Structural organization of the intermediate and light chain complex of Chlamydomonas ciliary I1 dynein

Abstract

Axonemal I1 dynein (dynein f) is the largest inner dynein arm in cilia and a key regulator of ciliary beating. It consists of two dynein heavy chains, and an intermediate chain/light chain (ICLC) complex. However, the structural organization of the nine ICLC subunits remains largely unknown. Here, we used biochemical and genetic approaches, and cryo-electron tomography imaging in Chlamydomonas to dissect the molecular architecture of the I1 dynein ICLC complex. Using a strain expressing SNAP-tagged IC140, tomography revealed the location of the IC140 N-terminus at the proximal apex of the ICLC structure. Mass spectrometry of a tctex2b mutant showed that TCTEX2B dynein light chain is required for the stable assembly of TCTEX1 and inner dynein arm interacting proteins IC97 and FAP120. The structural defects observed in tctex2b located these 4 subunits in the center and bottom regions of the ICLC structure, which overlaps with the location of the IC138 regulatory subcomplex, which contains IC138, IC97, FAP120 and LC7b. These results reveal the three-dimensional organization of the native ICLC complex and indicate potential protein-protein interactions that are involved in the pathway by which I1 regulates ciliary motility.

1. INTRODUCTION

Cilia and flagella are highly conserved organelles in organisms ranging from unicellular protists to humans.¹ They are essential for various biological processes involving cellular motility and signal transduction. In humans, defects in cilia can result in severe diseases collectively termed ciliopathies.^2,3 The microtubule-based core structure of motile cilia, the axoneme, has a “9+2” arrangement, e.g., nine peripheral doublet microtubules (DMTs) surround two singlet microtubules in the center known as the central pair complex (CPC) (Figure 1A). Two rows of axonemal dyneins, the outer and inner dynein arms (ODA and IDA), are arranged along the length of the A-tubule of the DMTs, and motility of the cilia is driven by ATP hydrolysis and force generation of these dynein motors.^4,5 Longitudinally, each DMT is built from 96-nm repeat units that contain four homogenous ODAs and seven heterogeneous IDAs that are distinct in protein composition and function (reviewed by King.SM⁶; Figure 1B; Video 1). During ciliary motility, the activities of these dynein arms are spatiotemporally regulated by multiple complexes: for example, the radial spoke (RS)–CPC network transmits signal(s) from the CPC to the dyneins through the RS, and perturbation of this network often leads to paralyzed cilia.^7–11 The I1 dynein (or dynein f) and the nexin-dynein regulatory complex (N-DRC) are positioned in the proximal and distal region of the 96-nm repeat, respectively. Mutants with structural defects in both I1 dynein or the N-DRC display impaired motility. ^{12, 13} The N-DRC is also thought to restricts inter-doublet sliding to convert the dynein force into bending of the axoneme.^5,13–15

FIGURE 1.

Structure of wild-type and the ida7-1 mutant cilium. A, cross-sectional diagram of the “9+2” Chlamydomonas cilium, consisting of 9 doublet microtubules (DMTs) and a central pair complex (CPC) with two singlet microtubules. B, the axonemal 96-nm repeat diagrammed in cross-section (left) and longitudinal (right) orientations. The I1 dynein (I1/f) is a two-headed (I1α- and I1β) inner dynein arm (IDA) located at the proximal end of the axonemal repeat between radial spoke RS1 and the outer dynein arm (ODA) row. The tether and tether head (T/TH, red) and modifier of inner arms (MIA, green) complexes connect to the I1-dynein motor domains and the intermediate and light chain complex (ICLC), respectively. Ciliary polarity is indicated by “+” and “−” symbols. C-H, tomographic slices (C, F, cross-sectional; D, G, longitudinal) and isosurface renderings (E, H, longitudinal) of the axonemal 96-nm repeats from wild type (C-E) and ida7-1 (F–H) mutant. I1-dynein (orange arrowheads, C, D) is absent from the mutant axoneme (white arrowheads, F, G). Thin blue lines in C and F indicate the locations of the slices in D and G, respectively. Other labels: A_t/B_t, A- and B-tubule; a-g, single-headed IDAs; N-DRC, nexin-dynein regulatory complex; RS2/3S, radial spokes RS2 and RS3short. Scale bar: 20 nm in C (valid for C, D, F and G).

Our previous cryo-electron tomography (cryo-ET) studies showed that the I1 dynein (dynein f) is a regulatory hub that interacts with several neighboring structures, such as the ODA, IDA a, and two I1 dynein-associated structures, i.e. the I1 tether/tether head (T/TH) and the modifier of inner arms (MIA) complexes (Figure 1B).^16–19 The I1 dynein is the only IDA consisting of two dynein heavy chains, which form a three-lobed structure with two dynein head domains and a large intermediate chain/light chain (ICLC) complex.¹⁶ Genetic and biochemical studies in Chlamydomonas identified nine I1-dynein components in addition to two dynein heavy chains (DHC1 and DHC10): two intermediate chains (IC140 and IC138), five light chains (LC7a, LC7b, LC8, TCTEX1 and TCTEX2B), and two inner dynein arm interacting proteins (IC97 and FAP120) (Table 1). In vitro assays suggest that I1 dynein regulates microtubule sliding through reversible phosphorylation of IC138, and that IC138 hyperphosphorylation correlates with reduced microtubule sliding velocity.^20–24 However, the mechanistic link between IC138 phosphorylation and altered activity of I1 dyneins (and other dyneins) remains unclear. Analysis of the IC138 null mutant, bop5-2, revealed an IC138 subcomplex that is composed of IC138, IC97, FAP120, and LC7b, and localizes to a distinct domain at the bottom of the ICLC structure.^{16, 25} However, the detailed in situ spatial organization of most I1 subunits within the I1 dynein, particularly within the ICLC complex, remains unclear.

TABLE 1.

