A cytoplasmic protein kinase couples engagement of Chlamydomonas ciliary receptors to cAMP-dependent cellular responses

The primary cilium transduces cues from the extracellular milieu into cellular responses in the cytoplasm. In some ciliary signaling systems, changes in concentrations of cyclic nucleotides within cilia that are triggered by receptor activation lead to responses within the organelles that activate action potentials in the neuronal plasma membrane (vision and olfaction) (Dell'Orco et al., 2021; Dhallan et al., 1990; Nakamura, 2000), or delivery to the cytoplasm of an active transcription factor (the Hedgehog developmental pathway) (Bangs and Anderson, 2017; Nachury and Mick, 2019; Wen et al., 2010). In many others, including ciliary regulation of adipogenesis and ciliary regulation of glucagon and insulin secretion, cAMP diffusing from the cilium is a candidate carrier of cilium-to-cytoplasm information, but definitive signals are still unknown (Hilgendorf, 2021; Hilgendorf et al., 2019; Siljee et al., 2018; Wachten and Mick, 2021; Wu et al., 2021).

We are using ciliary adhesion-induced signaling during fertilization in the bi-ciliated, unicellular green alga Chlamydomonas reinhardtii (hereafter Chlamydomonas) as a model system to investigate mechanisms of cilium-to-cytoplasm communication. When plus and minus gametes encounter each other, their cilia adhere through the adhesion receptor SAG1 on plus cilia and SAD1 on minus cilia. SAG1–SAD1 engagement leads to an almost immediate ∼10-fold increase in total cellular cAMP, and consequent activation of an array of responses in the cell bodies that prepare the gametes for cell–cell fusion (Fig. 1A) (Pasquale and Goodenough, 1987; Pijst et al., 1984; Saito et al., 1993; Snell and Goodenough, 2009; Zhang et al., 1991). As with many metazoan systems, the mechanisms that couple ciliary signaling to cytoplasmic responses in Chlamydomonas are unknown. Here, we identify a protein kinase, Gamete-Specific Protein Kinase (GSPK), that is essential for the large, ciliary adhesion-induced increase in cellular cAMP. Surprisingly, GSPK is present in the cytoplasm and the entire cellular complement of the protein is phosphorylated within 1 min after SAG1–SAD1 engagement in the cilia. Moreover, even though all other cell body events triggered by ciliary adhesion can be induced in gametes of a single mating type by addition of a cell-permeable analog of cAMP, GSPK fails to be phosphorylated upon addition of the analog. Thus, GSPK is responsive to a non-cAMP-mediated signal generated by ciliary adhesion and central for transduction of ciliary adhesion into the large increase in cAMP required for fusion-essential cell body responses.

To identify gamete-specific proteins that function during ciliary signaling, we tested for a fertilization phenotype in several Chlamydomonas CLiP library mutants that were annotated to contain mutations in genes that exhibit gamete-specific expression profiles (Fig. 1B) (Li et al., 2016; Ning et al., 2013). One mutant strain [hereafter termed gspk(−)] with a mutation in Cre02.g104450 (which encodes GSPK) (Fig. S1A, Table S1), was strongly impaired in gamete fusion (Fig. 1C), as were two independent gspk CLiP mutant strains, gspk-2 and gspk-3 (Fig. 1C; Figs S1B,C, Table S1). Fusion between wild-type plus [WT(+)] gametes and gspk(−) gametes at 10 min after mixing was reduced over 50-fold compared to WT(+) and WT(−) gametes and increased slightly by 60 min (Fig. 1C). The fusion phenotype co-segregated with the genotype and was present in both gspk(+) and gspk(−) gametes (Fig. 1D; Fig. S1D). A separate CRISPR-generated (Greiner et al., 2017; Kelterborn et al., 2022) gspk(+) strain also exhibited the fusion phenotype (Fig. 1D; Fig. S2). As expected, gamete activation and cell–cell fusion by gspk mutants was rescued by introduction of a transgene encoding an epitope-tagged form of GSPK, GSPK-HA (Fig. 1E; Fig. S3). Consistent with the downregulation of GSPK transcripts during prolonged dibutyryl (db)-cAMP-induced gamete activation, the levels of the GSPK–HA protein were also substantially reduced after prolonged activation (Fig. 1F). Cell fractionation and immunoblotting indicated that GSPK was present in cell bodies, with little if any detectable in cilia (Fig. 1G). Finally, consistent with the published annotation (Goodstein et al., 2012), analysis of the GSPK protein sequence indicated that it indeed possessed all of the canonical protein kinase subdomains, with highest similarity to mixed lineage protein kinases (Fig. S4).

