Loss of Spatial Control of the Mitotic Spindle Apparatus in a Chlamydomonas Reinhardtii Mutant Strain Lacking Basal Bodies

Home , Spindle apparatus

Loss of Spatial Control of the Mitotic Spindle Apparatus in a Chlamydomonas reinhardtii Mutant Strain Lacking Basal Bodies

Linda L. Ehler, Jeffrey A. Holmes' and Susan K. Dutcher Department of Molecular, Cellular, and Developmental Biology, University of Colorado, Boulder, Colorado 80309-0347 Manuscript received April 27, 1995 Accepted for publication August 4, 1995

ABSTRACT The bld2-1 mutation in the green alga Chlamydomonas reinhardtii is the only known mutation that results in the loss of centrioles/basal bodies and the loss of coordination between spindle position and cleavage furrow position during cell division. Based on several different assays, bld2-1 cells lack basal bodies in >99% of cells. The stereotypical cytoskeletal morphology and precise positioning of the cleavage furrow observed in wild-type cellsis disrupted in bld2-1 cells. The positionsof the mitotic spindle and of the cleavage furrow are not correlated with respect to each other or with a specific cellular landmark during cell division in bld2-1 cells. Actin has a variable distribution during mitosis in bld2-1 cells, but this aberrant distribution is not correlated with the spindle positioning defect. In both wild- type and bld2-1 cells, the position of the cleavage furrow is coincident with a specialized set of microtubules found in green algae knownas the rootlet microtubules.We propose that the rootlet microtubules perform the functionsof astral microtubules and that functional centrioles are necessarythe organiza- for tion of the cytoskeletal superstructure critical for correct spindle and cleavage furrow placement in Chlamydomonas.

HE position of the mitotic spindle apparatus must Physarum and Naegleria cells do not have centrioles T be coordinated bothtemporally and spatially with during mycelial growth (FULTONand DINGLE1971), the cleavage furrow to ensure correct partitioning of and higher plant cells do not have centrioles. In addi- the cellular contents.Incorrect placement of these tion, the Drosophila haploid tissue culture cell line structures with respect to each other would result in 1182-4, which was isolated from dead embryos of the cellular catastrophe. The elaborate mechanisms that female sterile strain mh1182, lacks centrioles when ex- exist for a cell to monitor the correct replication and amined by serial thin-section electron microscopy. The segregation of itsDNA must act in concertwith cleavage cell line undergoes growth and mitotic divisions, but it furrow positioning. Furthermore, correct segregation spontaneously diploidizes, has a high levelof aneu- of developmental determinants and implementation of ploidy and contains large numbers of multinucleate developmental programs may require precise cleavage cells (DEBEC et al. 1982; DEBEC1984; SZOLLOSIet aZ. furrow position within cells undergoing development. 1986). It is unclear whether these defects arise from For example, inCaenarhabditis elegans, the specific orien- the lack of centrioles or from the haploid/aneuploid tation of the cleavage furrow is necessary for the correct nature of the cell line. In contrast,when the centrioles localization of P granules to only one daughter cell in and most of the pericentriolar material are surgically thesubsequent divisionof the P1 cell (HYMANand removed from cells of the monkey cell line BSG1, an WHITE1987). organized array of microtubules is regenerated but cen- The mitotic spindles of many cells are nucleated by trioles are not, and the cells do not complete cell divi- the centrosomes, which contain a pair of centrioles/ sion (MANIOTIS and SCHLIWA1991). It is unclear basal bodies andamorphous pericentriolar material whether the inability of these surgically altered cells to (PCM) that usually nucleates amicrotubule array divide is due to a lack of centrioles or to a missing (GOULDand BORIS 1977). Although centrosomes ap- component(s) of the PCM. The evidence for the role pear to be essential for formation of the spindle and of centrioles in cell division remains inconclusive. the mitotic asters, it is unclear whether the centrioles The exact mechanisms involved in positioning the are necessary for this function. Mouse meiotic spindles mitotic apparatus are unknown. One current model is do not have centrioles (CALARCO-GILLAMet al. 1983), that the centrosomes migrate to a position previously established or maintained by a microfilament network. Corresponding authm: Susan K. Dutcher, Department of Molecular, The interaction with the actin cytoskeleton is proposed Cellular, and Developmental Biology, University of Colorado, Boul- to occur via the astral microtubules that radiate from der, CO 80309-0347. E-mail: [email protected] the centrosomes (HYMANand WHITE 1987; PALMERet 'Present address: Intercampus Program, Molecular Parasitology, 3333 California Street University of California, San Francisco, California, al. 1992).In the P blastomeres of C. elegans, the 941431204, centrosomes separate and migrate to positions directly

Genetics 141: 945-960 (November, 1995) 946 L. L. Ehler, J. A. Holmes and S. K. Dutcher opposite each other and subsequently both rotate 90 formation is dependent on microtubules; loss of micro- deg to align on the anterior-posterior axis. Rotation of tubules from treatment with colchicine before midana- the centrosomes shows sensitivityto microtubuleinhibi- phase prevents cleavage furrow initiation (BWS and tors (HYMANand WHITE 1987). Laser irradiation be- EVANS1940; HAMAGUCHI 1975).These results suggest tween thecentrosome and the cell cortex results in that theasters and probably the astral microtubules play the cessation of centrosomal rotation and implicates a a key role in positioning the cleavage furrow. particular cortical site in the rotational events (HYMAN We have characterized a C. reinhardtii mutant strain, 1989). Treatmentof P blastomeres with cytochalasin D, bld2-I, that lacks centrioles/basal bodies. The strain was an inhibitor of microfilament assembly, results in the originally isolated by GOODENOUGH and ST. CLAIK failure of centrosomes to rotate (HILLand STROME (1975), has a single genetic lesion, and represents the 1988, 1990; HYMAN1989). Actin and actin capping pro- first allele at a previously unidentified locus. Cells lack tein have been shown to transiently accumulate at a centrioles/basal bodies in >99% of the population and cortical site that also has astral microtubules in its vicin- GOODENOUCHand ST. CWR (1975) proposed that the ity during thetime of centrosomal rotation;blastomeres BLD2 gene product is required for the formation of that do not undergo centrosomal rotation do notaccu- doublet and triplet microtubules. In addition to the mulate capping protein atthis site (WADDLEet al. 1994). basal body defect, we have found that cellular organiza- In the budding yeast, Saccharomyces cerevisicce, charac- tion is disrupted. The placement of the mitotic spindle terization of strains that have abnormal placement of is no longer correlated spatially or temporally with the the mitotic apparatus has implicated astral microtu- cleavage furrow and is misplaced relative to the sterec- bules, cytoplasmic dynein, dynactin complex, and actin typed cytological position seen in wild-typecells. The microfilaments in establishing or maintaining the spin- position of the cleavage furrow is also aberrant, but it dle in the correct position before anaphase. At the re- always colocalizes with specialized structures known as strictive temperature, the o-tubulin mutant straintub2- rootlet microtubules. The rootlet microtubules in bld2- 401 loses astral microtubules preferentially, the mitotic 1 cells are disorganized and often mispositioned in rela- spindles misorient; anucleate and multinucleate cells tion to the mitotic spindle. These observations clearly accumulate (HUFFAKERet al. 1988; SULLIVANand HUE implicate the BLD2 gene product as a component in- FAKER 1992). Deletion mutations in the cytoplasmic volved in the correct positioning of the mitotic spindle. dynein heavy chain gene DHCI (DYNl) orflM1 result The role of the BLD2 gene product may be in the initia- in minor spindle orientation defects (ESHELet al. 1993; tion or structure of functional centrioles, which may be LI et al. 1993; MCMILLANand TATCHELL1994). Spindle necessary to maintain the elaborate cytoskeletal organi- misorientation in dhcl (dynl) disruption strains is ac- zation critical for correct cell division in C. reinhardtii. companied by elongated astral microtubules and an increase in cells that are binucleate. Act5 is a member MATERIALS AND METHODS of the actin-related protein family Arpl and is homolo- Strains and culture conditions: The 137c mt+ strain of C. gous to the actin-like protein found in the vertebrate reinhardtii was used as the wild-type strainfor this study unless dynactin complex. Deletions of the ACT5 gene result otherwise noted. The bldl-l and the bld2-1 strains were origi- in minor spindle placement defects (GILL et al. 1991; nallyisolated by GOODENOUGHand Sr. CIAIR (1975) and MUHUAet al. 1994). obtained from the Chlamydomonas Genetics Center. The Several lines of evidence suggest that the position of aflagellate strain 426 was isolated in a screen for cells that could not swim after heat shock treatment (42", 30 min) of the cleavage furrow is determined by the position of a member of the F, generation from a cross of strain 137c to the mitotic apparatus. The furrow always forms at right the polymorphic strain, CC1952 (GROSSet al. 1988).Cultures angles to the spindle, bisecting it and can be reposi- were usually grown under constant illumination on agar plates tioned by moving the mitotic apparatus before the end or in liquid medium at 21" as described pre\aster are equally effective in generating a Deflagellation of wild-type cells was performed as described furrow (reviewed in RAPPAPORT 1986). Cleavage furrow in WITMAYrl nl. (1972). Spindle Position 947

