The methylome of motile cilia

Abstract

Cilia are highly complex motile, sensory, and secretory organelles that contain perhaps 1000 or more distinct protein components, many of which are subject to various posttranslational modifications such as phosphorylation, N-terminal acetylation, and proteolytic processing. Another common modification is the addition of one or more methyl groups to the side chains of arginine and lysine residues. These tunable additions delocalize the side-chain charge, decrease hydrogen bond capacity, and increase both bulk and hydrophobicity. Methylation is usually mediated by S-adenosylmethionine (SAM)-dependent methyltransferases and reversed by demethylases. Previous studies have identified several ciliary proteins that are subject to methylation including axonemal dynein heavy chains that are modified by a cytosolic methyltransferase. Here, we have performed an extensive proteomic analysis of multiple independently derived cilia samples to assess the potential for SAM metabolism and the extent of methylation in these organelles. We find that cilia contain all the enzymes needed for generation of the SAM methyl donor and recycling of the S-adenosylhomocysteine and tetrahydrofolate byproducts. In addition, we find that at least 155 distinct ciliary proteins are methylated, in some cases at multiple sites. These data provide a comprehensive resource for studying the consequences of methyl marks on ciliary biology.

• Motile cilia are highly complex organelles containing 1000 or more different protein types many of which are subject to various posttranslational modifications. Several studies indicate that some ciliary components such as axonemal dynein motors are methylated at key sites that likely modify their function. However, the degree to which this modification impacts on ciliary proteins and the abundance of methylated sites was unknown.

• The authors assessed the extent of methylation in motile cilia and identified over 150 different methylated ciliary components. Furthermore, cilia contained all the enzymes needed for generation of the S-adenosyl methionine methyl donor and for recycling of the methylation byproducts.

• This analysis provides a comprehensive resource for studying the role of methyl marks in ciliary biology.

INTRODUCTION

Cilia are microtubule-based cellular extensions that play key roles in generating cell motility and fluid flow. They also provide dedicated signaling platforms (e.g., for the hedgehog pathway needed during mammalian development) (Reiter and Leroux, 2017), and for the processing and regulated secretion of bioactive products such as degradative enzymes (Wood et al., 2013) and peptidergic signaling factors (Luxmi et al., 2019, 2022). These organelles are very highly conserved dating to before the last eukaryotic common ancestor (Satir et al., 2008; Mitchell, 2017), and are found throughout the eukaryotes with a few notable exceptions such as red algae, angiosperms, and the amoeboid filasteria (Carvalho-Santos et al., 2011; Kumar et al., 2019). Cilia are very complex and proteomic studies in a broad array of organisms, such as Chlamydomonas reinhardtii (Pazour et al., 2005; Sakato-Antoku and King, 2022), Paramecium tetraurelia (Arnaiz et al., 2009), Trypanosoma brucei (Subota et al., 2014), Euglena gracilis (Hammond et al., 2021) and metazoans, for example, (Ostrowski et al., 2002; Ishikawa et al., 2012; Blackburn et al., 2017; van Dam et al., 2019) suggest cilia contain perhaps 1000 or more different protein types; in humans, this equates to about 5% of all identified genes being involved in ciliary formation and/or function.

The general architecture of motile cilia consists of an axoneme comprising nine outer doublet microtubules surrounding a central pair of microtubule singlets. This is encompassed by a membrane that is contiguous with the plasma membrane although it has a very distinct protein and lipid composition (Garcia et al., 2018; Kumar et al., 2019). Microtubule-associated motors – the axonemal dyneins – are arrayed along the ciliary length in two rows. These dyneins are controlled by additional axonemal structures including the nexin-dynein regulatory complex (N-DRC), radial spokes, and a complex pattern of repeating structures associated with the central pair microtubules (Nicastro et al., 2006; Grossman-Haham et al., 2021; Gui et al., 2021; Han et al., 2022) as well as by Ca²⁺ and redox signals (Kamiya and Witman, 1984; Harrison et al., 2002; Wakabayashi and King, 2006). Cilia are assembled at their distal tip (Johnson and Rosenbaum, 1992) and the needed components are, in general, trafficked from the ciliary base by the intraflagellar transport (IFT) system (Rosenbaum and Witman, 2002). The IFT particle scaffolds (or trains), onto which cargo components such as tubulins are loaded, are moved to the ciliary tip by specialized kinesin motor(s) and returned to the base by a dedicated dynein complex (reviewed in [Klena and Pigino, 2022]).

Many ciliary components are subject to various posttranslational modifications that alter their function. For example, axonemal tubulins can be acetylated, tyrosinated, detyrosinated, polyglutamylated, polyglycylated, and/or phosphorylated amongst other reported alterations (Janke and Magiera, 2020). Mass spectrometry and biochemical analyses of Chlamydomonas cilia has identified numerous other components that are subject to phosphorylation (Piperno and Luck, 1981; King and Witman, 1994; Boesger et al., 2009). In a similar vein, most axonemal dynein proteins are subject to N-terminal acetylation that in some cases requires the prior action of a methionine aminopeptidase (Sakato-Antoku et al., 2023).

Methylation is a highly abundant post-translational modification that is found throughout the bacteria, archaea, and eukaryotes. In addition to histones, many other protein types are known to be methylated such as bacterial flagellin, Rubisco, RNA-binding proteins, cytochrome c, and myosin (Murn and Shi, 2017). Unlike phosphorylation which acts as a binary switch with respect to side-chain charge and steric bulk, the effects of methylation are more nuanced due to the successive nature of methylation posttranslational modifications (Wesche et al., 2017; Luo, 2018; Figure 1A). For example, lysine side chains can be mono-, di-, or tri-methylated. Although this does not alter the +1 charge on the side chain per se at physiological pH, it has more subtle effects. Instead of being focused only on the N_ε atom, the charge becomes increasingly delocalized as more methyl groups are added; for trimethyl-lysine, the charge also becomes obligate as the side chain no longer exhibits a pK_a. In addition, the number of potential hydrogen bonds that can be formed decreases (eventually to zero) as methylation increases. These modifications also increase the bulk and, perhaps more importantly, the hydrophobicity of the side chain. Similarly, arginine residues can be mono- or di-methylated; in this case, dimethylation is either asymmetric with both methyl groups attached to the same side chain N atom, or symmetric with both N and N’ atoms of the guanidino group being monomethylated. Again, these additions delocalize the +1 charge, decrease hydrogen bond capacity, and increase bulk and hydrophobicity. Eukaryotes express a broad array of arginine- and lysine-methyltransferases which use S-adenosylmethionine (SAM) as the methyl donor to modify target proteins and also numerous demethylases that act to remove the methyl marks. Consequently, methylation is a tunable and reversible step-wise modification with the capacity to exert varied effects that can impact protein–protein interactions and/or enzymatic functions.

