Extragenic Bypass Suppressors of Mutations in the Essential Gene BLD2 Promote Assembly of Basal Bodies With Abnormal Microtubules in Chlamydomonas reinhardtii

Corresponding author: Susan K. Dutcher, Department of Genetics, Box 8232, 660 S. Euclid Ave., St. Louis, MO 63110., dutcher{at}genetics.wustl.edu (E-mail)

Communicating editor: R. S. HAWLEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

bld2-1 mutant Chlamydomonas reinhardtii strains assemble basal bodies with singlet microtubules; bld2-1 cells display flagellar assembly defects as well as positioning defects of the mitotic spindle and cleavage furrow. To further understand the role of the BLD2 gene, we have isolated three new bld2 alleles and three partially dominant extragenic suppressors, rgn1-1, rgn1-2, and rgn1-3. bld2 rgn1-1 strains have phenotypes intermediate between those of bld2 and wild-type strains with respect to flagellar number, microtubule rootlet organization, cleavage furrow positioning, and basal body structural phenotypes. Instead of the triplet microtubules of wild-type cells, bld2 rgn1-1 basal bodies have mixtures of no, singlet, doublet, and triplet microtubules. The bld2-4 allele was made by insertional mutagenesis and identified in a noncomplementation screen in a diploid strain. The bld2-4 allele has a lethal phenotype based on mitotic segregation in diploid strains and in haploid strains generated by meiotic recombination. The lethal phenotype in haploid strains is suppressed by rgn1-1; these suppressed strains have similar phenotypes to other bld2 rgn1-1 double mutants. It is likely that BLD2 is an essential gene that is needed for basal body assembly and function.

THE structure of basal bodies is highly conserved in a wide range of organisms. Basal bodies consist of nine sets of blades arranged as a cylinder, and each blade contains three microtubules. The tubules are termed A, B, and C, with A being the innermost tubule. A- and B-tubules are continuous with the doublet microtubules of the flagellum, and the C-tubule terminates at the beginning of the transition zone, which lies between the basal body and flagellum. Basal bodies have several associated appendages (Fig 1). At the proximal end of the basal body, closest to the cell center, is an amorphous, electron dense disc (RINGO 1967 ). Distal to the disc, where the triplet microtubules begin, is a cartwheel structure, with a central hub and nine spokes radiating off the hub. Each spoke connects to one of the nine A-tubules. At the distal end are appendages called transitional fibers that extend outward from each set of triplet microtubules. The transition zone contains nine doublet microtubules and has several distinct structures. One striking feature is the stellate fibers that form a nine-point star. Each point of the star is associated with a corresponding A-tubule of the doublet microtubules (RINGO 1967 ).

View larger version (42K): In this window In a new window Download PPT slide   Figure 1. Scheme of a basal body, transition zone, and flagellum. The basal body contains blades arranged cylindrically. Each blade is composed of a triplet microtubule. The tubules are designated A, B, and C, with the innermost tubule being the A-tubule. The C-tubule terminates at the transition zone, and the A- and B-tubules of the basal body are continuous with the doublet microtubules of the transition zone and the flagellum. In the center of the diagram is a longitudinal view of a basal body, transition zone, and a portion of a flagellum. The proximal portion of the basal body, the structure closest to the cell center, is depicted at the bottom of the diagram. (1–4) Cross sections through a basal body. (5–7) Cross sections through a transition zone. (8) Cross section through a flagellum. (1) At the proximal end of the basal body is an amorphous, electron-dense disc. (2) The proximal portion of the basal body contains a cartwheel structure in the center of the basal body with a central hub and spokes. Each spoke is connected to an A-tubule. (3) The middle portion of the basal body lacks appendages. A-, B-, and C-tubules are labeled. (4) At the distal end of the basal body are transitional fibers that extend radially from each blade. (5) At the proximal end of the transition zone, fibers connect each of the doublet microtubules to the flagellar membrane. (6) In the middle portion of the transition zone, stellate fibers form a nine-point star. Each point of the star is connected to the A-tubule. (7) In the more distal portion of the transition zone, the star has a distinctive electron-dense circle where the stellate fibers end as well as a small electron-dense spot at the star's center. A central pair of microtubules is present at the corresponding place in a more distal section of the transition zone that still contains stellate fibers (not shown). (8) A cross section of a flagellum contains nine doublet microtubules and a central pair of microtubules.

DIPPELL 1968 performed a detailed examination of Paramecium basal body duplication. Duplication occurs at a right angle to the mother basal body. The first event in basal body genesis is the appearance of a flat, amorphous disc. Singlet tubules then appear that correspond to the A-tubule. Before all the singlet tubules have formed, the formation of doublets begins, and before all doublets have formed, the formation of triplets begins. Thus, in a given cross section, it is possible to observe singlet, doublet, and triplet microtubules simultaneously. The cartwheel structure is assembled completely by the time the triplets have formed. The entire structure is then elongated.

Centrioles have the same structure as basal bodies, but have no transition zone. Basal bodies and centrioles can interconvert. In many organisms, basal bodies take on the role of centrioles during fertilization (reviewed in SCHATTEN 1994 ). It is the sperm basal body that serves as a centriole in the zygotic microtubule organizing center. The interphase basal bodies of Chlamydomonas, a biflagellate green alga, serve as centrioles during mitosis (COSS 1974 ).

Basal bodies and centrioles are likely to have several functions in the cell. They play roles in flagellar assembly and in the organization of centrosome and cytoskeletal elements. Basal bodies function in two ways to promote flagellar assembly. The nine triplet microtubules in a basal body are hypothesized to template the nine doublet microtubules of the flagellum, so that flagella have the same ninefold rotational symmetry as basal bodies. Second, basal bodies are believed to be docking sites for the machinery that transports flagellar components to the distal tip. One of the components of a heterotrimeric kinesin, p85 (FLA10), which is required for the intraflagellar transport, is concentrated in a tripartite-shaped pattern around the basal bodies as well as being found along the length of flagella and in the cell body (WALTHER et al. 1994 ; COLE et al. 1998 ). This distribution of the molecular motor is consistent with basal bodies serving as a docking site for the flagellar transport machinery.

Recent experiments by BOBINNEC et al. 1998A demonstrated that centrioles are likely to play a key role in centrosomal organization. Antibodies specific for glutamylated tubulin primarily stain centrioles (BOBINNEC et al. 1998B ). Injection of these antibodies into HeLa cells results in the disassembly of centrioles and a corresponding disruption in the organization of the pericentriolar material (BOBINNEC et al. 1998A ), indicating that centrioles may be required for the proper organization of the centrosome. Genetic analysis in Paramecium has found mutant strains that are unable to duplicate basal bodies (RUIZ et al. 1987 ). These cells have reduced cell size, altered shape, and secondary defects in the cytoskeleton. Genetic analysis in Chlamydomonas reinhardtii suggests a role of basal bodies in the organization of a set of specialized, stable microtubule structures known as microtubule rootlets. An interphase cell contains four microtubule rootlets arranged in a cruciate array originating at the basal bodies and radiating outward, just under the plasma membrane (RINGO 1967 ). Two of these rootlets are composed of two microtubules, and the other two are composed of four microtubules. The rootlets alternate to form a 4-2-4-2 pattern (GOODENOUGH and WEISS 1978 ; MOESTRUP 1978 ). These microtubules are important for the orientation of the cleavage furrow during mitosis (EHLER et al. 1995 ). In addition, the basal bodies may play a role in the organization of centrin fibers for nuclear placement (KIRK 1998 ). In wild-type cells, centrin staining of the nucleus-basal body connectors, basal bodies, flagella, and distal striated fiber results in a "diamond ring" pattern (WRIGHT et al. 1985 ; SALISBURY et al. 1988 ). In bld2-1 cells, a similar pattern, with fibers extending from the nucleus, is present in only 1/500 cells, and the remainder of cells lack centrin fibers that extend from the basal bodies to the nucleus (EHLER et al. 1995 ).