Mating of Chlamydomonas strains: Mating of flagellated Meioticprogeny experiment: After mating, zygotes were cells was performed as described by HARRls (1989).Aflagellate placed on solid agar medium for 18 hr in the light and then cells were induced to mate with the additionof dibutryl cyclic placed in the dark for 23days. Plates were then exposed to adenine monophosphate (CAMP)and the phosphodiesterase chloroform vapors for 45 sec; nonzygotic cells are killed by inhibitor 3-isobutyl-1-methylxanthine to themedium (PAS this treatment. Zygotes were placed in liquid medium and QUAI,E and GOODENOUGH1987; DUTCHER1995). allowed to germinate. The presence of flagella was monitored Mapping of the bZd2-1 allele: The bld2-1 allele is unlinked over time by microscopic examination of the cells. to all linkage groups except linkage group 111 (datanot shown). TheBIB2 locus maps to linkage group I11 in crosses RESULTS to NIT2, ACl7, and TUAl: bld2-1-acl7-1: 33:0:7; bld2-I-nit2: 66:0:3; bld2-1-tual-1: 46:O:ll. Threepoint crosses between Cellular organization and division in Chlamydomo- bld2-1, nit& and tual-l indicate that BLB2 is distal to both nas: The highly stereotyped organization of the Chlamy- NIT2 and TUAl. domonas cytoskeleton results in a precise, structured Reversion of the bZd2-1 allele: Ultraviolet irradiation (W) organization of cell division. The interphase microtu- and ethyl methyl sulfonate (EMS) mutageneses were performed as described in LUX and DUTCHER(1991), except that bule organizing center (MTOC) is located at the ante- cells were grown at 21" and four transfers were performed. rior of the cell and consists of the basal bodies, four In a separate mutagenesis, log phase cultures were exposed specialized microtubules known asrootlet microtubules, to a IS7Csgamma (7) irradiation source (J. L. S. Shepherd and the nucleus-basal body connector (Figure 1). Each and Associates, model 143) at 498R/min for 30 min. After basal body nucleates a flagellum and is associated with irradiation, cultures were treated as for the W and EMS two rootlet microtubules, one ofwhich contains two mutageneses. Microscopic techniques: Observations were made and im- tubules and the other four tubules (RINGO 1967;JOHN- ages photographed ona Zeiss Axiophot microscope equipped SON and PORTER1968; GOODENOUGHand WEISS 1978; with a X100 plan-Neofluor objective. Several different anti- MOESTRUP1978). The four rootletmicrotubules are po- bodies were used in this study. The monoclonal antibody 3A5 sitioned in a cruciate pattern; they extend toward the is specific for all forms of a-tubulin tested, was used at a posterior of the cell just underneath the plasma mem- 1:lO dilution, and was a gift from M. T. FULLER(Stanford University) andJ. R. MCINTOSH(University of Colorado, Boul- brane. The remaining cytoplasmic microtubules extend der). The monoclonal antibody 611B-1 is specific for ace- from the vicinity of the basal bodies and rootlets and tylated a-tubulin (LEDIZETand PIPERNO 1986),was used at a form a cupshaped pattern with the basal bodies as the 1:50 dilution, and was purchased from Sigma Chemical Co. base. During mitosis, the duplicated pairs of centro- The monoclonal antibody 17E10 is specific for centrin, was used at a dilution of 1:500, and was a gift from J. L. SALISBURY somes and their old rootlet microtubules migrate away (Mayo Clinic). The polyclonal anti-Volvox actinantibodies from each other to the future poles of the mitotic spin- were generated against the conserved amino-terminal deca- dle and the basal bodies become centrioles; the cyto- peptide of Volvocalean algal actin and have been shown by plasmic microtubules and flagella disassemble (LEDIZET Western blot analysis to specifically recognize the 43-kD C. and PIPERNO1986). The direction of migration of the reinhardtii actin protein with no cross reactivity (GARLINERet basal bodies and rootlet microtubules results in the posi- al. 1994). This antibody was used at a 1:20 dilution and was a gift from D. KIRK(Washington University). Centrin staining tioning of the mitotic spindle perpendicular to the ante- was performed as described in SALISBURYet al. (1988). Stain- rior-posterior axis in the anterior of the cell (Figure 2) ingfor tubulin alone wasas described by HOLMESand (GAFFELand EL-GAMMEL1990). During the migration DUTCHER (1989).Double labeling with actin and tubulin anti- of the basal bodies, the four-membered rootlet microtu- bodies was performed using the protocol described by HARPER bules bend at approximate right angles. The distal por- el al. (1992). Secondary antibodies were Texas Red-goat anti- mouse IgG (Molecular Probes), andFITC-goat an ti-rabbit IgG tions of the rootlets orient perpendicularly to the spin- (Zymed). Standard control experimentswere performed and dle and the proximal portions areparallel to the spindle no aberrantlabeling or cross reactivity of any secondary anti- (Figure 2, prophase).In anaphase, new microtubule bodies was observed (data not shown).DAPI was used in water arrays begin to form perpendicularly between the nu- at 0.2 pg/pl final concentration. clei; they originate from the four-membered rootlet mi- Identification of mitoticstages: Assessment of mitotic stages was based on the shape, position, and appearance of crotubules (JOHNSON and PORTER1968; GAFFELand EL- the DAPI-stained DNA, as wellas the conformation of the GAMMEL1990; SCHIBLERand HUANC1991). mitotic spindle. Designation of prometaphase normally in- These microtubules spread to the interior of the cell volves distinct assessment of the attachment of kinetochores by telophase and together with thefour-membered to spindle microtubules; this was not possible in these experi- rootlet microtubules form the phycoplast. The cleavage ments. Therefore, the assessment of prometaphase was determined by slight differences in DNA shape and spindle mor- furrow forms along the phycoplast in a position that is phology frommetaphase and is somewhat subjective. In correlated cytologically with the position of the distal mammalian cells, telophase is usually defined by the reforma- ends of the parental four-membered rootlet microtu- tion of the nuclear envelope. Because the nuclear envelope bules and proceeds from the anterior of the cell to the does not completely break down in C. reinhardtiz, no attempt posterior (HOLMESand DUTCHER1989). The precise was made to identify this stage. The telophase stage identified by previous authors by the absence of the mitotic spindle or positioning of the rootlet microtubules in relation to the close position of the nuclei to each other (GAFFELand the spindle is therefore correlated with the exact posi- EL-GAMMEL1990; HARPERet al. 1992) was scored as cytokinesis tioning of the cleavage furrow that divides the cell into in this study, as cleavage furrows are present in such cells. equal halves. 948 L. L. Ehler, J. A. Holmes and S. K. Dutcher

trans flagellum cis flagellum / \

FIGUREI.-Diagram of interphase cell morphology of C. rei'nhnrdtii. The microtubule organizing center consists of two flagella, two ma- ture basal bodies with associated probasal bodies nucleus-basal body (not shown), and two sets of rootlet microtubules. The nucleus-basal body connector extends from the basal bodies to the nucleus. The eyespot is always associated with the cis four-membered mi- microtubules crotubule rootlet. The pyrenoid serves as a cellular landmark for the stereotypical organization of the cells in these experiments. The cytoplasmic microtubules and the pericentriolar material are not included in this diagram.

interphase

Genetic characterizationof the bM2 strain: The hZd2- strains: Several methods were used to confirm that bZd2- l strain (previously designated bald-2) is aflagellate and 1 cells lack flagella and basal bodies. By light micro- was isolated in a screen forcells unable to mate because scopic examination, 6Zd2-l cells were completely afla- flagella are requiredfor mating in Chlamydomonas gellate (n = 5000). As mentioned above, mating be- (GOODENOUCHand ST. CLAIR1975). However, this re- tween Chlamydomonas cells normally requires the quirement can be bypassed by artificially raising the presence of flagella, and we failed to detect any mating intracellular CAMP levels (PASQUALEand GOODENOUCH between IO8 wild-type and loH 6Zd2-Z cells, which sug- 1987). Between 0.5-30% of the hZd2-1 cells mated with gested thatnone of the 6Zd2-1 cellshave flagella. wild-type cells under these conditions to form zygotes. GOODENOUCHand ST. CLAIR(1975) examined the bZd2- The aflagellate phenotype of bZd2-l cells showed Men- 1 strain by electron microscopy and found that most delian segregation in 550 tetrads and is therefore likely cells lack an assembled basal body. We examined two to result from a single gene mutation. Greater than aflagellate meiotic progeny from our backcrossed 96% of the meiotic progeny from these crosses were strains by serial thin-section electron microscopy and viable, which suggests that the mutation does not have no basal bodies were detected inthese cells or in -1000 a dominant effect on meiotic viability and that there random thin sections (data not shown). are no unlinked modifiers of a growth or cell cycle The examination of bZd2-l cells by immunofluores- phenotype in hZd2-l strains. The BIB2 locus mapped cence allowed us to examine a larger number of cells 9.5 cM to the left of the centromere of linkage group for the presence of basal bodies. Using indirect immu- 111 (see MATERIALS AND METHODS). The bZd2-l mutation nofluorescence with an antibody against acetylated a- also conferred a recessive meiotic phenotype; homozy- tubulin that stains basal bodies and rootlet microtu- gous bZd2-I zygotes failed to germinate at all tempera- bules (LEDMETand PIPERNO1986), we examined de- tures tested from 16 to 32". This phenotype cosegreg- flagellated wild-typecells, a control aflagellate strain ated with the aflagellate phenotype in 75 tetrads. The (426), and 6Zd2-l cells. Deflagellation removes flagella bld2-l mutation was recessive to the wild-type allele in at the cell wall and leaves the basal bodies intact. In stable heterozygous diploid strains for the phenotypes 90% of deflagellated wild-type cells and in strain 426, described below (data not shown). Dominanceof actin two dot-like structures at the anterior end of the cell localization and temporal control was not determined. and therootlet microtubules were stained with the anti- Penetrance of theaflagellate phenotype in bM2-2 body (Figure 3, a and b). In bZd2-1 cells, acetylated mi- Spindle Position 949

