Fig 1.
LF4A regulates cilia length and number in Tetrahymena.
(A) A neighbor joining phylogenetic tree of a subset of CMGC kinases. The human MAPK3 was used as an outgroup. The numbers on the branches represent bootstrap support values above 50%. Arrows mark the two LF4/MOK homologs of Tetrahymena. (B and C) A wild-type (B) and an LF4A-KO cell (C) stained with the anti-polyG tubulin antibodies to visualize cilia (green), anti-centrin antibody 20H5 to mark the basal bodies (red) and DAPI (blue). Arrowheads indicate the posterior-dorsal region that is less densely ciliated in LF4A-KO. Abbreviations: oa, oral apparatus; A, anterior cell end; P, posterior cell end. (D) A box and whisker plot of locomotory cilia length. Cilia from ten wildtype and ten LF4A-KO cells were measured. The average cilium length for the wild type (n = 216) is 5.88 μm, SD (standard deviation): 0.38 μm; LF4A-KO (n = 367) 7.74 μm, SD: 1.82 μm. p < 0.01. (E) The length of cilia during regeneration after deciliation by pH shock. Four to 6 cells, 50–150 cilia were measured at each time point. Asterisks (*) indicate significant differences (p < 0.01 at each indicated time point).
Fig 2.
LF4A-GFP localizes to the basal bodies and along cilia, is transported by IFT and affects IFT.
(A-B) A wild-type (negative control non-GFP expressing) cell (A) and a cell expressing LF4A-GFP under the native promoter (B) were subjected to simultaneous fixation/permeabilization (using a mixture of paraformaldehyde and Triton X-100) and stained with the anti-GFP antibodies (red) and DAPI (blue). In the negative control, the red background (anti-GFP) signal is in the cell body and occasionally at the tips of cilia (A inset). Specific LF4A-GFP signal is present near the basal bodies and along the shafts of locomotory cilia (B inset compare to A inset). (C) An LF4A-GFP expressing cell that was first permeabilized (with Triton X-100) and then fixed (with paraformaldehyde) prior to immunofluorescence. Compared to (B), the LF4A-GFP signal remained strong near the basal bodies but is decreased in cilia. (D) A TIRFM image of a live LF4A-GFP-expressing cell. Arrowheads indicate several basal bodies within a locomotory row, the box highlights an example of a ciliary shaft. (E) Five representative examples of kymographs obtained from the TIRFM videos of LF4A-GFP cells. In most cilia, LF4A-GFP particles are stationary and some move with velocities similar to IFT velocities (arrows) or diffuse (an example of a diffusion event is marked with a red line on the duplicate of the bottom right kymograph). (F) The IFT velocities (left) and IFT event frequencies (right) in cilia of either wild-type or LF4A-KO cells that express DYF1-GFP, an IFT reporter. (G) The IFT velocities (left) and event frequencies (right) in cilia of either wild-type or GFP-LF4A-overexpressing cells. Both strains were exposed to Cd2+ for 3 hours. In (F) and (G), the sample sizes are indicated, asterisks mark significant differences (one way Anova p<0.01), error bars represent SDs. (H) Examples of kymographs of GFP-DYF1 in cilia of either wild-type or LF4A-KO cells corresponding to the data shown in panel F. (H’) Examples of kymographs of GFP-DYF1 in either wild-type or GFP-LF4A overproducing cells corresponding to the data shown in panel G. Abbreviations: oa, oral apparatus; PFA, paraformaldehyde.
Fig 3.
Excessive LF4A kinase activity shortens cilia and decreases IFT.