FIGURE 1:

Methylation Products and the SAM-dependent Protein Methylation Pathway. (A) Space-filling models for methylated Lys and Arg residues colored by element (C, yellow; H, silver; N, blue; O, red). As methyl groups are added to the N_ε atom of the Lys side-chain, the +1 charge becomes more delocalized and for trimethyl Lys obligate. In addition, the number of possible hydrogen bonds decreases, and both side chain bulk and hydrophobicity increase. The Arg guanidino group can be monomethylated on either N atom. When a second methyl group is added, the result can be asymmetric methylation with both methyl groups on a single N, or symmetric with both N atoms carrying a single methyl addition. (B) Schematic illustrating S-adenosyl methionine metabolism. All the key enzymes (indicated in red) needed for SAM utilization, SAH hydrolysis, conversion of homocysteine to methionine, recycling of THF to 5-MeTHF, and the subsequent ATP-dependent regeneration of SAM are present in cilia. Protein-CH₃ represents a monomethylated product.

Although there have been several reports of ciliary protein methylation in Chlamydomonas and on their potential involvement in ciliary disassembly (Schneider et al., 2008; Werner-Peterson and Sloboda, 2013) and dynein motor function (Sakato-Antoku et al., 2024), there has not been a comprehensive analysis aimed at defining the extent to which cilia components are subject to this modification. Here, we have reexamined multiple Chlamydomonas cilia proteomes and several electrophoretically-isolated axonemal high molecular weight component proteomes obtained by high-resolution tandem mass spectrometry searching for arginine and lysine methylation. We find that these variable posttranslational modifications are highly abundant on many motile cilia components. As methylation has multiple and subtle effects on the properties of basic residues, the ciliary methylome described here provides a resource for studying how this tunable and reversible addition might alter or regulate the activities of various ciliary subsystems.

RESULTS AND DISCUSSION

Cilia contain the enzymes needed for S-Adenosyl methionine (SAM) metabolism

The key methyl donor for protein methylation reactions is SAM (see Figure 1B for an outline of SAM metabolism). This is generated from methionine and ATP by SAM synthetase with the concomitant release of phosphate and pyrophosphate. Following the methylation reaction where the methyl group is covalently attached to a Lys or Arg side chain by a protein methyltransferase, the S-adenosyl homocysteine (SAH) byproduct is hydrolyzed by SAH hydrolase to yield adenosine and homocysteine. The latter is then converted back to methionine by METE methionine synthase-mediated methyl transfer from 5-methyltetrahydrofolate (5-MeTHF). 5-MeTHF is generated from 5,10 methylene-tetrahydrofolate (5,10-MeTHF) by tetrahydrofolate reductase (THFR). In turn, 5,10-MeTHF is synthesized from serine and THF by serine hydroxymethyl transferases with glycine as a by-product. It was noted previously (Sloboda and Howard, 2009) that several of these enzymes were found in the original Chlamydomonas ciliary proteome (Pazour et al., 2005). Since then, numerous additional whole cilia proteome datasets have become available that provide further support for this original report (e.g., [Sakato-Antoku and King, 2022; Sakato-Antoku et al., 2024] and see The Chlamydomonas Flagellar Proteome Project for a comprehensive recent compilation of numerous studies). These new proteomes also extend the initial observations revealing that all the key enzymes needed for SAM synthesis, SAH recycling and THF metabolism are present in Chlamydomonas cilia (Figure 1B; Table 1). Therefore, for this metabolic cycle to function in cilia requires only initial starting amounts of methionine (or SAH) and THF and replenishable sources of ATP, NADH, and serine.

TABLE 1:

S-Adenosyl methionine metabolism enzymes found in cilia.

Enzyme	Cre#	Gene symbol	Function
S-Adenosyl methionine synthetase	Cre06.g250200	METM1	Converts methionine and ATP to SAM + Pi+ PPi
Cobalamin-independent methionine synthasea	Cre03.g180750	METE1	Methylates homocysteine to generate methionine. Requires 5-MeTHF
NAD(H)-dependent THFR	Cre10.g433600	—–	Converts 5,10-MeTHF to 5-MeTHF. Requires NADH and generates NAD+
SAH hydrolase	Cre03.g204250	SAH1	Hydrolyzes SAH to yield adenosine + homocysteine
Serine hydroxymethyl transferaseb	Cre06.g293950	SHMT2	Combines serine + THF to yield 5,10-MeTHF + glycine

^aIn the current version of Phytozome the annotation for cobalamin-dependence/independence is incorrect for both this protein and the product of the METH1 (Cre06.g250902) gene.

^bAlthough predicted to be mitochondrial by TargetP, a second serine hydroxymethyl transferase (SHMT1; Cre16.g664550) was also identified in several cilia proteomes.

Overview of the cilia protein methylome

To identify methylated proteins in Chlamydomonas cilia, we examined nine independently obtained proteomes; eight represented triplicate cilia samples from CC-124 and CC-125 wildtype vegetative and gametic cells that had been fractionated by detergent extraction to yield membrane plus matrix and axonemal samples (Sakato-Antoku and King, 2022), while the ninth proteome derived from analysis of whole cilia from the CC-125 strain (Sakato-Antoku et al., 2023, 2024). In addition, we also examined the proteins present in the high molecular weight regions (>250 kDa) of SDS–PAGE gels loaded with isolated ciliary axonemes from wildtype (CC-125) Chlamydomonas and the oda2 mutant which lacks outer arm dyneins. In combination, this led to the identification of 437 methylated sites on Arg/Lys residues in 155 different proteins (Figure 2, A and B). The distribution of methylated proteins in ciliary fractions generated by detergent and high salt extraction and the overall number of methylated sites per protein are shown in Figure 2, C and D, respectively; in the few cases where multiple Arg/Lys residues occurred within a short peptide region, the modified site(s) were assessed by manual inspection of the individual spectra for the methylated peptide.