FIGURE2.-Diagrammatic representation of the distribution of actin in relation to the spindle and rootlet microtubules during wild-type celldivision in Chlamydomonas. Interphase: the centrioles are found in a perinuclear location and function as basal bodies for the assem- rootlet rnicro- bly of flagella. Each basal body is associated with a two- and a four-membered microtubule root- actin let throughout the cellcycle. Preprophase: the DNA is condensed and the flagella are disassembled. The rootlets and basal bodies have not yet migrated. Actincolocalizes with the proximal portions of the rootlets and appears in a ring around the nucleus at the anterior of the cell. Prophase: the DNA is condensed butnot aligned at the metaphase plate.The spindle has begun to form but appears incomplete. Actin is distributed in a diffuse ring that does not ap telophase/cytokinesispre-prophase overlappear to the spindle. Thehas eyespot begunto breakdown. Metaphase: the DNA aligns along the metaphase plate; the spindle is fully formed. Actin is distributed in a ring around the spindle. Anaphase: the DNA sepa- rates; the spindle appears elongated. A small cleavage furrow is sometimes apparent and the actin appears distributed along the rootlets and the incipient fM-ow. Telophase/cytokinesis: the DNA moves back toward the furrow; the MTOCs appear fully formed with a ftdl complement of rootlets; the parental four-membered rootletsare anaphase prophaseanaphase coincident with the furrow. The phycoplast has formed and actin is concentrated at the furrow.

metaphase crotubules were stained, but no staining oftwo dot-like flagellar and cytoskeletal organizational phenotypes de- structures was detected in 2000 cells (Figure 3c). scribed in the following sections were the consequence To confirm these observations, wild-type cells were of the same mutation, we analyzed the coreversion of stained with an antibody against centrin. Centrin is a these phenotypes. We isolated revertant~/suppressors of calcium-bindingprotein that localizes tothe fibrous the aflagellate phenotype using enrichment screens for nucleus-basal body connector and the distal striated fi- cells that regained the ability to assemble functional fla- ber between basal bodies (SALISBURYet al. 1988). Anti- gella and therefore couldswim in liquid cultures. With- centrin stainingofwild-type cellsand cells from another out mutagenesis, no swimming cells were recovered in aflagellate strain, bldl-2, showed a pattern of two rib- 5 x 10'' bkl2-I cells. We obtained 15 revertants following bons that connect the nucleus tobasal the bodies where UV irradiation, three revertants from y irradiation and they are connected by an additional band of staining; 22 from treatmentwith EMS in four independentexperi- this pattern has been referred asto the "diamond ring" ments. Each revertant was crossed to a wild-type strain (Figure 3d) (WRIGHTet al. 1985). Cells without basal to determine if the reversion event was intragenic or bodies are unlikely to have distal striated fibers and an extragenic. In all 40 revertants, the reversions appeared extended nucleus-basal body connector. About0.5% of to result from an intragenic or closely linked event (Ta- the bld2-I cells observed have the wild-type diamond ble l). The likely intragenicnature of the reversion ring morphology (n= 2000; Figure 3e). It is likely that events was best illustrated in strains UV18R, UVl-IIR, there is sufficient basal body structure to assemble the and yl-lR, in which the second event is at most 0.1, centrin fibers in this subset of cells. 0.37, and 0.4 cM, respectively, from the original blit2- GOODENOUGHand ST. CLAIR (1975) observed thata I lesion (Table 1). The UV and y irradiation-induced small fraction ofbld2-I cells had rings of singlet microtu- mutationswere indistinguishable from wild-type cells bules; these cells may correspond to the cells with ex- based on swimming patterns, phototacticbehaviors, and tended centrin fibers. mitotic doubling times at temperatures from 16 to 32". Coreversion analysis of bM2-1: To determine if the Therefore, all phenotypes examined in Table 1 appear 950 L. L. Ehler, J. A. Holmes and S. K. Dutcher

FIGURE3.-lmn~r1noflr1orescent staining of interphase cells. (a-c) Cells were stained with anti-acetylated a-tubulin antibody and with FITGconjugated seconckary antibody. (a) Deflagellated wild-type cells.The acetylated microtubules appear in a cruciate pattern and two dot-like structures that are in positions that correlate to the expected position of basal bodies are clearly visible (n = 200). (b) Aflagellate strain 426. The acetylated microtubules appear in the wild-type cruciate pattern with the two dot-like structures clearly visible. (c) 01*12-1. The acetylated microtubules appear disorganized and abnormal in number and no dot-like structures are observed. (d and e)Cells were stained with anticentrin antibody and FITGconjugated secondary antibody. (d) bldl-I cells are aflagellate, but have basal bodies, and show the normal “diamond ring” pattern of wild-type centrin distribution. (e) bZd2-I cells show centrin staining, but very few cells have the “diamond ring” associated with extended centrin fibers (arrow). Bars = 10 pm. to arise from the 6112-1 allele. Interestingly, six of the up to 2 hr after mixing of the parentalcells. To increase revertants recovered after treatment with EMS acquired our ability to detect rescue of the aflagellate phenotype, an altered swimming phenotype that was tightly linked to a phototaxis assay was used that specifically monitored the Bld2- aflagellate phenotype (Table 1).This aberrant the behavior of dikaryons. Dikaryons weredistinguished swimming mav arise from flagella that are incorrectly from unmated haploid cells based on the observation templated or that are defective in structure. that zygotic cell walls adhere to one another to form a Rescue of the aflagellate phenotypein dikaryons: We pellicle. Phototactic dikaryons will form a pellicle that examined 6112-1 X RIJ12 transient dikaryons to deter- adheres to the glass wall of a test tube at the point of mine if the components that are necessary and suffi- light incidence. Zygotes formed from dikaryons that cient to assemble basal bodies were present in gametic were not phototactic formed a pellicle, but did not con- cells. If so, then wild-type cytoplasm might be able to centrate at the point of light incidence. The ability of rescue the aflagellate phenotype of hld2-I cells. During bLd2-I dikaryons to undergo phototaxis when mated to mating in Chlamydomonas, parental cells fuse within wild-type cells or to pf18 cells was tested. Cells with the 5-10 min and the resulting dikaryons remain in the Go pf18mutation have a paralyzed flagellar phenotype that interval of the cell cycle for “1.5-2 hr, during which is not rescued in dikaryons (STARLINGand RCVDALL the haploid nuclei fuse and the programof zygoticgene 1971). Homozygous pf18 dikaryons were not phototac- expression commences (ARMBURST et al. 1993). During tic while heterozygous dikaryons were, which suggests this interval, most mutant strains that have aberrant that a single pair of flagellawas sufficient for phototaxis motility or defects in the assembly offlagellar structures (Figure 4). Dikaryons formed in crosses of bld2-I with are rescued by thepresence ofwild-type cytoplasm wild-type cells performed phototaxis; the bld2-1 muta- (STARLINGand RZNDALI.1971; LUCK et al. 1977; tion did not have a negative effect on phototaxis. Dikar- DUTCHER~1 nl. 1984). yons formed in crosses of bld2-1 and pf18 were not pho- In populations of 6112-1 X BZB2 mating cells in which totactic, although dikaryons were formed as confirmed zygotes were later recovered at a frequency of lo%, no by the presence of the pellicle. This suggested that mo- quadriflagellate cells wereobserved by light microscopy tile flagella were not assembled and that the RLD2 cyto- Spindle Position 951

