Introduction

Sexual reproduction is almost universal among eukaryotic organisms. Mating type (or sex) is a key regulatory feature of gamete fusion. Most species have only two mating types (e.g., male and female, + and -, or a and α) and species usually use either self-or non-self-recognition mechanism to achieve mating-type recognition (Goodenough & Heitman, 2014). However, species in some lineages, such as some ciliates and basidiomycetes (Heitman, 2015; Phadke & Zufall, 2010), possess multiple mating types, and multiple-alleles self-incompatibility system were observed in some plants, such as Brassicaceae (Iwano & Takayama, 2012; Takayama & Isogai, 2005; Vekemans & Castric, 2021). This raises the interesting question of how mating types are recognized at the molecular level in multiple mating-type systems.

The model unicellular ciliate, Tetrahymena thermophila, has seven mating types (I–VII). Under starvation conditions, any cell of one mating type can mate with a cell of any of the other six mating types, but not with one of the same mating type (Figure 1A, Figure 1–figure supplement 1, Video 1–3) (Cervantes et al., 2013; Nanney, 1953; Orias et al., 2017; Yan et al., 2021). Mating-type recognition in Tetrahymena depends on direct cell–cell contact, which suggests that mating-type proteins localize to the cell surface. However, there is no direct evidence to indicate whether they are ciliary proteins or not. When one cell comes into contact with a cell of a different mating type, a mating-type-dependent recognition event enables both cells to enter a pre-conjugation stage (called costimulation) (Bruns & Palestine, 1975; Finley & Bruns, 1980). Even when cells of different mating types are mixed in unequal ratios (for example, 9 : 1), all cells become fully stimulated (Bruns & Palestine, 1975). Processes that take place during costimulation include Tip transformation (Wolfe & Grimes, 1979) and concanavalin A (Con-A) receptor appearance (Figure 1B) (Wolfe & Feng, 1988; Wolfe et al., 1986). In preparation for pairing, costimulated cells of the same and different mating type(s) adhere to form very loose pairs. Heterotypic cell pairs form a stable conjugation junction, whereas homotypic pairs separate very quickly (Video 2 and 3) (Kitamura et al., 1986).

Figure 1.

The mating-type system of T. thermophila was described by Nanney and collaborators in the early 1950s (reviewed in (Orias, 1981; Orias et al., 2017)). We previously showed that mating type is determined by a pair of mating-type genes that are organized in head-to-head orientation: MTA and MTB (Figure 1C) (Cervantes et al., 2013). Each gene has a terminal exon that encodes five transmembrane (TM) helices and a cysteine-rich growth factor receptor (GFR) domain. The region between the two terminal exons of the gene pair encodes the N-terminal mating-type-specific extracellular regions. Based on the mating-type-specific regions, the mating-type genes are called MTA1-MTB1 for Mating type I, MTA2-MTB2 for Mating type II and so on. We previously showed that ΔMTB cells do not form pairs or produce progeny and that ΔMTA cells retain mating-type specificity but pair extremely poorly and rarely produced progeny (Cervantes et al., 2013); we concluded that the two genes are non-redundant and both are essential for mating. However, challenges such as multiple mating types; the high molecular weight, membrane localization and extremely low expression levels of mating-type proteins; and difficulty in genetically manipulating the mating-type gene locus have so far prevented elucidation of the mode of action of the MTA and MTB proteins and of whether they mediate self-or non-self mating-type recognition.

In this study, we provide direct evidence that the MTA and MTB form an elaborate multifunctional protein complex that can identify cells of both self and non-self mating types to inhibit or activate mating, respectively.

Results

Mating-type recognition cannot be explained by a simple receptor–ligand model

Receptor–ligand interactions are a critical mechanism for intercellular communication that may regulate mating-type recognition in T. thermophila, irrespective of whether self-or non-self-recognition mechanisms are employed. Therefore, we first assessed whether mating-type recognition conforms to a straightforward receptor–ligand model (in which one mating-type protein acts as the receptor and the other as the ligand) and whether self or non-self is recognized (Figure 2–figure supplement 1A, 1 and 2). For this, we determined whether deletion of each mating-type gene affected the transmission and detection of mating signals to and from wild-type (WT) cells (Figure 2A shows the experimental procedure) by assessing the ability of cells to undergo costimulation (the prerequisite for mating). Our experiment allows us to test if cells missing one of the two mating-type proteins can still costimulate WT cells (Figure 2–figure supplement 1A, 3 and 6).

