Cloning and sequencing of the cyclin-related cdc13* gene and a cytological

study of its role in fission yeast mitosis

IAIN HAGAN, JACQUELINE HAYLES* and PAUL NURSE

ICRF Cell Cycle Control Laboratory, Microbiology Unit, Department of Biochemistry, University of Oxford, South Parks Road, Oxford, 0X1 JQU

•Author for correspondence

Summary

We have cloned and sequenced the cdclJ* genefrom fission yeast. When a major part of the cdclS*gene is deleted from the chromosome, cells arrest ininterphase, but partial loss of gene activity leads tocells containing condensed chromosomes, aberrantsepta and a microtubular cytoskeleton with charac-teristics of both G2 and M. Expression of thisphenotype is influenced by the nutritional status ofthe cell. Our results suggest that the cdclJ* genefunction is required for the control of the G2 to Mtransition. It appears to play a role in regulating theseparate pathways of events involved in thephysical process of mitosis, for example in thereorganization of the cytoskeleton on transition

from G2 to mitosis. The cdclJ* gene functioninteracts closely with both the yeast and humanhomologues of cdcZ*, suggesting that mammaliancells may contain a cdcl3*~ homologue. The geneencodes a putative polypeptide of 482 amino acids,and a central region of 176 amino acids of thispolypeptide is 50 % identical with sea urchin cyclin.Therefore, the cdclJ4" protein is cyclin related andcould act as a regulator or substrate of the p34cdc2

protein kinase, which initiates mitosis.

Key words: cell cycle, mitosis, Schizosaccharomyces poinbe,microtubules, cdcl3+, cyclins.

Introduction

The fission yeast Schizosaccharomyces pombe sharesmany mitotic features with other more complex eu-karyotes (Hirano & Yanagida, 1988). During the G2phase before mitosis, the cell contains a system ofcytoplasmic microtubules and a nucleus with littleobservable chromatin structure (Hagan & Hyams, 1988;Toda et al. 1981). Electron-microscopic studies suggestthat there is a microtubular organizing centre (MTOC)located on the nuclear membrane called the spindle polebody (SPB) (McCully & Robinow, 1971; Tanaka &Kanbe, 1986). On entry into mitosis the cytoplasmicmicrotubules disappear and are replaced by an intranu-clear mitotic spindle. This is generated between twoSPBs formed as a consequence of duplication of theoriginal single SPB. The chromatin takes on a granularappearance, and in the best light-microscopic prep-arations three partly condensed chromosomes can bevisualized (Robinow, 1977). The nuclear membraneexpands but does not break down, and the nucleusextends through the length of the cell as the mitoticspindle elongates. The two sets of sister chromatids aredrawn apart to the opposite ends of the cell and theextended nucleus splits into two daughter nuclei. After

Journal of Cell Science 91, 587-595 (1988)Printed in Great Britain © The Company of Biologists Limited 1988

anaphase the intranuclear mitotic spindle disappears anda new system of cytoplasmic microtubules is established,emanating from two cytoplasmic MTOCs located nearthe middle of the cell (Hagan & Hyams, 1988). An actinring and a septum are generated in the middle of the cellduring nuclear separation (Marks & Hyams, 1985) andsubsequently the cell divides into two by medial fission.

The fission yeast cell also undergoes a change in growthpattern during mitosis. In G2 phase the cell elongates bygrowth, usually at both tips, but on entry into mitosis thiselongation stops and cells enter the so-called constantvolume stage (Mitchison & Nurse, 1985). After mitosisand cell division have been completed the two daughtercells resume elongation at the two original growing tips.

Many gene functions have been described in fissionyeast, which are required for successful completion ofmitosis (Hirano & Yanagida, 1988). These have beendefined using conditional mutants, either heat- or cold-sensitive, which are unable to complete mitosis at therestrictive temperatures of 36°C or 20°C respectively.The mutants fall into two classes, those blocked in late G2just before mitosis and those blocked during the physicalprocess of mitosis. Mutants in the former class havecytologies typical of G2 phase and continue cell growth,producing highly elongated cells. Mutants in the latter

587

class have a variety of altered or defective nuclear andseptal structures consistent with a block in mitosis itself.These mutants generally do not become highlyelongated, presumably because they are arrested in theconstant volume stage. The temperature-sensitive mu-tant cdcl3-\\l is unusual because at the restrictivetemperature it produces highly elongated cells, whichcontain some cytological features typical of mitosis (Nas-myth & Nurse, 1981). Chromatin in the nucleus takes ona granular appearance and condensed chromosomes arealso occasionally observed. Many cells become septatedalthough the septa are aberrant in structure and do notlead to cell separation. Therefore, cdcl3-\\l arrestedcells exhibit features of both the G2 and mitotic phases.

The cdd3+ gene function is also of interest because itinteracts closely with the p34cdc2 protein kinase encodedby the cdc2+ gene, which plays a central role in theinitiation of mitosis both in fission yeast and other morecomplex eukaryotes (Lee & Nurse, 1987; Draetta et al.1987; Gautier et al. 1988). An example of this interactionis the suppression of the cdd3-\\l mutant phenotype byover-expressing the p34cdc2 protein kinase (Booher &Beach, 1987). These interactions and the cytologicalfeatures suggest that the cdcJ3+ gene product may play arole in the transition of the cell from G2 to mitosis. Toinvestigate this possibility further we have carried out adetailed analysis of the cdcl3-\\l mutant cytology, andhave cloned and sequenced the cdcl3+ gene to begin amolecular investigation of its function.

