PCD variants in DNAH5 do not affect airway cell differentiation fate. To link transcriptional profiles to specific cell types in PCD cells, we performed scRNAseq in airway cells from 4 unrelated patients with pathogenetic variants in DNAH5 (Figure 1A and Supplemental Table 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.180198DS1.). All patients were confirmed to have PCD using established genetic and clinical criteria (17) (Supplemental Table 1). For comparison, we obtained nasal cells from 5 unrelated healthy individuals who did not carry pathogenic alleles in 53 known PCD genes, confirmed by whole exome sequencing. As additional controls, we also obtained cells from the heterozygous mother of each participant with PCD.

Primary airway cell clustering. (A) Schematic of showing nasal cell brush biopsy, culture, and analysis. (B) UMAP for dimension reduction of cultured nasal cells showing unsupervised cell clustering and distribution of basal (Bas), secretory (Sec), multiciliated (Cil), dividing (Div), and ionocytes (Ion) (n = 5 samples, approximately 7,000 cells/sample). Colors represent different subclusters. (C) Heatmap of identified cell clusters shown in the UMAP, showing top expressed genes per cell subcluster. Unique transcriptional signatures differentiated the different subclusters. (D) UMAPs comparing cell distribution between normal control cells, heterozygous DNAH5 mothers, and homozygous DNAH5 patients’ cells show no differences in cell subclusters between groups (n = 5 for normal cells, n = 4 for heterozygous, and n = 4 PCD cells). (E) Cell numbers of the major cell types comparing the different experimental groups. (F) Dotplot comparing the different cell subclusters of each submitted sample between the indicated experimental groups.

The airway has a complex interplay between epithelial cells, the local immune cells, and the microbiome of the upper airway. Our goal was to identify epithelial cell responses to the gene variant and motile cilia dysfunction rather than the response to the local environment. Therefore, nasal cells were first expanded in culture at least 10-fold in medium containing antibiotics and antifungal drugs for up to 10 days to minimize the acute effects of the local airway environment. Cells were then differentiated into secretory and multiciliated cells using air-liquid interface (ALI) conditions for another 28 days (18) (Supplemental Figure 1A). Cultured PCD cells showed no cilia motility (as expected) compared with the presence of cilia activity in their heterozygous mothers and healthy controls (Supplemental Figure 1B).

Unsupervised scRNAseq clustering and uniform manifold approximation and projection (UMAP) identified expected airway epithelial cell types as defined by known marker genes, including basal (Bas), secretory (Sec), and multiciliated cells (Cil), as well as a small population of ionocytes (Iono) and dividing cells (Div) (Figure 1B). Within each cell type, multiple subclusters were identified with distinct markers (Figure 1C). These subclusters likely represent different stages of cell differentiation yet show differences in gene expression. There were no significant differences in cell numbers within each cell type (basal, secretory, ciliated) among samples from patients, their heterozygous parents, or controls (Figure 1, D and E, and Supplemental Figure 1C), indicating that the general program of airway differentiation are conserved in PCD DNAH5 variants. Examining the top markers per cell cluster and considering each of the sequenced samples, we found that the clusters’ identities were broadly preserved among the different samples (Figure 1F).

Transcriptional analysis identifies unique subcell clusters of ciliated cells. To identify potential differences among our study groups and the impact of the DNAH5 allele, we further analyzed the subclusters of multiciliated cell (Cil1–Cil5). We first determined if the cell subclusters can be defined by the known sequence of transcription factors expression and their relationship to the multiciliated cell assembly pathway (19) (Figure 2A). We defined multiciliated cells by the expression of the transcription factor FOXJ1 (Figure 2B). Unsupervised clustering identified 5 subclusters of multiciliated cells that could be distinguished by the pattern of expression of differentially expressed genes (Figure 1C). Using known transcription factors and motile cilia markers, we defined these subclusters as “basalociliated cells” (FOXJ1+ P63+) (Cil1), “secretociliated cells” (FOXJ1+ SLURP2+) (Cil2), “deuterostomal cells” (FOXJ1+ CCNO+ DEUP1+ PLK4+) (Cil3), early mature ciliated cells (FOXJ1+ DNAH7+, low CFAP141(c1ORF189)) (Cil4), and late mature ciliated cells (FOXJ1+ DNAH7+, high CFAP141) (Cil5) (Figure 2C and Figure 1C). Several microtubule inner protein (MIP) genes were highly expressed in late mature ciliated cells (Cil5) compared with early mature ciliated cells (Cil4) subclusters, including CFAP276 (C1ORF194), CIMIP1(C20ORF85), and CFAP141(C1ORF189). Several MIP proteins are found in the lumen of axonemal microtubules and are thought to organize the periodicity of the ciliary motors and their regulator components (20), which may suggest their requirement at later stages of cilia assembly.