Figure 2.

In T. thermophila, cells normally enter into the fully costimulation stage within ∼30 min after mixing starved WT cells of two different mating types, and start pair formation during the next ∼30 min (Figure 2B, black line). Cells that have already been costimulated immediately start forming pairs with other costimulated cells of a different mating type (Figure 2B, red line).

To our surprise, the rate of pair formation in WT cells pre-incubated with either ΔMTA cells (Figure 2B, green line) or ΔMTB cells (Figure 2B, blue line) did not increase (i.e., costimulation did not occur). These results were not changed by extending the pre-incubation time (Figure 2–figure supplement 1B). Therefore, both MTA and MTB proteins are essential for the mating-type signal; there is no simple receptor–ligand relationship.

In addition, WT cells were not costimulated even when they were simultaneously incubated with both ΔMTA and ΔMTB cells (Figure 2B, teal line), although, according to the receptor–ligand model, they should have received “MTA stimulation” from ΔMTB cells and “MTB stimulation” from ΔMTA cells. This result indicates that the absence of a mating-type protein in one cell cannot be complemented by its presence in another cell in the same culture; the MTA and MTB proteins must be in the same cell to transmit the mating-type signal. This finding suggests that mating-type recognition cannot be explained by a simple receptor–ligand model. It is possible that the MTA and MTB proteins form a complex which either serves as a recognizer (functioning as both ligand and receptor) or a co-receptor (which influences ligand–receptor activity). But, since MTA and MTB are the only genes with mating-type-specificity, it is unlikely that the complex is acting as a co-receptor.

Mating-type proteins differentially regulate two steps of costimulation

During costimulation, cells undergo a sequence of developmental events that remodel the anterior cell membrane and its associated cytoskeleton (Cole, 2013). Two hallmarks of this process are Tip transformation (in which the anterior tip of the cell becomes curved) and Con-A receptor appearance (receptors bound by the plant lectin concanavalin A, which binds to mannose containing glycoproteins). When WT cells of one mating type were mixed with WT cells of another mating type, the cell tips became transformed (Figure 2C, 5 and 9) and Con-A receptors appeared in almost all cells (Figure 2D, 4). When WT cells were pre-incubated with ΔMTA, ΔMTB, or both cell types, Tip transformation was not observed in any cell (Figure 2C, 6–8). Similarly, when cells of each mutant were pre-incubated with WT cells, Tip transformation was not detected (Figure 2C, 10 and 11). The outcome was slightly different for Con-A receptor appearance. Con-A receptors were not observed in WT cells pre-incubated with ΔMTB cells (Figure 2D, 6) or in cells of either mutant pre-incubated with WT cells (Figure 2D, 7 and 8). In contrast, when WT cells were exposed to ΔMTA or “ΔMTA cells plus ΔMTB cells”, the Con-A receptor was detected (Figure 2D, 5 and 9); this is consistent with ΔMTA cells retaining a very weak ability to pair (Cervantes et al., 2013). These results indicate that neither ΔMTA and ΔMTB cells can fully stimulate WT cells or be stimulated by WT cells. They also demonstrate that costimulation can be separated into two stages: (i) the appearance of Con-A receptors, which requires the expression of MTB only in partner cells, and (ii) morphological transformation of the cell tip, which requires both MTA and MTB.

Mating-type proteins form a complex with several coexpressed proteins

The inability of “ΔMTA cells plus ΔMTB cells” to costimulate WT cells of a different mating type indicates that MTA and MTB proteins can only function when expressed on the same cell (Figure 2B, teal; Figure 2C, 8). We previously found that CDK19 and CYC9 proteins are also essential for mating. ΔCDK19 and ΔCYC9 cells cannot mate with other cells, but they still express MTA and MTB proteins (Ma et al., 2020). In WT cells, the pairing rate was increased following pre-incubation with ΔCDK19 or ΔCYC9 cells of a different mating type (Ma et al., 2020). These results further support the idea that MTA and MTB must be expressed on the same cell to be fully functional, leading us to hypothesize that MTA and MTB proteins form a mating-type recognition complex (MTRC).

