The multifaceted roles of microtubule-associated proteins in the primary cilium and ciliopathies Free

Cilia are microtubule-based organelles that protrude from the surface of most eukaryotic cells. They are classified into non-motile cilia, which transduce signaling pathways essential for development and homeostasis, and motile cilia, which generate fluid flow in organs or facilitate the movement of single cells like sperm or unicellular organisms. All cilia share an evolutionarily conserved architecture, including the basal body, the microtubule-based axoneme and the ciliary membrane (Fig. 1). Their assembly and disassembly, which is linked to the cell cycle, involve coordinated activities of numerous molecules, organelles and signaling pathways (Breslow and Holland, 2019). Deregulation of cilium structure, composition and biogenesis is associated with both cancer and ciliopathies, the latter being characterized by multisystem pathologies of the eye, kidney, skeleton, brain, lung and other organs (Higgins et al., 2019; Reiter and Leroux, 2017). Defining how a functional cilium is built and maintained is crucial for understanding the pathogenesis of ciliary diseases.

Cilium biogenesis and function have been extensively studied across various cell types, tissues and organisms. These studies have defined the molecular players in cilium assembly, disassembly, signaling and trafficking, elucidating their molecular mechanisms of action. However, the mechanisms governing the structure of the axoneme, and its composition and biochemistry, which allow cilia to sense stimuli or facilitate movement, remain largely unexplored. Recently, advanced cell-biological, structural, proteomic and biochemical methods have begun to illuminate the unique complexity of the axoneme as a microtubule-based structure, providing insights into the formation of these diverse microtubule arrays within cells.

Much of our understanding of axoneme composition and its three-dimensional (3D) structure largely comes from studies on motile cilia in multiciliated epithelia of vertebrates and the flagella of ciliated organisms, such as Chlamydomonas, Tetrahymena and sea urchins. For more detailed information, readers can refer to comprehensive reviews on these subjects (Klena and Pigino, 2022; Lee and Ostrowski, 2021). In this Review, we will focus on the primary cilium, drawing parallels to motile cilia when relevant. We begin by presenting current knowledge about the organization, biogenesis and composition of the primary cilium, laying the foundation for an in-depth discussion of the axoneme. Subsequently, we discuss the recent advances regarding the architecture, biochemical and dynamic properties, and formation of the axoneme, as well as the role of microtubule-associated proteins (MAPs) in its regulation. We conclude by examining the multifaceted roles and localizations of ciliary MAPs and their impact on the pathobiology of ciliary diseases.

The primary cilium is compartmentalized into subdomains with distinct architecture, composition and function: the axoneme, basal body, transition zone and ciliary tip (Lee and Chung, 2015). The axoneme is the microtubule skeleton of the cilium, and hence supports its structure and provides tracks for ciliary transport complexes. During cilium formation, the nine-fold symmetric microtubule doublets of the axoneme are templated from the A- and B-tubules of the basal body microtubule triplets (Fig. 1). Distal to the basal body at the proximal part of the cilium is the transition zone, which is characterized by Y-shaped linkers connecting microtubule doublets of the axoneme to the ciliary membrane. The transition zone regulates the entry and exit of molecules and, thereby, ciliary composition (Dean et al., 2016; Garcia-Gonzalo et al., 2011; Goncalves and Pelletier, 2017; Szymanska and Johnson, 2012). The distal end of the cilium between the ciliary membrane and the plus-ends of axonemal microtubules is the ciliary tip, which has been implicated in the addition or removal of tubulin dimers to and from axonemal microtubules, inactivation of the anterograde intraflagellar transport (IFT) motor kinesin-2 and activation of the retrograde motor cytoplasmic dynein-2, as well as in regulation of Hedgehog (Hh) signaling (Soares et al., 2019).