Multiciliated cells single cell analysis and velocity trajectories. (A) Scheme depicting the canonical differentiation pathway of multiciliated cells. (B) FOXJ1, the ciliogenesis master regulator, is shown overlayed on the scRNAseq UMAP to show the multiciliated cell clusters. (C) UMAP of extracted FOXJ1+ cells shows 5 unique multiciliated cell subclusters, marked Cil1 for (FOXJ1+ P63+) basalociliated cells, Cil2 for (FOXJ1+ SLURP2+) secretociliated cells, Cil3 for (FOXJ1+ CCNO+ DEUP1+ PLK4+) deuterostomal cells, Cil4 for (FOXJ1+ DNAH7+, low c1orf189) early mature multiciliated cells, and Cil5 for (FOXJ1+ DNAH7+, high c1orf189) late mature multiciliated cells. (D) Violin plot showing known multiciliated cell differentiation transcriptional markers in each of the multiciliated cell subclusters. (E) UMAP showing expression levels of GMNC and E2F1 expression levels in airway cells. (F) Pseudotime analysis showing projected differentiation trajectory of cells overlayed on the UMAP of cell clusters showing no trajectories from secretory cells into ciliated cells. (G) Velocity analysis showing unsupervised high-dimensional vectors predicting future state of individual cells, demonstrating a direct relationship between differentiation of basal cells and multiciliated cells.

Analysis of the expression levels of defined ciliogenesis transcription factors suggested a sequential transition from Cil1 to Cil5 (Figure 2, D and E). Cil1 subcluster had a higher expression of the basal cell marker TP63. Moreover, there was a pattern of increasing expression of MYCL between Cil1 and Cil3, a transcription factor that was recently shown to be active at the branch point of basal to multiciliated cell fate (21). Pseudotime and RNA velocity analysis of all clusters indicated that a high percentage of multiciliated cells likely differentiate directly from basal progenitor cells through an intermediary dividing cell cluster (Div) without transitioning through secretory cells (Figure 2, F and G) (21). Multiciliated cells with homozygous variants in DNAH5 did not use different pathways of differentiation compared with the other groups.

DNAH5 variant airway cells have unique gene expression compared with the other groups. To identify pathways that are perturbed as the result of variants in DNAH5, we compared the differential gene expression among the 3 groups of cultured airway cells using unsupervised analysis. Differential gene expression was identified in all major cell types among PCD, maternal, and control cells (Supplemental Figure 2A). Examining the most differentially expressed genes between groups, we found that differences between PCD cells and control cells indicated a clear distinction (Figure 3A), with a high degree of concordance among different individuals with PCD (Figure 3B). Since our control samples had a male bias, we performed a subanalysis between the different groups, using samples from women only (including additional control samples from published data sets from women (22, 23), identifying similar results (Supplemental Figure 2B).

scRNAseq comparison between PCD and non-PCD cells. (A) Heat map showing the top differentially expressed genes in ciliated, secretory, and basal cells of PCD patient primary nasal cells compared with heterozygous maternal cells and healthy control cells demonstrating unique differences among the different groups. (B) Dotplot depicting the transcript expression levels of all samples analyzed in A showing the highest differentially expressed genes per sample. (C) Pathway analysis of the top differentially expressed genes in multiciliated cells, comparing cells from patients with PCD to cells from people in the healthy control group. (D) Box plots showing the differential gene expression in multiciliated cells of select NRF2 pathway genes between PCD and healthy control cells. In panels A, B, and C, n = 5 control samples, n = 4 heterozygous, and n = 4 PCD samples. In panel D, n = 5 control samples and n = 4 PCD samples. 3 technical replicates each. ****P < 0.0001. Error bars represent standard deviation unless noted otherwise.