I1-dynein proteins and MS analysis of Chlamydomonas WT and tctex2b axonemes

Group		Proteins†	Accession number	MW (kD)	Unique peptide NO.		Ratio‡
					WT	tctex2b	tctex2b/WT
I1/f dynein compo-nents	Heavy chains	I1α (DHC1)	EDO96546.1	523	282	265	0.72
		I1β (DHC10)	Q9MBF8.1	511	270	273	0.77
	Intermediate chains	IC140 (DIC3)	AAD45352.1	110	48	47	0.71
		IC138 (DIC4)	AAU93505.1	111	57	54	0.75
	Light chains	Tctex1 (DLT3)	XP_001702138.1	13	6	0	0.19
		Tctex2b (DLT4)	DAA05278.1	14	6	0	0
		LC7a (DLR1)	XP_001694381.1	12	8	10	0.94
		LC7b (DLR2)	XP_001694098.1	11	8	8	1.22
		LC8 (DLL1)	XP_001702907.1	10	7	7	0.89
	IDA interacting proteins	IC97 (DII6)	ACN22074.1	81	30	20	0.13
		FAP120 (DII7)	EDP07339.1	32	17	8	0.16
I1/f dynein-associated	T/TH	FAP43	A8JAF2.1	178	85	85	0.71
		FAP44	A8J1V4.1	186	87	91	0.94
		FAP244	EDP02827.1	179	56	54	0.86
	MIA	FAP73	BAM95826.1	36	24	27	0.58
		FAP100	XP_001699777.1	65	39	41	0.8

^†Protein nomenclature inside bracket is based on Hom et al.⁷²

^‡Protein abundance was estimated by normalized spectral index value. Proteins reduced in tctex2b were highlighted by bold font.

Previous studies have shown that the two dynein tail domains and the second large intermediate chain IC140 are essential for I1 dynein assembly and stable attachment to the axoneme.^26–29 Loss of any of these components results in the complete loss of the I1 dynein complex and severe motility defects. The TCTEX1 and TCTEX2B dynein light chains are considerably smaller subunits, yet loss of TCTEX2B results in slower swimming speed and reduced microtubule sliding velocity in Chlamydomonas,³⁰ and mutations in the TCTEX2B human homolog, TCTEX1D2, cause Jeune asphyxiating thoracic dystrophy due to impaired intraflagellar transportation (IFT) dynein function,³¹ indicating that these light chains are also required for proper regulation of ciliary motility.

The ida7-1 (lacks IC140) and the tctex2b mutants offer powerful tools to dissect the molecular architecture of the I1-dynein ICLC structure. In this study, we rescued ida7-1 with a SNAP tagged IC140 gene. Using cryo-ET and sub-tomogram averaging, we localized the N-terminus of IC140 at the proximal apical region of the ICLC structure. Our structural and biochemical studies of the tctex2b mutant localized TCTEX2B, TCTEX1, IC97, and FAP120 to the center and bottom of the ICLC bordering and partially including the previously localized IC138 subcomplex. Overall, these results provide detailed insight into the in situ three-dimensional organization and interactions of I1-dynein proteins, providing a structural basis for understanding how I1 dynein regulates ciliary motility.

2. MATERIALS AND METHODS

2.1. Strains and cell culture

The Chlamydomonas wild-type (cw15, CC-125) and ida7-1 (CC-3921) strains were obtained from the Chlamydomonas Resource Center (https://www.chlamycollection.org). The pf16D2 and pf16D2;PF16 (tctex2b) mutants were received from the Smith lab and generated as previously described.^{9, 30} Chlamydomonas cells were maintained on agar Tris-acetate-phosphate (TAP) plates under a 12:12 h light:dark cycle at 23°C. Cells were transferred from agar plates to liquid TAP medium and grown with aeration.

2.2. Generation and verification of the ida7-1;SNAP::IC140 strain

A GFP-tagged IC140 construct containing the AphVIII selection cassette³² was generated from the pCP3 clone.²⁷ The AphVIII cassette was inserted into pCP3 with SmalI and KpnI to produce the plasmid IC140-GFP-AphVIII (courtesy of Winfield S. Sale). The 3.28-kb NruI - EcoRI fragment from IC140-GFP-AphVIII was excised and subcloned into plasmid pSE280 to produce pCS5.1. A 2.068-kb NruI - RsrII fragment containing the SNAP tag in place of GFP in exon 2 was synthesized and cloned into the pUC57-Kan vector (Genscript, Piscataway, NJ) to make the clone pCS17. The NruI - RsrII fragment from pCS17 was excised and subcloned into the pCS5.1 plasmid to produce pCS18. A 2.71-kb PspXI-EcoRV fragment from pCS18 was subcloned into the IC140-GFP-AphVIII clone to produce pCS20, which contains the IC140 gene with the SNAP-tag in exon 2. PCR was used to verify the SNAP-tag containing fragments was inserted in the correct orientation (Table S3). All plasmids were propagated in NEB5 E. coli cells (New England Biolabs, Ipswich, MA, USA).

Plasmid pCS20 was used to transform the ida7-1 strain using the glass bead method.³³ Paromomycin-resistant colonies were selected and examined for presence of the SNAP sequences using the SNAP-F and SNAP-R primer sets (Table S3), assembly of the SNAP-IC140 fusion protein in axonemes (Figure 2B) and rescue of motility (Figure 2C). Several transformants were selected that displayed wild-type motility. The swimming velocities of the ida7-1 parental strain CC2677 (cw15, nit1-305), ida7-1 and transformant cells were measured as previously described.³⁴

FIGURE 2.

