Flagellar length in the biflagellate alga Chlamydomonas reinhardtii is under constant and tight regulation. A number of mutants with defects in flagellar length control have been previously identified. Mutations in the three long-flagella (lf) loci result in flagella that are up to three times longer than wild-type length. In this article, we describe the isolation of long-flagellar mutants caused by mutations in a new LF locus, LF4. lf4 mutations were shown to be epistatic to lf1, while lf2 was found to be epistatic to lf4 with regard to the flagellar regeneration defect. Mutations in lf4 were able to suppress the synthetic flagella-less phenotype of the lf1, lf2 double mutant. In addition, we have isolated four extragenic suppressor mutations that suppress the long-flagella phenotype of lf1, lf2, or lf3 double mutants.
THE ability to assess and to regulate size is an important factor in many cellular processes. The active control of flagellar length in the biflagellate alga, Chlamydomonas reinhardtii, provides an excellent opportunity to use genetics to dissect a size control mechanism (Lefebvreet al. 1995). Active regulation of flagellar length is demonstrated by several lines of evidence. First, flagella are disassembled and reassembled to proper size at specific stages during both vegetative growth and sexual differentiation (Johnson and Porter 1968; Cavalier-Smith 1974). Second, cells are able to regrow new flagella to their original length when their flagella are amputated (Lefebvreet al. 1978). Third, a cell maintains its two flagella at the same length. This is demonstrated when a single flagellum is amputated from a biflagellate cell. The intact flagellum is disassembled while the amputated one is reassembled until both flagella reach equal length at which time they elongate together until the proper length is attained (Rosenbaumet al. 1969).
The most striking evidence for active length control comes from the phenotypes of mutants which have lost flagellar length control. Three long-flagella (lf) loci (McVittie 1972; Jarviket al. 1976; Barselet al. 1987) have been previously described. Mutations in lf loci result in cells with flagella as long as 40 μm, whereas wild-type cells never have flagella longer than 16 μm. The fact that these mutations disrupt an active mechanism controlling length is shown by the following experiment. During the sexual cycle, gametes of opposite mating type fuse, forming a quadriflagellate dikaryon cell with two flagella contributed from each parent. In quadriflagellate cells formed by mating a wild-type strain with a lf mutant strain, the two long flagella shorten in minutes to wild-type length, indicating that size control is rapidly imposed on the flagellar pair donated by the long-flagella parent (Starling 1969). It is important to note that in this experiment flagellar length is not averaged, but the preset wild-type length is enforced on the mutant flagella.
To identify new genes involved in flagellar length control and to uncover possible genetic interactions among length control genes, we have isolated mutants in five new loci. We have identified multiple alleles at a new LF locus, LF4, using DNA insertional mutagenesis (Tam and Lefebvre 1993). These new lf mutants are phenotypically similar to mutations in the other three LF genes, in that the lf4 mutants assemble flagella that are two to three times wild-type length. However, all lf4 mutants differ in two important ways from the previously described lf mutants. All lf4 mutants are able to regenerate flagella rapidly after deflagellation, whereas lf1 and several alleles of lf2 regenerate flagella very slowly. Second, lf4 mutants in double mutant combination with any of the other three lf loci have the long-flagellar phenotype, while any double mutant combination of lf1, lf2, or lf3 alleles is flagella-less. lf4 mutations were shown to be epistatic to lf1, while lf2 was found to be epistatic to lf4 with regard to the flagellar regeneration defects. In addition, mutations in lf4 were able to suppress the flagella-less phenotype of the lf1, lf2 double mutant.
Mutations in four suppressor of long-flagella (slf) loci were isolated using a screen that takes advantage of the synthetic flagella-less phenotype observed for double lf mutants. In addition to suppressing the flagella-less phenotype of double lf mutants, all slf mutants were able to suppress the flagellar length defect of single lf mutations to wild-type length. Three of the slf suppressors showed allele-specificity but not locus-specificity while suppression by slf4-1(d) was neither allele nor locus-specific. All slf mutations were found to be dominant for their suppression phenotype. The genes identified by these mutations may encode proteins whose products play a role in flagellar length regulation.
