New mutations in flagellar motors identified by whole genome sequencing in Chlamydomonas

Background The building of a cilium or flagellum requires molecular motors and associated proteins that allow the relocation of proteins from the cell body to the distal end and the return of proteins to the cell body in a process termed intraflagellar transport (IFT). IFT trains are carried out by kinesin and back to the cell body by dynein. Methods We used whole genome sequencing to identify the causative mutations for two temperature-sensitive flagellar assembly mutants in Chlamydomonas and validated the changes using reversion analysis. We examined the effect of these mutations on the localization of IFT81, an IFT complex B protein, the cytoplasmic dynein heavy chain (DHC1b), and the dynein light intermediate chain (D1bLIC). Results The strains, fla18 and fla24, have mutations in kinesin-2 and cytoplasmic dynein, respectively. The fla18 mutation alters the same glutamic acid (E24G) mutated in the fla10-14 allele (E24K). The fla18 strain loses flagella at 32?C more rapidly than the E24K allele but less rapidly than the fla10-1 allele. The fla18 mutant loses its flagella by detachment rather than by shortening. The fla24 mutation falls in cytoplasmic dynein and changes a completely conserved amino acid (L3243P) in an alpha helix in the AAA5 domain. The fla24 mutant loses its flagella by shortening within 6 hours at 32?C. DHC1b protein is reduced by 18-fold and D1bLIC is reduced by 16-fold at 21?C compared to wild-type cells. We identified two pseudorevertants (L3243S and L3243R), which remain flagellated at 32?C. Although fla24 cells assemble full-length flagella at 21?C, IFT81 protein localization is dramatically altered. Instead of localizing at the basal body and along the flagella, IFT81 is concentrated at the proximal end of the flagella. The pseudorevertants show wild-type IFT81 localization at 21?C, but proximal end localization of IFT81 at 32?C. Conclusions The change in the AAA5 domain of the cytoplasmic dynein in fla24 may block the recycling of IFT trains after retrograde transport. It is clear that different alleles in the flagellar motors reveal different functions and roles. Multiple alleles will be important for understanding structure-function relationships.


Background
The building of a cilium or flagellum requires molecular motors and associated proteins in a process termed intraflagellar transport (IFT). IFT was first described as a bidirectional movement of 'granule-like' particles along the axoneme in Chlamydomonas reinhardtii [1]. Concurrently, a novel heterotrimeric kinesin was isolated from sea urchin embryos [2]. A temperature-sensitive mutation in the Chlamydomonas FLA10 gene shows it is needed for flagellar assembly [3], and that IFT is dependent upon FLA10 [4]. FLA10 encodes a subunit of the heterotrimeric kinesin first found in sea urchins [5,6]. The Chlamydomonas FLA8 and FLA3 genes encode the other kinesin-2 motor subunit and the kinesin-associated protein (KAP) subunit, respectively [7,8]. The IFT trains are composed of at least 19 proteins, which fall into two complexes, A and B, which are dissociated by salt [5,6]. Complex B contributes to anterograde transport away from the cell body [5], and complex A is involved in retrograde transport toward the cell body [9][10][11]. Anterograde movement requires kinesin-2 and retrograde movement requires cytoplasmic dynein. At the tip, the anterograde IFT particles rearrange into new trains with a different shape and size for retrograde IFT [12]. This simple picture is made more complex by examining the behavior of the BBSome in Caenorhabditis elegans [13]. The BBSome is a complex of seven proteins [14] that is postulated to be involved in the import of G protein coupled receptors in mammalian cells [15,16] and the export of cycling proteins in Chlamydomonas [17]. The BBSome may regulate anterograde IFT assembly and then rearrangement at the tip [13]. Cargo-specific adapter proteins may be important for IFT transport of cargo [18]. IFT is essential for mammalian development as mutants in various IFT proteins are lethal in mice (reviewed in Eggenschwiler and Anderson [19]). Mutations in several IFT proteins and cytoplasmic dynein cause asphyxiating thoracic dystrophy in humans [20][21][22][23].
