calaxin mutation causes OAD partial loss in zebrafish axoneme

Previously, we reported that calaxin^−/− zebrafish exhibit abnormal KV ciliary movement and consequent randomized left–right patterning (Sasaki et al, 2019). Calaxin is known as a component of the vertebrate DC, the linker structure between OAD and DMT. Thus, we speculate that the calaxin^−/− phenotype is caused by OAD defects.

To clarify this point, we investigated OAD localization in calaxin^−/− sperm flagella. Immunofluorescence microscopy (IFM) showed that calaxin^−/− sperm lacked Calaxin and OAD heavy chains Dnah8 and Dnah9 (Fig 1A, C, and D). Inner arm dynein (IAD) heavy chain Dnah2 showed normal localization (Fig 1B). Interestingly, the borders of OAD(+) and OAD(−) regions varied between calaxin^−/− sperms (Fig 1C and D; white lines and white dotted lines show OAD(+) and OAD(−) regions, respectively). These data suggest that calaxin mutation caused unstable OAD-DMT docking in sperm flagella and possibly in KV cilia. Consistent with our IFM data, a recent study of PCD patients also showed that CALAXIN mutation resulted in the absence of OAD from the distal axoneme (Hjeij et al, 2023). Moreover, our structural analysis of the calaxin^−/− axoneme showed that Calaxin is required to stabilize the OAD docking onto DMT in vertebrates (Yamaguchi et al, 2023).

Zebrafish Calaxin binds to calcium, which is disrupted by the E130A mutation

Calaxin is also known as an OAD-associated calcium sensor. In ascidian C. intestinalis, Calaxin plays an essential role in sperm waveform regulation by inhibiting axonemal dynein activity depending on calcium concentration (Mizuno et al, 2012). However, in vertebrates, the calcium-binding ability of Calaxin and its function have not been assessed.

To analyze this point, we focused on zebrafish E130A-Calaxin, which corresponds to Ciona E118A-Calaxin. Calaxin has three calcium-binding domains, EF-hands (Fig 1E). Although all three EF-hands of Ciona Calaxin bind to a calcium ion, E118A, a single substitution within the EF-hand2 leads to the complete loss of calcium-binding ability (Shojima et al, 2018).

To validate the calcium-binding ability of Calaxin in vertebrates, we purified zebrafish WT- and E130A-Calaxin and performed the isothermal titration calorimetry (ITC) experiment. The result of WT-Calaxin was fitted with a three-site sequential binding model, indicating that all three EF-hands bind to calcium. On the contrary, E130A-Calaxin showed no significant binding, as in Ciona E118A-Calaxin (Fig 1F). SDS–PAGE under the presence of calcium or EGTA also revealed that calcium increased the mobility of WT-Calaxin, but E130A-Calaxin showed only a slight change (Fig 1G). This behavior is identical to Calmodulin and its calcium-binding–deficient mutant, respectively (Geiser et al, 1991). To confirm that E130A-Calaxin retains the normal secondary structure, we assessed the interaction between recombinant Calaxin and calaxin^−/− axoneme. Recombinant WT- and E130A-Calaxin specifically bound to the OAD(+) region of calaxin^−/− sperm flagella, demonstrating that they folded properly (Fig S1A–C). armc4^−/− axoneme was used as the negative control because it completely lacks OAD (Yamaguchi et al, 2023). These results indicate zebrafish Calaxin binds to three calcium ions, and E130A-Calaxin completely loses its calcium-binding ability without disrupting the native conformation. EF-hand2 containing zebrafish E130 residue is highly conserved among Ciona, zebrafish, and mammals, suggesting the importance of this domain (Fig 1E). Here, we decided to use E130A-Calaxin to analyze the calcium-dependent function of Calaxin.

Calaxin stabilizes OAD independent of its calcium-binding ability

We previously showed the loss of Calaxin caused the reduction of OAD in sperm flagella and the abnormal Kupffer’s vesicle (KV) ciliary movement, resulting in laterality randomization (Sasaki et al, 2019; Yamaguchi et al, 2023). To analyze whether the calcium-binding ability of Calaxin is involved in these phenotypes, we performed rescue experiments using WT- and E130A-Calaxin.

