1 Cilia: An introduction

1.1 Cilium as an organelle

Cilia are thin, hair-like cell membrane extensions with a microtubule-based core (axoneme), specialized in structure and functions such as sensory perception and motility. Flagella, whip-like projections of the cell but longer than cilia, were considered different organelles until their internal structure was found to be identical to cilia. Ciliary/flagellar motility is the movement of cilia/flagella such that cells that harbour them can swim from one point to another or can ‘sweep’ fluid/particles around them. This motility is driven by axonemal dyneins situated on the tubulin-based structure. Intraciliary/intraflagellar motility is the bidirectional transport of particles along axonemal microtubules. Recently, the prevalence, utility, and importance of the cilium are being thoroughly investigated, thereby unravelling its genesis, architecture, composition as well as biological functions. Found widely in species ranging from protozoa to mammals (figure 1; Mitchison and Valente 2017), cilia display significantly varying extents of motility (figure 1). In multicellular organisms, cilia give rise to a diverse spectrum of functionality, specifically those including different tissue types and organ systems. Aberration in ciliary proteins results in a dysfunctional organelle, thereby leading to a series of ailments collectively known as ciliopathies that afflict the brain, heart, kidneys, respiratory tract, reproductive system, etc. Studies on ciliary composition and structure encompassed the use of both simple organisms such as Tetrahymena, Drosophila melanogaster, Caenorhabditis elegans, Chlamydomonas reinhardtii, and sperm of multicellular organisms such as rat and sea urchin (Afzelius 1959; Gibbons 1961). The use of complex model organisms, like Xenopus and zebrafish, has highlighted the subtle differences in the conserved structures of this organelle.

Figure 1
figure 1

Types of cilia in unicellular and multicellular organisms, their respective tissue types, and their functions. Note the axonemal structures for each type and the ciliary diseases that the respective cilia types cause when proteins that make these organelles are mutated leading to dysfunctional cilia.

1.2 Ciliary structure

All cilia are covered by the plasma membrane (ciliary membrane) and have an axoneme at their core that extends from the centriole-derived basal bodies (figure 2). The arrangement of the microtubules and associated proteins is quite heterogeneous throughout the axoneme (Mohri et al. 2012). These differences attribute stability yet flexibility to the organelle. Out of the three (A, B, and C) tubules present in a centriole (discussed below), the C-tubule terminates within a structure known as the transition zone (TZ) that lies between the basal body and the axonemal shaft. TZ has Y-shaped linkers between the ciliary membrane and the microtubules, and a structure called the ciliary necklace, one that is conspicuously absent from the sperm flagellum (Reiter et al. 2012; reviewed by Gonçalves and Pelletier 2017). It seems to be the selective gate preventing the free entry of soluble proteins (Takao and Verhey 2016). Using a combination of cryo-electron tomography and expansion microscopy techniques, Hoek et al. (2022) have shown the molecular architecture of the ciliary base in the green alga Chlamydomonas reinhardtii. Using cryo-ET and subtomogram averaging, the group has demonstrated that membrane binding is not a prerequisite for Y-link formation. Distal to the axonemal structure, a previously unseen helical ‘sleeve’ was observed decorating the microtubules, which the authors predict might designate sites of axoneme severing.

Figure 2
figure 2

Overall structure of the 9+2 cilium and its cross-section. The cilium is made up of the axoneme that is formed by nine microtubule doublets surrounded by the ciliary membrane. The other substructures (see the adjacent illustration of the cross-section) include the central pair, radial spokes, and nexin–dynein regulatory complex, typical of the 9+2 axoneme. Kinesin present along the axoneme moves the IFT complex and its cargo towards the ciliary tip (anterograde transport), whereas dynein moves the IFT complex and its cargo towards the cell body (retrograde transport). The transition zone present at the base of the cilium provides selective entry of proteins into the axoneme. Y-shaped linkers are occupied in the transition zone and connect the axoneme to the base of the ciliary membrane. The cilium is anchored on to the basal body (in gray). The ciliary pocket programs ciliary-associated vesicular trafficking. Note the A, B (axonemal doublets), and C (centriolar) microtubules.

Each of the A-tubules in the outer doublet is connected to nine lollipop-like structures called the radial spokes that project toward a structure called the central pair of singlet microtubule apparatus and its proteinaceous projections, which together are believed to regulate ciliary beating (figure 2; Satir et al. 2014; Teves et al. 2016). An excellent cryo-EM structure of the active central pair has recently been solved (Han et al. 2022). Cilia are classified based on the presence of this central pair (designated 9 + 2) and its absence (designated 9 + 0). Both 9+2- and 9+0-containing cilia are either motile or immotile, depending on the structural differences present in the axoneme (Satir and Christensen 2007). Extensive work on this ciliary architecture has revealed that the arrangement of the sub-ciliary structures like central pair, radial spokes, and dynein arms may be responsible for the observed differences among the types of motilities (figures 1 and 2; Teves et al. 2016; Kempeneers and Chilvers 2018; Kempeneers et al. 2019). This makes it important to characterize cilia in several organisms so that the scope of cross-species differences is evident.

Based on the function and axonemal format, the human body can generate four types of cilia: 9+0 motile, 9+0 non-motile, 9+2 motile, and 9+2 non-motile (figure 1). Of the four axonemal formats, the 9+2 motile cilia in vertebrates are present in the respiratory airways, the reproductive system, and the ependymal cells, whose ciliary motility renders fluidic movement (figure 1). Also known as primary cilia, 9+0 non-motile cilia are usually solitary and are deprived of the dynein arms, radial spokes, and central pair. These cilia are present on most differentiated cells in humans relying on the solitary cilium to monitor the environment. Recent studies on the sensory role of primary cilia have revealed their importance in development and tissue homeostasis. It has been found that some tissue-specific cell types also harbour 9+2 non-motile cilia lacking dynein arms; for example, olfactory neurons with multiple long chemoreceptive cilia, the vestibular and cochlear mechano-sensory cilia present in the inner ear of the embryo (figure 1; Satir and Christensen 2007; Kempeneers and Chilvers 2018; Kempeneers et al. 2019). Primary cilia present on sensory cells of mammalian olfactory and vision systems transduce extracellular signals that are chemical, physical (smell and visual), or electrical in nature (Falk et al. 2015). In particular, the specialized kinocilia present in the auditory hair cells of the inner ear use their sensory feature for responding to sound. Unlike non-motile cilia, the 9+2 motile cilia are usually (although not always) present as clusters (figure 1). The 9+0 motile nodal monocilia found at the embryonic node are solitary and show unique rotatory motion. They function in determining the left/right (L/R) body symmetry during early development by generating the leftward fluid flow (nodal flow) followed by the differential patterning of lateral plate mesoderm and then asymmetric organogenesis (McGrath et al. 2003; Tabin and Vogan 2003). The clockwise movement of this cilium present at the centre of the ventral node is also known as the left–right organizer (LRO). It acts to break the L/R symmetry in organisms such as zebrafish, Xenopus, and mouse (Nonaka et al. 1998; Essner et al. 2019; Schweickert et al. 2007). This cilium is different from the airway cilia that show planar beating. But, it is the 9+0 immotile sensory cilium present at the periphery of the ventral node that can sense the flow-dependent signal coming from calcium (Pennekamp et al. 2002; Field et al. 2011; Yoshiba et al. 2012). FOXJ1, a master regulator of the multiciliogenesis program (Yu et al. 2008), and NOTO, a homeodomain transcription factor expressed in the ventral node of the mouse embryo, play a crucial role in determining L/R patterning (Abdelkhalek et al. 2004; Didon et al. 2013). Microarray analysis of the NOTO targets identified about 716 candidate genes of which 59 overlap with lung candidate genes (Stauber et al. 2017). Mutational and rescue studies in mouse embryos showed that both function together in the localization of cilia at the posterior pole. NOTO acts upstream of FOXJ1, and mutation in FOXJ1 causes randomization of left/right body asymmetry (Beckers et al. 2007; Alten et al. 2012; Wallmeier et al. 2019). However, studies in urochordate and sea urchin embryos show the involvement of the 9+2 motile cilia in establishing L/R patterning (Nishide et al. 2012; Tisler et al. 2016). Further, the molecular mechanism of motility is central to understanding physiological processes carried out by the respiratory airways for clearing mucus, fertilization rendered by sperm, movement of the ovum in the fallopian tubes, and the directional flow of cerebrospinal fluid (CSF) generated by the ciliated ependymal cells (Banizs et al. 2005). In contrast, the 9+2 motile sperm flagellum and other multiciliated epithelial cells exhibit sigmoidal, i.e. symmetric waveform and planar waveform, respectively (figures 1 and 2; Hirokawa et al. 2006). Taken together, a detailed understanding of cilia will provide the foundation  in determining the cause and effect of ciliopathies.

