The versatile molecular complex component LC8 promotes several distinct steps of flagellar assembly

Abstract

LC8 is present in various molecular complexes. However, its role in these complexes remains unclear. We discovered that although LC8 is a subunit of the radial spoke (RS) complex in Chlamydomonas flagella, it was undetectable in the RS precursor that is converted into the mature RS at the tip of elongating axonemes. Interestingly, LC8 dimers bound in tandem to the N-terminal region of a spoke phosphoprotein, RS protein 3 (RSP3), that docks RSs to axonemes. LC8 enhanced the binding of RSP3 N-terminal fragments to purified axonemes. Likewise, the N-terminal fragments extracted from axonemes contained LC8 and putative spoke-docking proteins. Lastly, perturbations of RSP3's LC8-binding sites resulted in asynchronous flagella with hypophosphorylated RSP3 and defective associations between LC8, RSs, and axonemes. We propose that at the tip of flagella, an array of LC8 dimers binds to RSP3 in RS precursors, triggering phosphorylation, stalk base formation, and axoneme targeting. These multiple effects shed new light on fundamental questions about LC8-containing complexes and axoneme assembly.

Figure 1.

Sequence analysis of RSP3. (A) Sequence alignment of RSP3 orthologs. The TQT-like LC8-binding motifs (gray boxes) are adjacent to the axoneme-binding region (underlined) near the N terminus and upstream to the AH that binds the RIIa domain (black box at aa 161–178). As in many LC8-binding sites, in some of these sites in RSP3, T is replaced by aa residues with similar properties. Asterisks represent the predicted ERK1/2 phosphorylation sites. The secondary structure of the C. reinhardtii RSP3 sequence was predicted by the Jpred program. H.s., Homo sapiens; G.g., Gallus gallus (chicken); X.l., Xenopus laevis; C.i., Ciona intestinalis; C.r., Chlamydomonas reinhardtii; H, α helix; E, β strand; −, random coil. (B) Schematic depicting the region containing the TQT-like motifs in full-length Chlamydomonas RSP3. The tertiary structure was predicted using the Regional Order Neural Network Disorder Prediction tool. Arrows mark the position of the putative LC8-binding motifs. The white area represents <50% disorder; the black area represents >50% disorder. Axo, axoneme.

Figure 2.

Comigration of LC8 and the RSP3 N-terminal proteolytic fragments in electrophoresis. (A–C) Native gel electrophoresis (A), SDS-PAGE (B) and native/SDS-PAGE 2D electrophoresis (C) of the KI extract from pf28pf30 axonemes that were pretreated with the buffer only or with trypsin (Tryp.) at the ratio of 100:1 (+) or 100:2 (++) by weight. The blots were probed sequentially for LC8, the RSP3 N-terminal region, and RSP11, which contains an RIIa domain that binds to the AH adjacent to the last LC8-binding site. RS subcomplexes containing RSP3 N-terminal fragments were resolved into two particles in the native gel (A) and depicted as two lines in a gel strip schematic in C. Arrows indicate the directions of electrophoresis. The digested RSP3 N-terminal fragments migrated as ∼30- and 20-kD fragments in SDS-PAGE (B and C, asterisks). In the 2D gel, LC8 and RSP11 (lines) aligned vertically with the two RS subcomplexes. The LC8 spot that aligned with the 30-kD RSP3 fragment appeared more abundant than the LC8 aligned with the 20-kD RSP3 fragment; the two RSP11 spots appeared equally abundant. (D) Schematic depicting the predicted positions of the two proteolytic fragments near the RSP3 N terminus, relative to the binding sites for LC8 dimers (gray rectangles, 8₂), outer doublets, and the spokehead. The prediction was based on Western blots, the size of proteolytic fragments, and trypsin cleavage sequences. The actual cleavage sites were not mapped (indicated by a question mark).

Figure 3.

Copurification of recombinant LC8 and RSP3_1–160. (A and B) Ni-NTA (A) and S tag (B) affinity purification of the bacterial extracts containing His-tagged LC8 or S tag RSP3_1–160 or the mixture of both extracts. Protein samples were fractionated by SDS-PAGE and stained with Coomassie blue. Both fusion proteins in the extract mixtures were copurified by both matrices (arrowhead in eluate lane). The aberrant migration of the S tag eluate was a result of the high salt concentration in the elution buffer. Pre, bacterial extract; Post, flow through. Molecular mass is indicated in kilodaltons.

Figure 4.

LC8 enhances the binding of RSP3 N-terminal fragment to the spokeless pf14 axonemes. Bacterial extract containing RSP3_1–160, His-LC8, or both was incubated with different amounts (clear bar) of axonemes from pf14 or the WT strain. The axoneme (Axo.) pellets were fractionated by SDS-PAGE, and the Western blots were probed as indicated. pf14 axonemes pulled down more RSP3 fragments than WT axonemes (compare lanes 4 and 11, arrows), whereas addition of His-tagged LC8 greatly increased the amount of RSP3 bound by pf14 axonemes (compare lanes 3 and 4, arrows). A subunit of I2 and I3 inner dyneins, p28, was used to indicate the amounts of axonemes (clear bar). Endogenous untagged LC8 (arrowhead) was substantially less abundant in pf14 than WT.

Figure 5.

RSP3_1–170 in axonemes exists as homodimers. (A and B) RSP3 and HA Western blots of the axonemes from the strains expressing RSP3_1–170 that has a C-terminal tag containing three Cys residues (A) or three HA epitopes and 12 His residues (B). Monomeric RSP3_1–170 migrated as double bands (arrows). A portion of Cys-tagged RSP3_1–170 migrated as the polypeptides twice the size of monomers (A, asterisk) in the presence of 140 mM β-mercaptoethanol (β-ME). The putative dimer bands were absent when the reducing agent was threefold as concentrated. The dimer bands were not present in RSP3_1–170-HAHis axonemes (B, left) unless they were treated with the zero-length cross-linker EDC (asterisk in B, right). The amount of RSP3_1–170 in the axonemes was small compared with the control sample of axonemes from cells expressing HAHis-tagged RSP3_Δ171–244 lacking aa 171–244. Molecular mass is indicated in kilodaltons.