First, we assessed the sperm phenotype by generating transgenic zebrafish expressing WT- or E130A-Calaxin under beta-actin promoter (Tg(actb2:calaxin);calaxin^−/− and Tg(actb2:calaxin_E130A);calaxin^−/−). Hereafter, we call them Tg WT and Tg E130A, respectively. Transgenes were integrated into the calaxin^−/− zebrafish with the Tol2 transposon system (Kawakami, 2007), resulting in viable and fertile animals. IFM showed that both WT- and E130A-Calaxin were properly incorporated into the axoneme of transgenic sperm (Fig 1A). Despite the loss of calcium-binding ability, E130A-Calaxin, as well as WT-Calaxin, restored the loss of OAD in the distal region of calaxin^−/− sperm flagella (Fig 1C and D). These results show that Calaxin localizes to the axoneme and stabilizes OAD independent of its calcium-binding ability. Note that OAD was absent from all distal tips of sperm flagella obtained from both transgenic animals (Fig 1C and D, white boxes and magnified views). We speculate that this is caused by different expression regulations between the native calaxin promoter and the actb2 promoter we used for transgene expression.

Next, we also tested the KV ciliary phenotype. calaxin^−/− KV cilia exhibited a decreased ratio of regularly rotating cilia and a lower beating frequency. Injection of either calaxin WT or E130A mRNA into calaxin ^−/− embryos restored the motility of KV cilia. Importantly, the fraction of the regularly rotating cilia and the beating frequency were identical between WT and E130A mRNA injections, indicating that E130A mutation does not alter KV ciliary motility (Fig 2A–C and Video 1). In addition, Tg E130A zebrafish showed normal laterality (Fig 2D). These results indicate that Calaxin contributes to KV ciliary motility independent of its calcium-binding ability, possibly stabilizing OAD.

Demembranated sperm model reconstituted Calaxin-dependent calcium-induced waveform asymmetry

The inhibition of Calaxin disrupts sperm’s calcium-induced asymmetric beating (Mizuno et al, 2012), but it is undetermined whether Calaxin directly detects calcium or is a mediator of another calcium sensor. To test whether its calcium-binding ability is essential for an asymmetric beating, we established a demembranated sperm model according to the Chlamydomonas flagellum model (Brokaw, 1979; Kamiya & Witman, 1984). Briefly, sperm was demembranated with 0.5% NP-40, then reactivated by 50 μM ATP + 50 μM ADP in the presence of 0.1 mM EGTA (EGTA condition) or CaCl₂ (pCa4 condition). ADP was supplemented to activate all axonemal dynein subspecies (Yagi, 2000). A comparison of different ATP and ADP concentrations showed that a higher ATP concentration induced a faster beating frequency and that the supplementation of ADP reduced the variation of beating frequencies (Fig S2A). On the contrary, a higher ATP concentration resulted in a lower reactivation rate, especially in the pCa4 condition (Fig S2B). Thus, we chose the 50 μM + ATP 50 μM ADP condition to enable the stable beating frequency and the higher reactivation rate. WT, calaxin ^−/−, Tg WT, and Tg E130A sperm models were successfully reactivated.

In WT, Tg WT, and Tg E130A sperm models, the pCa4 condition induced more asymmetric waveform than the EGTA condition, but the asymmetry of Tg E130A was lower (Fig 3A–C and Video 2). Calaxin ^−/−sperm model contained a proximal region with high amplitude and a distal region with low amplitude (Fig 3D and Video 2). We investigate this phenotype in the later section. Beating frequencies of Tg WT and Tg E130A sperm models were similar in EGTA and pCa4 conditions, respectively. Sperm models showed a slower beating under the pCa4 condition, although it was not statistically significant in calaxin ^−/− and Tg WT (Fig S2C). To test whether Calaxin’s calcium-binding ability is necessary for calcium-induced waveform transformation, we decided to quantify the waveform asymmetry of WT, Tg WT, and Tg E130A sperm models.