1.3 Ciliary base

The ciliary base comprises two major components: the basal body, which nucleates axonemal microtubules, and the periciliary membrane, which serves as a diffusion barrier and separates the ciliary membrane from the plasma membrane. Formation of the basal body and the ciliary membrane are the two key steps towards the initiation of ciliogenesis. The ciliary base is also an important site for sorting the cargo destined for the ciliary compartment. All proteins required for ciliary growth are first transported to the ciliary base, sorted further, and then sent to different destinations of the axoneme (Sorokin 1962; Rieder et al. 1979).

2 Ciliogenesis: An introduction

The birth of a cilium is known as ciliogenesis. Even though most information on ciliogenesis was acquired using vast studies on Chlamydomonas flagella, the focus has recently shifted to mammalian models such as mice and zebrafish (Bangs and Anderson 2017; Fu et al. 2019; Haider et al. 2019). Most of this information comes from the identification of genes whose mutations have resulted in the disruption of ciliary functions including birth and assembly; some of these have been associated with human ciliopathies (Waters and Beales 2011). Genetic screens generated in Chlamydomonas, zebrafish, and Drosophila have helped enormously in identifying molecules whose gene mutations result in the disruption of the major steps of ciliogenesis (Satir and Christensen 2007). The primary cilium, being solitary, only needs a single basal body as the nucleation site, as opposed to multicilia, where the centriole needs to multiply several times to generate enough basal bodies. Studies are being conducted using primary cell cultures of ependymal brain cells, respiratory airway epithelial cells, and fallopian tube epithelial cells, right from tracing the events during the vigorous multiplication of centrioles to the formation of mature tufts of cilia. Although the complete molecular mechanism remains elusive, extensive efforts are ongoing to deduce the molecular players responsible for this transition. One important term is ciliary compartmentalization, which suggests an important role for sub-ciliary regions in regulating the structure and/or function of cilia. For the convenience of studying ciliogenesis, cilia could be viewed as consisting of three main sub-compartments: ciliary base, transition zone (TZ), and distal ciliary compartment (ciliary tip). Ciliogenesis and ciliary compartmentalization happen simultaneously. Taken together, the entire process involves the synthesis of precursors that contribute towards the making of the centriole, transportation and axoneme elongation, and a regulatory network. It is emphasized that the regulation of ciliary growth dynamics forms an important aspect of this process and is required for accurate cilia formation. Right from the development of progenitor cells and their differentiation into fully mature cilia, various pathways intersect to give rise to the developed multiciliated tissue. The following is a detailed description of the steps in the ciliogenesis process of 9+2 and 9+0 cilia.

3 Process of ciliogenesis

3.1 Centriole structure

Cilia are derived from centrioles, structures that are evolutionarily conserved and play a role in organizing microtubules and generating spindle fibres during cell division. Understandably, therefore, ciliogenesis and cell division are coupled events. A centriole is a barrel-shaped microtubular organelle composed of α- and β-tubulin heterodimers, ~150 different proteins with dimensions of about 400–500 nm length and 250 nm diameter (Andersen et al. 2003; Keller et al. 2005; Winey and O'Toole 2014). Each barrel is a cylindrical structure consisting of nine sets of microtubular triplets that display the 9-fold radial (cartwheel) symmetry (figure 3; Marshall and Rosenbaum 1999). The two centrioles (mother and daughter) are present at a perpendicular position to each other with a surrounding pericentriolar material (PCM), which is a fibrous meshwork and can bind to several proteins (figure 3; Schnackenberg and Palazzo 1999) – all this together forms the centrosome. The proximal and distal ends of each centriole perform distinct roles, the former is responsible for PCM recruitment (Woodruff et al. 2014), whereas the latter functions in the accurate docking of centrioles (figure 3; Tanos et al. 2013). Docking accurately positions both the centrioles in its cellular niche. While the mother centriole consists of distal and subdistal appendages (DA and SDA, respectively), the daughter centriole does not have any such structures (Paintrand et al. 1992). The PCM also acts as the centrosomal microtubule-organizing centre (MTOC). It can nucleate the microtubular array formed during the interphase of cell division and the spindle microtubules during mitosis (Euteneuer and McIntosh 1981). Electron microscopy studies have revealed the presence of small aggregates of electron-dense non-membrane-bound proteinaceous particles around the centrosome termed as centriolar satellites in some vertebrates (Lopes et al. 2011). These help in the recruitment of PCM material and proteins required for ciliogenesis and are especially known for the presence of the scaffolding protein, PCM1. The absence of PCM1 abrogates ciliogenesis in some cells, indicating an important role in cilium assembly (Lopes et al. 2011; Kodani et al. 2015).

Figure 3
figure 3

The structure of a centriole. Centrioles are the paired barrel-shaped structures. The procentriole is arranged at the proximal end of the mature parent centriole. At the distal end of the mature parent centriole are attached the distal and subdistal appendages. The A, B, and C microtubules are shaped like a cartwheel.

3.2 Centriole production (centriole genesis, or centriologenesis) and duplication (copy number)

Cell division requires centriole duplication that begins during the G1–S phase transition in cycling cells. This duplication is marked by the appearance of a new procentriole (a short daughter centriole) at the proximal end of each existing centriole in an orthogonal orientation. Polo-like kinase (PLK4) is known as the earliest marker that depicts the site for procentriole emergence (figure 4A; Bettencourt-Dias et al. 2005; Habedanck et al. 2005; Eckerdt et al. 2011). The recruitment of PLK4 is mediated by CEP152 and CEP192, which co-operate in promoting this process (Kim et al. 2013). During the G1–S phase transition, PLK4 further interacts with SCL/TAL1 interrupting locus (STIL, where SCL is stem cell leukaemia and TAL is T-cell acute lymphoblastic leukaemia), which initiates trans-autophosphorylation and activation of PLK4. This active PLK4 phosphorylates STIL which in turn triggers the recruitment of spindle assembly abnormal-6 (SAS-6; Arquint et al. 2015), leading to the formation of the cartwheel structure serving as a building block for the procentriole (Dzhindzhev et al. 2014). Another protein centrosomal P4.1-associated protein (CPAP), with its interacting partners CEP135, CEP120, and spindle and centriole-associated protein 1 (SPICE1), further elongates the procentriole by forming microtubules (figure 4A; Tang et al. 2009). The procentriole elongates during the S and G2 phases, finally detaching from the mother centriole in the late G2 phase. After mitosis, each cell inherits a pair consisting of the mother centriole and a newly duplicated daughter centriole. At this stage, the mother centriole acquires its unique fibrous DAs and SDAs. This basal body finally serves as a nucleation centre for a cilium.