Primary cilium assembly and disassembly are highly regulated, multistep processes coordinated with the cell cycle. In most animal cells, the primary cilium exists in quiescent or differentiated cells that are in G0/G1 and it is removed through resorption or excision when cells re-enter the cell cycle (Mirvis et al., 2018). At the onset of ciliogenesis, the mother centriole matures into a basal body, docking to the plasma membrane via distal appendages, also known as transition fibers (Ma et al., 2023; Tanos et al., 2013). To initiate axoneme elongation, the centriole distal cap complex, comprising CP110 (also known as CCP110) and Cep97, is removed from the basal body (Spektor et al., 2007). Given that cilia lack ribosomes, axonemal elongation requires transporting its building blocks, such as tubulin dimers, from the cytoplasm to the growing microtubule tips, which is facilitated by free diffusion and motor-based IFT. As the axoneme elongates, it protrudes from the plasma membrane, which forms the ciliary membrane. When the axoneme reaches its steady-state length, its microtubules continuously polymerize and depolymerize at their plus-ends. The composition of the primary cilium during its assembly and maintenance is mediated by various molecular machines, including the transition zone, IFT machinery and the so-called BBSome complex (Nachury, 2018; Nachury and Mick, 2019). Advances in proximity labeling and quantitative mass spectrometry approaches have significantly advanced our understanding of the primary cilium composition (see Box 1).

In this section, we highlight the architecture, composition, and both the biochemical and dynamic properties of the primary cilium axoneme, paving the way for a subsequent discussion on how MAPs regulate the axoneme.

The basic architecture of the axoneme is composed of parallel microtubules arranged in ninefold radial symmetry and is remarkably conserved across evolution. The plus-ends of the microtubules, exposing β-tubulin, are positioned at the ciliary tip, whereas the minus-ends are anchored to the basal body. The axoneme architecture is best understood in motile cilia, where nine double microtubules are radially arranged around a central pair of singlet microtubules (called the ‘9+2’ architecture) (Fig. 1). Microtubule doublets are formed by a complete A-tubule (13 protofilaments) and an incomplete B-tubule (10 protofilaments). The axoneme of a motile cilium contains specialized motility structures, including dynein arms, nexin links and radial spokes (Klena and Pigino, 2022; Satir et al., 2014). Additionally, its lumen is periodically decorated by protein complexes known as microtubule inner proteins (MIPs), which stabilize the microtubule doublets and prevent their breakage (Gui et al., 2021; Ishikawa and Marshall, 2017; Klena and Pigino, 2022; Owa et al., 2019). Recent structural characterization of different types of motile cilia axonemes has begun to elucidate the structure, position and higher organization of a growing list of MIPs.

The primary cilium consists of a radial array of nine doublet microtubules (called the ‘9+0’ architecture) and lacks the structural components required for motility (Fig. 1). Serial section electron tomography or cryo-electron microscopy (cryo-EM) combined with sub-tomogram averaging has revealed unprecedented insights into its 3D architecture in kidney epithelial cells. These studies showed that the architecture of the kidney primary cilium differs significantly from the classic 9+0 model and from the motile cilia axoneme (Kiesel et al., 2020; Klena and Pigino, 2022; Sun et al., 2019). The primary cilium axoneme lacks a defined geometry and contains twisted, loosely organized microtubule bundles, whereas an 9+0 arrangement of the microtubule doublets is present solely at the base of the cilium. Moving towards the tip, the microtubule number in these doublets decreases, with some microtubules collapsing inward, reducing the diameter of the cilium. As a result, the organization, quantity and length of axonemal microtubules vary throughout the distal three-quarters of the primary cilium. Notably, such deviations from the 9+0 model were previously described for the cilia of the kidney, brain, retina and olfactory system before the advent of cryo-EM technology (Allen, 1965; Dahl, 1963; Gluenz et al., 2010; Moran et al., 1982).

The primary cilium lacks the periodical decoration of axonemal microtubules by MIPs and large axonemal complexes, leading to questions about how the stabilization of its axoneme is achieved. Strikingly, cryo-ET studies have identified fibrous protein networks that crosslink axonemal microtubules and also link them to the ciliary membrane, as well as proteins bound to the luminal surfaces of microtubule singlets (Kiesel et al., 2020; Klena and Pigino, 2022; Sun et al., 2019). Although the identities of many proteins within these densities remain unknown, interactions of the singlet microtubules in the axoneme with actin filaments and with the end-binding 1 (EB1; also known as MAPRE1)-like protein have been reported. A comprehensive identification of the proteins within axonemal densities will further our understanding of the mechanisms governing axoneme stability and length regulation.

The biochemical and dynamic properties of axonemal microtubules differ from non-axonemal ones, even though both are built from α- and β-tubulin heterodimers. Such differences support the stability and mechanical strength of the cilium, as required for its sensory and motile roles. Whereas non-axonemal microtubules exhibit dynamic instability, stochastically switching between growth and shrinkage phases, this instability is suppressed in axonemal microtubules to maintain their assembly and disassembly in an equilibrium, which is critical for maintaining cilium length (Avasthi and Marshall, 2013; Hao et al., 2011; Hendel et al., 2018). Notably, the length of singlet microtubules in the distal segment has been shown to alter in response to external stimuli (Marshall and Rosenbaum, 2001; Mukhopadhyay et al., 2008; Rich and Clark, 2012).