Materials and methods

Yeast and bacterial strainsThe cdcl3-in (Nasmyth & Nurse, 1981), cdc25-22, cdcl5-140(Nurse et at. 1976) and the ade6-M21Q/ade6-M21(> ura4-D18/W4-D18 leu 1-32/'leu 1-32 h + /h~ strains were derivedfrom 972 h~ (Leupold, 1970). The ura4-D 18 allele is a deletionof the ura4+ locus (a gift from Christian Grimm and JurgKohli). The Escherichia coli strain JA226 (recBC leuB6 trpEShsdR~ hsdm+ lacYc600) was used to recover the plasmid fromyeast; E. coli DH5 (recAl cndAl gyr A96 thi hsdRH supE44relAl) was used for routine cloning procedures and E. coliJM109 (recAl endAl gyrA96 tin hsdR17 supE44 relAl A(lac-pnoAB) [F' traD36 proAB laclq lacZAMIS]) for cloning intopTZ18R/l9R vectors (Pharmacia) and for generating single-Stranded DNA.

MediaYeast phthallate minimal medium with 2 % and OS % glucose isdescribed by Nurse (1975) and Beach & Nurse (1981). Bacterialmedia was prepared according to Maniatis et al. (1982) andHanahan (1985).

Yeast and bacterial techniquesGenetical manipulations were performed as described by Gutzet al. (1984). The yeast transformation was carried out usingyeast protoplasts (Beach & Nurse, 1981). Plasmid was re-covered from the transformants by growing 20 ml of cells to A5950-5 (IX 107 cells ml" ). Cells were then harvested and resus-pended in 500 /il citrate phosphate buffer pH5-6, 40mM-EDTA, 3mgml~' Novozyme (Novo Biolabs). Cells wereincubated at 37°C for 30-60min, harvested, resuspended in

300//I 5 X T E (50mM-Tris, pH7-5, 5mM-EDTA) with 1%sodium dodecyl sulphate (SDS) to lyse the cells, 100 (x\ 5M-potassium acetate was added and the lysate incubated on ice for30 min. The lysate was then centrifuged at 4°C in an Eppendorfcentrifuge for 10 min. A 100-;tl sample of supernatant wasremoved and DNA recovered using glass powder according tothe "geneclean" protocol (Stratech Scientific Ltd.). The DNAwas then transformed into the E. coli strain JA226 usingstandard methods (Hanahan, 1985). The inclusion of the glasspowder DNA isolation step increased the transformationfrequency 5- to 10-fold.

DNA and RNA manipulationsRoutine cloning was performed according to Maniatis et al.(1982). The nested deletions of cdcl3+ described below weresequenced using the dideoxynucleotide chain terminationmethod (Sanger et al. 1977). Yeast RNA was prepared asdescribed by Kauffer et al. (1985) and yeast DNA was preparedas previously described (Beach et al. 1982). Southern andNorthern blot analysis was carried out using Gene Screen plusaccording to the NEN Genescreen plus protocol. The DNAprobe used for both Southern and Northern blot hybridizationwas the 3-2-kb Sall-BamHl fragment of plasmid ptZ18cdcl3labelled with [y-32P]dCTP using the random oligolucleotidelabelling procedure of Feinberg & Vogelstein (1983).

Computer analysis of the DNA sequenceThe sequenced region was analysed using the Staden pro-grammes analysing for reading frames and codon usage. PESTregions and intron consensus sequences were identified usingthe PC/gene programme (Intelligenetics Corporation, PaloAlto, CA) with parameters from previously establishedS.pombe introns (Hughes, 1988).

Cloning of the cdcl3 geneA cdcl3-\\l leul-32 strain was transformed with either of twoS. pombe gene banks constructed in the LEU2-based auton-omously replicating plasmids pDB262 or pYepl3 (Wright et al.1986). The insert DNA in the pYepl3 bank was prepared by apartial Sau3A digest of 5. pombe DNA whereas with thepDB262-based bank the insert was from Hindi 11 -digested yeastDNA. Plates were left for 24 h at 25°C and then shifted to 36°Cfor 5 days. Plasmids isolated from the surviving colonies wereanalysed for their ability to rescue thecdcl3-\\ l defect upon re-transformation into the mutant cells. One of the pYepl3 clones,pcdcl3+ contained a 72-kb insert (see Fig. 4) which rescuedthe defect. From this insert a 3-2-kb Sall-BamHl fragmentcontaining cdcl3+ was subcloned by complementation. Plas-mids pTZ18cdcl3 and pTZ19cdcl3 were constructed by inser-tion of this fragment into the Sall-BamHl sites of plasmidspTZ18R and pTZ19R respectively.

Construction o/cdcl3 gene internal deletionAn internal deletion of the cdc!3+ gene was constructed byremoving nucleotides 784-1598 and inserting at that positionthe 5. pombe ura 4+ gene. Plasmid ptZ19cdcl3 was digestedwith BamHl and Xhol and the larger fragment was isolated.The 5. pombe ura4+ gene was removed from plasmid pura4 bydigestion with Hindlll, the insert was blunt ended and A7iollinkers were ligated onto the fragment, which was then digestedwith Xhol. The BamHl-Xhol fragment and the ura4+ genefragment were then ligated with a Sall-BamHl fragment ofcdcl3, which has had the 5' sequence deleted to position 1599(see Fig. 5). This fragment was obtained by ZJ.volIl deletion ofpTZ19cdcl3. The resultant plasmid (ptZ19cdcl3A) is equival-ent to the plasmid pTZ19cdcl3 containing the cdcl3 3-2-kb

588 /. Hagan et al.

fragment from which a 814-bp internal fragment has beenremoved and replaced with the ura4+ gene.