MATERIALS AND METHODS
Cell strains and media: The C. reinhardtii strains used in this study are listed in Table 1. All cells were grown under constant light (unless otherwise noted) in M (Minimal) medium (Harris 1989) or TAP medium (Gorman and Levine 1965).
Genetic analysis: Techniques for mating and tetrad analysis were as previously described (Jameset al. 1988). Meiotic progeny were scored for flagellar length by resuspending a single colony of cells in 100 μl of M medium in individual wells of a 96-well microtiter plate, and then growing the cells under constant light overnight. Cells were then fixed in 1% glutaraldehyde. For routine phenotypic analysis, fixed cells were examined by phase-contrast microscopy (×40 objective), and flagellar lengths were compared using an ocular micrometer. Flagellar lengths were also measured on a Zeiss Axioplat microscope (Carl Zeiss, Inc., Thornwood, NY) equipped with an Argus 10 Image Analysis System. Flagellar regeneration was assayed using methods previously described (Lefebvreet al. 1978).
Diploids for complementation tests were constructed by mating strains containing the lf mutations of interest along with one of the complementing auxotrophic markers, arg7 and arg2, and then selecting for growth on M medium. To confirm that putative diploid strains were in fact diploid, two additional phenotypes were scored. Diploid cells were identified as having the minus mating type and larger cell size than the haploid parents. Six diploid strains were measured for each complementation test.
Genetic screens: DNA insertional mutagenesis was performed as described (Tam and Lefebvre 1993; Smith and Lefebvre 1996, 1997). Single colonies were resuspended into 96-well culture plates and screened for aberrantly swimming cells using a stereomicroscope (Zeiss DR-C at ×80 magnification) because excessive flagellar length causes impaired motility. These aberrantly swimming strains were then fixed in 1% glutaraldehyde and then screened by phase-contrast microscopy (×400) to identify those mutants with longer than wild-type length flagella.
To screen for suppressor mutations, double mutant strains (lf1-1 lf3-2; lf2-1 lf3-2; lf2-3 lf3-2; and lf1-1 lf3-1) were constructed by genetic crosses. A single colony isolate of each double mutant was used to start a 4 liter culture in M medium. These cultures were grown on a 12-hr, light/dark cycle with constant aeration until late log phase (~3–5 × 106 cells/ml). Cells were collected by low-speed centrifugation (~2000 g) and resuspended to a density of 108 cells/ml. Each suspension was placed in a small Petri dish, constantly stirred, and was irradiated for 180 sec with ultraviolet (UV) light from a General Electric G8T5 germicidal lamp at a distance of 18 cm. This UV exposure yielded ~30–50% survival. Immediately after mutagenesis, 200 μl aliquots of the cell suspension were added to 10 ml of M medium in glass culture tubes. The tubes were placed in the dark for 6 hr to prevent photorepair. The tubes were then placed in constant light for 2 wk, or until swimming cells could be seen at the meniscus of the medium. Phenotypic revertants were isolated by pipetting ~100 μl of liquid from the meniscus and plating cells to obtain single colonies (M medium). One colony was retained from each original culture tube to ensure that each phenotypic revertant examined was produced by an independent event.
Seven lf mutants generated by DNA insertional mutagenesis: A variety of C. reinhardtii flagellar mutants have been generated by DNA Insertional Mutagenesis (Tam and Lefebvre 1993). This procedure involves transforming nit1 cells with a plasmid spMN24 that contains the structural gene for nitrate reductase (NIT1) (Fernandezet al. 1989). Nonhomologous integration of the plasmid into the genome results in a mutagenic event in which the insertion of the transgene is often accompanied by a deletion and/or rearrangement of the genomic DNA at the site of insertion.
Seven lf mutants were isolated by this lab and by others (Table 1) by first screening for cells with motility defects because the lf4 mutation causes cells to swim poorly relative to wild-type cells. Next, motility-defective mutant strains were then individually screened using phase-contrast microscopy to identify those with extra-long flagella. While the length of the flagella on wild-type cells varies, no flagella exceeding 16 μm are ever observed in wild-type populations. In lf mutant populations, however, cells have flagella that are three times as long as the maximum wild-type value.