TABLE 1 Coreversion of bM2-1 phenotypes in intragenic revertant strains

Coreversion" Isolation No. tetrads Swimming name examined EyespotCytokinesis placement Rootlet MT phenotypeb Wl-1R 34 + + ND + Wl-2R 17 + + ND + W1-3R 31 + + ND + UV14R 15 + + ND + Wl-5R 29 + + ND + Wl-6R 49 + + ND + W1-7R 16 + + ND + WlSR 510 + + + + Wl-9R 17 + + ND + UVl-IOR 50 + + ND + UVI-1IR 136 + + + + W1-12R 17 + + ND + Wl-13R 52 + + ND + W2-1R 60 + + ND + W2-2R 31 + + ND + yl-1R 125 + + + + y2-1R 18 + + ND + y2-2R 34 + + ND + EMS7-1 25 + + + Aberrant' EMS7-2 35 + + + + EMS7-3 14 + + + + EMS74 17 + + + + EMS7-5 12 + + + + EMS7-6 18 + + + + EMS8-5 32 + + + + EMS86 16 + + + + EMS87 13 + + + + EMS9-1 15 + + + + EMS9-2 33 + + + + EMS9-3 19 + + + + EMS94 31 + + + Aberrantc EMS9-5 32 + + + Aberrant' EMS9-6 32 + + + + EMS103 33 + + + Aberrant" EMS 104 98 + + + Aberrant' EMSl 0-5 34 + + + Aberrant" EMSlO-7 13 + + + + EMSlO-8 15 + + + + EMSlO-9 17 + + + + EMSl 0-1 1 16 + + + + ND, not determined. a Revertant strains were crossedby wild-type cells and tetrad progeny were examined for the cosegregation of flagella, cytokinetic fidelity, eyespotplacement, rootlet number and placement, and any new phenotype exhibited by the revertant strain. Cytokinesis was observed by phase microscopy (n= loo), eyespot position was examined by DIC microscopy (n= loo), and rootlets were examined by immunofluorescence (n= 20) using the antibody against acetylated a-tubulin. In addition, the number of nuclei per cell (n = 100) and daughter cell size (n = 25) were also scored and were reverted in all strains examined. + indicates reversion. Swimming phenotypes were assessed as compared with wild type; + indicates no difference from wild type. 'Linkage of abnormal swimming to the BLDP locus was tested in crosses to the bld2-1 strain and in all cases the swimming phenotype was tightly linked (n = 20-30 tetrads each). plasm contributed from the pf18 parent was unable to eny: To determine if the components that are neces- rescue the aflagellate phenotype; however, we cannot sary and sufficient to assemble basal bodies were pres- rule out the unlikely possibility that nonmotile flagella ent in bZd2-1 cells during or directly after meiosis in the were assembled in a small, microscopically undetected initial mitotic divisions, we examined the phenotype fraction of the dikaryons. of cells that had just completed meiosis. Heterozygous Rescue of the atlagellate phenotype in meiotic prog- BLD2/bld2-1 zygotes were germinated and the presence 952 L. L. Ehler, J. A. Holmes and S. K. Dutcher

Parental Strains Dikaryon Observed Phototaxis Phenotype

wild-type x wild-type four normal flagella I + ly wild-type x pf18 I+ two normal flagella, two paralyzed flagella wild-typex bld2-1 1 + 1 normalflagellatwo