Examining the top upregulated genes in multiciliated PCD cells compared with control cells identified genes related to cilia, inflammation, and downstream effectors of the NRF2 oxidative stress response pathway (24, 25) (Figure 3C and Supplemental Table 2). The latter included NRF2-activated genes SLC7A2, ALDH3A1, GSTA1, and GSTA2, with GSTA1 and GSTA2 being among the highest differentially expressed genes in multiciliated cells (Figure 3D). Although DNAH5 expression is unique to multiciliated cells, secretory and basal cells also contributed to the group of differentially expressed genes between variant and control cells (Figure 3A). In secretory and basal cells, gene ontology terms and pathways associated with genes upregulated in PCD cells suggested induction of inflammation-related signaling. Compared with control basal and secretory cells, PCD cells had a decrease in TGF-β–related genes, suggesting that the roles of this pathway may be dysregulated in PCD (Supplemental Figure 2, C and D).

To confirm these results, we independently harvested and cultured primary nasal cells from additional individuals with PCD and individuals who were healthy and analyzed differential gene expression using bulk RNAseq. Specifically, primary culture nasal cells from patients with PCD homozygous for 2 additional but different homozygous pathogenic variants in DNAH5 and 1 in HYDIN were compared with cells from healthy individuals. Cells were cultured in vitro and allowed to redifferentiate into ciliated and nonciliated cells using ALI conditions before analysis (18). Like the scRNAseq analysis, bulk RNAseq analysis of these cultures also showed that genes related to NRF2 and glutathione-related pathways were enriched in the PCD cells compared with normal cells (Supplemental Table 3 and Supplemental Figure 3) (26).

Cells from mothers’ heterozygous variants in DNAH5 show transcriptional divergence. Unexpectedly, we found that the transcriptome of the heterozygous mothers cells differed from both control and PCD cells (Figure 3A). We observed differential gene expression in all major cell types in cultures of the 4 unrelated mothers with heterozygous DNAH5 compared with healthy individuals who had no pathogenetic variants in known PCD-associated genes (Figure 3A and Figure 4A). Differences in the expression profile of mothers’ heterozygous cells were also observed when only female samples from control and PCD cultures were included in the analysis (including additional control samples from women in publicly available data sets) (Supplemental Figure 2B). Interestingly, the gene expression profile of heterozygous cells was closer to PCD cells than to control non-PCD cells.

Differential gene expression of heterozygous maternal cells. (A) Dotplot showing the relative differential gene expression between heterozygous cells and healthy control cells without variants in known PCD genes. (B) Pathway analysis of differentially expressed genes in heterozygous DNAH5 multiciliated cells compared with control cells. (C) Dotplot showing differentially expressed inflammatory markers between heterozygous maternal and control cells within the different cell subclusters. n = 5 control samples and n = 4 heterozygous samples.

Examining the molecular pathways driven by the transcriptional changes in the mothers’ heterozygous cells compared with the control cells showed increased expression of genes related to inflammation and cytokine production (Figure 4B and Supplemental Figure 4). Inflammatory markers were found in all cell types but were particularly enriched in clusters Sec 4 and Bas 5 (Figure 4C), which may be due to effects of a cilia-related dysfunction on neighboring cells.

PCD cells have increased cellular stress markers. We considered that the differential gene expression and induction of the NRF2 pathway observed in multiciliated PCD cells could be either cell intrinsic due to the PCD variant or cell extrinsic due to changes induced by the airway inflamed environment. At the onset, we chose to examine the differential gene expression in cultured cells rather than freshly harvested cells to separate these effects. To further examine the consequence of the genetic variant, we generated iPScs from the peripheral leukocytes of one of the patients with PCD (16). Unlike the primary nasal PCD cells, the iPScs were never exposed to the airway environment. These cells were also differentiated to multiciliated cells using ALI conditions and underwent transcriptional analysis using scRNAseq. Unsupervised integrated analysis categorized all the major epithelial cell types (Figure 5A), including KRT5+ basal cells (Bas), SLURP2+/FOXJ1–/KRT5– secretory cells (Sec), FOXJ1+ multiciliated cells (Cil), and a small dividing cell population (Div). Ionocytes were not identified in these cells. As with primary airway cells, the cell type proportions were similar between DNAH5 iPScs and control iPScs.

Expression of GSTA2 and NRF2 markers in native and DNAH5 iPScs. (A) UMAPs comparing cell distribution between control iPS and DNAH5 iPScs showing no differences in cell subclusters between groups. Different colors represent different cell types. (1 donor, n = 3 technical replicates combined). (B) Pathway analysis of differentially expressed genes comparing nasal PCD cells and iPS PCD cells generated from the same patient with PCD. Diagram shows unique pathways upregulated in iPS PCD cells compared with control iPS cells, those upregulated in the nasal PCD cells compared with control nasal cells, and pathways that are shared between the iPS PCD cells and the nasal PCD cells (n = 1 donor for iPS PCD cells, n = 1 donor for control iPS cells, n = 1 donor for nasal PCD cells, n = 5 control nasal cells. Each in n = 3 technical replicates combined). (C) Violin plots showing the differential gene expression of select NRF2 pathway genes shared between multiciliated cells originating from induced pluripotent cells and native nasal PCD cells.

