwartz, 1995). The fact CENPs-A and -C are present throughout the cell cycle, their presence specifically correlates with mitotic centromeactivity indicates that they may participate in the very earliesignaling and/or marking events in interphase that eventudetermine mitotic centromere assembly and function. Yeremains unclear whether the interphase centromere reliesthese proteins merely to mark the chromosomal region destto assemble the kinetochore at mitosis or whether they hother unsuspected functions specific to interphase.

In order to probe centromere function throughout the ccycle, we have used the yeast interaction trap to searchproteins that interact with CENP-C. We have identified acharacterized an unanticipated interaction of CENP-C wHDaxx, a human protein whose mouse homolog has bproposed to modulate Fas-mediated apoptosis. Our finding HDaxx colocalizes with interphase centromeres suggests centromeres may play a role in regulating cellular responseapoptotic stimuli.

MATERIALS AND METHODS

DNA constructs used in the yeast interaction trap pTCATG (Pluta and Earnshaw, 1996) was digested with NdeI andtreated with T4 DNA polymerase to produce blunt ends, then digeswith BamHI, and the resulting insert containing the full-length CENPC open reading frame was cloned into EcoRI-digested, T4 DNA

e itsasean atal.,hisisereesereedon

ict al., Itsnce al.,hator

neraninrewn, etr,es,llynd

oth89;

thatbutrest

allyt it on

inedave

ell fornditheenthatthats to

ted-

polymerase-treated pEG202 (Gyuris et al., 1993), which was adigested with BamHI. The correct reading frame of the resultingconstruct, LEXA:CENP-C1-943 (in which the numbers refer to thamino acid residues of CENP-C), was confirmed by DNA sequenanalysis. The ~945 bp fragment produced by digesting pTCATG wNdeI, followed by treatment with T4 DNA polymerase and then BglII,was cloned into the same vector as above to create LEXA:CENP-C315. The carboxyl-terminal third of the CENP-C ORF was PCRamplified from pTCATG using Vent DNA Polymerase (New EnglanBiolabs) and oligonucleotide primers CENPC14 (5′-GG-GAATTCAGATCTACAAGAAGCTC-3′) and CENPC2831-2816 (5′-GGGAATTCCATCTTTTTATCTGAG-3′). The resulting fragmentwas gel-purified, digested with EcoRI (sites underlined) and clonedinto the EcoRI site of pEG202 to create LEXA:CENP-C635-943.Oligonucleotide primers CENPC12 (5′GGGAATTCAGTGGA-TCCTGGATTAC-3′) and CENPC15 (5′-CATTTCTCGAGTT-CATTCTTTGAGCTTC-3′) were used to PCR-amplify the middlethird of the CENP-C ORF from pTCATG. The resulting fragment waEcoRI- and XhoI-digested, gel-purified and cloned into EcoRI- andXhoI-digested pEG202 to create LEXA:CENP-C316-643. DNAconstructs expressing LEXA:CENP-C180-468 and LEXA:CENPC462-802 were provided by J. Tomkiel (Wayne State University).

To express HDaxx in yeast as an activation domain-tagged protethe full HDaxx ORF was PCR-amplified from pBS/D31-1 usingoligonucleotide primers CCBP22 (5′-GCCCGAATTCCCTATGGC-CACCGCTAAC-3′) and T7 (5′-GTAATACGACTCACTATAGGGC-3′), and the resulting 2.4 kb fragment was digested with EcoRI andSalI (which cuts in the library adaptor sequence at the 3′ end of theHDaxx ORF), gel-purified and cloned into EcoRI-and XhoI-digestedpJG4-5 (Gyuris et al., 1993) to produce pJG4-5/HDaxx1-740.

Yeast interaction trap screen and quantitative β-galactosidase assays The yeast interaction trap screen was performed essentially described (Golemis et al., 1996). ~20 µg of a poly(dT)-primed HeLacell cDNA library in pJG4-5 (Gyuris et al., 1993) was transformeinto yeast strain EGY48 which contained LEXA:CENP-C1-943 anpSH18-34, yielding ~5×106 primary transformants which werepooled, washed in sterile dH2O, resuspended in sterile 65% (v/v)glycerol, 0.1 M MgSO4, 25 mM Tris-HCl, pH 7.4, and frozen in smallaliquots at −70°C. ~3×107 cfu containing LEXA:CENP-C1-943,pSH18-34 and library plasmids were plated on Gal/Raf/CM-ura-hitrp-leu plates to screen for transcriptional activation of th3LexAop:Leu2gene engineered into the yeast strain. 156 Leu+ yeastcolonies were picked to a Glu/CM-ura-his-trp master plate, thereplica-plated to Glu/CM-ura-his-trp-leu, Gal/Raf/CM-ura-his-trpleu, Glu/Xgal/CM-ura-his-trp and Gal/Raf/Xgal/CM-ura-his-trpplates to test for galactose-dependent leu2 and lacZ expression. 80Leu+ colonies displayed the interaction phenotype when plated these media (growth on Gal/Raf-ura-his-trp-leu but not on Glu-urhis-trp-leu, and blue colony color on Gal/Raf/Xgal, but white oGlu/Xgal); of those, 32 were randomly selected for further study an23 were subsequently recovered as library plasmids in Escherichiacoli strain JBe15. Specificity of the interaction between the CENP-bait and the individual library plasmids was verified by transformineach isolated library plasmid into yeast strain EGY48 containinreporter plasmid pSH18-34 and either LEXA:CENP-C1-943 opRFHM1 (Golemis and Brent, 1992; Zervos et al., 1993), and testifor colony color on Gal/Raf/Xgal/CM-ura-his-trp plates and growthon Gal/Raf/CM-ura-his-trp-leu plates. Library plasmids giving thphenotype expected for specific interaction with the CENP-C bawere used to transform E. coli strain DH5α, from which they wereisolated for further study. β-Galactosidase activity of yeasttransformants (3-5 isolates of each strain) grown in liquid culture wdetermined using ONPG (o-nitrophenyl β-D-galactopyranoside) assubstrate (Reynolds and Lundblad, 1989).

2031CENP-C binding protein, HDaxx

aese

sedts

).

-

by

-

-rch).nndly-C

l.,sted-A,-

Sys

ose-ttrent

he

n.

on223

nthen

to

HeLa cDNA library screen and northern blot analysisThe ~1 kb insert from library plasmid pJG4-5/CBP1 was PCamplified using oligonucleotide primers BCO2 (5′-GAGAAGCCGACAACCTTGATTGGAG-3′) and 5′ATGJG4-5 (5′-TGATGCTGAGTGGAGATGCCTCC-3′), gel-purified and 32P-labeledby random priming (Feinberg and Vogelstein, 1983) for use as proin cDNA library screening and northern blot analysis. cDNA clonencoding the HDaxx message were isolated by screening ~1.6×105

pfu from an oligo(dT) + random-primed HeLa 5′-Stretch Plus cDNAλgt11 library (Clontech). Plaque lifts and hybridizations weperformed according to the manufacturer’s instructions. cDNA insewere excised with EcoRI from phage DNA prepared from plaquepurified positives and subcloned into the EcoRI site of Bluescript inboth orientations to produce pBS/D31-1 and pBS/D31-2. DNsequence analysis of this insert revealed that the bona fide 5′ end ofthe phage cDNA insert had not been recovered as part of this EcoRIfragment, and it was subsequently recovered by subcloning the ~bp fragment resulting from digestion of D31 phage DNA with NotI(which cuts in the library adaptor sequence) and PstI into NotI- andPstI-digested Bluescript.

Total RNA was isolated from HeLa (S3) cells grown in spinnculture (Chomczynski, 1996), electrophoresed on a 1formaldehyde-agarose gel, transferred in 20× SSC to a nitrocellulosefilter and hybridized as above. All filters were washed once fominutes at 25°C in 2× SSC, 0.1% SDS, then twice for 5 minutes 25°C in 0.2× SSC, 0.1% SDS, and twice for 15 minutes at 50°C0.1× SSC, 0.1% SDS.

DNA sequence determination and analysisDouble-stranded DNA was sequenced manually on both strands uSequenase Version 2.0 (US Biochemicals) and SequiTherm EXC(Epicentre Technologies) DNA Sequencing kits and [α-35S]dATP.DNA and protein sequences were analyzed using MacVector 6.0AssemblyLIGN 1.0.5. Homology searches of the NCBI noredundant protein and nucleotide databases were performed usinBLAST program (Altschul et al., 1990).

Bacterial fusion protein expression and antibodyproductionInserts from several members of the largest group of library plasmrecovered in the two-hybrid screen were PCR-amplified usoligonucleotide primers BCO2 and 5′ATGJG4-5 phosphorylated withT4 polynucleotide kinase and cloned into the NdeI site of pET16b(Novagen) which was treated with T4 DNA polymerase and cintestine alkaline phosphatase (Boehringer Mannheim). Corrorientation of inserts with respect to the T7 promoter was confirmby PCR amplification using oligonucleotide primers T7 and BCOHistidine-tagged fusion protein expression was induced in E. colistrain BL21(DE3) at 37°C for 2 hours with 1 mM IPTG, after whicthe cells were collected, boiled in SDS, then sonicated to produlysate from which fusion proteins were affinity-purified on His-Binresin (Novagen) according to the manufacturer’s instructions. ~1 of each purified protein was dialyzed against 10 mM Tris-HCl, p7.5, 100 mM NaCl, 0.2% SDS, then precipitated on ice with 20TCA. Following centrifugation, the pellets were washed with 90acetone, 10% 0.1 N HCl and air-dried. Pellets were resuspendePBS, pooled and used as immunogen by Covance Research Pro(Denver, PA) to generate a rabbit polyclonal antibody against amacids 510-740 of HDaxx (anti-HDaxx).

Cell culture, subcellular fractionation and western blottingSuspension cultures of HeLa (S3) were grown in RPMI 1640 with fetal calf serum; adherent HeLa (JW) and U2OS (humosteosarcoma) cells were grown in monolayer cultures on gcoverslips in DMEM with 10% fetal bovine serum. Subcellulafractionation and immunoblotting were carried out as describedEarnshaw and Rattner (1991) with the following modification

R-

beses

rerts

-

A

800

er%

r 5at in

singEL

andn-g the

idsing

alfected2.

hce admgH%%d inductsino

5%anlassr

bys.