(A-C) Cells (treated with 2.5 μg/ml CdCl2 for 6 hours) overexpressing either GFP (A), GFP-LF4A (B) or GFP-LF4AF82A (C) showing GFP fluorescence (green) and labeled by the anti-polyG tubulin antibodies to visualize cilia (red). The shortening of cilia is evident only in the cell expressing GFP-LF4A (B) where it is enriched at the basal bodies (asterisk). GFP-LF4AF82A accumulates at the tips of cilia (C). The insets in B-C show the green signal of GFP-LF4A alone at higher magnification. (D) GFP-LF4A-overexpressing cells after overnight induction with Cd2+, have short cilia, are large and irregular in shape, consistent with cytokinesis defects. (E) The results of an in vitro kinase assay with either GFP-LF4A or GFP-LF4F82A. Each protein was overexpressed in Tetrahymena and purified on anti-GFP-beads. The beads were incubated with the recombinant myelin basic protein (MBP) and ATP-γ-S. The phosphorylated products were detected on a western blot probed with the anti-thiophosphorylation antibody 51–8; the upper and middle panels show sections of the same blot containing autophosphorylated GFP-LF4A and MBP, respectively. The same amounts of IP inputs were analyzed on a separate western blot probed the anti-GFP antibodies and shown in the bottom panel (WB: GFP); the lower band in the right lane likely is a proteolytic degradation product of GFP-LF4A. (F) The lengths of locomotory cilia of cells that overproduce (for 3 hours) either GFP, GFP-LF4A or the kinase-weak GFP-LF4AF82A. The cilia lengths in the GFP and GFP-LF4AF82A cells were not significantly different (p > 0.01), while the GFP-LF4A cilia were significantly reduced (~75% of the length of GFP controls, one way Anova test, p < 0.01). Sample sizes are indicated, error bars represent SDs. (G) The anterograde and retrograde IFT velocities and frequencies in cilia of cells that express IFT140-GFP and overexpress either mCherry-LF4A or mCherry-LF4AF82A. The IFT speeds are significantly reduced in mCherry-LF4A as compared to mCherry-LF4AF82A overexpressing cells (exposed to added Cd2+ for 3 hours). The sample sizes (numbers of tracks measured) are indicated, asterisks indicate statistically significant differences (one way Anova, p<0.01), error bars represent SDs. (H) Examples of kymographs of cilia in cells expressing IFT140-GFP in different genetic backgrounds and conditions corresponding to the data shown in panel G. Scale bar: 1 μm x 1 s.
Fig 4.
Isolation of intragenic and extragenic suppressors of GFP-LF4A overexpression.
(A-C) The pipeline used for identification of intragenic and extragenic suppressors of overexpression of GFP-LF4A. (A) The structure of the transgene that was placed in the micronucleus. The transgene uses MTT1 promoter to express GFP-LF4A. A neo5 cassette is closely linked. The transgene replaces the endogenous LF4A. (B) An outline of the procedure for generating suppressors. A heterokaryon with the ovGFP-lf4a transgene in the micronucleus (solid black) was subjected to mutagenesis and a self-fertilizing cross. The homozygous progeny was selected based on paromomycin resistance conferred by neo5 and some progeny clones may carry a suppressor mutation (red stripes). The progeny cells were treated with Cd2+ in tubes kept in vertical position. The progeny clones that lack a suppressor mutation shorten cilia and sink to the tube bottom. The suppressors (F0) remain motile and accumulate near the top of the tube due to negative gravitaxis. (C) The principle of testing whether the suppression is intra- or extragenic (for details see S3 Fig). The F1 clones were subjected to a self-cross and the pm-r F2 progeny clones were isolated. An intragenic suppressor gives pm-r F2 progeny clones that are 100% motile (suppressed), as the suppression is linked to the transgene. An extragenic suppressor generates F2 pm-r clones that are either suppressed (motile) or not (paralyzed). (D-F) Cells that were exposed for 6 hours to Cd2+ to induce GFP-LF4A, subjected to immunofluorescence to reveal GFP (green) and polyG tubulin (red). Insets show the GFP-LF4A signal alone in examples of cilia at a higher magnification. (D) A non-mutagenized non-suppressed cell; GFP-LF4A is enriched at the bases of cilia. (E) An intragenic suppressor cell (SUP5); GFP-LF4A is enriched at the tips of both oral and locomotory cilia. (F) An extragenic suppressor cell (SUP1); GFP-LF4A is prominent at the bases, at the distal ends and along the ciliary shafts in short (presumably assembling) cilia. (G) The locomotory cilia length of F2 clones of four genotypes without and with 6 hours Cd2+ treatment. Sample sizes (number of cilia measured) are indicated, asterisks indicate a statistically significant difference (one way Anova p<0.01), error bars indicate SDs. (H) A 3D predicted structure of the kinase domain of LF4A based on homology-directed modeling using CDK of Cryptosporidium (Chain A of PDB 3NIZ) as template. (I) A zoomed-in view of a region of the structure showing three locations of substitutions (shown as sticks) found in the intragenic suppressors sup3 (E132K), sup4 (G13S) and sup5 (E160K).