Figure 4
figure 4

(A) Molecules important for centriole duplication and assembly in humans. The flow diagram shows the molecular pathway responsible for centriole production. The solid lines represent a physical association between the molecules and the arrow heads indicate the recruitment/regulation of the next molecule in the pathway. The pathways can be separated in three major parts. First, the PLK4-STIL association is required for the initiation of centriolar duplication. CEP192/CEP152/CEP63 is required for the recruitment of PLK4 to the parent centriole. This association results in PLK4 activation, which further phosphorylates STIL and itself, as represented by blue arrows. Second, the association with HsSAS-6 is required for the initiation of cartwheel assembly in the centriole. The black arrow indicates self-polymerization of HsSAS-6 after its activation. The third step includes the recruitment of CEP135 which further recruits a set of molecules required for centriole elongation. CPAP, in collaboration with CEP135, CEP120, and SPICE1, facilitates the formation of the procentriole’s microtubule wall. γ-tubulin helps in the microtubule stabilization. (B) Pathways for centriole migration and docking at plasma membrane (formation of basal body and cilium assembly). The mature mother centriole serves as a precursor for the basal body formation and cilium assembly. There are two main pathways known in mammals for centriole migration and/or docking at the plasma membrane: (1) In the extracellular pathway (top), the parent centriole migrates towards the plasma membrane and the axonemal elongation starts only after it has docked at the plasma membrane on the cell surface. Loss of CP110 from distal end of the parent centriole marks the initiation of cilium assembly. At similar time points, IFT particles are also recruited to the centriole for initiating the formation of the TZ. The TZ specific complexes are then localized to TZ. (2) The intracellular pathway (bottom) starts with the recruitment of pre-ciliary vesicles to the distal appendages of the mature centriole with the help of MYO5A. These are called distal appendage vesicles (DAVs), and they contain Rab-11 GTPase associated with Rabin8 that helps in the recruitment of Rab8. Protein EHD1 helps in fusion of these DAVs to form the ciliary vesicle (CV), and Rab8 is responsible for the growth of CV. The loss of CP110 happens at the same time marking the initiation of TZ formation. This results in maturation of ciliary vesicles forming the distinct domains of ciliary sheath and ciliary membrane. Axonemal extension takes place inside the ciliary sheath until the fusion of ciliary sheath with plasma membrane exposes the axoneme to the external environment.

The spatial, numerical, and temporal regulation of this process serves importantly to ensure copy number control where one mother centriole can only give rise to a single new procentriole. Studies in human somatic cells have shown that experimental depletion of existing centrioles triggers de novo centriole genesis (Uetake et al. 2007). An example of de novo centriole production is seen in terminally differentiated multiciliated cells (MCCs). These have completely different molecular dynamics as compared with monociliated cells. Being multiciliated, and to ensure multiciliogenesis, MCCs require several centrioles to be docked at the plasma membrane after exiting the cell cycle (at a point where normal cells suppress centriole duplication). It was in 1971 that Anderson and Brenner distinguished two pathways for centriologenesis in MCCs, namely, centriolar and acentriolar (major pathway) (figure 4B; Anderson and Brenner 1971). In the centriolar pathway, an existing centriole acts as a scaffold to produce multiple new procentrioles around it. The origin of the fundamental centriole that produces these new procentrioles is yet to be elucidated. In the acentriolar pathway, clusters of dense granules (intermediary structures) that may or may not be associated with the existing centriole spontaneously arise in the cytoplasm. New procentrioles are produced around these rather than the existing centriole. These dense granules further condense and give rise to novel structures called deuterosomes triggered by the expression of the Deup1 protein. Deup1 is a paralog (37% identical) of CEP63 (a human microcephaly protein) that regulates mother centriole-dependent centriole duplication (Zhao et al. 2013). Another protein, multicilin (encoded by the gene Mcidas), is largely expressed during multiciliated cell differentiation and positively regulates the expression of Deup1, Plk4, Cep152, Stil, and Sas-6 (figure 4A; Kim et al. 2018). High expression of Deup1 initiates deuterosome formation, thus making a scaffold for supporting multiple de novo centriole duplication. Deup1 binds to the CEP152 protein which in turn activates PLK4 resulting in the initiation of the process of centriolar assembly. These centrioles further mature to form basal bodies. Most of the key regulators (Plk4, Cep152, Stil, and Sas-6) are similar in both centriolar and acentriolar duplication pathways (figure 4B). Another example where the copy number of centrioles deviates is seen in the sexual reproduction of mammals wherein the oocytes lose the centrioles and the centrosome of the zygote comes entirely from sperm cells (Pimenta-Marques et al. 2016) (table 1).

Table 1 Protein molecules involved in centriole production and duplication, and important for ciliogenesis in humans and their respective homologs in other model organisms

3.3 Centrioles in flagella (Chlamydomonas reinhardtii)

Centrioles in Chlamydomonas play the role of basal bodies after attachment of DAs and ciliary vesicles docking to it. Here too, the basal bodies serve as the main nucleation centre for flagellar assembly. The 9+2 flagella are anchored onto the basal bodies and are formed in three cycles of semi-conservative division. Out of the three generations of basal bodies present during the interphase, a flagellum is extended from the first- (mother) and the second-generation (daughter) basal body, while that of the third generation – the pro-basal body – is associated with each of the two at the proximal end. The two pairs of basal bodies duplicate during the late prophase/early metaphase that follows and form spindle poles (Wingfield et al. 2018). A class of mutants in C. reinhardtii called uniflagellate have a mutation in the uni3 gene that encodes for δ-tubulin (makes the C-tubules). These cells preferentially assemble the flagella on the mother basal body; this means that more than one cycle of division is required for a basal body to mature for flagellar transformation (Beech et al. 1991). Another mutant called bld2 shows no presence of flagella. The gene BLD2 codes for ε-tubulin, which is required for forming the incomplete B-tubules (Dutcher and Trabuco 1998). Mutation in this gene results in the defective formation of the whole cartwheel structure of the basal body. The normal formation of the ring structure is controlled by the SAS-6 complex encoded by the bld12 gene. The Sas-6 spokes harbour CEP135 at their tips, which is encoded by the gene bld10. Bld12 mutants show defective ring structure in basal bodies, but bld10 mutants completely lack basal bodies (Hiraki et al. 2007). This shows that these two proteins are of utmost importance for the normal formation of the 9-fold symmetry of the basic cartwheel structure. These proteins are recruited to their location by CEP70, a protein present in the ring structure in pro-basal bodies. An amorphous ring structure serves as a precursor for the cartwheel formation and harbours CEP70. This protein is lost during the maturation of basal bodies (Shiratsuchi et al. 2011). Along with proper formation, the correct positioning of these basal bodies is also very important for flagellar assembly. The attached distal and proximal fibres with the basal body ensure that docking happens at the correct site. Distal fibres contain a protein called centrin that provides elastic properties to these fibres, ensuring correct spatial positioning of basal bodies. A mutant with a mutation in the VLF2 gene that encodes for centrin has defective basal body positioning with varied numbers of flagella (Wright et al. 1985; Silflow and Lefebvre 2001). The basal bodies also harbour proteins required for the formation of the TZ. The two major protein components of TZ are NPHP4 and CEP290. In Chlamydomonas, mutants for these two proteins show altered protein content in the axoneme resulting in defective biochemical and structural composition of the whole axoneme (Craige et al. 2010; Awata et al. 2014). The proteins destined for the axoneme are first recruited to the basal body and then allowed to enter the axoneme through the TZ. These proteins are carried towards the axoneme through the IFT system (discussed in section 3.5) (table 1).

3.4 Centriole migration and acquisition

After the entire centriole is formed, it is positioned correctly and is ready to function, a process known as centriolar maturation. This sets the stage for the initiation of cilium assembly. Sorokin (1962) described two different pathways for cilium assembly. One of the pathways elucidated in the fibroblast and smooth muscle cells includes small vesicles most likely derived from the Golgi bodies directly binding to the DAs of the mother centriole thereby forming a large membranous ciliary vesicle at the distal end. These small (pre-ciliary) vesicles are directed to the centriole by the coordinated action of dynein and myosin-Va (Wu et al. 2018). Later, it was demonstrated that one of the DA proteins, CEP164, is responsible for the docking of vesicles at the centriole (figure 4; Schmidt et al. 2012). With the help of the Eps15 homology domain (EHD) family of proteins, these pre-ciliary vesicles (PCVs) then fuse together to form the ciliary vesicles (CVs; Lu et al. 2015). In turn, these vesicles act as a protective sheath for the growth of nascent cilium inside the cell. Following the growth of this nascent cilium, CVs fuse with the plasma membrane, exposing the cilium for the very first time to the extracellular environment. This fusion results in the formation of the ciliary membrane and the adjacent ciliary pocket (a specific membrane domain found at the base of cilia). The other pathway has been reported in epithelial cells, where the distal appendage protein CEP164 facilitates the attachment of DAs to cell membrane without the involvement of CVs (Graser et al. 2007; Schmidt et al. 2012). This migration is facilitated by microtubule and actin cytoskeleton (Pitaval et al. 2010, 2017). It has been reported that since actin inhibitors can promote ciliogenesis, actin is absent from the site of docking (Kim et al. 2010; Francis et al. 2011). After migration, the fully mature mother centriole that docks at the plasma membrane and nucleates the axoneme serves as basal bodies for ciliary elongation. The basal body acquires additional structures like transition fibres and basal feet (like DAs and SDAs, respectively), as well as ciliary rootlets for support (Anderson and Brenner 1971). The DAs play a role in the docking of the centriole to the apical plasma membrane, while the SDAs help in microtubule anchoring (Tanos et al. 2013; Mazo et al. 2016; Breslow and Holland 2019).