A collection of temperature-sensitive mutants in Chlamydomonas that assemble flagella at the permissive temperature of 21°C, but lack flagella at the restrictive temperature of 32°C (Table 1) provides an important resource for the analysis of flagellar assembly [11,24,25]. Since many conditional mutants have reduced but sufficient function at the permissive temperature, this collection offers the opportunity to examine IFT in assembled flagella at the permissive temperature to ask about the effects of reduced function. For example, the temperaturesensitive allele in IFT172 suggests a role in remodeling IFT at the tip [26]. IFT is required to transport many of the flagellar proteins from the cytoplasm to the flagella. These include the inner dynein arm protein p28 that fails to be imported in the fla10-1 mutant [6]. Recent isobaric tags for relative and absolute quantitation (iTRAQ) experiments suggest that numerous proteins accumulate or are depleted in the presence of a mutant cytoplasmic dynein even when the length of the flagella has not changed [27], which shows the importance of retrograde movement for moving proteins back to the cell body.
The role of IFT differs between different axonemal proteins/cargos. Piperno et al [6] used temporary dikaryons, which are formed by the mating of two parental cells, to examine the kinetics and localization of proteins using antibodies to proteins in axonemal structures. The parental strains both carried the temperature-sensitive fla10-1 mutation in kinesin-2 [3] that stops IFT within 30 minutes after the shift to the restrictive temperature. One parent is otherwise wild-type, while the other parent has either an oda6 mutation that blocks assembly of the outer dynein arms [45] or an ida4 mutation that blocks assembly of a subset of inner dynein arms [46]. In ida4 x wild-type dikaryons at 21°C, IDA4 appear at the distal end of the mutant flagella by antibody staining and staining moved towards the proximal end with time after mating. In oda6 x wild-type dikaryons, ODA6 behave very differently. Staining appears along the entire length of the flagella 6 minutes after mating. The intensity increased with time. To test the role of IFT in the incorporation of dynein arm proteins, the parental cells were shifted to 32°C for 30 minutes to inactivate kinesin-2. The incorporation of IDA4 was blocked at the restrictive temperature, while ODA6 continued to be incorporated. Thus, the outer dynein arms appear to enter by diffusion or by a different motor complex [6], while the entry of the inner arm component requires kinesin-2. Transport of outer dynein arms also requires an adapter between the dynein arms and IFT. ODA16 functions as a cargo-specific adaptor between IFT particles and outer row dynein needed for efficient dynein transport into the flagellar compartment, as shown by its localization and interactions by immunoprecipitation and yeast two-hybrid experiments [18]. Recent results suggest that transport of tubulin into cilia is mediated by a weak affinity between tubulin and IFT81 and IFT74 [47].
Analysis of IFT using differential interference contrast (DIC) optics and kymographs showed that six of these conditional mutants have defects in the number of anterograde IFT particles or their velocity at 21°C, six have defects in either retrograde IFT particle number or velocity at 21°C, and four have no change in IFT particle number or velocity at 21°C (Table 1). As genes have been identified, it is clear that alleles in the same gene have slightly different properties; these differences must reflect the properties of the mutant alleles and not the function of the gene. For example, the fla1 and fla8 mutants both encode the other motor subunit of kinesin-2 [8] but show differences in the behavior of IFT particles [11]. These differences may reflect the degree of activity/ concentration of the mutant proteins at the permissive temperature.
In Chlamydomonas, seven genes needed for intraflagellar transport have been identified by conditional alleles (Table 1). Nonconditional mutations in nine additional genes as well as in three of the genes with conditional alleles have been identified. RNA depletion of two IFT genes has been analyzed in Chlamydomonas and result in short flagella (Table 1), which may occur because there is only partial knockdown of the genes. Eighteen of the IFT and motor protein genes have mutants or depletion results. In this report, we employed whole genome sequencing to identify the FLA18 and FLA24 genes. These genes encode a new allele in the FLA10 kinesin-2 motor subunit and a new allele in the cytoplasmic dynein DHC1b, respectively.

