A molecular ruler determines the repeat length in eukaryotic cilia and flagella

See allHide authors and affiliations

Science  14 Nov 2014:
Vol. 346, Issue 6211, pp. 857-860
DOI: 10.1126/science.1260214


Existence of cellular structures with specific size raises a fundamental question in biology: How do cells measure length? One conceptual answer to this question is by a molecular ruler, but examples of such rulers in eukaryotes are lacking. In this work, we identified a molecular ruler in eukaryotic cilia and flagella. Using cryo-electron tomography, we found that FAP59 and FAP172 form a 96–nanometer (nm)–long complex in Chlamydomonas flagella and that the absence of the complex disrupted 96-nm repeats of axonemes. Furthermore, lengthening of the FAP59/172 complex by domain duplication resulted in extension of the repeats up to 128 nm, as well as duplication of specific axonemal components. Thus, the FAP59/172 complex is the molecular ruler that determines the 96-nm repeat length and arrangements of components in cilia and flagella.

Molecular ruler rules cilia and flagella length

Cilia and flagella contain within their ultrastructure repeating structures at regularly spaced intervals. How does the cell measure length with nanometer precision? Oda et al. identify a flagella protein complex in Chlamydomonas that appears to act as a sort of molecula ruler to define repeat length. Genetic changes that would change the length of this protein led to corresponding changes in the length of repeats within the resulting flagella.

Science, this issue p. 857

Cilia and flagella are motile organelles that play crucial roles in the generation of fluid flow by beating motion and in the development of mammals (1, 2). Their doublet structure has 96-nm repeats, which are very accurately specified. One explanation for this precise length control is that a molecular ruler exists (3). In bacteria and bacteriophages, ruler molecules have been shown to determine the length of the injectisome needles and the tail structures (4, 5). In eukaryotes, however, titins and nebulins have been proposed to be the rulers for muscle (6, 7), but this has been controversial (810).

To investigate the mechanism of ciliary and flagellar assembly, we focused on two Chlamydomonas proteins, FAP59 and FAP172. A lack of these proteins results in axonemal disorganization in various mammals and zebrafish (1114). We first isolated a FAP172-missing mutant (strain 5-8) by screening Chlamydomonas motility mutants that lack FAP59 and FAP172 (table S1 and fig. S1, A and B; also see supplementary materials and methods). The mutant produced short and immotile flagella. Another four mutants that exhibit similar phenotypes to strain 5-8 were identified, including FAP172-deficient (pf7) or FAP59-deficient (pf8, HA24, and gam5) mutants (Fig. 1A, table S1, and fig. S1, B and C). Hereafter, we labeled pf8 as fap59 and pf7 as fap172. Transformation of hemagglutinin (HA)–tagged genes to fap59 and fap172 mutants recovered wild-type (WT) motility and flagella length, confirming the direct relation between mutations in the two genes and the phenotypes (Fig. 1B, table S1, and fig. S1D). In fap59 and fap172 flagella, inner dynein arm (IDA) and nexin-dynein regulatory complex (N-DRC) components were absent or severely reduced (Fig. 1C). On the other hand, FAP59 and FAP172 were retained by mutant strains deficient in major axonemal components (Fig. 1D and table S2). Biochemical characterization revealed that FAP59 and FAP172 are mainly localized in flagella and released by treatment with 0.2 to ~0.4 M NaI and that FAP59 protein is phosphorylated in flagella (fig. S1, E to G) (14, 15).

Fig. 1 Biochemical characterization of fap59 and fap172 mutants.

(A) Differential interference contrast (DIC) images of Chlamydomonas cells. Expression of HA-tagged FAP59 and FAP172 rescued fap59 and fap172 mutants, respectively, is shown. (B) Immunofluorescence of WT and rescued nucleoflagellar apparatuses. Immunostaining of HA-tagged FAP59 and FAP172 indicates that the two proteins are localized along the length of the flagella. (C) Immunoblots of axonemal proteins from WT, fap59, and fap172 cells. Flagellar localizations of FAP59 and FAP172 are dependent on each other. Components of the IDA (p28 and IC140) and N-DRC (DRC2 and DRC4) are missing or severely reduced in fap59 and fap172 mutants. Tektin, a DMT-associated protein, is also missing in the mutants. IC2, an intermediate chain of the ODAs, and RSP1, a component of the RSs, appear unchanged. Components of the DMT (Rib72 and FAP20) are not reduced, either. (D) Immunoblot analysis of axonemal proteins. Except for fap59 and fap172, all of the mutants retain FAP59 and FAP172 proteins. See table S2 for descriptions of the mutant strains used in this study. CBB, Coomassie Brilliant Blue. (E) Pull-down assay. BCCP tags were added to the C termini of FAP59 or FAP172, and the BCCP-tagged proteins were pulled-down by streptavidin beads. FAP172 was copurified with BCCP-tagged FAP59 and vice versa.