Mutations in several Chlamydomonas genes affect basal body assembly. uni3-1, a deletion of the gene encoding -tubulin in C. reinhardtii, affects basal body morphology (DUTCHER and TRABUCO 1998 ). The basal bodies of uni3-1 cells contain doublet instead of triplet microtubules; the C-tubule is missing. This basal body defect results in a flagellar phenotype. One-fourth of uni3-1 cells are biflagellate, one-fourth are uniflagellate, and one-half are aflagellate. uni3-1 also has a cleavage furrow-positioning defect. A population of uni3-1 cells has an increased number of cells with multiple nuclei and sister cells of unequal size (S. FROMHERZ, T. H. GIDDINGS, N. GOMEZ-OSPINA, N. LOW-NAM and S. K. DUTCHER, personal communication).

bld2-1 cells also have a basal body assembly defect. bld2-1 cells are aflagellate and lack fully assembled basal bodies/centrioles (GOODENOUGH and ST. CLAIR 1975 ; EHLER et al. 1995 ). The basal bodies that GOODENOUGH and ST. CLAIR 1975 observed had singlet instead of triplet microtubules. This observation led these authors to propose that the BLD2 gene product is required for the formation of doublet and triplet microtubules. EHLER et al. 1995 found that, in addition to the basal body defect, bld2-1 strains have coordination defects during cell division. bld2-1 strains show defects in timing; nearly half of bld2-1 cells initiate cytokinesis prematurely with respect to karyokinesis. In addition, the microtubule rootlets are often abnormal in number and placement within cells. The position of the cleavage furrow is still correlated with the position of one of the microtubule rootlets, but the cleavage furrow is not positioned properly with respect to the spindle or other cellular organelles. This indicates that the microtubule rootlets play a central role in designating the position of the cleavage furrow.

In this study, a screen for extragenic suppressors of the bld2 alleles was performed to find other genes involved in the assembly of doublet and triplet microtubules. Screens using the bld2-1 allele were biased toward intragenic suppressors (EHLER et al. 1995 ); therefore partial intragenic revertants of bld2-1 were isolated to use in suppressor screens. The basal body-assembly defect of bld2 alleles is partially suppressed by an extragenic suppressor; bld2 rgn1-1 basal bodies have mixtures of singlet, doublet, and triplet microtubules. A fourth allele of BLD2, bld2-4, was isolated in a noncomplementation screen. bld2-4 allele results in lethality when it is homozygous or hemizygous, and the lethality can be suppressed by the rgn1-1 allele. It is likely that the BLD2 gene product is essential for viability.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Chlamydomonas strains and culture conditions:
Chlamydomonas reinhardtii strains used in this study are listed in Table 1. 137c mt^- and 137c mt⁺ were used as wild-type strains, unless otherwise stated. Cultures were grown at either 21° or 25° under constant illumination. Generally, cells were grown in medium I of SAGER and GRANICK 1953 supplemented with 8 mM sodium acetate (rich medium). 5-Fluoroindole was used at a concentration of 7 µM (PALOMBELLA and DUTCHER 1998 ). Potassium chlorate was used at a concentration of 5 mM, with 2 mM urea as the sole nitrogen source (SCHNELL and LEFEBVRE 1993 ).

Genetic analysis:
Mating of flagellate strains and determination of the meiotic segregation of genetic markers was performed as described by HARRIS 1989 . Aflagellate strains were mated using the method of PASQUALE and GOODENOUGH 1987 and DUTCHER 1995B . Dominance was determined by constructing diploid strains using the nit2 and ac17-1 markers (PALOMBELLA and DUTCHER 1998 ). Disomy on a separate linkage group, linkage group VI, was verified by Southern blot analysis. A 2.7-kb KpnI-ApaI fragment from pThi4.9 (FERRIS 1995 ) was used as a probe to distinguish mt⁺ and mt^- loci (data not shown).

The ability of bld2 strains to germinate was determined in a colony-formation assay. Strains were mated using the method of DUTCHER 1995B . Mixtures of mated cells were plated to rich medium and were placed under constant illumination overnight. The plates were wrapped in aluminum foil for 5 days. A brief exposure to chloroform vapors kills vegetative cells but not zygotes. Plates were inverted over chloroform for 90 sec and were then rewrapped in foil for 1 more day. Plates were unwrapped and placed under constant illumination at 25°. Colonies were observed after 5–6 days.

Each of the 10 independent suppressors and revertants of bld2-2 (see below) was sequentially backcrossed to a wild-type strain (137c mt⁺) four times before the phenotypes described were assessed.

Analysis of progeny from triploid zygotes:
A set of seven successive crosses was performed with progeny produced by triploid zygotes obtained from crosses with the 4G strain and strain 3159 (NIT2 AC17 MAA2 MAA7; rgn1-1; nit1) following the addition of dibutyryl cAMP and 3-isobutyl-1-methylxanthine (IBMX) to promote mating. The viable progeny produced from triploid zygotes are likely to be aneuploid and to produce low levels of viable progeny in two to three subsequent crosses (DUTCHER and GIBBONS 1988 ). We crossed three independent Nit1⁺ progeny in three successive crosses to strain 3159 in the presence of dibutyryl cAMP and IBMX and selected Nit1⁺ progeny. All crosses were performed with a BLD2 allele because of the recessive meiotic phenotype of bld2 alleles (EHLER et al. 1995 ). We observed Nit1⁺ progeny that were flagellate in the third round of crosses. It is likely that this strain was disomic for linkage group III; three additional successive crosses were made to strain 3160 (nit2 AC17 MAA2 MAA7; rgn1-1; nit1) and flagellate Nit1⁺ Maa^- progeny were selected after each cross. To obtain euploid progeny a recombination event between NIT1 and NIT2 is needed. Four strains (3155–3158) were used to examine the phenotype of the 4GI allele in haploid strains.

Mutagenesis and isolation of pseudorevertants and suppressors:
Insertional mutations (TAM and LEFEBVRE 1993 ) were generated using the glass bead transformation method of KINDLE 1990 . Digestion with EcoRI of pMN56, which contains the NIT1 gene (NELSON et al. 1994 ), linearized the plasmid. Logarithmically growing cells were harvested by centrifugation and resuspended in autolysin for 30 min. Autolysin was prepared as described in DUTCHER 1995A . The cells were reharvested by centrifugation and resuspended in medium containing 4 mM sodium nitrate as the sole nitrogen source at a final density of 1 x 10⁸ cells per ml. A total of 300 µl cells, 300 µl glass beads, 100 µl 15% PEG 4000, and 1–2 µg DNA were vortexed at maximum speed for 30 sec. Cells were plated onto medium containing 4 mM sodium nitrate as the sole source of nitrogen. Top agar (0.5%) with 4 mM sodium nitrate was used in the plating to increase transformation efficiencies. Transformants were selected at 25°, multiple tranformants were tested subsequently for their ability to grow at 32°, and none could grow at this temperature.