Each lf mutant was backcrossed to a wild-type strain, and the segregation of length phenotype in the F1 progeny was analyzed. All seven mutants showed Mendelian segregation of the lf phenotype (two lf: two wild-type progeny per tetrad) indicating that in each case the length defect was caused by a single mutation (data not shown).
Seven new lf mutants define a single new complementation group, LF4: Stable vegetatively growing diploid strains can be constructed in C. reinhardtii by mating cells containing complementing auxotrophic markers (Ebersold 1956; Jameset al. 1989). All seven mutants were demonstrated to be recessive to the wild-type allele in stable diploid strains. Diploids were then constructed using all seven mutants in pairwise combinations. In every case tested, heterozygous diploid cells had long flagella (Table 2), indicating that the seven mutations define a single complementation group. Several strains containing mutant alleles from this complementation group were crossed to strains with mutations in one of the three previously identified lf loci: lf1, lf2, and lf3. Wild-type recombinants were recovered in all crosses (Table 3), demonstrating that the seven mutants are alleles of a new lf locus, LF4.
LF4 maps to linkage group XIV on the Chlamydomonas genetic map: The lf4 locus is linked to the linkage group XIV (LG XIV) marker ery2b (conferring erythromycin resistance) (O'Brien 1980). Results from two-factor crosses place lf4 6.7 cM from ery2b, and 25 cM from mam2 (Table 4). The LF4 centromere distance of 16.2 cM was determined from crosses with the centromere-linked marker, lf3 (Table 4). These results place LF4 on the right arm of LGXIV (Figure 1). Gene order was inferred from the centromere data, but not directly tested by three-factor crosses. The placement on LGXIV was confirmed by restricted fragment length polymorphism (RFLP) mapping (Ranumet al. 1988) using a fragment of DNA as a hybridization probe, flanking the site of plasmid insertion in lf4-1 (data not shown).
Phenotypic analysis of lf4: While flagella from a wild-type population may vary in length, no flagella exceed 16 μm in size (Figure 2A). In lf mutant populations, cells also have flagella with a wide variety of lengths, but many cells have flagella that exceed the wild-type maximum length, with some as long as 40 μm. The seven lf4 mutant populations also exhibit this same excess flagellar length phenotype (Figure 2, B–H).
One perhaps counter-intuitive aspect of the phenotype of several of the original lf mutations is that they are also defective in their ability to regenerate flagella following amputation, and this additional phenotype always cosegregates with the length defect (Barselet al. 1987). For example, both lf1 and lf2-3 are unable to regenerate flagella following amputation by mechanical shearing. When examined 2 hr after deflagellation, greater than 90% of the cells remained flagella-less (Figure 3, A–B).
All seven lf4 alleles were examined for their ability to regenerate flagella. In all experiments, the cells were examined immediately following shearing in a homogenizer to confirm that 100% of the cells had been deflagellated. After 2 hr, the wild-type control cells had regrown full-length flagella as expected (Rosenbaumet al. 1969). The lf4-3 allele was able to regenerate flagella within 2 hr following amputation when greater than 80% of the population had near wild-type length flagella (Figure 3C). During this same time period, all other lf4 alleles also regenerated flagella to at least wild-type length, with some cells having flagellar lengths exceeding 16 μm. Thus, each lf4 mutant regenerates flagella with at least wild-type rates of regeneration following amputation.
Double mutant phenotypes of lf4 with other lf loci: As previously described, any double mutant combination of lf1, lf2, or lf3 alleles results in a flagella-less phenotype (Barselet al. 1987). A similar phenotype has also been seen with many combinations of short-flagella mutants as well (Jarvik et al. 1984; Kuchka and Jarvik 1987). In contrast, double mutants containing any allele of lf4 and any of the other three lf loci have the lf phenotype (Figure 3, D–E).