IC140 subunit N-terminus is located at the proximal/apical area of ICLC. A, diagram of the SNAP-IC140 gene (exons in blue) used to rescue the ida7-1 mutant. B, immunoblots of axonemal proteins extracted from ida7-1, wild type (WT), and SNAP-IC140 rescued strain (ida7-1;SNAP::IC140) were probed with anti-IC140 antibodies; note the slightly higher molecular weight of the SNAP-tagged compared to WT IC140. C, the swimming velocities of the SNAP-IC140 rescued strain (ida7-1;SNAP::IC140) are restored to wild-type levels. Values represent mean ± SEM for n=30. D, SDS-polyacrylamide gel reveals a specific band (arrowhead) of appropriate relative motility in the ida7-1;SNAP::IC140 axonemes treated with streptavidin-Au but not in the control (-Au). Coomassie Brilliant Blue (CBB) staining of tubulin shown as loading control. E-L, tomographic slices (E-J) and isosurface renderings (K, L) of the axonemal 96-nm repeats from wild type (E-G, K) and ida7-1;SNAP::IC140 (H-J, L), viewed in cross-sectional (E, H), longitudinal front (F, I) and longitudinal bottom (G, J) orientations. An additional density in the rescued strain (yellow arrowheads) indicates the location of the IC140 N-terminus, not visible in wild type (white arrowheads). The wild-type average consists of all axonemal particles; for ida7-1;SNAP::IC140 one class average (18% of axonemal particles) is shown for better visualization of the Au label density. Thin blue lines in E,H indicate slice locations in F, G, I, J. Other labels: A_t/B_t, A- and B-tubule; a-g, single-headed IDAs; I1α/β, α- and β-head of the I1 dynein; ICLC, intermediate and light chain complex of the I1 dynein; N-DRC, nexin-dynein regulatory complex; ODA, outer dynein arm; T/TH, tether and tether head complex. Scale bar: 20 nm in E (valid for E-J).

2.3. Axoneme preparation

Chlamydomonas cells cultured in liquid TAP medium were harvested by centrifugation (2200 rpm for 5 min) and then washed twice with freshly made M-N/5 minimal medium.³⁵ Cilia were purified by the pH shock method as previously described.^{18, 36, 37} In brief, the pH shock buffer (10 mM HEPES, 1 mM SrCl₂, 4% sucrose, 1 mM DTT, pH 7.4) was added to the cell pellets, and the pH was reduced to 4.3 by adding 0.5 M acetic acid to the buffer. With gentle stirring for 80 s on ice, cilia were detached from most of the cells. Then, 1 M KOH was added to the solution to increase the pH to 7.2, followed by adding 5 mM MgSO₄, 1 mM EGTA, 0.1 mM EDTA and 100 μl of protease inhibitor (Sigma-Aldrich, St. Louis, MO, USA). The detached cilia were isolated from the cell bodies by centrifugation (1800 ×g for 10 min, 4°C). The cilia-containing supernatant was collected and further centrifuged (2400 ×g for 10 min, 4°C) with a 20% sucrose cushion to purify the cilia. This sucrose cushion purification was repeated twice. The ciliary membrane was removed by adding 1% IGEPAL CA-630 (Sigma-Aldrich, St. Louis, MO, USA) to the supernatant and incubating for 20 min at 4°C with gentle rotation, which was followed by another centrifugation (10,000 ×g for 10 min, 4°C) to collect the axonemes. The axonemal pellet was resuspended in HMEEK buffer (30 mM HEPES, 25 mM KCl, 5 mM MgSO₄, 0.1 mM EDTA and 1 mM EGTA, pH 7.2).

2.4. Gel electrophoresis and immunoblotting

SDS-PAGE, 2DE and immunoblotting were performed as previously described.¹⁸ In brief, total axonemal proteins (20 μg each) of the wild-type and tctex2b strains were separated on a 4-12% gradient SDS-polyacrylamide gel and then transferred to a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA, USA). The membrane was blotted with anti-CK1 (rabbit; 1:10,000 dilution), anti-PP2A (B-subunit) (rabbit; 1:500 dilution), anti-IC140 and anti-IC2 (mouse; 1:10,000 dilution; Sigma-Aldrich, St. Louis, MO, USA). To detect the immune-reactive bands, horseradish peroxidase-conjugated secondary antibody (1:10,000 dilution; Bio-Rad, Hercules, CA, USA) and the Clarity Western ECL Substrate Kit (Bio-Rad, Hercules, CA, USA) or the Thermo Scientific Pierce ECL substrate (Thermo Fisher Scientific, Hillsboro, OR, USA) were used. For the 2DE assay, 50 μg of axonemal proteins were first separated on a 7-cm immobilized pH 3-10 nonlinear gradient dry strip (GE Healthcare, Chicago, IL, USA) and then on a 10% SDS-polyacrylamide gel. Anti-IC138 (rabbit; 1:10,000 dilution) was used as the primary antibody to detect IC138 isoforms. The same secondary antibody and the Clarity ECL visualization kit were used for SDS-PAGE experiments. Both SDS-PAGE and 2DE analysis were replicated three times.

2.5. LC-MS/MS

Wild-type and tctex2b axonemal proteins (40 μg each) were separated by SDS-PAGE (4-12% gradient SDS-polyacrylamide gel) and stained with Coomassie Brilliant Blue. Protein gel pieces were digested overnight with Pierce™ trypsin (Thermo Fisher Scientific, Hillsboro, OR, USA) following reduction and alkylation with DTT and iodoacetamide (Sigma–Aldrich, St. Louis, MO, USA). The samples then underwent solid-phase extraction cleanup with an Oasis HLB μElution plate (Waters, Milford, MA, USA) and the resulting samples were analyzed by LC/MS/MS using an Orbitrap Fusion Lumos mass-spectrometer coupled to an Ultimate 3000 RSLC-Nano liquid chromatography system (Thermo Fisher Scientific, Hillsboro, OR, USA). Samples were injected onto a 75 μm i.d., 75-cm long EasySpray column and eluted with a gradient from 0-28% buffer B over 90 min. Buffer A contained 2% (v/v) acetonitrile and 0.1% formic acid in water, and buffer B contained 80% (v/v) acetonitrile, 10% (v/v) trifluoroethanol, and 0.1% formic acid in water. The mass spectrometer operated in positive ion mode with a source voltage of 1.8 kV and an ion transfer tube temperature of 275 °C. MS scans were acquired at 120,000 resolution in the Orbitrap and up to 10 MS/MS spectra were obtained in the ion trap for each full spectrum acquired using higher-energy collisional dissociation for ions with charges 2-7. Dynamic exclusion was set for 25 s after an ion was selected for fragmentation.