Constructioti o/cdcl3+ gene 3'and 5' terminal deletionsNested deletions of the 3-2-kb fragment containing the cdcl3+

gene were constructed by digesting plasmid pTZ18cdcl3 withBamHl and Sad, digesting plasmid pTZ19cdcl3 with Sailand Sphl and then treating the linear plasmids with exonucleaseIII followed by Si nuclease digestion (Henikoff, 1984). Prior torecircularization, BamHl linkers were ligated onto the bluntends of the pTZ18-based deletions and Sail linkers were ligatedonto the blunt ends of the pTZ19-based deletions, followed bydigestion with BwnHl or Sail respectively. In this way theparental 3-2-kb fragment and both the 3' and 5' deletions can beremoved from the pTZ vector sequences by digestion with Sailand BamHl. These partial deletions of the cdcl3+ gene-containing fragment were used for sequencing, for localizingthe functional portion of the clone by complementation, and forconstruction of the internal deletion described below.

One-step gene disruptionPurified Bum HI -Sal I fragment (1, 5, 10 or 20/Jg) fromptZ19cdcl3A was transformed into ade6-M210/ade6-M2l6ura4-D\8/ura4-D\8 leu 1-32/leu 1-32 h+/h~. Stable ura+

transformants were sporulated and diploid colonies selectedthat gave rise to elongated cells upon sporulation.

Southern blot analysis of the diploids established that one ofthe cdcl3+ genes had been replaced by the deleted fragment.Tetrad analysis performed on these diploids showed that theelongated cell phenotype segregated 2: 2 with wild type and thatthe wild-type spores were all ura~.

CytologyThe analysis of the deletion was performed by sporulating thediploid strain in minimal medium containing 0 -5% (w/v)glucose in which the ammonium chloride had been replacedwith 1 mM-sodium glutamate. The spores were germinated inminimal medium with either 2% or 0 '5% glucose sup-plemented with 250 mg I"1 adenine and 250mgl~ leucine. Allother cells were grown using the media described in the text.

Cells were prepared for immunofluorescence microscopy asdescribed previously (Hagan & Hyams, 1988) using the anti-yeast tubulin antibody YOLl/34 (Kilmartin et al. 1982).However, in the preparation of the cells containing the cdcl3+

partial deletion the sodium borohydride step was omitted asthere was a paucity of material and a high level of cell debrisfrom the spore walls, which made centrifugation difficult.

Results

Giemsa staining of cdc 13-W1 mutant cells arrested at therestrictive temperature of 37 °C revealed a granularity innuclear structure, with some cells showing three con-densed chromosomes (Nasmyth & Nurse, 1981). Manycells also became septated, but the septa were aberrantand irregular in structure, and failed to lead to cellseparation. In order to investigate the kinetics of appear-ance of condensed chromosomes and abnormal septatedcells, we used DAPI staining of DNA at various timesafter shift to 37°C. Using minimal medium with 2%glucose for growth (high glucose medium), cells withcondensed chromosomes begin to appear about 3 h aftershift to 37°C, and reach a peak level at around 4h(Fig. 1). About 10-20 % of the cells contained condensed

100 200Time (min)

300

Fig. 1. Chromosome condensation and septation in cdc13-117 cells after shift to the restrictive temperature are affectedby the nature of the growth medium. The figure shows thefrequency of cells with condensed chromosomes (circles) andseptation (triangles) in cdc 13-117 after a temperature shiftfrom the permissive temperature (29°C) to the restrictivetemperature (37°C) at time zero. Different results wereobtained depending upon the concentration of glucose in theminimal medium. Higher levels of chromosome condensationand septation were seen when cells were cultured in 0-5 %glucose (closed symbols) as opposed to the standard 2 %glucose (open symbols).

chromosomes at this time. Unexpectedly, the growthmedium dramatically affected the proportion of the totalcell population that accumulated condensed chromo-somes. When cells were grown in minimal medium with0-5% (w/v) glucose (low glucose medium), many morecells accumulated condensed chromosomes. The increasebegan at 2 h after shift to 37°C, and reached peak levels of65-90% at 4h (Figs 1, 2A). Similar results were seenwith the appearance of abnormal septated cells. In highglucose medium 10-15% of the cells become septatedand 60-80 % of the cells become septated in the lowglucose medium. The time course indicates that septationoccurs about 50 min after chromosome condensation.

These experiments indicate that the growth mediumcan influence the expression of the cdcl3-\ll mutantphenotype. They also suggest that the arrested mutantcells slowly proceed into mitosis, generating condensedchromosomes and aberrant septa. To investigate thecytology of the mutant phenotype more fully we haveexamined the cells using anti-tubulin immunofluor-escence. For comparison typical G2 phase cytoplasmicmicrotubules and nuclei are shown in Fig. 2B,C, whichillustrates cdc25-22 mutant cells (Nurse et al. 1976;Nasmyth & Nurse, 1981) arrested in late G2. At the onsetof mitosis in wild-type cells this network of cytoplasmicmicrotubules disappears and is replaced by an intranu-clear microtubular mitotic spindle. Initial experimentswith the cdcl3-\\l mutant indicated that examination ofthe microtubular structures was hampered by the ab-errant septa generated 3-4 h after shift to 37°C. To avoidthis complication, a double mutant strain cdd3-\YIcdd5-\¥) was used. The cdc!5-\40 mutation blocksseptation, and enables the microtubular structures to be