How a cell decides which pathway to undertake remains elusive. It can be dependent on the molecular requirements for different pathways or on the functional aspects of the cilia that are formed. The cells harbouring the motile cilia are mostly seen to undergo the extracellular pathway where the cilium extends beyond the cell surface. This helps sensory cilia sense and regulate the extracellular fluid flow (Sorokin 1968), extending very little of the cilium beyond the apical surface. These cilia are found emerging near the nucleus and Golgi body deep inside the cell (Sorokin 1962). Experimental alteration in the positioning of these cilia can result in the dysregulation of their function. A study wherein the submerged cilia were converted to apical cilia although showed poor sensory functions, did promote the regulation of fluid flow (Mazo et al. 2016). This implies that the cell type specificity for cilium positioning is correlated with the functional properties of the cell.

3.5 Transportation and compartmentalization of ciliary components to the site of assembly (protein trafficking)

Several proteins are required to build the cilium. Since cilia do not contain ribosomes, and therefore lack the protein synthesis machinery, proteins required for the elongation of cilia have to be imported from the cytoplasm (Rosenbaum and Child 1967; Long and Huang 2020). For the continuous assembly of cilia, these proteins must be transported to their site of assembly. This transport can either happen by diffusion or by a special translocation system called the intraflagellar transport (IFT), the pathways that contribute towards their formation being highly conserved across ciliated organisms (Kozminski et al. 1993; Rosenbaum and Witman 2002; Lechtreck 2015). Interestingly, soluble proteins smaller than 50 kDa and certain transmembrane proteins can freely diffuse into cilia (Kee et al. 2012; Breslow et al. 2013; Lin et al. 2013). Several studies have demonstrated the presence of a gate that serves as a barrier with a lower limit of the Stokes radius (~4.5 nm) for free diffusion into the cilia (Lin et al. 2013). For proteins with higher molecular weights such as outer dynein arms (~2 MDa), the TZ acts as a molecular barrier, therefore making IFT an important part of cilium assembly. It has been reported that for ciliary transmembrane-associated proteins, a targeting sequence known as ciliary localization signal (CLS) is required for the transport of these proteins into cilia (Berbari et al. 2008). However, in contrast to the nuclear localization signal (NLS), no unique consensus sequence is found for the transport of ciliary proteins, suggesting that other pathways are involved in the transport of these proteins. A rare exception was found in Plasmodium falciparum where there is no requirement of IFT for ciliogenesis as the axoneme assembles in the cytoplasm and directly transports proteins from cytoplasm to the tip of the growing cilium (Sinden et al. 1976; Briggs et al. 2004; Absalon et al. 2008). Another exception was found in Drosophila sperm flagella where necessary proteins for the formation of cilia are independent of the IFT (Han et al. 2003; Hao and Scholey 2009). First observed in the green alga C. reinhardtii, IFT provides the bidirectional movement of multimolecular protein complexes formed by strings of many IFT particles encompassing molecular motor proteins (Kozminski et al. 1993; Hao and Scholey 2009; Thomas et al. 2010). To be precise, its IFT system consists of 22 IFT protein particles and IFT motors (kinesin-2 for anterograde and dynein for retrograde movements) that come together to make the multisubunit long filamentous chains/strings called IFT trains for the transport of cargo to and fro from the axonemal tip. Particles that constitute the IFT are essential for flagellar assembly and maintenance. Jordan and Pigino (2019) used of in situ cryo-electron tomography (cryo-ET) and subtomogram averaging to investigate the structure and mechanism of the IFT depicting the movement of IFT trains along the microtubule doublets while transporting ciliary building blocks between the cell body and the ciliary tip. A comparison of the spatial arrangement of the various IFT components revealed that trains are assembled in a sequential manner (Hoek et al. 2022). The trains or strings that were complete were adjacent to the TZ at one end. The other end extended into the cytosol, and as it extends into the cytosol, the densities of dynein-1b and IFT-A keep reducing in a sequential order. This suggests that the IFT-B backbone is built first, followed by IFT-A and dynein-2 recruitment. Using cryo-ET and expansion microscopy techniques, the relatively small and flexible kinesin-2 was identified as the final component that is loaded onto the IFT trains close to the TZ (Hoek et al. 2022). These studies also indicated drastic differences between assembling and mature IFT trains. Assembling IFT trains were found to be more flexible than their mature axonemal counterparts and contained an extra density on IFT-B near the kinesin-2 binding site of unknown identity (Hoek et al. 2022). The importance of this process is evident wherein mutations in the motors DHC1B (a member of dynein superfamily) and kinesin-II cause defective or no flagella assembly, respectively (Silflow and Lefebvre 2001). The IFT protein subunits are formed in the cell body and then recruited by the basal body, where they start the construction of the IFT particles (Wingfield et al. 2017).

Further, the BBSome, a complex composed of eight proteins, works as an adapter in association with the IFT trains, thereby providing stabilization and maintenance. Mutation in these proteins causes Bardet-Biedl syndrome, a multi-organ disease specified by the defects in the sensory function of cilia (Wei et al. 2012; Williams et al. 2014). The axonemal assembly is normal in these mutants and so suggests the main role for the BBSome in ciliary signalling mechanisms (Lechtreck et al. 2009; Weihbrecht et al. 2017). Certain ciliary signalling proteins including G protein-coupled receptors (GPCRs), protein kinases, and ion channels have been found to transmit signals from the extracellular environment to the cell body. The BBS3 protein is thought to have a role in the interaction of ciliary signalling cargoes like phospholipase D (PLD). It has been shown that the BBSome relies on the BBS3 protein to enter the basal body via direct interaction with the BBS1 subunit of the BBSome in mammalian cells (Jin et al. 2010; Zhang et al. 2012; Mourão et al. 2014). However, in C. reinhardtii, GTPases ARL6/BBS3 recruit BBsomes from the cell body to the basal body through IFT22/RABL5 (Xue et al. 2020; Liu et al. 2021). Further, the BBSome splits from BBS3 and loads on to the anterograde IFT train and enters the cilium (Xue et al. 2020). The GTP-bound BBS3 then binds to the cell membrane, allowing it to laterally diffuse into the cilium (Liu and Lechtreck 2018). Through BBSome/IFT train remodelling, the BBSome is liberated from anterograde IFT trains at the ciliary tip. BBS3 binds to GTP and attracts the BBSome to the ciliary membrane, where it interacts with PLD. PLD-laden BBSomes are subsequently loaded onto retrograde IFT trains for ciliary escape by connecting with IFT-B1 (Liu and Lechtreck 2018; Liu et al. 2021; Xue et al. 2020). The entire protein transportation by the IFT/BBSome complex is divided into (i) IFT binding to cargo, (ii) anterograde movement, and (iii) retrograde movement (sections 3.5.1 to 3.5.3) (table 2).