Because lf1 and lf2-3 cells are defective in flagellar regeneration, while lf4 mutants regenerate normally, it was possible to determine the epistatic relationship among these mutants with regard to flagellar regeneration. The double mutants lf4-3 lf1-1, and lf4-3 lf2-3, were constructed and examined for the ability to regenerate flagella. The double mutant lf4-3 lf1-1 was able to regenerate flagella following amputation (Figure 3D), whereas the double mutant, lf4-3 lf2-3 was unable to regenerate flagella following deflagellation (Figure 3E). The regeneration phenotype of the lf4-3 lf1-1 double mutant combination is the same as the lf4-3 mutant alone (Figure 3C). The lf4-3 lf2-3 double mutant displayed the same regeneration phenotype as the lf2-3 mutant alone (Figure 3B). Thus, with regard to the flagellar regeneration phenotype, lf4 is epistatic to lf1, and lf2 is epistatic to lf4.
Triple mutant phenotype: A triple lf1-1 lf2-3 lf4-3 mutant was constructed by crossing a lf1-1 lf4-3 double mutant and a lf2-3 lf4-8 double mutant. Only progeny with long flagella were recovered, indicating that the triple mutant combination results in a long-flagella phenotype (data not shown). Thus the lf4 mutation can act as an extragenic suppressor of the flagella-less phenotype of the double lf1 lf2-3 mutant, as well as functioning as a suppressor of the lf1-1 regeneration defect.
Genetic screen for extragenic suppressors of lf mutants: To isolate extragenic suppressors of lf mutations, we capitalized on the synthetic flagella-less phenotype of double lf mutant strains. We were able to obtain phenotypic revertants of double lf mutants by isolating cells that swim up from a pellet of nonmotile cells. These revertants were then backcrossed to determine whether the phenotypic reversion was because of a reversion of one of the two lf loci, or whether it was because of an extragenic suppressor mutation. A similar screen was used to obtain suppressors of short-flagella mutations (Kuchka and Jarvik 1987).
Several double lf mutant strains were used as parents for the isolation of revertants: lf1-1 lf3-2; lf2-1 lf3-2; lf2-3 lf3-2; lf1-1 lf3-1. Cultures grown from a single colony were mutagenized with UV light (Jameset al. 1989), split into 10 separate 10 ml cultures, incubated in the dark for 6 hr to prevent photorepair, and then grown under constant light. Initially, all of the cells in the culture were nonmotile, accumulating at the bottom of the culture tube. After several weeks of growth, motile cells were observed at the top of the culture tube. In cultures not exposed to UV light, spontaneous motile revertants were also obtained. These motile revertants were isolated by pipetting the swimming cells from the meniscus. The cells were then plated to obtain single-colony isolates, and only one isolate from each 10 ml culture was characterized further.
A total of 49 swimming revertant strains were isolated from the four different double mutant strain backgrounds (Table 5). Of these independent revertants, 44 strains were isolated from UV-mutagenized double mutant populations, and five revertants arose in nonmutagenized cultures. Some of the revertant strains had cells with flagella of wild-type length and the rest had cells with long flagella. For example, 13 different swimming revertants were isolated from lf1-1 lf3-2 double mutant populations; two contained spontaneous mutations and 11 were from UV-mutagenized cultures. Eleven of the 13 had long flagella, while only two revertants had flagella of wild-type length.
Each revertant strain was backcrossed to wild type to determine whether the phenotypic reversion was caused by the action of a new extragenic suppressor mutation, or by an intragenic event involving one of the two original lf mutations. If motility was restored by an intragenic reversion event, the resulting backcross tetrads should each contain two progeny with wild-type flagellar length and two progeny with long flagella. On the other hand, if phenotypic reversion was caused by an extragenic suppressor mutation, then the segregation patterns in the tetrads from the backcross would be complex. In this case, three genes would segregate in the cross, producing eight different genotypes among the progeny. One of these eight genotypes would contain the original two lf mutations, along with the wild-type allele at the suppressor locus, producing the flagella-less phenotype. Therefore, the presence of an extragenic suppressor mutation in the motile revertant strain was indicated by the identification of flagella-less cells among the progeny of a backcross.