FIGURE 2:

Overview of Cilia Protein Methylation Parameters. (A) Plots illustrating the distribution of the five types of methylated residues identified; the mass spectra do not distinguish between asymmetric and symmetric dimethyl Arg. In addition to all combined sites, plots for both axonemal dyneins and nonaxonemal dynein components are shown. Totals represent the number of sites identified. (B) A total of 155 methylated ciliary proteins were identified. This plot illustrates their distribution in related groups. The most common categories are axonemal dynein proteins, FAPs not assigned to other categories, and proteins for which there is currently no, or at best extremely limited, annotation in Phytozome. (C) Proteomic analysis of methylated proteins from various ciliary fractions identifies four distinct groups: 1) those tightly associated with the axoneme; 2) detergent-soluble proteins in the membrane plus matrix; 3) proteins extracted from demembranated axonemes by 0.6 M KCl; and 4) those for which fractionation data is uncertain (i.e., similar numbers of peptides in more than one fraction) or unavailable (for proteins found only in the whole cilia proteome). (D) The number of methylation sites on Arg and Lys residues per protein is plotted. The majority of identified proteins are methylated on a single residue. Forty-seven proteins had at least one methyl modification on both residue types.

Of all the identified methylation sites, 161 modifications were found in axonemal dynein components or closely associated proteins; most of these sites were reported previously (Sakato-Antoku et al., 2024). In this group, Arg modifications were mainly monomethyl (75) with a total of only seven dimethyl Arg residues identified, whereas multimethyl Lys modifications were somewhat more abundant. In contrast, although non-axonemal dynein proteins had a similar overall proportion of modified Arg as compared to Lys residues, the amount of dimethylated forms was considerably increased (Figure 2A). These differences likely reflect the specificity of the various methyltransferases involved. All identified methylated sites on cilia proteins are shown in Table 2.

TABLE 2:

Methylated sites identified in chlamydomonas cilia proteins.^†

Cre#	Description	Kme	Kme2	Kme3	Rme	Rme2
	Membrane proteins
Cre09.g392867	Flagella major membrane glycoprotein FMG1B	K881, K1190, K2263, K2553, K2649, K2659, K2849, K3124, K3179, K3781, K3949, K3962	K2263	—	R485, R1526, R2515, R3051, R3700,	R2810
Cre18.g750047	Flagella major membrane glycoprotein FMG1A	K1031, K2146, K3029, K3274, K3753, K3954	K3550	—	R481, R3730,	—
Cre16.g650600	Mastigoneme-like protein MST1	K1176, K1567	K1176	—	R1171	—
Cre07.g340450	PKHD1 (similar to Fibrocystin)	K299	—	—	—	—
	Tubulins
Cre12.g542250	Beta-1 tubulin TUB1	K216	K216	K216	R213, R262	R213
Cre03.g190950	Alpha-1 tubulin TUA1	—	K430	—	R79, R105	—
	Kinesin
Cre09.g386700	Kinesin 13 (KIN13A) microtubule depolymerase	—	—	—	R509	R511
	Radial spoke and central pair proteins
Cre03.g201900	Radial spoke protein 1 (RSP1)	K315	—	—	—	—
Cre07.g330200	Radial spoke protein 9 (RSP9)	K125	—	—	R156	—
Cre01.g025400	Hydin	K1770, K3258, K3534, K3964, K4401	—	—	R178, R263, R2569, R3706, R3818	—
Cre06.g256450	FAP119 central pair protein (in C1a with PF6)	K260	K80	—	—	—
Cre06.g271150	FAP74 central pair protein (in C1d)	—	K132	—	—	—
Cre10.g434400	PF6 central pair protein (in C1a)	K4, K6, K7, K29, K116, K410	K4, K6, K410	K4	R820, R1149	—
Cre16.g654800	CPC2 central pair protein	K2271	—	—	R1200, R2116	R1208
	IFT
Cre11.g467739	IFT54 (in IFT-B2 complex)	—	K184,	K183, K188	—	—
Cre01.g027950	IFT74 (in IFT-B1 complex)	K79	—	—	R80, R87, R102	—
Cre06.g250300	IFT dynein DHC1b (DHC16)	K1776, K2006, K2340, K4231	—	K3438	R1585, R2164, R2685	—
Cre09.g398882	IFT dynein D1b LIC (DLI1)	—	K53	—	—	—
Cre02.g110950	IFT dynein D1bIC2 (DIC5, WDR34, FAP133)	K53	—	—	—	—
Cre10.g428664	IFT dynein D1bIC1 (DIC6, WDR60, FAP163)	—	—	—	R626	—
	Axonemal dyneins* and associated proteins
Cre12.g484250	I1/f alpha HC (DHC1)	K1682, K2582, K2951, K3983	—	K1371	R1069, R1284	R666
Cre09.g392282	Dynein d heavy chain (DHC2)	—	—	—	R81, R1541, R1594, R2090, R2629	R87
Cre02.g107350	Minor dynein (DHC4)	K3552	—	—	R4510	—
Cre02.g107050	Dynein b HC (DHC5)	K2106, K2924, K3001	—	—	R1495, R3704, R4140	R2753
Cre05.g244250	Dynein a HC (DHC6)	K2940, K3148	—	—	R134, R1664, R3847	—
Cre14.g627576	Dynein g HC (DHC7)	K2089, K2366, K3721, K4144	—	—	R436, R1383, R2267, R2327, R3576, R3769	—
Cre16.g685450	Dynein e HC (DHC8)	K1104, K1201, K2967	—	—	R1470, R1523, R1883, R3334, R4126, R4166	—
Cre02.g141606	Dynein c HC (DHC9)	K1174, K2789, K3222, K3639, K3699, K3773	—	—	R540, R1374, R3020, R3028, R4062,	—
Cre14.g624950	I1/f 1beta HC (DHC10)	K1541, K1709, K1878, K3283, K4059	K2453	K2459, K3193	R524, R2193, R2832, R3694, R3699	—
Cre12.g555950	Minor dynein (DHC11)	K2291, K4274	—	K3626	R4	R12, R2303
Cre03.g145127	OAD alpha HC (DHC13)	K110, K253, K1436, K1791, K2840, K3044, K3065	K3072	K3066, K3439, K3445	R18, R38, R647, R791, R1119, R1168, R1418, R1702, R1777, R2111, R2500, R3139, R3654, R3655, R3678, R3720, R3756, R3816, R3878, R4022, R4068	R1945
Cre09.g403800	OAD beta HC (DHC14)	K296, K527, K756, K1989, K2149, K2208, K2401, K2491, K2998, K3223, K3635, K4114	K653, K990, K1319, K2606, K3169, K3230, K3296, K3876	K2634, K3224	R1621, R1975, R2332, R2488, R2710, R2763, R3202, R3467, R4285,	R3959
Cre11.g476050	OAD gamma HC (DHC15)	K1533, K1744, K2952, K2997, K4228	K292	K3079	R213, R1317, R1373, R1904, R2713, R3678, R4424	—
Cre01.g029750	ODA5-associated adenylate kinase (DAAK1)	K521, K603, K1580	—	—	—	—
Cre12.g494800	IDA4/p28 (DII1)	—	—	—	R108	—
Cre11.g476850	DRC4 (N-DRC)	—	—	K4	—	—
	Centriole/basal body/transition zone
Cre12.g559250	Centriole protein 14 (POC14)	—	—	K253	—	—
Cre13.g607650	Centriole protein 15 (POC15). Many EF hands	—	—	—	—	R2914, R2919
Cre10.g432850	FAP77 (also in basal body as BUG1)	—	—	—	R196	—
Cre14.g632050	Retinitis pigmentosa GTPase regulator interacting protein 1 RPG1 (RPGRIP1)	—	K695	K743	R748	—
	FAPs
Cre17.g747247	FAP2 (TPRs, similar to kinesin light chain)	—	—	—	R605	—
Cre09.g390615	FAP12 Diacylglycerol lipase (LIP1)	K70, K141, K313, K401, K418	—	—	—	—
Cre10.g450450	FAP18 (rhomboid membrane protease)	K138	—	—	—	—
Cre09.g386736	FAP44 (WDR52)	K153	—	—	—	—
Cre17.g704300	FAP47 (weakly similar Hydrocephalus 3 has calponin homology domain)	K1012	—	—	—	—
Cre03.g171900	FAP56	K355	—	—	R402	R354
Cre17.g701250	FAP59 (CCDC39)	—	—	—	—	R52
Cre06.g296850	FAP81 (similar to deleted in lung and esophageal cancer; DLEC1)	—	K81	—	—	R86
Cre09.g389150	FAP93	K1077	—	—	—	—
Cre09.g393954	FAP106 (microtubule inner protein)	K26	—	—	—	—
Cre02.g145850	FAP112	K262	—	—	—	—
Cre17.g737100	FAP127	—	—	K72	—	—
Cre16.g693600	FAP137 Hydroxyproline rich	K333	—	—	R385	—
Cre09.g387912	FAP139 (CCDC139)	—	—	—	R106	—
Cre11.g468850	FAP152	—	K82	—	—	—
Cre08.g362100	FAP154 (PAS domain)	K2562	—	—	—	—
Cre14.g624900	FAP171 (armadillo folds)	—	K4	—	—	—
Cre09.g395769	FAP177	K508	K2172, K2175	—	R1995	—
Cre01.g032150	FAP187	—	—	—	R847	R881
Cre01.g004550	FAP190	—	—	—	R319	—
Cre17.g728650	FAP196 (WD repeats)	—	—	K147	—	—
Cre02.g078550	FAP210 (trichohyalin-plectin homology domain)	K95	—	—	—	—
Cre07.g350350	FAP217 (coiled coil)	—	—	—	—	R1139, R1141
Cre14.g618750	FAP246 (LRRs, transglutaminase-like and EF hands)	—	—	—	—	R997
Cre12.g518050	FAP261 translin-associated factor X interacting protein 1-like - involved in spermatogenesis	—	—	—	R596, R597	—
Cre17.g717150	FAP263 (CCDC113)	—	K130, K358	K358, K361	R136, R360	R360
Cre15.g637050	FAP274	K108	—	—	—	—