Vegetative gspk cells and naive gspk gametes were indistinguishable from WT in morphology, growth, and motility (data not shown), and the mutant gametes underwent initial ciliary adhesion to nearly the same extent as WT (Fig. 2A). We found, however, that gspk gametes were impaired in ciliary adhesion-induced cytoplasmic responses required for cell–cell fusion. Cell wall release was reduced nearly 3-fold and mating structure activation in gspk(+) gametes was almost completely blocked (Fig. 2B,C). Addition of db-cAMP to gspk gametes induced cell wall loss and mating structure activation similarly to WT gametes (Fig. 2B,C; Fig. S4) and also rescued fusion (Fig. 2D), indicating that the gspk mutants were capable of responding to the second messenger. Notably, and similar to earlier studies (Pasquale and Goodenough, 1987; Saito et al., 1993; Zhang and Snell, 1994), cAMP levels increased over 10-fold within 1–2 min after mixing WT(+) with fusion-defective hap2(−) gametes, but the increase was less than 2-fold at 1 min after gspk(−) and gspk(+) gametes were mixed. Furthermore, unlike in the control mixture, the (slight) cAMP increase was not sustained in the mutants (Fig. 2E). Thus, this cytoplasmic protein is required to transduce SAG1–SAD1 interactions in the cilia into responses in the cytoplasm, and GSPK function is required within the first minute after the ciliary signal is initiated.

We assessed whether intra-ciliary responses to ciliary receptor engagement were intact in the gspk mutants. Consistent with previous results (Wang et al., 2006; Wang and Snell, 2003), phosphorylation of the previously identified ciliary cGMP-dependent protein kinase (PKG) in isolated cilia was detected in samples of cilia isolated from adhering, but not non-adhering, WT gametes. Moreover, it was similarly phosphorylated in cilia samples isolated from adhering gspk gametes (Fig. 3A). Thus, GSPK is dispensable for this intraciliary biochemical response to SAG1–SAD1.

We also examined whether the previously described barrier to SAG1 entry onto cilia was removed during SAG1–SAD1 interactions between gspk mutant gametes. In naive gametes, only a small portion of total cellular SAG1 is present in cilia, and the remainder is in a non-exchanging, inactive pool on the surface of the cell body plasma membrane (Belzile et al., 2013; Cao et al., 2015; Hunnicutt et al., 1990; Musgrave et al., 1986; Ranjan et al., 2019). During ciliary adhesion, however, the ciliary barrier is relaxed to allow SAG1 entry. Moreover, the ciliary adhesion-induced increase in cellular cAMP induces rapid recruitment of SAG1 from an inactive pool on the plasma membrane onto the ciliary membrane as part of a mechanism to support and enhance ciliary adhesion (Cao et al., 2015; Goodenough, 1989; Hunnicutt et al., 1990; Ranjan et al., 2019; Saito et al., 1985). Immunoblotting of samples taken after mixing gspk/SAG1-HA(+) gametes and WT SAG1-HA(+) gametes with fusion-defective hap2(−) gametes showed that at 10 min after mixing, SAG1–HA had moved into the cilia of both the WT gamete mixtures and those of the gspk gametes (Fig. 2B). Thus, GSPK is dispensable for this second intraciliary response to ciliary adhesion, removal of the ciliary barrier.

Notably, though, at 45 min after mixing, SAG1–HA had continued to increase in the cilia of the WT gametes (Fig. 3B), but it had decreased in the cilia of the gspk/SAG1-HA gametes, even though levels of SAG1–HA remained high in both sets of cell bodies (Fig. 3B). Thus, consistent with the requirement for GSPK for the large increase in cellular cAMP, active SAG1 mobilization from the cell body and enrichment in the cilia also requires GSPK. Reflecting the loss of SAG1 from the gspk cilia, the clusters of adhering gspk plus and minus gametes that had initially formed had disassembled by 45 min, although the clusters had become even larger in mixtures of WT(+) and hap2(−) gametes (Fig. 3C). Taken together, these results indicate that the immediate intra-ciliary responses to ciliary adhesion are independent of GSPK, but the cytoplasmic responses to ciliary adhesion, including mobilization of cell body SAG1 to the cilia, depend on GSPK.