Partially purified HeLa chromosomes were prepared from a HeLculture that had been blocked in mitosis with 6 nM vinblastin(Sigma). HeLa nuclei were prepared using an unblocked log phaculture: the pellet resulting from the first spin following Dounce lysiconsisted of a crude nuclear fraction, while the supernatant containthe crude cytoplasmic fraction. Immunodetection on western blowas carried out using either 125I-labeled Protein A or horseradishperoxidase-conjugated anti-rabbit IgG and ECL (Amersham Corp.

GFP fusion constructs, transfections andimmunofluorescenceThe first 573 codons of HDaxx were PCR-amplified from pBS/D311 using primers CCBPBNATG and CCBP11 (5′-ACAGAGGAAGGGGTATC-3′), digested with BamHI and HindIII,and cloned into BglII- and HindIII-digested pEGFP-C1 (Clontech) togenerate pGFP:HDaxx1-573. pGFP:HDaxx1-740 was constructed replacing the small HindIII-EcoRI fragment of pGFP:HDaxx1-573with the 700 bp HindIII-EcoRI fragment of pBS/D31-1. The 680 bpHindIII-BamHI fragment from pBS/D31-1 was subcloned intoHindIII- and BamHI-digested pEGFP-C1 to make pGFP:HDaxx573740.

GFP constructs were introduced into HeLa or U2OS cells by lipidmediated transfection using Lipofectin reagent (Bethesda ReseaLabs; Pluta et al., 1992) or by electroporation (Mackay et al., 1993Indirect immunofluorescence detection of HDaxx and humacentromere proteins was performed on cells grown on coverslips athen fixed in 3% paraformaldehyde for 5 minutes as previousdescribed (Pluta et al., 1992; Earnshaw and Rattner, 1991). CENPwas detected with polyclonal rabbit serum anti-24-623 (Tomkiel et a1994) followed by biotinylated goat anti-rabbit IgG followed by TexaRed conjugated to streptavidin; human centromeres were detecusing human autoimmune serum GS, which recognizes CENPs -B and -C (Earnshaw and Rothfield, 1985), followed by FITCconjugated anti-human IgG; DNA was stained with 4′,6-diaminidino-2-phenyl-indole (DAPI). Slides were examined on a Leica DM IRBmicroscope and images were collected using a Photometric Sen camera driven by IP Labs image processing program.

RESULTS

Isolation of HDaxx The yeast interaction trap assay was used to screen a galactinducible HeLa cDNA expression library for proteins thainteract with CENP-C (Gyuris et al., 1993). Primary yeastransformants containing LEXA:CENP-C1-943, lacZ reporteplasmid pSH18-34 and HeLa cDNA library plasmids werplated on selective medium for galactose-dependetranscriptional activation of the LexAop-LEU2reporter geneengineered into the genome of yeast strain EGY48. 156 of tresulting Leu+ yeast colonies were tested for their ability toinduce galactose-dependent lacZ expression from the LexAop-lacZ reporter gene contained on pSH18-34 in this straiApproximately 50% of the Leu+ transformants displayed theexpected phenotype for interaction when plated on selectimedium (see Materials and Methods). Of those, 3transformants were randomly selected for further study and were subsequently recovered as plasmids in E. coli. 18 of theisolated library plasmids reproduced the specific interactiophenotype when retransformed into yeast co-expressing original CENP-C bait, but not in yeast co-expressing airrelevant bait, LEXA:Drosophilabicoid, thus fitting the initialcriteria for CENP-C interactors.

The 18 putative CENP-C interactors were categorized in

2032 A. F. Pluta and others

R T R G S R R Q I Q R L E Q L L A L Y V A E I R R L Q E K E L D L S E L D D 217

CCA GAC TCC GCA TAC CTG CAG GAG GCA CGG TTG AAG CGT AAG CTG ATC CGC CTC TTT GGG CGA CTA TGT GAG CTG AAA GAC TGC TCT TCA CTG ACC GGC CGT GTC ATA GAG CAG CGC ATC CCC 889 P D S A Y L Q E A R L K R K L I R L F G R L C E L K D C S S L T G R V I E Q R I P 258

TAC CGT GGC ACC CGC TAC CCA GAG GTT AAC AGG CGC ATT GAG CGG CTC ATC AAC AAG CCA GGG CCT GAT ACC TTC CCT GAC TAT GGG GAT GTG CTT CGG GCT GTA GAG AAG GCA GCT GCC CGA 1012 Y R G T R Y P E V N R R I E R L I N K P G P D T F P D Y G D V L R A V E K A A A R 299

CAC AGC CTT GGC CTC CCC CGA CAG CAG CTC CAG CTC ATG GCT CAG GAT GCC TTC CGA GAT GTG GGC ATC AGG TTA CAG GAG CGA CGT CAC CTC GAT CTC ATC TAC AAC TTT GGC TGC CAC CTC 1135 H S L G L P R Q Q L Q L M A Q D A F R D V G I R L Q E R R H L D L I Y N F G C H L 340

ACA GAT GAC TAT AGG CCA GGC GTT GAC CCT GCA CTA TCA GAT CCT GTG TTG GCC CGG CGC CTT CGG GAA AAC CGG AGT TTG GCC ATG AGT CGG CTG GAT GAG GTC ATC TCC AAA TAT GCA ATG 1258 T D D Y R P G V D P A L S D P V L A R R L R E N R S L A M S R L D E V I S K Y A M 381 TTG CAA GAC AAA AGT GAG GAG GGC GAG AGA AAA AAG AGA AGA GCT CGG CTC CAA GGC ACC TCT TCC CAC TCT GCA GAC ACC CCC GAA GCC TCC TTG GAT TCT GGT GAG GGC CCT AGT GGA ATG 1381 L Q D K S E E G E R K K R R A R L Q G T S S H S A D T P E A S L D S G E G P S G M 422

GCA TCC CAG GGG TGC CCT TCT GCC TCC AGA GCT GAG ACA GAT GAC GAA GAC GAT GAG GAG AGT GAT GAG GAA GAG GAG GAG GAG GAG GAA GAA GAA GAG GAG GAG GCC ACA GAT TCT GAA GAG 1504 A S Q G C P S A S R A E T D D E D D E E S D E E E E E E E E E E E E E A T D S E E 463

GAG GAG GAT CTG GAA CAG ATG CAG GAG GGT CAG GAG GAT GAT GAA GAG GAG GAC GAA GAG GAA GAA GCA GCA GCA GGT AAA GAT GGA GAC AAG AGC CCC ATG TCC TCA CTA CAG ATC TCC AAT 1627 E E D L E Q M Q E G Q E D D E E E D E E E E A A A G K D G D K S P M S S L Q I S N 504

GAA AAG AAC CTG GAA CCT GGC AAA CAG ATC AGC AGA TCT TCA GGG GAG CAG CAA AAC AAA GGA CGC ATA GTG TCA CCA TCG TTA CTG TCA GAA GAA CCC CTG GCC CCC TCC AGC ATA GAT GCT 1750 E K N L E P G K Q I S R S S G E Q Q N K G R I V S P S L L S E E P L A P S S I D A 545

GAA AGC AAT GGA GAA CAG CCT GAG GAG CTG ACC CTG GAG GAA GAA AGC CCT GTG TCT CAG CTC TTT GAG CTA GAG ATT GAA GCT TTG CCC CTG GAT ACC CCT TCC TCT GTG GAG ACG GAC ATT 1873 E S N G E Q P E E L T L E E E S P V S Q L F E L E I E A L P L D T P S S V E T D I 586

TCC TCT TCC AGG AAG CAA TCA GAG GAG CCC TTC ACC ACT GTC TTA GAG AAT GGA GCA GGC ATG GTC TCT TCT ACT TCC TTC AAT GGA GGC GTC TCT CCT CAC AAC TGG GGA GAT TCT GGT CCC 1996 S S S R K Q S E E P F T T V L E N G A G M V S S T S F N G G V S P H N W G D S G P 627

CCC TGC AAA AAA TCT CGG AAG GAG AAG AAG CAA ACA GGA TCA GGG CCA TTA GGA AAC AGC TAT GTG GAA AGG CAA AGG TCA GTG CAT GAG AAG AAT GGG AAA AAG ATA TGT ACC CTG CCC AGC 2119 P C K K S R K E K K Q T G S G P L G N S Y V E R Q R S V H E K N G K K I C T L P S 668

CCA CCT TCC CCC TTG GCT TCC TTG GCC CCA GTT GCT GAT TCC TCC ACG AGG GTG GAC TCT CCC AGC CAT GGC CTG GTG ACC AGC TCC CTC TGC ATC CCT TCT CCA GCC CGG CTG TCC CAA ACC 2242 P P S P L A S L A P V A D S S T R V D S P S H G L V T S S L C I P S P A R L S Q T 709

CCC CAT TCA CAG CCT CCT CGG CCT GGT ACT TGC AAG ACA AGT GTG GCC ACA CAA TGC GAT CCA GAA GAG ATC ATC GTG CTC TCA GAC TCT GAT TAG CTGCCTCCCCTTCTCCCTGCCTCCAGAATGTTCTG 2373 P H S Q P P R P G T C K T S V A T Q C D P E E I I V L S D S D * 740

GGATAACATTTGGAGGAAGGTGGGAAGCAGATGACTGAGGAAGGGATGGACTAAGCTAATCCCCTTTTGGTGGTGTTTCTTTAAAAAAAAAAAAAAAAAAAAAA

GGCGGAGGGAACCATGCGAGGTTCTGAGAATTGCGGCGAGGGTCGCCTCGAGAGACGGTTTCTGAGGAATTCTGAAATCCCCACCACTTCCTCCCTCCGGGGGATTTGATCCCCT ATG GCC ACC GCT AAC AGC ATC ATC GTG CTG GAT GAT 151 M A T A N S I I V L D D 12