pf 18 X bld2- 1 - two paralyzed flagellaa

FIGURE4.-Dikaryon phenotypes of bld2-1 strains. BLD2 cells were mated to bld2-1 cells and the resulting dikaryons were assayed for their ability to perform phototaxis. Formation of dikaryons was scored by the subsequent presence of zygotes. The ability of dikaryons to phototax was scored by the presence or absence of zygotes at the point of light incidence (see MATERIALS AND METHODS). “Due to low mating efficiency, this phenotype was not observed directly. of flagella was monitored over time by microscopic ex- germination. The phenotypic rescue in postmeiotic amination. Because each zygote will produce two zoo- cells suggests the BLD2 gene does not need to be tran- spores with the wild-type genotype and two zoospores scribed in each cell cycle. However, the introduction of with the bZd2-1 genotype, 50% of cells examined should wild-type cytoplasm to Go arrested cells was not suffi- have flagella. If there is rescue of the aflagellate pheno- cient to provide rescue (see above). One explanation type then >50% of cells will have flagella. After the for these observations is that the BLD2 gene product first mitotic division, >70% of the meiotic progeny had may be a component of a structure thatis not normally flagella (n = 500cells; Figure 5). The homozygous assembled in Go, such as the basalbody annulus BLB2 control meiotic progeny had a similar percentage (GWLD1975). of flagellated cells (n= 500). The presence of flagella Eyespot and rootlet microtubule placementin bld2-2 on meiotic progeny decreased over time, so that after cells: In each cellcycle, the eyespot is disassembled. approximately four to five generations, -45% of cells After mitosis, each daughter assembles a new eyespot were flagellated. No increase in the number of unifla- opposite the cleavage furrow in close association with gellate cells was observed. Therefore, genotypically bZd2- a newly formedrootlet microtubule (HOL~MESand 1 cells assembled basal bodies during meiosis from the DUTCHER1989). Bothwild-type and aflagellate bZdl-1 cytoplasm contributed from the BLD2 parent andmain- cells assembled new eyespots opposite the cleavage fur- tained them for a limited number of generations after row (n = 500). In contrast, bZd2-1 cells contained eyespots that were not on the cell equator or opposite the cleavage furrow in 45% of the dividing pairs (n= 500). Because the eyespot is found in an exact cellular loca- 90 loo I tion with respect to the four-membered rootletmicrotu- BLD2fBLD2 bules in wild-type cells, it was possible that the abnormal 0 positioning of the eyespot in bZd2-1 cells was associated flagella with a defect in the placement of the rootlet microtubules. The rootlet microtubules of bZd2-1 and wild-type 40 bM2-1IBLD2 cells were stained using the antibody against acetylated 30 I a-tubulin. The stereotyped cruciate organization of 2o I rootlet microtubules observed in wild-typecells lo I I (HOLMES andDUTCHER 1989) was absent in most bkd2- 0 5 10 15 20 25 30 35 40 45 50 55 Hours after germination 1 cells; both the placement and the number of rootlet microtubules were abnormal (Figure 3c). During mito- FIGURE5.-Rescue of the Bld- aflagellate phenotype in sis and cytokinesis of bld2-1 cells, the rootlet microtu- meiotic progeny. BLDZ/BLD2 and BWZ/bldZ-I zygotes were bules were often mispositioned (see below). placed in liquid medium and allowed to germinate. Zygotes germinated at -16 hr after exposure to light. The number Fidelity of cleavage furrow placement in bld2-I cells: of cells with flagella was monitored over time by light micros- The position of the cleavage furrow is always correlated copy. Each point represents an average of the numberof cells with the four-membered rootlet microtubules in wild- with flagella at the median time point for two experiments. typecells (JOHNSON and PORTER1968; HOLMESand The doublingtimes of individual meiotic progeny were monitored over the course of the experiment and were indistin- DUTCHER1989; GAFFELand EL-GAMMEL1990). There- guishable from each other. Five hundred cells were counted fore, it was possible either that theposition of the cleav- for each time point in each experiment. age furrow was disassociated from the rootlet microtu- Spindle Position 953 bules in bU2-I cells or thatthe cleavage furrow was a.wild-type mother cell mispositioned along with the mispositioned rootlet mi- wall crotubules. The location of the cleavage furrow in relation to the rootlet microtubules was determined using the antibody against acetylated a-tubulin; the position of the cleavage furrow always correlated with a rootlet microtubule in dividing bU2-I cells (n = 40). Light microscopic examination of 6Zd2-1 cells showed that the 3 cleavage furrow was placed such that unequally sized Isister cells # pairs n nonsister cells sister cells were often produced. The area of cells in of cells 2 wild-type and hld2-I cultures was measured and the difference in area between sister cells and between unre- 1 lated (nonsister) cells was compared (Figure 6). In wild- type populations the mean difference in area between sisters (6.4 pm') was much less than between unrelated difference in area ( !am2) cells (15.3 pm'). This suggests that the placement of the cleavage furrow in Chlamydomonas is nonrandom and generally results in two equal-sized cells. In hM2-I b. bld2-1 cells, the distribution of the mean differences in cell area between sister cells was similar to that of the mean difference between unrelated cells, which indicates that # pairs of cells 2 the areasof sister cells are not correlated to each other. Isister cells Therefore, cleavage furrows are misplaced in hld2-1 cells. o nonsister cells Spindle andcleavage furrow placement in bZd2-2 1 cells: We examined hM2-I cells to determine if the misplacement of the cleavage furrow was accompanied by a corresponding misplacement of the spindle, as oh served in a C. Plegnns par1 mutant strain (KEMPHUES 2 4 6 8 10 12 14 16 18 2022 24 26 pt nl. 1986). bZd2-1 cells exhibited a longer population difference in area ( pm2) doubling time in log phase cultures thanwild-type cells FIGURE6,"Fidelity of cleavage furrow placement in wild- (data not shown).A significantnumber of nondividing type and bld2-I cells. The difference in the area between sister bM2-I cells with 0 or 2 nuclei was observed byDAPI cells, whichare still within the mother cell wall, wascompared staining when compared with bldl-I cells (x' = 15.72, with that between randomly selected single cellsin the popula- P < 0.001; Table 2), which suggests that the spindle tion. The random population represents cells at all stages of the cell cycle as opposed to the specific stage of the sister may no longer be oriented correctly in relation to the cells and therefore results in a relatively broad distribution cleavage furrow. The pyrenoid, an easily visualized or- of cell sizes. (a) Wild-type cells. The mean difference in area ganelle located at the posterior end of the cell, was for sister cells was 6.4 pm' and the variance was 6.2 pm' (n used as a cellularlandmark. The orientation of the = 95); for nonsister cell pairs the mean was 15.3 pm' and the spindle was assessed by DAPI staining of the chromo- variance was 57.8 pm' (n = 87). (b) bld2-1 cells. The mean difference in area for sister cellswas 11.93 p.m' and the vari- somes, and the orientation of the cleavage furrow by ance was 38.6 pm' (n= 85); for nonsister cell pairs the mean phase microscopy. Only 25% of bld2-I cells examined was 13.9 pm' and the variance was 52.5 pm' (n = 83). Area had the normal orientationof the spindle and thecleav- = T7d, where 70 = width of cell at widest point and 1 = length age furrow (n= 81, Figure 7). Thespindle and cleavage of cell at longest point. furrow were misplaced with respect to each other in >36% of the scored cells. In relation to the pyrenoid, in relationship to microtubule structureswas not exam- the spindlewas oriented incorrectly in 48% of cells and ined directly (HARPER~1 nl. 1992). The distribution of the furrow was misplaced in 63% of cells. In the most actin in relation to the position of the spindle androot- extreme cases, the furrow was positioned parallel to the let microtubules duringmitosis and cytokinesis in wild- newly formed nuclei, such that the nuclei would have type and in OZd2-I cells was examined by immunofluo- been cleaved oranucleate cells would have been rescence to determine if actin localization is correlated formed (Figures 8 and 10, d and g). to the defect in positioning the spindle in hld2-1 cells. Distribution of actin in relationto microtubule struc- During preprophase in wild-type cells, the flagella tures in mitotic cells: Actin microfilament networks are were resorbed and cytoplasmic microtubules disassem- postulated to be involved in positioning the mitotic ap bled, the rootlet microtubules andbasal bodies had not paratus in other organisms. The dynamics of actinlocal- migrated, and the DNA was condensed. Actin colocalization during mitosis in Chlamydomonas have been ized with the proximal portions of the rootlet microtu- described previously; however, the distribution of actin bules and in some cases was also in a ring around the 954 L. L,. Ehler,J. A. Holmesand S. K. Dutcher TABLE 2 microtubule, along the rootlet microtubules only, or The number of nuclei per cell in bldl-1 and bld2-1 strains along rootlet microtubules and near the spindle but not in a ring (Figure 10, d-g). Ring patterns were often faint No. of nuclei or diffuse; sharp rings such as those seen in wild-type cells per cell Percent cells were rarely observed. In prometaphase and metaphase, with abnormal Strain 0 1 2 no. of nuclei X? the actin distribution was also variable. It appeared both along rootlet microtubules and in diffuse rings or half bldl-I 2.8 0 384 11 15.72 rings along the spindle or appeared near but not along 01d2-1548 8.514 37 the spindle (Figure 10, h-k and I-q). The classic actin The number of nuclei per cell in both bldl-1 and bld2-I “metaphase ring” of wild-type cells was rarely observed mutant strains was determined during logphase growth. Nu- and even then was less defined than in wild-type cells. clei were visualized using DAF’l to stain the DNA. Mitotic However, cells with a wild-type appearing actin distribu- cells were not counted. The chi-square statistic used was a contingency test. tion did not necessarily have correct spindle placement (Figure 10, h-k). In anaphase, actin was concentrated nucleus (Figure 9, a-c). In prophase, actin was distrik along the rootlet microtubules but could also still be local- uted in a loosely defined ring or half ring around the ized around part of the spindle (Figure 10, r-w). Al- forming spindleand colocalized with the rootlet micro- though there appeared to be no overall consensus distri- tubules that arced over the spindle (Figure 9, d-g). In bution of actin during most of mitosis inbM2-1 cells, actin prometaphase, actin was usually found in a full ring colocalized with the rootlet microtubules in all but one around the spindle and along therootlet microtubules cell where rootlet microtubules were detected (n = 27). that continued to arc over the spindle (Figure 9, h- The microtubule cytoskeleton of bZd2-I cells was also m). In metaphase, actin was distributed in a sharply disorganized during mitosis. The rootlet microtubules defined ring around the spindle (Figure 9, n-q). In were mispositioned and could be present anywhere in anaphase, a small cleavagefurrow was sometimes appar- the cell as opposed to the strict placement observed in ent; actin colocalized with the rootlet microtubules and wild-type cells. In particular, the rootlet microtubules theincipient furrow (Figure9, r-w). In cytokinesis, rarely arced over the spindle as in wild-type cells (com- actin colocalized to the rootlet microtubules and the pare Figure 9d with Figure 10h). In some cells, the phycoplast, which were coincident with the cleavage poles of the mitotic spindle were out of alignment with furrow (data not shown). one another, so that the half spindles met offcenter or The distribution of actin was altered during most stages at anangle in relation to each other. Rarely, monopolar of mitosis in bH2-1 cells. During preprophase, actin local- or tripolar spindles were observed. iiration was indistinguishable from that of wild-type cells The distribution of centrin reorganizes during mitosis (Figure 10, a-c). However, by prophase actin distribution (SALISBURYet al. 1988). In wild-typecells, the centrin became highly variable and did not appear as focused as localization is near thebasal bodies and thenuclear enve- in wild-type cells. Common localiiration patterns included lope. At metaphase, the majority of the staining is at half ringsalong one side ofthe spindle and along a rootlet the poles and centrin fibers emanate toward the spindle

normal DNA misplaced; furrow correct FICXIRE7.”Spindle and cleavage furrow misplacement in bM2-I cells. Diagrammatic represen- wild-type: 84% 2% tation of DNA position and cleavage furrow posi- bld2- 1 25% 12% tion in relation to the position of the pyrenoid in wild-type and bId2-1 cells. The position of the DNA represents the orientation of the mitotic spindle;therefore two nuclei were required to make theassessment. Twelve percent of wild-type cells were unscorable; n = 50 for wild type, n = 81 for bId2-1.

DNA correct: furrow misplaced DNA and furrow misplaced wild-type: 2% 0% bld2- 1: 27% 36% FIGURE8.-lm~nrlnofluo~~~sccnt imqcs ol'spintllc and clcavage furrow misplacement in M2-I cclls. (a and c) The cells were stained with anti-cu-tubulin antibody and with FITGconjugated secondary antibody. (b and d) Phase contrast images of cells. Note that the cleavage fllrrow was not perpendicular to the spindle in either cell. Bar, 10 pm. microtubules (SALISRURYPt nl. 1988). We have examined that maps to linkage group 111. Based on this map posi- the pattern of centrin staining during mitosis in bld2-I tion, the BLD2 gene product cannot be actin (C. SIL cells and found that the localization was similar to that FLOW, personal communication), cy-tubulin (JAMES et nl. in wild-type cells, although the staining that emanated 1993; A. M. PRERLEand S. K. DUTCHER,unpublished toward the spindle may be reduced (data not shown). results), 0-tubulin (BOLDUCet nl. 1988), y-tubulin (C. Karyokinesis and cytokinesis are mistimed in bld2-I SILFLOW,personal communication), or centrin (TAIL- cells: In wild-type cells, the initiation of the cleavage LON et nl. 1992; K. MILL.. and S. K. DUTCHER,unpub furrow can occur as early as late anaphase (Figure 9u); lished results). Attempts to isolate additional alleles more commonly the cleavage furrow initiates in telo- with similar phenotypes at the BIB2 locus using stan- phase (GAFFEI.and EL-GAMMEI.1990). bld2-1 cells show dard mutageneses and screening for aflagellate cells a defect in coordinating karyokinesis and cytokinesis; have been unsuccessful. Among 250 independent afla- cleavage furrows were observed at all nuclear stages of gellate isolates, no alleles at theBZB2 locus were recov- mitosis except preprophase (Figure 10, g and 0; Table ered (S. K. DUTCHER,unpublished results). This may 3).Assuming cleavagefurrows initiated in late anaphase indicate that the6112-1 allele results in a rare phenotype are normal, 46% of 0132-1 cells (n= 46) initiated cytoki- for this locus. It is possible that the majority of muta- nesis at inappropriate stages of karyokinesis compared tions at the BZB2 locus are lethal. Alternatively, it is with wild-type cells (Table 3). "hen a furrow was pres- possible that the bM2-I allele produces a recessive poi- ent duringan early stage of mitosis, actin was invariably son (antimorphic) gene product thatresults in a pheno- concentrated along it and usually colocalized with a type that other alleles do not display and that this phe- rootlet microtubule. Actin was often distributed along notype is different from the null phenotype. Screens or near the spindle. Incorrect spindle placement did using the aflagellate phenotype as a basis for identifjmg not depend on thepresence of cleavagefurrows during new b132 alleles would fail in these cases. We believe early mitotic stages. Some cells with furrows in an early it is unlikely that the 6132-1 allele represent5 the null mitotic stage had spindles positioned correctly, and phenotype for this locus. The BLD2 gene product may some cells without furrows in an early mitotic stage had be a structural component of the basal body or may be aberrantly positioned spindles. involved in the assembly of microtubule based structures. Although the BLD2 geneproduct could be a DISCUSSION transcription factor for genes that encode basal body Genetics of the BLD2locus: The bM2-I mutation rep proteins, this seems unlikely in view of the new pheno- resents a single allele at a previously unidentified locus types in six of the intragenic revertants. Therefore, we