It was expected that most revertants with a long-flagella phenotype would result from an intragenic event, either true reversion or intragenic suppression. Surprisingly, among the six different long-flagellar revertants that were analyzed further by backcrosses, the segregation data indicated that only two were caused by intragenic events. That is, when the motile revertant strain was crossed to a wild-type strain, all of the resulting tetrads contained two wild-type progeny and two progeny with long flagella. From this result, we can conclude that in these two revertants either an intragenic event or tightly linked extragenic event occurred. In the other four revertants with long flagella, the backcross segregation pattern was more complex and flagella-less progeny were generated. These results suggest that an extragenic event occurred to cause the phenotypic reversion in the parental strains.
Six revertants with wild-type flagellar length were also backcrossed. All six crosses resulted in complex segregation patterns that included tetrads containing flagella-less progeny. From these data we conclude that all six revertants with wild-type length flagella contain extragenic suppressor mutations.
Extragenic suppressor mutations were isolated from each of the four double lf mutant backgrounds: Four revertant strains with wild-type flagellar length, E13, F4, F3, and E30, were analyzed in greater detail after they were shown to contain extragenic mutations. The extragenic suppressor mutations present in these strains were designated slf. The genotypes and phenotypes of the triple mutants are listed in Table 6.
To examine the flagellar phenotype of the slf mutants, revertant strains were crossed to a wild-type strain. The genotypes of each of the four meiotic products from any tetrad with two flagella-less progeny could be inferred. The two flagella-less progeny were presumed to contain the original two lf mutations, along with the wild-type allele at the suppressor locus. The remaining two progeny from this tetrad would therefore contain the slf mutation along with the wild-type alleles at the two lf loci. By examining tetrads that contained at least two flagella-less progeny, the phenotype of cells containing the slf mutation alone was determined. It was established that all cells containing any of the four slf mutations alone had wild-type length flagella (Table 6).
As discussed above, some lf mutations, especially lf1 and lf2 alleles, have defects in flagellar regeneration after amputation. To determine whether the slf mutations share this phenotype, the kinetics of flagellar regeneration was examined for each slf mutant. We found that after 2 hr, all slf mutants regenerated flagella to wild-type length (data not shown) as well as wild-type cells. Thus, none of the four slf mutants were defective in flagellar regeneration.
Characterization of suppression: All four slf mutations were examined for the ability to suppress single lf mutations. Each suppressor was crossed to lf1 and to multiple alleles of both lf2 and lf3. All four slf mutations suppressed the length defects of lf1, lf2-3, and lf3-2. They differed in their ability to suppress lf2-2 and lf3-3. For example, while slf1-1(d) was able to suppress lf3-2, it did not suppress the lf3-3 length defect. In contrast, slf4-1(d) was able to suppress both lf3-2 and lf3-3. None of the slf mutations were able to suppress the length defects of either lf4-1 or lf4-8 (Table 7).
Although the slf mutations suppressed defects in flagellar length control, it was possible that the defects in regeneration seen for lf1 and lf2 mutations would not be suppressed. To test this possibility, specific slf, lf double mutants, in which the lf allele is defective in flagellar regeneration, were examined for the suppression of the regeneration defect. For all cases in which the suppressor mutation could suppress the flagellar length defect, it also suppressed the flagellar regeneration defect (data not shown).
The suppressor of long-flagella mutations are dominant in stable diploids: Because the slf mutants had no other phenotype than their ability to suppress lf mutations, diploids were constructed that were heterozygous for the slf mutation being tested and homozygous for the lf3-2 mutation. Using these strains, we were able to determine whether slf mutations were dominant or recessive. Each heterozygous slf diploid strain had wild-type length flagella, indicating that all four slf extragenic suppressor mutations were dominant for the wild-type allele with regard to their ability to suppress lf3-2 (Table 6).
Genetic locations of slf mutations: Two of the slf loci have been placed on the Chlamydomonas genetic map. The loci, slf4 and slf1, are both linked to lf3 on linkage group I (LGI) (Table 8). The slf1 locus maps 24 cM from lf3, and 2.8 cM from the arg7/arg2 locus (Figure 4). The slf4 locus is 22 cM from lf3, but linkage was not detected with arg7/arg2; slf4 is presumably on the opposite arm of LGI (Figure 4).