Next we examined whether FAP59 and FAP172 form a complex because the flagellar localization of each protein depends on the other (Fig. 1C) (13). We labeled the two proteins with biotin carboxyl carrier protein (BCCP) tags and performed pull-down assays using streptavidin beads (Fig. 1E). BCCP-tagged FAP59 pulled down FAP172 and vice versa, showing that FAP59 and FAP172 form a complex in flagella.

To examine the structural defects caused by a loss of the FAP59/172 complex, we used cryo-electron tomography (cryo-ET) to observe the mutant axonemes. In WT axonemes, the doublet microtubule (DMT) components are arranged with 96-nm periodicity, as represented by radial spokes (RSs) (Fig. 2, A and B) (16, 17). In fap59 and fap172, however, RSs were attached to DMTs with an irregular periodicity of ~32 nm instead of 96 nm (Fig. 2, C and D, and fig. S2, A and B). In agreement with our immunoblot analyses (Fig. 1C) and previous reports (12, 13), IDAs and the N-DRC were missing in the mutant axonemes (fig. S2C and fig. S3, A, B, and G to L). Thus, the FAP59/172 complex is essential for the establishment of 96-nm repeats in DMTs. The 24-nm repeats of outer dynein arms (ODAs) were unaltered in these mutants (fig. S3, C to E), suggesting that the arrangement of ODAs does not depend on the FAP59/172 complex. Therefore, we hypothesized that the FAP59/172 complex serves as a molecular ruler for the 96-nm repeats of DMTs.

Fig. 2 The 96-nm periodicity of radial spokes was lost in fap59 and fap172 axonemes.

(A) The 96-nm repeating units of DMTs in a Chlamydomonas WT axoneme. Each 96-nm repeat has two RSs (RS1 and RS2), seven IDAs (a to e, , , g, and an IC-LC complex of f), four ODAs, and one N-DRC. (B to E) Slices of tomograms showing positions of the RSs (white arrowheads). (B) The WT axoneme shows the regular 96-nm periodicity of the RSs. Distal ends are to the left. CP, central pair microtubules. (C and D) Both fap59 and fap172 axonemes show irregular alignment of the RSs with ~32-nm gaps. (E and F) sup-pf-5 axonemes show a normal 96-nm repeat, in spite of tektin protein deficiency. (F) Immunoblot analysis of N-DRC mutant axonemes. An absence of tektin proteins in the sup-pf-5 mutant and a reduction of tektin protein in the pf3 and ida6 mutants were observed. WB, Western blot.

To test an alternative hypothesis that tektin is a 96-nm molecular ruler (18, 19), we also examined a N-DRC mutant, sup-pf-5 (20, 21). Tektin is missing in sup-pf-5 flagella (Fig. 2F), but its axonemes retain the 96-nm periodicity of RSs (Fig. 2E). Thus, tektin does not define the 96-nm repeats of DMTs.

To examine the possible role of the FAP59/172 complex as a molecular ruler, we inserted BCCP tags into the two proteins at different sites and determined their three-dimensional (3D) positions on DMTs using cryo-ET and an enhanced streptavidin-labeling method (Fig. 3 and fig. S4) (22, 23). The labels on FAP59 and FAP172 were similarly positioned along the protofilament(s), and each label appeared once per 96-nm repeat (Fig. 3, B and C). The labels in the middle segment were located approximately halfway between the N and C termini. Thus, both FAP59 and FAP172 form 96-nm-long structures and lie along the long axes of DMTs, consistent with the idea that the FAP59/172 complex is a 96-nm molecular ruler.

Fig. 3 Both FAP59 and FAP172 take on a 96-nm-long extended conformation.

(A) Coiled-coil structure prediction of FAP59 and FAP172 using the CoilScan program (24). The predicted coiled-coil domains were roughly divided into three parts. (B and C) FAP59 and FAP172 were structurally tagged, and their positions were visualized by comparing the averaged subtomograms of the WT DMT (gray) with those of the labeled DMTs. The positions of tags (indicated by stars) are shown above each tomogram. Views on the right were rotated 90° around the vertical axis. (B) BCCP tags were added to the N terminus (red, after Met22) and the C terminus (yellow, before the stop codon) or inserted into the middle (orange, after Asp420) of FAP59. (C) BCCP tags were added to the N terminus (green, before Met1) and the C terminus (blue, before the stop codon) or inserted into the middle (light blue, after Leu381) of FAP172. Distal ends are to the left. The orientation of the proteins relative to the DMTs was determined in combination with the next experiment (Fig. 4).