GAT GAC GAA GAT GAA GCA GCT GCT CAG CCA GGG CCC TCC CAC CCA CTC CCC AAT GCG GCC TCA CCT GGG GCA GAA GCC CCT AGC TCC TCT GAG CCT CAT GGG GCC AGA GGA AGC AGT AGT TCG 274 D D E D E A A A Q P G P S H P L P N A A S P G A E A P S S S E P H G A R G S S S S 53

GGC GGC AAG AAA TGC TAC AAG CTG GAG AAT GAG AAG CTG TTC GAA GAG TTC CTT GAA CTT TGT AAG ATG CAG ACA GCA GAC CAC CCT GAG GTG GTC CCA TTC CTC TAT AAC CGG CAG CAA CGT 397 G G K K C Y K L E N E K L F E E F L E L C K M Q T A D H P E V V P F L Y N R Q Q R 94

GCC CAC TCT CTG TTT TTG GCC TCG GCG GAG TTC TGC AAC ATC CTC TCT AGG GTC CTG TCT CGG GCC CGG AGC CGG CCA GCC AAG CTC TAT GTC TAC ATC AAT GAG CTC TGC ACT GTT CTC AAG 520 A H S L F L A S A E F C N I L S R V L S R A R S R P A K L Y V Y I N E L C T V L K 135

GCC CAC TCA GCC AAA AAG AAG CTG AAC TTG GCC CCT GCC GCC ACC ACC TCC AAT GAG CCC TCT GGG AAT AAC CCT CCC ACA CAC CTC TCC TTG GAC CCC ACA AAT GCT GAA AAC ACT GCC TCT 643 A H S A K K K L N L A P A A T T S N E P S G N N P P T H L S L D P T N A E N T A S 176

CAG TCT CCA AGG ACC CGT GGT TCC CGG CGG CAG ATC CAG CGT TTG GAG CAG CTG CTG GCG CTC TAT GTG GCA GAG ATC CGG CGG CTG CAG GAA AAG GAG TTG GAT CTC TCA GAA TTG GAT GAC 766 Q S P

2477

B

Fig. 1.Northern blot analysis of HDaxx mRNA, and DNA and deduced amino acid sequenceof the human HDaxx cDNA. (A) A single size message of ~2.4 kilobases was detected on anorthern blot of total HeLa RNA hybridized with a 32P-labeled cDNA insert from CBP1,which encoded a peptide that interacted with CENP-C in the yeast two-hybrid assay. Sizes, inkilobases, were determined by running RNA size markers in adjacent lanes. (B) The sameprobe was used to isolate phage D31 from a λgt11 HeLa cDNA library. The entire 2,477 bpDNA sequence of the D31 insert is shown. The dashed underline indicates the DNA sequencecorresponding to the cDNA insert from CBP1. Conceptual translation of the major openreading frame, corresponding to the 740 residue HDaxx protein, is shown below the DNAsequence. Two regions with high probability of forming coiled coils (Lupas, 1996) areindicated by a single underline; two short basic motifs predicted to represent nuclearlocalization signals (Nakai and Kanehisa, 1992) are indicated by double underlines.

A

2033CENP-C binding protein, HDaxx

o ofnineoey

tionfshe

llestetf

the.

nx

ra

Ce

,

740

A. 5.00 4.00 3.00 2.00 1.00 0.00

-1.00 -2.00 -3.00 -4.00 -5.00

100 200 300 400 500 600 700

Hyd

roph

ilici

ty

Amino acid number

C.

GFP:HDaxx1-573

GFP:HDaxx573-740

GFP

GFP

GFP

GFP:HDaxx1-740* *

**

434 485 63611-17

anti-HDaxx* *

B.

Fig. 2.Functional and structural regions of the HDaxxopen reading frame and GFP:HDaxx expressionconstructs. (A) Kyte-Doolittle hydrophilicity plot ofdeduced HDaxx amino acid sequence showing thatHDaxx is a highly hydrophilic protein. (B) Schematicdrawing of HDaxx showing the relative positions ofvarious functional and structural domains. Numbersrefer to amino acid residues in the protein sequence.The black boxes represent the positions of acidic aminoacid clusters; asterisks indicate the positions of theputative nuclear localization sequences. The shadedregion represents the domain of HDaxx sufficient forinteraction with CENP-C: this corresponds to thesmallest cDNA insert that was recovered in the yeastinteraction trap assay. The region of HDaxx used toproduce the polyclonal antibody (amino acids 510-740)is indicated as anti-HDaxx. The entire region of HDaxxdesignated by the partially dashed line above theHDaxx open reading frame (amino acids 493-740) wasrecovered in a yeast two-hybrid screen for mouse Fasdeath domain interactors, while the solid part of thatline indicates the amino acids (625-740) sufficient for interaction with Fas (Yang et al., 1997). (C) The predicted structure of GFP:HDaxxfusion proteins expressed in vivo. The name of each construct indicates, in numbers, the amino acid residues of HDaxx expressed in the fusionprotein.

groups based on a comparison of their DNA sequences. largest group consisted of 11 non-identical library plasmithat were related to each other by DNA sequence. Becadatabase searches revealed that the gene encoded by this of library plasmids was novel, the full-length cDNA encodinthis interactor was isolated for further study. The cDNA inseof one member of this group of library plasmids, CBP1, wused as a probe for northern blot analysis of total HeLa Rand it detected a single message of ~2.4 kb (Fig. 1A). The sprobe was used to screen a λgt11 HeLa cDNA library, andamong the cross-hybridizing phage recovered was one, naD31, that contained an insert of ~2.5 kb. DNA sequenanalysis of this insert revealed a single long open reading fracapable of encoding a protein of 740 amino acids withcalculated molecular mass of 81.3 kDa, which we initialnamed CCBP (for CENP-Cbinding protein; Fig. 1B).

Conceptual translation of the CCBP open reading frampredicts a protein that is highly hydrophilic and acid(calculated pI = 4.6), and contains two regions with higconcentrations of glutamic and aspartic acid residues (Fig.The minor acidic region, near the amino terminus, is 7 amacids in length (100% glu + asp), while the major acidic regiin the middle third of the protein is 52 amino acids in leng(78% glu + asp). Computer analyses of the CCBP amino asequence also revealed two short basic motifs predictedrepresent nuclear localization signals (Nakai and Kaneh1992) as well as two regions with a high probability of formincoiled coils (Lupas, 1996; see Fig. 1B).

Routine searches of public databases with the CCBP amacid sequence eventually led to the surprising discovery tthis protein is highly related to mouse Daxx, a recenidentified Fas death domain-binding protein. The cDNA fmouse Daxx was cloned by its cross-hybridization withpartial human cDNA recovered in a yeast two-hybrid screfor human peptides that interacted with the intracelluldomain of murine Fas (Yang et al., 1997). This partial humcDNA encoded a peptide identical to the carboxyl-terminal 2amino acids of CCBP (Figs 2B, 3). Based on the seque

Thedsusegroupgrt

asNAame

medceme aly

eich

2).inoonthcid to

isa,g

inohattlyor aenaran48nce

identity of the partial human peptide, as well as its ability tinteract with Fas, CCBP appears to be the human homologmouse Daxx and we will hereafter refer to it as HDaxx (humaDaxx). Comparison of the mouse and human Daxx protesequences revealed an overall identity of 73% (Fig. 3). Thregion of highest amino acid conservation between the twproteins extends from their amino-termini to approximately thmajor acidic domains (83% identity). The regions immediatelfollowing the major acidic domains to the carboxyl-termini ofthese proteins showed somewhat lower sequence conserva(54% identity). The carboxyl-terminal 104 amino acids oHDaxx (residues 636-740), which are contained within thiregion of reduced sequence identity and do not include tmajor acidic domain, are sufficient for the interaction withCENP-C because these residues were encoded by the smainsert recovered in our yeast two-hybrid screen (Fig. 2B). Thcarboxyl-terminal 112 amino acids of HDaxx is also sufficienfor its interaction with the intracellular death domains omurine and human Fas (Yang et al., 1997).

The amino terminus of CENP-C interacts with HDaxx We also used the yeast interaction trap assay to determine region of CENP-C responsible for its interaction with HDaxxVarious baits, consisting of full-length CENP-C and fiveoverlapping domains constituting the entire CENP-C opereading frame, were assayed for their interaction with HDaxby monitoring galactose-inducible activation of the lacZreporter plasmid pSH18-34 in yeast (Fig. 4). Blue colony coloon Xgal-containing galactose plates was initially used as gross indicator of interactions between HDaxx and CENP-domains. These interactions were verified and their relativstrengths quantified by measuring β-galactosidase activity ofliquid yeast cultures (Fig. 4A). By this criterion, full-lengthHDaxx interacted with full-length CENP-C as well as with thetwo amino-terminal proximal CENP-C domain constructsencompassing amino acids 1-315 and 180-468.

