1747Research Article

IntroductionKinetochores attach chromosomes to the spindle apparatus, movechromosomes along the spindle, and govern the spindle assemblycheckpoint (SAC) that prevents precocious mitotic exit before theestablishment of proper connections between the chromosomes andthe spindle (Meraldi et al., 2006; Musacchio and Salmon, 2007).Many kinetochore components have been evolutionarily conservedbetween unicellular eukaryotes and metazoans, even though thehomology is sometimes difficult to discern (Meraldi et al., 2006;Przewloka et al., 2007; Schittenhelm et al., 2007). However, onefundamental difference between the kinetochores of Saccharomycescerevisiae and those of higher eukaryotes is that the latter containthe minus-end-directed motor cytoplasmic dynein, whereas yeastkinetochores do not. At least in some metazoan cell types, dyneinat the kinetochores helps move the chromosomes toward thespindle poles during prometaphase and anaphase (Savoian et al.,2000; Sharp et al., 2000). Furthermore, at metaphase, dyneintransports several proteins, including SAC components, off thekinetochores onto the microtubules to which they are attached, aprocess called ‘streaming’ or ‘shedding’. This transport is thoughtto be responsible for turning off the SAC when chromosomes areproperly bioriented (Howell et al., 2001; Wojcik et al., 2001).

Metazoan kinetochores must therefore contain molecules thatattract dynein during mitosis and regulate dynein function once ithas arrived at the kinetochores. The polypeptides Rough Deal,ZW10 and Zwilch form a complex (RZZ) at the kinetochore, whichis required for dynein recruitment (Karess, 2005; Starr et al., 1998;Williams et al., 2003). The involvement of RZZ in dynein’s

association with the kinetochore is probably mediated by dynactin(a large multiprotein assembly that can activate dynein), becausehuman ZW10 interacts with the p50 subunit of dynactin in the yeasttwo-hybrid system (Starr et al., 1998). A second interactor foundin the same two-hybrid screen was Zwint-1, which was subsequentlyshown to be a human kinetochore component needed for kinetochoretargeting of RZZ (Starr et al., 2000; Wang et al., 2004). Thekinetochore thus contains complicated molecular machinery whosemain purpose is to control the localization and regulate the functionof dynein.

Yet another human ZW10 interactor identified in the same yeasttwo-hybrid screen was originally annotated as a homolog ofXenopus mitotic phosphoprotein 43 (MP43), whose function at thattime was unknown (Starr et al., 2000). MP43 was subsequentlydiscovered to be one of the two human homologs of Aspergillusnidulans NudE, a protein that is important for the even distributionof nuclei along the hyphae (Efimov and Morris, 2000). Severallaboratories have since reported physical interactions betweenNudE (or the closely related Nudel) and dynein (Liu et al., 2000;Niethammer et al., 2000; Sasaki et al., 2000), and between NudEor Nudel and Lis1 (Efimov and Morris, 2000; Feng et al., 2000;Kitagawa et al., 2000; Sasaki et al., 2000; Sweeney et al., 2001),the latter being a WD40-containing, dynein-interacting protein thatis involved in neuronal migration and neural stem cell division (Shuet al., 2004), and which localizes both to microtubule plus ends (Liet al., 2005) and to the kinetochores of mitotic chromosomes (Silleret al., 2005; Tai et al., 2002). These observations all suggest thatNudE proteins might also be part of the dynein-related machinery

We examined the distribution of the dynein-associated proteinNudE in Drosophila larval brain neuroblasts and spermatocytes,and analyzed the phenotypic consequences of a nudE nullmutation. NudE can associate with kinetochores, spindles andthe nuclear envelope. In nudE mutant brain mitotic cells,centrosomes are often detached from the poles. Moreover, thecentrosomes of mutant primary spermatocytes do not migratefrom the cell cortex to the nuclear envelope, establishing a newrole for NudE. In mutant neuroblasts, chromosomes fail tocongress to a tight metaphase plate, and cell division arrestsbecause of spindle assembly checkpoint (SAC) activation. Thetargeting of NudE to mitotic kinetochores requires the dynein-interacting protein Lis1, and surprisingly Cenp-meta, a

Drosophila CENP-E homolog. NudE is non-essential for thetargeting of all mitotic kinetochore components tested. However,in the absence of NudE, the ‘shedding’ of proteins off thekinetochore is abrogated and the SAC cannot be turned off,implying that NudE regulates dynein function at thekinetochore.

Supplementary material available online athttp://jcs.biologists.org/cgi/content/full/122/11/1747/DC1

Key words: Mitosis, Meiosis, Kinetochores, Centrosomes, Spindleassembly checkpoint, Dynein

Summary

Roles of the Drosophila NudE protein in kinetochorefunction and centrosome migrationAlan Wainman1,*, Jacklyn Creque2,*, Byron Williams2, Erika V. Williams2, Silvia Bonaccorsi1, Maurizio Gatti1and Michael L. Goldberg2,‡

1Instituto Pasteur Fondazione Cenci Bolognetti and Istituto di Biologia e Patologia Molecolari del CNR, Dipartimento di Genetica e BiologiaMolecolare, Sapienza, Università di Roma, 00185 Rome, Italy2Department of Molecular Biology and Genetics, Biotechnology Building, Cornell University, Ithaca, NY 14853, USA*These authors contributed equally to this work‡Author for correspondence (e-mail: mlg11@cornell.edu)

Accepted 16 February 2009Journal of Cell Science 122, 1747-1758 Published by The Company of Biologists 2009doi:10.1242/jcs.041798

Jour

nal o

f Cel

l Sci

ence

1748

at the kinetochore. However, little information currently exists aboutthe function of NudE, particularly in terms of its roles in celldivision.

In this report, we examine the properties of NudE in Drosophila.We find NudE on kinetochores, spindles and the nuclear envelopeat various stages in the cell cycle or in various cell types.Furthermore, because Drosophila has only a single gene for thisprotein, we have been able to analyze the phenotypic consequencesof nudE mutation. We observed strong defects in spindleorganization and centrosome behavior in both mutant larval braincells and spermatocytes. In mutant brains, most dividing cells arearrested in metaphase because of activation of the SAC. Despitethe interaction of NudE with ZW10, the targeting of NudE to thekinetochore does not require the RZZ complex, but it does dependupon Lis1, and surprisingly, Cenp-meta, one of two Drosophilahomologs of the vertebrate kinetochore-associated kinesin-likeprotein CENP-E (Yucel et al., 2000). We found no evidence thatNudE is required for the targeting of any other SAC component tomitotic kinetochores. However, the shedding of various SACproteins onto kinetochore microtubules (KMTs) does not occur innudE mutants, so the SAC cannot be turned off, implying that NudEhelps to regulate dynein function at the kinetochore.

ResultsLocalization of NudETo explore the intracellular localization of NudE, we generated anantibody that recognizes this protein. Western blots of larvalhomogenates showed a prominent band of approximately 38 kDain the wild type, which is close to the expected size of two isoformsof the Drosophila NudE protein annotated in FlyBase (36.2 and37.7 kDa) (Fig. 1A). This band was not found in extracts of larvaehomozygous for a null mutation in the nudE gene (nudE39A, whosegenesis is described below), indicating that this band represents theNudE protein (Fig. 1A). The antibody also recognized a minorNudE-specific band of ~45 kDa that probably corresponds to a thirdannotated NudE isoform of 42.6 kDa.

Immunofluorescence experiments using purified NudE antibodyshowed that the localization of NudE changed during the course ofmitosis in third instar larval brains (Fig. 1B; supplementary materialFig. S1). In wild-type brains, NudE was found at discrete spots onthe chromosomes during prometaphase (Fig. 1B; supplementarymaterial Fig. S1). We interpret these spots as being the kinetochores,based on their position just distal to the centromere as defined bystaining for CID, the Drosophila centromere-specific histone H3variant (supplementary material Fig. S1) (Blower and Karpen,2001). When the chromosomes are aligned at the metaphase plate,NudE is found either at the kinetochores or simultaneously at thekinetochores and along the spindle (Fig. 1B; supplementary materialFig. S1). From precedents established with other outer kinetochoremarkers including the RZZ complex (reviewed by Karess, 2005),Mad2 (Howell et al., 2000), the p50 and Glued subunits of dynactin(Siller et al., 2005; Wojcik et al., 2001), and Lis1 (Siller et al., 2005),we believe the spindle localization reflects the poleward ‘streaming’of NudE from the kinetochores onto the KMTs to which they areattached. However, the proportion of metaphases exhibitingstreaming was much lower for the NudE protein (22% of 50 cells)than for the RZZ protein Rod (54% of 50 cells). Discrete NudEsignals on chromosomes or spindles were not seen during anaphaseor telophase (Fig. 1B), in contrast with Rod, which was almostalways found at anaphase or telophase kinetochores (100% of 30cells). In mitotic cells from the brains of third instar larvae