Cre16.g679550	FAP277	K5	—	—	—	—
Cre18.g749797	FAP281 (coiled coil protein)	—	—	—	R1723, R1728	R1728
Cre17.g740261	FAP365 (TPR repeats)	—	—	—	R761	R671
Cre11.g467580	FAP370	K589	—	—	R566	—
Cre11.g468900	FAP404 (Volvocales-specific)	—	—	—	R240	—
Cre10.g429750	FAP417	K705	—	—	—	—
	Kinases
Cre12.g538300	LF5 (FAP247) protein kinase	—	—	—	R554	—
Cre11.g481400	S/T/dual Kinase	—	K983	—	—	R630
Cre12.g555450	Inositol tetrakisphosphate 5 kinase	—	—	K77	—	R76
Cre03.g164250	Protein kinase + IQ motif (FAP262)	K901, K915, K921, K935, K1640	—	—	—	—
Cre12.g508900	Mitogen activated protein kinase 6 (MAPK6)	K110	—	—	R114	—
Cre02.g075900	Protein kinase (SNRK2C)	—	—	—	R155	—
	Putative NTPases
Cre13.g563700	FAP369. NTPase, alcohol dehydrogenase and WD repeat regions	—	—	K303	—	—
Cre06.g249900	FAP75 contains NAD(P) binding domain, NTPase Adenylate or UMP-CMP kinase	—	—	—	R948	—
Cre06.g269950	CDC48 AAA ATPase	K343	—	—	—	R345
	Metabolic and other enzymes
Cre03.g180750	Cobalamin-independent methionine synthase METE1	—	—	—	R537	—
Cre12.g551800	Bis (5’-nucleosyl) tetra-phosphatase (generates ADP)	—	—	—	R132, R134	—
Cre10.g433600	Methylene THFR	K10, K12, K13	—	—	—	—
Cre05.g234700	Pyruvate kinase 3 (PYK3)	—	—	—	R2261	—
Cre12.g513200	Enolase (ENO1)	K350	—	—	—	—
Cre06.g282800	Isocitrate lyase (ICL1)	K233	—	—	—	—
Cre16.g682725	Glutathione S-transferase	K124	—	—	—	—
Cre05.g233300	SAM-dependent methylase (dipthine synthase family)	K169, K192	—	—	R178, R180, R187	—
Cre12.g483800	Serine carboxypeptidase	—	—	—	R945, R947, R948	—
Cre14.g620300	Anthranilate synthase beta subunit	—	K73	—	—	—
Cre02.g115950	Glucan endo-1,3-beta-D-glucosidase	—	—	K136	—	—
	EF-hand proteins
Cre01.g004200	No annotation. Has EF hand	—	—	—	—	R1115
Cre06.g287100	Calmodulin-like. Two EF hands	—	—	—	—	R4
Cre07.g334450	Two EF hands	—	—	K55	—	R52
Cre13.g604050	EF hand motif. Pectin lyase fold	—	—	—	R2491	R2492
Cre16.g652400	FAP183 (EF hands)	K1812	—	—	—	—
Cre03.g178350	FAP272 (similar to calmodulin)	—	K135	—	—	—
	Induced upon gametogenesis
Cre06.g251050	Possible kinase. Coexpression cluster with mating activation induced in minus gametes	—	—	—	R907	—
Cre17.g698903	No annotation. Coexpression cluster with mating activation induced in minus gametes	—	—	K1568	R709	—
Cre03.g206700	No annotation. Coexpression cluster with mating activation induced in plus gametes	—	—	—	R12, R15	—
Cre12.g511350	No annotation. Coexpression cluster with mating activation induced in plus gametes	—	—	—	R614	—
	Miscellaneous annotated proteins
Cre09.g397142	Acid induced Ca2+ TRP channel ADF1	—	—	—	R1617	—
Cre09.g400850	Carbohydrate binding protein (CTL4)	K4305	—	—	—	—
Cre09.g398900	Vegetative cell wall protein CWP1	K432	—	—	—	—
Cre06.g258800	Vegetative cell wall protein CWP2	K1234	—	—	—	—
Cre03.g149800	MAP TOG1 (XMAP215)	K1171	—	—	R1158	—
Cre03.g204913	Shippo-1 related, sperm tail PG repeats	—	—	—	R884	R879
Cre01.g003950	Tiny macrocysts protein B-related	K1282	—	K587	—	R585
Cre02.g110450	Proteasome interacting thioredoxin	—	—	—	—	R64
Cre02.g114600	Peroxiredoxin PRX2	—	K110	—	—	—
Cre08.g378000	LSM and DFDF domains plus TFG box	—	—	—	—	R332, R336
Cre06.g297016	DNA excision repair (ERCC3)	—	K846	K850	—	—
Cre09.g393358	Vasa intronic gene homolog VIG1	—	—	—	R117, R120	R115
	Translation-related
Cre06.g263450	Eukaryotic translation initiation factor 1 alpha EEF1A3	K43, K306	K435,	K406, K435	R37, R253, R442	—
Cre04.g222700	Elongation factor 3 (EF3)	K1023	—	K1027	—	—
Cre04.g217550	Eukaryotic initiation factor 3 (eIF3C)	—	—	K242	—	—
Cre02.g097400	Eukaryotic translation initiation factor 5A EIF5A	K89	K89	—	R88	—
Cre06.g252850	Plant specific eukaryotic initiation factor 4b (eIF-4B)	—	—	—	—	R212
Cre12.g516200	Elongation Factor G (EFG2)	—	K511	—	R514	—
Cre02.g143200	Alanyl tRNA synthetase	K604	—	—	—	—
Cre03.g160500	Lysyl tRNA synthetase	K185	—	—	—	—
Cre08.g372100	HSP70A	K563	—	—	R475	—
Cre14.g613550	HSP70H (FAP367 or FAP203)	K139	—	—	R81	—
	No (or very limited) gene annotation
Cre01.g009850	No annotation. (Volvocales-specific)	—	—	—	R647	R647
Cre01.g049300	No annotation	—	—	—	—	R612
Cre02.g084150	No annotation	—	—	—	—	R1775
Cre02.g103800	No annotation. LRR protein	—	—	—	R386	—
Cre03.g152250	No annotation	—	—	K180	—	—
Cre03.g162150	No annotation	—	K1579	—	—	R1581
Cre03.g170900	No annotation. Aldehyde/histidinol dehydrogenase domain	K147	—	—	—	—
Cre04.g217952	No annotation. Contains zinc finger	—	—	—	—	R463, R571, R572
Cre04.g232202	No annotation	—	—		R3232	R3235
Cre06.g256600	No annotation	—	—	—	R731	—
Cre06.g278186	No annotation. Toll receptor homology and armadillo fold	—	—	—	—	R910
Cre06.g299100	No annotation	—	—	—	R249, R254	—
Cre07.g354650	No annotation. cNMP binding domain	—	—	—	R1280	—
Cre10.g419400	No annotation. Contains zinc finger	—	—	—	—	R611
Cre10.g427850	No annotation. PAS domain	K705	—	—	—	—
Cre10.g801108	No annotation	—	—	—	R98	—
Cre11.g477950	No annotation. Crystallin domain	K553	—	—	—	—
Cre12.g540050	No annotation. cNMP binding domain.	—	—	—	R59	R61
Cre12.g555378	No annotation	K446	—	—	—	R458, R869, R872
Cre15.g641950	No annotation	—	—	K561	R1095, R1098, R1103	—
Cre16.g696050	No annotation	—	—	—	—	R1630, R1631
Cre16.g669750	No annotation	—	—	—	—	R488
Cre16.g801909	No annotation may be a TRP channel	K208	—	—	R210	—
Cre17.g727550	No annotation	K1062	—	—	R1808	—
Cre17.g744747	No annotation	—	K18	—	—	R57
Cre18.g749997	No annotation	—	—	—	R30	—

