Interaction between ciliary component proteins from Chlamydomonas revealed by CRISPR/CAS9, cryo-electron tomography and mass spectrometry

To understand molecular mechanism of ciliary beating motion, knowledge of location, interaction and dynamics of >400 component proteins are indispensable. While recent progress of structural biology revealed conformation and localization of >100 proteins, we still need to investigate their networking, art of their interaction and assembly mechanism. We applied CRISPR/CAS9 genome editing technique to the green algae Chlamydomonas to engineer a deletion mutant of a ciliary component, FAP263, located at the distal protrusion, and examined it structurally by cryo-electron tomography (cryo-ET) and mass spectrometry (MS). Cryo-ET and atomic model fitting demonstrated that the FAP263 deletion mutant lacks additional components, FAP78, and FAP184. Unassigned density near FAP263 in the cryo-ET map of WT cilia is likely FAP151, as suggested by cross-linking mass spectrometry. Based on the structure, we modeled how these four proteins might form a complex. Furthermore, it was shown that dynein f phosphorylation is inhibited in the FAP263 mutant, indicating an important role of this protein complex for dynein f phosphorylation. Our study demonstrates a novel approach to investigate protein networking inside cilia.


Introduction
Motile cilia, beating organelle to cause either swimming of the cells or extracellular fluid flow, are composed of more than 400 proteins (Pazour et al, 2005).Ciliary motion is considered to be product of dynamic interactions and networking of these proteins.There has been wide variety of attempts to reveal the protein interactions, 3D arrangements and their dynamic changes.Recent progress of single particle cryo-EM analysis and folding prediction by Alphafold2 allowed visualization of >100 components which exist in the peripheral microtubule doublet with 96nm periodicity (Walton et al, 2023), components of the central pair apparatus with 32nm periodicity (Gui et al, 2022), and intraflagellar transport complexes (Hesketh et al, 2022).Cryo-electron tomography of intact cilia, combined with high resolution obtained by single particle cryo-EM, enabled modeling of conformational change of dynein and associated proteins during motion (Zimmermann et al, 2023).Further expansion of our knowledge to cover more component proteins at various states of beating motion and at various stages of ciliogenesis is awaited both for basic understanding of ciliary movement and medical investigation of ciliopathy (Wallmeier et al, 2020).
Both structural and functional studies utilize deletion mutants.By comparing the wild type and deletion mutants, or by comparing the wild type and a strain with a deletion mutant rescued by a tagged gene, location of the target protein can be detected based on decrease or increase of density in the 3D map, respectively.In cilia research, combination of deletion mutants and cryo-ET revealed functional roles and locations of a number of component proteins in intact cilia (Ishikawa, 2016).In the case target proteins are small and hard to detect as a loss of density in deletion mutants or in case deletion could cause collapse of other proteins in the complex, genetic tag to increase density helps efficiently.
Components of the radial spoke (Oda et al, 2014c), the dynein regulatory complex (Oda et al, 2014b), ruler proteins for the doublet microtubule (Oda et al, 2014a) and an inner dynein scaffold protein (Kutomi et al, 2021), were located by cryo-ET in this way.
However, availability of deletion mutants for motile cilia research was severely limited.
Chlamydomonas reinhardtii has been the most popular model organism for motile cilia research because of abundance of mutants of ciliary components, which were isolated based on motility defect.
Active mutagenesis of Chlamydomonas by targeting gene of interest, however, is complicated because homologous recombination of this species is not established.While there are wide variety of deletion mutants, first induced chemically or by radiation and later by random insertion (Li et al, 2016), systematic methods to delete targeted gene have been unavailable for years.Meanwhile Tetrahymena thermophila, another popular model organism for motile cilia research can be mutated in targeting genes using homologous recombination, which has been used for biochemical research by purifying mutated proteins (Ichikawa et al, 2015).For cellular studies such as cellular cryo-ET of cilia, mutation of Tetrahymena cannot be used easily, since there are 50 copies of each gene in the small nucleus and complete exclusion of the intrinsic sequence is not straightforward.
In this study, we generated deletion mutant of a ciliary component FAP263 from Chlamydomonas, which is located near the outer surface of the A-tubule and dynein f, using CRISPR/CAS9.We characterized these deletion mutants biochemically and structurally.Our study demonstrates proof-ofprinciple of combination of CRISPR/CAS9 and cryo-ET for cilia research, as well as its advantage to study influence of gene deletion to other component proteins.