Journal of Cell Science 122 (11)

Fig. 1. The nudE gene and gene product. (A) Specificity of the anti-NudEantibody. Western blot with the purified antibody shows two bands in wild-type third instar larval extracts not found in comparable extracts from nudE39A

homozygous mutant larvae. These bands correspond to the molecular massesof known Drosophila NudE isoforms. As a loading control, the same blot wasreprobed with antibody generated against the product of the CG7705 gene.(B) Distribution of NudE protein in wild-type larval brains. NudE is targeted tothe kinetochores at prometaphase and then ‘streams’ along KMTs duringmetaphase. During anaphase and telophase, NudE is not found in discretestructures. Scale bar: 5 μm. (C) Generation of a null mutation in nudE. The toprow shows the genomic region that includes the nudE gene, its neighborCG6707, and the P-element insertion P{Epgy2}EY09537. From left to right ofthe image corresponds to the telomere-to-centromere direction alongchromosome 3L. Exons are shaded boxes; the open reading frame is in black,and the 5� and 3� untranslated regions are gray. The deletion mutantDf(3L)nudE39A (abbreviated in the text as nudE39A) was created by impreciseexcision of P{Epgy2}EY0953; the parentheses indicate the extent of deletedDNA.

Jour

nal o

f Cel

l Sci

ence

1749NudE in Drosophila cell division

homozygous for a null mutation for nudE (described below), onlyvery weak, unlocalized signals were observed (supplementarymaterial Fig. S1), verifying that the structures observed with ouranti-NudE antibody are specific for the NudE protein.

Distribution of NudE during male meiosisBecause meiosis in Drosophila males is particularly amenable tocytological analysis (Cenci et al., 1994), we also investigated thedistribution of NudE in spermatocytes (Fig. 2). During meioticprophase, the protein is found at the nuclear envelope. Localizationto kinetochores and (presumably) KMTs was seen duringprometaphase or metaphase of meiosis and is similar to that seenat the equivalent mitotic stages; some diffuse spindle staining wasalso observed. Remarkably, the spindle staining during metaphaseI and anaphase I was restricted to the ‘spindle envelope’ and themicrotubules surrounded by this structure. The spindle envelopeis formed in male meiosis in many invertebrates, and consistsof endoplasmic-reticulum-derived parafusorial membranessurrounding both the nuclei and part of the spindle apparatus (Fuller,1993; Giansanti et al., 2007; Inoue et al., 2004; Wolf, 1995) (alsosee A. D. Tates, Cytodifferentiation during spermatogenesis inDrosophila melanogaster: an electron microscope study, PhDthesis, Rijksuniversiteit de Leiden, 1971). Little if any NudE signalwas visible at the kinetochores during anaphase I. Finally, attelophase I, NudE was targeted strongly to the nuclear envelope.We also observed a NudE signal at the equator of late telophases.However, this localization is probably unrelated to cell cleavage,because nudE mutant spermatocytes are not defective in cytokinesis.We saw no staining of any discrete structure in nudE null mutantspermatocytes (data not shown), attesting to the specificity of theNudE antibody staining seen in wild-type testes.

Phenotypic effects of NudE depletionTo explore the consequences of the absence of NudE, we generatedthe null allele nudE39A by imprecise excision of a nearby P element.This deletion entirely removes the first three exons of nudE, pluspart of the fourth exon (Fig. 1C). As a result, the first 37 codonsof the gene, including the initiation codon, are missing, indicatingthat nudE39A cannot encode NudE protein. The conclusion thatnudE39A is a null allele is further supported by the results of westernblots of extracts from mutant larvae (Fig. 1A), and by the absenceof discrete NudE immunofluorescence signals in nudE39A larvalbrains and testes (supplementary material Fig. S1). Animalshomozygous for the nudE39A mutation die as third instar larvae oras pupae, demonstrating that nudE is an essential gene inDrosophila. We presume that nudE39A homozygous zygotes surviveuntil late larval or pupal stages by using maternally deposited NudEprotein from their heterozygous mothers, as is the case for manycell cycle genes in flies (Gatti and Goldberg, 2001).

The mitotic progression of many cells in nudE39A larval brainsis arrested at prometaphase or metaphase. In orcein-stained nudE39A

mutant brains, the mitotic index (the ratio of mitotic cells tomicroscopic fields) is substantially elevated relative to that in wild-type brains (Table 1). The large majority of mutant mitotic cellswere in prometaphase or metaphase, and very few cells continueto anaphase. Chromosome overcondensation, an expectedconsequence of mitotic arrest (Gatti and Goldberg, 1991), wasobserved in most mutant brain cells (Table 1). The frequencies ofall these abnormalities were roughly equivalent in the brains ofnudE39A homozygotes and in the brains of larvae heterozygous fornudE39A and a chromosomal deletion [Df(3L)AC1] that removes

the nudE gene (Table 1). This latter result verifies that the phenotypeassociated with homozygosity for the nudE39A mutation is due tolesions at the nudE gene rather than elsewhere in the genome, andfurther indicates that this allele fits the classical genetic definitionof a null mutation.

The metaphase arrest seen in nudE39A mutant larval brainsprobably reflects activation of the SAC, which prevents cells frominitiating anaphase until all the chromosomes are properly alignedat the metaphase plate (reviewed by Malmanche et al., 2006). Thisconclusion arises from analysis of nudE39A zwilch1229 doublemutants, where zwilch encodes an RZZ component necessary forSAC function (Williams et al., 2003). In the double mutants, themitotic index and the proportion of mitotic figures in anaphase

Fig. 2. Distribution of NudE protein in wild-type spermatocytes. NudE ismostly localized to the nuclear envelope during meiosis, except when thenuclear envelope is disrupted. NudE accumulates on the kinetochores duringmeiotic prometaphase or metaphase; some ‘streaming’ along KMTs canoccasionally be seen (second row). During meiotic anaphase, NudE staining isrestricted to the spindle envelope region; note that the spindle itself extendsbeyond the spindle envelope toward the cell periphery. Scale bar: 10 μm.

Jour

nal o

f Cel

l Sci

ence

1750

relative to those in prometaphase or metaphase are nearly normal(Table 1). Thus, depletion of NudE causes mitotic arrest only whenthe SAC is operational.

Visualization of spindles with antibody against tubulin revealedthat nudE mutations also cause defects in mitotic spindlemorphology and chromosome congression. In contrast to the wildtype, where essentially all spindles appear to be normal, only a small

minority of nudE39A cells had straight, bipolar spindles (Fig. 3; Table2). The most frequent defects were broad, unfocused spindle poles,and, as seen by staining for the centrosomal component DSpd-2(Giansanti et al., 2008), the failure of centrosomes to remain attachedto the spindles. Smaller proportions of nudE39A brain cells showedother defects, including curved spindles, the presence of twocentrosomes at one pole but none at the other pole, andsupernumerary centrosomes. In the large majority of nudE39A

mitoses, the chromosomes failed to form a tight metaphase plate,suggesting problems in congression (Fig. 3). By countingprometaphase or metaphase cells with chromosomes that werealigned or unaligned, we verified this idea. In wild-type controls,42% of cells at this cell cycle stage (n=176) had uncongressedchromosomes, whereas 61% of such cells in nudE39A mutants(n=189) showed this phenotype; this difference is significant(P<0.0001).

Centrosome misbehavior in nudE mutant spermatocytesThe detachment of centrosomes from the spindle is also a significantfeature of meiosis in the testes of nudE39A males (Fig. 4). 88% ofcells at metaphase I showed spindle-detachment defects (n=77),whereas this phenotype was not observed in any wild-typespermatocytes at the same stage (n=46). Time-lapse movies revealedthe likely basis for the failure of centrosomes to associate with thespindle poles in nudE mutant spermatocytes (Fig. 5). Normally, justbefore the first meiotic division, the centrosomes located at oneside of the spermatocyte cortex migrate towards the nuclearenvelope; concomitantly they separate from each other and start tomigrate to the opposite cell poles (Gunsalus et al., 1995; Rebolloet al., 2004) (also see A. D. Tates, Cytodifferentiation duringspermatogenesis in Drosophila melanogaster: an electronmicroscope study, PhD thesis, Rijksuniversiteit de Leiden, 1971).(This migration does not occur in neuroblasts, because thecentrosomes are never associated with the plasma membrane.)However, in nudE mutants, the centrosomes do not move from thecortex to the nuclear envelope. Failure of this centrosomalmovement could easily produce the centrosomal detachmentphenotype shown in Fig. 4, if a bipolar spindle eventually formsaround the chromosomes at a location far from the cortex. Indeed,mutant metaphase I spermatocytes often exhibited spindle-likebipolar arrays of microtubules assembled around the chromosomesbut not connected to the membrane-bound asters (Figs 4 and 5;