Because most protein kinases undergo changes in phosphorylation state at steps in the signaling pathways in which they function, we assessed GSPK phosphorylation state in naive gametes, in gametes undergoing ciliary adhesion, and in non-adhering gametes of a single mating type undergoing activation induced by db-cAMP. SDS-PAGE and immunoblotting of lysates of naive GSPK-HA(−) gametes before and after incubation with the protein de-phosphorylating enzyme λ-phosphatase showed that the phosphatase incubation led to a shift in migration of GSPK. Thus, GSPK was basally phosphorylated (Fig. 4A) in naive gametes. Moreover, when GSPK-HA(+) gametes were mixed with hap2(−) gametes, GSPK–HA became phosphorylated within 1 min after mixing, and, remarkably, the entire cellular complement of GSPK–HA underwent this ciliary adhesion-triggered phosphorylation (Fig. 4B). Notably, experimentally activating GSPK-HA(+) gametes alone with db-cAMP failed to alter GSPK-HA phosphorylation state, indicating that phosphorylation is upstream of the ciliary adhesion-induced cAMP increase.

Our studies have uncovered a new participant in ciliary signaling in Chlamydomonas and showed that it functions at a previously unrecognized step in the ciliary signaling pathway. GSPK is not required for responses within the cilia to ciliary adhesion – phosphorylation of a ciliary PKG (Wang et al., 2006; Wang and Snell, 2003) and relaxation of the ciliary barrier to entry of SAG1 (Belzile et al., 2013; Pan and Snell, 2002) – but it is required for the adhesion-induced increase in cellular cAMP required for gamete activation. Most importantly, GSPK is a cytoplasmic protein and responds to ciliary signaling within 1 min by becoming phosphorylated. Unlike all other characterized ciliary adhesion-induced cytoplasmic responses, however, including cell wall loss, mating structure activation, and sustained cell body mobilization and ciliary enrichment of SAG1, the change in phosphorylation state of GSPK fails to occur upon experimental gamete activation induced by exogenously added cAMP. Our findings are at odds with previous suggestions that cAMP diffusing from cilia during ciliary adhesion is responsible for triggering the subsequent large increase in cell body cAMP and gamete activation (Pasquale and Goodenough, 1987; Pijst et al., 1984; Saito et al., 1993; Zhang and Snell, 1994; Zhang et al., 1991), and indicate that signaling within cilia triggers transmittance of a still unidentified signal to the cytoplasm that regulates GSPK function. A recent report showing that generation of ciliary, but not cytoplasmic cAMP, regulated the Hedgehog pathway in vertebrate cells (Truong et al., 2021) also indicates that ciliary and cytoplasmic concentrations of cAMP are tightly regulated. And, the recent report that binding of melanin-concentrating hormone to its ciliary receptor on cultured mouse hippocampal neurons triggers phosphorylation of the extracellular signal-related kinases Erk1/2, detectable within 3 min after hormone addition (Hsiao et al., 2021), suggests that cytoplasmic protein kinases also participate at very early steps in vertebrate ciliary signaling.

In future studies, it will be important to test for relationships between GSPK and the three other proteins previously shown to participate in gamete-activated ciliary signaling downstream of SAG1–SAD1 interactions – the ciliary PKG (Wang et al., 2006; Wang and Snell, 2003), a protein phosphatase 2A catalytic subunit (PP2A3) (Lin et al., 2013), and the anterograde intraflagellar transport (IFT) kinesin 2 family member FLA10 (Pan and Snell, 2002). Because all three are present in cilia, and because FLA10 is required for intraciliary phosphorylation of PKG, all 3 might function upstream of GSPK. It will also be important to learn whether the adhesion-induced phosphorylation of GSPK is essential for its function during gamete activation and whether its protein kinase activity is required to induce the increase in cellular cAMP through as yet unidentified adenylyl cyclases or phosphodiesterases. Perhaps of even more importance, though, will be to use this system to investigate the undefined signal transmitted from adhering cilia to the cytoplasm and the mechanism of its transport.