During testing of the various LEXA:CENP-C baits, wenoted that some had intrinsic transcription activation activity

2034

,

inos.meeondinx.or-dg

A. F. Pluta and others

resulting in low to mid-level expression of the lacZ reportgene under non-inducing conditions (glucose) as well as inabsence of specific activation-tagged prey. In order determine to what extent such self-activation could accountthe interaction phenotype observed above, the followiexperiment was performed. Full-length CENP-C, as well as two amino-terminal proximal CENP-C domain constructs thalso appeared to interact with HDaxx, were each transforminto a yeast strain that contained the lacZ reporter and eiactivation-tagged HDaxx1-740 or empty vector JG4-5 as prIndividual transformants were then grown in parallel in liquculture under conditions that repressed (glucose) or indu(galactose) expression of the prey. Quantitative β-galactosidaseassays were then performed to determine the relative increin lacZ expression when HDaxx was specifically expressedthe prey. Thus, β-galactosidase enzyme units measured cells grown in glucose (repressing conditions) reflect tintrinsic activation activity of a particular bait, while enzymunits measured for cells grown in galactose (induci

HDaxx 1 MATANSIIVLDDDDEDEAAAQPGPSHPLPNAAS ||| ||||||||||||||||||||: || ||mDaxx 1 MATDDSIIVLDDDDEDEAAAQPGPSNLPPNPAS

HDaxx 69 EFLELCKMQTADHPEVVPFLYNRQQRAHSLFLA ||||||| :|:||||||||| ||||:|:|||mDaxx 75 EFLELCKTETSDHPEVVPFLHKLQQRAQSVFLA

HDaxx 143 LNLAPAATTSNEPSGNNPPTHLSLDPTNAENTA |||||||:|: | || |||| | ||:||||mDaxx 149 LNLAPAASTTSEASGPNPPTEPPSDLTNTENTA

HDaxx 217 DPDSAYLQEARLKRKLIRLFGRLCELKDCSSLT ||||:||||||||||||||||||||||||||||mDaxx 223 DPDSSYLQEARLKRKLIRLFGRLCELKDCSSLT

HDaxx 291 RAVEKAAARHSLGLPRQQLQLMAQDAFRDVGIR |||||||:|||||||||||||:|||||||||:|mDaxx 297 RAVEKAATRHSLGLPRQQLQLLAQDAFRDVGVR

HDaxx 365 RSLAMSRLDEVISKYAMLQDKSEEGERKKRRAR |:||| |||||||||||:|||:|||||:|||||mDaxx 371 RTLAMNRLDEVISKYAMMQDKTEEGERQKRRAR

HDaxx 439 DD---------EESDEEEEEEEEEEEEEATDSE || ||| |||||||||| |||: |mDaxx 445 DDDDDDDDEDNEES--EEEEEEEEEEKEATEDE

HDaxx 504 NEKNLEPGKQISRSSGEQQNKGRIVSPSLLSEE : :| || : | || :| :|: mDaxx 511 HRRNSEPAEGLRTPEG-QQKRGLTETPASPPGA

HDaxx 578 TPSSVETDISSSRKQSEEPFTTVLENGAGMVSS | |||| ||:||: |:||||| :|:|mDaxx 583 -----ERDISSPRKKSEDSLPTILENGAAVVTS

HDaxx 652 QRSVHEKNGKKICTLPSPPSPLASLAPVADSST : :: |: : | | ||||:| ||||||mDaxx 652 PMAQQDS-GQNTSVQPMPSPPLASVASVADSST

HDaxx 726 TQCDPEEIIVLSDSD 740 |||||||||||||||mDaxx 725 TQCDPEEIIVLSDSD 739

Fig. 3.HDaxx is highly homologous to mouse Daxx. HDaxx and mClustal W(1.4) sequence alignment program (MacVector 6.0) withacids are indicated by (:). Overall identity is 73%. The amino terminimmediately following the major acidic domains show 54% identitysufficient for interaction with CENP-C. The same region plus the 7interaction with Fas (Yang et al., 1997).

er theto

forngtheated

therey.idced

ase as

forheeng

conditions) reflect the activation activity resulting from theinteraction of a bait with a specific prey.

Results of this experiment, shown in Fig. 4B, aresummarized as follows. First, while LEXA:CENP-C1-943produced negligible enzyme activity when grown in glucoseLEXA:CENP-C1-315 and LEXA:CENP-C180-468 eachshowed modest, but detectable enzyme activity when grownglucose, indicating the ability of these latter two baits tactivate transcription of the lacZ reporter by themselveSecond, all three of the baits produced increased enzyactivity when grown in galactose, conditions that inducexpression of the prey, which consisted of either the activatidomain itself (empty vector) or activation domain-taggeHDaxx. However, the most significant increases occurred cells co-expressing the CENP-C-derived bait and HDaxSpecifically, a ~130-fold increase in activity was measured fcells expressing both LEXA:CENP-C1-943 and HDaxx; a 40fold increase for cells expressing LEXA:CENP-C1-315 anHDaxx; and a 9-fold increase for cells expressin

PG------AEAPSSSEPHGARGSSSSGGKKCYKLENEKLFE 68 | :| |||: ||| || :|||||:||||||TGPGPGLSQQATGLSEPRVDGGSSNSGSRKCYKLDNEKLFE 74

SAEFCNILSRVLSRARSRPAKLYVYINELCTVLKAHSAKKK 142||||||||||||:|:| ||||:||||||||||||||| |||SAEFCNILSRVLARSRKRPAKIYVYINELCTVLKAHSIKKK 148

SQSPRTRGSRRQIQRLEQLLALYVAEIRRLQEKELDLSELD 216|:: |||||||||||||||||||||||||||||||||||||SEASRTRGSRRQIQRLEQLLALYVAEIRRLQEKELDLSELD 222

GRVIEQRIPYRGTRYPEVNRRIERLINKPGPDTFPDYGDVL 290|||||||||||||||||||||||||||||| ||||||||||GRVIEQRIPYRGTRYPEVNRRIERLINKPGLDTFPDYGDVL 296

LQERRHLDLIYNFGCHLTDDYRPGVDPALSDPVLARRLREN 364|||||||||||||||||||||||||||||||| ||||||||LQERRHLDLIYNFGCHLTDDYRPGVDPALSDPTLARRLREN 370

LQGTSSHSADTPEASLDSGEGPSGMASQGCPSASRAETDDE 438| ||: : :| |:|| :||||||||||| ||::|:|||||:LLGTAPQPSDPPQASSESGEGPSGMASQECPTTSKAETDDD 444

EEEDLEQMQEGQEDDEEEDEEEEAAAGKDGDKSPMSSLQIS 503:| ||||:|| | || ||| :| || |DE-DLEQLQEDQGGDE---EEE--GGDNEGNESPTSPSDFF 510

PLAPSSIDAESNGEQPEELTLEEESPVSQLFELEIEALPLD 577 | | | |||| ||| | | :||||||| |||:||||SLDPPSTDAESSGEQLLEPLLGDESPVSQLAELEMEALPE- 582

TSFNGGVSPHNWGDSGP PCKKSRK EKKQTGSGPLGNSYVER|| || || |:| |: ||:|: |||||| ||| |||||:TSVNGRVSSHTWRDASPPSKRFRKEKKQLGSGLLGNSYIKE 651

RVDSPSHGLVTSSLCIPSPARLSQTPHSQPPRPGTCKTSVA||||||| ||||||| |||: | |||::| | |||||RVDSPSHELVTSSLCSPSPSLLLQTPQAQSLRQCIYKTSVA 724

ouse Daxx protein sequences were aligned using identity matrix of the default settings. Identical amino acids are indicated by (|); similar aminoi of the proteins (first 440 residues) show 83% identity; the regions over an aligned length of 265 residues. Underlined amino acids are additional residues indicated by a dotted underline are sufficient for

2035CENP-C binding protein, HDaxx

wenLa

ic

red

ndlls

ofrsc

estxindosctsa

t isarent

al

al

A. LexA:CENP-C Baits + HDaxx1-740:

** * *AA1-943

1-315

180-468

315-643

462-802

635-943

Colony Color

β-galUnits

blue

blue

blue

white

white

white

99

100

21

1

37

LEXA:Bicoid

LEXA

white

white

4

3

Control Baits + HDaxx1-740:

B.

1-315 1-315

180-468 180-468

HDaxx1-740 JG4-5

HDaxx1-740 JG4-5

3.122.17

1.100.74

128.01 5.37

9.37 1.63

41 2

9 2

HDaxx 1-740 JG4-5

1-943 1-943

0.520.55

68.38 1.06

β-gal Units Fold Glu Gal Prey

132 2

IncreaseLexA:CENP-C Bait

Fig. 4.HDaxx interacts with the amino-terminal 315amino acids of CENP-C. (A) Five different CENP-C-derived baits and two control baits were tested fortheir ability to interact with full-length HDaxx in theyeast two-hybrid assay. Interaction was monitored bycolony color on X-gal-containing galactose plates andwas quantified by measuring β-galactosidase units forstrains grown in liquid cultures. The domains ofCENP-C that were tested are indicated schematically.Numbers refer to the amino acid residues of CENP-Cexpressed as LexA fusion proteins; relative positionsof putative nuclear localization sequences in theCENP-C open reading frame are marked by asterisks;the hatched box indicates the region of CENP-Cresponsible for targeting the protein to centromeres invivo (Yang et al., 1996). Blue colony color on X-galcontaining galactose plates indicates an interactionbetween the bait and prey proteins, while white colonycolor indicates the absence of an interaction. Therelative strength of protein interactions wasdetermined by measuring β-galactosidase enzymeunits resulting from expression of the lacZ reportergene. (B) EGY48 containing the lacZ reporter pSH18-34 was cotransformed with the indicated combinationsof CENP-C-derived baits and activation-tagged prey vector, JG4-5, containing or lacking full-length HDaxx. β-Galactosidase units weremeasured and averaged for 3-5 independent transformants, each of which was tested under conditions that repressed (glucose) and induced(galactose) expression of the activation-tagged prey. Fold increase is the ratio of β-galactosidase enzyme units determined for cells containing aparticular bait and prey in grown in galactose divided by that determined for the same cells grown in glucose, and indicates the relative strengthof interaction independent of intrinsic activation activity of the bait.

LEXA:CENP-C180-468 and HDaxx. In contrast, a constant fold increase in enzyme activity was measured for cells cexpressing any of the CENP-C-derived baits and the emprey vector JG4-5 when grown in inducing conditions. Thudespite this low intrinsic transcriptional activation activity othe two amino-terminal proximal CENP-C baits, the mosignificant increase in activation was observed when the fi315 amino acids of CENP-C was co-expressed with HDaindicating that this region is responsible for the interaction CENP-C with HDaxx.