Spindle Position 9.57

FIGURE10.-The distribution TUBULIN ACTIN DAP I LIGHT of tubdin and actin during mitosis of bM2-1 cells. Cells were stained with antiiu-tubulin, anti- acetylated a-tubulin, and anti-ac- PP tin primary antibodies and with Texas Red and FITC conjugated secondary antibodies (see MATERI- AIS AND METI.IOI)S). Bar, 10 pm. (a-c) Preprophase. Actin localization was similar tothat seen in wild-type at this stage. (d-g) Pro- phase. The secondpole of the spindle seen in d was out of the P plane of focus and was located at the anteriorof the cell, just to the left of the rootlet microtubule. Ac- tin was localized along the rootlet microtubule (arrows) that was associated with a prominent cleavage furrow (g. arrowhead)that did not bisect the spindle. This cy- PM tokinetic event would likely have resulted in the production of an anucleate cell. (h-k) Prometa- phase. Actin was present in a half ring around the spindle and also colocalized to a rootlet microtubule (arrows in h and i). The spindle \vas not positioned perpendic- M ular to the anterior-posterior axis of the cell as determined by the position of the pyrenoid (arrow in k). (I-q) Metaphase. (I) Anti-tubulin staining in the focal plane of a rootlet microtubule, (m) anti- actin in the same focal plane, (p) anti-tubulinstaining in the focal plane of the spindle, (4) anti-actin in the same focal plane. Actin colocalized with both rootlet microtubules (arrows); a cleavage furrow has initiated in this cell. (r- v) Anaphase (anterior view). (r) Anti-tubulin staining in the focal plane of a rootlet microtubule positioned along theaxis of the spin- A dle. The spindle was positioned so that only one pole was in the plane of focus. (s) Anti-actin in the same focal plane, (v) anti-tubulin in the focal plane of the rest of the spindle, and (w) anti-actin in the same focal plane. Actin appeared to localize along the rootlet microtubule and along theside of the cell (arrows in rand s); it was also present in a ring around the toppole of the spindle.

GUCHI 1975; DE\Y)KEPL d. 1989; HARRISand GEWALT centrosomes at the spindle poles, and are clearly not 1989; reviewed in R/WP,\POKT 1986; SA~TERWAITEand part of the spindle itself. In addition, the association POILARD1992; STROMI.:1993). The mitotic rootlet mi- between the rootlet microtubules and the position of crotubules of Chlamydomonas are similar to astral mi- the cleavage furrow is striking. In wild-type cells, actin crotubules and unlike the cytoplasmic microtubules colocalizes with the rootlets throughout mitosis and that disassemble in preprophase. The rootlet microtu- concentrates along them during anaphase and cytoki- bules are present throughout mitosis, originate in the nesis, and the position of the distal portions of the 958 L. L. Ehler, J. A. Holmes and S. K. Dutcher

TABLE 3 Mistiming of karyokinesis and cytokinesis in bld2-1 cells

No. of cells at mitotic stage"

Prophase Prometaphase Metaphase Anaphase Percent of total bld2-1 wt bld2-1 wt b1d2-1 wt bld2-1 wt bld2-1 wt Furrow present 9 0 5 0 7 0 3 2 47 8 No furrow 13 9 4 2 8 10 2 3 53 92 Wild-type (wt; n = 26) and bld2-1 cells (n = 51) were observed by indirect immunofluorescence using antibodies against both acetylated and nonacetylated a-tubulin and DNA was visualized by DAPI staining. The presenceor absence of cleavage furrows was determined for cells that contained mitotic spindles. " Cells in telophase normally may have initiated furrows, so this stage was not assessed independently from cytokinesis. rootlet microtubules invariably predicts the position of of a Saccharomyces cermisiae P-tubulin mutation tub2-401. the cleavage furrow. Although the overall distribution The loss ofastral microtubules observed at the restrictive of actin in bld2-1 cells during mitosis was variable, actin temperature in tub2401 strains is accompanied by mis- colocalized to the astral rootlet microtubules in 96% positioning of the spindle and the production of an- of cells, and cleavage furrows colocalized with rootlet ucleate and multinucleate cells; however, the actin mi- microtubules even when the rootlet microtubules were crofilament network is apparently functional, as the cells aberrantly placed.The colocalization of actin with root- continue the cell cycle (SULLIVANand HUFFAKER1992). let microtubules throughout mitosis indicated that the Many lines of evidence support the model thatactin association of rootlet microtubules with cleavage fur- microfilaments have an important role in the position- rows was not coincidental. This indicates that the root- ing of the mitotic apparatus via the astral microtubules let microtubules play a role in positioning the cleavage (HYMANand WHITE1987; PALMERet al. 1992; WADDLE furrow, and it is probable that the defect in cleavage et al. 1994). For example, at therestrictive temperature furrow positioning seen in bld2-1 cells is a result of the an actl-4 mutant strain of S. cmmisiae shows complete aberrant positioning of the astral rootlet microtubules loss of actin cables and randomization of the few corti- and not of a defect in actin localization. cal actin patches remaining (DUNNand SHORTLE1990; The centrosomaldefect in bld2-I results in the mispo- PALMERet al. 1992). Actin disorganization is accompa- sitioning of the mitotic spindle via the astral microtu- nied by the failure to maintain spindle position at the bules: In bld2-1 cells, the mitotic apparatus was posi- restrictive temperature (PALMERet al. 1992).Aside from tioned aberrantly in 248% of cells and mispositioning the positioning defect, the microtubule cytoskeleton of was observed in any stage of mitosis that had an assem- actl-4 cells appears normal during mitosis (PALMERet bled spindle. We suggest that this incorrect positioning al. 1992). Therefore, alterationof either astral microtu- resulted from centrosomes that are unable to maintain bules or actin microfilaments results in the misposition- the normal cytoskeletal organization required for cor- ing of the mitotic spindle, but the disruptionof one of rect spindle orientation. In wild-type cells, the rootlet the cytoskeletal networks does not necessarily affect the microtubules are maintained in a precise orientation integrity of the other. relative to the basal bodies (RINGO 1967; WEISS 1984; In contrast, the disorganization of rootlet microtu- GAFFELand EL-GAMMEL1990) and are normally linked bules and the spindle positioning defect in bld2-1 cells to the basal bodies via the striated root fibers (WEISS was accompanied by abnormal localization of actin. In 1984); this linkage is not possible in bld2-1 cells. This wild-type cells, migration of the centrosomes occurs in cytoskeletal disorganization could directly result in the preprophase and prophase (GAFFELand EL-GAMMEL. aberrant migration of the centrosomes, failure of the 1990); actin localizesmainly to the rootlets before spindle to maintain a correct position, or both. It is centrosome migration. In bld2-1 cells, there waswild- clear from experiments in other organisms that astral type localization of actin before centrosomal migration microtubules have a significant role in both establishing and abnormal actin localization was observed only after and maintaining spindle position (HYMANand WHITE the centrosomes attempted to migrate. This abnormal 1987; HUFFAKERet al. 1988; HYMAN1989; SLJLLJVAN and actin localization was observed as a failure either to HUFFAKER1992). The occasional displaced spindle make or maintain the actin metaphase ring. However, halves, and monopolar and tripolar spindles observed bld2-1 cells continued to localize actin to the rootlet during mitosis further indicated the deficiency of microtubules during mitosis and showed no defect in centrosomal organization in bld2-1 cells, whereas the the completion of cytokinesis, which indicates that the spindle itself appeared to function normally. The spin- actin cytoskeleton was at least partially functional. dle positioning and anucleate/binucleate phenotypes Therefore, if actin is involved in the positioning of the exhibited by bld2-1 cells are similar to the phenotypes spindle via the rootletmicrotubules in Chlamydomonas Spindle Position 959