For the chemical mutagenesis, 10⁸ cells were resuspended in sterile deionized water containing the mutagen. After mutagenesis, cells were immediately diluted into 20 ml of rich medium, and five to seven rounds of enrichment for pelleting cells were performed. Ethyl methanesulfonate (EMS; Sigma Chemical Co., St. Louis) was added to a final concentration of 10–30 µl/ml. Cells were incubated in EMS for 30–40 min and then were washed three times in 1% sodium thiosulfate to deactivate the EMS (LUX and DUTCHER 1991 ). Approximately 70% of the cells survived the EMS treatment. Diepoxybutane (Sigma) was added to a final concentration of 30 µl/ml. Cells were incubated in diepoxybutane for 60 min and were then washed three times in sterile, deionized water. Approximately 60% of the cells survived the diepoxybutane treatment. Methyl methanesulfonate (MMS; Sigma) was added to a final concentration of 200 µl/ml. Cells were incubated in MMS for 60 min and then were washed three times in sterile, deionized water. Approximately 10% of the cells survived the MMS treatment. Heat shock was as described in DUTCHER and TRABUCO 1998 . Approximately 90% of the cells survived the heat-shock treatment. Ultraviolet irradiation was as described previously (LUX and DUTCHER 1991 ). Approximately 80% of cells survived the ultraviolet irradiation.

The previous set of bld2-1 revertants was isolated in enrichment screens for swimming cells (EHLER et al. 1995 ). This type of screen is biased toward reversion events that result in a high percentage of flagellate cells. Using this criterion, no revertants or suppressors were identified without treatment with a mutagen. To identify partial revertants, a selection strategy with no mutagen was employed. Chlamydomonas initiates mating through flagellar contact between cells of the two mating types. This requirement for flagella can be bypassed by treatments that increase intracellular cAMP levels (PASQUALE and GOODENOUGH 1987 ). However, even under these conditions, cells with flagella have a mating advantage over aflagellate cells in the population. A selection for cells that mated relied on the observation that unmated cells are killed by exposure to chloroform while zygotes survive.

Light microscopy:
Both flagellar number counts and sister cell analysis were performed on cultures at low densities (<1 x 10⁶ cells/ml) to ensure that cells were in log phase growth. For flagellar number counts, cells were observed with phase optics using a Zeiss microscope with a x40 Neofluar objective. Primarily aflagellate strains [bld2-1, bld2-2, bld2-3, and bld2-4/bld2-2 (4G)] were observed live. All twitching cells clearly had a single flagellum, and spinning cells were designated as having a single flagellum. All other strains were fixed in 0.1% glutaraldehyde in rich medium before assessing numbers of flagella.

Sister cells were designated as two cells closely appressed and still contained within a mother cell wall. Phase images of sister cells were acquired with a Zeiss Axiophot microscope equipped with a x40 plan-Neofluar objective, a MicroImage i308 camera system (MicroImage Video Systems, Bechtelsville, PA) and Apple Video Player software. The area of the cells in pixels was determined by tracing and filling in the cells using Adobe PhotoShop (Adobe Systems, Mountain View, CA). Sixty-seven pixels were equivalent to 10 µm. The differences between the sizes of sister cells are presented as the area of the large cell divided by the area of the small cell. Fifty pairs of sister cells were analyzed for each strain. A permutation test in which two strains were compared for each test was used to determine if distributions were significantly different from one another. This analysis has the caveat that, for certain strains, images that were obtained later on any given day tended to have greater differences between sister cells.

Immunofluorescence:
The monoclonal antibody 6-11B-1 (Sigma) is specific for acetylated -tubulin (PIPERNO and FULLER 1985 ) and was used at a dilution of 1:80. A fluorescein-linked sheep anti-mouse antibody (Amersham Life Science, Arlington Heights, IL) was used as a secondary antibody at a dilution of 1:50. Labeling was performed as described in HOLMES and DUTCHER 1989 , except that cells were incubated in autolysin for 1 hr after fixation and were spotted onto polyethylenimine-coated coverslips (0.01%; Sigma), instead of polylysine-coated coverslips. A Leica DMRXA/RF4/V automated universal microscope equipped with a Cooke SensiCam high-performance digital camera and x100 PL APO and x63 PL APO objectives was used to visualize immunofluorescence. The Slidebook software package (Intelligent Imaging Innovations, Denver) was used to acquire, deconvolve, and project images. Because the flagella and microtubule rootlets often spanned several micrometers, projections were used to view multiple focal planes simultaneously.

Electron microscopy:
Cells were prepared for thin sectioning and electron microscopy essentially as described by WINEY et al. 1995 . Briefly, cultures were grown at 25° under constant illumination in rich medium. Cells were collected in log phase growth by centrifugation and were resuspended in medium diluted 1:5 that lacked acetate and a nitrogen source (M-N/5). Cells were reharvested by centrifugation and were resuspended in M-N/5 with 100 mM mannitol. Cells were harvested a final time by centrifugation. The pellet was transferred to sample holders and frozen in a Balzers HPM10 high-pressure freezer. Samples were freeze substituted in 0.1% tannic acid in acetone for 4 days at -80°. Samples were then fixed in 2% osmium tetroxide in acetone for 1 day at -20° and then overnight at 4°. Samples were rinsed in acetone and brought to room temperature. Spurrs resin (Polyscience, Warrington, PA) was used to embed the samples. Serial sections with a thickness of 50–60 nm were cut using a Reichert Ultracut microtome and were collected on Formvar-coated slot grids (1 x 2 mm). The sections were stained with 2% uranyl acetate in 70% methanol for 6 min and then in aqueous lead citrate for 4 min.

Sections were viewed in a Philips CM10 electron microscope (Philips Electronic Instruments, Mahwah, NJ), operating at either 80 or 100 kV. A Philips rotating specimen holder was used to view the samples. The long axis of the basal body was determined, and the specimen was tilted so that the triplet microtubules could be viewed in cross section. Micrographs were taken at a magnification of x39,000.

DNA isolation and hybridization conditions:
Genomic DNA was isolated as described by DUTCHER and TRABUCO 1998 . Hybridization conditions were as described by JOHNSON and DUTCHER 1991 . Probes were labeled using [³²P]dATP with the Multiprime DNA labeling system (Amersham Life Science). Plasmid DNA was isolated with either Promega (Madison, WI) Wizard or QIAGEN (Valencia, CA) plasmid kits.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of partial intragenic revertants of bld2-1:
In a previous screen for suppressors of bld2-1, only intragenic revertants were isolated (EHLER et al. 1995 ). Forty intragenic revertants of bld2-1 were generated using a variety of mutagens, and these revertants showed complete or nearly complete reversion of the bld2-1 defects. Although six of these revertants had motility defects, all of the revertants displayed wild-type numbers of flagella and wild-type cleavage furrow positioning. Partial intragenic revertants of bld2-1 were isolated in a mating selection (see MATERIALS AND METHODS) with the aim of identifying new bld2 alleles that could be used to isolate extragenic suppressors.

Two weak revertants, 2-2 and 19d, were isolated. Based on data presented below, 2-2 and 19d are alleles at the BLD2 locus and are designated as bld2-2 and bld2-3, respectively. One percent or less of bld2-2 and bld2-3 cells were uniflagellate, while the rest of the cells in these populations were aflagellate (Table 2). In both strains, the single flagellum was assembled on the basal body opposite the eyespot (n = 150 for each strain), which is known as the transbasal body (HUANG et al. 1983 ). Crosses were performed with bld2-2 or bld2-3 by a wild-type strain to determine if the low percentage of flagellate cells would segregate independently from the original bld2-1 mutation. These crosses did not yield any progeny that were completely aflagellate (Table 3); thus, the new mutations are closely linked to bld2-1. Stable diploid strains were constructed to determine whether the new mutations were dominant or recessive to the wild-type allele. Both bld2-2/BLD2 and bld2-3/BLD2 strains had wild-type numbers of swimming cells (Table 2), so the phenotypes of bld2-2 and bld2-3 are recessive.