Whole genome sequencing
Chlamydomonas genomic DNA preparation for whole genome sequencing was prepared as described previously [30]. Three micrograms of DNA were submitted to Genome Technology Access Core (St Louis, MO, USA) for library construction, Illumina sequencing (San Diego, CA, USA), and initial data analysis. For multiplex Illumina sequencing, 7-nucleotide indexes were added to individual DNAs during the library construction before the samples were subjected to sequencing. The fla18 and fla24 samples were tagged with TGAGGTT and GCTTAGA, respectively, and shared the same sequencing lane with two other samples. All resulting sequencing data were de-multiplexed before being subjected to sequence alignment and SNP calling.

dCAPS markers and segregation analysis
Restriction enzymes that provide differences between mutant and wild-type alleles are listed in Additional file 1: Table S1. For fla24, NEBCutter (New England BioLabs, Ipswich, MA, USA) was used to find the appropriate restriction enzyme. However, no restriction enzyme distinguishes between CC-125 and fla18. A dCAPS marker was designed using dCAPS Finder 2.0 (Washington University, St Louis, MO, USA) [48]. A forward primer (fla18-dcapF) introduces a mismatch immediately upstream of the point mutation that creates an MboII recognition site in the wild-type PCR product (GAAGA(N) 8 ) but not in the fla18 PCR product (GAGGA(N) 8 ). The 132 bp PCR product, when digested with MboII, generates 102 bp and 30 bp fragments from wild-type but is uncut in fla18.

Flagellar isolation
Flagella were isolated as described previously [49,50] with the addition of Protease Arrest (G-Biosciences, St Louis, MO, USA).

Flagellar counts
Cells were grown overnight in a 21°C lighted incubator to a density of approximately 2 × 10 6 cells/ml. Cells were then transferred to a 32°C lighted incubator and samples were taken every hour. Samples were prepared by spotting 19 μL of cells onto a microscope slide and adding 1 μL of 2% glutaraldehyde in 0.1 M phosphate buffer (pH 7.0) directly to the spotted cells. A total of 200 cells from each strain at each time point were scored using phase optics (40x) for the presence or absence of flagella. Flagellar length was monitored by immunofluorescence with monoclonal antibody to acetylated α-tubulin (Sigma-Aldrich, St Louis, MO, USA) at a dilution of 1:1,000.

Immunoblots and immunofluorescence
The antibodies were: ift81.3 (a gift from Dr Doug Cole) for immunofluorescence at 1:200; α-tubulin (DM1α from Sigma-Aldrich) used for immunoblots at 1:5,000, DHC1b and D1bLIC (a gift from Dr George Witman, University of Massachusetts, Worcester, MA, USA) were used for immunoblots at 1:2,000 and for immunofluorescence at 1:100; and ε-tubulin was used at 1:500 for immunofluorescence [51]. For immunoblots, flagellar proteins were isolated and resuspended in HEPES/Sr/DTT/sucrose buffer [52]. All the protein samples were stored at −80°C before use. Protein concentrations were ascertained by using Bio-Rad protein assay (Bio-Rad, Hercules, CA, USA) following the manufacturer's instruction. The proteins were boiled for 5 minutes and centrifuged for 1 minute before loading onto the gel. The proteins were size-fractionated on SDS-PAGE minigels (1.0 mm thick, 6% acrylamide gel (for DHC1b) or 10% acrylamide gel (for D1bLIC) prepared from 30% acrylamide and bis-acrylamide solution, 29:1 (BioRad, Hercules, CA, USA)) and transferred to Immobilon-P PVDF membranes (EMD Millipore, Billerica, MA, USA) in 25 mM Tris, 192 mM glycine buffer containing 20% methanol at 62v for one hour. The incubation with primary antibody was incubated overnight at 4°C. The blots were washed three times for 10 minutes each in PBST. The secondary antibody was incubated at room temperature for 1 hour. Goat anti-mouse HRP (BioRad) and goat anti-rabbit HRP (Sigma-Aldrich) were used at 1:5,000 dilution. SuperSignal West Femto Chemiluminescent Substrate (Thermo Scientific, Waltham, MA, USA) was used according to the manufacturer's instructions. The blots were imaged on a FluorChem H2 imager (Alpha Innotech, Santa Clara, CA, USA). Signal quantification analysis was performed by ImageJ (National Institutes of Health (NIH), Bethesda, MD, USA). For immunofluorescence, Alexa 488 goat anti-rabbit (Invitrogen, Grand Island, NY, USA) and Alexa 594 goat anti-mouse were used at a 1:500 dilution with published protocols [51], except newborn goat serum and fish gelatin were not used. All individual immunofluorescence channels were maintained with the same exposure time throughout the time courses.