The HDaxx-interacting region of CENP-C does not overlawith any of the characterized functional domains of CENP-which include a centrally-located region responsible ftargeting the protein to centromeres, an overlapping domthat has non-specific DNA-binding activity in vitro, and carboxyl-terminal domain with in vitro dimerization activity(Yang et al., 1996; Sugimoto et al., 1997). A phosphorylaticonsensus site for p34cdc2, SPSK located at residues 73-76, the only distinguishing feature of amino terminus of CENP(Saitoh et al., 1992), but phosphorylation of this site during tcell cycle has not been studied, and its functional significanis not known. It has been reported that the amino terminusCENP-C possesses oligomerization activity in vitro (Sugimoet al., 1997) as well as a novel instability activity that preventoxic accumulation of high levels of CENP-C when ectopicaexpressed in baby hamster kidney cells (Lanini and McKe1995). However, a similar instability activity has not beeobserved for CENP-C when it is overexpressed in human c(Yang et al., 1996).

HDaxx is a predominantly nuclear protein and isabsent from isolated mitotic chromosomesTo begin to address the biological significance of t

2-o-ptys,fstrst

xx,of

pC,oraina

onis-Chece oftots

llyon,nells

he

interaction detected in yeast between HDaxx and CENP-C, determined the subcellular distribution of HDaxx in humatissue culture cells. A log phase suspension culture of Hecells was mechanically separated into nuclear and cytoplasmfractions by differential centrifugation of hypotonicallyswollen, Dounce-homogenized cells. An independent cultuof HeLa cells was blocked in mitosis with vinblastine, ansimilarly treated to produce partially purified mitoticchromosomes. Aliquots of each fraction (nuclei, cytoplasm achromosomes) corresponding to an equivalent number of cewere immunoblotted with a rabbit polyclonal antibodyproduced against the carboxyl-terminal 230 amino acids HDaxx (anti-HDaxx; see Materials and Methods). To monitothe efficiency of the subcellular fractionation, CENP-C wadetected in a parallel blot of identical fractions with a specifipolyclonal antibody (Tomkiel et al., 1994).

HDaxx migrated in SDS polyacrylamide gels with a relativmolecular mass of ~116 kDa and fractionated almoquantitatively with nuclear proteins (Fig. 5). The minor HDaxsignal seen in the lane containing the cytoplasmic protefraction is probably due to contamination by nuclei damageduring purification, as a small amount of CENP-C was alsdetected in this fraction. Nuclear fractionation of HDaxx waunexpected since its putative homolog, mouse Daxx, interawith the cytoplasmic death domain of Fas (CD95), a plasmmembrane-bound cell surface receptor. However, this resulconsistent with the presence of two putative nuclelocalization signals within the HDaxx amino acid sequenc(Fig. 1B). HDaxx was not detectable to an appreciable extein fractions prepared from partially purified mitoticchromosomes, suggesting either that it is not a chromosomprotein during mitosis, that it is easily dissociated fromcondensed mitotic chromosomes during the biochemic

2036

esst,ith

Sus theted-Bof

tent

uredg

llsseder

re-ingivect toe

ofct

ofy

A. F. Pluta and others

Fig. 6.HDaxx displays a punctate staining pattern in some interphasenuclei. Indirect immunofluorescence localization of HDaxx ininterphase HeLa cells. (A) DNA stained with DAPI and (B)endogenous HDaxx detected with anti-HDaxx. Bar, 10 µm.

fractionation process, or that its association with chromosomis disrupted by treatment of cells with vinblastine. In contraCENP-C fractionated as expected with nuclei and wpartially purified chromosomes.

The aberrant migration of endogenous HDaxx in SDpolyacrylamide gels is most likely due to the anomaloelectrophoretic behavior of the clustered charges present inacidic domains of the protein, a property that has been nofor other acidic clusters, such as those found in CENP(Earnshaw, 1987; Earnshaw et al., 1987). The size endogenous HDaxx observed by western blotting is consiswith that obtained by in vitro translation of the HDaxx opereading frame and is similar to that observed for in vitrtranslated mouse Daxx (AFP, data not shown; Yang et 1997). The smaller molecular mass species detected withantibody probably represent proteolytic breakdown productthe full-length protein generated during preparation of tsamples rather than immunologically related proteins presin cells, as bands of similar size were observed in loexposures of 35S-labeled HDaxx produced by in vitrotranslation (data not shown).

HDaxx colocalizes with a subset of interphasecentromeres in human nucleiTo further investigate the intracellular distribution of HDaxindirect immunofluorescence was performed using anti-HDato detect the endogenous protein in fixed HeLa cells. example of the variety of HDaxx staining patterns observed

me

-Bed

lylls.er

,en(sees

ets

ingleitxoWeedPlly

-C

ed

osedct

ig.

Fig. 5.HDaxx fractionates with nuclei but not with mitoticchromosomes. Detection of HDaxx and CENP-C in an immunobloof HeLa proteins partially purified by subcellular fractionation.Lanes labeled N contain a crude nuclear protein extract; lanes labC contain crude cytoplasmic protein extract; lanes labeled X contacrude chromosomal protein extract. All lanes were loaded withextracts normalized for equal numbers of HeLa cells. Panel 1 wasprobed with pre-immune serum from a rabbit immunized withHDaxx-derived antigen; panel 2 was probed with immune rabbitserum recognizing amino acids 510-740 of HDaxx; panel 3, wasprobed with a rabbit serum recognizing amino acids 24-623 ofCENP-C (Tomkiel et al., 1994). Molecular mass standards (in kDaare indicated.

no-al., thiss ofheentng

x,xxAn in

interphase cells is shown in Fig. 6. As predicted by obiochemical fractionation experiments, HDaxx was detectexclusively in nuclei, where it gave a diffuse nuclear staininpattern that was usually excluded from nucleoli. Many cealso displayed a heterogeneous punctate pattern superimpoon this overall pattern of diffuse nuclear staining that was nevobserved for cells stained with the corresponding rabbit pimmune serum (data not shown). The same nuclear stainpatterns were observed by fluorescence microscopy of lHeLa cells transfected with GFP:HDaxx1-740, a construexpressing the full-length HDaxx open reading frame fusedGFP, indicating that neither the nuclear localization nor thpunctate staining observed with anti-HDaxx were artifacts the procedures used to fix or permeabilize cells for indireimmunofluorescence (Figs 7K, 9B).

The punctate staining pattern we observed in a numberinterphase cells stained with anti-HDaxx was highlreminiscent of that produced by human autoimmune serucontaining anti-centromere antibodies (ACA) that recognizthe three intrinsic human centromere proteins, CENPs-A, and -C (Earnshaw and Rothfield, 1985). We therefore usanti-HDaxx and human ACA to perform doubleimmunofluorescence labeling in order to simultaneouslocalize HDaxx and centromeres in human tissue culture ceIn general, anti-HDaxx detected fewer nuclear spots in eithHeLa or U2OS cells than did ACA (Fig. 7A-D and E-Hrespectively). In addition, we noted that the HDaxx spots oftappeared less discrete than those produced by human ACA for example, Fig. 7F). However, when the individual imageproduced by staining cells with anti-HDaxx and ACA wermerged, it was clear that many, if not most, of the spoobtained with the two anti-sera were superimposable indicatthat HDaxx colocalizes with centromeres in interphase nuc(Figs 7D,H, 8G-I). Interestingly, the number of spots thacolocalized varied from cell to cell. In some cells, many HDaxspots were often juxtaposed in very close proximity tcentromeres, but could not be superimposed with them. obtained similar results in fixed cells expressing transfectGFP:HDaxx1-740 in which HDaxx was detected by GFfluorescence and endogenous CENP-C was specificadetected by indirect immunofluorescence using anti-CENP(Fig. 7I-L). Thus, the intimate juxtaposition of HDaxx andCENP-C in human cells validates the interaction we detectbetween these two proteins in yeast.

Results from our biochemical fractionation of HeLa cells alssuggested that HDaxx might not be a component of condenmitotic chromosomes (Fig. 5). This was confirmed by indireimmunofluorescence localization of HDaxx in fixed mitoticHeLa cells. Cells in prophase (Fig. 8A-C) and metaphase (F

t

eledin

)

2037CENP-C binding protein, HDaxx

naei

yots.fortd

),edd

nxts

l

of).d

hots-ndede

Fig. 7.HDaxx colocalizes withcentromere proteins in humannuclei. (A-D) untransfectedHeLa nucleus; (E-H)untransfected U2OS nuclei; (I-L)nucleus of a transfected HeLacell expressing GFP:HDaxx1-740. (A,E,I) DNA staining byDAPI; (B and F) the distributionof endogenous HDaxx detectedby anti-HDaxx; (K) fluorescencelocalization of transfectedGFP:HDaxx1-740; (C and G)centromere staining by humanautoimmune serum GS; (J)CENP-C staining by anti-CENP-C. (D,H,L) merged images of thecorresponding middle panels;colocalization of HDaxx andcentromeres in these panelsappears as yellow/orange spots.Bars, 10 µm.

8D-F) displayed diffuse overall staining but none of thpunctate pattern seen in interphase cells and likewise shono evidence of HDaxx colocalization with centromeres. Thwas also the case for cells in anaphase and telophase (dashown). In contrast, the punctate pattern of HDaxx staining centromere colocalization was apparent in cells that had exmitosis and were inferred to be in G1 (by the presence of aresidual midbody, decondensed DNA and dispersal unduplicated centromeres) or early S phase (Fig. 8G-I).

The absence of punctate HDaxx staining in some interphnuclei and in mitotic cells prompted us to investigate whetlevels of this protein fluctuate during the cell cycle. Equivaleamounts of protein lysates prepared from HeLa cells that been synchronized in the cell cycle by mitotic shake-off weimmunoblotted with anti-HDaxx (Fig. 8J; Monteiro and Mica1996). The steady-state level of HDaxx did not vary appreciaas cells progressed from mitosis through G1, S and G2 of thecell cycle, and into the next mitosis. Thus we conclude that absence of HDaxx spots in some nuclei (as well as the presof spots in others) may be controlled by cell cycle-specific potranslational modifications of HDaxx (and/or proteins wiwhich it interacts), rather than by its specific degradation.