Cyclin-related cdc 13 gene in S. pombe 589

Fig. 2. cdcl3-\\l cells arrest with condensed chromosomes in contrast to a typical G2 arrest. A. Cells of the strain cdcl3-\\lcdc!5-\40 grown in minimal medium containing 0 5 % glucose stained with DAPI after growth and incubation at 37°C for4-25 h. B. cdc25-22 cells arrested in G2 after incubation at the restrictive temperature for an identical period stained with DAPIand viewed under combined fluorescence and phase contrast optics. C. Anti-tubulin staining of the cells shown in B. Theterminal phenotype of cdcl3-\\l cdcl5-\¥) cells with condensed chromosomes differs markedly from the somewhat diffuseextended nuclear staining seen in cdc2S-22 cells. The cytoplasmic microtubules in C are typical of cell cycle arrest in late Gz-Bar, 10 fun.

examined more easily. The double mutant strainexhibited all the characteristics of the cdcl3-\\l mutantwith respect to chromosome condensation, but did notseptate (Fig. 2). Cells were prepared for anti-tubulinimmunofluorescence after 4-25 h at the restrictive tem-perature. Surprisingly, the majority of cells exhibited thecytoplasmic array of microtubules typical of G2 phase,even though the same cells frequently also had condensed

chromosomes (Fig. 3A-J). These chromosomes werefound either clustered together in the centre of the cell orspread out along the length of the cell. A low proportionof cells did contain mitotic spindles, but in these cases anextensive array of cytoplasmic microtubules was alsoclearly present (Fig. 3K-N). During normal mitosis anextensive cytoplasmic array and a mitotic spindle arenever present in the same cell (Hagan & Hyams, 1988).

590 /. Hagan et al.

Fig. 3. cdcl3-\\l cdcl5-\40 cells arrest with both interphase and mitotic features at the restrictive temperature. The panelsshow paired YOLl/34 anti-tubulin immunofluorescence and DAPI staining of the same cells of the strain cdcl3-\\l cdcl5-\4Qwhich were grown and arrested at 37°C for 4-25 h in minimal medium containing 0-5 % glucose. The pairs A & B, C & D, E &F, G & H and 1 & J all show examples of cells containing condensed chromosomes and cytoplasmic niicrotubules. The pairs K& L and M & N both show cells with mitotic spindles and cytoplasmic microtubules. Interestingly, the lower chromosome in Nis some distance from the spindle in M suggesting that it is not associated with it. Bar, 5 fim.

Most of the mitotic spindles seen in the arrested cdcl3-117 mutant cells were short and were always present withcondensed chromosomes (Fig. 3K-N). However, thecondensed chromosomes did not appear to be associatedwith the spindle (Fig. 3N).

These data indicate that arrested cdcl3-\\l cellsexhibit features that are typical of both G2 and M phase.This is an unusual situation because in a normal cell thecytologies associated with each phase are mutually exclus-ive. In the cdcl3-\\l mutant the appearance of con-densed chromosomes and, in certain cells, the mitoticspindle together with an interphase array of cytoplasmicmicrotubules, suggest that a slow progression throughthe initial events of mitosis occurs even though cells havenot properly left the G2 state. Therefore, the cdcl3+ genefunction appears to play an important role at or just afterthis initiation of mitosis.

In order to investigate the cdcl3+ gene function morefully we cloned the gene by complementation asdescribed in Materials and methods. Two gene bankswere used, based on the plasmids pDB262 and pYepl3.Six colonies grew up from the pDB262 bank transform-ation and two from the pYepl3 bank transformation.Plasmids isolated from the six pDB262 clones all provedto contain the cdc2+ gene by restriction mapping. Thisgene has been shown to rescue cdcl3-\\l when presenton a plasmid (Booher & Beach, 1987). One of the twopYepl3 clones contained a plasmid with a 7-2-kb Sau3Apartial insert which rescued the cdcl3-\\l defect on re-transformation back into yeast. The plasmid was inte-

grated into this yeast chromosome via homologous re-combination. The integrated plasmid was very closelylinked to cdcl3+ because only three recombinants weregenerated between the cdcl3+ gene and the integratedplasmid out of 789 spores. Such close linkage indicatesthat the plasmid had integrated at the site of the cdcl3+

gene, and thus the insert contains the cdcJ3+ genefunction.

A 3-2-kb Sall-BamHI fragment, which still hadcdcl3-\\l rescue activity, was subcloned from the 7-2-kbpcdcl3 insert. Chromosomal Southern blots showed thatthis fragment was also found in the 5. pombe genome(data not shown). A series of 5' and 3' deletions of the3-2-kb fragment were tested for their ability to rescuecdcl3-\\l. The smallest inserts with activity are shown inFig. 4. The region of DNA that contained rescue activitywas completely sequenced in both directions usingSanger dideoxynucleotide procedures and the sequence isshown in Fig. 5.

Examination of the sequence revealed a potential openreading frame (ORF) starting at nucleotide 501 andextending to nucleotide 1947. A deletion within this ORF(see below) leads to cell cycle arrest, showing that afunctional part of cdcl3+ is located in this region. TheORF would encode a polypeptide of 482 amino acids inlength with a molecular weight of 56K (K=10 3 M r )(Fig. 6). This ORF shows good codon usage whencompared with other S. pombe cdc genes except for theregion between nucleotides 1300 and 1380. A search forintron consensus sequences (Hughes, 1988) revealed a

Cyclin-related cdcl3 gene in S. pombe 591

Sail Sail Cla\ EcoRl Xho\ tfmdIII EcoRl BamHl BamHl

200 bp

Fig. 4. The DNA fragment that rescues cdcU-UJ. Thefigure shows the structure of the cdcl3-\\l complementingclone. The entire 7-2-kb DNA fragment from pcdcl3 isrepresented although only the 3-2-kb BamHl-SaJI fragmentis shown to scale (solid line), and only the region sequencedin both directions shown in detail (bar). Nine deletions of the3-2-kb fragment are shown below, along with the indicationof the ability of each to rescue the cdcl3-\\l mutation on theright hand side ( + , rescue; —, no rescue).

potential intron in this region between nucleotides 1312and 1372. If this intron were spliced out, 20 amino acidswould be deleted and the molecular weight of thepolypeptide would be reduced to 54K. Also of interest isthe presence of a region located between amino acidresidues 176 and 226 which contains a PEST consensussequence (PESQ) suggesting that the polypeptide may berapidly turned over (Rogers et al. 1986).