Although slf2 and slf3 were not mapped, neither of these slf loci correspond to a known lf locus based on the following observations. Neither slf2 nor slf3 is linked to lf1 on LGII, to lf2 on LGXII/XIII or to lf3 on LGI. The slf2 locus is also unlinked to ery2b, an erythromycin-resistance marker linked to the LF4 locus on LGXIV. Linkage to LF4 could not be assayed directly because neither extragenic suppressor mutation suppressed lf4. These data eliminate the possibility that slf2-1(d) is a mutation of any of the four lf loci.
All seven lf insertional mutations examined define a single new complementation group, LF4. While the purpose of this study was to identify new genes involved in flagellar length control, it was expected that insertional alleles of lf1, lf2, and lf3 would also have been isolated by this screen. There are several possibilities to explain the absence of insertion mutations affecting these loci.
The first possibility is that the process of insertional mutagenesis generates predominantly null mutations, and null mutants of lf1, lf2, and lf3 may not have the lf phenotype. Tam and Lefebvre (1993) examined three flagellar insertion mutants at the molecular level. They found that not only were the mutations generated by the insertion of plasmid DNA, but this insertion was accompanied by deletions and rearrangements of the flanking genomic region. The previously identified alleles of lf1, lf2, and lf3 were generated with either chemical mutagens or UV irradiation. While nothing is known about the nature of the genetic lesions in these mutants, it is entirely possible that the null phenotype at these loci may be different from the long-flagella phenotype. If so, then insertional mutagenesis that produces mainly null mutations will not be useful for identifying mutations in lf1, lf2, and lf3.
A second possibility is that the LF4 sequence is a frequent target for insertional mutagenesis because it is a physically large locus relative to lf1, lf2, and lf3. Therefore, the LF4 gene would be more likely to be disrupted by the random insertion of plasmid DNA than would LF1, LF2, or LF3. Alternatively, the LF4 locus could represent a hot-spot for these integration events. These possibilities will be examined as a consequence of the molecular characterization of this locus. We have obtained clones of genomic DNA from the region flanking the insertion in lf4-1. These genomic clones are being used to examine the genomic DNA in all lf4 alleles to characterize the genetic lesions in these mutants.
Two classes of lf mutations: A priori, it is likely that two types of length-control mutations will be uncovered. First, mutations in the cellular mechanisms that detect the length of the flagella would cause the cells either to assemble flagella of an improper size or to lose control of size altogether. Second, mutations in genes that regulate the assembly of the flagella, probably under the control of the detection mechanism, could result in improper flagellar size. For example, if a mutation causes cells to be unable to turn off assembly once it was initiated, flagella of excess length would be produced. It is possible that the length-control loci may be separated into these two classes.
Flagellar assembly occurs under two conditions. During each cell cycle, flagella must be reassembled because Chlamydomonas cells resorb their flagella before division. In addition, cells can regrow their flagella after amputation. Some lf mutations separate the cellular control for these two assembly events because lf1-1 and lf2-3 are obviously able to assemble flagella as part of the cell cycle, albeit to an incorrect size, but these mutants are unable to regenerate their flagella after amputation. Once assembly in these mutants finally begins, flagella of wild-type length or greater are regenerated. Thus, lf1 and lf2 mutants are not defective in flagellar assembly per se, but in the regulation of flagellar assembly, particularly in response to amputation. This observation suggests that LF1 and LF2 may normally function to regulate flagellar assembly rather than to monitor flagellar length.
Genetic interactions between long-flagella loci: The synthetic phenotype of double lf mutants (lf1, lf2, and lf3) suggests that the gene products either interact to form a functional complex or function independently of each other in redundant pathways. In contrast, the lf4 mutations do not exhibit synthetic interactions with the other three lf loci. Double mutants containing an lf4 allele all have the lf phenotype of the single mutant (Figure 3). Analysis of double mutants established that lf4 is epistatic to lf1, and that lf2 is epistatic to lf4 with respect to flagellar regrowth. If the synthetic lf1, lf2 double mutant phenotype is the result of the lf1 and lf2 products failing to form an interaction necessary for function, one would predict that lf4 would show the same epistatic relationship to both genes. The fact that the epistatic interactions of lf4 with lf1 and lf2 differ suggests that lf1 and lf2 may function in redundant or parallel pathways.