Complementation tests of flagellar and meiotic phenotypes:
All three bld2 alleles are in the same complementation group with respect to the flagellar and meiotic defects of bld2-1. Diploid strains that are homozygous for bld2-1 fail to assemble flagella (EHLER et al. 1995 ), while 1–3% of the diploid strains homozygous for bld2-2 or bld2-3 had a single flagellum and resembled the haploid parental strains (Table 2). Diploid cells heteroallelic with the bld2-1 allele failed to assemble flagella while 2.8% of the heteroallelic bld2-2/bld2-3 cells assembled a single transflagellum.

The three bld2 alleles are in the same complementation group with respect to the meiotic defect. Zygotes that are homozygous for bld2-1 fail to complete meiosis (EHLER et al. 1995 ). The ability of homozygous bld2 zygotes to germinate was tested by a colony-formation assay (see MATERIALS AND METHODS). Exposure to chloroform kills nonzygotic cells and zygotes that germinate form colonies on plates. bld2-1/bld2-2 and bld2-2/bld2-3 zygotes did not form colonies (data not shown). The recombination and complementation data are consistent with these mutations being new alleles of BLD2. bld2-3 and bld2-2 were mapped with respect to genetic markers on linkage group III. These bld2 alleles lie between 2.2 and 4 map units from nit2-1 (Table 3).

The phenotypes of bld2-2 and bld2-3 strains are similar to bld2-1:
To determine whether the weak reversion of the flagellar phenotype is accompanied by a reversion of the other bld2-1 defects, cleavage furrow positioning, colony size, and microtubule rootlets were observed in bld2-2 and bld2-3 strains. Measuring the difference between sizes of recently divided sister cells monitored the cleavage furrow-placement defect of bld2-1 cells. The average difference in area between bld2-1 sister cells is greater than the average difference in area between wild-type sister cells (EHLER et al. 1995 ). To assess defects in cleavage furrow placement, the areas of recently divided sister cells in bld2-1, bld2-2, bld2-3, and wild-type strains were compared (see MATERIALS AND METHODS). The differences in areas between the two sister cells (large cell area - small cell area) were calculated. The distributions of bld2-1 and wild-type cells were similar to those described previously (EHLER et al. 1995 ). However, cell sizes varied from one set of measurements to the next, and the average cell size from one population was not necessarily the same as the average from another. To eliminate the factor of cell size, for all analysis in this study the data are presented as the large cell area divided by small cell area. The distributions of the cell measurements for bld2-2 and bld2-3 are indistinguishable from those of bld2-1, which indicates that there is not a significant suppression of the cleavage furrow-positioning defect in these strains (Fig 2). Cleavage furrow placement was also examined by observing the number of nuclei present in logarithmically growing cells. The majority of wild-type cells in such cultures have a single nucleus. The bld2 cultures have an increased frequency of cells with no nuclear DNA (3.5–8%) or with two or more nuclei (5.5–14%; Table 4). The mispositioning of the cleavage furrow results in an increased number of cells lacking nuclei or having too many nuclei.

View larger version (33K): In this window In a new window Download PPT slide   Figure 2. Differences in the areas of sister cells in bld2 strains. Histograms display the number of cells with a given large cell/small cell area value. Fifty pairs of sister cells were measured for each strain. A permutation test in which two strains were compared for each test was used to determine if distributions were significantly different from one another (PHILIP BEINEKE, personal communication). (A) Wild type. (B) bld2-1. (C) bld2-2. (D) bld2-3. (E) bld2-2/4GI (4G). (F) rgn1-1. (G) bld2-1 rgn1-1. (H) bld2-2 rgn1-1. (I) bld2-3 rgn1-1. (J) 4GI rgn1-1. The P values for any bld2 strain or 4G compared to any other bld2 strains were not significantly different from each other. The P value for any bld2 strain compared to wild type was 0.003. The rgn1-1 and bld2-2 rgn1-1 distributions were not significantly different from one another (P = 0.6). Both rgn1-1 and bld2-1 rgn1-1 distributions were significantly different than bld2-2 (P = 0.005 and P = 0.05) and wild type (P = 0.002 and P = 0.003). The bld2-1 rgn1-1 and bld2-3 rgn1-1 strains were not significantly different from bld2-1 and bld2-3 strains (P = 0.2 and P = 0.3).

The colony size and microtubule rootlets of bld2-2 and bld2-3 cells were also similar to those of bld2-1 cells. Colony size on solid medium was used as an indicator of growth rate. All bld2 strains produced smaller colonies than the wild-type strain (Fig 3). To observe microtubule rootlets, bld2 strains were stained with an antibody that is specific for acetylated -tubulin (PIPERNO and FULLER 1985 ). This antibody recognizes flagella, basal bodies, and microtubule rootlets in wild-type cells (LEDIZET and PIPERNO 1986 ). As in bld2-1 cells, the acetylated -tubulin staining in bld2-2 and bld2-3 cells was abnormal. Some cells in these populations had greater than four linear structures that stained with acetylated -tubulin antibodies in wild-type cells. It is likely that these structures are microtubule rootlets, and they are referred to as microtubule rootlets in the descriptions that follow. For all bld2 strains, the microtubule rootlets were often disorganized and present in abnormal numbers (Fig 4). There were cells in all bld2 populations that clearly had too many microtubule rootlets. Some bld2 cells had normal numbers of rootlets, but these rootlets were disorganized. Still other cells had little or no staining. This lack of staining could be due either to a lack of microtubule rootlets or to poor permeabilization of the cells. Some aflagellate cells had bright dots, which may have been basal bodies. Other cells had no dots and probably lacked basal bodies. The microtubule rootlets in the rare flagellate cells in a bld2-3 population remained abnormal (Fig 4). Cells that have a basal body that is competent to grow flagella do not necessarily assemble wild-type microtubule rootlets.

View larger version (49K): In this window In a new window Download PPT slide   Figure 3. Colony size of bld2 strains. Cells were grown for 6 days at 25° in constant light. (A) bld2-1, bld2-2, and bld2-3 strains all produce colonies smaller than those produced by wild-type strains. (B) The bld2-3 rgn1-1 strain had colonies intermediate in size to those of bld2-3 and wild-type strains. rgn1-1 colonies are similar in size to wild-type colonies. (C) 4G produced colonies smaller than those produced by the heterozygous diploid strain (bld2-2/BLD2) and bld2-2 strains produced colonies smaller than those produced by a haploid wild-type strain. The bld2/BLD2 diploid strain (strain 1925 from Table 1) was the parent strain used in the noncomplementation screen described below.

View larger version (69K): In this window In a new window Download PPT slide   Figure 4. Microtubule rootlets in bld2 strains. The microtubule rootlets of bld2 strains were disorganized, and the number of microtubule rootlets in these strains was often abnormal. (A–F) Immunofluorescent images of cells stained with an antibody that is specific for acetylated -tubulin. This antibody labels flagella, basal bodies, and microtubule rootlets (LEDIZET and PIPERNO 1986 ). All are projections of multiple deconvoluted images. (A–D and F) Images obtained with x100 objective. Bar in A, 10 µm. (E) Image obtained at x63 objective. Bar, 10 µm. (A) An aflagellate wild-type cell. The four microtubule rootlets form a cruciate pattern. The bright dots at the center are basal bodies (arrowheads). The projection consists of 28 images. (B) bld2-1 cells with disorganized microtubule rootlets. The staining is slightly brighter where microtubule rootlets cross, but there are no distinct dots that would be basal bodies. The projection consists of 10 images. (C) A bld2-2 cell with abnormal microtubule rootlets. The projection consists of 5 images. (D) A bld2-3 cell with many more microtubule rootlets than wild-type cells. The bright dots may be basal bodies. The projection consists of 8 images. (E) A bld2-3 cell with a single flagellum (arrow). The cell has numerous, disorganized microtubule rootlets. The projection consists of 18 images. (F) A 4G cell with disorganized microtubule rootlets that resemble those of the bld2 cells. The projection consists of 13 images.