Journal of Cell Science 122 (11)

Table 1. Mitotic parameters of nudE mutant larval brains

Genotype No. fields* Metaphase† Anaphase‡Metaphase/anaphase§ %OC¶ MI**

RelativeMI††

OreR (wild type) 630 1100 271 4.1 0.3 2.18 1.00nudE39A/nudE39A 936 6055 53 114.2 89.4 6.53 3.00nudE39A/Df(3L)AC1 255 1131 13 87.0 96.9 4.49 2.06

OreR (wild type) 690 266 74 3.6 ND 0.49 1.00nudE39A/nudE39A 227 434 6 72.1 ND 1.94 3.96

nudE39A zwilch1229/nudE39A zwilch1229 486 263 46 5.7 ND 0.63 1.29

*Number of microscopic fields viewed.†Number of cells in prometaphase or metaphase.‡Number of cells in anaphase or telophase.§Ratio of cells in prometaphase or metaphase to those in anaphase or telophase.¶Percentage of mitotic cells with overcondensed chromosomes. ND indicates not determined.**Mitotic index, the number of mitotic figures per microscopic field. The discrepancy between the MI in the two groups of samples resulted from the use of

microscopes with different-sized fields.††Relative mitotic index, with wild type set equal to 1.00.

Fig. 3. Spindle abnormalities in nudE mutant neuroblasts. The spindles ofwild-type (Oregon R) metaphase cells are bipolar, with each pole narrowlyfocused at a single centrosome (labeled by DSpd-2). The DNA forms a tightmetaphase plate equidistant from the poles. In nudE39A homozygous mutantneuroblasts, the poles are often wide and unfocused, and the centrosomes arefrequently found separated from the poles. Chromosome congression to themetaphase plate is aberrant. The presence of supernumerary centrosomes(bottom row) is also sometimes observed. Scale bar: 5 μm.

Jour

nal o

f Cel

l Sci

ence

1751NudE in Drosophila cell division

supplementary material Fig. S10). This ‘spindle-in-the-spindle’phenotype, which has been previously observed in asp mutants(Rebollo et al., 2004), was also found in lis1, but not in cenp-meta(cmet), mutants (Fig. 4).

In accordance with previous findings that the SAC is relativelyinactive in Drosophila spermatocytes (Basu et al., 1999; Rebollo

and Gonzalez, 2000), meiotic progression was not adverselyaffected in nudE39A mutant testes; the proportions of cells atanaphase or telophase are normal and there is no failure incytokinesis (data not shown). However, chromosome segregationis aberrant in many mutant cells at late anaphase and telophase,with DNA apportioned unequally to the two daughter cells, or withDNA remaining in the middle of the spindle (Table 3; supplementarymaterial Fig. S2). Owing to these difficulties in chromosomesegregation, ‘onion stage’ spermatids displayed nuclei of differentsizes. In addition, the sizes of the Nebenkern (mitochondrialderivatives) were uneven, suggesting that the cleavage furrows arenot always symmetrically positioned during the male meioticdivisions (supplementary material Fig. S2).

Interactions of NudE with other kinetochore componentsWe used a tandem-affinity purification (TAP-tagging) approach todetermine whether NudE associates with other proteins in stable,soluble complexes (Puig et al., 2001; Veraksa et al., 2005; Williamset al., 2007). Our results (supplementary material Fig. S3) clearlydemonstrate a strong association between NudE and Lis1, a findinganticipated by previous results in other organisms (Efimov andMorris, 2000; Feng et al., 2000; Kitagawa et al., 2000; Sasaki etal., 2000; Sweeney et al., 2001). The complex of NudE and Lis1formed both in normally growing tissue culture cells and in cellstreated with colchicine to enrich for those in mitosis. Using massspectrometry and western blotting, we asked whether NudE isolatedby TAP-tagging copurified with any of several other candidates,including ZW10, p50 dynamitin and Glued (components of thedynactin complex), dynein heavy chain and Cenp-meta. Exceptinga very weak signal seen on western blots using an antibody againstp50 dynamitin, the results failed to demonstrate the association ofNudE with any known kinetochore protein other than Lis1.

The availability of a NudE antibody allowed us to ask whetherthe targeting of NudE to kinetochores or to KMTs occurs in larvalbrain cells homozygous for loss-of-function alleles of genesencoding other outer kinetochore components. Because of the yeasttwo-hybrid interaction we had previously observed between NudEand ZW10 (Starr et al., 2000), we first examined NudE localizationin cells lacking ZW10 or its associated RZZ subunits Rod andZwilch. We found that NudE was targeted to the kinetochores inthe larval brains of animals homozygous for mutations in zw10,rod and zwilch. However, extensive observations failed to show the‘streaming’ of NudE onto the KMTs at any time betweenprometaphase and anaphase (Fig. 6). NudE exhibited the samebehavior (kinetochore targeting but lack of streaming) in mutantsfor dynein heavy chain (dhc64) and for NudC, a protein that formscomplexes with Lis1 (Cunniff et al., 1997; Morris et al., 1998)(supplementary material Fig. S4).

We obtained similar findings in mutants for the SAC componentBubR1 (supplementary material Fig. S4): although NudEaccumulated strongly at kinetochores during prometaphase and

Table 2. Mitotic aberrations in metaphase brain cells

Genotype

Normal(bipolar,straight

spindles)

Bipolar,curved

spindles

Bipolar,unfocused

polesBipolar, unattached

centrosomes

Bipolar, twocentrosomes on

half spindle

Threespindlepoles

No bipolarspindle

OreR (wild type) 103 0 5 2 0 0 0nudE39A/nudE39A 7 6 24* 36 4 1 1

The number of each phenotype counted in each class is given. *Two of the 24 cells in this class had three centrosomes; the remainder had two centrosomes.

Fig. 4. Prometaphase or metaphase I spindles in nudE, lis1 and cenp-metamutant spermatocytes. Spermatocytes from nudE39A and lis1K13201

homozygous mutant larvae show centrosomes [labeled with centrosomin(Cnn)] (Li and Kaufman, 1996) detached from the meiotic spindle; bycontrast, cenp-meta (cmetΔ) mutant spermatocytes form meiotic spindles withcentrosomes at the poles that are indistinguishable from spindles in the wild-type (Oregon-R) controls. Scale bar: 10 μm.

Jour

nal o

f Cel

l Sci

ence

1752

metaphase, we did not observe streaming in any metaphase cells(n>100). In fact, in contrast with the wild type, NudE often (63%of 30 cells) remained at kinetochores in bubR1 mutant cells duringanaphase and telophase (data not shown). Because the shedding ofany kinetochore component has not previously been reported to bedefective in mutants for SAC components, we examined bubR1mutant cells in more detail. Surprisingly, these cells had aberrantspindles that were very short and only thinly populated bymicrotubules. It is possible that these spindle defects might beresponsible for the lack of NudE streaming. However, because wealso found that Rod streams normally in bubR1 metaphases (datanot shown), shedding per se is not blocked by the lack of BubR1.

We found that mutations in three genes – bub3, lis1 and cenp-meta (cmet) – disrupted the targeting of NudE to the kinetochoresand KMTs in larval neuroblasts (Fig. 7) (Lei and Warrior, 2000;Lopes et al., 2005; Yucel et al., 2000). In the lis1 and bub3 mutants,some cells in the brain exhibited weak NudE kinetochore staining,whereas other cells showed no staining. However, in both cases,the proportion of NudE-negative cells increased with larval age (to>90% in the oldest larvae for both mutants). These results indicatethat the low NudE signals observed in younger third instar larvaeare caused by the perdurance of maternally supplied gene products,and suggest that Lis1 and Bub3 are needed for the recruitment ofNudE to kinetochores. The lack of NudE staining in cenp-metamutants was clear-cut (even in colchicine-treated brains; data notshown) and unanticipated, because no interactions between theseproteins have previously been reported.

To verify that the kinetochore targeting of NudE depends uponLis1 and Cenp-meta, we examined NudE localization in the testesof animals with mutations in these genes (supplementary materialFig. S5). NudE staining was not observed at the kinetochores inlis1 and cenp-meta mutant spermatocytes. Of particular interest,

NudE was strongly redirected to the vicinity of the spindle polesin the absence of Cenp-meta (compare with the wild-type imagesin Fig. 2). In lis1 mutant spermatocytes, NudE accumulatedanomalously in patches on the spindle envelope, particularly in theregion closest to the chromosomes.

Targeting of other kinetochore proteins in nudE mutantsWe examined the effects of homozygosity for the nudE39A nullmutation on the localization of other outer kinetochore components.