Results
To identify and understand the function of genes that can be mutated to a temperature-sensitive flagellar assembly phenotype, we are using whole genome sequencing. Mutants with an anterograde IFT defect (fla18) and a retrograde IFT defect (fla24) were chosen for analysis [11]. Each mutant was crossed to wild-type (CC-124 or CC-125) to verify that the phenotype segregated as a single mutation in meiotic progeny. In 56 and 130 tetrads, respectively, the aflagellate phenotype at 32°C segregated two wild-type and two mutant progeny, which suggests a single mutation or several tightly linked mutations. Each mutant was subjected to whole genome sequencing. The coverage ranged from 51-fold for fla24 to 94-fold for fla18 (Table 2).

FLA18 encodes a kinesin-2 subunit
The fla18 mutant strain was crossed to the highly polymorphic strain S1C5 and one fla18 meiotic progeny was subjected to whole genome sequencing [8]. About 71% of the 101 bp reads align to the reference genome (Table 2). A total of 43,103 SNPs/indels unique to the fla18 mutant strain were found after subtracting the SNPs/indels found in 15 other Chlamydomonas strains [53]. Among this set of SNPs/indels, 892 changes map to exons or intron/exon boundaries when synonymous changes are excluded ( Table 2).
In our study of transcript levels following pH shock at various time points during flagellar assembly [54], we found that all of the IFT genes are upregulated at least 2.5-fold within 1 hour of flagellar amputation. Given the fla18 mutant strain has a defect in anterograde IFT [11], we hypothesized that the FLA18 gene is likely to be among the 1,850 upregulated genes. A comparison between the list of genes that have SNPs/indels in fla18 and the list of upregulated genes identifies 85 SNPs/ indels in 59 genes. One of the changes is a glutamic acid (GAG) to glycine (GGG) in the FLA10 kinesin (E 24 G, Table 3). In the fla10-14 mutant strain, the same glutamic acid is replaced by lysine (E 24 K) [8]. We then verified that the temperature-sensitive phenotype was linked to FLA10 in 20 progeny from a cross of fla18 x S1C5 ( Table 3). The  Table 3). These values conform to other crosses where approximately 100 kb corresponds to 1 mu and indicate that fla18 is linked to the FLA10 gene.