Raw MS data files were converted to a peak list format and analyzed using the central proteomics facilities pipeline (CPFP), version 2.0.3.^{38, 39} Peptide identification was performed using the X!Tandem⁴⁰ and Open MS Search Algorithm (OMSSA)⁴¹ search engines against the Chlamydomonas reinhardtii protein database from the Joint Genome Institute (https://jgi.doe.gov), with common contaminants and reversed decoy sequences appended.⁴² Fragment and precursor tolerances of 10 ppm and 0.5 Da were specified, and three missed cleavages were allowed. Carbamidomethylation of cystein was set as a fixed modification and oxidation of methionine was set as a variable modification. Label-free quantitation of proteins across samples was performed using SINQ normalized spectral index Software.⁴³ The MS analysis was repeated three times for both wild-type and mutant axonemal samples prepared from different batches of culture. The values for unique peptide numbers and ratio of tctex2b/WT for the proteins listed in Table 1 were averaged using the second and third MS data only (Table 1), whereas the first MS run was excluded because of contaminants.

2.6. Streptavidin-gold labeling

In situ streptavidin-gold labeling was performed as previously described.³⁷ Briefly, axonemes isolated from the ida7-1;SNAP::IC140 strain were resuspended in 400 μl of HMEEK buffer (30 mM HEPES, 25 mM KCl, 5 mM MgSO₄, 0.1 mM EDTA and 1 mM EGTA, pH 7.2) and divided into two 1.5-ml Eppendorf tubes (200 μl each). One microliter of 1 mM BG-biotin (New England Biolabs, Ipswich, MA, USA) was added to each tube and incubated overnight with gentle rotation at 4°C. To wash away unbound BG-biotin, 1 ml of HMEEK was added to the tube and centrifuged (10,000 ×g for 2 min, 4°C). The wash step was repeated three times, and biotin-axonemes were resuspended in 200 μl of fresh HMEEK buffer. 5 μl of 80 μg/ml 1.4-nm-sized streptavidin nanogold particles (Nanoprobes Inc, Yaphank, NY, USA) was added to one tube and incubated at 4°C for 4 h; no streptavidin-gold was added to the sample in the second tube, which served as control. The labeled axonemes were washed by centrifugation (10,000 ×g for 2 min, 4 °C) and resuspended in HMEEK buffer.

2.7. Cryo-ET

Freshly isolated or labelled axonemes were mixed with 10-fold-concentrated BSA-coated gold (10 nm in diameter) solution⁴⁴ in a 3:1 (axoneme: gold) ratio. An aliquot (4 μl) of the mixed solution was applied to a glow-discharged (30 s at −35 mA) copper R2/2 holey carbon grid (Quantifoil Micro Tools GmbH, Jena, Germany). Excess liquid was blotted with Whatman filter paper (No. 1) for 2.0-2.5 s from the backside of the grid, which was then plunge-frozen into liquid ethane using a homemade plunge-freezer. Grids were stored in liquid nitrogen until imaged.

Vitrified grids were either loaded into a Tecnai F30 transmission electron microscope (Thermo Fisher Scientific, Hillsboro, OR, USA) (for samples: wild type, ida7-1, tctex2b, pf16D2) or were mounted in Autogrids™ (Thermo Fisher Scientific, Hillsboro, OR, USA) before loading into a Titan Krios transmission electron microscope (Thermo Fisher Scientific, Hillsboro, OR, USA) (for samples: wild type, ida7-1;SNAP::IC140 axonemes). The Tecnai F30 was equipped with a 2k x 2k charge-coupled device camera (Gatan, Pleasanton, CA, USA), and the Titan Krios was equipped with a 4k x 4k K2 direct detection camera (Gatan, Pleasanton, CA, USA) which was operated in counting mode (15 frames, 0.4 s exposure time per frame and a dose rate of 8 electrons/pixel/s for each tilt image). Both microscopes were operated at 300 kV and data were collected using the SerialEM software.⁴⁵ Tilt series images were acquired in low-dose mode from −60° to 60° with 1.5-2.5° increments, and a dose-symmetric tilting scheme⁴⁶ was applied for the data collected by the Titan Krios. The post-column energy filter (Gatan, Pleasanton, CA, USA) was operated in zero-loss mode with a 20-eV slit width. Magnification was set to 13,500 with an effective pixel size of 1.07 nm and to 26,000 with an effective pixel size of 0.55 nm for the Tecnai F30 and Titan Krios, respectively. The defocus was set to −8 nm for the Tecnai F30, and a Volta-Phase-Plate⁴⁷ was used with −0.5 μm defocus for the Titan Krios.

2.8. Image processing

The frames of wild-type and ida7-1;SNAP::IC140 data were motion corrected with a script extracted from the IMOD software;⁴⁸ the latter was also used to align the tilt serial images with the fiducial markers (10 nm gold particles) and to reconstruct the tomograms by weighted back-projection. Subtomogram averaging of the axonemal 96-nm repeats was performed with PEET software,⁴⁹ and a principal component analysis clustering method⁵⁰ was used for classification analyses. The three-dimensional structures of the axonemal averages were visualized with a UCSF Chimera software package.⁵¹ The biological replicates information about the number of tomograms, axonemal repeats and EM sessions of data recording for each strain used in this study, as well as the estimated resolution of the averages (using the 0.5 criterion of the Fourier shell correlation), are listed in Table S2.