Journal of Cell Science 122 (11)

Fig. 5. Meiotic spindle formation in wild-type and nudEmutant spermatocytes. Images are taken from time-lapsemovies of control (OR) and nudE39A homozygous mutantlarval spermatocytes expressing β-tubulin-GFP (time inminutes). In the wild-type panels, the centrosomes havealready reached the nuclear envelope at the start of filming.By contrast, the centrosomes fail to migrate from the plasmamembrane to the nuclear envelope in nudE spermatocytes. Asa result, a spindle-like microtubule array forms around thechromosomes, while the asters (arrows) remain associatedwith the plasma membrane. Scale bar: 10 μm.

Table 3. DNA segregation defects in spermatocytes

Genotype

NormalDNA

segregation

UnequalDNA

segregation

UnsegregatedDNA at the

spindleequator

DNAbridge

OreR (wild type) Meiosis I 107 0 1 0 Meiosis II 74 0 0 0nudE39A/nudE39A

Meiosis I 9 7 6 4 Meiosis II 1 7 0 1cmet /cmet Meiosis I 12 24 0 2 Meiosis II 26 1 0 0

Fig. 6. NudE protein is targeted to kinetochores in RZZ mutants, but fails tostream onto KMTs. Prometaphase or metaphase cells from zw10S1, rodX5 andzwilch1229 mutant brains are shown. CID labels the centromeres. Seesupplementary material Fig. S1 for corresponding wild-type controls. Scalebar: 5 μm.

Jour

nal o

f Cel

l Sci

ence

1753NudE in Drosophila cell division

Dynein heavy chain (Dhc), as well as the checkpoint proteins Bub3,Rod, ZW10 and Zwilch, all still associated with kinetochores inbrain cells from nudE mutant third instar larvae (supplementarymaterial Fig. S6; data not shown); the proper localization of SACcomponents to kinetochores is expected since the SAC is functionalin nudE mutant neuroblasts (see above). However, we failed toobserve the ‘streaming’ of Rod (Fig. 8), or of ZW10 or Zwilch (datanot shown) onto the KMTs, even though such movement is visiblein wild-type controls, specifically during metaphase (Fig. 8)(Williams et al., 2003). Since Dhc could be visualized at thekinetochores only in the presence of colchicine (which causesspindle disassembly) (supplementary material Fig. S6), we couldnot assess whether Dhc streaming was also abrogated in the nudEmutant cells. In nudE39A larval brains, the kinesin-like protein Cenp-meta and the SAC component Bub3 were targeted to thekinetochores (Fig. 9A and data not shown). These proteins nevermoved poleward along KMTs in nudE mutant cells, but neither ourown observations nor the published literature have revealed clearevidence for the streaming of Bub3 or of Cenp-meta or itsmammalian homolog CENP-E during mitosis in wild-type cells(Basu et al., 1998; Yao et al., 2000; Yucel et al., 2000).

Because the behavior of Cenp-meta in spermatocytes has notpreviously been described, we stained wild-type spermatocytes with

antibody we generated against Cenp-meta (supplementary materialFig. S7). The localization of Cenp-meta in these meiotic cells wasvery similar to that shown above for NudE: Cenp-meta associatedwith the nuclear envelope, the kinetochores and the spindle envelopeat equivalent times during meiosis. However, we did not observestreaming of Cenp-meta along the KMTs in spermatocytes duringmetaphase I; the signals for Cenp-meta on the kinetochores werevery clear, so streaming should have been visible if it does occur.With these baseline observations, we could then investigate thebehavior of Cenp-meta in nudE mutant larval testes. The behaviorof Cenp-meta in spermatocytes was different from that seen in larvalbrains of the same animals, because it failed to localize to thekinetochores or to the KMTs (Fig. 9B). To confirm this unexpectedfinding, we examined spermatocytes incubated with colchicine, atreatment that causes many proteins to accumulate at abnormallyhigh levels at centromeres. Even under these conditions, Cenp-metawas absent from kinetochores in ~70% (n=24) of mutant cells,whereas the remaining spermatocytes displayed only very weakkinetochore signals (supplementary material Fig. S8). Remarkably,Cenp-meta is redirected in untreated nudE spermatocytes to thespindle envelope, in a pattern very similar to that noted for NudEprotein in lis1 mutants (see supplementary material Fig. S5).

Asp localization in mitotic and meiotic spindles is independentof NudE functionBecause the centrosomal aspects of the nudE mutant phenotype arereminiscent of aberrations reported in asp mutants (Avides andGlover, 1999; Wakefield et al., 2001; Rebollo et al., 2004), we askedwhether NudE is required for proper Asp localization. We foundthat the Asp protein is properly localized in both mitotic and malemeiotic cells of nudE mutants. Asp is enriched at the spindle polesduring mitotic prophase in nudE39A neuroblasts (supplementary

Fig. 7. NudE protein fails to localize to the kinetochores of chromosomesduring prometaphase or metaphase in bub31, lis1K13201 and cmetΔ mutants.Very weak NudE staining in young bub3 mutant third instar larvae is visible inthe top row of the bub3 figures. See supplementary material Fig. S1 for wild-type controls. Scale bar: 5 μm.

Fig. 8. Rod is targeted to kinetochores, but fails to ‘stream’ onto KMTs innudE39A homozygous mutant neuroblasts. The top row shows the streaming ofRod from the kinetochores towards the poles in a wild-type (OR) metaphasecell. The two bottom rows illustrate the lack of streaming in nudE39A

metaphases. Scale bar: 5 μm.

Jour

nal o

f Cel

l Sci

ence

1754

material Fig. S9). In mutant prometaphase or metaphase neuroblasts,Asp was still associated with microtubule minus-ends, but itsdistribution was more diffuse, because the spindle poles wereunfocused. The spindle poles of the rare anaphases in nudE mutantbrains showed normal Asp accumulation (supplementary materialFig. S9). Asp was also found at the spindle poles in nudE39A

spermatocytes throughout meiosis, although the spindle poles wereless compact than in the wild type (supplementary material Fig.S10). Collectively, these results indicate that the aberrant centrosome

phenotype caused by nudE mutations is not due to any inability ofAsp to localize to the spindle poles.

DiscussionDuring various stages of the cell cycle in Drosophila, NudEassociates with kinetochores, KMTs and possibly other parts of thespindle apparatus, and (during meiosis) the nuclear envelope (Fig.1B; Fig. 2; supplementary material Fig. S1). A null mutation in thenudE gene disrupts cell cycle progression and the behavior of thecentrosomes, spindles and chromosomes during mitotic and meioticdivisions (Figs 3, 4 and 5; supplementary material Fig. S2).

NudE, centrosomes and the nuclear envelopeNeuroblasts and spermatocytes differ in the process of centrosomeseparation. During the spermatocyte growth phase, the centrosomesmigrate to the periphery of the cell and associate with the plasmamembrane. Shortly before the onset of the first meiotic division,the centrosomes together move towards the nucleus while nucleatingastral microtubules; the centrosomes then rotate around the nucleusuntil they reach the opposite cell poles (Baker et al., 2004; Fuller,1993; Rebollo et al., 2004) (also see A. D. Tates, Cytodifferentiationduring spermatogenesis in Drosophila melanogaster: an electronmicroscope study, PhD thesis, Rijksuniversiteit de Leiden, 1971).By contrast, neuroblast centrosomes separate during interphase, andthere is no evidence that these centrosomes associate with the plasmamembrane. By the time the centrosomes start nucleating astralmicrotubules in early prophase, they have already achieved ~70%of their final separation. Centrosome separation is then completedduring late prophase or prometaphase with the formation of a bipolarspindle with a centrosome at each pole (Rebollo et al., 2004; Rusanand Peifer, 2007).

Despite these differences in the normal mechanisms ofcentrosome separation, centrosomes in nudE mutant neuroblasts andspermatocytes are both detached from the spindle poles (Figs 3 and4; Table 2). Centrosome detachment from the spindle poles has beenpreviously observed in neuroblasts from lis1, glued and dyneinheavy chain (dhc64) mutants (Robinson et al., 1999; Siller et al.,2005; Wojcik et al., 2001), as well as in neuroblasts defective forAsp (Wakefield et al., 2001). In nudE mutant spermatocytes,centrosome detachment might be a secondary consequence of thetendency of centrosomes to remain associated with the plasmamembrane and thus fail to migrate towards the nucleus (Figs 4 and5). We also observe persistent association of the centrosomes withthe spermatocyte plasma membrane in lis1 mutants (Fig. 4); thissame phenomenon was also previously reported in asp mutant testes(Rebollo et al., 2004; Wakefield et al., 2001).