Chlamydomonas reinhardtii wild-type strains 21gr (mating type plus; mt+; CC-1690; designated WT(+), CMJ030 (mating type minus; mt−; CC-5325; designated WT(−), hap2 (40D4; CC5281) and SAG1-HA strains used in this study were grown in liquid tris-acetate phosphate medium (TAP) medium containing trace metals at 22°C with aeration, or on 1.5% TAP agar plates (Wang and Snell, 2003). Gametogenesis was induced by transferring vegetatively growing cells into nitrogen-free medium (M-N) along with continuous light incubation overnight with aeration or agitation (Snell and Roseman, 1979). Chlamydomonas CLiP library mutants (Li et al., 2019) were obtained from the Chlamydomonas Resource Center. Genomic DNA was isolated from single colonies using Clontech plant genomic DNA isolation reagent, and the insertion site of the CIB1 cassette in each mutant was verified by PCR using the primers listed in Table S1. Cell numbers were determined with a hemocytometer.

To prepare a plasmid containing a GSPK gene encoding an HA-tagged GSPK protein, a gene fragment of 8769 bp that included the full-length GSPK gene sequence (7272 bp) and an additional 850 bp 5′ to the annotated transcription start site predicted to include the endogenous promoter and an additional 647 bp 3′ to the stop codon was amplified from DNA of BAC clone 34G21 by PCR using primers possessing Xho1 and Not1 restriction sites at the 5′ and 3′ ends, respectively. The PCR product was cloned into a paromomycin resistance vector, pChlamiRNA3int (Chlamydomonas Resource Center) in between Xho1 and Not1 restriction sites by In-fusion HD EcoDry cloning plus kit. A gene fragment encoding three copies of the 9-amino-acid HA epitope followed by EcoR1 and XbaI restriction sites was inserted using QuikChange II XL Site-Directed Mutagenesis Kit (Table S2). The resulting GSPK-HA transgene plasmid (13,468 bp) was verified by sequencing. For Chlamydomonas transformation, purified BspH1-linearized pGSPK-HA was electroporated into 21gr mt+ and CC-5325 mt− strains (Shimogawara et al., 1998). Transformants that grew on TAP-paromomycin plates were picked into 96-well plates and screened for GSPK-HA by PCR using primer sets P1_Fwd and P2_Rev to confirm N-terminal end, P3_Fwd and P4_Rev to confirm the presence of in-frame HA-tag and P5_Fwd and P6_Rev to confirm the C-terminal end as shown in Fig. S2. PCR-positive transformants were screened for GSPK expression by immunoblotting. gspk(−) strains expressing GSPK–HA were obtained as progeny from crosses between GSPK-HA(+) cells and gspk(−) cells. Colonies were screened for mating type by bioassays and for the presence of the insertion cassette in the GSPK gene by PCR using primers listed in Table S1, followed by further selection for GSPK–HA expression by immunoblotting.

Protein concentrations were determined with the Bradford assay. For immunoblotting, samples were separated by SDS-PAGE on 4–12% SDS-MOPS gradient gels and transferred onto PVDF membranes as described previously (Belzile et al., 2013; Cao et al., 2015). Membranes were blocked by incubation in 3% fat-free dried milk for 1 h followed by 1 h of incubation in the primary antibody. Membranes were washed three times for 10 min with Tris-buffered saline with 0.1% Tween 20 (TBST) followed by incubation with secondary antibody. After three TBST washes and incubation in the chemiluminescent substrate, fluorescence signals were captured on a C-Digit blot scanner. Antibodies were rat anti-HA (1:3000 dilution; Roche, cat. no. 11867423001), mouse anti-α-tubulin (1:5000 dilution; Sigma, cat. no. T1026), goat anti-rat IgG HRP (1:5000 dilution; Merck, cat. no. AP136A), and goat anti-mouse-IgG HRP (1:5000 dilution; Sigma, cat. no. A9917).

For gamete activation, plus and minus gametes were mixed together or gametes of single mating types were incubated in M-N containing 15 mM db-cAMP and 150 µM papaverine (db-cAMP buffer) (Pasquale and Goodenough, 1987). Ciliary adhesion was quantified using an electronic particle counter (Snell and Moore, 1980; Snell and Roseman, 1979). Cell fractionation and assays for cell wall loss and gamete fusion were as described previously (Liu et al., 2008; Snell, 1982; Wang and Snell, 2003). Phosphorylation of PKG was assayed in vitro using a protein tyrosine kinase (PTK) assay (Wang and Snell, 2003). 20 µl of whole cilia (∼3 µg/µl protein) in 5% sucrose, 20 mM HEPES buffer were mixed with 20 µl of 2× PTK buffer (20 mM HEPES, pH 7.2, 10 mM MgCl₂, 2 mM dithiothreitol, 1 mM EDTA, 50 mM KCl, 2 mM ATP, 0.2% Nonidet P-40, 0.4 mM orthovanadate, 20 mM β-glycerolphosphate, and 2% Sigma plant protease inhibitor cocktail) in the presence of ATP for 10 min and the phosphorylated form of PKG was detected by use of 4–20% gradient SDS-PAGE gels and immunoblotting using anti-phospho-tyrosine antibody (anti-p-Tyr; 1:1000 dilution; Sigma, cat. no. 05-321).