To confirm that this mutation is the causative change in fla18, we isolated revertants of fla18 that regained the ability to swim at 32°C. Following UV mutagenesis, 46 independent swimming strains were isolated. A dCAPS marker that distinguishes the polymorphism in wild-type and fla18 was used to analyze 8 of the 46 revertants/suppressors. The restriction enzyme MboII cuts the wild-type PCR product, but not the fla18 PCR product. A representative gel of the PCR and digest is shown in Figure 1. Digestion by MboII indicates that these eight strains are likely to be true revertants. The restoration of the original codon was confirmed in these eight revertants by Sanger sequencing (Table 4). Thus, our revertant analysis indicates that mutation in the FLA10 kinesin is the causative mutation in fla18. To examine the temperature-sensitive phenotype of the fla18 mutant strain, it was shifted from the permissive temperature of 21°C to the restrictive temperature of 32°C, and the percentage of flagellated cells was determined. While wild-type cells remain approximately 80% flagellated, the fla18 cells lose their flagella gradually and at 7 hours most cells are aflagellate ( Figure 2A). The length of the flagella was measured and the length changed by only 2 μm ( Figure 2B). The cells may be losing their flagella by detachment rather than by shortening.

FLA24 encodes the cytoplasmic dynein heavy chain
The fla24 allele was mapped to chromosome 6 near the mating-type locus [31] and linkage to the mating-type locus was confirmed in an additional 243 tetrads (239:0:4; PD:NPD:T). There is only one change in fla24 in the mapped interval after subtracting changes found in other unrelated strains [53]. The candidate in the interval is DHC1b, the cytoplasmic dynein gene for retrograde IFT ( Table 2). The T to C mutation predicts a L 3242 P change. We used a PCR-based assay to examine linkage of the flagellar phenotype with the alteration in the cytoplasmic dynein gene. The PCR product generated is 303 bp long in both the fla24 and wild-type strains. The mutant product is cut by NciI but not cut by AlwN1, while the wild-type product is cut by AlwN1 but not by NciI. This change cosegregates with the flagellar assembly defect in 59 meiotic progeny. To ask if this change is responsible for the phenotype, we again used reversion/suppressor analysis. Following mutagenesis, 64 independent swimming strains were recovered. Nine of the strains are no longer cut by NciI. Of these, AlwN1 fails to cut three ( Figure 3, Table 5).  Table 3 Primers for mapping of fla18 in crosses with CC-1952 (S1C5) Markers without a restriction enzyme generate size polymorphisms due to CA repeats. All reactions were performed with an annealing temperature of 53°C.
Sanger sequencing verified that we recovered six true revertants that changed the proline at amino acid 3243 back to leucine, and three pseudorevertants that changed the proline to either serine (2) or to arginine (1).
Currently, there are five other mutant alleles identified in the cytoplasmic dynein gene in Chlamydomonas. Three show nonconditional phenotypes; they assemble very short flagella [9,10], and two temperature-sensitive alleles have been identified. The dhc1b ts allele assembles one-half length flagella (5.5 to 6 μm) at the permissive temperature of 18°C. Upon shifting cells to the restrictive temperature, the flagella shorten by about one-half in 2.5 hours and are very short by 24 hours [32]. The dhc1b-3 allele shows a very slow decrease in flagellar length and number upon shifting from 21°C to 34°C [27]. It takes nearly 4 days for the cells to shorten their flagella. We determined the number of cells with flagella in the fla24 allele ( Figure 2C). In contrast to the other conditional alleles, the fla24 allele shows much faster shortening and loss of flagella. By 2.5 hours, there is noticeable shortening, and by 6 hours the population is aflagellate ( Figure 2D). The true revertant and two pseudorevertants (P to S and P to R) remain flagellated at the restrictive temperature over the same time period as would be expected ( Figure 2C).
DHC1b and D1bLC are reduced in fla24 at the restrictive temperature DHC1b localizes to the basal body region and along the flagella in both wild-type and fla24 at 21°C. At this temperature, the signal intensities at the basal body and in the flagella are comparable between wild-type and fla24 cells ( Figure 4A). In the dhc1b-3 allele, the levels of DHC1b and the dynein light intermediate chain D1bLIC are reduced when cells were shifted to 34°C [27]. After the shift to 32°C, the wild-type cells maintain the same localization and similar signal intensity of DHC1b throughout the assay (5 hours). However, the intensity of DHC1b at the basal body and flagella in fla24 cells starts to show a reduction within 2 hours, before the majority of cells lost their flagella (Figures 2 and 4). The DHC1b signal remains detectable in the basal body area, but decreases with time at the restrictive temperature. By immunoblots, we showed that DHC1b is reduced by 18-fold in fla24 flagella compared to wildtype flagella ( Figure 4C). In contrast, the three revertants (D11, D12, and D42b) show similar signal intensities and locations to wild-type cells at both the permissive and restrictive temperatures ( Figure 4B and Additional file 2: Figure S1).