(as proposed by HARPER et al. 1992), it is likely that the in initiating or stabilizing the formation of microtubules lack of basal bodies disrupts the cytoskeletal architec- was proposed. However, it is clear from this study that ture needed to maintain the orientation of the astral centrioles/basal bodies are not necessary for the nucle- rootlet microtubules with the forming spindle, which ation of cytoplasmic, rootlet/astral, or spindle microtu- results in the mispositioning of the spindle and the bules during the mitotic cell cycle in Chlamydomonas. apparent abnormal actin distribution. We propose that Rather, it appears that the role of the centriole in some the lack of basal bodies in the mitotic centrosome leads celltypes may be to provide an organizedcytoskeletal to the observed defects in actin localization and spindle superstructure that is critical for the correct placement of and astral rootlet microtubule positioning. the spindle and the cleavage furrow during cell division. Karyokinesis and cytokinesis are uncoupled in bld2-2 We thank Dr. KENT MCDONAILIfor theelectron microscopy of cells: The presence of cleavage furrows during early bZd2-1 cells; Dr. JOHN HARPER fortechnical advice on actin staining; stages of mitosis in bZd2-1 cells suggests a defect in the ANDREA M. PREBLE,AMY R. SCHUTZ,KEVIN M11.1.s. and Dr. CAROLYN timing of cytokinesis withnuclear division. Uncoupling SII.FI.OWfor unpublished data; Dr. DAVIDKIRK, Dr. MARGARETFUL of karyokinesis and cytokinesis could result from the LER, Dr. JEFF SAI.ISBURY,and Dr. J. RICHARD MCINTOSHfor generous gifts of antibodies; and the membersof the DUTCHERlab for numer- loss of a checkpoint control mechanism or the disrup- ous and helpful comments on the manuscript. This work was begun tion/delay of a crucial event that occurs after the inde- with a Searle Scholar’s Award and a Creative Work and Research pendence of the two pathways has been established. In Award from the University of Colorado. It was supported by the Na- previous studies of mitosis in wild-type Chlamydomo- tional Science Foundation (grant DMB8916641) and the National nas, the addition of hydroxyurea blocks the completion Institutes of Health (grant GM-32843). J.A.H. and L.L.E. were supported in part by a training grant from the National Institutes of of DNA replication and the subsequent formation of a Health (5T32 GM-07135). spindle, butcleavage furrows are still initiated (HARPER andJOHN 1986). Thissuggests that cytokinesis proceeds independently ofDNA replication and karyokinesis LITERATURE CITED once a commitment to divide has been achieved. This ARMBURST,E.V., P. J. FERRISand U. W. GOODENOUCH,1993 A experiment also suggests that there is not a checkpoint mating type-linked gene cluster expressed in Chlamydomonas zygotes participates in the uniparental inheritanceof the chloro- function that links DNA replication and spindle forma- plast genome. Cell 74: 801-811. tion with cytokinesis. This result is unlike the situation BEAMS,H. W., and T. C. EVANS,1940 Some affects of colchicine upon in S. cermisiae, where both cytokinesis and karyokinesis the first cleavage in Arbacia punctulata. Biol. Bull. 79 188-198. BOLDUC,C., D. L. VINCENTand B. HUANG,1988 P-tubulin mutants are blocked by a checkpoint until the DNA is correctly of the unicellular green alga Chlamydomonas reinhardtii. Proc. replicated (HARTWELL and WEINERT1989). In addition, Natl. Acad. Sci. USA 85: 131-135. treatment of wild-type Chlamydomonas cells with vin- CAIARCO-GILIAM,P. D., M. C. SIEBERT,R. HUBBLE,T. MITCHISONand M. KIRSCHNER,1983 Centrosome development in early mouse blastine for 24hr shifts the majority of the population emblyos as defined by an autoantibody against pericentriolar to a 2C or greater ploidy ofDNA. Treated cells are material. Cell 35: 621-629. uninucleate and do not form spindles, but >30% of CHENG,N. N., C. M. KIRBY and IC J. KEMPHUES, 1995 Control of cleavage spindle orientation in Cuenorhabditis ekgans: the role of the cells initiate at least one cleavage furrow (A. R. the genes par-2 and par-3. Genetics 139: 549-559. SCHUTZ,L. L. EHLER,and S. K. DUTCHER,unpublished DEBEC,A,, 1984 Evolution of karyotype in haploid ccll lines of Bo- observations). This experiment suggests that there is scphila melanogaster. Exp. Cell Res. 151: 236-246. DEBEC,A., A. Szii1.1.6s1and D. SZOILOSI,1982 A Drosophila melano- not a microtubuledependent cell-cycle block to cytoki- gaster cell line lacking centrioles. Cell 44 133-138. nesis and provides additional evidence that there is not DEVOREJ. J.. G. W. CONRADand R. RAPPAPORT,1989 A model for a G2 or M phase checkpoint that links karyokinesis with astral stimulation of cytokinesis in animal cells. J. Cell Biol. 109: 2225-2232. cytokinesis in Chlamydomonas. Therefore, it is likely DUNN,T. M., and D. SHORTIE,1990 Null alleles of SAC7 suppress that the uncoupling of karyokinesis and cytokinesis ob- temperaturesensitive actin mutations in Sarrharomycps cereuisine. served in bld2-1 cells is due to a disruption/delay of an Mol. Cell. Biol. 10 2308-2314. DUTCHER,S. IC, 1995 Mating and tetrad analysis in Chlamydomonas event that occurs after the pathways have separated. reinhardtii. Methods Cell Biol. 47: 531 -540. One possible role of the centriole/basal body is to DUTCHER,S. IC, B. HUANG andD. J. L. LUCK,1984 Genetic disser- maintain cytoskeletal organization during cell division: tion of the central pair microtubules of the flagella of Chlamydo- manus reinhardtii. J. Cell Biol. 98: 229-236. The role of the centriole/basal body in cell division has ESHEL,D., L. A. URRESTARAZU,S. VISSERS,J. C. JAUNIAUX, J. C. VAN been under investigation for many years. The isolation of VIIET-REEDJIKet al., 1993 Cytoplasmic dynein is requiredfor cell lines lackingcentrioles showed that centrioles are not normal nuclear segregation in yeast. Proc. Natl. Acad. Sci. USA 90: 11172-11176. essential for cell division and called into question the idea FBIJX,M.-A,, C. ANTONY,M. WRIGHT and B. MARO, 1994 that they were necessary at all. Recently, in vitro studies Centrosome assembly in vitro: role of y-tubulin recruitment in have led to the hypothesis that centrioles provide a tem- Xenopus sperm aster formation. J. Cell Biol. 124 19-31. FULTON, C., and A. D. DINGLE,1971 Basal bodies but not centrioles, plating function necessaq for the assembly of microtu- in Naegleria. J. Cell Biol. 51: 826-836. bules in cells. Experiments using the Xenopus sperm GAFFEL,G. P., and S. EI,-GA“EI., 1990 Elucidation of the enigma of centrosome suggest that y-tubulin is necessaq but not the “metaphase band”of Chlamydomonas ra’nhardtii. Protoplasma 156 199-148. (F~LIXet sufficient for the nucleation of microtubules al. GARLINER,L. C., E. O’TOOLE,C. A. PERRONE,T. GIDDINGSand M.E. 1994; STEARNSand KIRSHNER 1994); a role of centrioles PORTER, 1994 Components of a “dynein regulatory complex” are 960 L. L. Ehler, J. A. Holmes and S. K. Dutcher