GSPK phosphorylation in lysates of naive, adhering or db-cAMP-activated gametes was assessed by changes in migration in SDS-PAGE and immunoblotting. Samples were prepared by addition of 4× SDS-PAGE sample buffer followed by immediate boiling. For phosphatase treatment, 2×10⁷ cells/ml in 1 ml in HEMDK buffer were briefly sonicated and 31 µl cell lysate, 1 µl λ-phosphatase (NEB, 400,000 U/µl), and 8 µl of phosphatase reaction buffer were incubated at 30°C for 30 min. Reactions were terminated by adding 40 µl of 4×SDS sample buffer followed by boiling. As controls, samples were incubated in the presence of a phosphatase inhibitor cocktail (Sigma).

cAMP amounts in adhering wild-type and GSPK mutant gametes were quantified by use of a cAMP Elisa kit. Equal numbers (100 µl, 2×10⁷ cells/ml in M-N) of the WT and gspk plus and minus gametes were separately mixed in 1.5 ml Eppendorf tubes to initiate ciliary adhesion, and at the times indicated the cells were harvested by centrifugation (6350 g; 4°C) and flash-frozen in liquid nitrogen. For assays, samples were resuspended in 100 µl of 0.1 M HCl and incubated at room temperature for 10 min followed by clarification by centrifugation (20,000 g; 4°C). Supernatants were transferred to fresh tubes for use in the assay, which was performed using the acetylation protocol according to the manufacturer's instructions. Absorbances at 405 nm of standards and experimental samples were determined using a microplate reader. Results shown are from six independent experiments and are plotted as pmol/ml cAMP produced in the WT and gspk mutant mixtures.

As described previously (Wilson et al., 1997), samples (∼200 µl, 5×10⁶ cells/ml) in M-N were seeded on coverslips coated with poly-L-lysine for 5 min followed by fixation with 4% paraformaldehyde solution freshly made in 10 mM HEPES, pH 7.4. Coverslips were washed with PBS for 3 min, immersed for 6 min in 80% acetone, 30 mM NaCl and 2 mM sodium phosphate buffer, pH 7.0, at −20°C, followed by immersion in 100% −20°C acetone 6 min. Filamentous actin in the fertilization tubules was visualized by staining with Alexa Flour 488–Phalloidin (Thermo Fisher Scientific, cat. no. A12379) for 15 min in the dark (Craig et al., 2019), followed by a 5-min PBS wash. Coverslips mounted on slides using ProLongTM Gold antifading agent were examined by a Hyd detector-equipped Leica TCS SP5 confocal microscope (1.4 numerical aperture, 63× oil immersion objective). Z-series images were summed to produce a projected image using Leica LAS X and cropped in Adobe Systems (USA) Illustrator.

Protein sequences were aligned with ClustalW, and the percentages of positions with identical or identical plus similar amino acid positions were calculated using BioEdit 7.2 software with a threshold of 75%. JPred4 was used to predict secondary structure (Drozdetskiy et al., 2015). N-terminal myristoylation sites were predicted using NMT – The MYR Predictor (Maurer-Stroh et al., 2002). All quantitative data represent at least three independent sets of experiments. Statistical significance of differences between groups was assessed by unpaired two-tailed Student's t-test. Data were analyzed using GraphPad Prism 9. All experiments were performed three to six times.

For the reagents used and their identifier numbers, please refer Table S3.

We are grateful to Dr Melanie Cobb of UT Southwestern Medical School, Dallas, TX and Dr Caren Chang, University of Maryland, College Park, MD for insightful discussions, our laboratory colleagues, Drs Jennifer Pinello and Jun Zhang for their constructive insights, and Amy Beaven and the CBMG Imaging Core Facility for use of the Leica TCS SP5 confocal microscope.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.259814.