3.5.1 IFT binding to cargo

In Chlamydomonas, the IFT trains pass through the TZ and then move towards the axonemal tip, dropping off the cargo in between at its exact locations. The flagellar assembly and disassembly are facilitated by the IFT along with a few other proteins. The X-chromosome-associated protein 5 (XAP5) has been identified as the transcription factor in C. reinhardtii for regulating the transcription of flagellar assembly genes by recruiting RNA polymerase II (Pol II) for transcriptional activation (Li et al. 2018). FAP256, a homolog of the mammalian protein CEP104, is a flagellar-tip-associated protein and seems to have a function in the regular assembly of the flagella (Tammana et al. 2013). Other proteins like katanin and aurora kinases participate in flagellar disassembly during cell division. Katanin is a microtubule-sequestering protein that does not allow further microtubule elongation beyond the transition zone (Lohret et al. 1998). Having served as the pioneer model system for investigations on the IFT and motors, most orthologs have been found in higher organisms (tables 2 and 3).

Table 2 Protein molecules from the IFT complexes important for ciliogenesis in humans and their respective homologs in other model organism
Table 3 Motor protein molecules important for ciliogenesis in humans and their respective homologs in other model organisms

Binding of the IFT complex to the cargo is the foremost and necessary step for the transportation of ciliary proteins into the growing cilium. Broadly, IFT trains are composed of two large complexes, labelled as IFT-A and IFT-B (Ishikawa and Marshall 2011; Taschner et al. 2012; Picariello et al. 2019). These multimeric protein complexes have distinct yet complementary roles and move together into the cilium. IFT-B consists of 16 subunits that transport proteins from the base to the tip of the cilium (anterograde movement), while IFT-A consists of 6 subunits that transport proteins from the tip to the cell body (retrograde movement) (Cole et al. 1998). IFT-B further assembles into IFT-B1 and IFT-B2 sub-complexes composed of at least 10 and 6 distinct subunits, respectively (Boldt et al. 2016; Katoh et al. 2016; reviewed by Taschner and Lorentzen 2016). They are arranged within the basal body region in the vicinity of transition fibres (part of the transition zone that connects the distal basal body to the base of the ciliary membrane). The IFT complexes are first gathered into IFT particles, and the latter connect to IFT motors forming IFT trains (Lechtreck 2015; Ishikawa and Marshall 2017; Prevo et al. 2017). Following this, binding of the BBSome protein to IFT trains occurs. However, the recruitment of the BBSome to the basal body for binding to IFT trains remains elusive. IFT complexes contain cargo-binding sites that bind to a variety of cargo molecules, mainly axonemal proteins such as ciliary tubulins and motor proteins (Wren et al. 2013; Craft et al. 2015). In silico analysis has shown several protein–protein interaction motifs present on the IFT molecule including tetratricopeptide repeats (TPR), WD40, and a coiled–coiled domain which helps in the binding of these IFTs to the cargo (Smith et al. 1999; Burkhard et al. 2001). For example, structural studies have shown that a calponin homology (CH) domain is present in the N-terminal region of IFT81 (part of IFT-B1) that harbours recognition sites for tubulin (Bhogaraju et al. 2013). Interaction of tubulin with IFT81 is enhanced by an IFT74 that has a high-affinity binding site at its C-terminal tail. Knockdown of IFT81 with point mutation and its rescue experiments indicate the combined role of these IFTs in tubulin-binding revealing its importance for ciliogenesis (Bhogaraju et al. 2013). Additionally, similar binding mechanisms are present in the IFT-B2 complex suggesting the transport of two tubulins per IFT-B complex (Taschner and Lorentzen 2016). The mode of transport of ciliary transmembrane proteins requires their synthesis in the endoplasmic reticulum (ER) followed by modification and sorting in the Golgi apparatus. After processing from the Golgi apparatus, proteins are carried to the ciliary base via a transport vesicle that moves into the cilium with the help of the IFT-B subunit and KIF17 motor protein (Long and Huang 2020). Axonemal proteins, such as radial spoke proteins (RSPs), enter into cilia as a multiprotein complex requiring pre-assembly at the ciliary base. Of the 23 known RSPs, the 12S radial spoke complex (component of 11 RSPs) is first assembled at the ciliary base and then transported to the tip of the cilia via IFTs, where they combine with the remaining RSPs to form the 20S radial spokes complex (Diener et al. 2011). Likewise, the dynein arms are transported by the same mechanism (described above) (Fowkes and Mitchell 1998).

3.5.2 Delivering of cargo proteins (anterograde movement) and regulation of IFT transport

Once the IFT complexes, IFTB/A (designated so because in the complex, IFT-B is present on the top and IFT-A at the bottom) assemble into IFT trains, the movement of cargo proteins is initiated and mediated by the motor protein kinesin-2 family (Cole et al. 1993; Kozminski et al. 1995; Engelke et al. 2019). The kinesin-2 motor family is known to be present in the heterotrimeric and homodimeric forms, both of which contribute to the anterograde IFT movement. The heterotrimeric kinesin-2 consists of two motor subunits of the kinesin family member, KIF3A and KIF3B, and an accessory subunit kinesin-associated protein (KAP3). It is the KIF3B subunit that modulates ciliary transportation by binding directly to IFT-B (Yamazaki et al. 1996; Mueller et al. 2005; Scholey 2013). The transport of IFT machinery is highly regulated at the base and inside the cilia, especially at the ciliary tip. Phosphorylation events serve as a molecular switch between the binding and unloading of the cargo proteins (Craige and Witman 2014; Liang et al. 2014). For example, dephosphorylation of the KIF3B subunit is required for binding to IFT-B at the ciliary base. As soon as it is organized, IFT trains release the cargo proteins into the cilium. Inside the cilium, phosphorylation by numerous local protein kinases induces a conformational change that leads to the disruption of the binding between IFT particles, motor protein, and cargo, thereby unloading the cargo at the ciliary tip. An instance of phosphorylation has been reported in Chlamydomonas wherein the FLA8 subunit (flagella assembly 8, homologue of KIF3B) of the kinesin-2 family serves as a substrate for calcium-dependent protein kinases (CDPKs) that are enriched at the flagellar tip during flagellar assembly. This disrupts the interaction between IFT-B and kinesin-2 such that the cargo does not bind again (Liang et al. 2014). On the other hand, in animals such as C. elegans and vertebrates, different subunits of kinesin-2 (OSM-3 in C. elegans and kif17 in vertebrates) help in building the cilium (Taschner et al. 2012). Full-length KIF17 and a construct of KIF17 without its motor domain accumulate at the ciliary tip (Jiang et al. 2015), suggesting that KIF17 can function as cargo as well. Further, a co-immunoprecipitation of KIF17 pulls down the IFT46-IFT56 dimer of the IFT-B complex. Knockout of IFT56 in the human RPE1 cell line (Funabashi et al. 2017) restricts the entry of KIF17 into cilia. Nevertheless, KIF17-deficient mice do not affect IFT-B localization and cells show normal ciliogenesis, indicating an ambiguous role for this protein (Yin et al. 2011). Mutational analysis in the subunits of the motor protein shows that inactivation of kinesin-2 blocks cilia assembly and IFT transport in various organisms, including Chlamydomonas, Drosophila, C. elegans, and mice (Kozminski et al. 1995; Marszalek et al. 1999; Snow et al. 2004; Chaya et al. 2014; Engelke et al. 2019).

3.5.3 Returning of IFT trains to the cell body (retrograde movement)

After the cargo is released into the cilium, the designated proteins diffuse to their docking site, and anterograde IFT trains immediately disassemble (by the molecular switch discussed earlier) at the ciliary tip to rearrange into retrograde trains. The retrograde movement of the IFT is mediated by the dynein motor protein, a multiprotein complex such as cytoplasmic dynein 2 that contributes to the returning of IFT trains from the ciliary tip to the cell body (Pazour et al. 1998; Porter et al. 1999; Perrone et al. 2003; Engel et al. 2012). Defects in the retrograde motor protein show the accumulation of IFT-B proteins on the tip of the cilium suggesting the important role of IFT-A in returning IFT-B to the cell body (Efimenko et al. 2006; Iomini et al. 2009). Overall, these IFT particles provide the assembly of cilia in a very coordinated manner and take part in the maintenance of the organelles in most ciliated organisms.