To test whether MTA and MTB proteins physically interact, an HA-tag coding sequence was ligated to the 3ʹ end of the MTA gene (Figure 3A); cellular proteins were co-precipitated with HA-tagged MTA and analyzed by immunoprecipitation-coupled mass spectrometry (IP-MS). As expected, the MTB protein co-purified with MTA (Figure 3B, C), as did another set of proteins, which we named MRC1–MRC6 (Figure 3B, C and Figure 3–Source data 1 and 2). Next, we produced strains expressing either HA-tagged MTB or MRC1 (Figure 3A), and found that each protein pulled down a subset of the proteins that co-purified with MTA (Figure 3B, C). Unfortunately, these pull-down experiments were not as successful as the MTA IPs, perhaps because of the higher molecular weight of MTB and MRC1. Taken together, our results suggest that MTA, MTB, and MRC1–MRC6 form the MTRC. Alternatively, MTA and MTB may interact with subsets of MRC proteins to form smaller complexes or alternative MTRCs. Different protein interactors were identified in extracts from cells at different mating stages. This may reflect conformational changes in the MTRC but the huge molecular weight of the complex and extremely low expression levels of its proteins make this possibility difficult to investigate.

Figure 3.

Mating-type-related factors in other species are often relatively small (Iwano & Takayama, 2012; Kahmann & Bölker, 1996), but those of T. thermophila are large proteins (predicted size, 92–212 kDa). Figure 3D shows the predicted domains of the MRC1–MRC6 proteins. Like MTA and MTB, MRC1 has five predicted TM helices and a GFR domain. MRC2 has eight TM helices and a pectin lyase-fold domain, suggesting a possible role in carbohydrate chain modification. MRC3 has two TM helices in the central region and an adjacent ∼35 amino acid (aa) poly-E region. MRC4 and MRC5 (previously named TPA9 (Wang et al., 1997; Wang & Takeyasu, 1997)) are both P-type ATPases that are likely to function as calcium-translocators. MRC6 has four TM helices and a P-loop containing nucleoside triphosphate hydrolases. Most of the MRC genes are highly coexpressed with MTA and MTB (Figure 3E). Examination of the whole genome sequences of strains with mating types II–VII (Wang et al., 2020), confirmed that only the sequences of MTA and MTB genes are mating-type-specific.

In addition to the MRC proteins, CDK19, CYC9, CIP1, and AKM3 were identified in IP-MS experiments, but with relatively few supporting peptides (Figure 3C). Our previous work showed that CDK19, CYC9, and CIP1 are components of a cyclin-dependent kinase complex. These genes are coexpressed with MTA and MTB (Figure 3E) and are important for mating, and the encoded proteins localize to the cell tip and pairing junction (Ma et al., 2020). AKM3 is predicted to be a K⁺ channel of unknown biological function. It is coexpressed with MTA and MTB (Figure 3E), and we found that the AKM3 deletion strain cannot pair. Therefore, these four proteins are also likely to interact with the mating-type proteins (perhaps indirectly and/or via weak interactions) and might be involved in downstream signaling following mating-type recognition.

Mating-type proteins localize to the cell surface but not to the cilia

We used the MTA7-HA strain to determine the localization of mating-type proteins. Cell fractionation (Figure 4–figure supplement 1) revealed that MTA7-HA is a membrane protein (Figure 4A). Biotinylation and isolation of cell surface proteins confirmed that MTA7-HA is exposed on the cell surface (Figure 4B). To investigate whether MTA7-HA localizes to the cilia membrane, we isolated and collected cilia (Figure 4C) and then analyzed cilia protein extracts by IP-coupled western blotting (WB) and MS. MS analysis identified typical ciliary proteins, such as inner and outer arm dynein proteins (Figure 4–Source data 1). However, both IP-WB (Figure 4D) and MS (Figure 4–Source data 1) consistently failed to identify MTA7-HA protein in the cilia fraction. These results conclusively indicate that MTA7-HA localizes to the cell surface, but not to cilia. Unfortunately, we failed to detect MTA7-HA by immunofluorescence staining of cells at any mating stage (starvation, costimulation, or conjugation), probably due to the epitope masking and extremely low expression level.