Our suggested ORF is consistent with the deletionanalysis, and with the size of the major RNA transcriptencoded by this region which is approximately 2-5 kb(Fig. 6). This hybridizes to the 3-2-kb Sall-BamHIfragment containing cdd3+ (Fig. 4), to a 513-nucleotidefragment spanning nucleotides 785-1297, internal to theORF, and to a single-stranded RNA probe complemen-tary to the direction of the ORF (data not shown). Thistranscript was not detected with the opposite-strandedRNA probe, or by probes made from regions flanking theSall-BamHI fragment, indicating that the 2-5-kb tran-script is encoded by sequences contained within theSall-BamHI fragment (data not shown).

To determine the effect of eliminating the cdd3+ geneproduct entirely from the cells we deleted part of thecdcl3+ gene. The sequence between nucleotides 785 and1599 was deleted and replaced with the 5. pombe ura4+

gene, eliminating 271 internal codons from the putative482 codons of the cdcl3+ gene. This construct was usedto disrupt the cdd3+ gene in a diploid strain using theone-step gene replacement method (Rothstein, 1983).

Sporulation of these strains revealed that half of thespores germinated and formed highly elongated cells,which did not divide. Germination of these spores ineither high or low glucose minimal medium yielded thesame phenotype. The elongated cells arrested with acytoplasmic microtubular array typical of G2 phase(Fig. 7). No condensed chromosomes or aberrant septawere observed. This indicated that a complete loss of thecdcl3+ gene product leads to arrest in G2 phase beforeentry into mitosis. Therefore the more complex pheno-type associated with the cdcl3-\\l allele suggests that

chromosome condensation and aberrant septation takeplace as a consequence of partial gene activity in thismutant. This partial activity does not lead to a complete'leak' past the mutant block point as observed in someother cdc mutants (Hagan & Hyams, 1988). Unlike theseother cases only a partial set of mitotic events areobserved and as a consequence the cells fail to completecell separation.

The ability of the plasmid containing the cdcl3+ geneto rescue the cdd3-\\l mutation was compared with thesuppression provided by plasmids containing the cdc2+

gene and its human homologue CDC2(Hs) (Lee & Nurse,1987). All three plasmids were introduced by transform-ation, and enabled the mutant to divide at the restrictivetemperature, although the cdcl3+ gene transformed cellsat about 20 % of the frequency found with the other twoplasmids, and also resulted in slower growth at thepermissive temperature. These results indicate that theclose interactions observed between cdd3+ and cdc2+ inyeast are conserved with the human homologueCDC2Hs.

Discussion

We have shown that removal of the cdc!3+ gene productentirely from the fission yeast cell leads to an arrest in G2during the cell cycle. However, arrest of the cells usingthe temperature-sensitive mutant cdd3-\\l leads to aterminal phenotype with characteristics of both G2 andM phase. The cells have a cytoplasmic array of micro-tubules, but in addition they slowly accumulate con-densed chromosomes and aberrant septa and, in a smallproportion of the cells, intranuclear mitotic spindles, allfeatures of mitosis. This mixing of G2 and M phasecharacteristics does not occur in wild-type cells, whichundergo an orderly transition from G2 to M. In wild-typecells cytoplasmic microtubules completely disappear,followed by the appearance of partly condensed chromo-somes, a mitotic spindle and eventually a septum.

Our interpretation of these results is that the cdc!3+

gene function is required just at the initiation of mitosis.In the complete absence of any cdd3+ gene product,mitosis cannot be initiated. With a partial loss of cdcI3+

activity, chromosome condensation and aberrant sep-tation can slowly take place, but in terms of microtubularorganization the cell never leaves the G2 state. Even whenmitotic spindles are observed in some cells, the cytoplas-mic microtubular array persists. We propose that thecdcl3+ gene function has a role in the control of thereorganization of microtubules from the G2 extranuclearcytoplasmic array to the intranuclear mitotic spindle.The fact that chromosome condensation and some formof septation can take place in the presence of a G2cytoskeleton suggests that there are separate pathways ofevents making up the physical process of mitosis, whichare differentially affected by partial loss of cdcl3+

function.It is also of interest that the cdcl3-\\l cells that do

produce a mitotic spindle all have condensed chromo-somes and these often lie a considerable distance from the

592 /. Hagan et al.

spindle. Failure to generate microtubular connectionsbetween the chromosomes and the spindle pole bodiescould lead to the observed separation of chromosomesand spindle, suggesting that kinetochore activation maynot have taken place. This separation is in markedcontrast to the distribution of chromosomes in nuc2t*mutants which are clustered around the centre of an earlymitotic spindle in a manner analogous to a classicalmetaphase plate (Hirano et al. 1988).

The growth medium of the cells was found to influencethe full expression of the cdcl3-\\l terminal phenotype.Growth medium does influence the timing of mitosis,which can be advanced in cells that are nutritionallystarved, and it has also been suggested that mitoticreactions are accompanied by a drop in cAMP levels(Hirano & Yanagida, 1988). It is possible that progress

through mitosis is facilitated by changes in the nutritionalstatus of the cell, and that these conditions are found incells growing in media that allow full expression of thecdcl3-\Yl phenotype.