3.6 Regulatory network

Right from the first step of centriole duplication, the entire process of ciliogenesis is strictly regulated. Its understanding comes chiefly from the loss of specific proteins such as CEP120 and neuralized-like protein 4 (NEURL4) from the daughter centriole that marks the beginning of mother centriole formation (Zou et al. 2005; Mahjoub et al. 2010; Li et al. 2012). The recruitment of DA proteins is regulated by C2 calcium-dependent domain containing 3 (C2CD3), orofaciodigital syndrome 1 (OFD1), and mitogen-activated protein kinase 15 (MAPK15; Kazatskaya et al. 2017), whereas the recruitment of SDA proteins is regulated by a different set of proteins consisting of trichoplein (TCHP) and coiled-coil and C2 domain-containing protein 2A (CC2D2A; Veleri et al. 2014). While the role of these proteins is known in ciliogenesis, their interaction with other regulatory proteins and how they work as a group to ensure synchronous maturation of the mother centriole with cell cycle is unclear. The PCVs arising from either the Golgi or endosomes are recruited to the mother centriole appendages with the help of kinesins and dyneins. Specific proteins such as kinesin family member C1 (KIFC1) and dynein-1 heavy chain (DYNC1H1) regulate the trafficking of these vesicles to the mother centriole (Lee et al. 2017; Wu et al. 2018). Along with these motor proteins, the actin network formed with the help of the actin-nucleator complex containing actin-related proteins 2/3 (ARP2/3) and the nucleation-promoting complex WASH (Wiskott Aldrich syndrome protein and scar homolog) results in the actin assembly with the centrosome acting as the actin-organizing centre (Antoniades et al. 2014). This observation is also supported by the localization of specific actin cytoskeleton-associated proteins (LIM-kinase 2, LIMK2 wherein LIM stands for LIN-11, Isl-1 and MEC-3; testicular protein kinase 1, TESK1) near the centriole (Kim et al. 2015a, b). Specific regulators (microtubule actin crosslinking factor 1, MACF1A) that maintain the actin cytoskeleton network further help in vesicle trafficking (May-Simera et al. 2016). After the formation of CVs by the fusion of numerous PCVs, the axoneme elongation process starts. The formation of CVs is triggered by the recruitment of Rab11-Rabin8-Rab8 signalling pathways. Proteins including CEP164, Chibby (CBY1), Abelson-helper integration site 1 (AHI1), and Talpid3 (KIAA0586) (Burke et al. 2014) serve as substrate and facilitate the interaction of Rab8 with the DAs on the mother centriole.

3.6.1 Regulators present at the ciliary base

The mother centriole-associated DAs fuse with the recycling endosomes containing the signalling molecule Rab-11 GTPase, which are then called distal appendage vesicles (DAVs). Rab-11 helps in the recruitment of Rab8, a protein that is important for the progression of ciliogenesis (Westlake et al. 2011). This activation is mediated through the recruitment of Rabin8, a guanine exchange factor (GEF) that is responsible for the final activation of Rab8. DAVs also contain some membrane-shaping proteins delivered by Rab-11 endosomes. These proteins are called EHD family proteins; EHD1 and EHD3 proteins work together to bring the DAVs proximal to each other. This results in the vesicle fusion mediated by SNAP29 (a member of the SNARE protein family). This fusion results in the formation of the CVs associated with the mother centriole (Lu et al. 2015). After this, the CVs extend because of Rab8 activation (figure 4B). It is also hypothesized that the EHD proteins help in the removal of a centriolar protein (CP110, a negative regulator of ciliogenesis) from the M-centriole that leads it to become a basal body. CP110 forms a complex with CEP97, CEP290, and KIF24, and promotes centriolar duplication during cell cycle stages ensuring that abnormal cilia are not formed. This complex is stabilized by a centrosomal deubiquitinase, ubiquitin carboxyl-terminal hydrolase 33 (USP33), that de-ubiquitinylates CP110. CP110 is degraded by abundant ubiquitylation with the help of the skp1-cul1-F-box-protein (SCFCyclinF) complex that is a positive regulator of ciliogenesis. This equilibrium between ubiquitylation and de-ubiquitylation regulates the level of centrosomal proteins and therefore ciliogenesis (Li et al. 2013). Another protein, tau tubulin kinase 2 (TTBK2; a serine-threonine kinase), has been shown to displace CP110 from the centriole to initiate basal body formation and hence ciliogenesis. It removes CP110 by phosphorylating one of the members of the complex CP110/CEP97/CEP290/KIF24 and further helps in the recruitment of IFTs for axoneme extension (figure 4B). Hence, TTBK2 also serves as an important player in the initiation and progression of ciliogenesis (Goetz et al. 2012). Another regulator is leucine-rich repeat kinase2 (LRRK2) that functions in commanding the microtubule dynamics through proper and direct interaction with microtubule components like TUBB (one of the ten β-tubulin-encoding genes present in the human genome). This also suggests the potential role of axonemal tubulin acetylation in ciliary biogenesis (Gillardon 2009; Kett et al. 2020; Law et al. 2014). Besides this, LRRK2 also helps in removing CP110 and recruiting TTBK2 at the mother centriole (Sobu et al. 2021). It is speculated that the EHD proteins along with the pre-ciliary membrane proteins help in the recruitment of the ubiquitination-related enzymes to the ciliary base. Further, these EHD proteins are also responsible for the transport of proteins of the TZ and IFT complex (specifically IFT20) to the ciliary base where further sorting of proteins that are destined to become the axoneme occurs (Lu et al. 2015). The ciliary base is an important site for ciliogenesis initiation as well as cargo sorting for ciliary proteins. All the proteins required for ciliary growth are first transported to the ciliary base and then sorted and sent to different destinations of the axoneme. This protein trafficking happens in two parts, viz., pre-ciliary and intraciliary transport. All the proteins (cytoplasmic and transmembrane) that are meant to be sent to the ciliary base are sorted at the trans-Golgi network based on specific ciliary targeting signals (CTSs) (Mazelova et al. 2009). All the ciliary-destined proteins are carried to the ciliary base by four different pathways based on their initial localization and abundance. For example, plasma membrane-based proteins do not require the formation of separate vesicles for transport. These proteins directly transfer to the ciliary base through lateral diffusion to the periciliary membrane.

3.6.2 Regulators of cilium disassembly

In contrast to the knowledge gathered on cilium assembly, ciliary disassembly/resorption remains less understood. Before mitosis, the centriole pair present on the apical membrane of the cell is disengaged from the cortex and moves to the spindle poles. The main rationale behind the resorption of cilia is the formation of mitotic spindles that ensure the proper organization and position of chromosomes during the cell cycle. The detachment of the centriole from the surface involves mainly two proteins, mitotic kinase Aurora A and histone deacetylases (HDAC2/6). Aurora A kinase is a key regulator of cell proliferation and participates in primary cilia resorption by phosphorylating and stimulating HDAC; the latter destabilizes and depolymerizes the tubulins of the axonemal microtubule, thereby promoting cilium disassembly (Pugacheva et al. 2007; Plotnikova et al. 2012; Ran et al. 2015). The activation of Aurora A is controlled by a complex signalling pathway initiated by the extracellular signal platelet-derived growth factor (PDGFRβ) that further activates the downstream enzyme, phospholipase C-gamma (PLCγ). Activation of PLCγ induces the release of intracellular Ca2+, whose influx activates the Ca2+-calmodulin complex (CaM) that binds to Aurora A and its binding partner, a scaffolding protein, human enhancer of filamentation 1 (HEF1) (Plotnikova et al. 2012). This multiprotein complex harbours the active Aurora kinase enzyme that causes phosphorylation of HDAC6. Phosphorylated HDAC6 destabilizes the axonemal microtubule by removing the acetylated tubulins present in cilia. Another protein pitchfork (Pifo) also promotes cilium disassembly by activation of Aurora A kinase. Experimental analysis shows that Pifo accumulates at the basal body and physically interacts with Aurora A kinase, culminating in the destabilization of the ciliary microtubules (Kinzel et al. 2010). Mutation in pitchfork R80K blocks Aurora A activation. However, the exact interacting site involved in the binding of pitchfork to Aurora A kinase is poorly understood. Apart from the two proteins discussed above, two kinesin superfamily proteins have also been implicated in disassembly of primary cilia before mitosis, KIF2A and KIF24. Depolymerization of microtubules by KIF2A is activated by a mitotic kinase PLK1 (polo-like kinase). PLK1 phosphorylates KIF2A at the T554 position which in turn activates the KIF2A and promotes the disassembly of ciliary microtubules in a growth signal-dependent manner (Miyamoto et al. 2015). Similarly, depolymerization of microtubules by KIF24 is enhanced by the activation of NEK-2, a protein kinase that is generally expressed in the S- and G2-phase and phosphorylates KIF24, ensuring the blockage of reciliation throughout the S/G2/M-phases of the cell cycle (Kim et al. 2015a, b).