One scenario that could provide a common explanation for theproblems of centrosome behavior in mutant neuroblasts andspermatocytes is that NudE and other dynein-associated proteins,as well as the microtubule-associated protein Asp, are all requiredto attach microtubule minus-ends to the centrosome. Roles forDrosophila Lis1 and dynactin in microtubule-centrosomeinteractions were proposed in a previous study (Siller et al., 2005).In addition, it has been suggested that Asp might have the samefunction as vertebrate NuMa (Wakefield et al., 2001), which isthought to anchor microtubule minus-ends at the centrosome(reviewed by Fant et al., 2004). Weakening of the connectionsbetween the centrosomes and the microtubule minus-ends could intheory lead to centrosome detachment from the spindle poles.Unfortunately, this attractive hypothesis is not supported by ourNudE immunolocalization results, because we do not see an

Journal of Cell Science 122 (11)

Fig. 9. Cenp-meta behavior in nudE mutants. (A) Cenp-meta associates withkinetochores in larval neuroblasts in the absence of NudE. Streaming of Cenp-meta along the KMTs is not seen either in wild-type (Oregon-R) or in nudEmutant brain cells. Scale bar: 5 μm. (B) Cenp-meta is not targeted tokinetochores in nudE mutant spermatocytes during prometaphase I ormetaphase I. Cenp-meta instead accumulates aberrantly in the middle regionof the spindle envelope, similarly to NudE localization in lis1 mutantspermatocytes (compare with supplementary material Fig. S5). Merged panelswithout tubulin staining are provided to verify that Cenp-meta does notassociate with kinetochores in nudE spermatocytes. Cenp-meta targeting to thenuclear membrane does not require NudE either during meiotic prophase (datanot shown) or during telophase (bottom row). Scale bar: 10 μm.

Jour

nal o

f Cel

l Sci

ence

1755NudE in Drosophila cell division

accumulation of NudE near centrosomes in either neuroblasts orspermatocytes. However, NudE streams along the kinetochoremicrotubules in both spermatocytes and neuroblasts, so the fact thatNudE helps mediate interactions between the minus ends of spindlemicrotubules and centrosomes cannot be excluded.

The strong NudE localization we observed at the spermatocytenuclear envelope (Fig. 2) provides another possible explanation forthe nudE mutant phenotype. In spermatocytes lacking NudE, theastral microtubules could fail to interact properly with the nuclearenvelope, disrupting forces that would normally pull centrosomesaway from the plasma membrane towards the nucleus. Aconceptually similar mechanism involving dynein-mediatedmicrotubule interactions with the cell cortex (rather than the nuclearenvelope) determines the orientation and asymmetrical localizationof the mitotic spindle during the first embryonic division ofCaenorhabditis elegans (Gonczy and Rose, 2005). A defectiveinteraction between the astral microtubules and the nuclear envelopecould also explain the abnormal centrosome localization in nudEmutant neuroblasts. However, since we did not observe NudEaccumulation at the neuroblast nuclear envelope, the current datado not allow a precise definition of the role of NudE in centrosomepositioning.

NudE, kinetochores, KMTs and the spindle checkpointThe absence of NudE from neuroblasts results in failure ofchromosomes to congress to a tightly organized metaphase plate,even though cell cycle progression is arrested at prometaphase ormetaphase (Fig. 3; Table 1). Stehman and co-workers (Stehman etal., 2007) have recently reported similar findings in mammaliantissue culture cells injected with an antibody against NudE/Nudelthat apparently blocks the interactions of these proteins withcytoplasmic dynein at kinetochores. Their detailed analysis showedthat the majority of injected cells had one or more misorientedkinetochore pairs, and that misorientation was usually associatedwith the absence of microtubule attachments at these kinetochores.In accordance with such results, it seems logical to assume that thecongression problems we observed result from the failure of NudEfunction specifically at kinetochores. It should be stressed, however,that the majority of chromosomes in both their study and ours domigrate to the metaphase plate, therefore NudE is not essential forestablishing most kinetochore microtubule attachments.

Mitotic arrest in nudE mutant neuroblasts is due to persistentactivation of the SAC (Table 1), therefore NudE cannot be requiredfor SAC function. Because NudE is present at multiple locationsand affects the mitotic spindle both in terms of microtubuleorganization and the ability of chromosomes to congress to themetaphase plate, it is difficult to pinpoint the precise reason thespindle checkpoint remains ‘on’ in these cells. One possibility isthat the checkpoint machinery detects the absence of one or moreconnections between kinetochores and KMTs in mutant cells,because, as mentioned above, anti-NudE/Nudel antibody appearsto interfere with these connections in mammalian cells (Stehmanet al., 2007). A second possibility is that generalized spindle defectsor defective kinetochore connections might reduce the tensionexerted across the chromosomes. In support of this idea, the averagedistance between sister kinetochores is shorter than normal inmammalian tissue culture cells injected with NudE/Nudel antibody(Stehman et al., 2007) and in Drosophila larval neuroblasts mutantfor lis1 (Siller et al., 2005). Since only a small minority of mitoticcells in nudE mutant brains had well-defined metaphase plates, wedid not make the same type of measurements, but our limited

observations support the idea of reduced tension acrosschromosomes.

The persistence of the metaphase arrest in nudE mutantneuroblasts might also be due to a third factor: nudE mutationsappear to block the ‘streaming’ or ‘shedding’ of outer kinetochorecomponents along the KMTs. Current models suggest that streamingis dependent on cytoplasmic dynein activity and serves to moveSAC proteins off the kinetochores, thus turning off the spindlecheckpoint and allowing the initiation of anaphase (reviewed byHowell et al., 2001; Wojcik et al., 2001). In support of the idea thatNudE streaming is part of the same mechanism, NudE fails to streamin neuroblasts mutant for the abnormal spindle (asp) gene(supplementary material Fig. S4). These cells display spindledefects that cause metaphase arrest, therefore SAC signalingremains on for extended periods (Basto et al., 2000). Neither ourobservations nor those made previously truly discriminate betweenthese three hypotheses for the metaphase arrest seen when dynein-associated proteins are removed or inactivated. Since any of thethree possibilities would be sufficient, a combination of thesemechanisms might contribute to the metaphase arrest in nudE mutantbrain cells.

NudE and its partners Lis1 and cytoplasmic dyneinOur TAP-tagging results indicate that NudE and Lis1 form a stable,soluble complex; furthermore, the nudE brain phenotypes describedhere affecting spindle structure, chromosome congression, thespindle checkpoint, and RZZ shedding are virtually identical to thosepreviously reported for lis1 mutant Drosophila neuroblasts (Silleret al., 2005). We have observed that Lis1 is needed for thekinetochore localization of NudE; although we have not been ableto detect reliably Lis1 at kinetochores using a variety of antibodies,it would not be surprising if Lis1 targeting to kinetochoresreciprocally requires NudE. In our TAP-tagging experiments, wehave seen no evidence for the association of NudE (or Lis1) withdynein, dynactin, NudC or ZW10, but others have reported co-immunoprecipitation or yeast two-hybrid results indicative of suchinteractions (Cunniff et al., 1997; Liu et al., 2000; Niethammer etal., 2000; Sasaki et al., 2000; Starr et al., 2000; Tai et al., 2002).Our inability to find other partners of NudE might be caused bythe disruption of important NudE domains in the TAP-taggingconstruct we used, but we believe it is more likely that theinteractions of NudE with these other proteins are simply weakeror less stable than its relationship with Lis1.

Our results do not agree with two recent reports in mammaliancells that NudE is required for the kinetochore localization ofcytoplasmic dynein (Stehman et al., 2007; Vergnolle and Taylor,2007). It is possible that dynein at the kinetochore is significantlyreduced in Drosophila nudE mutant cells, but that we were unableto observe this because we could only visualize Dhc at thekinetochores in cells exposed to colchicine (a treatment thatenhances the immunofluorescence signals of many kinetochoreproteins, presumably by interfering with shedding) (supplementarymaterial Fig. S6). Another possibility is that many intermolecularcontacts are required for the kinetochore targeting of dynein andother kinetochore components, so the dependencies might vary indifferent cell types or in different organisms. For example, dyneindoes not associate with the kinetochore in flies carrying zw10mutations (Starr et al., 1998); in a second example, the kinetochorerecruitments of ZW10 and NudE are mutually independent (thispaper) in spite of their apparent physical association as assayed inthe yeast two-hybrid system (Starr et al., 2000). In any event, the

Jour

nal o

f Cel

l Sci

ence

1756

lack of streaming observed in nudE mutant neuroblasts andspermatocytes coupled with the presence of at least some Dhc atthe kinetochores in these cells implies that NudE might affect notonly dynein targeting to the kinetochore, but also its function atthese structures.