^†Residue numbers refer to the protein sequences shown in the CC-4352 v6.1 Chlamydomonas reinhardtii genome release in Phytozome 13 (https://phytozome-next.jgi.doe.gov/).

^*Most sites identified in axonemal dynein HCs were reported previously (Sakato-Antoku et al., 2024). No methylated sites have been identified on the minor inner arm dyneins DHC3 (Cre06.g265950) and DHC12 (Cre06.g297850). Potential methylation of outer arm dynein LC1 and inner arm I1/f dynein IC97 was not confirmed following in-depth manual analysis of the raw MS/MS spectra.

As an example of the evidence for the modifications observed, unambiguous mass spectral identification of di-methylation on K₂₁₆ of β-tubulin based on near complete b- and y- fragment ion coverage is shown in Figure 3A and the location of all methylated residues within the α/β tubulin dimer in Figure 3B. In multiple cases, a given Arg or Lys residue was identified in more than one methylated state, for example, K₂₁₆ of β-tubulin and K₄ of the central pair protein PF6 were found as the mono-, di-, and trimethylated forms.

FIGURE 3:

Methylation of Axonemal Tubulins. (A) Mass spectral analysis of a dimethylated β-tubulin peptide. Annotated higher energy C-trap dissociation (HCD) MS/MS spectrum and the corresponding fragment ion coverage with the identified b- and y-type ions for the tryptic peptide (R)TLK_me2LTTPTFGDLNHLISAVMSGITccLR (where c represents Cys+57 due to the mass increase from iodoacetamide-mediated carbamidomethylation derivatization during sample processing). High abundance doubly charged ions present in the MS/MS spectrum are not included in the sequence representation. This spectrum and the m/z values unambiguously demonstrate dimethylation of K₂₁₆ of β-tubulin. (B) Ribbon diagram of an AlphaFold2 model of the α/β tubulin dimer from Chlamydomonas; α-tubulin, pink; β-tubulin, wheat. All identified methylated residues are indicated: monomethyl, marine blue; di-methyl, split-pea; tri-methyl, red; the added methyl groups are in green (C)/silver (H). The N- and C-termini are also indicated where visible. Modified residues occur on the outer surface, the luminal face and in regions involved in inter-protofilament interactions. The right panel view is related to the left panel by an ∼ 90^o rightward rotation about the vertical axis.

Classes of methylated ciliary proteins

Modified cilia-derived proteins can be divided into multiple classes based on their ciliary location as well as a variety of functional criteria (Figure 2B; Table 2). The major classes include axonemal dynein components (n = 16), a collection of unrelated proteins for which there is little or no functional annotation in Phytozome (n = 26), and a large number of flagellar-associated proteins (FAPs) many of which are also of unknown function (n = 33). Furthermore, multiple groups containing smaller numbers of modified proteins were also classified including tubulins, EF-hand proteins that predictably bind Ca²⁺, kinases, several NTPases, radial spoke/central pair components, a single microtubule-depolymerizing kinesin (KIN13), IFT proteins, and motors, and several metabolic and other enzymes. Also identified were some membrane-associated proteins in which methylation occurs on the domains exposed to the environment.

In addition, to the axonemal dynein HCs identified previously (Sakato-Antoku et al., 2024), in this extended analysis we obtained evidence that several other dynein related proteins are subject to methylation including the ODA5-associated adenylate kinase (encoded at DAAK1), the DRC4 component of the N-DRC, and the p28 inner arm dynein LC (encoded at IDA4/DII1); modification of the p28 dimer occurs in a region involved in association with monomeric inner arm dynein HCs (Figure 4A). Similarly, both ICs and the LIC of the dynein that powers retrograde IFT were also found to be modified in at least one sample. Intriguingly within the methylated FAPs, three (FAP106, FAP112, and FAP210) are associated with the luminal face of tubulins forming the B-tubule of the outer doublets (Ma et al., 2019). FAP106 for example, forms a structure bridging B-tubule wall protofilament B10 to the A-tubule protofilaments (A12 and A13) forming the shared doublet microtubule wall; the methylated Lys residue (K₂₆) protrudes into the B-tubule lumen (Figure 4B). Whether methylation of these inner proteins occurs within the microtubule B-tubule lumen as is thought to happen with acetylation on K₄₀ by the α-tubulin acetyltransferase αTAT1 (Coombes et al., 2016) or on the free monomers before microtubule assembly remains to be determined. In several cases (e.g., IFT54 and IFT74; Supplemental Figure S1), examination of AlphaFold structural models revealed that methylation occurs on residues located within inherently disordered regions or loops as is often the case for other posttranslational modifications.

FIGURE 4:

Methylation of Inner Arm Dynein p28 and Microtubule Inner Protein FAP106. (A) AlphaFold2 multimer model of the N-terminal 1000-residue region of DHC9 (dynein c; wheat) associated with a dimer of the p28 light chain (p28 and p28’ monomers in pale green and pale blue); the region of the trimeric complex involved in ATP-independent doublet microtubule associations is at left. The region boxed by the dash lines is enlarged at right illustrating the locations in the dimer of R₁₀₈ that is monomethylated (space-filling side chains in marine blue; methyl groups in green (C)/silver (H)). (B) Ribbon representation of the 48 nm axonemal doublet microtubule repeat unit (PDB 6U42; [Ma et al., 2019]). All components are shown in light silver/blue except for FAP106 (red) and its monomethylated K₂₆ residue (space-filling side chain in marine blue). The enlarged view at right reveals that mono-methylated K₂₆ is exposed to the lumen of the outer doublet B-tubule.

Stoichiometry of methyl modifications

A broad assessment of the stoichiometry of methyl modifications on many proteins is complicated by the observation that these Arg/Lys-containing peptides represent tryptic cleavage sites. We have observed that methylation (especially tri-methylation) inhibits trypsin digestion whereas the unmodified peptides are readily cleaved often to fragments too small for unambiguous identification. Even so, analysis of endoproteinase Asp-N digests (this protease cleaves to the N-terminal side of Asp and sometimes Glu residues) does allow for some degree of comparative stoichiometry analysis. Previously, we determined that methylation of axonemal dynein HCs in cytoplasm is substoichiometric at most sites, for example, the numerous modifications that occur within the AAA+ domains (Sakato-Antoku et al., 2024). The only apparently highly methylated region found was on helix 1 of the microtubule binding domains of the α and β outer arm dynein HCs where multiple methylated patterns of a K-K-X₅-K motif were identified but the unmodified form was not. In this broader analysis, we find multiple cases where only methylated peptides can be identified for a given protein region. For example, the N-terminal segment of the central pair protein PF6 was identified from endoproteinase Asp-N digests only as (M)APK_me3PKKPEAAPPPPPEPSP(D) or (M)APK_me2PK_meKPEAAPPPPPEPSP(D). Similarly, in FAP56 only the methylated peptide was found for the sequence (D)DFER_meK_meER(D), while conversely less than 20% of peptides covering a second methylated site (R)DRQNKYLTAVEIEEGR_meKKAIR(D) in the same protein were modified. Thus, there is a considerable range of methylation levels between different proteins and even within individual proteins that may represent fine-tuned regulatory processes or responses.