The loss of the cilium has been investigated in Chlamydomonas and in mammalian cells. When Mirvis et al. (2019) used live-cell imaging to characterize the individual events during loss of the cilium, they found rapid deciliation wherein the membrane and axoneme of the cilium were shed from the cell. While rapid deciliation was the most prominent, some gradual resorption was also observed. In some cells, a period of gradual resorption was followed by rapid deciliation. Deciliation or disassembly as seen in Chlamydomonas results in the intact cilia being shed in the medium that could be recovered. Such cilia (in both cases) contain both membrane and axoneme proteins. Interestingly, when the levels of katanin are modulated, the intracellular calcium rises and deciliation occurs. It is also known that excess katanin can alone promote cilia loss by deciliation, independently of calcium. Together, these results suggest that mammalian ciliary loss involves a tuneable decision between deciliation and resorption (Lohret et al. 1998; Mirvis et al. 2019).

4 Role of cilia in human health and diseases

4.1 Involvement of cilia in non-ciliopathic diseases

Primary cilia have a diverse array of receptors present on the ciliary membrane which efficiently coordinate through the various cellular signalling pathways for the proper functioning of the cell, which also makes them prone to diseases, commonly called ‘ciliopathies’. However, accumulating evidence shows the involvement of cilia in non-ciliopathic diseases like anosmia. Recently, infections by a novel coronavirus (SARS-COV2), which is the causative agent of COVID-19, reported patients with symptoms that included loss of smell (anosmia). Ciliary membrane receptors seem to be points of entry for the virus. Upon binding, the dysfunctional cilia render the patient with a loss of smell. TEM studies on influenza and the Sendai virus clearly depict the attachment of the virus particles to the ciliary membrane. As soon as the virus enters, the nucleocapsid of the virus partially unwinds and escalates the entry into the cilia (Dourmashkin and Tyrrell 1970). Also, research on human nasal epithelial cells infected with a variant of coronavirus (hCov) shows cilia dysfunction (Chilvers et al. 2001), thereby causing a defect in axonemal structure, and subsequently damaging the respiratory epithelium (Chilvers et al. 2003). ACE2 localization in the ciliated epithelial cell of respiratory, pulmonary, and urinary systems is probably seen as a host receptor for the entry of virus into the cells (Lukassen et al. 2020; Sungnak et al. 2020). Investigation using human airway epithelial cell culture revealed that the virus in the form of virions is trapped at the base of the cell (near the ciliary pocket) which marks the endocytosis pathway of virus entry to the cells (Sims et al. 2005). The ciliary pocket has abundant clathrin-coated pits, making cilia a potential target for viruses to enter the cell (Otsuka et al. 2015; Verhey and Yang 2016). Studies have indicated that despite the virus deciliating the organelle upon infection, with some delay, the cilia are known to re-assemble and return to normal functioning post recovery (Afzelius 1994). Moreover, the loss of several cytokines in COVID-19 infection impacts ciliary function mainly due to the loss of regulation in cilia length (Dummer et al. 2018). Greater understanding is needed to find out the molecular link between the virus and cilium biology to strengthen knowledge on the impact of cilia on human health and diseases.

4.2 Role of Sonic Hedgehog signalling in primary cilia

Sonic Hedgehog (SHH) signalling is the major pathway that regulates crucial events during the development process. It is noted that abnormal activation of this process has a significant association with glioma, basal cell carcinoma (BCC), prostate, and lung cancer (Goetz et al. 2009; Hassounah et al. 2012). The presence of the canonical SHH receptor has therefore a connection with the primary cilia. Almost 90% of BCC and 30% of medulloblastoma have a mutation in the genes involved in SHH signalling (Hassounah et al. 2012; Pugh et al. 2012). When the ligand SHH binds to the receptor patched (PTCH1) present on the plasma membrane, there is an accumulation of SMO (smoothened, a seven-pass transmembrane protein) and SUFU (suppressor of fused) on the primary ciliary membrane. Out of the three Gli transcription factors (Gli1, Gli2, and Gli3), Gli2 and Gli3 are converted by proteolytic cleavage to perform roles of transcriptional activation or repression. Active forms of SMO and SUFU recruit Gli1 to form a complex that together targets Gli2 and Gli3 to act as transcriptional activators upon translocation to the nucleus and mediate SHH signalling. In the absence of SHH, Gli2 and Gli3 act as transcriptional repressors resulting in the suppression of SHH signalling genes (Rohatgi et al. 2007; Wong et al. 2009). Depending on the initiating carcinogenic factor, SHH signalling has a dual role in promoting and preventing carcinoma (Han and Alvarez-Buylla 2010; Archer et al. 2012; Hassounah et al. 2012; Gerhardt et al. 2016). This was observed in BCC (a skin cancer) wherein cells with an activated SMO possess primary cilia, and removal of primary cilia inhibits the progression of BCC (Wong et al. 2009; Athar et al. 2014; Basset-Seguin et al. 2015). In contrast, however, cells with the active form of Gli2 possess primary cilia and the removal of primary cilia accelerates the progression of BCC (Wong et al. 2009). Similarly, in the case of medulloblastoma (malignant brain tumour), cilia are present when there is a loss of function in PTCH1 and SUFU. On the other hand, a gain of function in SMO and Gli causes an enhanced activity of SHH, thereby promoting medulloblastoma. In non-ciliated cells, amplification/mutation of Gli transcription factors without the activation of PTCH1 and SMO causes an activation of Gli2-R which acts as an antagonist of Gli2-A, causing the suppression of SHH medulloblastoma (Han and Alvarez-Buylla 2010).

Apart from the primary cilium, motile cilia also contain some components of SHH signalling, but in a non-canonical manner (Mao et al. 2018). Studies on human airway epithelia showed that SHH (Cigna et al. 2012; Peng et al. 2015; Henkin et al. 2017) is present in the basolateral side of epithelia. During early development, the primary cilium present on the airway cell produces SHH for its normal functioning. After completion of development, i.e., in mature airway epithelia, the primary cilium disappears and numerous motile cilia appear on the apical portion of cells (Sorokin 1968; Jain et al. 2010). Several efforts have been carried out to identify the components of SHH signalling based on arrayed siRNA libraries (Roosing et al. 2015). Recently, a whole-genome-wide CRISPR screen technology was used to identify the factors required for ciliary signalling, revealing the relation of certain disorders caused by ciliary signalling (Breslow et al. 2018; Pusapati et al. 2018). The novel ciliary components (FAM92A, TTC23, TXNDC15, and TEDC proteins) that were candidate ciliopathy genes came to light after the CRISPR-based screening. For FAM92A, a mutant cell pool was generated to identify the precise role in ciliogenesis, and it has been shown that the formation of cilia is drastically reduced in Fam92A-knockout pools (Breslow et al. 2018). Further, localization studies using FAM92A antibodies found it to be present in the TZ (Breslow et al. 2018). FAM92A was found to co-localize with its partners, viz., CBY1 and DZIP1L (Glazer et al. 2010; Burke et al. 2014; Li et al. 2021). TTC23, on the other hand, has been found to be associated with IQCE (IQ domain-containing protein E) and EFCAB7 (EF-hand calcium-binding domain-containing protein 7) previously shown to localize to the Ellis-van Creveld (EvC) zone (proximal region of cilium; Ruiz-Perez and Goodship 2009; Dorn et al. 2012; Pusapati et al. 2014). Although knockout of TTC23 does not show any defect in the cilium assembly, it certainly reduces the localization of IQCE and EFCAB7 proteins to the EvC zone (Breslow et al. 2018). Thioredoxin-domain containing 15 (TXNDC15) has been shown to be associated with the Meckel syndrome (Shaheen et al. 2016). Knockout studies of TXNDC15 have shown decreased levels of the ciliary GTPase ARL13B, indicating a clear defect in SHH signalling (Breslow et al. 2018). Tubulin ε and δ complex 1 and 2 (TEDC1 and 2) proteins have been found to associate with ε-tubulin, δ-tubulin, α/β-tubulin, and CENPJ (centriolar regulator of microtubule dynamics). Mutation in the cenpj gene disrupts the triplet microtubules present in the flagella of Drosophila spermatocytes (Zheng et al. 2016). Knockout of both TedC1 and TedC2 showed complete loss of the centrioles. These mutants lack cilia and show strong defects in the SHH signalling (Breslow et al. 2018). Thus, the CRISPR screening platform has led to the identification of genes involved in ciliary signalling with direct relevance to human health and diseases.