The differences in gene expression between the multiciliated cell clusters from DNAH5 iPScs and control iPScs mirrored the changes observed in the primary nasal cells (Figure 5, B and C). To confirm changes in NRF2 target genes identified in PCD nasal cells, we compared the expression levels of these targets in the DNAH5 iPScs (Figure 5C). We found increased expression of genes related to oxidative stress, including increased expression of GSTA1 and GSTA2 (Figure 5C). These results indicate that the increased expression of pathways related to cellular stress, especially NRF2 downstream effectors, are cell intrinsic and due to a direct effect of the DNAH5 variants. This indicates that the gene expression changes are less likely to be caused by environment due to an infected or inflamed airway in the patients’ nasal passages.

GSTs and NRF2 pathway transcripts are increased in PCD. We focused on GSTA2 and GSTA1, which were among the highest differentially expressed transcripts in PCD multiciliated cells. The GSTA2 transcripts were highly expressed in multiciliated airway cells while GSTA1 transcripts were expressed in both multiciliated and secretory cells, as shown by comparing expression overlay of GSTA1 and GSTA2 to FOXJ1 on the UMAP data (Figure 6A). GSTA2 expression was previously reported in multiciliated cells in the fallopian tubes (27) and was suggested to be a specific marker of multiciliated cells (28). However, closer inspection of the mean expression levels of GSTA2 showed very low transcript levels in secretory cells as well (Figure 6B).

Expression of GSTA2 and NRF2 markers in multiciliated cells. (A) The expression levels of GSTA1 and GSTA2 projected on UMAPs of airway cells compared with the expression of FOXJ1 as a marker of multiciliated cells. (B) Median expression of GSTA2 in airway cells. (C) Immunoblot of NRF2 pathway targets in cells from 2 patients with PCD. In A and B, n = 5 normal samples, 3 technical replicates each. ****P < 0.0001, respectively. Error bars represent SD unless noted otherwise.

Increased GSTA1, GSTA2, and NRF2 target gene transcripts in multiciliated PCD cells pointed to a wider cellular mechanism of defense against cellular stress in PCD that includes oxidative and electrophilic stress. Indeed, NQO1 protein levels, a key canonical NRF2 target gene (29), was increased at the protein levels in PCD cells (Figure 6C). To confirm these results, we used independent cultures of normal and PCD cells analyzed by a sensitive, targeted mass spectrometry method to identify predefined NRF2 targets (Optimized internal standard–triggered parallel reaction monitoring, OIS-PRM) (30–32). OIS-PRM uses a panel of proteins that are known NRF2 targets but are not specific to airway epithelial cells and did not include specific probes for GSTA2 or GSTA1. Using OIS-PRM, we detected increased protein levels of several NRF2 targets within the log2-fold range considered significant using this technique (Table 1 and Supplemental Table 4) (30, 32). These increased levels paralleled the elevated transcripts levels detected by scRNAseq, confirming the activation of the NRF2 pathway in PCD. Thus, there may be a pattern of gene expression directly related to cilia dysfunction and PCD pathobiology due to the DNAH5 variant.

Expression levels of predefined NRF2 targets

Motile cilia have a dedicated glutathione pathway. Since GSTA2 was highly expressed in ciliated cells and was increased in PCD, we determined its cellular localization and relation to motile cilia function. Immunofluorescent staining of normal primary culture nasal cells localized GSTA2 to the motile cilia (Figure 7A and Supplemental Video 1), while GSTA1 was diffusely expressed in the cytoplasm (Figure 7B). To confirm localization of GSTA2 to motile cilia, we expressed either an N- or C-terminal GFP fusion GSTA2 protein in normal airway cells using lentiviral-mediated transduction, driven by a FOXJ1 promotor. Cells underwent differentiation using ALI conditions and were evaluated for expression of GFP. Live imaging of GFP-GSTA2 transduced cells showed localization of GFP in motile cilia (Supplemental Videos 2 and 3). Moreover, immunofluorescent staining of primary culture airway cells showed localization in motile cilia of tagged GSTA2 when using a GFP antibody (Supplemental Figure 5A). We also found that the substrate of GSTA2, glutathione, was present in the cilia and the cytoplasm (Figure 7C). GSTA2 mRNA expression increased as multiciliated cells underwent differentiation, showing a temporal relation to ciliogenesis when compared with FOXJ1 (Figure 7D). Similarly, GSTA2 protein levels increased while more cilia emerged during ciliogenesis, as indicated by the levels of cilia marker acetylated α-tubulin (Figure 7E), further demonstrating a close relationship between GSTA2 expression and motile cilia.