Figure 6.

The pull-down of RSP3_1–178 contains LC8, RSPs, and putative RS-docking proteins. (A and B) Western blots (A) and silver-stained protein gels of the Ni-NTA pull-down (B) from the KI axonemal extracts of pf14 and the pf14 transformant that expresses RSP3_1–178-HAHis. The RSPs in the RSP3_1–178 pull-down were LC8 and the RIIa domain–containing RSP7 (B, right) and RSP11 (A and B, right blots). An unknown 85-kD doublet (arrow) and CaM-IP2, -IP3, and -IP4 (dots) in the CSC were also in the pull-down but were scarce or undetectable in the pull-down from pf14. CaM and IC140 in the inner dynein arm I1 were not pulled down appreciably. CaM-IP4 was identified based on its copurification with CaM-IP2 and CaM-IP3 and its molecular mass. Pre, axonemal extract; Post, flow-through; E, eluate; Tub, tubulin.

Figure 7.

LC8 is undetectable in the RS precursors. The flagellar membrane matrix from pf28pf30, which lacks LC8-containing axonemal dyneins, was first fractionated through a sucrose gradient. The 20S and 12S fractions were subjected to further purification through a MonoQ column. The protein blots of RS-positive fractions from the anionic chromatography were probed for RSP3 and LC8. LC8 was detectable in the purified 20S complex but not in the 12S complex. RSP3 was used as a loading control. A fraction of the 12S RSP3 migrated faster (line) than the 20S RSP3.

Figure 8.

The mutants defective in three to five LC8-binding motifs in RSP3 exhibit severe motility deficiencies. (A–D) Representative recording of the liquid cultures of 1-5AAA, 3-5AAA, and 1-2AAA mutants. The videos were recorded at 200× magnification and 16 frames/s, and the motile cells were tracked using the MetaMorph program to reveal their trajectories. The tracking showed that the motile cells in the cultures of 3-5AAA and 1-5AAA strains moved slowly and locally with irregular trajectories. 1-2AAA cells swam in linear or helical trajectories as WT. Bars, 100 µm.

Figure 9.

The axonemes with 3-5AAA mutations are deficient in the RS, LC8, and RSP3 phosphorylation. (A) Representative Western blots of axonemes from two mutant strains of each group, as indicated. The three lanes with decreasing amounts of WT axonemes and the blot of IC140 in I1 dynein showed protein loads. All mutated RSP3 polypeptides could be assembled into the axonemes. The amounts of RSP3 and LC8 in 1-5AAA axonemes appeared less abundant than in the others (compare arrows). (B) Western blots of the NaCl extract and the subsequent KI extract of 3-5AAA axonemes. Contrary to WT control, a fraction of the RSs from the axonemes of transformants was extracted by 0.6 M NaCl (arrows, left) along with dyneins, represented by IC140. LC8 in the KI extracts that contained most RSP3 polypeptides were less abundant in the 3-5AAA samples (arrows, right) than the WT control. (A and B) Gradient gels were used to resolve all relevant proteins. (C) Western blots of the 6% SDS-PAGE revealed hypophosphorylated RSP3 in the axonemes of RSP3 and LC8 mutants. The tagged RSP3 defective in the last three LC8-binding motifs migrates as double bands (dots and line), the lower band (line) migrating faster than the tagged RSP3 in WT control strain and 1-2AAA strain (top). Likewise, a fraction of the untagged RSP3 polypeptides in pf27 and fla14-3 migrated faster than the untagged WT RSP3 (bottom). In addition, RS abundance was reduced in these two mutants (compared RSP3 and IC140 blots). To probe for phosphorylation, protein loads were adjusted to make each lane, except pf14, contain similar amounts of RSP3. The upper RSP3 band (dot) in the RSP3 mutant or the LC8 mutant fla14-3 was recognized by a p-Thr-Pro antibody, whereas the lower band (line) was not. In pf27, neither RSP3 band was recognized by the antibody.

Figure 10.

Models depicting the effects of LC8 on the RS complex. (A) Multiple LC8 dimers (blue rectangles, 8₂) associate with a dimeric RSP3 (black lines) at the region N terminal (N’) upstream of the AH. The sequences between the LC8-binding sites may loop out of the stack and become phosphorylated (red asterisks). (B) The dimeric RSP3 N terminus and the stack of LC8 form the basal part of the RS. Phosphorylation near the LC8-binding sites may alter local conformation, promoting the interactions of the nearby AH with homodimeric RIIa domains (R) and the interaction of the axonemal binding region with the spoke-docking complex (SDC), possibly the CSC, or a FAP206-containing complex. Only a fraction of the 9 + 2 axoneme and relevant RSPs are illustrated. CP, central pair. (C) The two-stage assembly process of the RS complex in the cell body (left) and flagellum (right). LC8 dimer and HSP40 dimer do not join the RS until the L-shaped 12S RS precursors are transported into flagella by kinesin-driven IFT. Association of HSP40 with the 12S RS transforms the spokehead, whereas binding of LC8 to the N-terminal region of RSP3 in the 12S RS triggers formation of the stalk base, phosphorylation (Phos.) of RSP3 (asterisks), and docking of the T-shaped 20S RS to the axoneme. The unbound 20S RSs are delivered back to the cell body by dynein-driven IFT. The unknown transport mechanisms of LC8 and HSP40 are not depicted.