Axonemal microtubules are remarkably stable against depolymerization induced by various mechanical and chemical treatments, including colchicine (Tilney and Gibbins, 1968). Two seminal studies provided different explanations for the reasons underlying this stability. Tilney and Gibbins proposed that stability depends on the associations or connections between tubules (doublets versus singlets) and the number of associated proteins (Tilney and Gibbins, 1968). Conversely, Behnke and Forer posited that inherent differences in tubulin account for this stability (Behnke and Forer, 1967). A recent study that compared in vitro dynamic and biophysical properties of the microtubules assembled from Chlamydomonas axonemal tubulins and bovine brain tubulin supported the latter mechanism (Orbach and Howard, 2019). Consistent with their highly stable nature, these authors find that growing tips of axonemal microtubules exhibit long curved protofilaments in vitro. Furthermore, axonemal microtubules have a low average dynamicity, manifesting as a slower growth rate and decreased catastrophe frequency. However, they still undergo dynamic instability and display fast growth phases. Although this study identified low dynamicity and higher stability of microtubules as two factors that contribute to axoneme length stability, it did not rule out the contributions of different tubulin isoforms, post-translational modifications (PTMs) and axoneme-associated proteins (Orbach and Howard, 2019). Of note, axonemal microtubules are highly modified by various PTMs. While B-tubules are enriched in glutamylation, glycylation and detyrosination, acetylation occurs on both A- and B-tubule of the microtubule doublets (Wloga et al., 2017). These PTMs either directly alter the biophysical and dynamic properties of axonemal microtubules or modulate their interactions with other proteins. For an in-depth discussion of the PTMs and enzymes that catalyze their addition, we refer readers to comprehensive reviews (Guichard et al., 2023; Janke and Magiera, 2020; Wloga et al., 2017).

MAPs bind directly to microtubules and play essential roles in their assembly into a wide variety of structures, from the highly dynamic mitotic spindle to the stable ciliary axoneme. Based on their regulation of microtubules, they are classified into (1) motor proteins that generate forces and facilitate movement, (2) enzymes that depolymerize, destabilize or break microtubules, (3) microtubule nucleators, (4) end-binding proteins, (5) stabilizers, and (6) cross-linkers (Fig. 2) (Bodakuntla et al., 2019; Conkar and Firat-Karalar, 2020; Goodson and Jonasson, 2018). Many MAPs that control interphase microtubule organization, mitotic and central spindles and motile cilia have been extensively studied. However, the MAPs that localize and function at the primary cilium remain less defined. Over the past decade, the biochemical and functional characterization of known ciliary MAPs has provided new insights into the mechanisms governing the nucleation, elongation and maintenance of axonemal microtubules from the basal body. Most ciliary MAP studies have centered on motor proteins involved in IFT-mediated bidirectional transport of cargoes. For in-depth discussions on IFT motors, we refer readers to recent comprehensive reviews (Ishikawa and Marshall, 2017; Morthorst et al., 2018; Prevo et al., 2017; Webb et al., 2020). Here, we focus on non-IFT ciliary MAPs and their roles at the primary cilium, especially those regulating the initiation of axonemal assembly, its elongation and the maintenance of length and stability (Fig. 3).