Calaxin E130A mutant diminished calcium-induced waveform asymmetry

The planar ciliary beating can be divided into the static component and the dynamic component corresponding to the time-averaged basal shape and the dynamic beating, respectively (Eshel & Brokaw, 1986; Geyer et al, 2016). To quantify the sperm waveform asymmetry, we calculated the spatial average of the static component in each flagellum, which we call “basal curvature” below (Fig S2E). We traced sperm flagella and calculated their tangent angles. Because zebrafish sperm have round heads, we could not calculate the tangent angle relative to the direction of the sperm head as in sea urchin sperm (Brokaw, 1979). Instead, we selected sperm with their heads fixed to the glass slide and freely swimming flagella and calculated the tangent angle relative to the recorded frame. To obtain the time and spatial average, we plotted tangent angles from ≧5 traces over one beat cycle followed by the fitting with the linear equation. Basal curvature was defined as the slope of the fitted equation.

The quantification revealed that the Tg E130A sperm model failed to increase the basal curvature with statistical significance in response to calcium, whereas WT and Tg WT significantly elevated it in the pCa4 condition. The basal curvature of Tg E130A in the pCa4 condition was also significantly lower than Tg WT in the same condition (Fig 3E). To test whether beatings are successfully separated into dynamic and static components, we calculated the dynamic component by subtracting the basal curvature from the tangent angles. All of WT, Tg WT, and Tg E130A showed uniform basal curvature and symmetric, sinusoidal dynamic components, which confirmed that their beatings were successfully separated into the dynamic and static components (Fig 3A–C, dotted lines, Fig S2F). Because the calculated dynamic components of Tg WT and Tg E130A were symmetric, the loss of OADs at the distal tips of their flagella did not strongly affect the basal curvature. We also calculated the asymmetry index as in Ciona sperm (Mizuno et al, 2012) (Fig S2D). The result showed a similar trend as the basal curvature: the pCa4 condition led to higher asymmetry indices, with Tg E130A exhibiting a weaker elevation compared with WT and Tg WT. However, these differences did not reach statistical significance, possibly because of the limited sample size. Note that we used an alternative definition of “P-bend” and “R-bend”: we defined the beating with higher curvature as “P-bend” and the other as “R-bend,” deviating from the original definition based on the sperm swimming path, as we observed sperm fixed to the glass slide. These results demonstrate that Calaxin’s calcium-binding ability is necessary to generate the calcium-induced asymmetric waveform.

calaxin^−/− sperm model has a highly asymmetric and inactive distal region

calaxin ^−/− sperm model contained an active proximal region and an inactive distal region. To further investigate this phenotype, we compared WT and calaxin ^−/− sperm models. The distal region of the calaxin ^−/− sperm model showed higher asymmetry than that of WT in the pCa4 condition, but both genotypes showed similar waveforms in the proximal region (Fig 3A and D and Video 2). Moreover, the dynamic component of the calaxin ^−/− sperm model was not uniform, indicating that the proximal and distal regions show different beating patterns (Fig S2F). This distal region seemed to coincide with the distal OAD(−) IAD(+) region shown in the IFM experiment. The length of the inactive region differed between flagella, possibly reflecting the variable length of the OAD(−) region (Fig 1C and D). From these results, we speculated that IAD affects waveform asymmetry calcium-dependently.

IAD contributes to calcium-dependent flagellar asymmetry

To further test the contribution of IAD to the asymmetric waveform, we observed the armc4^−/− zebrafish sperm model and quantified the basal curvature. Armc4 is a major component of the OAD-DC (Gui et al, 2021), and armc4^−/− zebrafish sperm axoneme does not contain OAD, whereas it retains IAD and other DMT appendages (Yamaguchi et al, 2023). Armc4^−/− sperm model showed highly curved flagella in the EGTA condition, and the pCa4 condition further increased the curvature (Fig S3A and B and Video 3). Most of the armc4^−/− sperm model was immotile or only showed irregular, vibrating movement (Fig S3C and Video 3). The addition of orthovanadate, a dynein ATPase inhibitor, resulted in straight and immotile flagella in both EGTA and pCa4 conditions, suggesting that ATP-dependent mechanisms generate such curvatures (Fig S3A and B and Video 3). To quantify the basal curvature of immotile or irregularly beating flagella, we randomly selected one frame from each video and traced the flagellum. These results reinforce the idea that IAD alters waveform asymmetry calcium-dependently.