4.3 Role of cilia in neurodegenerative diseases

LRRK2 (leucine-rich repeat kinase 2), a large multifunctional kinase, is responsible for neurodegeneration such as autosomal-dominant Parkinson’s disease. It has been shown that in cultured cells, the pathogenic form of LRRK2 phosphorylates the switch II domain of Rab GTPases which are considered master regulators of intracellular vesicle trafficking (Steger et al. 2017). Rab-interacting lysosomal protein-like proteins 1 and 2 (RILPL1 and 2) localize in both primary cilium and centrosome (mother centriole) and are known to regulate the protein concentration of the primary ciliary membrane. Depletion of these two proteins causes an imbalance in the concentration of signalling molecules in the ciliary membrane and prevents the proper organization of epithelial cells in 3D (ALI) cultures (Schaub and Stearns 2013). RILPL1 and RILPL2 are the primary ciliogenesis regulators that help in the formation of cilia; the phosphorylated form of Rab GTPases interacts with RILPL1 and RILPL2 and thus blocks the formation of primary cilia (Schaub and Stearns 2013). This was the first study to show the link between Parkinson’s disease and cilia. Experimental analysis has shown that phosphorylation of Rab GTPase proteins and the dephosphorylation by a phosphatase, viz., protein phosphatase Mg2+/Mn2+-dependent1H (PPM1H) is crucial for ciliary functioning. Precisely, PPM1H, a family of phosphatase enzymes, has been shown to counter this process by specifically dephosphorylation of the subset of Rab proteins, namely, Rab8A, Rab8B, Rab10, and Rab35 (Steger et al. 2017; Dhekne et al. 2018; Schmidt et al. 2022). In conclusion, cells expressing hyperactive LRRK2 kinase mutants cause depletion of PPM1H and subsequently block the formation of primary cilia. Hence, this needs to be further ascertained with greater insight into these processes so that one can apply these as therapeutics to treat Parkinson’s disease (Berndsen et al. 2019).

4.4 Role of cilia in skeletal dysplasia

As mentioned earlier, the IFT machinery is a highly conserved mechanism for the motor-driven transport of cargo within cilia. WDR35, also known as IFT121, is an IFT machinery protein (Quidwai et al. 2021). It is a component of the IFT-A complex best known for its role in ciliary retrograde transport. Mutation in the Wdr35 gene causes mutant cilia but fails to enrich diverse classes of ciliary membrane proteins, resulting in severe disorders in both humans and mice including a range of potentially lethal skeletal dysplasia. The absence of this protein not only disrupts the IFT, but also stops certain ciliary proteins and their associated membranes from entering cilia. It also causes the accumulation of transport intermediates from its cellular manufacturing site to the cilia (Quidwai et al. 2021). Analysis of the structure of WDR35 and other IFT proteins has shown that they are very similar to a protein complex called COPI, which is involved in transporting membrane proteins around the cell. The role of COPI is to coat the vesicles (small lipid bubbles) used by cells to store newly synthesized proteins until they are destined for their sites. This process has been referred to as vesicular transport wherein COPI has a major role to play. Quidwai et al. (2021) found a similar role for WDR35 and other IFT proteins. For example, these proteins can bind to specific types of lipid molecules, suggesting a role in vesicular transport. Likewise, when mouse cells were mutated to stop producing WDR35, coatless vesicles accumulated around the base of the cilia. Adding back normal WDR35 production to these mutant cells rescued these defects giving rise to functional cilia (Quidwai et al. 2021). These results indicate a COPI-like complex system to deliver proteins to growing cilia. Therapy on these lines could prove useful in human skeletal dysplasia.

5 Concluding remarks

This review is an attempt to describe the complexity with which the cilium appears and disappears in cells. From precursor generation and their migration inside the cell to ciliary elongation and maintenance at the membrane, ciliary programs are highly regulated. There are divergent ciliogenesis pathways and patterns in various cell types, despite conserved ciliary structures and components. This begs for further explorations showing the extent of conservation of these ciliary processes among different species (vertebrates, invertebrates, protozoans, C. reinhardtii), tissues (hair cells, ependymal cells, sperm, kidney) and cilia types (9+0, 9+2, motile, immotile). Since the cilia progress from centrioles, a significant direction of study would be showing the interdependence of the two organelles on each other and the level of decision-making by the cell for maintaining the location, number, structure, length, and duration of cilia for itself.

A unique question that comes up in this direction is: Is the architecture, proteome, and regulation of the basal bodies the same or different for the 9+0 and 9+2 cilia? It is interesting to note that there are differences in the cilium basal bodies of 9+0 and 9+2 cilia. The basal body plays a very important role in the making of a cilium and provides a much-needed direction to the microtubule in forming the ciliary axoneme. Although the basic structure and function of all basal bodies do not differ significantly, could there be differences between species? Could there be differences between basal bodies in various tissues, especially ones in cells that make the 9+0 cilium versus those that make the 9+2 cilium types? Such an understanding across species and ciliary types within the same organism are sparse. For example, analysis of the accessory structure of the murine basal body reveals that in most of the 9+0 ciliated cells, the basal body is connected to the proximal end of the daughter centriole via a linkage that involves the rootlet and rootletin (Bahe et al. 2005). Rootlets are a thick striated bunch of filaments that extend close to the nucleus. Rootletin (also known as ciliary rootlet coiled-coiled or crocc protein) is a component of the ciliary rootlet (Yang et al. 2002). Both rootlet and rootletin are known to be associated with basal bodies (Klotz et al. 1986) in which rootletin in connection with centrosomal Nek2-associated protein 1 (C-Nap1, wherein Nek stands for NIMA-related kinase) forms fibres that connect the mother and daughter centriole and help in centrosome cohesion (Yang et al. 2006). This mother centriole now matures after the ciliary vesicle docks to the distal part and onto the plasma membrane to form a cilium. However, in contrast to 9+2 multiciliated cells, the basal body parts are not connected with the daughter centriole, which itself is a drastic difference between the two types (Jord et al. 2014). In addition, the proteomes of the basal bodies of 9+0 and 9+2 are similar but not identical between the two types of cilia of the same organism.

Incidentally, the TZs are short or long depending on cell types. For example, TZ is short in fibroblast cells and long in photoreceptor cells (Garcia and Reiter 2016). On the other hand, it may be emphasized that the central pair is not in continuation with microtubules of the basal body. Although it is unknown if the basal body components help in stabilizing the central pair, in an invertebrate like Drosophila, one such basal body component, viz. BLD10/CEP135, stabilizes the central pair microtubule of the ciliary axoneme (Carvalho-Santos et al. 2012). Future implications for research on basal bodies of distinct types of cilia will provide better insight on ciliopathy diseases that are particularly linked with basal body components.