Motile cilia dedicated glutathione pathway. (A) Immunofluorescent staining of GSTA2 localization to motile cilia of airway cells. (B) Immunofluorescent detection of cytoplasmic GSTA1 in the cytoplasm of multiciliated cells. (C) Immunodetection of glutathione in cilia and cytoplasm of multiciliated cells. (D) Levels of GSTA2 and FOXJ1 during the differentiation of basal cells (ALI day 0) to multiciliated cells, detected by RT-qPCR (n = 3 replicates, error bars represent standard error). (E) Immunoblot detection of GSTA1 and GSTA2 compared with acetylated α-tubulin (ac-TUB) used as a cilia marker during airway epithelial cell differentiation. (F) Immunogold labeling of multiciliated airway cells showing GSTA2 along the cilia axoneme and basal-body microtubules. (G) Scanning electron microscopy of multiciliated cells treated with butter to demembranate cilia showing GSTA2 along the cilia microtubules as well as on the ciliary membrane. (H) CBF and (I) displacement of microbeads across the apical surface of culture primary airway cells following transduction of GSTA2 shRNA. (J) Superoxide levels in multiciliated cells detected by MitoSox. (K) MFI of ciliary axonemes showing increased labeling of GSTA2 in PCD and iPS PCD cells compared with control cells (n = 3 replicates each). (L) MFI of ciliary axonemes showing increased labeling of GSTA2 in CCDC39 and RSPH1 PCD cells compared to control cells, and no increase in DNAAF5 and SPAG1 PCD cells. In A–L, n = 3 replicates each. Representative images are shown. **P < 0.01, ****P < 0.0001. Error bars represent SD unless noted otherwise. Scale bars = 10μM.

Ciliary GSTA2 suggests a local role in maintaining homeostasis during changes in cilia activity or to protect the axonemes from external damage. To determine whether ciliary GSTA2 localizes to the ciliary microtubules or to the ciliary membrane, we used immunogold-labeled GSTA2 in cultures imaged by electron microscopy. Transmission electron microscopy (TEM) detected labeled GSTA2 along the cilia microtubules (Figure 7F, upper panels) and to a lesser extent along the basal body microtubules (Figure 7F, lower panels).

To determine whether GSTA2 was only restricted to the microtubules, we used scanning electron microscopy (SEM) to image ciliated cells and untreated cells, we used scanning electron microscopy after stripping the cell membranes with a mild detergent. GSTA2 immunogold-labelled cells were compared with those incubated with gold-labeled secondary antibody only (Figure 7G). Like the TEM imaging, the majority of GSTA2-labeling was observed primarily along the ciliary axonemal microtubules (Figure 7G, upper panels) while a smaller fraction was observed on the ciliary membrane (Figure 7G, lower panels).

GSTA2 is required for maintaining normal cilia motility. Localization of GSTA2 along the cilia and increased expression in PCD cells strongly suggested a role in motility. To determine a requirement of GSTA2 in normal cilia motility, we knocked down GSTA2 expression in primary human airway cells using shRNA employing previously described methods (Supplemental Figure 5B and Supplemental Table 5) (33). Compared with nontargeted shRNA sequences, GSTA2-specific shRNA sequences reduced ciliary expression of GSTA2 as determined by the mean fluorescent intensity (MFI) of GSTA2 within cilia, which correlated with the degree of decrease in GSTA2 mRNA (Supplemental Figure 5, B, C, and D).

Cilia function was decreased in the GSTA2 shRNA treated compared with nontargeted control cells. GSTA2 knockdown (KD) resulted in a significant decrease in cilia beat frequency (CBF) and ciliary transport of microbeads on the surface of ciliated cells compared with a nontargeted control sequence (Figure 7, H and I). Transport was partially recovered after treating cells with catalase, an enzyme that neutralizes reactive oxygen species. The partial recovery of cilia motility after catalase treatment of GSTA2 KD confirmed a role for GSTA2 in modulating oxidative burden in ciliated cells, likely within the ciliary axoneme. Since the mitochondria are a major contributor of endogenous reactive species (34), we measured the levels of superoxide activity in cells from a participant with PCD. The MitoSox fluorescent indicator level was higher in live cultured PCD nasal cells compared with control non-PCD cells (Figure 7J).