Independent domains of HDaxx form nuclear spotsin vivo To begin a functional dissection of HDaxx, the full-length opreading frame and amino- and carboxyl-terminal deletiowere fused to green fluorescent protein (GFP) of A. victoria(Fig. 2C) and their expression in live HeLa cells followintransfection was monitored by fluorescence microscopy. Ctransfected with the empty GFP vector displayed fluoresce

ewedis

ta notandited

of

asehernt

hadre

l,bly

theencest-

th

enns

gellsnce

throughout the entire cell, with no specific accumulation inuclei and never in nuclear spots (Fig. 9A). In contrast, diffuse fluorescent signal accumulated exclusively in the nuclof cells transfected with GFP:HDaxx1-740 andGFP:HDaxx573-740 (Fig. 9B and E, respectively) and mancells expressing those constructs also displayed nuclear spThis pattern was the same as that which we observed endogenous HDaxx detected in fixed cells by indirecimmunofluorescence (Fig. 7A-H). The nuclear spots produceby expression of GFP:HDaxx573-740 colocalized withinterphase centromeres stained with ACA (data not shownthough to a lesser extent than did the nuclear spots producby expression of GFP:HDaxx1-740 (Fig. 7I-L). The decreasecentromere colocalization of GFP:HDaxx573-740 (whichcontains the CENP-C-interacting domain of HDaxx) may, ipart, be due to improper conformation of this partial HDaxpeptide, which contains the 27 kDa GFP moiety fused to iamino terminus. Alternatively, efficient centromerelocalization of HDaxx may require that the carboxyl-terminadomain be physically coupled to the rest of the protein.

The distribution of GFP:HDaxx1-573 in live transfectedcells was predominantly nuclear though a variable amount cytoplasmic fluorescence was always apparent (Fig. 9C,DSome cells expressing GFP:HDaxx1-573 also displayenuclear spots (Fig. 9D) which we found did not colocalize witinterphase centromeres detected with human ACA (data nshown). Thus, HDaxx lacking the carboxyl-terminal residuethat interact with CENP-C in yeast accumulates in noncentromeric nuclear spots. The nature of these spots, awhether they correspond to the non-centromeric spots observby staining cells for endogenous HDaxx remains to b

2038

pes,ingnestionear

A. F. Pluta and others

Fig. 8.HDaxx is present throughout the cell cycle, but its associationwith centromeres is restricted to interphase. Endogenous HDaxx andintrinsic centromere proteins were simultaneously localized byindirect immunofluorescence detection in: (A-C) a mitotic HeLa cellin prophase, (D-F) a mitotic HeLa cell in metaphase, and (G-I) aninterphase HeLa cell in G1 or early S phase. (A,D,G) DNA stainingby DAPI; (B,E,H) HDaxx staining by anti-HDaxx; (C,F,I)centromere staining by human autoimmune serum GS. HDaxx ispresent throughout the cell cycle at a constant steady state level.(J) The 116 kDa region of an immunoblot of probed with anti-HDaxx in which each lane contains 30 µg total protein from HeLacell populations synchronized in the cell cycle by mitotic shake-off(Monteiro and Mical, 1996). M, mitotic cells; numbers refer to thetime, in hours, when protein extracts were prepared after mitotic cellscollected by shake-off were plated; Ex, exponentially growing cells.Bars, 10 µm. Fig. 9.Localization of GFP and GFP:HDaxx fusion proteins in live

HeLa cells. The distribution of GFP and GFP:HDaxx fusion proteinswas observed by fluorescence microscopy of live HeLa cellstransfected with (A) the empty GFP vector, pEGFP-C1, (B)pGFP:HDaxx1-740, (C,D) pGFP:HDaxx1-573, and (E)pGFP:HDaxx573-740. (F) Immunoblot of total protein lysatescollected from cell populations transfected with lane 1, pEGFP-C1;lane 2, pGFP:HDaxx1-740; lane 3, pGFP:HDaxx1-573; lane 4,pGFP:HDaxx573-740. GFP and GFP:HDaxx fusion proteins weredetected with an antibody specific for GFP. Molecular massstandards (in kDa) are indicated. Bar, 10 µm.

determined. In no case did we observe any obvious phenotysuch as the induction of cell death or cell-cycle arrest, resultfrom the over-expression of any of the GFP:HDaxx fusioconstructs in human cells. Collectively, these results suggthat independent domains of HDaxx mediate its associatwith interphase centromeres and with other unknown nuclsubstructures.

2039CENP-C binding protein, HDaxx

eeyisin

athant

N-2lyus).de be

dtheas-ctylLa

sein

byfueneaxxtictoanell- ofseouss.inirst,urasenotareas.edhells.tol no

e

enre

entin

Immunoblot analysis of whole cell lysates collected frotransfected HeLa cell populations revealed that all construexpressed GFP fusion peptides of the expected sizes (Fig.Interestingly, GFP:HDaxx573-740 expressed in HeLa cewas always detected as a broadly migrating band immunoblots, suggesting that the carboxyl terminus of HDamay be a target for post-translational modification in vivo.

DISCUSSION

The human centromere is a dynamic chromosomal structhat undergoes striking morphological and functiontransformations throughout the cell cycle, the becharacterized of which are accompanied by its transiassociation with specific proteins during mitosis. Here we repthe unexpected interphase-specific association of humcentromeres with HDaxx, a protein whose murine homolog wpreviously identified through its ability to bind the death domaof the cell surface receptor Fas (CD95). The associationHDaxx with interphase centromeres is mediated by interaction with the amino terminus of the intrinsic centromeprotein CENP-C, with which it interacts in the yeast interactitrap assay. Targeting of HDaxx to interphase centromeredetermined by residues in its carboxyl terminus, which are aresponsible for its interaction with CENP-C and with thintracellular death domain of Fas. These results suggepreviously unsuspected role for interphase centromeres inregulation of cellular responses to death stimuli.

CENP-C is one of the few known proteins that colocalizwith centromeric heterochromatin at all stages of the cell cyThis juxtaposition is best characterized in condensed mitochromosomes, where CENP-C is found in the inner kinetochplate (Saitoh et al., 1992). While the structural and functiostatus of centromeres during interphase is less clear, presence of CENP-C at interphase centromeres is neverthimportant for proper kinetochore structure and function in tsubsequent mitosis. This was first demonstrated by antibmicroinjection studies, which showed that the structuintegrity of the kinetochore as well as mitotic progression wedisrupted when antibodies to CENP-C were injected ininterphase cells, but not when they were injected into mitocells (Tomkiel et al., 1994). Because little or no CENP-C coube detected in the antibody-injected cells, it was suggestedthese effects resulted from depletion of CENP-C frointerphase centromeres, indicating that interphase centromare not functionally quiescent at this point in the cell cycle.

Similar results were obtained in a chicken cell lincontaining a conditional mutation in the cognate chicken gefor CENP-C that causes the efficient removal of CENP-C frocentromeres and results in a complete block in the metaphto anaphase transition (Fukagawa and Brown, 199Interestingly, these mitotically arrested cells subsequently dby apoptosis. Targeted disruption of the mouse CENP-C gwas also recently reported to result in disruption of mitosis aearly embryo death (Kalitsis et al., 1998). The apoptoresponse to the loss of CENP-C in the chicken cells suggthe perturbation of a regulatory function that links mitotprogression with cell viability. While it was argued that CENC is not necessary for the cell to successfully traveinterphase, the resulting metaphase block is nonetheconsistent with an important function, or perhaps interacti

mcts

9F).llsonxx

turealst-entortanasin ofitsre

ons islsoe

st a the

escle.ticorenalthe

elessheodyralretoticld

thatmeres

enemase7).iedenendticestsicP-rselesson,

that must occur in interphase but is not manifest until thsubsequent mitosis. How such functions are controlled so thare not expressed inappropriately is not known, but it reasonable to presume that cell cycle-specific protein-proteinteractions and/or protein modifications could be involved.

Although indirect, three lines of evidence suggest thHDaxx, which interacts with CENP-C and colocalizes withuman interphase centromeres, is the functional humhomolog of mouse Daxx, a novel signaling protein thaenhances Fas-mediated apoptosis by activating the Junterminal kinase pathway. First, the carboxyl-terminal 11amino acids of both mouse Daxx and HDaxx interact strongwith mouse Fas intracellular region, and the carboxyl terminof HDaxx also interacts with human FasIC (Yang et al., 1997Second, the two proteins are 73% identical, and nucleotisequence conservation enabled the mouse Daxx cDNA tocloned by hybridization with the 3′ end of the human DaxxcDNA (Yang et al., 1997). Finally, the most compellingevidence for functional conservation between HDaxx anmouse Daxx was the demonstration that overexpression of carboxyl terminus of mouse Daxx with Fas suppressed Fmediated apoptosis in HeLa cells, a dominant-negative effethat was reversed when full-length mouse Daxx, the carboxterminus of mouse Daxx and Fas were co-expressed in Hecells, suggesting that the carboxyl terminus of the mouprotein competed with an endogenous human prote(presumably HDaxx) for binding to Fas (Yang et al., 1997).

We have demonstrated by biochemical fractionation and immunolocalization that the majority, and possibly all, oendogenous HDaxx is found in the nuclei of human tissculture cells. This result is consistent with HDaxx’s interactiowith CENP-C and its colocalization with human interphascentromeres, but it presents a paradox if HDaxx and mouse Dare true functional homologs, since transduction of apoptosignals by Daxx is thought to occur through its direct binding the cytoplasmic death domain of Fas at the plasma membr(Yang et al., 1997). However, the lack of direct evidence for cesurface localization of mouse Daxx as well as the presencethree putative nuclear localization signals within the mouDaxx amino acid sequence raises the possibility that endogenmouse Daxx, like HDaxx, also normally resides in the nucleuWe can envision two scenarios by which HDaxx could gaaccess to the cytoplasm to bind cell surface death domains. Fthe interaction of HDaxx with Fas death domains could occby default after nuclear envelope breakdown at prometaphwhen, as we have shown, HDaxx is present in the cell but associated with centromeres or confined to the nuclecompartment. Alternatively, HDaxx may shuttle between thnucleus and cytosol, perhaps in response to activation of FKiriakidou et al. (1997) recently reported that an epitope-taggprotein essentially identical to HDaxx was detected in both tnucleus and cytoplasm when over-expressed in COS ceAlthough we cannot completely rule out the possibility thacytoplasmic HDaxx exists in human cells or that a minor pobecomes cytoplasmic when over-expressed, we have seenevidence for significant amounts of full-length HDaxx in thcytoplasm of either transfected or untransfected HeLa cells.