The kinesin super family 13 contains the best-characterized group of microtubule depolymerizers known for regulating cilium biogenesis and length. Kinesin-13 family members are known as catastrophe factors because they catalytically depolymerize microtubules by inducing conformational changes that accelerate tubulin removal from microtubule ends hundredfold (Fig. 3C) (Asenjo et al., 2013; Desai et al., 1999; Hunter et al., 2003; Moores et al., 2002). Although kinesin-13 members depolymerize microtubules from both ends in vitro, their action on microtubule ends depends on their cellular localization and interaction partners (Desai et al., 1999; Welburn and Cheeseman, 2012). Recent atomic force microscopy studies, with single-protofilament resolution, have revealed that Kif2C (also known as MCAK) depolymerizes isolated axonemal microtubule doublets from both ends in vitro, with a faster rate for B tubule depolymerization than for the A tubule (Wijeratne et al., 2022). In organisms, such as Leishmania, Chlamydomonas and Tetrahymena, the sole kinesin-13 homolog localizes to the flagella/cilia and depolymerizes axonemal microtubules from plus- and minus-ends during flagellar assembly and disassembly (Blaineau et al., 2007; Piao et al., 2009; Vasudevan et al., 2015). In contrast to what is seen for motile cilia, mammalian kinesin-13 family members, including Kif2A, Kif2B, Kif2C and Kif24, localize to centrioles, but not primary cilia; however, they play important roles during primary cilium disassembly (Kobayashi et al., 2011; Miyamoto et al., 2015). It has been suggested that Kif2A depolymerizes cytoplasmic microtubules from the mother centriole, facilitating cilium disassembly upon cell cycle re-entry (Miyamoto et al., 2015). Consistent with this, aberrant turnover of Kif2A at the basal body leads to defective primary cilium growth in quiescent cells (Vyslouzil et al., 2023 preprint). Kif24 also plays a role in cilium disassembly during cell cycle re-entry, with its microtubule-depolymerizing activity targeting centriolar microtubules (Kim et al., 2015; Kobayashi et al., 2011). Moreover, Kif24 prevents untimely cilium assembly in dividing cells by recruiting the CP110–Cep97 centriole capping complex (Kobayashi et al., 2011). Notably, in vitro binding assays show that Kif24 has a much lower depolymerizing capacity compared to Kif2A, Kif2B and Kif2C (Kobayashi et al., 2011). The interplay between Kif24 and Kif2A during cilium disassembly and how they specifically regulate different microtubules remains to be investigated.

Kif7, a member of the kinesin-4 family, localizes to the ciliary tip where it regulates axoneme length during ciliogenesis and facilitates the recruitment and regulation of the Gli–Sufu complex during Hh signaling (Cheung et al., 2009; Endoh-Yamagami et al., 2009; He et al., 2014; Liem et al., 2009). By binding to the plus-ends of microtubules, Kif7 slows growth rates and increases catastrophe frequencies of microtubules; this explains the longer cilia observed in Kif7-depleted cells (He et al., 2014). Initially, it was proposed that the ciliary functions of Kif7 depended on its enrichment at the ciliary tip and that this enrichment was due to its preferential binding to the GTP caps of growing microtubule ends in vitro (Jiang et al., 2019). However, recent studies utilizing Kif7-truncation mutants in cells and total internal reflection fluorescence microscopy experiments in vitro have suggested that binding of Kif7 to microtubules is not necessary for its ciliary tip localization (Blasius et al., 2021; Yue et al., 2022). This discrepancy might arise from variations in Kif7 activity in vitro compared to in cells. Alternatively, modulation of the phosphorylation status of Kif7 might influence its ciliary recruitment (Chen and Jiang, 2013; Liu et al., 2014).

Microtubule-severing enzymes, or severases, are AAA-ATPases that cut microtubules into shorter filaments, thereby regulating their disassembly, amplification and length (Kuo and Howard, 2021). The best-studied severase that acts on axonemal microtubules is katanin (Fig. 3F) (Kuo and Howard, 2021; McNally and Vale, 1993). In vitro studies have shown that katanin forms oligomeric complexes on the microtubule lattice, stimulating its ATPase activity and promoting microtubule disassembly (Hartman and Vale, 1999). Notably, katanin has a higher affinity towards microtubules that are enriched with PTMs, such as axonemal microtubules (Sharma et al., 2007; Sudo and Baas, 2010). The catalytic ATPase subunit A (p60; also known as KATNA1 in mammals) and regulatory subunit B (p80; also known as KATNB1 in mammals) of katanin have important roles in assembly, maintenance, disassembly and signaling functions of cilia (Lynn et al., 2021). Specifically, the p60 subunit plays a key role in detaching ciliary axonemes from basal bodies, thereby facilitating the disassembly of both primary and motile cilia (Mirvis et al., 2019; Parker et al., 2010; Rasi et al., 2009). Notably, the Chlamydomonas pf19 mutant, which has a point mutation in p60 that results in a lack of microtubule-severing activity, does not exhibit a stress-induced deflagellation defect. This highlights the multifaceted ciliary roles of katanin beyond just disassembly (Dymek and Smith, 2012). In line with this, the p60 subunit is implicated in the assembly and maintenance of motile cilia, as well as in the formation of their central pair of microtubules (Banks et al., 2018; Liu et al., 2021; Ververis et al., 2016; Willsey et al., 2018). In contrast, the p80 subunit has a distinct role in human cortical development, where it controls the number of centrioles and cilia and the activation of the Hh pathway (Hu et al., 2014). Despite its link to multiple ciliary processes, several aspects of katanin function are still unclear, including its requirement for deciliation and its exact role during deciliation, as well as its potential indirect effects on cilia by severing cytoplasmic microtubules, which could modulate the pool of cytoplasmic tubulin or interfere with ciliary protein transport.