Over-expression of the cdc2+ gene product rescues thecdcl3-\\l mutant phenotype and has led to the sugges-tion that the cdc2+ and cdcJ3+ gene products interactclosely (Booher & Beach, 1987). Our results support thisnotion. The S. ponibe cdc2+ gene on a multi-copynumber plasmid rescues cdd3-\\l, as does the humanCDC2 (Hs) homologue. The ability of the human gene torescue indicates that the domain of the CDC2 (Hs)protein responsible for the interaction has been conservedand suggests that a cdd3+ homologue may also be foundin mammalian cells. Our sequencing has identified aputative polypeptide of 54-56K encoded by the cdcl3+

Tl lUlTTAGATTTACTAACCTCATTTAAAGTCAACArTTACCU 1L111 111 1GTTCTCCTGTTAATATAGGAAAATCGTACATCGATTCTACATATAGCTT10 20 30 40 SO t o 10 10 90 100 110

CATTTCAAAGTTCAUT 1 1H.1U1 !CTM:CTTTCCTATAT^^rrCTTGGACTTCATCCCAATCCCCTTACTTTATTCAAAAAAAGAAU rLriAAAAGCTCTTCCCAAATT120 130 140 150 110 170 160 190 200 210 220

CTACATTCCCTCAAACCCTATTATTTTCTCCTrTATACGTACGACAAACATACAl111 f rCCTTGGATTTTATCCTTAAAA 11 U 1GTATAAACTTTCAAGTATAACTCAT230 240 250 2(0 270 2 80 2 90 300 310 330 320

TCTArrrTATCTTCTCCGTCCCCTACTACCCCTTCACCTTTTTAGAC^34 0 350 360 370 360 390 400 410 420 4 30 440

H T T R R L T R Q H L L A M T L GTATTTTAArrrcoc T TCCTCTCTTTCAATCCCCCACA CAAT n a ^ r r r c ^

450 4(0 470 4B0 490 500 510 52 0 530 540 550

H N D E N H P S M H I A R A K 5 S ^ H S S E H 3 L V H G K K A T V S 3 T NAACAATGAC GAAAATCATCCTTCAAACCAT A rTGCCCCTGCAAAAAGCTCTTTGCACTCTTCAGAAAATTCTTTA^

5(0 570 510 5*0 (00 (10 (20 (30 (40 650 ((0

CGTrCCTAACAACCGTCATGCGTTGGATt»TGTTTCCAATTTTCACA^(70 (B0 (90 700 710 720 730 740 750 7(0 770

120 830 140 850 ICO 870 110

CCTCAACAL1 lUlCAAI^JU^TCACACTCTGTTTCAACCCATCCCCTTGATCCl 1 ILCATAACCATCAAGCAACTArKCAAAAAAATTAAAGAAAGATCrrCATCAACC

CGTTCTTTCGAAAGATATTCCCAAACrrCACCGTGATACTCTTr^r^CTCCCC^1000 1010 1020 1030 1040 1050 10(0 1070 1080 1090 1100

1110 1120 1130 1140 1150 11(0 1170 1180 1190 1200 1210

CTGCTTCC^1220 1230 1240 1250 12(0 1270 1280 1290 1300 1310 1320

L H K L O I. S K Y E E V M C P S V Q N F V Y M A D G GACC*aCtAT»C«CCTCATGTCCCCTTCACTCCAAAALl"l lUlATATATGGCGCATGCTGCCT

1330 1340 1350 13(0 1370 1360 1390 1400 1410 1420 1430

Y D E E E I L Q A E R Y I L R V L E F N L A Y P M P M M F L R R I S K A DATGATCAA^GGAAATTCTTCAAGCCCTCCGCTACATTTrGCGTCrcCTAG

1440 14 50 14(0 1470 1460 14 90 1500 1510 1520 1530 1540

F Y D I Q H K L L P Y P P S Q Q C A A A M Y

1550 15(0 1570 1560 1590 1(00 1610 1(20 1630 1(40 1(50

S G Y E E Y Q L I S V V K K H I H Y

1710 1720 1730 1740 17 50 17(0

L Q K P V Q H E A r F K K Y A S K K F H K A S L F V R D W I K K H S I P LTACAAAAGCCTCTTCAACATt^CCTTTTTTCAAa^grATGCCTCCAA^

G D D A D E D Y T F H K Q K R I Q H D H K H E E W *CCCCATCACCCTCATGAAGATTATACTTTTCACAAGCAAAAACGTATACAACATCACATCAAACATGAAGA^

18S0 1890 1900 1910 1920 1930 1940 1950 19(0 1970 1910

TCTTTCAATTGTC^ATATTCCAACrTCTTCAAi GCT1990 2000 2010 2020 2030 2040 2050 2060 2070 2060 2090

GCTT TTCTTTTTCTT TT GTT CT ATT ATT AT ATCC TCGGGTACTTGAAGATTG CCAGATGCATACTACATGCTGGTCTCATTATGAAACCAAAATCATTTCCACAAGTCTA2100 2110 2120 2130 2140 2150 21(0 2170 2180 2190 2200

ATTATIG<^TGCCTTTAAATTATCCCACCTTTAAG1CTTAGATTTGATT^2210 2220 2230 2240 2250 22(0 2270 2260 2290 2300 2310

AATCATGCTGCTTTAAACAACTACGAGTATTGTGCTATTATTTCTTTITrAa^232 0 2330 2340 2 350 2360 2370 2360 2 390 24 00 2410 2420

CGATTCCCAATCCCATA TCCTACACCATGAGTC GGAGTAATATTAATAA T AAT AATt TAATGATAGTTCCTTGACCTGAATC2430 2440 2450 24(0 2470 2490 2490 2500