rgn1-1 is a suppressor of bld2-1, bld2-2, and bld2-3:
To isolate extragenic suppressors of bld2-2, a screen for mutations that suppress the aflagellate phenotype of bld2-2 was performed. A bld2-2 NIT2 ac17-1 haploid strain was mutagenized with MMS or with ultraviolet irradiation, and swimming cells were enriched for by transferring the upper 10% of liquid cultures from four to eight times. Revertants of the bld2-2 allele were rare. No revertants were recovered from ultraviolet irradiation of 10⁹ cells, which suggests that revertants of this allele are more infrequent than for the bld2-1 allele (EHLER et al. 1995 ). Ten independent strains with swimming cells were isolated in the screen with MMS. Five of these strains are likely to contain intragenic mutations. In crosses to wild-type strains, no Bld2^- progeny were recovered in >100 tetrads for each of these 5 strains. These strains showed complete reversion of the flagellar phenotype and will not be described further. Two strains were recovered that appeared to have a diploid complement of DNA based on their meiotic behavior in crosses. These two strains are likely to contain an extragenic suppressor, because Bld2^- progeny were recovered, but the suppressor mutations have not been characterized. These strains are not described further. Three of the suppressed strains contained an extragenic suppressor and appeared to be haploid strains based on their meiotic behavior. These extragenic suppressors are all in the same recombination group. No recombinants were observed in pairwise crosses (Table 3). The locus has been designated RGN1 (for Rogaine), and the three alleles are rgn1-1, rgn1-2, and rgn1-3. Each rgn1-1 allele behaves as a single gene mutation in crosses homozygous for the bld2-2 allele. Of these extragenic suppressors, only rgn1-1 was characterized in greater detail, but rgn1-2 and rgn1-3 showed similar flagellar phenotypes.

rgn1-1 partially suppresses the flagellar defect of bld2-2 strains (Table 2). rgn1-1 bld2 strains have increased numbers of cells with one or two flagella (36–48%). To determine whether this suppressor was allele specific, bld2-1 rgn1-1 and bld2-3 rgn1-1 strains were examined. The flagellar defect is also partially suppressed in these strains (Table 2). bld2 rgn1-1 strains had an additional abnormal class of flagellate cells; 1% of bld2 rgn1-1 cells had three or four flagella. rgn1-1 is phenotypically wild type with respect to flagellar number, colony size, and microtubule rootlet morphology in interphase cells (Table 2).

To determine if other Bld2^- phenotypes were suppressed by rgn1-1, cleavage furrow positioning, colony size, and microtubule rootlets were examined. rgn1-1 BLD2 cells display an intermediate cleavage furrow placement defect; the difference in sister cell size for this strain is significantly different from both bld2-2 and wild-type cells (Fig 2). bld2-2 rgn1-1 strains have an intermediate cleavage furrow defect that is similar to that of rgn1-1 cells (Fig 2). There was no detectable suppression of the cleavage furrow-placement phenotype in bld2-1 rgn1-1 and bld2-3 rgn1-1 cells. The distributions for these strains were not significantly different than those of the bld2 strains based on a permutation test. All bld2 rgn1-1 strains had increased numbers of cells with abnormal numbers of nuclei compared to wild-type strains (Table 4). Colony sizes of bld2-1 rgn1-1, bld2-2 rgn1-1, and bld2-3 rgn1-1 strains were intermediate between those of the bld2 and wild-type/rgn1-1 strains (Fig 3).

rgn1-1 was partially dominant for the suppression of both the flagellar and the meiotic defects of bld2 strains to the wild-type RGN1 allele (Table 2). In diploid strains homozygous or heteroallelic for bld2 alleles and heterozygous for the rgn1-1 mutation, suppression of the flagellar assembly defect was observed. In three of the four diploid strains, there was a reduction in the number of cells with a single flagellum compared to the homozygous diploid strains. As noted earlier, bld2-1/bld2-2 and bld2-3/bld2-2 zygotes failed to germinate. Under the same conditions, bld2-1/bld2-2 rgn1-1/RGN1 and bld2-3/bld2-2 rgn1-1/RGN1 zygotes formed colonies. rgn1-1 suppressed the meiotic defect although the degree of suppression was not quantitated.

The microtubule rootlet phenotypes differed among the three different suppressor alleles. bld2-1 rgn1-1 and bld2-2 rgn1-1 strains had abnormal microtubule rootlets in the majority of aflagellate cells (Fig 5). There were abnormal numbers of microtubule rootlets, and they were often disorganized. On the other hand, the majority of uniflagellate and biflagellate bld2-1 rgn1-1 and bld2-2 rgn1-1 cells had four properly placed microtubule rootlets. Only a few flagellate cells in these populations had an abnormal number of microtubule rootlets, but they remained properly organized. The rootlets originated at the base of the flagellum(a) in the region of the basal bodies (Fig 5). In contrast, all bld2-3 rgn1-1 cells had abnormal microtubule rootlets. The majority of aflagellate bld2-3 rgn1-1 cells had disorganized microtubule rootlets, and flagellate bld2-3 rgn1-1 cells had too many microtubule rootlets (Fig 5). The extra microtubule rootlets in flagellate bld2-3 rgn1-1 cells originated in the basal body region of the cell. The rgn1-1 BLD2 cells had normal microtubule rootlets (Fig 5). Thus, the ability to assemble flagella in bld2 rgn1-1 strains was correlated with an organization of microtubule rootlets that was more like wild-type cells.

View larger version (72K): In this window In a new window Download PPT slide   Figure 5. Microtubule rootlets in bld2 rgn1-1 cells. (A–I) Immunofluorescence images of cells stained with an antibody that is specific for acetylated -tubulin. All are projections of multiple deconvoluted images. Images were obtained with a x100 objective. Bar, 10 µm. (A) Top view of a biflagellate wild-type cell with four microtubule rootlets. The projection consists of 23 images. (B) Side view of a biflagellate rgn-1 cell with four microtubule rootlets. Projection consists of 33 images. (C) An aflagellate bld2-1 rgn1-1 cell with abnormal microtubule rootlets. The bright spots may be basal bodies. The projection consists of 24 images. (D) A biflagellate bld2-1 rgn1-1 cell with four microtubule rootlets. The projection consists of 17 images. (E) Top of a biflagellate bld2-1 rgn1-1 cell with five microtubule rootlets. The projection consists of 25 images. (F) Side view of a uniflagellate bld2-2 rgn1-1 with six microtubule rootlets. The projection consists of 38 images. (G) An aflagellate bld2-3 rgn1-1 cell with disorganized rootlet microtubules. The projection consists of 30 images. (H) Top view of a uniflagellate bld2-3 rgn1-1 cell with at least eight microtubule rootlets. The projection consists of 23 images. (I) A quadriflagellate bld2-1 rgn1-1 cell. Two MTOCs are present in the cell, each organizing its own set of flagella and microtubule rootlets. The projection consists of 43 images.

The microtubule rootlets of bld2 rgn1-1 cells with three or four flagella were distinctly organized. These cells appeared to have two microtubule organizing centers; flagella originated from two locations within these cells, and each set of flagella had its own set of microtubule rootlets (Fig 5). These cells are likely to have arisen after a failure to properly segregate basal bodies in the previous cell division.

bld2 rgn1-1 basal bodies have singlet, doublet, and triplet microtubules:
To determine the extent of suppression of basal body assembly defects, bld2 rgn1-1 strains were examined by electron microscopy. Various known structures within basal bodies and the transition zone were used for orientation in serial sections of basal bodies; these include the stellate fibers in the transition zone (star), transitional fibers, and cartwheel structure (Fig 1). To ensure that mature basal bodies and not developing or immature basal bodies were examined, all basal bodies described below were continuous with transition zones. These basal bodies are likely to represent the class of basal bodies that template flagella. In most cases, this was confirmed by the observation of the flagellar axoneme in additional serial sections.