Spastin is another microtubule-severing enzyme that controls ciliary length and Hh signaling by regulating ciliary microtubule dynamics in neural stem cells (Jeong et al., 2019). The mechanism by which spastin catalytically severs microtubules is similar to that of katanin. Although spastin and katanin have similar diffusion rates on microtubules, their binding affinities differ based on the nucleotide state of their catalytic subunits (Eckert et al., 2012). Notably, pathogenic variants of spastin are linked to hereditary spastic paraplegia, a neurological disorder resulting from the dysfunction of cortical motor neurons (Blackstone et al., 2011). Given that the ciliary functions of spastin contribute to neural development, it will be important to further investigate how spastin and katanin act on axonemal microtubules and regulate neuronal cilia (Plaud et al., 2017; Sherwood et al., 2004).

Polymerizers promote the active addition of tubulin to the ends of growing microtubule protofilaments. These proteins modulate microtubule dynamics by directly binding to tubulin via their tumor overexpressed gene (TOG) domains (Farmer and Zanic, 2021). TOG domains bind to tubulin heterodimers using an oblong-shaped solenoid-like structure that is formed by six HEAT repeats (Al-Bassam et al., 2007; Slep and Vale, 2007). Variations in the number, architecture and tubulin-binding activity of TOG domains across family members lead to unique regulatory effects on microtubule dynamics (Cook et al., 2019; Farmer and Zanic, 2021; Leano et al., 2013). XMAP215 (also known as ch-TOG and CKAP5) uses a pentameric TOG-domain array to act as a processive polymerase, catalyzing the addition of up to 25 tubulin dimers to the growing plus ends (Brouhard et al., 2008). This process can accelerate microtubule growth rates up to 10-fold, as observed in single-molecule assays (Brouhard et al., 2008). In contrast, CLASP family members utilize a trimeric TOG-domain array to stabilize dynamic microtubules by both suppressing catastrophes and promoting rescues (Lawrence et al., 2020).

The search for TOG-containing proteins that function at the cilia led to the identification of Crescerin (TOGARAM1 in vertebrates; SHF1 in Chlamydomonas), which has four TOG domains (Fig. 3B) (Das et al., 2015). In vitro studies have shown that the TOG2 and TOG4 domains of Crescerin have notable microtubule-polymerizing and possible nucleation activities, whereas the TOG1 and TOG3 domains do not (Das et al., 2015). These findings are consistent with the localization of Crescerin orthologs to the axonemes of cilia or flagella in Caenorhabditis elegans, Chlamydomonas and vertebrates, where they modulate their length and architecture (Bacaj et al., 2008; Das et al., 2015; Perlaza et al., 2022). In C. elegans, Crescerin localizes to amphid and phasmid sensory cilia, and the loss-of-function mutant has cilia that are shorter, have disorganized bundles at their proximal ends and lack discernible microtubules at their distal ends (Bacaj et al., 2008; Das et al., 2015). Furthermore, pathogenic variants in TOGARAM1 have been associated with shorter primary cilia in individuals with a neurodevelopmental ciliopathy termed Joubert syndrome (JBTS) (Latour et al., 2020; Morbidoni et al., 2021). Besides TOGARAM1, mutations in other ciliary MAPs, form part of a disease module as they have also been linked to JBTS, underscoring the role of axonemal defects in JBTS (Latour et al., 2020). TOGARAM1 cooperates with another MAP, ARMC9, to regulate cilium length, stability and disassembly (Latour et al., 2020). Intriguingly, recent characterization of a Chlamydomonas SHF1 loss-of-function mutant challenged the proposed microtubule polymerase role of Crescerin at the cilium (Perlaza et al., 2022). Through kinetic and computational analysis of the cilium assembly defects in these SHF1 mutants an alternative hypothesis emerged – that SHF1 might provide a high-affinity binding site for tubulin within the IFT complex, thereby controlling the transport of tubulin dimers from the cytoplasm to the cilium (Perlaza et al., 2022). Collectively, these findings define microtubule polymerization and tubulin transport as the two mechanisms, by which Crescerin regulates axonemal microtubule dynamics.