Fig 5.
The extragenic suppressor clone SUP1 has a mutation in CDKR1.
(A-B) Immunofluorescence images of (unsuppressed and suppressed) pools of F2 progeny derived from a single sup1/SUP1+ (after an overnight exposure to Cd2+) that were subsequently subjected to whole genome sequencing. Note the patterns of GFP-LF4A inside cilia: the base accumulation in the non-suppressed pool and the ciliary shaft and tip signals in the suppressed pool (see insets for higher magnifications). (C) The results of variant subtraction and filtering based on the alignment of sequencing reads to the macronuclear reference genome. Among the 278 variants consistent with nitrosoguanidine (MNNG) mutagenesis, three candidate variants affect a gene product and have a high fraction of reads supporting the alternative base in the mutant pool (see S1 Table). (D) An IGV browser view of the macronuclear genome sequence of TTHERM_01080590 (CDKR1) that contains the variant scf_8254401:105680 T to C. This point mutation, supported by 100% of the sequencing reads from the mutant pool, changes the stop codon and adds a short peptide to the C-terminus of the predicted product. (E) An allelic composition contrast analysis of the variant co-segregation across all micronuclear chromosomes. The normalized linkage scores show the difference in the allelic composition between the mutant and the wild-type pool at each variant site. This reveals a cluster of variant co-segregation at 9–10 Mb on the micronuclear chromosome 3.
Fig 6.
A loss of CDKR1 lengthens both the locomotory and oral cilia.
(A-D’) Wildtype (A and A’), LF4A-KO (B and B’), CDKR1-KO (C and C’) and double knockout LF4A-KO_CDKR1-KO (D and D’) cells that are either in interphase (top panels) or dividing (bottom panels). The cells were stained with the anti-polyG tubulin (green) and anti-centrin antibodies (red). (E) The locomotory cilia lengths in different backgrounds. The cilia length in LF4A-KO, CDKR1-KO and the double knockout strain are similar (p > 0.01) and all are significantly longer than the wild-type cilia (one way Anova, p < 0.01). Sample sizes are indicated, error bars indicate SDs.
Fig 7.
A complete loss of CDKR1 rescues the cilia shortening induced by GFP-LF4A overexpression and reduces the LF4A kinase activity.
GFP-overproducing cells that have either wild-type CDKR1 (A, C, E) or are CDKR1-KO (B, D, F), imaged before (A and B) or after a 6 hours exposure to Cd2+ (C, D, E and F). In A-D, the cells were stained with the anti-polyG antibodies (green) and anti-centrin antibodies (red). In (E) and (F), the cells show a GFP-LF4A signal (green) and are stained with anti-polyG antibodies (red). In cells lacking CDKR1, overproduced GFP-LF4A accumulated at the distal ends of cilia (F), indicating a reduced LF4A kinase activity. (G) The growth rates of multiple strains that overproduce GFP-LF4A and are either wild-type or CDKR1-KO. The cells were inoculated in SPPA media without Cd2+ (each data point averages 6 cultures), and an overexpression of GFP-LF4A was induced at 12 hours (each data point averages 3 cultures hereafter). (H) An in vitro kinase activity of overproduced GFP-LF4A isolated from two strains that are either wild-type (CDKR1+) or CDKR1-KO. The phosphorylated products were detected on a western blot probed with the anti-thiophosphorylation antibody 51–8 (the upper and middle panels show areas of the same blot containing the autophosphorylated GFP-LF4A and MBP, respectively). The same amounts of IP inputs were analyzed on a western blot probed with the anti-GFP antibodies shown in the bottom panel (WB: GFP) the multiple bands likely are proteolytic degradation products of GFP-LF4A. (I) Top: a graphical summary of the phenotypes and genotypes. Bottom: a scheme of the likely pathway involving CDKR1, LF4A and another kinase, likely an RCK that acts downstream of CDKR1 in oral cilia.