Figure 4.

To further examine localization of the mating-type proteins, eGFP-tagged MTB2 was overexpressed from an exogenous locus (Figure 4E). This strain (which has a mating type VI background) mated normally with WT cells of all mating types except for VI and II. This result indicates the overexpressed MTB2-eGFP protein is fully functional for mating. Interestingly, cells of this strain can also mate with one another (self-mating); a similar phenotype was previously reported for strains expressing multiple mating-type proteins (Lin & Yao, 2020). The overexpressed MTB2-eGFP protein was detected on the cell surface in a linear pattern radiating from the cell tip to the cell body along the ciliary rows (Figure 4F, costimulated cell; Figure 4G, mating pair), although signals are also apparent between ciliary rows. Co-staining with a tubulin dye showed that the MTB2-eGFP protein is adjacent to, rather than co-localizing with, the base of cilia (Figure 4F, G, enlarged). Confocal images from the interior of the cell and through the cilia showed that MTB2-eGFP localizes to the cell surface (and also to intracellular structures, probably the endoplasmic reticulum (ER) and Golgi), but not to the cilia (Figure 4–figure supplement 2), confirming our results with the MTA7-HA protein (Figure 4A-D). MS analysis showed that overexpressed MTB2-eGFP protein was not present in isolated cilia of the MTB2-eGFP strain (Figure 4–Source data 2). Therefore, mating-type proteins localize to the cell surface, as might be expected since mating-type recognition depends on cell–cell contact.

Mating-type proteins influence non-self-recognition

The mating-type-specific region of the MTA and MTB gene pair is the only known genetic locus with mating-type specificity; therefore, we next tested whether this region influences self-and/or non-self mating-type recognition. For this, we expressed the extracellular region of MTA (MTAxc) or MTB (MTBxc) in an insect cell secreted expression system, purified the recombinant proteins (Figure 5–figure supplement 1), and then tested their effect on mating behavior.

Figure 5.

First, we tested whether MTAxc and MTBxc can influence mating in cells with a different mating-type specificity (i.e., non-self-recognition). When WT cells were incubated with MTAxc (and/or MTBxc) of a different mating type, markers of costimulation were not observed (Figure 5–figure supplement 2). Surprisingly, treated cells had a significantly increased pairing rate. Compared with controls (Figure 5A–C, black), WT cells (VI and VII) pre-treated with MTAxc or MTBxc of different mating types (VII and VI, respectively) had an increased pairing rate (Figure 5A–C, green or blue), with a stronger effect after pre-treatment with both MTAxc and MTBxc (Figure 5A–C, teal). Dose–effect assays showed that pairing rates increased with increasing MTA6xc and MTB6xc concentrations between 3 pg/ml and 30 pg/ml, with the effect becoming saturated or weaker at higher concentrations (Figure 6A–F). MTAxc and MTBxc also stimulated mating for all other WT mating types (Figure 6G, H), indicating that this is a general effect. Based on these results, we conclude that the mating-type-specific regions of MTA and MTB proteins are involved in non-self mating-type recognition.

Figure 6.

Mating-type proteins also influence self-recognition

We used similar methods to examine whether pre-treatment with a cognate mating-type-specific region (i.e., self-recognition) affects mating. Treatment of WT cells (VI and VII) with MTAxc and/or MTBxc of the same mating type decreased the pairing rate (Figure 5D, E, green or blue). No obvious difference was found between treatments with either MTAxc or MTBxc. Meanwhile, a significant synergistic effect was observed (Figure 5D, E, teal). For all treatments (single or combined), the pairing rate was similar by 4 h (reaching >80%; Figure 5D, F), indicating that the initial inhibitory effect on pairing was eventually overcome. Negative regulation by MTAxc and MTBxc was also observed for other mating-type combinations (Figure 6I, J). These results support the idea that the mating-type-specific regions of MTA and MTB proteins mediate both non-self (between cells of different mating types) and self (between cells of same mating type) recognition.