Another TOG domain-containing ciliary protein is Cep104, which contains only a single TOG domain. It localizes to the tips of the cilium or flagella in mammalian cells, Chlamydomonas and Tetrahymena, regulating their length and architecture (Frikstad et al., 2019; Louka et al., 2018; Satish Tammana et al., 2013). Structural studies and in vitro assays have revealed that the Cep104 TOG domain binds directly to free tubulin but only exhibits a mild microtubule-polymerizing activity (Al-Jassar et al., 2017; Rezabkova et al., 2016; Widlund et al., 2011). Nevertheless, Cep104 forms homodimers in solution through its coiled-coil domain and interacts directly with the microtubule plus end-tracking protein EB1 via its SXIP motif (Rezabkova et al., 2016). Consequently, the two TOG domains in the dimer and EB1 might synergize to enhance the microtubule-polymerizing activity of Cep104 in cells, as is the case for the XMAP125–EB1 interaction (Zanic et al., 2013). In phenotypic rescue experiments, a Cep104 TOG domain mutant failed to compensate for the shorter primary cilia phenotype observed in Cep104-depleted mammalian cells (Yamazoe et al., 2020). This finding suggests that Cep104 might contribute to axoneme length regulation either by using the potential microtubule-polymerizing activity of its TOG domain or by transporting tubulin dimers to the cilia for polymerization of axonemal microtubules.

In addition to tubulin and EB1, Cep104 interacts with the centriole capping complex component CEP97 and the MAP CSPP1 via its so-called jelly-roll fold domain, as well as with the kinase Nek1 and CP110 via its zinc-finger region (Al-Jassar et al., 2017; Frikstad et al., 2019; Jiang et al., 2012; Rezabkova et al., 2016; Yamazoe et al., 2020). The interactions between Cep104 and these ciliogenesis factors suggest its potential involvement in multiple stages of ciliogenesis. A recent study in mammalian cells has demonstrated that depleting Cep104 does not affect the removal of CP110 from the mother centriole, implying that it has a role downstream of initiation of axoneme assembly (Yamazoe et al., 2020). Supporting this, Cep104 colocalizes with CSPP1 at the ciliary tip, and the two proteins cooperate during axoneme elongation (Frikstad et al., 2019). In Tetrahymena, orthologs of Cep104 along with the ciliary MAPs Crescerin and ARMC9 determine the geometry of the distal segment by controlling the positions of specific microtubule ends (Louka et al., 2018). Specifically, this study suggests that Cep104 promotes A-tubule elongation and connects the central-pair microtubules to the membranes at the distal end, whereas Crescerin acts as a positive regulator of B-tubule length (Louka et al., 2018). Single-molecule reconstitution experiments with full-length TOGARAM1, Cep104 and other ciliary MAPs will clarify how they modulate axonemal microtubule dynamics, either individually or collectively.

MAPs stabilize microtubules by shifting the equilibrium from soluble to polymerized tubulin by decreasing or inhibiting dynamics of polymerized microtubules. CSAP, SAXO-1, ENKD1, MAP4, CSPP1, SSNA1 and CCDC66 are ciliary MAPs described for their ability to cross-link microtubules into arrays (bundling) and/or stabilization, either in cells or in vitro (Fig. 3E). Although loss-of-function and gain-of-function studies have identified these MAPs as regulators of cilium biogenesis and function, the exact mechanisms by which they modulate axonemal microtubules remain unclear for most of them, with CSPP1 being a notable exception. In this section, we will first briefly introduce these MAPs and then discuss CSPP1 in more detail.

SAXO-1 is a MAP that is related to MAP6 and localizes to the ciliary axoneme, promoting its elongation (Dacheux et al., 2015). ENKD1 also acts as a positive regulator of ciliary length and Hh signaling (Tiryaki et al., 2022). Besides binding and stabilizing microtubules, ENKD1 competes with CEP97 for binding to CP110, facilitating axonemal assembly (Song et al., 2022). Conversely, MAP4, a member of the Tau protein family, is a negative regulator of cilium length; it localizes to axonemes and preferentially restricts growth in the ciliary distal segment relative to that in the proximal, highly polyglutamylated region (Ghossoub et al., 2013; Kanamaru et al., 2022). Additionally, MAP4 disrupts the recruitment of the SEPT2–SEPT7–SEPT9 septin complex to the axoneme, thereby opposing the role of septins in promoting ciliogenesis and axoneme length (Ghossoub et al., 2013).