The localization of D1bLIC in wild-type cells appears similar to that of DHC1b. The signal intensities of D1bLIC remain high in wild-type cells at 21°C and for at least 5 hours after cells are shifted to 32°C ( Figure 5A). However, in fla24 cells, even though D1bLIC shows similar localization, the signal intensity is greatly reduced when compared to wild-type cells at 21°C ( Figure 5A). The signal shows further reduction as cells lose their flagella ( Figure 5A, 32°C, 2 to 5 hours). The level of D1bLIC  is reduced, as shown by immunoblots of flagellar proteins at 21°C by 16-fold ( Figure 4C). An immunoblot of D1bLIC using whole cell extract from wild-type and fla24 cells at 21°C indicates that similar to flagellar D1bLIC, the amount of D1bLIC in whole cell extract is reduced approximately 16-fold in the mutant ( Figure 5C). We also observed approximately 3-fold reduction when wildtype cells were switched from 21°C to 32°C for 5 hours ( Figure 5C). Consistent with the observation by immunofluorescence ( Figure 5A), we were unable to detect the D1bLIC signal by immunoblot after fla24 cells were shifted to 32°C for 5 hours ( Figure 5C). Similar to observation of DHC1b, the signal intensities of D1bLIC remain high in all three revertants ( Figure 5B and Additional file 3: Figure S2).
Localization of IFT81 is perturbed at permissive temperature in fla24 but not in fla18 At the permissive temperature of 21°C, Iomini et al.
found that fla18 IFT trains show a reduced anterograde velocity, and the fla24 IFT particles show a reduced retrograde velocity [11]. These mutant strains have    defects in the anterograde and retrograde motors, respectively. Therefore, we asked the whether the localization of IFT81, a complex B protein, is affected in these strains.
In the fla18 mutant strain at the permissive temperature, no change of localization or reduction of intensity of IFT81 was observed when compared to wild-type cells ( Figure 6). After cells were shifted to the restrictive temperature, the IFT81 signals remained in the basal body region and in the flagella until the cells became aflagellate at 6 hours. IFT81 appears unperturbed by the fla18 mutation.
In the fla24 mutant strain at the permissive temperature, the majority of the IFT81 protein is not localized at the basal body as observed in wild-type and the other mutants, but accumulates in the proximal ends (0.4 to 1.4 μm; n = 20) of flagella regardless of their flagellar length ( Figure 7A). This accumulation remains unchanged after cells were shifted to restrictive temperature (32°C) until flagella are lost. In the true revertant (D11) of fla24, IFT81 localizes to the basal body region and flagella at both temperatures, as observed in wild-type cells ( Figure 7B and Additional file 4: Figure S3). While the localizations of IFT81 in the two pseudorevertants (D12; P 3243 S and D42b; P 3243 R) at permissive temperature are identical to the pattern in wild-type cells, they show accumulation of IFT81 at the proximal ends of flagella after 6 hours and 4 hours at the restrictive temperature ( Figure 7B and Additional file 4: Figure S3), respectively. This suggests that the serine or arginine at the conserved leucine position does not completely restore function.