Discussion

Although the basic biological features of the T. thermophila mating-type system were discovered over half a century ago, the mechanism for mating-type recognition remains unclear. Here, we identified a novel MTRC that contains MTA, MTB, and several other proteins and provide evidence that MTA and MTB mediate both self and non-self mating-type recognition.

In most species, mating signals are associated with relatively small proteins or molecules (Iwano & Takayama, 2012; Kahmann & Bölker, 1996). However, in T. thermophila, the MTRC is likely to be over a million Daltons in size. This large size is consistent with its multiple functions: An elaborate MTRC is needed to inhibit mating between cells of the same mating type while also activating mating between cells of different mating types. This is likely to be a highly complex process, given that there are seven mating types in all. Further structural studies of this protein complex may reveal the specific self-and non-self-recognition mechanisms.

An open question is why T. thermophila should use such a dual approach to achieving mating-type recognition. Recent research on basidiomycetes and flowering plants (other species with multiple mating types) has shown that their mating-type recognition (or self-incompatibility) mechanisms involve either self-or non-self-recognition (Fraser & Heitman, 2003; Iwano & Takayama, 2012; Vekemans & Castric, 2021). Our previous evolutionary study showed that the length of the mating-type-specific region differs significantly among different Tetrahymena species (Yan et al., 2021). Based on the massive difference in length, it is reasonable to speculate that different species might use different mechanisms for mating-type recognition. Therefore, dramatic evolution of the mating-type recognition mechanism seems to have occurred relatively soon after the emergence of the Tetrahymena genus. Further detailed functional and evolutionary studies may reveal how and why this recognition mechanism evolved and how its evolution contributed to speciation.

We still do not know what intracellular signals are transduced when the MTRCs on two cells interact. Mating-type self-recognition might generate an inhibitory signal or might simply inactivate the MTRC. In many species, such as Papaver rhoeas and Ciona intestinalis, interaction between mating-type proteins of the same mating type induces changes in cytoplasmic Ca²⁺ concentration that cause self-incompatibility (Giamarchi et al., 2006; Harada et al., 2008; Wu et al., 2011). A similar system may exist in T. thermophila, since the MRC4 and MRC5 proteins are predicted to be Ca²⁺-translocators. Interaction between MTRCs on cells of different mating types (non-self-recognition) should result in their activation to allow the cells to initiate pairing. We propose that the activation signal involves the CDK19 complex (a cyclin-dependent kinase complex) (Ma et al., 2020) and AKM3 (a K⁺ channel; Figure 3C) because they probably interact with the MTRC. We expect future studies to lead to the discovery of more detailed mechanisms for mating-type recognition and initiation of conjugation involving these proteins.

Paramecium tetraurelia, a closely related Oligohymenophorean ciliate, has only two mating types, which are determined by the expression or non-expression of a Tetrahymena-MTA/B-like protein called mtA (Singh et al., 2014; Yan et al., 2021). An intriguing question is whether mtA also engages in the formation of a MTRC with other proteins by serving as a recognizer rather than mediating a straightforward receptor–ligand interaction. Future investigations of Paramecium tetraurelia may shed light on the origins and evolutionary aspects of this distinctive mating system. Moreover, due to the extremely long evolutionary distance, recognition mechanisms discovered in model species fall short of explaining many of the intricate biological events in protists. Insight into the detailed function of MTRC could contribute to our understanding of cell–cell recognition processes in other species, such as Toxoplasma and Plasmodium.