Calaxin works as a calcium-dependent dynein regulator and a calcium-independent dynein stabilizer

Our results showed that zebrafish Calaxin has three calcium-binding sites that are necessary for the calcium-induced asymmetric beating but dispensable to stabilize OAD. Only the sperm model expressing calcium-binding–deficient E130A-Calaxin showed an incompletely asymmetric waveform in response to calcium, whereas those expressing WT-Calaxin exhibited a normal asymmetric beating. Combining this with the ITC experiment, our results directly linked Calaxin’s calcium-binding ability to the calcium-induced asymmetric beating. Considering that Ciona Calaxin inhibits OAD activity in response to calcium (Mizuno et al, 2012), it appears Calaxin binds to calcium and inhibits OAD to elevate the waveform asymmetry when intraciliary calcium concentration increases.

We found that the Tg E130A sperm model exhibited a uniform and symmetric dynamic component even in the presence of calcium, suggesting that Calaxin’s calcium-binding ability does not influence the dynamic beating (Fig S2F). This is inconsistent with the previously proposed model, where Calaxin influences the bend formation and propagation and triggers an asymmetric beating calcium-dependently (Mizuno et al, 2012; Shiba et al, 2023). Instead, our results imply that Calaxin’s calcium-binding ability is necessary to regulate the basal curvature but dispensable for the bend propagation. The difference may result from different model species, zebrafish we used and C. intestinalis in existing studies. Alternatively, the E130A-Calaxin we used might more selectively disrupt the calcium-dependent functions of Calaxin while maintaining its calcium-independent functions.

Our results also suggested that Calaxin is not the sole calcium-dependent axonemal dynein regulator. Both WT- and E130A-Calaxin–expressing sperm models showed a slower beating under the presence of calcium (Fig S2C). Moreover, calcium-induced weak asymmetry was observed in the E130A-Calaxin–expressing sperm model, although it was not statistically significant (Fig 3C and E). These results imply that there are calcium-dependent regulators other than Calaxin. IAD is one of the candidates because calcium elevated the asymmetry of the OAD(−) armc4^−/− sperm model and the distal region of the calaxin ^−/− sperm model (Figs 3D and S3A and B). Axoneme also contains other calcium-binding proteins, such as calmodulin (Pazour et al, 2005).

Another important finding is that Calaxin works as a calcium-independent OAD stabilizer and a calcium-dependent dynein regulator. Sperm flagella derived from calaxin ^−/− zebrafish partly lack OAD heavy chains Dnah8 and Dnah9. Intriguingly, we have previously shown that OADs are also reduced in the proximal region of calaxin^−/− flagella (Yamaguchi et al, 2023). This indicates that Calaxin knockout destabilizes OADs in any region, and some other factors may accelerate the phenotype in the distal region. Because IFT machinery is known to disappear as the spermatozoa mature (San Agustin et al, 2015), we speculate that the supply of OAD during spermiogenesis decreases as the sperm flagella elongate, leading to the shortage of OAD in the distal flagella to compensate for the instability of calaxin^−/− OADs. E130A-Calaxin rescued this phenotype (Fig 1C and D), indicating that Calaxin stabilizes OAD independent of calcium. Further supporting this idea, the expression of E130A-Calaxin rescued the abnormal motility of KV cilia and the laterality randomization of the calaxin ^−/− mutant (Fig 2A–D), which possibly resulted from the reduction of OAD. It is likely that the Calaxin knockout reduced OAD and the reduction resulted in incomplete bend propagation and abnormal movements of KV cilia and reactivated sperm flagella. The axoneme from human PCD patients also shows the reduction of OAD (Hjeij et al, 2023), suggesting such stabilizing function is conserved among vertebrates. It also implies that Calaxin does not work as a calcium-dependent dynein regulator in KV cilia. Consistently, calcium depletion does not alter KV ciliary motility (Yuan et al, 2015) and KV cilia exhibit uniform rotation without waveform switching.