Discussion
Conditional mutants have been extremely useful in the study of essential genes in many cellular processes from ribosome assembly to cell division to secretion to synaptic vesicles. A collection of temperature-sensitive flagellar assembly mutants in Chlamydomonas has allowed the analysis of intraflagellar transport; they have documented that anterograde movement requires kinesin and IFT complex B, and that retrograde movement requires cytoplasmic dynein and IFT complex A. In addition, remodeling of IFT trains at the tip requires IFT172, a protein in complex B, and a temperature-sensitive mutation in this protein leads to a retrograde defect (Table 1). Of the 21 extant conditional mutants, previous work together with this work has identified the causative lesion in 12 of them. All of these genes encode either IFT components or motor proteins (Table 1). Six have mutations in one of the three kinesin motor genes and four other mutations are in the cytoplasmic dynein motor. This bias suggests that conditionality may be more easily achieved in the motor subunits than in the IFT components.
Whole genome sequencing in Chlamydomonas has been fruitful when the gene is mapped to a region or chromosome [30]. We have developed a collection of changes in other wild-type and mutant strains that can be used to eliminate non-causative candidates [53]. In addition, the transcriptional profiles during regeneration of flagella are useful for identifying candidate flagellar assembly genes [54]. For fla18, we narrowed the list of candidates from 892 to 85 (Table 2) by combining the data from whole genome sequencing and transcriptional profiles. If we had used this strategy for fla24, only three of the 62 genomewide candidates showed increased levels of transcript during regeneration. For fla9 (unpublished data at Cilia), this strategy would have narrowed the 78 genome-wide candidates to eight. The use of both data sets may help to obviate the need for genome-wide fine mapping of flagellar assembly mutants.
The mutation in fla18 affects the same amino acid that is mutated in the fla10-14 strain. In fla10-14, the glutamic acid is changed to lysine but in fla18 (now renamed fla10-16) it becomes a glycine. The two alleles have different kinetics of flagellar loss [8]; the E 24 K allele takes over 12 hours to see loss of 50% of the flagella compared to the E 24 G allele that takes only 6 hours to see complete loss ( Figure 2A). This glutamic acid is conserved in all kinesin-2 molecules across the ciliated phylogenetic tree (n = 75, data not shown). As speculated previously [8], it seems likely that this amino acid may interact with the P-loop and be important for motor activity. Interestingly, the fla2 mutant shows a 'fragile' phenotype [24]. Upon shifting cells to the restrictive temperature, the flagella detach rather than shorten. We observe similar detachment with the fla18 allele. Since fla18 greatly reduced anterograde IFT velocity, it is interesting to speculate that either a component is transported that maintains the integrity of the flagellar axoneme or a signal to maintain the integrity fails. This phenotype is allele-specific, which supports the idea that different alleles may provide different information about the functions of anterograde IFT.
The fla1 mutation (now fla8-2) and the fla8-1 mutation are in the second motor subunit of kinesin-2 [8]; they were thought to affect different phases of IFT [11]. The fla10-1 and fla8-1 alleles show similar phenotypes with normal anterograde velocity but a reduced ratio of anterograde to retrograde particles, while fla8-2 and fla18 show similar phenotypes with reduced anterograde fla24 is a mutation in the cytoplasmic dynein. The mutant fla24 allele has several helpful properties that will allow dissection of a successful IFT cycle. The IFT trains must be assembled at the basal bodies, turn around at the tip to change from anterograde to retrograde movement, and then be reloaded at the base for anterograde transport. In our study of fla15 (IFT144) and fla17 (IFT139) IFT complex A mutants, we observed that diploid cells heterozygous for fla24 and either fla15 or fla17 were aflagellate at 32°C but flagellated at 21°C [31]. It is not unexpected that defects in complex A might show an enhancement of a cytoplasmic dynein mutant phenotype.
Since fla24 is compromised by reduced mutant Complex B proteins, we considered that the IFT dynein function may also be sensitized to dynein inhibitors since the retrograde velocity is reduced to 0.9 μm/second from 3.1 μm/second for wild-type cells and the frequency of retrograde particles is reduced [11]. Ciliobrevin D is a small molecule that inhibits cytoplasmic dynein [55]. We asked if fla24 cells were more sensitive to ciliobrevin D than wild-type cells. Surprisingly, with the addition of 100 μM ciliobrevin D, fla24 cells showed no effect on flagellar length after 30 minutes (data not shown), although this concentration has been shown to reduce the retrograde particle frequency after 5 minutes [56]. Further experiments to examine IFT particle rates with the mutant and inhibitor will shed more light on synthetic interactions.