It is curious that essentially the same region of HDaxx has befound to interact with CENP-C and Fas, two proteins that shano recognizable common features and that reside in differsubcellular compartments. It is possible that subdomains with

2040

nr

.

:

e.

se.

innd

-

th

c

ct.

n

e

A. F. Pluta and others

the carboxyl terminus of HDaxx have unique binding specificitiwhose activities are differentially regulated in the cell. In thregard, our demonstration that the carboxyl terminus of HDaappears to be a target for post-translational modification mprovide a mechanism for the regulation of HDaxx interactiowith CENP-C and/or Fas. Although there is evidence indicatithat HDaxx can modulate the apoptotic response under cercircumstances, conflicting reports question the generality oparallel Daxx-mediated apoptotic pathway downstream of Fthat is distinct from the FADD-mediated pathway (Yang et a1997; Zhang et al., 1998; Wajant et al., 1998). Thus, odemonstration that HDaxx is a nuclear protein and thattransiently associates with interphase centromeres suggestsits roles in vivo may be more complex than previously suspect

The interphase centromere has been a notoriously elusubject for study not only because it resides in a structuraundistinguished collection of chromatin in the nucleus, bbecause it is not well-characterized biochemically. AClocalization and fluorescence in situ hybridization studies haindicated that the only subcellular structures with whicinterphase centromeres routinely associate are the nucenvelope and nucleoli (Brenner et al., 1981; Moroi et al., 198In a previous search for nuclear proteins that interact wCENP-C, we found that the nucleolar transcription factUBF/NOR-90 specifically binds the carboxyl terminus oCENP-C (Pluta et al., 1995). While the functional significanof that interaction remains unclear, it is consistent with tultrastructural and biochemical detection of human CENPsisolated nucleoli and supports the argument that the dominautoantigen in scleroderma spectrum disease is macromolecular complex consisting of centromeres anucleoli (Tan et al., 1988; Ochs and Press, 1992). It was recereported that proteins encoded by two genes associated early-onset familial Alzheimer’s disease, PS1 and PS2,colocalize with centromere antigens recognized by human Ain interphase nuclei (Li et al., 1997). It is not known which, any, of the intrinsic CENPs are involved in this colocalizatio

Our identification of a protein that associates with an intrinscomponent of the centromere specifically during interphaprovides not only a unique opportunity to functionallcharacterize interphase centromeres, but a potential link betwcentromeres and a cell death pathway. If HDaxx, like mouDaxx, potentiates Fas-mediated apoptosis through direct bindof a cell surface death receptor, centromere association duinterphase may serve a cell cycle-regulated sequestering functo control cellular responses to apoptotic stimuli. Thus, perhain addition to sensing spindle damage during mitosis by thtransient interactions with MAD2 and BUB1, centromeres malso serve as cell cycle sensors of death stimuli by their transassociation with HDaxx in interphase. Alternatively, HDaxmay have additional/separate functions within the nucleperhaps linking chromosomal events during interphase with Jun-N-terminal kinase signalling pathway that is involved in ccycle regulation rather than the apoptotic response.

We thank Dr R. Brent for generously providing the yeast twhybrid reagents; Dr J. Tomkiel for DNA constructs; Dr J. Boeke fE. coli JBe15; Dr M. Monteiro for providing protein lysates fromsynchronized HeLa cells; and A. Cox and L. Ostrowski for excelletechnical assistance. A.F.P is grateful to Drs M. Monteiro, L. CasioRosen and A. Rosen for their support and encouragement du

esisxxayn

ngtainf aasl.,ur

it thated. sivellyutAvehlear1).ithorf

cehe inant

andntlywith

CAifn. icseyeenseingringtionps

eirayientxus,theell

o-or

ntla-ring

different stages of this work. This work was supported in part by aArthritis Investigator Award to A.F.P. The GenBank accession numbefor the sequence of HDaxx reported in this paper is AF050179.

REFERENCES

Allshire, R. C. (1997). Centromeres, checkpoints and chromatid cohesionCurr. Opin. Genet. Dev.7, 264-273.

Altschul, S. F., Gish, W., Miller, W., Myers, E. W. and Lipman, D. J.(1990).Basic local alignment search tool. J. Mol. Biol.215, 403-10.

Brenner, S., Pepper, D., Berns, M. W., Tan, E. and Brinkley, B. R.(1981).Kinetochore structure, duplication and distribution in mammalian cellsanalysis by human autoantibodies from scleroderma patients. J. Cell Biol.91, 95-102.

Brinkley, B. R., Ouspenski, I. and Zinkowski, R. P.(1992). Structure andmolecular organization of the centromere-kinetochore complex. Trends CellBiol. 2, 15-21.

Brown, M. T. (1995). Sequence similarities between the yeast chromosomsegregation protein Mif2 and the mammalian centromere protein CENP-CGene160, 111-116.

Campbell, M. S. and Gorbsky, G. J.(1995). Microinjection of mitotic cellswith the 3F3/2 anti-phosphoepitope antibody delays the onset of anaphaJ. Cell Biol.129, 1195-1204.

Casiano, C. A., Landberg, G., Ochs, R. L. and Tan, E. M.(1993).Autoantibodies to a novel cell cycle-regulated protein that accumulates the nuclear matrix during S phase and is localized in the kinetochores aspindle midzone during mitosis. J. Cell Sci.1993, 106.

Chomczynski, P.(1996). Guanidine methods for total RNA preparation. InCurrent Protocols in Molecular Biology(ed. F. A. Ausubel, R. Brent, R. E.Kingston, D. D. Moore, J. G. Seidman, J. A. Smith and K. Struhl), pp. 4.2.34.2.4. New York: John Wiley & Sons.

Cooke, C. A., Bernat, R. L. and Earnshaw, W. C.(1990). CENP-B: A majorhuman centromere protein located beneath the kinetochore. J. Cell Biol.110,1475-1488.

Earnshaw, W. C. and Rothfield, N.(1985). Identification of a family ofhuman centromere proteins using autoimmune sera from patients wiscleroderma. Chromosoma91, 313-321.

Earnshaw, W. C. (1987). Anionic regions in nuclear proteins. J. Cell Biol.105, 1479-1482.

Earnshaw, W. C., Sullivan, K. F., Machlin, P. S., Cooke, C. A., Kaiser, D.A., Pollard, T. D., Rothfield, N. F. and Cleveland, D. W.(1987). Molecularcloning of cDNA for CENP-B, the major human centromere autoantigen. J.Cell Biol. 104, 817-829.

Earnshaw, W. C., Ratrie, H. and Stetten, G.(1989). Visualization ofcentromere proteins CENP-B and CENP-C on a stable dicentrichromosome in cytological spreads. Chromosoma (Berl.). 98, 1-12.

Earnshaw, W. C., Bernat, R. L., Cooke, C. A. and Rothfield, N. F.(1991).Role of the centromere/ kinetochore in cell cycle control. Cold SpringHarbor Symp. Quant. Biol.56, 675-685.

Earnshaw, W. C. and Cooke, C. A.(1991). Analysis of the distribution ofthe INCENPs throughout mitosis reveals the existence of three distinsubstages of metaphase and early events in cleavage furrow formationJ.Cell Sci.98, 443-461.

Earnshaw, W. C. and Rattner, J. B.(1991). The use of autoantibodies in thestudy of nuclear and chromosomal organization. Meth. Cell Biol. 35, 135-175.

Feinberg, A. P. and Vogelstein, B.(1983). A technique for radiolabeling DNArestriction endonuclease fragments to high specific activity. Anal. Biochem.132, 6-13.

Fukagawa, T. and Brown, W. R. A.(1997). Efficient conditional mutationof the vertebrate CENP-C gene. Hum. Mol. Genet.6, 2301-2308.

Golemis, E. A. and Brent, R.(1992). Fused protein domains inhibit DNAbinding by LexA. Mol. Cell. Biol.12, 3006-3014.

Golemis, E., Gyuris, J. and Brent, R.(1996). Interaction trap/two-hybridsystem to identify interacting proteins. In Current Protocols in MolecularBiology (ed. F. A. Ausubel, R. Brent, R. E. Kingston, D. D. Moore, J. G.Seidman, J. A. Smith and K. Struhl), pp. 20.1.1-20.1.28. New York: JohWiley & Sons.

Gyuris, J., Golemis, E., Chertkov, H. and Brent, R.(1993). Cdi1, a human G1and S phase protein phosphatase that associates with Cdk2. Cell 75, 791-803.

Halverson, D., Baum, M., Stryker, J., Carbon, J. and Clarke, L.(1997). Acentromere DNA-binding protein from fission yeast affects chromosomsegregation and has homology to human CENP-B. J. Cell Biol.136, 487-500.

2041CENP-C binding protein, HDaxx

e

.

e

-ell

:

re

r

r

of

r

d

e

Kalitsis, P., Fowler, K. J., Earle, E., Hill, J. and Choo, K. H. A.(1998).Targeted disruption of mouse centromere protein C gene leads to midisarray and early embryo death. Proc. Nat. Acad. Sci. USA95, 1136-1141.

Kapoor, M., Montes de Oca Luna, R., Lozano, G., Cummings, C.,Brinkley, B. R. and May, G. S.(1997). The CENP-B gene is not essentiain mice. Mol. Biol. Cell8, 267a.

Kiriakidou, M., Driscoll, D. A., Lopez-Guisa, J. M. and Strauss, J. F. I.(1997). Cloning and expression of primate Daxx cDNAs and mappingthe human gene to chromosome 6p21.3 in the MHC region. DNA Cell Biol.16, 1289-1298.