How Calaxin-dependent dynein inhibition is transformed into an asymmetric waveform remains to be determined. One hypothesis is that the localization of OAD components is not uniform, resulting in the uneven inhibition of dynein activity. However, structural studies of sperm flagella did not show such asymmetry (Chen et al, 2023; Yamaguchi et al, 2023). If Calaxin localizes to specific DMTs, it is likely that loss of Calaxin does not affect overall OAD conformation. Instead, calaxin ^−/− sperm partly lacks OAD in their distal region, suggesting that Calaxin exists on all DMTs. Another possibility is that asymmetric axonemal structures such as o-SUB5-6 in sea urchin flagella (Lin et al, 2012) transform the global inhibition of dynein activity into an asymmetric waveform.

Another limitation of this study comes from the sperm model we used to precisely control the calcium condition. In WT, although the waveform of the sperm model was similar to that of intact sperm, it exhibited a much slower beating compared with intact sperm (Fig S2C). In armc4^−/− sperm, the frequency of beating flagella was also strongly reduced (Fig S3C) (Yamaguchi et al, 2023). One possibility is that the reactivation solution does not completely mimic the intraciliary environment. Alternatively, the loss of plasma membrane may accelerate the disruption of axonemal dyneins under the depletion of Calaxin or Armc4. It suggests the demembranated sperm model is not a completely physiological system, and analyzing intact sperm obtained from mutants we used may contribute to further understanding.

Zebrafish maintenance

Zebrafish were maintained at 28.5°C. Embryos and larvae were raised at the same condition in 1/3 Ringer’s solution. Embryos were staged as previously described (Kimmel et al, 1995).

Plasmids

calaxin was amplified from Danio rerio testis cDNA with calaxin_cloning.F (GCAACATTCAACCAAAGACTAAAAAGAG) and calaxin_cloning.R (GAGGGAAAACAACAAATTGTCAAAC) primers and cloned into the pCR-Blunt vector. pCR-Blunt-calaxin was digested with EcoRI and subcloned into the pCS2+ EcoRI site. For protein expression, testis cDNA was amplified with calaxin_EcoRI.F (GGGAATTCCCATGCTGAAAATGTCGGCGATG) and calaxin_EcoRI.R (GGGAATTCTTATTCTTTGCAGTGTTCGTGTTTCTG) primers, digested with EcoRI, and cloned into the pGEX 6P-2 (28-9546-50; Cytiva) EcoRI site. E130A mutation was introduced by PCR with calaxin_E130Amut.F (CACGAGAAGCCATGTTTCAGATGCTGAAAGACAGC) and calaxin_E130Amut.R (ACATGGCTTCTCGTGAGATGTATCCATCCCCG) primers. To generate a Tol2 donor plasmid, a 5.3-kb promoter element from the β-actin gene (Higashijima et al, 1997) was cloned from the zebrafish genome with actb2p5304_clo.F (AATTCCAGTTTGAAGAAACTTTTC) and actb2p5304_clo.R (GGCTGAACTGTAAAAGAAAGG) primers. The amplified fragment was combined with Tol2 elements and inserted into the pBlueScript SK(+) vector (Kawakami, 2007). The resulting pBlueScript SK(+)-Tol2-actb plasmid was amplified with pBSKTol2actb(calaxin).F and pBSKTol2actb(calaxin).R primers to add Kozak (GCCACC) sequence. Calaxin ORFs were amplified from pGEX 6P-2 calaxin WT or E130A with calaxin_ORF.F (ATGCTGAAAATGTCGGCGAT) and calaxin_ORF.R (TTATTCTTTGCAGTGTTCGTGTTTCT) primers. These fragments were ligated with NEBuilder (E2621; New England Biolabs). All plasmids were sequenced before they were used.

mRNA synthesis

calaxin WT and E130A mRNA were transcribed from pCS2+-calaxin WT and E130A plasmids using the mMESSAGE mMACHINE SP6 kit (AM1340; Invitrogen) following the manufacturer’s instruction. The template plasmid was digested with DNase I, and mRNA was purified by ethanol precipitation. mRNA was quantified by Abs260 and stored at −80°C before use.