Numerous methyltransferases are required for ciliary protein methylation

Analysis of the Chlamydomonas ciliary methylome reveals extensive modifications of many proteins on both Arg and Lys residues. As methyltransferases normally exhibit considerable specificity for either one residue or the other, this implies there are likely multiple enzymes dedicated to modifying particular groups or classes of ciliary proteins. These modifications could occur in the cytoplasm and the methylated product then trafficked into the cilium, or they might happen within the organelle as predicted from the presence of SAM regenerating enzymes. There is currently evidence for both modalities with respect to both Arg and Lys methylation. Recently, we found that one arginine methyltransferase (PF22; the mammalian ortholog is DNAAF3) is involved in modifying axonemal dyneins (Sakato-Antoku et al., 2024); this enzyme is present in cytoplasm but apparently completely excluded from cilia (Mitchison et al., 2012), which is consistent with our observation that axonemal dynein HCs derived from wild-type cytoplasmic extracts are methylated. For example, the H1 helix of the outer arm β HC microtubule-binding domain is methylated at the same sites in dyneins derived from cytoplasm and from cilia. Whether all identified sites are modified in cytoplasm is more difficult to assess as the additions are substoichiometric and the depth of sequence coverage is much greater for axoneme-derived samples. Furthermore, axonemal dynein HCs in cytoplasm of the pf22 null mutant lack detectable methylated Arg residues consistent with the loss of the PF22 methyltransferase. However, these same pf22-derived HCs do retain some modified Lys residues strongly suggesting that at least one additional cytosolic lysine methyltransferase acts on these motor proteins (Sakato-Antoku et al., 2024). In contrast, (Schneider et al., 2008) found that the METE methionine synthase and several unidentified proteins detected with an antibody against asymmetric dimethyl-Arg were present in resorbing cilia but were unmodified (i.e., undetectable with the antidimethyl Arg antibody) in growing or full-length steady-state organelles. This observation thereby predicts that methylation occurs within the cilium and provides a link, although it does not establish causation, between methylation status and ciliary disassembly.

In addition to internal ciliary components, we also identify here both Arg and Lys methylation of several ciliary membrane proteins on domains that are exposed to the external environment. As a source of SAM is required for these modifications, they most likely occurred within the lumen of the secretory pathway, which necessitates at least two additional methyltransferases that might modify proteins destined for cilia as they transit the endoplasmic reticulum and/or Golgi apparatus.

There are two major classes of SAM-dependent protein methyltransferase that are based on distinct secondary structures – SET (suppressor of variation, enhancer of zeste, and trithorax) and Rossmann-fold domains (Dillon et al., 2005; Chouhan et al., 2019). Mass spectrometry and immunoblot analysis using specific antibodies provides some evidence for multiple methyltransferases within cilia (Table 3). For example, four protein arginine methyltransferases – PRMT1, PRMT2 (previously annotated as PRMT10), PRMT3, and PRMT5 – were identified immunologically (Mizuno and Sloboda, 2017); two of these, PRMT1 and PRMT3, appeared concentrated at the ciliary base. Currently, there is no mass spectrometry evidence for PRMT5 in cilia, but some peptides from PRMT1, PRMT2, and PRMT3 and also PRMT7 have been found. Thus, it appears that cilia do indeed contain several distinct protein arginine methyltransferases.

TABLE 3:

Methyltransferases potentially involved in modifying ciliary proteins.^†

Gene	Cre#	Evidence for cilia localization		Properties
In Cilia		Antibody *	MS/MS +
PRMT1	Cre03.g172550	Yes	Yes	Arginine methyltransferase.
PRMT2	Cre12.g558100	Yes	Yes	Arginine methyltransferase. Previously called PRMT10.
PRMT3	Cre16.g685900	Yes	Yes	Arginine methyltransferase.
PRMT5	Cre03.g176550	Yes	No	Arginine methyltransferase.
PRMT7	Cre02.g141626	No	Yes	Arginine methyltransferase.
—–	Cre01.g014100	No	Yes	FAM86A class I lysine methyltransferase that can generate tri-methyl Lys. N-terminal armadillo fold.
HLM19	Cre12.g523650	No	Yes	SET domain. Annotated as histone lysine methyltransferase.
FAP393	Cre06.g260150	No	Yes	Methyltransferase of undefined specificity.
—–	Cre08.g362700	No	Yes	Methyltransferase of undefined specificity.
Cytoplasm only		Tagged gene expression	MS/MS +
PF22 (DNAAF3)	Cre01.g001657	No	No	Functions in cytoplasm to methylate axonemal dynein heavy chains on Arg (and possibly some Lys) residues.
Unknown	Unknown	N/A	N/A	Cytoplasmic methyltransferase that modifies Lys residues on dynein heavy chains.

^†Several methyltransferases annotated as involved in tRNA methylation have been identified in ciliary proteomes including Cre10.g422000, Cre01.g027350, Cre06.g273400, Cre13.g566300 and Cre08.g359100. Evidence for these being in cilia is generally considered weak. In addition, other methyltransferases involved in modifying non-protein substrates such as thiopurines (Cre04.g215600) and sterols (Cre12.g500500) have been found. Although Cre01.g004350 (HLM1) is also present in these proteomes and annotated in Phytozome as a histone lysine methyltransferase, BLAST searches suggest it is related to a BRO1 domain-containing protein that interacts with programmed cell death protein 6 (PCD6).

^*Antibody localization data are from (Mizuno and Sloboda, 2017). The location of PF22 is from (Mitchison et al., 2012).

⁺Mass spectrometry data from multiple cilia proteomes are compiled at The Chlamydomonas Flagellar Proteome Project.

Numerous ciliary proteins are modified on Lys residues and to date two methyltransferases which are thought to specifically modify this residue have been identified in cilia samples (see Table 3). In addition, there are two putative methyltransferases of undefined specificity (Table 3) that have been found in these organelles, and there is also evidence in multiple cilia proteomes for a SAM-dependent tetrapyrrole methylase (Cre05.g233300) that is a member of the diphthine synthase superfamily; canonical diphthine synthase is involved in the methylation of a modified His residue in elongation factor 2 (Hörberg et al., 2018).