Materials and Methods

Biological methods

Strains used in this study are summarized in Appendix 1–table 1. All cell growth, starvation, and pairing experiments were conducted at 30°C. Cells were grown in Super Proteose Peptone (SPP) medium (1% Proteose Peptone, 0.1% yeast extract, 0.2% glucose, 0.003% Sequestrene) or Neff medium (0.25% Proteose Peptone, 0.25% yeast extract, 0.5% glucose, 0.003% Sequestrene). Cells were starved in 10 mM Tris-Cl (pH 7.4) for ∼16 h before all pairing experiments. For normal pairing assays, equal numbers of starved cells of different mating types (at ∼2×10⁶ cells/ml) were mixed. To obtain costimulated (pre-incubated) cells, two starved strains were mixed at a 9 : 1 ratio for ∼30 min (unless otherwise stated). Before mixing costimulated cells, any potentially pairing cells were separated by shaking. Figure 2A shows the setup of costimulation experiments. For all mating experiments (whether or not they involved mutant cells), the starting WT cell density was ∼2×10⁶ cells/ml. The following formula was used to calculate pairing ratios and correct for the presence of mutant cells:

Somatic gene deletion, truncation, and protein tagging

To construct deletion strains, a ∼1 kb fragment upstream of the gene’s open reading frame (ORF) (#1), a ∼0.5 kb fragment downstream of the gene’s ORF (#2), and a ∼1 kb fragment downstream of #2 (#3) were amplified. Fragments #2 and #3 were joined to the NEO4 cassette (Cd²⁺-inducible MTT1 promoter linked to the neomycin resistance gene) by fusion PCR and then cloned into the pBlueScript SK (+) vector together with fragment #1. In this way, #1-#2-NEO4-#3 constructs were obtained for the next transformation. HA-tagged strains were constructed in a similar way, except that fragment #1 was upstream of the stop codon or upstream of the terminal intron. To obtain the MTB2-eGFP construct, MTB2-coding sequences replaced the MTT1-coding sequence (Figure 4E) and the construct was made using the large DNA fragment assembly method (Jiang et al., 2022). Constructs were introduced into starved WT cells by biolistic transformation to obtain deletion strains (Mochizuki, 2008). Positive clones were selected in SPP medium containing decreasing Cd²⁺ concentrations (from 1 μg/ml to 0.05 μg/ml) and increasing paromomycin concentrations (from 120 μg/ml to 40 mg/ml) until all WT somatic chromosomes had been replaced by mutant ones, as determined by PCR using checking primers. All primers used are listed in Appendix 1–table 2.

Immunoprecipitation and mass spectrometry

The IP method was adapted from a published method (Tian et al., 2017). To pull down HA-tagged proteins from T. thermophila, cells were harvested from 500 ml cultures (density ∼3×10⁶ cells/ml). Cells were then treated for 20 min with paraformaldehyde (PFA) (at a final concentration of 0.3%) to stabilize protein–protein interactions, washed with PB buffer (2.7 mM KCl, 8 mM Na₂HPO₄, 1.5 mM K₂HPO₄), and blocked with 125 mM glycine. Cells were then resuspended in lysis buffer (1% Triton X-100, 30 mM Tris-HCl, 20 mM KCl, 2 mM MgCl₂, 1 mM phenylmethylsulfonyl fluoride, 150 mM NaCl, cOmplete proteinase inhibitor [Roche Diagnostics, Indianapolis, IN, USA]), lysed by ultrasonic treatment and incubated with EZview anti-HA agarose beads (Sigma-Aldrich, St Louis, MO, USA) for 2.5 h at 4°C. The beads were washed with wash buffer (1% Triton X-100, 600 mM NaCl, 30 mM Tris-HCl, 20 mM KCl, 2 mM MgCl₂, cOmplete proteinase inhibitor) to remove nonspecific-binding proteins and then HA-tagged proteins were eluted with HA peptides (Sigma-Aldrich). WT samples (not HA tagged) were run in parallel for each sample. In total, data for 13 WT controls were combined to identify non-specific binding proteins.

For MS, the EASY-nLC chromatography system (Thermo Scientific, Rockford, IL, USA) was coupled on-line to an Orbitrap Elite instrument (Thermo Scientific) via a Nanospray Flex Ion Source (Thermo Scientific). Xtract software (Thermo Scientific) and Proteome Discoverer 2.1 software were used for MS data analysis based on a database that combines the 2014 version of whole genome protein annotation (http://ciliate.org/index.php/home/downloads, which contains the whole length sequence of MTA6 and MTB6) and mating-type-specific regions of all other mating-type proteins. IP data were analyzed using CRAPome (Mellacheruvu et al., 2013).