Mutagenesis

To produce Tg WT or Tg E130A, calaxin ^−/− embryos were injected at the one-cell stage with 20 pg Tol2 transposase mRNA and 30 pg pBlueScript SK(+)-Tol2-actb-calaxin WT or E130A plasmid. Injected embryos were raised and crossed with calaxin ^−/− fish. F1 fish were screened for integration by PCR with calaxin_check.F (GGAGAGCAGGCAGAGAGAAAG) and calaxin_D164_check.R (GGTTTTCATCCATGACCGTCTTCTC) primers. PCR-positive male F1 fish were further subjected to sperm immunofluorescence with anti-Calaxin antibody, and those with Calaxin-positive sperm were used for experiments.

Sperm collection

To collect sperm, adult zebrafish were anesthetized with tricaine. Sperm were collected by pushing their abdomen or euthanizing and dissecting their testis in 5 μl PBS. From the pushing method, 0.1–0.5 μl of sperm was obtained from each fish.

Immunofluorescence

The collected sperm were diluted with Hank’s buffer so that 10–20 sperm/view were observed at x200 magnification. 40 μl of the sperm solution was placed on the glass slide coated with 0.01% polyethyleneimine for 10 min. After the solution was removed, the attached sperm were fixed and demembranated with Hank’s buffer containing 4% PFA and 0.05% NP-40. The slide was washed and dehydrated with acetone followed by methanol at −20°C. Then, it was rehydrated with PBS, blocked with blocking buffer (PBS/0.05% Tween-20, 2% cold water fish gelatin, 4% normal goat serum), and incubated with primary antibodies (blocking buffer with 1/2,500 anti-acetylated alpha-tubulin antibody and 1/50 polyclonal antibodies) at 4°C overnight. After the wash with PBS/0.05% Tween-20 (PBST), it was incubated with secondary antibodies (blocking buffer with 2.5 μg/ml DAPI and 1/250 secondary antibodies corresponding to primary antibodies) at RT for 2 h and washed twice with PBST and once with PBS. Images were obtained with a fluorescence microscope (BX60; Olympus) equipped with a 40x NA0.75 objective lens and a CCD camera (ORCA-R2; Hamamatsu and Micro-Manager software).

Zebrafish demembranated sperm model

Zebrafish sperm were suspended in 5 μl PBS. To demembranate the sperm, 1 μl of the suspension was added to 20 μl HMDK (30 mM Hepes–NaOH, pH 7.2, 5 mM MgSO₄, 1 mM DTT, 50 mM KAcO) with 0.1 mM EGTA and 0.5% NP-40. For activation and observation, 1 μl demembranated sperm and 10 μl of activation solution (HMDK, 0.5% NP-40, 0.1 mM EGTA or CaCl₂, 50 μM ATP, 50 μM ADP) were mixed for activation and recorded with a phase-contrast microscope (BX53; Olympus) equipped with a 100x NA1.35 objective lens and a high-speed camera (EoSens MC1362; Mikrotron and StreamPix 5 software) at 100 fps, 640 × 480 pixels. For inhibition of dynein with vanadate, 10 μM sodium orthovanadate was added to the activation solution. Demembranated sperm was kept on ice and used within 1 h after the sperm collection.

Waveform analysis

Recorded videos of flagella were converted into 2,560 × 1920 and manually traced with ImageJ. Traces were analyzed with self-made Python programs (Fig S2E). First, the LineTrace_4x.py program calculated the tangent angle and the cumulative length from the head on every 10 pixels (∼0.4 μm). To obtain the tangent angle, the tangent line was drawn as an average of 20 pixels (∼0.8 μm). The tangent angle is the angle between the tangent line and the horizontal axis of the image. Next, tangent angles from ≧5 time points in one beating cycle were combined and fitted with the linear equation by the PlotLine.py program. The tangent angle plot and the fitted equation were shifted so that the equation starts from (x, y) = (0, 0). “Basal curvature” was defined as the slope of the fitted equation. To calculate the asymmetry index, sequential 20 pixels were fitted with a circle and the curvature was defined as 1/r, where r is the radius of the fitted circle. The asymmetry index was defined as the ratio of the maximal to the minimal value of the curvature at 10 μm from the head. We assumed the bend with the maximal curvature is “P-bend” and the one with the minimal curvature is “R-bend.” Python scripts we used are available on GitHub (https://github.com/motohiromorikawa/Calaxin-analysis).