The microtubule blades of bld2-1 rgn1-1, bld2-2 rgn1-1, and bld2-3 rgn1-1 basal bodies consisted of singlet, doublet, and triplet microtubules. The following observations are based on the examination of serial sections from 9 bld2-1 rgn1-1, 11 bld2-2 rgn1-1, and 11 bld2-3 rgn1-1 basal bodies. The number of singlet, doublet, and triplet blades varied from one basal body to another, but most basal bodies had mixtures of these classes (Fig 6). Occasionally basal bodies with only seven or eight blades were observed (Fig 7). The spacing of the blades in bld2 rgn1-1 basal bodies was roughly wild type. The diameters of the inner circumference of the basal body cylinders for wild-type and bld2 rgn1-1 strains ranged from 130 to 160 nm, with the average diameter being 140 nm. Larger gaps were present between blades with singlet or doublet microtubules than blades with triplet microtubules. The appendages of bld2 rgn1-1 basal bodies appeared normal. bld2 rgn1-1 basal bodies had wild-type cartwheel structures and transitional fibers.

View larger version (166K): In this window In a new window Download PPT slide   Figure 6. bld2 rgn1-1 basal bodies have mixtures of singlet, doublet, and triplet microtubules. Bar, 0.1 µm. (A) A wild-type basal body with triplet microtubules. Transitional fibers are present on each of the triplet microtubules. (B) A rgn1-1 basal body with triplet microtubules. (C) A bld2-1 rgn1-1 basal body with singlet, doublet, and triplet microtubules. Transitional fibers are present on some doublet and triplet microtubules. (D) A bld2-1 rgn1-1 basal body with doublet and triplet microtubules. Transitional fibers are present on some doublet microtubules.

View larger version (89K): In this window In a new window Download PPT slide   Figure 7. Three sections of bld2-3 rgn1-1 basal body with seven microtubule blades at the proximal end. All sections from the basal body are not shown, and sections are not continuous. Bar, 0.1 µm. (A) The proximal end of the basal body contains the cartwheel structure with hub and spokes. Only seven microtubule blades are present. A gap occurs where missing blades should be (arrowheads). The spokes of the cartwheel are present in positions that correspond to the missing blades. (B) A section in the middle of the basal body. The distal striated fiber is visible in the upper left quadrant of the micrograph. A gap is present where the right blade was missing in A. There may be a singlet microtubule where the left blade was absent in A. (C) The distal portion of the basal body. Transitional fibers are present on some blades. All nine blades of the basal body are present.

Occasionally, the number of tubules in a blade or the number of blades was different along the longitudinal axis of a basal body. For example, a blade with a singlet microtubule in proximal sections may have a doublet microtubule in distal sections or in the transition zone (Fig 8). There were also specimens in which different numbers of blades were present in different parts of the same basal body. For example, a basal body with seven blades at its proximal end may have nine blades at its distal end (Fig 7). In every case we observed, when there were differences from one part of the basal body to the next, there were higher numbers of tubules or blades closer to the transition zone. Thirteen of the 17 bld2 rgn1-1 transition zones also appeared normal; they had recognizable nine-point stars. Four basal bodies were missing blades at the transition zone. The stellate fibers of these transition zones were incomplete or missing where blades were absent (data not shown). Axonemes templated by three of these basal bodies had reduced numbers of doublet microtubules (PREBLE et al. 2000 ). BLD2 rgn1-1 basal bodies were phenotypically wild type.

View larger version (176K): In this window In a new window Download PPT slide   Figure 8. A bld2-2 rgn1-1 basal body with singlet microtubules. Bar, 0.1 µm. Sections from a basal body and transition zone are shown. Not all sections are presented, and sections are not continuous. (A) The middle portion of the basal body. The distal striated fiber is present in the upper portion of the micrograph. Singlet (arrowheads), doublet, and triplet microtubules are present. (B) The distal end of the basal body. Transitional fibers are present on some blades. Single microtubules are still present. (C) Stellate fibers of the transition zone are partially visible. Singlet microtubules are still present. The singlet microtubule on the left is attached to a C-shaped structure that may be a partial microtubule. (D) The stellate fibers of the transition zone are clearly visible. Each point of the star has a doublet microtubule, including positions where singlet microtubules are present in proximal sections.

bld2 alleles are rare among aflagellate strains:
It is often useful to know the phenotype associated with null alleles; therefore additional new alleles at the BLD2 locus were sought in screens for aflagellate strains. A wild-type strain was mutagenized with a variety of mutagens that included EMS, ultraviolet irradiation, diepoxybutane, and a short exposure to 42° (heat shock). A total of 250 independent aflagellate strains were isolated (Table 5). Each of the strains lacked flagella on the basis of light microscopic examination of at least 500 cells. Eleven strains produced unequal-sized sister cells by casual examination. To determine if the new mutations were linked to the BLD2 locus, each strain was mated to a bld2-1 strain in the presence of dibutyryl cAMP and IBMX. Because the frequency of mating between two aflagellate strains is low, random meiotic progeny were examined. The presence of meiotic progeny with two flagella indicated the mutations were separable from the bld2-1 allele. None of the 250 alleles was tightly linked to bld2-1.

Noncomplementation screen for new alleles of BLD2:
Because most aflagellate haploid mutant strains harbor mutations that did not map to the BLD2 locus, we performed a noncomplementation screen to isolate new bld2 alleles. The diploid strain maa2-8 MAA7 bld2-2 NIT2 ac17-1/MAA2 maa7-4 BLD2 nit2-1 AC17; nit1-1/nit1-1 was used for the screen (Fig 9A). This strain is heterozygous for bld2-2, and cells swim normally in liquid medium. Mutations that remove the function of the wild-type copy of BLD2 should reveal the Bld2^- phenotype, so that strains with new bld2 mutations would be unable to swim. To ensure that this Bld2^- phenotype resulted from a new bld2 mutation rather than from mitotic recombination or chromosome loss of linkage group III, we constructed a diploid strain with both proximal and distal markers. maa2-8 and maa7-4 define two closely linked mutations that confer resistance to 5-fluoroindole. They are recessive and complement (PALOMBELLA and DUTCHER 1998 ). A maa2-8 MAA7/MAA2 maa7-4 strain is sensitive to 5-fluoroindole, but maa2-8/maa2-8 and maa7-4/maa7-4 strains are resistant to 5-fluoroindole. Thus, a strain that is aflagellate due to mitotic recombination or chromosome loss events would become resistant to 5-fluoroindole (Fig 9C). A strain that becomes aflagellate due to a mutation should remain 5-fluoroindole sensitive, like the parental strain (Fig 9B). In addition to new alleles, noncomplementation screens can yield dominant mutations and dominant enhancers (KENNISON and TAMKUM 1988 ; HAYS et al. 1989 ; STEARNS et al. 1989; SIMON et al. 1991 ).