Extragenic suppressor mutations: Extragenic suppressors can be grouped into three classes: informational, interactive, and bypass. The four slf extragenic suppressors are unlikely to be interactive suppressors because they do not act in an allele-specific and locus-specific fashion; therefore, the slf mutants are probably either informational or bypass suppressors of long-flagella mutations. Several observations are consistent with the possibility that the slf mutants are informational suppressors. The slf mutants have wild-type flagellar length and are able to regenerate flagella with wild-type kinetics following deflagellation. The lack of any other flagellar phenotype beside the suppression is consistent with an informational suppressor model. In addition, the fact that all four slf mutations are dominant is consistent with the possibility that they are informational suppressors.
Other characteristics of the suppressor mutations do not readily fit an informational suppressor model. For example, if one of the slf mutations is an amber suppressor, it will only suppress mutations caused by an inframe UAG codon. If a second slf mutation is an ochre suppressor, it will only suppress mutations caused by an in-frame UAA codon. The set of alleles that each suppressor can suppress should be unique and nonoverlapping. The data in Table 7 show that the sets of alleles suppressed by each slf mutation have some alleles in common and some alleles that are unique to a particular slf mutation. For instance, slf1-1(d) can only suppress lf1-1, lf2-3, and lf3-2, but not lf2-2 and lf3-3, whereas slf4-1(d) can suppress all five of these alleles. This overlap in suppression sets can be explained if some slf mutations act as informational suppressors and others function as bypass suppressors. In addition, these suppressors are unlikely to be omnipotent suppressors because they cannot suppress the arg2/arg7 mutation used to build the stable diploids, nor can sfl1 or slf3 suppress pf14, an ochre mutation (data not shown) (Williamset al. 1989).
The lack of allele-specificity for suppression is also consistent with the possibility that some, or all, of the slf mutations operate by bypassing the function of the mutant lf loci because bypass suppressors may or may not be allele-specific. We do know that it is possible to obtain bypass suppressors of flagellar length control mutations because the lf4 mutations function as putative bypass suppressors. lf4 mutations can suppress both the flagellar regeneration defect of lf1 and the synthetic flagella-less phenotype of lf1-1 lf2-3. These mutations appear to function as bypass suppressors for two reasons. First, the data are consistent with the possibility that the lf4 gene functions in the same pathway as the lf1 gene. While the original three lf loci have synthetic double mutant phenotypes, the lf4 alleles have no synthetic interactions with the other lf loci. This lack of interaction is consistent with the hypothesis that lf4 functions in the same pathway as the other lf loci. The lf4 lf1 double mutant has the same flagellar-regeneration phenotype as the lf4 single mutant; therefore, lf4 is epistatic to lf1. In other words, mutations in lf4 are able to suppress the regeneration defect of lf1. If the suppression was due to a restoration of an interaction between the LF4 and LF1 gene products, then the extent of suppression would be highly dependent on specific alleles of lf4 tested. Because suppression of lf1 was not found to be specific for individual lf4 alleles, these mutations most likely act as bypass suppressors of the flagellar regeneration defect. These results fit with a model in which the lf4 gene product acts downstream of lf1.
Second, as discussed earlier, the lf4 mutations are probably null alleles. The lack of a functional lf4 gene product is also consistent with a model of bypass suppression. Therefore, it appears that both lf4 and some slf mutations may be functioning as bypass suppressors of the original three lf loci.
The authors thank Andrea Kernan, Greg Pazour, and Elizabeth Smith for graciously providing lf mutant strains. We thank Craig Amundsen and Shinichiro Enomoto for critical readings of the manuscript. This work was supported by National Institute of General Medical Sciences grant 34437 to P.A.L. C.M.A. was supported by National Science Foundation grant DIR-9113444.
Communicating editor: V. G. Finnerty
- Received June 9, 1997.
- Accepted November 3, 1997.
- Copyright © 1998 by the Genetics Society of America