Results and discussion
We made deletion mutant of FAP263 by inserting stop codon to Chlamydomonas genome by CRISPR/CAS9 following the protocol of (Shin et al, 2016).Mutation was confirmed by PCR and sequencing (Fig. 1).
After back-crossing the mutant to WT to minimize a risk of off-target effects, we further structurally analyzed FAP263 deletion mutant by cryo-ET and subtomogram averaging (Figs.2, 3A, Supplementary Fig. 1).Protein components can be located in the cryo-ET map by fitting high-resolution single particle cryo-EM structure of split doublet microtubule from Chlamydomonas cilia (Walton et al, 2023).There is an area, where density exists in WT, but not in the FAP263 deletion mutant (Fig. 2; Supplementary Fig. 1) which likely corresponds to proteins lost by FAP263 deletion.In the fitted atomic model from single particle cryo-EM, this lost area corresponds to FAP78/FAP184 and FAP263 (Fig. 3B).Therefore this density can be explained as a complex of FAP263, FAP78 and FAP184.
However we found a small unassigned region as well (Fig. 3B).This unassigned density is positioned between FAP78 and FAP263.To find candidates for an additional protein located in this region, we performed cross-linking mass spectrometry (Leitner et al, 2014) and looked for proteins that were identified in earlier proteomics work on Chlamydomonas cilia (Pazour et al, 2005), but were not localized in the cryo-EM structure.The cross-link data points to a putative interaction between FAP78 and the protein FAP151 (Fig. 4), which was not involved in the list of single particle cryo-EM analysis.
It should be pointed out that the confidence of this identification is relatively low, which may be attributed to the low abundance of the protein complex in cilia, and/or specific mass spectrometric properties of the cross-linkned peptides.Nevertheless, we modeled FAP151 by Alphafold2 and fitted to the unassigned density (Fig. 3B).FAP151 fits well to other components and is likely a piece of this complex (Fig. 3CD).Since FAP78 is solved in full-length by single particle cryo-EM (Walton et al, 2023), this density may contain a part of FAP78 as well.
Furthermore, we performed phosphoproteomics experiments to study the role of protein phosphorylation on putative functions of the subcomplex involving FAP263.Comparative mass spectrometric analysis of phosphopeptides from wild type and FAP263 mutant samples after enrichment with titanium dioxide (Leitner et al, 2010) resulted in a high coverage of the Chlamydomonas phosphoproteome.Four proteins in the distal protrusion have more than one assigned spectrum for peptides that are detected phosphorylated in WT but not in the FAP263 mutant (Table 2).Among dynein isoforms, dynein f (DHC10) is located close to FAP263 and likely influenced by its deletion for phosphorylation.Indeed dynein f shows phosphorylation in WT, but not upon FAP263 deletion (Table 1).This suggests that the FAP78/FAP151/FAP184/FAP263 complex is responsible for dynein f phosphorylation.Among these four proteins, FAP78 is likely to have kinase activity, according to classification by Panther (https://phytozomenext.jgi.doe.gov/report/gene/Creinhardtii_v5_6/Cre12.g536600).However, phosphorylation by another protein located near this complex, for example a Nima kinase CNK4, cannot be excluded.
Although we have not seen visual difference of swimming property between WT and the FAP263 deletion mutant, change of beating frequency and amplitude caused by mutation of Ccdc113/Ccdc96 (corresponding to FAP263/184) in Tetrahymena was reported (Bazan et al, 2021).While phosphorylation of IC138, associated with dynein f, was reported (Bower et al, 2009), phosphorylation of dynein f itself has not been studied.Functional roles of this complex is still to be investigated.
In this study, we demonstrated CRISPR/CAS9 of Chlamydomonas is useful for structural research of cilia.The FAP263 deletion causes loss of FAP78, FAP151, and FAP184.This approach can be applied to solve various unanswered questions, such as precise location of various species of dyneins, role of individual dyneins and regulatory proteins in ciliary motion, as well as ciliogenesis.

Figure caption
Fig. 1 Sequence of the FAP263 deletion mutant by CRISPR/CAS9.crRNA design crRNAs were chosen using Benchling (Hirano et al, 2019).On-target and off-target scores calculated by Benchling against Chlamydomonas r. genome were aimed to be above 60 and 90, respectively.crRNAs were purchased from Integrated DNA Technology's (IDT) online custom Alt-R® CRISPR-Cas9 guide RNA tool.

Design of repair template and preparation
The repair template was designed to have two homology arms upstream and downstream of the cutting side, each with a length of 25 bp.In the middle of the homology arms is a FLAGtag with two stop codons.The FLAG-tag serves the detection of mutant colonies by polymerase chain reaction (PCR).The repair template was ordered as a forward and reverse oligonucleotide on Microsynth AG.The oligonucleotides were ordered with 3 phosphorothioate bonds on both 5' and 3'.To anneal the oligonucleotides, 1 µL of forward and 1 µL of reverse oligonucleotides were mixed in 18 µL IDT RNA duplex buffer.Afterwards the mix was heated to 95 °C for 2 minutes and cooled down at 0.1 °C/sec to room temperature.