3. RESULTS

3.1. Location of the IC140 N-terminus

We used cryo-ET and sub-tomogram averaging to compare the 96-nm axonemal repeats from Chlamydomonas wild type and the ida7-1 mutant, and found that the I1-dynein complex is not assembled in ida7-1 axonemes (Figure 1C–H), which is consistent with previous biochemical and TEM data.²⁷ The absence of the I1 dynein also appears to destabilize the I1 dynein-associated structures, i.e. the T/TH and MIA complexes, which exhibited altered morphologies or were blurred in the average of ida7-1 compared to wild type (Figure 1C–H).

To determine the location of IC140 within the ICLC complex, we used in situ SNAP-tag labeling.³⁷ ida7-1 was rescued using a plasmid with the IC140 gene that had a SNAP-tag inserted into the second exon (Figure 2A). The N-terminally tagged IC140 protein was assembled into the ida7-1;SNAP::IC140 axonemes (Figure 2B), and the motility defects of ida7-1 were fully rescued (Figure 2C), demonstrating that the SNAP-tag did not affect IC140 function, I1-dynein function, nor axonemal assembly. To evaluate streptavidin-gold binding to the SNAP-tag in situ, ida7-1;SNAP::IC140 axonemes were conjugated with BG-biotin, treated with or without streptavidin-gold, and separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Gel staining (with a silver-enhancement kit) revealed a specific band with an appropriate molecular weight (~220 kDa) in the streptavidin-gold treated sample (Figure 2D), indicating that streptavidin-gold could ligate with the SNAP-tagged IC140 in intact axonemes.

In the averaged 96-nm repeats of labeled ida7-1;SNAP::IC140 axonemes, the streptavidin-gold particle was visualized as an additional density at the proximal apical region of the ICLC complex (Figure 2H–J,L), marking the location of the IC140 N-terminus. This density was not observed in the corresponding position of wild-type axonemes (Figure 2E–G,K). The position of the IC140 N-terminus close to the ODA row suggests that the IC140 protein accounts for the apical density of the ICLC complex, which includes a small structure that connects the ICLC complex to the A-tubule (Video 1).

3.2. TCTEX1, IC97 and FAP120 are reduced in tctex2b axonemes

Previous insertional mutagenesis of PF16 (pf16D2), which encodes a central pair complex protein, affected the expression of two genes in Chlamydomonas, PF16 and TCTEX2B, which is ~2 kb downstream of the PF16 gene.^{9, 30} Rescue of pf16D2 using wild-type PF16 yielded the tctex2b (pf16D2;PF16) mutant, which was reported to be a null mutant for TCTEX2B.^{9, 30} DiBella et al also showed that TCTEX2B - in contrast to IC140 - is not required for assembly of the I1-dynein complex, but is essential for normal ciliary motility and I1-dynein stability, because the mutant phenotype includes significantly slower swimming speed and reduced microtubule sliding velocity.³⁰ To further understand the effect of TCTEX2B loss on ciliary structure and motility, we compared the proteomes of wild-type and tctex2b mutant axonemes using mass spectrometry (MS).

We found that in tctex2b mutant axonemes, TCTEX2B was completely missing, along with significant reduction of three other I1-dynein components, TCTEX1, IC97, and FAP120. In contrast, the remaining five ICLC subunits, IC140, IC138, LC7a, LC7b and LC8, assembled close to wild-type level (Table 1). DiBella et al. previous reported that TCTEX2B is essential for stably incorporating TCTEX1 into the I1 dynein³⁰, which was confirmed by our data. In addition, our results suggest that TCTEX2B is also required for the stable assembly of IC97 and FAP120 into the axoneme. In contrast, IC97 and FAP120 do not appear to be required for the stable assembly of TCTEX2B (and TCTEX1), because axonemes of the bop5-2 mutant, which lacks the IC138 subcomplex (i.e. IC138, IC97, FAP120, and LC7b), still assembled TCTEX2B and TCTEX1 into their axonemes.²⁵

3.3. Tctex2b has structural defects in the ICLC center and bottom regions

To identify the location of the TCTEX2B-dependent subcomplex (containing TCTEX2B, TCTEX1, IC97, and FAP120) within the ICLC, we compared the 3D structures of the 96-nm axonemal repeats from Chlamydomonas wild type and tctex2b using cryo-ET and sub-tomogram averaging. In the average of all tctex2b axonemal repeats, a density at the bottom of the ICLC complex - proximal of the central ICLC protrusion - was clearly reduced compared to the wild-type ICLC (Figure 3A–D), indicating that all or some of the four subunits that were greatly reduced in tctex2b (i.e. TCTEX2B, TCTEX1, IC97, and FAP120) reside in this area in wild-type axonemes. Our previous cryo-ET study of the bop5-2 mutant showed that the IC138 subcomplex (containing IC138, IC97, FAP120, and LC7b) localizes to the central protrusion and bottom region of the ICLC (Figure 3A),¹⁶ including the area defective in tctex2b. These results are consistent, because the two largest TCTEX2B-dependent components, IC97 (81 KDa) and FAP120 (32 KDa), are also part of the IC138 subcomplex. The remaining I1 dynein and associated structures, such as the dynein motor heads, the T/TH and MIA complexes, and the apical density of the ICLC (largely composed of IC140), resembled those of the wild-type ICLC structure (Figure 3A–D). This agrees well with our mass-spectrometry results (Table 1), and suggests stable assembly of these proteins and corresponding structures in the tctex2b axoneme.

FIGURE 3.