To test whether the FAP59/172 complex defines the 96-nm DMT repeats, we expressed longer versions of the complex in fap59 fap172 double mutants and used cryo-ET to observe their axonemal ultrastructures (Fig. 4). The FAP59/172 complex was lengthened by duplicating the coiled-coil domains at the N terminal, at the middle, or near the C-terminal part. The short flagella phenotype of the double mutant was rescued by expressing the elongated proteins (fig. S5, A to E). Cryo-ET of the axonemes of the rescued strains—named FAP59/172-EL1, -EL2, and -EL3—revealed that the repeat length was extended to ~128 nm (FAP59/172-EL1 and -EL3) or ~120 nm (FAP59/172-EL2) (Fig. 4, C to H). Extension of the repeats by 32 nm in FAP59/172-EL1 and -EL3, and by 24 nm in FAP59/172-EL2, is approximately proportional to the number of amino acids duplicated in each mutant (~1.3 Å per residue).

Fig. 4 Generation of artificial longer repeats by lengthening the FAP59/172 sequences.

(A) Slices of a tomogram showing the 96-nm periodicity of the RSs and (B) 3D visualization of averaged 96-nm repeats from WT axonemes. The coiled-coil domains used for elongation of FAP59 and FAP172 are colored (lower right). Arrowheads indicate the positions of the RSs. (C and D) In the FAP59/172-EL1, the N-terminal coiled-coil domain was duplicated. (C) In the FAP59/172-EL1 axoneme, the RSs showed a 128-nm periodicity, and three RSs were present in one repeat. (D) 3D structure of the averaged 128-nm repeats from FAP59/172-EL1 DMTs. (E and F) In FAP59/172-EL2, the middle coiled-coil domain was duplicated. (E) FAP59/172-EL2 DMTs show 120-nm repeats and have two RSs in one repeat. (G and H) In the FAP59/172-EL3, the near–C-terminal coiled-coil domain was duplicated. (C and E) Normal 96-nm repeats with two RSs and corrupted repeats appear occasionally in EL axonemes (red asterisks) (see table S3). (D and H) Although arrangements of ODAs are normal in all of the EL mutants (fig. S4C), shapes of ODAs appear disordered in FAP59/172-EL1 and -EL3. This is probably because there is a phase mismatch between ODA’s 24-nm periodicity and RS’s 128-nm periodicity.

Elongation of the repeats was accompanied by duplication of specific structures in the flagella. In FAP59/172-EL1, the IDA f (Fig. 4D, deep blue), IDA a (blue), the intermediate chain–light chain (IC-LC) complex (light blue), and RS1 (RS1’) were duplicated. In FAP59/172-EL2, the gap between IDAs f and d was widened (Fig. 4F, square brackets), and weak densities of additional IDA-like structures (purple) were observed. In FAP59/172-EL3, the single-headed IDAs c and e (Fig. 4H, green), N-DRC (orange), and RS2 (RS2’) were duplicated. These results strongly support the idea that each domain of the FAP59/FAP172 complex recruits specific axonemal components to the DMTs.

Taken together, our results indicate that the FAP59/FAP172 complex uses very intricate mechanisms to construct the 96-nm repeat of DMT, rather than simply serving as a physical ruler. We propose “positive” and “negative” scaffolding mechanisms for the correct alignment of RSs, IDAs, and N-DRCs (fig. S6). In the absence of the FAP59/172 complex, an excessive number of RSs bind along the specific protofilament(s) of the A-tubule, indicating that binding of the RSs to the A-tubule does not depend on the FAP59/172 complex. Therefore, the FAP59/172 complex can be seen as a kind of negative regulator for RSs, as the complex masks most of the RS-binding regions. At the same time, the complex leaves “holes” that allow for binding of RSs in appropriate locations. In contrast, the absence of the FAP59/172 complex causes loss of IDAs and the N-DRC, suggesting that the complex works as a positive regulator and provides anchoring sites for IDAs and N-DRCs.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S6

Tables S1 to S3

References (2559)

References and Notes

  1. Acknowledgments: We thank M. Hirono (University of Tokyo) and T. Yagi (Prefectural University of Hiroshima) for technical assistance and valuable advice. This work was supported by CREST, Japan Science and Technology Agency (to M.K.), the Takeda Science Foundation (to M.K. and T.O.), the Kazato Research Foundation (to T.O.), and Japan Society for the Promotion of Science KAKENHI grant 20770119 (to H.Y.). We declare no competing financial interests. The electron microscopy maps of averaged DMT are available in the EM Data Bank ( under accession numbers EMD-6108 to EMD-6117.
View Abstract

Stay Connected to Science

Navigate This Article