Membrane protein extraction

Figure 4–figure supplement 1A shows the workflow used for membrane protein extraction. Cells were collected, resuspended in 20 ml lysis buffer (150 mM NaCl, 25 mM HEPES, 10% glycerol, 2 mM PMSF, 2.6 μg/ml aprotinin, 1.4 μg/ml pepstatin, 10 μg/ml leupeptin, pH 7.4), and lysed by high-pressure homogenization. The lysate was clarified first at low speed (14,000 rpm, 4°C, 15 min), and then at high speed (150,000 g, 4°C, 1 h). The pellet was resuspended in 5 ml lysis buffer containing 1% DDM (Anatrace, Maumee, OH, USA) and rotated for 2 h at 4°C to extract the membrane proteins. Undissolved material was removed by centrifugation (14,000 rpm, 4°C, 30 min). The membrane protein extract was incubated with EZview anti-HA agarose beads for 2.5 h at 4°C and then washed with 5 ml lysis buffer.

Biotinylation and isolation of cell surface proteins

Pierce Cell Surface Protein Biotinylation and Isolation Kit (Thermo Scientific) was used to biotinylate and isolate cell surface proteins. For this, 75 ml cells (density ∼3×10⁶ cells/ml) were harvested and washed once with BupH phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 4.3 mM Na₂HPO₄, 1.4 mM KH₂PO₄). Cells were then resuspended in 75 ml PBS containing 0.72 mg/ml Sulfo-NHS-SS-biotin and incubated at room temperature for 10 min. After two washes with 50 ml ice-cold BupH Tris buffer, cells were resuspended in 3 ml lysis buffer (PBS containing 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride and cOmplete proteinase inhibitor [Roche Diagnostics]), lysed by ultrasonic treatment and incubated with 1.2 ml NeutrAvidin Agarose for 0.5 h at room temperature. The resin was washed four times with 0.5 ml wash buffer and then cell surface proteins were eluted with 1.2 ml elution buffer (with 10 mM DTT). Before WB, cell surface protein samples were concentrated into 0.1 ml volumes using a 30 kDa centrifugal concentrator (Merck Millipore).

Ciliary protein collection

To remove cilia, 500 ml cells (density ∼3×10⁶ cells/ml) were harvested at room temperature and resuspended in 25 ml 10 mM Tris-Cl (pH 7.4), to which 50 ml medium A (10 mM EDTA₂Na, 50 mM sodium acetate, pH 6.0) was added. After 30 s, 25 ml cold distilled water was added; 1 min later, 0.25 ml 0.4 M CaCl₂ was added and incubated for 15 s. The cilia were detached from the calcium-shocked cells by vortexing three times for 5 s at 15 s intervals. To collect the cilia, cell bodies were removed by two rounds of centrifugation at 1500 rpm for 5 min at 4°C, and then cilia were collected by centrifugation at 15000 rpm for 15 min at 4°C.

Cytological methods

For tubulin staining, cells were collected and fixed in PHEM buffer (30 mM PIPES, 14 mM HEPES, 5 mM EGTA and 2 mM MgSO₄) containing 1% PFA and incubated for 30 min at 4°C. After three washes with PBS (10 min each), Tubulin-Atto 594 was added and incubated for 1 h at 25°C. Finally, cells were washed three times with PBS (10 min each). Fluorescein-labeled Con-A labeling was performed as previously reported (Ma et al., 2020). In brief, cells were fixed and stained with fluorescein-labeled Con-A (Vector Laboratories, Burlingame, CA, USA) at 37.5 μg/ml for 5 min and then washed three times with PB. For analysis of Tip transformation, cells were observed and photographed as soon as possible after fixation with 1% PFA. To distinguish between cell strains in a pairing mixture, starved cells of one strain were labeled with 500 nM MitoTracker (Invitrogen, Eugene, OR, USA), followed by two washes with 10 mM Tris-Cl (pH 7.4) before mixing.