TCTEX2B loss causes structural defects in the center and at the bottom of the ICLC. A-D, tomographic slices (left) and isosurface renderings (right) of the axonemal 96-nm repeats viewed in longitudinal (A, C) and cross-sectional (B, D) orientations averaged from all axonemal particles (100%) from wild type (A, B) and tctex2b (C, D). Density of the ICLC bottom region is reduced in the mutant (light orange arrowheads, C, D). The location of the IC138 subcomplex (purple outline, A) is adapted from our previous cryo-ET study of the Chlamydomonas bop5-2 mutant that lacks the IC138 subcomplex.¹⁶ Note that the slight reduction of IDA b in tctex2b is not relevant to TCTEX2B mutation but derived from its proximal-distal asymmetrical distribution along the axoneme.⁷¹ E-H, classification analysis of tctex2b showed that 75% of the axonemal repeats (Class 1, E, F) had a defect at the ICLC bottom region (white arrowheads, E, F), and 25% (Class 2, G, H) lacked density in the center of the ICLC complex (white arrowheads G, H). I, isosurface renderings of wild-type I1 dynein in longitudinal front (left) and 40°-rotated (right) orientations depict the structural defects (light orange densities) observed in tctex2b. To visualize I1 dynein and associated structures (T/TH and MIA complexes), other densities were made transparent. Thin blue line in A indicates locations of the slices viewed in cross-sectional orientation. Other labels: A_t/B_t, A- and B-tubule; a-e, single-headed IDAs; I1α/β, the α- and β-head of the I1 dynein; ICLC, intermediate and light chain complex of the I1 dynein; ODA, outer dynein arm; RS1, radial spoke RS1. Scale bar: 20 nm in A (valid for EM images in A-H).

However, the proteome analysis also showed that the TCTEX2B-dependent subunits (TCTEX1, IC97, FAP120) were only reduced, but not completely missing from the tctex2b axonemes. Therefore, some of the averaged axonemal repeats from tctex2b could still contain some or all of the TCTEX2B-dependent components, resulting in structural heterogeneity and blurring/weakening of the averaged density rather than completely missing density. To sort the axonemal repeat units into structurally more homogeneous groups, we applied a classification approach focusing on different regions of the ICLC structure using masks. The classification of wild-type axonemes resulted in only one class, with all axonemal particles having an intact ICLC. However, the tctex2b axonemal particles were grouped into two classes (Figure 3E–I): Class average #1 (with 75% of the particles) showed a structural defect in the proximal bottom region of the ICLC that was already observed by the average of all tctex2b repeats, only the defect in the class average was more pronounced including part of the central ICLC protrusion (Figure 3E,F,I). In contrast, class average #2 (with 25% of the particles) revealed an additional defect in the center of the ICLC at the apical border of the IC138 subcomplex (Figure 3G–I).

To verify the structural defects caused by the missing of TCTEX2B, we performed the same classification analysis for the pf16D2 mutant, which is the parental strain of tctex2b, lacking both the PF16 and TCTEX2B proteins.⁹ Similar to tctex2b, axonemal particles from pf16D2 were grouped into Class 1 (85%) and Class 2 (15%) defects (Figure S1). These results suggest that TCTEX2B, TCTEX1, IC97, and FAP120 localize to the proximal bottom and central regions of the ICLC complex.

3.4. DMTs 3-4 doublet-specific structure and IC138 hyperphosphorylation in tctex2b

In our previous cryo-ET study of the I1 dynein, we identified a doublet-specific feature in the ICLC complex of DMTs 3 and 4 of Chlamydomonas flagella: A small ring-shaped structure connects the ICLC central and distal protrusions only on these two doublets¹⁶ (see also cyan-colored density in Figure 4). Since the central protrusion and the bottom region of the ICLC are connected, and both regions showed structural defects in the Class 1 average of tctex2b axonemes (Figure 3E,F), we examined whether this doublet-specific structure was also affected in tctex2b. However, despite the defect of the central protrusion, the tctex2b DMTs 3-4 still possessed the doublet-specific ring-structure (compare Figure 4E,F), indicating that the interaction of the ring-structure with the distal ICLC protrusion is sufficient for its stable assembly into DMTs 3-4.

FIGURE 4.

tctex2b mutant axonemes retain DMTs 3-4-specific structures and contain hyperphosphorylated IC138. A-D, isosurface renderings (A, B; longitudinal front view) and tomographic slices (C, D; longitudinal bottom view) of wild-type I1 dynein averaged from DMTs 3-4 (A, C) and DMTs 1,2,5-9 (B, D) show the DMTs 3-4-specific structure (cyan arrowheads, A, C), reported by Heuser et al.¹⁶ Location of the inset in (A) is indicated by a black dotted box. Note that the DMTs 3-4-specific structure connects to the I1 distal protrusion density (black arrowhead). E and F, tomographic slices of tctex2b I1 dynein averaged from DMTs 3-4 (E) and DMTs 1,2,5-9 (F) show the retention of the DMTs 3-4-specific structure (cyan arrowhead, E) despite the reduction in ICLC density (light orange arrowheads, E, F). G, two-dimensional (2D) gel immunoblots of axonemal proteins extracted from wild type and tctex2b were probed with anti-IC138 antibodies. Note that the IC138 isoforms have shifted from the basic (arrowheads) to the acidic side of the gel in tctex2b. H, immunoblots of axonemal proteins extracted from wild type and tctex2b were probed with antibodies to casein kinase 1 (CK1), phosphatase 2A (PP2A) and intermediate chain 2 (IC2, control) of the outer dynein arm. Other labels: At, A-tubule; I1α/β, α- and β-head of the I1 dynein; N-DRC, nexin-dynein regulatory complex; ODA, outer dynein arm; RS1/2, radial spokes RS1 and RS2; T/TH, tether and tether head complex. Scale bar: 20 nm in C (valid for C-F).