Fig. 5. DNA sequence of the regionthat rescues cdcl3-\\l and theputative amino acid sequence of thecdcl3+ gene product. DNA sequenceobtained by sequencing the 2503-bpregion of the 3-2-kb BamHl-Sa!\fragment in the region shown inFig. 4. The putative cdcl3+ openreading frame (ORF) extends frombase 551 to 1947. The single aminoacid code corresponding to the codingregions of the gene is shown above thesequence. The underlined amino acidsequence corresponds to the regionthat would be lost if the putativeintron spanning bases 1313-1372 wereto be excised. The intron consensussequences used to construct thisputative gene are underlined. Bothstrands were identical apart fromresidue 83, which lies outside the openreading frame, which read as an A inone direction and a G in the other.We assume that this is due to amutation in the bacterial host.

Cyclin-related cdcl3 geiie in S. pombe 593

gene. The predicted sequence of this polypeptide isclosely related to the sequence of the sea urchin Arbaciapunctulata cyclin (Pines & Hunt, 1987). The cyclins area class of proteins that change in level during the first fewmitotic divisions in eggs of various organisms, and arethought to be an essential component for entry intomitosis (see review by Pines & Hunt, 1987). The seaurchin B cyclin is 409 amino acids in length comparedwith the 482 amino acids of the cdd3+ gene product.Overall identity in amino acids is 23 % between the twoproteins, but most of this similarity is confined to thecarboxy-terminal 60% of cdcl3+. As can be seen inFig. 8, there is a 37 % identity between residues 187-482,with a rather higher level between residues 187-363 of50% identity. These regions include consensus se-

T S G T C E J T J O L I S V V K K H I H T L J O J K P V O N D A r T M H r n s iM T H | T | 3 K J T | S E D H L R P I V Q K I v j o j I L L K D D S A 3 0 K T S I A I V K Tl K Y I G I S S

. o D D » D Schizosaccharomyces pombe cdcl3H

i o a s » E Arbacia punctulata cyclin

Fig. 8. Comparison of polypeptide sequences of S. pombecdcl3+ and .4. punctulata cyclin. The polypeptide sequencesfor S. pombe cdcl3+ between amino acids 187 and 462, andfor A. punctulata cyclin between amino acids 127 and 409 areshown. Identities are boxed.

2-5 k b -

Fig. 6. An RNA of approximately2 5 kb is produced from the cdcI3+

gene. A Northern blot of total poly(A) +

selected S. pombe mRNA was probedusing oligo-labelled 3-2-kbBamHl-Sall fragment fromptZ19cdcl3. One major band was seenat2-5kb.

Fig. 7. Germinating spores containinga partial gene deletion of the cdcl3 genearrest cell cycle progression withinterphase cytoplasmic microtubulesand interphase nuclei. Atypical cell,germinated from a spore deleted forpart of the cdcl3+ gene as described inthe text, is shown. A, YOL 1/34 anti-tubulin immunofluorescence; B, DAPIstaining of the same cell. X 10 000.

quences found in all cyclins and there are also othermotifs typical of cyclins such as double lysines in theamino-terminal region (Tim Hunt, personal communi-cation).

These similarities suggest that cdcl3+ encodes acyclin-related protein, and strengthens the connectionbetween cell cycle control elements in fission yeast andvertebrate cells established by studies with cdc2+ (Lee &Nurse, 1987; Draetta et al. 1987).

It will now be of considerable interest to determine therole that the cdcl3+ cyclin-like gene product plays ininteracting with p34cdc2 and in the regulation of mitosis inboth yeast and vertebrate cells.

After submission of this manuscript the cloning andsequencing of cdcl3+ was also reported by Booher &Beach (1988). Our sequences are essentially identicalapart from our report of a histidine at amino acid position479 and their report of an aspartate.

We are endebted to Ry Young for his valuable contributionsto this work, and we thank Stuart MacNeill, Shelley Sazer andViesturs Simanis for useful discussions, Tim Hunt for dis-cussions concerning cyclins and Jenny Stephens for typing. Weare grateful to ICRF for supporting this work and to the MRCfor a fellowship to I.H.

References

BEACH, D. & NURSE, P. (1981). High frequency transformation ofthe fission yeast Schizosaccharomyces pombe Nature, Loud. 290,140-142.

BEACH, D., PIPER, P. & NURSE, P. (1982). Construction of a

Schizosaccharvmyces pombe gene bank in a yeast bacterial shuttlevector and its use to isolate genes by complementation. Molec. ?e».Genet. 187, 326-329.

BOOHER, R. & BEACH, D. (1987). Interaction between cdcl3 + andcdc2 + in the control of mitosis in fission yeast: dissociation of theGl and G2 roles of the cdc2 + protein kinase. EMBOJ. 6,3441-3447.

BOOHER, R. & BEACH, D. (1988). Involvement of cdclj+\n mitoticcontrol in Schizosacchammyces pombe: possible interaction of thegene product with microtubules. EMBOJ. 7, 2321-2327.

594 /. Hagan et al.

DRAETTA, G., BRIZUELA, L., POTASHKIN, J. & BEACH, D. (1987).Identification of p34 and pl3, human homologs of the cell cycleregulators of fission yeast encoded by cdc2 + and sucl +. Cell 50,319-325.

FEINBERC, A. P. & VOCELSTEIN, B. (1983). A technique forradiolabelling DNA restriction endonuclease fragments to highspecific activity. Analyt. Biochem. 132, 2-13.