Often posttranslational modifications occur within short linear sequence motifs such as those recognized by various kinases, for example, (Johnson et al., 2023). Analysis of the regions surrounding methylated sites within axonemal dyneins, and subsets of cilia components has not identified clear motifs that might direct the action of particular methyltransferases (Figure 5). This global analysis is however hindered by the possibility that multiple, currently unidentified, enzymes of differing specificities might be involved in these modifications. Furthermore, to date no annotated protein demethylases have been identified in any Chlamydomonas ciliary proteome. This raises the possibility that, although generally considered a reversible reaction, methylation of proteins in cilia may be irreversible at least while they remain within the organelle.

FIGURE 5:

Motif Analysis of Ciliary Methylation Sites. Methylated regions (including five residues either side of the modified Arg or Lys) from various ciliary protein classes were aligned and displayed using WebLogo. (A) Di-methyl Lys (n = 11) from whole cilia proteins, (B) tri-methyl Lys (n = 17) from whole cilia proteins, (C) mono-methyl Lys from outer arm dynein HCs (n = 24), and (D) mono- and di-methyl Arg from outer arm dynein HCs (n = 38). No specific short linear motifs with embedded methylation sites are readily evident.

In conclusion, we describe here the identification of extensive methyl posttranslational modifications on numerous protein components of motile cilia. Available data indicates that methylation can occur either in the cytoplasm or secretory pathway on proteins destined for cilia or within the organelle itself. The nuanced and tunable effects of methyl additions on Arg and Lys residues suggests this may represent a mechanism for regulating numerous aspects of cilia protein function that have to date received little consideration. Thus, our data provide a resource for the further investigation of the role(s) of methyl marks in ciliary biology. There are numerous approaches that can be used in the future to dissect the functional significance of ciliary protein methylation including site-directed mutagenesis of Arg/Lys residues located at key positions involved in enzymatic activities and/or protein–protein interactions, and analysis of the ciliary phenotype(s) of mutant strains lacking one or more of the methyltransferases identified in cilia or which are known to modify ciliary proteins in cytoplasm.

MATERIALS AND METHODS

Mass spectrometry datasets and searches

Multiple previously described ciliary proteomes from Chlamydomonas were reanalyzed for the presence of methylated residues using the Chlamydomonas reinhardtii v5.6 and CC-4352 v6.1 genome releases in Phytozome 13. One large dataset derived from cilia isolated from the wild-type strains CC-124 (mating type minus) and CC-125 (mating type plus) vegetative and gametic cells that had been fractionated into a detergent-soluble extract and a detergent-insoluble axoneme pellet. These samples were prepared in triplicate and digested with trypsin before mass spectrometry as described previously (Sakato-Antoku and King, 2022).

Posttranslational modifications that were searched in the raw data from these samples using Mascot in Proteome Discoverer included fixed carbamidomethylation on Cys, N-terminal acetylation, pyroglutamylation on Gln, oxidation on Met, phosphorylation on Ser, Thr and Tyr, and methylation on Arg and Lys.

An additional independently prepared CC-125 whole cilia proteome and several electrophoretically purified high molecular weight SDS–PAGE region samples derived from CC-125 and the outer dynein arm-less mutant oda2 (CC-2230) were digested with trypsin or endoproteinase Asp-N and analyzed using an Orbitrap Eclipse Tribrid mass spectrometer as described in (Sakato-Antoku et al., 2023, 2024). For these datasets, modifications were identified using the Andromeda search engine embedded within MaxQuant and the searches included fixed carbamidomethylation on Cys, Met oxidation, mono-, di-, and tri-methylation, and a series of N-terminal alterations including acetylation, arginylation, methylation, myristoylation, palmitoylation, and ubiquitination.

All mass spectrometry results were visualized and analyzed using Scaffold v5 (Proteome Software). In addition to the above datasets, other cilia proteomic data examined to assess the presence or absence of SAM metabolism enzymes and putative protein methyltransferases and demethylases were from (Pazour et al., 2005; Jordan et al., 2018; Wang et al., 2017; Picariello et al., 2019; Zhao et al., 2019) and are compiled and available at The Chlamydomonas Flagellar Proteome Project.

Structural modeling and graphical displays

Structural models of the Chlamydomonas α/β tubulin dimer, the p28 dimer associated with the N-terminal region of DHC9 (dynein c), IFT54, and IFT74 were generated using AlphaFold 2 (Jumper et al., 2021) running on the Colabfold server (Mirdita et al., 2022). Tubulin and p28/DHC9 models and the single particle cryoelectron microscopy near-atomic reconstruction of the Chlamydomonas outer doublet microtubule 48 nm repeat (PDB 6U42 [Ma et al., 2019]) were displayed using the PyMOL molecular graphics system (Schrödinger, LLC). Space filling models for methylated Arg and Lys residues were generated using the builder interface within PyMOL. All structural displays were created using ray tracing.

Pie charts and graphs were prepared using GraphPad Prism v.7. To search for possible methylation motifs, sequences surrounding the identified methylation sites were aligned manually and displayed using WebLogo v.3.7.12.

All figures were constructed using Adobe Illustrator and/or Photoshop.

Data availability

Sequence data are available at Phytozome https://phytozome-next.jgi.doe.gov/. Cilia proteomic data are available at Dryad with the dataset identifiers fn2z34txn and mw6m90635. Additional mass spectrometry data from methylation searches of cilia samples have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD045935.

Abbreviations used:

AAA+

ATPase associated with cellular activities domain

FAP

flagellar associated protein

HC

heavy chain

IC

intermediate chain

LC

light chain

LIC

light intermediate chain

K _me

monomethyl lysine

K _me2

dimethyl lysine

K _me3

trimethyl lysine

5-MeTHF

5-methyltetrahydrofolate

5,10-MeTHF

5,10-methylenetetrahydrofolate

NTPase

nucleoside triphosphatase

R _me

monomethyl arginine

R _me2

dimethyl arginine

SAH

S-adenosyl homocysteine

SAM

S-adenosyl methionine

THF

tetrahydrofolate.