located at the junction between the radial spokes and the dynein LUX, F. G., 111, and S. K. DUTCHER,1991 Genetic interactions at the arms in ChZamydonrmas flagella. J. Cell Biol. 127: 1311-1326. FLAlO locus: suppressors and synthetic phenotypes that affect GIIL,S. R., T. A. SCHOER,I. SZ-, E. R. STEUER,M. P. SHEETZet al., the cell cycle and flagellar function in Chlamydomonas reinhardtii. 1991 Dynactin, aconserved, ubiquitously expressed compo- Genetics 128: 549-561. nent of an activator of vesicle motility mediated by cytoplasmic MANIOTIS, A,, and M. SCHIJWA,1991 Microsurgical removal of dynein. J. Cell Biol. 115 1639-1650. centrosomes blocks cell reproduction and centriole generation GOWENOUCH,U. W., and H. S. ST. CIAIR,1975 bald-2: a mutation in BSC-1 cells. Cell 67: 495-504. affecting the formation of doublet and triplet sets of microtu- MCMIILAN,J. N., and K. TATCHEIL,1994 The JNMl gene in the bules in Chlamydomonas reinhardtii. J. Cell Biol. 66: 480-491. yeast Saccharomyces cereuisiaeis required for nuclear migration and GOODENOUGH,U. W., and R. L. WEISS,1978 Interrelationships be- spindle orientation during the mitotic cell cycle.]. Cell Biol. 125: tween microtubules,a striated fiber, and the gameticmating 143-158. structure of Chlamydomonas reinhardtii. J. Cell Biol. 76: 480-483. MOESTRUP,O., 1978 On the phylogenetic validity of the flagellar GOULD,R. R., 1975 The basal bodies of Chlamydomonas reinhardtii: apparatus in green algae and otherchlorophyll aand chlorophyll formation from probasal bodies, isolation, and partial character- b containing plants. Biosystems 10: 117-144. ization. J. Cell Biol. 65: 65-74. MUtIITA, L., T. S. KARPOVA and J. A. COOPER,1994 A yeast actin- G~uI.o,R. R., and G. G. BORIS, 1977 The pericentriolar material related protein homologous to that in vertebrate dynactin corn- in Chinese hamster ovary cells nucleates microtubule formation. plex is important for spindle orientation andnuclear migration. J. Cell Biol. 73: 601-615. Cell 78: 669-679. GROSS,C. H., L. P. W. RANUM and P. A. LEFEBVRE,1988 Extensive PAI.MER,R. E., D. S. SUI.I.IVAN,T. HUFFAKERand D. KosrrrAvn. 1992 restriction fragment length polymorphisms in a new isolate of Role of astral microtubules and actin in spindle orientation and Chlamydomonas reinhardtii. Curr. Genet. 13: 503-508. migration in the budding yeast, Saccharomyces ct.rpoisia~.J. Cell HAMAGUC;HI,Y., 1975 Microinjection of colchicine into sea urchin Biol. 119: 583-593. eggs. Dev. Growth Differ. 17: 111-1 17. PASQIJALE,S. M., and U. W. GOODFNOIJGH,1987 Cyclic AMP func- tions as a primary sexual signal in gametes of Ch1amydomona.r HARPER, J. D. I., and P. C. L. JOHN, 1986 Coordination of division events in the Chlamydomonascell cycle. Protoplasma 131: 118-130. reinhardtii. J. Cell Biol. 105: 2279-2292. RAPPAPORT, R., 1961 Experiments concerning the cleavage in sand HARPER, J. D. I., D. W. McCumu, M. A. SANDERS,J. 1,. SALISBURY, dollar eggs. J. Exp. Zool. 148: 81 -89. and P. C. JOHN, 1992 Actin dynamics during the cell cycle in RAPPAPORT, R., 1971 Reversalof cleavage inhibition in enchintr Chlamydomonas reinhardtii. Cell Motil. Cytoskeleton 22: 117-126. derm eggs. J. Exp. Zool. 176: 249-255. HARRISA. K., and S. L. GEWALT,1989 Simulation testing of mecha- RAPPAPORT, R., 1973 Onthe rate of movement of the cleavage nisms for inducing theformation of the contractilering in cytoki- stimulus in sand dollar eggs. J. Exp. Zool. 183: 115-119. nesis. J. Cell Biol. 109: 2215-2224. RAPPAPORT, R., 1986 Establishment of the mechanism of cytokinesis HARRIS, E. H., 1989 The Chlamydomonas Sourcebwk, a Comprehensive in animal cells. Int. Rev. Cytol. 105: 245-281. Guide to Bioloa and Laboratq Usr. Academic Press, San Diego, CA. RINCO, D. L., 1967 Flagellar motion and fine structure of the flagel- HAR~WEI.~.,L. H., and T. A. WEINERT,1989 Checkpoints:controls lar apparatus in Chlamydomonas. J. Cell Biol. 33: 543-571. that ensure the order of cell cycle events. Science 246 629-633. SAGER,R., and S. GRANICK,1954 Nutritional control of sexuality in HILL,D. P., and S. SIKOME,1988 An analysis of the role of microfil- Chlamydomonas reinhardtii. J. Gen. Physiol. 37: 729-742. aments in the establishment and maintenance of asymmetry in SALISBURY,J. L., A. T. BARONand M. A. SANDERS,1988 The centrin- Camorhabditis elegans zygotes. Dev. Biol. 124: 75-84. based cytoskeleton of Chlamydomonas reinhurdtii: distribution in HII.I.,D. P., and S. STROME,1990 Brief cytochalasin-induced disrup interphase and mitotic cells. J. Cell Biol. 107: 635-641. tion of microfilaments during a critical interval in 1-cell C. elegans SATTERWAITE,L. L., and T. D. POIJAM),1992 Cytokinesis. Curr. embryos alters the partitioning of developmental instructions to Opin. Cell Bid. 4: 43-52. the 2cell embryo. Development 108: 159-172. SCHIBLER,M. J, and B. HUANG,1991 The col"4 and rol"15 /3-tubulin HOLMES,J. A,, and S. K. DUTCHER,1989 Cellular asymmetry in Chla- mutations in Chlamydomonas rrinhardtii confer altered sensitivities mydomonas reinhardtii. J. Cell Sci. 94: 273-285. to microtubule inhibitors and herbicides by enhancing microtw HUFFAKER,T. C., J. H. THOMASand D. BOTSTEIN,1988 Diverse ef- bule stability. J. Cell Biol. 113: 605-614. fects of beta-tubulin mutations on microtubule formation and STARLING,D., and J. RANDAH., 1971 The flagella of temporary dikar- function. J. Cell Biol. 106: 1997-2010. yons of Chlamydomonas rfinhardtii. Genet. Res. 18: 107-113. HYMAN,A. A,, 1989 Centrosomemovement in the early divisions STEAKNS,T., and M. KIRSGHNER, 1994 In vitro reconstitution of of Camorhabditis ekgans: a cortical site determining centrosome centrosome assembly and function: thecentral role of y-tubulin. position. J. Cell Biol. 109: 1185-1193. Cell 76: 623-637. HYMAN,A. A., and J. G. WHITE,1987 Determination of cell division STROME,S., 1993 Determination of cleavage planes. Cell 72: 3-6. axes in the early embryogenesis of Caenorhabditis elegans. J. Cell SUI.I.WAN,D., and T. HUFFAKER,1992 Astral microtubules are not Biol. 105: 2123-2135. required for anaphase h in Sarrharomyces crrmisiar. J. Cell Bid. JAMES,S. W., C. D. SILFLOW,P. STROOMand P. A. LEFEBVRE,1993 A 119: 379-388. mutation in the al-tubulin gene of Chlamydomonas reinhardtii SZ~~LL~SI,A,, H. RIS,D. SXOI.LOSIand A. DEBEC,1986 A centriole- confers resistance to anti-microtubule herbicides.J. Cell Sci. 106: free Drosophila cell line: a high voltage EM study. Eur. J. Cell 209-218. Biol. 40: 100-104. JOHNSON,U. G., and K. R. PORTER, 1968 Fine structure of cell divi- TAII.ON,B. E., S. A. AIDER, J.P. SUHANand J. W. JARVIK,1992 Muta- tional analysis of centrin: an EF-hand protein associated with sion in Chlamydomonas reinhardtii. J. Cell Biol. 38: 403-425. three distinct contractile fibers in the basal body apparatns of KEMPHIIES, K. J., N. WOI.F,W. B. WOOD and D. HIRSH,1986 Two Chlamydomonas. J. Cell Biol. 119: 1613-1624. loci required for cytoplasmic organization in early embryos of WADDLE,J. A,, J. A. COOPERand R. H. WATERSTON,1994 Transient Cmorhabditis ekgans. Dev. Bid. 113: 449-460. localized accumulation of actin in Cmorhabditis ekguns blasto- KEMPHUES. K. J., J. R. PRIESS,D. G. MORTONand N. S. CHENG,1988 meres with oriented asymmetric divisions. Development 120: Identification of genes required for cytoplasmic localization in 231 7-2328. early C. elegans embryos. Cell 52: 311-320. WEISS,R. I,,, 1984 Ultrastructure of the flagellar roots in Cldamydo- LEDMET, M., and G. PIPERNO,1986 Cytoplasmic microtubules con- monas gametes. J. Cell Sci. 67: 133- 143. taining acetylated alpha-tubulin in Chlamydomonas reinhardtii; spa- WITMAN,G. B., K. CARI.SON,J. BERI.INERand J. L. ROSENBAIJM,1972 tial arrangement and properties. J. Cell Biol. 103: 13-22. Chlamydomonas flagella. I. Isolation and electrophoretic analysis LI, Y. Y., E. YEH,T. HAYS and K. BLOOM,1993 Disruption of mitotic of microtubules, matrix, membranes, and mastigonemes. J. Cell spindle orientation in a yeast dynein mutant. Proc. Natl. Acad. Biol. 54: 507-539. Sci. USA 90 10096-11000. WRIGHT,R. I,., J. L. SALISBURYand J. W. JARVIK,1985 A nucleus-basd LUCK,D., G. PIPERNO,Z. RAMANIS and B. HUANC, 1977 Flagellar body connector in Chlamydomonas relnhardtii that may function in mutants of Chlamydomonas: studies of radial spoke-defective basal body localizationor seg7-egati0n.J.Cell Biol. 101: 1903- 1912. strains by dikaryon and revertant analysis. Proc. Natl. Acad. Sci. USA 74: 3456-3460. Communicating editor: E. W. JONES