Our MS result showed that two IC138 subcomplex components, IC97 and FAP120, were reduced in the tctex2b mutant, whereas IC138 itself remained at wild-type levels (Table 1). The IC138 phosphorylation state is thought to be a regulatory switch, and IC138 hyperphosphorylation has previously been shown to reduce microtubule sliding velocity, indicates impaired I1-dynein function and motility.^12,20 Therefore, we investigated the IC138 phosphorylation status in the tctex2b mutant using a 2D-gel immunoblot assay. Axonemal proteins from wild type and tctex2b were separated by 2D gel electrophoresis and then probed with anti-IC138 antibody. Previous studies have shown that the spot locations of IC138 isoforms shift from basic to the more acidic side of 2D gels as their phosphorylation level increases.^{5, 21} Compared to wild type, more IC138 isoforms were shifted to the acidic side, suggesting that IC138 showed a higher phosphorylation level in the tctex2b mutant (Figure 4G). In contrast, the IC138 phosphorylation enzyme casein kinase 1 (CK1) and dephosphorylation enzyme phosphatase protein 2A (PP2A, B-subunit),^52,53 appeared to retain wild-type levels in tctex2b (Figure 4H). The high phosphorylation level of IC138 is consistent with the reduced microtubule sliding velocity seen in tctex2b mutant axonemes.³⁰ Our data suggest that not only IC138, but additional subunits located in the proximal bottom and central regions of the ICLC complex (i.e. TCTEX2B, TCTEX1, IC97, FAP120) are required for proper function of I1 dynein, dynein motor activity and ciliary motility.

4. DISCUSSION

Cilia are composed of hundreds of different proteins, e.g., more than 700 proteins in Chlamydomonas (http://labs.umassmed.edu/chlamyfp/index.php)⁵⁴, and often the mutation of a single protein results in impaired ciliary motility; yet the location of most ciliary proteins in the intact organelle remains unclear.⁵⁵ Recent studies have combined biochemical, genetic, and high resolution imaging approaches, such as cryo-ET and sub-tomogram averaging, to locate proteins, visualize their 3D structure and characterize their function(s) within cilia, including elucidating the conformational changes of axonemal structures in actively beating cilia.^{5, 13, 18, 37, 56–58} Using these approaches, we investigated the protein composition and structural differences between Chlamydomonas wild type and I1-dynein ICLC mutants, ida7-1 and tctex2b (Table S1). Our data together with published work support a more detailed model of the in situ 3D arrangement of the eleven known I1-dynein proteins (Figure 5).

FIGURE 5.

Model of the structural organization of the I1 ICLC. A-C, isosurface renderings (A, B) and schematic drawing (C) of Chlamydomonas wild-type I1 dynein viewed from longitudinal front (A, C) and back (B) orientations. Based on wild type-mutant comparisons, the ICLC proteins are organized into three major groups: (i) the IC140/LC7a/LC8 density (orange; the N-terminus of IC140 is indicated with a small orange circle in A); (ii) the IC138/LC7b density (blue); and (iii) the IC97/FAP120 density (yellow). TCTEX2B may interact with TCTEX1, and TCTEX2B/TCTEX1 (yellow density with dotted line) is predicted to interact with both the IC97/FAP120 and IC140/LC7a/LC8 group. D-F, schematic drawings show the observed structural defects in the I1-dynein mutants. The entire I1 dynein is missing in ida7-1 (D); loss of TCTEX2B causes defects in the center and bottom regions of the ICLC complex and hyperphosphorylated IC138 (star) (E); the IC138 subcomplex is missing in bop5-2 axonemes, but TCTEX1 and TCTEX2B are present, possibly through a connection to the IC140 subunit.^16,25 Other labels: At, A-tubule; a-c, single-headed IDAs; I1α/β, α- and β-head of the I1 dynein; IC140-N, N-terminus of IC140; IDA, inner dynein arm; MIA, modifier of inner dynein arms; ODA, outer dynein arm; OID, outer-inner dynein linker; RS1, radial spoke RS1; T/TH, tether and tether head complex.

I1 dynein has a “tri-lobed” overall structure that includes two dynein motor heads associated with the T/TH complex, and an ICLC complex that connects to several neighboring structures, i.e. the A-tubule via the ICLC-to-A_t docking complex, the ODAs via the OID linker, the MIA complex at its distal end, and IDA a on its bottom side (Figure 5A,B).^16–18 Here, our SNAP-tag labeling data revealed that the IC140 N-terminus is located at the proximal apex of the ICLC complex, which allows interactions with multiple other structures, such as the tails of the dynein heavy chains, the ODAs via the OID linker, the IC138 subcomplex at the ICLC bottom, the ICLC-to-A_t docking complex, and the MIA complex (Figure 5A–C; Video 1). These interactions may serve different functions, such as being required for the preassembly of the 20S I1-dynein complex in the cytoplasm, docking to and transport by the IFT, facilitate site-specific docking of the preassembled I1 dynein onto the axoneme, and/or transmit signal(s) that regulate ODA and/or I1-dynein activity and ciliary beating.^{27, 59–61}

The boundaries of the IC138 subcomplex at the bottom of the ICLC structure were previously defined using cryo-ET of the bop5-2 mutant (Figure 5C,F).¹⁶ Because LC7a and LC8 are present in bop5-2 axonemes,²⁵ we predict that these two dynein light chain proteins are likely associated with IC140 and located in the apical region of the ICLC as shown in our summary model of the I1 ICLC organization (Figure 5C).

Reductions in TCTEX1, LC97, and FAP120 in the tctex2b mutant (Table 1) suggest that TCTEX2B may occupy a position that maintains the stability of these proteins. TCTEX1 may be closely associated with TCTEX2B as they have similar sedimentation profiles, distinct from those of IC140 and DHC1 in bop5-2 axoneme sucrose gradient assays.²⁵ In addition, TCTEX2B was demonstrated to be responsible for the instability of TCTEX1 in the tctex2b mutant.³⁰ Based on our cryo-ET and classification analyses of tctex2b (Figure 3), the position of TCTEX1/TCTEX2B is likely to be adjacent to the Class 1 and Class 2 defective densities, which consist largely of IC97 and FAP120 (Figure 5C). In addition, the direct attachment of TCTEX1/TCTEX2B to the IC140 subcomplex ensures their presence in bop5-2 axonemes (Figure 5F). Therefore, TCTCEX1/TCTEX2B are predicted to be located at the center of the ICLC complex, connecting to multiple ICLC subunits (Figure 5C).