View larger version (38K): In this window In a new window Download PPT slide   Figure 9. Scheme for isolating and observing phenotypes of new alleles at the BLD2 locus. (A) The diploid strain maa2-2 MAA7 bld2-2 NIT2 ac17/MAA2 maa7-3 BLD2 mit2-1 AC17; nit/nit1 was used for the screen. The strain is heterozygous for bld2-2, so it swims. The maa2-8 and maa7-3 alleles are recessive noncomplementing alleles that confer resistance to 5-fluoroindole. These alleles allow for the selection and/or monitoring of mitotic segregation events. A maa2-2 MAA7/MAA2 maa7-3 strain would be 5-fluoroindole sensitive, but strains homozygous or hemizygous for either maa2-8 or maa7-3 would be 5-fluoroindole resistant. The diploid strain was mutagenized by insertional mutagenesis with the NIT1 gene. Nit+ transformants were screened for the inability to swim. (B) Strains with a new mutation at BLD2 that result in a nonswimming phenotype would be 5-fluoroindole sensitive. (C) Nonswimming strains can also arise from mitotic recombination or chromosome loss events that result in bld2-2 becoming homozygous or heterozygous, but these strains would be 5-fluoroindole resistant. (D) Unlinked dominant or dominant enhancer mutations were identified as mitotic segregants that become able to swim. A possible mitotic segregant is shown. (E–G) Generation of strains that are homozygous or hemizygous for the NIT1 insertion. Mitotic segregation events were selected on medium containing 5-fluoroindole and homozygous or hemizygous for the bld2-2 allele. (F) Nit1+ Nit2- colonies were sensitive to chlorate and resistant to 5-fluoroindole. They contain mitotic recombination events distal to the BLD2 locus. (G) Nit2- colonies should be resistant to chlorate and 5-fluoroindole and be homozygous or hemizygous for the 4GI allele.

The diploid strain was transformed with the plasmid pMN56, which contains the NIT1 gene (NELSON et al. 1994 ). A total of 3000 Nit1⁺ transformants were screened for the inability to swim in liquid medium, and five nonswimming strains were recovered. Each of these strains was sensitive to 5-fluoroindole, so their nonswimming phenotype was due to a new mutation and not to mitotic recombination or to chromosome loss events. When plated for single colonies, four of these strains resulted in both swimming and nonswimming colonies. These four strains are likely to contain either dominant mutations or dominant enhancers of bld2-2. The swimming cells are likely to arise via mitotic recombination or chromosome loss events that result in the loss of the new mutation (Fig 9D). In all four cases, some of the swimming cells are no longer Nit1⁺. These four strains were not pursued at this time. One of the five strains, 4G, was consistently aflagellate and was investigated further (Fig 9C). The insertional allele is designated 4GI.

4GI is linked to BLD2:
To determine whether the 4GI allele is an allele of the BLD2 locus, the chromosomal location and phenotype of the strain were examined by mitotic and meiotic recombination. Mitotic segregants of the 4G strain were selected on medium containing chlorate (Fig 9E and Fig G). Nit2⁺ Nit1⁺ cells die on chlorate, a suicide substrate for nitrate reductase, while cells that are either Nit1^- or Nit2^- survive on chlorate (SCHNELL and LEFEBVRE 1993 ). Sixty chlorate-resistant colonies were isolated, and all 60 were homozygous or hemizygous for the distal mutations that confer resistance to 5-fluoroindole (Fig 9E). Thus, 4GI maps to linkage group III and is proximal to MAA2 and MAA7.

The 4GI allele has a lethal phenotype when homozygous or hemizygous:
With the goal of observing the phenotype of the 4GI allele in the absence of the bld2-2 allele, the phenotypes of strains obtained by chlorate selection were examined. Segregants that are Nit1^- Nit2⁺, which are able to grow on nitrite medium but not on nitrate medium (Fig 9E), would be homozygous or hemizygous for the bld2-2 allele. Segregants that are Nit1⁺ Nit2^-, which are unable to grow on either nitrite or nitrate media (Fig 9G), would be homozygous or hemizygous for the 4GI allele. These two classes would be expected to be generated at equal frequencies. However, all 60 chlorate-resistant isolates were Nit1^- Nit2⁺, and none were Nit1⁺ Nit2^-. This result suggests that strains that are homozygous or hemizygous for 4GI cannot be isolated from the 4G strain and that the 4GI allele may confer a lethal phenotype. Southern blot analysis of the Nit1^- Nit2⁺ mitotic segregants indicated that all copies of the inserted NIT1 gene were lost, and thus the insertion occurred at a single site, which was linked to the BLD2 locus (Fig 10). To confirm that Nit1⁺ Nit2^- colonies could not be isolated from this strain, additional mitotic segregation events on linkage group III were selected on medium containing 5-fluoroindole (Fig 9, E–G). In total, 87 5-fluoroindole-resistant colonies were obtained. Sixty-one percent of the isolates were Nit1^- Nit2⁺ (Fig 9E), but none of the isolates were Nit1⁺ Nit2^- (Fig 9G). The remainder were Nit1⁺ Nit2⁺ (Fig 9F), indicating that the recombination events occurred distally to the BLD2 locus. These 5-fluoroindole-resistant isolates were the result of at least 13 independent events. Given that about one-third of the isolates from selection on 5-fluoroindole with the parental bld2-2/BLD2 diploid (Fig 9A) were Nit2^-, the probability of obtaining 13 5-fluoroindole-resistant colonies that were Nit2⁺ and none that were Nit2^- by chance is 1.5 x 10^-7. The mitotic segregation results strongly suggest that the 4GI allele confers a lethal phenotype to hemizygous or homozygous strains.

View larger version (58K): In this window In a new window Download PPT slide   Figure 10. Digests of 4G genomic DNA were probed with the vector probe. (A–F) Segregation of vector sequences in 5-fluoroindole-selected mitotic segregation events. Genomic DNA was digested with PvuII and probed with vector sequences. (A, B, E, and F) Nit1+ Nit2+ segregants that are 5-fluoroindole resistant and chlorate sensitive. The vector sequences are present. (C and D) Nit1- Nit2+ segregants that are 5-fluoroindole resistant and chlorate resistant. (G) Genomic DNA digested with an enzyme that does not cut within pMN56 (NdeI) results in a single band.

rgn1-1 suppresses the lethality of 4GI, and the double mutant has a meiotic defect:
To further examine the possible lethal phenotype and the map location of 4GI in haploid strains, we attempted to germinate the 4G diploid strain following mating to a haploid strain using colchicine to block nuclear fusion, which results in haploid progeny from stable diploid strains (DUTCHER 1988 ). The presence of colchicine blocked mating completely. We resorted to analyzing the meiotic progeny from triploid zygotes through seven successive generations to ensure that the progeny carried the rgn1-1 mutation and that they were haploid and not aneuploid (see MATERIALS AND METHODS for complete description; DUTCHER and GIBBONS 1988 ). Four of these multiply backcrossed strains (strains 3155–3158) were crossed to strain 3161 (Table 1; ac17 nit2 BLD2; RGN1). If the 4GI RGN1 genotype was inviable, then we anticipated that about one-quarter of the progeny would die. If the 4GI rgn1 and the 4GI RGN1 genotypes were inviable, then we anticipated that about one-half of the progeny would die. All zygotes germinated (n = 736), but most of the progeny died. Only 30 of the 2944 dissected zygospores produced colonies. We observed 99% inviability rather than the expected 25% or 50% inviability, which suggested that the 4GI allele conferred lethality to the meiotic progeny regardless of the specific haploid genotype. To better analyze the phenotypes of the viable progeny obtained from this cross, 86 additional viable progeny were collected from plating germinated zygotes. Of the 116 progeny, 27 were Nit1⁺ Nit2⁺ and assembled some flagella; 83 of the progeny were Nit^- and were fully flagellate; 5 of the progeny were Nit1^- Nit2⁺ and were fully flagellate; and 1 of the progeny was Nit^- and assembled some flagella. Genotypes for these four classes are proposed in Table 6.