NudE and Cenp-metaOur results provide clear evidence for a close but unanticipatedrelationship at the kinetochore between NudE and Cenp-meta, oneof two Drosophila homologs of the mammalian kinesin-like proteinCENP-E. NudE fails to target kinetochores in cenp-meta mutantbrain cells and spermatocytes. In the reciprocal experiments, Cenp-meta associates with kinetochores in nudE mutant brain cells butnot in nudE mutant spermatocytes (Figs 7 and 9; supplementarymaterial Fig. S5), except very weakly in a small percentage ofspermatocytes treated with colchicine (supplementary material Fig.S8). It is presently unclear whether these two proteins contact eachother directly. Our TAP-tagging experiments (supplementarymaterial Fig. S3) provide no evidence for a stable, soluble complexcontaining both NudE and Cenp-meta; however, interactions couldbe weak or occur only in the context of the kinetochore. Oneinteresting possibility is that the interactions between NudE andCenp-meta could be mediated by a Drosophila ortholog of thekinetochore protein CENP-F (also known as mitosin), since thereare reports in the literature that the kinetochore targeting of NudE,Nudel and Lis1 are all dependent on CENP-F (Vergnolle and Taylor,2007), whereas the amounts of CENP-E and cytoplasmic dyneinat the kinetochores are significantly reduced in CENP-F-depletedcells (Yan et al., 2003). Unfortunately, BLAST searches do notreveal the identity of a clear-cut homolog of CENP-F in flies.

Of particular interest, the finding that NudE is not targeted tokinetochores in cenp-meta mutants raises questions about the originof the phenotypes previously described in these mutants. The mitoticaberrations seen in larval neuroblasts that are associated withmutations in both nudE and cenp-meta are similar: Chromosomesexhibit problems in congression to the metaphase plate, and thereis an increase in the mitotic index because of the accumulation ofcells in prometaphase or metaphase (Maia et al., 2007; Williams etal., 2003; Yucel et al., 2000). It is quite possible that many of theproblems in congression previously attributed to the lack of Cenp-meta at the kinetochore are actually caused by the absence of NudE(and thus the misregulation of cytoplasmic dynein). However, oneperplexing attribute of cenp-meta mutants cannot be ascribed to thelack of NudE at the kinetochore. Although the elevated mitotic indexand the delay of anaphase onset in cenp-meta mutant neuroblastsindicate activity of the SAC, depolymerization of the spindle in themutant cells with colchicine or nocodazole leads to precocious sisterchromatid separation and other events indicative of the failure ofSAC signaling (Williams et al., 2003). The reason for thisparadoxical behavior of the SAC in cenp-meta mutants is notcurrently understood, but since colchicine-treated nudE mutant cellsremain arrested in prometaphase or metaphase with the sisterchromatids of their chromosomes remaining attached, the failureof SAC signaling in similarly treated cenp-meta larval neuroblastscannot be due to the absence of NudE from the kinetochores.

Materials and MethodsGenetic stocks and manipulationsDrosophila melanogaster stocks were raised on yeast-glucose-agar medium at23±2°C in a 12 hour L:12 hour D photoperiod. Oregon R was the wild-type strain.The deletion Df(3L)AC1 (67A2-67D3) was obtained from the Drosophila stock center(Bloomington IN; stock no. 997). Strains containing the mutations zw10S1, rodX5,

zwilch1229, dynein heavy chain (Dhc64C6-10), bub31, bubR11, lis1K13201, cmetΔ (a nullallele of cenp-meta), and asp1 have been previously described (Basu et al., 1999; Leiand Warrior, 2000; Lopes et al., 2005; McGrail et al., 1995; Ripoll et al., 1985; Scaerouet al., 1999; Siller et al., 2005; Williams et al., 1992; Williams et al., 2003; Yucel etal., 2000). The nudC9jE8 allele was the kind gift of Rahul Warrior (University ofCalifornia, Irvine, CA).

To generate a null allele of nudE, imprecise excisions were generated from thehomozygous viable P element insertion P{EPgy2}EY09537 (~600 bp upstream ofthe nudE gene) using standard techniques (e.g. Williams et al., 2007). We verifiedthe extent of the deletion in the strain now known as nudE39E by direct DNAsequencing of a PCR product spanning the excision.

Immunostaining and microscopy of mitotic and meiotic figuresThird instar larval brains as well as testes from third instar larvae or early pupae werefixed using formaldehyde and stained as previously described (Bonaccorsi et al., 2000;Cenci et al., 1994; Williams et al., 1992; Williams et al., 2003). In some cases, slideswere exposed to an additional incubation of goat serum before the addition of secondaryantibody so as to lower background. Primary antibodies used include anti-α-tubulin(Sigma Aldrich, St Louis, MO) at a 1:1000 dilution, purified anti-NudE at 1:200 (thispaper; see below), anti-DSpd-2 [a centrosomal marker (Giansanti et al., 2008)] at1:2000, anti-Rod preparation BE40 at 1:300 [a gift from Roger Karess (Scaerou etal., 2001)]; anti-Cenp-meta at 1:100 (see below), and anti-centrosomin (CNN) at 1:1000(Megraw et al., 1999). Secondary antibodies used include FITC-conjugated anti-rabbitIgG and IgM at a 1:20 dilution (Jackson ImmunoResearch, West Grove, PA) and AlexaFluor 555-conjugated anti-rabbit IgG at 1:300 (Invitrogen, Carlsbad, CA).

For time-lapse imaging studies of spermatocytes, testes from third instar larvae orearly pupae were prepared as described by Inoue and colleagues (Inoue et al., 2004),using flies harboring a β-tubulin-GFP construct also described in the same reference(a kind gift from Matthew Savoian, Cancer Research UK, Cambridge, UK). Theimaging of living onion-stage spermatids (supplementary material Fig. S2) wasaccording to published methods (Regan and Fuller, 1990).

To make anti-NudE, the entire coding region was PCR-amplified from the cDNAclone LD32494 (Berkeley Drosophila Genome Project) and inserted into the vectorpMAL-C2 (New England Biolabs, Beverly, MA) in-frame with the gene for maltosebinding protein (MBP). The resulting NudE-MBP fusion protein was purified asdescribed (Williams et al., 2003), and then injected into rabbits (Cocalico Biologicals,Reamstown, PA). Crude serum was affinity purified against the MBP-NudE fusionprotein coupled to CnBr-activated Sepharose beads. Anti-Cenp-meta was made withthe same constructs and according to the same protocol described (Yucel et al., 2000);injection of the fusion protein and isolation of crude serum was performed by CocalicoBiologicals. Results obtained with this antibody on western blots (not shown) and inimmunofluorescence studies (controls in Fig. 9A and data not shown) were identicalto those previously reported for a different antibody made in the same way (Yucelet al., 2000; Williams et al., 2003).

Purification of NudE-containing complexes by TAP-taggingThe entire coding sequence of nudE was cloned into pMK33-NTAP (Veraksa et al.,2005), so as to fuse a TAP tag in frame to the N-terminus of NudE. This constructwas transfected into Drosophila Kc tissue culture cells using Cellfectin (Invitrogen),and used to generate stable Hygromycin-resistant cell lines expressing TAP-NudE.The production of TAP-NudE was assayed on western blots using HRP-conjugatedanti-Protein-A antibody (Rockland, Gilbertsville, PA). Protein complexes from oneliter of TAP-NudE cells were isolated following established procedures (Puig et al.,2001), using a lysis buffer for making Drosophila extracts (Veraksa et al., 2005).After purification using IgG-Sepharose and calmodulin-Sepharose beads (GEHealthcare, Piscataway, NJ), proteins in the final eluate were precipitated withtrichloroacetic acid, solubilized in Laemmli sample buffer (Bio-Rad, Hercules, CA)and subjected to SDS-PAGE. Bands were excised, digested with trypsin, andanalyzed by matrix-assisted laser desorption/ionization mass spectrometry (MALDI;Cornell Bioresource Center, Ithaca, NY). Samples from the NudE complex purificationwere also analyzed by western blotting for the presence of Lis1, p50 dynamitin, Cenp-meta (Williams et al., 2003), Glued and ZW10 (Williams et al., 1992).

Western blotting of fractions from TAP-tagging experiments were performed usingstandard procedures (Williams et al., 2007) and the following diluted antibodies:purified rabbit anti-NudE at 1:1000 (this paper); rabbit anti-ZW10 1:3000 (Williamset al., 1992); rabbit anti-Cenp-meta 1:6000 (Williams et al., 2003); rabbit anti-Glued1:2000 [(Waterman-Storer and Holzbaur, 1996); a kind gift from Erica Holzbaur(University of Pennsylvania, Philadelphia, PA)]; rat anti-Lis1 1:2000 (from R.Warrior); and rat anti-p50 dynamitin (R. Warrior).

We thank Rahul Warrior and Erica Holzbaur for generously providingantibodies and mutant strains. This work was supported by NIH grantGM48430 to M.L.G. and grants from Centro di Eccellenza di Biologiae Medicina Molecolare (BEMM) to M.G.; A.W. was supported by aEuropean Community Training and Mobility of Researchers grant(HPRN-CT-2002-00260) to M.G. Deposited in PMC for release after12 months.