Like the increase in GSTA2 transcripts in DNAH5 variant cells, endogenous GSTA2 levels in the cilia of these cells were increased compared with controls, measured by the MFI of GSTA2 in DNAH5 PCD cells compared with control cells (Figure 7K and Supplemental Figure 5E). Likewise, the MFI of GSTA2 was higher in iPS DNAH5 cells compared with control normal iPS cells (Figure 7K and Supplemental Figure 5E).

We hypothesized that the increased level of GSTA2 within cilia is a physiologic response to increased retained activity of dynein motors within the cilia of PCD cells. For example, in the DNAH5 variant cells, the inner dynein arm motors are intact, and only the outer dynein motor arms are missing. We therefore measured the levels of GSTA2 in the cilia of human PCD cells with other pathogenic variants that have different effects on outer and inner dynein motor complexes, including (a) the assembly factor DNAAF5 (HEAT2) that lacks both outer and inner dynein arm motor complexes (35), (b) pathogenic variants of CCDC39 that have absent inner dynein motors but retain outer dynein motor complexes (36), and (c) RSPH1 that retains both inner and outer dynein motor arm complexes but has dysmotile cilia due to a defect in the radial spokes (37). Compared with control cells, CCDC39 PCD and RSPH1 PCD cells had a higher GSTA2 MFI within cilia, while DNAAF5 PCD cells (missing all ciliary motors) had lower GSTA2 MFI (Figure 7L and Supplemental Figure 5E). These data suggest that GSTA2 was responding to compensatory persistent dynein complex motor activity within remaining compartments of the dyskinetic cilia.

Ciliary GSTA2 is evolutionarily conserved. To discover if GSTA2 expression was evolutionarily conserved, we analyzed published databases of isolated cilia of the algae Chlamydomonas reinhardtii (38, 39) and identified 9 GSTs. Three are orthologs of GSTA (α) (Cre16.g688550, Cre16.g670973, and Cre16.g682725). Three are orthologs of GSTT (θ) (Cre17.g708300, Cre15.g636800, and Cre15.g636750), and Cre02.g142200, which is likely to also be a θ ortholog. Cre17.g742300 and Cre17.g742450 are orthologs of GSTP (π) (Supplemental Figure 6A). Using proteomic analysis of isolated cilia that separate the axonemal and membrane/matrix fraction (M+M) (40), 6 GST species are present in cilia and are predominantly found in the M+M (Supplemental Table 6) as opposed to the axonemal-only fraction of isolated cilia (40). The ratio of axonemal to M+M for the GSTA orthologs ranges from 0.73 to 1 (Supplemental Table 6). These findings suggest that a small fraction of the Chlamydomonas GST orthologs binds to the microtubules, while the majority reside in the matrix fraction (41), similar to findings in human cilia.

To confirm findings in PCD cells, we performed proteomic analysis of Chlamydomonas mutants. Like human DNAH5 variants, Chlamydomonas outer dynein arm docking complex mutants oda1 and oda3 (orthologs of the human genes ODAD1 and ODAD3), lack the outer dynein arms and have decreased ciliary motility (42). Indeed, proteomic analysis of cilia axonemes isolates (without the membrane-matrix fraction) from oda1 and oda3 mutants, shows that the numbers of peptides from the outer dynein arms are about 3%–5% of the WT control (Supplemental Table 6). Peptides of only 1 of the 9 GSTs are present in these preparations (Cre17.g742300) since axonemes without the membrane/matrix fraction were prepared. Most GST proteins are likely to be lost during the cilia isolation and processing with nonionic detergent. Interestingly, Cre17.g742300 was increased in the mutants compared with WT. We also identified increased levels of ROS-responsive proteins in the docking complex mutant cilia isolates (43), including 2-cys-peroxiredoxin (PRX2), FAP102, and a thioredoxin (TRX) (Supplemental Table 6).

Moreover, like findings of decreased cilia motility in GSTA2 KD in human cells, an insertional mutant in the GSTA ortholog (Cre16.g682725) in Chlamydomonas showed reduced swimming velocity as well as reduced CBF compared with a control strain (Supplemental Figure 6, B and C). These results confirm an evolutionary requirement of GSTA for normal cilia motility.