Expression and purification of the extracellular region of mating-type proteins

Coding sequences of the extracellular region of mating-type proteins (MTA6xc, MTB6xc, MTA7xc, and MTB7xc) were codon-optimized and synthesized for expression in an insect-cell system (Trichoplusia ni Hi5 cells). Codon-optimized sequences were cloned into pFastBac vectors containing an N-terminal hemolin signal peptide sequence and a C-terminal 10×His tag sequence. The obtained constructs were transformed into competent DH10Bac cells and individual bacmids were transfected into Spodoptera frugiperda Sf9 cells. Recombinant Baculoviruses were collected after 4 days and used to infect Trichoplusia ni Hi5 cells for protein expression. Proteins were harvested 60 h after infection and purified with Ni-NTA Superflow resin (Qiagen), anion-exchange chromatography (Source 15Q, GE Healthcare), and size-exclusion chromatography (Superdex-200 Increase 10/300, GE Healthcare).

Bioinformatics analysis

All microarray data were derived from TetraFGD (Xiong et al., 2011) (http://tfgd.ihb.ac.cn/). DNA sequencing data for mating type II–VII cells are derived from a previous report (Wang et al., 2020). Compute pI/Mw (https://web.expasy.org/compute_pi/) was used to predict protein molecular weight. InterProScan (http://www.ebi.ac.uk/interpro/) was used for function and domain annotation (Jones et al., 2014).

Statistical analysis

For mating experiments, more than 100 unpaired cells or cell pairs were counted, with three to five independent replicates. GraphPad software (version 8.0.2) was used for statistical analysis based on ANOVA (matched, Fisher’s LSD test).

Acknowledgements

We thank Eduardo Orias (University of California Santa Barbara), Eileen P. Hamilton (University of California Santa Barbara), and Kazufumi Mochizuki (University of Montpellier) for their help and suggestions about experiment design and manuscript writing, and Yunfei Wei (Huazhong University of Science and Technology) for tubulin staining using Tubulin-Atto 594. We also thank members of the Protist 10,000 Genomes Project (P10K) consortium for helpful suggestions. We would like to thank Min Wang at the Analysis and Testing Center of Institute of Hydrobiology, Chinese Academy of Sciences for her help with mass spectrometry. The bioinformatics analysis was supported by the Wuhan Branch, Supercomputing Center, Chinese Academy of Sciences, China. Culture and maintenance of Tetrahymena cells were supported by the National Aquatic Biological Resource Center (NABRC).

Funding information

This work was funded by Bureau of Frontier Sciences and Education, Chinese Academy of Sciences (No. ZDBS-LY-SM026), National Natural Science Foundation of China (No. 32130011, 32200344) and China Postdoctoral Science Foundation (No. 2021M703433).

Author Contributions

G.Y. performed the experiments and analyzed the data. Y.M and H.C helped to construct mutants and perform cellular experiments. Y.W. and F.T. performed protein expression and purification experiments. J.Z helped to maintenance of Tetrahymena strains. G.Y. wrote the manuscript. All of the authors helped to conceptualize the work, and edit the text and the figures. W.M. and P.Y. supervised the work.

Conflict of Interest

Authors declare that they have no competing interests.

Supplementary Information

Figure 1 – figure supplement 1.

Figure 2 – figure supplement 1.

Figure 4 – figure supplement 1.

Figure 4 – figure supplement 2.

Figure 5 – figure supplement 1.

Figure 5 – figure supplement 2.

Video 1. Mating behavior of T. thermophila.

Video 2. Mating behavior of T. thermophila. To distinguish cells of different mating types, smaller mating type VI cells and larger mating type VII cells were used in this experiment.

Video 3. Cells of different mating types form a pair, whereas cells of the same mating type become separated after a short contact. To distinguish cells of different mating types, smaller mating type VI cells and larger mating type VII cells were used in this experiment.

Figure 3–Source data 1. IP-MS results. Figure 3–Source data 2. Gene identifiers.

Figure 4–Source data 1. MS analysis of MTA7-HA cilia protein.

Figure 4–Source data 2. MS analysis of MTB2-eGFP cilia protein.

Appendix 1—table 1. Strains used in this study.

Appendix 1—table 2. Primers used in this study.