CSPP1 was identified as the first luminal MAP of the primary cilium using cryo-ET analysis in vitro and nano-meter scale 3D microscopy in cells (van den Berg et al., 2023). It localizes to the basal bodies, axoneme, ciliary tip and centriolar satellites in quiescent cells and is required for axoneme elongation and Hh signaling (Box 2; Fig. 3) (Frikstad et al., 2019). In vitro reconstitution experiments have demonstrated that CSPP1 stabilizes microtubule plus-ends from the luminal side, inhibiting microtubule growth and shrinkage, and promoting pausing in vitro (van den Berg et al., 2023). CSPP1 also recognizes and stabilizes damaged microtubule lattices (van den Berg et al., 2023). Detailed cryo-ET analysis confirmed that CSPP1 binds only to the inner surface of the microtubules, excluding tapered and curved microtubule ends where Cep104 and TOGARAM1 reside. Consequently, it remains elusive how CSPP1 can interact with these TOG proteins. Thus far, it has not been possible to detect whether CSPP1 molecules protrude from the microtubule lumen using tomography. An alternative interaction site might be the microtubule lattice rather than the tip. Laser ablation experiments have shown that CSPP1 recognizes and stabilizes damaged microtubule lattices (Fig. 3G) (van den Berg et al., 2023). Such damage could occur in axonemes when kinesins walk, which can randomly pull tubulin subunits from the lattice, similar to the pulling by spastin (Kuo et al., 2022). Importantly, the biochemical activities of CSPP1 mirror those of microtubule-stabilizing compounds such as taxanes, which also bind to the luminal side of microtubules and stabilize them (Steinmetz and Prota, 2018). This parallel offers the potential for pharmacological intervention in ciliopathies.

CCDC66 is another MAP that localizes to centriolar satellites, the basal body, axoneme and ciliary tip, and it interacts with CSPP1, Cep104, ARMC9 and TOGARAM1 as part of a JBTS-linked module (Box 2, Fig. 3). CCDC66 is important for cilium assembly, regulation of its content and activation of the Hh pathway (Conkar et al., 2017; Odabasi et al., 2023). Overexpression of Cep104 has been shown to compensate for the shortened cilia in CCDC66-depleted cells, indicating that they might work together during axonemal elongation (Odabasi et al., 2023). Notably, cells depleted of CCDC66 exhibit significant fluctuations in ciliary length, pointing to its role in axoneme stability. In vitro studies suggest that CCDC66 stabilizes microtubules through its bundling activity (Batman et al., 2022; Jujimon et al., 2022). Bundled microtubules are considered to resist dilution- and cold-induced disassembly, as observed for kinetochore microtubules and the central spindle (Brandt and Lee, 1994; Walczak and Shaw, 2010). Similarly, axonemal microtubules of the primary cilium crosslink (bundle) with fibrous protein networks, potentially providing axonemal singlets with stability and contributing to ciliary elasticity (Sun et al., 2019). Further in vitro single-molecule assays are required to elucidate whether CCDC66 directly affects microtubule dynamics, has a stabilizing effect similar to CSPP1, or requires cooperation with Cep104 to modulate microtubule dynamics.

In summary, this section has explored the diverse roles of MAPs within the ciliary axoneme, discussing their biochemical and cellular functions. Understanding the mechanistic underpinnings of axonemal architecture and stability is crucial because disruptions in these properties are directly associated with human diseases. Our next section will expand on this by focusing on the relationship between these MAPs and ciliopathies.