Journal of Cell Science 122 (11)

Jour

nal o

f Cel

l Sci

ence

1757NudE in Drosophila cell division

ReferencesAvides, M. C. and Glover, D. M. (1999). Abnormal spindle protein, Asp, and the integrity

of mitotic centrosomal microtubule organizing centers. Science 283, 1733-1735.Baker, J. D., Adhikarakunnathu, S. and Kernan, M. J. (2004). Mechanosensory-

defective, male-sterile unc mutants identify a novel basal body protein required forciliogenesis in Drosophila. Development 131, 3411-3422.

Basto, R., Gomes, R. and Karess, R. E. (2000). Rough deal and Zw10 are required forthe metaphase checkpoint in Drosophila. Nat. Cell Biol. 2, 939-943.

Basu, J., Logarinho, E., Herrmann, S., Bousbaa, H., Li, Z., Chan, G. K., Yen, T. J.,Sunkel, C. E. and Goldberg, M. L. (1998). Localization of the Drosophila checkpointcontrol protein Bub3 to the kinetochore requires Bub1 but not Zw10 or Rod. Chromosoma107, 376-385.

Basu, J., Bousbaa, H., Logarinho, E., Li, Z., Williams, B. C., Lopes, C., Sunkel, C. E.and Goldberg, M. L. (1999). Mutations in the essential spindle checkpoint gene bub1cause chromosome missegregation and fail to block apoptosis in Drosophila. J. CellBiol. 146, 13-28.

Blower, M. D. and Karpen, G. H. (2001). The role of Drosophila CID in kinetochoreformation, cell-cycle progression and heterochromatin interactions. Nat. Cell Biol. 3,730-739.

Bonaccorsi, S., Giansanti, M. G. and Gatti, M. (2000). Spindle assembly in Drosophilaneuroblasts and ganglion mother cells. Nat. Cell Biol. 2, 54-56.

Cenci, G., Bonaccorsi, S., Pisano, C., Verni, F. and Gatti, M. (1994). Chromatin andmicrotubule organization during premeiotic, meiotic and early postmeiotic stages ofDrosophila melanogaster spermatogenesis. J. Cell Sci. 107, 3521-3534.

Cunniff, J., Chiu, Y. H., Morris, N. R. and Warrior, R. (1997). Characterization of DnudC,the Drosophila homolog of an Aspergillus gene that functions in nuclear motility. Mech.Dev. 66, 55-68.

Efimov, V. P. and Morris, N. R. (2000). The LIS1-related NUDF protein of Aspergillusnidulans interacts with the coiled-coil domain of the NUDE/RO11 protein. J. Cell Biol.150, 681-688.

Fant, X., Merdes, A. and Haren, L. (2004). Cell and molecular biology of spindle polesand NuMA. Int. Rev. Cytol. 238, 1-57.

Feng, Y., Olson, E. C., Stukenberg, P. T., Flanagan, L. A., Kirschner, M. W. and Walsh,C. A. (2000). LIS1 regulates CNS lamination by interacting with mNudE, a centralcomponent of the centrosome. Neuron 28, 665-679.

Fuller, M. T. (1993). Spermatogenesis. In The Development of Drosophila Melanogaster(ed. M. Bate and A. Martinez-Arias), pp. 71-147. Cold Spring Harbor, NY: Cold SpringHarbor Press.

Gatti, M. and Goldberg, M. L. (1991). Mutations affecting cell division in Drosophila.Methods Cell Biol. 35, 543-586.

Giansanti, M. G., Belloni, G. and Gatti, M. (2007). Rab11 is required for membranetrafficking and actomyosin ring constriction in meiotic cytokinesis of Drosophila males.Mol. Biol. Cell 18, 5034-5047.

Giansanti, M. G., Bucciarelli, E., Bonaccorsi, S. and Gatti, M. (2008). Drosophila SPD-2 is an essential centriole component required for PCM recruitment and astral-microtubule nucleation. Curr. Biol. 18, 303-309.

Gonczy, P. and Rose, L. S. (2005). Asymmetric cell division and axis formation in theembryo. WormBook, 1-20.

Gunsalus, K. C., Bonaccorsi, S., Williams, E., Verni, F., Gatti, M. and Goldberg, M.L. (1995). Mutations in twinstar, a Drosophila gene encoding a cofilin/ADF homologue,result in defects in centrosome migration and cytokinesis. J. Cell Biol. 131, 1243-1259.

Howell, B. J., Hoffman, D. B., Fang, G., Murray, A. W. and Salmon, E. D. (2000).Visualization of Mad2 dynamics at kinetochores, along spindle fibers, and at spindlepoles in living cells. J. Cell Biol. 150, 1233-1250.

Howell, B. J., McEwen, B. F., Canman, J. C., Hoffman, D. B., Farrar, E. M., Rieder,C. L. and Salmon, E. D. (2001). Cytoplasmic dynein/dynactin drives kinetochore proteintransport to the spindle poles and has a role in mitotic spindle checkpoint inactivation.J. Cell Biol. 155, 1159-1172.

Inoue, Y. H., Savoian, M. S., Suzuki, T., Mathe, E., Yamamoto, M. T. and Glover, D.M. (2004). Mutations in orbit/mast reveal that the central spindle is comprised of twomicrotubule populations, those that initiate cleavage and those that propagate furrowingression. J. Cell Biol. 166, 49-60.

Karess, R. (2005). Rod-Zw10-Zwilch: a key player in the spindle checkpoint. Trends CellBiol. 15, 386-392.

Kitagawa, M., Umezu, M., Aoki, J., Koizumi, H., Arai, H. and Inoue, K. (2000). Directassociation of LIS1, the lissencephaly gene product, with a mammalian homologue ofa fungal nuclear distribution protein, rNUDE. FEBS Lett. 479, 57-62.

Lei, Y. and Warrior, R. (2000). The Drosophila Lissencephaly1 (DLis1) gene is requiredfor nuclear migration. Dev. Biol. 226, 57-72.

Li, J., Lee, W. L. and Cooper, J. A. (2005). NudEL targets dynein to microtubule endsthrough LIS1. Nat. Cell Biol. 7, 686-690.

Li, K. and Kaufman, T. C. (1996). The homeotic target gene centrosomin encodes anessential centrosomal component. Cell 85, 585-596.

Liu, Z., Steward, R. and Luo, L. (2000). Drosophila Lis1 is required for neuroblastproliferation, dendritic elaboration and axonal transport. Nat. Cell Biol. 2, 776-783.

Lopes, C. S., Sampaio, P., Williams, B., Goldberg, M. and Sunkel, C. E. (2005). TheDrosophila Bub3 protein is required for the mitotic checkpoint and for normalaccumulation of cyclins during G2 and early stages of mitosis. J. Cell Sci. 118, 187-198.

Maia, A. F., Lopes, C. S. and Sunkel, C. E. (2007). BubR1 and CENP-E have antagonisticeffects upon the stability of microtubule-kinetochore attachments in Drosophila S2 cellmitosis. Cell Cycle 6, 1367-1378.

Malmanche, N., Maia, A. and Sunkel, C. E. (2006). The spindle assembly checkpoint:preventing chromosome mis-segregation during mitosis and meiosis. FEBS Lett. 580,2888-2895.

McGrail, M., Gepner, J., Silvanovich, A., Ludmann, S., Serr, M. and Hays, T. S. (1995).Regulation of cytoplasmic dynein function in vivo by the Drosophila Glued complex.J. Cell Biol. 131, 411-425.

Megraw, T. L., Li, K., Kao, L. R. and Kaufman, T. C. (1999). The centrosomin proteinis required for centrosome assembly and function during cleavage in Drosophila.Development 126, 2829-2839.

Meraldi, P., McAinsh, A. D., Rheinbay, E. and Sorger, P. K. (2006). Phylogenetic andstructural analysis of centromeric DNA and kinetochore proteins. Genome Biol. 7,R23.

Morris, S. M., Albrecht, U., Reiner, O., Eichele, G. and Yu-Lee, L. Y. (1998). Thelissencephaly gene product Lis1, a protein involved in neuronal migration, interacts witha nuclear movement protein, NudC. Curr. Biol. 8, 603-606.

Musacchio, A. and Salmon, E. D. (2007). The spindle-assembly checkpoint in space andtime. Nat. Rev. Mol. Cell. Biol. 8, 379-393.

Niethammer, M., Smith, D. S., Ayala, R., Peng, J., Ko, J., Lee, M. S., Morabito, M.and Tsai, L. H. (2000). NUDEL is a novel Cdk5 substrate that associates with LIS1and cytoplasmic dynein. Neuron 28, 697-711.