GAUTIER, J., NORBURY, C , LOHKA, M., NURSE, P. & MALLER, J.(1988). Purified maturation promoting factor contains the productof a Xenopus homolog of the fission yeast cell cycle control genecdc2+. Cell 54, 433-439.

Gurz, H., HESLOT, H., LEUPOLD, U. & LOPRENO, N. (1984).Schizosaccharomyces pombe. In Handbook of Genetics (ed. R. C.King), pp. 395-446. New York: Plenum Press.

HAGAN, I. M. & HYAMS, J. S. (1988). The use of cell division cyclemutants to investigate the control microtubule distribution in thefission yeast Schizosaccharomyces pombe. J. Cell Sci. 89, 343-357.

HANAHAN, D. (1985). Techniques for transformation of E. coli. InDNA Cloning: a Practical Approach, vol. 1 (ed. D. M. Glover),pp. 109-135. Oxford, Washington: IRL Press.

HENIKOFF, S. (1984). Unidirectional digestion with exonuclease IIIcreates targeted breakpoints for DNA sequencing. Gene 28,351-359.

HIRANO, T., HIRAKOA, Y. & YANAGIDA, M. (1988). A temperature-sensitive mutation of the Schizosaccharomyces pombe gene nuc2 +

that encodes a nuclear scaffold-like protein blocks spindleelongation in mitotic apparatus. J. Cell Biol. 106, 1171-1183.

HIRANO, T. & YANAGIDA, M. (1988). Controlling elements in the celldivision cycle of Schizosaccharomyces pombe. In The Molecularand Cell biology of Yeasts (ed. G. F. Walton & T. Yarranton) (inpress). Glasgow: Blackie & Son.

HUGHES, D. (1988). Molecular genetics of the cdc27 gene ofSchizosaccharomyces pombe. Ph.D Thesis, Edinburgh.

KAUFFER, N. F., SIMANIS, V. & NURSE, P. (1985). The fission yeastSchizosaccharomyces pombe correctly excises a mammalian RNAtranscript intervening sequence. Nature, Loud. 318, 78—80.

KILMARTIN, J. V., WRIGHT, B. & MILSTEIN, C. (1982). Ratmonoclonal anti-tubulin antibodies derived using a newnonsecreting rat line.J. Cell Biol. 93, 576-582.

LEE, M. G. & NURSE, P. (1987). Complementation used to clone ahuman homolog of the fission yeast cell cycle control gene cdc2.Nature, Land. 327, 31-36.

LEUPOLD, U. (1970). Genetic methods for Schizosaccharomvcespombe. Methods Cell Physiol. 4, 169-177.

MCCULLY, E. K. & ROBINOW, C. F. (1971). Mitosis in the fissionyeast Schizosaccharomyces pombe: A comparative study with lightand electron microscopy. J. Cell Sci. 9, 475-507.

MANIATIS, T., FRITSCH, E. F. & SAMBROOK, J. (1982). Molecular

cloning. A laboratory manual. New York: Cold Spring HarborLaboratory Press.

MARKS, J. & HYAMS, J. S. (1985). Localisation of F-actin throughthe cell division cycle of Schizosaccharomvces pombe. Eur. J. CellBiol. 39, 27-32.

MITCHISON, J. M. & NURSE, P. (1985). Growth in cell length in thefission yeast Schizosaccharomyces pombe. J. Cell Sci. 75, 357-376.

NASMYTH, K. & NURSE, P. (1981). Cell division cycle mutantsaltered in DNA replication and mitosis in the fission yeastSchizosaccharomyces pombe. Molec. gen. Genet. 182, 119-124.

NURSE, P. (1975). Genetic control of cell size at division in yeast.Nature, Land. 256, 547-551.

NURSE, P., THURIAUX, P. & NASMYTH, K. (1976). Genetic control ofthe cell division cycle in the fission yeast Schizosaccharomycespombe. Molec gen. Genet. 146, 167-178.

PINES, J. & HUNT, T. (1987). Molecular cloning and characterisationof the raRNA for cychn from sea urchin eggs. EMBO J. 6,2987-2995.

ROBINOW, C. F. (1977). The number of chromosomes inSchizosaccharomyces pombe: light microscopy of stainedpreparations. Genetics 87, 491-497.

ROGERS, S., WELLS, R. & RECHSTEINER, M. (1986). Amino acidsequences common to rapidly degraded proteins: the PESThypothesis. Science 234, 364-369.

ROTHSTEIN, R. J. (1983). One step gene disruption in yeast. Meth.Enzym. 101, 202-211.

SANGER, F., NICKLEN, S. & COULSON, A. R. (1977). DNA

sequencing with chain terminating inhibitors. Proc. natn. Acad.Sci. U.SA. 74, 5463-5469.

TANAKA, K. & KANBE, T. (1986). Mitosis in the fission yeastSchizosaccharomyces pombe as revealed by freeze-substitutionelectron microscopy, j Cell Sci. 80, 253-268.

TODA, T., YAMAMOTO, M. & YANAGIDA, M. (1981). Sequential

alterations in the nuclear chromatin region during mitosis of thefission yeast Schizosaccharomyces pombe: video fluorescencemicroscopy of synchronously growing wild type and cold-sensitiveede mutants by using a DNA-binding fluorescence probe. J. CellSci. 52, 271-287.

WRIGHT, A., MAUNDRELL, K., HEYER, W.-D., BEACH, D. & NURSE,

P. (1986). Vectors for the construction of gene banks andintegration of cloned genes in Schizosaccharomyces pombe andSaccharomyces cerevisiae. Plasmid 15, 156-158.

(Received I August 1988 -Accepted, in revised form, 23 August I9S8)

Cyclin-related cdcl3 gene in S. pombe 595