Mutations in genes encoding ciliary MAPs have been linked to various ciliopathies, which arise from the dysfunction of primary cilium and/or motile cilia. Ciliopathies comprise a spectrum of severity and include around 35 distinct disorders, with common manifestations including polycystic kidney disease, retinal degeneration, obesity, skeletal abnormalities and brain anomalies (Reiter and Leroux, 2017). Our discussion is centered on a recently described MAP module that includes JBTS-associated proteins Cep104, CSPP1, TOGARAM1, ARMC9 and CCDC66 (Box 2, Fig. 3) (Latour et al., 2020). For context, JBTS is a neurodevelopmental ciliopathy mainly characterized by hypotonia, ataxia, abnormal eye movements and cognitive impairments. Some of the rarer symptoms associated with JBTS are retinal dystrophy, cystic kidneys, liver fibrosis and even sinusitis (Gana et al., 2022). Whereas the transition zone and the Hh signaling pathway were previously thought to be primary contributors to JBTS, recent research has highlighted the importance of axonemal length and organization, and thus, the role of MAPs in the disease. To elucidate this connection, interactions among the JBTS-module MAP proteins were defined using yeast two-hybrid, immunoprecipitation, tandem affinity purification and/or proximity labeling screens (Al-Jassar et al., 2017; Breslow et al., 2018; Conkar et al., 2019; Frikstad et al., 2019; Latour et al., 2020; Odabasi et al., 2023). Functional studies identified the roles of these MAPs in diverse ciliary processes ranging from regulating axonemal microtubule stability, length and PTMs to influencing primary cilium biogenesis and its signaling function (Fig. 3). In Tetrahymena, the Cep104, Armc9 and TOGARAM1 orthologs localize to the motile cilia tip, where they regulate the length and stability of microtubule ends (Louka et al., 2018). This function has not yet been investigated for their mammalian orthologs. Thus far, it is only known that CSPP1 localizes to the tip of motile cilia in mammalian multiciliated epithelia (Frikstad et al., 2019). Hence, the role of these MAPs in mammalian motile cilia remains unexplored, and better understanding their function could provide insights into how these MAPs distinctively or commonly contribute to both sensory and motile ciliopathies. Beyond cilia, ciliary MAPs are found in non-ciliary cellular locations, where they are involved in cell division by regulating mitotic and central spindle microtubules, and in cell adhesion by organizing cell–cell adhesion sites (Box 2, Fig. 3). Given their ubiquitous expression and affinity for microtubules, future research might reveal other potential microtubule-dependent functions for these proteins. Importantly, such roles might underpin the non-ciliary defects seen in associated ciliopathies, for example cystogenesis arising from defective cell division plane or defective cell-to-cell adhesion. This underscores their importance for cataloging of the genetic and phenotypic causes of JBTS and other ciliopathies. Future studies should clarify the connection between the myriad of cellular defects resulting from ciliopathy gene dysfunction and disease phenotypes.

Here, we reviewed the architecture, composition and biogenesis of the primary cilium axoneme, as well as the various MAPs within the cilium, highlighting their mode of action and roles in the assembly, maintenance and disassembly of the axoneme. The characterization of these ciliary MAPs has provided important insights into the regulation of axonemal length and stability, and might pave the way for identifying therapeutic targets for associated ciliopathies. Furthermore, identifying the complete set of ciliary MAPs that appear as densities in the structural studies of the primary cilium, could provide deeper insight into the structural and functional complexities of the axoneme. Outstanding questions also include how and when MAPs are recruited to the axoneme, as the molecular pathways that govern their ciliary localization and interactions remain unknown, even for the known MAPs discussed here. As has been proven fruitful for motile cilia, employing advanced structural and imaging approaches combined with in vitro reconstitution experiments and cellular and organismal models will enable a more rigorous dissection of the primary cilium axoneme in health and in disease.

Moreover, the diversity in cellular localization, interactions and functions of ciliary MAPs requires the development of new tools and approaches to explore their role in the pathogenesis of ciliopathies and potential therapeutic strategies. The current methods are insufficient to distinguish the relative contributions of ciliary and non-ciliary defects to disease. In this context, spatial and temporal manipulation of MAP localization and activity, complemented by a systems biology approach, will be necessary to pinpoint ciliary and non-ciliary binding partners and correlate localization of MAPs with their mechanisms and functions. Besides the diversity of MAPs, the structure and function of cilia also vary considerably depending on the cellular context. The adaptations of the axoneme and its MAPs to meet the specific needs of different cell types and tissues remain unexplored, impeding progress in deciphering the basis of phenotypic heterogeneity of ciliopathies. Consequently, future research aimed at uncovering the roles of ciliary MAPs across diverse cell and tissue types will be crucial for advancing efforts to combat ciliopathies.

We acknowledge Seyma Cengiz-Emek for insightful discussions regarding this work.