Przewloka, M. R., Zhang, W., Costa, P., Archambault, V., D’Avino, P. P., Lilley, K.S., Laue, E. D., McAinsh, A. D. and Glover, D. M. (2007). Molecular analysis ofcore kinetochore composition and assembly in Drosophila melanogaster. PLoS ONE2, e478.

Puig, O., Caspary, F., Rigaut, G., Rutz, B., Bouveret, E., Bragado-Nilsson, E., Wilm,M. and Seraphin, B. (2001). The tandem affinity purification (TAP) method: a generalprocedure of protein complex purification. Methods 24, 218-229.

Rebollo, E. and Gonzalez, C. (2000). Visualizing the spindle checkpoint in Drosophilaspermatocytes. EMBO Rep. 1, 65-70.

Rebollo, E., Llamazares, S., Reina, J. and Gonzalez, C. (2004). Contribution ofnoncentrosomal microtubules to spindle assembly in Drosophila spermatocytes. PLoSBiol. 2, E8.

Regan, C. L. and Fuller, M. T. (1990). Interacting genes that affect microtubule functionin Drosophila melanogaster: two classes of mutation revert the failure to complementbetween haync2 and mutations in tubulin genes. Genetics 125, 77-90.

Ripoll, P., Pimpinelli, S., Valdivia, M. M. and Avila, J. (1985). A cell division mutantof Drosophila with a functionally abnormal spindle. Cell 41, 907-912.

Robinson, J. T., Wojcik, E. J., Sanders, M. A., McGrail, M. and Hays, T. S. (1999).Cytoplasmic dynein is required for the nuclear attachment and migration of centrosomesduring mitosis in Drosophila. J. Cell Biol. 146, 597-608.

Rusan, N. M. and Peifer, M. (2007). A role for a novel centrosome cycle in asymmetriccell division. J. Cell Biol. 177, 13-20.

Sasaki, S., Shionoya, A., Ishida, M., Gambello, M. J., Yingling, J., Wynshaw-Boris,A. and Hirotsune, S. (2000). A LIS1/NUDEL/cytoplasmic dynein heavy chain complexin the developing and adult nervous system. Neuron 28, 681-696.

Savoian, M. S., Goldberg, M. L. and Rieder, C. L. (2000). The rate of polewardchromosome motion is attenuated in Drosophila zw10 and rod mutants. Nat. Cell Biol.2, 948-952.

Scaerou, F., Aguilera, I., Saunders, R., Kane, N., Blottiere, L. and Karess, R. (1999).The rough deal protein is a new kinetochore component required for accurate chromosomesegregation in Drosophila. J. Cell Sci. 112, 3757-3768.

Scaerou, F., Starr, D. A., Piano, F., Papoulas, O., Karess, R. E. and Goldberg, M. L.(2001). The ZW10 and Rough Deal checkpoint proteins function together in a large,evolutionarily conserved complex targeted to the kinetochore. J. Cell Sci. 114, 3103-3114.

Schittenhelm, R. B., Heeger, S., Althoff, F., Walter, A., Heidmann, S., Mechtler, K.and Lehner, C. F. (2007). Spatial organization of a ubiquitous eukaryotic kinetochoreprotein network in Drosophila chromosomes. Chromosoma 116, 385-402.

Sharp, D. J., Rogers, G. C. and Scholey, J. M. (2000). Cytoplasmic dynein is requiredfor poleward chromosome movement during mitosis in Drosophila embryos. Nat. CellBiol. 2, 922-930.

Shu, T., Ayala, R., Nguyen, M. D., Xie, Z., Gleeson, J. G. and Tsai, L. H. (2004). Ndel1operates in a common pathway with LIS1 and cytoplasmic dynein to regulate corticalneuronal positioning. Neuron 44, 263-277.

Siller, K. H., Serr, M., Steward, R., Hays, T. S. and Doe, C. Q. (2005). Live imagingof Drosophila brain neuroblasts reveals a role for Lis1/dynactin in spindle assembly andmitotic checkpoint control. Mol. Biol. Cell 16, 5127-5140.

Starr, D. A., Williams, B. C., Hays, T. S. and Goldberg, M. L. (1998). ZW10 helpsrecruit dynactin and dynein to the kinetochore. J. Cell Biol. 142, 763-774.

Starr, D. A., Saffery, R., Li, Z., Simpson, A. E., Choo, K. H., Yen, T. J. and Goldberg,M. L. (2000). HZwint-1, a novel human kinetochore component that interacts withHZW10. J. Cell Sci. 113, 1939-1950.

Stehman, S. A., Chen, Y., McKenney, R. J. and Vallee, R. B. (2007). NudE and NudELare required for mitotic progression and are involved in dynein recruitment tokinetochores. J. Cell Biol. 178, 583-594.

Sweeney, K. J., Prokscha, A. and Eichele, G. (2001). NudE-L, a novel Lis1-interactingprotein, belongs to a family of vertebrate coiled-coil proteins. Mech. Dev. 101, 21-33.

Tai, C. Y., Dujardin, D. L., Faulkner, N. E. and Vallee, R. B. (2002). Role of dynein,dynactin, and CLIP-170 interactions in LIS1 kinetochore function. J. Cell Biol. 156,959-968.

Veraksa, A., Bauer, A. and Artavanis-Tsakonas, S. (2005). Analyzing protein complexesin Drosophila with tandem affinity purification-mass spectrometry. Dev. Dyn. 232, 827-834.

Jour

nal o

f Cel

l Sci

ence

1758

Vergnolle, M. A. and Taylor, S. S. (2007). Cenp-F links kinetochores toNdel1/Nde1/Lis1/dynein microtubule motor complexes. Curr. Biol. 17, 1173-1179.

Wakefield, J. G., Bonaccorsi, S. and Gatti, M. (2001). The Drosophila protein asp isinvolved in microtubule organization during spindle formation and cytokinesis. J. CellBiol. 153, 637-648.

Wang, H., Hu, X., Ding, X., Dou, Z., Yang, Z., Shaw, A. W., Teng, M., Cleveland, D.W., Goldberg, M. L., Niu, L. et al. (2004). Human Zwint-1 specifies localization ofZeste White 10 to kinetochores and is essential for mitotic checkpoint signaling. J. Biol.Chem. 279, 54590-54598.

Waterman-Storer, C. M. and Holzbaur, E. L. (1996). The product of the Drosophilagene, Glued, is the functional homologue of the p150Glued component of the vertebratedynactin complex. J. Biol. Chem. 271, 1153-1159.

Williams, B. C., Karr, T. L., Montgomery, J. M. and Goldberg, M. L. (1992). TheDrosophila l(1)zw10 gene product, required for accurate mitotic chromosome segregation,is redistributed at anaphase onset. J. Cell Biol. 118, 759-773.

Williams, B. C., Li, Z., Liu, S., Williams, E. V., Leung, G., Yen, T. J. and Goldberg,M. L. (2003). Zwilch, a new component of the ZW10/ROD complex required forkinetochore functions. Mol. Biol. Cell 14, 1379-1391.

Williams, B., Leung, G., Maiato, H., Wong, A., Li, Z., Williams, E. V., Kirkpatrick,C., Aquadro, C. F., Rieder, C. L. and Goldberg, M. L. (2007). Mitch a rapidly evolvingcomponent of the Ndc80 kinetochore complex required for correct chromosomesegregation in Drosophila. J. Cell Sci. 120, 3522-3533.

Wojcik, E., Basto, R., Serr, M., Scaerou, F., Karess, R. and Hays, T. (2001). Kinetochoredynein: its dynamics and role in the transport of the Rough deal checkpoint protein. Nat.Cell Biol. 3, 1001-1007.

Wolf, K. (1995). Spindle membranes and spindle architecture in invertebrates. Micron 26,69-98.

Yan, X., Li, F., Liang, Y., Shen, Y., Zhao, X., Huang, Q. and Zhu, X. (2003). HumanNudel and NudE as regulators of cytoplasmic dynein in poleward protein transport alongthe mitotic spindle. Mol. Cell. Biol. 23, 1239-1250.

Yao, X., Abrieu, A., Zheng, Y., Sullivan, K. F. and Cleveland, D. W. (2000). CENP-Eforms a link between attachment of spindle microtubules to kinetochores and the mitoticcheckpoint. Nat. Cell Biol. 2, 484-491.

Yucel, J. K., Marszalek, J. D., McIntosh, J. R., Goldstein, L. S., Cleveland, D. W. andPhilp, A. V. (2000). CENP-meta, an essential kinetochore kinesin required for themaintenance of metaphase chromosome alignment in Drosophila. J. Cell Biol. 150, 1-11.

Journal of Cell Science 122 (11)

Jour

nal o

f Cel

l Sci

ence