The complex of outer-arm dynein light chain-1 and the microtubule-binding domain of the γ heavy chain shows how axonemal dynein tunes ciliary beating

Axonemal dynein is a microtubule-based molecular motor that drives ciliary/flagellar beating in eukaryotes. In axonemal dynein, the outer-arm dynein (OAD) complex, which comprises three heavy chains (α, β, and γ), produces the main driving force for ciliary/flagellar motility. It has recently been shown that axonemal dynein light chain-1 (LC1) binds to the microtubule-binding domain (MTBD) of OADγ, leading to a decrease in its microtubule-binding affinity. However, it remains unclear how LC1 interacts with the MTBD and controls the microtubule-binding affinity of OADγ. Here, we have used X-ray crystallography and pulldown assays to examine the interaction between LC1 and the MTBD, identifying two important sites of interaction in the MTBD. Solving the LC1-MTBD complex from Chlamydomonas reinhardtii at 1.7 Å resolution, we observed that one site is located in the H5 helix and that the other is located in the flap region that is unique to some axonemal dynein MTBDs. Mutational analysis of key residues in these sites indicated that the H5 helix is the main LC1-binding site. We modeled the ternary structure of the LC1-MTBD complex bound to microtubules based on the known dynein-microtubule complex. This enabled us to propose a structural basis for both formations of the ternary LC1-MTBD-microtubule complex and LC1-mediated tuning of MTBD binding to the microtubule, suggesting a molecular model for how axonemal dynein senses the curvature of the axoneme and tunes ciliary/flagellar beating.

Mutational studies have shown that expression of an LC1 mutant leads to dominant-negative effects on swimming velocity and beat frequency in  Table 1). The asymmetric unit contains one molecule each of LC1 and the MTBD, suggesting that LC1 is bound to the MTBD with a stoichiometric ratio of 1:1, as predicted previously (17). The MTBD interacted with the hydrophobic core of LC1 mainly via two regions, the H5 helix and the flap region (Fig. 1B).
Next, we compared published NMR and X-ray structures (PDB ID: 1M9L and 5YXM) of LC1 alone (22, 23) to the newly obtained MTBD-complexed LC1 structure ( Fig. 2A). The RMSD of Cα atoms between the new structure and the NMR or X-ray structure was 2.92 Å (134 aa) or 0.48 Å (194 aa), respectively, suggesting that the structure of LC1 in the MTBD complex is more similar to the X-ray structure than to the NMR structure. This is understandable because the X-ray structure is likely to mimic the complexed conformation due to overlap between interaction sites with the next crystallographic molecule and interaction sites with the MTBD (76.9% residues of shared interaction sites, Fig. S1).
We also compared the structure of the MTBD in the LC1-MTBD complex to structures of the MTBD from Chlamydomonas dynein-c (PDB ID: 2RR7), mouse cytoplasmic dynein 1 (PDB ID: 5AYH) and human cytoplasmic dynein 1 (PDB ID: 3J1T) (Fig. 2B). Surprisingly, the conformations of the axonemal dynein-specific flap region differed significantly between the LC1complexed MTBD and the single MTBD of dynein-c. In the LC1-MTBD structure, the flap was kinked approximately 90° around the axis parallel to the stalk as compared with that of dynein-c, indicating that the new conformation enables the flap to interact with LC1. We were also able to identify the structural state of the MTBD in terms of its affinity for microtubules, because the registry of the helical packing of stalk is exchangeable and linked to the microtubulebinding affinity. The RMSD of Cα atoms between the LC1-complexed MTBD and the MTBD of dynein-c or cytoplasmic dynein 1 (β-registry) was 2.13 Å (117 aa) or 1.74 Å (118 aa), respectively (11,12). By contrast, between the complex and the MTBD of human cytoplasmic dynein 1 (αregistry) was 3.37 Å (111 aa) (21), suggesting that the LC1-complexed MTBD structure is in the β- The MTBD binds to LC1 through the H5 helix and flap region with several hydrophobic and hydrophilic interactions (Fig. 3A, B, and Table S1).
Complex formation involves a total of 41 residues (21 in LC1 and 20 in the MTBD) using a wide range of molecular surfaces. We selected 10 residues of LC1 for mutational analyses to assess the effect of electrostatic interaction (His31 and Our study indicates that LC1 is not able to bind directly to the microtubule track in its current binding geometry. Here it should be noted that LC1 is also able to interact with the MTBD stably even in α-registry. When we screened suitable constructs for crystallization, we prepared several variants with different coiled-coil lengths, and one of which fixed the α-

Structure determination
The diffraction images were processed by using HKL2000 software (28). The structure was solved by molecular replacement with Phaser-MR in the Phenix program suite (29) using the LC1 structure (PDB ID: 5YXM) as a starting model (23).
Structure refinement was performed by using phenix.refine in the Phenix program suite (29). The final structure was validated using MolProbity (30).
The crystallographic data and refinement statistics are summarized in Table 1.   c. R free is the R-factor computed for the test set of reflections that were omitted from the refinement process.