KV ciliary motility

Zebrafish Kupffer’s vesicle was observed as previously described (Sasaki et al, 2019). Briefly, 8–10 somite embryo was fixed on a glass dish by 1.0% low melting point agar in 1/3 Ringer’s solution, recorded with a bright-field microscope (DMI6000B; Leica) equipped with a 100x NA1.4 objective lens and a high-speed camera (HAS-L1; DITECT and HAS-L1 Basic software) at 1,000 fps. For mRNA-injected embryos, 10 pg of mRNA was injected at the one-cell stage.

Purification and ITC experiment of Calaxin

E. coli BL21 cells transformed with pGEX 6P-2 calaxin WT or E130A were cultured in LB medium at 37°C to reach Abs 0.5 at 600 nm, followed by induction with 0.1 mM IPTG. For WT, cells were cultured at 37°C for 3 h after the induction. For E130A, they were cultured at 20°C overnight. Harvested cell pellets were sonicated, clarified, and glutathione-affinity–purified (GST-Accept, 09277; Nacalai Tesque) in lysis buffer (PBS with 150 mM NaCl and 1 mM DTT) supplemented with 0.2x protease inhibitor cocktail (03969; Nacalai Tesque). The column was sequentially washed with lysis buffer with 0.01% NP-40 and 0.2x protease inhibitor, lysis buffer with 0.01% NP-40, 1 mM EDTA, and 0.2x protease inhibitor, lysis buffer with 0.01% NP-40 and 1 mM EDTA, lysis buffer with 0.005% NP-40, and lysis buffer with 0.001% NP-40. After the wash, lysis buffer with 0.001% NP-40 and HRV 3C Protease (7360; TaKaRa) was added for on-column digestion at 4°C overnight. The eluted protein was applied to TALON Metal Affinity Resin (635501; TaKaRa) to remove the protease. The resulting protein was concentrated, and the buffer was exchanged for storage buffer (PBS with 150 mM NaCl) by Vivaspin 2–10K (VS0201; Sartorius), then stored at −80°C. To check the quality and mobility of recombinant proteins, SDS–PAGE was conducted with Extra PAGE One Precast Gel (13064-64; Nacalai Tesque). Proteins were mixed with the sample buffer containing 1 mM CaCl₂ or EGTA and denatured at 95°C for 5 min before electrophoresis. Before the ITC experiment, Calaxin was supplemented with 0.5 mM EGTA to remove residual calcium and buffer-exchanged on Superdex 200 10/300 GL (17517510; Cytiva) equilibrated with ITC buffer (25 mM MOPS-KOH, pH 7.8, 200 mM NaCl, 1 mM DTT). MicroCal iTC200 and Origin software (GE Life Sciences) were used for the ITC experiment and analysis as previously described (Mizuno et al, 2009).

Antibodies

To visualize the axoneme, a mouse monoclonal anti-acetylated alpha-tubulin antibody (T6793; Sigma-Aldrich) was used. Other polyclonal primary antibodies are as follows: guinea pig anti-Calaxin antibody (Sasaki et al, 2019) (referred to as anti-Efcab1 antibody), rabbit anti-Dnah2 antibody (Yamaguchi et al, 2018), rabbit anti-Dnah8 antibody (Yamaguchi et al, 2018), and rabbit anti-Dnah9 antibody (Yamaguchi et al, 2023). As secondary antibodies, goat anti-rabbit IgG Alexa Fluor 488 (A-11008; Invitrogen), goat anti-guinea pig IgG Alexa Fluor 488 (A-11073; Invitrogen), and goat anti-mouse IgG Alexa Fluor 555 (A-21422; Invitrogen) were used.

Statistics

Statistical analyses were conducted with R version 4.1.2. Statistical significances were tested by an unpaired two-tailed t test. For multiple comparisons, P-values were adjusted by the Holm–Bonferroni method. P-value < 0.05 was considered to be significant.