Structural comparison of tctex2b (Figure 5E) and bop5-2 (Figure 5F) allows the assignment of IC138 to a position between the IC140 and IC97/FAP120 densities. Cross-linking experiments revealed that LC7b interacts with the WD-repeat domains of IC138, indicating a tight association between these two proteins.²³ This association is also consistent with our MS analysis of tectex2b, in which IC138 and LC7b are present at wild-type levels, whereas levels of the two other components of the IC138 subcomplex, LC97 and FAP120, were significantly reduced (Table 1). These results lead to the definition of IC138/LC7b density within the ICLC complex (Figure 5C) and thus indicate a direct interaction between IC97/FAP120, and in the region where the IDA a motor domain interacts with the ICLC (Figure 5A).

Given the conservation of proteins and structural features in I1 dynein, the Chlamydomonas ICLC model described above is likely conserved in the motile cilia of other organisms. The eleven Chlamydomonas I1-dynein proteins are found in eukaryotes.^{54, 55} Our cryo-ET studies also revealed that I1 dynein shares common structural features among axonemes from various species, including Chlamydomonas, Tetrahymena, sea urchin, and human.^{18, 62} I1 dynein also appears to be present in the eel sperm axoneme, which lacks a large subset of axonemal structures, such as ODAs, RSs, and CPC.⁶³

Cryo-EM single particle reconstruction revealed the high-resolution structures of recombinantly expressed human cytoplasmic⁶⁴ and IFT⁶⁵ dynein complexes, as well as axonemal ODAs docked to isolated doublet microtubules.^66–68 These dynein complexes share common organizational features of the tail-bound IC-LC complex, which is important for dynein heavy chain dimerization and cargo-binding, or microtubule-docking in case of axonemal ODAs. Specifically, the IC C-terminal WD40 domain binds to the dynein heavy chain and the extended N-terminus binds to dimers of light chains (LC7, LC8 and TCTEX). These light chains are conserved in axonemal I1 dynein, suggesting that they may also form dimers and bridge between the N-termini of IC138 and IC140. However, previous analysis of a truncated IC138 mutant, bop5-1, revealed that LC7b but not LC7a, was missing from the mutant I1 ICLC,²³ suggesting different binding affinities of these light chains to the IC. Other factors that imply a different structural organization of the I1 dynein compared to cytoplasmic, IFT and axonemal outer dynein complexes include the presence of additional subunits, i.e. I1 ICLC specific interacting proteins FAP120 and IC97, a comparably small direct connection to the A-tubule (“cargo”), and multiple connections with surrounding axonemal complexes, i.e. the MIA complex, IDA a, OID-linker, and T/TH.¹⁷ A high-resolution study of I1 dynein would be needed to resolve its atomic structure and detailed subunit interactions, such as IC/LC dimerization. However, isolation of intact and homogeneous I1 dynein complexes has been difficult, and cryo-EM single particle reconstruction of I1 dynein docked to isolated doublet microtubules, is more challenging than for ODAs, because the outer row consists just of ODAs repeating with 24 nm periodicity, whereas the inner row is more heterogeneous with I1 dynein repeating with 96 nm periodicity.

Within the Chlamydomonas ICLC complex, IC138 plays a pivotal role in the regulation of ciliary motility by I1 dynein. Phosphorylation of IC138 is diagnostic of dynein activity^55,69,70 and hyperphosphorylated IC138 correlates with inhibition of dynein function in mutants lacking the CPC, RS, I1β dynein motor domain, and the I1-associated T/TH and MIA complexes.^5,17,18,21 However, the molecular mechanism underlying the CPC–RS–I1 signaling cascade has remained unknown, in part because it involves several multi-subunit complexes for which we know little about subunit-organization and protein-protein interactions. The hyperphosphorylation of IC138 and motility defects seen in tctex2b are due to the missing of a small I1-dynein light chain, TCTEX2B (14 kDa), which caused reduction of three additional proteins (TCTEX1, LC97, FAP120) and structural defects in the ICLC that are considerably smaller than observed in previously studied I1-dynein mutants. Our data suggest that three subunit groups, i.e., IC140/LC7a/LC8, IC138/LC7b and IC97/FAP120, are positioned in order, from the apical to the bottom regions of the ICLC complex, with TCTEX1/TCTEX2B located in a central position, linking these groups (Figure 5C). In wild-type axonemes, RS1 indirectly associates with the IC97/FAP120 density through interactions with IDA a (Figure 5A–C; Video 1), but this association is physically disrupted in the tctex2b mutant as a result of the Class 1 defect (Figure 3I). This suggests that these protein interactions are part of the CPC–RS–I1 signaling cascade, and disruption results in tctex2b’s severe motility defects. Thus, our results show that full-length TCTEX2B is critical for the stable assembly of a TCTEX2B-dependent complex (including TCTEX1, LC97, and FAP120), which plays a functional role in the CPC-RS-I1 signal pathway that regulates cilia motility.

Abbreviations:

A_t

A-tubule

B_t

B-tubule

CPC

central pair complex

Cryo-ET

cryo-electron tomography

DHC

dynein heavy chain

DMT

doublet microtubule

FAP

flagellar associated protein

I1α/β

α- and β-head of I1 dynein

ICLC

intermediate chain and light chain complex of I1 dynein

IDA

inner dynein arm

IFT

intraflagellar transportation

MIA

modifier of inner arms complex

N-DRC

nexin-dynein regulatory complex

ODA

outer dynein arm

OID

outer-inner dynein linker

RS

radial spoke

T/TH

tether and tether head complex

DATA AVAILABILITY

The averaged I1-dynein densities in this study have been deposited in EMDataResource with the accession codes EMD-20566 (wild type), EMD-20567 (ida7-1) and EMD-20568 (tctex2b).