Upon the shift of fla24 cells to 32°C, the flagella are lost within 4 hours ( Figure 2C); this is quite rapid compared to the other DHC1B alleles. The amount of DHC1b in flagella is greatly reduced as observed by immunofluorescence and immunoblot (Figure 4). The fla24 cells have three interesting phenotypes at 21°C that may suggest roles for the AAA5 domain of cytoplasmic dynein. First, the retrograde velocity and number of particles are reduced [11]. Second, the level of light intermediate chain (D1bLIC) is reduced as indicated by immunofluorescence and immunoblots ( Figures 4C and 5). Third, the IFT81 protein distribution is dramatically altered; instead of localizing to the basal body and along the flagella, IFT81 has left the basal body region and is concentrated in the proximal approximate 1 μm of the flagella ( Figure 7A). Unlike the dhc1b-3 allele that shows a reversal in the direction of phototaxis [27], we never observed a change in the phototaxis phenotype of the fla24 mutant over a 6hour period (data not shown). Again, it is clear that different alleles have different phenotypic properties.
The cytoplasmic dynein molecule is composed of a central ATP-hydrolyzing ring that has six AAA modules arranged around the ring's central pore. The fla24 mutation falls into an alpha-helix in the AAA5 domain that is extended into the alpha helical strut/buttress [57,58] (Figure 8). The strut/ buttress is postulated to have a high degree of plasticity that may be important for its function in communicating between the microtubule-binding domain (MTBD) of the stalk and the AAA ring. The distal region of the strut interacts with the middle of the stalk and a deletion of the distal end of the strut removes this interaction. Allosteric communication between AAA1 and MTBD is postulated to be relayed through the C-sequence, the strut, and the stalk [57].
Two in-frame deletions of 6 and 7 amino acids in the cytoplasmic dynein of Neurospora crassa affect the strut [60]. The 3739 Δ6 amino acid deletion removes part of the first coil of the strut and is postulated to affect communication between the MTBD and the nucleotide status of the AAA1 domain. The dynein localizes distally in this mutant. The 3756 Δ7 amino acid deletion is in the first coil of the strut as well but causes aggregation of the dynein. It is postulated that this mutation may lock the structure of the dynein. The cytoplasmic dynein mutants in Neurospora also demonstrate that different alleles have very different phenotypic effects [52].
The reduced retrograde velocity of IFT trains [31], the reduction in D1bLIC at the basal bodies, and the accumulation of IFT81 in the fla24 mutant suggest that the mutant has a defect in moving along the microtubules and in remodeling the IFT trains at the proximal end of the flagella via cargo binding. In our screen for suppressor and revertants of fla24, we have identified at least three genes that suppress the restrictive temperature flagellar assembly defect. These strains may provide information about interactions with the dynein heavy chain that will speak to how the strut and other structures communicate flagellar assembly defect (data not shown).

Conclusions
Whole genome sequencing provides a fast and inexpensive means to identify chemically induced mutations in Chlamydomonas [30,53]. Identification of the remaining seven temperature-sensitive mutations will be greatly helped by this technique. The identification of multiple mutant alleles in kinesin and in cytoplasmic dynein that have different phenotypes will greatly help studies of the function of these proteins. The fla18 allele is unique among the motor mutants in that it appears to deflagellate after the temperature shift instead of shortening. The fla24 allele results in an abnormal localization of IFT81 near the basal bodies at the permissive temperature, which may suggest a defect in recycling IFT trains. Additional files : Table S1. Primers for reversion analysis of fla18 and fla24.