Knehr, M., Poppe, M., Schroeter, D., Eickelbaum, W., Finze, E.-M.,Kiesewetter, U.-L., Enulescu, M., Arand, M. and Paweletz, N.(1996).Cellular expression of human centromere protein C demonstrates a cybehavior with highest abundance in the G1 phase. Proc. Nat. Acad. Sci. USA93, 10234-10239.

Lanini, L. and McKeon, F. (1995). Domains required for CENP-C assemblat the kinetochore. Mol. Biol. Cell6, 1049-1059.

Lee, J.-K., Huberman, J. A. and Hurwitz, J. (1997). Purification andcharacterization of a CENP-B homologue protein that binds to tcentromeric K-type repeat DNA of Schizosaccharomyces pombe. Proc. Nat.Acad. Sci. USA94, 8427-8432.

Li, J., Xu, M., Zhou, H., Ma, J. and Potter, H.(1997). Alzheimer presenilinsin the nuclear membrane, interphase kinetochores, and centrosomes sua role in chromosome segregation. Cell 90, 917-927.

Li, Y. and Benezra, R.(1996). Identification of a human mitotic checkpoingene:hsMAD2. Science274, 246-248.

Lupas, A. (1996). Prediction and analysis of coiled-coil structures. Meth.Enzymol.266, 513-525.

Mackay, A., Eckley, D. M., Chue, C. and Earnshaw, W. C.(1993).Molecular analysis of the INCENPs (inner centromere proteins): sepadomains are required for association with microtubules during interphand with the central spindle during anaphase. J. Cell Biol.123, 373-385.

Matsumoto, H., Masukata, H., Muro, Y., Nozaki, N. and Okazaki, T.(1989a). A human centromere antigen (CENP-B) interacts with a shspecific sequence in alphoid DNA, a human centromeric satellite. J. CellBiol. 109, 1963-1973.

Matsumoto, H., Sugimoto, K. and Okazaki, T.(1989b). Alphoid satelliteDNA is tightly associated with centromere antigens in human chromosomthroughout the cell cycle. Exp. Cell Res.181, 181-196.

Matsumoto, H., Yoda, K., Ikeno, M., Kitagawa, K., Muro, Y. and Okazaki,T. (1993). Properties of CENP-B and its target sequence in a satellite DIn Chromosome Segregation and Anneuploidy(ed. B. K. Vig), pp. 31-43.Berlin: Springer-Verlag.

Meluh, P. B. and Koshland, D.(1995). Evidence that the MIF2 gene oSaccharomyces cerevisiae encodes a centromere protein with homolothe mammalian centromere protein CENP-C. Mol. Biol. Cell6, 793-807.

Monteiro, M. J. and Mical, T. I. (1996). Resolution of kinase activities duringthe HeLa cell cycle: Identification of kinases with cyclic activities. Exp. CellRes.223, 443-451.

Moroi, Y., Peebles, C., Fritzler, M. J., Steigerwald, J. and Tan, E. M.(1980). Autoantibody to centromere (kinetochore) in scleroderma seProc. Nat. Acad. Sci. USA77, 1627-1631.

Moroi, Y., Hartman, A. L., Nakane, P. K. and Tan, E. M. (1981).Distribution of kinetochore (centromere) antigen in mammalian cell nuclJ. Cell Biol.90, 254-259.

Nakai, K. and Kanehisa, M.(1992). A knowledge base for predicting proteinlocalization sites in eukaryotic cells. Genomics14, 897-911.

Ochs, R. L. and Press, R. I.(1992). Centromere autoantigens are associatwith the nucleolus. Exp. Cell Res.200, 339-350.

Page, S. L., Earnshaw, W. C., Choo, K. H. A. and Shaffer, L. G.(1995).Further evidence that CENP-C is a necessary component of accentromeres: studies of a dic(X;15) with simultaneous immunofluoresceand FISH. Hum. Mol. Genet.4, 289-294.

Pluta, A. F., Saitoh, N., Goldberg, I. and Earnshaw, W. C.(1992).Identification of a subdomain of CENP-B that is necessary and sufficifor localization to the human centromere. J. Cell Biol.116, 1081-1093.

Pluta, A. F., Mackay, A. M., Ainsztein, A. M., Goldberg, I. G. andEarnshaw, W. C. (1995). The centromere: hub of chromosomal activitieScience270, 1591-1594.

Pluta, A. F. and Earnshaw, W. C.(1996). Specific interaction between humakinetochore protein CENP-C and a nucleolar transcriptional regulatorJ.Biol. Chem.271, 18767-18774.

Rattner, J. B., Rao, A., Fritzler, M. J., Valencia, D. W. and Yen, T. J.(1993).

totic

l

of

clic

y

he

ggest

t

ratease

ort

es

NA.

fgy to

ra.

ei.

ed

tivence

ent

s.

n.

CENP-F is a.ca 400 kDa kinetochore protein that exhibits a cell-cycldependent localization. Cell Motil. Cytoskel.26, 214-226.

Reynolds, A. and Lundblad, V. (1989). Yeast vectors and assays forexpression of cloned genes. In Current Protocols in Molecular Biology(ed.F. A. Ausubel, R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. ASmith and K. Struhl), pp. 13.6.1-13.6.4. New York: John Wiley & Sons.

Saitoh, H., Tomkiel, J., Cooke, C. A., Ratrie III, H., Maurer, M., Rothfield,N. F. and Earnshaw, W. C. (1992). CENP-C, an autoantigen inscleroderma, is a component of the human inner kinetochore plate. Cell 70,115-125.

Schaar, B. T., Chan, G. K. T., Maddox, P., Salmon, E. D. and Yen, T. J.(1997). CENP-E function at kinetochores is essential for chromosomalignment. J. Cell Biol.139, 1373-1382.

Stoler, S., Keith, K. C., Curnick, K. E. and Fitzgerald-Hayes, M.(1995).A mutation in CSE4, an essential gene encoding a novel chromatinassociated protein in yeast, causes chromosome nondisjunction and ccycle arrest at mitosis. Genes Dev.9, 573-586.

Sugimoto, K., Kuriyama, K., Shibata, A. and Himeno, M. (1997).Characterization of internal DNA-binding and C-terminal dimerizationdomains of human centromere/kinetochore autoantigen CENP-C in vitrorole of DNA-binding and self-associating activities in kinetochoreorganization. Chromosome Res.5, 132-141.

Sullivan, B. A. and Schwartz, S.(1995). Identification of centromericantigens in dicentric Robertsonian translocations: CENP-C and CENP-E anecessary components of functional centromeres. Hum. Mol. Genet.4,2189-2197.

Sullivan, K. F., Hechenberger, M. and Masri, K.(1994). Human CENP-Acontains a histone H3 related histone fold domain that is required fotargeting to the centromere. J. Cell Biol.127, 581-592.

Tan, E. M., Chan, E. K. L., Sullivan, K. F. and Rubin, R. L. (1988).Antinuclear antibodies (ANAs): diagnostically specific immune markers andclues toward the understanding of systemic autoimmunity. Clin. Immunol.Immunopathol.47, 121-141.

Taylor, S. S. and McKeon, F.(1997). Kinetochore localization of murineBub1 is required for normal mitotic timing and checkpoint response tospindle damage. Cell 89, 727-735.

Tomkiel, J., Cooke, C. A., Saitoh, H., Bernat, R. L. and Earnshaw, W. C.(1994). CENP-C is required for maintaining proper kinetochore size and foa timely transition to anaphase. J. Cell Biol.125, 531-545.

Wajant, H., Johannes, F.-J., Haas, E., Siemienski, K., Schwenzer, R.,Schubert, G., Weiss, T., Grell, M. and Scheurich, P.(1998). Dominant-negative FADD inhibits TNFR60-, Fas/Apo1- and TRAIL-R/Apo2-mediated cell death but not gene induction. Curr. Biol. 8, 113-116.

Warburton, P., Cooke, C. A., Bourassa, S., Vafa, O., Sullivan, B. A.,Stetten, G., Gimelli, G., Warburton, D., Tyler-Smith, C., Sullivan, K. F.,Poirer, G. G. and Earnshaw, W. C.(1997). Immunolocalization of CENP-A suggests a distinct nucleosome structure at the inner kinetochore plateactive centromeres. Curr. Biol. 7, 901-904.

Wood, K. W., Sakowicz, R., Goldstein, L. S. B. and Cleveland, D. W.(1997). CENP-E is a plus end-directed kinetochore motor required fometaphase chromosome alignment. Cell 91, 357-366.

Wordeman, L., Steurer, E., Sheetz, M. and Mitchison, T.(1991). Chemicalsubdomains within the kinetochore domain of isolated CHO mitoticchromosomes. J. Cell Biol.114, 285-294.

Wordeman, L. and Mitchison, T. J. (1995). Identification and partialcharacterization of mitotic centromere-associated kinesin, a kinesin-relateprotein that associates with centromeres during mitosis. J. Cell Biol.128,95-105.

Yang, C. H., Tomkiel, J., Saitoh, H., Johnson, D. H. and Earnshaw, W. C.(1996). Identification of overlapping DNA-binding and centromere-targeting domains in the human kinetochore protein CENP-C. Mol. Cell.Biol. 16, 3576-3586.

Yang, X., Khosravi-Far, R., Chang, H. Y. and Baltimore, D.(1997). Daxx,a novel Fas-binding protein that activates JNK and apoptosis. Cell 89, 1067-1076.

Zervos, A. S., Gyuris, J. and R., B.(1993). Mxi1, a protein that specificallyinteracts with Max to bind Myc-Max recognition sites. Cell 72, 223-232.

Zhang, J., Cado, D., Chen, A., Kabra, N. H. and Winoto, A.(1998). Fas-mediated apoptosis and activation-induced T-cell proliferation are defectivin mice lacking FADD/Mort1. Nature 392, 296-299.

Zhu, X., Mancini, M. A., Chang, K.-H., Liu, C.-Y., Chen, C.-F., Shan, B.,Jones, D., Yang-Feng, T. L. and Lee, W.-H.(1995). Characterization of anovel 350-kilodalton nuclear phosphoprotein that is specifically involved inmitotic-phase progression. Mol. Cell. Biol.15, 5017-5029.