Additional crosses determined the genotypes for a subset of the progeny in Table 6. Ten of the 83 Nit^- progeny were crossed to strain 3163. Each of these crosses showed good viability; >90% of the meiotic progeny survived. Three of the five strains contained the rgn1-1 mutation, and two did not as evidenced by the presence of aflagellate (bld2-2) and partially suppressed (bld2-2 rgn1-1) progeny. Five progeny were flagellated and were likely to be BLD2 and the genotype at the RGN1 locus was not scored. The excellent viability observed in these crosses suggests that the inviability in the meiotic crosses arises from the 4GI mutation and not from aneuploidy in the progeny. The proposed genotype of the single Nit^- colony that assembled some flagella was verified in a cross by a NIT2 BLD2 RGN1 AC17; nit1 strain (3162) to ask if the 4GI and NIT1 mutations were present. High frequencies of inviable meiotic progeny were produced. Consequently, zygotes were plated on nitrate medium to select for zygospores that had the NIT1 NIT2 genotype. Twelve Nit1⁺ colonies were obtained, which showed that the NIT1 allele was present. Each of the colonies showed the partially flagellate phenotype, which suggested that the 4GI mutation was present (Table 2; 4GI rgn1-1). The Nit1^- strain as well as the 2 of the 27 Nit1⁺ Nit2⁺ progeny were crossed by NIT2 bld2-2; rgn1-1; nit1 cells (strain 3159) and plated to nitrite medium to select for zygospores that contained the NIT2 allele. Among 15 colonies examined from each of the three crosses, none showed an aflagellate (Bld2^-) phenotype, which suggests that the rgn1-1 allele was present in these progeny. There were progeny that had 25% aflagellate cells and progeny that had 90% aflagellate cells. The lack of gross aneuploidy in the progeny with the 4GI allele is supported by the 2⁺:2^- segregation of the unlinked ACT2 locus (Table 6). We recovered fewer Nit1⁺ Nit2⁺ progeny than would have been expected from the number of Nit1^- Nit2^- progeny. This may reflect the slower growth of the Nit1⁺ Nit2⁺ progeny among the random meiotic progeny or incomplete suppression of the lethality by rgn1-1. These crosses suggest that the 4GI mutation fails to complement bld2-2 and maps 4 cM from NIT2, which is the same approximate location as BLD2.

The 4G strain has flagellar, basal body, and cytokinesis phenotypes similar to other bld2 strains:
The phenotype of the 4G diploid strain was similar to that of bld2-1, bld2-2, and bld2-3 haploid strains. Like the bld2 strains, the 4G strain was aflagellate (Table 2). Cleavage furrow placement, colony size, and microtubule rootlet phenotypes were also similar to those of the bld2 strains. The distribution of sister cell sizes was indistinguishable from the bld2-2 distribution (Fig 2). The 4G strain produced small colonies when compared to the heterozygous diploid strain (Fig 3). The population doubling time of the 4G strain was qualitatively different from that of bld2 strains; it took longer for 4G cultures to reach the same cell density as bld2/bld2 diploid cultures. Staining with the antibody that recognizes acetylated -tubulin revealed that the 4G strain had abnormal microtubule rootlets similar to those of bld2 strains (Fig 4). The microtubule rootlets in 4G cells were disorganized, and they were present in abnormal numbers.

The haploid 4GI rgn1-1 strain showed a lower level of suppression relative to the other bld2 rgn1-1 strains. Only 9% of the cells in the double mutant assembled flagella (Table 2). The cleavage furrow defect was similar to other bld2 rgn1-1 strains (Fig 2I). 4GI rgn1-1 colonies were smaller than other bld2 rgn1-1 colonies (data not shown). Rootlet microtubules in the flagellated cells were similar to those seen in other bld2 rgn1-1 strains (data not shown). Nine serially sectioned basal bodies from 4GI rgn1-1 haploid strains were examined. The microtubule blade phenotype varied from basal body to basal body as observed for other alleles. Four of the basal bodies consisted only of doublet blades; four others had both doublet and triplet blades, and one had only triplets. Unlike the basal bodies from other bld2 rgn1-1 alleles, no singlet or absent blades were observed. In most if not all basal bodies, the cartwheel, transitional fibers, and stellate fibers were present and appeared normal.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Is the BLD2 locus essential?
Few mutations have been identified that affect the assembly pathway of basal bodies. The bld2-1 mutation is one of these. This allele has the phenotype that many cells lack or have very short basal bodies, and those cells have basal bodies with only singlet microtubules (GOODENOUGH and ST. CLAIR 1975 ). It is striking that we as well as others have been unable to isolate additional alleles at this locus among a large collection of aflagellate strains (Table 5; GOODENOUGH et al. 1974 ; ADAMS et al. 1982 ; PIPERNO et al. 1998 ). This result suggests that the null phenotype at the BLD2 locus is unlikely to be solely the absence of flagella.

Several lines of evidence suggest that 4GI is a new allele at the BLD2 locus and that its lethal phenotype represents the null phenotype of the BLD2 gene. First, 4GI maps to the same region of linkage group III as the BLD2 locus (Table 2). Second, the 4G strain shares phenotypes with bld2-1 strains beyond the one used in the noncomplementation screen, which was the absence of flagella. Most aflagellate strains do not show the same spectrum of cleavage furrow and spindle placement phenotypes as the bld2 strains (Table 5). Among our collection of aflagellate strains, only 7% showed cleavage furrow-placement defects. The 4G diploid strain is aflagellate, produces small colonies, has cleavage furrow-placement defects, and has abnormal microtubule rootlets. Dominant enhancers might not be expected to show the same constellation of phenotypes as new alleles. Indeed, the microtubule rootlets of one of the other four transformed strains were examined, and this strain had wild-type microtubule rootlets (data not shown).

Two lines of evidence suggest that the 4GI allele has a lethal phenotype. Examination of the mitotic segregants of the 4G strain isolated by chlorate or 5-fluoroindole selection suggests that diploid strains hemizygous or homozygous for the 4GI mutation are inviable. Our resolution of triploid zygotes provides further evidence for the lethal nature of this mutation (Table 6).

It is likely that 4GI is a new bld2 allele and this allele results in lethality. Is the lethality of 4GI due to the loss of the BLD2 gene or another closely linked gene that is also lost in this insertion allele? Because insertions of transforming DNA in Chlamydomonas are often associated with deletions of DNA, we were initially reluctant to conclude that the BLD2 locus had an essential function. It is possible that an essential gene(s) adjacent to the BLD2 gene was removed, and the lethal phenotype associated with the 4GI mutation is the result of the deletion of another gene. However, the observation that rgn1-1 can suppress the lethality of the 4GI mutation argues strongly that the lethal phenotype does not result from the deletion of an adjacent gene, but from the inactivation or loss of the BLD2 gene. rgn1-1 fails to suppress other mutations that affect basal bodies or flagella; these include uni3-1, vfl1-1, vfl2-1, vfl2-1, or vfl5 (data not shown). It remains a formal possibility that the lethality does not arise from the loss of the BLD2, but from another linked locus. The lethality of the 4GI allele can be tested definitively once the BLD2 gene is cloned. It will be possible to ask if the same DNA that rescues the flagellar phenotypes is sufficient to rescue the lethality. A chromosome walk that encompasses the BLD2 locus has recently been completed (C. RACKLEY, A. M. PREBLE, J. STANGA and S. K. DUTCHER, unpublished results). This walk should make it possible to clone the BLD2 locus and test this hypothesis.

Despite repeated crosses to a variety of strains, we found that strains bearing the 4GI mutation primarily produced inviable meiotic progeny. As described previously, homozygous bld2/bld2 zygotes have a recessive meiotic phenotype; they fail to produce viable m