Paromomycin cassette for selection
The plasmid pSI103-1, which confers paromomycin resistance, was ordered from Chlamydomonas Resource Center.The resistance cassette was amplified using chemically competent E. coli.The plasmid was extracted with a plasmid extraction kit from QIAGEN.
Afterwards, the DNA is linearized by KpnI-HF from Biolabs.The linearized DNA was concentrated to 1 µg/µL by ethanol precipitation.Transfer on TAP agar plates After recovery we centrifuged the cells at 800 g for 5 minutes and plated them on 1.5% TAP agar plates with 10 µg/mL paromomycin.Colonies can be picked after 5-7 days and transferred to 96 well plate with TAP.Confirmation of mutant colonies was done by PCR using Phusion™ High-Fidelity DNA Polymerase.Screening for mutants were done by gel electrophoresis.To confirm mutants, they were sent for sequencing at Microsynth AG.

Cell culture and harvesting cilia
Chlamydomonas cell culture and cilia isolation were followed by Witman's protocol (Witman, 1986).

Cryo-ET
Cryo-ET grid preparation and data acquisition were after our previous work (Zimmermann et al, 2023), using one-sided manual blotting and freezing by Cryo-plunge (Gatan, USA) and the Titan Krios G3 transmission electron microscope (TFS, USA) with Quantum energy filter (Gatan).Subtomogram average using pseudo nine-fold symmetry and 96nm periodicity of cilia was carried on using our algorithm published previously (Bui & Ishikawa, 2013;Zimmermann et al, 2023).

Cross-linking mass spectrometry
C. reinhardtii strain cc124-was cultured for 3 days.Cilia were isolated by dibucaine.Isolated cilia by dibucaine were treated with 1% OGP in equal volume to remove cell membrane.
cOmplete™ Proteasehemmer-Cocktail by Roche were used to stop protein degradation.Protein concentration was measured using BCA assay and adjusted to 0.5-2 mg/ml with a total amount of 50-100 µg protein.Cross-linking experiments were performed at room temperature with the amine-reactive disuccinimidyl suberate in isotopically light and heavy form for 1 hour (DSS-d0/d12, Creative Molecules).Afterwards the protocol by (Leitner et al., 2014) was used to process the samples further.Samples were analyzed by liquid chromatography-tandem mass spectrometry on an Orbitrap Fusion Lumos instrument (ThermoFisher Scientific), and MS data was analyzed by xQuest.

Phosphoproteomics
C. reinhardtii strain cc124-was used as wild type control for this experiment.Three replicates of the cc124-strain and three replicates of the FAP263 mutant strain were cultured in 300 ml of TAP medium.Each culture was maintained for a duration of three days.Isolation of cilia and sample preparation were done as described above.To prevent dephosphorylation, PhosSTOP™ from Sigma Aldrich was added to sample.Amount of protein was adjusted to performed following (Leitner et al., 2010) using Titansphere TiO material (GL Sciences) for enrichment.Samples were analyzed by liquid chromatography-tandem mass spectrometry on an Orbitrap Fusion Lumos instrument (ThermoFisher Scientific), and MS data was analyzed by FragPipe/MSFragger.For data analysis a cutoff of Peptide Prophet Probability > 0.95 was chosen.For comparison of phosphorylation states between wild type and FAP263 mutant the total amount of peptide spectrum matches of phospho peptides were counted.
Preparation of Cas9/gRNA RNP gRNA was assembled by mixing 5 μL of 100 μM IDT Alt-R® crRNA with 5 μL of 100 μM IDT Alt-R® tracrRNA.The mix was incubated at 95 °C for 2 minutes and cooled down at 0.1 °C/sec to room temperature.4.5 μL of 10 μM gRNA was mixed with 4.5 μL of 10 μM IDT Alt-R® Cas9, 1.5 µL of 10x NEB 3.1 and 4.5 µL of ddH2O.The RNP system was incubated at 37 °C for 15 minutes.Transformation by electroporation 150 mL of cw92 cells were grown for 3 days in TAP medium.Cells were centrifuged at 800 g for 5 minutes.The cells were washed once with electroporation buffer (30 mM Hepes, 5 mM MgSO4, 50 mM K-acetate, 1 mM Ca-acetate, 60 mM Sucrose).Afterwards the cells were pelleted again and resuspended in 2-3 times the volume of cells in electroporation buffer.Cells were diluted to a concentration of 3 * 10 8 cells/mL.Then they were incubated at 40 °C for 30 minutes at 120 rpm. 100 µL of heat shocked cells, 15 µL of RNP, 4.5 µL of repair template and 1 µg of pSI103-1 were mixed in a cuvette fromBio-Rad (Catalog No. 165-2086).The mixture was electroporated using ECM® 630 from BTX -Harvard Apparatus with the following conditions: 410 V low voltage, Resistor 25 Ohm, Capacitance 600 uF.After electroporation cells were incubated for 1 hour at 15 °C.The cells were transferred into 10 mL of 60 mM TAP sucrose for recovery overnight under constant light and shaking.

Table 2 .
Phosphorylation of possible distal protrusion proteins.